EPA
Integrated Science Assessment for
Paniculate Matter
'
-------
&EPA
United States December 2009
Pr0tectl°nEPA/600/R-08/139F
Integrated Science Assessment for
Particulate Matter
National Center for Environmental Assessment-RTF Division
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC
-------
Disclaimer
This document has been reviewed in accordance with U.S. Environmental Protection Agency
policy and approved for publication. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
December 2009
-------
Table of Contents
PM ISA PROJECT TEAM XLII
AUTHORS, CONTRIBUTORS, REVIEWERS XLV
CLEAN AIR SCIENTIFIC ADVISORY COMMITTEE FOR PARTICULATE MATTER NAAQS LI
ACRONYMS AND ABBREVIATIONS LIN
CHAPTER 1. INTRODUCTION 1-2
1.1. Legislative Requirements 1-4
1 .2. History of Reviews of the NAAQS for PM
1.3. ISA Development
1.4. Document Organization
1.5. EPA Framework for Causal Determination
1.5.1. Scientific Evidence Used in Establishing Causality
1.5.2. Association and Causation
1 .5.3. Evaluating Evidence for Inferring Causation
1 .5.4. Application of Framework for Causal Determination
1 .5.5. First Step— Determination of Causality
1 .5.6. Second Step— Evaluation of Response
1.5.6.1. Effects on Human Populations
1.5.6.2. Effects on Public Welfare
1 .5.7. Concepts in Evaluating Adversity of Health Effects
1.6. Summary
Chapter 1 References
1-5
1-9
1-13
1-14
1-15
1-15
1-15
1-19
1-20
1-22
1-22
1-24
1-24
1-24
1-26
CHAPTER 2. INTEGRATIVE HEALTH AND WELFARE EFFECTS OVERVIEW 2-28
2.1.
2.2.
2.3.
Concentrations and Sources of Atmospheric PM
2.1.1. Ambient PM Variability and Correlations
2. 1 . 1 . 1 . Spatial Variability across the U.S.
2. 1 . 1 .2. Spatial Variability on the Urban and Neighborhood Scales
2.1.2. Trends and Temporal Variability
2.1.3. Correlations between Copollutants
2.1.4. Measurement Techniques
2.1.5. PM Formation in the Atmosphere and Removal
2.1.6. Source Contributions to PM
2.1.7. Policy-Relevant Background
Human Exposure
2.2.1 . Spatial Scales of PM Exposure Assessment
2.2.2. Exposure to PM Components and Copollutants
2.2.3. Implications for Epidemiologic Studies
Health Effects
2.3.1. Exposure to PM9S
2.3.1.1. Effects of Short-Term Exposure to PM9S
2.3.1.2. Effects of Long-Term Exposure to PM2s
2-29
2-29
2-29
2-30
2-30
2-31
2-31
2-31
2-32
2-33
2-33
2-33
2-34
2-34
2-35
2-36
2-36
2-38
December 2009
-------
2.3.2. Integration of PM25 Health Effects 2-40
2.3.3. Exposure to PM10.25 2-45
2.3.3.1. Effects of Short-Term Exposure to PM10.25 2-45
2.3.4. Integration of PM10.25 Effects 2-46
2.3.5. Exposure to UFPs 2-48
2.3.5.1. Effects of Short-Term Exposure to UFPs 2-48
2.3.6. Integration of UFP Effects 2-49
2.4. Policy Relevant Considerations 2-50
2.4.1. Potentially Susceptible Populations 2-50
2.4.2. Lag Structure of PM-Morbidity and PM-Mortality Associations 2-51
2.4.2.1. PM-Cardiovascular Morbidity Associations 2-51
2.4.2.2. PM-Respiratory Morbidity Associations 2-51
2.4.2.3. PM-Mortality Associations 2-52
2.4.3. PM Concentration-Response Relationship 2-52
2.4.4. PM Sources and Constituents Linked to Health Effects 2-53
2.5. Welfare Effects 2-54
2.5.1. Summary of Effects on Visibility 2-54
2.5.2. Summary of Effects on Climate 2-55
2.5.3. Summary of Ecological Effects of PM 2-56
2.5.4. Summary of Effects on Materials 2-57
2.6. Summary of Health Effects and Welfare Effects Causal Determinations 2-58
Chapter 2 References 2-61
CHAPTERS. SOURCE TO HUMAN EXPOSURE 3-1
3.1. Introduction 3-1
3.2. Overview of Basic Aerosol Properties 3-1
3.3. Sources, Emissions and Deposition of Primary and Secondary PM 3-6
3.3.1. Emissions of Primary PM and Precursors to Secondary PM 3-8
3.3.2. Formation of Secondary PM 3-10
3.3.2.1. Formation of Nitrate and Sulfate 3-10
3.3.2.2. Formation of Secondary Organic Aerosol 3-10
3.3.2.3. Formation of New Particles 3-12
3.3.3. Mobile Source Emissions 3-13
3.3.3.1. Emissions from Gasoline Fueled Engines 3-13
3.3.3.2. Emissions from Diesel Fueled Engines 3-13
3.3.4. Deposition of PM 3-14
3.3.4.1. Deposition Forms 3-16
3.3.4.2. Methods for Estimating Dry Deposition 3-17
3.3.4.3. Factors Affecting Dry Deposition Rates and Totals 3-18
3.4. Monitoring of PM 3-20
3.4.1. Ambient Measurement Techniques 3-20
3.4.1.1. PMMass 3-20
3.4.1.2. PMSpeciation 3-23
3.4.1.3. Multiple-Component Measurements on Individual Particles 3-28
3.4.1.4. UFPs: Mass, Surface Area, and Number 3-29
3.4.1.5. PM Size Distribution 3-29
3.4.1.6. Satellite Measurement 3-29
3.4.2. Ambient Network Design 3-30
3.4.2.1. Monitor Siting Requirements 3-30
3.4.2.2. Spatial and Temporal Coverage 3-31
3.4.2.3. Network Application for Exposure Assessment with Respect to
Susceptible Populations 3-36
3.5. Ambient PM Concentrations 3-40
December 2009
-------
3.5.1. Spatial Distribution 3-41
3.5.1.1. Variability across the U.S. 3-42
3.5.1.2. Urban-Scale Variability 3-60
3.5.1.3. Neighborhood-Scale Variability 3-85
3.5.2. Temporal Variability 3-91
3.5.2.1. Regional Trends 3-91
3.5.2.2. Seasonal Variations 3-96
3.5.2.3. Hourly Variability 3-97
3.5.3. Statistical Associations with Copollutants 3-100
3.6. Mathematical Modeling of PM 3-104
3.6.1. Estimating Source Contributions to PM Using Receptor Models 3-104
3.6.1.1. Receptor Models 3-104
3.6.2. Chemistry Transport Models 3-109
3.6.2.1. Global Scale 3-111
3.6.2.2. Regional Scale 3-111
3.6.2.3. Local or Neighborhood Scale 3-113
3.6.3. Air Quality Model Evaluation for Air Concentrations 3-113
3.6.3.1. Ground-based Comparisons of Photochemical Dynamics 3-120
3.6.3.2. Predicted Chemistry for Nitrates and Related Compounds 3-120
3.6.4. Evaluating Concentrations and Deposition of PM Components with CTMs 3-126
3.6.4.1. Global CTM Performance 3-126
3.6.4.2. Regional CTM Performance 3-127
3.7. Background PM 3-139
3.7.1. Contributors to PRB Concentrations of PM 3-139
3.7.1.1. Estimates of PRB Concentrations in Previous Assessments 3-140
3.7.1.2. Chemistry Transport Models for Predicting PRB Concentrations 3-142
3.8. Issues in Exposure Assessment for PM and its Components 3-152
3.8.1. General Exposure Concepts 3-153
3.8.2. Personal and Microenvironmental Exposure Monitoring 3-155
3.8.2.1. New Developments in Personal Exposure Monitoring Instrumentation 3-155
3.8.2.2. New Developments in Microenvironmental Exposure Monitoring
Instrumentation 3-156
3.8.3. Exposure Modeling 3-157
3.8.3.1. Time-Weighted Microenvironmental Models 3-157
3.8.3.2. Stochastic Population Exposure Models 3-158
3.8.3.3. Dispersion Models 3-160
3.8.3.4. Land Use Regression and GIS-Based Models 3-160
3.8.4. Exposure Assessment Studies 3-162
3.8.4.1. Micro-to-Neighborhood Scale Ambient PM Exposure 3-162
3.8.4.2. Ambient PM Exposure Estimates from Central Site Monitoring Data 3-165
3.8.4.3. Infiltration 3-168
3.8.5. Multicomponent and Multipollutant PM Exposures 3-170
3.8.5.1. Exposure Issues Related to PM Composition 3-170
3.8.5.2. Exposure to PM and Copollutants 3-175
3.8.6. Implications of Exposure Assessment Issues for Interpretation of Epidemiologic
Studies 3-176
3.8.6.1. Measurement Error 3-176
3.8.6.2. Model-Related Errors 3-177
3.8.6.3. Spatial Variability 3-179
3.8.6.4. Temporal Variability 3-181
3.8.6.5. Use of Surrogates for PM Exposure 3-183
3.8.6.6. Compositional Differences 3-184
3.8.6.7. Conclusions 3-184
3.9. Summary and Conclusions 3-185
3.9.1. Concentrations and Sources of Atmospheric PM 3-185
3.9.1.1. PM Source Characteristics 3-185
December 2009
-------
3.9.1.2. Measurement Techniques
3.9.1.3. Ambient PM Variability and Correlations
3.9.1.4. Temporal Variability
3.9.1.5. Correlations between Copollutants
3.9.1.6. Source Contributions to PM
3.9.1.7. Policy-Relevant Background
3.9.2. Human Exposure
3.9.2.1. Characterizing Human Exposure
3.9.2.2. Spatial Scales of PM Exposure Assessment
3.9.2.3. Multicomponent and Multipollutant PM Exposures
3.9.2.4. Implications for Epidemiologic Studies
Chapter 3 References
CHAPTER 4. DOSIMETRY
4.1. Introduction
4.1.1. Size Characterization of Inhaled Particles
4.1.2. Structure of the Respiratory Tract
4.2. Particle Deposition
4.2.1. Mechanisms of Deposition
4.2.2. Deposition Patterns
4.2.2. 1 . Total Respiratory Tract Deposition
4.2.2.2. Extrathoracic Region
4.2.2.3. Tracheobronchial and Alveolar Region
4.2.2.4. Localized Deposition Sites
4.2.3. Interspecies Patterns of Deposition
4.2.4. Biological Factors Modulating Deposition
4.2.4.1. Physical Activity
4.2.4.2. Age
4.2.4.3. Gender
4.2.4.4. Anatomical Variability
4.2.4.5. Respiratory Tract Disease
4.2.4.6. Hygroscopicity of Aerosols
4.2.5. Summary
4.3. Clearance of Poorly Soluble Particles
4.3.1 . Clearance Mechanisms and Kinetics
4.3.1.1. Extrathoracic Region
4.3.1.2. Tracheobronchial Region
4.3.1.3. Alveolar Region
4.3.2. Interspecies Patterns of Clearance and Retention
4.3.3. Particle Translocation
4.3.3.1. Alveolar Region
4.3.3.2. Olfactory Region
4.3.4. Factors Modulating Clearance
4.3.4.1. Age
4.3.4.2. Gender
4.3.4.3. Respiratory Tract Disease
4.3.4.4. Particle Overload
4.3.5. Summary
4.4. Clearance of Soluble Materials
4.4.1 . Clearance Mechanisms and Kinetics
4.4.2. Factors Modulating Clearance
4.4.2.1. Age
4.4.2.2. Physical Activity
4.4.2.3. Disease
4.4.2.4. Concurrent Exposures
3-185
3-186
3-187
3-188
3-188
3-189
3-189
3-189
3-190
3-191
3-191
3-193
4-1
4-1
4-2
4-3
4-5
4-6
4-7
4-8
4-9
4-10
4-10
4-11
4-11
4-12
4-13
4-14
4-14
4-15
4-16
4-16
4-17
4-17
4-17
4-18
4-19
4-19
4-20
4-21
4-22
4-23
4-23
4-24
4-24
4-25
4-25
4-26
4-26
4-27
4-28
4-28
4-28
4-29
December 2009
-------
4.4.3. Summary 4-29
Chapter 4 References 4-30
CHAPTER 5. POSSIBLE PATHWAYS/ MODES OF ACTION 5-1
5.1. Pulmonary Effects 5-1
5.1.1. Reactive Oxygen Species 5-1
5.1.2. Activation of Cell Signaling Pathways 5-3
5.1.3. Pulmonary Inflammation 5-4
5.1.4. Respiratory Tract Barrier Function 5-6
5.1.5. Antioxidant Defenses and Adaptive Responses 5-6
5.1.6. Pulmonary Function 5-7
5.1.7. Allergic Disorders 5-8
5.1.8. Impaired Lung Defense Mechanisms 5-8
5.1.9. Resolution of Inflammation/Progression or Exacerbation of Disease 5-8
5.1.9.1. Factors Affecting the Retention of PM 5-9
5.1.9.2. Factors Affecting the Balance of Pro/Anti-lnflammatory Mediators,
Oxidants/Anti-Oxidants and Proteases/Anti-Proteases 5-9
5.1.9.3. Pre-Existing Disease 5-10
5.1.10. Pulmonary DMA Damage 5-10
5.1.11. Epigenetic Changes 5-10
5.1.12. Lung Development 5-11
5.2. Systemic Inflammation 5-12
5.2.1. Endothelial Dysfunction and Altered Vasoreactivity 5-13
5.2.2. Activation of Coagulation and Acute Phase Response 5-14
5.2.3. Atherosclerosis 5-15
5.3. Activation of the Autonomic Nervous System by Pulmonary Reflexes 5-16
5.4. Translocation of UFPs or Soluble PM Components 5-17
5.5. Disease of the Cardiovascular and Other Organ Systems 5-18
5.6. Acute and Chronic Responses 5-19
5.7. Results of New Inhalation Studies which Contribute to Modes of Action 5-19
5.8. Gaps in Knowledge 5-22
Chapter 5 References 5-23
CHAPTERS. INTEGRATED HEALTH EFFECTS OF SHORT-TERM PM EXPOSURE 6-1
6.1. Introduction 6-1
6.2. Cardiovascular and Systemic Effects 6-2
6.2.1. Heart Rate and Heart Rate Variability 6-2
6.2.1.1. Epidemiologic Studies 6-2
6.2.1.2. Controlled Human Exposure Studies 6-8
6.2.1.3. lexicological Studies 6-10
6.2.2. Arrhythmia 6-13
6.2.2.1. Epidemiologic Studies 6-13
6.2.2.2. Toxicological Studies 6-18
6.2.3. Ischemia 6-20
6.2.3.1. Epidemiologic Studies 6-20
6.2.3.2. Controlled Human Exposure Studies 6-22
6.2.3.3. Toxicological Studies 6-23
6.2.4. Vasomotor Function 6-24
6.2.4.1. Epidemiologic Studies 6-24
6.2.4.2. Controlled Human Exposure Studies 6-26
6.2.4.3. Toxicological Studies 6-29
December 2009 vii
-------
6.2.5. Blood Pressure 6-33
6.2.5.1. Epidemiologic Studies 6-33
6.2.5.2. Controlled Human Exposure Studies 6-36
6.2.5.3. Toxicological Studies 6-37
6.2.6. Cardiac Contractility 6-38
6.2.6.1. Toxicological Studies 6-38
6.2.7. Systemic Inflammation 6-39
6.2.7.1. Epidemiologic Studies 6-40
6.2.7.2. Controlled Human Exposure Studies 6-44
6.2.7.3. Toxicological Studies 6-46
6.2.8. Hemostasis, Thrombosis and Coagulation Factors 6-47
6.2.8.1. Epidemiologic Studies 6-47
6.2.8.2. Controlled Human Exposure Studies 6-48
6.2.8.3. Toxicological Studies 6-50
6.2.9. Systemic and Cardiovascular Oxidative Stress 6-52
6.2.9.1. Epidemiologic Studies 6-52
6.2.9.2. Controlled Human Exposure Studies 6-53
6.2.9.3. Toxicological Studies 6-54
6.2.10. Hospital Admissions and Emergency Department Visits 6-56
6.2.10.1. All Cardiovascular Disease 6-60
6.2.10.2. Cardiac Diseases 6-64
6.2.10.3. Ischemic Heart Disease 6-64
6.2.10.4. Acute Myocardial Infarction 6-67
6.2.10.5. Congestive Heart Failure 6-68
6.2.10.6. Cardiac Arrhythmias 6-69
6.2.10.7. Cerebrovascular Disease 6-70
6.2.10.8. Peripheral Vascular Disease 6-72
6.2.10.9. Copollutant Models 6-72
6.2.10.10. Concentration Response 6-75
6.2.10.11. Out of Hospital Cardiac Arrest 6-76
6.2.11. Cardiovascular Mortality 6-77
6.2.12. Summary and Causal Determinations 6-78
6.2.12.1. PM25 6-78
6.2.12.2. PM10.25 6-81
6.2.12.3. UFPs 6-83
6.3. Respiratory Effects 6-84
6.3.1. Respiratory Symptoms and Medication Use 6-84
6.3.1.1. Epidemiologic Studies 6-84
6.3.1.2. Controlled Human Exposure Studies 6-93
6.3.2. Pulmonary Function 6-94
6.3.2.1. Epidemiologic Studies 6-95
6.3.2.2. Controlled Human Exposure Studies 6-98
6.3.2.3. Toxicological Studies 6-99
6.3.3. Pulmonary Inflammation 6-101
6.3.3.1. Epidemiologic Studies 6-101
6.3.3.2. Controlled Human Exposure Studies 6-104
6.3.3.3. Toxicological Studies 6-106
6.3.4. Pulmonary Oxidative Responses 6-110
6.3.4.1. Controlled Human Exposure Studies 6-111
6.3.4.2. Toxicological Studies 6-112
6.3.5. Pulmonary Injury 6-114
6.3.5.1. Epidemiologic Studies 6-114
6.3.5.2. Controlled Human Exposure Studies 6-114
6.3.5.3. Toxicological Studies 6-115
6.3.6. Allergic Responses 6-122
6.3.6.1. Epidemiologic Studies 6-122
6.3.6.2. Controlled Human Exposure Studies 6-122
6.3.6.3. Toxicological Studies 6-123
December 2009 viii
-------
6.3.7. Host Defense
6.3.7.1. Epidemiologic Studies
6.3.7.2. Toxicological Studies
6.3.8. Respiratory ED Visits, Hospital Admissions and Physician Visits
6.3.8.1. All Respiratory Diseases
6.3.8.2. Asthma
6.3.8.3. Chronic Obstructive Pulmonary Disease
6.3.8.4. Pneumonia and Respiratory Infections
6.3.8.5. Copollutant Models
6.3.9. Respiratory Mortality
6.3.10. Summary and Causal Determinations
6.3.10.1. PM,,
6.3.10.2. PMm.9,
6.3.10.3. UFPs
6.4. Central Nervous System Effects
6.4.1. Epidemiologic Studies
6.4.2. Controlled Human Exposure Studies
6.4.3. Toxicological Studies
6.4.3.1. Urban Air
6.4.3.2. CAPs
6.4.3.3. Diesel Exhaust
6.4.3.4. Summary of Toxicological Study Findings of CMS Effects
6.4.4. Summary and Causal Determination
6.5. Mortality
6.5.1 . Summary of Findings from 2004 PM AQCD
6.5.2. Associations of Mortality and Short-Term Exposure to PM
6.5.2.1. PMin
6.5.2.2. PM,fi
6.5.2.3. Thoracic Coarse Particles (PMin-2s)
6.5.2.4. Ultrafine Particles
6.5.2.5. Chemical Components of PM
6.5.2.6. Source-Apportioned PM Analyses
6.5.2.7. Investigation of Concentration-Response Relationship
6.5.3. Summary and Causal Determinations
6.5.3.1. PM7,
6.5.3.2. PMin.?s
6.5.3.3. UFPs
6.6. Attribution of Ambient PM Health Effects to Specific Constituents or Sources
6.6.1. Evaluation Approach
6.6.2. Findings
6.6.2. 1 . Epidemiologic Studies
6.6.2.2. Controlled Human Exposure Studies
6.6.2.3. Toxicological Studies
6.6.3. Summary by Health Effects
6.6.4. Conclusion
Chapter 6 References
CHAPTER?. INTEGRATED HEALTH EFFECTS OF LONG-TERM PM EXPOSURE
7.1. Introduction
7.2. Cardiovascular and Systemic Effects
7.2.1. Atherosclerosis
7.2.1.1. Epidemiologic Studies
7.2.1.2. Toxicological Studies
7.2.2. Venous Thromboembolism
7.2.2. 1 . Epidemiologic Studies
6-129
6-129
6-129
6-132
6-133
6-137
6-142
6-143
6-147
6-149
6-149
6-149
6-152
6-153
6-154
6-154
6-155
6-155
6-155
6-156
6-156
6-157
6-157
6-158
6-158
6-159
6-160
6-174
6-184
6-190
6-191
6-196
6-197
6-200
6-200
6-201
6-202
6-202
6-202
6-203
6-203
6-206
6-206
6-210
6-211
6-213
7-1
7-1
7-1
7-2
7-2
7-4
7-6
7-7
December 2009
-------
7.2.3. Metabolic Syndromes
7.2.3.1. Epidemiologic Studies
7.2.3.2. Toxicological Studies
7.2.4. Systemic Inflammation, Immune Function, and Blood Coagulation
7.2.4. 1 . Epidemiologic Studies
7.2.4.2. Toxicological Studies
7.2.5. Renal and Vascular Function
7.2.5.1. Epidemiologic Studies
7.2.5.2. Toxicological Studies
7.2.6. Autonomic Function
7.2.6.1. Toxicological Studies
7.2.7. Cardiac changes
7.2.7.1. Toxicological studies
7.2.8. Left Ventricular Mass and Function
7.2.9. Clinical Outcomes in Epidemiologic Studies
7.2.10. Cardiovascular Mortality
7.2.11. Summary and Causal Determinations
7.2.11. 1.PM,,
7.2.11.2. PMin.™
7.2.11.3. UFPs
7.3. Respiratory Effects
7.3.1 . Respiratory Symptoms and Disease Incidence
7.3.1.1. Epidemiologic Studies
7.3.2. Pulmonary Function
7.3.2.1. Epidemiologic Studies
7.3.2.2. Toxicological Studies
7.3.3. Pulmonary Inflammation
7.3.3.1. Epidemiologic Studies
7.3.3.2. Toxicological Studies
7.3.4. Pulmonary Oxidative Response
7.3.4.1. Toxicological Studies
7.3.5. Pulmonary Injury
7.3.5.1. Toxicological Studies
7.3.6. Allergic Responses
7.3.6.1. Epidemiologic Studies
7.3.6.2. Toxicological Studies
7.3.7. Host Defense
7.3.7.1. Epidemiologic Studies
7.3.7.2. Toxicological Studies
7.3.8. Respiratory Mortality
7.3.9. Summary and Causal Determinations
7.3.9.1. PM7,
7.3.9.2. PMin.?s
7.3.9.3. UFPs
7.4. Reproductive, Developmental, Prenatal and Neonatal Outcomes
7.4.1. Epidemiologic Studies
7.4.1.1. Low Birth Weight
7.4.1.2. Preterm Birth
7.4.1.3. Growth Restriction
7.4.1.4. Birth Defects
7.4.1.5. Infant Mortality
7.4.1.6. Decrements in Sperm Quality
7.4.2. Toxicological Studies
7.4.2.1. Female Reproductive Effects
7.4.2.2. Male Reproductive Effects
7.4.2.3. Multiple Generation Effects
7.4.2.4. Receptor Mediated Effects
7-7
7-7
7-7
7-8
7-8
7-8
7-9
7-10
7-11
7-12
7-12
7-12
7-12
7-13
7-13
7-17
7-18
7-18
7-19
7-19
7-20
7-20
7-20
7-26
7-26
7-30
7-32
7-32
7-32
7-34
7-34
7-35
7-35
7-38
7-38
7-39
7-40
7-40
7-40
7-41
7-42
7-42
7-43
7-44
7-44
7-44
7-45
7-48
7-51
7-52
7-53
7-58
7-58
7-59
7-60
7-62
7-63
December 2009
-------
7.4.2.5. Developmental Effects
7.4.3. Summary and Causal Determinations
7.4.3.1. PM,,
7.4.3.2. PMin.,,
7.4.3.3. UFPs
7.5. Cancer, Mutagenicity, and Genotoxicity
7.5.1. Epidemiologic Studies
7.5.1.1. Lung Cancer Mortality and Incidence
7.5.1.2. Other Cancers
7.5.1.3. Markers of Exposure or Susceptibility
7.5.2. Toxicological Studies
7.5.2.1. Mutagenesis and Genotoxicity
7.5.2.2. Carcinogenesis
7.5.2.3. Summary of Toxicological Studies
7.5.3. Epigenetic Studies and Other Heritable DMA mutations
7.5.4. Summary and Causal Determinations
7.5.4.1. PM,,
7.5.4.2. PMin.,,
7.5.4.3. UFPs
7.6. Mortality
7.6.1 . Recent Studies of Long-Term Exposure to PM and Mortality
7.6.2. Composition and Source-Oriented Analyses of PM
7.6.3. Within-City Effects of PM Exposure
7.6.4. Effects of Different Long-term Exposure Windows
7.6.5. Summary and Causal Determinations
7.6.5.1. PM,,
7.6.5.2. PMin.,,
7.6.5.3. UFPs
Chapter 7 References
CHAPTERS. POPULATIONS SUSCEPTIBLE TO PM-RELATED HEALTH EFFECTS
8.1. Potentially Susceptible Populations
8.1.1. Lifestaqe
8.1.1.1. Older Adults
8.1.1.2. Children
8.1.2. Pregnancy and Developmental Effects
8.1.3. Gender
8.1.4. Race/Ethnicity
8.1.5. Gene-Environment Interaction
8.1.6. Pre-Existing Disease
8. 1 .6. 1 . Cardiovascular Diseases
8.1.6.2. Respiratory Illnesses
8.1.6.3. Respiratory Contributions to Cardiovascular Effects
8.1.6.4. Diabetes and Obesity
8.1.7. Socioeconomic Status
8.1.8. Summary
Chapter 8 References
CHAPTER 9. WELFARE EFFECTS
9.1. Introduction
9.2. Effects on Visibility
9.2.1. Introduction
9.2.2. Backqround
9.2.2.1. Non-PM Visibility Effects
7-63
7-67
7-67
7-68
7-68
7-68
7-69
7-70
7-73
7-73
7-75
7-76
7-79
7-80
7-80
7-81
7-81
7-82
7-82
7-82
7-84
7-89
7-90
7-92
7-95
7-95
7-97
7-97
7-98
8-1
8-3
8-3
8-3
8-5
8-5
8-6
8-7
8-7
8-9
8-9
8-12
8-13
8-13
8-14
8-15
8-17
9-1
9-1
9-1
9-1
9-2
9-5
December 2009
-------
9.2.3.
9.2.4.
9.2.5.
9.3. Effects
9.3.1.
9.3.2.
9.3.3.
9.3.4.
9.3.5.
9.3.6.
9.3.7.
9.3.8.
9.3.9.
9.3.10.
9.2.2.2. PM Visibility Effects
9.2.2.3. Direct Optical Measurements
9.2.2.4. Value of Good Visual Air Quality
Monitoring and Assessment
9.2.3.1. Aerosol Properties
9.2.3.2. Spatial Patterns
9.2.3.3. Urban and Regional Patterns
9.2.3.4. Temporal Trends
9.2.3.5. Causes of Haze
Urban Visibility Valuation and Preference
9.2.4. 1 . Urban Visibility Preference Studies
9.2.4.2. Denver, Colorado Urban Visibility Preference Study
9.2.4.3. Phoenix, Arizona Urban Visibility Preference Study
9.2.4.4. British Columbia, Canada Urban Visibility Preference Study
9.2.4.5. Washington, DC Urban Visibility Preference Studies
9.2.4.6. Urban Visibility Valuation Studies
Summary of Effects on Visibility
on Climate
The Climate Effects of Aerosols
Overview of Aerosol Measurement Capabilities
9.3.2.1. Satellite Remote Sensing
9.3.2.2. Focused Field Campaigns
9.3.2.3. Ground-Based In Situ Measurement Networks
9.3.2.4. In Situ Aerosol Profiling Programs
9.3.2.5. Ground-Based Remote Sensing Measurement Networks
9.3.2.6. Synergy of Measurements and Model Simulations
Assessments of Aerosol Characterization and Climate Forcing
9.3.3.1. The Use of Measured Aerosol Properties to Improve Models
9.3.3.2. Intercomparisons of Satellite Measurements and Model Simulation of
Aerosol Optical Depth
9.3.3.3. Satellite-Based Estimates of Aerosol Direct Radiative Forcing
9.3.3.4. Satellite-Based Estimates of Anthropogenic Component of Aerosol
Direct Radiative Forcing
9.3.3.5. Aerosol-Cloud Interactions and Indirect Forcing
9.3.3.6. Remote Sensing of Aerosol-Cloud Interactions and Indirect Forcing
9.3.3.7. In Situ Studies of Aerosol-Cloud Interactions
Outstanding Issues
Concluding Remarks
Modeling the Effect of Aerosols on Climate
9.3.6.1. Introduction
9.3.6.2. Modeling of Atmospheric Aerosols
9.3.6.3. Calculating Aerosol Direct Radiative Forcing
9.3.6.4. Calculating Aerosol Indirect Forcing
9.3.6.5. Aerosol in the Climate Models
9.3.6.6. Impacts of Aerosols on Climate Model Simulations
9.3.6.7. Outstanding Issues
9.3.6.8. Conclusions
Fire as a Special Source of PM Welfare Effects
Radiative Effects of Volcanic Aerosols
9.3.8.1. Explosive Volcanic Activity
Other Special Sources and Effects
9.3.9.1. Glaciers and Snowpack
9.3.9.2. Radiative Forcing by Anthropogenic Surface Albedo Change: BC in
Snow and Ice
9.3.9.3. Effects on Local and Regional Climate
Summary of Effects on Climate
9-5
9-8
9-10
9-10
9-11
9-16
9-23
9-31
9-37
9-65
9-67
9-68
9-69
9-69
9-69
9-71
9-72
9-74
9-74
9-82
9-82
9-88
9-89
9-91
9-95
9-96
9-99
9-100
9-103
9-105
9-111
9-112
9-113
9-116
9-117
9-120
9-121
9-121
9-124
9-129
9-137
9-145
9-153
9-157
9-158
9-159
9-160
9-160
9-164
9-167
9-169
9-170
9-171
9.4. Ecological Effects of PM 9-172
December 2009 xii
-------
9.4.1. Introduction
9.4.1.1. Ecosystem Scale, Function, and Structure
9.4.1.2. Ecosystem Services
9.4.2. Deposition of PM
9.4.2. 1 . Forms of Deposition
9.4.2.2. Components of PM Deposition
9.4.2.3. Magnitude of Dry Deposition
9.4.3. Direct Effects of PM on Vegetation
9.4.3.1. Effects of Coarse-mode Particles
9.4.4. PM and Altered Radiative Flux
9.4.5. Effects of Trace Metals on Ecosystems
9.4.5.1. Effects on Soil Chemistry
9.4.5.2. Effects on Soil Microbes and Plant Uptake via Soil
9.4.5.3. Plant Response to Metals
9.4.5.4. Effects on Aquatic Ecosystems
9.4.5.5. Effects on Animals
9.4.5.6. Biomagnification across Trophic Levels
9.4.5.7. Effects near Smelters and Roadsides
9.4.5.8. Toxicity to Mosses and Lichens
9.4.6. Organic Compounds
9.4.7. Summary of Ecological Effects of PM
9.5. Effects on Materials
9.5.1. Effects on Paint
9.5.2. Effects on Metal Surfaces
9.5.3. Effects on Stone
9.5.4. Summary of Effects on Materials
Chapter 9 References
ANNEX A. ATMOSPHERIC SCIENCE
Annex A References
ANNEX B. DOSIMETRY
Annex B References
ANNEX C. CONTROLLED HUMAN EXPOSURE STUDIES
Annex C References
ANNEX D. TOXICOLOGICAL STUDIES
Annex D References
ANNEX E. EPIDEMIOLOGIC STUDIES
Annex E References
ANNEX F. SOURCE APPORTIONMENT STUDIES
Annex F References
9-172
9-173
9-174
9-174
9-175
9-175
9-179
9-182
9-182
9-183
9-183
9-185
9-186
9-189
9-192
9-192
9-193
9-194
9-196
9-196
9-199
9-201
9-202
9-202
9-203
9-203
9-204
A-1
A-353
B-1
B-7
C-1
C-14
D-1
D-172
E-1
E-524
F-1
F-11
December 2009 xiii
-------
List of Tables
Table 1-1. Summary of NAAQS promulgated for PM, 1971-2006. 1-6
Table 1 -2. Aspects to aid in judging causality. 1 -20
Table 1-3. Weight of evidence for causal determination. 1-22
Table 2-1. Summary of causal determinations for short-term exposure to PM2 5. 2-36
Table 2-2. Summary of causal determinations for long-term exposure to PM2 5. 2-38
Table 2-3. Summary of causal determinations for short-term exposure to PM10.2.5. 2-45
Table 2-4. Summary of causal determinations for short-term exposure to UFPs. 2-48
Table 2-5. Summary of causality determination for welfare effects. 2-54
Table 2-6. Summary of PM causal determinations by exposure duration and health outcome. 2-59
Table 2-7. Summary of PM causal determinations for welfare effects 2-60
Table 3-1. Characteristics of ambient fine (ultrafine plus accumulation-mode) and coarse particles. 3-4
Table 3-2. Constituents of atmospheric particles and their major sources. 3-7
Table 3-3. Proximity to PM2 5 and PM10 monitors for total population3 by city. 3-35
Table 3-4. Proximity to PM2 5 and PM10 monitors for children age 0-4 yr, children age 5-17 yr, and
adults age 65 yr and older.3 The figures presented here are cumulative for the 15
CSAs/CBSAs examined in Chapter 3. 3-37
Table 3-5. Proximity to PM2 5 and PM10 monitors for adults age 65 yr and older3 by city. 3-38
Table 3-6. Proximity to PM25 and PM10 monitors based on the population identified as white, black,
Hispanic, or non-Hispanic3. 3-39
Table 3-7. Proximity to PM25 and PM10 monitors based on the population below or above the poverty
line, population over age 25 with less than high school education, population over 25 with
high school education, and population over 25 with college education or more3. 3-40
Table 3-8. PM25 distributions derived from AQS data (concentration in ug/m3). 3-44
Table 3-9. PM10.2.5 distributions derived from AQS data (concentration in ug/m3). 3-47
Table 3-10. PM10 distributions derived from AQS data (concentration in ug/m3). 3-49
Table 3-11. Inter-sampler comparison statistics for each pair of 24-h PM2 5 monitors reporting to AQS
for Boston, MA. 3-63
Table 3-12. Inter-sampler comparison statistics for each pair of 24-h PM2 5 monitors reporting to AQS
for Pittsburgh, PA. 3-67
Table 3-13. Inter-sampler comparison statistics for each pair of 24-h PM2 5 monitors reporting to AQS
for Los Angeles, CA. 3-69
Table 3-14. Inter-sampler comparison statistics for each pair of 24-h PM10 monitors reporting to AQS for
Boston, MA. 3-75
Table 3-15. Inter-sampler comparison statistics for each pair of 24-h PM10 monitors reporting to AQS for
Pittsburgh, PA. 3-78
Table 3-16. Inter-sampler comparison statistics for each pair of 24-h PM10 monitors reporting to AQS for
Los Angeles, CA. 3-81
December 2009 xiv
-------
Table 3-1 7.
Table 3-1 8.
Table 3-1 9.
Table 3-20.
Table 3-21.
Table 3-22.
Table 3-23.
Table 3-24.
Table 4-1 .
Table 6-1 .
Table 6-2.
Table 6-3.
Table 6-4.
Table 6-5.
Table 6-6.
Table 6-7.
Table 6-8.
Table 6-9.
Table 6-10.
Table 6-11.
Table 6-1 2.
Table 6-1 3.
Table 6-1 4.
Table 6-1 5.
Example of emissions factors (ng/km) for trace elements under variable speed and steady
speed drivinq conditions for PM emitted bv diesel and qasoline engines.
Estimates of annual average natural background concentrations of PM2.5 and PM10 (ug/m3)
from Triionisetal. (1990, 157058).
Annual and quarterly mean PM2 5 concentrations (ug/m3) measured at IMPROVE sites in
2004.
Annual and quarterly mean PM2.5 concentrations (ug/m3) for the CMAQ "base case" at
IMPROVE sites in 2004.
Annual and quarterly mean PM2 5 concentrations (ug/m3) for the CMAQ PRB simulations at
IMPROVE sites in 2004.
Annual and quarterly mean of the CMAQ-predicted base case PM25 concentrations (ug/m3)
in the U.S. EPA CONUS reqions in 2004.
Annual and quarterly mean of the CMAQ-predicted PRB PM25 concentrations (ug/m3) in
the U.S. EPA CONUS reqions in 2004.
Examples of studies comparing near-road personal exposures with fixed site ambient
concentrations.
Breathing patterns with activity level in adult human male.
Characteristics of epidemiologic studies investigating associations between PM and
changes in HRV.
Epidemiologic studies of ventricular arrhythmia and ambient PM concentration, in patients
with implantable cardioverter defibrillators.
PM Concentrations reported in epidemiologic studies ECG changes suggestive of
ischemia.
PM concentrations reported in epidemioloqic studies of vasomotor function.
Mean PM concentrations reported in epidemioloqic studies of blood pressure.
PM concentrations reported in epidemiologic studies of inflammation, hemostasis,
thrombosis, coagulation factors and oxidative stress.
Description of ICD-9 and ICD-10 codes for diseases of the circulatory system.
Characterization of ambient PM concentrations in epidemiologic studies of hospital
admission and ED visits for cardiovascular diseases.
PM concentrations reported in studies of out-of-hospital cardiac arrest.
Characterization of ambient PM concentrations from epidemiologic studies of respiratory
morbidity and short-term exposures in asthmatic children and adults.
PAMCHAR PMm-?
-------
Table 6-17. Effect modification of composition on the estimated percent increase in mortality with a 10
|jg/m3 increase in PM2.5. 6-194
Table 6-18. Study-specific PM2.5 factor/source categories associated with health effects. 6-207
Table 7-1. Characterization of ambient PM concentrations from studies of subclinical measures of
cardiovascular diseases and long-term exposure. 7-11
Table 7-2. Characterization of ambient PM concentrations from studies of clinical cardiovascular
diseases and long-term exposure. 7-14
Table 7-3. Characterization of ambient PM concentrations from studies of respiratory
symptoms/disease and long-term exposures. 7-22
Table 7-4. Characterization of ambient PM concentrations from studies of FEVi and long-term
exposures. 7-27
Table 7-5. Characterization of ambient PM concentrations from studies of reproductive,
developmental, prenatal and neonatal outcomes and long-term exposure. 7-46
Table 7-6. Characterization of ambient PM concentrations from recent studies of cancer and long-term
exposures to PM. 7-70
Table 7-7. Associations* between ambient PM concentrations from select studies of lung cancer
mortality and incidence. 7-72
Table 7-8. Characterization of ambient PM concentrations from studies of mortality and long-term
exposures to PM. 7-83
Table 7-9. Comparison of results from ACS intra-urban analysis of Los Angeles and New York City
using kriging or land use regression to estimate exposure. 7-91
Table 7-10. Distribution of the effect of a hypothetical reduction of 10 ug/m3 PM10 in 2000 on all-cause
mortality 2000-2009 in Switzerland. 7-94
Table 8-1. Definitions of susceptible and vulnerable in the PM literature. 8-2
Table 8-2. Susceptibility factors evaluated. 8-3
Table 8-3. Percent of the U.S. population with respiratory diseases, cardiovascular diseases, and
diabetes. 8-11
Table 9-1. Regional Planning Organization websites with visibility characterization and source
attribution assessment information. 9-17
Table 9-2. Summary of urban visibility preference studies. 9-67
Table 9-3. Top-of-atmosphere, cloud-free, instantaneous direct aerosol radiative forcing dependence
on aerosol and surface properties. 9-81
Table 9-4. Summary of major satellite measurements currently available for the tropospheric aerosol
characterization and radiative forcing research. 9-83
Table 9-5. List of major intensive field experiments that are relevant to aerosol research in a variety of
aerosol regimes around the globe conducted in the past two decades. 9-93
Table 9-6. Summary of major U.S. surface in situ and remote sensing networks for the tropospheric
aerosol characterization and radiative forcing research. 9-94
Table 9-7. Summary of approaches to estimating the aerosol direct radiative forcing in three
categories: (1) satellite retrievals; (2) satellite-model integrations; and (3) model
simulations. 9-106
Table 9-8. Summary of seasonal and annual average clear-sky DRF (W/m2) at the TOA and the
surface (SFC) over global OCEAN derived with different methods and data. 9-109
December 2009 xvi
-------
Table 9-9.
Table 9-10.
Table 9-11.
Table 9-12.
Table 9-13.
Table 9-14.
Summary of seasonal and annual average clear-sky DRF (W/m ) at the TOA and the
surface (SFC) over global LAND derived with different methods and data.
Estimates of anthropogenic components of aerosol optical depth (Tant) and clear-sky DRF
at the TOA from model simulations.
Anthropogenic emissions of aerosols and precursors for 2000 and 1750.
Summary of statistics of AeroCom Experiment A results from 16 global models..
, 2- .
S04 mass loading, MEE and AOD at 550 nm, shortwave radiative forcing at the top of the
atmosphere, and normalized forcing with respect to AOD and mass.
Particulate organic matter (POM) and BC mass loading, AOD at 550 nm, shortwave
radiative forcing at the top of the atmosphere, and normalized forcing with respect to AOD
and mass.
_9-126
_9-127
9-132
Table 9-1 5.
Table 9-1 6.
Table 9-1 7.
Table 9-1 8.
Table 9-1 9.
Table 9-20.
Table 9-21.
Table A-1 .
Table A-2.
Table A-3.
Table A-4.
Table A-5.
Table A-6.
Table A-7.
Table A-8.
Table A-9.
Table A-1 0.
Table A-1 1.
Table A-1 2.
Table A-1 3.
Table A-1 4.
Table A-1 5.
Differences in present day and pre-industrial outgoing solar radiation (W/m2) in the different
experiments.
Forcinqs used in IPCC AR4 simulations of 20th century climate chanqe.
Climate forcings (1880-2003) used to drive GISS climate simulations, along with the
surface air temperature changes obtained for several periods.
Overview of the different aerosol indirect effects and their sign of the net radiative flux
chanqe at the top of the atmosphere (TOA).
Overview of the different aerosol indirect effects and their implications for the global mean
net shortwave radiation of the surface Fsfc (columns 2-4) and for precipitation (columns
5-7).
Recent studies hiqhliqhtinq POP occurrence and fate in the maior arctic compartments.
Factors potentially important in estimatinq Hq exposure.
Summary of inteqrated and continuous samplers included in the field comparison.
Summary of PM?^ and PMm FRMand FEM samplers.
Measurement and analytical specifications for filter analysis of mass, elements, ions, and
carbon.
Measurement and analytical specifications for filter analysis of orqanic species.
Measurement and analytical specifications for continuous mass and mass surrogate
instruments.
Measurement and analytical specifications for continuous elements.
Measurement and analytical specifications for continuous N0r.
Measurement and analytical specifications for continuous Sd2".
Measurement and analytical specifications for ions other than NOf and Sd2".
Measurement and analytical specifications for continuous carbon.
Summary of mass measurement comparisons.
Summary of element and liquid water content measurement comparisons.
Summary of PM? ^ NOV measurement comparisons.
Summary of PM? ^ Sd2" measurement comparisons
Summary of PM? $ carbon measurement comparisons.
9-141
9-145
9-154
9-166
9-166
9-168
9-177
A-1
A-5
A-7
A-9
A-11
A-1 4
A-1 5
A-1 7
A-1 9
A-20
A-23
A-31
A-32
A-38
A-41
December 2009
XVII
-------
TableA-16.
TableA-17.
TableA-18.
TableA-19.
Table A-20.
Table A-21.
Table A-22.
Table A-23.
Table A-24.
Table A-25.
Table A-26.
Table A-27.
Table A-28.
Table A-29.
Table A-30.
Table A-31.
Table A-32.
Table A-33.
Table A-34.
Table A-35.
Table A-36.
Table A-37.
Table A-38.
Summary of particle mass spectrometer measurement comparisons.
Summary of key parameters for TD-GC/MS and pyrolysis-GC/MS.
Relevant Spatial Scales for PMm, PM^, and PMm-?<; Measurement
Maior routine operatinq air monitorinq networks3
Inter-sampler correlation statistics for each pair of PM2 5 monitors reporting to AQS for
Atlanta, GA.
Inter-sampler correlation statistics for each pair of PM2 5 monitors reporting to AQS for
Birminqham, AL.
Inter-sampler correlation statistics for each pair of PM2 5 monitors reporting to AQS for
Boston, MA.
Inter-sampler correlation statistics for each pair of PM2 5 monitors reporting to AQS for
Chicaqo, IL.
Inter-sampler correlation statistics for each pair of PM2 5 monitors reporting to AQS for
Denver, CO.
Inter-sampler correlation statistics for each pair of PM2 5 monitors reporting to AQS for
Detroit, Ml.
Inter-sampler correlation statistics for each pair of PM2 5 monitors reporting to AQS for
Houston, TX.
Inter-sampler correlation statistics for each pair of PM2 5 monitors reporting to AQS for Los
Anqeles, CA.
Inter-sampler correlation statistics for each pair of PM25 monitors reporting to AQS for New
York, NY.
Inter-sampler correlation statistics for each pair of PM2 5 monitors reporting to AQS for
Philadelphia, PA.
Inter-sampler correlation statistics for each pair of PM2 5 monitors reporting to AQS for
Phoenix, AZ.
Inter-sampler correlation statistics for each pair of PM2 5 monitors reporting to AQS for
Pittsburgh, PA.
Inter-sampler correlation statistics for each pair of PM2 5 monitors reporting to AQS for
Riverside, CA.
Inter-sampler correlation statistics for each pair of PM2 5 monitors reporting to AQS for
Seattle, WA.
Inter-sampler correlation statistics for each pair of PM2 5 monitors reporting to AQS for St.
Louis, MO.
Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Atlanta, GA.
Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Birminqham, AL.
Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Boston, MA.
Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Chicaqo, IL.
A-48
A-52
A-54
A-56
A-98
A-102
A-106
A-111
A-116
A-120
A-123
A-127
A-131
A-136
A-140
A-144
A-148
A-152
A-155
A-159
A-163
A-166
A-170
December 2009
XVIII
-------
Table A-39.
Table A-40.
Table A-41.
Table A-42.
Table A-43.
Table A-44.
Table A-45.
Table A-46.
Table A-47.
Table A-48.
Table A-49.
Table A-50.
Table A-51.
Table A-52.
Table A-53.
Table A-54.
Table A-55.
Table A-56.
Table A-57.
Table A-58.
Table A-59.
Table A-60.
Table A-61.
Table A-62.
Table A-63.
Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Denver, CO.
Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Detroit, Ml.
Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Houston, TX.
Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for Los
Anqeles, CA.
Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for New
York, NY.
Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Philadelphia, PA.
Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Phoenix, AZ.
Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Pittsburgh, PA.
Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Riverside, CA.
Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for
Seattle, WA.
Inter-sampler correlation statistics for each pair of PM10 monitors reporting to AQS for St.
Louis, MO.
Correlation coefficients of hourly and daily average particle number, surface and volume
concentrations in selected particle size ranges.
Different receptor models used in the Supersite source apportionment studies: chemical
mass balance.
Different receptor models used in the Supersites source apportionment studies: factor
analysis.
Different receptor models used in the Supersites source apportionment studies: tracer-
based methods.
Different receptor models used in the Supersites source apportionment studies:
meteorology-based methods.
Source Profiles: Part I
PM7^ receptor model results (pg/m3)
PMm receptor model results (mass percent)
Exposure Assessment Study Summaries
Examples of studies showing developments with UFP sampling methods since the 2004
PMAQCD.
Summary of in-vehicle studies of exposure assessment.
Summary of personal PM exposure studies with no indoor source during 2002-2008.
Summary of PM species exposure studies.
Summary of personal PM exposure source apportionment studies.
A-174
A-178
A-181
A-184
A-188
A-191
A-195
A-200
A-204
A-207
A-211
A-212
A-273
A-274
A-278
A-280
A-286
A-289
A-290
A-291
A-317
A-318
A-320
A-324
A-337
December 2009
XIX
-------
Table A-64.
Table A-65.
Table B-1 .
Table B- 2.
Table B- 3.
Table B-4.
Table B-5.
Table C-1.
Table C-2.
Table C- 3.
Table D-1.
Table D-2.
Table D-3.
Table D-4.
Table D-5.
Table D-6.
Table D-7.
Table D-8.
Table E-1
Table E-2.
Table E-3.
Table E-4.
Table E-5.
Table E-6.
Table E-7.
Table E-8.
Table E-9.
Table E-1 0.
Table E-1 1.
Table E-1 2.
Table E-1 3.
Table E-1 4.
Table E-1 5.
Table E-1 6.
Table E-1 7.
Summary of PM infiltration studies.
Summary of PM - copollutant exposure studies.
Ultrafine disposition in humans.
Ultrafine disposition in animals.
In vitro studies of ultrafine disposition.
Olfactory particle translocation.
Studies of respiratory tract mucosal and macrophaqe clearance as a function of aqe.
Cardiovascular effects.
Respiratory effects.
Central nervous system effects.
Cardiovascular effects.
Respiratory effects: in vitro studies.
Respiratory effects: in vivo studies.
Effects related to immunity and allerqy.
Effects of the central nervous system.
Reproductive and developmental effects.
Mutaqenic/qenotoxic effects in bacterial cultures.
Mutaqenicity and qenotoxicity data summary: In vitro and in vivo.
Short-term exposure - cardiovascular morbidity outcomes: PMm
Short-term exposure - cardiovascular morbidity studies: PMio-2.5.
Short-term exposure - cardiovascular morbidity studies: PlVhs (including PM
components/sources).
Short-term exposure-cardiovascular morbidity studies: Other size fractions.
Short-term exposure-cardiovascular: ED/HA PMm
Short-term exposure-cardiovascular-ED/HA - PMin.?^.
Short-term exposure - cardiovascular: ED/HA PM?^ (includinq PM components/sources)
Short-term exposure-cardiovascular-ED/HA-other size fractions.
Short-term exposure-respiratory morbidity outcomes -PMio.
Short-term exposure - respiratory morbidity outcomes - PMio-2.5.
Short-term exposure - respiratory morbidity outcomes - PM2 5 (including
components/sources).
Short-term exposure-respiratory-ED/HA-PMio.
Short-term exposure-respiratory-ED/HA-PMio-2.5.
Short-term exposure-respiratory-ED/HA-PMzs (includinq PM components/sources).
Short-term exposure-respiratory-ED/HA-Other Size Fractions.
Short-term exposure-mortality - PMio.
Short-term exposure-mortality - PMio-2.5.
A-339
A-350
B-1
B-2
B-3
B-3
B-5
C-1
C-9
C-1 3
D-1
D-31
D-81
D-1 16
D-1 55
D-1 57
D-1 63
D-1 66
E-1
E-1 9
E-21
E-79
E-85
E-1 04
E-1 07
E-1 30
E-1 34
E-1 69
E-1 73
E-230
E-256
E-262
E-284
E-293
E-343
December 2009 xx
-------
TableE-18.
TableE-19.
Table E-20.
Table E-21.
Table E-22.
Table E-23.
Table E-24.
Table E-25.
Table E-26.
Table E-27.
Table E-28.
Table E-29.
Table E-30.
Table E-31.
Table E-32.
Table E-33.
Table E-34.
Table F-1 .
Table F-2.
Table F-3.
Short-term exposure-mortality - PM^ (includinq PM components/sources).
Short-term exposure-mortality - other PM size fractions.
Lonq-term exposure - cardiovascular morbidity outcomes - PMm.
Lonq-term effects-cardiovascular- PMzs (includinq PM components/sources).
Lonq-term exposure - respiratory morbidity outcomes - PMm.
Lonq-term exposure - respiratory morbidity outcomes - PMm-zs.
Long-term exposure - respiratory morbidity outcomes - PMz.5 (including PM
components/sources).
Lonq-term exposure - respiratory morbidity outcomes - other PM size fractions.
Lonq-term exposure - cancer outcomes - PMm.
Lonq-term exposure - cancer outcomes - PMz.5 (includinq PM components/sources).
Lonq-term exposure - cancer outcomes - other PM size fractions.
Lonq-term exposure - reproductive outcomes - PMm.
Lonq-term exposure-mortality - PMm.
Lonq-term exposure-mortality - PMm.
Lonq-term exposure-mortality - PMio-2.5.
Lonq-term exposure-mortality - PMzs (includinq PM components/sources).
Lonq-term exposure - central nervous system outcomes - PM.
Epidemioloqic studies of ambient PM sources, factors, or constituents
Human clinical studies of ambient PM sources, factors, or constituents
Toxicoloqical studies of ambient PM sources, factors, or constituents
E-346
E-360
E-363
E-372
E-383
E-407
E-412
E-432
E-434
E-437
E-440
E-441
E-491
E-498
E-499
E-500
E-521
F-1
F-6
F-7
December 2009 xxi
-------
List of Figures
Figure 1-1.
Figure 2-1 .
Figure 2-2.
Figure 2-3.
Figure 3-1 .
Figure 3-2.
Figure 3-3.
Figure 3-4.
Figure 3-5.
Figure 3-6.
Figure 3-7.
Figure 3-8.
Figure 3-9.
Figure 3-10.
Figure 3-11.
Figure 3-1 2.
Figure 3-1 3.
Figure 3-1 4.
Figure 3-1 5.
Figure 3-1 6.
Figure 3-1 8.
Identification of studies for inclusion in the ISA.
Summary of effect estimates (per 10 ug/m3) by increasing concentration from U.S. studies
examining the association between short-term exposure to PM25 and cardiovascular and
respiratory effects, and mortality
Summary of effect estimates (per 10 ug/m3) by increasing concentration from U.S. studies
examining the association between long-term exposure to PM25 and cardiovascular and
respiratory effects, and mortality
Summary of U.S. studies examining the association between short-term exposure to PM10.
?<; and cardiovascular morbidity/mortality and respiratory morbidity/mortality
Particle size distributions by number and volume.
X-ray spectra and scanning electron microscopy images of individual particles.
Detailed source categorization of anthropogenic emissions of primary PM2.5, PM10 and
gaseous precursor species S02, NOX, NH3 and VOCs for 2002 in units of million metric
tons (MMT).
Primary emissions and formation of SOA through gas, cloud and condensed phase
reactions.
Schematic of the resistance-in-series analogy for atmospheric deposition.
The relationship between particle diameter and Vri for particles.
PM?<; monitor distribution in comparison with population density, Boston CSA.
PMm monitor distribution in comparison with population density, Boston CSA.
Three-yr avg 24-h PM25 concentration by county derived from FRM or FRM-like data,
2005-2007.
Three-yr avg 24-h PM10.2.5 concentration by county derived from co-located low volume
FRM PMm and PM,^ monitors, 2005-2007.
Three-yr avg 24-h PM10 concentration by county derived from FRM or FEM monitors,
2005-2007.
Three-yr avg 24-h PM25 OC concentrations measured at CSN sites across the U.S., 2005-
2007.
Three-yr avg 24-h PM25 EC concentrations measured at CSN sites across the U.S., 2005-
2007.
Three-yr avg 24-h PM25 S042" concentrations measured at CSN sites across the U.S.,
2005-2007.
Three-yr avg 24-h PM25 N03" concentrations measured at CSN sites across the U.S.,
2005-2007.
Three-yr avg 24-h PM25 NH4+ concentrations measured at CSN sites across the U.S.,
2005-2007.
Seasonally-stratified 3-yr avg PM2 5 speciation estimates for 2005-2007 derived using the
SANDWICH method.
1-11
2-41
2-42
2-47
3-2
3-5
3-9
3-12
3-15
3-19
3-33
3-34
3-43
3-46
3-48
3-52
3-53
3-54
3-55
3-56
3-58
December 2009 xxii
-------
Figure 3-1 9.
Figure 3-20.
Figure 3-21.
Figure 3-22.
Figure 3-23.
Figure 3-24.
Figure 3-25.
Figure 3-26.
Figure 3-27.
Figure 3-28.
Figure 3-29.
Figure 3-30.
Figure 3-31.
Figure 3-32.
Figure 3-33.
Figure 3-34.
Figure 3-35.
Figure 3-36.
Figure 3-37.
Figure 3-38.
Figure 3-39.
Figure 3-40.
Figure 3-41 .
Locations of PM,^ monitors and maior highways, Boston, MA.
Seasonal distribution of 24-h avq PM^ concentrations bv site for Boston, MA, 2005-2007.
Locations of PM?^ monitors and maior highways, Pittsburgh, PA.
Seasonal distribution of 24-h avg PM25 concentrations by site for Pittsburgh, PA, 2005-
2007.
Locations of PM? ^ monitors and maior highways, Los Angeles, CA.
Seasonal distribution of 24-h avg PM2 5 concentrations by site for Los Angeles, CA, 2005-
2007.
Inter-sampler correlations for 24-h PM2 5 as a function of distance between monitors in
Boston, MA.
Inter-sampler correlations for 24-h PM2 5 as a function of distance between monitors in
Pittsburgh, PA.
Inter-sampler correlations for 24-h PM2 5 as a function of distance between monitors in Los
Angeles, CA.
Seasonal distribution of 24-h avg PMm-?<; concentrations by site
Locations of PMm monitors and maior highways, Boston, MA.
Seasonal distribution of 24-h avg PMm concentrations by site for Boston, MA, 2005-2007.
Locations of PMm monitors and maior highways, Pittsburgh, PA.
Seasonal distribution of 24-h avg PM10 concentrations by site for Pittsburgh, PA, 2005-
2007.
Locations of PMm monitors and maior highways, Los Angeles, CA.
Seasonal distribution of 24-h avg PM10 concentrations by site for Los Angeles, CA, 2005-
2007.
Inter-sampler correlations for 24-h PM10 as a function of distance between monitors in
Boston, MA.
Inter-sampler correlations for 24-h PM10 as a function of distance between monitors in
Pittsburgh, PA.
Inter-sampler correlations for 24-h PM10 as a function of distance between monitors in Los
Angeles, CA.
Bin-wise Spearman correlation coefficients in aerosol particle number concentrations
between the Ift (urban background) and the Eisenbahn-strasse (city/urban center) sites in
Leipzig, Germany.
Dimensionless concentration as a function of height at windward and leeward locations and
street canyon aspect ratios (H/W).
Inter-sampler correlations for 24-h PM2 5 and PM10 as a function of distance between
monitors for samplers located within 4 km (neighborhood scale).
Particle size distributions measured at various distances from the 71 0 freeway in Los
3-61
3-62
3-65
3-66
3-68
3-69
3-71
3-71
3-72
3-73
3-74
3-75
3-76
3-77
3-80
3-81
3-82
3-83
3-83
3-84
3-86
3-87
Angeles, CA (top), and particle number concentration as a function of distance from the
710 freeway (bottom). 3-88
Figure 3-42. Mass distributions for BaP at a high traffic urban center (HTC), high traffic urban periphery
(HTP), low traffic urban center (LTC), low traffic urban periphery (LTP), and low traffic
industrial urban periphery (LTIP) in Seville, Spain. 3-90
December 2009 xxiii
-------
Figure 3-43. Mass distributions for 16 PAHs at a high traffic city center in Seville, Spain. 3-91
Figure 3-44. Ambient 24-h PM2.5 concentrations in the U.S., 1999-2007, showing A) ambient
concentrations, B) number of trends sites above the 24-h NAAQSand C) trends by U.S.
EPA Region. 3-92
Figure 3-45. Ambient annual PM2.5 concentrations in the U.S., 1999-2007, showing A) ambient
concentrations, B) number of trends sites above the annual NAAQS and C) trends by U.S.
EPA Region. 3-93
Figure 3-46. Ambient 24-h PM10 concentrations in the U.S., 1988-2007, showing A) ambient
concentrations, B) number of trends sites above the 24-h NAAQSand C) trends by U.S.
EPA Region. 3-94
Figure 3-47. Regional and seasonal trends in annual PM2 5 compostion from 2002 to 2007 derived using
the SANDWICH method. 3-95
Figure 3-48. UFP size distribution at highway (site A) and background (site B) sites in Los Angeles, CA,
during summer and winter seasons, with winter broken into day and evening distributions. 3-97
Figure 3-49. Diel plot generated from hourly FRM-like PM2.5 data ()4)/m3) stratified by weekday (left) and
weekend (right) for Pittsburgh, PA, and Seattle, WA, 2005-2007. 3-98
Figure 3-50. Diel plots generated from hourly FEM PM10 data ()4)/m3) stratified by weekday (left) and
weekend (right) for Chicago, IL, and Phoenix, AZ, 2005-2007. 3-99
Figure 3-51. Average diel variation in total particle number (ToN) and total particle volume (ToV) on
weekdays (left column) and Sundays (right column) from two sites in Denmark: one in a
busy street canyon (Jagtv) and one measuring urban background (HC0). 3-100
Figure 3-52. Distribution of correlations between 24-h avg PM25 and co-located 24-h avg PM10, PM10.2.5,
S02, N02 and CO and daily max 8-h avg 03 for the U.S. stratified by season (2005-2007). 3-101
Figure 3-53. Distribution of correlations between 24-h avg PM10 and co-located 24-h avg PM25, PM10.2.5,
S02, N02 and CO and daily max 8-h avg 03 for the U.S. stratified by season (2005-2007). 3-102
Figure 3-54. Schematic of organic composition of particulate emissions from gasoline-fueled vehicles. 3-105
Figure 3-55. Source category contributions to PM2 5 at a number of sites in the East derived using PMF. 3-108
Figure 3-56. Pearson correlation coefficients for source category contributions to PM2 5 between the 10
Regional Air Pollution Study/Regional Air Monitoring System (RAPS/RAMS) monitoring
sites in St. Louis. 3-109
Figure 3-57. Pearson correlation coefficients for source contributions to PM10.2.5 between the 10
Regional Air Pollution Study/Regional Air Monitoring System (RAPS/RAMS) monitoring
sites in St. Louis. 3-109
Figure 3-58. Eight km southeast U.S. CMAQ-UCD domain zoomed over Tampa Bay, FL. 3-115
Figure 3-59. Two km southeast U.S. CMAQ-UCD domain zoomed over Tampa Bay, FL. 3-115
Figure 3-60. Hourly average CMAQ-UCD predictions and measured observations of NO (top), N02
(middle), and total NOX (bottom) concentrations for May 1 -31, 2002. 3-116
Figure 3-61. CMAQ-UCD predictions and measured observations of ethene concentrations at Sydney,
FL for May 1-31, 2002. 3-117
Figure 3-62. CMAQ-UCD predictions and measured observations of isoprene concentrations at Sydney,
FL for May 1-31, 2002. 3-118
Figure 3-63. CMAQ-UCD predictions and measured observations of PM25 concentrations at Sidney, FL
for May 1 -31, 2002. 3-119
December 2009 xxiv
-------
Figure 3-64.
Figure 3-65.
Figure 3-66.
Figure 3-67.
Figure 3-68.
Figure 3-69.
Figure 3-70.
Figure 3-71.
Figure 3-72.
Figure 3-73.
Figure 3-74.
Figure 3-75.
Figure 3-76.
Figure 3-77.
Figure 3-78.
Figure 3-79.
Figure 3-80.
Figure 3-81.
Figure 3-82.
Figure 3-83.
Figure 3-84.
Figure 3-85.
Figure 3-86
Figure 3-87.
CMAQ-UCD predictions of HN03 concentrations and corresponding measured
observations at Sydney, FL, for May 1 -31 , 2002.
CMAQ-UCD predictions of NH3 concentrations and corresponding measured observations
at Sydney, FL, for May 1-31, 2002.
CMAQ-UCD predictions of pN03" concentrations and corresponding measured
observations at Sydney, FL, for 1-31 May, 2002.
CMAQ-UCD predictions of the ratio of HN03 to total N03 and corresponding measured
observations at Sydney, FL, for May 1 -31 , 2002.
CMAQ-UCD predicted size and chemical-form fractions of total N03" for days in May 2002
with measured observations.
Scatter plot of total nitrate (HN03 plus pN03") wet deposition (mg N/m2/yr) of the model
mean versus measurements for the North American Deposition Proqram (NADP) network.
Scatter plot of total S042" wet deposition (mg S/m2/yr) of the model mean versus
measurements for the National Atmospheric Deposition Proqram (NADP) network.
CMAQ modeling domains for the OAQPS risk and exposure assessments: 36 km outer
parent domain in black; 12 km western U.S. (WUS) domain in red; 12 km eastern U.S.
(EUS) domain in blue.
12-km EUS Summer sulfate PM.
12-km EUS Winter nitrate PM.
12-km EUS Winter total nitrate (HNCK + total pNOO.
12-km EUS annual sulfate wet deposition.
12-km EUS annual nitrate wet deposition.
CMAQ vs. measured air concentrations from east-coast sites in the IMPROVE, CSN
(labeled STN), and CASTNet sites in the summer of 2002 for sulfate (left) and ammonium
(riant).
Comparison of CMAQ-predicted and NADP-measured NH/ wet deposition
CMAQ-predicted (red symbols and lines) and 12-h measured (blue symbols and lines) NH3
and S042" surface concentrations at high and low concentration grid cells in North Carolina
in July 2004.
Surface grid cell (layer 1) analysis of the sensitivity of NHX deposition and transport to the
chanqe in NH3 Vri in CMAQ.
Total column analysis for NH3 (left) and NHX (right) showing modeled NH3 emissions,
transformation, and transport throuqhout the mixed layer and up to the free troposphere.
Range of influence (where 50% of emitted NH3 deposits) from the high concentration
Sampson County qrid cell in the June 2002 CMAQ simulation of Vri sensitivities.
Areal extent of the change in NHX range of influence as predicted by CMAQ for the
Sampson County high concentration grid cell (center of range circles) in June 2002 using
the base case and sensitivity case Vri.
IMPROVE monitorinq site locations.
12-km EUS Summer S042"PM.
12-km EUS Winter NOV PM.
12-km EUS Winter total nitrate (HN03 + total particulate N03").
3-121
3-122
3-123
3-124
3-125
3-127
3-127
3-128
3-129
3-130
3-131
3-132
3-133
3-134
3-134
3-135
3-136
3-136
3-137
3-138
3-141
3-144
3-145
3-145
December 2009 xxv
-------
Figure 3-88.
Figure 3-89.
Figure 3-90.
Figure 3-91 .
Figure 3-92.
Figure 3-93.
Figure 3-94.
Figure 3-95.
Figure 3-96.
Figure 3-97.
Figure 3-98.
Figure 3-99.
Figure 3-1 00.
Figure 3-1 01.
Figure 4-1 .
Figure 4-2.
Figure 4-3.
Monthly average of PM25 concentrations measured at IMPROVE sites in the East and
Midwest for 2004.
Monthly average of PM2.5 concentrations measured at IMPROVE sites in the West for
2004.
Distribution of PM2 5 concentrations measured at IMPROVE sites in the East and Midwest
for 2004.
Distribution of PM^ concentrations measured at IMPROVE sites in the West for 2004.
Model of total personal exposure to PM as a function of ambient and nonambient sources.
Distribution of time sample population spends in various environments, from the National
Human Activity Pattern Survey.
Total exposure to S042" as a function of measured ambient S042" concentration, from the
Vancouver study.
Estimated ambient exposure to PM2 5 as a function of measured ambient PM2 5
concentration, from the Vancouver study
Total exposure to PM2 5 as a function of measured ambient PM2 5 concentration, from the
Vancouver study.
Finf as a function of particle size.
Apportionment of aliphatic carbon, carbonyl, and S042" components of outdoor, indoor, and
personal PM?^ samples, for Los Anqeles (top), Elizabeth (center), and Houston (bottom).
Apportionment of infiltrated PM from mechanical generation (top), primary combustion
(center), and secondary combustion (bottom).
Results of the positive matrix factorization model showing differences in the mass of
outdoor PM and PM that has infiltrated indoors based on source category.
Grid resolution of the CMAQ model in Philadelphia compared with distribution of census
tracts in which exposure assessment is performed.
Diaqrammatic representation of respiratory tract regions in humans.
Structure of lower airways with progression from the large airways to the alveolus.
Comparison of total and regional deposition results from the ICRP and MPPD models for a
resting breathing pattern (VT = 625 ml, f = 12 min"1) and corrected for particle inhalability.
3-146
3-147
3-148
3-149
3-154
3-159
3-166
3-166
3-167
3-170
3-172
3-173
3-174
3-178
4-3
4-4
4-7
Figure 4-4. Comparison of total and regional deposition results from the ICRP and MPPD models for a
light exercise breathing pattern (VT = 1250 ml, f = 20 min"1) and corrected for particle
inhalability. 4-8
Figure 4-5. Total lung deposition measured in healthy adults (UF, 11 M, 11 F, 31 ± 4 yr; fine and
coarse, 11 M, 11 F, 25 ± 4 yr) during controlled breathing on a mouthpiece. 4-9
Figure 4-6. Total deposition of hygroscopic sodium chloride and hydrophobic aluminosilicate aerosols
during oral breathing (VT= 1.0 L; f = 15 min"1). 4-16
Figure 4-7. Retention of poorly soluble particles (0.5-5 urn) in the alveolar region of the lung over time
in various mammalian species. 4-20
Figure 5-1. PM oxidative potential. 5-2
Figure 5-2 PM stimulates pulmonary cells to produce ROS/RNS. 5-3
Figure 5-3. PM activates cell signaling pathways leading to pulmonary inflammation. 5-5
Figure 5-4. Potential pathways for the effects of PM on the respiratory system. 5-7
December 2009 xxvi
-------
Figure 5-5. Potential pathways for the effects of PM on the cardiovascular system. 5-13
Figure 6-1. Excess risk estimates per 10 ug/m3 increase in 24-h avg PM2.5, PM10.2.5, and PM10
concentration for CVD ED visits and HAs. 6-62
Figure 6-2. Excess risk estimates per 10 ug/m3 increase in 24-h avg (unless otherwise noted) PM2.5,
PM10.2.5, and PM10 concentration for Ml and IHD ED visits and HAs. 6-66
Excess risk estimates per 10 ug/m3 increase in 24-h avg PM2.5, PM10.2.5, and PM10
concentration for CHF ED visits and HAs. 6-69
Figure 6-4. Excess risk estimates per 10 ug/m3 increase in 24-h avg PM2 5 and PM10 concentration for
CBVD ED visits and HAs. 6-72
Figure 6-5. Excess risk estimates per 10 ug/m3 increase in 24-h avg PM2.5, and PM10.2.5 for
cardiovascular disease ED visits or HAs, adjusted for co-pollutants. 6-74
Figure 6-6. Combined random-effect estimate of the concentration-response relationship between Ml
emergency hospital admissions and PM10, computed by fitting a piecewise linear spline,
with slope changes at 20 ug/m3 and 50 ug/m3. 6-76
Figure 6-7. Respiratory symptoms and/or medication use among asthmatic children following acute
exposure to PM. 6-86
Figure 6-8. Respiratory symptoms and/or medication use among asthmatic adults following acute
exposure to particles. 6-92
Figure 6-9. Respiratory symptoms following acute exposure to particles and additional criteria
pollutants. 6-93
Figure 6-10. Excess risk estimates per 10 ug/m3 24-h avg PM2.5, PM10.2.5, and PM10 concentration for
ED visits and HAs for respiratory diseases in children. 6-134
Figure 6-11. Excess risks estimates per 10 ug/m3 increase in 24-h avg PM2.5, PM10.2.5, and PM10forED
visits and HAs for respiratory diseases among adults. 6-138
Figure 6-12. Excess risk estimates per 10 ug/m3 increase in 24-h avg PM2.5, PM10.2.5, and PM10 for
asthma ED visits and HAs. 6-140
Figure 6-13. Excess risks estimates per 10 ug/m3 increase in 24-h avg PM2.5, PM10.2.5, and PM10 for
COPD ED visits and HAs among older adults (65+ yr, unless other age group is noted). 6-143
Figure 6-14. Excess risks estimates per 10 ug/m3 increase in 24-h avg PM2.5, PM10.2.5, and PM10 for
respiratory infection ED visits* and HAs. 6-145
Figure 6-15. Excess risk estimates per 10 ug/m3 increase in 24-h avg PM2.5, and PM10.2.5 for respiratory
disease ED visits or HAs, adjusted for co-pollutants. 6-148
Figure 6-16. National and regional estimates of smooth seasonal effects for PM10 at a 1 -day lag and
their sensitivity to the degrees of freedom assigned to the smooth function of time in the
updated NMMAPS data 1987-2000. 6-162
Figure 6-17. Percent increase in the daily number of deaths, for all ages, associated with a 10-ug/m3
increase in PM10: lag 1 (A) and lags 0 and 1 (B) for all three centers. 6-166
Figure 6-18. Effect modification by city characteristics in 20 U.S. cities. 6-168
Figure 6-19. Percent excess risk in mortality (all-cause [nonaccidental] and cause-specific) per
10 ug/m3increase in PM10by individual-level characteristics. 6-170
Figure 6-20. Percent excess risk in mortality (all-cause [nonaccidental] and cause-specific) per
10 ug/m3increase in PM10 by location of death and by season. 6-171
December 2009 xxvii
-------
Figure 6-21. Percent increase in mortality (all-cause [nonaccidental] and cause-specific) per 10 ug/m3
increase in PM10 by contributing causes of death. 6-172
Figure 6-22. Summary of percent increase in all-cause (nonaccidental) mortality from recent multicity
studies per 10 ug/m3 increase in PM10. 6-174
Figure 6-23. Percent increase in all-cause (nonaccidental) and cause-specific mortality per 10 ug/m3
increase in the average of 0- and 1-day lagged PM25, combined by climatic regions. 6-177
Figure 6-24. Empirical Bayes-adjusted city-specific percent increase in total (nonaccidental),
cardiovascular, and respiratory mortality per 10 ug/m3 increase in the average of 0- and 1-
day lagged PM2.5 by decreasing mean 24-h avg PM2.5 concentrations. 6-178
Figure 6-25. Summary of percent increase in all-cause (nonaccidental) mortality per 10 ug/m3 increase
in PM2.5 by various effect modifiers. 6-181
Figure 6-26. Summary of percent increase in all-cause (nonaccidental) and cause-specific mortality per
10 ug/m3 increase in PM25from recent multicity studies. 6-183
Figure 6-28. Percent increase in all-cause (nonaccidental) and cause-specific mortality per 10 ug/m3
increase in the average of 0- and 1-day lagged PM10.2.5, combined by climatic regions. 6-186
Figure 6-29. Empirical Bayes-adjusted city-specific percent increase in total (nonaccidental),
cardiovascular, and respiratory mortality per 10 ug/m3 increase in the average of 0- and 1-
day lagged PM10.2.5 by decreasing 98th percentile of mean 24-h avg PM10.2.5 concentrations. 6-187
Figure 6-30. Summary of percent increase in total (nonaccidental) and cause-specific mortality per
10 ug/m3 increase in PM10.2.5 for all U.S.-, Canadian-, and international-based studies. 6-190
Figure 6-31. Percent increase in PM10 risk estimates (point estimates and 95% CIs) associated with a
5th-95th percentile increase in PM25 and PM25 chemical components. 6-192
Figure 6-32. Sensitivity of the percent increase in PM10 risk estimates (point estimates and 95% CIs)
associated with an interquartile increase in Ni. 6-193
Figure 6-33. Percent excess risk (Cl) of total (nonaccidental) mortality per IQR of concentrations. 6-196
Figure 6-34. Relative risk and Cl of cardiovascular mortality associated with estimated PM25 source
contributions. 6-197
Figure 6-35. Concentration-response curves (spline model) for all-cause, cardiovascular, respiratory and
other cause mortality from the 20 NMMAPS cities. 6-198
Figure 6-36. Percent increase in the risk of death on days with PM10 concentrations in the ranges of
15-24, 25-34, 35-44, and 45 ug/m3 and greater, compared to a reference of days when
concentrations were below 15 ug/m3. 6-199
Figure 6-37. Combined concentration-response curves (spline model) for all-cause, cardiovascular, and
respiratory mortality from the 22 APHEA cities. 6-200
Figure 7-1. Risk estimates for the associations of clinical outcomes with long-term exposure to ambient
PM2.5and PM10. 7-15
Figure 7-2. Adjusted ORs and 95% CIs of symptoms and respiratory diseases associated with a
decline of 10 ug/m3 PM10 levels in Swiss Surveillance Program of Childhood Allergy and
Respiratory Symptoms. 7-23
Figure 7-3. Effect of PM2 5 on the association of lung function with asthma. 7-24
Figure 7-4. Proportion of 18-yr olds with an FE\A below 80% of the predicted value plotted against the
average levels of pollutants from 1994 through 2000 in the 12 southern California
communities of the Children's Health Study. 7-28
December 2009 xxviii
-------
Figure 7-5. Percent increase in postneonatal mortality per 10 ug/m3 in PM10, comparing risk for total
and respiratory mortality. 7-58
Figure 7-6. Mortality risk estimates associated with long-term exposure to PM25 from the Harvard Six
Cities Study (SCS) and the American Cancer Society Study (ACS). 7-85
Figure 7-7. Mortality risk estimates, long-term exposure to PM25 in recent cohort studies. 7-86
Figure 7-8. Plots of the relative risk of death from cardiovascular disease from the Women's Health
Initiative study displaying the between-city and within-city contributions to the overall
association between PM25 and cardiovascular mortality windows of exposure-effects. 7-92
Figure 7-9. The model-averaged estimated effect of a 10-ug/m3 increase in PM2 5 on all-cause mortality
at different lags (in years) between exposure and death. 7-93
Figure 7-10. Time course of relative risk of death after a sudden decrease in air pollution exposure
during the year 2000, assuming a steady state model (solid line) and a dynamic model
(bold dashed line). 7-94
Figure 7-11. Experts' mean effect estimates and uncertainty distributions for the PM2 5 mortality
concentration-response coefficient for a 1 ug/m3 change in annual average PM25. 7-96
Figure 9-1. Important factors involved in seeing a scenic vista are outlined. 9-3
Figure 9-2. Schematic of remote-area (top) and urban (bottom) nighttime sky visibility showing the
effects of PM and light pollution. 9-4
Figure 9-3. Effect of relative humidity on light scattering by mixtures of ammonium nitrate and
ammonium sulfate. 9-6
Figure 9-4. Estimated fractions of total particulate nitrate during each field campaign comprised of
ammonium nitrate, reacted sea salt nitrate (shown as NaN03), and reacted soil dust nitrate
(shown as Ca(N03)2). 9-12
Figure 9-5. A scatter plot of the original IMPROVE algorithm estimated particle light scattering versus
measured particle light scattering. 9-15
Figure 9-6. Scatter plot of the revised algorithm estimates of light scattering versus measured light
scattering. 9-15
Figure 9-7. IMPROVE network PM species estimated light extinction for 2000 (left) and for 2004 (right). 9-18
Figure 9-8. Mean estimated light extinction from PM speciation measurements for the first (top left),
second (top right), third (bottom left), and fourth (bottom right) calendar quarters of 2004. 9-18
Figure 9-9. Percent contributions of ammonium nitrate (left column) and ammonium sulfate (right
column) to particulate light extinction for each calendar quarter of 2004 (first through fourth
quarter arranged from top to bottom). 9-20
Figure 9-10. Percent contributions of organic mass (left column) and EC (right column) to particulate
light extinction for each calendar quarter of 2004 (first through fourth quarter arranged from
top to bottom). 9-21
Figure 9-11. Percent contributions of coarse mass (left column) and fine soil (right column) to particulate
light extinction for each calendar quarter of 2004 (first through fourth quarter arranged from
top to bottom). 9-22
Figure 9-12. IMPROVE Mean PM25 mass concentration determined by summing the major components
for the 2000-2004. 9-24
Figure 9-13. IMPROVE and CSN (STN) mean PM25 mass concentration determined by summing the
major components for 2000-2004. 9-24
Figure 9-14. IMPROVE mean ammonium nitrate concentrations for 2000-2004. 9-25
December 2009 xxix
-------
Figure 9-1 5.
Figure 9-1 6.
Figure 9-1 7.
Figure 9-1 8.
Figure 9-1 9.
Figure 9-20.
Figure 9-21.
Figure 9-22.
Figure 9-23.
Figure 9-24.
Figure 9-25.
Figure 9-26.
Figure 9-27.
Figure 9-28.
Figure 9-29.
Figure 9-30.
Figure 9-31.
Figure 9-32.
Figure 9-33.
IMPROVE and CSN (STN) mean ammonium nitrate concentrations for 2000-2004.
IMPROVE mean ammonium sulfate concentrations for 2000-2004.
IMPROVE and CSN (STN) mean ammonium sulfate concentrations for 2000-2004.
IMPROVE monitored mean orqanic mass concentrations for 2000-2004.
IMPROVE and CSN (STN) mean orqanic mass concentrations for 2000-2004.
IMPROVE mean EC concentrations for 2000-2004.
IMPROVE and CSN (STN) mean EC concentrations for 2000-2004.
IMPROVE mean fine soil concentrations for 2000-2004.
IMPROVE and CSN (STN) fine soil concentrations, 2000-2004.
Regional and local contributions to annual average PM2 5 by particulate S042", nitrate and
total carbon (i.e., organic plus EC) for select urban areas based on paired IMPROVE and
CSN monitorinq sites.
IMPROVE mean coarse mass concentrations for 2000-2004.
Ten-year (1995-2004) haze trends for the mean of the 20% best annual haze conditions.
Ten-year (1995-2004) haze trends for the mean of the 20% worst annual haze conditions.
Ten-year trends in the 80th percentile particulate S042" concentration based on IMPROVE
and CASTNet monitoring and net S02 emissions from the National Emissions Trends
(NET) data base by reqion of the U.S.
Map of 1 0-yr trends (1 994-2003) in haze by particulate nitrate contribution to haze for the
worst 20% annual haze periods.
Contributions of the Pacific Coast area to the ammonium sulfate (ug/m3) at 84 remote-area
monitoring sites in western U.S. based on trajectory regression for all sample periods from
2000-2002
Shows the IMPROVE monitoring sites in the WRAP region with at least three years of valid
data and identifies the six sites selected to demonstrate the apportionment tools.
Particulate S042" (a) and nitrate (b) source attribution by region using CAMx modeling for
six western remote area monitorinq sites
Monthly averaged model predicted organic mass concentration apportioned into primary
PM and anthropogenic and biogenic secondary PM categories for the Olympic NP (top)
and San Gorqonio W (bottom) monitorinq sites.
9-25
9-26
9-26
9-28
9-28
9-29
9-29
9-30
9-30
9-32
9-33
9-34
9-35
9-36
9-37
9-39
9-41
9-43
9-44
Figure 9-34. Monthly averaged model predicted organic mass concentration apportioned into primary
PM and anthropogenic and biogenic secondary PM categories for the Yellowstone NP (top)
and Grand Canyon (Hopi Point) (bottom) monitoring sites. 9-45
Figure 9-35. Monthly averaged model predicted organic mass concentration apportioned into primary
PM and anthropogenic and biogenic secondary PM categories for the Badland NP (top)
and Salt Creek W (bottom) monitoring sites. 9-46
Figure 9-36. Comparison of carbon concentrations between Seattle (Puget Sound site) and Mt. Rainer
(left) and between Phoenix and Tonto (right) showing the background site concentration
(gray) and the urban excess concentration (black) for total, fossil and contemporary carbon
during the summer and winter studies. 9-47
Figure 9-37. Average contemporary fraction of PM2.s carbon for the summer (top) and winter (bottom),
estimated from IMPROVE monitoring data (June 2004-February 2006) based on EC/TC
ratios. 9-48
December 2009 xxx
-------
Figure 9-38. Results of the weighted emissions potential tool applied to primary OC emissions (top) and
EC emissions (bottom) for the baseline and projected 2018 emissions inventories for
Olympic NP. 9-50
Figure 9-39. Results of the weighted emissions potential tool applied to primary OC emissions (top) and
EC emissions (bottom) for the baseline and projected 2018 emissions inventories for San
Gorgonio W. 9-51
Figure 9-40. Results of the weighted emissions potential tool applied to primary OC emissions (top) and
EC emissions (bottom) for the baseline and projected 2018 emissions inventories for
Yellowstone NP. 9-53
Figure 9-41. Results of the weighted emissions potential tool applied to primary OC emissions (top) and
EC emissions (bottom) for the baseline and projected 2018 emissions inventories for Grand
Canyon NP. 9-54
Figure 9-42. Results of the weighted emissions potential tool applied to primary OC emissions (top) and
EC emissions (bottom) for the baseline and projected 2018 emissions inventories for
Badlands NP. 9-55
Figure 9-43. Results of the weighted emissions potential tool applied to primary OC emissions (top) and
EC emissions (bottom) for the baseline and projected 2018 emissions inventories for Salt
Creek W. 9-56
Figure 9-44. BRAVO study haze contributions for Big Bend NP, TX during a 4-mo period in 1999. 9-58
Figure 9-45. Maps of spatial patterns for average annual particulate nitrate measurements (top), and for
ammonia emissions for April 2002 from the WRAP emissions inventory (bottom). 9-59
Figure 9-46. Maps of spatial patterns of annual NO (left) and N02 (right) emissions for 2002 from the
WRAP emissions inventory. 9-60
Figure 9-47. Midwest ammonia monitoring network. 9-61
Figure 9-48. Upwind transport probability fields associated with high particulate nitrate concentrations
measured at Toronto, Canada; Boundary Water Canoe Area, MN; Shenandoah NP, VA;
Lye Brook, VT; and Great Smoky Mountains NP, TN. 9-62
Figure 9-49. Trajectory probability fields for periods with high particulate S042" measured at Underbill,
VT and Brigantine, NJ (shown as white stars) associated with oil-burning trace components
(left) and with coal-burning trace components (right). 9-63
Figure 9-50. Scatter plots of particulate S042" (left) and particulate S042" and organic mass (right)
versus nephelometer measured particle light scattering for Acadia NP, ME. 9-64
Figure 9-51. CMAQ air quality modeling projections of visibility responses on the 20% worst haze days
at Great Smoky Mountains NP, NC (top) and Swanquarter W, NC (bottom) to 30%
reductions. 9-65
Figure 9-52. Aerosol radiative forcing. 9-76
Figure 9-53. Global average radiative forcing (RF) estimates and uncertainty ranges in 2005, relative to
the pre-industrial climate. 9-78
Figure 9-54. Probability distribution functions (PDFs) for anthropogenic aerosol and GHG RFs. 9-79
Figure 9-55. The clear-sky forcing efficiency ET, defined as the diurnally averaged aerosol direct
radiative effect (W/m2) per unit AOD at 550 nm, calculated at both TOA and the surface, for
typical aerosol types over different geographical regions. 9-80
Figure 9-56. A composite of MODIS/Terra observed aerosol optical depth (at 550 nm, green light near
the peak of human vision) and fine-mode fraction that shows spatial and seasonal
variations of aerosol types. 9-87
December 2009 xxxi
-------
Figure 9-57. Oregon fire on September 4, 2003, as observed by MISR: (a) MISR nadir view of the fire
plume, with five patch locations numbered and wind-vectors superposed in yellow; (b)
MISR aerosol optical depth at 558 nm; and (c) MISR stereo height without wind correction
for the same region. 9-88
Figure 9-58. Global maps at 18 km resolution showing monthly average (a) AOD at 865 nm and (b)
Angstrom exponent of AOD over water surfaces only for June, 1997, derived from radiance
measurements by the POLDER. 9-89
Figure 9-59. A dust event that originated in the Sahara desert on 17 August 2007 and was transported
to the Gulf of Mexico. 9-91
Figure 9-60. A constellation of five spacecraft that overfly the Equatorat about 1:30 p.m., the so-called
A-Train, carries sensors having complementary capabilities, offering unprecedented
opportunities to study aerosols from space in multiple dimensions. 9-92
Figure 9-61. Geographical coverage of active AERONET sites in 2006. 9-95
Figure 9-62. Comparison of the mean concentration (ug/m3) and standard deviation of the modeled
(STEM) aerosol chemical components with shipboard measurements during INDOEX,
ACE-Asia, and ICARTT. 9-98
Figure 9-63. Correlations between one-hour PMu surface measurements in the U.S. and southern
Canada reported to AIRNOW and MODIS satellite AOD values for the period between 4
July and 1 September 2009. 9-99
Figure 9-64. Location of aerosol chemical composition measurements with aerosol mass spectrometers. 9-101
Figure 9-65. Scatterplots of the submicrometer POM measured during NEAQS versus A) acetylene and
B) iso-propyl nitrate. 9-102
Figure 9-66. Comparison of annual mean aerosol optical depth (AOD). 9-104
Figure 9-67. Percentage contributions of individual aerosol components. 9-105
Figure 9-68. Geographical patterns of seasonally (MAM) averaged aerosol optical depth at 550 nm (left
panel) and the diurnally averaged clear-sky aerosol direct radiative (solar spectrum) forcing
(W/m2) at the TOA (right panel) derived from satellite (Terra) retrievals. 9-107
Figure 9-69. Summary of observation- and model-based (denoted as OBS and MOD, respectively)
estimates of clear-sky, annual average DRF at the TOA and at the surface. 9-108
Figure 9-70. Scatter plots showing mean cloud drop effective radius (re) versus aerosol extinction
coefficient (unit: km-1) for various liquid water path (LWP) bands on April 3,1998 at ARM
SGPsite. 9-115
Figure 9-71. Sampling the Arctic Haze. Pollution and smoke aerosols can travel long distances, from
mid-latitudes to the Arctic, causing "Arctic Haze." 9-123
Figure 9-72. Global annual averaged AOD (upper panel) and aerosol mass loading (lower panel) with
their components simulated by 15 models in AeroCom- A (excluding one model which only
reported mass). 9-130
Figure 9-73. Aerosol direct radiative forcing in various climate and aerosol models. 9-134
Figure 9-74. Aerosol optical thickness and anthropogenic shortwave all-sky radiative forcing from the
AeroCom study. 9-135
Figure 9-75. Radiative forcing from the cloud albedo effect (1st aerosol indirect effect) in the global
climate models used from IPCC (2007, 092765), Chapter 2, Figure 2.14, of the IPCC AR4. 9-136
December 2009 xxxii
-------
Figure 9-76. Anthropogenic impact on cloud cover, planetary albedo, radiative flux at the surface (while
holding sea surface temperatures and sea ice fixed) and surface air temperature change
from the direct aerosol forcing (top row), the first indirect effect (second row) and the
second indirect effect (third row). 9-139
Figure 9-77. Global average present-day short wave cloud forcing at TOA (top) and change in whole sky
net outgoing shortwave radiation (bottom) between the present-day and pre-industrial
simulations for each model in each experiment. 9-140
Figure 9-78.
Figure 9-79.
Figure 9-80.
Figure 9-81.
Figure 9-82.
Figure 9-83.
Figure 9-84.
Figure 9-85.
Figure A-1 .
Figure A-2.
Figure A-3.
Figure A-4.
Figure A-5.
Figure A-6.
Figure A-7.
Figure A-8.
Figure A-9.
Figure A-1 0.
Figure A-1 1.
Figure A-1 2.
Figure A-1 3.
Figure A-1 4.
Figure A-1 5.
Figure A-1 6.
Figure A-1 7.
Figure A-1 8.
Direct radiative forcing by anthropogenic aerosols in the GISS model (including sulfates,
BC, OC and nitrates).
Percentage of aerosol optical depth in the GISS, left, based on Liu et al. (2006, 190422),
provided by A. Lads, GISS, and GFDL, riqht, from Ginoux et al. (2006, 190582).
Most probable aerosol altitude (in pressure, hPa) from the GISS model in January (top) and
July (bottom).
Time dependence of aerosol optical thickness (left) and climate forcing (right). Note that as
specified, the aerosol trends are all "flat" from 1990-2000.
Change in global mean ocean temperature (left axis) and ocean heat content (right axis) for
the top 3000 m due to different forcinqs in the GFDL model.
Visible (wavelength 0.55 urn) optical depth estimates of stratospheric S042" aerosols
formed in the aftermath of explosive volcanic eruptions that occurred between 1860 and
2000.
The transfer of POPs between the maior abiotic compartments of the Arctic.
Relationship of plant nutrients and trace metals with veqetation.
PM?^ monitor distribution in comparison with population density, Atlanta, GA.
PMm monitor distribution in comparison with population density, Atlanta, GA.
PM?^ monitor distribution in comparison with population density, Birminqham, AL.
PMm monitor distribution in comparison with population density, Birminqham, AL.
PM^ monitor distribution in comparison with population density, Boston, MA.
PMm monitor distribution in comparison with population density, Boston, MA.
PM?<; monitor distribution in comparison with population density, Chicaqo, IL.
PMm monitor distribution in comparison with population density, Chicaqo, IL.
PM?<; monitor distribution in comparison with population density, Denver, CO.
PMm monitor distribution in comparison with population density, Denver, CO.
PM?<; monitor distribution in comparison with population density, Detroit, Ml.
PMm monitor distribution in comparison with population density, Detroit, Ml.
PM?<; monitor distribution in comparison with population density, Houston, TX.
PMm monitor distribution in comparison with population density, Houston, TX.
PM?<; monitor distribution in comparison with population density, Los Anqeles, CA.
PMm monitor distribution in comparison with population density, Los Anqeles, CA.
PM? i monitor distribution in comparison with population density, New York, NY.
PMm monitor distribution in comparison with population density, New York, NY.
9-148
9-151
9-153
9-155
9-156
9-163
9-169
9-185
A-59
A-60
A-61
A-62
A-63
A-64
A-65
A-66
A-67
A-68
A-69
A-70
A-71
A-72
A-73
A-74
A-75
A-76
December 2009 xxxiii
-------
Figure A-1 9.
Figure A-20.
Figure A-21.
Figure A-22.
Figure A-23.
Figure A-24.
Figure A-25.
Figure A-26.
Figure A-27.
Figure A-28.
Figure A-29.
Figure A-30.
Figure A-31.
Figure A-32.
Figure A-33.
Figure A-34.
Figure A-35.
Figure A-36.
Figure A-37.
Figure A-38.
Figure A-39.
Figure A-40.
Figure A-41 .
Figure A-42.
Figure A-43.
Figure A-44.
Figure A-45.
Figure A-46.
Figure A-47.
PM^ monitor distribution in comparison with population density, Philadelphia, PA.
PMm monitor distribution in comparison with population density, Philadelphia, PA.
PM?^ monitor distribution in comparison with population density, Phoenix, AZ.
PMm monitor distribution in comparison with population density, Phoenix, AZ.
PM?^ monitor distribution in comparison with population density, Pittsburgh, PA.
PMm monitor distribution in comparison with population density, Pittsburgh, PA.
PM?^ monitor distribution in comparison with population density, Riverside, CA.
PMm monitor distribution in comparison with population density, Riverside, CA.
PM?<; monitor distribution in comparison with population density, Seattle, WA.
PMm monitor distribution in comparison with population density, Seattle, WA.
PM?<; monitor distribution in comparison with population density, St. Louis, MO.
PMm monitor distribution in comparison with population density, St. Louis, MO.
Three-yravg of 24-h PM25 Cu concentrations measured at CSN sites across the U.S.,
2005-2007.
Three-yr avg of 24-h PM25 Fe concentrations measured at CSN sites across the U.S.,
2005-2007
Three-yr avg of 24-h PM25 Ni concentrations measured at CSN sites across the U.S.,
2005-2007
Three-yr avg of 24-h PM25 Pb concentrations measured at CSN sites across the U.S.,
2005-2007
Three-yravg of 24-h PM25 Se concentrations measured at CSN sites across the U.S.,
2005-2007
Three-yr avg of 24-h PM25 V concentrations measured at CSN sites across the U.S., 2005-
2007
PM?<; monitor distribution and maior highways, Atlanta, GA.
Box plots illustrating the seasonal distribution of 24-h avg PM25 concentrations for Atlanta,
GA.
PM?<; inter-sampler correlations as a function of distance between monitors for Atlanta, GA.
PM?<; monitor distribution and maior highways, Birmingham, AL.
Box plots illustrating the seasonal distribution of 24-h avg PM25 concentrations for
Birmingham, AL.
PM25 inter-sampler correlations as a function of distance between monitors for
Birmingham, AL.
PM?<; monitor distribution and maior highways, Boston, MA.
Box plots illustrating the seasonal distribution of 24-h avg PM25 concentrations for Boston,
MA.
PM?<; inter-sampler correlations as a function of distance between monitors for Boston, MA.
PM?<; monitor distribution and maior highways, Chicago, IL.
Box plots illustrating the seasonal distribution of 24-h avg PM25 concentrations for Chicago,
IL.
A-77
A-78
A-79
A-80
A-81
A-82
A-83
A-84
A-85
A-86
A-87
A-88
A-89
A-90
A-91
A-92
A-93
A-94
A-96
A-97
A-99
A-1 00
A-1 01
A-1 03
A-1 04
A-1 05
A-1 07
A-1 08
A-1 10
December 2009 xxxiv
-------
Figure A-48. PM2 5 inter-sampler correlations as a function of distance between monitors for Chicago, IL. A-113
Figure A-49. PM2.5 monitor distribution and major highways, Denver, CO. A-114
Figure A-50. Box plots illustrating the seasonal distribution of 24-h avg PM2 5 concentrations for Denver,
CO. A-115
Figure A-51. PM2.5 inter-sampler correlations as a function of distance between monitors for Denver, CO. A-117
Figure A-52. PM25 monitor distribution and major highways, Detroit, Ml. A-118
Figure A-53. Box plots illustrating the seasonal distribution of 24-h avg PM2 5 concentrations for Detroit,
Ml. A-119
Figure A-54. PM25 inter-sampler correlations as a function of distance between monitors for Detroit, Ml. A-121
Figure A-55. PM25 monitor distribution and major highways, Houston, TX. A-122
Figure A-56. Box plots illustrating the seasonal distribution of 24-h avg PM25 concentrations for Houston,
TX. A-123
Figure A-57. PM2 5 inter-sampler correlations as a function of distance between monitors for Houston,
TX. A-124
Figure A-58. PM2 5 monitor distribution and major highways, Los Angeles, CA. A-125
Figure A-59. Box plots illustrating the seasonal distribution of 24-h avg PM2 5 concentrations for Los
Angeles, CA. A-126
Figure A-60. PM2 5 inter-sampler correlations as a function of distance between monitors for Los
Angeles, CA. A-127
Figure A-61. PM2 5 monitor distribution and major highways, New York City, NY. A-128
Figure A-62. Box plots illustrating the seasonal distribution of 24-h avg PM2 5 concentrations for New
York, NY. A-130
Figure A-63 PM2 5 inter-sampler correlations as a function of distance between monitors for New York,
NY. A-133
Figure A-64. PM2 5 monitor distribution and major highways, Philadelphia, PA. A-134
Figure A-65. Box plots illustrating the seasonal distribution of 24-h avg PM2 5 concentrations for
Philadelphia, PA. A-135
Figure A-66. PM2 5 inter-sampler correlations as a function of distance between monitors for
Philadelphia, PA. A-137
Figure A-67. PM25 monitor distribution and major highways, Phoenix, AZ. A-138
Figure A-68. Box plots illustrating the seasonal distribution of 24-h avg PM2 5 concentrations for Phoenix,
AZ. A-139
Figure A-69. PM2 5 inter-sampler correlations as a function of distance between monitors for Phoenix,
AZ. A-141
Figure A-70. PM25 monitor distribution and major highways, Pittsburgh, PA. A-142
Figure A-71. Box plots illustrating the seasonal distribution of 24-h avg PM2 5 concentrations for
Pitsburgh, PA. A-143
Figure A-72. PM2 5 inter-sampler correlations as a function of distance between monitors for Pittsburgh,
PA. A-145
Figure A-73. PM25 monitor distribution and major highways, Riverside, CA. A-146
December 2009 xxxv
-------
Figure A-74. Box plots illustrating the seasonal distribution of 24-h avg PM2 5 concentrations for
Figure A-75.
Figure A-76.
Figure A-77.
Figure A-78.
Figure A-79.
Figure A-80.
Figure A-81
Figure A-82.
Figure A-83.
Figure A-84.
Figure A-85.
Figure A-86.
Figure A-87
Figure A-88.
Figure A-89.
Figure A-90
Figure A-91.
Figure A-92.
Figure A-93.
Figure A-94.
Figure A-95.
Figure A-96.
Figure A-97.
Figure A-98.
Figure A-99.
Figure A-1 00.
Riverside, CA.
PM2 5 inter-sampler correlations as a function of distance between monitors for Riverside
CA.
PM?^ monitor distribution and maior highways, Seattle, WA.
Box plots illustrating the seasonal distribution of 24-h avg PM25 concentrations for Seattle,
WA.
PM?^ inter-sampler correlations as a function of distance between monitors for Seattle, WA.
PM?<; monitor distribution and maior highways, St. Louis, MO.
Box plots illustrating the seasonal distribution of 24-h avg PM25 concentrations for St.
Louis, MO.
PM2 5 inter-sampler correlations as a function of distance between monitors for St. Louis,
MO.
PMm monitor distribution and maior highways, Atlanta, GA.
Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for Atlanta,
GA.
PMm inter-sampler correlations as a function of distance between monitors for Atlanta, GA.
PMm monitor distribution and maior highways, Birmingham, AL.
Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Birmingham, AL.
PM10 inter-sampler correlations as a function of distance between monitors for Birmingham,
AL.
PMm monitor distribution and maior highways, Boston, MA.
Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for Boston,
MA.
PMm inter-sampler correlations as a function of distance between monitors for Boston, MA.
PMm monitor distribution and maior highways, Chicago, IL.
Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for Chicago,
IL.
PMm inter-sampler correlations as a function of distance between monitors for Chicago, IL.
PMm monitor distribution and maior highways, Denver, CO.
Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for Denver,
CO.
PMm inter-sampler correlations as a function of distance between monitors for Denver, CO.
PMm monitor distribution and maior highways, Detroit, Ml.
Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for Detroit,
Ml.
PMm inter-sampler correlations as a function of distance between monitors for Detroit, Ml.
PMm monitor distribution and maior highways, Houston, TX.
A-1 47
A-1 49
A-1 50
A-1 51
A-1 52
A-1 53
A-1 54
A-1 56
A-1 57
A-1 58
A-1 60
A-1 61
A-1 62
A-1 64
A-1 65
A-1 66
A-1 67
A-1 68
A-1 69
A-1 71
A-1 72
A-1 73
A-1 75
A-1 76
A-1 77
A-1 78
A-1 79
December 2009 xxxvi
-------
Figure A-101. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for Houston,
Figure A-1 02.
Figure A-1 03.
Figure A-1 04.
Figure A-1 05.
Figure A-1 06.
Figure A-1 07.
Figure A-1 08.
Figure A-1 09.
Figure A-1 10.
Figure A-1 11.
Figure A-1 12.
Figure A-1 13.
Figure A-1 14.
Figure A-1 15.
Figure A-1 16.
Figure A-1 17.
Figure A-1 18.
Figure A-1 19.
Figure A-1 20.
Figure A-1 21.
Figure A-1 22.
Figure A-1 23.
Figure A-1 24.
Figure A-1 25.
Figure A-1 26.
TX.
PMio inter-sampler correlations as a function of distance between monitors for Houston, TX.
PMm monitor distribution and maior highways, Los Anqeles, CA.
Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for Los
Anqeles, CA.
PMio inter-sampler correlations as a function of distance between monitors for Los Angeles,
CA.
PMm monitor distribution and maior highways, New York, NY.
Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for New
York, NY.
PM10 inter-sampler correlations as a function of distance between monitors for New York,
NY.
PMm monitor distribution and maior highways, Philadelphia, PA.
Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Philadelphia, PA.
PM10 inter-sampler correlations as a function of distance between monitors for Philadelphia,
PA.
PMm monitor distribution and maior highways, Phoenix, AZ.
Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for Phoenix,
AZ.
PMm inter-sampler correlations as a function of distance between monitors for Phoenix, AZ.
PMm monitor distribution and maior highways, Pittsburgh, PA.
Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Pittsburgh, PA.
PM10 inter-sampler correlations as a function of distance between monitors for Pittsburgh,
PA.
PMm monitor distribution and maior highways, Riverside, CA.
Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for
Riverside, CA.
PM10 inter-sampler correlations as a function of distance between monitors for Riverside,
CA.
PMm monitor distribution and maior highways, Seattle, WA.
Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for Seattle,
WA.
PMm inter-sampler correlations as a function of distance between monitors for Seattle, WA.
PMm monitor distribution and maior highways, St. Louis, MO.
Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations for St. Louis,
MO.
PM10 inter-sampler correlations as a function of distance between monitors for St. Louis,
MO.
A-1 80
A-1 82
A-1 83
A-1 84
A-1 85
A-1 86
A-1 87
A-1 88
A-1 89
A-1 90
A-1 91
A-1 92
A-1 94
A-1 97
A-1 98
A-1 99
A-201
A-202
A-203
A-205
A-206
A-207
A-208
A-209
A-210
A-212
December 2009 xxxvii
-------
Figure A-127. Seasonally averaged PM2 5 speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in Atlanta, GA. A-213
Figure A-128. Seasonally averaged PM2 5 speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in Birmingham, AL. A-214
Figure A-129. Seasonally averaged PM2 5 speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in Boston, MA. A-215
Figure A-130. Seasonally averaged PM2 5 speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in Chicago, IL. A-216
Figure A-131. Seasonally averaged PM2 5 speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter, derived using the SANDWICH method in Denver, CO. A-217
Figure A-132. Seasonally averaged PM2 5 speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in Detroit, Ml. A-218
Figure A-133. Seasonally averaged PM2 5 speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in Houston, TX. A-219
Figure A-134. Seasonally averaged PM2 5 speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in Los Angeles, CA. A-220
Figure A-135. Seasonally averaged PM2 5 speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in New York, NY. A-221
Figure A-136. Seasonally averaged PM2 5 speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in Philadelphia, PA. A-222
Figure A-137. Seasonally averaged PM2 5 speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in Phoenix, AZ. A-223
Figure A-138. Seasonally averaged PM2 5 speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in Pittsburgh, PA. A-224
Figure A-139. Seasonally averaged PM2 5 speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in Riverside, CA. A-225
Figure A-140. Seasonally averaged PM2 5 speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in Seattle, WA. A-226
Figure A-141. Seasonally averaged PM2 5 speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in St. Louis, MO. A-227
Figure A-142. Seasonal patterns in PM2 5 chemical composition from city-wide monthly average values for
Atlanta, GA, 2005-2007. A-228
Figure A-143. Seasonal patterns in PM2 5 chemical composition from city-wide monthly average values for
Birmingham, AL, 2005-2007. A-228
Figure A-144. Seasonal patterns in PM2 5 chemical composition from city-wide monthly average values for
Boston, MA, 2005-2007. A-229
Figure A-145. Seasonal patterns in PM2 5 chemical composition from city-wide monthly average values for
Chicago, IL, 2005-2007. A-229
Figure A-146. Seasonal patterns in PM2 5 chemical composition from city-wide monthly average values for
Denver, CO, 2005-2007. A-230
Figure A-147. Seasonal patterns in PM2 5 chemical composition from city-wide monthly average values for
Detroit, Ml, 2005-2007. A-230
December 2009 xxxviii
-------
Figure A-148. Seasonal patterns in PM2 5 chemical composition from city-wide monthly average values for
Houston, TX, 2005-2007. A-231
Figure A-149. Seasonal patterns in PM2 5 chemical composition from city-wide monthly average values for
Los Angeles, CA, 2005-2007. A-231
Figure A-150. Seasonal patterns in PM2 5 chemical composition from city-wide monthly average values for
New York, NY, 2005-2007. A-232
Figure A-151. Seasonal patterns in PM2 5 chemical composition from city-wide monthly average values for
Philadelphia, PA, 2005-2007. A-232
Figure A-152. Seasonal patterns in PM2 5 chemical composition from city-wide monthly average values for
Phoenix, AZ, 2005-2007. A-233
Figure A-153. Seasonal patterns in PM2 5 chemical composition from city-wide monthly average values for
Pittsburgh, PA, 2005-2007. A-233
Figure A-154. Seasonal patterns in PM2 5 chemical composition from city-wide monthly average values for
Riverside, CA, 2005-2007. A-234
Figure A-155. Seasonal patterns in PM2 5 chemical composition from city-wide monthly average values for
Seattle, WA, 2005-2007. A-234
Figure A-156. Seasonal patterns in PM2 5 chemical composition from city-wide monthly average values for
St. Louis, MO, 2005-2007. A-235
Figure A-157. Diel plots generated from all available hourly FRM-like PM25 data, stratified by weekday
(left) and weekend (right), in Atlanta, GA. A-235
Figure A-158. Diel plots generated from all available hourly FRM-like PM25 data, stratified by weekday
(left) and weekend (right), in Chicago, IL. A-236
Figure A-159. Diel plots generated from all available hourly FRM-like PM25 data, stratified by weekday
(left) and weekend (right), in Houston, TX. A-236
Figure A-160. Diel plots generated from all available hourly FRM-like PM25 data, stratified by weekday
(left) and weekend (right), in New York, NY. A-237
Figure A-161. Diel plots generated from all available hourly FRM-like PM25 data, stratified by weekday
(left) and weekend (right), in Pittsburgh, PA. A-237
Figure A-162. Diel plots generated from all available hourly FRM-like PM25 data, stratified by weekday
(left) and weekend (right), in Seattle, WA. A-238
Figure A-163. Diel plots generated from all available hourly FRM-like PM25 data, stratified by weekday
(left) and weekend (right), in St. Louis, MO. A-238
Figure A-164. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by weekday
(left) and weekend (right), in Atlanta, GA. A-239
Figure A-165. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by weekday
(left) and weekend (right), in Chicago, IL. A-239
Figure A-166. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by weekday
(left) and weekend (right), in Denver, CO. A-240
Figure A-167. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by weekday
(left) and weekend (right), in Detroit, Ml. A-240
Figure A-168. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by weekday
(left) and weekend (right), in Los Angeles, CA. A-241
December 2009 xxxix
-------
Figure A-169. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by weekday
(left) and weekend (right), in Philadelphia, PA. A-241
Figure A-170. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by weekday
(left) and weekend (right), in Phoenix, AZ. A-242
Figure A-171. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by weekday
(left) and weekend (right), in Pittsburgh, PA. A-242
Figure A-172. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by weekday
(left) and weekend (right), in Riverside, CA. A-243
Figure A-173. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by weekday
(left) and weekend (right), in Seattle, WA. A-243
Figure A-174. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by weekday
(left) and weekend (right), in St. Louis, MO. A-244
Figure A-175. Correlations between 24-h PM2 5 and co-located 24-h avg PM10, PM10.2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for Atlanta, GA, stratified by season (2005-2007). A-245
Figure A-176. Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PM10.2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for Birmingham, AL, stratified by season (2005-2007). A-246
Figure A-177. Correlations between 24-h PM2 5 and co-located 24-h avg PM10, PM10.2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for Boston, MA, stratified by season (2005-2007). A-247
Figure A-178. Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PM10-2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for Chicago, IL, stratified by season (2005-2007). A-248
Figure A-179. Correlations between 24-h PM2 5 and co-located 24-h avg PM10, PM10.2.5, S02, N02 and CO
and daily maximum 8-h avg 03for Denver, CO, stratified by season (2005-2007). A-249
Figure A-180. Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PM10-2.5, S02, N02 and CO
and daily maximum 8-h avg 03for Detroit, Ml, stratified by season (2005-2007). A-250
Figure A-181. Correlations between 24-h PM2 5 and co-located 24-h avg PM10, PM10.2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for Houston, TX, stratified by season (2005-2007). A-251
Figure A-182. Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PM10-2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for Los Angeles, CA, stratified by season (2005-2007). A-252
Figure A-183. Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PM10-2.5, S02, N02 and CO
and daily maximum 8-h avg 03for New York, NY, stratified by season (2005-2007). A-253
Figure A-184. Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PM10-2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for Philadelphia, PA, stratified by season (2005-2007). A-254
Figure A-185. Correlations between 24-h PM2 5 and co-located 24-h avg PM10, PM10.2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for Phoenix, AZ, stratified by season (2005-2007). A-255
Figure A-186. Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PM10-2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for Pittsburgh, PA, stratified by season (2005-2007). A-256
Figure A-187. Correlations between 24-h PM2 5 and co-located 24-h avg PM10, PM10.2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for Riverside, CA, stratified by season (2005-2007). A-257
Figure A-188. Correlations between 24-h PM2.5 and co-located 24-h avg PM10, PM10-2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for St. Louis, MO, stratified by season (2005-2007). A-258
Figure A-189. Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10-2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for Atlanta, GA, stratified by season (2005-2007). A-259
December 2009
-------
Figure A-190. Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10-2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for Birmingham, AL, stratified by season (2005-2007). A-260
Figure A-191. Correlations between 24-h PM10 and co-located 24-h avg PM2 5, PM10.2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for Boston, MA, stratified by season (2005-2007). A-261
Figure A-192. Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10.2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for Chicago, IL, stratified by season (2005-2007). A-262
Figure A-193. Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10.2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for Denver, CO, stratified by season (2005-2007). A-263
Figure A-194. Correlations between 24-h PM10 and co-located 24-h avg PM2 5, PM10.2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for Detroit, Ml, stratified by season (2005-2007). A-264
Figure A-195. Correlations between 24-h PM10 and co-located 24-h avg PM2 5, PM10.2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for Houston, TX, stratified by season (2005-2007). A-265
Figure A-196. Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10-2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for Los Angeles, CA, stratified by season (2005-2007). A-266
Figure A-197. Correlations between 24-h PM10 and co-located 24-h avg PM2 5, PM10.2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for New York, NY, stratified by season (2005-2007). A-267
Figure A-198. Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10-2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for Philadelphia, PA, stratified by season (2005-2007). A-268
Figure A-199. Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10-2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for Phoenix, AZ, stratified by season (2005-2007). A-269
Figure A-200. Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10-2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for Pittsburgh, PA, stratified by season (2005-2007). A-270
Figure A-201. Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10-2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for Riverside, CA, stratified by season (2005-2007). A-271
Figure A-202. Correlations between 24-h PM10 and co-located 24-h avg PM2.5, PM10-2.5, S02, N02 and CO
and daily maximum 8-h avg 03 for St. Louis, MO, stratified by season (2005-2007). A-272
December 2009 xli
-------
PM ISA Project Team
Executive Direction
Dr. John Vandenberg (Director)—National Center for Environmental Assessment-RTF Division, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Ms. Debra Walsh (Deputy Director)—National Center for Environmental Assessment-RTP Division, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Mary Ross (Branch Chief)—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Scientific Staff
Dr. Lindsay Wichers Stanek (PM Team Leader)—National Center for Environmental Assessment,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jeffrey Arnold— National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC (now at Institute for Water Resources, U.S. Army Corps of
Engineers, Washington, D.C.)
Dr. Christal Bowman—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. James S. Brown—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Barbara Buckley—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Mr. Allen Davis—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Jean-Jacques Dubois—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Steven J. Dutton—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Erin Hines—National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Douglas Johns—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Ellen Kirrane—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Dennis Kotchmar—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Thomas Long—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
December 2009 xlii
-------
Dr. Thomas Luben—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Qingyu Meng—Oak Ridge Institute for Science and Education, Postdoctoral Research Fellow to
National Center for Environmental Assessment, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Dr. Kristopher Novak—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Joseph Pinto—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Jennifer Richmond-Bryant—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Mary Ross—National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Jason Sacks—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. David Svendsgaard—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Lisa Vinikoor—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. William Wilson—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Lori White—National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC (now at National Institute for Environmental Health Sciences, Research
Triangle Park, NC)
Technical Support Staff
Mattie Arnold—Senior Environmental Employee Program, National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Laeda Baston—Senior Environmental Employee Program, National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Kimberly Branch—Student Services Contractor, National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Ken Breito—Senior Environmental Employee Program, National Center for Environmental Assessment,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Eleanor Jamison—Senior Environmental Employee Program, National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Ryan Jones—Oak Ridge Institute for Science and Education, at National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Erica Lee—Oak Ridge Institute for Science and Education, at National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Barbara Liljequist—Senior Environmental Employee Program, National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
December 2009 xliii
-------
Ellen Lorang—National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Kelsey Matson—Student Services Contractor, National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Sandy Pham—Student Services Contractor, National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Olivia Phillpott—Senior Environmental Employee Program, National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Deborah Wales—National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Erica Wilson—Oak Ridge Institute for Science and Education, at National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Richard Wilson—National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Barbara Wright—Senior Environmental Employee Program, National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
December 2009 xliv
-------
Authors, Contributors, Reviewers
AUTHORS
Dr. Lindsay Wichers Stanek (PM Team Leader)—National Center for Environmental Assessment,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Jeffrey Arnold—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC (now at Institute for Water Resources, U.S. Army Corps of
Engineers, Washington, D.C)
Dr. Christal Bowman—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. James S. Brown—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Barbara Buckley—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Mr. Allen Davis—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Jean-Jacques Dubois—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Steven J. Dutton—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Tara Greaver—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Erin Hines—National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Douglas Johns—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Ellen Kirrane—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Dennis Kotchmar—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Thomas Long—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Thomas Luben—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Qingyu Meng—Oak Ridge Institute for Science and Education, Postdoctoral Research Fellow to
National Center for Environmental Assessment, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Dr. Kristopher Novak—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Joseph Pinto—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
December 2009 xlv
-------
Dr. Jennifer Richmond-Bryant—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Mary Ross—National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Jason Sacks—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Timothy J. Sullivan—E&S Environmental Chemistry, Inc., Corvallis, OR
Dr. David Svendsgaard—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Lisa Vinikoor—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. William Wilson—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Lori White— National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC (now at National Institute for Environmental Health Sciences, Research
Triangle Park, NC)
Dr. Christy Avery—University of North Carolina, Chapel Hill, NC
Dr. Kathleen Belanger —Center for Perinatal, Pediatric and Environmental Epidemiology, Yale
University, New Haven, CT
Dr. Michelle Bell—School of Forestry & Environmental Studies, Yale University, New Haven, CT
Dr. William D. Bennett—Center for Environmental Medicine, Asthma and Lung Biology, University of
North Carolina, Chapel Hill, NC
Dr. Matthew J. Campen—Lovelace Respiratory Research Institute, Albuquerque, NM
Dr. Leland B. Deck— Stratus Consulting, Inc., Washington, DC
Dr. Janneane F. Gent—Center for Perinatal, Pediatric and Environmental Epidemiology, Yale University,
New Haven, CT
Dr. Yuh-Chin Tony Huang—Department of Medicine, Division of Pulmonary Medicine, Duke University
Medical Center, Durham, NC
Dr. Kazuhiko Ito—Nelson Institute of Environmental Medicine, NYU School of Medicine, Tuxedo, NY
Mr. Marc Jackson—Integrated Laboratory Systems, Inc., Research Triangle Park, NC
Dr. Michael Kleinman—Department of Community and Environmental Medicine, University of
California, Irvine
Dr. Sergey Napelenok—National Exposure Research Laboratory, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Marc Pitchford—National Oceanic and Atmospheric Administration, Las Vegas, NV
Dr. Les Recio—Genetic Toxicology Division, Integrated Laboratory Systems, Inc., Research Triangle
Park, NC
Dr. David Quincy Rich—Department of Epidemiology, University of Medicine and Dentistry of New
Jersey, Piscataway, NJ
Dr. Timothy Sullivan— E&S Environmental Chemistry, Inc., Corvallis, OR
Dr. George Thurston—Department of Environmental Medicine, NYU, Tuxedo, NY
December 2009 xlvi
-------
Dr. Gregory Wellenius—Cardiovascular Epidemiology Research Unit, Beth Israel Deaconess Medical
Center, Boston, MA
Dr. Eric Whitsel—Departments of Epidemiology and Medicine, University of North Carolina, Chapel
Hill, NC
CONTRIBUTORS
Dr. Philip Bromberg—Department of Medicine, University of North Carolina, Chapel Hill, NC
Mr. Michael Burr—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Mr. Turhan Carroll—Student Services Contractor, National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Ms. Rosana Datti—Student Services Contractor, National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Mr. Neil Frank—Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Jonathan Krug—Student Services Contractor, National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Ms. Katie Lane—Student Services Contractor, National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Mr. Phil Lorang—Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Ms. Christina Miller—National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Ms. Irina Mordukhovich—Oak Ridge Institute for Science and Education, at National Center for
Environmental Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Elizabeth Oesterling Owens—National Center for Environmental Assessment, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Adam Reff—Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Ms. Victoria Sandiford— Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Mark Schmidt—Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Ms. Angelina Schultz—Student Services Contractor, National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Ms. Kirsten Simmons—Student Services Contractor, National Center for Environmental Assessment,
U.S. Environmental Protection Agency, Research Triangle Park, NC
Ms. Genee Smith—Oak Ridge Institute for Science and Education, at National Center for Environmental
Assessment, U.S. Environmental Protection Agency, Research Triangle Park, NC
Mr. Kurt Susdorf—Student Services Contractor, National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Barbara Turpin—Department of Environmental Sciences, Rutgers University, New Brunswick, NJ
December 2009 xlvii
-------
Ms. Lauren Turtle—Student Services Contractor, National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Ms. Rebecca Yang—Student Services Contractor, National Center for Environmental Assessment, U.S.
Environmental Protection Agency, Research Triangle Park, NC
PEER REVIEWERS
Dr. Sara Dubowsky Adar, Department of Epidemiology, University of Washington, Seattle, WA
Mr. Chad Bailey, Office of Transportation and Air Quality, Ann Arbor, MI
Mr. Richard Baldauf, Office of Transportation and Air Quality, Ann Arbor, MI
Dr. Prakash Bhave, National Exposure Research Laboratory, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. George Bowker, Office of Atmospheric Programs, U.S. Environmental Protection Agency,
Washington, D.C.
Dr. Judith Chow, Division of Atmospheric Sciences, Desert Research Institute, Reno, NV
Dr. Dan Costa, U.S. Environmental Protection Agency, Research Triangle Park, NC
Dr. Ila Cote, National Center for Environmental Assessment, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Robert Devlin, National Health and Environmental Effects Research Laboratory, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. David DeMarini, National Health and Environmental Effects Research Laboratory, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Neil Donahue, Department of Chemical Engineering, Carnegie Mellon University, Pittsburgh, PA
Dr. Aimen Farraj, National Health and Environmental Effects Research Laboratory, U.S. Environmental
Protection Agency, Research Triangle Park, NC
Dr. Mark Frampton, Department of Environmental Medicine, University of Rochester Medical Center,
Rochester, NY
Mr. Neil Frank, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Tyler Fox, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Jim Gauderman, Department of Environmental Medicine, Department of Preventive Medicine,
University of Southern California, Los Angeles, CA
Dr. Barbara Glenn, National Center for Environmental Research, U.S. Environmental Protection Agency,
Washington, D.C.
Dr. Terry Gordon, School of Medicine, New York University, Tuxedo, NY
Mr. Tim Hanley, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Jack Harkema, Department of Pathobiology and Diagnostic Investigation, Michigan State University,
East Lansing, MI
Ms. Beth Hassett-Sipple, Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC
December 2009 xlviii
-------
Dr. Amy Herring, Department of Biostatistics, University of North Carolina, Chapel Hill, NC
Dr. Israel Jirak, Department of Meteorology, Embry-Riddle Aeronautical University, Prescott, AZ
Dr. Mike Kleeman, Department of Civil and Environmental Engineering, University of California, Davis,
CA
Dr. Petros Koutrakis, Exposure, Epidemiology and Risk Program, Harvard School of Public Health,
Boston, MA
Dr. Sagar Krupa, Department of Plant Pathology, University of Minnesota, St. Paul, MN
Mr. John Langstaff, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Meredith Lassiter, Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Mr. Phil Lorang, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Karen Martin, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Ms. Connie Meacham, National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Mr. Tom Pace, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Jennifer Peel, Department of Environmental and Radiological Health Sciences, College of Veterinary
Medicine and Biomedical Sciences, Colorado State University, Fort Collins, CO
Dr. Zackary Pekar, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Rob Pinder, National Exposure Research Laboratory, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Norm Possiel, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Sanjay Rajagopalan, Division of Cardiovascular Medicine, Ohio State University, Columbus, OH
Dr. Pradeep Rajan, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Venkatesh Rao, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Ms. Joann Rice, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Mr. Harvey Richmond, Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Ms. Victoria Sandiford, Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Stefanie Sarnat, Department of Environmental and Occupational Health, Emory University, Atlanta,
GA
Dr. Frances Silverman, Gage Occupational and Environmental Health, University of Toronto, Toronto,
ON
December 2009 xlix
-------
Mr. Steven Silverman, Office of General Council, U.S. Environmental Protection Agency, Washington,
D.C.
Dr. Barbara Turpin, Department of Environmental Sciences, Rutgers University, New Brunswick, NJ
Dr. Robert Vanderpool, National Exposure Research Laboratory, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. John Vandenberg (Director)—National Center for Environmental Assessment-RTP Division, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Dr. Alan Vette, National Exposure Research Laboratory, U.S. Environmental Protection Agency, Research
Triangle Park, NC
Ms. Debra Walsh (Deputy Director)—National Center for Environmental Assessment-RTP Division, U.S.
Environmental Protection Agency, Research Triangle Park, NC
Mr. Tim Watkins, National Exposure Research Laboratory, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Christopher Weaver, National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Mr. Lewis Weinstock, Office of Air Quality Planning and Standards, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Ms. Karen Wesson, Office of Air Quality Planning and Standards, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. Jason West, Department of Environmental Sciences and Engineering, University of North Carolina,
Chapel Hill, NC
Mr. Ronald Williams, National Exposure Research Laboratory, U.S. Environmental Protection Agency,
Research Triangle Park, NC
Dr. George Woodall, National Center for Environmental Assessment, U.S. Environmental Protection
Agency, Research Triangle Park, NC
Dr. Antonella Zanobetti, Department of Environmental Health, Harvard School of Public Health, Boston,
MA
December 2009
-------
Clean Air Scientific Advisory Committee
for Particulate Matter NAAQS
CHAIRPERSON
Dr. Jonathan Samet, Department of Preventive Medicine, Keck School of Medicine, University of
Southern California, Los Angeles, CA
MEMBERS
Dr. Lowell Ashbaugh, Crocker Nuclear Lab, University of California, Davis, CA
Dr. Ed Avol, Department of Preventive Medicine, Keck School of Medicine, University of Southern
California, Los Angeles, CA
Dr. Joseph Brain*, Department of Environmental Health, Harvard School of Public Health, Harvard
University, Boston, MA
Dr. Wayne Cascio, Brody School of Medicine, East Carolina University, Greenville, NC
Dr. Ellis B. Cowling*, Colleges of Natural Resources and Agriculture and Life Sciences, North Carolina
State University, Raleigh, NC
Dr. James Crapo*, Department of Medicine, National Jewish Medical and Research Center, Denver, CO
Dr. Douglas Crawford-Brown, Department of Environmental Sciences and Engineering, University of
North Carolina at Chapel Hill, Chapel Hill, NC
Dr. H. Christopher Frey*, Department of Civil, Construction and Environmental Engineering, College of
Engineering, North Carolina State University, Raleigh, NC
Dr. David Grantz, Botany and Plant Sciences and Air Pollution Research Center, Riverside Campus and
Kearney Agricultural Center, University of California, Parlier, CA
Dr. Joseph Helble, Thayer School of Engineering, Dartmouth College, Hanover, NH
Dr. Rogene Henderson**, Lovelace Respiratory Research Institute, Albuquerque, NM
Dr. Philip Hopke, Department of Chemical Engineering, Clarkson University, Potsdam, NY
Dr. Donna Kenski*, Lake Michigan Air Directors Consortium, Rosemont, IL
Dr. Morton Lippmann, Nelson Institute of Environmental Medicine, New York University School of
Medicine, Tuxedo, NY
Dr. Helen Suh Macintosh, Environmental Health, School of Public Health, Harvard University, Boston,
MA
Dr. William Malm, National Park Service Air Resources Division, Cooperative Institute for Research in
the Atmosphere, Colorado State University, Fort Collins, CO
Mr. Charles Thomas (Tom) Moore, Jr., Western Regional Air Partnership, Western Governors'
Association, Fort Collins, CO
December 2009
-------
Dr. Robert F. Phalen, Center for Occupation & Environment Health, College of Medicine, Department of
Community and Environmental Medicine, Air Pollution Health Effects Laboratory, University of
California Irvine, Irvine, CA
Dr. Kent Pinkerton, Center for Health and the Environment, University of California, Davis, CA
Mr. Richard L. Poirot, Air Pollution Control Division, Department of Environmental Conservation,
Vermont Agency of Natural Resources, Waterbury, VT
Dr. Armistead (Ted) Russell*, Department of Civil and Environmental Engineering, Georgia Institute of
Technology, Atlanta, GA
Dr. Frank Speizer, Channing Laboratory, Harvard Medical School, Boston, MA
Dr. Sverre Vedal, Department of Environmental and Occupational Health Sciences, School of Public
Health and Community Medicine, University of Washington, Seattle, WA
* Members of the statutory Clean Air Scientific Advisory Committee (CAS AC) appointed by the EPA
Administrator.
**As immediate past CASAC Chair, Dr. Henderson is invited to participate in CASAC advisory activities
for FY 2009.
SCIENCE ADVISORY BOARD STAFF
Dr. Holly Stallworth, Economist and Designated Federal Officer, Clean Air Scientific Advisory
Committee, Environmental Economics Advisory Committee, Washington, D.C.
December 2009
-------
Acronyms and Abbreviations
a alpha, ambient exposure factor
a-HCH alpha-hexachlorocyclohexane
A Angstrom
A surface albedo
AAC abdominal aortic calcium
AAS atomic absorption spectrophotometry
AB Alcian Blue stain
ABC Asian Brown Cloud
ABI ankle-arm or resting blood pressure index
AC air conditioning
Ace acenaphthene
ACE-1 angiotensin converting enzyme-1
ACEAsia (Asian Pacific Regional) Aerosol Characterization Experiment
ACGIH American Conference of Governmental Industrial Hygienists
ACh acetylcholine
Acl acenaphthylene
ACP accumulation mode particle
ACS American Cancer Society
Ad4BP adrenal-4-binding protein
ADEOS-1 Advanced Earth Observing Satellite-1
A-DEP automobile diesel exhaust particles
ADM angular distribution model(s), angular dependence model
ADMA asymmetric dimethylarginine
AD-Net Asian Dust Network
Ae AERONET
AeroCom Aerosol Comparisons between Observations and Models
AERONET NASA AERosol RObotic NETwork
AF atrial fibrillation
AGA appropriate for gestational age
AGE advanced glycation end product
AHR airway hyperresponsiveness, airway hyperreactivity
AhR arylhydrocarbon receptor
December 2009
-------
AH SMOG
AI
AIC
AIM
AIOP
AIRS
Al
ALI
AM
AM,AMF
AMAP
AMDP
AMI
AMS
Angll
ANOVA
ANP
ANS
Ant
AOD
AP-1
APC
APCS
APEX
APHEA
APO
ApoE
APS
aPTT
AQCD
AQI
AQM
AQS
Aqua
AR4
California Seventh Day Adventist study
aerosol index
Akaike's information criterion
ambient ion monitor
2003 Aerosol Intensive Operating Period
Aerometric Information Retrieval System
aluminum
air liquid interface
alveolar macrophage(s)
arbuscular mycorrhizal
Arctic Monitoring and Assessment Programme
annual maximum of daily precipitation
acute myocardial infarction
aerosol mass spectrometry
angiotensin II
analysis of variance
atrial natriuretic peptide
autonomic nervous system
anthracene
aerosol optical depth
activator protein 1
antigen presenting cell(s)
Absolute Principal Components Scores
Air Pollutants Exposure Model
Air pollution and Health: a European Approach
apocynin
apolipoprotein E
aerodynamic particle sizer, aerosol polarimetry sensor
activated partial thromboplastin time
Air Quality Criteria Document
Air Quality Index
air quality model
U.S. EPA Air Quality System database
NASA satellite
Fourth Assessment Report (AR4) from the IPCC
December 2009
liv
-------
ARCTAS
ARD
ARDS
ARI
ARIC
ARIES
ARM
ARQM
ARS
As
ASDNN5
ASOS
ATOFMS
ATP
A- Train
ATS
AURA
avg
AVHRR
B
P-HCH
3PHSD
PTGF
Ba
BaA
BAD
BAL
BALB/c
Arctic Research of the Composition of the Troposphere from Aircraft
and Satellites
Air Resources Division
adult respiratory distress syndrome
acute respiratory infection
Atherosclerosis Risk in Communities study
Aerosol Research and Inhalation Epidemiology Study
Atmospheric Radiation Measurement program
Air Quality Research Branch (Meteorological Service of Canada
Toronto)
Air Resource Specialists
arsenic
mean of the standard deviation in all 5-min segments of a EKG 24 h
recording
Automated Surface Observing System
aerosol time-of-flight mass spectrometry
adenosine triphosphate
a group of 5 afternoon overpass satellites (Aura, PARASOL,
CALIPSO, CloudSat, Aqua)
American Thoracic Society
NASA satellite
average
Advanced Very High Resolution Radiometer
beta, beta coefficient, slope
beta-hexachlorocyclohexane(s)
3 p-hydroxysteroid dehydrogenase
P transforming growth factor
absorption by gases coefficient
absorption by particles coefficient
light extinction coefficient
scattering by gases coefficient
sum of light scattering by (aerosol) particles coefficient
barium
benz[a]anthracene
bronchial artery diameter
bronchoalveolar lavage
albino inbred mouse strain
December 2009
-------
BALF bronchoalveolar lavage fluid
BALT bronchus-associated lymphoid tissues
BAM beta attenuation monitor
BaP benzo[a]pyrene
BASIC Brain Attack Surveillance in Corpus Christi
BASE-A Burning Airborne and Spaceborne Experiment - Amazon and Brazil
BbF benzo[b]fluoranthene
BC black carbon
BCC-CMI Beijing Climate Center - Carbon Mitigation Initiative
BCCR Bjerknes Centre for Climate Research
BeP benz[e]pyrene
BghiP, BpPe benzo[g,h,i]perylene
BGT beta-gauge technique
BH4 tetrahydrobiopterin
bhp brake horsepower
BkF benzo[k]fluoranthene
BMI body mass index
BMP bone morphogenetic protein
BN/BR Brown Norway rat strain
BNP brain natriuretic peptide, B-type natriuretic peptide
BOSS BYU Organic Sampling System
BP blood pressure
BPM blowing PM2.5
BpPe benzo[ghi]perylene
BPQ benz(ayrene (BaP)-quinone
Br bromine
BRAVO Big Bend Regional Aerosol and Visibility Observational (Study)
BrdU bromodeoxyuridine
BS black smoke
BUG bucillamine (N-[2-mercapto-2-methylpropionyl]-L-cysteine)
B VAIT B-Vitamin Atherosclerosis Intervention Trial
BYU Brigham Young University
C carbon
C4 Center of Clouds, Chemistry and Climate
12C carbon-12
December 2009
Ivi
-------
C60(OH)24
Ca
CAA
CAAA
CAAM
CAC
CaCO3
CAD
CALINE
CALIOP
CALIPSO
CAM
CAMM
CAMP
CAMx
Ca(NO3)2
CAP
CAPMoN
CASAC
CaSO4
CASTNet
CATS
CB
CB-Fe
CB(P)
CBSA
CB-V
CBVD
CC16
CCCma
CCM3
CCN
CCPM
carbon-13
carbon-14
water-soluble fullerene
calcium
Clean Air Act
1977 Clean Air Act Amendments
continuous ambient mass monitor
coronary artery calcification
calcium carbonate
coronary artery disease
California Line Source Dispersion Model
Cloud and Aerosol Lidar with Orthogonal Polarization
Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations
Community Atmosphere Model
continuous ambient mass monitor
Childhood Asthma Management Program
comprehensive air quality model with extensions
calcium nitrate
concentrated ambient particle
Canadian Air and Precipitation Monitoring Network
Clean Air Scientific Advisory Committee
calcium sulfate
Clean Air Status and Trends Network
cumulative air toxics surface
carbon black, chronic bronchitis
carbon black particles artificially coated with Fe(II) salt.
carbon black (particles)
Core-Based Statistical Area
carbon black particles artificially coated with a targeted concentration
of Vanadium (IV) salt
cerebrovascular disease
Clara cell protein, Clara cell 16 protein
Canadian Centre for Climate Modeling and Analysis
Community Climate Model
cloud condensation nuclei
continuous coarse particle monitor
December 2009
-------
CCSM3
CCSP
Cd
CDC
CDE
CDNC
CDPHE
Ce
CEN
CenRAP
CERES
CERFACS
CF
CFA
CFD
CFR
CGCM3.1
cGMP
CH2C12
CH2O
CH4
CHAD
CHD
CHF
CHL
CHO
Chr
CHS
CI
GIF
CUT
CIMT
Cl
CL
CLAMS
Community climate system model, version 3
Climate Change Science Program
cadmium
Centers for Disease Control and Prevention
conjugated diene
cloud droplet number concentration
Colorado Department of Public Health and Environment
cerium
European Committee for Standardization
Central Regional Air Planning Association
Clouds and the Earth's Radiant Energy System
European Centre for Research and Advanced Training in Scientific
Computation
coronary flow, cystic fibrosis
coal fly ash
cystic fibrosis disease
Code of Federal Regulations
Coupled global climate model
cyclic guanosine monophosphate
methylene chloride
formaldehyde
methane
Consolidated Human Activity Database
chronic heart disease
congestive heart failure
crown heel length
Chinese hamster ovary cells
chrysene
Children's Health Study
confidence interval
carbon-impregnated charcoal filter
Chemical Industry Institute of Toxicology
carotid intimal-medial thickness
chlorine
chemiluminescence
Chesapeake Lighthouse and Aircraft Measurements for Satellites
December 2009
-------
CM
CMAQ
CMAR
CMB
CMD
ONES
CNP
CNRM
CNRM
CNS
Co
CO
C02
COD
COH, CoH
CONUS
COPD
CoPP
COX-2
CPC
CPZ
Cr
C-R
CRP
Cs
137Cs
CS
CSA
CSC
CSE
cSHMT
CSIRO
CSN
CTM
conditioned medium, cell culture medium
Community Multi-scale Air Quality modeling system
CSIRO Marine and Atmospheric Research
chemical mass balance
count median diameter
Centre National d"Etudes Spatiales
carbon nano particle
Centre National de Recherches Meteorologiques
Center National Weather Research
central nervous system
cobalt
carbon monoxide
carbon dioxide
coefficient of divergence
coefficient of haze
continental United States
carboxyl group
chronic obstructive pulmonary disease
cobalt protoporphyrin
cyclooxygenase 2 enzyme
condensation particle counter
capsazepine
chromium
concentration-response (relationship)
C-reactive protein
cesium
cesium-137
cigarette smoke
Combined Statistical Area
cigarette smoke condensates
cigarette smoke extract
cytosolic serine hydroxymethyltransferase
Commonwealth Scientific and Industrial Research Organization
Chemical Speciation Network
chemistry-transport model, chemical transport model
December 2009
lix
-------
Cu copper
CuSO4 copper sulfate
Cu/Zn SOD Cu/Zn superoxide dismutase
CUP Current Use Pesticide
CV cardiovascular, coefficient of variation
CVD cardiovascular disease(s)
CVM contingent valuation method
GYP cytochrome P450
CYP 1A1 cytochrome P450 1A1
A delta, change, difference
AFEVj change in forced expiratory volume in one second
d50 50 percent cut point or 50 percent diameter
dae aerodynamic diameter of a particle
D diameter
Da Dalton
DAAC Distributed Active Archive Center
DAASS Dry Ambient Aerosol Size Spectrometer
DABEX Dust and Biomass-burning Experiment (in West Africa)
DAR denuded aortic ring
DAX-1 x-chromosome gene-1
DBA dibenzo(a,hnthracene
DBF diastolic blood pressure
DC dendritic cell
DC diesel exhaust particles + cigarette smoke condensates
DCS Douglas aircraft
DCF direct climate forcing, 2',7'-dichlorofluorescin
DDT dichlorodiphenyltrichloroethane
DE diesel exhaust
DEE diesel exhaust extract
DEP diesel exhaust particle
DEPAL diesel exhaust particles aliphatic (extract)
DEPAR diesel exhaust particles aromatic (extract)
DEPE diesel exhaust particles extract
DEPM diesel exhaust particles methanol (extract)
DEPME diesel exhaust particles methylene chloride extract
December 2009
-------
DEPPO diesel exhaust particles polar (extract)
Dex dexamethasone
dF\d change fold, unit change in property
DFO desferrioxamine (Desferral) an iron chelator
DFX deferasirox (Exjade) an oral iron chelator
DHR dihydrorhodamine 123
DLCO carbon monoxide diffusing capacity
DMEM Dulbecco's modified Eagle's medium (culture medium)
DMSO dimethyl sulfoxide
DMT1 divalent metal transporter-1 protein
DMTU dimethylthiourea
DNA deoxyribonucleic acid
DOE U.S. Department of Energy
dpc days post conception
DPC dodecylphosphocholine
DPCC 1,2-dipalmitoyl-SN-glycero-3-phosphocholine
DPI diphenyleneiodonium
DPM diesel particulate matter
DPPC dipalmitoylphosphatidylcholine
DRE direct radiative effects
DRF direct radiative forcing
DRUM Davis Rotating Uniform size-cut Monitor
DS diffusion screens
DSP daily sperm production
DTMA Dynamic mechanical thermal analysis
DTPA diethylene triamine pentaacetic acid
DU dust
dv deciview(s)
DVT deep vein thrombosis
EAD electrical aerosol detector
EANET Acid Deposition Monitoring Network in East Asia
EARLINET European Aerosol Research Lidar Network
EarthCARE Earth Clouds, Aerosols and Radiation Explorer
EAST-AIRE East Asian Study for Tropospheric Aerosols
EBCT electron beam computed tomography
December 2009
Ixi
-------
EC
ECE-1
EGG, EKG
ECHAM5
ECHO-G
ECRHS
EC/TC
ED
EDGAR
EDTA
ED-XRF
EGM
EGU
EHC-93
EKG, EGG
ELISA
EMECAS
EMEP
eNO
eNOS
EOS
EPA
ER
ERBS
ERK1/2
ESRL
ESTR
ET
ET
ET
ETA
ETB
ETS
EU
elemental carbon
endothelin converting enzyme-1
electrocardiogram
European Centre Hamburg with Hamburg Aerosol Module
(ECHAM4 + HOPE-G):
European Community Respiratory Health Survey
ratio of elemental carbon to total carbon
emergency room, emergency department
Emissions Database for Global Atmospheric Research
ethylenediaminetetraacetic acid
energy dispersive X-ray fluorescence
electrogram
electricity-generating unit
Ottawa dust; urban air particulate matter PMi0, collected in 1993 in
Ottawa, Canada
electrocardiogram
enzyme-linked immunosorbent assay
Spanish Multi-centric Study on the Relation between Air Pollution
and Health
European Monitoring and Evaluation Programme
exhaled nitric oxide
endothelial nitric oxide synthase
Earth Observing System
U.S. Environmental Protection Agency
estrogen receptor
Earth Radiation Budget Satellite
ERK-1 (MAPK p42) and ERK-2 (MAPK p44)
Earth System Research Laboratory
expanded simple tandem repeat
forcing efficiency
extrathoracic region
endothelin
endothelin A receptor subtype
endothelin B receptor subtype
environmental tobacco smoke
endotoxin units
December 2009
Ixii
-------
F breathing frequency
f the ratio of ambient aerosol mass (wet) to dry aerosol mass M.
fasp(RH) total light scattering coefficient at given relative humidity(RH) values
faf anthropogenic fraction of fine-mode fraction
ff fine mode fraction
f(RH) the unitless water growth term that depends on relative humidity
F fine particles
F344 Fisher 344 strain of rats
Fa adjusted forcings
FA filtered air
FAC ferric ammonium citrate
FBI Federal Bureau of Investigation
FBS fetal bovine serum
PCS fetal calf serum
FDMS Filter Dynamics Measurement System
FDMS-TEOM Filter Dynamics Measurement System - Tapered Element Oscillating
Microbalance
Fe effective (Fe) forcings
Fe iron
Fe2(SO4)3 ferric sulfate
FeCl3 ferric chloride
FEF forced expiratory flow
FEF25_75 mean forced expiratory flow over the middle half of the forced vital
capacity
FEF50o/0 mid-expiratory flow
FEM Federal Equivalent Method
FeNO fractional exhaled nitric oxide
FERA Fire and Environmental Research Applications (Team)
FEV] forced expiratory volume in one second
FGA one fibrinogen alpha chain
FOB one fibrinogen beta chain
FGOALS-gl.O Flexible Global Ocean-Atmosphere-Land System Model
Fi instantaneous forcing
FID flame ionization detection
FIMS fast integrated mobility scanners
Fjnf infiltration factors
December 2009
Ixiii
-------
FKHR Proapoptotic Factor FOXO1
Fie fluorine
Flu fluoranthene
FMD flow-mediated dilation
FPG formamidopyrimidine-DNA glycosylase
f-PM, FPM fine particulate matter
FR Federal Register
FRM Federal Reference Method
FROSTFIRE The landscape-scale prescribed research burn in the boreal forest of
interior Alaska, July 1999; conducted by FERA.
Fs SST forcing(s), forcing driven by sea surface temperature (SST)
Fsfc mean net solar flux at the (Earth) surface
FT free troposphere
FTIR Fourier transform infrared spectrometry
F/UFP mix of fine and ultrafine particles, all < 2.5 |im
FVC forced vital capacity
yGCS gamma glutamylcysteine sythetase
Ga gallium
GAM generalized additive model
GATOR Gas, Aerosol, Transport, and Radiation model
GATORG Gas, Aerosol, Transport, Radiation, and General circulation model
GAW Global Atmospheric Watch network
GBS group B streptococcus
GC gas chromatography
GCM(s) general circulation model(s), global climate model
GCMOM General Circulation, Mesoscale and Ocean Model
GC/MS gas chromatography/mass spectrometry
GCS gamma glutamylcysteine sythetase
GD gestational day
GDF growth differentiation factor (e.g., GDF-9)
GEE generalized estimating equations, gasoline engine exhaust
GEIA Global Emissions Inventory Activity
GEM gaseous elemental mercury
GEOS-Chem Goddard Earth Observing System-CHEMistry
GFAAS graphite furnace atomic absorption spectrometry
GFAP glial fibrillatory acidic protein
December 2009
Ixiv
-------
GFDL Geophysical Fluid Dynamics Laboratory
GFDL-CM2.X GFDL Climate Models
GFED Global Fire Emission Database
GGT gamma-glutamyltranspeptidase
GHG greenhouse gas
GIS Geographic Information System
GISS Goddard Institute for Space Studies
GISS-AOM GISS Atmosphere-Ocean Model climate prediction model
GISS-EH GISS AOM for sea ice model
GIS S-ER GIS S AOM for liquid sea model
GLAS Geoscience Laser Altimeter System
GLM generalized linear models
GM geometric mean
GM-CSF granulocyte macrophage colony-stimulating factor
GMD Global Monitoring Division
GMS Greater Mekong Subregion
GOCART Goddard Chemistry Aerosol Radiation and Transport
GOES Geostationary Operational Environmental Satellite
GoMACCS Gulf of Mexico Atmospheric Composition and Climate Study
GPS Global Positioning System
GSD geometric standard deviation
GSFC NASA Goddard Space Flight Center
GSH glutathione
GSFLGSSG ratio of reduced glutathione to glutathione disulfide (oxidized
glutathione)
GSO, GSNO S-Nitrosoglutathione
GSSG glutathione disulfide; oxidized glutathione
GST glutathione-S-transferase
GSTM1 glutathione S-transferase polymorphism Ml
GSTP1 glutathione-S-transferase polymorphism PI
GSTT1 glutathione-S-transferase polymorphism Tl
GWP global warming potential
h hour
H atomic hydrogen, hydrogen radical, height, heart rate, high dose, high
exposure
H+ hydrogen ion
December 2009
Ixv
-------
HR heart rate
H2 molecular hydrogen
H2CO formaldehyde
H2O water
H2O2 hydrogen peroxide
H2S hydrogen sulfide
H2SO4 sulfuric acid
H9c2 rat embryonic cardiomyocytes cell line
HA hospital admission
HAEC Human Aortic Endothelial Cell
HAPC Harvard ambient particle concentrator
HBE, HBEC Human Bronchial Epithelial cells
HC hydrocarbon(s); head circumference
HCB hexachlorobenzene
HCH hexachlorocyclohexane(s) (e.g. a-HCH, P-HCH)
HDL high density lipoprotein
HEAPSS Health Effects of Air Pollution among Susceptible Subpopulations
study
HEI Health Effects Institute
HEPA high efficiency particle air (filter)
HERO Health and Environmental Research Online, NCEA Database System
HF heart failure, high frequency (HRV parameter), high (dose/exposure)
filtered
HFCD High-Fat Chow Diet
HFE HFE gene, HFE protein
Hg mercury
Hg(0) gaseous elemental mercury
Hg(II) gaseous divalent (oxidized) mercury
HH hereditary hemochromatosis
HNRS Hans Nixdorf Recall Study
HO-1 heme oxygenase-1
hOGGl 8-hydroxyguanine DNA-glycosylase
HOPE-G Hamburg Atmosphere-Ocean Coupled Circulation Model
hPA hectopascal
hPAEC human pulmonary artery endothelial cells
hPBMC human peripheral blood mononuclear cells
December 2009
Ixvi
-------
HPLC
HPMF
hPMVEC
HR
HRV
HSD
HSP-70
HSPH
HSRL
HUVEC
hv
HWS
Hz
1C
ICAM-1
ICARTT
ICAS
ICD
ICD-9
ICD-10
ICESat
ICP-AES
ICP-MS
ICR
ICRP
IDP
IFN-y
IPS
Ig
IGS
IHD
IIASA
IL
iMDDC
high pressure liquid chromatography
high particulate matter filtered
human pulmonary microvascular endothelial cells
heart rate, hazard ratio, high level DE
heart rate variability
17p-hydroxysteroid dehydrogenase
heat shock protein
Harvard School of Public Health
High Spectral Resolution Lidar
human umbilical vein endothelial cells
photon
hardwood smoke
hertz
ion chromatography
intercellular adhesion molecule-1
International Consortium for Atmospheric Research on Transport and
Transformation
Inner-City Asthma Study
implantable/implanted cardioverter defibrillator
International Classification of Disease 9th revision
International Classification of Disease 1 Oth revision
Ice, Cloud and land Elevation Satellite
inductively coupled plasma-atomic emission spectroscopy
inductively-coupled plasma-mass spectrometry
imprinting control region, mouse strain
International Commission on Radiological Protection
indenof 1,2,3-c,d]pyrene
interferon-gamma
Integrated Forest Study
immunoglobulin (e.g., IgE)
International Genetic Standard
ischemic heart disease
International Institute for Applied Systems Analysis
interleukin
immature monocyte-derived dendritic cells
December 2009
Ixvii
-------
IMPACT Interactive Modeling Project for Atmospheric Chemistry and
Transport
IMPROVE Interagency Monitoring of Protected Visual Environment
IN ice nuclei
INAA instrumental neutron activation analysis
INCA Interactions between Chemistry and Aerosol
INDOEX Indian Ocean Experiment
INGV-SXG Istituto Nazionale di Geofisica e Vulcanologia coupled to SINTEX-G
INM-CM3.0 Institute of Numerical Mathematics climate model
iNOS inducible nitric oxide synthase
INTEX Intercontinental Chemical Transport Experiment
I/O indoor-outdoor ratio
IOM Institute of Medicine
i.p. intraperitoneal
IP inhalable particle
IPCC Intergovernmental Panel on Climate Change
IPSL-CM4 Institut Pierre Simon Laplace climate model
IQR interquartile range
Ir iridium
IR incidence rate, infrared radiation
IRE iron responsive element
IRMS isotope ratio mass spectrometer
ISA Integrated Science Assessment
ISO International Standards Organization
ISO isoprene, 2-methyl analog of 1,3-butadiene
IT intratracheal, intratracheally
IUGG International Union of Geodesy and Geophysics
IUGR intrauterine growth restriction, intrauterine growth retardation
i.v. intravenous
JNK c-jun N-terminal kinase
KB kappa B
K potassium
KG local neutrophil chemoattractant protein
kHz kilohertz
kJ kilojoules
KLH keyhole limpet hemocyanin
December 2009
Ixviii
-------
km kilometer
km"1 inverse kilometer
Kow octanol-water partition coefficient
L, dL, mL, uL Liter, deciLiter, milliLiter, microLiter
L low
La lanthanum
LAC light-absorbing carbon
LACE98 Lindenberg Aerosol Characterization Experiment 1998
LBA-SMOCC Large-Scale Atmosphere-Biosphere Experiment in Amazon
LEW low birth weight
LC lethal concentration
LC50 median lethal concentration
LDH lactate dehydrogenase
LDL low-density lipoprotein
LDLR low-density lipoprotein receptor
LDVP left developing ventricular pressure
LES large eddy simulations model
LF low frequency an HRV parameter
LF/HF ratio of LF to HF an HRV parameter
LIBS laser induced breakdown spectroscopy
LIF leukemia inhibitory factor
LITE Lidar In-space Technology Experiment
LMD Laboratoire de Meteorologie Dynamique
LMDz LMD with Zoom
LMDZ-INCA LMDZ INteractive Chemistry and Aerosols model
LMDZ-LOA LMDZ with Laboratoire d'Optique Atmospherique model
L-NAME arginine analog; N(G)-nitro-L- arginine methyl ester
L-NMMA N(G)-mono-methyl-L-arginine
LnRMSSD natural log of RMSSD; measure of HRV
InSDNN natural log of the standard deviation of NN intervals in an EKG
LOA Laboratoire d'Optique Atmospherique
LOESS locally weighted scatterplot smoothing
LOSU level of scientific understanding
Lpm liters per minute (L/min)
LPMF low particulate matter filtered
December 2009
Ixix
-------
LPO plasma lipid peroxides
LPS lipopolysaccharide
LRAT long range atmospheric transport
LROT long range oceanic transport
LSCE Laboratoire des Sciences du Climat et de l'Environnement
LSDF low-sulfur diesel fuel
LTB4 leukotriene B4
LTE4 leukotriene E4
LUA NRW The North Rhine-Westphalia State Environment Agency
LUDEP LUng Dose Evaluation Program
LUR land use regression
LV left ventricle
LVEDP left-ventricular end-diastolic pressure,
LVSP left-ventricular systolic pressure, left ventricular developed pressure
L/W ratio of lumen to wall
LWC liquid water content
LWDE Low Whole Diesel Exhaust
LWP liquid water path
jig microgram
Hg/m3 micrograms per cubic meter
|im micrometer, micron
m, cm, \im, nm meter(s), centimeter(s), micrometer(s), nanometer(s)
M, mM, uM, nM, pM Molar, milliMolar, microMolar, nanoMolar, picoMolar
M dry aerosol mass, medium dose/exposure
ma moving average
MAM March-April-May
MAN Maritime Aerosol Network
MANE-VU Mid-Atlantic/Northeast Visibility Union
MAP mitogen-activated protein, mean arterial pressure
MAPK mitogen-activated protein kinase(s), MAP kinase
MARAMA Mid Atlantic Regional Air Management Association
MATCH Model of Atmospheric Transport and Chemistry
max maximum
MBP major basic protein
MCAPS Medicare Air Pollution Study
December 2009
Ixx
-------
Mch methacholine
MCN mixed carbon nanoparticle
MCP-1 monocyte chemoattractant protein 1
MC V mean corpuscular volume
MD mineral dust
MDA malondialdehyde
MDCT multidetector computed tomography
ME Multilinear Engine
MEE mass extinction efficiency
MEF maximal expiratory flow
MEF50 maximum expiratory flow rate at 50% of vital capacity
MeHg methyl mercury
MENTOR Modeling Environment for Total Risk Studies
MEP motorcycle exhaust particulate(s)
MEPE motorcycle exhaust particulate extract (particle-free)
MESA Multi-Ethnic Study of Atherosclerosis
MFFSR multifilter rotating shadowband radiometer
mg/m3 milligrams per cubic meter
Mg magnesium
MI myocardial infarction
MIROC3 .x Model for Interdisciplinary Research on Climate
MILAGRO Megacity Initiative: Local and Global Research Observations, study
of air pollution in Mexico City
min minute(s), minimum
MINOS MPI Mediterranean INtensive Oxidant Study
MIP-2 macrophage inflammatory protein-2
MIRAGE Megacities Impact on Regional and Global Environment program
MIS mullerian inhibiting substance
MISR Multi-angle Imaging SpectroRadiometer
Mm megameter
Mm"1 inverse megameter
MM monocyte-derived macrophages
MM5 mesoscale model
MMAD mass median aerodynamic diameter
MMD mass median diameter
MMEF maximal mid-expiratory flow
December 2009
Ixxi
-------
mmHg millimeters of mercury
MMP mitochondria membrane potential
MMP(2,9) matrix metalloproteinase (2, or 9)
MMT million metric tons
Mn manganese
MN micronuclei
MnSO4 manganese sulfate
MnSOD manganese superoxide dismutase
MnTBAP manganese tetrakis (4-benzoic acid) porphyrin
mo month
MOA mode(s) of action
MODIS MODerate resolution Imaging Spectroradiometer
MOUDI Micro-Orifice Uniform Deposit Impactor
MOZART MOdel for Ozone and Related chemical Tracers
MP mid polar, myelopeptide
MFC mean platelet component
MPF median power frequency
MPG N-(2-mercaptopropionyl) glycine
MPI Max Planck Institute for Meteorology
MPLNET Micro-Pulse Lidar Network
MPO myeloperoxidase
MPPD Multiple-Path Particle Dosimetry model
MPV mean platelet volume
MRI Meteorological Research Institute
MRI-CGCM MRI coupled general circulation model
mRNA messenger RNA
MRPO Midwest Regional Planning Organization
ms millisecond
MSA metropolitan statistical area
MSH melanocyte stimulating hormone
MSHA Mount St. Helen ash
MSU monosodium urate crystals
MT metric ton
MTHFR methylenetetrahydrofolate reductase
MTT methyl thiazol tetrazolium
December 2009
Ixxii
-------
MV motor vehicle
MWNT multiple-wall nanotube
M/Z mass-to-charge ratio
N nitrogen
N2O nitrous oxide
Na sodium
Na2SO4 sodium sulfate
NAAQS National Ambient Air Quality Standards
NAC N-acetylcysteine, a thiol antioxidant
NaCl sodium chloride
NADPH reduced form of nicotinamide adenine dinucleotide phosphate
NAG N-acetyl-p-D-glucosaminidase
Na,K-ATPase sodium-potassium adenosine triphosphatase
NAMS National Ambient Monitoring Stations
NaN3 sodium azide
NaNO3 sodium nitrate
nano-BAM low pressure-drop ultrafine particle impactor coupled with a Beta
Attenuation Monitor
NAPAP National Acid Precipitation Assessment Program
NAPCA National Air Pollution Control Administration
NAS National Academy of Sciences
NASA U. S. National Aeronautics and Space Administration
NASDA National Space Development Agency, Japan
NATA U. S. EPA's National Air Toxics Assessment
2-NB 2-nitrobenzanthrone
NC total (particle) number concentration
NC AR National Center for Atmospheric Research
NCC-MPSP negatively charged carboxylate-modified polystyrene particle(s)
NCD Normal Chow Diet
NCEA National Center for Environmental Assessment
NCHS National Center for Health Statistics
NCICAS National Cooperative Inner-City Asthma Study
NCore National Core
Nd drop number concentration
Nd: YAG neodymium-doped yttrium aluminum garnet laser
NDDN National Dry Deposition Network
December 2009
Ixxiii
-------
NEAQS NOAA New England Air Quality Study
NEI National Emissions Inventory
NESCAUM Northeast States for Coordinated Air Use Management
NET National Emissions Trends database
NFicB nuclear factor kappa-B
NG neutrophil granulocytes
NH northern hemisphere
NH3 ammonia
NH4+ ammonium ion
NH4NO3 ammonium nitrate
(NH4)2SO4 ammonium sulfate
NHANES National Health and Nutrition Examination Survey
NHBE(C) normal human bronchial epithelial cells
NHPAE normal human pulmonary artery endothelial cells
NHS Nurses' Health Study
Ni nickel
NIOSH National Institute for Occupational Safety and Health
NIST National Institute of Standards and Technology
NMHC non-methane volatile hydrocarbon
NMMAPS U. S. National Morbidity, Mortality, and Air Pollution Study
NO nitric oxide
NO2, NO2 nitrogen dioxide, nitrogen dioxide radical
NO3~ nitrate
NOAA National Oceanic and Atmospheric Administration
NOAEL no observed adverse effect level
NOS nitric oxide synthase
NOS3 nitric oxide synthase 3
NOx nitrogen oxides, oxides of nitrogen (NO + NO2)
NP National Park
NPM non-blowing PM2 5
NPOESS National Polar-orbiting Operational Environment Satellite System
NFS National Park Service, U.S. Department of the Interior
NR not reported
NR5A1 nuclear receptor subfamily 5, group A, member 1
NRC National Research Council
December 2009
Ixxiv
-------
NRPB
NSA
NT
NWS
NYHA
O
O2
03
OAQPS
OC
OCM
OE
OGG1
OH, OH'
8-OHdG
OM
OMI
OMM
OR
OSM
OSPM
OVA
oxLDL
8-oxodG
ox-PAPC
P450
P450cl7
P450scc
P90
P
P
PA
PAF
PAH
PAI
National Radiological Protection Board
North Slope Alaska
neurotrophin, nitrotyrosine
National Weather Service
New York Heart Association
oxygen
molecular oxygen
ozone
Office of Air Quality Planning and Standards
organic carbon
organic carbon mass
organic extracts
8 oxo-guanine repair enzyme
hydroxyl group, hydroxyl radical
8-hydroxydeoxyguanosine
organic matter
Ozone Monitoring Instrument
organic molecular marker
odds ratio(s)
oncostatin M, a cytokine
Operational Street Pollution Model
ovalbumin
oxidation of LDL, marker of oxidative stress
8-oxo-7-hydrodeoxyguanosine
oxidized l-palmitoyl-2-arachidonoyl-sn-glycero-3-phosphorylcholine
cytochrome P450
cytochrome P450 17-a-hydroxylase
cytochrome P450 cholesterol side chain cleavage enzyme
90th percentile; Printex 90
probability value
phosphorus
photoacoustic analyzer, physical activity, plasminogen activator,
pulmonary arterial, alveolar pressure
platelet-activating factor
polycyclic aromatic hydrocarbon(s)
plasminogen activator inhibitor, (e.g. PAI-1)
December 2009
Ixxv
-------
PALMS NOAA Particle Analysis by Laser Mass Spectrometry instrument
PAMCHAR Chemical and Biological Characterisation of Ambient Air Coarse,
Fine, and Ultrafine Particles for Human Health Risk Assessment in
Europe
PAMS Photochemical Assessment Monitoring Stations network
PAR photosynthetically active radiation
PAR(s) Pulmonary Artery Rings
PARASOL Polarization and Directionality of the Earth's Reflectances, coupled
with observations from a Lidar, a CNES satellite
PARP poly(ADP-ribose) polymerase
PAS Periodic Acid Schiff stain
Pb lead
207Pb lead-207
PBDE polybrominated diphenyl ether
PEL planetary boundary layer
PBMC peripheral blood mononuclear cell
PBMM peripheral blood monocyte-derived macrophages
PBP primary biological particle(s)
PBS phosphate buffered saline
PC synthetic carboxylate-modified particles
PCA principal component analysis
PCA-MPSP positively-charged amine modified polystyrene particle
PCB polychlorinated biphenyl(s)
PCDD polychlorinated dibenzo-p-dioxin
PCIS Personal Cascade Impactor Sampler
PCM NCAR Parallel Climate Model
PCPSP positively charged polystyrene particle
PCR polymerase chain reaction
PDF probability distribution functions
pDR personal DataRam
PE post exposure, post exercise, phenylephrine
PEACE Pollution Effects on Asthmatic Children in Europe study
PEC particulate elemental carbon
PEC AM-1 platelet endothelial cell adhesion molecule 1
PEF peak expiratory flow (L/min)
PEFR peak expiratory flow rate
December 2009
Ixxvi
-------
PEFT
PEM
PEM-West
Penh
Per
PESA
PFDE
PGE2
PGI2
Phe
PI
PICT
PILS
PILS-IC
PIXE
PKA
PLS
PM
PMx
PMx-y
FMo.i
PM2.5
PM10
PM10.2.5
PMA
PMF
time to peak flow
personal exposure monitor
NASA Pacific Exploratory Missions in the western Pacific
enhanced pause
perylene
particle elastic scattering analysis
particle free diesel exhaust
prostaglandin E2
prostacyclin
phenanthrene
post instillation, posterior interval, pulmonary inflammation
pollution-induced community tolerance
Particle Into Liquid Sampler
Particle Into Liquid Sampler-Ion Chromatography
Particle Induced X-ray Emission
protein kinase A
partial least squares, projection to latent structures
particulate matter
particulate matter of a specific size range. X refers to the diameter at
which the sampler collects 50% of the particles and rejects 50% of the
particles. Collection efficiency increases for particles with smaller
diameters and decreases for particles with larger diameters. The
variation of collection efficiency with size is given by a collection
efficiency curve. The definition of PMx is frequently abbreviated as
"particles with a nominal mean aerodynamic diameter less than or
equal to x |im.
particulate matter with a nominal mean diameter greater than x |im
and less than y |im where x and y are the numeric mean aerodynamic
or mobility diameters (|im).
particulate matter with a nominal mean mobility diameter less than or
equal to 0.1 |im (referred to as ultrafine PM)
particulate matter with a nominal mean aerodynamic diameter less
than or equal to 2.5 |im (referred to as fine PM)
particulate matter with a nominal mean aerodynamic diameter less
than or equal to 10 |im
particulate matter with a nominal mean aerodynamic diameter greater
than 2.5 [im and less than or equal to 10 [im (referred to as thoracic
coarse particulate matter or the course fraction of PM10)
Concentration may be measured with a dichotomous sampler or
calculated as the difference between measured PMi0 and measured
PM25 concentrations.
phorbol 12-myristate 13-acetate
particulate matter filtrate, positive matrix factorization
December 2009
Ixxvii
-------
PM-HD
PM-LD
PMN
PN
PNC
PND, pnd
PNMD
PNN
pNNSO
PNNL
pNO3
POA
POC
POLDER
POM
POP
PP
PP
ppb
PPFL
ppm
ppt
PRB
PRE
PRELC
PRIDE
PS
PSAS
PSO
pS04
PSS
PSU
PT
PIT
participate matter at high concentration
participate matter at low concentration
polymorphonuclear leukocytes
particle number
particle number concentration, particle number count
post-natal day
particle number median diameter
proportion of interval differences of successive normal-beat intervals
inEKG
proportion of interval differences of successive normal-beat intervals
greater than 50 ms in an EKG
Pacific Northwest National Laboratory
particulate nitrate
primary organic aerosol
particulate organic carbon
POLarization and Directionality of the Earth's Reflectance
particulate organic matter
persistent organic pollutant
particle density
pulse pressure
parts per billion
percent predicted lung function
parts per million
parts per trillion
policy-relevant background
AeroCom Experiment
Primary Rat Epithelial Lung Cells
Puerto Rico Dust Experiment
public school
The French National Program on Air Pollution Health Effects
Public Service Company of Oklahoma
particulate sulfate
physiologic saline solution
Pennsylvania State University
prothrombin time
partial thomboplastin time
December 2009
Ixxviii
-------
PTV programmable temperature vaporization
PVD peripheral vascular disease
Pyr pyrene
<2 cardiac output
Q coronary flow of the heart
QAI Q A interval
QBQ backup quartz-fiber filter behind a quartz-fiber filter
QEEG quantitative electroencephalography
Qext the extinction coefficient (a function of particle size distribution and
refractive index)
r correlation coefficient
R2 coefficient of determination
RAIN Regional Aerosol Intensive Network
RAMS real-time total ambient mass sampler
RANTES regulated upon activation, normal T cell expressed and secreted
RAPS/RAMS Regional Air Pollution Study / Regional Air Monitoring Study
RAR rapidly activating receptor(s)
RASMC rat aortic smooth muscle cells
RAW 264.7 mouse macrophage cell line
RBC red blood cell
RD respiratory disease
REALM Regional East Atmospheric Lidar Mesonet
RF radiative forcing(s)
reff particle effective radius
RFL Fetal Lung Fibroblasts
RH relative humidity
RHMVE rat heart micro-vessel endothelial cell
RHR Regional Haze Rule
RLF rat lung fibroblasts
RMC rat cardiomyocyte(s)
RME rapeseed oil methyl ester
RMSSD root mean squared differences of successive normal-beat to normal-
beat (NN or RR) time intervals between each QRS complex in the
EKG
RMV respiratory minute volume
RNA ribonucleic acid
December 2009
Ixxix
-------
RNS
RO
ROCK
ROFA
ROFA-L
ROI
ROS
RPO
RR
RS
RSV
RTI
RTM
RTF
RV
RVCFB
RVCM
a
la
°g
s
S
SAB
SAFARI
SAGE
SALIA
SAM
SAMUM
SAP2.3
Sb
SB
SBL
SBP
reactive nitrogen species
residual oil
rho associated kinase
residual oil fly ash (particles)
residual oil fly ash leachate
reactive oxygen intermediates
reactive oxygen species
Regional Planning Organizations
risk ratio, relative risk, normal-to-normal (NN or RR) time interval
between each QRS complex in the EKG
resuspended soil
respiratory syncytial virus
respiratory tract infection
Radiative Transfer Model
Research Triangle Park, North Carolina
right ventricular
right ventricular cardio fibroblasts
right ventricular cardiomyopathy, rat ventricular cardiomyocytes,
reduced volume culture medium
sigma, standard deviation
one sigma; one standard deviation
sigma-g; geometric standard deviation
second
sulfur
(EPA) Science Advisory Board
South African Fire-Atmosphere Research Initiative
Stratospheric Aerosol and Gas Experiment
German study on the Influence of Air Pollution on Lung Function,
Inflammation, and Aging
Stratospheric Aerosol Measurement
Saharan Mineral Dust Experiment
Synthesis and Assessment Product 2.3
antimony
strand breaks
stable boundary layer
systolic blood pressure
December 2009
Ixxx
-------
Sc scandium
SC summer curbside particles
SCAB California South Coast Air Basin
SCAR Smoke/Sulfates, Clouds and Radiation
SCARPOL Swiss Study on Childhood Allergy and Respiratory Symptoms with
Respect to Air Pollution
sCD40L soluble CD40 ligand
SCE sister chromatid exchange
SCS Harvard Six Cities Study
SD standard deviation; Sprague-Dawley rat
SDANN5 standard deviation of the average of normal to normal (N:N) intervals
in all 5-min intervals in a 24-h period
SDNN standard deviation normal-to-normal (NN or RR) time interval
between each QRS complex in the EKG
SDNN24HR standard deviation of the average of all normal to normal intervals in
a 24-h period
Se selenium
se standard error
SEARCH Southeastern Aerosol Research and Characterization
sem standard error of mean
SEM scanning electron microscopy
SES socioeconomic status, sample equilibration system
SF-1 steroidogenic factor -1
SF-UFID suspension, particle free ultrafine industrial exhaust
SGA small for gestational age
sGC soluble guanylate cyclase
SGP Southern Great Plains
-SH sulfhydryl group
SH Mount Saint Helen's ash
SH, SHR spontaneously hypertensive disease model rat
SHADE Saharan Dust Experiment
SHEDS Stochastic Human Exposure and Dose Simulation model
Si silicon
sIC AM-1 soluble intercellular adhesion molecule
SIDS sudden infant death syndrome
SiO2 silicone dioxide
SIPS State Implementation Plan
December 2009
Ixxxi
-------
SJV San Joaquin Valley
SLAMS State and Local Air Monitoring Stations
SME soybean oil methyl ester
SMOCC Smoke Aerosols, Clouds, Rainfall and Climate
SMOKE Spare-Matrix Operator Kernel Emissions system
SMPS scanning mobility particle sizer
SMPS-APS scanning mobility particle sizer- aerodynamic particle sizer
SMRA small mesenteric rat arteries
SNP single-nucleotide polymorphism, sodium nitroprusside
SNS sympathetic nervous system
SO2 sulfur dioxide
SO3 sulfur trioxide
SO42~ sulfate
SOA secondary organic aerosol
SOC semi-volatile organic compound
SOD superoxide dismutase
SOPHIA Study of Particulates and Health in Atlanta
SOX sulfur oxides, oxides of sulfur
SP surfactant protein (e.g., SPA, SPD)
SPA surfactant protein A
SPD surfactant protein D
SPEW Speciated Pollutant Emission Wizard
SPG Southern Great Plains site
SPM suspended particulate matter
SPPJNTARS Spectral Radiation-Transport Model for Aerosol Species
SRM-154b NIST standard reference material 154b; (TiO2 Titanium dioxide)
SRM1648 NIST standard reference material 1648; (urban particulate matter)
SRM-1649 NIST standard reference material 1649 (Washington, D.C. urban air
particulate matter, urban dust)
SRM-1650 NIST standard reference material 1650 (diesel exhaust particulate
matter)
SRM-1879 NIST standard reference material 1859; (silicone dioxide, respirable
cristobalite [respirable crystalline silica])
SRM-2975 NIST standard reference material 2975 (diesel exhaust particulate
matter)
s-ROFA soluble portion of residual oil fly ash
SS secondary sulfate, sea salt
December 2009
Ixxxii
-------
SSA
SSR
SST
STEM
STN
STP
STZ
SUB
SURFRAD
SVA
sVCAM-1
SVEB
SWNT
SXRF
SZA
T
T
TAR
TARC
TARFOX
TAT
TB
TEA
TBAP
TEARS
TBQ
Tc
Tc-DMTA
Tc-DTPA
Too
TD
TD-GC/MS
TEAC
TEOM
TexAQS
99m.
99m.
99M.
single-scattering albedo
standardized sex ratio
sea surface temperature
Sulfur / Sulfate Transport Eulerian Model
EPA Speciation Trend Network
standard temperature and pressure
Streptozotocin
summer urban background particles
NOAA GMD Surface Radiation network
supraventricular arrhythmia
soluble vascular adhesion molecule 1
supraventricular ectopic beats
singlewalled nanotube
Synchroton X-ray fluorescence
solar zenith angle
photochemical lifetime
body temperature
IPCC 3rd Assessment Report
thymus and activation-regulated chemokine
Tropospheric Aerosol Radiative Forcing Observational Experiment
thrombin-anti-thrombin complexes
tracheobronchial
thiobarbituric acid
tetrakis(4-benzoic acid) porphyrin
thiobarbituric acid reactive substances
backup quartz-fiber filter behind a Teflon-membrane filter
Technetium-99m
99MTc Dynamic mechanical thermal analysis
99MTc-diethylenetriaminepentaacetic acid
core temperature
thermal desorption, tire debris extracted in methanol
thermal desorption-gas chromatography/mass spectrometry
Trolox Equivalent Antioxidant Capacity assay
Tapered Element Oscillating Microbalance
Texas Air Quality Field Study
December 2009
Ixxxiii
-------
TF
TFPI
Tg
TG
TGF
TGFP
Th
Thl
Th2
tHcy
Ti
TIA
TiFe
TIMP-2
Ti02
TK
TM
TM5
TMTU
TNF-a
TOA
TOP-SIMS
TOMS
TOT/GC
TOYS
tPA, t-PA
TRACE
TRP
TRPV1
TR-XRF
TSA
TSP
TSS
TVOC
TWP
tissue factor
tissue factor pathway inhibitor
teragram
terminal ganglion (neurons)
transforming growth factor
P transforming growth factor
thorium
T helper cell type 1
T helper cell type 2
total homocysteine
titanium
transient ischemic attack
iron-loaded fine titanium oxide
tissue inhibitor of MMP
titanium dioxide
thymidine kinase
transition metals
Thematic Mapper, a sensor on LandsatS satellite
tetramethylthiourea
tumor necrosis factor alpha
top of the atmosphere
time-of-flight - secondary ion mass spectrometry
Total Ozone Mapping Spectrometer
thermal optical transmission analyzer coupled with gas
chromatography
TIROS-N Operational Vertical Sounder
tissue plasminogen activator
Transition Region and Coronal Explorer
transient receptor potential
transient receptor potential vanilloid-1 receptor
total reflection X-ray fluorescence
tricho statin A
total suspended particulate
WRAP Technical Support System website
total VOC
Tropical West Pacific island
December 2009
Ixxxiv
-------
TXB2
U
UACR
UAE2
UAP
UF
UFAA
HFC
UfCB
UFDG
UFID
UFP
UFPM
UFTiO2
UIO
U.K.
UKMO
ULAQ
ULTRA
UMI
UNEP
UP
UPM
UPSP
URI
URS
U.S.
U.S.C.
UV
V
V, mV, \i
VAQ
VCAM-l
Vd
VEAPS
thromboxane B-2
uranium
urinary albumin / creatinine ratio
United Arab Emirates Unified Aerosol Experiment
urban ambient particle
ultrafine, uncertainty factor
ultrafine ambient air
ultrafine carbon
ultrafine carbon black
ultrafine diesel engine exhaust
ultrafine industrial exhaust
ultrafine particle
ultrafine particulate matter
ultrafine titanium dioxide
University of Oslo
United Kingdom
United Kingdom Meteorological Office
University of IL'Aquila.
Exposure and Risk Assessment for Fine and Ultrafine Particles in
Ambient Air
University of Michigan
United Nations Environmental Programme
urban particle
ultrafine particulate matter
unmodified polystyrene particle(s)
upper respiratory infection
upper respiratory symptoms
United States of America
U.S. Code
ultraviolet radiation
vanadium
volt, millivolt, microvolt
visual air quality
vascular adhesion molecule 1
deposition velocity
Vitamin E Atherosclerosis Progression Study
December 2009
Ixxxv
-------
VEGF vascular endothelial growth factor
VIEWS Visibility Information Exchange Web Site
VISTAS Visibility Improvement State and Tribal Association of the Southeast
VOC volatile organic compound
VOSO4 vanadyl sulfate
VPB ventricular premature beat
VR visual range
VR1 vanilloid receptor 1
VSCC very sharp cut cyclone
VSMC Vascular Smooth Muscle Cells
VT tidal volume
vWF von Willebrand factor
W Wilderness
WACAP Western Airborne Contaminates Assessment Project
WBC white blood cell(s)
WC winter curbside particles
WHI Women's Health Initiative
WHI OS Women's Health Initiative Observational Study
wk week(s)
WKY Wistar-Kyoto rat strain
W/m2, W m"2 watts per square meter
WMO World Meteorological Organization
Wnt wingless gene family
WRAP Western Regional Air Partnership
WRF Weather Research and Forecasting model
WS wood smoke
WSOC water soluble organic carbon
WUB winter urban background particles
XAD polystyrene-divinyl benzene
XPS X-ray photoelectron spectroscopy
Y yttrium
yr year
Z radar reflectivity (measured in dBZ [decibels of Z, where Z represents
the energy reflected back to the radar.])
Zn zinc
ZnO zinc oxide
December 2009
Ixxxvi
-------
ZnS zinc sulfide
ZnSO4 zinc sulfate
Zr zirconium
December 2009 Ixxxvii
-------
Chapter 1. Introduction
This Integrated Science Assessment (ISA) is a review, synthesis, and evaluation of the most
policy-relevant evidence, and communicates critical science judgments relevant to the National
Ambient Air Quality Standards (NAAQS) review. As such, the ISA forms the scientific foundation
for the review of the primary (health-based) and secondary (welfare-based) NAAQS for particulate
matter (PM). The ISA accurately reflects "the latest scientific knowledge useful in indicating the
kind and extent of identifiable effects on public health which may be expected from the presence of
[a] pollutant in ambient air" (42 U.S.C. 7408). Key information and judgments formerly contained in
an Air Quality Criteria Document (AQCD) for PM are incorporated in this assessment. Additional
details of the pertinent literature published since the last review, as well as selected older studies of
particular interest, are included in a series of annexes. This ISA thus serves to update and revise the
evaluation of the scientific evidence available at the time of the previous review of the NAAQS for
PM that was concluded in 2006.
The Integrated Review Plan for the National Ambient Air Quality Standards for Paniculate
Matter identifies a series of policy-relevant questions that provide a framework for this assessment
of the scientific evidence (U.S. EPA, 2008, 157072). These questions frame the entire review of the
NAAQS for PM, and thus are informed by both science and policy considerations. The ISA
organizes and presents the scientific evidence such that, when considered along with findings from
risk analyses and policy considerations, will help the EPA address these questions during the
NAAQS review for PM. In evaluating the health evidence, the focus of this assessment will be on
scientific evidence that is most relevant to the following questions that have been taken directly from
the Integrated Review Plan:
• Has new information altered the body of scientific support for the occurrence of health
effects following short- and/or long-term exposure to levels of fine and thoracic coarse
particles found in the ambient air?
• Has new information altered conclusions from previous reviews regarding the
plausibility of adverse health effects associated with exposures to PM2.5, PMi0, PMi0_2.5,
or alternative PM indicators that might be considered?
• What evidence is available from recent studies focused on specific size fractions,
chemical components, sources, or environments (e.g., urban and non-urban areas) of PM
to inform our understanding of the nature of PM exposures that are linked to various
health outcomes?
• To what extent is key scientific evidence becoming available to improve our
understanding of the health effects associated with various time periods of PM
exposures, including not only short-term (daily or multi-day) and chronic (months to
years) exposures, but also peak PM exposures (<24 hours)? To what extent is critical
research becoming available that could improve our understanding of the relationship
between various health endpoints and different lag periods (e.g., <1 day, single day,
multi-day distributed lags)?
• What data are available to improve our understanding of spatial and/or temporal
heterogeneity of PM exposures considering different size fractions and/or components?
• At what levels of PM exposure do health effects of concern occur? Is there evidence for
the occurrence of adverse health effects at levels of PM lower than those observed
previously? If so, at what levels and what are the important uncertainties associated with
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009 1-1
-------
that evidence? What is the nature of the dose-response relationships of PM for the
various health effects evaluated?
• What evidence is available linking particle number concentration with adverse health
effects of UF particles?
• Do risk/exposure estimates suggest that exposures of concern for PM-induced health
effects will occur with current ambient levels of PM or with levels that just meet the
current standards? If so, are these risks/exposures of sufficient magnitude such that the
health effects might reasonably be judged to be important from a public health
perspective? What are the important uncertainties associated with these risk/exposure
estimates?
• To what extent is key evidence becoming available that could inform our understanding
of subpopulations that are particularly sensitive or vulnerable to PM exposures? In the
last review, sensitive or vulnerable subpopulations that appeared to be at greater risk for
PM-related effects included individuals with pre-existing heart and lung diseases, older
adults, and children. Has new evidence become available to suggest additional sensitive
subpopulations should be given increased focus in this review (e.g., fetuses, neonates,
genetically susceptible subpopulations)?
• To what extent is key evidence becoming available to inform our understanding of
populations that are particularly vulnerable to PM exposures? Specifically, is there new
or emerging evidence to inform our understanding of geographical, spatial, SES, and
environmental justice considerations?
• To what extent have important uncertainties identified in the last review been reduced
and/or have new uncertainties emerged?
• To what extent is new information available to inform our understanding of non-PM-
exposure factors that might influence the associations between PM levels and health
effects being considered (e.g., weather-related factors; behavioral factors such as
heating/air conditioning use; driving patterns; and time-activity patterns)?
In evaluating evidence on welfare effects of PM, the focus will be on evidence that can help
inform these questions from the Integrated Review Plan:
• What new evidence is available on the relationship between PM mass/size fraction
and/or specific PM components and visibility impairment and climate-related and other
welfare effects?
• To what extent has key scientific evidence now become available to improve our
understanding of the nature and magnitude of visibility, climate, and ecosystem
responses to PM and the variability associated with those responses (including ecosystem
type, climatic conditions, environmental effects and interactions with other
environmental factors and pollutants)?
• Do the evidence, the air quality assessment, and the risk/exposure assessment provide
support for considering alternative averaging times?
• At what levels of ambient PM do visibility impairment and/or environmental effects of
concern occur? Is there evidence for the occurrence of adverse visibility and other
welfare-related effects at levels of PM lower than those observed previously? If so, at
what levels and what are the important uncertainties associated with the evidence?
• Do the analyses suggest that PM-induced visibility impairment and/or other welfare-
effects will occur with current ambient levels of PM or with levels that just meet the
current standards? If so, are these effects of sufficient magnitude and/or frequency such
December 2009 1-2
-------
that these effects might reasonably be judged to be important from a public welfare
perspective? What are the uncertainties associated with these estimates?
What new evidence and/or techniques are available to quantify the benefits of improved
visibility and/or other welfare-related effects?
To what extent have important uncertainties identified in the last review been reduced
and/or have new uncertainties emerged?
1.1. Legislative Requirements
Two sections of the United States (U.S.) Clean Air Act (CAA, the Act) govern the
establishment and revision of the NAAQS. Section 108 of the Act (42 U.S.C. 7408) directs the
Administrator to identify and list "air pollutants" that "in his judgment, may reasonably be
anticipated to endanger public health and welfare" and whose "presence... in the ambient air results
from numerous or diverse mobile or stationary sources" and to issue air quality criteria for those that
are listed (42 U.S.C. 7408). Air quality criteria are intended to "accurately reflect the latest scientific
knowledge useful in indicating the kind and extent of identifiable effects on public health or welfare
which may be expected from the presence of [a] pollutant in ambient air..." 42 U.S.C. 7408(b).
Section 109 of the Act (42 U.S.C. 7409) directs the Administrator to propose and promulgate
"primary" and "secondary" NAAQS for pollutants listed under Section 108. 42 U.S.C. 7409(a).
Section 109(b)(l) defines a primary standard as one "the attainment and maintenance of which in the
judgment of the Administrator, based on such criteria and allowing an adequate margin of safety, are
requisite to protect the public health."! 42 U.S.C. 7409(b)(l). A secondary standard, as defined in
Section 109(b)(2), must "specify a level of air quality the attainment and maintenance of which, in
the judgment of the Administrator, based on such criteria, is required to protect the public welfare
from any known or anticipated adverse effects associated with the presence of [the] pollutant in the
ambient air."2 42 U.S.C. 7409(b)(2).
The requirement that primary standards include an adequate margin of safety was intended to
address uncertainties associated with inconclusive scientific and technical information available at
the time of standard setting. It was also intended to provide a reasonable degree of protection against
hazards that research has not yet identified. See Lead Industries Association v. EPA, 647 F.2d 1130,
1154 (D.C. Cir 1980), cert, denied, 449 U.S. 1042 (\9%Q); American Petroleum Institute v. Costle,
665 F.2d 1176, 1186 (D.C. Cir. 1981), cert, denied, 455 U.S. 1034 (1982); American Farm Bureau
Federation v. EPA, 559 F. 3d 512, 533 (D.C. Cir. 2009). Both kinds of uncertainties are components
of the risk associated with pollution at levels below those at which human health effects can be said
to occur with reasonable scientific certainty. Thus, in selecting primary standards that include an
adequate margin of safety, the Administrator is seeking not only to prevent pollution levels that have
been demonstrated to be harmful, but also to prevent lower pollutant levels that may pose an
unacceptable risk of harm, even if the risk is not precisely identified as to nature or degree.
In selecting a margin of safety, the EPA considers such factors as the nature and severity of the
health effects involved, the size of the sensitive population(s) at risk, and the kind and degree of the
uncertainties that must be addressed. The selection of any particular approach to providing an
adequate margin of safety is a policy choice left specifically to the Administrator's judgment. See
Lead Industries Association v. EPA, supra, 647 F.2d 1161-62.
In setting standards that are "requisite" to protect public health and welfare, as provided in
Section 109(b), the Administrator's task is to establish standards that are neither more nor less
1 The legislative history of Section 109 indicates that a primary standard is to be set at "the maximum permissible ambient air
level... which will protect the health of any [sensitive] group of the population," and that for this purpose "reference should be made to a
representative sample of persons comprising the sensitive group rather than to a single person in such a group" [S. Rep. No. 91-1196,
91st Cong., 2d Sess. 10 (1970)].
2 Welfare effects as defined in Section 302(h) [42 U.S.C. 7602(h)] include, but are not limited to, "effects on soils, water, crops, vegetation,
man-made materials, animals, wildlife, weather, visibility and climate, damage to and deterioration of property, and hazards to
transportation, as well as effects on economic values and on personal comfort and well-being."
December 2009 1-3
-------
stringent than necessary. In so doing, EPA may not consider the costs of implementing the standards.
See generally Whitman v. American Trucking Associations, 531 U.S. 457, 465-472, 475-76 (2001).
Section 109(d)(l) requires that "not later than December 31, 1980, and at 5-yr
intervals thereafter, the Administrator shall complete a thorough review of the criteria
published under Section 108 and the national ambient air quality standards... and shall make such
revisions in such criteria and standards and promulgate such new standards as may be
appropriate... " 42 U.S.C. 7409(d)(l). Section 109(d)(2) requires that an independent scientific
review... committee "shall complete a review of the criteria and the national primary and secondary
ambient air quality standards... and shall recommend to the Administrator any new standards and
revisions of existing criteria and standards as may be appropriate..." 42 U.S.C. 7409(d)(2). Since the
early 1980s, this independent review function has been performed by the Clean Air Scientific
Advisory Committee (CASAC).
1.2. History of Reviews of the NAAQS for PM
PM is the generic term for a broad class of chemically and physically diverse substances that
exist as discrete particles (liquid droplets or solids) over a wide range of sizes. Particles originate
from a variety of anthropogenic stationary and mobile sources, as well as from natural sources.
Particles may be emitted directly or formed in the atmosphere by transformations of gaseous
emissions such as sulfur oxides (SOX), nitrogen oxides (NOX), and volatile organic compounds
(VOC). The chemical and physical properties of PM vary greatly with time, region, meteorology,
and source category, thus complicating the assessment of health and welfare effects. Table 1-1
summarizes the NAAQS that have been promulgated for PM to date. These reviews are briefly
described below, and further details are provided in the Integrated Review Plan (U.S. EPA, 2008,
157072).
EPA first established NAAQS for PM in 1971 (36 FR 8186, April 30, 1971), based on the
original criteria document (NAPCA, 1969, 014684). The reference method specified for determining
attainment of the original standards was the high-volume sampler, which collects PM up to a
nominal size of 25-45 micrometers ((irn) (referred to as total suspended particulates [TSP]). The
primary standards (measured by the indicator TSP) were 260 (ig/m3, 24-h avg, not to be exceeded
more than once per year, and 75 (ig/m3, annual geometric mean. The secondary standard was
150 (ig/m3, 24-h avg, not to be exceeded more than once per year. In October 1979 (44 FR 56730,
October 2, 1979), EPA announced the first periodic review of the air quality criteria and NAAQS for
PM, and significant revisions to the original standards were promulgated in 1987 (52 FR 24634,
July 1, 1987). In that decision, EPA changed the indicator for particles from TSP to PMi0, the latter
including particles with a mean aerodynamic diameter1 < 10 (im, which delineated that subset of
inhalable particles small enough to penetrate to the thoracic region (including the tracheobronchial
and alveolar regions) of the respiratory tract (referred to as thoracic particles). EPA also revised the
level and form of the primary standards by (1) replacing the 24-h TSP standard with a 24-h PMi0
standard of 150 (ig/m with no more than one expected exceedence per year; and (2) replacing the
annual TSP standard with a PMi0 standard of 50 (ig/m3, annual arithmetic mean, averaged over 3 yr.
1 The more precise term is 50% cut point or 50% diameter (d50). This is the aerodynamic particle diameter for which the efficiency of
particle collection is 50%. Larger particles are not excluded altogether, but are collected with substantially decreasing efficiency and
smaller particles are collected with increasing (up to 100%) efficiency.
December 2009 1-4
-------
Table 1-1. Summary of NAAQS promulgated for PM, 1971-2006.
Year (Final Rule) Indicator
TSP (Total
1971 (36 FR 81 86) Suspended
Particulates)
1987 (52 FR 24634) PM10
PM25
1997 (62 FR 38652)
Avg Time Level
94 h 260 pg/m3 (primary)
^4n 1 50 pg/m3 (secondary)
Annual 75 pg/m3 (primary)
24 h 150 pg/m3
Annual 50 pg/m3
24 h 65 pg/m3
Annual 15 pg/m3
o/i h -icn i i,wm3
Form
Not to be exceeded more than once per yr
Annual geometric mean
Not to be exceeded more than once per yr on average over a 3-yr period
Annual arithmetic mean, averaged over 3 yr
98th percentile, averaged over 3 yr
Annual arithmetic mean, averaged over 3 yr1
Initially promulgated 99th percentile, averaged over 3 yr; when 1997 standards
PM,r
24 h
150 pg/m
were vacated in 1999, the form of 1987 standards remained in place (not to be
exceeded more than once per yr on average over a 3-yr period)
Annual 50 pg/m
Annual arithmetic mean, averaged over 3 yr
2006 (71 FR 61144)
PM25
24 h
98th percentile, averaged over 3 yr
Annual
15 pg/rri
Annual arithmetic mean, averaged over 3 yr
PM,,
24 h
150 pg/m3
Not to be exceeded more than once per yr on average over a 3-yr period
Note: When not specified, primary and secondary standards are identical.
The secondary standard was revised by replacing it with 24-h and annual standards identical in
all respects to the primary standards. The revisions also included a new reference method for the
measurement of PMi0 in the ambient air and rules for determining attainment of the new standards.
On judicial review, the revised standards were upheld in all respects. See Natural Resources Defense
Council v. Administrator, 902 F. 2d 962 (D.C. Cir. 1990), cert, denied, 498 U.S. 1082 (1991).
In April 1994, EPA announced its plans for the second periodic review of the air quality
criteria and NAAQS for PM, and promulgated significant revisions to the NAAQS in 1997 (62 FR
38652, July 18, 1997). In that decision, EPA revised the PM NAAQS in several respects. Most
significantly, EPA determined that the fine and coarse3 fractions of PM10 should be considered
separately. The Administrator's decision to modify the standards was based on evidence that serious
health effects were associated with short- and long-term exposure to fine particles in areas that met
the existing PMi0 standards. EPA accordingly added new standards, using PM2.5 as the indicator for
fine particles (with PM2.5 referring to particles with a nominal mean aerodynamic diameter
< 2.5 (im), and PMi0 as the indicator for thoracic coarse particles or coarse-fraction particles
(generally including particles with a nominal mean aerodynamic diameter >2.5 (im and < 10 (im, or
PMio_2.5). The EPA established two new PM2.5 standards: an annual standard of 15 (ig/m3, based on
the 3-yr avg of annual arithmetic mean PM25 concentrations from single or multiple
community-oriented monitors; and a 24-h standard of 65 (ig/m3, based on the 3-yr avg of the 98th
percentile of 24-h PM2.5 concentrations at each population-oriented monitor within an area. Also,
EPA established a new reference method for measuring PM2 5 in the ambient air and adopted
protocols for determining attainment of the new standards. To continue to address thoracic coarse
particles, EPA retained the annual PMi0 standard, while revising the form, but not the level, of the
1 The level of the 1997 annual PM2.5 standard was to be compared to measurements made at the community-oriented monitoring site
recording the highest level, or, if specific constraints were met, measurements from multiple community-oriented monitoring sites could
be averaged ("spatial averaging"). This approach was judged to be consistent with the short-term epidemiologic studies on which the
annual PM2.5 standard was primarily based, in which air quality data were generally averaged across multiple monitors in an area or were
taken from a single monitor that was selected to represent community-wide exposures, not localized "hot spots" (62 FR 38672). These
criteria and constraints were intended to ensure that spatial averaging would not result in inequities in the level of protection afforded by
the PM2.5 standards. Community-oriented monitoring sites were specified to be consistent with the intent that a spatially averaged annual
standard provide protection for persons living in smaller communities, as well as those in larger population centers.
In the revisions to the PM NAAQS finalized in 2006, EPA tightened the constraints on the spatial averaging criteria by further limiting
the conditions under which some areas may average measurements from multiple community-oriented monitors to determine compliance
(71 FR 61165-61167, October 17, 2006).
3 See definitions of "fine" and "coarse" particles in Section 3.2.
December 2009
1-5
-------
24-h PM10 standard to be based on the 99th percentile of 24-h PM10 concentrations at each monitor
in an area. The EPA revised the secondary standards by making them identical in all respects to the
primary standards.
Following promulgation of the 1997 PM NAAQS, petitions for review were filed by a large
number of parties, addressing a broad range of issues. In May 1999, a three-judge panel of the U.S.
Court of Appeals for the District of Columbia Circuit issued an initial decision that upheld EPA's
decision to establish fine particle standards, holding that "the growing empirical evidence
demonstrating a relationship between fine particle pollution and adverse health effects amply
justifies establishment of new fine particle standards. "American Trucking Associations v. EPA (175
F. 3d 1027, 1055-56 (D.C. Cir. 1999); rehearing granted in part and denied in part, 195 F. 3d 4 (D.C.
Cir. 1999), affirmed in part and reversed in part, Whitman v. American Trucking Associations 531
U.S. 457 (2001). The panel also found "ample support" for EPA's decision to regulate coarse particle
pollution, but vacated the 1997 PMi0 standards, concluding that EPA had not provided a reasonable
explanation justifying use of PMi0 as an indicator for coarse particles (175 F. 3d at 1054-55).
Pursuant to the court's decision, EPA removed the vacated 1997 PMi0 standards from the Code of
Federal Regulations. The pre-existing 1987 PMi0 standards remained in place (65 FR 80776,
December 22, 2000). The Court also upheld EPA's determination not to establish more stringent
secondary standards for fine particles to address effects on visibility (175 F. 3d at 1027).
More generally, the panel held (over one judge's dissent) that EPA's approach to establishing
the level of the standards in 1997, both for the PM and ozone (O3) NAAQS promulgated on the same
day, effected "an unconstitutional delegation of legislative authority" (Id. at 1034-40). Although the
panel stated that "the factors EPA uses in determining the degree of public health concern associated
with different levels of ozone and PM are reasonable," it remanded the rule to EPA, stating that when
EPA considers these factors for potential non-threshold pollutants "what EPA lacks is any
determinate criterion for drawing lines" to determine where the standards should be set. Consistent
with EPA's long-standing interpretation and D.C. Circuit precedent, the panel also reaffirmed its
prior holdings that in setting NAAQS EPA is "not permitted to consider the cost of implementing
those standards" (Id. at 1040-41).
On EPA's petition for rehearing, the panel adhered to its position on these points. American
Trucking Associations v. EPA, 195 F. 3d 4 (D.C. Cir. 1999). The full Court of Appeals denied EPA's
suggestion for rehearing en bane, with five judges dissenting (Id. at 13).
Both sides filed cross appeals on these issues to the U.S. Supreme Court, and the Court
granted certiorari. In February 2001, the Supreme Court issued a unanimous decision upholding
EPA's position on both the constitutional and cost issues. Whitman v. American Trucking
Associations, 531 U.S. 457, 464, 475-76. On the constitutional issue, the Court held that the statutory
requirement that NAAQS be "requisite" to protect public health with an adequate margin of safety
sufficiently guided EPA's discretion, affirming EPA's approach of setting standards that are neither
more nor less stringent than necessary. The Supreme Court remanded the case to the Court of
Appeals for resolution of any remaining issues that had not been addressed in that court's earlier
rulings (Id. at 475-76). In March 2002, the Court of Appeals rejected all remaining challenges to the
standards, holding under the traditional standard of judicial review that PM2.5 standards were
reasonably supported by the administrative record and were not "arbitrary and capricious" American
Trucking Associations v. EPA, 283 F. 3d 355, 369-72 (D.C. Cir. 2002).
In October 1997, EPA published its plans for the third periodic review of the air quality criteria
and NAAQS for PM (62 FR 55201). After CASAC and public review, EPA finalized the 2004 PM
AQCD (U.S. EPA, 2004, 056905) and 2005 Staff Paper (U.S. EPA, 2005, 090209). For the primary
fine particle standards, most CASAC PM Panel members favored the option of revising the level of
the 24-h PM2.5 standard in the range of 35 to 30 (ig/m3 with a 98th percentile form, in concert with
revising the level of the annual PM2.5 standard in the range of 14 to 13 (ig/m3 (Henderson, 2005,
188316). Most of the members of the CASAC PM Panel also strongly supported establishing anew,
secondary PM2 5 standard to protect urban visibility and recommended establishing a sub-daily (4- to
8-h averaging time) PM25 standard within the range of 20 to 30 ug/m3 with a form within the range
of the 92nd to 98th percentile (Henderson, 2005, 188316). For thoracic coarse particles, there was
general concurrence among CASAC PM Panel members to revise the PMi0 standards by establishing
a primary standard specifically targeted to address particles in the size range of 2.5 to 10 um
(PMio_2.5). The CASAC PM Panel was also in general agreement "that coarse particles in urban or
industrial areas are likely to be enriched by anthropogenic pollutants that tend to be inherently more
toxic than the windblown crustal material which typically dominates coarse particle mass in arid
December 2009 1-6
-------
rural areas." Based on its review of the Staff Paper, there was general agreement among the CASAC
PM Panel members that a 24-h PMi0_2.5 standard with a level in the range of 50 to 70 (ig/m3, with a
98th percentile form, was reasonably justified and that a PMi0_2.5 standard with an annual averaging
time was not warranted (Henderson, 2005, 156537). On January 17, 2006, EPA proposed to revise
the NAAQS for PM (71 FR 2620). For fine particles, EPA proposed to retain PM2.5 as the indicator,
to retain standards for 24-h and annual exposures, and to revise the form of the annual standard to
tighten conditions for demonstrating compliance using spatially averaged monitoring. EPA also
proposed to revise the level of the 24-h PM25 standard to 35 (ig/m3 to provide increased protection
against health effects associated with short-term PM2 5 exposures, including premature mortality and
increased hospital admission and emergency room visits, but proposed to retain the level of the
annual PM2.5 standard at 15 (ig/m3, continuing protection against health effects associated with
long-term exposure including premature mortality and development of chronic respiratory disease.
With regard to the primary standards for thoracic coarse particles, EPA proposed to revise the 24-h
PMio standard in part by establishing a new indicator for thoracic coarse particles (particles generally
between 2.5 and 10 (im in diameter), qualified so as to include any ambient mix of PMi0_2.5 that was
dominated by resuspended dust from high density traffic on paved roads and PM generated by
industrial sources and construction sources, and proposed to exclude any ambient mix of PMi0_2.5
that was dominated by rural windblown dust and soils and PMi0_2.s generated by agricultural and
mining sources. EPA also proposed a detailed monitoring regime in conjunction with this proposed
indicator (71 FR 2710, 2731-42). The EPA proposed to set a 24-h standard (using the proposed
indicator) at a level of 70 (ig/m to continue to provide a level of protection against health effects
associated with short-term exposure (including hospital admissions for cardiopulmonary diseases,
increased respiratory symptoms and possibly premature mortality) in those areas where the proposed
indicator was found, generally equivalent to the level of protection provided by the existing 24-h
PMio standard. Also, EPA proposed to revoke, upon finalization of a primary 24-h standard for
thoracic coarse particles, the 24-h PMio standard as well as the annual PMio standard.
EPA proposed to revise the secondary standards by making them identical to the suite of
proposed primary standards for fine and coarse particles, providing protection against PM-related
public welfare effects including visibility impairment, effects on vegetation and ecosystems, and
materials damage and soiling. EPA also solicited comment on adding a new sub-daily PM2 5
secondary standard to address visibility impairment in urban areas.
CASAC provided additional advice to EPA in a letter to the Administrator requesting
reconsideration of CASAC's recommendations for both the primary and secondary PM25 standards,
as well as standards for thoracic coarse particles (Henderson, 2006, 156538).
On September 21, 2006, EPA announced its final decisions to revise the primary and
secondary NAAQS for PM to provide increased protection of public health and welfare, respectively
(71 FR 61144). With regard to the primary and secondary standards for fine particles, EPA revised
the level of the 24-h PM25 standard to 35 (ig/m3, retained the level of the annual PM25 standard at
15 (ig/m3, and revised the form of the annual PM25 standard by narrowing the constraints on the
optional use of spatial averaging. EPA established the secondary standard for fine particles identical
to the primary standards. With regard to the primary and secondary standards for thoracic coarse
particles, EPA retained PMio as the indicator for coarse particles, retained the level and form of the
24-h PM10 standard (so the standard remains 150 (ig/m3 with a one expected exceedence form) and
revoked the annual standard because available evidence generally did not support a link between
long-term exposure to current ambient levels of coarse particles and health or welfare effects.
Following promulgation of the revised PM NAAQS in 2006, several parties filed petitions for
review with respect to: (1) selecting the level of the annual primary PM25 standard; (2) setting the
secondary PM2 5 standards identical to the primary standards; (3) retaining PMio as the indicator for
coarse particles and retaining the level and form of the PM10 24-h standard; and (4) revoking the
PMio annual standard. On judicial review, the D.C. Circuit remanded the annual standard for fine
particles to EPA because EPA failed to adequately explain why the annual PM2 5 standard provided
the requisite protection from both short- and long-term exposures to fine particles including
protection for vulnerable subpopulations. With respect to protection from short-term exposures, in
1997 EPA determined that the annual standard was the generally controlling standard for lowering
both short- and long-term PM2 5 concentrations and the 24-h standard was set to "provide an
adequate margin of safety against infrequent or isolated peak concentrations that could occur in areas
that attain the annual standard" (62 FR 38676-77, July 18, 1997). In the 2006 decision, the
Administrator considered it appropriate to use a somewhat different evidence-based approach from
December 2009 1-7
-------
that used in 1997 to set the level of the 24-h and annual PM2.5 standards. In that decision, the
Administrator relied upon evidence from the short-term exposure PM2.5 studies as the principal basis
for selecting the proposed level of the 24-h standard and relied upon evidence from the long-term
exposure PM2.5 studies as the principal basis for selecting the level of the annual standard. The court
found EPA failed to adequately explain this change in approach in light of CASAC and staff's
recommendations to do otherwise. The court also found that EPA had failed to adequately explain
why a short-term 24-h standard by itself would provide the protection needed from short-term
exposures. American Farm Bureau Federation v. EPA, 559 F. 3d 512, 520-24 (D.C. Cir. 2009). With
respect to protection from long-term exposure, the court found that EPA failed to adequately explain
how the current standard provided "an adequate margin of safety for vulnerable subpopulations, such
as children, the elderly, or those with conditions that exposure them to greater risk from fine
particles". Specifically, EPA did not provide a reasonable explanation of why certain studies,
including a study of children in Southern California showing lung damage from long-term exposure,
did not call for a more stringent annual standard (Id. at 522-23).
The court also remanded the secondary standard for fine particles, based on EPA's failure to
adequately explain why setting the secondary NAAQS equivalent to the primary standards provided
the required protection for public welfare including protection from visibility impairment. The court
found that EPA failed to identify a target level of visibility impairment that would be requisite to
protect public welfare. This was contrary to the statute and resulted in a lack of a reasoned basis for
the final decision. In addition, EPA's near exclusive reliance on a comparison of numbers of counties
that would be in nonattainment under various types of standards was an inadequate basis for making
a decision. It did not take into account the relative visibility protection of different standards, as well
as the failure of a 24-h standard to address regional differences in humidity and its effect on visibility
(Id. at 528-31).
The court upheld EPA's decision to retain the 24-h PMi0 standard to provide protection for
coarse particle exposures and to revoke the annual PMi0 standard. The court found that EPA
reasonably included all coarse PM within the standard, both urban and non-urban, to provide
nationwide protection for exposure to coarse PM. It rejected arguments that the evidence showed
there are no risks from exposure to non-urban coarse PM (Id. at 531-33). The court further found that
EPA had a reasonable basis to not set separate standards for urban and non-urban coarse PM, namely
the inability to reasonably define what ambient mixes would be included under either "urban" or
"non-urban." In addition, the court found that record evidence supported EPA's cautious decision to
provide "some protection from exposure to thoracic coarse particles... in all areas." The court also
upheld EPA's decision to use PMi0 as the indicator for coarse particles and to retain the level of the
standard at 150 (ig/m3. EPA's final rule acknowledged that evidence of harm from urban-type coarse
PM is stronger than for other types, and targeted protection at areas where urban-type coarse PM is
most likely present. The targeting is done by using the indicator PMi0 for coarse particles. PM10
includes both coarse PM and fine PM. Urban and industrial areas tend to have higher levels of fine
PM than rural areas, so that in those areas less coarse PM is allowed - the desired targeting.
Conversely, fine PM levels tend to be lower in rural areas, so more coarse particles are allowed in
those areas - again the desired targeting. Likewise, the court concluded that the EPA's choice of the
level for the PMi0 standard was reasonable for many of the same reasons (Id. at 533-36). The court
also upheld EPA's decision to revoke the annual PM10 standard (Id. at 537-38).
1.3. ISA Development
EPA initiated the current formal review of the NAAQS for PM on June 28, 2007 with a call for
information from the public (72 FR 35462). In addition to the call for information, publications were
identified through an ongoing literature search process that includes extensive computer database
mining on specific topics. Literature searches were conducted routinely to identify studies published
since the last review, focusing on publications from 2002 to May 2009. Search strategies were
iteratively modified in an effort to optimize the identification of pertinent publications. Additional
papers were identified for inclusion in several ways: review of pre-publication tables of contents for
journals in which relevant papers may be published; independent identification of relevant literature
by expert authors; and identification by the public and CASAC during the external review process.
Generally, only information that had undergone scientific peer review and had been published or
December 2009 1-8
-------
accepted for publication was considered. All relevant epidemiologic, controlled human exposure,
animal toxicological, and welfare effects studies published since the last review were considered,
including those related to exposure-response relationships, mode(s) of action (MOA), or susceptible
populations.
In general, in assessing the scientific quality and relevance of health and environmental effects
studies, the following considerations have been taken into account when selecting studies for
inclusion in the ISA or its annexes. The selection process for studies included in this ISA is shown in
Figure 1-1.
• Are the study populations, subjects, or animal models adequately selected and are they
sufficiently well defined to allow for meaningful comparisons between study or exposure
groups?
• Are the statistical analyses appropriate, properly performed, and properly interpreted?
Are likely covariates adequately controlled or taken into account in the study design and
statistical analysis?
• Are the PM aerometric data, exposure, or dose metrics of adequate quality and
sufficiently representative of information regarding ambient PM?
• Are the health or welfare effect measurements meaningful and reliable?
In selecting epidemiologic studies, EPA considered whether a given study contained
information on associations with short- or long-term PM exposures at or near ambient levels of PM;
evaluated health effects of PM size fractions, components or source-related indicators; considered
approaches to evaluate issues related to potential confounding by other pollutants; assessed potential
effect modifiers; and evaluated important methodological issues (e.g., lag or time period between
exposure and effects, model specifications, thresholds, mortality displacement) related to
interpretation of the health evidence. Among the epidemiologic studies selected, particular emphasis
was placed on those studies most relevant to the review of the NAAQS. Specifically, studies
conducted in the U.S. or Canada were discussed in more detail than those from other geographical
regions. Particular emphasis was placed on: (1) recent multicity studies that employ standardized
analysis methods for evaluating effects of PM and that provide overall estimates for effects based on
combined analyses of information pooled across multiple cities; (2) studies that help understand
quantitative relationships between exposure concentrations and effects; (3) recent studies (published
since the last PM NAAQS review) that provide evidence on effects in susceptible populations; and
(4) studies that consider and report PM as a component of a complex mixture of air pollutants.
December 2009 1-9
-------
Continuous,
comprehensive
literature review
of peer-reviewed
journal articles
Studies that do
not address
exposure and/or
effects of air
pollutant(s) under
review are
excluded.
Studies added
to the docket
during public
comment period.
Studies identified
during EPA
sponsored kickoff
meeting (including
studies in
preparation).
KEY DEFINITIONS
INFORMATIVE studies are well-designed,
properly implemented, thoroughly described.
HIGHLY INFORMATIVE studies reduce
uncertainty on critical issues, may include
analyses of confounding or effect modification
by copollutants or other variables, analyses of
concentration-response or dose-response
relationships, analyses related to time
between exposure and response, and offer
innovation in method or design.
POLICY-RELEVANT studies may include
those conducted at or near ambient concen-
trations and studies conducted in U.S. and
Canadian airsheds.
V J
nformative
studies
are identified
Studies are
evaluated for inclusion
in the ISA and/
or Annexes.
Selection of
studies
discussed and
additional studies
identified during
CASAC peer
review of draft
document.
Policy relevant and highly informative studies discussed in the ISA text include
those that provide a basis for or describe the association between the criteria
pollutant and effects. Studies summarized in tables and figures are included
because they are sufficiently comparable to be displayed together. A study
highlighted in the ISA text does not necessarily appear in a summary table or
figure.
ANNEXES
All newly identified informative studies are included in the Annexes. Older, key
studies included in previous assessments may be included as well.
Figure 1 -1. Identification of studies for inclusion in the ISA.
December 2009
1-10
-------
Criteria for the selection of research evaluating controlled human exposure or animal
toxicological studies included a focus on studies conducted using relevant pollutant exposures. For
both types of studies, relevant pollutant exposures are considered to be those generally within one or
two orders of magnitude of ambient PM concentrations. Studies in which higher doses were used
may also be considered if they provide information relevant to understanding MO As or mechanisms,
as noted below.
Evaluation of controlled human exposure studies focused on those that approximated expected
human exposure conditions in terms of concentration and duration. In the selection of controlled
human exposure studies, emphasis is placed on studies that: (1) investigate potentially susceptible
populations such as people with cardiovascular diseases or asthmatics, particularly studies that
compare responses in susceptible individuals with those in age-matched healthy controls; (2) address
issues such as concentration-response or time-course of responses; (3) investigate exposure to PM
separately and in combination with other pollutants such as O3; (4) include control exposures to
filtered air; and (5) have sufficient statistical power to assess findings.
For selecting toxicological studies for highlighting in the text, emphasis is placed on inhalation
studies conducted at concentrations <2 mg/m3 and those studies that approximate expected human
dose conditions in terms of concentration, size distributions, and duration, which will depend on the
toxicokinetics and biological sensitivity of the particular laboratory animals examined. Studies that
elucidated MOAs and/or susceptibility, particularly if the studies were conducted under
atmospherically relevant conditions, were emphasized whenever possible. A limited number of
toxicological studies were included that employed intratracheal (IT) instillation techniques, mainly
for PMio_2.5 studies in rodents, that explored new emerging areas of investigation (e.g., vasomotor
function), or that evaluated specific potential MOA or mechanisms of response. The sources,
transport, and fate of fibers and unique nano-materials (viz., dots, hollow spheres, rods, fibers, tubes)
are not reviewed herein because the in vivo disposition of these unique nanomaterials is not
necessarily relevant to the behavior of ultrafme (UF) aerosols in the urban environment that are
created by combustion sources and photochemical formation of secondary organic aerosols. In
considering the potential effects of different components of PM, EPA has focused on studies that
have assessed effects for a range of PM sources or components, including those using source
apportionment methods or comparing effects for numerous PM components, and not on studies of
individual constituents or species. Studies of ubiquitous PM sources as part of a mixture (i.e., diesel
exhaust, gasoline exhaust, wood smoke) are included, provided they meet the other remaining
selection criteria. Those studies of mixtures that are not a significant source of ambient PM, such as
environmental tobacco smoke (ETS), are not included.
These criteria provide benchmarks for evaluating various studies and for focusing on the
policy relevant studies in assessing the body of health and welfare effects evidence. Detailed critical
analysis of all PM health and welfare effects studies, especially in relation to the above
considerations, is beyond the scope of this document. Of most relevance for evaluation of studies is
whether they provide useful qualitative or quantitative information on exposure-effect or
exposure-response relationships for effects associated with current ambient air concentrations of PM
that can inform decisions on whether to retain or revise the standards.
In developing the PM ISA, EPA began by reviewing and summarizing the evidence on (1)
atmospheric sciences and exposure; (2) the health effects evidence from in vivo and in vitro animal
toxicological, controlled human exposure, and epidemiologic studies; and (3) the welfare effects of
PM, including visibility, climate, and ecological effects. In June 2008, EPA held a workshop to
obtain review of the scientific content of initial draft materials or sections for the draft ISA and its
annexes, that primarily contain summary information. The purpose of the initial peer review
workshop was to ensure that the ISA is up to date and focused on the most policy-relevant findings,
and to assist EPA with integration of evidence within and across disciplines. Following the peer
review workshop, EPA addressed comments from the peer review workshop and completed the
initial integration and synthesis of the evidence.
The integration of evidence on health or welfare effects involves collaboration between
scientists from various disciplines. As described in the section below, the ISA organization is based
on health or welfare effect categories. As an example, an evaluation of health effects evidence would
include summaries of findings from epidemiologic, controlled human exposure, and toxicological
studies, and integration of the results to draw conclusions based on the causal framework described
below. Using the causal framework described in Section 1.5, EPA scientists consider aspects such as
strength, consistency, coherence and biological plausibility of the evidence, and develop draft
December 2009 1-11
-------
causality judgments on the nature of the relationships. The draft integrative synthesis sections and
conclusions are reviewed by EPA internal experts and, as appropriate, by outside expert authors. In
practice, causality determinations often entail an iterative process of review and evaluation of the
evidence. The draft ISA is released for review by the CASAC and the public. Comments on the
characterization of the science as well as the implementation of the causal framework are carefully
considered in revising and completing the ISA.
PMio health studies are included in this assessment because they provide important evidence
regarding the health effects of PM in general. However, the ISA draws no conclusions regarding
causality for short- or long-term exposure to PMi0, as PMi0 is comprised of both fine and thoracic
coarse particles. As a result, causality determinations are limited to PM2.5, PMi0_2.5, and UF particle
(UFP) size fractions. In the cases where it was determined that PMip is dominated by fine or thoracic
coarse PM in specific study locations, these health studies are used in supporting the causality
determinations for PM2.5 or PM10_2.5. Epidemiologic studies of short-term exposure to PM10 are also
relied upon to examine potential effect modifiers, potential confounding by copollutants, and the
influence of different modeling approaches on PM-mortality risk estimates, as well as the
concentration-response relationship between PM and mortality. Therefore, to the extent possible, the
findings of PMio studies are considered insofar as they provide information relevant to the review of
the NAAQS for fine and thoracic coarse particles.
1.4. Document Organization
This ISA is composed of nine chapters. This introductory chapter presents background
information, and provides an overview of EPA's framework for making causal judgments. Key
findings and conclusions for consideration in the review of the NAAQS for PM from the
atmospheric sciences, ambient air data analyses, exposure assessment, dosimetry, health and welfare
effects, including judgments on causality for the health and welfare effects of PM exposure, are
presented in Chapter 2. More detailed summaries, evaluations and integration of the evidence are
included in Chapters 3 through 9.
Chapter 3 highlights key concepts or issues relevant to understanding the atmospheric
chemistry, sources, and exposure of and to PM following a "source-to-exposure" paradigm.
Chapter 4 summarizes key concepts and recent findings on the dosimetry of PM, and Chapter 5
discusses possible pathways and MOA for the effects of PM. Chapters 6 and 7 evaluate and integrate
epidemiologic, controlled human exposure, and animal toxicological information relevant to the
review of the primary NAAQS for PM. Health effects related to short-term exposures (hours to days)
to PM are the focus of Chapter 6. Chapter 7 evaluates health evidence related to long-term exposures
(months to years) to PM. Chapters 6 and 7 are organized by health outcome categories, such as
cardiovascular or respiratory effects, and each section includes effects of the various types of PM
studied. For each health outcome category, summary sections then integrate the findings to draw
conclusions on the evidence for the main size classes of PM (i.e., PM25, PMi0_2.5, and UFP).
Chapter 6 also includes a summary and synthesis of the recent health evidence that uses systematic
approaches to assess health effects of sources and constituents of ambient PM; most such studies
have evaluated effects of short-term exposure. Chapter 8 evaluates evidence related to populations
potentially susceptible to PM-related effects.
Chapter 9 evaluates welfare effects evidence that is relevant to the review of the secondary
NAAQS for PM. This chapter includes consideration of effects of PM on visibility impairment,
materials damage, effects of PM on climate, and ecological effects of PM that were not addressed in
the Integrated Science Assessment for Oxides of Nitrogen and Sulfur—Ecological Criteria (NOXSOX
ISA) (U.S. EPA, 2008, 157074). The chapter also presents key conclusions and scientific judgments
regarding causality for welfare effects of PM. In 2008, EPA completed the NOXSOX ISA (U.S. EPA,
2008, 157074). that focused on ecological effects related to the deposition of nitrogen (N)- and sulfur
(S)-containing compounds. The 2008 NOXSOX ISA included ecological effects from particle-phase
compounds (e.g., nitrates and sulfates), primarily effects from acidification and N-nutrient
enrichment and eutrophication. In this ISA, the focus is on recent data for direct welfare effects of
particle-phase NOX and SOX in the ambient air - primarily visibility impairment, damage to
materials, and positive and negative climate interactions - not the welfare effects related to
deposition of particle-phase NOX and SOX.
December 2009 1-12
-------
A series of annexes supplement this ISA. The annexes provide additional details of the
pertinent literature published since the last review, as well as selected older studies of particular
interest. These annexes contain information on:
• atmospheric chemistry of PM, sampling and analytic methods for measurement of PM,
concentrations, emissions, sources and human exposure to PM (Annex A);
• studies on the dosimetry of PM (Annex B);
• controlled human exposure studies of health effects related to exposure to PM (Annex
C);
• toxicological studies of health effects related to exposure to PM in laboratory animals
and cell cultures (Annex D);
• epidemiologic studies of health effects from short- and long-term exposure to PM
(Annex E); and
• studies that evaluate PM-induced health effects attributable to specific constituents or
sources (Annex F).
Within Annexes B through F, detailed information about methods and results of health studies
is summarized in tabular format, and generally includes information about: concentrations of PM
and averaging times; study methods employed; results; and quantitative results for relationships
between effects and exposure to PM. As noted in the section above, the most pertinent results of this
body of studies are brought into the ISA.
1.5. EPA Framework for Causal Determination
The EPA has developed a consistent and transparent basis to evaluate the causal nature of air
pollution-induced health or environmental effects. The framework described below establishes
uniform language concerning causality and brings more specificity to the findings. It drew
standardized language from across the federal government and wider scientific community,
especially from the recent National Academy of Sciences (NAS) Institute of Medicine (IOM)
document, Improving the Presumptive Disability Decision-Making Process for Veterans (IOM, 2008,
156586). the most recent comprehensive work on evaluating causality.
This introductory section focuses on the evaluation of health effects evidence; while focusing
on human health outcomes, the concepts are also generally relevant to causality determination for
welfare effects. This section:
• describes the kinds of scientific evidence used in establishing a general causal
relationship between exposure and health effects;
• defines cause, in contrast to statistical association;
• discusses the sources of evidence necessary to reach a conclusion about the existence of
a causal relationship;
• highlights the issue of multifactorial causation;
• identifies issues and approaches related to uncertainty; and
• provides a framework for classifying and characterizing the weight of evidence in
support of a general causal relationship.
Approaches to assessing the separate and combined lines of evidence (e.g., epidemiologic,
controlled human exposure, and animal toxicological studies) have been formulated by a number of
December 2009 1-13
-------
regulatory and science agencies, including the IOM of the NAS (IOM, 2008, 156586). International
Agency for Research on Cancer (IARC, 2006, 093206). EPA Guidelines for Carcinogen Risk
Assessment (U.S. EPA, 2005, 086237). Centers for Disease Control and Prevention (CDC, 2004,
056384). and National Acid Precipitation Assessment Program (NAPAP, 1991, 095894). These
formalized approaches offer guidance for assessing causality. The frameworks are similar in nature,
although adapted to different purposes, and have proven effective in providing a uniform structure
and language for causal determinations. Moreover, these frameworks have supported
decision-making under conditions of uncertainty.
1.5.1. Scientific Evidence Used in Establishing Causality
Causality determinations are based on the evaluation and synthesis of evidence from across
scientific disciplines; the type of evidence that is most important for such determinations will vary
by assessment. The most direct evidence of a causal relationship between pollutant exposures and
human health effects comes from controlled human exposure studies. This type of study
experimentally evaluates the health effects of administered exposures in human volunteers under
highly-controlled laboratory conditions.
In most epidemiologic or observational studies of humans, the investigator does not control
exposures or intervene with the study population. Broadly, observational studies can describe
associations between exposures and effects. These studies fall into several categories:
cross-sectional, prospective cohort, and time-series studies. "Natural experiments" offer the
opportunity to investigate changes in health with a change in exposure; these include comparisons of
health effects before and after a change in population exposures, such as the closure of a pollution
source.
Experimental animal data can help characterize effects of concern, exposure-response
relationships, susceptible populations, MO As and enhance understanding of biological plausibility of
observed effects. In the absence of controlled human exposure or epidemiologic data, animal data
alone may be sufficient to support a likely causal determination, assuming that similar responses are
expected in humans.
1.5.2. Association and Causation
"Cause" is a significant, effectual relationship between an agent and an effect on health or
public welfare. "Association" is the statistical dependence among events, characteristics, or other
variables. An association is prima facie evidence for causation; alone, however, it is insufficient
proof of a causal relationship between exposure and disease or health effect. Determining whether an
observed association is causal rather than spurious involves consideration of a number of factors, as
described below. Much of the newly available health information evaluated in this ISA comes from
epidemiologic studies that report a statistical association between ambient exposure and health
outcomes.
Many of the health and environmental outcomes reported in these studies have complex
etiologies. Diseases such as asthma, coronary artery disease or cancer are typically initiated by a web
of multiple agents. Outcomes depend on a variety of factors, such as age, genetic susceptibility,
nutritional status, immune competence, and social factors (Gee and Payne-Sturges, 2004, 093070;
IOM, 2008, 156586). Effects on ecosystems are also multifactorial with a complex web of causation.
Further, exposure to a combination of agents could cause synergistic or antagonistic effects. Thus,
the observed risk represents the net effect of many actions and counteractions.
1.5.3. Evaluating Evidence for Inferring Causation
Moving from association to causation involves elimination of alternative explanations for the
association. In estimating the causal influence of an exposure on health or environmental effects, it is
recognized that scientific findings include uncertainty. Uncertainty can be defined as a state of
having limited knowledge where it is impossible to exactly describe an existing state or future
outcome; the lack of knowledge about the correct value for a specific measure or estimate.
Uncertainty characterization and uncertainty assessment are two activities that lead to different
degrees of sophistication in describing uncertainty. Uncertainty characterization generally involves a
December 2009 1-14
-------
qualitative discussion of the thought processes that lead to the selection and rejection of specific
data, estimates, scenarios, etc. Uncertainty assessment is more quantitative. The process begins with
simpler measures (e.g., ranges) and simpler analytical techniques and progresses, to the extent
needed to support the decision for which the assessment is conducted, to more complex measures
and techniques. Data will not be available for all aspects of an assessment, and those data that are
available may be of questionable or unknown quality. In these situations, evaluation of uncertainty
can include professional judgment or inferences based on analogy with similar situations. The net
result is that the assessments will be based on a number of assumptions with varying degrees of
uncertainty. Uncertainties commonly encountered in evaluating health evidence for the criteria air
pollutants are outlined below for epidemiologic and experimental studies. Various approaches to
characterizing uncertainty include classical statistical methods, sensitivity analysis, or probabilistic
uncertainty analysis, in order of increasing complexity and data requirements. The ISA generally
evaluates uncertainties qualitatively in assessing the evidence from across studies; in some situations
quantitative analysis approaches, such as meta-regression may be used.
It is important to note here that, although the following discussion refers primarily to health
effect studies, many parallels exist with welfare effects studies. Controlled exposure studies have
been conducted in which plant species have been directly exposed to air pollutants, and the strengths
and limitations of that body of studies mirror those of the controlled human exposure studies
discussed below. Ecological field or natural gradient studies are similar to epidemiologic studies, for
example, in the study of free-living populations and in the challenges faced in distinguishing effects
of pollutants within a mixture.
Controlled human exposure studies evaluate the effects of exposures to a variety of pollutants
in a highly-controlled laboratory setting. Also referred to as human clinical studies, these
experiments allow investigators to expose subjects to fixed concentrations of air pollutants under
carefully regulated environmental conditions and activity levels. Controlled human exposures to PM
typically involve exposing subjects either at rest or while engaged in intermittent exercise in a
whole-body exposure chamber, although mouthpiece and facemask systems can also be used. A
variety of different types of particles are used in these studies including ambient outdoor particles,
concentrated ambient particles (CAPs), diesel exhaust (DE) from a diesel engine, wood smoke
generated in a wood stove, laboratory generated model particles (e.g., elemental carbon [EC] or zinc
oxide [ZnO]), or particles collected on a filter, resuspended in saline, and administered either
through IT instillation or inhalation. The recovery of particles on filters is variable and some
components, such as organics, may be too volatile to be collected. Exposures to artificially generated
particles may provide important information on the health effects of PM, but are not truly
representative of ambient air pollution particles. The direct exposure of humans to ambient air
pollution particles may be complicated by factors that cannot be controlled such as coexposures to
other air pollutants (e.g., O3, SO2, and NO2). In concentrating ambient particles, gaseous copollutants
are not proportionately concentrated and interactions between PM and the copollutants cannot be
investigated unless the latter are re-introduced. These limitations as well as daily variability in
concentration and composition can make it difficult to compare the results from controlled human
exposure studies employing particles from different sources.
In some instances, controlled human exposure studies can also be used to characterize
concentration-response relationships at pollutant concentrations relevant to ambient conditions.
Controlled human exposures are typically conducted using a randomized crossover design with
subjects exposed both to PM and a clean air control. In this way, subjects serve as their own controls,
effectively controlling for many potential confounders. However, controlled human exposure studies
are limited by a number of factors including a small sample size and short exposure times. These
laboratory studies are often conducted at PM concentrations much higher than those typically
observed under ambient conditions, which may result in an overestimate of the acute response to
exposure in the general population. Although the repetitive nature of ambient PM exposures may
lead to cumulative health effects, this type of exposure is not practical to replicate in a laboratory
setting. In addition, while subjects do serve as their own controls, personal exposure to pollutants in
the hours and days preceding the controlled exposures may vary significantly between and within
individuals. Finally, controlled human exposure studies require investigators to adhere to stringent
health criteria for a subject to be included in the study, and therefore the results cannot necessarily be
generalized to an entire population. Although some controlled human exposure studies have included
health comprised individuals such as asthmatics or individuals with chronic obstructive pulmonary
disease (COPD) or coronary artery disease, these individuals must also be relatively healthy and do
December 2009 1-15
-------
not represent the most sensitive individuals in the population. Thus, a lack of observation of effects
from controlled human exposure studies does not necessarily mean that a causal relationship does
not exist. While controlled human exposure studies provide important information on the biological
plausibility of associations observed between air pollutant exposure and health outcomes in
epidemiologic studies, observed effects in these studies may underestimate the response in certain
subpopulations.
Epidemiologic studies provide important information on the associations between health
effects and exposure of human populations to ambient air pollution. In the evaluation of
epidemiologic evidence, one important consideration is potential confounding. Confounding is "... a
confusion of effects. Specifically, the apparent effect of the exposure of interest is distorted because
the effect of an extraneous factor is mistaken for or mixed with the actual exposure effect (which
may be null)" (Rothman and Greenland, 1998, 086599). One approach to remove spurious
associations from possible confounders is to control for characteristics that may differ between
exposed and unexposed persons; this is frequently termed "adjustment." Appropriate statistical
adjustment for confounders requires identifying and measuring all reasonably expected confounders.
Deciding which variables to control for in a statistical analysis of the association between exposure
and disease or health outcome depends on knowledge about possible mechanisms and the
distributions of these factors in the population under study. In addition, scientific judgment is needed
regarding likely sources and magnitude of confounding, together with consideration of how well the
existing constellation of study designs, results, and analyses address this potential threat to
inferential validity. One key consideration in this review is evaluation of the potential contribution of
PM to health effects when it is a component of a complex air pollutant mixture. Reported PM effect
estimates in epidemiologic studies may reflect independent PM effects on respiratory and
cardiovascular health. Ambient PM may also be serving as an indicator of complex ambient air
pollution mixtures that share the same source as PM (i.e., combustion of S-containing fuels or motor
vehicle emissions). Alternatively, copollutants may mediate the effects of PM or PM may influence
the toxicity of copollutants.
Another important consideration in the evaluation of epidemiologic evidence is effect
modification. "Effect-measure modification differs from confounding in several ways. The main
difference is that, whereas confounding is a bias that the investigator hopes to prevent or remove
from the effect estimate, effect-measure modification is a property of the effect under study ... In
epidemiologic analysis one tries to eliminate confounding but one tries to detect and estimate effect-
measure modification" (Rothman and Greenland, 1998, 086599). Examples of effect modifiers in
some of the studies evaluated in this ISA include environmental variables (e.g., temperature or
humidity), individual risk factors (e.g., education, cigarette smoking status, age), and community
factors (e.g., percent of population > 65 years old). It is often possible to stratify the relationship
between health outcome and exposure by one or more of these risk factor variables. Effect modifiers
may be encountered (a) within single-city time-series studies; or (b) across cities in a two-stage
hierarchical model or meta-analysis.
Several statistical methods are available to detect and control for potential confounders, with
none of them being completely satisfactory. Multivariable regression models constitute one tool for
estimating the association between exposure and outcome after adjusting for characteristics of
participants that might confound the results. The use of multipollutant regression models has been
the prevailing approach for controlling potential confounding by copollutants in air pollution health
effects studies. Finding the pollutant likely responsible for the health outcome from multipollutant
regression models is made difficult by the possibility that one or more air pollutants may be acting as
a surrogate for an unmeasured or poorly-measured pollutant or for a particular mixture of pollutants.
In addition, more than one pollutant may exert similar health effects, resulting in independently
observed associations for multiple pollutants. Further, the correlation between the air pollutant of
interest and various copollutants may make it difficult to discern associations between different
pollutant exposures and health effects. Thus, results of models that attempt to distinguish gaseous
and particle effects must be interpreted with caution. The number and degree of diversity of
covariates, as well as their relevance to the potential confounders, remain matters of scientific
judgment. Despite these limitations, the use of multipollutant models is still the prevailing approach
employed in most air pollution epidemiologic studies, and provides some insight into the potential
for confounding or interaction among pollutants.
Adjustment for potential confounders can be influenced by differential exposure measurement
error. There are several components that contribute to exposure measurement error in epidemiologic
December 2009 1-16
-------
studies, including the difference between true and measured ambient concentrations, the difference
between average personal exposure to ambient pollutants and ambient concentrations at central
monitoring sites, and the use of average population exposure rather than individual exposure
estimates. Previous AQCDs have examined the role of measurement error in time-series
epidemiologic studies using simulated data and mathematical analyses and suggested that "transfer
of effects" would only occur under unusual circumstances (i.e., "true" predictors having high
positive or negative correlation; substantial measurement error; or extremely negatively correlated
measurement errors) (U.S. EPA, 2004, 056905).
Confidence that unmeasured confounders are not producing the findings is increased when
multiple studies are conducted in various settings using different subjects or exposures; each of
which might eliminate another source of confounding from consideration. Thus, multicity studies
which use a consistent method to analyze data from across locations with different levels of
covariates can provide insight on potential confounding in associations. Intervention studies, because
of their quasi-experimental nature, can be particularly useful in characterizing causation.
In addition to controlled human exposure and epidemiologic studies, the tools of experimental
biology have been valuable for developing insights into human physiology and pathology. Animal
toxicological studies explore the effects of pollutants on human health, especially through the study
of model systems in other species. These studies evaluate the effects of exposures to a variety of
pollutants in a highly controlled laboratory setting, and allow exploration of MO As or mechanisms
by which a pollutant may cause effects. Background knowledge of the biological mechanisms by
which an exposure might or might not cause disease can prove crucial in establishing, or negating, a
causal claim. There are, however, uncertainties associated with quantitative extrapolations between
laboratory animals and humans on the pathophysiological effects of any pollutant. Animal species
can differ from each other in fundamental aspects of physiology and anatomy (e.g., metabolism,
airway branching, hormonal regulation) that may limit extrapolation. The differences between
humans and rodents with regard to pollutant absorption and distribution profiles based on breathing
pattern, exposure dose, and differences in lung structure and anatomy all have to be taken into
consideration.
A relatively new tool available for experimental studies of PM exposure is the particle
concentrator. Particle concentrators enable human subjects, animals, or cell culture systems to be
exposed to atmospheric PM at concentrations greater than that observed under ambient conditions.
As ambient PM is just one component of a complex mixture that interacts with gases and other
aerosols, CAPs systems provide a method of exposing subjects to the particle phase. There are
several instrument systems used to concentrate ambient PM in controlled human or animal exposure
studies (Gordon et al, 1999, 001176: Maciejczyk and Chen, 2005, 087456: Sioutas et al, 1995,
001629: Sioutas et al., 1999, 001633). Gases (such as O3 and SO2) are not concentrated nor is
PM10_2.5 (except for the coarse particle concentrator) and only certain systems are capable of
concentrating UFPs. In UF CAPs systems, increased number fraction of organic carbon and PAHs,
along with decreased relative percentage of EC particles have been reported in concentrated PM
compared to ambient PM (Su et al., 2006, 157021). These data suggest that for UF concentrators, the
CAPs do not accurately reflect atmospheric UFP composition.
The ability to extrapolate between species has not generally changed since the 2004 PM
AQCD but some considerations related to coarse particles merit attention. The inhalability of
particles >2.5 um in diameter is considerably lower in rats than in humans; however, once inhaled,
deposition in the extrathoracic region is near 100% percent for particles >5 um for most laboratory
animal species (rat, mouse, hamster, guinea pig, and dogs). By contrast, penetration of thoracic
coarse particles into the lower respiratory tract is greater in humans than rodents due to the
moderately less efficient nasal deposition of humans and oronasal breathing (especially during
exercise). The extent to which coarse particle deposition in the lower respiratory tract differs
between the species is highly dependent on the activity level of the human exposure scenario in
contrast with the resting exposure conditions common to rodent exposures. Endotracheal exposures
of rodents may be needed to achieve coarse particle tissue doses in the lower respiratory tract of
rodents similar to those experienced by humans. For particles <1 urn, including UFPs, deposition is
expected to be relatively similar between the species.
There are also differences between species in both the rates of particle clearance from and
retention in the lung. The clearance rate of particles from the ciliated airways of rats is considerably
greater than humans. There is also evidence of prolonged particle retention in the smaller
bronchioles of humans that does not appear to exist or has not been observed in rats. Under most
December 2009 1-17
-------
circumstances, clearance from the alveolar region of rats is also more rapid than observed in humans.
Thus, these combined effects contribute to a greater particle burden in the lower respiratory tract of
humans relative to rats. An important consideration in studies where are rats chronically exposed to
high concentrations of insoluble particles, is the potential for "overload conditions." Rats, unlike
other laboratory animals or humans, may experience a reduction in their alveolar clearance rates and
an accumulation of interstitial particle burden and under these conditions, the relevance of tissue
burdens and responses to humans is questionable. Considering interspecies differences in both
deposition and clearance, greater exposure concentrations are required to achieve coarse particle
tissue doses in the lower respiratory tract of rodents similar to those experienced by humans.
1.5.4. Application of Framework for Causal Determination
EPA uses a two-step approach to evaluate the scientific evidence on health or environmental
effects of criteria pollutants. The first step determines the weight of evidence in support of causation
and characterizes the strength of any resulting causal classification. The second step includes further
evaluation of the quantitative evidence regarding the concentration-response relationships and the
loads or levels, duration and pattern of exposures at which effects are observed.
To aid judgment, various "aspects" of causality have been discussed by many philosophers
and scientists. The most widely cited aspects of causality in epidemiology, and public health, in
general, were articulated by Sir Austin Bradford Hill (1965, 071664) and have been widely used
(CDC, 2004, 056384; IARC, 2006, 093206; IOM, 2008, 156586; NRC, 2004, 156814; U.S. EPA,
2005, 086237). Several adaptations of the Hill aspects have been used in aiding causality judgments
in the ecological sciences (Adams, 2003, 156192; Collier, 2003, 155736; Fox, 1991, 156444;
Gerritsen et al, 1998, 156465).
These aspects (Hill, 1965, 071664) have been modified (Table 1-2) for use in causal
determinations specific to health and welfare effects or pollutant exposures.2 Some aspects are more
likely than others to be relevant for evaluating evidence on the health or environmental effects of
criteria air pollutants. For example, the analogy aspect does not always apply and specificity would
not be expected for multi-etiologic health outcomes such as asthma or cardiovascular disease, or
ecological effects related to acidification. Aspects that usually play a larger role in determination of
causality are consistency of results across studies, coherence of effects observed in different study
types or disciplines, biological plausibility, exposure-response relationship, and evidence from
"natural" experiments.
Although these aspects provide a framework for assessing the evidence, they do not lend
themselves to being considered in terms of simple formulas or fixed rules of evidence leading to
conclusions about causality (Hill, 1965, 071664). For example, one cannot simply count the number
of studies reporting statistically significant results or statistically nonsignificant results and reach
credible conclusions about the relative weight of the evidence and the likelihood of causality. In
addition, it is important to note that the aspects in Table 1-2 cannot be used as a strict checklist, but
rather to determine the weight of the evidence for inferring causality. While these aspects are
particularly salient in this assessment, it is also important to recognize that no one aspect is either
necessary or sufficient for drawing inferences of causality.
1 The "aspects" described by Hill (1965, 071664) have become, in the subsequent literature, more commonly described as "criteria." The
original term "aspects" is used here to avoid confusion with 'criteria' as it is used, with different meaning, in the Clean Air Act.
2 The Hill aspects were developed for interpretation of epidemiologic results. They have been modified here for use with a broader array of
data, i.e., epidemiologic, controlled human exposure, and animal toxicological studies, as well as in vitro data, and to be more consistent
with EPA's Guidelines for Carcinogen Risk Assessment.
December 2009 1-18
-------
Table 1-2. Aspects to aid in judging causality.
Aspect
Description
CONSISTENCY OF THE
OBSERVED ASSOCIATION
An inference of causality is strengthened when a pattern of elevated risks is observed across
several independent studies, conducted in multiple locations by multiple investigators. The
reproducibility of findings constitutes one of the strongest arguments for causality. If there are
discordant results among investigations, possible reasons such as differences in exposure,
confounding factors, and the power of the study are considered.
COHERENCE An inference of causality from epidemiologic associations may be strengthened by other lines of
evidence (e.g., controlled human exposure and animal toxicological studies) that support a
cause-and-effect interpretation of the association. Causality is also supported when epidemiologic
associations are reported across study designs and across related health outcomes. Evidence on
ecological or welfare effects may be drawn from a variety of experimental approaches (e.g.,
greenhouse, laboratory, and field) and subdisciplines of ecology (e.g., community ecology,
biogeochemistry and paleological/ historical reconstructions). The coherence of evidence from
various fields greatly adds to the strength of an inference of causality. The absence of other lines of
evidence, however, is not a reason to reject causality
An inference of causality tends to be strengthened by consistency with data from experimental
studies or other sources demonstrating plausible biological mechanisms. A proposed mechanistic
linking between an effect, and exposure to the agent, is an important source of support for causality,
especially when data establishing the existence and functioning of those mechanistic links are
available. A lack of biological understanding, however, is not a reason to reject causality.
BIOLOGICAL PLAUSIBILITY
BIOLOGICAL GRADIENT
(EXPOSURE-RESPONSE
RELATIONSHIP)
Awell characterized exposure-response relationship (e.g., increasing effects associated with greater
exposure) strongly suggests cause and effect, especially when such relationships are also observed
for duration of exposure (e.g., increasing effects observed following longer exposure times). There
are, however, many possible reasons that a study may fail to detect an exposure-respor
relationship. Thus, although the presence of a biological gradient may support causality,
absence of an exposure-response relationship does not exclude a causal relationship.
Ionse
the
STRENGTH OF THE
OBSERVED ASSOCIATION
The finding of large, precise risks increases confidence that the association is not likely due to
chance, bias, or other factors. However, given a truly causal agent, a small magnitude in the effect
could follow from a lower level of exposure, a lower potency, or the prevalence of other agents
causing similar effects. While large effects support causality, modest effects therefore do not
preclude it.
The strongest evidence for causality can be provided when a change in exposure brings about a
change in occurrence or frequency of health or welfare effects.
Evidence of a temporal sequence between the introduction of an agent and appearance of the effect
constitutes another argument in favor of causality.
As originally intended, this refers to increased inference of causality if one cause is associated with
a single effect or disease (Hill, 1965,071664). Based on the current understanding this is now
considered one of the weaker guidelines for causality; for example, many agents cause respiratory
disease and respiratory disease has multiple causes. At the scale of ecosystems, as in
epidemiology, complexity is such that single agents causing single effects, and single effects
following single causes, are extremely unlikely. The ability to demonstrate specificity under certain
conditions remains, however, a powerful attribute of experimental studies. Thus, although the
presence of specificity may support causality, its absence does not exclude it.
ANALOGY Structure activity relationships and information on the agent's structural analogs can provide insight
into whether an association is causal. Similarly, information on mode of action for a chemical, as one
of many structural analogs, can inform decisions regarding likely causality.
EXPERIMENTAL EVIDENCE
TEMPORAL RELATIONSHIP OF
THE OBSERVED ASSOCIATION
SPECIFICITY OF THE
OBSERVED ASSOCIATION
1.5.5. First Step—Determination of Causality
In the ISA, EPA assesses the results of recent relevant publications, building upon evidence
available during the previous NAAQS review, to draw conclusions on the causal relationships
between relevant pollutant exposures and health or environmental effects. This ISA uses a five-level
December 2009
1-19
-------
hierarchy that classifies the weight of evidence for causation, not just association1. In developing this
hierarchy, EPA has drawn on the work of previous evaluations, most prominently the lOM's
Improving the Presumptive Disability Decision-Making Process for Veterans (IOM, 2008, 156586],
EPA's Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005, 086237) and the U.S. Surgeon
General's smoking reports (CDC, 2004, 056384). This weight of evidence evaluation is based on
various lines of evidence from across the health and environmental effects disciplines. These
separate judgments are integrated into a qualitative statement about the overall weight of the
evidence and causality. The five descriptors for causal determination are described in Table 1-3.
For PM, this determination of causality step involved a rather complex evaluation of evidence
for different PM indices, different types of health or environmental effects, and for short- and long-
term exposure periods. There were insufficient data on peak (i.e., <24 h) exposures for any PM size
fraction with health effects to make causality determinations for this exposure category. Causality
determinations were made for the PM measure (PM2.5s PM10_2.5, and UFPs, to the extent evidence was
available for each measure), the overall effect category, and the exposure duration. As noted above,
to the extent possible, results of PMi0 studies are considered in causality determinations for PM2.5
and PM 10-2.5. In the evaluation of health effects findings in Chapter 6 (for short-term exposure) and
Chapter 7 (for long-term exposure), evidence was evaluated for health outcome categories, such as
cardiovascular effects, and then conclusions were drawn based upon the integration of evidence from
across disciplines (e.g., epidemiology, controlled human exposure, and toxicology) and also across
the suite of related individual health outcomes. Chapters 6 and 7 initially summarize and evaluate
findings for individual health outcomes, then integrate the results in summary sections to draw
conclusions on causality for each PM indicator. The causality narratives present the weight of
evidence that highlights the quality and breadth of the data, including any limitations or
uncertainties. In the integrative synthesis and conclusions in Chapter 2, the ISA presents causality
determinations and a summary of the underlying basis for those determinations for the PM indicator
(e.g., PM2.5), for the exposure time period (e.g., short- and long-term exposure) and for the major
health effect categories.
1 It should be noted that the CDC and IOM frameworks use a four-category hierarchy for the strength of the evidence. A five-level
hierarchy is used here to be consistent with the EPA Guidelines for Carcinogen Risk Assessment and to provide a more nuanced set of
categories.
December 2009 1-20
-------
Table 1-3. Weight of evidence for causal determination.
Determination
Health Effects
Ecological and Welfare Effects
CAUSAL Evidence is sufficient to conclude that there is a causal
RELATIONSHIP relationship with relevant pollutant exposures. That is, the
pollutant has been shown to result in health effects in
studies in which chance, bias, and confounding could be
ruled out with reasonable confidence. For example: a)
controlled human exposure studies that demonstrate
consistent effects; or b) observational studies that cannot
be explained by plausible alternatives or are supported by
other lines of evidence (e.g., animal studies or mode of
action information). Evidence includes replicated and
consistent high-quality studies by multiple investigators.
Evidence is sufficient to conclude that there is a causal
relationship with relevant pollutant exposures. That is, the
pollutant has been shown to result in effects in studies in
which chance, bias, and confounding could be ruled out
with reasonable confidence. Controlled exposure studies
(laboratory or small- to medium-scale field studies)
provide the strongest evidence for causality, but the scope
of inference may be limited. Generally, determination is
based on multiple studies conducted by multiple research
groups, and evidence that is considered sufficient to infer
a causal relationship is usually obtained from the joint
consideration of many lines of evidence that reinforce
each other.
LIKELY TO BE A Evidence is sufficient to conclude that a causal
CAUSAL relationship is likely to exist with relevant pollutant
RELATIONSHIP exposures, but important uncertainties remain. That is,
the pollutant has been shown to result in health effects in
studies in which chance and bias can be ruled out with
reasonable confidence but potential issues remain. For
example: a) observational studies show an association,
but copollutant exposures are difficult to address and/or
other lines of evidence (controlled human exposure,
animal, or mode of action information) are limited or
inconsistent; or b) animal toxicological evidence from
multiple studies from different laboratories that
demonstrate effects, but limited or no human data are
available. Evidence generally includes replicated and
high-quality studies by multiple investigators.
Evidence is sufficient to conclude that there is a likely
causal association with relevant pollutant exposures. That
is, an association has been observed between the
pollutant and the outcome in studies in which chance, bias
and confounding are minimized, but uncertainties remain.
For example, field studies show a relationship, but
suspected interacting factors cannot be controlled, and
other lines of evidence are limited or inconsistent.
Generally, determination is based on multiple studies in
multiple research groups.
SUGGESTIVE OF Evidence is suggestive of a causal relationship with
A CAUSAL relevant pollutant exposures, but is limited because
RELATIONSHIP chance, bias and confounding cannot be ruled out. For
example, at least one high-quality epidemiologic study
shows an association with a given health outcome but the other studies are inconsistent.
results of other studies are inconsistent.
Evidence is suggestive of a causal relationship with
relevant pollutant exposures, but chance, bias and
confounding cannot be ruled out. For example, at least
one high-quality study shows an effect, but the results of
INADEQUATE TO Evidence is inadequate to determine that a causal
INFER A CAUSAL relationship exists with relevant pollutant exposures. The
RELATIONSHIP available studies are of insufficient quantity, quality,
consistency or statistical power to permit a conclusion
regarding the presence or absence of an effect.
The available studies are of insufficient quality,
consistency or statistical power to permit a conclusion
regarding the presence or absence of an effect.
NOT LIKELY TO Evidence is suggestive of no causal relationship with
BE A CAUSAL relevant pollutant exposures. Several adequate studies,
RELATIONSHIP covering the full range of levels of exposure that human
beings are known to encounter and considering
susceptible populations, are mutually consistent in not
showing an effect at any level of exposure.
Several adequate studies, examining relationships with
relevant exposures, are consistent in failing to show an
effect at any level of exposure.
1.5.6. Second Step—Evaluation of Response
Beyond judgments regarding causality are questions relevant to quantifying health or
environmental risks based on our understanding of the quantitative relationships between pollutant
exposures and health or welfare effects.
1.5.6.1. Effects on Human Populations
Once a determination is made regarding the causal relationship between the pollutant and
outcome category, important questions regarding quantitative relationships include:
December 2009
1-21
-------
• What is the concentration-response or dose-response relationship in the human
population?
• What exposure conditions (dose or exposure, duration and pattern) are important?
• What subpopulations appear to be differentially affected (i.e., more susceptible to
effects)?
To address these questions, in the second step of the EPA framework, the entirety of
quantitative evidence is evaluated to best quantify those concentration-response relationships that
exist. This requires evaluation of pollutant concentrations and exposure durations at which effects
were observed for exposed populations including potentially susceptible populations. This
integration of evidence results in identification of a study or set of studies that best approximates the
concentration-response relationships between health outcomes and PM indicators for the U.S.
population or subpopulations, given the current state of knowledge and the uncertainties that
surrounded these estimates.
To accomplish this, evidence from multiple and diverse types of studies is considered. To the
extent available, the ISA evaluates results from across epidemiologic studies that use various
methods to evaluate the form of relationships between PM and health outcomes, and draws
conclusions on the most well-supported shape of these relationships. Controlled human exposure
studies can also provide data on the concentration-response relationship. Animal data may inform
evaluation of concentration-response relationships, particularly relative to MOAs, and characteristics
of susceptible populations. For some health outcomes, the probability and severity of health effects
and associated uncertainties can be characterized. Chapter 2 presents the integrated findings
informative for evaluation of population risks.
An important consideration in characterizing the public health impacts associated with
exposure to a pollutant is whether the concentration-response relationship is linear across the full
concentration range encountered, or if nonlinear relationships exist along any part of this range. Of
particular interest is the shape of the concentration-response curve at and below the level of the
current standards. The shape of the concentration-response curve varies, depending on the type of
health outcome, underlying biological mechanisms and dose. At the human population level,
however, various sources of variability and uncertainty tend to smooth and "linearize" the
concentration-response function (such as the low data density in the lower concentration range,
possible influence of measurement error, and individual differences in susceptibility to air pollution
health effects). In addition, many chemicals and agents may act by perturbing naturally occurring
background processes that lead to disease, which also linearizes population concentration-response
relationships (Clewell and Crump, 2005, 156359: Crump et al., 1976, 003192: Hoel, 1980,
These attributes of population dose-response may explain why the available human data at ambient
concentrations for some environmental pollutants (e.g., PM, O3, lead [Pb], ETS, radiation) do not
exhibit evident thresholds for health effects, even though likely mechanisms include nonlinear
processes for some key events. These attributes of human population dose-response relationships
have been extensively discussed in the broader epidemiologic literature (Rothman and Greenland,
1998. 0865991
Publication bias is a source of uncertainty regarding the magnitude of health risk estimates. It
is well understood that studies reporting non-null findings are more likely to be published than
reports of null findings, and publication bias can also result in overestimation of effect estimate sizes
(loannidis, 2008, 188317). For example, effect estimates from single-city epidemiologic studies have
been found to be generally larger than those from multicity studies (Anderson et al., 2005, 087916).
Finally, identification of the susceptible population groups contributes to an understanding of
the public health impact of pollutant exposures. Epidemiologic studies can help identify susceptible
populations by evaluating health responses in the study population. Examples include stratified
analyses for subsets of the population under study, or testing for interactions or effect modification
by factors such as gender, age group, or health status. Experimental studies using animal models of
susceptibility or disease can also inform the extent to which health risks are likely greater in specific
population subgroups. Further discussion of these groups is in Chapter 8.
December 2009 1-22
-------
1.5.6.2. Effects on Public Welfare
Key questions for understanding the quantitative relationships between exposure (or
concentration or deposition) to a pollutant and risk to the public welfare (e.g., ecosystems, visibility,
materials, climate):
• What elements of the ecosystem (e.g., types, regions, taxonomic groups, populations,
functions, etc.) appear to be affected, or are more sensitive to effects?
• Under what exposure conditions (amount deposited or concentration, duration and
pattern) are effects seen?
• What is the shape of the concentration-response or exposure-response relationship?
Evaluations of causality typically consider the probability of welfare effects changing in
response to exposure. A challenge to the quantification of exposure-response relationships for
ecological effects is the variability across ecosystems. Ecological responses are evaluated within the
range of observations, so a quantitative relationship may be determined for a given ecological system
and scale. However, there is great regional and local variability in ecosystems. Thus, exposure-
response relationships are often available site by site, rather than at the national or even regional
scale. For example, an ecological response to deposition of a given pollutant can differ greatly
between ecosystems. Where results from greenhouse or animal ecotoxicological studies are
available, they may be used to aid in characterizing exposure-response relationships, particularly
relative to mechanisms of action, and characteristics of sensitive biota.
1.5.7. Concepts in Evaluating Adversity of Health Effects
In evaluating the health evidence, a number of factors can be considered in determining the
extent to which health effects are "adverse" for health outcomes such as changes in lung function.
What constitutes an adverse health effect may vary between populations. Some changes in healthy
individuals may not be considered adverse while those of a similar type and magnitude are
potentially adverse in more susceptible individuals.
The American Thoracic Society (ATS) published an official statement titled What Constitutes
an Adverse Health Effect of Air Pollution? (ATS, 2000, 011738). This statement updated the
guidance for defining adverse respiratory health effects that had been published 15 years earlier
(ATS, 1985, 006522). taking into account new investigative approaches used to identify the effects
of air pollution and reflecting concern for impacts of air pollution on specific susceptible groups. In
the 2000 update, there was an increased focus on quality of life measures as indicators of adversity
and a more specific consideration of population risk. Exposure to air pollution that increases the risk
of an adverse effect to the entire population is viewed as adverse, even though it may not increase
the risk of any identifiable individual to an unacceptable level; estimated mean population effects do
not reflect more severe effects in individuals. For example, a population of asthmatics could have a
distribution of lung function such that no identifiable individual has a level associated with
significant impairment. Exposure to air pollution could shift the distribution such that no identifiable
individual experiences clinically-relevant effects; this shift toward decreased lung function, however,
would be considered adverse because individuals within the population would have diminished
reserve function and, therefore, would be at increased risk to further environmental insult.
1.6. Summary
This ISA is a review, synthesis, and evaluation of the most policy-relevant science, and
communicates critical science judgments relevant to the NAAQS review. It reviews the most
policy-relevant evidence from environmental effects studies and includes information on
atmospheric chemistry, PM sources and emissions, exposure, and dosimetry. This ISA incorporates
clarification and revisions based on advice and comments provided by EPA's CASAC (Samet, 2009,
190992; Samet, 2009, 199522). Annexes to the IS A provide additional details of the literature
published since the last review. A framework for making critical judgments concerning causality
December 2009 1-23
-------
appears in this chapter. It relies on a widely accepted set of principles and standardized language to
express evaluation of the evidence. This approach can bring rigor and clarity to current and future
assessments. This ISA should assist EPA and others, now and in the future, to accurately represent
what is presently known—and what remains unknown—concerning the effects of PM on human
health and public welfare.
December 2009 1-24
-------
Chapter 1 References
Adams SM. (2003). Establishing causality between environmental stressors and effects on aquatic ecosystems. Hum Ecol
Risk Assess, 9: 17-35. 156192
Anderson HR; Atkinson RW; Peacock JL; Sweeting MJ; Marston L. (2005). Ambient particulate matter and health effects:
Publication bias in studies of short-term associations. Epidemiology, 16: 155-163. 087916
ATS. (1985). American Thoracic Society: Guidelines as to what constitutes an adverse respiratory health effect, with
special reference to epidemiologic studies of air pollution. Am Rev Respir Dis, 131: 666-668. 006522
ATS. (2000). What constitutes an adverse health effect of air pollution? Official statement of the American Thoracic
Society. Am J Respir Grit Care Med, 161: 665-673. 011738
CDC. (2004). The health consequences of smoking: A report of the Surgeon General. Centers for Disease Control and
Prevention, U.S. Department of Health and Human Services. Washington, DC. 056384
Clewell HJ; Crump KS. (2005). Quantitative estimates of risk for noncancer endpoints. Risk Anal, 25: 285-289. 156359
Collier TK. (2003). Forensic ecotoxicology: Establishing causality between contaminants and biological effects in field
studies. Hum Ecol Risk Assess, 9: 259 - 266. 155736
Crump KS; Hoel DG; Langley CH; Peto R. (1976). Fundamental carcinogenic processes and their implications for low
dose risk assessment. Cancer Res, 36: 2973-2979. 003192
Fox GA. (1991). Practical causal inference for ecoepidemiologists. J Toxicol Environ Health, 33: 359-373. 156444
Gee GC; Payne-Sturges DC. (2004). Environmental health disparities: A framework integrating psychosocial and
environmental concepts. Environ Health Perspect, 112: 1645-1653. 093070
Gerritsen J; Carlson RE; Dycus DL; Faulkner C; Gibson GR. (1998). Lake and reservoir bioassessment and biocriteria:
Technical guidance document. US Environmental Protection Agency, Office of Water. Washington, DC. EPA 841-
B-98-007. http://www.epa.gov/owow/monitoring/tech/lakes.html. 156465
Gordon T; Gerber H; Fang CP; Chen LC. (1999). A centrifugal particle concentrator for use in inhalation toxicology. Inhal
Toxicol, 11: 71-87. 001176
Henderson R. (2005). Clean Air Scientific Advisory Committee (CASAC) review of the EPA staff recommendations
concerning a Potential Thoracic Coarse PM standard in the Review of the National Ambient Air Quality Standards
for Particulate Matter: Policy Assessment of Scientific and Technical Information. U.S. Environmental Protection
Agency. Washington, DC. 156537
Henderson R. (2005). EPA's Review of the National Ambient Air Quality Standards for Particulate Matter (Second Draft
PM Staff Paper, January 2005): A review by the Particulate Matter Review Panel of the EPA Clean Air Scientific
Advisory Committee. U.S. Environmental Protection Agency. Washington D.C.. 188316
Henderson R. (2006). Clean Air Scientific Advisory Committee recommendations concerning the proposed National
Ambient Air Quality Standards for particulate matter. U.S. Environmental Protection Agency. Washington, DC.
156538
Hill AB. (1965). The environment and disease: Association or causation. J R Soc Med, 58: 295-300. 071664
Hoel DG. (1980). Incorporation of background in dose-response models. Fed Proc, 39: 73-75. 156555
IARC. (2006). IARC monographs on the evaluation of carcinogenic risks to humans: Preamble. International Agency for
Research on Cancer. Lyon,France . http://monographs.iarc.fr/ENG/Preamble/CurrentPreamble.pdf. 093206
loannidis JPA. (2008). Why most discovered true associations are inflated. Epidemiology, 19: 640-648. 188317
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009 1-25
-------
IOM. (2008). Improving the Presumptive Disability Decision-Making Process for Veterans; Committee on Evaluation of
the Presumptive Disability Decision-Making Process for Veterans, Board on Military and Veterans Health.
Washington, DC: Institue of Medicine of the National Academies, National Academies Press. 156586
Maciejczyk P; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice: VIII
source-related daily variations in in vitro responses to CAPs. Inhal Toxicol, 17: 243-253. 087456
NAPAP. (1991). The experience and legacy of NAPAP report of the Oversight Review Board. National Acid Precipitation
Assessment Program. Washington, DC. 095894
NAPCA. (1969). Air quality criteria for particulate matter. National Air Pollution Control Administration. Washington, DC.
014684
NRC. (2004). Research priorities for airborne particulate matter: IV Continuing research progress. National Academies
Press. Washington, DC. 156814
RothmanKJ; Greenland S. (1998). Modern epidemiology. Philadelphia, PA: Lippincott-Raven Publishers. 086599
Samet JM. (2009). CASAC Review of integrated science assessment for particulate matter (second external review draft,
July 2009). U.S. Environmental Protection Agency. Washington, D.C.. EPA-CASAC-10-001.199522
Samet JM. (2009). Review of EPA's Integrated Science Assessment for Particulate Matter (First External Review Draft,
December 2008). Clean Air Scientific Advisory Committee (CASAC) and members of the CASAC Particulate
Matter (PM) Review Panel, Science Advisory Board, U.S. Environmental Protection Agency. Washington, D.C..
EPA-CASAC-09-008. 190992
Sioutas C; Kim S; Chang M. (1999). Development and evaluation of a prototype ultrafine particle concentrator. J Aerosol
Sci, 30: 1001-1017.001633
Sioutas D; Koutrakis P; Ferguson ST; Burton RM. (1995). Development and evaluation of a prototype ambient particle
concentrator for inhalation exposure studies. Inhal Toxicol, 7: 633-644. 001629
Su Y; Sipin MF; Spencer MT; Qin X; Moffet RC; Shields LG; Prather KA; Venkatachari P; Jeong C-H; Kim E; Hopke PK;
Gelein RM; Utell MJ; Oberdorster G; Berntsen J; Devlin RB; Chen LC. (2006). Real-time characterization of the
composition of individual particles emitted from ultrafine particle concentrators. Aerosol Sci Technol, 40: 437-455.
157021
U.S. EPA. (2004). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research Triangle
Park, NC. EPA/600/P-99/002aF-bF. 056905
U. S. EPA. (2005). Guidelines for carcinogen risk assessment, Risk Assessment Forum Report. U. S. Environmental
Protection Agency. Washington, DC. EPA/630/P-03/001F. http://cfpub.epa.gov/ncea/index.cfm. 086237
U.S. EPA. (2005). Review of the national ambient air quality standards for particulate matter: Policy assessment of
scientific and technical information OAQPS staff paper. U.S. Environmental Protection Agency. Washington, DC.
EPA/452/R-05-005a. http://www.epa.gov/ttn/naaqs/standards/pm/data/pmstaffpaper_20051221 .pdf. 090209
U.S. EPA. (2008). Integrated review plan for the national ambient air quality standards for particulate matter. U.S.
Environmental Protection Agency, Office of Research and Development, National Center for Environmental
Assessment. Research Triangle Park, NC. 157072
U.S. EPA. (2008). Integrated science assessment for oxides of nitrogen and sulfur: Ecological criteria. U.S. Environmental
Protection Agency. Research Triangle Park, NC. EPA/600/R-08/082F. 157074
December 2009 1-26
-------
Chapter 2. Integrative Health and
Welfare Effects Overview
The subsequent chapters of this ISA will present the most policy-relevant information related
to this review of the NAAQS for PM. This chapter integrates the key findings from the disciplines
evaluated in this current assessment of the PM scientific literature, which includes the atmospheric
sciences, ambient air data analyses, exposure assessment, dosimetry, health studies
(e.g., toxicological, controlled human exposure, and epidemiologic), and welfare effects. The EPA
framework for causal determinations described in Chapter 1 has been applied to the body of
scientific evidence in order to collectively examine the health or welfare effects attributed to PM
exposure in a two-step process.
As described in Chapter 1, EPA assesses the results of recent relevant publications, building
upon evidence available during the previous NAAQS reviews, to draw conclusions on the causal
relationships between relevant pollutant exposures and health or environmental effects. This ISA
uses a five-level hierarchy that classifies the weight of evidence for causation:
• Causal relationship
• Likely to be a causal relationship
• Suggestive of a causal relationship
• Inadequate to infer a causal relationship
• Not likely to be a causal relationship
Beyond judgments regarding causality are questions relevant to quantifying health or
environmental risks based on our understanding of the quantitative relationships between pollutant
exposures and health or welfare effects. Once a determination is made regarding the causal
relationship between the pollutant and outcome category, important questions regarding quantitative
relationships include:
• What is the concentration-response or dose-response relationship?
• Under what exposure conditions (amount deposited, dose or concentration, duration and
pattern) are effects observed?
• What populations appear to be differentially affected (i.e., more susceptible) to effects?
• What elements of the ecosystem (e.g., types, regions, taxonomic groups, populations,
functions, etc.) appear to be affected, or are more sensitive to effects?
To address these questions, in the second step of the EPA framework, the entirety of
quantitative evidence is evaluated to identify and characterize potential concentration-response
relationships. This requires evaluation of levels of pollutant and exposure durations at which effects
were observed for exposed populations including potentially susceptible populations.
This chapter summarizes and integrates the newly available scientific evidence that best
informs consideration of the policy-relevant questions that frame this assessment, presented in
Chapter 1. Section 2.1 discusses the trends in ambient concentrations and sources of PM and
provides a brief summary of ambient air quality. Section 2.2 presents the evidence regarding
personal exposure to ambient PM in outdoor and indoor microenvironments, and it discusses the
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009 2-1
-------
relationship between ambient PM concentrations and exposure to PM from ambient sources.
Section 2.3 integrates the evidence for studies that examine the health effects associated with short-
and long-term exposure to PM and discusses important uncertainties identified in the interpretation
of the scientific evidence. Section 2.4 provides a discussion of policy-relevant considerations, such
as potentially susceptible populations, lag structure, and the PM concentration-response relationship,
and PM sources and constituents linked to health effects. Section 2.5 summarizes the evidence for
welfare effects related to PM exposure. Finally, Section 2.6 provides all of the causal determinations
reached for each of the health outcomes and PM exposure durations evaluated in this ISA.
2.1. Concentrations and Sources of Atmospheric PM
2.1.1. Ambient PM Variability and Correlations
Recently, advances in understanding the spatiotemporal distribution of PM mass and its
constituents have been made, particularly with regard to PM2.5 and its components as well as
ultrafme particles (UFPs). Emphasis in this ISA is placed on the period from 2005-2007,
incorporating the most recent validated EPA Air Quality System (AQS) data. The AQS is EPA's
repository for ambient monitoring data reported by the national, and state and local air monitoring
networks. Measurements of PM2.5 and PMi0 are reported into AQS, while PMi0_2.5 concentrations are
obtained as the difference between PMi0 and PM2 5 (after converting PMi0 concentrations from STP
to local conditions; Section 3.5). Note, however, that a majority of U.S. counties were not
represented in AQS because their population fell below the regulatory monitoring threshold.
Moreover, monitors reporting to AQS were not uniformly distributed across the U.S. or within
counties, and conclusions drawn from AQS data may not apply equally to all parts of a geographic
region. Furthermore, biases can exist for some PM constituents (and hence total mass) owing to
volatilization losses of nitrates and other semi-volatile compounds, and, conversely, to retention of
particle-bound water by hygroscopic species. The degree of spatial variability in PM was likely to be
region-specific and strongly influenced by local sources and meteorological and topographic
conditions.
2.1.1.1. Spatial Variability across the U.S.
AQS data for daily average concentrations of PM25 for 2005-2007 showed considerable
variability across the U.S. (Section 3.5.1.1). Counties with the highest average concentrations of
PM2.5 (>18 (ig/m3) were reported for several counties in the San Joaquin Valley and inland southern
California as well as Jefferson County, AL (containing Birmingham) and Allegheny County, PA
(containing Pittsburgh). Relatively few regulatory monitoring sites have the appropriate co-located
monitors for computing PMi0_2.5, resulting in poor geographic coverage on a national scale
(Figure 3-10). Although the general understanding of PM differential settling leads to an expectation
of greater spatial heterogeneity in the PMi0_2.s fraction, deposition of particles as a function of size
depends strongly on local meteorological conditions. Better geographic coverage is available for
PMio, where the highest reported annual average concentrations (>50 (ig/m3) occurred in southern
California, southern Arizona and central New Mexico. The size distribution of PM varied
substantially by location, with a generally larger fraction of PMi0 mass in the PMi0_2.s size range in
western cities (e.g., Phoenix and Denver) and a larger fraction of PMi0 in the PM25 size range in
eastern U.S. cities (e.g., Pittsburgh and Philadelphia). UFPs are not measured as part of AQS or any
other routine regulatory network in the U.S. Therefore, limited information is available regarding
regional variability in the spatiotemporal distribution of UFPs.
Spatial variability in PM2 5 components obtained from the Chemical Speciation Network
(CSN) varied considerably by species from 2005-2007 (Figures 3-12 through 3-18). The highest
annual average organic carbon (OC) concentrations were observed in the western and southeastern
U.S. OC concentrations in the western U.S. peaked in the fall and winter, while OC concentrations in
the Southeast peaked anytime between spring and fall. Elemental carbon (EC) exhibited less
seasonality than OC and showed lowest seasonal variability in the eastern half of the U.S. The
December 2009 2-2
-------
highest annual average EC concentrations were present in Los Angeles, Pittsburgh, New York, and
El Paso. Concentrations of sulfate (SO42~) were higher in the eastern U.S. as a result of higher SO2
emissions in the East compared with the West. There is also considerable seasonal variability with
higher SO42~ concentrations in the summer months when the oxidation of SO2 proceeds at a faster
rate than during the winter. Nitrate (NO3~) concentrations were highest in California and during the
winter in the Upper Midwest. In general, NO3~ was higher in the winter across the country, in part as
a result of temperature-driven partitioning and volatilization. Exceptions existed in Los Angeles and
Riverside, CA, where high NO3~ concentrations appeared year-round. There is variation in both
PM2.5 mass and composition among cities, some of which might be due to regional differences in
meteorology, sources, and topography.
2.1.1.2. Spatial Variability on the Urban and Neighborhood Scales
In general, PM2 5 has a longer atmospheric lifetime than PMi0_2.5. As a result, PM2 5 is more
homogeneously distributed than PMi0_2.5, whose concentrations more closely reflect proximity to
local sources (Section 3.5.1.2). Because PMi0 encompasses PMi0_2.5 in addition to PM25, it also
exhibits more spatial heterogeneity than PM2 5. Urban- and neighborhood-scale variability in PM
mass and composition was examined by focusing on 15 metropolitan areas, which were chosen
based on their geographic distribution and coverage in recent health effects studies. The urban areas
selected were Atlanta, Birmingham, Boston, Chicago, Denver, Detroit, Houston, Los Angeles, New
York, Philadelphia, Phoenix, Pittsburgh, Riverside, Seattle and St. Louis. Inter-monitor correlation
remained higher over long distances for PM25 as compared with PMi0 in these 15 urban areas. To a
large extent, greater variation in PM2 5 and PMi0 concentrations within cities was observed in areas
with lower ratios of PM25 to PM10. When the data was limited to only sampler pairs with less than
4 km separation (i.e., on a neighborhood scale), inter-sampler correlations remained higher for PM2 5
than for PMi0. The average inter-sampler correlation was 0.93 for PM25, while it dropped to 0.70 for
PMio (Section 3.5.1.3). Insufficient data were available in the 15 metropolitan areas to perform
similar analyses for PMi0_2.5 using co-located, low volume FRM monitors.
As previously mentioned, UFPs are not measured as part of AQS or any other routine
regulatory network in the U.S. Therefore, information about the spatial variability of UFPs is sparse;
however, their number concentrations are expected to be highly spatially and temporally variable.
This has been shown on the urban scale in studies in which UFP number concentrations drop off
quickly with distance from roads compared to accumulation mode particle numbers.
2.1.2. Trends and Temporal Variability
Overall, PM25 concentrations decreased from 1999 (the beginning of nationwide monitoring
for PM2.5) to 2007 in all ten EPA Regions, with the 3-yr avg of the 98th percentile of 24-h PM2.5
concentrations dropping 10% over this time period. However from 2002-2007, concentrations of
PM25 were nearly constant with decreases observed in only some EPA Regions (Section 3.5.2.1).
Concentrations of PM25 components were only available for 2002-2007 using CSN data and showed
little decline over this time period. This trend in PM2 5 components is consistent with trends in PM2 5
mass concentration observed after 2002 (shown in Figures 3-44 through 3-47). Concentrations of
PMi0 also declined from 1988 to 2007 in all ten EPA Regions.
Using hourly PM observations in the 15 metropolitan areas, diel variation showed average
hourly peaks that differ by size fraction and region (Section 3.5.2.3). For both PM25 and PMi0, a
morning peak was typically observed starting at approximately 6:00 a.m., corresponding with the
start of morning rush hour. There was also an evening concentration peak that was broader than the
morning peak and extended into the overnight period, reflecting the concentration increase caused by
the usual collapse of the mixing layer after sundown. The magnitude and duration of these peaks
varied considerably by metropolitan area investigated.
UFPs were found to exhibit similar two-peaked diel patterns in Los Angeles and the San
Joaquin Valley of CA and Rochester, NY as well as in Kawasaki City, Japan, and Copenhagen,
Denmark. The morning peak in UFPs likely represents primary source emissions, such as rush-hour
traffic, while the afternoon peak likely represents the combination of primary source emissions and
nucleation of new particles.
December 2009 2-3
-------
2.1.3. Correlations between Copollutants
Correlations between PM and gaseous copollutants, including SO2, NO2, carbon monoxide
(CO) and O3, varied both seasonally and spatially between and within metropolitan areas
(Section 3.5.3). On average, PM2.5 and PMi0 were correlated with each other better than with the
gaseous copollutants. Although data are limited for PMi0_2.5, the available data suggest a stronger
correlation between PM10 and PM 10.2.5 than between PM2 5 and PM10_2.5 on a national basis.There was
relatively little seasonal variability in the mean correlation between PM in both size fractions and
SO2 and NO2. CO, however, showed higher correlations with PM2 5 and PMi0 on average in the
winter compared with the other seasons. This seasonality results in part because a larger fraction of
PM is primary in origin during the winter. To the extent that this primary component of PM is
associated with common combustion sources of NO2 and CO, then higher correlations with these
gaseous copollutants are to be expected. Increased atmospheric stability in colder months also results
in higher correlations between primary pollutants (Section 3.5).
The correlation between daily maximum 8-h avg O3 and 24-h avg PM2 5 showed the highest
degree of seasonal variability with positive correlations on average in summer (avg = 0.56) and
negative correlations on average in the winter (avg = -0.30). During the transition seasons, spring
and fall, correlations were mixed but on average were still positive. PM2 5 is both primary and
secondary in origin, whereas O3 is only secondary. Photochemical production of O3 and secondary
PM in the planetary boundary layer (PBL) is much slower during the winter than during other
seasons. Primary pollutant concentrations (e.g., primary PM2 5 components, NO and NO2) in many
urban areas are elevated in winter as the result of heating emissions, cold starts and low mixing
heights. O3 in the PBL during winter is mainly associated with air subsiding from above the
boundary layer following the passage of cold fronts, and this subsiding air has much lower PM
concentrations than are present in the PBL. Therefore, a negative association between O3 andPM25
is frequently observed in the winter. During summer, both O3 and secondary PM2 5 are produced in
the PBL and in the lower free troposphere at faster rates compared to winter, and so they tend to be
positively correlated.
2.1.4. Measurement Techniques
The federal reference methods (FRMs) for PM2 5 and PM10 are based on criteria outlined in the
Code of Federal Regulations. They are, however, subject to several limitations that should be kept in
mind when using compliance monitoring data for health studies. For example, FRM techniques are
subject to the loss of semi-volatile species such as organic compounds and ammonium nitrate
(especially in the West). Since FRMs based on gravimetry use 24-h integrated filter samples to
collect PM mass, no information is available for variations over shorter averaging times from these
instruments. However, methods have been developed to measure real-time PM mass concentrations.
Real-time (or continuous and semi-continuous) measurement techniques are also available for PM
species, such as particle into liquid sampler (PILS) for multiple ions analysis and aerosol mass
spectrometer (AMS) for multiple components analysis (Section 3.4.1). Advances have also been
achieved in PM organic speciation. New 24-h FRMs and Federal Equivalent Methods (FEMs) based
on gravimetry and continuous FEMs for PMi0_2.5 are available. FRMs for PMi0_2.5 rely on calculating
the difference between co-located PM10 and PM2 5 measurements while a dichotomous sampler is
designated as an FEM.
2.1.5. PM Formation in the Atmosphere and Removal
PM in the atmosphere contains both primary (i.e., emitted directly by sources) and secondary
components, which can be anthropogenic or natural in origin. Secondary PM components can be
produced by the oxidation of precursor gases such as SO2 and NOX to acids followed by
neutralization with ammonia (NH3) and the partial oxidation of organic compounds. In addition to
being emitted as primary particles, UFPs are produced by the nucleation of H2SO4 vapor, H2O vapor,
and perhaps NH3 and certain organic compounds. Over most of the earth's surface, nucleation is
probably the major mechanism forming new UFPs. New UFP formation has been observed in
environments ranging from relatively unpolluted marine and continental environments to polluted
December 2009 2-4
-------
urban areas as an ongoing background process and during nucleation events. However, as noted
above, a large percentage of UFPs come from combustion-related sources such as motor vehicles.
Developments in the chemistry of formation of secondary organic aerosol (SOA) indicate that
oligomers are likely a major component of OC in aerosol samples. Recent observations also suggest
that small but significant quantities of SOA are formed from the oxidation of isoprene in addition to
the oxidation of terpenes and organic hydrocarbons with six or more carbon atoms. Gasoline engines
have been found to emit a mix of nucleation-mode heavy and large poly cyclic aromatic
hydrocarbons on which unspent fuel and trace metals can condense, while diesel particles are
composed of a soot nucleus on which sulfates and hydrocarbons can condense. To the extent that the
primary component of organic aerosol is overestimated in emissions from combustion sources, the
semi-volatile components are underestimated. This situation results from the lack of capture of
evaporated semi-volatile components upon dilution in common emissions tests. As a result, near-
traffic sources of precursors to SOA would be underestimated. The oxidation of these precursors
results in more oxidized forms of SOA than previously considered, in both near source urban
environments and further downwind. Primary organic aerosol can also be further oxidized to forms
that have many characteristics in common with oxidized SOA formed from gaseous precursors.
Organic peroxides constitute a significant fraction of SOA and represent an important class of
reactive oxygen species (ROS) that have high oxidizing potential. More information on sources,
emissions and deposition of PM are included in Section 3.3.
Wet and dry deposition are important processes for removing PM and other pollutants from the
atmosphere on urban, regional, and global scales. Wet deposition includes incorporation of particles
into cloud droplets that fall as rain (rainout) and collisions with falling rain (washout). Other
hydrometeors (snow, ice) can also serve the same purpose. Dry deposition involves transfer of
particles through gravitational settling and/or by impaction on surfaces by turbulent motions. The
effects of deposition of PM on ecosystems and materials are discussed in Section 2.5 and in
Chapter 9.
2.1.6. Source Contributions to PM
Results of receptor modeling calculations indicate that PM2.5 is produced mainly by
combustion of fossil fuel, either by stationary sources or by transportation. A relatively small number
of broadly defined source categories, compared to the total number of chemical species that typically
are measured in ambient monitoring source receptor studies, account for the majority of the observed
PM mass. Some ambiguity is inherent in identifying source categories. For example, quite different
mobile sources such as trucks, farm equipment, and locomotives rely on diesel engines and ancillary
data is often required to resolve these sources. A compilation of study results shows that secondary
SO42~ (derived mainly from SO2 emitted by Electricity Generating Units [EGUs]), NO3~ (from the
oxidation of NOX emitted mainly from transportation sources and EGUs), and primary mobile source
categories, constitute most of PM2.5 (and PMi0) in the East. PMi0_2.5 is mainly primary in origin,
having been emitted as fully formed particles derived from abrasion and crushing processes, soil
disturbances, plant and insect fragments, pollens and other microorganisms, desiccation of marine
aerosol emitted from bursting bubbles, and hygroscopic fine PM expanding with humidity to coarse
mode. Gases such as HNO3 can also condense directly onto preexisting coarse particles. Suspended
primary coarse PM can contain Fe, Si, Al, and base cations from soil, plant and insect fragments,
pollen, fungal spores, bacteria, and viruses, as well as fly ash, brake lining particles, debris, and
automobile tire fragments. Quoted uncertainties in the source apportionment of constituents in
ambient aerosol samples typically range from 10 to 50%. An intercomparison of source
apportionment techniques indicated that the same major source categories of PM25 were consistently
identified by several independent groups working with the same data sets. Soil-, sulfate-, residual
oil-, and salt-associated mass were most clearly identified by the groups. Other sources with more
ambiguous signatures, such as vegetative burning and traffic-related emissions were less consistently
identified.
Spatial variability in source contributions across urban areas is an important consideration in
assessing the likelihood of exposure error in epidemiologic studies relating health outcomes to
sources. Concepts similar to those for using ambient concentrations as surrogates for personal
exposures apply here. Some source attribution studies for PM2 5 indicate that intra-urban variability
increases in the following order: regional sources (e.g., secondary SO42~ originating from EGUs)
< area sources (e.g., on-road mobile sources) < point sources (e.g., metals from stacks of smelters).
December 2009 2-5
-------
Although limited information was available for PM10_2.5, it does indicate a similar ordering, but
without a regional component (resulting from the short lifetime of PMi0_2.5 compared to transport
times on the regional scale). More discussion on source contributions to PM is available in
Section 3.6.
2.1.7. Policy-Relevant Background
The background concentrations of PM that are useful for risk and policy assessments, which
inform decisions about the NAAQS are referred to as policy-relevant background (PRB)
concentrations. PRB concentrations have historically been defined by EPA as those concentrations
that would occur in the U.S. in the absence of anthropogenic emissions in continental North America
defined here as the U.S., Canada, and Mexico. For this document, PRB concentrations include
contributions from natural sources everywhere in the world and from anthropogenic sources outside
continental North America. Background concentrations so defined facilitated separation of pollution
that can be controlled by U.S. regulations or through international agreements with neighboring
countries from those that were judged to be generally uncontrollable by the U.S. Over time,
consideration of potential broader ranging international agreements may lead to alternative
determinations of which PM source contributions should be considered by EPA as part of PRB.
Contributions to PRB concentrations of PM include both primary and secondary natural and
anthropogenic components. For this document, PRB concentrations of PM2.5 for the continental U.S.
were estimated using EPA's Community Multi-scale Air Quality (CMAQ) modeling system, a
deterministic, chemical-transport model (CTM), using output from GEOS-Chem a global-scale
model for CMAQ boundary conditions. PRB concentrations of PM2.5 were estimated to be less than
1 (ig/m3 on an annual basis, with maximum daily average values in a range from 3.1 to 20 (ig/m3 and
having a peak of 63 (ig/m3 at the nine national park sites across the U.S. used to evaluate model
performance for this analysis. A description of the models and evaluation of their performance is
given in Section 3.6 and further details about the calculations of PRB concentrations are given in
Section 3.7.
2.2. Human Exposure
This section summarizes the findings from the recent exposure assessment literature. This
summary is intended to support the interpretation of the findings from epidemiologic studies and
reflects the material presented in Section 3.8. Attention is given to how concentration metrics can be
used in exposure assessment and what errors and uncertainties are incurred for different approaches.
Understanding of exposure errors is important because exposure error can potentially bias an
estimate of a health effect or increase the size of confidence intervals around a health effect estimate.
2.2.1. Spatial Scales of PM Exposure Assessment
Assessing population-level exposure at the urban scale is particularly relevant for time-series
epidemiologic studies, which provide information on the relationship between health effects and
community-average exposure, rather than an individual's exposure. PM concentrations measured at a
central-site ambient monitor are used as surrogates for personal PM exposure. However, the
correlation between the PM concentration measured at central-site ambient monitor(s) and the
unknown true community average concentration depends on the spatial distribution of PM, the
location of the monitoring site(s) chosen to represent the community average, and division of the
community by terrain features or local sources into several sub-communities that differ in the
temporal pattern of pollution. Concentrations of SO42~ and some components of SOA measured at
central-site monitors are expected to be uniform in urban areas because of the regional nature of their
sources. However, this is not true for primary components like EC whose sources are strongly
spatially variable in urban areas.
At micro-to-neighborhood scales, heterogeneity of sources and topography contribute to
variability in exposure. This is particularly true for PMi0_2.5 and for UFPs, which have spatially
December 2009 2-6
-------
variable urban sources and loss processes (mainly gravitational settling for PM10_2.5 and coagulation
for UFPs) that also limit their transport from sources more readily than for PM2 5. Personal activity
patterns also vary across urban areas and across regions. Some studies, conducted mainly in Europe,
have found personal PM2.5 and PMi0 exposures for pedestrians in street canyons to be higher than
ambient concentrations measured by urban central site ambient monitors. Likewise,
microenvironmental UFP concentrations were observed to be substantially higher in near-road
environments, street canyons, and tunnels when compared with urban background concentrations.
In-vehicle UFP and PM2.5 exposures can also be important. As a result, concentrations measured by
ambient monitors likely do not reflect the contributions of UFP or PM25 exposures to individuals
while commuting.
There is significant variability within and across regions of the country with respect to indoor
exposures to ambient PM. Infiltrated ambient PM concentrations depend in part on the ventilation
properties of the building or vehicle in which the person is exposed. PM infiltration factors depend
on particle size, chemical composition, season, and region of the country. Infiltration can best be
modeled dynamically rather than being represented by a single value. Season is important to PM
infiltration because it affects the ventilation practices (e.g., open windows) used. In addition, ambient
temperature and humidity conditions affect the transport, dispersion, and size distribution of PM.
Residential air exchange rates have been observed to be higher in the summer for regions with low
air conditioning usage. Regional differences in air exchange rates (Southwest < Southeast
< Northeast < Northwest) also reflect ventilation practices. Differential infiltration occurs as a
function of PM size and composition (the latter of which is described below). PM infiltration is
larger for accumulation mode particles than for UFPs and PMi0_2 5. Differential infiltration by size
fraction can affect exposure estimates if not accurately characterized.
2.2.2. Exposure to PM Components and Copollutants
Emission inventories and source apportionment studies suggest that sources of PM exposure
vary by region. Comparison of studies performed in the eastern U.S. with studies performed in the
western U.S. suggest that the contribution of SO42~ to exposure is higher for the East (16-46%)
compared with the West (-4%) and that motor vehicle emissions and secondary NO3~ are larger
sources of exposure for the West (-9%) as compared with the East (-4%). Results of source
apportionment studies of exposure to SO42~ indicate that SO42~ exposures are mainly attributable to
ambient sources. Source apportionment for OC and EC is difficult because they originate from both
indoor and outdoor sources. Exposure to OC of indoor and outdoor origin can be distinguished by
the presence of aliphatic C-H groups generated indoors, since outdoor concentrations of aliphatic
C-H are low. Studies of personal exposure to ambient trace metal have shown significant variation
among cities and over seasons. This is in response to geographic and seasonal variability in sources
including incinerator operation, fossil fuel combustion, biomass combustion (wildfires), and the
resuspension of crustal materials in the built environment. Differential infiltration is also affected by
variations in particle composition and volatility. For example, EC infiltrates more readily than OC.
This can lead to outdoor-indoor differentials in PM composition.
Some studies have explored the relationship between PM and copollutant gases and suggested
that certain gases can serve as surrogates for describing exposure to other air pollutants. The findings
indicate that ambient concentrations of gaseous copollutants can act as surrogates for personal
exposure to ambient PM. Several studies have concluded that ambient concentrations of O3, NO2,
and SO2 are associated with the ambient component of personal exposure to total PM25. If
associations between ambient gases and personal exposure to PM2 5 of ambient origin exist, such
associations are complex and vary by season and location.
2.2.3. Implications for Epidemiologic Studies
In epidemiologic studies, exposure may be estimated using various approaches, most of which
rely on measurements obtained using central site monitors. The magnitude and direction of the
biases introduced through error in exposure measurement depend on the extent to which the error is
associated with the measured PM concentration. In general, when exposure error is not strongly
correlated with the measured PM concentration, bias is toward the null and effect estimates are
December 2009 2-7
-------
underestimated. Moreover, lack of information regarding exposure measurement error can also add
uncertainty to the health effects estimate.
One important factor to be considered is the spatial variation in PM concentrations. The degree
of urban-scale spatial variability in PM concentrations varies across the country and by size fraction.
PM2.5 concentrations are relatively well-correlated across monitors in the urban areas examined for
this assessment. The limited available evidence indicates that there is greater spatial variability in
PMio_2.5 concentrations than PM2.5 concentrations, resulting in increased exposure error for the larger
size fraction. Likewise, studies have shown UFPs to be more spatially variable across urban areas
compared to PM2.5. Even if PM2.5, PMi0_2.5, or UFP concentrations measured at sites within an urban
area are generally highly correlated, significant spatial variation in their concentrations can occur on
any given day. In addition, there can be differential exposure errors for PM components (e.g., SO42~,
OC, EC). Current information suggests that UFPs, PMi0_2.5s and some PM components are more
spatially variable than PM25. Spatial variability of these PM indicators adds uncertainty to exposure
estimates.
Overall, recent studies generally confirm and build upon the key conclusions of the 2004 PM
AQCD: separation of total PM exposures into ambient and nonambient components reduces
potential uncertainties in the analysis and interpretation of PM health effects data; and ambient PM
concentration can be used as a surrogate for ambient PM exposure in community time-series
epidemiologic studies because the change in ambient PM concentration should be reflected in the
change in the health risk coefficient. The use of the community average ambient PM25 concentration
as a surrogate for the community average personal exposure to ambient PM2 5 is not expected to
change the principal conclusions from time-series and most panel epidemiologic studies that use
community average health and pollution data. Several recent studies support this by showing how
the ambient component of personal exposure to PM25 could be estimated using various tracer and
source apportionment techniques and by showing that the ambient component is highly correlated
with ambient concentrations of PM25. These studies show that the non-ambient component of
personal exposure to PM2 5 is largely uncorrelated with ambient PM2 5 concentrations. A few panel
epidemiologic studies have included personal as well as ambient monitoring data, and generally
reported associations with all types of PM measurements. Epidemiologic studies of long-term
exposure typically exploit the differences in PM concentration across space, as well as time, to
estimate the effect of PM on the health outcome of interest. Long-term exposure estimates are most
accurate for pollutants that do not vary substantially within the geographic area studied.
2.3. Health Effects
This section evaluates the evidence from toxicological, controlled human exposure, and
epidemiologic studies that examined the health effects associated with short- and long-term exposure
to PM (i.e., PM25, PMio_2.5 and UFPs). The results from the health studies evaluated in combination
with the evidence from atmospheric chemistry and exposure assessment studies contribute to the
causal determinations made for the health outcomes discussed in this assessment (a description of
the causal framework can be found in Section 1.5.4). In the following sections a discussion of the
causal determinations will be presented by PM size fraction and exposure duration (i.e., short- or
long-term exposure) for the health effects for which sufficient evidence was available to conclude a
causal, likely to be causal or suggestive relationship. Although not presented in depth in this chapter,
a detailed discussion of the underlying evidence used to formulate each causal determination can be
found in Chapters 6 and 7.
December 2009 2-8
-------
2.3.1. Exposure to PM2.s
2.3.1.1. Effects of Short-Term Exposure to PM2.5
Table 2-1. Summary of causal determinations for short-term exposure to PM2.6.
Size Fraction
Outcome
Causality Determination
Cardiovascular Effects
Causal
PM2
Respiratory Effects
Likely to be causal
Mortality
Causal
Cardiovascular Effects
Epidemiologic studies that examined the effect of PM2.5 on cardiovascular emergency
department (ED) visits and hospital admissions reported consistent positive associations
(predominantly for ischemic heart disease [IHD] and congestive heart failure [CHF]), with the
majority of studies reporting increases ranging from 0.5 to 3.4% per 10 ug/m3 increase in PM25.
These effects were observed in study locations with mean1 24-h avg PM2.5 concentrations ranging
from 7-18 ug/m3 (Section 6.2.10). The largest U.S.-based multicity study evaluated, Medicare Air
Pollution Study (MCAPS), provided evidence of regional heterogeneity (e.g., the largest excess risks
occurred in the Northeast [1.08%]) and seasonal variation (e.g., the largest excess risks occurred
during the winter season [1.49%]) in PM2.5 cardiovascular disease (CVD) risk estimates, which is
consistent with the null findings of several single-city studies conducted in the western U.S. These
associations are supported by multicity epidemiologic studies that observed consistent positive
associations between short-term exposure to PM2 5 and cardiovascular mortality and also reported
regional and seasonal variability in risk estimates. The multicity studies evaluated reported
consistent increases in cardiovascular mortality ranging from 0.47 to 0.85% in study locations with
mean 24-h avg PM25 concentrations above 12.8 ug/m3 (Table 6-14).
Controlled human exposure studies have demonstrated PM2 5-induced changes in various
measures of cardiovascular function among healthy and health-compromised adults. The most
consistent evidence is for altered vasomotor function following exposure to diesel exhaust (DE) or
CAPs with O3 (Section 6.2.4.2). Although these findings provide biological plausibility for the
observations from epidemiologic studies, the fresh DE used in the controlled human exposure
studies evaluated contains gaseous components (e.g., CO, NOX), and therefore, the possibility that
some of the changes in vasomotor function might be due to gaseous components cannot be ruled out.
Furthermore, the prevalence of UFPs in fresh DE limits the ability to conclusively attribute the
observed effects to either the UF fraction or PM25 as a whole. An evaluation of toxicological studies
found evidence for altered vessel tone and microvascular reactivity, which provide coherence and
biological plausibility for the vasomotor effects that have been observed in both the controlled
human exposure and epidemiologic studies (Section 6.2.4.3). However, most of these toxicological
studies exposed animals via intratracheal (IT) instillation or using relatively high inhalation
concentrations.
In addition to the effects observed on vasomotor function, myocardial ischemia has been
observed across disciplines through PM2 5 effects on ST-segment depression, with toxicological
studies providing biological plausibility by demonstrating reduced blood flow during ischemia
(Section 6.2.3). There is also a growing body of evidence from controlled human exposure and
toxicological studies demonstrating PM2 5-induced changes on heart rate variability (HRV) and
1 In this context mean represents the arithmetic mean of 24-h avg PM concentrations.
December 2009
2-9
-------
markers of systemic oxidative stress (Sections 6.2.1 and 6.2.9, respectively). Additional but
inconsistent effects of PM2.5 on blood pressure (BP), blood coagulation markers, and markers of
systemic inflammation have also been reported across disciplines. Toxicological studies have
provided biologically plausible mechanisms (e.g., increased right ventricular pressure and
diminished cardiac contractility) for the associations observed between PM2 5 and CHF in
epidemiologic studies.
Together, the collective evidence from epidemiologic, controlled human exposure, and
toxicoiogicai studies is sufficient to conclude that a causal relationship exists between short-
term exposures to PM and cardiovascular effects.
Respiratory Effects
The recent epidemiologic studies evaluated report consistent positive associations between
short-term exposure to PM2.5 and respiratory ED visits and hospital admissions for chronic
obstructive pulmonary disease (COPD) and respiratory infections (Section 6.3). Positive associations
were also observed for asthma ED visits and hospital admissions for adults and children combined,
but effect estimates are imprecise and not consistently positive for children alone. Most studies
reported effects in the range of ~1% to 4% increase in respiratory hospital admissions and ED visits
and were observed in study locations with mean 24-h avg PM2.5 concentrations ranging from
6.1-22 (ig/m3. Additionally, multicity epidemiologic studies reported consistent positive associations
between short-term exposure to PM2 5 and respiratory mortality as well as regional and seasonal
variability in risk estimates. The multicity studies evaluated reported consistent, precise increases in
respiratory mortality ranging from 1.67 to 2.20% in study locations with mean 24-h avg PM25
concentrations above 12.8 ug/m3 (Table 6-14). Evidence for PM25-related respiratory effects was
also observed in panel studies, which indicate associations with respiratory symptoms, pulmonary
function, and pulmonary inflammation among asthmatic children. Although not consistently
observed, some controlled human exposure studies have reported small decrements in various
measures of pulmonary function following controlled exposures to PM25 (Section 6.3.2.2).
Controlled human exposure studies using adult volunteers have demonstrated increased
markers of pulmonary inflammation following exposure to a variety of different particle types;
oxidative responses to DE and wood smoke; and exacerbations of allergic responses and allergic
sensitization following exposure to DE particles (Section 6.3). Toxicological studies have provided
additional support for PM2 5-related respiratory effects through inhalation exposures of animals to
CAPs, DE, other traffic-related PM and wood smoke. These studies reported an array of respiratory
effects including altered pulmonary function, mild pulmonary inflammation and injury, oxidative
responses, airway hyperresponsiveness (AHR) in allergic and non-allergic animals, exacerbations of
allergic responses, and increased susceptibility to infections (Section 6.3).
Overall, the evidence for an effect of PM25 on respiratory outcomes is somewhat restricted by
limited coherence between some of the findings from epidemiologic and controlled human exposure
studies for the specific health outcomes reported and the sub-populations in which those health
outcomes occur. Epidemiologic studies have reported variable results among specific respiratory
outcomes, specifically in asthmatics (e.g., increased respiratory symptoms in asthmatic children, but
not increased asthma hospital admissions and ED visits) (Section 6.3.8). Additionally, respiratory
effects have not been consistently demonstrated following controlled exposures to PM2 5 among
asthmatics or individuals with COPD. Collectively, the epidemiologic, controlled human exposure,
and toxicoiogicai studies evaluated demonstrate a wide range of respiratory responses, and although
results are not fully consistent and coherent across studies the evidence is sufficient to conclude that
a causal relationship is likely to exist between short-term exposures to PM and
respiratory effects.
Mortality
An evaluation of the epidemiologic literature indicates consistent positive associations
between short-term exposure to PM2 5 and all-cause, cardiovascular-, and respiratory-related
mortality (Section 6.5.2.2.). The evaluation of multicity studies found that consistent and precise risk
estimates for all-cause (nonaccidental) mortality that ranged from 0.29 to 1.21% per 10 (ig/m3
December 2009 2-10
-------
increase in PM2.5 at lags of 1 and 0-1 days. In these study locations, mean 24-h avg PM2.5
concentrations were 12.8 (ig/m3 and above (Table 6-14). Cardiovascular-related mortality risk
estimates were found to be similar to those for all-cause mortality; whereas, the risk estimates for
respiratory-related mortality were consistently larger (i.e., 1.01-2.2%) using the same lag periods and
averaging indices. The studies evaluated that examined the relationship between short-term exposure
to PM2.5 and cardiovascular effects (Section 6.2) provide coherence and biological plausibility for
PM2.5-induced cardiovascular mortality, which represents the largest component of total
(nonaccidental) mortality (~ 35%) (American Heart Association, 2009, 198920). However, as noted
in Section 6.3, there is limited coherence between some of the respiratory morbidity findings from
epidemiologic and controlled human exposure studies for the specific health outcomes reported and
the subpopultions in which those health outcomes occur, complicating the interpretation of the PM25
respiratory mortality effects observed. Regional and seasonal patterns in PM2 5 risk estimates were
observed with the greatest effect estimates occurring in the eastern U.S. and during the spring. Of the
studies evaluated only Burnett et al. (2004, 086247), a Canadian multicity study, analyzed gaseous
pollutants and found mixed results, with possible confounding of PM2 5 risk estimates by NO2.
Although the recently evaluated U.S.-based multicity studies did not analyze potential confounding
of PM25 risk estimates by gaseous pollutants, evidence from the limited number of single-city
studies evaluated in the 2004 PM AQCD (U.S. EPA, 2004, 056905) suggest that gaseous
copollutants do not confound the PM2 5-mortality association. This is further supported by studies
that examined the PMio-mortality relationship. An examination of effect modifiers (e.g.,
demographic and socioeconomic factors), specifically air conditioning use as an indicator for
decreased pollutant penetration indoors, has suggested that PM2 5 risk estimates increase as the
percent of the population with access to air conditioning decreases. Collectively, the epidemiologic
literature provides evidence that a causal relationship exists between short-term exposures
to PM2.5 and mortality.
2.3.1.2. Effects of Long-Term Exposure to PM2.5
Table 2-2. Summary of causal determinations for long-term exposure to PM2.6.
Size Fraction
PM2.5
Outcome
Cardiovascular Effects
Respiratory Effects
Mortality
Reproductive and Developmental
Cancer, Mutagenicity, and Genotoxicity
Causality Determination
Causal
Likely to be causal
Causal
Suggestive
Suggestive
Cardiovascular Effects
The strongest evidence for cardiovascular health effects related to long-term exposure to PM2 5
comes from large, multicity U.S.-based studies, which provide consistent evidence of an association
between long-term exposure to PM25 and cardiovascular mortality (Section 7.2.10). These
associations are supported by a large U.S.-based epidemiologic study (i.e., Women's Health Initiative
[WHI] study) that reports associations between PM2 5 and CVDs among post-menopausal women
using a 1-yr avg PM25 concentration (mean = 13.5 (ig/m3) (Section 7.2). However, epidemiologic
studies that examined subclinical markers of CVD report inconsistent findings. Epidemiologic
studies have also provided some evidence for potential modification of the PM25-CVD association
when examining individual-level data, specifically smoking status and the use of anti-
December 2009 2-11
-------
hyperlipidemics. Although epidemiologic studies have not consistently detected effects on markers
of atherosclerosis due to long-term exposure to PM2.5, toxicological studies have provided strong
evidence for accelerated development of atherosclerosis in ApoE~'~ mice exposed to CAPs and have
shown effects on coagulation, experimentally-induced hypertension, and vascular reactivity (Section
7.2.1.2). Evidence from toxicological studies provides biological plausibility and coherence with
studies of short-term exposure and cardiovascular morbidity and mortality, as well as with studies
that examined long-term exposure to PM2.5 and cardiovascular mortality. Taken together, the
evidence from epidemiologic and toxicological studies is sufficient to conclude that 3 C3US3I
relationship exists between long-term exposures to PM and cardiovascular effects.
Respiratory Effects
Recent epidemiologic studies conducted in the U.S. and abroad provide evidence of
associations between long-term exposure to PM2.5 and decrements in lung function growth, increased
respiratory symptoms, and asthma development in study locations with mean PM2.5 concentrations
ranging from 13.8 to 30 ug/m3 during the study periods (Section 7.3.1.1 and Section 7.3.2.1). These
results are supported by studies that observed associations between long-term exposure to PMi0 and
an increase in respiratory symptoms and reductions in lung function growth in areas where PMi0 is
dominated by PM2 5. However, the evidence to support an association with long-term exposure to
PM2 5 and respiratory mortality is limited (Figure 7-8). Subchronic and chronic toxicological studies
of CAPs, DE, roadway air and woodsmoke provide coherence and biological plausibility for the
effects observed in the epidemiologic studies. These toxicological studies have presented some
evidence for altered pulmonary function, mild inflammation, oxidative responses, immune
suppression, and histopathological changes including mucus cell hyperplasia (Section 7.3).
Exacerbated allergic responses have been demonstrated in animals exposed to DE and wood smoke.
In addition, pre- and postnatal exposure to ambient levels of urban particles was found to affect lung
development in an animal model. This finding is important because impaired lung development is
one mechanism by which PM exposure may decrease lung function growth in children. Collectively,
the evidence from epidemiologic and toxicological studies is sufficient to conclude that 3 C3US3I
relationship is likely to exist between long-term exposures to PM25 and respiratory
effects.
Mortality
The recent epidemiologic literature reports associations between long-term PM2 5 exposure and
increased risk of mortality. Mean PM25 concentrations ranged from 13.2 to 29 (ig/m3 during the
study period in these areas (Section 7.6). When evaluating cause-specific mortality, the strongest
evidence can be found when examining associations between PM2 5 and cardiovascular mortality,
and positive associations were also reported between PM2 5 and lung cancer mortality (Figure 7-8).
The cardiovascular mortality association has been confirmed further by the extended Harvard Six
Cities and American Cancer Society studies, which both report strong associations between long-
term exposure to PM2 5 and cardiopulmonary and IHD mortality (Figure 7-7). Additional new
evidence from a study that used the WHI cohort found a particularly strong association between
long-term exposure to PM2 5 and CVD mortality in post-menopausal women. Fewer studies have
evaluated the respiratory component of cardiopulmonary mortality, and, as a result, the evidence to
support an association with long-term exposure to PM2 5 and respiratory mortality is limited (Figure
7-8). The evidence for cardiovascular and respiratory morbidity due to short- and long-term exposure
to PM2 5 provides biological plausibility for cardiovascular- and respiratory-related mortality.
Collectively, the evidence is sufficient to conclude that 3 C3US3I relstionship exists between
long-term exposures to PM25 and mortality.
December 2009 2-12
-------
Reproductive and Developmental Effects
Evidence is accumulating for PM2.5 effects on low birth weight and infant mortality, especially
due to respiratory causes during the post-neonatal period. The mean PM2.5 concentrations during the
study periods ranged from 5.3-27.4 ug/m3 (Section 7.4), with effects becoming more precise and
consistently positive in locations with mean PM2.5 concentrations of 15 ug/m3 and above
(Section 7.4). Exposure to PM2.5 was usually associated with greater reductions in birth weight than
exposure to PMi0. The evidence from a few U.S. studies that investigated PMi0 effects on fetal
growth, which reported similar decrements in birth weight, provide consistency for the PM2 5
associations observed and strengthen the interpretation that particle exposure may be causally related
to reductions in birth weight. The epidemiologic literature does not consistently report associations
between long-term exposure to PM and preterm birth, growth restriction, birth defects or decreased
sperm quality. Toxicological evidence supports an association between PM2 5 and PMi0 exposure and
adverse reproductive and developmental outcomes, but provide little mechanistic information or
biological plausibility for an association between long-term PM exposure and adverse birth
outcomes (e.g., low birth weight or infant mortality). New evidence from animal toxicological
studies on heritable mutations is of great interest, and warrants further investigation. Overall, the
epidemiologic and toxicological evidence is suggestive of a causal relationship between long-
term exposures to PM25 and reproductive and developmental outcomes.
Cancer, Mutagenicity, and Genotoxicity
Multiple epidemiologic studies have shown a consistent positive association between PM2 5
and lung cancer mortality, but studies have generally not reported associations between PM2 5 and
lung cancer incidence (Section 7.5). Animal toxicological studies have examined the potential
relationship between PM and cancer, but have not focused on specific size fractions of PM. Instead
they have examined ambient PM, wood smoke, and DEP A number of studies indicate that ambient
urban PM, emissions from wood/biomass burning, emissions from coal combustion, and gasoline
and DE are mutagenic, and that PAHs are genotoxic. These findings are consistent with earlier
studies that concluded that ambient PM and PM from specific combustion sources are mutagenic and
genotoxic and provide biological plausibility for the results observed in the epidemiologic studies. A
limited number of epidemiologic and toxicological studies examined epigenetic effects, and
demonstrate that PM induces some changes in methylation. However, it has yet to be determined
how these alterations in the genome could influence the initiation and promotion of cancer.
Additionally, inflammation and immune suppression induced by exposure to PM may confer
susceptibility to cancer. Collectively, the evidence from epidemiologic studies, primarily those of
lung cancer mortality, along with the toxicological studies that show some evidence of the mutagenic
and genotoxic effects of PM is suggestive of a causal relationship between long-term
exposures to PM25 and cancer.
2.3.2. Integration of PM2.5 Health Effects
In epidemiologic studies, short-term exposure to PM2 5 is associated with a broad range of
respiratory and cardiovascular effects, as well as mortality. For cardiovascular effects and mortality,
the evidence supports the existence of a causal relationship with short-term PM2 5 exposure; while the
evidence indicates that a causal relationship is likely to exist between short-term PM2 5 exposure and
respiratory effects. The effect estimates from recent and older U.S. and Canadian-based
epidemiologic studies that examined the relationship between short-term exposure to PM2 5 and
health outcomes with mean 24-h avg PM25 concentrations <17 ug/m3 are shown in Figure 2-1. A
number of different health effects are included in Figure 2-1 to provide an integration of the range of
effects by mean concentration, with a focus on cardiovascular and respiratory effects and all-cause
(nonaccidental) mortality (i.e., health effects categories with at least a suggestive causal
determination). A pattern of consistent positive associations with mortality and morbidity effects can
be seen in this figure. Mean PM25 concentrations ranged from 6.1 to 16.8 ug/m3.in these study
locations.
December 2009 2-13
-------
Study
Outcome
Mean3 98tha
Effect Estimate (95% Cl)
Chimonas & Gessner (2007.093261)
Lisabeth et al. (2008, 155939)
Slaughter et al. (2005, 073854)
Rabinovitch et al. (2006, 088031)
Chen etal. (2004.087262)
Chen etal. (2005,087555)
Fung etal. (2006, 089789)
Villeneuve et al. (2003, 055051)
Stieb etal. (2000.011675)
Villeneuve et al. (2006, 090191)
Lin etal. (2005. 087828)
Mar etal. (2004. 057309)
Rich etal. (2005. 079620)
Dockery etal. (2005, 078995)
Rabinovitch et al. (2004, 096753)
Pope etal. (2006.091246)
Slaughter etal. (2005. 073854)
Pope etal. (2008. 191969)
Zanobetti and Schwartz (2006, 090195)
Peters etal. (2001.016546)
Delfino etal. (1997. 082687)
Sullivan etal. (2005,050854)
Burnett etal. (2004.086247)
Bell etal. (2008. 156266)
Wilson etal. (2007,1 571 49f
Zanobetti & Schwartz (2009, 188462)
Burnett and Goldberg (2003, 042798)
Dominici et al. (2006, 088398)
Fairley (2003, 042850)
Zhang etal. (2009. 191970)
O'Connor et al. (2008, 156818)
Klemm and Mason (2003, 042801)
Franklin etal. (2008, 097426)
NYDOH (2006, 090132)
Ito etal. (2007. 156594)
Franklin etal. (2007,091257)
Rich etal. (2006.089814)
Symons etal. (2006,091258)
Sheppard (2003, 042826)
NYDOH (2006, 090132)
Burnett etal. (1997.084194)
b Study did not present mean; median presented.
Asthma HA 6.1
LRI HA 6.1
IschemicStroke/TIAHA 7.0°
Asthma Exacerbation 7.3°
Asthma Medication Use 7.4
COPD HA 7.7
Respiratory HA 7.7
Respiratory HA 7.7
Nonaccidental Mortality 7.9
CVD ED Visits 8.5
Respiratory ED Visits 8.5
Hemhrgc Stroke HA 8.5
Ischemic Stroke HA 8.5
TIAHA 8.5
RTI HA 9.6
Respiratory Symptoms (any) 9.8C
Respiratory Symptoms (any) 9.8°
Ventricular Arrhythmia 9.8°
Ventricular Arrhythmia 10.3°
Asthma Exacerbation 10.6°
IHDHA 10.7C
CVD HA 10.8
Respiratory ED Visits 10.8
CHFHA 10.8
Ml HA 11.1e
Pneumonia HA 11.1°
Ml 12.1
Respiratory HA (summer) 12.1
Ml 12.8
Nonaccidental Mortality 12.8
Respiratory HA 12.9"
CVD HA 12.9"
CVD Mortality 13.0.
Nonaccidental Mortality 13.2"
Nonaccidental Mortality 13.3
CBVDHA 13.3
PVDHA 13.3
IHDHA 13.3
Dysrhythmia HA 13.3
CHFHA 13.3
COPD HA 13.3
RTIHA 13.3
Nonaccidental Mortality 13.6
ST Segment Depression 13.9
Wheeze/Cough 14.0C.
Nonaccidental Mortality 14.7°''
Nonaccidental mortality 14.8
Asthma ED Visits 15.0"
Asthma HA 15.1
Non-accidental Mortality 15.6
Ventricular Arrhythmia 16.2°
CHFHA 16.5"
Asthma HA 16.7
Asthma ED Visits 167
Respiratory HA (summer) 1 6.8
CVD HA (summer) 16.8
' Averaged annual values for years in study
provided by study author.
23.6r -" — •
17.2r -*•-
07 Or L^^A
273'- (-^
t
on or" * *
29.6r — »~
29.6'. -»*•
*-•-
OP. of -.
38.0f »
34.2r *
34.2' »
rj-t p* t || |__
-31 .O. I 1 1
34.3' »
38.9' »
34.8' *
34.8' *
34.8r fc
34.8' k
34.8' i*
34.8' k
34.8' *
59.0' L+-
37.6f -»-. —
39.09 *i
h
43.0' b
39.0f -•-
45.8' i.
50 1f » i
46.6r -+-
/I7/I' ) (
171' i t
rom data
1 ! 1 1 I 1 1 1 I 1
.
Mean value slightly different from those reported in the published
study or not reported in the published study; mean was either provided
by study authors or calculated from data provided by study authors.
s Mean value not reported in study; median presented.
98th percentile of PM2 5 distribution was either provided by study
authors or calculated from data provided by study authors.
3 98th estimated from data in study.
Air quality data obtained from original study by rt c
Schwartz et al. (1996, 077325) u-°
J Mean PM2 5 concentration reported is for lag 0-2.
k Bronx; TEOM data.
' Manhattan; TEOM data.
m Study does not present an overall effect estimate; the
vertical lines represent the effect estimate for each of the
areas of Phoenix examined.
0.8 1.0 1.2
Relative Risk / Odds Ratio
1.4
Figure 2-1. Summary of effect estimates (per 10 ug/m3) by increasing concentration from U.S.
studies examining the association between short-term exposure to PM2.6 and
cardiovascular and respiratory effects, and mortality, conducted in locations
where the reported mean 24-h avg PM2.6 concentrations were <17 ug/m3.
December 2009
2-14
-------
Long-term exposure to PM2.5 has been associated with health outcomes similar to those found
in the short-term exposure studies, specifically for respiratory and cardiovascular effects and
mortality. As found for short-term PM2.5 exposure, the evidence indicates that a causal relationship
exists between long-term PM2.5 exposure and cardiovascular effects and mortality, and that a causal
relationship is likely to exist between long-term PM2.5 exposure and effects on the respiratory
system.
Figure 2-2 highlights the findings of epidemiologic studies where the long-term mean PM2 5
concentrations were < 29 ug/m3. A range of health outcomes are displayed (including cardiovascular
mortality, all-cause mortality, infant mortaltiy, and bronchitis) ordered by mean concentration. The
range of mean PM25 concentrations in these studies was 10.7-29 ug/m3 during the study periods.
Additional studies not included in this figure that focus on subclinical outcomes, such as changes in
lung function or atherosclerotic markers also report effects in areas with similar concentrations
(Sections 7.2 and 7.3). Although not highlighted in the summary figure, long-term PM2 5 exposure
studies also provide evidence for reproductive and developmental effects (i.e., low birth weight) and
cancer (i.e., lung cancer mortality) in response to to exposure to PM2 5.
Study
Zeqeretal. (2008. 191951)
Zegeretal. (2008, 191951)
Eftimetal. (2008, 099104)
MrPnnnoll otal fOC\m ClAQAQCh
Zegeretal. (2008, 191951)
Krewski et al. (2009, 191193)
Eftimetal. (2008.099104)
Lipfertetal. (2006, 088756)
FVirkprv pt al f1QQfi n4R91Cfi
Woodruff et al. (2008, 098386)
Laden etal. (2006, 087605)
Wnnrlriiff ptal C9HHR HQR^RQ
Enstrom (2005. 087356)
Chen et al (2005 087942)
* Mean estimated from data in stu
Outcome
All-Cause Mortality, Central U.S.
All-Cause Mortality, Western U.S.
All-Cause Mortality, ACS Sites
All-Cause Mortality, Eastern U.S.
All-Cause Mortality
All-Cause Mortality, Harv 6-Cities
All-Cause Mortality
Infant Mortality (Respiratory)
All-Cause Mortality
All-Cause Mortality
PHD Mnrtalitv MalpQ
dy
Mearv
10.7
13.1 ~J
13.6
•10 Q
14.0
14.0
14.1
14.3
1/1 c
14.8 J
14.8*
10°
23.4
290
""10
1 1
0.7 0.9
Effect Estimate (95% Cl)
— •—
L
-
—~-
L_
~
1 1 1 1 1 1
1.1 1.3 1.5 1.7 1.9 2.1
Relative Risk
Figure 2-2.
Summary of effect estimates (per 10 ug/m3) by increasing concentration from U.S.
studies examining the association between long-term exposure to PM2.s and
cardiovascular and respiratory effects, and mortality conducted in locations
where the mean annual PM2.s concentration were <17 ug/m3.
The observations from both the short- and long-term exposure studies are supported by
experimental findings of PM25-induced subclinical and clinical cardiovascular effects.
Epidemiologic studies have shown an increase in ED visits and hospital admissions for IHD upon
exposure to PM2 5. These effects are coherent with the changes in vasomotor function and ST-
segment depression observed in both toxicological and controlled human exposure studies. It has
been postulated that exposure to PM2 5 can lead to myocardial ischemia through an effect on the
autonomic nervous system or by altering vasomotor function. PM-induced systemic inflammation,
oxidative stress and/or endothelial dysfunction may contribute to altered vasomotor function. These
effects have been demonstrated in recent animal toxicological studies, along with altered
microvascular reactivity, altered vessel tone, and reduced blood flow during ischemia. Toxicological
studies demonstrating increased right ventricular pressure and diminished cardiac contractility also
provide biological plausibility for the associations observed between PM2 5 and CHF in
epidemiologic studies.
Thus, the overall evidence from the short-term epidemiologic, controlled human exposure, and
toxicological studies evaluated provide coherence and biological plausibility for cardiovascular
effects related to myocardial ischemia and CHF. Coherence in the cardiovascular effects observed
can be found in long-term exposure studies, especially for CVDs among post-menopausal women.
December 2009
2-15
-------
Additional studies provide limited evidence for subclinical measures of atherosclerosis in
epidemiologic studies with stronger evidence from toxicological studies that have demonstrated
accelerated development of atherosclerosis in ApoE"" mice exposed to PM2.5 CAPs along with
effects on coagulation, experimentally-induced hypertension, and vascular reactivity. Repeated acute
responses to PM may lead to cumulative effects that manifest as chronic disease, such as
atherosclerosis. Contributing factors to atherosclerosis development include systemic inflammation,
endothelial dysfunction, and oxidative stress all of which are associated with PM2.5 exposure.
However, it has not yet been determined whether PM initiates or promotes atherosclerosis. The
evidence from both short- and long-term exposure studies on cardiovascular morbidity provide
coherence and biological plausibility for the cardiovascular mortality effects observed when
examining both exposure durations. In addition, cardiovascular hospital admission and mortality
studies that examined the PMi0 concentration-response relationship found evidence of a log-linear
no-threshold relationship between PM exposure and cardiovascular-related morbidity (Section 6.2)
and mortality (Section 6.5).
Epidemiologic studies have also reported respiratory effects related to short-term exposure to
PM2.5, which include increased ED visits and hospital admissions, as well as alterations in lung
function and respiratory symptoms in asthmatic children. These respiratory effects were found to be
generally robust to the inclusion of gaseous pollutants in copollutant models with the strongest
evidence from the higher powered studies (Figure 6-9 and Figure 6-15). Consistent positive
associations were also reported between short-term exposure to PM2 5 and respiratory mortality in
epidemiologic studies. However, uncertainties exist in the PM2 5-respiratory mortality associations
reported due to the limited number of studies that examined potential confounders of the PM25-
respiratory mortality relationship, and the limited information regarding the biological plausibility of
the clinical and subclinical respiratory outcomes observed in the epidemiologic and controlled
human exposure studies (Section 6.3) resulting in the progression to PM25-induced respiratory
mortality. Important new findings, which support the PM2 5-induced respiratory effects mentioned
above, include associations with post-neonatal (between 1 mo and 1 yr of age) respiratory mortality.
Controlled human exposure studies provide some support for the respiratory findings from
epidemiologic studies, with demonstrated increases in pulmonary inflammation following short-term
exposure. However, there is limited and inconsistent evidence of effects in response to controlled
exposures to PM2 5 on respiratory symptoms or pulmonary function among healthy adults or adults
with respiratory disease. Long-term exposure epidemiologic studies provide additional evidence for
PM2^-induced respiratory morbidity, but little evidence for an association with respiratory mortality.
These epidemiologic morbidity studies have found decrements in lung function growth, as well as
increased respiratory symptoms, and asthma. Toxicological studies provide coherence and biological
plausibility for the respiratory effects observed in response to short and long-term exposures to PM
by demonstrating a wide array of biological responses including: altered pulmonary function, mild
pulmonary inflammation and injury, oxidative responses, and histopathological changes in animals
exposed by inhalation to PM25 derived from a wide variety of sources. In some cases, prolonged
exposures led to adaptive responses. Important evidence was also found in an animal model for
altered lung development following pre- and post-natal exposure to urban air, which may provide a
mechanism to explain the reduction in lung function growth observed in children in response to
long-term exposure to PM.
Additional respiratory-related effects have been tied to allergic responses. Epidemiologic
studies have provided evidence for increased hospital admissions for allergic symptoms (e.g.,
allergic rhinitis) in response to short- and long-term exposure to PM2 5. Panel studies also positively
associate long-term exposure to PM2 5 and PMip with indicators of allergic sensitization. Controlled
human exposure and toxicological studies provide coherence for the exacerbation of allergic
symptoms, by showing that PM2 5 can promote allergic responses and intensify existing allergies.
Allergic responses require repeated exposures to antigen over time and co-exposure to an adjuvant
(possibly DE particles or UF CAPs) can enhance this response. Allergic sensitization often underlies
allergic asthma, characterized by inflammation and AHR. In this way, repeated or chronic exposures
involving multifactorial responses (immune system activation, oxidative stress, inflammation) can
lead to irreversible outcomes. Epidemiologic studies have also reported evidence for increased
hospital admissions for respiratory infections in response to both short- and long-term exposures to
PM2 5. Toxicological studies suggest that PM impairs innate immunity, which is the first line of
defense against infection, providing coherence for the respiratory infection effects observed in
epidemiologic studies.
December 2009 2-16
-------
The difference in effects observed across studies and between cities may be attributed, at least
in part, to the differences in PM composition across the U.S. Differences in PM toxicity may result
from regionally varying PM composition and size distribution, which in turn reflects differences in
sources and PM volatility. A person's exposure to ambient PM will also vary due to regional
differences in personal activity patterns, microenvironmental characteristics and the spatial
variability of PM concentrations in urban areas. Regional differences in PM2.5 composition are
outlined briefly in Section 2.1 above and in more detail in Section 3.5. An examination of data from
the CSN indicates that East-West gradients exist for a number of PM components. Specifically, SO42"
concentrations are higher in the East, OC constitutes a larger fraction of PM in the West, and NO3"
concentrations are highest in the valleys of central California and during the winter in the Midwest.
However, the available evidence and the limited amount of city-specific speciated PM2.5 data does
not allow conclusions to be drawn that specifically differentiate effects of PM in different locations.
It remains a challenge to determine relationships between specific constituents, combinations
of constituents, or sources of PM25 and the various health effects observed. Source apportionment
studies of PM2.5 have attempted to decipher some of these relationships and in the process have
identified associations between multiple sources and various respiratory and cardiovascular health
effects, as well as mortality. Although different source apportionment methods have been used across
these studies, the methods used have been evaluated and found generally to identify the same
sources and associations between sources and health effects (Section 6.6). While uncertainty
remains, it has been recognized that many sources and components of PM2 5 contribute to health
effects. Overall, the results displayed in Table 6-17 indicate that many constituents of PM25 can be
linked with multiple health effects, and the evidence is not yet sufficient to allow differentiation of
those constituents or sources that are more closely related to specific health outcomes.
Variability in the associations observed across PM2 5 epidemiologic studies may be due in part
to exposure error related to the use of county-level air quality data. Because western U.S. counties
tend to be much larger and more topographically diverse than eastern U.S. counties, the day-to-day
variations in concentration at one site, or even for the average of several sites, may not correlate well
with the day-to-day variations in all parts of the county. For example, site-to-site correlations as a
function of distance between sites (Section 3.5.1.2) fall off rapidly with distance in Los Angeles, but
high correlations extend to larger distances in eastern cities such as Boston and Pittsburgh. These
differences may be attributed to a number of factors including topography, the built environment,
climate, source characteristics, ventilation usage, and personal activity patterns. For instance,
regional differences in climate and infrastructure can affect time spent outdoors or indoors, air
conditioning usage, and personal activity patterns. Characteristics of housing stock may also cause
regional differences in effect estimates because new homes tend to have lower infiltration factors
than older homes. Biases and uncertainties in exposure estimates resulting from these aspects can, in
turn, cause bias and uncertainty in associated health effects estimates.
The new evidence reviewed in this ISA greatly expands upon the evidence available in the
2004 PM AQCD particularly in providing greater understanding of the underlying mechanisms for
PM2 5 induced cardiovascular and respiratory effects for both short- and long-term exposures. Recent
studies have provided new evidence linking long-term exposure to PM2 5 with cardiovascular
outcomes that has expanded upon the continuum of effects ranging from the more subtle subclinical
measures to cardiopulmonary mortality.
December 2009 2-17
-------
2.3.3. Exposure to PMi0-2.5
2.3.3.1. Effects of Short-Term Exposure to PM10.2.5
Table 2-3. Summary of causal determinations for short-term exposure to PMi0.2.6-
Size Fraction
Outcome
Causality Determination
Cardiovascular Effects
Suggestive
PMi,
Respiratory Effects
Suggestive
Mortality
Suggestive
Cardiovascular Effects
Generally positive associations were reported between short-term exposure to PMi0_2.5 and
hospital admissions or ED visits for cardiovascular causes. These results are supported by a large
U.S. multicity study of older adults that reported PMi0_2.5 associations with CVD hospital
admissions, and only a slight reduction in the PMi0_2.5 risk estimate when included in a copollutant
model with PM2.5 (Section 6.2.10). The PMi0_2.5 associations with cardiovascular hospital admissions
and ED visits were observed in study locations with mean 24-h avg PMi0_2.5 concentrations ranging
from 7.4 to 13 (ig/m3. These results are supported by the associations observed between PMi0_2.5 and
cardiovascular mortality in areas with 24-h avg PM10_2.5 concentrations ranging from 6.1-16.4 (ig/m3
(Section 6.2.11). The results of the epidemiologic studies were further confirmed by studies that
examined dust storm events, which contain high concentrations of crustal material, and found an
increase in cardiovascular-related ED visits and hospital admissions. Additional epidemiologic
studies have reported PMi0_2.5 associations with other cardiovascular health effects including
supraventricular ectopy and changes in HRV (Section 6.2.1.1). Although limited in number, studies
of controlled human exposures provide some evidence to support the alterations in HRV observed in
the epidemiologic studies (Section 6.2.1.2). The few toxicological studies that examined the effect of
PMio_2.5 on cardiovascular health effects used IT instillation due to the technical challenges in
exposing rodents via inhalation to PMi0_2.5, and, as a result, provide only limited evidence on the
biological plausibility of PMi0_2.5 induced cardiovascular effects. The potential for PMi0_2.5 to elicit an
effect is supported by dosimetry studies, which show that a large proportion of inhaled particles in
the 3-6 micron (dae) range can reach and deposit in the lower respiratory tract, particularly the
tracheobronchial (TB) airways (Figures 4-3 and 4-4). Collectively, the evidence from epidemiologic
studies, along with the more limited evidence from controlled human exposure and toxicological
studies is suggestive of a causal relationship between short-term exposures to PMi025 and
cardiovascular effects.
Respiratory Effects
A number of recent epidemiologic studies conducted in Canada and France found consistent,
positive associations between respiratory ED visits and hospital admissions and short-term exposure
to PMio_2.5in studies with mean 24-h avg concentrations ranging from 5.6-16.2 ug/m3 (Section 6.3.8)
In these studies, the strongest relationships were observed among children, with less consistent
evidence for adults and older adults (i.e., > 65). In a large multicity study of older adults, PMi0_2.5
was positively associated with respiratory hospital admissions in both single and copollutant models
with PM25. In addition, a U.S.-based multicity study found evidence for an increase in respiratory
mortality upon short-term exposure to PM10_2.5, but these associations have not been consistently
December 2009
2-18
-------
observed in single-city studies (Section 6.3.9). A limited number of epidemiologic studies have
focused on specific respiratory morbidity outcomes, and found no evidence of an association with
lower respiratory symptoms, wheeze, and medication use (Section 6.3.1.1). While controlled human
exposure studies have not observed an effect on lung function or respiratory symptoms in healthy or
asthmatic adults in response to short-term exposure to PMi0_2.5, healthy volunteers have exhibited an
increase in markers of pulmonary inflammation. Toxicological studies using inhalation exposures are
still lacking, but pulmonary injury has been observed in animals after IT instillation exposure
(Section 6.3.5.3). In some cases, PMi0_2.5 was found to be more potent than PM2.5 and effects were
not attributable to endotoxin. Both rural and urban PMi0_2.5 have induced inflammation and injury
responses in rats or mice exposed via IT instillation, making it difficult to distinguish the health
effects of PM 10-2.5 from different environments. Overall, epidemiologic studies, along with the
limited number of controlled human exposure and toxicological studies that examined PMi0_2.5
respiratory effects provide evidence that is suggestive of a causal relationship between short-
term exposures to PM 10-2.5 and respiratory effects.
Mortality
The majority of studies evaluated in this review provide some evidence for mortality
associations with PMi0_2.5 in areas with mean 24-h avg concentrations ranging from 6.1-16.4 ug/m3.
However, uncertainty surrounds the PMi0_2.5 associations reported in the studies evaluated due to the
different methods used to estimate PMi0_2.5 concentrations across studies (e.g., direct measurement of
PM10_2.5 using dichotomous samplers, calculating the difference between PM10 and PM2.5
concentrations). In addition, only a limited number of PMi0_2.5 studies have investigated potential
confounding by gaseous copollutants or the influence of model specification on PMi0_2.5 risk
estimates.
Anew U.S.-based multicity study, which estimated PMi0_2.5 concentrations by calculating the
difference between the county-average PMi0 and PM2.5, found associations between PMi0_2.5 and
mortality across the U.S., including evidence for regional variability in PM10_2.5 risk estimates
(Section 6.5.2.3). Additionally, the U.S.-based multicity study provides preliminary evidence for
greater effects occurring during the warmer months (i.e., spring and summer). A multicity Canadian
study provides additional evidence for an association between short-term exposure to PMio-2.5 and
mortality (Section 6.5.2.3). Although consistent positive associations have been observed across both
multi- and single-city studies, more data are needed to adequately characterize the chemical and
biological components that may modify the potential toxicity of PMio_2.s and compare the different
methods used to estimate exposure. Overall, the evidence evaluated JS Suggestive Of 3 Causal
relationship between short-term exposures to PM,,, and mortality.
2.3.4. Integration of PM10.2.s Effects
Epidemiologic, controlled human exposure, and toxicological studies have provided evidence that is
suggestive for relationships between short-term exposure to PMi0_2.5 and cardiovascular effects,
respiratory effects, and mortality. Conclusions regarding causation for the various health effects and
outcomes were made for PMi0_2.5 as a whole regardless of origin, since PMi0_2.5-related effects have
been demonstrated for a number of different environments (e.g., cities reflecting a wide range of
environmental conditions). Associations between short-term exposure to PMi0_2.5 and cardiovascular
and respiratory effects, and mortality have been observed in locations with mean PMi0_2.5
concentrations ranging from 5.6 to 33.2 (ig/m3, and maximumPMi0_2.5 concentrations ranging from
24.6 to 418.0 (ig/m ) (Figure 2-3). A number of different health effects are included in Figure 2-3 to
provide an integration of the range of effects by mean concentration, with a focus on cardiovascular
and respiratory effects, and mortality (i.e., health effects categories with at least a suggestive causal
determination). To date, a sufficient amount of evidence does not exist in order to draw conclusions
regarding the health effects and outcomes associated with long-term exposure to PMi0_2.5.
In epidemiologic studies, associations between short-term exposure to PMi0_2.5 and
cardiovascular outcomes (i.e., IHD hospital admissions, supraventricular ectopy, and changes in
HRV) have been found that are similar in magnitude to those observed in PM2.5 studies. Controlled
human exposure studies have also observed alterations in HRV, providing consistency and coherence
December 2009 2-19
-------
for the effects observed in the epidemiologic studies. To date, only a limited number of toxicological
studies have been conducted to examine the effects of PMi0_2.s on cardiovascular effects. All of these
studies involved IT instillation due to the technical challenges of using PMi0_2.5 for rodent inhalation
studies. As a result, the toxicological studies evaluated provide limited biological plausibility for the
PMio_2.5 effects observed in the epidemiologic and controlled human exposure studies.
Study
Outcome
Chen etal. (2004.087262) COPDHA
Fung etal. (2006,089789) RDHA
Chen etal. (2005. 087942) RDHA
Villeneuve et al. (2003, 055051) Nonaccidental Mortality
Lipfert et al. (2000, 004088) CVD Mortality
Peters et al. (2001 , 016546) Ml
Tolbert et al. (2007, 090316) CVD ED Visits
RD ED Visits
Klemm et al. (2003, 042801) Nonaccidental Mortality
Metzger et al. (2007, 092856) Ventricular Arrhythmia
Peel et al. (2005, 056305) Asthma ED Visits
COPD ED Visits
RD ED Visits
Pneumonia ED Visits
URI ED Visits
Metzger etal. (2004,044222) CHF ED Visits
I HD ED Visits
Klemm et al. (2004, 056585) Nonaccidental Mortality
Mar et al. (2004, 057309) Symptoms (any)
Asthma Symptoms
Lin etal. (2005. 087828) RTI HA
Burnett et al. (2004, 086247) Non-accidental Mortality
Burnett et al. (1 997, 084194) CVD HA
Respiratory HA
Fairley (2003, 042850) Nonaccidental Mortality
Zanobetti & Schwartz (2009, 188462) Nonaccidental Mortality
Lin etal. (2002. 026067) Asthma HA (boys)
Lin et al. (2002, 026067; 2004, 056067) Asthma HA (girls)
Peng et al. (2008, 156850) RD HA
CVD HA
Burnett and Goldberg (2003, 042798) Nonaccidental Mortality
Ito (2003, 042856) Nonaccidental Mortality
CHF HA
IHDHA
COPD HA
Pneumonia HA
Thurston et al. (1 994, 043921) Respiratory HA
Sheppard (2003, 042826) Asthma HA
Ostro et al. (2003, 042824) CVD Mortality
Mar et al. (2003, 042841) CVD Mortality
b Study did not present mean; median presented.
Mean3
5.6
5.6
5.6
6.1
6.9d
7.4
9.0
9.0
9.0°
9.6
9.7
9.7
9.7
9.7
9.7
9.7"
9.7d
9.9
10.8°
10.8C
10.9
11.4
11.5"
11. 5d
11.7"
11.8
12.2
12.2
12.3d
12.3"
12.6
13.3"
13.3"
13.3d
13.3"
13.3d
14.4C
16.2
30.5
33.2
Max3
24.6
27.1
24.6
72.0
28.3
50.3
50.3
30.0
50.3
34.2e
•349°
34.2e
34.2°
34.2°
34.2e
34.2°
25.2
50.9°
en oe
45.0
151.0
56.1
56.1
55.2
88.3°
68.0
68.0
81.3°
81 .3e
99.0
50.0
50.0
50.0
50.0
50.0
33.0
88.0
418.0
158.6
\
0,75
Effect Estimate (95% Cl)
i •
Jm-
t
•*—
-Ł_
A 1
(
I ^
Ł Ł
fe
| ,
I »
1.
J*.
L*.
L*.
I
1 I I I 1
1,00 1,25 1.50 1.75 2.00
Mean value slightly different from those reported in the published study; mean was either
provided by study authors or calculated from data provided by study authors.
' Maximum PMio-25 concentration provided by study authors or calculated from data provided by
study authors.
Relative Risk / Odds Ratio
Figure 2-3. Summary of U.S. studies examining the association between short-term
exposure to PMi0-2.5 and cardiovascular morbidity/mortality and respiratory
morbidity/mortality. All effect estimates have been standardized to reflect a
10 ug/m increase in mean 24-h avg PMi0.2.s concentration and ordered by
increasing concentration.
Limited evidence is available from epidemiologic studies for respiratory health effects and
outcomes in response to short-term exposure to PMi0_2.5. An increase in respiratory hospital
admissions and ED visits has been observed, but primarily in studies conducted in Canada and
Europe. In addition, associations are not reported for lower respiratory symptoms, wheeze, or
medication use. Controlled human exposure studies have not observed an effect on lung function or
respiratory symptoms in healthy or asthmatic adults, but healthy volunteers have exhibited
pulmonary inflammation. The toxicological studies (all IT instillation) provide evidence of
December 2009
2-20
-------
pulmonary injury and inflammation. In some cases, PM10_2.5 was found to be more potent than PM2.5
and effects were not solely attributable to endotoxin.
Currently, a national network is not in place to monitor PMi0_2.5 concentrations. As a result,
uncertainties surround the concentration at which the observed associations occur. Ambient
concentrations of PMi0_2.5 are generally determined by the subtraction of PMi0 and PM25
measurements, using various methods. For example, some epidemiologic studies estimate PMi0_2.5
by taking the difference between collocated PM10 and PM2.5 monitors while other studies have taken
the difference between county average PMi0 and PM2.5 concentrations. Moreover, there are potential
differences among operational flow rates and temperatures for PMi0 and PM2.5 monitors used to
calculate PMio_2.5. Therefore, there is greater error in ambient exposure to PM 10-2.5 compared to PM25.
This would tend to increase uncertainty and make it more difficult to detect effects of PMi0_2.5 in
epidemiologic studies. In addition, the various differences between eastern and western U.S. counties
can lead to exposure misclassification, and the potential underestimation of effects in western
counties (as discussed for PM25 in Section 2.3.2).
It is also important to note that the chemical composition of PMi0_2.5 can vary considerably by
location, but city-specific speciated PMi0_2.5 data are limited. PMi0_2.5 may contain Fe, Si, Al, and
base cations from soil, plant and insect fragments, pollen, fungal spores, bacteria, and viruses, as
well as fly ash, brake lining particles, debris, and automobile tire fragments.
The 2004 PM AQCD presented the limited amount of evidence available that examined the
potential association between exposure to PMi0_2.5 and health effects and outcomes. The current
evidence, primarily from epidemiologic studies, builds upon the results from the 2004 PM AQCD
and indicates that short-term exposure to PMi0_2.5 is associated with effects on both the
cardiovascular and respiratory systems. However, variability in the chemical and biological
composition of PMi0_2.5, limited evidence regarding effects of the various components of PMi0_2.5,
and lack of clearly defined biological mechanisms for PM10_2.5-related effects are important sources
of uncertainty.
2.3.5. Exposure to UFPs
2.3.5.1. Effects of Short-Term Exposure to UFPs
Table 2-4. Summary of causal determinations for short-term exposure to UFPs.
Size Fraction Outcome Causality Determination
Cardiovascular Effects Suggestive
UFPs
Respiratory Effects Suggestive
Cardiovascular Effects
Controlled human exposure studies provide the majority of the evidence for cardiovascular
health effects in response to short-term exposure to UFPs. While there are a limited number of
studies that have examined the association between UFPs and cardiovascular morbidity, there is a
larger body of evidence from studies that exposed subjects to fresh DE, which is typically dominated
by UFPs. These studies have consistently demonstrated changes in vasomotor function following
exposure to atmospheres containing relatively high concentrations of particles (Section 6.2.4.2).
Markers of systemic oxidative stress have also been observed to increase after exposure to various
particle types that are predominantly in the UFP size range. In addition, alterations in HRV
parameters have been observed in response to controlled human exposure to UF CAPs, with
inconsistent evidence for changes in markers of blood coagulation following exposure to UF CAPs
December 2009 2-21
-------
and DE (Sections 6.2.1.2 and 6.2.8.2). A few toxicological studies have also found consistent
changes in vasomotor function, which provides coherence with the effects demonstrated in the
controlled human exposure studies (Section 6.2.4.3). Additional UFP-induced effects observed in
toxicological studies include alterations in HRV, with less consistent effects observed for systemic
inflammation and blood coagulation. Only a few epidemiologic studies have examined the effect of
UFPs on cardiovascular morbidity and collectively they found inconsistent evidence for an
association between UFPs and CVD hospital admissions, but some positive associations for
subclinical cardiovascular measures (i.e., arrhythmias and supraventricular beats) (Section 6.2.2.1).
These studies were conducted in the U.S. and Europe in areas with mean particle number
concentration ranging from -8,500 to 36,000 particles/cm3. However, UFP number concentrations
are highly variable (i.e., concentrations drop off quickly from the road compared to accumulation
mode particles), and therefore, more subject to exposure error than accumulation mode particles. In
conclusion, the evidence from the studies evaluated JS Suggestive Of 3 C3USal relationship
between short-term exposures to UFPs and cardiovascular effects.
Respiratory Effects
A limited number of epidemiologic studies have examined the potential association between
short-term exposure to UFPs and respiratory morbidity. Of the studies evaluated, there is limited, and
inconsistent evidence for an association between short-term exposure to UFPs and respiratory
symptoms, as well as asthma hospital admissions in locations a median particle number
concentration of-6,200 to a mean of 38,000 particles/cm3 (Section 6.3.10). The spatial and temporal
variability of UFPs also affects these associations. Toxicological studies have reported respiratory
effects including oxidative, inflammatory, and allergic responses using a number of different UFP
types (Section 6.3). Although controlled human exposure studies have not extensively examined the
effect of UFPs on respiratory outcomes, a few studies have observed small UFP-induced
asymptomatic decreases in pulmonary function. Markers of pulmonary inflammation have been
observed to increase in healthy adults following controlled exposures to UFPs, particularly in studies
using fresh DE. However, it is important to note that for both controlled human exposure and animal
toxicological studies of exposures to fresh DE, the relative contributions of gaseous copollutants to
the respiratory effects observed remain unresolved. Thus, the current collective evidence JS
suggestive of a causal relationship between short-term exposures to UFPs and
respiratory effects.
2.3.6. Integration of UFP Effects
The controlled human exposure studies evaluated have consistently demonstrated effects on
vasomotor function and systemic oxidative stress with additional evidence for alterations in HRV
parameters in response to exposure to UF CAPs. The toxicological studies provide coherence for the
changes in vasomotor function observed in the controlled human exposure studies. Epidemiologic
studies are limited because a national network is not in place to measure UFP in the U.S. UFP
concentrations are spatially and temporally variable, which would increase uncertainty and make it
difficult to detect associations between health effects and UFPs in epidemiologic studies. In addition,
data on the composition of UFPs, the spatial and temporal evolution of UFP size distribution and
chemical composition, and potential effects of UFP constituents are sparse.
More limited evidence is available regarding the effect of UFPs on respiratory effects.
Controlled human exposure studies have not extensively examined the effect of UFPs on respiratory
measurements, but a few studies have observed small decrements in pulmonary function and
increases in pulmonary inflammation. Additional effects including oxidative, inflammatory, and pro-
allergic outcomes have been demonstrated in toxicological studies. Epidemiologic studies have
found limited and inconsistent evidence for associations between UFPs and respiratory effects.
Overall, a limited number of studies have examined the association between exposure to UFPs
and morbidity and mortality. Of the studies evaluated, controlled human exposure and toxicological
studies provide the most evidence for UFP-induced cardiovascular and respiratory effects; however,
many studies focus on exposure to DE. As a result, it is unclear if the effects observed are due to
UFP, larger particles (i.e., PM2.s), or the gaseous components of DE. Additionally, UF CAPs systems
December 2009 2-22
-------
are limited as the atmospheric UFP composition is modified when concentrated, which adds
uncertainty to the health effects observed in controlled human exposure studies (Section 1.5.3).
2.4. Policy Relevant Considerations
2.4.1. Potentially Susceptible Populations
Upon evaluating the association between short- and long-term exposure to PM and various
health outcomes, studies also attempted to identify populations that are more susceptible to PM (i.e.,
populations that have a greater likelihood of experiencing health effects related to exposure to an air
pollutant (e.g., PM) due to a variety of factors including, but not limited to: genetic or developmental
factors, race, gender, life stage, lifestyle (e.g., smoking status and nutrition) or preexisting disease; as
well as, population-level factors that can increase an individual's exposure to an air pollutant (e.g.,
PM) such as socioeconomic status [SES], which encompasses reduced access to health care, low
educational attainment, residential location, and other factors). These studies did so by conducting
stratified analyses; by examining effects in individuals with an underlying health condition; or by
developing animal models that mimic the pathophysiologic conditions associated with an adverse
health effect. In addition, numerous studies that focus on only one potentially susceptible population
provide supporting evidence on whether a population is susceptible to PM exposure. These studies
identified a multitude of factors that could potentially contribute to whether an individual is
susceptible to PM (Table 8-2). Although studies have primarily used exposures to PM2.5 or PMi0, the
available evidence suggests that the identified factors may also enhance susceptibility to PMi0_2.5.
The examination of susceptible populations to PM exposure allows for the NAAQS to provide an
adequate margin of safety for both the general population and for susceptible populations.
During specific periods of life (i.e., childhood and advanced age), individuals may be more
susceptible to environmental exposures, which in turn can render them more susceptible to PM-
related health effects. An evaluation of age-related health effects suggests that older adults have
heightened responses for cardiovascular morbidity with PM exposure. In addition, epidemiologic
and toxicological studies provide evidence that indicates children are at an increased risk of PM-
related respiratory effects. It should be noted that the health effects observed in children could be
initiated by exposures to PM that occurred during key windows of development, such as in utero.
Epidemiologic studies that focus on exposures during development have reported inconsistent
findings (Section 7.4), but a recent toxicological study suggests that inflammatory responses in
pregnant women due to exposure to PM could result in health effects in the developing fetus.
Epidemiologic studies have also examined whether additional factors, such as gender, race, or
ethnicity modify the association between PM and morbidity and mortality outcomes. Although
gender and race do not seem to modify PM risk estimates, limited evidence from two studies
conducted in California suggest that Hispanic ethnicity may modify the association between PM and
mortality.
Recent epidemiologic and toxicological studies provided evidence that individuals with null
alleles or polymorphisms in genes that mediate the antioxidant response to oxidative stress (i.e.,
GSTM1), regulate enzyme activity (i.e., MTHFR and cSHMT), or regulate levels of procoagulants
(i.e., fibrinogen) are more susceptible to PM exposure. However, some studies have shown that
polymorphisms in genes (e.g., HFE) can have a protective effect against effects of PM exposure.
Additionally, preliminary evidence suggests that PM exposure can impart epigenetic effects (i.e.,
DNA methylation); however, this requires further investigation.
Collectively, the evidence from epidemiologic and toxicological, and to a lesser extent,
controlled human exposure studies, indicate increased susceptibility of individuals with underlying
CVDs and respiratory illnesses (i.e., asthma) to PM exposure. Controlled human exposure and
toxicological studies provide additional evidence for increased PM-related cardiovascular effects in
individuals with underlying respiratory health conditions.
Recently studies have begun to examine the influence of preexisting chronic inflammatory
conditions, such as diabetes and obesity, on PM-related health effects. These studies have found
some evidence for increased associations for cardiovascular outcomes along with pathophysiologic
alterations in markers of inflammation, oxidative stress, and acute phase response. However, more
December 2009 2-23
-------
research is needed to thoroughly examine the affect of PM exposure on obese individuals and to
identify the biological pathway(s) that could increase the susceptibility of diabetic and obese
individuals to PM.
There is also evidence that SES, measured using surrogates such as educational attainment or
residential location, modifies the association between PM and morbidity and mortality outcomes. In
addition, nutritional status, another surrogate measure of SES, has been shown to have protective
effects against PM exposure in individuals that have a higher intake of some vitamins and nutrients.
Overall, the epidemiologic, controlled human exposure, and toxicological studies evaluated in
this review provide evidence for increased susceptibility for various populations, including children
and older adults, people with pre-existing cardiopulmonary diseases, and people with lower SES.
2.4.2. Lag Structure of PM-Morbidity and PM-Mortality Associations
Epidemiologic studies have evaluated the time-frame in which exposure to PM can impart a
health effect. PM exposure-response relationships can potentially be influenced by a multitude of
factors, such as the underlying susceptibility of an individual (e.g., age, pre-existing diseases), which
could increase or decrease the lag times observed.
An attempt has been made to identify whether certain lag periods are more strongly associated
with specific health outcomes. The epidemiologic evidence evaluated in the 2004 PM AQCD
supported the use of lags of 0-1 days for cardiovascular effects and longer moving averages or
distributed lags for respiratory diseases (U.S. EPA, 2004, 056905). However, currently, little
consensus exists as to the most appropriate a priori lag times to use when examining morbidity and
mortality outcomes. As a result, many investigators have chosen to examine the lag structure of
associations between PM concentration and health outcome instead of focusing on a priori lag times.
This approach is informative because if effects are cumulative, higher overall risks may exist than
would be observed for any given single-day lag.
2.4.2.1. PM-Cardiovascular Morbidity Associations
Most of the studies evaluated that examined the association between cardiovascular hospital
admissions and ED visits report associations with short-term PM exposure at lags 0- to 2-days, with
more limited evidence for snorter durations (i.e., hours) between exposure and response for some
health effects (e.g., onset of MI) (Section 6.2.10). However, these studies have rarely examined
alternative lag structures. Controlled human exposure and toxicological studies provide biological
plausibility for the health effects observed in the epidemiologic studies at immediate or concurrent
day lags. Although the majority of the evidence supports shorter lag times for cardiovascular health
effects, a recent study has provided preliminary evidence suggesting that longer lag times (i.e., 14-
day distributed lag model) may be plausible for non-ischemic cardiovascular conditions
(Section 6.2.10). Panel studies of short-term exposure to PM and cardiovascular endpoints have also
examined the time frame from exposure to health effect using a wide range of lag times. Studies of
ECG changes indicating ischemia show effects at lags from several hours to 2 days, while lag times
ranging from hours to several week moving averages have been observed in studies of arrhythmias,
vasomotor function and blood markers of inflammation, coagulation and oxidative stress
(Section 6.2). The longer lags observed in these panel studies may be explained if the effects of PM
are cumulative. Although few studies of cumulative effects have been conducted, toxicological
studies have demonstrated PM-dependent progression of atherosclerosis. It should be noted that PM
exposure could also lead to an acute event (e.g., infarction or stroke) in individuals with
atherosclerosis that may have progressed in response to cumulative PM exposure. Therefore, effects
have been observed at a range of lag periods from a few hours to several days with no clear evidence
for any lag period having stronger associations then another.
2.4.2.2. PM-Respiratory Morbidity Associations
Generally, recent studies of respiratory hospital admissions that evaluate multiple lags, have
found effect sizes to be larger when using longer moving averages or distributed lag models. For
example, when examining hospital admissions for all respiratory diseases among older adults, the
strongest associations were observed when using PM concentrations 2 days prior to the hospital
December 2009 2-24
-------
admission (Section 6.3.8). Longer lag periods were also found to be most strongly associated with
asthma hospital admissions and ED visits in children (3-5 days) with some evidence for more
immediate effects in older adults (lags of 0 and 1 day), but these observations were not consistent
across studies (Section 6.3.8). These variable results could be due to the biological complexity of
asthma, which inhibits the identification of a specific lag period. The longer lag times identified in
the epidemiologic studies evaluated are biologically plausible considering that PM effects on allergic
sensitization and lung immune defenses have been observed in controlled human exposure and
toxicological studies. These effects could lead to respiratory illnesses over a longer time course (e.g.,
within several days respiratory infection may become evident, resulting in respiratory symptoms or a
hospital admission). However, inflammatory responses, which contribute to some forms of asthma,
may result in symptoms requiring medical care within a shorter time frame (e.g., 0-1 days).
2.4.2.3. PM-Mortality Associations
Epidemiologic studies that focused on the association between short-term PM exposure and
mortality (i.e., all-cause, cardiovascular, and respiratory) mostly examined a priori lag structures of
either 1 or 0-1 days. Although mortality studies do not often examine alternative lag structures, the
selection of the aforementioned a priori lag days has been confirmed in additional studies, with the
strongest PM-mortality associations consistently being observed at lag 1 and 0-1-days (Section 6.5).
However, of note is recent evidence for larger effect estimates when using a distributed lag model.
Epidemiologic studies that examined the association between long-term exposure to PM and
mortality have also attempted to identify the latency period from PM exposure to death
(Section 7.6.4). Results of the lag comparisons from several cohort studies indicate that the effects of
changes in exposure on mortality are seen within five years, with the strongest evidence for effects
observed within the first two years. Additionally, there is evidence, albeit from one study, that the
mortality effect had larger cumulative effects spread over the follow-up year and three preceding
years.
2.4.3. PM Concentration-Response Relationship
An important consideration in characterizing the PM-morbidity and mortality association is
whether the concentration-response relationship is linear across the full concentration range that is
encountered or if there are concentration ranges where there are departures from linearity
(i.e., nonlinearity). In this ISA studies have been identified that attempt to characterize the shape of
the concentration-response curve along with possible PM "thresholds" (i.e., levels which PM
concentrations must exceed in order to elicit a health response). The epidemiologic studies evaluated
that examined the shape of the concentration-response curve and the potential presence of a
threshold have focused on cardiovascular hospital admissions and ED visits and mortality associated
with short-term exposure to PMi0 and mortality associated with long-term exposure to PM2.5.
A limited number of studies have been identified that examined the shape of the PM-
cardiovascular hospital admission and ED visit concentration-response relationship. Of these studies,
some conducted an exploratory analysis during model selection to determine if a linear curve most
adequately represented the concentration-response relationship; whereas, only one study conducted
an extensive analysis to examine the shape of the concentration-response curve at different
concentrations (Section 6.2.10.10). Overall, the limited evidence from the studies evaluated supports
the use of a no-threshold, log-linear model, which is consistent with the observations made in studies
that examined the PM-mortality relationship.
Although multiple studies have previously examined the PM-mortality concentration-response
relationship and whether a threshold exists, more complex statistical analyses continue to be
developed to analyze this association. Using a variety of methods and models, most of the studies
evaluated support the use of a no-threshold, log-linear model; however, one study did observe
heterogeneity in the shape of the concentration-response curve across cities (Section 6.5). Overall,
the studies evaluated further support the use of a no-threshold log-linear model, but additional issues
such as the influence of heterogeneity in estimates between cities, and the effect of seasonal and
regional differences in PM on the concentration-response relationship still require further
investigation.
December 2009 2-25
-------
In addition to examining the concentration-response relationship between short-term exposure
to PM and mortality, Schwartz et al. (2008, 156963) conducted an analysis of the shape of the
concentration-response relationship associated with long-term exposure to PM. Using a variety of
statistical methods, the concentration-response curve was found to be indistinguishable from linear,
and, therefore, little evidence was observed to suggest that a threshold exists in the association
between long-term exposure to PM2.5 and the risk of death (Section 7.6).
2.4.4. PM Sources and Constituents Linked to Health Effects
Recent epidemiologic, toxicological, and controlled human exposure studies have evaluated
the health effects associated with ambient PM constituents and sources, using a variety of
quantitative methods applied to a broad set of PM constituents, rather than selecting a few
constituents a priori (Section 6.6). There is some evidence for trends and patterns that link particular
ambient PM constituents or sources with specific health outcomes, but there is insufficient evidence
to determine whether these patterns are consistent or robust.
For cardiovascular effects, multiple outcomes have been linked to a PM2 5 crustal/soil/road
dust source, including cardiovascular mortality and ST-segment changes. Additional studies have
reported associations between other sources (i.e., traffic and wood smoke/vegetative burning) and
cardiovascular outcomes (i.e., mortality and ED visits). Studies that only examined the effects of
individual PM2.5 constituents found evidence for an association between EC and cardiovascular
hospital admissions and cardiovascular mortality. Many studies have also observed associations
between other sources (i.e., salt, secondary SO4 "/long-range transport, other metals) and
cardiovascular effects, but at this time, there does not appear to be a consistent trend or pattern of
effects for those factors.
There is less consistent evidence for associations between PM sources and respiratory health
effects, which may be partially due to the fact that fewer source apportionment studies have been
conducted that examined respiratory-related outcomes (e.g., hospital admissions) and measures (e.g.,
lung function). However, there is some evidence for associations between respiratory ED visits and
decrements in lung function with secondary SO42~ PM2.5. In addition, crustal/soil/road dust and
traffic sources of PM have been found to be associated with increased respiratory symptoms in
asthmatic children and decreased PEF in asthmatic adults. Inconsistent results were observed in
those PM2 5 studies that used individual constituents to examine associations with respiratory
morbidity and mortality, although Cu, Pb, OC, and Zn were related to respiratory health effects in
two or more studies.
A few studies have identified PM2 5 sources associated with total mortality. These studies
found an association between mortality and the PM2 5 sources: secondary SO4 /long-range transport,
traffic, and salt. In addition, studies have evaluated whether the variation in associations between
PM2 5 and mortality or PM10 and mortality reflects differences in PM2 5 constituents. PM10-mortality
effect estimates were greater in areas with a higher proportion of Ni in PM2 5, but the overall PMi0-
mortality association was diminished when New York City was excluded in sensitivity analyses in
two of the studies. V was also found to modify PMio-mortality effect estimates. When examining the
effect of species-to-PM2 5 mass proportion on PM2 5-mortality effect estimates, Ni, but not V, was
also found to modify the association.
Overall, the results indicate that many constituents of PM can be linked with differing health
effects and the evidence is not yet sufficient to allow differentiation of those constituents or sources
that are more closely related to specific health outcomes. These findings are consistent with the
conclusions of the 2004 PM AQCD (U.S. EPA, 2004, 056905) (i.e., that a number of source types,
including motor vehicle emissions, coal combustion, oil burning, and vegetative burning, are
associated with health effects). Although the crustal factor of fine particles was not associated with
mortality in the 2004 PM AQCD (U.S. EPA, 2004, 056905). recent studies have suggested that PM
(both PM2 5 and PMi0_2.5) from crustal, soil or road dust sources or PM tracers linked to these sources
are associated with cardiovascular effects. In addition, PM25 secondary SO42~has been associated
with both cardiovascular and respiratory effects.
December 2009 2-26
-------
2.5. Welfare Effects
This section presents key conclusions and scientific judgments regarding causality for welfare
effects of PM as discussed in Chapter 9. The effects of particulate NOX and SOX have recently been
evaluated in the ISA for Oxides of Nitrogen and Sulfur - Ecological Criteria (U.S. EPA, 2008,
157074). That ISA focused on the effects from deposition of gas- and particle-phase pollutants
related to ambient NOX and SOX concentrations that can lead to acidification and nutrient
enrichment. Thus, emphasis in Chapter 9 is placed on the effects of airborne PM, including NOX and
SOX, on visibility and climate, and on the effects of deposition of PM constituents other than NOX
and SOX, primarily metals and carbonaceous compounds. EPA's framework for causality, described
in Chapter 1, was applied and the causal determinations are highlighted.
Table 2-5. Summary of causality determination for welfare effects.
Welfare Effects
Causality Determination
Effects on Visibility
Causal
Effects on Climate
Causal
Ecological Effects
Likely to be causal
Effects on Materials
Causal
2.5.1. Summary of Effects on Visibility
Visibility impairment is caused by light scattering and absorption by suspended particles and
gases. There is strong and consistent evidence that PM is the overwhelming source of visibility
impairment in both urban and remote areas. EC and some crustal minerals are the only commonly
occurring airborne particle components that absorb light. All particles scatter light, and generally
light scattering by particles is the largest of the four light extinction components (i.e., absorption and
scattering by gases and particles). Although a larger particle scatters more light than a similarly
shaped smaller particle of the same composition, the light scattered per unit of mass is greatest for
particles with diameters from -0.3-1.0 um.
For studies where detailed data on particle composition by size are available, accurate
calculations of light extinction can be made. However, routinely available PM speciation data can be
used to make reasonable estimates of light extinction using relatively simple algorithms that multiply
the concentrations of each of the major PM species by its dry extinction efficiency and by a water
growth term that accounts for particle size change as a function of relative humidity for hygroscopic
species (e.g., sulfate, nitrate, and sea salt). This permits the visibility impairment associated with
each of the major PM components to be separately approximated from PM speciation monitoring
data.
Direct optical measurement of light extinction measured by transmissometer, or by combining
the PM light scattering measured by integrating nephelometers with the PM light absorption
measured by an aethalometer, offer a number of advantages compared to algorithm estimates of light
extinction based on PM composition and relative humidity data. The direct measurements are not
subject to the uncertainties associated with assumed scattering and absorption efficiencies used in the
PM algorithm approach. The direct measurements have higher time resolution (i.e., minutes to
hours), which is more commensurate with visibility effects compared with calculated light extinction
using routinely available PM speciation data (i.e., 24-h duration).
Particulate sulfate and nitrate have comparable light extinction efficiencies (haze impacts per
unit mass concentration) at any relative humidity value. Their light scattering per unit mass
concentration increases with increasing relative humidity, and at sufficiently high humidity values
(RH>85%) they are the most efficient particulate species contributing to haze. Particulate sulfate is
December 2009
2-27
-------
the dominant source of regional haze in the eastern U.S. (>50% of the participate light extinction)
and an important contributor to haze elsewhere in the country (>20% of particulate light extinction).
Particulate nitrate is a minor component of remote-area regional haze in the non-California western
and eastern U.S., but an important contributor in much of California and in the upper Midwestern
U.S., especially during winter when it is the dominant contributor to particulate light extinction.
EC and OC have the highest dry extinction efficiencies of the major PM species and are
responsible for a large fraction of the haze, especially in the northwestern U.S., though absolute
concentrations are as high in the eastern U.S. Smoke plume impacts from large wildfires dominate
many of the worst haze periods in the western U.S. Carbonaceous PM is generally the largest
component of urban excess PM2.5 (i.e., the difference between urban and regional background
concentration). Western urban areas have more than twice the average concentrations of
carbonaceous PM than remote areas sites in the same region. In eastern urban areas PM2.5 is
dominated by about equal concentrations of carbonaceous and sulfate components, though the
usually high relative humidity in the East causes the hydrated sulfate particles to be responsible for
about twice as much of the urban haze as that caused by the carbonaceous PM.
PM2.5 crustal material (referred to as fine soil) and PMi0_2.s are significant contributors to haze
for remote areas sites in the arid southwestern U.S. where they contribute a quarter to a third of the
haze, with PMi0_2.s usually contributing twice that of fine soil. Coarse mass concentrations are as
high in the Central Great Plains as in the deserts though there are no corresponding high
concentrations of fine soil as in the Southwest. Also the relative contribution to haze by the high
coarse mass in the Great Plains is much smaller because of the generally higher haze values caused
by the high concentrations of sulfate and nitrate PM in that region.
Visibility has direct significance to people's enjoyment of daily activities and their overall
sense of wellbeing. For example, psychological research has demonstrated that people are
emotionally affected by poor VAQ such that their overall sense of wellbeing is diminished. Urban
visibility has been examined in two types of studies directly relevant to the NAAQS review process:
urban visibility preference studies and urban visibility valuation studies. Both types of studies are
designed to evaluate individuals' desire for good VAQ where they live, using different metrics.
Urban visibility preference studies examine individuals' preferences by investigating the amount of
visibility degradation considered unacceptable, while economic studies examine the value an
individual places on improving VAQ by eliciting how much the individual would be willing to pay
for different amounts of VAQ improvement.
There are three urban visibility preference studies and two additional pilot studies that have
been conducted to date that provide useful information on individuals' preferences for good VAQ in
the urban setting. The completed studies were conducted in Denver, Colorado, two cities in British
Columbia, Canada, and Phoenix, AZ. The additional studies were conducted in Washington, DC. The
range of median preference values for an acceptable amount of visibility degradation from the 4
urban areas was approximately 19-33 dv. Measured in terms of visual range (VR), these median
acceptable values were between approximately 59 and 20 km.
The economic importance of urban visibility has been examined by a number of studies
designed to quantify the benefits (or willingness to pay) associated with potential improvements in
urban visibility. Urban visibility valuation research was described in the 2004 PM AQCD (U.S. EPA,
2004, 056905) and the 2005 PM Staff Paper (U.S. EPA, 2005, 090209). Since the mid-1990s, little
new information has become available regarding urban visibility valuation (Section 9.2.4).
Collectively, the evidence is sufficient to conclude that 3 C3USal relationship exists
between PM and visibility impairment.
2.5.2. Summary of Effects on Climate
Aerosols affect climate through direct and indirect effects. The direct effect is primarily
realized as planet brightening when seen from space because most aerosols scatter most of the
visible spectrum light that reaches them. The Intergovernmental Panel on Climate Change (IPCC)
Fourth Assessment Report (AR4) (IPCC, 2007, 092765). hereafter IPCC AR4, reported that the
radiative forcing from this direct effect was -0.5 (±0.4) W/m2 and identified the level of scientific
understanding of this effect as 'Medium-low'. The global mean direct radiative forcing effect from
individual components of aerosols was estimated for the first time in the IPCC AR4 where they were
reported to be (all in W/m2 units): -0.4 (±0.2) for sulfate, -0.05 (±0.05) for fossil fuel-derived organic
December 2009 2-28
-------
carbon, +0.2 (±0.15) for fossil fuel-derived black carbon (BC), +0.03 (±0.12) for biomass burning,
-0.1 (±0.1) for nitrates, and -0.1 (±0.2) for mineral dust. Global loadings of anthropogenic dust and
nitrates remain very troublesome to estimate, making the radiative forcing estimates for these
constituents particularly uncertain.
Numerical modeling of aerosol effects on climate has sustained remarkable progress since the
time of the 2004 PM AQCD (U.S. EPA, 2004, 056905). PM AQCD, though model solutions still
display large heterogeneity in their estimates of the direct radiative forcing effect from
anthropogenic aerosols. The clear-sky direct radiative forcing over ocean due to anthropogenic
aerosols is estimated from satellite instruments to be on the order of-1.1 (±0.37) W/m2 while model
estimates are -0.6 W/m2. The models' low bias over ocean is carried through for the global average:
global average direct radiative forcing from anthropogenic aerosols is estimated from measurements
to range from -0.9 to -1.9 W/m2, larger than the estimate of-0.8 W/m2 from the models.
Aerosol indirect effects on climate are primarily realized as an increase in cloud brightness
(termed the 'first indirect' or Twomey effect), changes in precipitation, and possible changes in cloud
lifetime. The IPCC AR4 reported that the radiative forcing from the Twomey effect was -0.7 (range:
-1.1 to +4) and identified the level of scientific understanding of this effect as "Low" in part owing
to the very large unknowns concerning aerosol size distributions and important interactions with
clouds. Other indirect effects from aerosols are not considered to be radiative forcing.
Taken together, direct and indirect effects from aerosols increase Earth's shortwave albedo or
reflectance thereby reducing the radiative flux reaching the surface from the Sun. This produces net
climate cooling from aerosols. The current scientific consensus reported by IPCC AR4 is that the
direct and indirect radiative forcing from anthropogenic aerosols computed at the top of the
atmosphere, on a global average, is about -1.3 (range: -2.2 to -0.5) W/m2. While the overall global
average effect of aerosols at the top of the atmosphere and at the surface is negative, absorption and
scattering by aerosols within the atmospheric column warms the atmosphere between the Earth's
surface and top of the atmosphere. In part, this is owing to differences in the distribution of aerosol
type and size within the vertical atmospheric column since aerosol type and size distributions
strongly affect the aerosol scattering and reradiation efficiencies at different altitudes and
atmospheric temperatures. And, although the magnitude of the overall negative radiative forcing at
the top of the atmosphere appears large in comparison to the analogous IPCC AR4 estimate of
positive radiative forcing from anthropogenic GHG of about +2.9 (± 0.3) W/m2, the horizontal,
vertical, and temporal distributions and the physical lifetimes of these two very different radiative
forcing agents are not similar; therefore, the effects do not simply off-set one another.
Overall, the evidence is sufficient to conclude that 3 C3USal relationship exists between
PM and effects on climate, including both direct effects on radiative forcing and indirect
effects that involve cloud feedbacks that influence precipitation formation and cloud
lifetimes.
2.5.3. Summary of Ecological Effects of PM
Ecological effects of PM include direct effects to metabolic processes of plant foliage;
contribution to total metal loading resulting in alteration of soil biogeochemistry and microbiology,
plant growth and animal growth and reproduction; and contribution to total organics loading
resulting in bioaccumulation and biomagnification across trophic levels. These effects were well-
characterized in the 2004 PM AQCD (U.S. EPA, 2004, 056905). Thus, the summary below builds
upon the conclusions provided in that review.
PM deposition comprises a heterogeneous mixture of particles differing in origin, size, and
chemical composition. Exposure to a given concentration of PM may, depending on the mix of
deposited particles, lead to a variety of phytotoxic responses and ecosystem effects. Moreover, many
of the ecological effects of PM are due to the chemical constituents (e.g., metals, organics, and ions)
and their contribution to total loading within an ecosystem.
Investigations of the direct effects of PM deposition on foliage have suggested little or no
effects on foliar processes, unless deposition levels were higher than is typically found in the
ambient environment. However, consistent and coherent evidence of direct effects of PM has been
found in heavily polluted areas adjacent to industrial point sources such as limestone quarries,
cement kilns, and metal smelters (Sections 9.4.3 and 9.4.5.7). Where toxic responses have been
December 2009 2-29
-------
documented, they generally have been associated with the acidity, trace metal content, surfactant
properties, or salinity of the deposited materials.
An important characteristic of fine particles is their ability to affect the flux of solar radiation
passing through the atmosphere, which can be considered in both its direct and diffuse components.
Foliar interception by canopy elements occurs for both up- and down-welling radiation. Therefore,
the effect of atmospheric PM on atmospheric turbidity influences canopy processes both by radiation
attenuation and by changing the efficiency of radiation interception in the canopy through
conversion of direct to diffuse radiation. Crop yields can be sensitive to the amount of radiation
received, and crop losses have been attributed to increased regional haze in some areas of the world
such as China (Section 9.4.4). On the other hand, diffuse radiation is more uniformly distributed
throughout the canopy and may increase canopy photosynthetic productivity by distributing radiation
to lower leaves. The enrichment in photosynthetically active radiation (PAR) present in diffuse
radiation may offset a portion of the effect of an increased atmospheric albedo due to atmospheric
particles. Further research is needed to determine the effects of PM alteration of radiative flux on the
growth of vegetation in the U.S.
The deposition of PM onto vegetation and soil, depending on its chemical composition, can
produce responses within an ecosystem. The ecosystem response to pollutant deposition is a direct
function of the level of sensitivity of the ecosystem and its ability to ameliorate resulting change.
Many of the most important ecosystem effects of PM deposition occur in the soil. Upon entering the
soil environment, PM pollutants can alter ecological processes of energy flow and nutrient cycling,
inhibit nutrient uptake, change ecosystem structure, and affect ecosystem biodiversity. The soil
environment is one of the most dynamic sites of biological interaction in nature. It is inhabited by
microbial communities of bacteria, fungi, and actinomycetes, in addition to plant roots and soil
macro-fauna. These organisms are essential participants in the nutrient cycles that make elements
available for plant uptake. Changes in the soil environment can be important in determining plant
and ultimately ecosystem response to PM inputs.
There is strong and consistent evidence from field and laboratory experiments that metal
components of PM alter numerous aspects of ecosystem structure and function. Changes in the soil
chemistry, microbial communities and nutrient cycling, can result from the deposition of trace
metals. Exposures to trace metals are highly variable, depending on whether deposition is by wet or
dry processes. Although metals can cause phytotoxicity at high concentrations, few heavy metals
(e.g., Cu, Ni, Zn) have been documented to cause direct phytotoxicity under field conditions.
Exposure to coarse particles and elements such as Fe and Mg are more likely to occur via dry
deposition, while fine particles, which are more often deposited by wet deposition, are more likely to
contain elements such as Ca, Cr, Pb, Ni, and V. Ecosystems immediately downwind of major
emissions sources can receive locally heavy deposition inputs. Phytochelatins produced by plants as
a response to sublethal concentrations of heavy metals are indicators of metal stress to plants.
Increased concentrations of phytochelatins across regions and at greater elevation have been
associated with increased amounts of forest injury in the northeastern U.S.
Overall, the ecological evidence is sufficient to conclude that a Causal relationship JS likely
to exist between deposition of PM and a variety of effects on individual organisms and
ecosystems, based on information from the previous review and limited new findings in
this review. However, in many cases, it is difficult to characterize the nature and magnitude of
effects and to quantify relationships between ambient concentrations of PM and ecosystem response
due to significant data gaps and uncertainties as well as considerable variability that exists in the
components of PM and their various ecological effects.
2.5.4. Summary of Effects on Materials
Building materials (metals, stones, cements, and paints) undergo natural weathering processes
from exposure to environmental elements (wind, moisture, temperature fluctuations, sunlight, etc.).
Metals form a protective film of oxidized metal (e.g., rust) that slows environmentally induced
corrosion. However, the natural process of metal corrosion is enhanced by exposure to
anthropogenic pollutants. For example, formation of hygroscopic salts increases the duration of
surface wetness and enhances corrosion.
A significant detrimental effect of particle pollution is the soiling of painted surfaces and other
building materials. Soiling changes the reflectance of opaque materials and reduces the transmission
December 2009 2-30
-------
of light through transparent materials. Soiling is a degradation process that requires remediation by
cleaning or washing, and, depending on the soiled surface, repainting. Particulate deposition can
result in increased cleaning frequency of the exposed surface and may reduce the usefulness of the
soiled material.
Attempts have been made to quantify the pollutant exposure levels at which materials damage
and soiling have been perceived. However, to date, insufficient data are available to advance the
knowledge regarding perception thresholds with respect to pollutant concentration, particle size, and
chemical composition. Nevertheless, the evidence is sufficient to conclude that 3 C3US3I
relationship exists between PM and effects on materials.
2.6. Summary of Health Effects and Welfare Effects
Causal Determinations
This chapter has provided an overview of the underlying evidence used in making the causal
determinations for the health and welfare effects and PM size fractions evaluated. This review builds
upon the main conclusions of the last PM AQCD (U.S. EPA, 2004, 056905):
• "A growing body of evidence both from epidemiological and toxicological studies...
supports the general conclusion that PM2.5 (or one or more PM2.5 components), acting
alone and/or in combination with gaseous copollutants, are likely causally related to
cardiovascular and respiratory mortality and morbidity." (pg 9-79)
• "A much more limited body of evidence is suggestive of associations between short-term
(but not long-term) exposures to ambient coarse-fraction thoracic particles... and various
mortality and morbidity effects observed at times in some locations. This suggests that
PMio_2.5, or some constituent component(s) of PMi0_2.5, may contribute under some
circumstances to increased human health risks... with somewhat stronger evidence for...
associations with morbidity (especially respiratory) endpoints than for mortality." (pg
9-79 and 9-80)
• "Impairment of visibility in rural and urban areas is directly related to ambient
concentrations of fine particles, as modulated by particle composition, size, and
hygroscopic characteristics, and by relative humidity." (pg 9-99)
• "Available evidence, ranging from satellite to in situ measurements of aerosol effects on
incoming solar radiation and cloud properties, is strongly indicative of an important role
in climate for aerosols, but this role is still poorly quantified." (pg 9-111)
The evaluation of the epidemiologic, toxicological, and controlled human exposure studies
published since the completion of the 2004 PM AQCD have provided additional evidence for
PM-related health effects. Table 2-6 provides an overview of the causal determinations for all PM
size fractions and health effects. Causal determinations for PM and welfare effects, including
visibility, climate, ecological effects, and materials are included in Table 2-7. Detailed discussions of
the scientific evidence and rationale for these causal determinations are provided in the subsequent
chapters of this ISA.
December 2009 2-31
-------
Table 2-6. Summary of PM causal determinations by exposure duration and health outcome.
Size Fraction Exposure Outcome
Cardiovascular Effects
Respiratory Effects
Central Nervous System
Mortality
PM2.5 Cardiovascular Effects
Respiratory Effects
Long-term Mortality
Reproductive and Developmental
Cancer, Mutagenicity, Genotoxicity
Cardiovascular Effects
Respiratory Effects
Central Nervous System
Mortality
PM-io-2.5 Cardiovascular Effects
Respiratory Effects
Long-term Mortality
Reproductive and Developmental
Cancer, Mutagenicity, Genotoxicity
Cardiovascular Effects
Respiratory Effects
Central Nervous System
Mortality
UFPs Cardiovascular Effects
Respiratory Effects
Long-term Mortality
Reproductive and Developmental
Cancer, Mutagenicity, Genotoxicity
Causality Determination
Causal
Likely to be causal
Inadequate
Causal
Causal
Likely to be Causal
Causal
Suggestive
Suggestive
Suggestive
Suggestive
Inadequate
Suggestive
Inadequate
Inadequate
Inadequate
Inadequate
Inadequate
Suggestive
Suggestive
Inadequate
Inadequate
Inadequate
Inadequate
Inadequate
Inadequate
Inadequate
December 2009
2-32
-------
Table 2-7. Summary of PM causal determinations for welfare effects
Welfare Effects
Causality Determination
Effects on Visibility
Causal
Effects on Climate
Causal
Ecological Effects
Likely to be causal
Effects on Materials
Causal
December 2009
2-33
-------
Chapter 2 References
American Heart Association. (2009). Cardiovascular disease statistics. Retrieved 17-NOV-09, from
http: //www. americanheart. org/pre senter. j html? identifier=447 8. 198920
ATSDR. (2006). A study of ambient air contaminants and asthma in New York City: Part A and B. Agency for Toxic
Substances and Disease Registry; Public Health Service; U.S. Department of Health and Human Services. Atlanta,
GA.http://permanent.access.gpo.gov/lps88357/ASTHMA_BRONX_FINAL_REPORT.pdf. 090132
Bell ML; Ebisu K; Peng RD; Walker J; Samet JM; Zeger SL; Dominic F. (2008). Seasonal and regional short-term effects
of fine particles on hospital admissions in 202 U.S. counties, 1999-2005. Am JEpidemiol, 168: 1301-1310. 156266
Burnett RT; Cakmak S; Brook JR; Krewski D. (1997). The role of particulate size and chemistry in the association between
summertime ambient air pollution and hospitalization for cardiorespiratory diseases. Environ Health Perspect, 105:
614-620. 084194
Burnett RT; Goldberg MS. (2003). Size-fractionated particulate mass and daily mortality in eight Canadian cities. Health
Effects Institute. Boston, MA. 042798
Burnett RT; Stieb D; Brook JR; Cakmak S; Dales R; Raizenne M; Vincent R; Dann T. (2004). Associations between short-
term changes in nitrogen dioxide and mortality in Canadian cities. Arch Environ Occup Health, 59: 228-236.
086247
Chen LH; Knutsen SF; Shavlik D; Beeson WL; Petersen F; Ghamsary M; Abbey D. (2005). The association between fatal
coronary heart disease and ambient particulate air pollution: Are females at greater risk?. Environ Health Perspect,
113: 1723-1729. 087942
Chen Y; Yang Q; Krewski D; Burnett RT; Shi Y; McGrail KM. (2005). The effect of coarse ambient particulate matter on
first, second, and overall hospital admissions for respiratory disease among the elderly. Inhal Toxicol, 17: 649-655.
087555
Chen Y; Yang Q; Krewski D; Shi Y; Burnett RT; McGrail K. (2004). Influence of relatively low level of particulate air
pollution on hospitalization for COPD in elderly people. Inhal Toxicol, 16: 21-25. 087262
Chimonas MA; Gessner BD. (2007). Airborne particulate matter from primarily geologic, non-industrial sources at levels
below National Ambient Air Quality Standards is associated with outpatient visits for asthma and quick-relief
medication prescriptions among children less than 20 years old enrolled in Medicaid in Anchorage, Alaska. Environ
Res, 103:397-404. 093261
Delfino RJ; Murphy-Moulton AM; Burnett RT; Brook JR; Becklake MR. (1997). Effects of air pollution on emergency
room visits for respiratory illnesses in Montreal, Quebec. Am J Respir Crit Care Med, 155: 568-576. 082687
Dockery DW; Damokosh AI; Neas LM; Raizenne M; Spengler JD; Koutrakis P; Ware JH; Speizer FE. (1996). Health
effects of acid aerosols on North American children: respiratory symptoms and illness. Environ Health Perspect,
104: 500-505. 046219
Dockery DW; Luttmann-Gibson H; Rich DQ; Link MS; Mittleman MA; Gold DR; Koutrakis P; Schwartz JD; Verrier RL.
(2005). Association of air pollution with increased incidence of ventricular tachyarrhythmias recorded by implanted
cardioverter defibrillators. Environ Health Perspect, 113: 670-674. 078995
Dominici F; Peng RD; Bell ML; Pham L; McDermott A; Zeger SL; Samet JL. (2006). Fine particulate air pollution and
hospital admission for cardiovascular and respiratory diseases. JAMA, 295: 1127-1134. 088398
Eftim SE; Samet JM; Janes H; McDermott A; Dominici F. (2008). Fine particulate matter and mortality: a comparison of
the six cities and American Cancer Society cohorts with a medicare cohort. Epidemiology, 19: 209-216. 099104
Enstrom JE. (2005). Fine particulate air pollution and total mortality among elderly Californians, 1973-2002. Inhal Toxicol,
17: 803-816. 087356
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009 2-34
-------
Fairley D. (2003). Mortality and air pollution for Santa Clara County, California, 1989-1996, In: Revised analyses of time-
series studies of air pollution and health. Special report. Health Effects Institute. Boston,
MA.http://www.healtheffects.org/Pubs/TimeSeries.pdf. 042850
Franklin M; Koutrakis P; Schwartz J . (2008). The role of particle composition on the association between PM2.5 and
mortality. Epidemiology, 19: 680-689. 097426
Franklin M; Zeka A; Schwartz J. (2007). Association between PM2.5 and all-cause and specific-cause mortality in 27 US
communities. JExpo Sci Environ Epidemiol, 17: 279-287. 091257
Fung KY; Khan S; Krewski D; Chen Y. (2006). Association between air pollution and multiple respiratory hospitalizations
among the elderly in Vancouver, Canada. Inhal Toxicol, 18: 1005-1011. 089789
Goss CH; Newsom SA; Schildcrout JS; Sheppard L; Kaufman JD. (2004). Effect of ambient air pollution on pulmonary
exacerbations and lung function in cystic fibrosis. Am J Respir Crit Care Med, 169: 816-821. 055624
IPCC. (2007). IPCC Fourth Assessment Report (AR4): Climate Change 2007: Working Group I Report: The Physical
Science Basis. Cambridge, UK and New York, NY: Intergovernmental Panel on Climate Change, Cambridge
University Press. 092765
Ito K. (2003). Associations of particulate matter components with daily mortality and morbidity in Detroit, Michigan, In:
Revised analyses of time-series studies of air pollution and health. Special report. Health Effects Institute. Boston,
MA. R828112. http: //www.healtheffects.org/Pubs/TimeSeries.pdf. 042856
Ito K; Thurston GD; Silverman RA. (2007). Characterization of PM2.5, gaseous pollutants, and meteorological interactions
in the context of time-series health effects models. J Expo Sci Environ Epidemiol, 17 Suppl 2: S45-S60. 156594
Kim JJ; Smorodinsky S; Lipsett M; Singer BC; Hodgson AT; Ostro B. (2004). Traffic-related air pollution near busy roads:
the East Bay children's Respiratory Health Study. Am J Respir Crit Care Med, 170: 520-526. 087383
Klemm RJ; Lipfert FW; Wyzga RE; Gust C. (2004). Daily mortality and air pollution in Atlanta: two years of data from
ARIES. Inhal Toxicol, 16 Suppl 1: 131-141. 056585
Klemm RJ; Mason R. (2003). Replication of reanalysis of Harvard Six-City mortality study. In HEI Special Report:
Revised Analyses of Time-Series Studies of Air Pollution and Health, Part II (pp. 165-172). Boston, MA: Health
Effects Institute. 042801
Krewski D; Jerrett M; Burnett RT; Ma R; Hughes E; Shi Y; Turner MC; Pope AC III; Thurston G; Calle EE; Thun MJ.
(2009). Extended follow-up and spatial analysis of the American Cancer Society study linking particulate air
pollution and mortality. Health Effects Institute. Cambridge, MA. Report Nr. 140. 191193
Laden F; Schwartz J; Speizer FE; Dockery DW. (2006). Reduction in fine particulate air pollution and mortality: extended
follow-up of the Harvard Six Cities study. Am J Respir Crit Care Med, 173: 667-672. 087605
Lin M; Chen Y; Burnett RT; Villeneuve PJ; Krewski D. (2002). The influence of ambient coarse particulate matter on
asthma hospitalization in children: case-crossover and time-series analyses. Environ Health Perspect, 110: 575-581.
026067
Lin M; Stieb DM; Chen Y. (2005). Coarse particulate matter and hospitalization for respiratory infections in children
younger than 15 years in Toronto: a case-crossover analysis. Pediatrics, 116: 235-240. 087828
Lipfert FW; Baty JD; Miller JP; Wyzga RE. (2006). PM2.5 constituents and related air quality variables as predictors of
survival in a cohort of U.S. military veterans. Inhal Toxicol, 18: 645-657. 088756
Lipfert FW; Morris SC; Wyzga RE. (2000). Daily mortality in the Philadelphia metropolitan area and size-classified
particulate matter. J Air Waste Manag Assoc, 50: 1501-1513. 004088
Lisabeth LD; Escobar JD; Dvonch JT; Sanchez BN; Majersik JJ; Brown DL; Smith MA; Morgenstern LB. (2008). Ambient
air pollution and risk for ischemic stroke and transient ischemic attack. Ann Neurol, 64: 53-59. 155939
Mar TF; Larson TV; Stier RA; Claiborn C; Koenig JQ. (2004). An analysis of the association between respiratory
symptoms in subjects with asthma and daily air pollution in Spokane, Washington. Inhal Toxicol, 16: 809-815.
057309
Mar TF; Norris GA; Larson TV; Wilson WE; Koenig JQ. (2003). Air pollution and cardiovascular mortality in Phoenix,
1995-1997. Health Effects Institute. Cambridge, MA. 042841
December 2009 2-35
-------
McConnell R; Berhane K; Gilliland F; Molitor J; Thomas D; Lurmann F; Avol E; Gauderman WJ; Peters JM. (2003).
Prospective study of air pollution and bronchitic symptoms in children with asthma. Am J Respir Crit Care Med,
168: 790-797. 049490
Metzger KB; Klein M; Flanders WD; Peel JL; Mulholland JA; Langberg JJ; Tolbert PE. (2007). Ambient air pollution and
cardiac arrhythmias in patients with implantable defibrillators. Epidemiology, 18: 585-592. 092856
Metzger KB; Tolbert PE; Klein M; Peel JL; Flanders WD; Todd KH; Mulholland JA; Ryan PB; Frumkin H. (2004).
Ambient air pollution and cardiovascular emergency department visits. Epidemiology, 15: 46-56. 044222
Miller KA; Siscovick DS; Sheppard L; Shepherd K; Sullivan JH; Anderson GL; Kaufman JD. (2007). Long-term exposure
to air pollution and incidence of cardiovascular events in women. N Engl J Med, 356: 447-458. 090130
O'Connor GT; Neas L; Vaughn B; Kattan M; Mitchell H; Grain EF; Evans R 3rd; Gruchalla R; Morgan W; Stout J; Adams
GK; Lippmann M. (2008). Acute respiratory health effects of air pollution on children with asthma in US inner
cities. JAllergy Clinlmmunol, 121: 1133-1139. 156818
Ostro BD; Broadwin R; Lipsett MJ. (2003). Coarse particles and daily mortality in Coachella Valley, California. Health
Effects Institute. Boston, MA. 042824
Peel JL; Tolbert PE; Klein M; Metzger KB; Flanders WD; Knox T; Mulholland JA; Ryan PB; Frumkin H. (2005). Ambient
air pollution and respiratory emergency department visits. Epidemiology, 16: 164-174. 056305
Peng RD; Chang HH; Bell ML; McDermott A; Zeger SL; Samet JM; Dominici F. (2008). Coarse particulate matter air
pollution and hospital admissions for cardiovascular and respiratory diseases among Medicare patients. JAMA,
299: 2172-2179. 156850
Peters A; Dockery DW; Muller JE; Mittleman MA. (2001). Increased particulate air pollution and the triggering of
myocardial infarction. Circulation, 103: 2810-2815. 016546
Pope C; Renlund D; Kfoury A; May H; Home B. (2008). Relation of heart failure hospitalization to exposure to fine
particulate air pollution. Am J Cardiol, 102: 1230-1234. 191969
Pope CAIII; Muhlestein JB; May HT; Renlund DG; Anderson JL; Home BD. (2006). Ischemic heart disease events
triggered by short-term exposure to fine particulate air pollution. Circulation, 114: 2443-2448. 091246
Rabinovitch N; Strand M; Gelfand EW (2006). Particulate levels are associated with early asthma worsening in children
with persistent disease. Am J Respir Crit Care Med, 173: 1098-1105. 088031
Rabinovitch N; Zhang LN; Murphy JR; Vedal S; Dutton SJ; Gelfand EW. (2004). Effects of wintertime ambient air
pollutants on asthma exacerbations in urban minority children with moderate to severe disease. J Allergy Clin
Immunol, 114: 1131-1137.096753
Renwick LC; Brown D; Clouter A; Donaldson K. (2004). Increased inflammation and altered macrophage chemotactic
responses caused by two ultrafine particle types. Occup Environ Med, 61: 442-447. 056067
Rich DQ; Kim MH; Turner JR; Mittleman MA; Schwartz J; Catalano PJ; Dockery DW. (2006). Association of ventricular
arrhythmias detected by implantable cardioverter defibrillator and ambient air pollutants in the St Louis, Missouri
metropolitan area. Occup Environ Med, 63: 591-596. 089814
Rich DQ; Schwartz J; Mittleman MA; Link M; Luttmann-Gibson H; Catalano PJ; Speizer FE; Dockery DW. (2005).
Association of short-term ambient air pollution concentrations and ventricular arrhythmias. Am J Epidemiol, 161:
1123-1132.079620
Schwartz J; Coull B; Laden F; Ryan L. (2008). The effect of dose and timing of dose on the association between airborne
particles and survival. Environ Health Perspect, 116: 64-69. 156963
Schwartz J; Dockery DW; Neas LM. (1996). Is daily mortality associated specifically with fine particles?. J Air Waste
Manag Assoc, 46: 927-939. 077325
Sheppard L; Levy D; Norris G; Larson TV; Koenig JQ. (2003). Effects of ambient air pollution and nonelderly asthma
hospital admissions in Seattle, Washington, 1987-1994. Epidemiology, 10: 23-30. 042826
Slaughter JC; Kim E; Sheppard L; Sullivan JH; Larson TV; Claiborn C. (2005). Association between particulate matter and
emergency room visits, hospital admissions and mortality in Spokane, Washington. J Expo Sci Environ Epidemiol,
15: 153-159.073854
December 2009 2-36
-------
Stieb DM; Beveridge RC; Brook JR; Smith-Doiron M; Burnett RT; Dales RE; Beaulieu S; Judek S; Mamedov A. (2000).
Air pollution, aeroallergens and cardiorespiratory emergency department visits in Saint John, Canada. J Expo Sci
Environ Epidemiol, 10: 461-477. 011675
Sullivan J; Sheppard L; Schreuder A; Ishikawa N; Siscovick D; Kaufman J. (2005). Relation between short-term fine-
particulate matter exposure and onset of myocardial infarction. Epidemiology, 16: 41-48. 050854
Symons JM; Wang L; Guallar E; Howell E; Dominici F; Schwab M; Ange BA; Samet J; Ondov J; Harrison D; Geyh A.
(2006). A case-crossover study of fine particulate matter air pollution and onset of congestive heart failure symptom
exacerbation leading to hospitalization. Am J Epidemiol, 164: 421-433. 091258
Thurston GD; Ito K; Hayes CG; Bates DV; Lippmann M. (1994). Respiratory hospital admissions and summertime haze air
pollution in Toronto, Ontario: consideration of the role of acid aerosols. Environ Res, 65: 271-290. 043921
Tolbert PE; Klein M; Peel JL; Sarnat SE; Sarnat JA. (2007). Multipollutant modeling issues in a study of ambient air
quality and emergency department visits in Atlanta. J Expo Sci Environ Epidemiol, 17: S29-S35. 090316
U.S. EPA. (2004). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research Triangle
Park, NC. EPA/600/P-99/002aF-bF. 056905
U.S. EPA. (2005). Review of the national ambient air quality standards for particulate matter: Policy assessment of
scientific and technical information OAQPS staff paper. U.S. Environmental Protection Agency. Washington, DC.
EPA/452/R-05-005a. http://www.epa.gov/ttn/naaqs/standards/pm/data/pmstaffpaper_20051221 .pdf. 090209
U.S. EPA. (2008). Integrated science assessment for oxides of nitrogen and sulfur: Ecological criteria. U.S. Environmental
Protection Agency. Research Triangle Park, NC. EPA/600/R-08/082F. 157074
Villeneuve PJ; Burnett RT; Shi Y; Krewski D; Goldberg MS; Hertzman C; Chen Y; Brook J. (2003). A time-series study of
air pollution, socioeconomic status, and mortality in Vancouver, Canada. J Expo Sci Environ Epidemiol, 13: 427-
435. 055051
Villeneuve PJ; Chen L; Stieb D; Rowe BH. (2006). Associations between outdoor air pollution and emergency department
visits for stroke in Edmonton, Canada. Eur J Epidemiol, 21: 689-700. 090191
Wilson WE; Mar TF; Koenig JQ. (2007). Influence of exposure error and effect modification by socioeconomic status on
the association of acute cardiovascular mortality with particulate matter in Phoenix. J Expo Sci Environ Epidemiol,
17: S11-S19. 157149
Woodruff TJ; Darrow LA; Parker JD. (2008). Air pollution and postneonatal infant mortality in the United States, 1999-
2002. Environ Health Perspect, 116: 110-115. 098386
Zanobetti A; Schwartz J. (2006). Air pollution and emergency admissions in Boston, MA. J Epidemiol Community Health,
60: 890-895. 090195
Zanobetti A; Schwartz J. (2009). The effect of fine and coarse particulate air pollution on mortality: A national analysis.
Environ Health Perspect, 117: 1-40. 188462
Zeger S; Dominici F; McDermott A; Samet J. (2008). Mortality in the Medicare population and chronic exposure to fine
particulate air pollution in urban centers (2000-2005). Environ Health Perspect, 116: 1614-1619. 191951
Zhang Z; Whitsel E; Quibrera P; Smith R; Liao D; Anderson G; Prineas R. (2009). Ambient fine particulate matter
exposure and myocardial ischemia in the Environmental Epidemiology of Arrhythmogenesis in the Women's Health
Initiative (EEAWHI) study. Environ Health Perspect, 117: 751-756. 191970
December 2009 2-37
-------
Chapter 3. Source to Human Exposure
3.1. Introduction
This chapter describes basic concepts and new and established findings in atmospheric
sciences and human exposure assessment relevant to PM to establish a foundation for the health and
ecological effects discussed in subsequent chapters. Information in this chapter builds on previous
AQCDs for PM using new data and re-interpretations of extant studies as well. This includes new
knowledge of PM chemistry, the latest developments in monitoring methodologies, recent national
and local measurements and trends in PM concentrations as a function of size range and
composition, advances in receptor and chemistry-transport modeling, revised estimates of policy-
relevant background PM, and recent work on exposure assessment.
The chapter and its associated annex material are organized as follows: Section 3.2 presents an
overview of basic information related to the size distribution and composition of airborne particles.
Section 3.3 provides a brief description of the sources, emissions, and deposition of PM, including
discussions of possible mechanisms of secondary PM formation from gaseous precursors and of the
atmospheric processes that deposit PM to the earth's surface. Issues related to the measurement of
PM and its chemical components and to monitors and networks in the U.S. are covered in
Section 3.4; supplementary material on these topics is contained in Annex A, Section A. 1. Analyses
of data for ambient concentrations of PM and its components are characterized in Section 3.5, and
supplementary information can be found in Annex A, Section A.2. Section 3.6 describes methods for
determining source contributions to ambient samples by receptor models and presents results from
recent receptor modeling studies. In addition, the construction of chemistry-transport models
(CTMs) to determine pollutant concentrations is described in Section 3.6. Supplementary
information about receptor model methods and results is given in Annex A, Section A.3. Policy
relevant background concentrations of PM, i.e., those concentrations defined to result from natural
sources everywhere in the world together with anthropogenic sources outside of Canada, the United
States, and Mexico, are presented in Section 3.7. Issues related to human exposure assessment
including sources of exposure and implications for epidemiologic studies are discussed in
Section 3.8. Supplementary information on exposure studies is included in Annex A, Section A.4.
Finally, the summary and conclusions from Chapter 3 are presented in Section 3.9.
3.2. Overview of Basic Aerosol Properties
Unlike gas-phase pollutants such as SO2, CO, H2CO and O3, which are well-defined chemical
entities, atmospheric PM varies in size, shape, and chemical composition. Atmospheric chemical and
microphysical processing of direct emissions of PM and its precursors together with mechanical
generation of particles tend to produce distinct lognormal modes (Whitby, 1978, 071181) as shown
in Figure 3-1. To the extent that information is available, discussions in this and subsequent chapters
will focus on particles in specific size ranges (i.e., PM2.5, PMi0_2.5 and PMi0). The subscripts after PM
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009 3-1
-------
refer to the aerodynamic diameter1 (dae) in micrometers ((irn) of 50% cut points of sampling devices.
For example, EPA defines PMi0 as particles collected by a sampler with an upper 50% cut point of
10 (im dae and a specific, fairly sharp, penetration curve, as defined in the Code of Federal
Regulations (40 CFR Part 58). PM2.5 is defined in an analogous way. Ultrafme particles (UFPs),
defined here as particles with a diameter < 0.1 (im (typically based on physical size, thermal
diffusivity, or electrical mobility), will also be discussed.
The terms "fine particles" and "coarse particles" have lost the precise meaning as defined in
Whitby (1978, 071181). where "fine particles" refers to all particles in the nucleation, Aitken, and
accumulation modes; and "coarse particles" characterizes all particles larger than these. UFPs
correspond loosely to the nucleation plus Aitken modes (in earlier literature, these modes were not
separated and the combination, unresolved by older instruments, was called the Aitken mode). Now,
the term "fine particles" is most often associated with the PM2.5 fraction, which includes the
nucleation, Aitken and accumulation modes and some particles from the lower-size tail of the coarse
particle mode between about 1 and 2.5 (im aerodynamic diameter. "Thoracic coarse" is frequently
used in reference to PMi0_2.5, which does not include the low-end tail of the coarse particle mode.
With high relative humidity, larger particles in the accumulation mode could also extend into the 1 to
3 (im size range. These relationships can be seen in Figure 3-1, which shows the number distribution
for UPFs and the volume distribution (or mass distribution if particle density is constant across the
size range) for fine and (thoracic) coarse particles. The figure is arranged this way because particle
number is most highly concentrated in the ultrafine (UF) size range but volume (or mass) is most
concentrated in the larger size ranges.
Fine Particles
Coarse Particles
O.O1
0.1 1 10
Diameter (micrometers)
Source:Reprinted with Permission of Cambridge University Press from Pandis (2004,1568381.
Figure 3-1. Particle size distributions by number and volume. Dashed lines refer to values in
individual modes and solid lines to their sum. Note that ultrafine particles are a
subset of fine particles.
1 Aerodynamic diameter is the diameter of a unit density (1 g/cm3) sphere that has the same gravitational settling velocity as the particle of
interest and is a useful metric for characterizing particles >~ 1 um. For sub-micron particles, forces other than gravity increase in
importance in determining a particle's motion and air can no longer be considered a continuum. Aerodynamic diameter is frequently
reported down to ~ 0.1 um where the assumptions used in its derivation no longer hold. A useful metric for characterizing particles <~
0.5 um is the mobility diameter defined as the diameter of a particle having the same diffusivity or electrical mobility in air as the
particle of interest. In the region between ~ 0.5-1.0 um, aerodynamic and mobility diameters are not necessarily the same. The question
of how best to merge these diameters is unresolved and depends on the particle properties of interest.
December 2009
3-2
-------
Characterizing particle size is important because different size particles penetrate to different
regions of the human respiratory tract. Thoracic particles refer to particles that travel past the larynx
to reach the lung airways and the gas-exchange region of the lung, and respirable particles are those
that reach the gas-exchange region. The selection of PMi0 as an indicator of thoracic particles was
based in large part on dosimetry (U.S. EPA, 1996, 079380). However, the selection of PM2.5 to
characterize respirable particles was driven mainly by considerations related to measurement
techniques available at the time rather than dosimetry. Currently, cut points other than 2.5 (im are
attainable and frequently put into use. For example, the American Conference of Governmental
Industrial Hygienists (ACGIH, 2005, 156188). the International Standards Organization, and the
European Standardization Committee have adopted a 50% cut point of 4 (im as an indicator of
respirable particles. Most commonly, however, PM2.5 is used as an indicator of respirable particles,
PMio_2.5 is used as an indicator of the thoracic component of coarse particles that is sometimes
referred to as thoracic coarse (noting that it excludes some coarse particles below 2.5 (im and above
10 (im), and PMi0 is used as an indicator of thoracic particles.
As can be seen from Table 3-1, particles in individual size modes are characterized by rather
distinct sources, composition, chemical properties, lifetimes in the atmosphere (T) and distances over
which they can travel. Whereas particles in the smaller size modes are formed mainly by combustion
processes and by nucleation and condensation of gases, coarse particles are generated mainly by
mechanical activity, such as by the action of wind on either the ground or the sea surface or by
construction or by resuspension by traffic. Particles in the UF size range are either emitted directly to
the atmosphere or are formed by nucleation of gaseous constituents in the atmosphere. The
properties of fibers and engineered nano-objects and nanometer scale products (e.g., dots, hollow
spheres, rods, fibers, and tubes) are not reviewed in this chapter because these classes of objects are
found mainly in certain occupational settings rather than in ambient air.
December 2009 3-3
-------
Table 3-1. Characteristics of ambient fine (ultrafine plus accumulation-mode) and coarse particles.
Fine
Coarse
Ultrafine
Accumulation
Formation
Processes
Combustion, high-temperature processes, and atmospheric reactions
Break-up of large solids/droplets
Nucleation of atmospheric gases
including H2S04, NH3 and some organic
Formed by compounds
Condensation of gases
Condensation of gases
Coagulation of smaller particles
Reactions of gases in or on particles
Evaporation of fog and cloud droplets in which gases
have dissolved and reacted
Mechanical disruption (crushing, grinding,
abrasion of surfaces)
Evaporation of sprays
Suspension of dusts
Reactions of gases in or on particles
Sulfate
EC
Composed of Metal compounds
Organic compounds with very low
saturation vapor pressure at ambient
temperature
Sulfate, nitrate, ammonium, and hydrogen ions
EC
Large variety of organic compounds
Metals: compounds of Pb, Cd, V, Ni, Cu, Zn, Mn, Fe, etc.
Nitrates/chlorides/sulfates from
HN03/HCI/S02 reactions with coarse particles
Oxides of crustal elements (Si, Al, Ti, Fe)
CaC03, CaS04, NaCI, sea salt
Bacteria, pollen, mold, fungal spores, plant
Particle-bound water and animal debris
Bacteria, viruses
Solubility
Not well characterized
Largely soluble, hygroscopic, and deliquescent
Largely insoluble and nonhygroscopic
Sources
High temperature combustion
Atmospheric reactions of primary,
gaseous compounds.
Combustion of fossil and biomass fuels, and high
temperature industrial processes, smelters, refineries,
steel mills etc.
Atmospheric oxidation of N02, S02, and organic
compounds, including biogenic organic species
(e.g.,terpenes)
Resuspension of particles deposited onto
roads
Tire, brake pad, and road wear debris
Suspension from disturbed soil (e.g., farming,
mining, unpaved roads)
Construction and demolition
Fly ash from uncontrolled combustion of coal,
oil, and wood
Ocean spray
Atmospheric
half-life
Minutes to hours
Days to weeks
Minutes to hours
Removal
Processes
Grows into accumulation mode
Diffuses to raindrops and other surfaces
Forms cloud droplets and rains out
Dry deposition
Dry deposition by fallout
Scavenging by falling rain drops
Travel distance <1 to 10s of km
100s to 1000s of km
<1 to 10s of km (100s to 1,000s of km in dust
storms for the small size tail)
Source: Adapted with Permission of the Air & Waste Management Association from Wilson and Suh (1997, 077408)
December 2009
3-4
-------
Source: National Exposure Research Laboratory.
Figure 3-2. X-ray spectra and scanning electron microscopy images of individual particles.
These include: (1) an aluminum-silicate fly ash sphere emitted from a coal-fired
power plant; (2) an iron oxide sphere emitted from a steel manufacturing facility;
(3) an aluminum-silicate particle, probably of crustal origin; (4) a carbon soot
aggregate from a diesel engine consisting of many sub-micron size carbon
particles; (5) a sodium chloride crystal, potentially of marine origin; and (6) a
partially collapsed pollen particle. The polycarbonate filter substrate used to
collect the particles is visible in the background of each image and contributes to
the carbon peak in each spectrum.
Particles appear in a wide variety of shapes such as spheres, ellipsoids, cubes, and irregular or
fractal geometries. This is one reason why a standard metric such as aerodynamic diameter is useful
for describing the mechanical properties of the particles. The shape of particles is important for
determining the optical properties of the particles. The directionality of sunlight scattered by certain
shapes of particles, such as plates, also depends strongly on their physical orientation while
suspended. The shape of particles also affects the surface area of the particles in contact with the
surface it is deposited on, including cell membranes.
Images of six types of individual particles obtained using scanning electron microscope
(SEM) and their corresponding x-ray spectra showing their major elemental composition are shown
December 2009
3-5
-------
in Figure 3-2. The images show particles sitting on a thin polycarbonate film with pores and a 1 (im
scale bar for size reference. Air is pulled through the filter pores with a vacuum pump and the
particles are left behind on the surface. The metal dominated spherical particles (image 2) were
formed at high temperatures and were quickly cooled. Particles which are liquid such as sulfate are
also spherical. Sodium chloride (NaCl) crystals are cubic (image 5); this particular particle could be
marine sea salt due to the proximity to the ocean where the sample was collected. Other particles,
such as the carbon chain agglomerates from diesel engines have much more irregular and complex
shapes (image 4). Note that these particles were placed under vacuum resulting in volatilization of
water and other volatile components and partial collapse of the pollen grain (image 6). Changes in
composition and possibly morphology could occur during the sampling, collection and analysis of
aerosol samples. For example, particles may be coated with semi-volatile material that evolves off
the particles under vacuum and under the electron beam. Similarly, particles in ambient air are
generally mixtures or agglomerates of particles coming from multiple sources and have a diverse
chemical make-up.
3.3. Sources, Emissions and Deposition of Primary and
Secondary PM
PM is composed of both primary (derived directly from emissions) or secondary (derived from
atmospheric reactions involving gaseous precursors) components. Table 3-4 summarizes
anthropogenic and natural sources for the major primary and secondary aerosol constituents of fine
and coarse particles. Anthropogenic sources can be further divided into stationary and mobile
sources. Stationary sources include fuel combustion for electrical utilities, residential space heating
and cooking; industrial processes; construction and demolition; metal, mineral, and petrochemical
processing; wood products processing; mills and elevators used in agriculture; erosion from tilled
lands; waste disposal and recycling; and biomass combustion. Biomass combustion encompasses
many emission activities including burning of wood for fuel, burning of vegetation to clear land for
agriculture and construction, to dispose of agricultural and domestic waste, to control the growth of
animal or plant pests, and to manage forest resources (prescribed burning). Wildlands also burn due
to lightning strikes and arson. Mobile or transportation-related sources include direct emissions of
primary PM and secondary PM precursors from highway vehicles and non-road sources as well as
fugitive dust from paved and unpaved roads. Also shown in Table 3-2 are sources for several
precursor gases, the oxidation of which can form secondary PM. An overview of estimates of
emissions of primary PM and precursors to secondary PM from major sources is given in
Section 3.3.1. The transformations from gaseous precursors shown in Table 3-2 to secondary PM are
described in Section 3.3.2.
In general, the sources of fine PM are very different from those of coarse PM. Some of the
mass in the fine size fraction forms during combustion from material that has volatilized in
combustion chambers and then recondensed before emission to the atmosphere. Some ambient PM2.s
forms in the atmosphere from photochemical reactions involving precursor gases. Included in this
category is the formation of new UFPs by (1) homogeneous nucleation of precursor gases and (2) the
condensation of gases on pre-existing particles. Biological material also exists in the fine fraction
including many types of microorganisms, especially viruses and bacteria and fragments of pollens
and fungal spores. PMi0_2.5 is mainly primary in origin, as it is produced by surface abrasion or by
suspension of biological material and fragments of living things (e.g., plant and insect debris). In
addition, atmospheric reaction products condense on coarse particles. Because precursor gases
undergo mixing during transport from their sources and chemical reactions, and the oxidation of
different gases can produce the same reaction products, it is difficult to identify individual sources of
secondary PM. Transport and transformation of precursors can occur over distances of hundreds of
kilometers. PMi0_2.5 has a shorter lifetime in the atmosphere, so its effects tend to be more localized.
However, intercontinental transport of dust from African and Asian deserts occurs and some of this
material is in the PMi0_2.5 size range. Major intercontinental dust events are highly episodic but small
contributions can be present at other times (see Section 3.7).
December 2009 3-6
-------
Table 3-2. Constituents of atmospheric particles and their major sources.
Primary (PM <2.5 urn)
Primary (PM >2.5 urn)
Secondary PM Precursors
(PM<2.5um)
Aerosol species Natural Anthropogenic
Natural
Anthropogenic
Natural
Anthropogenic
Sulfate (S042~)
Sea spray
Fossil fuel
combustion
Sea spray
Oxidation of reduced
sulfur gases emitted
by the oceans and Oxidation of S02
wetlands and S02 emitted from fossil fuel
and H2S emitted by combustion
volcanism and forest
fires
Nitrate (N03~)
Minerals
Ammonium (NH4*)
Organic carbon (00)
EC
Metals
Bioaerosols
Mobile source
exhaust
Fugitive dust from
paved and unpaved
Erosion and re- roads, agriculture,
entrainment forestry,
construction, and
demolition
Mobile source
exhaust
Prescribed burning,
wood burning,
\Ajwfiroc. mobile source
Wlldfires exhaust, cooking, tire
wear and industrial
processes
Mobile source
exhaust (mainly
Wildfires diesel), wood
biomass burning,
and cooking
Fossil fuel
combustion, smelting
Volcanicactivity Jg*",
processes, and
brake wear
Viruses and
bacteria
-
Erosion and re-
entrainment
-
Soil humic matter
-
Erosion, re-
entrainment, and
organic debris
Plant and insect
fragments, pollen,
fungal spores, and
bacterial
agglomerates
-
Fugitive dust, paved
and unpaved road
dust, agriculture,
forestry,
construction, and
demolition
-
Tire and asphalt
wear, paved and
unpaved road dust
Tire and asphalt
wear, paved and
unpaved road dust
-
-
Oxidation of NOX
produced by soils,
forest fires, and
lightning
-
Emissions of NH3
from wild animals,
and undisturbed soil
Oxidation of
hydrocarbons
emitted by
vegetation (terpenes,
waxes) and wild fires
-
.—
-
Oxidation of NOX
emitted from fossil fuel
combustion and in
motor vehicle exhaust
-
Emissions of NH3 from
motor vehicles, animal
husbandry, sewage,
and fertilized land
Oxidation of
hydrocarbons emitted
by motor vehicles,
prescribed burning,
wood burning, solvent
use and industrial
processes
-
-
-
Dash (—) indicates either very minor source or no known source of component.
Source: U.S. EPA (2004, 0569051.
Only major sources for each constituent within each broad category shown at the top of Table
3-2 are listed. Not all sources are equal in magnitude. Chemical characterizations of primary
particulate emissions for a wide variety of natural and anthropogenic sources (as shown in Table 3-2)
were given in Chapter 5 of the 1996 PM AQCD (U.S. EPA, 1996, 079380). Summary tables of the
composition of source emissions presented in the 1996 PM AQCD (U.S. EPA, 1996, 079380) and
updates to that information are provided in Appendix 3D to the 2004 PM AQCD (U.S. EPA, 2004,
056905). Source composition profiles are archived by the EPA at
http://www.epa.gov/ttn/chief/software/speciate/. The profiles of source composition were based in
large measure on the results of studies that collected source signatures for use in source
apportionment studies.
December 2009
3-7
-------
Natural sources of primary PM include windblown dust from undisturbed land, sea spray, and
biological material. The oxidation of a fraction of terpenes emitted by vegetation and reduced sulfur
species from anaerobic environments leads to secondary PM formation. Ammonium (NH4+) ions,
which play a major role in regulating the pH of particles, are derived from emissions of NH3 gas.
Source categories for NH3 have been divided into emissions from undisturbed soils (natural) and
emissions that are related to human activities (e.g., fertilized lands, domestic and farm animal waste).
There is ongoing debate about characterizing emissions from wildfires as either natural or
anthropogenic. Wildfires have been listed in Table 3-2 as natural in origin, but land management
practices and other human actions affect the occurrence and scope of wildfires. For example, fire
suppression practices allow the buildup of combustible fuels and increase the susceptibility of forests
to more severe and infrequent fires from whatever cause, including lightning strikes. Similarly,
prescribed burning is listed as anthropogenic, but can be viewed as a substitute for wildfires that
would otherwise occur eventually on the same land.
3.3.1. Emissions of Primary PM and Precursors to Secondary PM
U.S. national average emissions of primary PM2.5, PMi0 and gaseous precursor species (SO2,
NOX, NH3 and VOCs) from different source categories are shown in Figure 3-3. Note that the entries
refer mainly to anthropogenic sources, with little information about natural sources. However, for
categories such as VOCs, the contribution from biogenic emissions of isoprene and terpenes can be
quite large. The entries are continually undergoing revision and are subject to varying degrees of
uncertainty. For example, almost all of the sulfur in fuel is released as volatile components (SO2 or
SO3) during combustion. Hence, sulfur emissions can be calculated on the basis of sulfur content in
fuels to a greater accuracy than can be done for other pollutants like nitrogen oxides or primary PM.
There have been notable downward revisions to the inventories since 2002 in the emissions of dust
from roads. These have resulted in large measure from incorporation of emissions test data with
updated methods for measuring dust emissions. Also, the spatial and temporal characterization of
wildfire emissions has improved since 2002 by integrating satellite-derived fire detection and state-
of-the-art fuels characterization and consumption models (Pouliot et al., 2008, 156883). Emission
measurements from high-temperature combustion sources are sensitive to the dilution, temperature,
and pre-treatment of dilution air (England et al., 2007, 156420; England et al., 2007, 156421; Sheya
etal, 2008, 156977).
To a large extent, especially with regard to the contribution of road dust to PM, refinements in
emission estimates have been guided by the use of receptor modeling. See Section 3.6.1 for a
description of receptor modeling techniques and the 2004 PM AQCD (U.S. EPA, 2004, 056905) for
the role of receptor models in refining emissions estimates. Note that since the estimates given in
Figure 3-3 are U.S. national averages, they may not accurately reflect the contribution of specific
local sources determining a person's exposures to PM at any given time and location.
As can be seen from a comparison of the total U.S. emissions (in million metric tons) shown in
Figure 3-3, estimates of total emissions of potential precursors to secondary PM formation including
SO2, NOX, NH3 and VOCs are considerably larger than those for primary PM sources. However,
translating the emissions of precursors into production rates of secondary PM or using these
emissions as a guide to estimate PM composition is highly problematic. A significant fraction of
gaseous precursors are lost before they could be converted to PM. Dry deposition and precipitation
scavenging of some of these gaseous precursors and their intermediate oxidation products occur
before they are transformed in the atmosphere, and most VOCs emitted are oxidized to carbon
dioxide (CO2) rather than to PM. Some of these precursors are also transported outside the United
States. Even if gaseous precursors are converted to PM components, the effects of atmospheric
transformations must also be considered (discussed in Section 3.3.2 below). As a result of these
transformations, ratios of masses of particle phase products to each other will not be the same as
those in the emissions inventories for their precursors. For example, SO2, NOX and NH3 are
converted to secondary PM as SO42", NO3" and NH4+. The ratios of molecular weights of PM
products (SO42~, NO3" and NH4+) to gaseous precursors (SO2, NOX and NH3) are 1.5, 1.35, and 1.07,
respectively. Estimating a conversion factor for carbon in VOCs is less straightforward. The
oxidation of VOCs leads to the formation of secondary OC in PM. As the result of atmospheric
transformations, nitrogen and oxygen are added to the carbon originally present in VOCs. Turpin and
Lim (2001, 017093) recommend adjustment factors ranging from 1.4 to 2.0 to account for the
presence of oxygen and nitrogen in organic compounds in OC in the aerosol phase. Because of all
December 2009 3-8
-------
the above issues, the resultant mass and composition of ambient PM is quite different from what
might be inferred from examining the emissions inventories alone.
Fertilizer &
Lweslock
0%
hdust/
CormVRes
Fuels
Non-Road
2%
PM25(5.4MMT)
2% 1%
PMm (19.9 MMT}
Resdenlal
On-R=ad Woo
-------
3.3.2. Formation of Secondary PM
Precursors to secondary PM have natural and anthropogenic sources, just as primary PM has
natural and anthropogenic sources. A substantial fraction of the fine particle mass, especially during
the warmer months of the year, is secondary in nature, formed as the result of atmospheric reactions
involving both inorganic and organic gaseous precursors. The major atmospheric chemical
transformations leading to the formation of particulate nitrate (pNO3) and particulate sulfate (pSO4)
are relatively well understood; whereas those involving the formation of secondary organic aerosol
(SOA) are less well understood and are subject to much current investigation. A large number of
organic precursors are involved and many of the kinetic details still need to be determined. Also,
many of the products of the oxidation of hydrocarbons have yet to be identified. However, there has
been substantial progress made in understanding the chemistry of SOA formation in the past few
years.
3.3.2.1. Formation of Nitrate and Sulfate
The basic mechanism of the gas and aqueous phase oxidation of NO2 and SO2 has long been
studied and can be found in numerous texts on atmospheric chemistry, e.g., Seinfeld and Pandis
(1998, 018352). Finlayson-Pitts and Pitts (2000, 055565). Jacob (1999, 091122). and Jacobson
(2002, 090667). The reader is referred to the 2004 PM AQCD (U.S. EPA, 2004, 056905) as well as
the 2008 NOX and SOX ISAs (U.S. EPA, 2008, 157073: U.S. EPA, 2008, 157074: U.S. EPA, 2008,
157075) where these processes are described in great detail.
3.3.2.2. Formation of Secondary Organic Aerosol
Some key new findings have altered perceptions of SOA formation since the 2004 PM AQCD
(see especially the reviews by Kroll and Seinfeld (2008, 155910) and Rudich et al. (2007, 156059)).
New measurement techniques for estimating the speciation and water solubility of organic aerosols
have noted the dominant contribution of oxygenated species in atmospheric particles. Recent
measurements show that the abundance of oxidized SOA exceeds that of more reduced hydrocarbon
like organic aerosol in Pittsburgh (Zhang et al., 2005, 157185) and in about 30 other cities across the
Northern Hemisphere (Zhang et al., 2007, 101119). Based on aircraft and ship-based sampling of
organic aerosols in coastal waters downwind of northeastern U.S. cities, de Gouw et al. (2008,
191757) reported that 40-70% of measured organic mass was water soluble and estimated that
approximately 37% of SOA is attributable to aromatic precursors, based on PM yields estimated for
NOx-limited conditions. However, the remaining mass of estimated SOA (63%) was unexplained,
possibly due to oxidation of semivolatile precursors not measured by standard gas chromatography.
Aerosol yields from the oxidation of aromatic compounds have been reported to be higher when
reactions with NOX are not dominant, suggesting that transport of less reactive compounds (e.g.,
benzene) out of source regions with high NOX levels could result in greater overall SOA yields than
previously estimated (Ng et al., 2007, 199528). Furthermore, Zhang et al. (2007, 189998) noted that
the most common mass spectrum of oxygenated OA measured in ambient air resembles mass spectra
measured in irradiated diesel exhaust reported by Robinson et al. (2007, 156053) and Sage et al.
(2008, 191758).
Typical dilution sampling of combustion sources employ dilution rate, temperature, pressure,
and background aerosol concentrations that can differ substantially from ambient conditions. Lipsky
and Robinson (2006, 189891) and Robinson et al. (2007, 156053) showed that under higher dilution
conditions, the fraction of diesel engine organic emissions that volatilizes is higher than that
measured using common test methods.
Murphy and Pandis (2009, 190095) pointed out the importance of characterizing the volatility
distribution of emissions of organic species from combustion sources for more accurately predicting
the abundance and oxidation state of SOA in both urban and surrounding regional background
environments. They note that in urban areas, volatile emissions can be photochemically oxidized to
more non-volatile compounds which then condense, forming oxidized SOA. Braun (2009, 189997)
suggested that the weathering of diesel exhaust particles involves the desorption of semi-volatile
organic compounds followed by the decomposition and reaction of the amorphous non-volatile
carbon. These reactions would result in the formation of a number of functional groups on the
December 2009 3-10
-------
surface of the carbon core including quinones, carbohydroxide and carboxyl groups as well as
sulfate. In general, all the above studies underscore the importance of accurately describing the
phase distribution of semivolatile organic compounds emitted by combustion sources under
atmospheric conditions, and of atmospheric photochemical reactions in modifying the composition
of emissions.
Until a few years ago, the oxidation of terpenes and aromatic compounds were considered as
sources of SOA, and the oxidation of isoprene was not considered a source of SOA. However,
observations of 2-methyl tetrols in ambient samples from a number of different environments
suggest that small but not insignificant quantities of SOA are formed from isoprene oxidation
(Claeys et al, 2004, 058608). Laboratory studies also indicate the formation of 2-methyl tetrols from
isoprene oxidation (Edney et al., 2005, 155760; Kleindienst et al., 2006, 156650). Xia and Hopke
(2006, 179947) observed the seasonal variations for the two major diastereoisomers produced during
the oxidation of isoprene with highest concentrations occurring during summer and lowest
concentrations occurring during winter. During summer, the maximum contribution of these two
diastereoisomers to OC was 2.8%, however it is not clear if more SOA could have been produced
from isoprene oxidation.
Kroll and Seinfeld (2008, 155910) and Rudich et al. (2007, 156059) noted that the
composition of SOA evolves from repeated cycles of volatilization and condensation of chemical
reaction products in both the particle and gas phases. Rudich et al. (2007, 156059) focused on the
oxidation of particle phase species by reaction with gas phase oxidants. Kroll and Seinfeld (2008,
155910) identified three factors that determine the SOA forming potential of organic compounds in
the atmosphere:
1. Oxidation reactions of gas-phase organic species. These species include alkanes, alkenes,
aromatics, cyclic olefms, isoprene and terpenes. Note that oxidation reactions can either
lower volatility by addition of functional groups or increase volatility by cleavage of
carbon-carbon bonds;
2. Reactions in the particle, or condensed, phase that can change volatility either by oxidation
or formation of high-molecular-weight species. These reactions can lead to the formation of
oligomers, thereby decreasing volatility or to the formation of more volatile products; and
3. Ongoing reactions that result from the varied volatility of oxidation products.
Other detailed work has focused on the formation of higher molecular weight particle-phase
oligomers (Gao et al., 2004, 156460; Kalberer et al., 2004, 156619; Tolocka et al., 2004, 087578).
the importance of cloud processing in the evolution of SOA (Blando and Turpin, 2000, 155692;
Gelencser and Varga, 2005, 156463). and the role of acid seeds in oligomer formation (Tolocka et
al., 2004, 087578). These results imply that ambient samples could contain mixtures of SOA from
different sources at different stages of processing, some with common reaction products making
source identification of SOA problematic. Figure 3-4 shows a schematic of processes involved in the
formation of SOA.
December 2009 3-11
-------
PRIMARY
SECONDARY
Organic
Compounds
Gas phase
reactions
of alkanes,
aromatics, alkenes,
olefins, etc
and -OH, O3, NO2
products
Aqueous
phase
reactions
of dicarbonyls,
organic acids
and -OH, -NO2, O3
Aerosol
phase
reactions
products
products
Gas/particle partitioning
Nucleation, sorption, condensation/evaporation
Figure 3-4. Primary emissions and formation of SOA through gas, cloud and condensed
phase reactions.
It should be noted that many of the products of terpene oxidation are oxidative in nature, and
are not merely nonreactive oxidation products. Organic peroxides represent an important class of
reactive oxygen species (ROS) that have high oxidizing potential and could cause oxidative stress in
cells on which these species deposit. For example, Docherty (2005, 087613) found evidence for the
substantial production of organic hydroperoxides in secondary organic aerosol (SOA) resulting from
the reaction of monoterpenes with O3. Analysis of the SOA formed in their environmental chamber
indicated that the SOA was mainly organic hydroperoxides. In particular, they obtained yields of
47% and 85% of organic peroxides from the oxidation of a- and |3-pinene. The hydroperoxides then
react with aldehydes in particles to form peroxyhemiacetals, which can either rearrange to form other
compounds such as alcohols and acids or revert back to the hydroperoxides. The aldehydes are also
produced in large measure during the ozonolysis of the monoterpenes. Monoterpenes also react with
OH radicals resulting, however, in the production of more lower molecular weight products than in
their reaction with O3. Various terpenoid compounds are used in a number of household products and
can be oxidized by ozone that has infiltrated from outdoors. The oxidation of terpenoid compounds
indoors produces UFPs as described in the 2006 O3 AQCD (U.S. EPA, 2006,
3.3.2.3. Formation of New Particles
In addition to being emitted by high temperature combustion sources, new particles can form
by nucleation of atmospheric gases. New particle formation has been observed in environments
ranging from relatively unpolluted marine and continental environments to polluted urban areas
(Kulmala et al, 2004, 089159). These new, nucleation mode particles are formed from molecular
clusters. Competition between condensation of gases and clusters onto existing particles and
coagulation of clusters determines which process will dominate (McMurry et al., 2005, 191759).
Because of this competition, it is expected that particle number concentrations are dominated by
primary anthropogenic emissions in highly polluted settings and by nucleation in remote continental
sites. However, nucleation still occurs in urban environments and can still be the major source under
certain conditions. The composition of UFPs will differ depending on the nature of their sources.
December 2009
3-12
-------
Nucleation is observed in the morning and extending into the afternoon, and occurs at higher rates
during summer than during winter, consistent with a photochemical process. New particle formation
events have been observed to occur over distances of several hundred kilometers in what have been
called regional nucleation events (Shi et al., 2007, 191760).
The major gas phase nucleating species involved are sulfuric acid vapor and water vapor.
Kuang et al. (2008, 191196) suggested that the rate of nucleation is second order with respect to
H2SO4 vapor depending on mechanism. However, other studies (e.g., Kulmala et al., 2007, 097838)
have suggested that the nucleation rate is first order with respect to H2SO4 vapor. These differences
imply that a number of mechanistic details still remain to be determined, including the interactions
with other species. However, this disparity is small compared to classical thermodynamic binary
nucleation theory involving H2SO4 and water vapor, in which the nucleation rate is given by H2SO4
vapor to at least the 10th power (Kulmala et al., 1998, 129411). H2SO4 vapor is produced by the gas
phase oxidation of SO2 by OH radicals (U.S. EPA, 2008, 157075). Ammonia (Gaydos et al., 2005,
191762) and organic acids and bases (amines) are also involved to some extent (Smith et al., 2008,
199529). The formation of UFPs indoors from the oxidation of terpenoids by O3 as mentioned above
also indicates that nucleation occurs indoors as well (U.S. EPA, 2006, 088089).
3.3.3. Mobile Source Emissions
3.3.3.1. Emissions from Gasoline Fueled Engines
PM emitted from gasoline fueled engines is a mix of OC, EC and small quantities of trace
metals and sulfates, with OC constituting anywhere from 26-88% of PM (Cadle et al., 1999, 007636:
Geller et al., 2006, 139644; Schauer et al., 2002, 035332). Most of the compounds in OC have yet to
be characterized. High molecular weight and large PAHs have been identified in gasoline fueled
vehicle emissions (Phuleria et al., 2006, 156867; Riddle et al., 2007, 115272). EPA exhaust emission
standards do not control PM from gasoline vehicles as stringently as diesel vehicles. PM emissions
from gasoline fueled vehicles decreased greatly as other exhaust emissions (primarily hydrocarbons
and carbon monoxide) were controlled by improvements in the catalytic converter and better control
of air-to-fuel mixture ratios for the engine intake. When leaded gasoline was used in pre-1975 model
year vehicles, gasoline engine PM emissions were relatively large (about 300 mg/mile) and consisted
largely of lead salts from combustion of the lead additive. A current gasoline fueled vehicle emits far
lower PM, about 1-10 mg/mile. Emissions of gasoline PM increase at colder ambient temperatures
and recent rulemaking for air toxics will result in significant reduction of gasoline PM at colder
temperatures. Further details about the composition of motor vehicle emissions in the context of
source apportionment modeling can be found in Section 3.6.1.
3.3.3.2. Emissions from Diesel Fueled Engines
Marti Maricq (2007, 155973) presents a conceptual model of diesel PM as a mix of
nucleation-mode SO42~ and hydrocarbons from unspent fuel and soot embedded with trace metals on
which SO42~ and hydrocarbons condense. PM emissions from pre-2007 diesels consist largely of EC
(about 70% by mass) and OC (high molecular weight compounds derived from both diesel fuel and
lubricating oil) which is responsible for about 25% of the PM (Maricq, 2007, 155973). The EC is
non-volatile, while the organic material present exhibits temperature-dependent evaporation in
similar fashion to a mixture of C24-C32 alkanes (Sakurai et al., 2003, 113924). Sulfates constitute
about 5% of the PM. A small fraction of the diesel fuel sulfur (typically about 1-2%) is oxidized to
sulfate. Trace elements (such as Zn and halogens, mainly from lubricating oil; and others) are also
present. Mass spectra of organic diesel particles from pre-2007 engines appear to be largely similar
to engine lube oil with minor contributions from unburned diesel fuel.
Effective with the 2007 model year for on-road diesel heavy-duty highway truck engines, the
new EPA PM standard (0.01 g per brake horsepower-hour [g/bhp-h]) reduced PM emission limits by
90% from the prior standard (0.10 g/bhp-h). By comparison, uncontrolled heavy-duty diesels (pre-
1988 model years) emitted about 1-2 g/bhp-h of PM. The 2007 standard resulted in the introduction
of new emission control technology, mainly the diesel particulate filter (DPF). Other elements of the
December 2009 3-13
-------
new control technology also include water-cooled exhaust gas recirculation (mostly for control of
NOX), a diesel oxidation catalyst (DOC) used in some vehicles and improved fuel injection systems.
Besides the large reduction in diesel PM on a mass basis, the composition of diesel PM
changed greatly. EPA regulations required that diesel fuel for on-road vehicles contain no more than
15 ppm sulfur as of January 2007 (Lim et al., 2007, 155931). Prior to that, the limit on diesel fuel
sulfur established in 1995 for on-road vehicles was 500 ppm. The HEI-ACES study characterizes
emissions from four engines and shows that PM emissions are about 0.001 g/bhp-h or 90% below
the level of the current (2007) emission standard (Shimpi et al., 2009, 189888). This study also
characterizes the composition of the much lower mass of PM emitted with this new technology. PM
samples collected over a composite type test consisted of 53% sulfate, 30% OC, 13% EC and 4%
other components, including metals. A substantial fraction of sulfur is converted to sulfate over the
diesel particulate filter resulting in the higher fractional content of sulfate emissions. However, due
to the much lower mass of PM being emitted (over a 90% reduction compared to earlier diesels) as
well as the low sulfur content of the fuel, the total mass of sulfate emitted is somewhat less than that
from earlier diesels. This work also shows that UFP number emissions are lower (about 90% lower)
and that a number of other emissions are also controlled, including PAHs, nitro-PAHs, carbonyls
(such as aldehdyes), and metals.
Pre-2007 engines can be retrofitted with exhaust aftertreatment devices, including DPFs,
DOCs, and selective catalytic reduction (SCR) systems to reduce emissions (U.S. EPA, 2009,
189885). Hu et al. (2009, 189886) examined emissions of various metals (V, Pt) from various diesel
retrofit systems including those using V-SCR and a zeolite-based SCR with a DPF. Pakbin et al.
(2009, 189893) shows significant reductions in emissions of various PAHs for diesels with SCR
retrofit systems for NOX control. Biswas et al. (2008, 189969) examined PM size distribution and
composition (including semi-volatiles and non-volatiles) from several advanced technology diesels
including those with SCR. They showed major reductions in PM number in most driving conditions
but did not show a reduction in PM number under cruise conditions. Biswas et al. (2009, 189880)
examined four heavy-duty diesel vehicles with various retrofits showing, in general, large reductions
in PM but, in some cases, somewhat higher emissions (or smaller decreases than expected) for EC
andOC.
In general, under light load conditions such as idle, diesel PM has a higher percentage of OC
emissions than at high load conditions. Under lighter loads, organic compounds are not oxidized as
effectively as under high loads. Under higher loads, PM contains more EC than under light loads.
Also, newer model year diesels through the 2006 model year tend to have a higher fraction of PM
that is EC than older models.
Emissions have been measured with the new technology engines under a number of driving
cycles besides the Federal Test Procedure. The HEI ACES study examined emissions during a range
of test procedures. In general, the low-load test cycle resulted in lower exhaust temperatures and
higher emission than did the high-load cycles. Regeneration events also produced short-term
increases in particle emissions. However, particle emission measurements on the ACES engines were
consistently lower than those on a typical 2004 engine (Shimpi et al., 2009, 189888).
EPA standards will result in non-road diesels also having technology like catalyzed diesel
particulate filters starting in 2012. Similar standards have also been promulgated for locomotives
powered by diesel engines. Some work has been done with prototype SCR systems for diesel NOX
control such as would be used for the 2010 diesel NOX standard. This standard will also result in
reductions for NOX similar to those seen for PM in the 2007 standard.
There is no information on emissions from diesel engines with this new technology at
temperatures of < 10° C. In general, the ratio of emissions under cold start conditions at low
temperatures to emissions at 24° C is significantly higher for diesel engines with new technology
compared to emissions from non-catalyst systems. Note that the engines with the newer technology
require time to allow the catalyst to reach normal operating temperatures for full emission
reductions. During the period of catalyst heating, particles will be trapped in the DPF, but volatile
components can pass through. As a result of this particle-trapping, post-2007 model year engines still
emit less than the older engines, even under cold start conditions.
3.3.4. Deposition of PM
Wet and dry deposition are important processes for removing PM and other pollutants from the
atmosphere on urban, regional, and global scales. The conceptual model for dry deposition is to view
December 2009 3-14
-------
the flux deposited on a surface as the product of a concentration (mass or moles of a pollutant/m3)
times a deposition velocity (Vd) (m/s). Therefore, deposition has the units of mass per unit time per
unit area, or flux. The general approach used to estimate Vd for gases or very small particles is the
resistance-in-series method represented by Equation 3-1:
Vd=l/(Ra+Rh+Rc)
Equation 3-1
where Ra, Rb, and Rc represent the resistance due to atmospheric turbulence, transport in the fluid
sublayer very near the elements of the surface, such as leaves or soil, and the resistance to uptake of
the surface itself, respectively. Typically, these resistances are empirically derived and can vary as a
function of wind speed, solar radiation, plant characteristics, precipitation/moisture, and soil/air
temperature. These processes are shown schematically in Figure 3-5. This approach works for a
range of substances, although it is inappropriate for species with substantial re-emissions from the
surface or for species where deposition to the surface depends on concentrations at the surface itself.
The approach is also modified somewhat for aerosols where Rb and Rc are replaced with a surface Vd
to account for gravitational settling.
aerodynamic
laminar' sub-layer
Rb
Atmospheric
Resistances
Canopy
Resistances
Resistance analogy for the deposition of atmospheric pollutants
Source: Courtesy of T. Pierce, USEPA/ ORD / NERL/Atmospheric Modeling Division.
Figure 3-5. Schematic of the resistance-in-series analogy for atmospheric deposition.
Wesley and Hicks (2000, 025018) listed several shortcomings of the then-current knowledge
of dry deposition. Among those shortcomings were difficulties in representing dry deposition over
December 2009
3-15
-------
varying terrain where horizontal advection plays a significant role in determining the magnitude of
Ra and difficulties in adequately determining Vd for extremely stable conditions such as those
occurring at night; see the discussion by Mahrt (1998, 048210). Under optimal conditions, when a
model is exercised over a relatively small area where dry deposition measurements have been made,
models still generally showed uncertainties on the order of ± 30% (e.g., Brook et al., 1996, 024023;
Massman et al., 1994, 043681: Padro, 1996, 052446: Wesely and Hicks, 2000, 025018). Wesley and
Hicks (2000, 025018) concluded that an important result of those comparisons was that the level of
sophistication of most dry deposition models was relatively low, and that deposition estimates,
therefore, must rely heavily on empirical data. Still larger uncertainties exist when the surface
features in the built environment are not well known or when the surface comprises a patchwork of
different surface types, as is common in the eastern U.S.
3.3.4.1. Deposition Forms
Wet Deposition
Wet deposition results from the incorporation of atmospheric particles and gases into cloud
droplets and their subsequent precipitation as rain or snow, or from the scavenging of particles and
gases by raindrops or snowflakes as they fall (Lovett, 1994, 024049). Wet deposition depends on
precipitation amount and ambient pollutant concentrations. Vegetation surface properties have little
effect on wet deposition, although leaves can retain liquid and solubilized PM.
Landscape characteristics can affect wet deposition via orographic effects and by the closer
aerodynamic coupling to the atmosphere of tall forest canopies as compared to the shorter shrub and
herbaceous canopies. Following wet deposition, humidity and temperature conditions further affect
the extent of drying versus concentrating of solutions on foliar surfaces, which influence the rate of
metabolic uptake of surface solutes (Swietlik and Faust, 1984, 046678). The net consequence of
these factors on direct physical effects of wet deposited PM on leaves is not known (U.S. EPA, 2004,
056905).
Rainfall introduces new wet deposition and also redistributes throughout the canopy
previously dry-deposited particles (Peters and Eiden, 1992, 045277). The concentrations of
suspended and dissolved materials are typically highest at the onset of precipitation and decline with
duration of individual precipitation events (Hansen et al., 1994, 046634). Sustained rainfall removes
much of the accumulation of dry-deposited particles from foliar surfaces, reducing direct foliar
effects and combining the associated chemical burden with the wet-deposited material (Lovett, 1994,
024049) for transfer to the soil. Intense rainfall may contribute substantial total particulate inputs to
the soil, but it also removes bioavailable or injurious pollutants from foliar surfaces. This washing
effect, combined with differential foliar uptake and foliar leaching of different chemical constituents
from particles, alters the composition of the rainwater that reaches the soil and the pollutant burden
that is taken-up by plants. Once in the soil, these particle constituents may affect biogeochemical
cycles of major, minor, and trace elements. Low intensity precipitation events, in contrast, may
deposit significantly more particulate pollutants to foliar-surfaces than high intensity precipitation
events. Additionally, low-intensity events may enhance foliar uptake through the hydrating of some
previously dry-deposited particles (U.S. EPA, 2004, 056905)
Dry Deposition
Dry particulate deposition, especially of heavy metals, base cations, and organic contaminants,
is a complex and poorly characterized process. It appears to be controlled primarily by such
variables as atmospheric stability, macro- and micro-surface roughness, particle diameter, and
surface characteristics (Hosker and Lindberg, 1982, 019118). The range of particle sizes, the
diversity of canopy surfaces, and the variety of chemical constituents in airborne particles have made
it difficult to predict and to estimate dry particulate deposition (U.S. EPA, 2004, 056905).
Dry deposition of atmospheric particles to plant and soil surfaces affects all exposed surfaces.
Larger particles >5 um diameter are dry-deposited mainly by gravitational sedimentation and inertial
impaction. Smaller particles, especially those with diameters between 0.2 and 2.0 um, are not readily
December 2009 3-16
-------
dry-deposited and may travel long distances in the atmosphere until their eventual deposition, most
often via precipitation. Plant parts of all types, along with exposed soil and water surfaces, receive
steady deposits of dry dusts, EC, and heterogeneous secondary particles formed from gaseous
precursors (U.S. EPA, 1982, 017610).
Estimates of regional particulate dry deposition infer fluxes from the product of variable and
uncertain measured or modeled parti culate concentrations in the atmosphere and even more variable
and uncertain estimates of Vd parameterized for a variety of specific surfaces (e.g., Brook et al,
1996, 024023). Even for specific sites and well-defined particles, uncertainties are large. Modeling
the dry deposition of particles to vegetation is at a relatively early stage of development, and it is not
currently possible to identify a best or most generally applicable modeling approach (U.S. EPA,
2004, 056905).
Deposition from Clouds and Fog
The occurrence of cloud and fog deposition tends to be geographically restricted to coastal and
high mountain areas. Several factors make it particularly effective for the delivery of dissolved and
suspended particles to vegetation. Concentrations of particulate-derived materials are often many-
fold higher in cloud or fog water than in precipitation or ambient air due to orographic effects and
gas-liquid partitioning. In addition, fog and cloud water deliver particulate chemical species in a
bioavailable-hydrated form to foliar surfaces. This enhances deposition by sedimentation and
impaction of submicron aerosol particles that exhibit low Vd before fog droplet formation (Fowler et
al., 1989, 002515). Deposition to vegetation in fog droplets is proportional to wind speed, droplet
size, concentration, and fog density. In some areas, typically along foggy coastlines or at high
elevations, this deposition represents a substantial fraction of total deposition to foliar surfaces
(Fowler et al., 1991, 046630).
3.3.4.2. Methods for Estimating Dry Deposition
Methods for estimating dry deposition of particles are more restricted than for gaseous species
and fall into two major categories: surface analysis methods, which include all types of estimates of
contaminant accumulation on surfaces of interest, and atmospheric deposition rate methods, which
use measurements of contaminant concentrations in the atmosphere and descriptions of surrounding
surface elements to estimate deposition rates (Davidson and Wu, 1990, 036799). Surface extraction
or washing methods characterize the accumulation of particles on natural surfaces of interest or on
experimental surrogate surfaces. These techniques rely on methods designed specifically to remove
only surface-deposited material. Total surface rinsate may be equated to accumulated deposition or
to the difference in concentrations in rinsate between exposed and control (sheltered) surfaces and
may be used to refine estimates of deposition. Foliar extraction techniques may underestimate
deposition to leaves because of uptake and translocation processes that remove pollutants from the
leaf surface (Garten and Hanson, 1990, 036803: Taylor et al., 1988, 019289). Foliar extraction
methods also cannot distinguish gas from particle-phase sources (Bytnerowicz et al., 1987, 036493;
Dasch, 1987, 036496; Kelly, 1988, 037379; Lindberg and Lovett, 1985, 036530; Van Aalst, 1982,
036481).
The National Dry Deposition Network was established in 1986 to document the magnitude,
spatial variability, and trends in dry deposition across the United States. Currently, the network
operates as a component of the CASTNet (Clarke et al., 1997, 025022). A significant limitation on
current capacity to estimate regional effects of NOX and SOX deposition is inadequate knowledge of
the mechanisms and factors governing particle dry deposition to diverse surfaces (U.S. EPA, 2004,
056905).
Collection and analysis of stem flow and throughfall can also provide useful estimates of
particulate deposition when compared to directly sampled precipitation. The method is most precise
for particle deposition when gaseous deposition is a small component of the total dry deposition and
when leaching or uptake of compounds of interest out of or into the foliage is not a significant
fraction of the deposition because these lead to positive and negative artifacts in the calculated totals.
Foliar washing, whether using precipitation or experimental lavage, is one of the best available
methods to determine dry deposition to vegetated ecosystems. Major limitations include the site
specificity of the measurements and the restriction to elements that are largely conserved within the
December 2009 3-17
-------
vegetative system. Surrogate surfaces have not been found that can adequately replicate essential
features of natural surfaces, and therefore do not produce reliable estimates of particle deposition to
the landscape.
Micrometeorological methods employ eddy covariance, eddy accumulation, or flux gradient
protocols for quantifying dry deposition. These techniques require measurements of particulate
concentrations and of atmospheric transport processes. They are currently well developed for ideal
conditions of flat, homogeneous, and extensive landscapes and for chemical species for which
accurate and rapid sensors are available. Additional studies are needed to extend these techniques to
more complex terrain and more chemical species.
The eddy covariance technique measures vertical fluxes of gases and fine particles from
calculations of the mean covariance between the vertical component of wind velocity and pollutant
concentration (Wesely et al., 1982, 036564) using sensors acquiring concentration data at 5-20 Hz.
For the flux gradient or profile techniques, vertical fluxes are calculated from a concentration
difference and an eddy exchange coefficient determined at discrete heights (Erisman et al., 1988,
036510; Huebert et al., 1988, 036569). Most measurements of eddy transport of PM have used
chemical sensors (rather than mass or particle counting) to focus on specific PM components. These
techniques have not been well developed for generalized particles and may be less suitable for coarse
particles that are transported efficiently in high frequency eddies (Gallagher et al., 1988, 046631).
3.3.4.3. Factors Affecting Dry Deposition Rates and Totals
In the size range of-0.1-1.0 um where Vd is relatively independent of particle diameter as
shown in Figure 3-6, parti culate deposition is controlled by roughness of the surface and by the
stability and turbulence of the atmospheric surface layer. Impaction and interception dominate over
diffusion as dry deposition processes, and the Vd is considerably lower than for particles that are
either smaller than ~0.1 um or larger than ~1.0 um.
Deposition of particles between 1 and 10 um diameter is strongly dependent on particle size as
shown in Figure 3-6. Larger particles within this size range are collected more efficiently at typical
wind speeds than are smaller particles (Clough, 1975, 070850). suggesting the importance of
impaction. Impaction is related to wind speed, the square of the particle diameter, and the inverse of
the deposition surface cross-section. As a depositing particle's trajectory deviates from the
streamlines of the air in which it is suspended, increasing either wind speed or the ratio of particle
size to the deposition surface cross-section increases the probability of collision.
Empirical estimates of Vd for fine particles under wind tunnel and field conditions are often
several-fold greater than predicted by theory (Unsworth and Wilshaw, 1989, 046682). A large
number of transport phenomena, including streamlining of foliar obstacles, turbulence structure near
surfaces, and various phoretic transport mechanisms are not well characterized (U.S. EPA, 2004,
056905). The discrepancy between estimated and predicted values of Vd may reflect model
limitations or experimental limitations in the specification of the effective size and number of
deposition obstacles. Previous reviews (e.g., U.S. EPA, 1996, 079380: U.S. EPA, 2004, 056905)
suggest the following generalizations: (1) particles >10 urn exhibit variable Vd between 0.5 and
1.1 cm/s depending on friction velocities, whereas a minimum particle Vd of 0.03 cm/s exists for
particles in the size range 0.1 to 1.0 um; (2) the Vd of particles is approximately a linear function of
friction velocity; and (3) deposition of particles from the atmosphere to a forest canopy is from 2 to
16 times greater than deposition in adjacent open terrain like grasslands or other low vegetation.
December 2009 3-18
-------
O
o
O
8
o.
CD
Q
1,000
100
10
1 —
0.1 —
0,01 —
0.001
0.000
' Stokes Law
» Bro^nian Diffusion
• Peter* and Eiden
• Little and Mfen (1677)
0.001
I
0.1
T
0.01 0.1 1 10
Particle Diameter (Mm)
100
Source: U.S. EPA (2004, 0569051.
Figure 3-6. The relationship between particle diameter and Vd for particles. Values measured
in wind tunnels by Little and Wiffen (1977, 070869) over short grass with wind
speed of 2.5 mi/s closely approximate the theoretical distribution determined by
Peters and Eiden (1992, 045277) for a tall spruce forest. These distributions
reflect the interaction of Brownian diffusivity (descending dashed line), which
decreases with particle size and sedimentation velocity (ascending dotted line
derived from Stokes Law), which increases with particle size. Intermediate-sized
particles (0.1-1.0 urn) are influenced strongly by both particle size and
sedimentation velocity, and deposition is relatively independent of size.
Leaf Surface Effects on Vd
The chemical composition of a particle is not usually considered to be a primary determinant
of its Vd. Rather, the plant leaf surface has an important influence on the Vd of particles, and
therefore on the flux of dry deposition to the terrestrial environment. Relevant leaf surface properties
include stickiness, microscale roughness, and cross-sectional area. These properties affect the
probability of impaction and particle bounce. The efficiency of deposition to vegetation also varies
with leaf shape. Particles impact more frequently on the adaxial (upper) surface than on the abaxial
(lower) surface. Most particles accumulate in the midvein, central portion of leaves. The greatest
particle loading on dicotyledonous leaves is frequently on the adaxial surface at the base of the
blade, just above the petiole junction. Precipitation washing probably plays an important role in this
distribution pattern (U.S. EPA, 2004, 056905).
Lead particles have been shown to accumulate to a greater extent on older as compared with
younger needles and twigs of white pine, suggesting that wind and rain may be insufficient to fully
wash the foliage. Fungal mycelia (derived from windborne spores) were frequently observed in
intimate contact with other particles on leaves, which may reflect minimal re-entrainment of the
spore due to shelter by the particles, mycelia development near sources of soluble nutrients provided
by the particles, or simply co-deposition (Smith and Staskawicz, 1977, 046675).
Leaves with complex shapes tend to collect more particles than do those with shapes that are
more regular. For example, conifer needles are more efficient than broad leaves in collecting
particles by impaction as a result of the small cross-section of the needles relative to the larger leaf
laminae of broadleaves allowing for greater penetration of wind into conifer canopies than broadleaf
ones (U.S. EPA, 2004, 056905).
December 2009
3-19
-------
Canopy Surface Effects on Vd
Surface roughness increases particulate deposition, and Vd is usually greater for a forest than
for a nonforested area and greater for a field than for a water surface. Different size particles have
different transport properties and Vd. The upwind leading edges of forests, hedgerows, and
individual plants are primary sites of coarse particle deposition. Impaction at high wind speed and
the sedimentation that follows the reduction in wind speed and carrying capacity of the air in these
areas lead to preferential deposition of larger particles (U.S. EPA, 2004, 056905).
Air movement is slowed in proximity to vegetated surfaces. Canopies of uneven age or with a
diversity of species are typically aerodynamically rougher and receive larger inputs of dry-deposited
pollutants than do smooth, low, or monoculture vegetation (Garner et al., 1989, 042085; U.S. EPA,
2004, 056905). Canopies on slopes facing the prevailing winds receive larger inputs of pollutants
than more sheltered, interior canopy regions.
All foliar surfaces within a forest canopy are not equally exposed to particle deposition. Upper
canopy foliage tends to receive maximum exposure to coarse and fine particles, but foliage within
the canopy tends to receive primarily fine aerosol exposures. The dry deposition of fine-mode
particles and unreactive gases tends to be more evenly distributed throughout the canopy.
Both uptake and release of PM constituents can occur within the canopy. The leaf surface is a
region of leaching and uptake. Exchange also occurs with epiphytic organisms and bark and through
solubilization of previously dry-deposited PM. Vegetation emits a variety of particles and particulate
precursor materials.
3.4. Monitoring of PM
3.4.1. Ambient Measurement Techniques
3.4.1.1. PMMass
Federal reference methods (FRMs) and federal equivelance methods (FEMs) for PM were
discussed in detail in the 2004 PM AQCD (U.S. EPA, 2004, 056905). Issues discussed there include
the definition or description of FRMs and FEMs for PM2.5 and PMi0, and measurement methods for
PM^.2.5. Also included are detailed descriptions of the WINS impactor, virtual and cascade
multistage impactors for PMi0_2.5 measurement, high-volume and low-volume PMi0 samplers, and
real-time or continuous methods for PM2.5 and PMi0 including:
• Tapered Element Oscillating Microbalance (TEOM) operated at various temperatures;
• Sample Equilibration System (SES)-TEOM;
• Differential TEOM;
• (3-Gauge Techniques (BGT);
• Piezoelectric Microbalance;
• Real-Time Total Ambient Mass Sampler (RAMS);
• Continuous Ambient Mass Monitor (CAMM);
• Continuous Coarse Particle Monitor (CCPM);
• Micro-orifice Uniform Deposit Impactor (MOUDI);
December 2009 3-20
-------
• Multichannel diffusion denuder sampling system (BOSS); and
• Light scattering photometric instruments.
In this section, FRMs and FEMs for PM2.5, PMi0_2.5, and PMi0 will be revisited and evaluated
based on the cumulative understanding of these methods with a focus on evaluations performed
following the 2004 PM AQCD (U.S. EPA, 2004, 056905). followed by the discussion of new
techniques under development or evaluation.
Federal Reference Method and Federal Equivalent Method
FRM and FEM PM samplers are designed to measure the mass concentrations of ambient
particulate matter. The FRMs for measuring PM2 5, PMi0_2.5, and PMi0 are specified in CFR 40 Part
50, Appendices L, O and J, respectively. The FRM for PM2.5 is a hybrid-based method that specifies
certain aspects of the method (e.g., component dimensions and tolerances, sample handling and
analysis) by design specifications and other aspects (e.g., flow control) by performance
specifications (U.S. EPA, 2004, 056905). The PM10 FRM is performance-based; particles are
inertially separated with a penetration efficiency of 50% at 10 ± 0.5 um aerodynamic diameter. The
required collection efficiency as a function of particle size larger and smaller than 10 um
aerodynamic diameter is explicitly specified by a penetration curve in the CFR. Particles larger than
10 um in aerodynamic diameter are collected on the filter with diminishing collection efficiency as
particle size increases. Likewise, particles smaller than 10 um in aerodynamic diameter are
collected on the filter with increasing collection efficiency as particle size decreases. The FRM for
PMio_2.5 concentration is computed as the numeric difference between concurrent and co-located
PMio and PM2.5 concentrations obtained from low-volume FRM samplers of the same make and
model. It should be noted that while the FRM for PM2 5 and PMi0_2.5 reports data under local
conditions, the FRM for PMi0 reports data corrected to standard temperature (298 K) and pressure
(101.3 kPa)(STP).
A very sharp cut cyclone (VSCC) was approved in 2004 as a Class II PM2.5 FEM method
(Kenny et al., 2004, 155895). The VSCC provides superior performance over long sampling periods
under heavy loading and was also incorporated as an optional second-stage separator for the PM2 5
FRM (71 FR 61214, October 17, 2006). In 2006, EPA finalized new performance criteria (40 CFR
Part 53) for the approval of FEMs as Class II equivalent methods when based on integrated filter
sampling and as Class III equivalent methods when based on continuous technologies that can
provide at least hourly data reporting. The performance criteria include evaluating additive bias
(intercept) and multiplicative bias (slope) as well as correlation with co-located candidate and FRM
methods at field studies covering multiple seasons and sampling locations. As a result of these new
performance criteria, EPA has recently approved two filter-based PM2.5 and two filter-based PMi0_2.s
Class II FEMs based on the virtual impactor techniques (dichotomous sampler), and five PM2 5 and
one PMio_2.5 Class III continuous FEMs based on BGT or TEOM techniques. The most recent list of
FRMs and FEMs can be found in Annex A, Table A-2 and on the following EPA web site:
http://www.epa.gov/ttn/amtic/files/ambient/criteria/reference-equivalent-methods-list.pdf
Evaluations of FRMs and FEMs were conducted both in supersite studies and in other research
studies (Ayers, 2004, 097440: Brown et al., 2006, 097665: Butler et al., 2003, 156313: Cabada et al.,
2004, 148859: Chang and Tsai, 2003, 155718: Charron et al., 2004, 053849: Grover et al., 2005,
090044: Hains et al., 2007, 091039: Hering et al., 2004, 155837: Jaques et al., 2004, 155878: Krieger
et al., 2007, 129657: Lee et al., 2005, 155925: Lee et al., 2005, 128139: Price et al., 2003, 098082:
Rees et al., 2004, 097164: Russell et al., 2004, 082453: Salminen and Karlsson, 2003, 156070:
Schwab et al., 2004, 098450: Schwab et al., 2006, 098449: Solomon et al., 2003, 156994: Tsai et al.,
2006, 098312: Vega et al., 2003, 105974: Wilson et al., 2006, 091142: Yi et al., 2004, 156169: Zhu et
al., 2007, 098367) (see Annex A, Tables A-3, A-5 and A-ll). In general, the co-located FRMs
showed very good precision with coefficient of variation (CV) <5%. For different co-located FRMs,
the regression slope of one sampler on another is commonly close to unity with R2 >0.95. The PM25
and PMio concentrations measured by dichotomous samplers were within 10% of the FRM methods,
and the differences can be attributed to the sampling artifacts of semi-volatile components; see
Section 3.4.1.2 for details. The precision of various TEOMs ranges from 10-30%. The concentration
measured by the TEOM operated at 50°C was consistently lower than those measured by the TEOM
operated at 30°C. The differences between these monitors were also found to be a function of season
December 2009 3-21
-------
and location. BGTs were highly correlated with FRMs but BGT mass could be higher than the FRM
mass (30% higher at the Fresno supersite) (Chow et al., 2008, 156355). Additionally, a number of
techniques have been developed to reduce positive and negative sampling artifacts. These are
described in the ISA for NOX and SOX - Ecological Criteria (U.S. EPA, 2008, 157074).
Several papers (Buser et al., 2007, 156310: Buser et al., 2007, 156311: Buser et al., 2007,
156312) published since the 2004 PM AQCD (U.S. EPA, 2004, 056905) claim that the EPA FRM
samplers for PM10 "oversample certain agricultural and other source emissions." These claims are
based on the erroneous assumption that the "true" PMi0 concentration is what would be given by a
PMio sampler that excluded all particles greater than 10 (im aerodynamic diameter and included all
particles less than 10 (im aerodynamic diameter. The legal definitions for PM2.5 and PMi0, as defined
in the CFR (40 CFR Part 58), include both a 50% cut-point and a penetration curve. For PMi0, the
50% cut-point of 10 ± 0.5 (im aerodynamic diameter means that 50% of particles with aerodynamic
diameter of 10 ± 0.5 (im are removed by the inlet and 50% pass through the inlet where they are
collected on the filter. The penetration curve specifies, as a function of particle size, the fraction of
particles larger than 10 (im that pass through the inlet and the fraction of particles less than 10 (im
that are intercepted by the inlet. No effort was made in the development of the FRM to have the
PMio sampler collect all particles less than 10 (im and no particles greater than 10 (im since the
sampler was designed to collect a fraction of atmospheric particles similar to the "inhalable" or
thoracic fraction, i.e., those particles that would pass through the nose and throat and reach the lungs
(Miller et al., 1979, 070577: U.S. EPA, 2004, 056905). Thus, the FRM PM10 sampler correctly and
intentionally collects particles greater than 10 (im.
For PM2.5 and PMio, it has long been known that FRMs are subject to sampling artifacts
including particle bounce on heavily-loaded impaction substrates and the loss of semi-volatile
components of PM (e.g., NH4NO3, and some organics). Although there are no standard reference
materials that provide a test of accuracy of the sampling method for airborne PM mass, in
comparison with other sampling techniques that can measure both semi-volatile and nonvolatile PM,
FRMs reported PM2.5 or PMio mass concentrations biased low by as much as 10-30% (Chow et al.,
2008, 156355). The bias of the FRMs depends on the composition of ambient PM and the sampling
conditions (e.g., ambient temperature and relative humidity), which vary from day to day and from
season to season. Another limitation of the current FRM sampling protocol is that filter samples are
typically collected every 3 or 6 days (only -150 of the 900+ PM2.5 FRMs operating in 2007 were
scheduled to sample every day). Under this operating condition, the concentration-response
relationship in air pollution health studies (especially in time-series studies) cannot be fully
evaluated in terms of lag structures and distributed lags between ambient concentration and health
outcome (Lippmann, 2009, 190083: Solomon and Hopke, 2008, 156997).
Development and Evaluation of New Techniques
Several new innovations have recently emerged to measure both fine and coarse PM fractions
in the ambient air. These techniques include the Filter Dynamics Measurement System-TEOM
(FDMS-TEOM) (Grover et al., 2006, 138080) for real-time measurement of PM2.5 or PMi0 and
several new methods for measurement of PM 10.2.5. In addition, several new techniques exist for
measuring UFPs (discussed later in Section 3.4.1.4) and for estimating PM mass concentration
indirectly using particle size (discussed later in Section 3.4.1.5).
Real-time Measurement of PM2.5 or PM10 using the FDMS-TEOM
The FDMS-TEOM incorporates self-referencing capability to the traditional TEOM by
alternating measurement of ambient air and chilled clean air (particles and semivolatile gases are
removed by filtration at 4°C after which the clean air is reheated to 30°C) in 6-min intervals. As
clean air flows over the sample filter, the semivolatile PM on the sample filter is evaporated. Thus,
the instrument provides direct measurements of the nonvolatile particle mass and incorporates an
adjustment for the semivolatile NH4NO3 and organic material. In a comparison between the TEOM,
FDMS-TEOM, and FRM mass, the PM2.5 concentration measured by the TEOM operated at 50°C
was consistently lower than those measured by the TEOM operated at 30°C. The TEOM operated at
30°C provided concentrations 50% lower than the FDMS-TEOM, and the FDMS-TEOM provided
concentrations 10-30% higher than the FRM mass (Chow et al., 2008, 156355: Schwab et al., 2006,
098449).
December 2009 3-22
-------
Techniques for Measurement ofPM10.2.s
Methods developed to measure PM10_2.5 are based on three measurement techniques: (1) virtual
impactors using low-volume, high-volume, and real-time techniques; (2) cascade impactors; and
(3) passive samplers.
A low-volume dichotomous sampler (operated at 16.7 L/min), based on virtual impaction, was
described in the 2004 PM AQCD (U.S. EPA, 2004, 056905). Since then, a high-volume dichotomous
sampler was developed to operate at 1,000 L/min (Sardar et al, 2006, 156071). The sampler was
evaluated in the field by comparison with a MOUDI sampler, and the measured PM10_2.5 mass
concentrations were within 10%. The high-volume dichotomous sampler provides sufficient mass
collection for comprehensive standard chemical analyses over short sampling intervals. Using the
virtual impactor technique in conjunction with a TEOM or BGT as the detector, continuous PMio_2.s
measurement techniques were developed (Misra et al., 2001, 018998; Misra et al., 2003, 195001;
Solomon and Sioutas, 2008, 190139). The TEOM method was highly correlated with the PMio_2.5
FRM, but mass concentrations measured by the TEOM were 20-30% lower (Solomon and Sioutas,
2008, 190139). The BGT method also agreed well with the PM10_25 FRM (slopes = 0.88-1.17, and
R2>0.95) (Solomon and Sioutas, 2008, 190139).
Case et al. (2008, 155149) evaluated a cascade PM sampler designed to collect PMi0_2.5 on a
foam impactor. Particle bouncing on impactors has long been a concern for PM collection. Porous
foam was used to serve as the impactor substrate to reduce particle bounce and to collect relatively
large amounts of particles (Demokritou et al., 2004, 186901; Demokritou et al., 2004, 190115;
Huang et al., 2005, 186991; Kuo et al., 2005, 186997). The sampler was operated at 5 L/min, and it
agreed with a low-volume dichotomous sampler within ±20%. The precision of the sampler was
20% as determined by the CV.
An inexpensive passive sampler for PMi0_25 was also developed (Leith et al., 2007, 098241;
Ott et al., 2008, 195004; Wagner and Leith, 2001, 190153; Wagner and Leith, 2001, 190154). The
passive sampler collects particles by gravity, diffusion, and convective diffusion onto a glass
coverslip, and then an image analysis is conducted on the collected particles to estimate mass flux as
a function of aerodynamic diameter. Leith et al. (2007, 098241) conducted a field evaluation of the
passive sampler, and the difference between a FRM and the co-located passive sampler was within
la of concentrations measured with PMi0_2.5 FRM samplers. Ott et al. (2008, 119394) reported the
precision of the sampler was 11.6% (CV), and the detection limit was 2.3 |j,g/m3 for a 5-day sample.
3.4.1.2. PM Speciation
The following sections describe recent developments regarding measurement techniques to
ascertain quantities of particle-bound water, cations and anions, elemental composition, carbon, and
organic species.
Particle-Bound Water
Particle-bound water is an important component of ambient PM (U.S. EPA, 2004, 056905).
Recently, a differential method was developed to measure particle-bound water (Santarpia et al.,
2004, 156944; Stanier et al., 2004, 095955). The dry ambient aerosol size spectrometer (DAASS)
can measure particle-bound water in the particle size range from 3 nm-10 (im (Stanier et al., 2004,
095955). by alternatively measuring ambient PM size distribution at low relative humidity (RH) and
ambient RH. A comparison of the two size distributions provides information on the water
absorption and change in particle size due to RH. Khlystov et al. (2005, 156635) reported that the
particle-bound water, measured by DAASS, was underestimated for particles <200 nm and
overestimated for particles >200 nm compared with thermodynamic models. The loss of
semi-volatile components during measurement may bias the particle-bound water measurement
results. Methods and analytical specifications for particle-bound water are listed in Annex A,
Table A-12.
December 2009 3-23
-------
Cations and Anions
The measurement of cations and anions including SO42~, NO3~, NH4+, Cl", Na+, and K+ still
relies primarily on filter-based collection, water based extraction and ion chromatography (1C) based
chemical speciation and quantification. In addition, denuders are frequently used in the sampling
system to adjust for sampling artifacts. These methods have been reviewed in the 2004 PM AQCD
(U.S. EPA, 2004, 056905). Filter-denuder based integrated sampling methods for SO42~, NO3~, and
NH4+ have been detailed in the 2008 SOX - Health ISA (U.S. EPA, 2008, 157075) and the 2008 NOX
and SOX - Ecological ISA (U.S. EPA, 2008, 1570741
Recent developments in multiple ion measurements have focused on the coupling of 1C and a
sample dissolution system, represented by the Particle into Liquid Sampler-Ion Chromatography
(PILS-IC) and the Ambient Ion Monitor (AIM) (Orsini et al., 2003, 156008; Weber et al., 2001,
024640). When ambient PM passes through the PILS-IC system, water droplets are generated by
mixing ambient PM with saturated water vapor and collected by impaction. The resulting liquid
stream is then introduced into the 1C system for ion speciation and quantification. Hourly
concentrations of multiple ions can be obtained with the system, with a CV of 10%. For the AIM
system, a parallel plate denuder is used to remove the interfering gases, and then particles enter a
super-saturation chamber to form droplets. The collected droplets are then introduced into the 1C for
analysis. The AIM system can provide hourly concentrations for multiple ions. The particle mass
spectrometer is another advance in multiple PM component measurements, but most of these types
of measurements are semi-quantitative and will be detailed later in Section 3.4.1.3. Note that
measurement and analytical specifications for ions other than SO42~ and NO3~ are listed in Annex A,
Table A-9.
Sulfate
Methods used for continuous (sampling interval of minutes) measurements of SO42~ include
Aerosol Mass Spectrometry (AMS) (Drewnick et al., 2003, 099160; Hogrefe et al., 2004, 156560).
PILS-IC (Weber et al., 2001, 024640). flash volatilization techniques (Bae, 2007, 155669;
Stolzenburg and Hering, 2000, 013289) and the Harvard School of Public Health (HSPH) tube
furnace to convert SO4 ~ to SO2 for detection by a SO2 analyzer (Allen et al., 2001, 156205). These
methods are described in detail by Drewnick et al. (2003, 099160). along with an inter-sampler
comparison that found overall agreement within 2.9% for all continuous instruments with R2 of 0.87
or better. When compared with filter samples, Drewnick et al. (2003, 099160) showed differences
were less than 25% for the AMS, PILS, flash vaporization, and HSPH continuous SO42~ monitors.
The Thermo 5020 particulate sulfate analyzer (based on the HSPH technique) compared within 80%
of 24-h filter-based measurement at a rural site in New York (Schwab et al., 2006, 098449). Annex
A, Tables A-8 and A-14, list detailed methods and analytical specifications for sampling SO42~.
Nitrate
In addition to the nylon filter-based method and the new developments mentioned for SO42~,
methods based on flash volatilization-chemiluminescence analysis and catalytic
conversion-chemiluminescence analysis have also been developed for continuous NO3~
measurement (averaging time 30 s-10 min). For the flash volatilization system (Fine et al., 2003,
155775; Stolzenburg and Hering, 2000, 013289; Stolzenburg et al., 2003, 156102). particles are
collected by a humidified impaction process and analyzed in place by flash vaporization and
chemiluminescent detection of the evolved NOX. For the catalytic conversion-chemiluminescence
analysis system (Weber et al., 2003, 157129). NO3~ was measured by conversion of particle NO3~
into NO, and then detected with the chemiluminescence method. Field and lab comparisons were
conducted to compare the different instruments mentioned above. Although the R&P 8400N ambient
particulate NO3~ monitor, which is based on the Stolzenburg flash vaporization technique, could
provide 10-min resolution data and showed excellent precision (with a CV <10%) (Harrison et al.,
2004, 136787; Hogrefe et al., 2004, 156560; Long and McClenny, 2006, 098214; Rattigan et al.,
2006, 115897). it consistently reported NO3~ concentrations -30% lower than the denuder-filter
systems in both the Baltimore supersite and the multiyear field study in New York (Harrison et al.,
2004, 136787; Hogrefe et al., 2004, 156560; Rattigan et al., 2006, 115897). In the New York
measurement campaign, an AMS was also co-located with other instruments to obtain the real-time
NO3~ information. AMS did not always agree well with the denuder-filter system for reasons not
December 2009 3-24
-------
entirely apparent. However, Bae et al. (2007, 156244) reported that some organic compounds can
also produce signals at mass-to-charge ratio m/z = 30, which is one of the characteristic m/z for
NO3~. Therefore, the disagreement between the AMS and the filter-based method could be a result of
the interference of organic compounds using the AMS. Annex A, Tables A-7 and A-13, list methods
and analytical specifications for sampling NO3~.
Ammonium
Several continuous and semi-continuous instruments can be used to monitor ambient
ammonium concentrations (Al-Horr et al., 2003, 153951; Bae, 2007, 155669) including many listed
above for SO42~ and NO3~. Bae et al. (2007, 155669) conducted an inter-comparison of three
semi-continuous instruments during the New York multiyear air sampling campaign: a PILS-IC, an
AMS, and a wet scrubbing-long path absorption photometer. Bae et al. (2007, 155669) reported the
inter-sampler coefficients of determination (R2) between these instruments were above 0.75, and the
slopes (with zero intercept) were between 0.71 and 1.04. Annex A, Table A-9 describes measurement
of ions other than NO3~ and SO42~, including NH4+.
Elemental Composition
Techniques for measuring the elemental composition of PM samples were reviewed in the
2004 PM AQCD (U.S. EPA, 2004, 056905). These methods include:
• Energy dispersive X-ray fluorescence (ED-XRF);
• Synchrotron X-ray fluorescence (SXRF);
• Particle-induced X-ray emission (PIXE);
• Particle elastic scattering analysis (PESA);
• Total reflection X-ray fluorescence (TR-XRF);
• Instrumental neutron activation analysis (INAA);
• Atomic absorption spectrophotometry (AAS);
• Inductively-coupled plasma-atomic emission spectroscopy (ICP-AES);
• Inductively-coupled plasma-mass spectrometry (ICP-MS); and
• Scanning electron microscopy (SEM).
Recent development in this area focused on the semi-continuous measurement methods, in
which elements were analyzed in the lab using the methods mentioned above on time-resolved
and/or size resolved samples (Kidwell and Ondov, 2004, 155898). The concentrated slurry/graphite
furnace atomic absorption spectrometry (GFAAS) method collects ambient PM as a slurry using
impactors, and then the collected PM is analyzed by AAS in the lab. Laser induced breakdown
spectroscopy (LIBS) was used to measure seven metals at the Pittsburgh supersite. LIBS
concentrates ambient PM using a virtual impactor into a sample cell, and then a Nd: YAG
laser-spectrometer is used to identify and quantify different elements. A full listing of measurement
techniques and analytical specifications for trace elements is provided in Annex A, Table A-6.
Elemental and Organic Carbon
The large variety of aspects of carbon analyses were reviewed in the 2004 PM AQCD
(U.S. EPA, 2004, 056905). Measurement and analytical specifications for carbon measurements are
listed in Annex A, Tables A-10 and A-15. Aspects of the measurements include sampling artifacts
December 2009 3-25
-------
associated with the integrated filter-based OC and EC sampling methods, the IMPROVE vs. CSN
thermal optical protocols (i.e., different thermal optical methods) and optical techniques to measure
light-absorption or BC. One significant change taking place in the CSN is that the method for carbon
measurements is being changed from the CSN method to a method designed to be consistent with
the IMPROVE carbon analysis protocol. This is a phased process that began in May of 2007 with the
conversion of 56 stations. Phase 2 of the carbon sampler conversion occurred in April of 2009 with
another 62 stations. The balance of the CSN is scheduled to be converted to IMPROVE-like
sampling and IMPROVE analysis in late 2009 (Henderson, 2005, 156537). The CSN network was
implemented to support the PM2.5 NAAQS and provides data for PM2.5 mass, SO42~, NO3~, NH4+, Na,
K, EC, OC, and select trace elements (Al through Pb) at many sites across the U.S. This conversion
will increase consistency between these two networks. Also, since the release of the 2004 PM AQCD
(U.S. EPA, 2004, 056905). more studies have been conducted to extend the understanding of
sampling artifact issues (Chow et al., 2008, 156355; Watson et al., 2005, 157125). evaluate different
thermal and optical procedures (Chen et al., 2004, 199501; Chow et al., 2004, 156347; Chow et al.,
2005, 155728; Chow et al., 2007, 156354; Conny et al., 2003, 145948; Han et al., 2007, 155823;
Subramanian et al., 2004, 081203; Watson et al., 2005, 157125). develop reference materials
(Klouda et al., 2005, 130382; Lee, 2007, 155926). create water soluble organic carbon (WSOC)
measurement techniques (Andracchio et al., 2002, 155657; Yang et al., 2003, 156167). develop
semi-continuous/continuous/real-time carbon measurement techniques (Chow et al., 2008, 156355;
Watson et al., 2005, 157125). and introduce isotope identification into the OC/EC measurement
(Huang et al., 2006, 097654).
OC sampling artifact issues were further addressed in various studies (Arhami et al., 2006,
156224; Bae et al., 2004, 156243; Chow et al., 2005, 155728; Fan et al., 2003, 058628; Fan et al.,
2004, 155770; Grover et al., 2008, 156502; Lim et al., 2003, 037037; Mader et al., 2003, 155955;
Matsumoto et al., 2003, 124293; Muller et al., 2004, 097109; Offenberg et al., 2007, 098101; Olson
and Norris, 2005, 156005; Park et al., 2006, 098104; Rice, 2004, 156049; Subramanian et al., 2004,
081203; ten Brink et al., 2004, 097110; ten Brink et al., 2005, 156115; Viana et al., 2006, 179987).
and were well summarized by Watson et al. (2005, 157125) and Chow et al. (2008, 156355). There
are two commonly used methods to correct OC sampling artifacts: the filter with backup filter
system (TBQ: placing a backup quartz-fiber filter behind the front Teflon-membrane filter; QBQ:
placing a backup quartz-fiber filter behind the front quartz-fiber filter); and the
denuder-filter-adsorbent system. Subramanian et al. (2004, 081203) and Chow et al. (2006, 099031)
reported that during the Pittsburgh and Fresno supersite studies the positive artifact (organic gases
condensed on filters) from TBQ (24-34%, up to 4 (ig/m3 OC) was nearly twice that from QBQ
(13-17%). With the denuder-filter-adsorbent system, the negative artifact (OC evaporating from the
filter) was 5-10%. Watson and Chow (2002, 037873) reported that the XAD-coated denuder could
function as efficiently as a parallel plate denuder using carbon-impregnated charcoal filters (GIF)
with frequent denuder changes. Huebert and Charlson (2000, 156577) reported that using tandem
filter packs may hinder a quantitative analysis of the artifacts.
Different temperature protocols and optical correction methods in thermal-optical analyses
were further evaluated by Watson et al. (2005, 157125). Chow et al. (2004, 156347; 2005, 155728;
2007, 156354). Subramanian et al. (2006, 156107). Conny et al. (2003, 145948). Han et al. (2007,
155823). Chen et al. (2004, 199501) and (Conny et al., 2009, 191999). Solomon et al. (2003,
156994) reported a 20-50% difference for OC and a 20-200% difference for EC using 11 filter
samples and 4 different analytical protocols. In an assessment of the different thermal-optical
analysis protocols used around the world, Watson et al. (2005, 157125) reported that differences of a
factor of 2 to 7 in EC between different methods could be observed, and a factor of 2 was common,
while the relative differences in OC between different methods were small. As Watson et al. (2005,
157125) stated, there are 12 major differences among the thermal methods: (1) analysis atmosphere;
(2) temperature ramping rates; (3) temperature plateaus; (4) residence time at each plateau; (5)
optical pyrolysis monitoring configuration and wavelength; (6) standardization; (7) oxidation and
reduction catalysts; (8) sample aliquot and size; (9) evolved carbon detection method; (10) carrier
gas flow through or across the sample; (11) location of the temperature monitor relative to the
sample; and (12) oven flushing conditions. Chow et al. (2004, 156347) and Chen et al. (2004,
199501) addressed the difference between optical transmission and optical reflectance methods for
charring correction, and they reported that the charring OC on the surface of or inside a filter
dominated the differences between these two correction methods. The differences between different
sampling and measurement methods are also applied to the in-situ/semi-continuous methods, since
December 2009 3-26
-------
most of these methods are also based on thermal-optical analysis of collected filters. Most of these
methods agree with integrated filter methods within 30%.
The differences observed between methods for OC and EC come largely from how OC and
EC are defined. They are defined on an operational basis, as there are no standard reference
materials. Initial efforts have been made to produce OC/EC reference materials at the National
Institute of Science and Technology (NIST) (Klouda et al., 2005, 130382: Lee, 2007, 155926).
Klouda et al. (2005, 130382) described the development of Reference Material 8785: Air Particulate
Matter on Filter Media. Each reference filter is uniquely identified by its air PM number and its
gravimetrically determined mass of fine Standard Reference Material (SRM) 1649a, and each filter
has values assigned for total carbon, EC, and organic carbon mass fractions measured according to
both IMPROVE and NIOSH protocols. Lee et al. (2007, 155926) reported a method to create a
reference filter with a known amount of OC (as potassium hydrogen phthalate), and EC (as carbon
black hydrosol).
Measurement methods for WSOC have been developed recently (Miyazaki et al., 2006,
156767: Sullivan and Weber, 2006, 157031: Sullivan et al., 2004, 157029: Sullivan et al., 2006,
157030: Sullivan et al., 2007, 100083: Yu et al., 2004, 156172). WSOC can be measured on
integrated filter samples, or in-situ measurement can be conducted by coupling with the PILS-IC
(Sullivan et al., 2004, 157029). For integrated filter samples, filters are extracted with deionized
water and followed by oxidation of total WSOC to CO2. CO2 can then be detected by either infrared
spectroscopy (IR) (Decesari et al., 2000, 155748: Kiss et al., 2002, 156646: Yang et al., 2003,
156167). FID (Yang et al., 2003, 156167). or pyrolysis gas chromatography/mass spectrometry
(GC/MS) (Gelencser et al., 2000, 155785). A correlation coefficient of 0.84 was reported by Sullivan
et al. (Sullivan et al., 2004, 157029) between in-situ and filter based measurement of WSOC.
Further development and evaluation has been conducted on the measurement of BC with light
absorption instruments (Andreae and Gelencser, 2006, 156215: Amort et al., 2003, 037711: Bae et
al., 2004, 156243: Borak et al., 2003, 156284: Cyrys et al., 2003, 049634: Kurniawan and Schmidt-
Ott, 2006, 098823: Park et al., 2006, 098104: Saathoff et al., 2003, 156066: Sadezky et al., 2005,
097499: Slowik et al., 2007, 096177: Taha et al., 2007, 096277: Virkkula et al., 2007, 157098:
Wallace, 2000, 000803: Weingartner et al., 2003, 156149: Williams et al., 2006, 157148: Wu et al.,
2005, 157155). These instruments include the aethalometer, particle absorption photometer, and
photoacoustic analyzer. However, these instruments are subject to interferences by particle
scattering, interactions with the filter substrate, particle loading on filters, and other pollutants (e.g.,
NO2). Uncertainties of up to 50% were observed in the studies mentioned above by comparing these
methods with integrated filter methods and thermal analysis methods.
Huang et al. (2006, 097654) reported the measurement of a stable isotope, 13C, in OC and EC
with a thermal optical transmission analyzer coupled with gas chromatography-isotope ratio mass
spectrometer (TOT-GC-IRMS). The ratio of 13C/ 2C in OC and EC can provide useful information on
OC/EC source categories and origin. The method was applied to Pacific2001 aerosol samples from
the greater Vancouver area in Canada and produced a precision of-0.03%. Gustafsson et al. (2009,
192000) applied the radiocarbon measurement technique and quantified the source contributions of
carbonaceous aerosols to the Indian Ocean "brown cloud," with particular relevance for
understanding and mitigating the climate effects of EC/BC.
Organic Speciation
Organic matter makes up a substantial fraction of PM in all regions of the U.S. (U.S. EPA,
2004, 056905). and 10-40% of the total organic matter is currently quantifiable at the individual
compound level (Poschl, 2005, 156882). Recent advancements in traditional solvent extraction
GC/MS and high pressure liquid chromatography (HPLC) as well as application of thermal
desorption (TD) techniques are helping to expand the understanding of the composition of organic
matter as well as improving detection limits for quantification of organic molecular marker (OMM)
compounds (Robinson et al., 2006, 156918: Schnelle-Kreis et al., 2005, 112944: Sheesley et al.,
2007, 112017: Shrivastava et al., 2007, 111594). In addition, information about organic functional
groups can be obtained with Fourier transform infrared spectrometry (FTIR) (Tsai and Kuo, 2006,
156127).
Recent advancements in GC/MS technology including inert electron ionization sources and
improved instrument sensitivity and scan rates for better OMM quantification, have increased its
December 2009 3-27
-------
application in organic aerosol characterization studies (Cass, 1998, 155716; Dutton et al., 2009,
194887; Fraser et al., 2003, 042231; Graham et al., 2003, 156489; Hays et al., 2002, 026104;
Robinson et al., 2006, 156918; Schauer et al., 1996, 051162; Sheesley et al., 2007, 112017;
Subramanian et al., 2006, 156107; Watson et al., 1998, 012257; Zheng et al., 2002, 026100; Zheng et
al., 2006, 157189). Incorporation of high volume injection using programmable temperature
vaporization (PTV) (Engewald et al., 1999, 155765) has further lowered detection limits for trace
level OMM compounds. High volume injection has the added benefit of preventing the loss of
semivolatile compounds (Swartz et al., 2003, 157035). and has been applied for analysis of PAHs
using low volume samplers (down to 5 L/min), allowing for smaller required mass loadings (Bruno
et al., 2007, 155706; Crimmins and Baker, 2006, 097008). Since last review, HPLC analysis with
fluorescence detection has also been used frequently for quantification of semivolatile organic
compounds in both the particle and gas phase (Albinet et al., 2007, 154426; Barreto et al., 2007,
155676; Chow, 2007, 157209; Eiguren-Fernandez et al., 2003, 142609; Goriaux et al., 2006, 156484;
Murahashi, 2003, 096539; Ryno et al., 2006, 156065; Stracquadanio et al., 2005, 156104; Temime-
Roussel et al., 2004, 098530; Temime-Roussel et al., 2004, 098521). Lengthy extraction and analysis
times remain a limiting factor for these methods.
TD techniques bypass one of the time consuming steps in traditional solvent extraction
analysis for nonpolar organic compounds (n-alkanes, branched alkanes, cyclohexanes, hopanes,
steranes, alkenes, phthalates and PAHs). This is achieved by vaporizing and analyzing organic
constituents directly from the collection substrate, thereby bypassing the extraction step (Chow,
2007, 157209). Methods exist for both off-line TD analysis of previously collected filter samples and
semi-continuous TD analysis. Annex A, Table A-17 is adapted from Chow et al. (2007, 157209) and
summarizes recent TD-GC/MS studies. The most common off-line method is TD-GC/MS (Hays and
Lavrich, 2007, 155831). Continuous or semi-continuous methods have been developed for direct
analysis of individual organic constituents by coupling TD with various forms of mass spectrometry
(Smith et al., 2004, 156090; Tobias and Ziemann, 1999, 157053; Tobias et al., 2000, 156121; Voisin
et al., 2003, 156141; Williams et al., 2006, 156157). A comparison of measurement and analytical
specifications for filter analysis using solvent extraction and TD methods for organic speciation are
summarized in Annex A, Table A-17.
3.4.1.3. Multiple-Component Measurements on Individual Particles
The 2004 PM AQCD (U.S. EPA, 2004, 056905) discussed the aerosol time-of-flight mass
spectrometry (ATOFMS). Recently, the ATOFMS and several other aerosol mass spectrometry
methods have been further developed. Both lab and field comparisons have been conducted to
evaluate the reliability of these types of instruments.
There are four types of commonly used aerosol mass spectrometry: (1) particle analysis by
laser MS (PALMS; National Oceanic and Atmospheric Administration [NOAA]); (2) rapid single
particle mass spectrometer (RSMS; University of Delaware); (3) aerosol time-of-flight MS
(ATOFMS; TSI, Inc.); and (4) AMS (Aerodyne) (Chow et al., 2008, 156355; Nash et al., 2006,
199502). The differences between these instruments primarily come from the particle sizing methods
of mass spectrometers, as shown in Annex A, Table A-16. Although the technique varies, the
underlying principle is to fragment each particle into ions, using either a high-power laser or a heated
surface, and then a mass spectrometer to measure the mass to charge ratio of each ion fragment in a
vacuum.
These instruments were evaluated at the Atlanta, Houston, Fresno, Pittsburgh, New York, and
Baltimore supersites (Bein et al., 2005, 156265; Drewnick et al., 2004, 155755; Drewnick et al.,
2004, 155754; Hogrefe et al., 2004, 156560; Jimenez et al., 2003, 156611; Lake et al., 2003, 156669;
Lake et al., 2004, 088411; Middlebrook et al., 2003, 042932; Phares et al., 2003, 156866; Qin and
Prather, 2006, 156895; Wenzel et al., 2003, 157139). Measurements of the gross composition and
abundance of particles by these instruments were generally semi-quantitative, with the exception of
AMS. Particles of similar composition (e.g., OC/SO42~, Na/K/SO4 ~, soot/hydrocarbon, and mineral
particle types) were characterized by these instruments during the studies mentioned above. NO3~
and SO4 ~ concentrations measured with AMS were comparable with other continuous and
filter-based methods, as mentioned in Section 3.4.1.2. In addition, concentrations of different particle
types can be obtained by the co-location of these aerosol mass spectrometers and other particle
sizing instruments, such as particle counters or the Micro-Orifice Uniform Deposit Impactor
(MOUDI).
December 2009 3-28
-------
3.4.1.4. UFPs: Mass, Surface Area, and Number
Instruments for measuring UPFs developed during the past decade permit measurement of size
distributions of particles down to 3 nm in diameter with mobility particle sizers. Concentrations
down to this size range can be obtained by a MOUDI. The recently developed low pressure-drop
UFP impactor coupled with a (3 Attenuation Monitor (nano-BAM) can also provide UFP (<150 nm)
mass concentrations (Chakrabarti et al, 2004, 157426). A high correlation coefficient was observed
between MOUDIs and nano-BAMs, with a correlation of 0.96. A 50% cut point (d50) of 13-200 nm
can be achieved by a high-volume slot-type UFP virtual impactor (Middha and Wexler, 2006,
155982).
Methods are also being developed to measure the surface area of UFPs. Particle surface area is
usually measured by attaching labeled (radioactive or electrical labeling) molecules to particles and
detecting the radioactive or electrical properties of the attached molecules. Wilson et al. (2007,
098398) suggested that the electrical aerosol detector (BAD, based on diffusion charging)
measurement might be a useful indicator of the particle surface area deposited in the lung. This
method can be potentially useful for examining the association between health effects and particle
surface areas.
Developments involving the condensation particle counter include use of de-ionized water as a
condensation medium in lieu of butanol or n-propanol in condensation particle counters (Hering et
al., 2005, 155838: Hermann et al., 2007, 155840: Petaja et al., 2006, 156021). This development
makes the condensation particle counter (CPC) easier to use in field studies because water does not
have some of the same chemical properties (with respect to hazard and odor) as butanol or
n-propanol. The performance of this CPC was reported to be similar to the conventional butanol
based CPC (Hering et al., 2005, 155838). Use of a battery of water and butanol-based CPCs was
demonstrated to detect a range of solubilities in nucleation-mode particles (Kulmala et al., 2007,
155911). Additionally, CPCs have been used to measure particles in the smaller end of the UF scale
through adjustment of CPC cut-off diameters through tuning the temperature difference between the
CPC saturator and condenser (Kulmala et al., 2007, 155911) and improved charge reduction
techniques (Winkler et al., 2008, 156160). The latter method was effective in reducing the size of
particles detected by a CPC to <2 nm. These studies include assessment of errors related to these
developments with the CPC and generally show that counting efficiencies with these devices is
upwards of 95% (Hermann et al., 2007, 155840). Additionally, recent advancements have been made
in development of fast scanning methods for UFP size distributions, including diffusion screens (DS)
(Feldpausch et al., 2006, 155773) and fast integrated mobility scanners (FIMS) (Olfert et al., 2008,
156004).
3.4.1.5. PM Size Distribution
Along with particle density and shape (U.S. EPA, 2004, 056905). the particle size distribution
can be used to estimate PM mass concentrations. For particles >0.1 (im, several instruments,
including DRUM, MOUDIs, and aerodynamic particle sizer (APS), are available to measure
mass-based or count-based particle size distribution. An APS incorporating very sharp cut points
between 0.1 and 10 urn is now available (Peters, 2006, 156860: Zeng, 2006, 098375). For particles
in this range, inertial forces are used to separate particles based on impaction. For particles <0.1 (im,
particles can be separated by their electrical mobility, and as a result, electrical mobility diameter is
often used to describe UFP size distribution in lieu of aerodynamic diameter. It has been necessary to
develop techniques to convert mobility diameters, measured by the scanning mobility particle sizer
(SMPS) or the Engine Exhaust Particle Sizer (EEPS), to aerodynamic diameters, measured by the
APS, or vice versa, in order to merge the distributions spanning the UF, accumulation, and coarse
modes. A variety of techniques for combining SMPS and APS diameters have been reported in the
literature (Hand and Kreidenweis, 2002, 155824: Khlystov et al., 2004, 155897: Morawska et al.,
1999, 007609: Morawska et al., 2007, 155990: Shen et al., 2002, 156086: TSI, 2005, 157196).
However, each of these techniques incurs some uncertainty of which the user must be aware.
3.4.1.6. Satellite Measurement
Instruments sensing back scattered solar radiation on satellites have made it possible to derive
information about tropospheric aerosol properties on the global scale. The satellite borne instruments
December 2009 3-29
-------
vary in their complexity and in the aerosol properties they can measure. Satellite instruments
measure radiance (or brightness temperature) that can then be used to provide information on the
aerosol column amount, or the aerosol optical depth (AOD). Depending on the wavelengths sampled
and the spectral resolution of the instruments, information about the composition of particles of
diameter <2 (im and particles of diameter >2\am can be obtained. Data from two main instruments,
the moderate resolution imaging spectroradiometer (MODIS) and the multiangle imaging
spectroradiometer (MISR) have been used to estimate surface PM in the U.S. MODIS measures the
intensity of back scattered sunlight at seven wavelengths through the visible to the near infrared at
one viewing direction; and MISR measures the intensity at four wavelengths (from the visible to the
near IR) and the same ground pixel at nine viewing angles. The spatial resolution of reported AOD is
17.6x17.6 km for MISR and either 10^10 or 1x1 km for MODIS, depending on retrieval algorithm.
Since both instruments are located on the same satellite, their times of overpass are the same, about
1330 local time. Due to precession of the satellite's orbit, the satellite does not pass over the same
path every day, and instruments cannot sense aerosol properties beneath cloud tops.
The problem of using satellite data to retrieve properties of the atmospheric aerosol is complex
because the surface contribution to satellite measured reflectance must be separated from the aerosol
signal. Difficulties can arise when attempting to derive aerosol information over land surfaces
because of uncertainties in surface reflectivity, similarities between aerosol and surface composition,
and high signal-to-noise ratio when viewing AOD over reflective surfaces such as desert and snow.
To overcome this difficulty, data from MODIS have been applied over dark land surfaces and
ongoing improvements in retrieval algorithms are being developed. Instruments such as the MISR
that sense at multiple viewing angles can better cope with the problems over land surfaces because
they can use the information on the angular dependence of reflection from the surface and the
atmosphere to distinguish between their signals. Not only can total AODs be derived, but fractional
AODs that reflect external mixtures characterized by particle shape, effective radius, and single
scattering albedo can also be derived. These properties can then be used to infer particle
composition. Retrievals over the oceans have had less difficulty because the optical properties of
sunlight reflected from the sea surface are much better known, and reflectivities are low over most
zenith angles at less than gazing incidence.
Kokhanovsky et al. (2007, 190009) examined the errors associated with MODIS, MISR, and a
number of other satellite instruments with respect to associated retrieval algorithms for retrievals
over Central Europe. They found a correlation coefficient between MODIS and MISR AOD of 0.62.
Both MODIS and MISR AOD tended to underestimate ground-based AOD measurements from
AERONET (NASA's AErosol RObotic NETwork) slightly with MODIS generally retrieving higher
AODs than MISR. Chu et al. (2003, 190049) found correlations between MODIS and AERONET
AODs ranging from 0.82-0.91; and Kahn et al. (2005, 189961) found correlations of 0.7-0.9 between
MISR and AERONET AODs.
Further complexity is added when attempting to relate surface PM2 5 to aerosol optical depths.
The detailed comparisons of surface measurement and satellite measurements are given in Chapter 9.
3.4.2. Ambient Network Design
3.4.2.1. Monitor Siting Requirements
The EPA Air Quality System database (AQS) contains measurements of air pollutant
concentrations in the 50 states, plus the District of Columbia, Puerto Rico, and the Virgin Islands, for
the 6 criteria air pollutants as well as a more limited dataset of hazardous air pollutants. In 2007,
there were 4,693 PMi0 monitors and 2,194 PM2.5 monitors reporting values to the AQS. Where
SLAMS PMio and PM25 monitoring is required, at least one of the sites must be a maximum
concentration site for that specific area. The appropriate spatial scales for PM2 5, PMi0_2.5, and PMi0
monitoring differ given the contrasting spatial gradients of coarse PM relative to fine. The relevant
scales for each size classification are provided in Annex A, Table A-18.
Criteria for siting ambient monitors for PM at national monitoring networks are summarized
below by PM size, and details are given in the CFR 40 Part 58 Appendix D, and
SLAMS/NAMS/PAMS Network Review Guidance (U.S. EPA, 1998, 093211). Table A-19 in Annex
A provides a summary of the number of sites and operating specifications of these networks. Probing
December 2009 3-30
-------
and monitoring path siting criteria for any specific monitoring site are given in CFR 40 Part 58
Appendix E, including horizontal and vertical placement, spacing from minor source, spacing from
obstructions, spacing from trees, and spacing from roadways.
PM2.5
The minimum number of PM2.5 monitors required in a metropolitan statistical area is
determined by the population and the air quality in the area, as specified in Appendix D of 40 CFR
Part 58. The required minimum number of PM25 monitors ranges from 0 to 3 in any given
metropolitan statistical area. Continuous PM2.5 monitors must be operated in no fewer than one-half
of the minimum required sites in each area. Most PM25 monitoring in urban areas should be
representative of a neighborhood scale (for trends and compliance with standards). Urban or regional
scale sites are located to characterize regional transport of PM25. In certain instances where
population-oriented micro- or middle-scale PM2 5 monitoring are determined by the Regional
Administrator to represent many such locations throughout a metropolitan area, these smaller scales
can be considered to represent community-wide air quality. PM2 5 measurements are obtained at local
temperature and pressure across the NAMS/SLAMS networks (40 CFR Part 58).
PM25 chemical speciation monitoring is currently conducted at 197 CSN sites
(http://www.epa.gov/ttn/amtic/specgen.html). Within the CSN network, 53 locations are recognized
as the Speciation Trends Network (STN) operating on a sample schedule of one in every three days,
while the rest of the CSN typically operates every sixth day.
PMlO-2.5
PM10_2.5 monitoring has not been required at SLAMS sites, but will be required at NCore1
Stations (which is a sub-set of the SLAMS) by January 1, 2011. Middle and neighborhood scale
measurements are the most important station classifications for PMi0_2.5 to assess the variation in
coarse particle concentrations that would be expected across populated areas that are in proximity to
large emissions sources. PMi0_2.5 chemical speciation monitoring and analyses will also be required
at NCore sites by January 1, 2011. EPA has already approved FRMs and FEMs for PMi0_2.5 mass;
however, methods for PM10_2.5 speciation are still being developed (Henderson, 2009, 192001). PM10_
2.5 measurements are obtained at local temperature and pressure by recalculating the co-located PMi0
for local conditions.
PM
10
As for PM25, the minimum number of PMi0 monitors required in a metropolitan statistical area
is determined by the population and the air quality in the area, as specified in Appendix D of 40 CFR
Part 58. The required minimum number of PMi0 monitors ranges from 0 to 8 in any given
metropolitan statistical area. Except for some circumstances where microscale (<100 m, for
maximum PMi0 exposure) monitoring may be appropriate, the most important scales to characterize
the emissions of PMi0 effectively from both mobile and stationary sources are the middle scale (for
short-term public exposure) and neighborhood scale (for trends and compliance with standards).
PMio measurements are obtained at standard temperature and pressure across the NAMS/SLAMS
networks (40 CFR Part 58).
3.4.2.2. Spatial and Temporal Coverage
Locations of PM2.5 and PM10 Monitors in Selected Metropolitan Areas in the U.S.
Fifteen metropolitan regions were chosen for closer investigation of monitor siting based on
their distribution across the nation and relevance to health studies analyzed in subsequent chapters of
this ISA. These regions were: Atlanta, Birmingham, Boston, Chicago, Denver, Detroit, Houston, Los
Angeles, New York City, Philadelphia, Phoenix, Pittsburgh, Riverside, Seattle, and St. Louis. Core-
1 For more information on NCore, see the NCore web site at: http://www.epa.gov/ttn/amtic/ncore/index.html.
December 2009 3-31
-------
Based Statistical Areas (CBSAs) and Combined Statistical Areas (CSAs), as defined by the U.S.
Census Bureau (http://www.census.gov/), were used to determine which counties, and hence which
monitors, to include for each metropolitan region.1 Figure 3-7 and Figure 3-8 display PM2.s and PMi0
monitor density, respectively, with respect to population density in Boston. Annex A includes similar
information for all fifteen metropolitan regions (Figure A-l through Figure A-30).
1 A CBSA represents a county-based region surrounding an urban center of at least 10,000 people determined using 2000 census data and
replaces the older Metropolitan Statistical Area (MSA) definition from 1990. The CSA represents an aggregate of adjacent CBSAs tied
by specific commuting behaviors. The broader CSA definition was used when selecting monitors for the cities listed above with the
exception of Los Angeles, Riverside and Phoenix. Los Angeles and Riverside are contained within the same CSA, so the smaller CBSA
definition was used to delineate these two cities. Phoenix is not contained within a CSA, so the smaller CBSA definition was used for
this city as well.
December 2009 3-32
-------
Boston Combined Statistical Area
i Kilometers
0 5 10 20 30 40 50
ol\
r
2005 Population Density
[ | Boston PM2.5 Monitors (15km buffer)
Population per Sq Km
^B 0-251
^H 252 - 502
503 - 2508
2509-5016
^B 5017-12539
^H 12540-50155
: Kilometers
0 15 30 60 90 120 150
Figure 3-7. PM2.s monitor distribution in comparison with population density, Boston CSA.
December 2009
3-33
-------
Boston Combined Statistical Area
i Kilometers
0 5 10 20 30 40 50
ol\
r
2005 Population Density
[ | Boston PMio Monitors (15 km buffer)
Population per Sq Km
^B 0-251
^H 252 - 502
503 - 2508
2509-5016
^B 5017-12539
^H 12540-50155
: Kilometers
0 15 30 60 90 120 150
Figure 3-8. PM10 monitor distribution in comparison with population density, Boston CSA.
December 2009
3-34
-------
Table 3-3.
Proximity to PM^and PMio monitors for total population3 by city.
Proximity to PM Monitors'5
Region
Total CSA/CBSA
N
<1km
N %
<5km
N %
<10km
N %
<15km
N %
PROXIMITY TO PM2.5 MONITORS
Atlanta
Birmingham
Boston
Chicago
Denver
Detroit
Houston
Los Angeles
New York
Philadelphia
Phoenix
Pittsburgh
Riverside
Seattle
St. Louis
5,316,742
1,166,100
7,502,707
9,754,262
2,952,039
5,553,465
5,503,320
13,061,361
22,050,940
6,388,913
3,818,147
2,515,383
3,781,063
3,962,434
2,869,955
23,461 0.44
12,925 1.11
185,457 2.47
177,076 1.82
40,601 1.38
54,997 0.99
11,586 0.21
115,477 0.88
717,094 3.25
117,389 1.84
37,133 0.97
40,574 1.61
43,739 1.16
13,723 0.35
37,329 1.30
581,461 10.94
240,383 20.61
1,877,180 25.02
3,091,573 31.69
649,953 22.02
1,174,733 21.15
213,708 3.88
2,579,809 19.75
8,107,764 36.77
1,878,373 29.40
490,072 12.84
587,148 23.34
723,829 19.14
287,373 7.25
563,176 19.62
1,990,477 37.44
666,926 57.19
3,356,019 44.73
6,473,463 66.37
1,548,976 52.47
2,791,555 50.27
905,007 16.44
7,544,466 57.76
13,493,867 61.19
3,517,321 55.05
1,099,069 28.79
1,331,230 52.92
1,855,296 49.07
931,630 23.51
1,338,349 46.63
3,179,844 59.81
848,447 72.76
4,641,175 61.86
8,185,010 83.91
2,252,657 76.31
3,845,190 69.24
1,599,079 29.06
10,792,727 82.63
16,571,764 75.15
4,393,136 68.76
1,739,542 45.56
1,883,301 74.87
2,344,394 62.00
1,561,792 39.41
1,760,985 61.36
PROXIMITY TO PM10 MONITORS
Atlanta
Birmingham
Boston
Chicago
Denver
Detroit
Houston
Los Angeles
New York
Philadelphia
Phoenix
Pittsburgh
Riverside
Seattle
St. Louis
5,316,742
1,166,100
7,502,707
9,754,262
2,952,039
5,553,465
5,503,320
13,061,361
22,050,940
6,388,913
3,818,147
2,515,383
3,781,063
3,962,434
2,869,955
30,973 0.58
23,943 2.05
63,614 0.85
55,642 0.57
38,449 1.30
14,050 0.25
36,795 0.67
52,052 0.40
19,842 0.09
23,988 0.38
99,520 2.61
65,906 2.62
61,356 1.62
4,851 0.12
27,872 0.97
416,440 7.83
251,310 21.55
1,090,172 14.53
844,714 8.66
521,201 17.66
309,623 5.58
832,767 15.13
1,404,389 10.75
292,105 1.32
376,966 5.90
1,255,430 32.88
706,413 28.08
895,615 23.69
220,539 5.57
380,411 13.25
1,090,497 20.51
473,054 40.57
2,087,770 27.83
2,374,972 24.35
1,146,286 38.83
748,971 13.49
2,227,314 40.47
4,899,254 37.51
592,631 2.69
1,091,532 17.08
2,615,738 68.51
1,291,700 51.35
2,360,272 62.42
709,887 17.92
891,695 31.07
1,837,983 34.57
638,472 54.75
2,939,870 39.18
3,844,297 39.41
1,799,187 60.95
1,300,995 23.43
3,141,150 57.08
9,075,863 69.49
773,962 3.51
2,238,309 35.03
3,416,682 89.49
1,705,451 67.80
2,922,799 77.30
1,211,430 30.57
1,212,543 42.25
3Based on 2005 population totals.
Percentages are given with respect to the total population per city provided.
December 2009
3-35
-------
Table 3-3 shows the population density around PM2.5 and PM10 monitors for the total
population for each CS A/CBS A individually. Population totals within various distances of PM
monitors were calculated assuming equal internal population distribution for individual census
blocks. Between-city disparities in population density were large and were dependent primarily on
the location and number of PM monitoring sites per CSA/CBSA. For PM2.5, Los Angeles (83%) and
Denver (76%) had the largest proportion of the total population within 15 km of a monitor. Houston
(29%) had the least population coverage with their PM2.5 monitors. For PM10, Phoenix (89%) had the
largest proportion of the total population within 15 km of a monitor. Detroit (23%), Boston (39%),
Seattle (31%), and Philadelphia (35%) had the smallest proportions of the population within 15 km
of a PMio monitor. Proximity to monitoring stations is considered further in Section 3.5 and Section
3.8 regarding spatial variability within cities. Figure 3-7 shows that the PM2.5 network more closely
samples near population centers in the Boston CSA compared with the PMi0 network shown in
Figure 3-8, although both PM2 5 and PM10 networks place at least one monitor in the city center.
3.4.2.3. Network Application for Exposure Assessment with Respect to Susceptible
Populations
Subject Age
Table 3-4 breaks down the population density around PM2 5 and PMi0 monitors for
sub-populations of children age 0-4 yr, children age 5-17 yr, and elderly adults age 65 yr and over
cumulatively for the 15 CS As/CBS As examined. Table 3-5 shows the distribution for adults age 65
years and over for each CSA/CBSA individually. This detail of information is not provided for the 0-
to 4-yr and 5- to 17-yr age groups because variation in percentage within a certain radius of the
monitor was generally fairly low for each city across the child age groups when compared to total
population. In the cases of Denver, Detroit, Phoenix, Riverside, and St. Louis for PM25 and
Birmingham, Denver, Riverside, and St. Louis for PMi0, the elderly population's distribution around
the samplers varied more from the total population compared to other age groups. When all
CS As/CBS As were considered cumulatively, the percentage of the population within 15 km of a
monitor was similar for all age groups for both PM2 5 and PMi0. Between-city disparities in elderly
population density within a sampler radius were larger. For PM2 5, Chicago (87%) and Denver (84%)
had the largest proportion of the elderly population within 15 km of a monitor. Houston (31%) had
the least population coverage with their PM2 5 monitors. For PMi0, Phoenix (90%) had the largest
proportion of the total population within 15 km of a monitor. New York (4%), Detroit (27%), Seattle
(32%), and Philadelphia (39%) had the smallest proportions of the population within 15 km of a
PMio monitor. These differences may reflect overall density of the samplers within a given city, with
PM25 monitors more numerous than PM10 monitors in most of the CS As/CBS As, and retirement and
settlement trends among elderly adults.
December 2009 3-36
-------
Table 3-4. Proximity to PM2.6 and PMio monitors for children age 0-4 yr, children age 5-17 yr, and
adults age 65 yrand older.3 The figures presented here are cumulative for the 15
CSAs/CBSAs examined in Chapter 3.
Age
Grouping
Proximity to PM Monitors'5
Total CSA/CBSA
N
<1
N
km
%
<5km
N
%
<10km
N
%
<15km
N
%
PROXIMITY TO PM2.s MONITORS
0-4
5-17
>65
6,400,785
17,212,825
10,391,023
109,466
275,427
175,113
1.71
1.60
1.69
1,603,000
4,164,132
2,570,909
25.04
24.19
24.74
3,361,922
8,814,179
5,483,776
52.52
51.21
52.77
4,462,403
11,813,997
7,288,049
69.72
68.63
70.14
PROXIMITY TO PM10 MONITORS
0-4
5-17
>65
6,400,785
17,212,825
10,391,023
44,384
110,882
68,367
0.69
0.64
0.66
695,120
1,756,246
1,056,375
10.86
10.20
10.17
1,725,419
4,441,239
2,631,243
26.96
25.80
25.32
2,636,782
6,942,001
4,041,802
41.19
40.33
38.90
Based on 2000 population totals.
Percentages are given with respect to the total population per city provided.
Race and Hispanic Origin
Table 3-6 shows the percent of the population self-identified as white or black and having a
Hispanic or non-Hispanic origin within 1,5, 10, or 15 km distances from PM25 and PMio monitors
cumulatively across the fifteen CSAs/CBSAs. For PM2.5, blacks and Hispanics had similar
percentages of the population within 15 km of a monitor (86% and 82%, respectively), while a
smaller proportion of whites and non-Hispanics were within that same distance (63% and 67%,
respectively), across the fifteen CSAs/CBSAs studied. For PM^ Hispanics (54%) represented the
subpopulation with the largest percentage of total population within 15 km of a monitor across the
fifteen CSAs/CBSAs studied. The percentage of blacks within that same distance was marginally
lower (48%), whereas the percentage of whites and non-Hispanics within 15 km of a monitor was
approximately two-thirds that of Hispanics (35% and 37%, respectively). Higher percentages of
individual ethnic subpopulations within 15 km of a PM2.5 monitor most likely represents the fact that
more PM2 5 monitors are currently deployed compared with PMio monitors. While no ethnic
subpopulation appears to be well represented at the neighborhood scale, greater percentages of the
black and Hispanic populations (1% each) are within 1 km of a PMio monitor than the corresponding
white and non-Hispanic populations (0.5% each). Likewise, 2.5% of the black population and 2.8%
of the Hispanic population reside within 1 km of a PM25 monitor compared to 1.4% of the white
population and 1.5% of the non-Hispanic population. Furthermore, it is notable that at any scale
shown in Table 3-6 for both PM25 and PMio monitors, those self-identified as black or Hispanic
actually have greater representation by the monitors than those identified as white or non-Hispanic.
December 2009
3-37
-------
Table 3-5.
Proximity to PM2.eand PMio monitors for adults age 65 yr and older3 by city.
Proximity to PM Monitors'5
Age Grouping
Total CSA/CBSA
N
<1km
N
%
<5km
N
%
<10km
N
%
<15
N
km
%
PROXIMITY TO PM2.5 MONITORS
Atlanta
Birmingham
Boston
Chicago
Denver
Detroit
Houston
Los Angeles
New York
Philadelphia
Phoenix
Pittsburgh
Riverside
Seattle
St. Louis
362,201
145,905
945,790
1,018,983
232,974
626,216
377,586
1,207,436
2,710,675
834,110
388,150
449,544
342,334
390,372
358,747
1,757
1,619
18,821
18,539
3,891
5,765
1,010
9,653
78,918
13,323
2,738
8,933
3,024
1,721
5,401
0.49
1.11
1.99
1.82
1.67
0.92
0.27
0.80
2.91
1.60
0.71
1.99
0.88
0.44
1.51
36,772
29,952
224,628
348,656
59,625
138,672
14,911
229,893
921 ,599
251 ,459
39,833
111,050
50,901
29,429
83,528
10.15
20.53
23.75
34.22
25.59
22.14
3.95
19.04
34.00
30.15
10.26
24.70
14.87
7.54
23.28
136,179
84,223
438,920
713,194
140,523
345,808
66,741
688,844
1,619,177
487,003
90,304
249,269
129,836
101,223
192,532
37.60
57.72
46.41
69.99
60.32
55.22
17.68
57.05
59.73
58.39
23.27
55.45
37.93
25.93
53.67
207,122
106,488
606,231
883,112
196,361
469,462
117,661
984,889
2,048,842
605,663
142,084
347,711
170,933
156,562
244,929
57.18
72.98
64.10
86.67
84.28
74.97
31.16
81.57
75.58
72.61
36.61
77.35
49.93
40.11
68.27
PROXIMITY TO PM10 MONITORS
Atlanta
Birmingham
Boston
Chicago
Denver
Detroit
Houston
Los Angeles
New York
Philadelphia
Phoenix
Pittsburgh
Riverside
Seattle
St. Louis
362,201
145,905
945,790
1,018,983
232,974
626,216
377,586
1,207,436
2,710,675
834,110
388,150
449,544
342,334
390,372
358,747
2,115
3,663
6,852
7,619
3,675
1,555
2,085
4,693
2,463
2,740
8,605
13,302
4,181
503
4,316
0.58
2.51
0.72
0.75
1.58
0.25
0.55
0.39
0.09
0.33
2.22
2.96
1.22
0.13
1.20
35,448
35,628
124,911
107,540
43,658
41 ,833
57,413
126,696
37,580
49,413
119,306
133,285
65,499
22,333
55,833
9.79
24.42
13.21
10.55
18.74
6.68
15.21
10.49
1.39
5.92
30.74
29.65
19.13
5.72
15.56
93,903
66,839
262,854
291 ,705
107,548
99,680
166,715
422,725
80,222
154,535
267,456
243,723
182,615
72,979
117,743
25.93
45.81
27.79
28.63
46.16
15.92
44.15
35.01
2.96
18.53
68.91
54.22
53.34
18.69
32.82
139,240
86,299
385,046
441 ,771
168,447
167,760
219,615
810,078
104,951
322,700
348,464
314,941
236,900
123,054
172,535
38.44
59.15
40.71
43.35
72.30
26.79
58.16
67.09
3.87
38.69
89.78
70.06
69.20
31.52
48.09
+Based on 2000 population totals.
Percentages are given with respect to the
total population per city provided.
December 2009
3-38
-------
Table 3-6. Proximity to PM2.6 and PMio monitors based on the population identified as white, black,
Hispanic, or non-Hispanic3. The figures presented here are cumulative for the 15
CSAs/CBSAs examined in Chapter 3.
Proximity to PM Monitors'5
Race or
Hispanic
Origin
Total CSA/CBSA
N
< 1 km < 5 km
N % N %
<10km <15km
N % N %
PROXIMITY TO PM2.5 MONITORS
White
Black
Hispanic
Non-Hispanic
61,936,855
12,668,004
15,916,208
74,611,962
863,823
320,447
445,126
1,135,999
1.39
2.53
2.80
1.52
12,257,978
4,780,620
5,782,482
16,553,574
19.79
37.74
36.33
22.19
27,553,900
9,241,172
10,661,947
36,318,474
44.49
72.95
66.99
48.68
39,030,037
10,906,346
13,094,618
49,629,054
63.02
86.09
82.27
66.52
PROXIMITY TO PM10 MONITORS
White
Black
Hispanic
Non-Hispanic
61,936,855
12,668,004
15,916,208
74,611,962
325,771
134,174
169,305
421,917
0.53
1.06
1.06
0.57
5,554,906
1,611,263
2,496,959
6,767,187
8.97
12.72
15.69
9.07
14,041,215
3,867,436
5,905,322
17,261,734
22.67
30.53
37.10
23.14
21,913,907
6,020,348
8,589,819
27,254,421
35.38
47.52
53.97
36.53
"Based on 2000 population totals
Percentages are given with respect to the total population per city provided.
Socioeconomic Status
Table 3-7 shows the percent of the population below and above the poverty level and the
percent of the population over age 25 years stratified by education level that reside within 1 km,
5 km, 10 km, and 15 km of a PM2.5 and PMi0 monitor cumulatively across the 15 CSAs/CBSAs. For
PM2.5, 80% of the population below poverty level and 77% of the population with less than high
school education are within 15 km of a monitor for the 15 CSAs/CBSAs studied. Populations of
those above the poverty line, those with a high school or more education, and those with a college
education within 15 km of a monitor were less than the low SES groups (67% for each). For PMi0,
47% of the population below the poverty level and 45% of the population with less than a high
school education are within 15 km of a monitor, whereas the percentage of those above the poverty
line (39%), those with a high school or more education (38%), and those with a college education
(35%) were slightly less. Higher percentages of individual SES subpopulations within a given
distance of PM2.5 monitors relative to PMi0 monitors likely reflect the fact that more PM2.5 monitors
are currently deployed within the 15 CSAs/CBSAs studied compared with PMio monitors. Lower
SES groups are not shown to be well-represented at the neighborhood scale, with 1.2% of the
population below the poverty level and 1.0% of the population with less than a high school education
residing within 1 km of a PMio monitor. Likewise, 3.1% of the population below the poverty level
and 2.4% of the population with less than high school education reside within 1 km of a PM25
monitor. However, the populations of low SES groups are more represented at the neighborhood
scale than those for higher SES groups. For example, only about 1.5-1.7% of those above the
poverty line, those with a high school or more education, or those with a college education are within
1 km of a PM2 5 monitor. Moreover, it is notable that at any scale shown in Table 3-7 and for both
PM2 5 and PMio, those living under the poverty line and those age 25 years and older with less than
high school education have greater representation by the monitors than those above the poverty line
or those age 25 and older with high school or college education.
December 2009
3-39
-------
Table 3-7. Proximity to PM2.6 and PMio monitors based on the population below or above the
poverty line, population over age 25 with less than high school education, population
over 25 with high school education, and population over 25 with college education or
more3. The figures presented here are cumulative for the 15 CSAs/CBSAs
examined in Chapter 3.
Proximity to PM Monitors'5
SES
CSM2BSA *1km *5km
N N % N %
< 10 km < 15 km
N % N %
PROXIMITY TO PM2.s MONITORS
Below poverty line
Above poverty line
Less than HS
education
HS education
College education
10,645,411
85,551,420
11,606,042
30,583,598
16,433,811
330,970
1,297,591
276,942
444,262
280,810
3.11
1.52
2.39
1.45
1.71
3,951,549
19,094,985
3,806,208
6,940,261
3,451,717
37.12
22.32
32.80
22.69
21.00
7,107,192
41,736,460
7,225,291
15,152,047
7,776,218
66.76
48.79
62.25
49.54
47.32
8,528,731
57,070,312
8,930,174
20,489,904
11,000,917
80.12
66.71
76.94
67.00
66.94
PROXIMITY TO PM10 MONITORS
Below poverty line
Above poverty line
Less than HS
education
HS education
College education
10,645,411
85,551,420
11,606,042
30,583,598
16,433,811
132,979
485,874
112,901
185,439
59,892
1.25
0.57
0.97
0.61
0.36
1,626,694
8,171,398
1,544,594
2,975,200
1,229,885
15.28
9.55
13.31
9.73
7.48
3,504,957
21,096,615
3,537,414
7,580,000
3,447,148
32.92
24.66
30.48
24.78
20.98
5,024,714
33,034,279
5,186,441
11,747,653
5,722,347
47.20
38.61
44.69
38.41
34.82
"Based on 2000 population totals
Percentages are given with respect to the total population per city provided.
3.5. Ambient PM Concentrations
This section describes measurements of ambient PM mass and composition made since the
2004 PM AQCD (U.S. EPA, 2004, 056905) including analyses using AQS data as well as published
findings. Emphasis is placed on the period from 2005-2007 which incorporates the most recent
validated AQS data available at the time this document was prepared.
When the 2004 PM AQCD (U.S. EPA, 2004, 056905) was written, the full nationwide PM2.5
compliance monitoring network had only recently been deployed, providing three years (1999-2001)
of completed measurements. Based on observations from these first three years, the 2004 PM AQCD
(U.S. EPA, 2004, 056905) found that PM2.5 in eastern cities was generally more highly correlated
across monitoring sites than in western cities. The higher spatial correlations in the eastern cities
resulted from the more regionally dispersed sources of PM25 in the East. Although PM2.5
concentrations at sites within an urban area can be highly correlated, significant differences in
concentrations can occur on any given day. The ratio of PM25 to PMio was found to be higher in the
East than in the West in general, and values for this ratio are consistent with those found in numerous
earlier studies presented in the 1996 PM AQCD (U.S. EPA, 1996, 079380). Differences in the
composition of PM25 between eastern and western cities were also found to be consistent with
differences found in the 1996 PM AQCD (U.S. EPA, 1996, 079380). Much more limited data were
December 2009
3-40
-------
available for describing the spatial variability of coarse particulate mass measured as PM10_2 5, UFPs,
and PM composition. The 2004 PM AQCD (U.S. EPA, 2004, 056905) noted that components
produced by area (e.g., traffic) and point sources are more spatially variable than regionally
dispersed components (e.g., secondary SO42~). Spatial variability will affect estimates of community-
scale human exposure and caution should be exercised in extrapolating conclusions from one area to
another, particularly on a regional scale.
For this PM ISA, the PM2.5 monitoring network has been active for 8 or 9 years depending on
location. Observations and analyses based on PM2.5 measurements reported to AQS are included in
this chapter. Furthermore, by selecting locations where PMi0 and PM2.5 measurements are
co-located, information about the spatiotemporal distribution of the PMi0_2.5 size fraction is
investigated. Given the form of the current standard and the relative abundance of PMi0 monitors in
the AQS network, PMi0 mass concentrations are also included in this section with the understanding
that PM10 includes mass contributions from the smaller size fractions and therefore overlaps with
PM2 5, PMio_2.5 and UFP mass concentrations. Although compliance monitoring does not apply for
UFPs because there is no ambient standard for them, new observational information is available
from detailed studies in several cities. Similarly, advancements have been made in understanding PM
composition from the CSN and IMPROVE networks. Descriptions of UFPs and speciated PM are
covered throughout this section where information is available.
Unless otherwise specified, the PM2 5, PM10_2.5 and PM10 data utilized in this section comes
from the AQS. Based on the population and exposure requirements for monitor siting in 40 CFR Part
58 described in Section 3.4.2, monitors reporting to the AQS are not uniformly distributed across the
U.S. Monitors are far more abundant in urban areas than rural ones, so actual rural spatiotemporal
distributions may differ considerably from those reported here. Furthermore, biases exist for some
PM constituents (and hence, total mass) owing to volatilization losses of NO3~ and other
semi-volatile compounds and, conversely, retention of particle-bound water with hygroscopic
species. The magnitude of these effects is likely to be region-specific.
Spatial distributions of PM across a range of geographic scales are covered in Section 3.5.1.
Temporal behavior including trends, seasonality and hourly variability are covered in Section 3.5.2.
Finally, statistical associations between different size fractions of PM and copollutants including CO,
NO2, O3 and SO2 are covered in Section 3.5.3.
3.5.1. Spatial Distribution
Spatial scales of interest for PM range from global and continental scales (>1000 km) down to
micro scale (-5-100 m). Variation in PM concentration depends on the spatial scale and magnitude
of PM sources, formation and removal mechanisms, and transport and dispersion of PM. These
different sources and processes can cause substantial variation in particle size distribution and
chemistry. This section addresses the spatial variability of PM by focusing primarily on AQS data
across three different scales: variability across the U.S., urban-scale variability and neighborhood-
scale variability. These sections are further subdivided to the extent possible into PM size fractions
and composition.
December 2009 3-41
-------
3.5.1.1. Variability across the U.S.
PM;
2.5
Figure 3-9 shows the 3-yr mean of the 24-h PM2.5 concentrations by county across the U.S. for
2005-2007. The data used in generating this map are from FRM or FRM-like1 data obtained fromthe
AQS database after applying a completeness criterion of 75% per quarter (i.e., 11 out of 15 quarterly
measurements for a l-in-6 day sampling schedule). Counties shown in white did not contain
sufficient PM2.5 data between 2005-2007 to meet the completeness criterion as a result of either a
lack of monitoring sites or a lack of adequate or complete data from existing monitoring sites within
the county. Of the 3,225 U.S. counties, 540 (17%) had PM2.5 data meeting the completeness criterion
in all three years (2005-2007). These 540 counties represent roughly 63% of the U.S. population.
The fraction of the population residing within each county-average concentration range is shown on
the left-hand margin of Figure 3-9. Given the number of counties with no data, the varying size of
counties, and the non-uniform spacing of the monitors and population within each reporting county,
this should only be taken as a rough estimate of the relationship between population and average
ambient concentrations. As seen in Figure 3-9,Kern County, CA reported the highest 3-yr avg 24-h
PM2.5 concentration in excess of 20 ug/m3. Average concentrations between 18 and 20 (ig/m were
reported for several counties in the San Joaquin Valley and inland southern California as well as
Jefferson County, AL containing Birmingham and Allegheny County, PA containing Pittsburgh.
1 FRM-like refers to PM2.5 concehntration data associated with the parameter code "88502 - Acceptable PM2.5 AQI and Speciation Mass"
in the AQS. These data were collected by continuous instruments which are not approved as FRM or FEM, and consequently EPA does
not use these data for regulatory purposes. These data are denoted as "FRM-like" because state and local monitoring agencies have
individually decided that the continuous instruments reporting these data have a degree of agreement with FRM/FEM methods that is
sufficient in their opinion for the data to be used in public advisories regarding current air quality. In some cases, these data include
statistical adjustments by the state/local monitoring agency based on one-time or ongoing correlation analysis with co-located FRM/FEM
monitors, intended to improve the "FRM-likeness" of the continuous concentration data (e.g., Bortnick et al., 2002, 156285). State/local
monitoring agency decisions about whether to adjust continuous PM2.5 data and whether their raw or adjusted continuous PM2.5 data
should be associated with parameter code 88502 were informed by non-binding EPA guidance issued in 2006 (Technical Note on
Reporting PM2 5 Continuous Monitoring and Speciation Data to AQS
http://www.epa.gov/ttn/amtic/files/ambient/pm25/datamang/contrept.pdf).
December 2009 3-42
-------
Concentration Range
• >20.1 Lig/m3[l county]
18.1 - 20.0 ng/m3 [1 counties]
15.1 - 18.1 Lig/m3 [53 counties]
• 12.1 - 15.0 Lig/m3 [242 counties]
• < 12.0 ng/m3 [237 counties]
n No data
Figure 3-9. Three-yr avg 24-h PM2.s concentration by county derived from FRM or FRM-like
data, 2005-2007. The population bar shows the number of people residing within
counties that reported county-wide average concentrations within the specified
ranges.
Table 3-8 contains summary statistics for PM2.5 reported to AQS for the period 2005-2007. All
24-h FRM and 1-h FRM-like data reported to AQS and meeting the completeness criterion outlined
above are included in the table. The table provides a distributional comparison between annual, 24-h
and 1-h averaging times, calendar years (2005, 2006 and 2007) and seasons: winter (December-
February), spring (March-May), summer (June-August), and fall (September-November). In
addition, 15 CSAs/CBSAs were chosen for their importance in recent PM health studies, as
described in Section 3.4, and have been included individually in the table.
The distribution of PM2.5 annual averages (calculated without seasonal weighting and
presented in Table 3-8) was generated from 2,382 individual annual means reported by 794 24-h
FRM monitors reporting to AQS between 2005 and 2007. The mean of the annual averages was
12 (ig/m3, equivalent to the mean of the individual 24-h avg. The maximum annual average PM2.5
concentration calculated from 24-h FRM data over these 3 yr was 23 (ig/m3 in Bakersfield, CA (AQS
monitor ID: 060290010) during 2007. This site is located in the heavily populated portion of the San
Joaquin Valley where air pollution frequently becomes trapped at ground level due to local
topography. The distribution of the 24-h and 1-h avg, both generated from the same 1-h FRM-like
data, are comparable up to the 90th percentile. The 1-h avg is 3 (ig/m3 higher than the 24-h avg at the
95th percentile and 7 (ig/m3 higher at the 99th percentile. This deviation between 1-h and 24-h
averaging times is a result of short duration spikes in PM2.5 mass lasting long enough to influence the
upper percentiles of the 1-h distribution but not necessarily the 24-h avg distribution. Exceptional
events were not removed from this data set and are responsible for at least some of the higher
December 2009
3-43
-------
concentrations observed. For example, the maximum 1-h reading of 828 (ig/m3 was reported by a
monitor in Boise, ID (AQS monitor ID: 160010011) on July 4, 2007. Nine of the top 12 1-h PM2.5
concentrations reported across the country also occurred on July 4th, implicating fireworks as the
common source for these high values.
Table 3-8. PM2.6 distributions derived from AQS data (concentration in ug/m3).
n Meai
1
Percentiles
5
10
25
50
75
90
95
99
2005-2007 PM2.S FOR DIFFERENT AVERAGING PERIODS
Annual avga (24-h FRM)
24-h avg (24-h FRM)
24-havg(1-hFRM-like)
1-havg(1-hFRM-like)
2,382 12
349,028 12
183,057 10
4,403,817 10
5
2
1
0
7
4
2
1
8
4
3
2
10
7
5
4
12
10
8
8
14
16
13
13
16
23
19
21
17
28
24
27
19
39
35
42
23
193
126
828
PM2.5 ANNUAL AND SEASONAL STRATIFICATION USING 24-H AVG FRM DATA
2005
2006
2007
Winter (December-February)
Spring (March-May)
Summer (June-August)
Fall (September-November)
2005-2007 PM2.5 IN INDIVIDUAL
Atlanta
Birmingham
Boston
Chicago
Denver
Detroit
Houston
Los Angeles
New York
Philadelphia
Phoenix
Pittsburgh
Riverside
Seattle
St. Louis
AIMSCSAs/CBSAs
Notinthe15CSAs/CBSAs
114,346 13
113,197 12
121,485 12
86,286 12
88,489 11
86,830 14
87,423 12
2
2
2
2
2
2
2
4
4
4
4
3
4
3
5
4
4
5
4
5
4
7
7
7
7
6
8
6
11
10
10
10
9
12
10
17
15
16
15
14
19
15
24
21
22
22
20
26
22
30
26
27
27
24
31
26
42
36
40
44
33
40
39
133
193
145
193
145
133
126
CSAS/CBSAS USING 24-H AVG FRM DATA
4,939 15
4,869 16
8,464 10
10,308 14
4,192 9
5,223 14
1,342 15
6,600 15
15,826 13
7,541 14
1,634 10
5,783 16
2,751 17
1,297 9
6,887 14
87,656 14
261,372 12
4
4
2
3
2
2
4
3
2
3
2
3
3
2
3
2
2
6
6
3
4
3
3
6
5
4
4
3
5
5
3
5
4
3
7
7
4
6
4
5
8
6
4
5
4
6
6
3
6
5
4
10
10
5
8
6
7
10
9
6
8
6
9
10
4
9
7
6
14
15
9
13
8
12
14
13
10
12
9
13
14
7
13
12
10
19
21
13
18
10
19
18
18
17
18
12
20
21
10
18
17
15
25
29
20
25
14
26
23
25
24
25
17
29
31
20
24
25
22
29
34
24
31
18
31
26
32
29
30
21
36
40
29
29
30
27
37
47
32
42
31
45
34
50
39
38
32
52
58
43
40
42
38
145
64
50
65
61
82
44
133
58
63
77
101
106
68
50
145
193
3Straight annual average without quarterly weighting.
December 2009
3-44
-------
The distribution of the 24-h FRM PM2.5 data was similar across the 3 years (2005-2007)
investigated. Summer (June-August) had the highest mean and median relative to other seasons, but
only by a small margin. For the 99th percentile, winter (December-February) was slightly higher
than the other seasons. This is consistent with wintertime stagnation events resulting in short-term
elevated PM2.5 concentrations. Of the 15 CS As/CBS As investigated, the highest mean of 24-h PM2.5
concentrations was reported for Riverside (17 (ig/m3), Birmingham (16 (ig/m3) and Pittsburgh
(16 (ig/m3); the lowest was reported for Denver (9 (ig/m3) and Seattle (9 (ig/m ).
PMlO-2.5
Since PMi0_2.5 is not routinely measured and reported to AQS, co-located PMi0 and PM2.5
measurements from the AQS network were used to investigate the spatial distribution in PMi0_2.5.
Only low-volume FRM or FRM-like samplers were considered in calculating PMi0_2.5 to avoid
complications with vastly different sampling protocols (e.g., flow rates) between the independent
PMio and PM2.5 measurements. The same 11+ days per quarter completeness criterion discussed
above was applied to the PM10 and PM2.5 measurements. The PM2 5 concentrations are reported to
AQS at local conditions whereas the PMi0 concentrations are reported at standard conditions.
Therefore, prior to calculating PMi0_2.5 by subtraction, the PMi0 AQS data were adjusted to local
conditions on a daily basis using temperature and pressure measurements from the nearest National
Weather Service station. Figure 3-10 shows the 3-yr mean of the 24-h PMi0_2.5 concentration by
county across the U.S. for 2005-2007. There is considerably less coverage for PMi0_2.5 than for PM25
or PM10 alone since only a small subset of PM monitors are co-located and low-volume. The 40
counties included in Figure 3-10 incorporate less than 5% of the U.S. population. Of the 3,225 U.S.
counties, only 40 (1%) met the completeness and co-location criteria in all 3 yr (2005-2007), and
therefore the available measurements do not provide sufficient information to adequately
characterize regional-scale coarse PM spatial concentration distributions.
Table 3-9 contains summary statistics for PMi0_2.5 for the period 2005-2007 similar to those
reported in Table 3-8 for PM25. Only six of the 15 CS As/CBS As had sufficient data for inclusion in
Table 3-9. Although fewer monitoring sites within these CSAs/CBSAs were used for PMi0_2.5 than
for PM25, Table 3-8 and Table 3-9 provide a rough comparison of the PM present in the fine and
thoracic coarse modes for these six cities. The eastern cities including Atlanta, Boston, Chicago and
New York all had a higher fraction in the fine mode with the greatest ratio of fine to thoracic coarse
in Chicago (14 (ig/m3 PM25 5 (ig/m3 PMi0_2s ratio = 2.8). In contrast, Denver (9 (ig/m3 PM25,
20 (ig/myPM10_25, ratio = 0.45) and Phoenix (10 (ig/m3 PM25, 22 (ig/m3 PM10_25, ratio = 0.45) had a
higher fraction in the thoracic coarse mode. Given the limited information available from AQS for
PMio_2.5 and the current NAAQS for PMio, the next section characterizes the more prevalent
data, acknowledging that PMio incorporates both thoracic coarse and fine particles.
December 2009 3-45
-------
Concentration Range
> 26 \igltt? [1 county]
* 21 - 25 \ig/K? [3 counties]
16 - 20 |Jg/m3 [3 counties]
11 - 15^g/m3[17 counties]
" < 10 Lig/m3 [16 counties]
D No data
Figure 3-10. Three-yr avg 24-h PM10.2.s concentration by county derived from co-located low
volume FRM PM10 and PM2.6 monitors, 2005-2007. The population bar shows the
number of people residing within counties that reported county-wide average
concentrations within the specified ranges.
December 2009
3-46
-------
Table 3-9. PMi0.2.6 distributions derived from AQS data (concentration in ug/m3).
n
IV
1
Percentiles
5 10
25
50
75
90
95
99
— Max
2005-2007 PM10-2.s FOR DIFFERENT AVERAGING PERIODS
Annual avga (low volume FRM)
24-h avg (low volume FRM)
130
12,027
12
13
PM10-2.s ANNUAL AND SEASONAL STRATIFICATION USING
2005
2006
2007
Winter (December-February)
Spring (March-May)
Summer (June-August)
Fall (September-November)
3,990
4,037
4,000
2,942
3,088
2,968
3,029
12
13
13
11
13
14
14
3
-3
5 6
1 2
9
6
11
10
14
17
19
26
23
33
39
54
43
246
24-H AVG LOW VOLUME FRM DATA
-5
-2
-2
-5
-2
-2
-2
0 2
1 2
1 3
-1 1
1 2
3 5
1 3
5
6
6
4
5
8
6
10
10
11
8
10
12
11
16
17
18
15
17
18
18
26
27
26
27
26
25
28
33
34
33
34
33
31
34
52
56
56
56
62
44
60
246
182
148
246
151
93
148
2005-2007 PM10-2.s IN INDIVIDUAL CSAS/CBSAS USING 24-H AVG LOW VOLUME FRM DATA"
Atlanta
Boston
Chicago
Denver
New York
Phoenix
All 6 CSAs/CBSAs
Not in the 6 CSAs/CBSAs
167
340
161
353
338
163
1,522
10,505
10
7
5
20
9
22
12
13
-4
-2
-8
0
-16
-3
-6
-2
1 2
1 2
-4 -3
4 6
-2 1
8 11
0 2
1 2
5
4
1
11
5
16
5
6
9
6
4
19
8
20
10
10
13
9
8
28
12
29
17
17
18
12
14
36
17
35
27
26
21
16
19
42
23
46
34
33
30
25
37
59
34
67
51
56
46
27
37
78
56
70
78
246
3Straight annual average without quarterly weighting.
bNo co-located low-volume FRM PM10and FRM-like PM25 monitors available for Birmingham, Detroit, Houston, Los Angeles, Philadelphia, Pittsburgh, Riverside, Seattle or St. I
PM10
Figure 3-11 shows the 3-yr mean of the 24-h PMi0 concentrations by county across the U.S.
for 2005-2007. Both FRM and FEM PMi0 data reported to AQS were included and the same
11+ days per quarter completeness criterion described above for PM2.5 was applied. The highest
3-yr avg for PMi0 (>50 (ig/m3) occurred in inland southern California and the populous counties of
southern Arizona and central New Mexico. Of the 3,225 U.S. counties, 676 (12%) contained PM10
data meeting the completeness criterion in all three years; these 676 counties incorporate
approximately 43% of the U.S. population.
December 2009
3-47
-------
Concentration Range
• > 51 \ig/m3 [1 counties]
"41-50 ^g/m3 [15 counties]
31-40 \ig/m3 [378 counties]
• 21 - 30 >ig/m3[ 162 counties]
* < 20 \ig/m3 [114 counties]
n No data
Figure 3-11. Three-yr avg 24-h PM10 concentration by county derived from FRM or FEM
monitors, 2005-2007. The population bar shows the number of people residing
within counties that reported county-wide average concentrations within the
specified ranges.
Table 3-10 contains summary statistics for PMi0 reported to AQS for the period 2005-2007.
Both 24-h FRM and 1-h FEM data are included in the table. To facilitate a distributional comparison
between averaging times, annual, 24-h and 1-h averaging times using the FRM and FEM data have
been included separately in Table 3-10. As in the earlier tables, the data is also stratified by year and
season and includes the 15 CSAs/CBSAs individually.
December 2009
3-48
-------
Table 3-10. PMio distributions derived from AQS data (concentration in
n
Mea
1
ug/m3).
Percentiles
5 10
25
50
75
90
95
99
2005-2007 PM10 FOR DIFFERENT AVERAGING PERIODS
Annual avga (24-h FRM and 1-h FEM)
24-havg(24-hFRMand1-hFEM)
24-h avg (24-h FRM)
24-h avg (1-h FEM)
1-h avg (1-h FEM)
2022
326,675
167,310
156,931
3,767,533
25
26
25
26
27
10
3
2
4
1
14 16
6 9
6 9
7 9
4 6
19
14
14
14
11
23
21
21
21
19
28
32
31
32
32
35
46
45
48
51
44
59
57
62
69
60
97
91
105
145
85
8299
8299
979
8540
PM10 ANNUAL AND SEASONAL STRATIFICATION USING 24-H AVG FRM AND FEM DATA
2005
2006
2007
Winter (December-February)
Spring (March-May)
Summer (June-August)
Fall (September-November)
2005-2007 PM10 IN INDIVIDUAL
Atlanta
Birmingham
Boston
Chicago
Denver
Detroit
Houston
Los Angeles
New York
Philadelphia
Phoenix
Pittsburgh
Riverside
Seattle
St. Louis
AIMSCSAs/CBSAs
Notinthe15CSAs/CBSAs
107,524
109,505
109,646
80,959
82,772
81,351
81,593
25
26
26
23
25
29
26
2
3
4
2
2
6
3
6 9
6 9
7 9
5 7
6 8
10 12
7 9
13
13
14
11
13
18
14
21
21
21
17
20
25
21
31
32
32
27
31
35
32
46
46
47
42
45
49
48
58
59
60
57
58
60
62
93
101
99
99
96
92
102
1441
8299
2253
8299
2253
1839
1212
CSAS/CBSAS USING 24-H AVG FRM AND FEM DATA
1,868
5,478
1,412
6,165
4,706
1,407
1,397
2,020
514
4,207
12,005
12,677
4,327
2,136
2,464
62,783
263,892
24
34
17
26
28
30
31
27
19
19
52
24
35
19
33
32
24
6
6
2
6
5
7
7
4
2
4
7
4
4
5
6
5
2
9 11
9 12
5 7
9 11
10 12
10 12
10 12
8 11
6 7
7 9
14 19
7 9
8 11
7 9
10 12
8 10
6 8
16
19
10
16
18
18
17
18
11
12
29
13
19
12
18
16
13
23
28
15
23
25
26
23
25
17
17
44
19
30
17
28
25
20
31
43
22
32
35
38
34
33
25
24
65
31
45
23
42
39
30
39
64
30
45
47
53
56
42
35
34
91
45
64
31
59
60
43
44
82
36
55
54
64
80
51
40
40
112
57
75
37
74
77
54
57
120
50
78
75
81
137
74
51
52
166
83
111
52
114
120
88
108
241
58
214
118
182
248
489
83
84
2253
157
1212
79
315
2253
8299
3Straight annual average without quarterly weighting
December 2009
3-49
-------
The maximum annual average PM10 concentration calculated from 24-h FRM data over these
three years was 85 (ig/m3 in Stanfield, AZ (AQS monitor ID: 040213008) during 2007. Stanfield is a
small agricultural town (2007 population = 1074) approximately 64 km south of Phoenix and is in a
region heavily influenced by windblown dust. Many of the maximum 24-h and 1-h avg PMi0
concentrations in Table 3-10 exceed 1,000 (ig/m3, but these represent rare events given the much
lower 99th percentiles. Exceptional events were not removed from this data set and are responsible
for at least some of the higher concentrations observed.
The distribution of the 24-h FRM and FEM PMi0 data was similar across the three years
(2005-2007) investigated. Summer (June-August) had the highest mean and median relative to other
seasons, consistent with PM2.5 observations. Of the 15 CSAs/CBSAs investigated, the highest mean
of 24-h PMio concentrations was reported for Phoenix (52 (ig/m3), considerably higher than the
means for the other CSAs/CBSAs investigated. The lowest was reported for Boston (17 (ig/m3) with
New York, Philadelphia and Seattle only slightly higher (19 (ig/m ).
On average using the 2005-2007 data for PM2.5 in Table 3-8 and PMi0 in Table 3-10, the
distribution between fine and coarse PM varies substantially by location. A larger fraction of PM
mass is present in the thoracic coarse mode in Phoenix and Denver (3-yr mean PM2.5/PMi0 ratios of
0.19 and 0.32, respectively). In contrast, a larger fraction is present in the fine mode in Philadelphia
(0.74), New York (0.68) and Pittsburgh (0.67). Comparisons of PM2.5 to PMi0 as reported to AQS
should be used with caution, however, since PM2 5 concentrations are reported for local conditions
while PMio concentrations are converted to STP before reporting. Nevertheless, these findings are
consistent with those in Table 3-9 for PMi0_2.5 in the subset of 6 cities with available co-located low-
volume PM data that have been properly adjusted for temperature and pressure. These findings are
also consistent with those reported in the 2004 PM AQCD (U.S. EPA, 2004, 056905) where ratios of
PM2 5 to PMio were observed to be highest in the northeast (0.70), southeast (0.70), and industrial
Midwest (0.70) and lower in the upper Midwest (0.53), northwest (0.50), southern California (0.47)
and southwest (0.38).
UFPs
Little is known about the spatiotemporal distribution or composition of UFPs on a regional
scale. New particle formation has been observed in environments ranging from relatively unpolluted
marine and continental environments to polluted urban areas as an ongoing background process and
during nucleation events (Kulmala et al, 2004, 089159). During nucleation events, which may last
for several hours, UFP number concentrations can exceed 104 per cm3 over distances of several
hundred kilometers (Kulmala et al., 2004, 089159; Qian et al., 2007, 116435). These events occur
throughout the year on 5-40% of days, depending on location (Qian et al., 2007, 116435). Cloud
condensation nuclei, with diameters between ~10 and -100 nm have been monitored for several
years at a number of nonurban sites in the U.S. (http://cmdl.noaa.gov/aero/data/). Average particle
number counts at these sites in the U.S. range from several hundred to several thousand per cm3. The
particles are formed by nucleation of atmospheric gases with additional contribution from primary
emissions in these environments (Pierce and Adams, 2009, 191189).
In an urban setting, a large percentage of UFPs come from combustion-related emissions from
mobile sources (Sioutas et al., 2005, 088428). UFP number concentrations drop off quickly with
distance from the roadway (Levy et al., 2003, 052661; Reponen et al., 2003, 088425; Zhu et al.,
2005, 157191). and therefore concentrations can be highly heterogeneous in the near-road
environment depending on traffic, meteorological and topographic conditions (Baldauf et al., 2008,
190239). Studies characterizing spatial variability in UFPs are currently limited to a handful of close
proximity locations and therefore are discussed in Sections 3.5.1.2 and 3.5.1.3 in the context of
urban- and neighborhood-scale variability. Further elaboration on the composition of UFPs is
included below.
PM Constituents
Only PM25 is collected routinely at CSN network sites so the majority of this section on PM
constituents is devoted to PM2 5 composition. PM10_2.5 and UFP composition is discussed to the
extent possible below. Figure 3-12 through Figure 3-16 contain U.S. concentration maps for OC, EC,
SO42~, NO3", and NH4+ mass from PM2 5 measurements taken as part of the CSN network for the
period 2005-2007. Data used in these figures are as reported to AQS: no correction was applied to
OC for non-carbon mass and NO3" represents total particulate nitrate. Figure 3-12 shows regions of
December 2009 3-50
-------
high PM2.5 OC mass concentration with annual average concentrations greater than 5 (ig/m3 in the
western and the southeastern U.S. Concentrations at the western monitors peak in the fall and winter
while those in the Southeast peak anywhere from spring through fall. The central and northeastern
portions of the U.S. generally contain lower measured OC. Bell et al. (2007, 155683) present a
similar map for estimated organic carbon mass (OCM) from 2000-2005 calculated by multiplying
the blank corrected OC measurement by 1.4 to account for non-carbon mass. There are a range of
estimates in the literature for suggested scaling factors (Turpin and Lim, 2001, 017093). depending
predominantly on how highly oxygenated the aerosol is (Pang et al., 2006, 156012). Fresh PM, more
common in urban regions, has undergone limited chemical transformation. As the aerosol is
transported to rural regions, it becomes more oxygenated. Turpin and Lim (2001, 017093)
recommended ratios of 1.6 ± 0.2 for urban and 2.1 ± 0.2 for non-urban aerosols. Estimates range
from 1.6 to 2.6 for rural IMPROVE monitors (El-Zanan et al., 2005, 155764). Therefore, applying
one correction factor of 1.4 across the entire U.S. will lead to an underestimate of the OCM in rural
regions. Therefore, the OC data in Figure 3-12 is presented as measured with a national blank
correction, but no adjustment to OCM.
Figure 3-13 contains a similar map for PM2.5 EC mass concentration that exhibits smaller
seasonal variability than OC, particularly in the eastern half of the U.S. There are isolated monitors
spread throughout the country that measure high annual average EC concentrations. These EC 'hot
spots' are primarily associated with larger metropolitan areas such as Los Angeles, Pittsburgh, and
New York, but El Paso, TX, also reported high annual average EC concentrations (driven by a
wintertime average concentration greater than 2 (ig/m3). In a similar analysis for EC by Bell et al.
(2007, 155683) for 2000-2005 data, there were also high wintertime EC concentrations in eastern
Kentucky and western Montana. These particular locations do not stand out in the 2005-2007 data in
Figure 3-13.
Figure 3-14 contains a map for PM2.5 SO42~ mass concentration which shows that SO42~ is
more prevalent in the eastern U.S. owing to the strong west-to-east gradient in SO2 emissions. This
gradient is magnified in the summer months when more sunlight is available for photochemical
formation of SO42~. In contrast, PM2 5 NO3~ mass concentration in Figure 3-15 is highest in the west,
particularly in California. There are also elevated concentrations of NO3~ in the upper Midwest. The
seasonal plots show generally higher NO3~ in the wintertime as a result of temperature driven
partitioning. Exceptions exist in Los Angeles and Riverside where high NO3~ readings appear year-
round. The PM25 NH4+ mass concentration maps in Figure 3-16 shows spatial patterns related to
both SO42~ and NOjT resulting from its presence in both (NH4)2SO4 and NH4NO3. Figure A-31
through Figure A-36 in Annex A show similar U.S. concentration maps for PM25 Cu, Fe, Ni, Pb, Se
and V mass concentrations as measured by XRF. There is considerably less seasonal variation in the
concentration profile for these metals than OC or the ions.
For the 15 metropolitan areas identified earlier, the contribution of the major component
classes to total PM2 5 mass was derived using the measured sulfate, adjusted NO3~, derived water,
inferred carbonaceous mass approach (SANDWICH) (Frank, 2006, 098909). This approach uses the
measured FRM PM2 5 mass and co-located CSN chemical constituents to perform a mass balance-
based estimation of the PM25 mass fraction attributed to SO42~, NO3~, EC, OCM, and crustal
material. SO42~ and NO3~ include associated NH4+ mass and estimated particle-bound water.
Furthermore, NO3~ is assumed to be fully neutralized as NH4NO3 and has been adjusted to represent
the amount retained by the FRM monitor. EC is taken as measured, and the crustal component is
derived from common oxides contained in the Earth's crust (Pettijohn, 1957, 156862). but can also
include significant anthropogenic contributions, such as coal fly ash that are unrelated to soil
resuspension. Finally, OCM is estimated using mass balance by subtracting the sum of all other
constituents from the FRM PM2 5 mass. The SANDWICH method takes into account passive
collection of semi-volatile or handling-related mass on the FRM filters in the mass balance
calculation. The magnitude of this artifact is assigned a nominal value of 0.5 (ig/m3, which is derived
from limited analysis of FRM field blanks. Other constituents such as salt and other metallic oxides,
however, are not included in these calculations and therefore the OCM fraction estimated by mass
balance represents an upper bound on the FRM retained OCM. The calculations and assumptions
that go into the SANDWICH method are discussed in detail in Frank (2006, 098909)with further
information available on EPA's AirExplorer web site
http://www.epa.gov/cgi-bin/htmSOL/mxplorer/query spe.hsql
December 2009 3-51
-------
oc
Annual
tX^v^S
Off)
«.o
o
Concentration (ng/m3): <1.25 >1.25-2.5C >2.50 -3.75 >3.75-5.jO >5.00
Spring Summer
Fall
Winter
Figure 3-12. Three-yr avg 24-h PM2s OC concentrations measured at CSN sites across the
U.S., 2005-2007.
December 2009
3-52
-------
EC
Annual
Concentration (ng/m3): <0.5
Spring
• • •
>0.5-1.G M.0-1.5 >1.5-2.0 >2.0
Summer
. •. "s° ^'^j.0
•• '' ,-&
.
Fall
Winter
• ,*
Figure 3-13. Three-yr avg 24-h PM2s EC concentrations measured at CSN sites across the
U.S., 2005-2007.
December 2009
3-53
-------
so/
Annual
Concentration (ng/m3): <2
*
>2-4
c
>4-6
>B
Spring
Summer
Fall
Winter
Figure 3-14. Three-yr avg 24-h PM2s S042~ concentrations measured at CSN sites across the
U.S., 2005-2007.
December 2009
3-54
-------
N03-
Annual
Concentration (ng/m3): <1
Spring
Fall
>2-3 >3-4 >4
Summer
Winter
Figure 3-15. Three-yr avg 24-h PM2.s N0s~ concentrations measured at CSN sites across the
U.S., 2005-2007.
December 2009
3-55
-------
NH4+
Annual
Concentration (ng/m3): <1
Spring
>2-3 . X3-4 >4
Summer
Fall
Winter
*. .
Figure 3-16. Three-yr avg 24-h PM2.s NH4+ concentrations measured at CSN sites across the
U.S., 2005-2007.
December 2009
3-56
-------
Figure 3-17 shows the PM2.5 compositional breakdown for the 15 CSAs/CBSAs. All available
monitoring sites with co-located FRM PM2.5 and CSN speciation data reporting in all four seasons
for at least one calendar year from 2005-2007 were included. Furthermore, each season was required
to contain five reported values for mass and the major PM2.5 constituents. This resulted in a varying
number of sites (ranging from one to seven, as indicated in the caption to Figure 3-17) used to create
the averages shown in the figure. Variability in PM2.5 composition within each CSA/CBSA where
multiple monitors were available and trends in composition over time are discussed in subsequent
sections.
Sulfate f
Nitrate
EC
OCM
Crustal
Figure 3-17. Three-yr avg PM2.s speciation estimates for 2005-2007 derived using the
SANDWICH method. For the following 15 CSAs/CBSAs (with the number of sites
per CSA/CBSA listed in parenthesis): Atlanta, GA (1); Birmingham, AL (3);
Boston, MA (4); Chicago, IL (7); Denver, CO (2); Detroit, Ml (4); Houston, TX (1);
Los Angeles, CA (1); New York City, NY (7); Philadelphia, PA (6); Phoenix, AZ (2);
Pittsburgh, PA (4); Riverside, CA (1); Seattle, WA (4); and St. Louis, MO (3). S04*~
and N0s~ estimates include NH4+ and particle bound water and the circle area is
scaled in proportion to FRM PM2.s mass as indicated in the legend.
On an annual average basis, SO42~ is a dominant PM component in the eastern U.S. cities. For
the presented cities, this includes everything east of Houston where the SO42~ fraction of PM25
ranges from 42% in Chicago to 56% in Pittsburgh on an annual average basis. OCM is the next
largest component in the east ranging from 27% in Pittsburgh to 42% in Birmingham. In the west,
OCM is the largest constituent on an annual basis, ranging from 34% in Los Angeles to 58% in
Seattle. SO42~, NO3~ and crustal material are also important in many of the included western cities. In
the west, fractional SO42~ ranges from 18% in Denver to 32% in Los Angeles while fractional NO3~
is relatively large in Riverside (22%), Los Angeles (19%) and Denver (15%) and less important on
an annual basis in Phoenix (1%) and Seattle (2%). Crustal material is particularly prevalent in
Phoenix (28%). EC makes up a smaller fraction of the PM25 (4-11%), but it is consistently present in
all included cities regardless of region.
December 2009
3-57
-------
Spring
Summer
Figure 3-18. Seasonally-stratified 3-yr avg PM2 5 speciation estimates for 2005-2007 derived
using the SANDWICH method. For the following 15 CSAs/CBSAs: Atlanta, GA;
Birmingham, AL; Boston, MA; Chicago, IL; Denver, CO; Detroit, Ml; Houston, TX;
Los Angeles, CA; New York City, NY; Philadelphia, PA; Phoenix, AZ; Pittsburgh,
PA; Riverside, CA; Seattle, WA; and St. Louis, MO. S042" and N03~ estimates
include NH4+ and particle bound water and the circle area is scaled in proportion
to FRM PM2.s mass as indicated in the legend.
The seasonal variation in PM2.5 composition across the 15 CSAs/CBSAs is shown in Figure
3-18 where the seasons are defined as before. SO42~ dominates in most metropolitan areas in the
summertime, while NO3~ becomes important in the colder wintertime months. Notable exceptions
include Denver, Phoenix, Riverside, and Seattle where summertime SO42~ makes up a smaller
fraction of the PM2.5 mass compared with other regions. Likewise, NO3~ is less pronounced in the
wintertime in Atlanta, Birmingham, Houston, Phoenix, and Seattle compared with other regions. Los
Angeles and Riverside exhibit elevated NO3~ from fall through spring. Crustal material is a
substantial summertime component in Houston (26%), and is generally low elsewhere in the East in
all seasons. In the West, crustal material represents a substantial component year-round in Phoenix
and Denver.
The only PM size fraction routinely collected at CSN network sites is PM25, resulting in less
available information on speciated PM10_2.5. Edgerton et al. (2005, 088686; 2009, 180385) published
speciated measurements for PM2 5 and PMi0_2.5 obtained using dichotomous samplers from four
locations included in the Southeastern Aerosol Research and Characterization (SEARCH) study:
Yorkville, GA, Centreville, AL, Birmingham, AL and Atlanta, GA. Samples were collected between
1999 and 2003 on a l-in-3 day or l-in-6 day schedule, depending on site. Speciated measurements
for both PM25 and PMi0_2.5 included SO42~, NO3", NH4+, and major metal oxides (MMO). In addition,
OC and either black carbon (BC) or EC were reported for PM2 5 over the entire study period and for
PMio_2.5 for a subset of samples extending from April 2003 to April 2004.
For the Atlanta and Birmingham SEARCH sites, the annual average NO3" mass fraction was
approximately equal for PM25 (5.6% and 5.0%, respectively, for Atlanta and Birmingham) and PMi0_
2.5 (4.9% and 3.3%). Likewise, the OC mass fraction was approximately equal for PM25 (26% and
5%) and PMi0_25 (24% and 27%). MMO contributed an order of magnitude smaller mass fraction to
PM2.5 (2.6% and 4.7%) than PM10_2.5 (38% and 35%). In contrast, SO?" contributed an order of
December 2009
3-58
-------
magnitude greater mass fraction to PM25 (25.1% and 24.1%) than PM10_25 (2.8 and 2.1%). BC also
contributed a larger mass fraction of PM2.5 (8.6% and 10.5%) than EC did for PMi0.2.5 (2.9% and
2.4%). Based on these findings, MMO are present primarily in the thoracic coarse mode, while SO42~
and EC/BC are present primarily in the fine mode. NO3~ and OC are present in both modes in
approximately equal mass fractions. These results are specific to Atlanta and Birmingham and may
not represent other geographic regions.
Information about the composition of ambient UFPs directly emitted by sources is still sparse
compared to that for the larger size modes. However, their composition is expected to reflect that of
their sources. As noted in Section 3.3 (and references therein), particle number emissions from motor
vehicles are dominantly in the UF size range. The composition of gasoline vehicle emissions consists
mainly of a mix of OC, EC and small quantities of trace metals and sulfates, with OC constituting
anywhere from 26-88% of PM. Diesel PM is generally comprised of an EC and trace metal ash core
onto which organic material and nucleation-mode SO42~condense. With the introduction of new
diesel emissions standards in 2007, total emissions have decreased dramatically, particularly for
carbon. In areas where atmospheric nucleation is the dominant source of UFPs, sulfate along with
ammonium, and secondary organic compounds are the likely major components of UFPs.
In a study conducted at several urban sites in Southern California, Cass et al. (2000 020680)
found that the composition of UFPs ranged from 32-67% OC, 3.5-17.5% EC, 1-18% SO4 , 0-19%
NO3~, 0-9% NH4+, 1-26% metal oxides, 0-2% Na, and 0-2% Cl. Thus carbon, in various forms, was
found to be the major contributor to the mass of UFPs. However, ammonium was found to contribute
33% of the mass of UFPs at one site in Riverside. Fe was the most abundant metal found in the
UFPs. Chung et al. (2001, 017105) found that carbon was the major component of the mass of UFPs
in a study conducted during January of 1999 in Bakersfield, CA. However, in the study of Chung et
al. (2001, 017105). the contribution of carbonaceous species (OC and EC; typically 20-30%) was
much lower than that found in the cities in Southern California. They found that Ca was the
dominant cation, accounting for about 20% of the mass of UFPs in their samples. Sizable
contributions from Si (0-4%) and Al (6-14%) were also found. MOUDIs are used to collect size-
segregated filter samples in the UFP compositional analyses described above. Coarse particle bounce
is a concern when using MOUDIs and further studies, including scanning electron microscopy, may
be needed to quantify the effect of this sampling artifact on UFP compositional analyses.
Herner et al. (2005, 135983) reported a gradual increase in OC mass fraction as particle size
decreases from 1 (im (20% OC) to 100 nm (80% OC) in the San Joaquin Valley of California. Sardar
et al. (2005, 180086) found OC to be the major component of UFPs at four locations in California,
with higher OC mass fraction in the wintertime relative to summertime. EC and SO42~ were also
present in the UFP samples, but at much smaller mass fractions. EC was present year-round,
whereas SO42~had a summertime increase. More detailed chemical characterization of the OC
fraction of ambient UFPs is extremely limited, but recent studies have identified specific organic
molecular markers affiliated with motor vehicle emissions including hopanes and PAHs (Fine et al.,
2004, 141283; Ning et al., 2007, 156809; Phuleria et al., 2007, 117816).
As noted in the 2004 PM AQCD (U.S. EPA, 2004, 056905). primary biological aerosol
particles (PBAP), which include microorganisms, fragments of living things, and organic compounds
of miscellaneous origin in surface deposits on filters, are not distinguishable in analyses of total OC.
A clear distinction should be made between PBAP and primary OC that is produced by organisms
(e.g., waxes coating the surfaces of organisms) and precursors to secondary OC such as isoprene and
terpenes. Indeed, the fields of view of many photomicrographs of PM samples obtained by scanning
electron microscopy are often dominated by large numbers of pollen spores, plant and insect
fragments, and microorganisms. Bioaerosols such as pollen, fungal spores, and most bacteria are
expected to be found mainly in the coarse size fraction (see Figure 3-2 for an illustrative example of
a pollen particle). However, allergens from pollens can also be found in respirable particles
(Edgerton et al., 2009, 180385; Taylor, 2002, 025693). Matthias-Maser et al. (2000, 155972)
summarized information about the size distribution of PBAP in and around Mainz, Germany in what
is perhaps the most complete study of this sort. Matthias-Maser found that PBAP constituted up to
30% of total particle number and volume in the approximate size range from 0.35-50 (im on an
annual basis. Additionally, whereas the contribution of PBAP to the total aerosol volume did not
change appreciably with season, the contribution of PBAP to total particle number ranged from
about 10% in December and March to about 25% in June and October. Bauer et al. (2008, 189986)
measured contributions of fungal spores to OC at an urban and a suburban site in Vienna, Austria in
spring and summer. Fungal spores at the suburban site contributed on average 10% to OC in PMi0
December 2009 3-59
-------
and 5% at the urban site. At the suburban site, in summer, fungal spores accounted on average for
60% of the OC (0.56 (ig/m3) in PMi0_2.i (2.6 (ig/m3). The contribution to PM2.i was estimated to be
about 10% that in PMi0_21. Womiloju et al. (2003, 179954) estimated that fungal spores contribute
14-22% of OC in PM2 5 in and around Toronto.
Edgerton et al. (2009, 180385) found that PBAP contributed 60-70% of OC (average
~1.7 (ig/m3) in PMi0_2.5 at an urban and a rural site in Alabama in fall of 2000 and spring of 2001.
The percentage contributions were similar at both sites and higher concentrations were found in
spring than in fall. Although results for the U.S. are more limited, they are broadly consistent with
the results of the other studies in illustrating the importance of PBAP, at least for fungal spores in
OC.
3.5.1.2. Urban-Scale Variability
PM2.5
Data from the 15 CSAs/CBSAs were used to investigate urban-scale variability in PM
reported to AQS. PM2.5 has a longer residence time in the atmosphere compared to PMi0_2.5 resulting
from a slower Vd. As a result, PM2.5 exhibits increased spatial homogeneity with relatively less
localized influence from point sources. Maps of PM2.5 monitor locations and box plots of seasonal
PM2.5 mass concentration data are provided for Boston (Figure 3-19 and Figure 3-20), Pittsburgh
(Figure 3-21 and Figure 3-22), and Los Angeles (Figure 3-23 and Figure 3-24). Figures A-37 through
A-80 in Annex A contain similar information for all 15 CSAs/CBSAs under investigation. With very
few exceptions, the PM2 5 concentration is quite uniformly distributed across the monitors. Los
Angeles has one monitor (Site I) that reported noticeably less PM2 5 in all four seasons than the rest
of the monitors in the region. This monitor is located at Lancaster CA, separated from the rest of the
Los Angeles region by the San Gabriel Mountains. In general, however, PM2 5 varies approximately
the same magnitude between monitors as it does between seasons for the 15 selected cities.
Table 3-11 through Table 3-13 contain pair-wise monitor site comparison statistics for PM25 in
Boston, Pittsburgh, and Los Angeles, respectively. Tables A-20 through A-34 in Annex A contain the
same statistics for PM25 measured within all 15 of the CSAs/CBSAs investigated. Comparison
statistics shown include the Pearson correlation coefficient (R), the 90th percentile of the absolute
difference in concentrations (P90), the coefficient of divergence (COD) and the number of paired
observations (n). The COD provides an indication of the variability across the monitoring sites in
each CSA/CBSA and is defined as follows:
Equation 3-2
where X^ and X& represent observed concentrations averaged over some measurement averaging
period i (hourly, daily, etc.) at sites j and k, and/? is the number of paired observations. A COD of 0
indicates there are no differences between concentrations at paired sites (spatial homogeneity), while
a COD approaching 1 indicates extreme spatial heterogeneity.
Temporal correlations between 24-h PM25 concentrations in Boston range from 0.61 to 0.97 in
Table 3-11. The lowest correlation in this CSA was between Site A located in Fall River, MA, 1 km
from the Narragansett Bay and Site L located in Nashua, NH, on the bank of the Merrimack River,
120 km north. The highest correlation was between Sites P and R, located less than a kilometer apart
in Providence, RI.
In Pittsburgh, 24-h PM25 correlations range from 0.65 to 0.97. The lowest correlation in this
CSA was between Sites B and D, located diametrically opposite downtown Pittsburgh and 33 km
apart. The highest correlation was for Sites K and D, located 21 km apart and both west of
downtown. The prevailing wind in Pittsburgh is from the west, which explains the higher correlation
between the two upwind sites.
In Los Angeles, 24-h PM25 correlations range from 0.21 to 0.96. The lowest correlation was
between Sites I and J located 123 km apart and separated by the San Gabriel Mountains as discussed
December 2009 3-60
-------
earlier. The highest correlation was for sites G and H, located 3.7 km apart and both in Long Beach,
CA. Therefore, while distance between monitors plays an important role in how well any two
monitors correlate, other factors such as meteorology and topography can be important as well.
Boston Combined Statistical Area
Ql
r
Boston CSA
• PMa.s Monitors
Interstate Highways
— Major Highways
0 10 20
40
60
80
100
zzi Kilometers
Figure 3-19. Locations of PM2.s monitors and major highways, Boston, MA.
December 2009
3-61
-------
ABCDEFGH J
Mean 9.1 9.1 8.9 9.4 9.6 11.7 11.6 10.5 12.1 10.7
Obs 341 350 342 355 357 349 398 349 1015 335
SD 6.0 6.5 6.3 6.6 6.6 7.0 6.8 6.9 6.9 7.2
40 -
AQSStelD
Site A 25-005-1004
SiteB 25-009-2006
SiteC 25-009-5005 .„
—. 30 -
SiteD 25-009-6001 ^
SiteE 25-023-0004 'g,
SiteF 25-025-0002 —
SiteG 25-025-0027 o
SiteH 25-025-0042 Ł 20 -
Sitel 25-025-0043 Ł
SiteJ 25-027-0016 JJ
c
0
1 <
...
1 1 1 1
1 ^winter P
2-spring
3=summer
-1-fall ^ ~
1234 1234 1234 1234 1234 1234 1234 1234 1234 1234
KLMNOPQRS
Mean 11.4 7.2 10.0 9.7 8.9 10.1 11.9 10.5 9.7
Obs 345 183 361 362 362 1027 321 313 998
SD 7.4 5.3 6.7 6.0 5.8 6.6 6.9 6.5 6.5
40-
AQSSitelD
SiteK 25O27O023
SiteL 33001-2004
SiteM 33011-1015
30-
SiteN 33013-1006 "
SiteO 33O15O014 -^
O)
SiteP 44007O022 ^
SiteQ 44007O026 <=
O
SiteR 44007O028 '^20-
SiteS 44007-1010 Z
d
o
o I
0 | j ||
10-
1=winter
2=spring
3= summer
4=fall „.
1234 1234 1234 1234 1234 1234 1234 1234 1234
Figure 3-20. Seasonal distribution of 24-h avg PM2.s concentrations by site for Boston, MA,
2005-2007. Box plots show the median and interquartile range with whiskers
extending to the 5th and 95th percentiles at each site during (1) winter
(December-February), (2) spring (March-May), (3) summer (June-August) and (4)
fall (September-November).
December 2009
3-62
-------
Table 3-11. Inter-sampler comparison statistics for each pair of 24-h PM2.6 monitors reporting to
AQS for Boston, MA.
Site ABCDEFGH I J
A 1.00 0.80 0.77 071 0.84 0.79 0.78 0.79 0.79 0.77
(6.6,0.21) (6.2, 0.22) (6.9, 0.23) (4.8,0.19) (8.1,0.23) (7.7, 0.24) (6.8, 0.22) (7.9, 0.25) (7.5, 0.24)
326 318 323 329 318 319 325 338 310
1.00 0.92 0.87 0.87 0.90 0.90 0.90 0.90 0.85
(4.1,0.17) (4.1,0.18) (4.7,0.19) (6.3, 0.21) (6.2, 0.23) (4.9,0.19) (7.1,0.26) (5.5, 0.21)
328 331 339 326 323 333 343 317
1.00 0.90 0.85 0.90 0.89 0.90 0.88 0.86
(0.0, 0.00) (3.5,0.17) (5.3,0.21) (6.3, 0.23) (6.3, 0.24) (5.0, 0.20) (6.8, 0.26) (6.2,0.21)
342 321 331 316 318 326 336 311
LEGEND
Pearson R
E (P90, COD)
n
(0.0, 0.00)
355
(5.6, 0.20)
336
1.00
(0.0, 0.00)
(5.8,0.21)
324
0.90
(5.9,0.19)
(5.8, 0.22)
329
0.90
(5.8,0.21)
(4.6,0.19)
332
0.89
(5.0,0.19)
(7.0, 0.26)
345
0.87
(6.9, 0.24)
(5.8,0.19)
313
0.87
(5.4, 0.20)
1.00 0.94 0.94 0.92 0.92
(3.8,0.14) (3.5,0.15) (4.5,0.17) (5.4,0.18)
349 324 324 339 310
1.00 0.94 0.94 0.89
(4.0,0.16) (4.3,0.15) (5.7, 0.20)
325 338 308
1.00 0.93 0.89
(4.7,0.19) (5.0,0.17)
349 342 318
1.00 0.86
1015 330
1.00
335
December 2009 3-63
-------
Site
A
B
C
D
E
F
G
H
1
J
K
L
M
N
0
P
Q
R
S
K
0.77
(8.1,0.23)
320
0.86
(6.6,0.21)
329
0.86
(6.9,0.21)
321
0.88
(6.4,0.19)
325
0.87
(6.3, 0.20)
333
0.91
(4.7,0.17)
323
0.90
(5.0,0.19)
320
0.90
(4.4,0.17)
327
0.87
(6.1,0.20)
341
0.95
(3.0,0.14)
316
1.00
(0.0, 0.00)
346
L
0.61
(8.3, 0.29)
173
0.80
(6.2, 0.23)
175
0.89
(4.8, 0.23)
173
0.79
(5.7, 0.25)
174
0.72
(8.3, 0.27)
179
0.78
(9.6, 0.33)
168
0.77
(9.0, 0.33)
172
0.75
(9.4, 0.30)
175
0.75
(10.0,0.36)
181
0.73
(9.2, 0.28)
167
0.71
(10.3,0.31)
170
1.00
(0.0, 0.00)
183
LEGEND
Pearson R
(P90, COD)
n
M
0.71
(8.0, 0.23)
324
0.87
(5.3,0.19)
331
0.93
(4.4,0.17)
323
0.91
(3.5,0.16)
329
0.83
(5.8,0.17)
338
0.90
(5.3,0.18)
323
0.90
(5.3,0.19)
326
0.88
(4.9,0.18)
332
0.86
(6.7, 0.22)
352
0.87
(5.2,0.18)
314
0.88
(6.0,0.16)
326
0.89
(6.7, 0.24)
176
1.00
(0.0, 0.00)
361
N
0.68
(7.9, 0.23)
334
0.83
(6.0, 0.21)
341
0.90
(4.6,0.19)
335
0.85
(4.7,0.19)
339
0.79
(6.3, 0.20)
347
0.85
(6.4, 0.20)
334
0.85
(6.3, 0.20)
335
0.83
(5.6, 0.21)
341
0.82
(7.2, 0.23)
356
0.84
(5.9, 0.20)
326
0.85
(6.5,0.19)
337
0.91
(5.9, 0.23)
181
0.94
(3.8, 0.13)
341
1.00
(0.0, 0.00)
362
0
0.73
(7.0, 0.22)
331
0.88
(4.7,0.18)
336
0.93
(3.8,0.18)
328
0.86
(4.2,0.18)
334
0.84
(4.8,0.18)
343
0.85
(7.5, 0.22)
330
0.87
(7.0, 0.22)
329
0.84
(6.8,0.21)
336
0.83
(8.2, 0.25)
357
0.80
(7.5, 0.22)
323
0.81
(8.2, 0.22)
332
0.90
(4.8,0.21)
177
0.90
(4.6,0.16)
336
0.90
(4.4,0.17)
346
1.00
(0.0, 0.00)
362
P
0.87
(5.3,0.18)
326
0.86
(5.6,0.19)
335
0.83
(5.9,0.21)
329
0.80
(6.2, 0.20)
342
0.91
(4.5,0.17)
343
0.89
(5.2,0.16)
336
0.88
(5.5,0.17)
383
0.89
(4.5,0.16)
335
0.88
(6.1,0.20)
957
0.90
(5.0,0.17)
321
0.89
(5.2,0.16)
331
0.68
(10.0,0.29)
181
0.83
(5.5,0.16)
345
0.77
(6.7,0.19)
347
0.80
(5.8,0.19)
348
1.00
(0.0, 0.00)
1027
Q
0.81
(7.2, 0.23)
292
0.80
(7.9, 0.26)
300
0.79
(7.8, 0.26)
290
0.75
(7.8, 0.25)
300
0.86
(6.3, 0.22)
306
0.86
(6.0,0.16)
295
0.86
(5.3,0.17)
296
0.86
(6.0,0.19)
299
0.84
(6.0, 0.16)
314
0.86
(5.9, 0.20)
283
0.86
(5.8, 0.18)
296
0.63
(12.1,0.35)
153
0.81
(7.4, 0.20)
300
0.75
(8.1,0.22)
309
0.75
(8.8, 0.25)
304
0.95
(3.6,0.14)
307
1.00
(0.0, 0.00)
321
R
0.85
(5.6, 0.20)
285
0.85
(5.7,0.21)
288
0.81
(6.2, 0.23)
281
0.79
(6.2,0.21)
287
0.88
(4.9, 0.18)
295
0.88
(4.9, 0.16)
281
0.87
(5.2, 0.17)
282
0.87
(4.5, 0.16)
289
0.85
(6.0,0.18)
306
0.87
(5.3, 0.17)
272
0.87
(5.5,0.16)
286
0.72
(9.1,0.30)
149
0.82
(5.8,0.17)
288
0.78
(6.4, 0.20)
297
0.79
(6.8,0.21)
292
0.97
(2.0, 0.09)
299
0.92
(3.1,0.13)
268
1.00
(0.0, 0.00)
313
S
0.86
(5.2,0.18)
306
0.85
(6.0,0.19)
314
0.82
(6.0,0.21)
309
0.80
(5.8, 0.20)
321
0.91
(3.9,0.17)
324
0.89
(5.5,0.17)
316
0.88
(5.7,0.19)
356
0.88
(5.1,0.17)
314
0.87
(6.3,0.21)
936
0.88
(5.2,0.18)
302
0.88
(5.5,0.18)
313
0.69
(9.8, 0.29)
164
0.84
(5.1,0.16)
326
0.78
(6.2,0.19)
327
0.80
(6.0,0.19)
330
0.97
(2.1,0.08)
943
0.94
(4.0,0.16)
290
0.94
(2.7,0.12)
280
1.00
(0.0, 0.00)
998
December 2009
3-64
-------
A
01
Pittsburgh Combined Statistical Area
Pittsburgh CSA
• PM2.5 Monitors
Interstate Highways
Major Highways
0 10 20 40 60 80
100
—i Kilometers
Figure 3-21. Locations of PM2.s monitors and major highways, Pittsburgh, PA.
December 2009
3-65
-------
Mean 15.1 19.8 13.2 13.6 15.1 16.4
Ots 1063 1066 306 165 332 337
SD 8.9 14.7 8.0 8.5 9.0 9.5
70-
AQS Site ID
Site A 42-003-0008
SteB 42-003-0064 60"
SiteC 42-003-0067
SiteD 42-003-0095 ~ 50_
SitEE 42-003-1008 -5
Cft
SteF 42-003-1301 .a
C 40 -
O
ro
Ł
3J
u
c
O
u
20 -
1 0 ™
1 ^winter
2=spring
3=summer
4=fall 0 -
1234 234 234 1234 1234 234
G H j K L
Mean 15.3 16.4 15.5 14.8 13.4 15.4
Obs 171 328 354 345 966 350
SD 8.3 9.3 8.6 8.0 8.6 8.7
70 -
AQS Site ID
SiteG 42-003-3007
SteH 42-007-0014 60 "
Sitel 42-125-0005
SiteJ 42-125-0200 ~ cn
F
SteK 42-125-5001 -C
cn
SteL 42-129-0008 ^
c 40 -
O
I 30-
u
c
O
u
20 -
1 -winter
3=summer
4=fall o -
1234 1234 1234 1234 1234 1234
Figure 3-22. Seasonal distribution of 24-h avg PM2.s concentrations by site for Pittsburgh, PA,
2005-2007. Box plots show the median and interquartile range with whiskers
extending to the 5th and 95th percentiles at each site during (1) winter
(December-February), (2) spring (March-May), (3) summer (June-August) and (4)
fall (September-November).
December 2009
3-66
-------
Table 3-12. Inter-sampler comparison statistics for each pair of 24-h PM26 monitors reporting to
AQS for Pittsburgh, PA.
ABCDEFGH I JKL
A 1.00 0.79 0.95 0.92 0.93 0.95 0.95 0.85 0.90 0.93 091 0.88
(0.0,0.00) (15.9,0.19) (5.6,0.13) (4.7,0.11) (4.7,0.11) (4.9,0.10) (3.8,0.10) (6.4,0.13) (6.4,0.13) (5.0,0.12) (6.0,0.13) (5.6,0.12)
1063 1035 298 164 323 329 170 319 344 337 934 340
_B 1.00 071 0.65 0.80 0.85 0.76 0.69 071 0.68 0.68 0.67
(0.0,0.00) (16.9,0.24) (17.4,0.25) (14.4,0.19) (12.5,0.14) (15.7,0.20) (17.0,0.19) (15.7,0.21) (17.8,0.23) (19.3,0.25) (15.9,0.21)
1066 303 165 329 335 171 324 350 341 938 346
_C 1.00 0.93 0.90 091 0.94 0.80 0.93 0.96 0.95 0.91
(0.0,000) (2.8,0.09) (6.6,0.16) (8.7,0.17) (6.0,014) (9.4,0.19) (6.7,0.15) (4.6,012) (4.5,0.10) (6.5,0.15)
306 144 282 282 148 268 290 286 270 286
_D 1.00 0.84 0.87 091 0.79 0.89 091 0.97 0.85
(0.0,0.00) (6.4,0.15) (8.5,0.16) (5.8,0.13) (9.2,0.17) (5.9,0.13) (4.6,011) (3.1,0.08) (6.5,0.15)
165 153 161 158 156 158 155 146 157
_E 1.00 0.90 0.90 0.84 0.85 0.86 0.88 0.83
(0.0,0.00) (6.4,0.13) (6.5,0.13) (6.8,0.14) (8.3,0.16) (7.7,0.16) (7.6,0.15) (7.3,0.15)
332 313 157 295 320 315 290 318
LEGEND
1.00 0.91 0.82 0.88 0.88 0.89 0.86
Yearson K (0.0,0.00) (6.7,0.13) (7.4,0.14) (7.1,0.15) (7.9,0.15) (8.8,0.17) (7.0,0.14)
(P90, COD) 337 yff 3Q2 327 319 296 322
n 1.00 0.78 0.94 0.93 0.90 0.91
(7.3,0.16) (4.0,0.10) (5.0,0.11) (6.6,0.15) (5.0,0.13)
159 163 159 149
1.00 0.80 0.78 0.82
(8.4,0.15) (8.2,0.17) (9.0,0.18) (9.2,0.18)
328 317 309
1.00 0.93
(5.0,011) (7.2,0.16)
354 334 310
1.00
(00,0.00) (5.5,0.12) (5.9,0.13)
345 302
(0.0, 0.00) (6.9, 0.15)
December 2009 3-67
-------
Los Angeles Core Based Statistical Area
A
Ql
r
\,
| Los Angeles CBSA
• PM2.5 Monitors
Interstate Highways
Major Highways
0 10 20 40 60 80
100
—i Kilometers
Figure 3-23. Locations of PM2.6 monitors and major highways, Los Angeles, CA.
December 2009
3-68
-------
SiteA
SteB
SiteC
SteD
SiteE
SiteF
SiteG
SiteH
Stel
SteJ
SteK
AQSSitelD
06-037-0002
06-037-1002
06-037-1103
06-037-1201
06-037-1301
06-037-2005
06-037^1002
06-037-4004
06-037-9033
06-059-0007
06-059-2022
1=winter
2=spring
3=summer
4=fall
ABCDEFGHI
Mean 16.1 17.0 16.7 13.3 16.7 14.3 14.7 14.2 8.2
Obs 862 308 1004 291 327 334 946 990 221
SD 10.8 10.2 9.8 7.5 9.3 8.9 8.4 7.7 3.8
60 -
50 -
IE
C
O
ro 30-
c
CJ
t=
O 20 -
1 0 -
o -
1 .
t ii
w
J K
14.4 10.9
999 318
8.5 6.4
!
I11
|1
1234 1234 1234 1234 1234 1234 1234 1234 1234 1234 1234
Figure 3-24. Seasonal distribution of 24-h avg PM2.s concentrations by site for Los Angeles,
CA, 2005-2007. Box plots show the median and interquartile range with whiskers
extending to the 5th and 95th percentiles at each site during (1) winter
(December-February), (2) spring (March-May), (3) summer (June-August) and (4)
fall (September-November).
Table 3-13. Inter-sampler comparison statistics for each pair of 24-h PM2.6 monitors reporting to
AQS for Los Angeles, CA.
A
A 1.00
(0.0, 0.00)
862
B
C
D
E
F
G
H
I
J
K
B
0.86
(9.0, 0.18)
252
1.00
(0.0, 0.00)
308
LEGEND
Pearson R
(P90, COD)
n
C D
0.87 0.81
(7.7,0.16) (9.0,0.19)
803 238
0.92 0.87
(5.5,0.11) (9.1,0.19)
293 250
1.00 0.80
(0.0, 0.00) (9.6, 0.20)
1004 274
1.00
(0.0, 0.00)
291
E
0.80
(9.7,0.21)
262
0.83
(9.0,0.15)
278
0.89
(5.8,0.11)
315
0.69
(10.9,0.23)
263
1.00
(0.0, 0.00)
327
F
0.88
(5.8,0.14)
269
0.88
(7.6,0.15)
279
0.92
(6.4,0.13)
319
0.77
(7.4,0.18)
263
0.79
(9.1,0.19)
301
1.00
(0.0, 0.00)
334
G
0.68
(11.5,0.22)
761
0.77
(9.8, 0.17)
268
0.84
(9.0, 0.15)
880
0.63
(11.3,0.22)
256
0.95
(5.9,0.11)
289
0.70
(10.5,0.18)
290
1.00
(0.0, 0.00)
946
H
0.64
(12.4,0.23)
793
0.73
(11.6,0.18)
282
0.79
(10.0,0.17)
913
0.60
(11.1,0.22)
268
0.92
(7.6, 0.13)
301
0.70
(9.2, 0.19)
302
0.96
(4.0, 0.09)
859
1.00
(0.0, 0.00)
990
I
0.30
(18.0,0.36)
179
0.31
(24.1,0.38)
177
0.29
(18.6,0.38)
213
0.41
(14.8,0.31)
164
0.34
(19.7,0.39)
192
0.33
(14.8,0.34)
184
0.23
(17.0,0.35)
194
0.26
(15.3,0.34)
208
1.00
(0.0, 0.00)
221
J
0.70
(10.5,0.21)
804
0.74
(11.9,0.19)
292
0.82
(9.4,0.16)
920
0.64
(9.6,0.21)
274
0.88
(8.2,0.15)
307
0.69
(9.8,0.19)
311
0.92
(5.4,0.12)
882
0.91
(5.9,0.12)
914
0.21
(18.3, 0.35)
205
1.00
(0.0, 0.00)
999
K
0.82
(11.4,0.23)
259
0.71
(15.0,0.27)
277
0.78
(13.2,0.25)
305
0.60
(11.6,0.23)
261
0.76
(13.7,0.27)
291
0.72
(9.9,0.21)
293
0.78
(11.0,0.21)
277
0.78
(9.5,0.21)
294
0.31
(9.7, 0.28)
180
0.84
(9.8, 0.19)
298
1.00
(0.0, 0.00)
318
To further investigate the relationship between correlation and distance, Figure 3-25 through
Figure 3-27 plot inter-sampler correlation as a function of distance between monitors for PM2.5 in
December 2009
3-69
-------
Boston, Pittsburgh, and Los Angeles. These three cities were selected to illustrate how this
relationship varies across urban areas with different topography and climatology as well as different
PM2.5 sources, compositions and monitor densities. Plots are provided in Annex A for all 15
CSAs/CBSAs under investigation beginning with Figure A-39. The Boston data exhibit the strongest
relationship between inter-sampler correlation and distance, with average inter-sampler correlation
remaining higher than 80% when samplers are 95 km apart (R2 = 0.55). This small amount of
variability is expected given the consistency between distributions shown in the corresponding box
plots (Figure 3-20). The Pittsburgh data show some reductions in inter-sampler correlations at short
distances, with the samplers at Sites B and G having only 76% correlation with a distance of less
than 4 km. Site B is located in Liberty, PA, a mountainous suburb of Pittsburgh where emissions
from steel manufacturing and frequent stable conditions in the planetary boundary layer cause
localized events of elevated concentration. In contrast, Site G is in the neighboring town of Clairton,
PA, located at a lower elevation on the bank of the opposite side of the Monongahela River from
Liberty. On average, inter-sampler correlation remained higher than 80% when samplers were
separated by 61 km, but in this case with much greater scatter (R2 = 0.22) than observed in the
Boston data. This scatter is driven by the measurements at Site B; Figure 3-22 shows an elevated
mean and variability for this site compared with other monitors situated around the Pittsburgh CSA.
When data from Site B are removed, the inter-sampler correlation vs. distance plot for Pittsburgh
PM2.5 resembles the one from Boston (with R2 increasing to 0.68). The Los Angeles data exhibit a
much steeper slope, with average inter-sampler correlation remaining higher than 80% when
samplers are 29 km apart (R2 = 0.74). This suggests that other factors, such as mountainous
topography separating monitors, the distribution of traffic, re suspension of crustal components, and
occurrence of stable boundary layers, may cause more spatial variation in the PM2.5 concentration
profile within the Los Angeles region when compared with other parts of the country. The Site I
monitor, separated from the rest of the Los Angeles region by the San Gabriel Mountains as
mentioned above, provides the low correlations grouped in the lower right portion of Figure 3-27. It
should also be noted in examining Figure 3-19, Figure 3-21, and Figure 3-23 that some monitors are
often located close to major interstate highways while others in the same urban area are not. These
differences in proximity of monitors to nearby major roads may also result in lower inter-monitor
correlations.
December 2009 3-70
-------
0.8
0.6
10 20 30 40 50 60
Distance Between Samplers (km)
70 80 90 100
Figure 3-25. Inter-sampler correlations for 24-h PM2.s as a function of distance between
monitors in Boston, MA.
**
0.8
0.6
«»»««» «
10 20 30 40 50 60
Distance Between Samplers (km)
70 80
100
Figure 3-26. Inter-sampler correlations for 24-h PM2.s as a function of distance between
monitors in Pittsburgh, PA.
December 2009
3-71
-------
0.8 -
0.6 -
g
8
0.4 -
0.2 -
10 20 30 40 50 60 70
Distance Between Samplers (km)
80
90
100
Figure 3-27. Inter-sampler correlations for 24-h PM2.s as a function of distance between
monitors in Los Angeles, CA.
PMlO-2.5
Given the limited number of co-located low-volume FRM PM10 and FRM PM2.5 monitors,
only a very limited investigation into the intra-urban spatial variability of PMi0_2.5 was possible using
AQS data. Of the 15 cities under investigation, only six (Atlanta, Boston, Chicago, Denver, New
York and Phoenix) contained data sufficient for calculating PMi0_2.5 according to the data
completeness and monitor specification requirements discussed earlier. Figure 3-28 contains box
plots of PMio_2.5 for one or two available sites per CSA/CBSA providing adequate PMi0_2.5
concentration data. For Boston, the correlation between the two sites for PM10_2.5 was 0.45 compared
with 0.73 for PM2 5 alone and 0.84 for PM10 alone (using the same two monitoring sites). For New
York, the correlation was slightly higher for the two sites: 0.74 for PMi0_2.5 compared with 0.93 for
PM2.5 alone and 0.82 for PMi0 alone. The COD for PMi0_2.5 also increases in both cities compared
with PM2 5 and PMi0 alone, suggesting less spatial homogeneity for thoracic coarse particles
compared with fine particles. Wilson and Suh (1997, 077408) reported PMi0_2.5 correlations between
eight sites in Philadelphia ranging from 0.14 to 0.63 with an average of 0.38. This was considerably
less than the corresponding average correlation for PM25 (r = 0.90) and PM10 (r = 0.87) from the
same study. Thornburg et al. (2009, 190999) also reported a high degree of spatial variability in
PMio_2.5 in Detroit with between-monitor correlations ranging from 0.03 to 0.76. These results
suggest that local sources can have a substantial impact on PMi0_2.5 concentrations, resulting in a
higher degree of spatial variability in PMi0_2.5 relative to PM25 or PMi0.
December 2009
3-72
-------
Atlanta
Boston
Chicago
Denver
New York
Phoenix
1234 123
Figure 3-28. Seasonal distribution of 24-h avg PMi0.2.s concentrations by site for Atlanta, GA;
Boston, MA; Chicago, IL; Denver, CO; New York City, NY; and Phoenix, AZ;
2005-2007. Box plots show the median and interquartile range with whiskers
extending to the 5th and 95th percentiles at each site during (1) winter
(December-February), (2) spring (March-May), (3) summer (June-August) and (4)
fall (September-November). Note the different concentration scales on the y-axes.
PM10
PM10 mass concentration has been shown to vary as much as a factor of five over urban-scale
distances of 100 km or less, and by a factor of 2 or more on scales as small as 30 km in an analysis
of California air quality (Alexis et al., 2001, 079886). This can be attributed to the rapid Vd and
resulting short atmospheric lifetime of the coarse-mode particles making up much of PMi0 mass. As
a result, local emission sources often dominate PMi0 annual average mass at certain monitors. Data
from the 15 CS As/CBS As were used to investigate urban variability in PMi0 reported to the AQS
database.
Maps of PMio monitor locations and box plots of seasonal PMi0 mass concentration data are
provided for Boston (Figure 3-29 and Figure 3-30), Pittsburgh (Figure 3-31 and Figure 3-32), and
Los Angeles (Figure 3-33 and Figure 3-34) similar to the PM2.5 maps and box plots shown earlier in
Figure 3-19 through Figure 3-24. Annex A, Figures A-82 through A-125 incorporate similar
information for all 15 CSAs/CBSAs. Table 3-14 through Table 3-16 contain pair wise, within-city
comparison statistics (R, P90, COD and n, as defined above) for PMi0 measured at the available
monitors in Boston (Table 3-14), Pittsburgh (Table 3-15) and Los Angeles (Table 3-16); all 15
CSAs/CBSAs are included in Annex A, Tables A-35 through A-49.
Boston is an example of a city with a wide range in concentrations measured at different sites.
Inter-monitor variation in PMi0 is frequently larger than the seasonal variation measured at any given
site. Pairwise correlations between monitors in Boston range from 0.45 to 0.95 in Table 3-14.
Pittsburgh is an example of a city with a large number of PMi0 monitors providing consistent values
with a select few reporting higher concentrations (sites D, H, I and K in Figure 3-32). This illustrates
the potential influence of localized point or area sources or topography. Correlations between
monitors in Pittsburgh range from 0.47 to 0.97 in Table 3-15. Los Angeles shows a high degree of
between-season and within-season variability, which is on the order of the between-monitor
variation. Correlations between monitors in Los Angeles range from 0.29 to 0.93 in Table 3-16. Once
again, the lowest correlations are with the monitor separated from the other monitors by the San
Gabriel Mountains (Site E in Figure 3-33).
December 2009
3-73
-------
Boston Combined Statistical Area
A
01
Boston CSA
• PMio Monitors
Interstate Highways
Major Highways
0 10 20 40 60 80
100
^Kilometers
Figure 3-29. Locations of PM10 monitors and major highways, Boston, MA.
December 2009
3-74
-------
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
SiteG
SiteH
AQS Site ID
25-025-0042
25-027-0023
33-011-0020
33-015-0014
44-003-0002
44-007-0022
44-007-0026
44-007-0027
1 ^winter
2=spring
3=summer
4=fall
ABCDEFGH
Mean 16.4 21,6 16.5 14.8 10.7 17.0 21.3 18.8
Obs 191 174 182 175 171 182 169 168
SD 7.7 11.9 9.2 8.2 7.0 8.7 10.8 9.4
60 -
50 -
40 -
30 -
20-
10-
o -
1
1
1234 1234 1234 1234 1234 1234 1234
234
Figure 3-30. Seasonal distribution of 24-h avg PMi0 concentrations by site for Boston, MA,
2005-2007. Box plots show the median and interquartile range with whiskers
extending to the 5th and 95th percentiles at each site during (1) winter
(December-February), (2) spring (March-May), (3) summer (June-August) and (4)
fall (September-November).
Table 3-14. Inter-sampler comparison statistics for each pair of 24-h PMio monitors reporting to AQS
for Boston, MA.
Site A B
A 1.00 0.69
(0.0,0.00) (15.0,0.22)
191 169
B 1.00
(0.0, 0.00)
174
C
D
E LEGEND
Pearson R
(P90, COD)
F n
G
C
0.69
(12.0,0.20)
179
0.66
(17.0,0.24)
167
1.00
(0.0, 0.00)
182
D
0.73
(10.0,0.22)
173
0.56
(19.0,0.28)
161
0.72
(10.0, 0.22)
170
1.00
(0.0, 0.00)
175
E
0.71
(13.0,0.30)
171
0.45
(24.0, 0.39)
158
0.47
(17.0,0.33)
168
0.63
(11.0,0.29)
163
1.00
(0.0, 0.00)
171
F
0.84
(8.0, 0.14)
182
0.69
(15.0,0.21)
169
0.62
(12.0,0.21)
179
0.68
(10.0,0.23)
173
0.84
(13.0,0.29)
171
1.00
(0.0, 0.00)
182
G
0.70
(150,0.20)
169
0.77
(12.0,0.17)
156
0.64
(16.0,0.26)
166
0.59
(19.0,0.30)
161
0.58
(22.0, 0.38)
161
0.81
(11.0,0.16)
169
1.00
(0.0, 0.00)
H
0.79
(10.0,0.17)
167
0.65
(16.0,0.20)
154
0.59
(16.0,0.24)
164
0.69
(13.0,0.26)
158
0.80
(15.0,0.33)
157
0.95
(5.0,0.11)
167
0.79
(10.0,0.13)
169 154
H
1.00
(0.0, 0.00)
168
December 2009
3-75
-------
A
01
Pittsburgh Combined Statistical Area
0 10 20
40 60
Pittsburgh CSA
PMio Monitors
Interstate Highways
Major Highways
80
100
m Kilometers
Figure 3-31. Locations of PM10 monitors and major highways, Pittsburgh, PA.
December 2009
3-76
-------
Mean 21.2 18.2 21.8 27.7 23.4 19.6 18.4 29.7 25.2
Obs 1077 1019 1083 1087 176 179 978 182 1022
SD 12.9 11.4 12.3 20.3 11. 9.9 11.7 16.8 19.3
90-
AQSStelD
SfcA 42-003-0002 80-
SitB 42-0034021
SiteC 42-OB-0031 70-
SiteD 42-003-0064 ~
SteE 42-003-0092 ^ 60"
Site F 42-003-0095 3
SiteG 42-003-0116 "^ 50"
SiteH 42-003-1301 ~
03 40-
c
S 30-
c
o
u 20-
10-
1=winter
2=spring 0 "
3=summer
4=fall _t o -
1
f
234 234 234 1234 234 234 234 234 1234
J K L M N O P Q
Wban 20.1 36,5 26.3 26.4 21,7 19.7 27.3 21.1
Obs 177 1061 1051 1069 1092 167 178 1079
SD 10.3 26,7 16.3 15,0 12,4 11.0 12.0 11,4
120-
AQS Site ID , , o .
Site I 42-003-3006
SiteJ 42-003-3007 10°"
SiteK 42-003-7004
--, 90 -
SteL 42-007-0014 "p
SiteM 42-073-0015 "g, 80-
SteN 42-125-0005 3
c 70-
SteO 42-125-5001 o
SiteP 42-1294007 S 60-
SteQ 42-129-0008 Ł
u
C
O 40-
30-
20-1
1=winter B
2=spring 1 0 - T
3 -summer
4=fall 0 -
ill
i
F ff
1234 1234 1234 1234 1234 123'
1234 1234
Figure 3-32. Seasonal distribution of 24-h avg PMi0 concentrations by site for Pittsburgh, PA,
2005-2007. Box plots show the median and interquartile range with whiskers
extending to the 5th and 95th percentiles at each site during (1) winter
(December-February), (2) spring (March-May), (3) summer (June-August) and (4)
fall (September-November).
December 2009
3-77
-------
Table 3-15. Inter-sampler comparison statistics for each pair of 24-h PMio monitors reporting to AQS
for Pittsburgh, PA.
Site ABODE FGH I
A 1.00 0.93 0.93 0.80 0.92 0.89 0.93 0.79 0.86
(8.0,0.14) (23.0, 0.21) (8.0,0.12) (14.0,0.18) (8.0,0.14) (16.0,0.17) (18.0,0.18)
1077 1002 1065 1070 175 178 960 181 1005
1.00 0.96 0.80 091 0.92 0.97 081 0.89
(0.0, 0.00) (8.0,0.15) (29.0, 0.24) (11.0,0.20) (6.0,0.16) (5.0,0.10) (25.0, 0.29) (22.0, 0.20)
1019 1007 1012 163 166 911 169 954
1.00 081 0.94 0.93 0.94 0.77 0.87
(0.0, 0.00) (23.0, 0.20) (6.0,0.11) (7.0,0.12) (8.0,0.13) (21.0,0.22) (19.0,0.17)
1083 1075 173 176 966 179 1010
1.00 0.72 0.66 0.76 0.83 0.88
(0.0, 0.00) (21.0,0.20) (26.0, 0.24) (27.0, 0.24) (14.0,0.18) (16.0,0.14)
1087 176 179 970 182 1014
1.00 0.90 0.90 0.78 0.77
(0.0, 0.00) (10.0,0.14) (10.0,0.17) (20.0, 0.20) (20.0,0.19)
LEGEND 176 173 154 175 166
Pearson R 1.00 0.94 0.70 0.74
(P90, COD) (0.0, 0.00) (7.0,0.12) (25.0, 0.27) (25.0, 0.22)
n 179 157 178 168
1.00 0.70 0.87
(22.0, 0.28) (20.0,0.19)
978 160 910
0.76
(0.0, 0.00) (17.0,0.20)
182 171
1.00
1022
December 2009 3-78
-------
J
A 0.84
(14.0,0.20)
176
B 0.93
(7.0,0.16)
164
C 0.90
(8.0,0.13)
174
D 0.73
(24.0, 0.22)
177
E 0.86
(10.0,0.16)
171
F 0.90
(7.0,0.12)
174
G 0.92
(7.0,0.13)
156
H 0.74
(23.0, 0.26)
176
I 0.79
(22.0, 0.20)
166
J 1.00
(0.0, 0.00)
177
K
L
M
N
0
P
Q
K
0.76
(40.0, 0.30)
1044
0.76
(43.0, 0.36)
986
0.75
(39.0, 0.30)
1049
0.84
(24.0, 0.22)
1055
0.65
(36.0, 0.29)
169
0.57
(41.0,0.34)
172
0.73
(45.0, 0.35)
955
0.68
(26.0, 0.22)
175
0.83
(30.0, 0.25)
992
0.66
(44.5, 0.33)
170
1.00
(0.0, 0.00)
1061
L
0.88
(15.0,0.18)
1033
0.88
(19.0, 0.23)
982
0.88
(14.0,0.17)
1039
0.80
(20.0,0.18)
1043
0.83
(16.0,0.16)
169
0.82
(20.0, 0.20)
172
0.87
(18.0,0.21)
938
0.77
(15.0,0.18)
175
0.82
(16.0,0.17)
978
0.79
(18.0,0.20)
170
0.74
(31.0,0.26)
1017
1.00
(0.0, 0.00)
1051
M
0.85
(16.0,0.19)
1052
0.81
(20.0, 0.26)
994
0.83
(15.0,0.19)
1057
0.78
(20.0, 0.20)
1061
0.80
(14.0,0.17)
172
0.75
(19.0,0.22)
175
0.78
(19.0,0.24)
952
0.78
(17.0,0.18)
178
0.78
(18.0,0.20)
998
0.72
(18.0,0.22)
173
0.75
(33.0, 0.24)
1035
0.87
(13.0,0.16)
1025
1.00
(0.0, 0.00)
1069
N
0.86
(11.0,0.16)
1074
0.91
(10.0,0.16)
1016
0.89
(9.0,0.12)
1080
0.76
(25.0, 0.20)
1084
0.84
(12.0,0.14)
176
0.86
(11.0,0.14)
179
0.89
(9.0,0.15)
975
0.74
(21.0,0.22)
182
0.81
(20.0, 0.17)
1019
0.88
(8.0,0.13)
177
0.70
(40.0, 0.30)
1058
0.85
(16.0,0.17)
1048
0.74
(18.0,0.21)
1067
1.00
(0.0, 0.00)
1092
0
0.77
(16.0,0.22)
166
0.76
(12.0,0.19)
157
0.78
(12.0,0.18)
164
0.57
(28.0, 0.26)
167
0.77
(14.0,0.19)
161
0.83
(9.0,0.15)
164
0.81
(11.0,0.17)
146
0.60
(27.0, 0.29)
167
0.66
(26.0, 0.24)
158
0.78
(11.0,0.17)
163
0.47
(44.0, 0.36)
160
0.70
(22.0, 0.24)
160
0.64
(19.0,0.26)
163
0.72
(13.0,0.18)
167
1.00
(0.0, 0.00)
167
P
0.78
(15.0,0.19)
177
0.83
(18.0,0.28)
165
0.88
(13.0,0.19)
175
0.64
(20.0, 0.25)
178
0.84
(13.0,0.16)
172
0.84
(16.0,0.22)
175
0.84
(17.0,0.26)
157
0.65
(19.0,0.22)
177
0.69
(21.0,0.25)
167
0.86
(16.0,0.21)
173
0.58
(34.0, 0.30)
171
0.74
(17.0,0.21)
171
0.67
(17.0,0.22)
174
0.86
(14.0,0.20)
178
0.75
(18.0,0.25)
163
1.00
(0.0, 0.00)
178
Q
0.86
(11.0,0.15)
1061
0.88
(10.0,0.18)
1003
0.90
(9.0,0.12)
1067
0.74
(26.0,0.21)
1071
0.85
(11.0,0.15)
174
0.86
(9.0,0.14)
177
0.86
(10.0,0.16)
967
0.76
(21.5,0.24)
180
0.78
(22.0, 0.19)
1009
0.86
(8.0,0.15)
175
0.68
(43.0, 0.30)
1048
0.80
(18.0,0.19)
1035
0.77
(18.0,0.19)
1053
0.86
(10.0,0.14)
1076
0.69
(14.0,0.19)
165
0.84
(15.0,0.21)
176
1.00
(0.0, 0.00)
1079
December 2009 3-79
-------
Los Angeles Core Based Statistical Area
01
r
Los Angeles CBSA
• PMio Monitors
Interstate Highways
Major Highways
0 10 20 40 60
80
100
—i Kilometers
Figure 3-33. Locations of PM10 monitors and major highways, Los Angeles, CA.
December 2009
3-80
-------
AQS Site ID
Site A 06-037-0002
SiteB 06-037-1103
SiteC 06-037-4002
SiteD 06-037-6012
Site E 06-037-9033
Site F 06-059-0007
SiteG 06-059-2022
Mean 35.3
Obs 169
SD 19.8
c
o
90 -
80 -
70-
60-
50 -
C 40-
01
u
030-
20 -
1=winter
2=spring
3=summer
4=fall
1 0 -
0 -
31.1
175
13.3
C
31.5
178
19.6
D
27.3
176
18.1
E
23.7
985
12.1
F
33.5
175
37.6
G
21.6
162
9.4
1234 1234 1234 1234 1234 12
34 1234
Figure 3-34. Seasonal distribution of 24-h avg PM10 concentrations by site for Los Angeles,
CA, 2005-2007. Box plots show the median and interquartile range with whiskers
extending to the 5th and 95th percentiles at each site during (1) winter
(December-February), (2) spring (March-May), (3) summer (June-August) and (4)
fall (September-November).
Table 3-16. Inter-sampler comparison statistics for each pair of 24-h PM10 monitors reporting to AQS
for Los Angeles, CA.
Site A
A 1.00
(0.0, 0.00)
169
B
C
LEGEND
Pearson R
D (P90, COD)
n
E
F
G
B C
0.73 0.44
(17.0,0.17) (27.0,0.24)
153 154
1.00 0.61
(0.0,0.00) (14.0,0.14)
175 159
1.00
(0.0, 0.00)
178
D
0.73
(24.0, 0.22)
157
0.57
(21.0,0.24)
159
0.65
(27.0, 0.28)
158
1.00
(0.0, 0.00)
176
E
0.47
(28.0, 0.26)
169
0.52
(23.0, 0.23)
173
0.43
(22.0, 0.24)
176
0.70
(16.0, 0.20)
175
1.00
(0.0, 0.00)
985
F
0.41
(29.0, 0.24)
155
0.42
(15.0,0.16)
162
0.93
(11.0,0.11)
159
0.65
(26.0, 0.28)
161
0.29
(26.0, 0.25)
173
1.00
(0.0, 0.00)
175
G
0.65
(30.0, 0.28)
143
0.73
(20.0, 0.23)
149
0.73
(21.0,0.22)
148
0.57
(19.5,0.24)
150
0.38
(20.0, 0.24)
159
0.65
(21.5,0.22)
150
1.00
(0.0, 0.00)
162
December 2009
3-81
-------
Figure 3-35 through Figure 3-37 illustrate the relationship between inter-sampler correlation
and distance between sites for PMi0 measurements obtained in Boston, Pittsburgh and Los Angeles.
Annex A contains similar plots for all 15 CS As/CBS As under investigation beginning with Figure A-
84. In each plot, substantially more scatter is observed when compared to those for PM2.5 (Figure
3-25 through Figure 3-27). This is consistent with the variability observed in the seasonal box plots
of concentration shown in Figure 3-30, Figure 3-32, and Figure 3-34. The Boston data exhibit the
strongest relationship between inter-sampler correlation and distance, with average inter-sampler
correlation remaining higher than 80% when samplers are 44 km apart (R2 = 0.61). The lowest
correlations on this plot originate from comparisons between Site B (rural Worcester, MA) and
samplers located at Sites E (West Greenwich, RI) and G (Providence, RI). Boston is subject to long
i0
range transport of SO42~, which is a regional pollutant and is a major component of PM2.5 and PM
in the eastern U.S. The Pittsburgh data shows some lower inter-sampler correlations, with one
sampler pair having only 66% correlation within a distance of 2 km. On average, inter-sampler
correlation remained higher than 80% when samplers were also separated by 44 km, but in this case
with much greater scatter (R2 = 0.28) than observed in the Boston data. As seen for the Pittsburgh
PMio box plots in Figure 3-32, sites D, H, I, and K have elevated means and high variability that is
driving the observed scatter. These four sites are all located in mountainous suburbs of Pittsburgh
(North Braddock, PA, Liberty, PA, Lincoln Boro, PA, and Beaver Falls, PA, respectively), where
emissions from steel manufacturing and frequent stable conditions in the planetary boundary layer
cause localized events of elevated concentration. When those four sites are removed, scatter
decreases greatly (R2 = 0.56). The Los Angeles data exhibit a much steeper slope, with average
inter-sampler correlation remaining higher than 80% when samplers are only 30 km apart
(R2 = 0.56). The lower inter-sampler correlations in part reflect the fact that some of these
monitoring sites are separated from each other by hills or, in the case of one sited at Lancaster, CA
(Site I), by the San Gabriel Mountains. The Los Angeles data exhibit greater scatter than the
Pittsburgh data. However, the smallest inter-sampler separation distance is 23 km, and there are
relatively fewer PMi0 samplers. Given the present data, it is not possible to judge how data would
correlate on smaller spatial scales.
0.8
0.6
0.4
0.2
10
20
30 40 50 60 70
Distance Between Samplers (km)
80
90
100
Figure 3-35. Inter-sampler correlations for 24-h PM10 as a function of distance between
monitors in Boston, MA.
December 2009
3-82
-------
0.8
0.6
0.4
0.2
V - ** *
0 10 20
30 40 50 60 70
Distance Between Samplers (km)
80 90 100
Figure 3-36. Inter-sampler correlations for 24-h PMi0 as a function of distance between
monitors in Pittsburgh, PA.
0.8
0.6
0.4
0.2
4> 4>
4> 4>
0 10 20
30 40 50 60 70
Distance Between Samplers (km)
80 90 100
Figure 3-37. Inter-sampler correlations for 24-h PMio as a function of distance between
monitors in Los Angeles, CA.
December 2009
3-83
-------
UFPs
Relatively few studies compare UFP measurements at multiple locations within an urban
center. An early study by Buzorius et al. (1999, 081205) suggested spatial homogeneity in total
particle number concentrations between multiple locations in Helsinki, Finland. They found
correlations in 10-min average at three sites within the city as high as 0.84. The sites, however, were
relatively close together (2 km) and all near the same roadway. There was a high degree of
correlation between traffic intensity and total aerosol number concentrations, suggesting that traffic
was the primary source of the measured particles and the driving force behind the high correlations.
Weekend correlations (0.28-0.47) and correlations with a fourth monitor located 22 km outside the
city (0.05-0.64) were much lower.
Tuch et al. (2006, 157060) found more spatial heterogeneity in UFP concentrations measured
for an entire year at two locations 1.5 km apart in Leipzig, Germany. Figure 3-38 shows the
correlation as a function of particle size (mobility diameter) dropping off as the particle size
decreases from 0.5 at 100 nm down to 0.2 at 3 nm. Table A-50 in Annex A contains correlation
coefficients of hourly and daily average particle number, surface area and volume concentrations as a
function of particle diameter adapted from the Tuch et al. (2006, 157060) study. For all days
(N = 5481 hourly observations), the correlation between UFPs (10-100 nm) measured at the two
sites was 0.31.
1.0
c
•8 0.6 H
1
o
o
c 0-6 -
0.4
O
c
05
c5
8.
w
0.2 -
0.0
10 100
diameter [nm]
Source: Reprinted with Permission of Nature Publishing Group from Tuch et al. (2006,1570601
Figure 3-38. Bin-wise Spearman correlation coefficients in aerosol particle number
concentrations between the Ift (urban background) and the Eisenbahn-strasse
(city/urban center) sites in Leipzig, Germany.
The two sites represented in Figure 3-38 and Table A-44 were relatively close to each other,
but one was located in a mixed semi-industrial region while the other was in a street canyon in a
residential neighborhood near busy roadways. This suggests a high degree of spatial heterogeneity in
UFPs driven primarily by differences in nearby source characteristics. Sioutas et al. (2005, 088428)
reviewed studies of the distribution of UFPs and came to the similar conclusion that mobile sources
make a large contribution to UFPs and therefore UFP concentrations can exhibit substantial
variability in space and time. This is to be expected since UFP concentrations drop off much quicker
with distance from roadways than larger particle sizes (Levy et al., 2003, 052661; Reponen et al.,
2003, 088425; Zhu et al., 2005, 157191). Hagler et al. (2009, 191185) showed similar exponential
decreases in UFP number concentrations with distance from the road for multiple locations in the
U.S. Neighborhood-scale variability and near-roadway concentration gradients for UFPs are
discussed further in Section 3.5.1.3.
December 2009
3-84
-------
PM Constituents
The pie charts showing PM2.5 composition that were generated using the SANDWICH method
for the 15 CSAs/CBSAs presented earlier in Figure 3-17 and Figure 3-18 represent the average of all
available monitors within each region. Individual pie charts for each monitor are included in Figures
A-127 through A-141 in Annex A and provide an indication of the urban-scale spatial variability in
PM2.5 composition. In the instances where multiple monitors were available, there was a fair degree
of spatial homogeneity in PM2.5 bulk chemistry within each metropolitan area. Some notable
exceptions exist, however. Birmingham and Detroit show variation in the amount of crustal material,
both spatially and seasonally. Denver exhibits some spatial variation in NO3~ during the winter, the
season with the highest measured PM2 5 mass. Several sites in New York and one in Pittsburgh have
elevated fractions of EC relative to the other sites within the respective cities, and several sites in
New York have been shown to have elevated Ni concentrations in PM2 5 samples when compared
with surrounding areas (Peltier and Lippmann, 2009, 197455: Peltier et al., 2008, 197452). In
Phoenix, high winter PM2 5 mass is site specific and appears to be associated with high OC; the
crustal component also varies and is inversely proportional to total measured mass.
3.5.1.3. Neighborhood-Scale Variability
Neighborhood scale spatial variability in the particle concentration profile is affected by land
and building topography, meteorology, particle size distribution, particle composition, and particle
volatility. Population density at the neighborhood scale is also an important determinant of the
spatial distribution of PM concentration because population density impacts source prevalence,
source magnitude, topographical-driven ventilation, and heat island effects (Crist et al., 2008,
156372: Makar et al., 2006, 155959: Mfula et al., 2005, 123359: Rigby and Toumi, 2008, 156050).
A number of computational and wind tunnel modeling street canyon studies have
demonstrated the potential variability in pollutant concentrations within a street canyon (Borrego et
al., 2006, 155697: Chang and Meroney, 2003, 090298: Kastner-Klein and Plate, 1999, 001961: So
et al., 2005, 110746: Xiaomin et al., 2006, 156165). Influential parameters include street canyon
height to width ratio (H/W), source positioning, wind speed and direction, building shape and
upstream configuration of buildings. Figure 3-39 shows pollutant concentrations obtained from wind
tunnel and computational fluid dynamics simulations of transport and dispersion in an infinitely long
street canyon with a line source centered at the bottom of the canyon (Xiaomin et al., 2006, 156165).
When the canyon height was equal to the street width (typical of moderate density suburban or urban
fringe residential neighborhoods) and lower background wind speed existed, concentrations on the
leeward canyon wall were four times those of the windward wall near ground level. When the
canyon height was twice the street width (typical of higher-density urban planning) and background
winds were somewhat higher, near ground level concentrations on the windward canyon wall were
roughly three times higher than those measured at the leeward wall. Baldauf et al. (2008, 191017:
2009, 191766) noted that the presence of noise barriers, vegetation, or changes in topography
adjacent to the road can also alter particle dispersion characteristics. Specifically, depressed road
segments, where the road bed is below the surrounding terrain, leads to increased air turbulence and
pollutant dispersion. These results suggest that micro- and neighborhood-scale variation related to
urban topography may have a significant impact on pollutant concentrations at this scale.
December 2009 3-85
-------
windward, |
measured
windward, _ _
simulated
I leeward,
measured
- - leeward,
simulated
1.2
0.8
0,4
0.2
1.2
to)
20 40 60 80
Dimensionless concentration
100
0.8
0.6
0.4
0.2
0
0 100 200 300 400 500 600
-------
1.00
0.90
0.80
0.70
0.60
c
•3 0.50
o
o
0.40
0.30
0.20
0.10
0.00
•o- . . o
00 b- - .
RPM25 = -0.0163D+ 10
O R2 = 0.2178
RPM10 = -0.0748D
R2 = 0.0346
. O
o o
<
•o- •
0.5
1.5 2 2.5
Distance Between Samplers (km)
3.5
Figure 3-40. Inter-sampler correlations for 24-h PM2.s and PMi0 as a function of distance
between monitors for samplers located within 4 km (neighborhood scale).
Isakov et al. (2007, 156588) compared PM2.s concentrations from a central monitoring site in
Wilmington, DE with PM0.3 measurements taken on a mobile platform driven through mostly quite
residential streets within a 4 km><4 km grid containing the central monitor. Correlations were
generally high (average r = 0.87) over all time periods and locations monitored, consistent with the
range of correlations for PM2.5 shown in Figure 3-40.
PMlO-2.5
Neighborhood-scale variability in PMi0_2.s was investigated by Chen et al. (2007, 147318) in
the Raleigh/Durham area of NC. The average correlation between 26 residential monitors located
throughout the region and a centrally located monitor representing a maximum inter-sampler range
of 60 km was found to be 0.75 for PMi0_2.5 compared with 0.92 and 0.94 for PM2.5 and PMio,
respectively. Based on this study, neighborhood-scale variability is greater for PMi0_2.5 than for PM2.5
or PM10, matching the conclusion drawn above on the broader urban-scale.
UFPs
Moore et al. (2009, 191004) monitored UFP concentrations throughout the Ports of Los
Angeles and Long Beach, through which an interstate highway runs and found that concentrations
varied by a factor of 5-7 across sites with substantial differences in the daily concentration time
series at each site. Such variability reflects diversity of the sources (some near the Interstate, some
near the Port), and the influence of changing meteorology over an urban area. In a mobile platform
sampling study, Westerdahl et al. (2005, 086502) and Fruin et al. (2008, 097183) also reported
substantial peaks in UFP concentration when sampling at highways in comparison with a
background site (the University of Southern California) using the same data set.
Near roadway environments can exhibit high concentration gradients, particularly for UFPs.
Ntziachristos et al. (2007, 089164) observed that the near-road particle size distribution was
December 2009
3-87
-------
substantially higher in the UF mobility diameter range and that these results were very sensitive to
meteorology (rain) and time of day. Baldauf et al. (2008, 190239) reported elevated UFP number
concentrations downwind of a highway in Raleigh, NC, when compared to measurements
approximately 100 m upwind of the road. Hagler et al. (2009, 191185) noted a 5-12% decrease in
number concentrations per 10m distance from the road for a number of studies in the U.S. with
unobstructed air flow.
After initial emission from a motor vehicle, the evolution of the PM distribution within the
plume is a function of (1) the turbulence that dilutes the plume and (2) evaporation or condensation
of the volatile portion of the aerosol that results from rapid cooling of the exhaust. Figure 3-41
shows the size distribution measured by Zhu et al. (2002, 041553) at distances of 17-300 m away
from the roadway (in this case, Highway 710 in Los Angeles) and at an upwind site. It can be seen
that a mode originally measured around 9 nm increases in diameter and decreases in number
concentration as distance from the highway increases. Smaller secondary modes appear around 30 m
from the roadway with multiple modes at some particle sizes. By 150 m away from the highway, the
size distribution flattens with a small mode around 50 nm. It is clear from the bottom figure that the
number concentration of larger particles (i.e., 100-220 nm) does not vary as much as UFPs (<100
nm) with increasing distance downwind from the roadway.
10 100
Particle Diameter, Dp (nm)
6-25 nm
25-50 nm
50-100 nm
V ----- V --
100-220 nm
0 100 200 300
Distance down wind from the 710 freeway (m)
Source: Reprinted with Permission of Elsevier Ltd. From Zhu et al. (2002, 0415531.
Figure 3-41. Particle size distributions measured at various distances from the 710 freeway in
Los Angeles, CA (top), and particle number concentration as a function of
distance from the 710 freeway (bottom).
December 2009
3-88
-------
Zhou and Levy (2007, 098633) performed a meta-analysis of traffic-related air pollution
literature and found that background pollution and meteorology can have important impacts on the
size of the elevated concentration region around the highway. Zhu et al. (2002, 041553) and Zhang et
al. (2005, 157185) noted in field measurements of UFPs that small particles can be lost due to
evaporation or to coagulation during Brownian diffusion to form bigger particles, resulting in an
upward shift in mode diameter with distance from the roadway. Studies of particle sizes on roads
(Kittelson et al., 2006, 156649; Kittelson et al., 2006, 156648). in tunnels (Venkataraman et al.,
1994, 002475). and upwind and downwind of roads (Zhu et al., 2002, 041553) suggest that for well-
maintained spark-ignition vehicles, a large fraction of the mass of particles emitted from the vehicles
are in the nuclei mode (i.e., smaller than accumulation mode). High-speed highway driving may be
associated with a larger fraction of particle mass being emitted in the UF size range, while lower
speed operation results in a higher mass fraction in the accumulation mode (Cadle et al., 2001,
017192). In situations in which the dilution rates are lower than in a short tunnel or downwind of a
road way, condensation of vapors can give rise to particles in the accumulation mode (Kittelson,
1998, 051098). Diesel engines, in particular, emit black carbon in the lower end of the accumulation
mode, with number emissions dominated by semi-volatile material in the nuclei mode (Kittelson,
1998, 051098: Kittelson et al., 2006, 156649: Kittelson et al., 2006, 156648). Sharp gradients in
black carbon mass have been observed along roadways with high diesel traffic (Zhu et al., 2002,
041553). As the traffic pollution moves downwind, the UFPs may grow into the accumulation mode
by coagulation or condensation. In addition to Gaussian dispersion and wind eddies caused by the
presence of natural and anthropogenic barriers, Sahlodin et al. (2007, 114058) demonstrated that
turbulence produced by vehicles can result in modification of the plume emanating from the
highway. Hence, on-road turbulence could potentially alter the aerosol size distribution. This added
turbulence could cause some evaporation of tiny nucleation particles that have not absorbed or
adsorbed onto soot nuclei, which may affect the rate of coagulation (Jacobson et al., 2005, 191187).
The roadway configuration may also affect particle transport and dispersion. Depressed road
sections, where the road bed is below the surrounding terrain, leads to increased air turbulence and
mixing as air flows up and out of the road depression. This configuration can result in lower
particulate concentrations and flatten concentration decay curves away from the road. On the other
hand, configurations with the road bed at-grade with surrounding terrain, or elevated above the
surrounding terrain with solid fill material resulted in the highest pollutant concentrations and
sharpest concentration gradients downwind from the road.
PM Constituents
The composition of PM will also vary on the neighborhood-scale in response to local sources
and differential dispersion, resulting in variable spatial distribution of individual components.
Krudysz et al. (2008, 190064) investigated spatial variation in size-fractionated (<0.25 (im,
0.25-2.5 (im, >2.5 (im) PM composition data at four sites located within 3-6 km of each other in the
Long Beach, CA area. Inter-site R2 values in the 0.25-2.5 (im size range were higher for mass
(ranging from 0.56-0.91) than for EC (0.02-0.71) for pair wise site comparisons. Spatial
heterogeneity in all size ranges investigated was also found for several elements associated with
motor vehicle emissions and resuspended road dust including Cu, Mg, Ba, Ca and Al. Viana et al.
(2008, 156135) observed higher concentrations of crustal elements in PM2.5 and PMi0 samples in
rural neighborhoods and higher concentrations of combustion-derived PM2.5 and PMi0, such as EC
and NO3~, in higher density urban areas. Gutierrez-Daban et al. (2005, 155818) examined the mass
distribution of various PAHs under different traffic and urban density conditions. Figure 3-42
displays the distributions for benz[a]pyrene (BaP) at high and low traffic sites at the urban center,
periphery, and industrial areas in Seville, Spain (Gutierrez-Daban et al., 2005, 155818).
Concentrations were nearly an order of magnitude lower for the low traffic urban periphery location
when compared with the high traffic or industrial locations. Particles smaller than ~ 600 nm had
roughly an order of magnitude higher concentration than those at larger sizes and tended to have a
larger spread in concentrations among sampling sites. Figure 3-43 shows the distributions for sixteen
PAHs at a high traffic location at the city center in Seville, Spain (Gutierrez-Daban et al., 2005,
155818). PAH species varied in concentration by up to two orders of magnitude for each particle size
bin, and the highest concentrations of individual PAHs were generally found for particles smaller
than approximately 600 nm. Olson and McDow (2009, 191188) reported decreases by a factor of
1.04-2.37 in select PAH and organic source marker concentrations when comparing measurements
10m and 275 m from a highway in Raleigh, North Carolina. Phuleria et al. (2006, 156867) sampled
December 2009 3-89
-------
UFPs and PM2.5 concentrations and PAH species at the mouth of the Caldecott Tunnel in Orinda, CA
and found that the two size classes were highly correlated (R2 = 0.97). Given the size differentials of
each size bin presented in the Gutierrez-Daban et al. (2005, 155818) study, it is possible that the
PM2.5 sampled at the tunnel mouth in the latter study represented secondary PM2.5 that grew from
UFP emissions trapped within the tunnel.
10000
1000
100
10
• HTC
• HTP
ALTC
XLTP
XLTIP
<0.6
1.3-0.6
2.7-1.3 4.9-2.7 10-4.9 >10
Size bin (urn)
Source: Adapted with Permission of Springer-Verlag from Gutierrez-Daban et al. (2005,155818).
Figure 3-42.
Mass distributions for BaP at a high traffic urban center (HTC), high traffic urban
periphery (HTP), low traffic urban center (LTC), low traffic urban periphery (LTP),
and low traffic industrial urban periphery (LTIP) in Seville, Spain.
December 2009
3-90
-------
10000 T
1000
10
»Naph
• Ace
AAcey
xFlu
xPhen
• Ant
+ Flua
-Pyr
-BaA
oChry
DBbF
ABkF
BaP
InP
ODbA
Bper
0.6
1.3-0.6
2.7-1.3 4.9-2.7
Size bin (urn)
10-4.9
Source: Adapted with Permission of Springer-Verlag from Gutierrez-Daban et al. (2005,155818).
Figure 3-43. Mass distributions for 16 PAHs at a high traffic city center in Seville, Spain.
3.5.2. Temporal Variability
Temporal variability is another important factor in characterizing PM. This section addresses
trends as well as seasonal and hourly variability. Trends in PM2.5 and PM10 are addressed in
Section 3.5.2.1 based on AQS data. Seasonality is coupled with spatial variability and has been
discussed in the regional context above. Section 3.5.2.2 below briefly investigates the seasonality on
a finer time scale, thereby addressing issues relating to the seasonal definitions used earlier.
Section 0 addresses hourly patterns, an issue particularly important to understanding the behavior of
PM concentrations in reference to sources, human activity patterns and exposure. Hourly patterns are
investigated using AQS data on a national basis for PM2.5 and PM10. Data for UFPs and PM
constituents are presented where available.
3.5.2.1. Regional Trends
This section summarizes available information on trends in PM mass and composition. Mass
concentration trends are based on AQS data and incorporate 9 years (1999-2007) of PM2.5 data and
20 years (1988-2007) of PMi0 data. Composition trends are based on six years of available CSN data
(2002-2007). Several monitoring sites were excluded from the following trend analyses to provide a
consistent basis for comparison over the desired years of monitoring. This included exclusion of sites
when there was no corresponding site in later or earlier years. Region-average trends were calculated
to facilitate presentation and extrapolation of the results. These region-averages, however, may not
necessarily represent the trends that are being observed at any individual monitor or geographical
location within the specified region.
December 2009
3-91
-------
PM2.5
Figure 3-44 shows trends in U.S. ambient 24-h PM2.5 concentrations from 1999-2007. In the
period 2005-2007, the 3-yr avg of the 98th percentile of 24-h PM2.5 concentrations fell 10% relative
to the 1999-2001 period (see Figure 3-44A). The number of sites reporting values greater than the
24-h NAAQS was shown to decline 40% in Figure 3-44B. Figure 3-44C illustrates the downward
trend in the 98th percentile of 24-h PM2.5 concentrations for three consecutive calendar years in all
U.S. EPA regions. This trend is most pronounced in Region 9 incorporating Arizona, California and
Nevada where this value dropped 25% from the 1999-2001 period to the 2005-2007 period.
60
A) Ambient Concentrations
5 t
s-s
45
30
T5 ^±:
• I»
'§
5 15
0
90% of sites have
concentrations below this line
NAAQS = 35 jg/m:
10% of sites have
concentrations below this line
'99-'01 '00'02 '01-'03 '02-'04 '03-'05 '04-'06 '05-'07
Averaging period
400
B) Number of Trend Sites Above NAAQS
OJ
a,
8300
- .
i a
Ji
ro if:
s ° Ł
'? O) —
^ a> *r
SŁ^
200
o J> 100
o S
•99-01 '00-'02 '01-'03 '02-'04 '03-'05 '04-'06 '05-'07
Averaging period
C) Trends by EPA Region3
II
ro -—-
IS 45
CJ d>
s.
O. Oi
fee o
NAAQS . 35 ug/m3
-R1
-R2
-R3
-R4
-R5
-R6
R7
R8
-R9
-R10
-Nat'l
'99-'01 'OG-'02 '01-'03 '02-P04 W05 '04-'06 '05-07
Averaging period
'Coverage: 697 monitoring sites
in the EPA Regions (out of a total
of 831 sites measuring PM2.5 in
2007) that have sufficient data to
assess PM25 trends since 1999.
EPA Regions
© O °
O °0
Source: U.S. EPA (2008, 157076)
Figure 3-44. Ambient 24-h PM2.s concentrations in the U.S., 1999-2007, showing A) ambient
concentrations, B) number of trends sites above the 24-h NAAQS and C) trends
by U.S. EPA Region.
Figure 3-45 contains similar trend information for the annual PM2.5 NAAQS. The seasonally
weighted 3-yr avg PM2 5 concentrations for the years 2005-2007 were at the lowest since national
monitoring began in 1999 (Figure 3-45A). The seasonally weighted 3-yr avg fell 10% between the
1999-2001 averaging period and the 2005-2007 averaging period. The number of sites reporting
concentrations above the annual average PM2 5 NAAQS fell 56% over these same periods in Figure
3-45B. Figure 3-45C illustrates the annual trends in PM25 by U.S. EPA region. Declines were the
greatest in Region 9 again where annual PM25 concentrations fell 20% from the 1999-2001
averaging period to the 2005-2007 averaging period.
December 2009
3-92
-------
S1 s
S --e
20
A) Ambient Concentrations
03 ^
!jk
•5 s E
90% of sites have concentrations below this line
_NA_A_a_S_=_1_5_ug_/nr;
10% of sites have concentrations below this line
•99-01 '00-'02 -01-'03 '02-'04 '03-05 '04-06 '05-'07
Averaging period
300
B) Number of Trend Sites Above NAAQS
,.
"
250
t §
C o ^
200
Ł Ł 150
IP
s i'
g
100
Inn
'99-'01 '00-'02 '01-'03 '02-'04 '03-'05 "04-'06 '05-'07
Averaging period
20
||
ly weighted annu;
ration for three co
lendar years (pg/r
g t 3 •
1 i
00 «
C) Trends by EPA Region3
NAAQS = 15 U
g/m3
-031^^1^=
^
-.
- - —
— _
-R1
R2
R4
R5
R6
R7
R8
-R9
-R10
— Nat'l
'99-TJ1 '00-'02 '01-'03 '02-04 '03-TJ5 '04-'06 '05-'07
Averaging period
''Coverage: 697 monitoring
sites in the EPA Regions (out
of a total of 802 sites
measuring PMj5 in 2007) that
have sufficient data to assess
PM2.5 trends since 1999.
EPA Regions
0
i
Source: U.S. EPA (2008, 157076).
Figure 3-45. Ambient annual PM2.s concentrations in the U.S., 1999-2007, showing A) ambient
concentrations, B) number of trends sites above the annual NAAQS and C)
trends by U.S. EPA Region.
PM10
Figure 3-46 shows trends in U.S. ambient 24-h PMi0 concentrations from 1988-2007. In 2007,
the U.S. national average second highest PMi0 concentration was 37% lower than in 1988 (Figure
3-46A). Of 281 sites used in this trend analysis, the number reporting concentrations above the 24-h
PMio NAAQS (150 (ig/m3) fell from 23 in 1988 to 5 in 2007 with a max of 29 in 1989 (Figure
3-46B). Figure 3-46C shows trends in the second highest 24-h PM10 concentrations broken down by
U.S. EPA region. All regions exhibit an overall decrease from 1988-2007. Largest decreases occurred
in EPA Region 10, which incorporates Washington, Oregon, Idaho and Alaska. Most of the decrease
occurred between 1988-1995.
December 2009
3-93
-------
A) Ambient Concentrations
= 150pg/m3
90% of sites have concentrations below this line
10% of sites have
concentrations below this line
'90 '92 '94 '96 '98 '00
Year
02 '04 '06
gS 35
B) Number of Trend Sites Above NAAQS
20
Ł 11 10
55 " 5
|i
3S 0
llM^i^Ul
'90 '92 '94 '96 '98 '00 '02 '04 '06
Year
C) Trends by EPA Region
'90 '92 '94 '96 '98 '00 '02 '04 '06
Year
Coverage: 274 monitoring sites
in the EPA Regions (out of a total
of 879 sites measuring PMio in
2007) that have sufficient data to
assess PM-io trends since 1988.
EPA Regions
Source: U.S. EPA (2008, 157076).
Figure 3-46.
Ambient 24-h PMio concentrations in the U.S., 1988-2007, showing A) ambient
concentrations, B) number of trends sites above the 24-h NAAQS and C) trends
by U.S. EPA Region.
PM Constituents
The SANDWICH method discussed in Section 3.5.1.1 for estimating PM2.5 composition from
FRM mass measurements and CSN bulk composition measurements was used to evaluate trends in
PM2.5 constituents. Figure 3-47 includes stacked bar charts of PM2.5 composition from 2002 to 2007
stratified by region and season. The regions used in Figure 3-47 were selected based on common
aerosol characteristics including trends, seasonality, size distributions and/or composition as
described in chapter 6 of the 1996 PM AQCD (U.S. EPA, 1996, 079380) and differ from the EPA
regions used in the preceding figures. Figure 3-47 is based on 42 monitoring locations reporting
complete CSN data with 2002 being the first year with sufficient speciation data. The Southwest
region incorporating Arizona, New Mexico and parts of Texas and Oklahoma did not contain any
complete data and therefore is not represented in this analysis. Two seasons representing different
temperature ranges-cool (October-April) and warm (May-September)-were considered in the
figure since many PM2.5 components exhibit strong seasonal dependence.
December 2009
3-94
-------
Cool
Warm
20-
16-
Northwest
20-
16-
12-
20-
fl6-
112
*& 8
CO
S 4
0
02 03 04 05 06 07 02 03 04 05 06 07
North Central
20-
16-
12-
I
4-
0
02 03 04 05 06 07 02 03 04 05 06 07
Midwest
20-
F 16-
112-
S 8-
CO
5 4-
n-
20-
16-
-
-
-
1
-
12-
8-
4-
n -
— I
-
_
-
-
-
m
-
Cool
Warm
Northeast
20-
I16:
% 8-
TO
5 4-
n -
20-
16-
.
—
—
Dn
n
=
8:
4-
=,
.
p.
-
-
==
-
—
.
02 03 04 05 06 07 02 03 04 05 06 07
Southeast
20-
16-
•12-
4-
20
E~16
8-
4
0
02 03 04 05 06 07 02 03 04 05 06 07
Southern California
20-
16-
12-
4-
0
02 03 04 05 06 07 02 03 04 05 06 07
02 03 04 05 06 07
02 03 04 05 06 07
I I Sulfate I H Nitrate
I I Organic Carbon
Elemental Carbon
I Crustal
Northwest
North Central
Midwest
.*.'.'*
f, . Northeast
Southern • *
California Southwest . • .
•
Southeast
Source: U.S. EPA (2008,1911901.
Figure 3-47. Regional and seasonal trends in annual PM2.s compostion from 2002 to 2007
derived using the SANDWICH method. Data from the 42 monitoring locations
shown on the map were stratified by region and season including cool months
(October-April) and warm months (May-September). S042" and N03~ estimates
include NH4+ and particle bound water.
Most of the components showed little discernable trend over the 6-yr period. SO42~ showed a
peak during the warm months of 2005 in the Southeast, Northeast and Midwest, partly due to
atypical weather conditions (U.S. EPA, 2008, 191190). However, no trend over the 6-yr time period
is present for SO42~ in any of the regions or seasons. The same is true for EC and crustal material. A
slight decline in OC was observed for the Northeast during warm months and in Southern California
year-round. The largest decline was for NO3" in Southern California during both cool and warm
December 2009
3-95
-------
months. A smaller decline in NO3" is also observed in the other regions with the exception of the
Northwest where no discernible trend is present. This analysis is limited in time and space by the
availability of CSN data so a high degree of uncertainty remains regarding PM2.5 compositional
trends. However, with the exception of NO3" concentrations in Southern California, no major
changes in PM2.5 composition are evident based on available CSN data from 2002-2007. This is
consistent with Figure 3-44 and Figure 3-45 where the downward trend in PM25 mass begins to level
off after 2002.
3.5.2.2. Seasonal Variations
Many of the figures and tables presented in the preceding sections have included a seasonal
break-down based on the following climatological seasons: winter (December-February), spring
(March-May), summer (June-August) and fall (September-November). Figures A-142 through A-156
in Annex A show bar charts of PM2.5 composition by individual month, illustrating intra-annual
variability on a finer time scale. The same 15 CS As/CBS As are investigated and included in these
plots; they are generated from the same data used in the seasonal and annual pie charts based on the
SANDWICH method discussed in Section 3.5.1.1 and illustrated in Figure 3-17 and Figure 3-18.
Monthly plots for most of the areas reveal heterogeneity in PM composition within the
3-month long seasonal bins defined earlier. This is especially true in the spring and fall when daily
average weather conditions (e.g., temperature) are changing most rapidly, driving fluctuation in
PM2 5 composition on relatively short timescales in many cities. For example, the NO3~ mass in Los
Angeles (Figure A-129) and Riverside (Figure A-134) can vary from a small fraction to the most
prevalent fraction of PM25 mass in a month's time based on the 3-yr aggregate data. Therefore,
selecting a different delineation point for the seasons can have an influence on the seasonal
composition analysis, specifically for constituents that fluctuate rapidly (e.g., NO3~).
Relatively little is known about the seasonal variability in UFPs. Kuhn et al. (2005, 129448)
and Zhu et al. (2004, 156184) found that the concentrations in the UF mode in Los Angeles, CA can
be much higher during winter, particularly during evenings, because atmospheric dilution is reduced
in response to lower mixing heights (Figure 3-48). Jeong et al. (2004, 180350) made similar
observations in Rochester, NY, suggesting an inverse relationship between temperature and UFP
formation in the 11-470 nm size range. Singh et al. (2006, 190136) reported higher particle number
concentrations during winter months, relative to summer and spring, at urban sites in Southern
California, and that afternoon particle number concentrations in warm months either occurred during
a peak in ozone concentrations or followed shortly thereafter, suggesting a role for photochemistry in
addition to meteorological changes in the formation of aerosols. The study also reported increased
concentrations of 60-200 nm particles during a labor strike at the Port of Long Beach, suggesting
contributions from idling ships. Moore et al. (2009, 191004) also report higher particle number
concentrations during cooler months at 14 sites in Long Beach, San Pedro, and Wilmington, CA, a
location with diverse industrial and transportation sources. However, they noted substantial
heterogeneity in seasonal trends between sites with seasonal numerical size distributions not
generalizable across the study area with a maximum monitor separation of under 10 km.
December 2009 3-96
-------
, — ,
5
"^Q.
f
§
"O
180000
160000
140000
120000
100000
80000
60000
40000
20000
n
1 ' ' '
/ ''"-, Site A Summer -•-
-
i. / '"-_ —
V..,'""
o^— - ^~z % v^_ x^
r"^~""r^:^?^:
10
20
30
50
100
25000
— 20000
'1
^ 15000
M- 10000
~° 5000
Site B Winter Even. —
Site B Winter Day
Site B Summer
10 20 30 50 100
Particle diameter dp [nm]
Source:Reprinted with Permission of Elsevier Ltd. from Kuhn et al. (2005,1294481.
Figure 3-48. UFP size distribution at highway (site A) and background (site B) sites in Los
Angeles, CA, during summer and winter seasons, with winter broken into day
and evening distributions.
Studies reporting higher cold-season particle number concentrations are consistent with
vehicle emission studies that found particle emission rates elevated during lower ambient
temperatures (Baldauf et al., 2005, 191184: Mathis et al., 2005, 155970: U.S. EPA, 2008, 191767).
Mathis et al. (2005, 155970) found that cold-start conditions produce roughly an order of magnitude
greater PM number emissions in gasoline engines and more than two orders of magnitude higher PM
number emissions in diesel engines when compared with warm start conditions.
3.5.2.3. Hourly Variability
Hourly PM2.5 and PMi0 measurements are conducted at many sites using beta gauge or TEOM
monitors. Many of the hourly measurements for PM10 have FRM or FEM status. All available hourly
data from FRM, FEM and FRM-like monitors in the 15 CS As/CBS As discussed earlier were used to
investigate diel variation in PM. Of the 15 CS As/CBS As, Atlanta, Chicago, Pittsburgh, Seattle and
St. Louis had qualifying hourly PM2 5 and PMi0 data available. Houston and New York had only
qualifying PM2.5 data. Denver, Detroit, Los Angeles, Philadelphia, Phoenix, and Riverside had only
qualifying hourly PMi0 data. Birmingham and Boston had no qualifying hourly PM2.5 or PMi0 data.
Diel plots for PM25 stratified by weekdays and weekends for seven of the 15 CS As/CBS As
with available data between 2005 and 2007 are included in Annex A, Figures A-157 through A-163.
In most cities investigated, a morning PM25 peak is present starting at approximately 6:00 a.m.,
corresponding with the start of the morning rush hour just before the break-up of overnight
stagnation. In Pittsburgh, dispersion behavior during the night results in elevated PM2 5
December 2009
3-97
-------
concentrations throughout the night that blend in with any morning peak. With the exception of
Pittsburgh, all seven metropolitan areas show two distinct daily peaks on both the weekdays and
weekends. The evening PM2.5 concentration peak is broader than the morning peak and extends to
overnight hours, reflecting the concentration increase caused by a drop in boundary layer height at
night. Figure 3-49 compares the two-peak diel distribution in PM2.s for Seattle with the one-peak
distribution in PM2.5 for Pittsburgh Since these figures represent the distribution of hourly
observations over a 3-yr period, any fluctuations or changes in the timing of the daily peaks would
result in a broadening of the curves shown in the diel plot.
Pittsburgh
Weekday (N = 981)
Weekend (N = 407)
en
77 -i
58 -
38 -
19 -
77 -i
58 -
38 -
19 -
• Median
Mean
90th & 10th
95th & 5th
12 18 24
12 18 24
Seattle
Weekday (N = 5775)
Weekend (N = 2332)
Ł
O)
LO
c\i
^
CL
77 i
58 -
38 -
19 -
77 -i
58
38
19 -
• Median
Mean
90th & 10th
95th & 5th
12 18 24
12 18 24
Figure 3-49. Diel plot generated from hourly FRM-like PM2s data (iig/m3) stratified by weekday
(left) and weekend (right) for Pittsburgh, PA, and Seattle, WA, 2005-2007.
Included are the number of monitor days (N) and the median, mean, 5th, 10th,
90th and 95th percentiles for each hour of the day shown on the horizontal axis.
December 2009
3-9
-------
Annex A, Figures A-164 through A-174 show diel patterns for PM10 stratified by weekdays
and weekends for eleven of the 15 CSAs/CBSAs with available data between 2005 and 2007. All
cities show a gradual morning increase in mean PMi0 starting at approximately 6:00 a.m. on
weekdays, corresponding with the start of the morning rush hour before the break-up of overnight
stagnation. The magnitude and duration of this peak, however, varies considerably by area. Phoenix
shows the most pronounced morning PMi0 peak concentration, which drops off during the day and
reappears in the evening. In contrast, Chicago shows a less pronounced peak with the PM10
concentration remaining elevated throughout the day. Figure 3-50 shows the diel plots of PMi0 for
Chicago and Phoenix. In both instances, the weekend diel pattern is similar in shape to the weekday
pattern with less pronounced peaks. Once again, any fluctuations in the timing of the daily peaks
could result in a broadening of the peaks in the 3-yr composite diel figures.
Chicago
Weekday (N = 1971)
291 -
218 -
145 -
73-
291 -
218 -
145
73
Weekend (N = 793)
Median
Mean
90th & 1Oth
95th & 5th
1
12 18 24
12 18 24
Phoenix
Weekday (N = 1532)
Weekend JN = 618)
291 -,
"E 21S "
145 -
Ł 73-
291 -i
218 -
145
73
12 18 24
Median
Mean
90th & 1Oth
95th & 5th
12 18 24
Figure 3-50. Diel plots generated from hourly FEM PMi0 data (ng/m3) stratified by weekday
(left) and weekend (right) for Chicago, IL, and Phoenix, AZ, 2005-2007. Included
are the number of monitor days (N) and the median, mean, 5th, 10th, 90th and
95th percentiles for each hour of the day shown on the horizontal axis.
UFPs in urban environments have been shown to exhibit a similar two-peaked diel pattern in
Los Angeles (Moore et al, 2007, 122445; Sardar et al., 2005, 180086) and the San Joaquin Valley
(Herner et al., 2005, 135983) in CA, Rochester, NY (Jeong et al., 2004, 180350). Raleigh, NC
(Baldauf et al., 2008, 190239) as well as in Kawasaki City, Japan (Hasegawa et al., 2005, 157355)
and Copenhagen, Denmark (Ketzel et al., 2003, 131251). Figure 3-51 from the Denmark study
December 2009
3-99
-------
shows a large peak in total particle number (dominated by UFPs) corresponding with the morning
rush hour. The morning peak is absent on Sundays, however. Many studies also show a broad
afternoon UFP concentration peak, which likely originates from a combination of evening rush-hour
traffic, decreased atmospheric dilution and formation of UFPs through nucleation involving products
of active photochemistry. Nucleation likely plays an important role since the afternoon peak is
present on weekends whereas the morning traffic related peak is absent. This is consistent with
observations of particle counts in Atlanta peaking during the mid-afternoon for particles <10 nm
(Woo et al., 2001, 011702) resulting from nucleation.
Weekdays
Sundays
4CODO
35C-D3
30000
26CDO
20000
•ŁCDO
• 0000
5CDO
•3
r*,
-Jagtv
_i k.
lOOQO
Source: Adapted with Permission of Elsevier Science Ltd. From Ketzel et al. (2003,1312511
Figure 3-51. Average diel variation in total particle number (ToN) and total particle volume
(ToV) on weekdays (left column) and Sundays (right column) from two sites in
Denmark: one in a busy street canyon (Jagtv) and one measuring urban
background (HC0).
Hourly variability in particle-phase OC and EC were investigated by Bae et al. (2004, 156243)
in the urban St. Louis atmosphere. OC diel patterns were similar during weekdays and weekends
with a broad morning and evening concentration peak most likely reflecting daily fluctuations in
atmospheric mixing height. Weekend EC diel patterns were similar to those for OC, but the weekday
patterns showed more abrupt EC concentration peaks in the morning and afternoon, coinciding with
rush-hour traffic. The divergent weekday patterns between OC and EC suggests motor vehicles or
other EC sources with temporal profiles tracking traffic patterns are primarily responsible for the
daily fluctuations in EC concentrations in St. Louis.
3.5.3. Statistical Associations with Copollutants
Associations between different PM size fractions and between PM and other copollutants
including SO2, NO2, CO and O3 are investigated in this section. AQS data were obtained from all
available co-located monitors across the U.S. after application of a completeness criterion of 11 or
December 2009
3-100
-------
more observations per quarter. Pearson correlation coefficients (r) were calculated using 2005-2007
data. The results are displayed graphically in Figure 3-52 for correlations with PM2.5 mass
concentration and Figure 3-53 for correlations with PMi0 mass concentration. The different PM size
fractions are compared and contrasted in this section using temporal correlations which should not
be confused with average PM mass fraction (e.g., PM2.5/PMi0) comparisons discussed in
Section 3.5.1.1.
Winter
Spring
PM10 (daily avg) -
PM10.25 (daily avg) -
SO2 (daily avg) -
NO2 (daily avg) -
CO (daily avg) -
O3 (daily max 8-h) -
-1.0-0.8-0.6-0.4-0.2 0.0 0.2 0.4 0.6 0.8 1.0 -1.0-0.8-0.6-0.4-0.2 0.0 0.2 0.4 0.6 0.8 1.0
Summer Fall
PM10 (daily avg) -
PM10.2.6 (daily avg) -
SO2 (daily avg) -
NO2 (daily avg) -
CO (daily avg) -
O3 (daily max 8-h) -
-1.0-0.8-0.6-0.4-0.2 0.0 0.2 0.4 0.6 0.8 1.0 -1.0-0.8-0.6-0.4-0.2 0.0 0.2 0.4 0.6 0.8 1.0
correlation (r) with PM25 (daily avg) correlation (r) with PM25 (daily avg)
Figure 3-52. Distribution of correlations between 24-h avg PM2.s and co-located 24-h avg
PMio-2.6, S02, N02 and CO and daily max 8-h avg 03 for the U.S. stratified by
season (2005-2007). Statistics shown include the mean (green star), median (red
line), inner quartile range (box), 5th/95th percentiles (whiskers) and outliers
(black circles).
December 2009
3-101
-------
Winter
Spring
PM25 (daily avg) -
PM10.25 (daily avg) -
SO2 (daily avg) -
NO2 (daily avg) -
CO (daily avg) -
O3 (daily max 8-h) -
-1.0-0.8-0.6-0.4-0.2 0.0 0.2 0.4 0.6 0.8 1.0 -1.0-0.8-0.6-0.4-0.2 0.0 0.2 0.4 0.6 0.8 1.0
Summer Fall
PM25 (daily avg) -
PM10.25 (daily avg) -
SO2 (daily avg) -
NO2 (daily avg) -
CO (daily avg) -
O3 (daily max 8-h) -
-1.0-0.8-0.6-0.4-0.2 0.0 0.2 0.4 0.6 0.8 1.0 -1.0-0.8-0.6-0.4-0.2 0.0 0.2 0.4 0.6 0.8 1.0
correlation (r) with PM10 (daily avg) correlation (r) with PM10 (daily avg)
Figure 3-53. Distribution of correlations between 24-h avg PMi0 and co-located 24-h avg PM2.6,
PMio-2.6, S02, N02 and CO and daily max 8-h avg 63 for the U.S. stratified by
season (2005-2007). Statistics shown include the mean (green star), median (red
line), inner quartile range (box), 5th/95th percentiles (whiskers) and outliers
(black circles).
For both PM2.5 and PMi0 national composite copollutant correlations, there is considerable
spread in the observed correlations in all four seasons. On average, PM2.5 and PMi0 correlate with
each other better than with the gaseous copollutants. The correlations between PM2.5 and PM10 are all
positive but span the range from just above zero to near one. This illustrates the wide variability in
correlation between these two PM metrics. Fewer points are available for correlation with PMi0_2.5
because only data from low-volume FRM/FRM-like samplers were used to calculate PMi0_2.5. The
available data suggest a stronger correlation between PMi0 and PMi0_2.5 than between PM2.5 and
PMio_2.5 on a national basis.
Correlations among copollutants for individual CSAs/CBSAs are included in Annex A,
Figure A-175 through Figure A-188 for PM2.5 and Figure A-189 through Figure A-202 for PM10.
Each data point in these figures represents a co-located monitor pair. Seattle did not have sufficient
co-located data to be included with the other CSAs/CBSAs. As can be seen from the individual
CSAs/CBSAs, there can be considerable variation in the correlations even within an individual urban
area. For example, correlations between 24-h average PM2.5 and PMi0 concentrations measured at
the five co-located monitor pairs in Boston between 2005 and 2007 ranged from 0.42 to 0.88 during
winter and reach as high as 0.98 during the summer (Figure A-177).
Few locations within the 15 CSAs/CBSAs contained adequate data for calculating correlations
with PMio_2.s using low-volume PM data: Boston and New York had two locations each, Atlanta,
Chicago, Denver and Phoenix had only one location and the remaining CSAs/CBSAs had no
locations. Correlations between PM2 5 and PMi0_2.5 varied substantially by CSA/CBSA and season
and no general patterns were observed in the limited data set analyzed here. In contrast, correlations
between PMi0 and PM10_2.5 did show some trends by location and season and were greater in all
locations than correlations between PM2 5 and PMi0_2.5. The highest correlations between PMi0 and
December 2009
3-102
-------
PM10_2.5 were observed in Denver and Phoenix with correlations above 0.88 during all seasons.
Atlanta, Boston, Chicago and New York all had lower correlations between PMi0 and PMi0_2.s
(0.30< r <0.88), particularly during the fall (0.30< r <0.56). The lowest correlations between PMi0
and PMio_2.5 were observed in New York in the fall and Boston in the summer where they dropped to
0.30 and 0.38, respectively, at one of the two monitor locations in each city. In the four eastern
CSAs/CBSAs investigated, correlations between PMi0 and PMi0_2.5 were highest in the spring, in
agreement with the national averages shown in Figure 3-53. In Denver and Phoenix, there was less
seasonal dependence in the correlation.
A similar analysis of correlations between PM size fractions by region was reported in Table
3-1 of the 2004 PM AQCD (U.S. EPA, 2004, 056905) and Figure 2-20 of the 2005 OAQPS Staff
Paper (U.S. EPA, 2005, 090209). In all regions, correlations between PMi0 and PMi0_2.5 were greater
than those for PM2.5 and PMi0_2.5. Correlations between PMi0 and PMi0_2.5 were found to be largest
in the southwest and the upper Midwest and smallest in the southeast and the northeast. While these
regional analyses used different data inclusion criteria for estimating PMi0_2.5 than the criteria used in
this analysis for the individual CSAs/CBSAs, the results are generally consistent: higher
correlations between PMi0 and PMi0_2.5 mass concentrations compared with PM2.5 and PMi0_2.5 mass
concentrations in all regions and higher correlations between PMi0 and PMi0_2.5 mass concentrations
in the west compared to the east.
The correlation between PM and the gaseous pollutants included in Figure 3-52 and Figure
3-53 also showed a large range in values based on the national composite data. There was little
seasonal variability in the mean correlation between PM and SO2. NO2 and CO, however, showed
higher correlations with PM on average in winter than in the other seasons. This is possibly driven
by meteorology with increased frequency of stagnation events in colder months as well as potential
concurrent increases in emissions of these compounds from motor vehicles with colder temperatures.
The correlation between daily max 8-h avg O3 and 24-h avg PM showed the highest degree of
seasonal variability with positive correlations on average in summer (0.56 for PM2.5 and 0.39 for
PMio) and negative correlations on average in winter (-0.30 for PM2.s and -0.18 for PMi0). During
the transition seasons, spring and fall, correlations were mixed but on average were still positive.
PM2.5 is both primary and secondary in origin, whereas O3 is only secondary. Photochemical
production of O3 and secondary PM in the planetary boundary layer (PBL) is much slower during
the winter than during other seasons. Primary pollutant concentrations (e.g., primary PM2.5
components, NO and NO2) in many urban areas are elevated in winter as the result of heating
emissions, cold starts and low mixing heights. O3 in the PBL during winter is mainly associated with
air subsiding from above the boundary layer following the passage of cold fronts, and this subsiding
air has much lower PM concentrations than are present in the PBL. Therefore, a negative association
between O3 and PM is frequently observed in the winter. During summer, both O3 and secondary
PM2.5 are produced in the PBL and in the lower free troposphere at faster rates compared to winter,
and so they tend to be positively correlated. Bell et al. (2007, 093256) also observed wintertime
minima in same-day correlations between 24-h avg PM (both PM2.5 and PMio) and 24-h avg O3
using data from 98 U.S. urban communities. The average correlations were positive in winter, unlike
those shown in Figure 3-53. Furthermore, the highest national average correlations were in spring
and fall in the Bell et al. (2007, 093256) analysis rather than summer as observed in Figure 3-52 and
Figure 3-53. This discrepancy could be a result of the different averaging times used for O3 or the
selection of different monitoring networks and/or time periods.
For the PM2.5 city-specific correlations shown in Annex A, Figure A-175 through Figure A-
188, all selected cities with sufficient data showed negative correlations in the wintertime with daily
max 8-h avg O3 (including Birmingham, Boston, Chicago, Denver, Houston, Los Angeles,
Philadelphia, Phoenix, Pittsburgh, Riverside and St. Louis). The remaining four CSAs/CBSAs had
insufficient data. In Baltimore, Sarnat et al. (2001, 019401) found a significant (at the p <0.05 level)
positive (0.67) and negative (-0.72) correlation between daily PM25 and O3 in the summer (June
19-August 23, 1998) and winter (February 2-March 13, 1999), respectively. For PMi0, the city-
specific correlations with max 8-h avg O3 shown in Annex A, Figure A-189 through Figure A-202
were more variable. Birmingham, Boston, and St. Louis all showed positive wintertime correlations
between PMio and daily maximum 8-h avg O3 while Denver, Detroit, Houston, Los Angeles and
Phoenix showed negative wintertime correlations. The remaining seven CSAs/CBSAs had
insufficient data. These copollutant correlations illustrate the importance of considering seasonality
when assessing temporal relationships between air pollutants, particularly PM and O3.
December 2009 3-103
-------
3.6. Mathematical Modeling of PM
There are two main classes of models used to study atmospheric PM, receptor models and
CTMs. Receptor models are statistical models whereas CTMs are numerical models, i.e., they
approximate derivatives by finite difference approximations. Finite element models are also
numerical models but have not been used as extensively for applications described here, and so are
not discussed further. Receptor models are diagnostic in their approach, in that they try to derive
source contributions at monitoring locations using either ambient data alone or in combination with
data for the chemical composition of sources or in combination with meteorological data. Three-
dimensional CTMs are formulated in a prognostic, or predictive manner, that is, they attempt to
predict species concentrations by solving a set of coupled, non-linear partial differential equations
(continuity equations) for chemical species that include terms based on emissions inventories,
atmospheric transport, chemical transformations, and deposition. Monitoring data is used to evaluate
the performance of CTMs. Each of these approaches has its own advantages and disadvantages.
3.6.1. Estimating Source Contributions to PM Using Receptor Models
Methods for analyzing the composition of ambient PM samples in terms of contributions from
different sources are reviewed in this section. Associations between exposures to ambient PM, as
represented by ambient monitors, and health outcomes have been extensively studied. Some health
studies, described in Section 6.6, have used source apportionment modeling to evaluate relationships
between health outcomes and PM (mainly PM2.5) from different sources. This section is intended to
provide background concerning the uses of source apportionment techniques in such studies.
Understanding the contribution of different emissions sources to ambient PM has also been used
extensively in evaluating air quality data for use in developing control strategies.
3.6.1.1. Receptor Models
Receptor models have been used mainly as part of the development of air quality management
plans. However, there have been several publications relating apportioned source types based on
receptor models to human health effects. Discussions in this section will focus mainly on those
methods that have been used to relate health outcomes to sources. More complete descriptions of a
large number of types of receptor models currently in use are given in Watson et al. (2008, 157128).
who summarize the properties of these methods, including the strengths and weaknesses. This
compilation of receptor models, broken down into different approaches (i.e., chemical mass balance,
factor analysis, tracer-based, meteorology based) is included in Tables A-51 through A-54 in
Annex A.
Receptor models such as the chemical mass balance (CMB) model (Watson et al., 1990,
004848) relate source category contributions to ambient PM concentrations based on analyses of the
compositional profiles of ambient and source emissions samples. It uses as its basis a mass balance
equation that represents all chemical species in an aerosol sample as linear combinations of
contributions from a fixed number of independent sources plus an error term representing the portion
of the measurement that cannot be fit by the model.
The compositional profiles used in receptor models can be extensive (see for example the
SPECIATE data base, http://www.epa.gov/ttnchiel/software/speciate/index.html) for a
comprehensive collection of results from a large number of studies. As an example, several studies
have identified EC and over 100 organic carbon compounds in gasoline PM emissions, including
alkanes, PAHs, oxy-PAHs, steranes, hopanes, and organic acids (Maricq, 2007, 155973; Schauer et
al., 1999, 010582; Schauer et al., 2002, 035332). This breakdown in identifiable groups of organic
compounds is illustrated in Figure 3-54 and Table 3-17 shows emissions factors for trace elements.
Data for the compositional profiles for several other important sources of PM that could be used for
CMB modeling are shown in Table A-55 in Annex A.
Source categories are amenable to refinement and to analysis as information on tracers
becomes available. For example, PBAP have long been known to be significant constituents of the
atmospheric aerosol, but not many studies have evaluated their contributions, largely because of the
lack of suitable tracers and the additional equipment needs for sampling and analysis of bioaerosols.
December 2009 3-104
-------
Bauer et al. (2008, 189986) reviewed studies estimating the contribution of fungal spores to PM2.5
and PM 10-2.5 as fungal spores were expected to be major contributors to PBAP. They proposed the
use of arabitol and mannitol as unique tracers with an estimated accuracy of ± 50% to apportion the
contribution of fungal spores to OC in both PM2.s and PMi0_2.5. They estimated 24-h avg
contributions of- 40% to OC in PMi0_2.5 during spring and summer in Vienna, with a smaller
contribution to PM2.s.
Fine Organic Extractable and
Compounds Editable
142mg/l
Not Extractable
or Elutable
Extractable
and Elutable
/
11 8 mg/l
Unresolved
Resolved
Resolved
Organics
18mg/l
Identified
Organics
5 mg/l
-Oxy-PAH
s Steranes
~~ Hopanes
Figure 3-54.
Source: Reprinted with Permission of Elsevier Science Ltd. From Fraser et al. (1999, 0108191
Schematic of organic composition of participate emissions from gasoline-fueled
vehicles.
One recently-identified concern in the application of CMB-based receptor models with
detailed organic marker compounds is the photochemical stability of those species. Robinson et al.
(2006, 156918) reported evidence of significant summertime photooxidation of hopanes and long-
chain alkenoic acids, low-volatility compounds often used as mobile source and cooking emissions,
respectively. Seasonal differences in hopanes/EC ratios differed in a manner consistent with
oxidation. Photochemical loss of particle-phase marker species mass complicates the interpretation
of model results, as long-range transport and photochemistry may result in the loss of markers for
distant sources. Furthermore, photochemical breakdown of organic marker species may cause losses
in CMB model performance criteria and possible bias in source contribution estimates. The
photooxidation of condensed-phase organic compounds also may affect the polarity and volatility of
these compounds.
In other methods, various forms of factor analysis are used that rely on the varying mix of
species present in ambient observations of compositional data to derive the source contributions.
Standard factor analytic approaches such as Principal Component Analysis (PCA) have been used,
but PCA alone can apportion only the variance, not the mass, in an aerosol composition data set.
Additional steps in particular the identification of source tracers is required in Absolute Principal
Components Scores (APCS) to apportion mass from PCA (Miller et al., 2002, 030661; Thurston and
Spengler, 1985, 056074). However, it can be difficult to find suitable tracers for some sources
because many elements are emitted by more than one source. In Positive Matrix Factorization (PMF)
(Paatero and Tapper, 1994, 086998). the ambient compositional data matrix is decomposed into the
product of a matrix representing the source contributions and one representing the source profiles.
Solutions are obtained by minimizing an object function with respect to these two matrices, and
solutions are subject to non-negativity constraints. PMF also allows for the treatment of missing data
and data near or below detection limits by weighting elements inversely according to their
uncertainties. The PMF approach requires a large number of samples (n typically >50) and are most
often applied to time series data, whereas CMB can be applied to a single sample. Both the CMB
and the PMF approaches find solutions based on least squares fitting and minimization of an object
function. Both methods provide error estimates for the solutions based on estimates of the errors in
the input parameters. It should be noted, though, that the error estimates for both methods often
contain subjective judgments about the magnitude of the analytical and monitoring errors.
December 2009
3-105
-------
Table 3-17. Example of emissions factors (ng/km) for trace elements under variable speed and
steady speed driving conditions for PM emitted by diesel and gasoline engines. Note
that emissions are highly variable.
Diesel
Gasoline
Al
Ca
Fe
K
Mg
Na
Ba
Be
Cr
Cu
Li
Mn
Ni
Pb
S
Ti
V
Zn
Transient
9108 (5224)
69,443 (23,640)
22,910(21,448)
4672 (752)
3087 (461
7736(1751)
583 (349)
26(12)
634 (354)
1944(679)
13 (0.2)
368(183)
2310(656)
793 (593)
23,750 (5295)
2036 (320)
28 (9.4)
21,118(4422)
Steady State
2706
16,128
2036
1191
997
1945
73
23
93
627
7.9
76
644
79
6713
345
11
5620
Transient
2273 (545)
18,247(3044)
10,266(9928)
1935(558)
5183(1706)
2237(1125)
331 (55)
6.7(1.1)
138 (6.7)
1745(1803)
3.0(1.4)
152(85)
107 (0.7)
237 (2.3)
8705 (3375)
118(9.3)
15(11)
4650(1225)
Steady State
252
2324
138
117
183
321
4.8
1.5
8.6
16
0.9
3.4
21
11
349
24
1.8
198
Standard deviations are presented in parenthesis when multiple tests have been averaged.
Source: Adapted with Permission of Elsevier Ltd. from Geller et al. (2006,1396441
The nature of the solutions in terms of source categories is different in the CMB and PMF
approaches. In the CMB approach, the composition of the source emissions is assumed to be known
based on measurements. These assumptions may or may not reflect the composition of emissions
affecting a particular site at any given time or place. However, there may be variations in the
composition of individual source categories (e.g., soils, motor vehicle emissions) across a given
airshed and even in the composition of the same source with time. Source profiles can also be altered
between emission and receptor locations resulting from atmospheric reactions, depending on the
source type and species under analysis. The CMB technique was developed for apportioning source
categories of primary PM and was not formulated to include sources of secondary PM. CMB might
not explain all the mass or produce a valid result unless there is information for the composition of
all major sources affecting a given site, and there is confidence that the existing source profiles are
specific to those sources. For example, Volckens et al. (2008, 105465) describe PAH emission
profiles from hand-held gasoline lawn and garden equipment as found in some CMB source profiles
for motor vehicles.
In PMF, the source solutions are more general in that they contain information about the
entrainment of emissions from additional sources during transport, the time dependence of the
composition of emissions from particular sources, the formation of secondary species and local
differences in source compositions. PMF differs from CMB because it derives the mix of factors
from measured data. However, the procedure used to find a solution results in some rotational
ambiguity (Paatero and Tapper, 1994, 086998). The assignment of sources to PMF factors depends
largely on past experience and judgments. Judgments are based to large extent on comparison with
data for source profiles and also on the factors that could modify the assignments. These issues are
December 2009
3-106
-------
alleviated to some degree by incorporating information about local wind fields and other physical
parameters.
The UNMIX model takes a geometric approach that exploits the covariance of the ambient
data to determine the number of independent sources, the composition and contributions of the
sources, and corresponding uncertainties (Henry, 1997, 020941). UNMIX uses PC A to find edges in
m-dimensional space, where m is the number of ambient species. Success of the UNMIX model
hinges on the ability to find these "edges" in the ambient data from which the number of source
types and the source compositions are extracted. In simplest terms, the approach can be seen to be
similar to that for deriving ternary mixing diagrams, except there is extension to higher
dimensionality. Measurement errors in the ambient data "fuzz" the edges, making them difficult to
find. UNMIX employs an "edge-finding" algorithm to find the best edges in the presence of error.
UNMIX does not make explicit use of errors or uncertainties in the ambient concentrations, unlike
the methods outlined above. Rather they are implicitly incorporated into the analyses. PMF and
UNMIX have also used data for particle size distributions to obtain further information about
sources.
Partial least squares (PLS) is another mathematical model related to PCA which has been used
in a limited number of PM toxicology studies to establish a relationship between PM constituents
and health outcomes (McDonald et al., 2004, 087458: Seagrave et al, 2006, 091291: Veranth et al.,
2006, 087479). Although not really a receptor model and not designed as such, PLS shares some
similarities with certain receptor models; and more importantly attempts to link PM components
with health outcomes. Unlike PCA and other receptor models discussed in this section, PLS
incorporates both predictor variables (e.g., PM component concentrations) and outcome variables
(e.g., toxicological responses) into one coupled regression model. Like PCA, PLS groups the
observable variables into a reduced number of latent variables, thereby reducing the dimensionality
of the model. Typically, PM toxicology studies have been limited to two-component models (two
latent variables on the predictor side compared with two on the outcome side), thereby producing a
2x2 loading plot revealing relationships between predictors and outcomes. PLS is particularly useful
when there are more predictor variables than observations, which is a situation that other
multivariate factor analysis approaches do not handle well. However, since PLS is a variance based
approach, it shares the same shortcomings discussed earlier for PCA. PLS has also traditionally been
limited to two-component applications even though this is not a strict mathematical limitation.
Results from Receptor Models
Results from receptor modeling calculations indicate that PM2.5 is most often produced mainly
by fossil fuel combustion. Fugitive dust, found mainly in the PM10_2.5 size range, represents the
largest source of measured ambient PMi0 in many locations in the western U.S. Quoted uncertainties
in the source apportionment of constituents in ambient aerosol samples typically range from 10 to
50%. It is apparent that a relatively small number of broadly defined source categories, compared to
the total number of chemical species that typically are measured in ambient monitoring-source
receptor model studies, are needed to account for the majority of the observed mass of PM in these
studies. Trying to be more specific about contributions from source categories could result in
ambiguity. For example, some stationary sources (e.g., agriculture use engines) and quite different
mobile sources (e.g., trucks and locomotives) rely on diesel power and ancillary data is required to
resolve contributions from these sources. Compilations of source attribution studies using CMB for
PMi0 have appeared in the 2004 PM AQCD (U.S. EPA, 2004, 056905) and using PMF for PM2 5 in
Engel-Cox and Weber (2007, 156419). Results of the compilation by Engel-Cox and Weber (2007,
156419) for the eastern U.S. are shown in Figure 3-53. There are only three main source categories
in the figure constituting most of the PM2.5 mass (i.e., sulfate, nitrate, mobile). Two of these are
predominantly secondary and not identified by sources of precursors. Tables A-56 and A-57 in
Annex A list results of other receptor modeling studies for PM2 5 and PMi0, many of which are in the
western U.S.
December 2009 3-107
-------
TOTAL Sulfitc
Mobile Lu
Source
Industrial CrustaktSalt Othc
Source: Reprinted with Permission of Air & Waste Management Association from Engel-Cox and Weber (2007, 156419).
Figure 3-55. Source category contributions to PM2.s at a number of sites in the East derived
using PMF.
Spatial Variability in Source Contributions to PM Based on Receptor Models
Spatial variability in source contributions across urban areas is an important consideration in
assessing the likelihood of exposure measurement error in epidemiologic studies relating health
endpoints to sources. Arguments similar to those for using ambient concentrations as surrogates for
personal exposures apply here. Studies for PM2.5 (Kim et al., 2005, 083181; Wongphatarakul et al.,
1998, 049281) indicate that intra-urban variability increases in the following order: regional sources
(e.g., secondary SO42~ originating from EGUs) < area sources (e.g., on-road mobile sources) < point
sources (e.g., stacks). This point is illustrated in Figure 3-56. The only study available for PMi0_2.5
(Hwang et al., 2008, 134420) indicates a similar ordering, but without a regional component
(resulting from the short lifetime of coarse PM compared to transport times on the regional scale) as
shown in Figure 3-57.
December 2009
3-108
-------
0.2
0.0
-0.2
Sec. -
Source: Reprinted with Permission of ACS from Kim et al. (2005, 083181)
Figure 3-56. Pearson correlation coefficients for source category contributions to PM2.s
between the 10 Regional Air Pollution Study/Regional Air Monitoring System
(RAPS/RAMS) monitoring sites in St. Louis.
.c
JS 0.2
Source: Reprinted with Permission of ACS from Hwang et al. (2008,1945331
Figure 3-57. Pearson correlation coefficients for source contributions to PMi0.2.s between the
10 Regional Air Pollution Study/Regional Air Monitoring System (RAPS/RAMS)
monitoring sites in St. Louis.
3.6.2. Chemistry Transport Models
CTMs are the prime tools used to compute the interactions among atmospheric pollutants and
their transformation products, the production of secondary aerosols, the evolution of particle size
distribution, and transport and deposition of pollutants. CTMs are driven by emissions inventories
for primary species such as NOX, SOX, NH3, VOCs, and primary PM, and by meteorological fields
produced by other numerical prediction models. Values for meteorological state variables such as
winds and temperatures are taken from operational analyses, reanalyses, or weather circulation
models. In most cases, these are off-line meteorological analyses, meaning that they are not modified
December 2009
3-109
-------
by radiatively active species generated by the air quality model (AQM). Work to integrate
meteorology and chemistry was done in the mid-1990s by Lu et al. (1997, 048202 and references
therein; 1997, 191768) although limits to computing power prevented their wide-spread application.
More recently, new, integrated models of meteorology and chemistry are now available as well; see,
for example, Binkowski et al. (2007, 090563) and the Weather Research and Forecast model with
chemistry (WRF-Chem) (http://ruc.fsl.noaa.gov/wrf/WGll).
CTMs have been developed for application over a wide range of spatial scales ranging up from
neighborhood to global. CTMs are used to: (1) obtain better understanding of the processes
controlling the formation, transport, and destruction of gas- and particle-phase criteria and hazardous
air pollutants; (2) understand the relations between concentrations of secondary pollutant products
and concentrations of their precursors; (3) understand relations among the concentration patterns of
various pollutants that may exert adverse effects; and (4) evaluate how changes in emissions
propagate through the atmospheric system to secondary products and deposition.
Emissions of precursor compounds can be divided into anthropogenic and natural source
categories. Natural sources can be further divided into biogenic from vegetation, microbes, and
animals, and abiotic from biomass burning, lightning, and geogenic sources. However, the
distinction between natural sources and anthropogenic sources is often difficult to make in practice,
as human activities affect directly or indirectly emissions from what would have been considered
natural sources during the preindustrial era. Thus, emissions from plants and animals used in
agriculture have been referred to as anthropogenic or biogenic in different applications. Wildfire
emissions may be considered natural, except that forest management practices can lead to buildup of
fuels on the forest floor, thereby altering the frequency and severity of forest fires.
The initial conditions, or starting concentration fields of all species computed by a model, and
the boundary conditions, or concentrations of species along the horizontal and upper boundaries of
the model domain throughout the simulation, must be specified at the beginning of the simulation.
Both initial and boundary conditions can be estimated from models or data or, more generally, model
+ data hybrids. Because data for vertical profiles of most species of interest are very sparse, results
of model simulations over larger, usually global, domains are often used. As might be expected, the
influence of boundary conditions depends on the lifetime of the species under consideration and the
time scales for transport from the boundaries to the interior of the model.
Each of the model components described above has associated uncertainties and the relative
importance of these uncertainties varies with the modeling application. The largest errors in
photochemical modeling are still thought to arise from the meteorological and emissions inputs to
the model (Russell and Dennis, 2000, 035563). While the effects of poorly specified boundary
conditions propagate through the model's domain, the effects of these errors remain undetermined.
Because many meteorological processes occur on spatial scales smaller than the model's vertical or
horizontal grid spacing and thus are not calculated explicitly, parameterizations of these processes
must be used. These parameterizations introduce additional uncertainty. Because the chemical
production and loss terms in the continuity equations for individual species are numerically coupled,
the chemical calculations must be performed iteratively until calculated concentrations converge to
within some preset criterion. The number of iterations and the convergence criteria chosen also can
introduce error.
CTMs in current use mostly have one of two forms. The first, grid-based or Eulerian air
quality models subdivide the region to be modeled, the modeling domain, into a three-dimensional
array of grid cells. Spatial derivatives in the species continuity equations are cast in finite-difference
form over this grid and a system of equations for the concentrations of all the chemical species in the
model are solved numerically at each grid point. Time-dependent continuity or mass conservation
equations are solved for each species including terms for transport, chemical production and
destruction, and emissions and deposition (if relevant), in each grid cell. Chemical processes are
simulated with ordinary differential equations, and transport processes are simulated with partial
differential equations. Because of a number of factors such as the different time scales inherent in
different processes, the coupled, nonlinear nature of the chemical process terms, and computer
storage limitations, not all of the terms in the equations are solved simultaneously in three
dimensions. Instead, operator splitting, in which terms in the continuity equation involving
individual processes are solved sequentially, is used.
In the second common CTM formulation, trajectory or Lagrangian models, a number of
hypothetical air parcels are specified as though following wind trajectories. In these models, the
December 2009 3-110
-------
original system of partial differential equations is transformed into a system of ordinary differential
equations.
A less common approach is to use a hybrid Lagrangian-Eulerian model, in which certain
aspects of atmospheric chemistry and transport are treated with a Lagrangian approach and others
are treaded in an Eulerian manner (e.g., Stein et al., 2000, 048341).
Each approach has advantages and disadvantages. The Eulerian approach is more general in
that it includes processes that mix air parcels and allows integrations to be carried out for long
periods during which individual air parcels lose their identity. There are, however, techniques for
including the effects of mixing in Lagrangian models such as FLEXPART (Zanis et al., 2003,
053423). ATTILA (Reithmeier and Sausen, 2002, 053447). and CLaMS (McKenna et al., 2002,
053445). Because both the accuracy and the computational intensity of Eulerian models depend
strongly on the size of the horizontal and vertical grid spacing, speed and fidelity to actual
atmospheric conditions must sometimes be traded-off; that is to say, while finer grid spacing will
often capture effects missed at larger grid intervals, models set up in this way require longer to solve.
In a similar manner, the accuracy of Lagrangian models depends on the number of air parcels
deployed; thus they, too, become computationally intensive when higher-order accuracy is desired.
More detailed discussion of CTM applications appears in the 2008 ISA for NOX and SOX -
Ecological Criteria (U.S. EPA, 2008, 157074).
3.6.2.1. Global Scale
Global-scale CTMs are used to address issues associated with climate change and
stratospheric O3 depletion to characterize long-range air pollution transport, and to provide boundary
conditions for the regional-scale models. The CTMs include parameterizations of atmospheric
transport; the transfer of solar radiation through the atmosphere; chemical reactions; and removal to
the surface by turbulent motions and precipitation for emitted pollutants. The upper boundaries of
the CTMs extend anywhere from the tropopause (~8 km at the poles to -16 km in the tropics) to the
mesopause at -80 km in order to obtain more realistic boundary conditions for problems involving
stratospheric dynamics and chemistry.
Global simulations are typically conducted with a horizontal grid spacing of 200 km or more,
although some models such as GEOS-Chem have been run at grid spacings of about 100 km (e.g.,
Wu et al., 2008, 190039) and efforts are being made to achieve even higher spatial resolution.
Simulations of the effects of long-range transport at particular locations link multiple horizontal
resolutions from the global to the local scale. Finer resolution can only improve scientific
understanding to the extent that the governing processes are more accurately described at that scale.
Consequently, there is a crucial need for observations at the appropriate scales to evaluate the
scientific understanding represented by the models.
3.6.2.2. Regional Scale
Most major regional-scale air-related modeling efforts at EPA use the Community Multi-scale
Air Quality modeling system (CMAQ) (Byun and Ching, 1999, 156314; Byun and Schere, 2006,
090560). A number of other modeling platforms using Lagrangian and Eulerian frameworks were
reviewed in the 2006 O3 AQCD (U.S. EPA, 2006, 088089) and in Russell and Dennis (2000,
035563). The capabilities of a number of CTMs designed to study local- and regional-scale air
pollution problems were summarized by Russell and Dennis (2000, 035563). Evaluations of the
performance of CMAQ are given in Arnold et al. (2003, 087579). Eder and Yu (2005, 089229).
Appel et al. (2005, 089227). and Fuentes and Raftery (2005, 087580). CMAQ's horizontal domain
can extend from a few hundred kilometers on a side to the entire hemisphere. CMAQ is most often
driven by the MM5 mesoscale meteorological model (Seaman, 2000, 035562). though it may be
driven by other meteorological models including WRF and the Regional Atmospheric Modeling
System (RAMS); see http://atmet.com. Simulations of pollution episodes over regional domains
have been performed with a horizontal resolution as low as 1 km; see the application and general
survey results reported in Ching et al. (2006, 090300). However, simulations at such high resolutions
require better parameterizations of meteorological processes such as boundary layer fluxes, deep
convection and clouds (Seaman, 2000, 035562). Finer spatial resolution is necessary to resolve
features such as urban heat island circulation; sea, bay, and land breezes; mountain and valley
breezes; and the nocturnal low-level jet, all of which can affect pollutant concentrations.
December 2009 3-111
-------
The most common approach to setting up the horizontal domain is to nest a finer grid within a
larger domain of coarser resolution. However, there are other strategies such as the stretched grid
and the adaptive grid. In a stretched grid, the grid's resolution continuously varies throughout the
domain, thereby eliminating any potential problems with the sudden change from one resolution to
another at the boundary. Caution should be exercised in using such a formulation because certain
parameterizations like those for convection might be valid on a relatively coarse grid scale but may
not be valid on finer scales. Adaptive grids are not fixed at the start of the simulation, but instead
adapt to the needs of the simulation as it evolves. They have the advantage that they can resolve
processes at relevant spatial scales. However, they can be very slow if the situation to be modeled is
complex. Additionally, if adaptive grids are used for separate meteorological, emissions, and
photochemical models, there is no reason a priori why the resolution of each grid should match, and
the gains realized from increased resolution in one model will be wasted in the transition to another
model. The use of finer horizontal resolution in CTMs will necessitate finer-scale inventories of land
use and better knowledge of the exact paths of roads, locations of factories, and, in general, better
methods for locating sources and estimating their emissions.
The vertical resolution of these CTMs is variable and usually configured to have more layers
in the PEL and fewer higher up. Because the height of the boundary layer is of critical importance in
simulations of air quality, improved resolution of the boundary layer height would likely improve air
quality simulations. Additionally, current CTMs do not adequately resolve fine-scale features such as
the nocturnal low-level jet in part because little is known about the nighttime boundary layer.
CTMs require time-dependent, three-dimensional wind fields for the period of simulation. The
winds may be generated either by a model using initial fields alone or with four-dimensional data
assimilation to improve the model's performance; i.e., model equations can be updated periodically
to bring results into agreement with observations. Modeling series durations can range from
simulations of several days duration, the typical time scale for individual O3 episodes, to several
months or multiple seasons of the year. The current trend in modeling applications is towards annual
simulations. This trend is driven in part by the need to improve understanding of observations of
periods of high wintertime PM (Blanchard et al., 2002, 047598) and the need to simulate O3 episodes
occurring in spring, fall, and winter.
Chemical kinetics mechanisms representing the important reactions occurring in the
atmosphere are used in CTMs to estimate the rates of chemical formation and destruction of each
pollutant simulated as a function of time. Mechanisms that treat the reactions of all individual
reactive species explicitly are computationally too demanding to be incorporated into CTMs for
regulatory use. Similarly, very extensive "master mechanisms" (Derwent et al., 2001, 047912) that
include approximately 10,500 reactions involving 3,603 chemical species (Derwent et al., 2001,
047912) can be combined into mechanisms that group together compounds with similar chemistry.
Because of different approaches to the lumping of organic compounds into surrogate groups for
computational efficiency, chemical mechanisms can produce different results under similar
conditions. The Carbon Bond chemical mechanisms starting with CB-IV (Gery et al., 1989, 043039),
the RADM II mechanism (Stockwell et al., 1990, 043095). the SAPRC (e.g., Carter, 1990, 042893:
Wang et al., 2000, 048357: Wang et al., 2000, 048365). and the RACM mechanisms can be used in
CMAQ. Jimenez et al. (2003, 156611) provided brief descriptions of the features of the main
mechanisms in use and compared concentrations of several key species predicted by seven chemical
mechanisms in a box-model simulation over 24 h.
CMAQ and other state-of-the-science CTMs incorporate processes and interactions of aerosol-
phase chemistry (Binkowski and Roselle, 2003, 191769: Gaydos et al., 2007, 139738: Zhang and
Wexler, 2008, 191770). There have also been several attempts to study the feedbacks of chemistry on
atmospheric dynamics using meteorological models like MM5 and WRF (Grell et al., 2000, 048047:
Liu et al., 2001, 048201: Lu et al., 1997, 048202: Park et al., 2001, 044169). This coupling is
necessary to accurately simulate feedbacks which may be caused by the heavy aerosol loading found
in forest fire plumes (Lu et al., 1997, 048202: Park et al., 2001, 044169) or in heavily polluted areas.
Photolysis rates in CMAQ can now be calculated interactively with model produced O3, NO2, and
aerosol fields (Binkowski et al., 2007, 090563).
Spatial and temporal characterizations of anthropogenic and biogenic precursor emissions
must be specified as inputs to a CTM. Emissions inventories have been compiled on grids of varying
resolution for many hydrocarbons, aldehydes, ketones, CO, NH3, and NOX. Emissions inventories
for many species require the application of algorithms for calculating the dependence of emissions
on physical variables, such as temperature, and to convert the inventories into formatted emission
December 2009 3-112
-------
files which can be used by a CTM. For example, preprocessing of emissions data for CMAQ often is
done by the Spare-Matrix Operator Kernel Emissions (SMOKE) system (http://smoke-model.org).
For many species, information concerning the temporal variability of emissions is lacking, so long-
term annual averages are used in short-term, episodic simulations. Annual emissions estimates are
often modified by the emissions model to produce emissions more characteristic of the time of day
and season. Significant errors in emissions can occur if inappropriate time dependence is used.
Additional complexity arises in model calculations because different chemical mechanisms can
include different species, and inventories constructed for use with one mechanism must be adjusted
to reflect these differences in another.
3.6.2.3. Local or Neighborhood Scale
The grid spacing in regional CTMs, usually between 1 and 12 km2, is usually too coarse to
resolve spatial variations on the neighborhood scale. The interface between regional scale models
and models of smaller exposure scales is provided by smaller scale dispersion models. Several
models could be used to simulate concentration fields near roads, each with its own set of strengths
and weaknesses. The California Department of Transportation's most recent line dispersion model is
CALINE4; see http://www.dot.ca.gov/hq/env/air/pages/calinesw.htm. The CALINE family of
models is not supported by the California Department of Transportation for modeling of highway-
source PM, however, but only for roadway CO, although PM work with CALINE has been
performed for more than ten years; see Wu et al. (2009, 191773) and references therein.
In addition, AERMOD (http://www.epa.gov/scram001/dispersionjrefrec.htm) is a steady-
state plume model formulated as a replacement to the ISC3 dispersion model. In the stable boundary
layer (SBL), it assumes the concentration distribution to be Gaussian in both the vertical and
horizontal dimensions. In the convective boundary layer, the horizontal distribution is also assumed
to be Gaussian, but the vertical distribution is described with a bi-Gaussian probability density
function (pdf). AERMOD has provisions that can be applied to flat and complex terrain and multiple
source types (including, point, area and volume sources) in both urban and rural areas. It
incorporates air dispersion based on the structure of turbulence in the PEL and scaling concepts and
is meant to treat surface and elevated sources, in both simple and complex terrain in rural and urban
areas. The dispersion of emissions from line sources like highways in AEROMOD is handled as a
source with dimensions set using an area or volume source algorithm in the model; however, actual
emissions are usually not in steady state and there are different functional relationships between
buoyant plume rise in point and line sources. Moreover, most simple dispersion models including
AERMOD are designed without chemical mechanisms and so cannot produce secondary pollutants
from their primary emissions .
There are also non-steady state models that incorporate plume rise explicitly from different
types of sources. For example, CALPUFF (http://www.src.corn/calpuff/calpuff 1 .htm), which is
EPA's recommended dispersion model for transport in ranges >50 km, is a non-steady-state puff
dispersion model that simulates the effects of time- and space-varying meteorological conditions on
pollution transport, transformation, and removal and has provisions for calculating dispersion from
surface sources. However, CALPUFF was not designed to treat the dispersion of emissions from
roads, and like AERMOD does not include production of secondary pollutants. The distinction
between a steady-state and time varying model could be unimportant for long time scales; however,
at short time scales, the temporal variability in traffic emissions could result in underestimation of
peak concentration and exposures.
3.6.3. Air Quality Model Evaluation for Air Concentrations
Urban and regional air quality is determined by a complex system of coupled chemical and
physical processes including emissions of pollutants and pollutant precursors, complex chemical
reactions, physical transport and diffusion, and wet and dry deposition. NOX in these systems has
long been known to (1) act nonlinearly in the production of O3 and other secondary pollutants
(Dodge, 1977, 038646): and (2) involve complicated cross-media environmental issues, such as
acidic or nutrient deposition to sensitive biota and degradation of visibility.
NOY species emitted and transformed from emissions control the production and fate of both
O3 and aerosols by sustaining or suppressing OH cycling. Correctly characterizing the interrelated
December 2009 3-113
-------
NOY and OH dynamics for O3 formation and fate in the polluted troposphere depends on new
techniques using combinations of several NOY species for diagnostically probing the complex
atmospheric dynamics in typical urban and regional airsheds.
Evaluation results from a recent EPA exercise of CMAQ in the Tampa Bay, FL, airshed are
presented here as an example of the present level of skill of state-of-the-science AQMs for predicting
atmospheric concentrations of some of the relevant species for this PM NAAQS assessment. This
modeling series exercised CMAQ version 4.4 and with the University of California at Davis (UCD)
sectional aerosol module in place of the standard CMAQ modal aerosol module and was driven by
meteorology from MM5 v3.6 and with NEI emissions as augmented by continuous emissions
monitoring data where available. The UCD size-segregated module was preferred for this application
because of the importance of sea salt particles in the bay airshed. Testing of this new engineering
extension to CMAQ (termed CMAQ-UCD below) revealed that its performance was very similar to
that of CMAQ's standard modal module; hence, model behavior and performance reported here can
stand as a general indication of CMAQ's skill.
The CTM was run with 21 vertical layers for the month of May 2002. For this evaluation,
CMAQ-UCD was run in a one-way nested series of three domains with 32 km, 8 km, and 2 km
horizontal grid spacings from the CONUS (32 km) to central Florida and the eastern Gulf of Mexico
(2 km). Depictions of the 8 km and 2 km domains used here zoomed over the central Tampa area are
shown in Figure 3-58 and Figure 3-59.
December 2009 3-114
-------
8KM RESOLUTION; ZOOME
Figure 3-58. Eight km southeast U.S. CMAQ-UCD domain zoomed over Tampa Bay, FL.
KM RESOLUTION; ZOOMED2
I
Figure 3-59. Two km southeast U.S. CMAQ-UCD domain zoomed over Tampa Bay, FL.
December 2009
3-115
-------
ST
012345678 91011121314151617181920212223
Hours (EST)
012345
10 11 12 13 14 15 16 17 18 19 20 21 22 23
Hours (EST)
012345
10 11 12 13 14 15 16 17 18 19 20 21 22 23
Hours (EST)
Figure 3-60. Hourly average CMAQ-UCD predictions and measured observations of NO (top),
N02 (middle), and total NOX (bottom) concentrations for May 1-31, 2002. Green
squares = 8 km solution, red diamonds = 2 km solution, blue circles =
observations.
December 2009
3-116
-------
1 -May 2-May 3-May 4-May 5-May 6-May 7-May 8-May 9-May 10-
May
Hours (date mark = 0000 EST)
4-r
3--
11- 12- 13- 14- 15- 16- 17- 18- 19- 20-
May May May May May May May May May May
Hours (date mark = 0000 EST)
21- 22- 23- 24- 25- 26- 27- 28- 29- 30- 31-
May May May May May May May May May May May
Hours (date mark = 0000 EST)
Figure 3-61. CMAQ-UCD predictions and measured observations of ethene concentrations at
Sydney, FLfor May 1-31, 2002. Green squares = 8 km solution, blue circles =
observations.
December 2009
3-117
-------
6--
Ul M
E 4 - •
2--
1 -May 2-May 3-May 4-May 5-May 6-May 7-May 8-May 9-May 10-
May
Hours (date mark = 0000 EST)
11 -May 12-May 13-May 14-May 15-May 16-May 17-May 18-May 19-May 20-May
Hours (date mark = 0000 EST)
4T
21- 22- 23- 24- 25- 26- 27- 28- 29- 30- 31-
May May May May May May May May May May May
Hours (date mark = 0000 EST)
Figure 3-62. CMAQ-UCD predictions and measured observations of isoprene concentrations
at Sydney, FL for May 1 -31, 2002. Green squares = 8 km solution, blue circles =
observations.
December 2009
3-118
-------
60 T
10
1 -May 2-May 3-May 4-May 5-May 6-May 7-May 8-May 9-May 10-
May
Hours (date mark = 0000 EST)
11 -May 12-May 13-May 14-May 15-May 16-May 17-May 18-May 19-May 20-May
Hours (date mark = 0000 EST)
60 T
— 40 -
I
3.30 +
o I . i i i . i . i . i i i . i . i . i . i i i . i . i . i i i . i . i . Y i I i I . I
21- 22- 23- 24- 25- 26- 27- 28- 29- 30- 31-
May May May May May May May May May May May
Hours (date mark = 0000 EST)
Figure 3-63. CMAQ-UCD predictions and measured observations of PIVk.e concentrations at
Sidney, FLfor May 1-31, 2002. Green squares = 8 km solution, blue circles =
observations.
December 2009
3-119
-------
3.6.3.1. Ground-based Comparisons of Photochemical Dynamics
Errors in the NOX concentrations in the model, most likely from on-road emissions, affected
NOX predictions shown in Figure 3-60, but CMAQ-UCD's general responses were reasonable. The
model also replicated anthropogenic and biogenic VOC emissions well; see Figure 3-61 and Figure
3-62, respectively. After initial errors leading to underprediction in the first 21 days, CMAQ-UCD's
predictions of hourly PM2.5 concentrations and trends over the whole month also replicated the
observed concentrations well (Figure 3-63).
3.6.3.2. Predicted Chemistry for Nitrates and Related Compounds
Particulate NO3~ (pNO3~) plays a crucial and complex role in the health of aquatic and
estuarine ecosystems and human drinking water systems. Gas-phase NO3~ replacement of CPon sea
salt particles is often favored thermodynamically and the Vd of the coarse pNO3~ formed through this
replacement is more than an order of magnitude greater than for fine pNO3~. Over open bodies of salt
water such as the Gulf of Mexico and Tampa Bay, FL, pNO3~ from this reaction dominates dry
deposition and is estimated to be of the same order as pNO3~ wet deposition.
However, total NO3~ concentrations are driven, buffered, and altered by a wide range of
photochemical gas-phase reactions, heterogeneous reactions, and aerosol dynamics, making them
especially difficult to model. Because pNO3~ is derived mostly from gas-phase HNO3 and will
interact with Na+, NH4+, Cl~, and SO42~, all these species and the physical parameters governing their
creation, transport, transformation, and fate must be accurately replicated to predict pNO3~ with high
fidelity. This has historically been a difficult problem for numerical process models, owing in large
part to the pervasive dearth of reliable ambient measurements of NOjT in its various forms.
Normalized mean error (NME) for the large-scale Eulerian CTM-predicted pNO3~ has typically been
on the order of a factor of 3 greater than the NME for particulate SO42~ (pSO42~) (Odman et al, 2002,
092474: Pun et al 200+3, 047775).
SO42~, NH4+, Na+, and Crwere all predicted to within a factor of 2 and with no significant bias
during the photochemical day in the 8 km CMAQ-UCD solution, although a significant bias in Na+
and CPwas evident in the 2 km solution for two near-water sites. This grid-size dependent bias is
still being explored. Size segregation maxima were correct to within two size bins every day for
which there were observations for both SO42~and NH4+ (0.2 to 1.0 urn), and Na+ and Cl
(2.0-10.0 urn). Cl~ concentrations were greatly overpredicted during dark hours, but were nearer to
observed values during the photochemical day. CMAQ-UCD performance for HNO3 and NH3 are
shown in Figure 3-64 and Figure 3-65, respectively.
Figure 3-66 shows that CMAQ-UCD systematically underpredicted the hourly time series of
measured pNO3~ concentrations at the Sydney supersite, the only location with discrete pNO3~ data.
These time series data establish that CMAQ-UCD's largest errors were on 4 days in the first 2 wk of
the month, but that the total peak pNO3~ concentrations were nearly all underpredicted.
Since pNO3~ is derived in large part from gas-phase HNO3, its underprediction may be due to
an underprediction of HNO3 concentrations or an underrepresentation of the gas- to aerosol-phase
change. At Sydney, FL, in fact, both these conditions held. Figure 3-64 depicts the model's bias for
HNO3 underprediction in both the 8 km and 2 km solutions, except for four days of very large peak
overpredictions. This pattern of underpredictions was especially evident overnight. On 8 other days
the model overpredicted the one hour peak concentration as well, though not so substantially, but the
chief effect was still one of an artificial and inappropriate N limitation in the model.
A time series molar equivalent ratio of HNO3 to total NO3 depicts which phase stores the NO3~
and how that storage ratio changes over time. Figure 3-67 shows that at Sydney, FL, CMAQ-UCD
stored too much NO3~ in the gas phase as HNO3 (and recall that the daytime HNO3 concentrations
were sometimes overpredicted by the model) and too little in the gas phase overnight, when the
model was regularly low against the measurements; compare Figure 3-64 and Figure 3-65. Note
again here the similarity of the 8 km and 2 km solutions in this comparison.
Interestingly, the 23-h integrated data did not reveal this important difference in NO3~ form
between the model and measurements as Figure 3-68 shows in the stacked bar percentage plots of
fine and coarse pNO3~ together with gas-phase HNO3. Both the 8 km (Figure 3-68, middle panel)
and the 2 km (Figure 3-68, bottom panel) solutions predicted distributions between the two general
ranges of aerosol size, and between gas and aerosol phases, with good fidelity to the daily
observations (Figure 3-68, top panel) at Sydney, FL. This result illustrates that while discrete time
December 2009 3-120
-------
series data are crucial for diagnosing model behavior, on the integrated total daily and longer basis
used for computing total annual N loads, CMAQ-UCD predicted approximately the correct
distributions for pNO3~, even though the total NO3~ concentration prediction was biased low.
1-May 2-May 3-May 4-May 5-May 6-May 7-May 8-May 9-May 10-
May
Hours (Date Mark = 0000 EST)
11-May 12-May 13-May 14-May 15-May 16-May 17-May 18-May 19-May 20-May
Hours (Date Mark = 0000 EST)
21-May22-May23-May24-May25-May26-May27-May28-May29-May30-May31-May
Hours (Date Mark = 0000 EST)
Figure 3-64. CMAQ-UCD predictions of HMOs' concentrations and corresponding measured
observations at Sydney, FL, for May 1-31, 2002. Green x = 8 km solution, red
diamonds = 2 km solution, blue circles = observations.
December 2009
3-121
-------
1 -May 2-May 3-May 4-May 5-May 6-May 7-May 8-May 9-May 10-
May
Hours (date mark = 0000 EST)
22 T
11- 12- 13- 14- 15- 16- 17- 18- 19- 20-
May May May May May May May May May May
Hours (date mark = 0000 EST)
21 -May22-May23-May24-May25-May26-May27-May28-May29-May30-May31 -May
Hours (date mark = 0000 EST)
Figure 3-65. CMAQ-UCD predictions of NH3 concentrations and corresponding measured
observations at Sydney, FL, for May 1-31, 2002. Green x = 8 km solution, red
diamonds = 2 km solution, blue circles = observations.
December 2009
3-122
-------
1-May 2-May 3-May 4-May 5-May 6-May 7-May 8-May 9-May
Hours (date mark = 0000 EST)
10-
May
11-May 12-May 13-May 14-May 15-May 16-May 17-May 18-May 19-May 20-May
Hours (date mark = 0000 EST)
21-May22-May23-May24-May25-May26-May27-May28-May29-May30-May31-May
Hours (date mark = 0000 EST)
Figure 3-66. CMAQ-UCD predictions of pN03" concentrations and corresponding measured
observations at Sydney, FL, for 1-31 May, 2002. Green x = 8 km solution, red
diamonds = 2 km solution, blue circles = observations.
December 2009
3-123
-------
1.Oi-
l-May 2-May 3-May 4-May 5-May 6-May 7-May 8-May 9-May 10-May
Hours (date mark = 0000 EST)
0.0
11-May 12-May 13-May 14-May 15-May 16-May 17-May 18-May 19-May 20-May
Hours (date mark = 0000 EST)
1.0 T
21-May22-May23-May24-May25-May26-May27-May28-May29-May30-May31-May
Hours (date mark = 0000 EST)
Figure 3-67. CMAQ-UCD predictions of the ratio of HN03 to total N03 and corresponding
measured observations at Sydney, FL, for May 1-31, 2002. Green x = 8 km
solution, red diamonds = 2 km solution, blue circles = observations.
December 2009
3-124
-------
MOUCH aerosol NO3 and ARA HNO3 (23 h daily mean)
5/1 5/3 5/5 5/7 5/9 5/11 5/13 5/15 5/17 5/19 5/21 5/23 5/25 5/27 5/29 5/31
Date, 2002
• <25un 125-IOuii DHNO3
CMAQ-UCD 8 km solution (23 h daily mean)
5/1 5/3 5/5 5/7 5/9 5/11 5/13 5/15 5/I7 5/19 5/21 5/23 5/25 5/27 5/29 5/31
Date, 2002
• < 2.5 mi 125-10 urn DHNO3
CMAQ-UCD 2 km solution (23 HOT Means)
5/1 5/3 5/5 5/7 5/9 5/11 5/13 5/15 5/17 5/19 5/21 5/23 5/25 5/27 5/29 5/31
Date. 2002
• <2.5 urn •2.5-10UI11 BHNO3
Figure 3-68. CMAQ-UCD predicted size and chemical-form fractions of total N03" for days in
May 2002 with measured observations. Measured concentrations (top panel);
8 km solution (middle panel); 2 km solution (bottom panel). Red bars = pN03" <2.5
urn; blue bars = pN03" 2.5-10 urn; green bars = HN03.
December 2009
3-125
-------
While inorganic aerosol anion totals were dominated by NO3 in the coarse fraction and by
SO42~ in the fine fraction, there was sufficient NHX (NHX = NH3 + NH4+) at Sydney, FL, to form fine
aerosol NH4NO3 in some circumstances. Figure 3-65 depicts the hourly mass concentration of NH3
at Sydney, FL, showing again the strong similarity of the 8 km and 2 km solutions. Each solution,
however, underpredicted the measured NH3 concentrations consistently, and especially for the nine
very large excursions of 10-20 ug/m3 during the month.
Overall, CMAQ-UCD was found to be operationally sound in this evaluation of its 8 km and
2 km solutions for the Tampa Bay airshed using the ground-based and aloft data (not shown here)
from the May 2002 field intensive. Moreover, results from diagnostic tests of the model's
photochemical dynamics were generally in excellent agreement with results from the ambient
atmosphere. However, CMAQ-UCD was biased low in this application for total NO3 and for NO3
present as gas-phase HNO3. In addition, the model was biased low for the HOX radical reservoir
species CH2O and H2O2 (not shown here), though this bias appeared to have been limited to these
species. Performance of the new UCD aerosol module was judged to be entirely adequate, allocating
aerosols by chemical makeup to the appropriate size fractions. Model performance for fine-mode
aerosols was also judged to be fully adequate.
3.6.4. Evaluating Concentrations and Deposition of PM Components
with CTMs
3.6.4.1. Global CTM Performance
The wet and dry deposition processes described in Section 3.3 are of necessity highly
parameterized in all CTMs. While all current models implement resistance schemes for dry
deposition, the Vd generated from different models can vary highly across terrain types (Stevenson et
al., 2006, 089222).The accuracy of wet deposition in global CTMs is tied to spatial and temporal
distribution of model precipitation and the treatment of chemical scavenging. Dentener et al. (2006,
088434) compared wet deposition across 23 models with available measurements around the globe.
Figure 3-69 and Figure 3-70 extract results of a comparison of the 23-model mean versus
observations over the eastern U.S. for pNO3~ and pSO42~ deposition, respectively. The mean model
results were strongly correlated with the observations (r >0.8), and usually captured the magnitude of
wet deposition to within a factor of two over the eastern U.S. Dentener et al. (2006, 088434)
concluded that 60-70% of the participating models captured the measurements to within 50% in
regions with quality controlled observations.
December 2009 3-126
-------
600
400
200
200 400 600
Measurement
Source: Adapted with Permission of American Geophysical Union from Dentener et al. (2006, 088434).
Figure 3-69. Scatter plot of total nitrate (HN03 plus pN03") wet deposition (mg N/m2/yr) of the
model mean versus measurements for the North American Deposition Program
(NADP) network. Dashed lines indicate a factor of two. The gray line is a linear
regression through zero.
1000
800
600
400
200
I *
' I t'Jm *
• / . • **"
0 C±
'''"•••|"j*i ."
V V f* •
\v • ',.
I ** • . ^
> ""'"
<*-'
200
400 600
Measurement
800
1000
Source: Adapted with Permission of American Geophysical Union from Dentener et al. (2006, 088434).
Figure 3-70. Scatter plot of total S042" wet deposition (mg S/m2/yr) of the model mean versus
measurements for the National Atmospheric Deposition Program (NADP)
network. Dashed lines indicate a factor of two. The gray line is a linear
regression through zero.
3.6.4.2. Regional CTM Performance
Regional CTM performance for concentration and deposition of some of the most relevant PM
species is illustrated here with examples from CMAQ version 4.6.1 as configured and run for
December 2009
3-127
-------
exposure and risk assessments reported in the Risk and Exposure Assessment for the Review of the
Secondary National Ambient Air Quality Standards for Oxides of Nitrogen and Oxides of Sulfur
(U.S. EPA, 2009, 191774); additional details on the model configuration and application are found
there. A map of the 36 km parent domain and two 12 km (east and west) progeny domains appears in
Figure 3-71.
Figure 3-71. CMAQ modeling domains for the OAQPS risk and exposure assessments: 36 km
outer parent domain in black; 12 km western U.S. (WUS) domain in red; 12 km
eastern U.S. (EUS) domain in blue.
Comparisons from the 2002 annual run of CMAQ for the exposure assessment are shown here
against measured concentrations and deposition totals from nodes in three networks: IMPROVE,
CSN (labeled STN in the plots) and CASTNet. Comparisons were made as model-observation pairs
at all sites having sufficient data for the seasonal or the 2002 annual time period in the two 12 km
east and west domains and were evaluated with the following descriptive statistics: correlation, root
mean square error, normalized mean bias, and normalized mean error.
Summertime pSO42~ concentrations are well predicted by CMAQ, to within a factor of 2 at
nearly every point, and with R2 >0.8 across all three networks (Figure 3-72). This result tracks the
generally well-predicted SO42~ concentrations found in earlier CMAQ evaluations: see Eder and Yu
(2005, 089229). Mebust et al. (2003, 156749) and Tesche et al. (2006, 157050). Since pSO42~
concentrations are strongly a function of precipitation, care must be taken to ensure that the
meteorological solution driving individual CMAQ chemical applications produces precipitation
fields with low bias as discussed by Appel et al. (2008, 155660).
December 2009
3-128
-------
2002ac_met2v33_12kmE SO4 for June to August 2002
CM _
O
a IMPROVE (2002ac_met2v33_12kmE)
A STN(2002ac_met2v33_12kmE)
CASTNet (2002ac_met2v33_12kmE)
Monthly Average
SO4 ( ug/m3
2002ac met2v33 12kmE
CORR RMSE NMB NME
IMPROVE 0.96 0.84 -11.7 17.2
STN 0.83 1.57 0.4 20.1
CASTNet 0.97 0.83 -8.6 12.5
Figure 3-72.
Observation
12-km EUS Summer sulfate PM. Each data point represents a paired monthly
averaged (June/July/August) observation and CMAQ prediction at a particular
IMPROVE, STN, and CASTNet site. Solid lines indicate a factor of two around the
1:1 line shown between them.
Wintertime pNO3 (Figure 3-73) and total NO3 (HNO3+pNO3 ) (Figure 3-74) concentrations
are predicted less well by CMAQ, but NO3 is a pervasively difficult species to measure and model.
Still, at the CASTNet nodes where the total NO3 concentrations are higher than they are at all but a
few of the remote IMPROVE sites, CMAQ predicts concentrations for nearly every node to within a
factor of 2 and with an R2 >0.8. These CMAQ-predicted concentrations, coupled with modeled cloud
and precipitation fields produce wet deposition fields for SO42~ and NO3~ in the east domain as
shown in Figure 3-75 and Figure 3-76, respectively.
December 2009
3-129
-------
2002ac_met2v33_12kmE NO3 for December to February 2002
oo -
CD -
O
OJ -
IMPROVE (2002ac met2v33 12kmE)
A STN(2002ac_met2v33_12kmE)
Monthly Average
A NO3 ( ug/m3 )
2002ac met2v33 12kmE
CORR RMSE NMB NME
IMPROVE 0.73 0.69 0.5 47
STN 0.67 1.32 -6.9 35.7
Observation
Figure 3-73. 12-km EUS Winter nitrate PM. Each data point represents a paired monthly
averaged (December/January/February) observation and CMAQ prediction at a
particular IMPROVE and STN site. Solid lines indicate a factor of two around the
1:1 line shown between them.
December 2009
3-130
-------
2002ac_met2v33_12kmE TNO3 for December to February 2002
CD -
1 *
O
CO
(M -
o -I
n CASTNet (2002ac met2v33 12kmE)
Monthly Average
TNO3 ( ug/m3 )
2002ac_met2v33_12kmE
CORR RMSE NMB NME
CASTNet 0.85 0.88 -1.6 21.4
\
2
3
i
4
Observation
5
6
7
Figure 3-74. 12-km EUS Winter total nitrate (HMOs + total pNCV). Each data point represents a
paired monthly averaged (December/January/February) observation and CMAQ
prediction at a particular CASTNet site. Solid lines indicate a factor of two around
the 1:1 line shown between them.
December 2009
3-131
-------
2002ac met2v33 12km E SO4 for 20020101 to 20021231
co
o
CO
LD
C\]
o
C\]
o
o .
LO -
Q NADP_dep(2002ac_met2v33_12kmE)
IA
RMSE =
RMSEs =
RMSEu =
MB
ME
MdnB =
MdnE =
Period Accumulated
SO4 ( kg/ha
Observation
Figure 3-75. 12-km EUS annual sulfate wet deposition. Each data point represents an annual
average paired observation and CMAQ prediction at a particular NADP site. Solid
lines indicate the factor of 2 around the 1:1 line shown between them.
December 2009
3-132
-------
2002ac met2v33 12kltlE NO3 for 20020101 to 20021231
O
O
n NADP_dep(2002ac_met2v33_12kmE)
Period Accumulated
NO3 ( kg/ha
15
Observation
Figure 3-76. 12-km EUS annual nitrate wet deposition. Each data point represents an annual
average paired observation and CMAQ prediction at a particular NADP site. Solid
lines indicate a factor of two around the 1:1 line shown between them.
Importantly, CMAQ captured the chief spatial patterns and magnitudes of air concentrations
and wet deposition relevant to computing concentration and deposition budgets, as shown in Figure
3-77 for concentrations and Figure 3-78 for deposition. More specifically, CMAQ's predictions of
NH3 and SO42~ for both high and low concentration sites are well within the range of observed
measurements (Figure 3-79).
December 2009
3-133
-------
iv17 SO4 for June to August 2002
J4c_emisv17 NH4 for June to August 2001
MPROV= (J4c_emisv17
STN |J4c_.srnlsv17)
CASTNe: [J4c_emisv17)
monthly average
SO4 (ug/m3}
HPQ - None
Slate = All Site = All
6 8 10
Observation
234
Observation
Figure 3-77. CMAQ vs. measured air concentrations from east-coast sites in the IMPROVE,
CSN (labeled STN), and CASTNet sites in the summer of 2002 for sulfate (left) and
ammonium (right). Solid lines indicate a factor of 2 around the 1:1 line shown
between them.
CMAQ Ammonium Ion Wet Deposition
WET NH4 DEPOSITION (KG/HA)
CMAQ (Ml
VS. NADP (2001-2003 AVERAGED)
LIMITED TO SITES IN THE EASTERN U.S.
ANNUAL
__ _ 01234567
NADP Ammonium Ion Wet Deposition NADP
LEGEND
REGRESSION THROUGH ORIGIN ^—^—
RUNNING MEDIAN SMOOTH UNE
NADP SITES IN CHESAPEAKE BAY *
200 2003200
Figure 3-78. Comparison of CMAQ-predicted and NADP-measured NH4+ wet deposition : (top
left) CMAQ prediction; (bottom left) NADP-measurements; (right) regression and
smoothed median line through CMAQ predictions and NADP measurements with
sites in the Chesapeake Bay watershed highlighted.
December 2009
3-134
-------
Kenansville Ammonia July 2004
12-hour Averages: 6am-6pm
184 186 188 190 192 194 196 198 200 202 204 206 208 210 212 214
Julian Day: TickMark at Midnight (July 2004)
-Kenansville NH3 -B-CMAQ-J4C NH3
Kenansville Sulfate July 2004
12-hour Averages: 6am-6pm
184 186 188 190 192 194 196 198 200 202 204 206 208 210 212 214
Julian Day: TickMark at Midnight (July 2004)
- Kenansville SO4 -•- CMAQ-J4C ASO4
Millbrook (Raleigh) Ammonia July 2004
12-hour Averages: 6am-6pm
190 192 194 196 198 200 202 204 206 208 210 212 214
Julian Day: TickMark at Midnight (July 2004)
CMAQ-J4C NH3
Millbrook (Raleigh) Sulfate July 2004
12-hour Averages: 6am-6pm
190 192 194 196 198 200 202 204 206 208 210 212 214
Julian Day: TickMark at Midnight (July 2004)
CMAQ-J4C ASO4
Figure 3-79. CMAQ-predicted (red symbols and lines) and 12-h measured (blue symbols and
lines) NH3 and S042~ surface concentrations at high and low concentration grid
cells in North Carolina in July 2004. (top left) High concentration NH3 in
Kenansville; (top right) high concentration S042~ in Kenansville; (bottom left) low
concentration NH3 in Raleigh; (bottom right) low concentration S042~ in Raleigh.
Deposition velocities are difficult to estimate for reasons described in Section 3.3.3. Recent
work in EPA's Atmospheric Modeling and Analysis Division with CMAQ showed that the original
Vd for NH3 was very likely too high and should be nearer to the values for SO2 deposition, or even
lower over some land use surface types. A sensitivity study with the model was performed to test the
effects of changing Vd for NH3 on the fraction of NH3 available for transport away from grid cells
with high emissions concentrations. Comparisons were made for the surface grid cells and total
column NH3 concentrations.
In the highest emissions grid cells during June 2002, the surface NHX budget was dominated
by turbulent transport or vertical mixing moving a majority of the surface NH3 emissions up and
away from the surface into the mixed layer. Figure 3-80 depicts the NHX budget under the base case
(Base Vd) and the sensitivity case (SO2 Vd) for which the NH3 Vd was set equal to the SO2 Vd. Lower
NH3 Vd decreased NHX deposition to the surface from 15 to 8%, leaving more NHX for transport
horizontally, 22% up from 20% in the base case, and vertically, 69% up from 64% in the base case.
Typically, -67% of surface emissions were moved aloft where most was advected away from the
high emissions grid cell, with a small fraction converted to pNH4+ and an even smaller fraction wet-
deposited to the surface. The total column analyses for NH3 and NHX are shown in Figure 3-81.
December 2009
3-135
-------
0.04%
38m Gas to
Surface partiriP
69-64%.
Vertical
Diffusior
TTTlun*
Free Troposphere
Mixed Layer (-2km)
22-20%
Horizontal
Advection
8-15%
Dry Deposition
NH3
Emissions
NH3 Layer 1 Analysis
Right Number: BaseVd
Left Number: SO2Vd
Figure 3-80. Surface grid cell (layer 1) analysis of the sensitivity of NHX deposition and
transport to the change in NH3 Vd in CMAQ.
2.1-2.2%
Gas to Particle
Conversion
(unusually
low)
N
Enii
NH3Cc
C
•fopoj Model - 16km
Free Troposphere
89-82%
Horizontal
'Advection Mixed Layer( -2km)
\ * 0.55-0.54%
1 Wet Deposition
TT T
n3 Dry Deposition
ssions 8-15%
lunin Analysis
Right Number: BaseVd
Left Number: SOjVd
N
Emi
NHX(.
\ \ v
H3 Dry
ssions
Column /
91-84%
Horizontal
"" Advection
~0.58-0.57%
i'et Deposition
Deposition
8-15%
Analysis
Figure 3-81. Total column analysis for NH3 (left) and NHX (right) showing modeled NH3
emissions, transformation, and transport throughout the mixed layer and up to
the free troposphere.
Local total deposition (wet + dry) is a significant but not dominant loss pathway for surface
NH3 emissions. In these simulations, CMAQ deposited -25% of the NH3 emissions from the single
high concentration grid cell in Sampson County, NC, back into that grid cell. By far, the largest
contribution to the local deposition total was dry deposition. Dry-to-wet deposition ratios for the
Sampson County high emissions grid cell and surrounding surface grid cells ranged from 2 to 10.
Deposition to grid cells farther away from the high concentration, immediately surrounding
grid cells, was significantly affected by the change in NH3 Vd tested in this case. Figure 3-82 depicts
December 2009
3-136
-------
the range of influence of the high concentration grid cell, where that range is defined to be the
distance by which 50% of the emissions attributable to that grid cell have deposited. The range of
influence of the high concentration Sampson County grid cell was extended in the Vd sensitivity
tested here from -180 km in the base case to -400 km in the case using the lower, more realistic V
for NH3. The areal extent of this difference in range of influence is mapped in Figure 3-83.
June 2002 NHx Range of Influence: BaseVd vs. SO2Vd
Sampson County (single cell)
Distance from Center (km)
- -•- -Wet+Dry Dep BaseVd — •
— o— Wet+Dry Dep SO2Vd — a
Advection BaseVd
Advection SO2Vd
Figure 3-82. Range of influence (where 50% of emitted NH3 deposits) from the high
concentration Sampson County grid cell in the June 2002 CMAQ simulation of Vd
sensitivities. Base case and sensitivity case total deposition (blue symbols and
lines); base case and sensitivity case advection totals (red symbols and lines).
December 2009
3-137
-------
Figure 3-83. Areal extent of the change in NHX range of influence as predicted by CMAQ for
the Sampson County high concentration grid cell (center of range circles) in June
2002 using the base case and sensitivity case Vd.
December 2009
3-138
-------
3.7. Background PM
The background concentrations of PM that are useful for risk and policy assessments
informing decisions about the NAAQS are referred to as policy-relevant background (PRB)
concentrations. PRB concentrations have historically been defined by EPA as those concentrations
that would occur in the U.S. in the absence of anthropogenic emissions in continental North America
defined here as the U.S., Canada, and Mexico. For this document, PRB concentrations include
contributions from natural sources everywhere in the world and from anthropogenic sources outside
continental North America. Background concentrations so defined facilitated separation of pollution
that can be controlled by U.S. regulations or through international agreements with neighboring
countries from those that were judged to be generally uncontrollable by the U.S. Over time,
consideration of potential broader ranging international agreements may lead to alternative
determinations of which PM source contributions should be considered by EPA as part of PRB.
3.7.1. Contributors to PRB Concentrations of PM
Contributions to PRB concentrations of PM include both primary and secondary natural and
anthropogenic components. Natural sources include wind erosion of natural surfaces (Gillette and
Hanson, 1989, 030212); volcanic production of SO42~; PBAP; wildfires producing EC, OC, and
inorganic and organic PM precursors; and SOA produced by oxidation of biogenic hydrocarbons
such as isoprene and terpenes. However, human intervention can be involved in the formation of
SOA, as production of natural SOA depends to a large extent on the presence of anthropogenic NOX.
As described earlier in Section 3.3, prescribed fires are considered part of PRB. In addition to
emissions from forest fires in the U.S., emissions from forest fires in other countries can be
transported to the U.S. For example, Boreal forest fires in Canada (Mathur, 2008, 156742) and
Siberia (Generoso et al., 2007, 155786) and tropical forest fires in the Yucatan Peninsula and Central
America (Wang et al., 2006, 157109) have affected PM concentrations in the U.S. PRB PM varies
across the contiguous Unites States (CONUS) by region and season as a function of the complex
mechanisms of transport, dispersion, deposition, and reentrainment.
Dust from the Sahara desert and the Sahel in North Africa (Chiapello et al., 2005, 156339)
affects mainly the eastern U.S.; dust from the Gobi and Taklimikan deserts in Asia (VanCuren and
Cahill, 2002, 157087; Yu et al., 2008, 157168) have the largest effects in the western U.S. but also
affect air quality in the eastern U.S. Husar et al. (2001, 024947) report that the average PMi0
concentration at 25 reporting stations throughout the northwestern U.S. reached 65 ug/m3 during an
episode in the last week in April 1998, compared to an average of 10-25 ug/m3 during the rest of
April and May. This was accompanied by visual reports of milky-white discoloration of the normally
blue sky in non-urban areas along the west coast.
PRB contributions to PM2.5, PMi0_2.5, and PMi0 can also be viewed as coming from two
conceptually separate components: a reasonably consistent "baseline" component and an episodic
component. The baseline component consists of contributions that are generally well characterized
by a reasonably consistent distribution of daily values each year, although there is variability by
region and season. The episodic component consists of infrequent, sporadic contributions from
natural high-concentration events occurring over shorter periods of time (e.g., hours to several days)
both within North America (e.g., volcanic eruptions, large forest fires, dust storms) and outside
North America (e.g., transport related to dust storms from deserts in North Africa and China and
storms at sea). These episodic natural events, as well as events like the uncontrolled biomass burning
in Central America, are essentially uncontrollable and do not necessarily occur in all years.
In-situ measurements provide evidence for the transport of anthropogenic PM from Asia on
Mt. Batchelor, OR (Jaffe et al., 2003, 052229). These data show sporadic but well correlated
increases in CO, O3, total Hg, and aerosol backscatter associated with air coming from Asia. The
ITCT-2K2 campaign also found evidence for the oxidation of SO2 to H2SO4 during trans-Pacific
transport of Asian emissions. If particulate SO42~ were to be formed in the polluted boundary layer
where it originated, it would likely be deposited prior to transport across the Pacific Ocean (Brock et
al., 2004, 156295). Thus, primary species emitted directly and secondary species formed during
transport contribute to PRB concentrations. Satellite data have provided images to track clouds of
dust and pollution across the oceans and have been used for some quantitative estimation of the flux
of material leaving continents. Yu et al. (2008, 157168) used optical thickness data to estimate
December 2009 3-139
-------
column loadings from the MODIS along with satellite assimilated wind fields to estimate the
transport of PM from Asia. Three-dimensional, global-scale CTMs have also been used to estimate
intercontinental transport of PM pollution (UNCEC, 2007, 157078) and trans-Pacific transport of
mineral dust from Asian deserts (Fairlie et al., 2007, 141923) and the Sahara Desert (McKendry et
al., 2007, 156748).
Estimates for the contribution of PBAP are highly problematic. Heald and Spracklen (2009,
190014) estimated the contribution of fungal spores to PM2.5 based on GEOS-Chem simulations of
mannitol, considered to be a unique tracer for fungal spores (Bauer et al., 2008, 189986). They
estimated an annual mean contribution of fungal spores to OC ranging from <0.1 (ig/m in the desert
Southwest to ~ 0.5 (ig/m3 in the more humid Southeast. It should be noted that these are model
derived estimates that still require evaluation against measurements in the U.S. They do, however,
provide the only quantitative estimates of PBAP concentrations across the continental U.S.
3.7.1.1. Estimates of PRB Concentrations in Previous Assessments
Estimates of PRB concentrations reported in the 1996 PM AQCD (U.S. EPA, 1996, 079380)
and earlier PM AQCDs were based in large measure on estimates by Trijonis et al. (1990, 157058)
for the National Acid Precipitation Assessment Program (NAPAP) as shown in Table 3-18. The
importance of different sources is likely to be quite different for natural background compared to
current conditions in the US, resulting in large changes in relations among different size fractions.
For example, PMi0_2.5 might be expected to dominate under certain conditions in the absence of
primary and secondary PM2.5 from anthropogenic sources. Different approaches for estimating PRB
concentrations in the western and eastern U.S. were taken in the 2004 PM AQCD (U.S. EPA, 2004,
056905). Data obtained at IMPROVE monitoring sites in the western U.S. shown in Figure 3-84
were chosen as estimates of the distribution of daily average PRB concentrations in the West because
they were thought to be among the least likely influenced by regional pollution sources especially at
the upper end of the concentration distribution. This conclusion was drawn from back trajectory
analyses and examination of the trace elemental composition at IMPROVE sites. Because of likely
unresolved contamination from pollution sources at other IMPROVE sites, it was recommended to
use averaged data from these sites throughout the West. Concentrations distributions from 1988
through 2001 can be found in Appendix 3E of the 2004 PM AQCD (U.S. EPA, 2004, 056905).
Median concentrations were ~3 (ig/m3. Little interannual variability was observed below the 90th
percentile values. However, at the upper end of the concentration distribution substantial interannual
variability was observed due mainly to forest fires and dust transport from Asian deserts. It was also
recognized that this method would likely overestimate PRB concentrations.
Table 3-18. Estimates of annual average natural background concentrations of PM2.6 and P
(ug/m3) from Trijonis et al. (1990, 157058). Estimates of PMi0.2.s were obtained
subtraction.
PM2.5 PMlO PMlO-2.5
East 2-5 5-11 <1-9
West 1-4 4-8 <1-7
December 2009 3-140
-------
Redwood f
fr*
C^nyotj'lan^s"}.
Figure 3-84. IMPROVE monitoring site locations.
Table 3-19 shows annual and quarterly average PM2.5 concentrations measured at the
IMPROVE sites shown in Figure 3-84 for 2004. Annual average concentrations tend to be slightly
higher in the East, particularly in Brigantine and Dolly Sods. When the data are broken down by
season, a more complex picture emerges. The highest concentrations in the East and Midwest are
found during the 3rd calendar quarter, whereas in the West highest quarterly averages can occur
during other quarters. As can also be seen from a comparison with values shown in Table 3-18, PM2.5
values measured in the East are much higher than the PRB concentration estimates by Trijonis et al.
(1990, 157058) for the NAPAP
December 2009
3-141
-------
Table 3-19. Annual and quarterly mean PM26 concentrations (ug/m3) measured at IMPROVE sites in
2004.
Mean January-March April-June July-September October-December
EAST
Acadia
Brigantine
Dolly Sods
4.5
9.5
9.5
3.9
8.1
6.7
4.6
11.3
9.8
6.0
11.6
15.5
3.5
7.3
5.7
MIDWEST
Voyageurs
3.8
4.1
3.1
4.2
3.6
nor
Bridger
Canyonlands
Gila
Glacier
Redwood
2.1
2.6
2.9
4.8
3.5
1.2
2.2
2.0
4.6
2.7
3.1
3.2
4.0
4.2
3.6
2.8
2.9
3.8
5.3
3.7
1.3
2.1
1.8
5.0
3.9
Thus, estimating daily average PRB concentrations in the eastern U.S. using observations is
highly problematic because of the widespread mixing of precursors and anthropogenic PM generated
in the East. Therefore, results from receptor modeling studies using PMF by Song et al. (2001,
036064) were used in the 2004 PM AQCD (U.S. EPA, 2004, 056905) for the East to separate
contributions from likely regional pollution sources from natural and imported pollution. The
"background" sources contribute about 7% to annual average PM2.5 concentrations at Underbill, VT
and about 12% at Brigantine, NJ, i.e., values between 1 and 2 ug/m3. These sites were chosen
because they were outside of urban areas making it easier to separate pollution from background
components. However, some contribution of regional anthropogenic pollution was still present.
The PM Staff Paper (U.S. EPA, 2006, 157071) adopted a different approach for estimating
PRB concentrations. This approach separated out components mainly thought to be emitted by
regional pollution sources such as SO4 ~, which are obtained directly from observations at many
more IMPROVE sites than are shown in Figure 3-84, and to use the remaining PM components in
both the East and the West to estimate PRB. Removing regional pollution from data obtained at
IMPROVE sites is problematic as it involves assumptions about the relative contributions of regional
pollution and background sources. Although sulfate in the East is mainly produced by regional
pollution sources, it is not the only component with a regional pollution source. By comparison,
sulfate is a very minor component of PM2.5 in the West, leading to substantial overestimates in
populated states like California. Annual mean estimates in the continental U.S. ranged from
2.5 ug/m3 in the Central West (ID, MT, WY, ND, SD, CO, UT, NZ, AZ) to 5.2 ug/m3 for the
Southwest Coast (most of California), the latter value likely reflecting contributions from local, non-
sulfate pollution.
In general, the methods outlined in both these documents are of limited utility for two reasons:
(1) they lack detailed spatial coverage across the whole U.S., since both methods rely on monitoring
data that are limited both spatially and temporally; and (2) PM measurements from even the limited,
remote sites used in the previous estimates of PRB can not be completely devoid of contributions
from anthropogenic PM. Because of these limitations, numerical modeling can provide superior PM
background estimates, as described just below.
3.7.1.2. Chemistry Transport Models for Predicting PRB Concentrations
CTMs can be used to estimate the PRB concentrations of atmospheric components including
PM using a "zero-out" approach in which anthropogenic emissions inside continental North America
are set to zero while global biogenic emissions and anthropogenic emissions outside continental
December 2009 3-142
-------
North America remain. Numerical modeling can provide more precision in the estimate of PRB PM
than measurements since even the most remote measurement sites like some of those in the
IMPROVE network (see the discussion in Section 3.7.1.1) will necessarily be affected by non-local
non-biogenic pollution, thereby confusing the contributions from these sources. Numerical models
are also capable of supplying estimates of PRB concentrations at much higher spatial and temporal
resolution than can be obtained by relying on measurements obtained at even the most remote
monitoring sites. In this approach, the monitoring data are used to evaluate the CTM's performance.
For this assessment, the global-scale circulation model GEOS-Chem was coupled with the
regional scale air quality model CMAQ (Section 3.6.2.2) to simulate one year of air quality data over
the CONUS in two series of runs, the first annual series with all anthropogenic and biogenic
emissions included and the second annual series with the zero-out approach employed.
The global-scale scale circulation model was set up as follows. GEOS-Chem, version 7, was
used, with modifications to include aromatic and biogenic SOA formation; emissions were computed
from a variety of sources including the Global Emissions Inventory Activity (GEIA) (Benkovitz et
al., 1996, 156267), and Emissions Database for Global Atmospheric Research, version 2 (EDGAR)
(Olivier et al., 1996, 156828: Olivier et al., 1999, 1568291
Particularized emissions in specific areas used the European Monitoring and Evaluation
Program (EMEP) for Europe (Auvray and Bey, 2005, 156237). BRAVO (Kuhns and Knipping,
2005, 156663) for Mexico, and Streets et al. (2006, 157019) for Asia. Emissions from these studies
were supplemented with data from Martin et al. (2002, 089380) for additional NOX emissions from
biofuels, lightning, and ship traffic, Bond et al. (2004, 056389) for global primary organic aerosols,
and Cooke et al. (1999, 156365) and Park et al. (2003, 156842) for U.S. primary organic aerosols.
Biomass burning emissions are not climatological, but were computed with GFEDv2 (Giglio et al.,
2006, 156469; van der Werf et al., 2006, 157084); monthly values computed using active fire
observations from MODIS; global dust fields computed off-line using GOCART (see emissions from
DEAD: http://dust.ess.uci.edu/dead/) to make annual adjustments to photolysis rates and
heterogeneous-phase chemistry.
The regional CTM was set up as follows. CMAQ, version 4.7, (excluding the dynamic coarse
mode updates) was used with the SAPRC_99 chemical mechanism and AERO5 aerosol module;
emissions were processed through SMOKE (http://smoke-model.org). version 2.4, based on the 2004
projections from the NEI with specific CEM, biogenics, and fire updates; MM5, version 3.7.4, was
used with the Asymmetric Convective Mixing, version 2.2, PEL scheme; and data nudging was used
to analyze fields for winds and temperature.
Model Evaluation
Details from evaluations of the performance of a number of CMAQ applications are given in
Arnold et al. (2003, 087579). Eder and Yu (2006, 142721). Appel et al. (2005, 089227). and Fuentes
and Raftery (2005, 087580).
In an annual simulation series for 2002 using CMAQ, version 4.6.1, in two 12-km domains for
the CONUS (Figure 3-85), predicted concentrations of summertime particulate SO4, often a major
determinant of surface-layer PM concentrations, were well-predicted by CMAQ at a 12-km grid
spacing, to within a factor of 2 at nearly every point of comparison and with R2 >0.8 across all three
national networks (CASTNet, IMPROVE and CSN); a more detailed description is included in the
2008 NOxSOx ISA (U.S. EPA, 2008, 157074). This result for CMAQ, version 4.6.1, for 2002 tracks
the generally well-predicted SO42~ concentrations found in most earlier CMAQ evaluations: see
Mebust et al. (2003, 156749). Eder and Yu (2006, 142721). and Tesche et al. (2006, 157050). Since
particulate SO42~ concentrations are strongly a function of precipitation, care must be taken to ensure
that the meteorological solution driving individual CMAQ chemical applications produces
precipitation fields with low bias as discussed by Appel et al. (2008, 155660).
December 2009 3-143
-------
2002ac m»l2«3 izkmE SCM lor June lo August 2002
a IMPROVE (2002ac me12v33 lEkmE)
fl STN(2002ac mat2v33 12
CASTNet{2002ac me!2v33 I
Ctl
o .
Monthly Average
SO4 ( ug<'m3)
IMPROVE
STN
CASTNet
CQRfl HWSE
096 084
0,83 1 57
0.9? 0 83
NMB
-11 7
0,4
-86
NME
172
201
125
Observation
Figure 3-85. 12-km EUS Summer S042~ PM. Each data point represents a paired monthly
averaged (June/July/August) observation and CMAQ prediction at a particular
IMPROVE, STN, and CASTNet site. Solid lines indicate the factor of 2 around the
1:1 line shown between them.
Wintertime participate NO3~ (Figure 3-86) and total NO3 (HNO3 + participate NO3~) (Figure
3-87) concentrations are predicted as well by CMAQ; however, NO3~ is a pervasively difficult
species to measure and model. Still, at the CASTNet nodes where the total NO3~ concentrations are
higher than they are at all but a few of the remote IMPROVE sites, CMAQ predicts concentrations
for nearly every node to within a factor of 2 and with an R2 >0.8.
A "base case" in which conditions for 2004 including all the anthropogenic and natural
sources both within and outside of continental North America was run for comparison with
measurements. APRB simulation was also run by shutting off the anthropogenic sources of primary
PM and precursors to secondary PM inside continental North America.
December 2009
3-144
-------
Figure 3-86 12-km EUS Winter NOs" PM. Each data point represents a paired monthly
averaged (December/January/February) observation and CMAQ prediction at a
particular IMPROVE and STN site. Solid lines indicate the factor of 2 around the
1:1 line shown between them.
20UZ»:_mcl2v33_12kmE TN03 (or Of cf mtwr to I rtmjarj ZOQ2
Figure 3-87. 12-km EUS Winter total nitrate (HN03 + total particulate N03~). Each data point
represents a paired monthly averaged (December/January/February) observation
and CMAQ prediction at a particular CASTNet site. Solid lines indicate the factor
of 2 around the 1:1 line shown between them.
December 2009
3-145
-------
Figure 3-88 and Figure 3-89 show monthly average concentrations, and Figure 3-90 and
Figure 3-91 show 24-h avg concentration distributions for 2004 predicted by CMAQ for the base
case and for PRB. Measurements are also included for the five western and four eastern/midwestern
IMPROVE sites shown in Figure 3-84. Wildfires could have affected the grid cell containing the
midwestern Voyageurs site resulting in the high PRB values found for the July average compared to
the PRB values for the rest of the year. The "base case" simulations tend to underestimate
concentrations throughout most western sites as shown in Figure 3-89 and Figure 3-91. These
underestimates are still within the range of a few (ig/m3. However, the base case simulation also
greatly over-predicts PM2.5 concentrations at the upper end of the distribution at the Redwoods site
(Figure 3-91). This over-prediction results from emissions from wildfires in northern California that
are included in the grid cell containing the Redwoods site, but may not have affected the site.
However, wildfires indicated by MODIS would have affected other areas either close to these sites
or could have affected other locations in between the IMPROVE sites. The simulated monthly
average PRB concentrations in the east/midwest range from a minimum of 0.6 (ig/m3 at Acadia
National Park (NP) in July to 3.7 (ig/m3 at Voyageurs NP in July. However, most values are
<1 (ig/m3. The monthly average PRB concentrations calculated for the West tend to be lower than for
the East and range from 0.2 (ig/m3 at Bridger and Glacier NPs in January and February, respectively,
to 8.7 (ig/m3 at Redwoods NP in November. Excluding values at Redwoods NP which greatly exceed
measurements, the highest monthly average concentration was 3.7 (ig/m3 at Voyageurs NP in the
East/Midwest and 2.4 (ig/m3 at Gila NP in the West.
Acadia
Brigantine
20
16;
12
8
20
16
12
8
4
1 2 3 4 6 6 7 8 9 10 11 12
1 2 3 4 5 6 7 8 9 10 11 12
Month
Dolly Sods
Voyageurs
20
16
12
8
4
0
1 2 3 4 6 6 7 8 0 10 11 12
Month
20
16
12
8
4
0
• Base
1 2 3 4 E 6 7 8 9 10 11 12
Month
,-PRB
Figure 3-88. Monthly average of PM2.s concentrations measured at IMPROVE sites in the East
and Midwest for 2004. Also shown are distributions of PM2.s concentrations
calculated by CMAQ for the base case and for PRB.
December 2009
3-146
-------
Bridgar
Canyonlands
I
12
10
8
6
4
2
0
12
10
8
i
1 6
2
1 2 3 4 5 6 7 8 9 10 11 12
Month
Gila
10
8
6
4
2
1 2 3 4 5 6 7 8 9 10 11 12
Month
1 2 3 4 6 6 7 8 8 10 11 12
Month
I
12]
10-
8
e
4
2
1 2 3 4 E 8 7 8 8 10 11 12
Month
•Monltorcd -^•^Bws
PRB
Figure 3-89. Monthly average of PM2.s concentrations measured at IMPROVE sites in the West
for 2004. Also shown are distributions of PM2.s concentrations calculated by
CMAQ for the base case and for PRB.
December 2009
3-147
-------
Aeadia
Brigantino
*•••••
I
40
35
30
25
20
16
10
6
0
mln 10 20 30 40 60 60 70 60 90 86 88 m«
mlnlO 20 30 40 60 60 70 80 90 96 99mix
PnmUlt
Dolly So*
Voyigwra
mln 10 20 30 40 60 60 70 90 90 86 88mm
10 20 30 40 60 60 70 80 90 86 99 m«
Pmindli
•PRB
Figure 3-90. Distribution of PM2.s concentrations measured at IMPROVE sites in the East and
Midwest for 2004. Also shown are distributions of PM2.s concentrations
calculated by CMAQ for the base case and for PRB.
December 2009
3-148
-------
Brldger
Canyonlindi
mn 10 20 30 40 50 60 70 80 90 96 99 max
10 20 30 40 50 60 70 80 90 96 99
Oil!
mln 10 20 30 40 60 60 70 80 80 86 89 mn
mln 10 20 30 40 50 60 70 80 80 86 88 mix
mln 10 20 30 40 60 80 70 60 90 96 99mm
• Monitored —•— DIM
• PRB
Figure 3-91. Distribution of PM2.s concentrations measured at IMPROVE sites in the West for
2004. Also shown are distributions of PM2.s concentrations calculated by CMAQ
for the base case and for PRB. Note the scale change on the y-axis for Redwoods
NP.
Table 3-20 gives the annual and quarterly average CMAQ predictions at IMPROVE sites for
the "base case" and the ratio of CMAQ predictions to the measured concentrations at those sites in
2004. CMAQ performance for the annual average concentrations and most of the seasonal averages
is very good in the East and Midwest, generally falling within 35%. In the West, CMAQ's prediction
of PM2.5 mass averages at these remote sites is generally too low in all seasons, often by 50%. Air
quality model predictions in the mountainous West are often not as good as those over the flatter
December 2009
3-149
-------
terrain in the East and Midwest because the model's grid spacing (36 km in this case) smoothes over
significant variation at the surface that results in differences at such remote sites. However, the
model's trend relative to the geospatial difference is correct: the predicted PM2.5 concentrations are
lower at the western sites than they are at the eastern sites, just as the measurements are. Table 3-21
shows corresponding annual and quarterly mean PRB PM2 5 concentrations at IMPROVE sites.
Table 3-20. Annual and quarterly mean PM2 6 concentrations (ug/m3) for the CMAQ "base case" at
IMPROVE sites in 2004.
Annual; mod/obs
Jan-March; mod/obs
Apr-Jun; mod/obs
Jul-Sep; mod/obs
Oct-Dec; mod/obs
EAST
Acadia
Brigantine
Dolly Sods
4.7; 1.04
10.2; 1.07
9.8; 1.03
5.6; 1.44
10.9; 1.35
8.3; 1.24
4.0; 0.87
10.3; 0.91
8.6; 0.88
4.6; 0.77
10.2; 0.88
14.0; 0.90
4.6; 1.31
9.4; 1.29
8.0; 1.40
MIDWEST
Voyageurs
4.0; 1.05
4.9; 1.19
2.2; 0.71
3.9; 0.93
4.9; 1.36
nor
Bridger
Canyonlands
Gila
Glacier
Redwood
1.6; 0.76
1.6; 0.62
1.6; 0.55
2.2; 0.45
4.6; 1.31
1.3; 1.08
1.9; 0.86
1.4; 0.70
1.8; 0.39
4.0; 1.48
1.6; 0.52
1.4; 0.44
2.2; 0.55
2.1; 0.50
3.0; 0.83
1.8; 0.64
1.5; 0.52
1.7; 0.45
2.1; 0.40
2.9; 0.78
1.7; 1.30
1.6; .76
1.1; 0.61
2.8; 0.56
8.4; 2.15
Table 3-22 illustrates CMAQ predictions of seasonal variation in the base case PM2.5
concentrations across regions of the CONUS, while Table 3-23 shows CMAQ predictions of the
seasonal variation in regional PRB PM2.5 concentrations. Highest base case PM2 5 concentrations
were observed for the Northeast, Southeast, and industrial Midwest, with highest concentrations
during the fall and winter (and comparably high concentrations in the summer for the Industrial
Midwest). PRB PM2 5 concentrations were highest on an annual basis in the Southeast, and peaking
during the winter. In the summer, PRB PM2 5 is roughly comparable for the Northwest and Southern
California and elevated, but slightly lower for the Southwest. These results also likely indicate the
influence of sources that are more strongly related to hot and dry conditions such as wildfires and
dust suspension.
December 2009
3-150
-------
Table 3-21. Annual and quarterly mean PM26 concentrations (ug/m3) for the CMAQ PRB simulations
at IMPROVE sites in 2004.
Annual
January -March
April-June
July-September
October-December
EAST
Acadia
Brigantine
Dolly Sods
0.70
0.77
0.79
0.76
0.86
0.88
0.76
0.91
0.83
0.65
0.70
0.75
0.65
0.63
0.66
MIDWEST
Voyageurs
1.2
0.83
0.91
2.0
0.93
nor
Bridger
Canyonlands
Gila
Glacier
Redwood
0.57
0.49
0.74
0.91
2.8
0.33
0.38
0.42
0.36
2.4
0.57
0.54
1.4
0.87
1.5
0.76
0.68
0.80
1.1
1.1
0.61
0.35
0.32
1.3
6.1
Table 3-22. Annual and quarterly mean of the CMAQ-predicted base case PM2 5 concentrations
(ug/m3) in the U.S. EPA CONUS regions in 2004.
Annual January-March
April-June
July-September
October-December
Northeast
9.76
10.74
8.38
9.55
10.38
Southeast
10.05
12.28
7.72
9.78
10.42
Industrial Midwest
11.38
12.22
9.37
11.8
12.00
Upper Midwest
6.70
8.83
4.95
5.34
7.67
Southwest
3.30
4.08
2.77
3.31
3.03
Northwest
2.72
2.49
2.21
2.71
3.44
Southern California
4.43
4.64
3.93
4.34
4.82
December 2009
3-151
-------
Table 3-23. Annual and quarterly mean of the CMAQ-predicted PRB PM2 5 concentrations (ug/m3) in
the U.S. EPA CONUS regions in 2004.
Annual January-March April-June July-September October-December
Northeast 0.74 0.85 0.78 0.67 0.68
Southeast 1.72 2.43 1.41 1.41 1.64
Industrial Midwest 0.86 0.89 0.89 0.94 0.73
Upper Midwest 0.84 0.79 0.93 0.99 0.66
Southwest 0.62 0.61 0.76 0.70 0.40
Northwest 1.01 0.48 0.81 1.42 1.32
Southern California 0.84 0.54 0.92 1.21 0.67
3.8. Issues in Exposure Assessment for PM and its
Components
The purpose of this section is to present the latest exposure assessment studies to characterize
the exposure of individuals and populations to PM of ambient origin. Such information will aid the
interpretation of epidemiologic studies described in subsequent chapters of this ISA. This
section includes descriptions of modeling and monitoring techniques used to capture personal PM
exposure, observations reported in the literature at various relevant spatial scales, observations
related to PM composition and PM in a mix of copollutants, and the effect of exposure estimates on
epidemiologic results. Attention is given to use of community-based monitors at urban spatial scales
and use of personal and microenvironmental exposure data to present how each metric can be used
in exposure assessment and what errors and uncertainties exist for each approach. Understanding of
exposure errors is important because exposure error can potentially bias an estimate of a health effect
endpoint, or increase the size of confidence intervals around a health effect estimate. Typically,
exposure error biases analyses of health effects towards the null (i.e., no relationship between
exposure and health effect).
The information presented in this section builds upon the key findings of the 2004 PM AQCD
(U.S. EPA, 2004, 056905). One key finding was that separation of total PM exposures into ambient
and nonambient components reduces potential uncertainties in the analysis and interpretation of PM
health effects data. At the time of the 2004 PM AQCD (U.S. EPA, 2004, 056905). one study reported
that individual daily values of both the total and nonambient personal PM exposure were poorly
correlated with the daily ambient PM concentrations, while individual daily values of ambient PM
exposure and daily ambient PM concentrations were highly correlated. In pooled studies (different
subjects measured on different days), individual, daily values of the total PM exposure were
generally shown to be poorly correlated with the daily ambient PM concentrations. In longitudinal
studies (each subject measured for multiple days), individual, daily values of the total PM personal
exposure and the daily ambient PM concentrations were found to be highly correlated for some, but
not all, subjects. Using the PTEAM study data, the 2004 PM AQCD (U.S. EPA, 2004, 056905) also
analyzed exposure measurement errors in the context of time-series epidemiology to show that the
error introduced by using ambient PM concentrations as a surrogate for ambient PM exposures
negatively biases the estimation of health risk coefficients by the ratio of ambient PM exposure to
ambient PM concentration. However, it was concluded that the health risk coefficient determined
using ambient PM concentrations provides the correct information on the change in health risks that
would be produced by a change in ambient concentrations.
Personal exposure to PM can vary considerably depending on PM size and composition,
source strength and proximity, season, time of day, region of the country, population density of the
environment, personal activity patterns, and ventilation of indoor environments. Table A-58 of
Annex A, which summarizes findings from U.S. panel studies of personal exposure to PM with no
December 2009 3-152
-------
indoor sources published between 2002 and 2008 broken down by region of the country, illustrates
this variability. For example, Table A-58 presents 24-h personal PM2.s exposures that range from
roughly 1-55 ug/m3 with highest exposures in the southwest and northeastern regions of the country
and during the summer season. Section 3.8 is designed to present current theory and field results
regarding exposure to PM. To illustrate the concept of personal exposure within various
microenvironments, a general exposure model is presented in Section 3.8.1. New developments in
techniques for measuring personal and indoor PM are presented in Section 3.8.2, followed by
exposure modeling techniques in Section 3.8.3. In Section 3.8.4, exposure assessment field studies in
the literature are presented. Attention is given to ambient exposure over multiple spatial scales
including near-road, in-vehicle, and indoor environments. Section 3.8.5 presents issues related to PM
composition and PM in multipollutant mixtures. Finally, implications of exposure assessment issues
for epidemiologic studies are presented in Section 3.8.6.
3.8.1. General Exposure Concepts
A theoretical model of personal exposure is presented to highlight what is measurable and
what uncertainties exist in this framework. An individual's time-integrated total exposure to airborne
PM can be described based on a compartmentalization of the person's activities during a given time
period:
ET = \Csdt
Equation 3-3
where ET = total exposure over a time period of interest, C, = airborne PM concentration at
microenvironmentj, and dt = portion of the time-period spent in microenvironmentj. Equation 3-3
can be decomposed into a microenvironmental model that accounts for exposure to PM of ambient
(Ea) and nonambient (Ena) origin of the form:
Equation 3-4
Figure 3-92 illustrates Equation 3-4. Examples of ambient PM sources include industrial and
mobile source emissions, resuspended dust, biomass combustion, and secondary formation.
Examples of nonambient sources include smoking, cooking, home heating, cleaning, and indoor air
chemistry. PM concentrations generated by ambient and nonambient sources are subject to spatial
and temporal variability that can affect estimates of exposure and resulting health effects. Exposure
factors affecting interpretation of epidemiology are discussed in detail in Section 3.8.6.
December 2009 3-153
-------
in
Ambient
Exposure
Ambient
PM, C while
Outdoors
Ambient
PMthat
Infiltrates
Indoors
Total Personal
Exposure to PM
PEM
I
Nonambient
Exposure
Indoor Sources
(cooking,
cleaning)
Personal Sources
(smoking,
hobby)
Source: Adapted with permission of Nature Publishing Group from Wilson and Brauer (2006, 088933).
Figure 3-92. Model of total personal exposure to PM as a function of ambient and nonambient
sources.
This assessment focuses on the ambient component of exposure because this is more relevant
to the NAAQS review. Ea can be decomposed into the fraction of time spent in various outdoor and
indoor microenvironments (Wallace et al., 2006, 089190: Wilson et al., 2000, 010288):
Equation 3-5
where/= fraction of the relevant time period (equivalent to dt in Equation 3-2), subscript o = index
of outdoor microenvironments, subscript / = index of indoor microenvironments, subscript o,i =
index of outdoor microenvironments adjacent to a given indoor microenvironment /', and F^j =
infiltration factor for indoor microenvironment z. Equation 3-5 is subject to the constraint S/0 +
1/i = 1, and each term on the right hand side of the equation has a summation because it reflects
various microenvironmental exposures. Here, "indoors" refers to being inside any aspect of the built
environment, e.g., home, office buildings, enclosed vehicles (automobiles, trains, buses), and
recreational facilities (movies, restaurants, bars). "Outdoor" exposure can occur in parks or yards, on
sidewalks, and on bicycles or motorcycles.
Finf represents the equilibrium fraction of the PM concentration outside the microenvironment
that penetrates inside the microenvironment and remains suspended. It is a function of the
microenvironmental air exchange characteristics and the particle properties. Assuming steady state
conditions, the infiltration factor is a function of the penetration, P, of PM (a fractional quantity
representing the portion of outdoor PM that passes through the building envelope), the air exchange
rate, a, of the indoor microenvironment, and the rate of PM loss, k, within the indoor
microenvironment: Fmf = Pal(a+K). Determination of Ea can be complicated by PM loss through
chemical and physical processes in microenvironments, and the composition of PM can be modified
during infiltration of outdoor air into microenvironments (Meng et al., 2007, 194618: Sarnat et al.,
2006, 089166).
In the context of interpreting epidemiologic studies of the effects of ambient pollutants on
human health, the association between Ea and concentrations from a central site monitor, Ca, is more
relevant than the association between ET and Ca because nonambient PM is uncorrelated with Ca, as
discussed in Section 3.8.4. In ecologic studies of large panels or cohorts, Ca is often used in lieu of
outdoor microenvironmental data to represent these exposures based on the availability of data. Thus
it is often assumed that C0 = Ca and that the fraction of time spent outdoors can be expressed
cumulatively as/0; the indoor terms still retain a summation because infiltration differs among
December 2009
3-154
-------
different microenvironments. Under these assumptions, an individual's exposure to ambient PM,
first given in Equation 3-5, can be re-expressed as a function of Ca. The following approximation has
been employed in the literature to describe ambient exposure based on these assumptions (Wallace et
al., 2006, 089190: Wilson and Brauer, 2006, 088933: Wilson et al., 2000, 010288):
Equation 3-6
Particle size, particle composition, meteorology, urban and natural topography, and other factors
determine whether or not Equation 3-6 is a reasonable approximation for Equation 3-5. Errors and
uncertainties inherent in the use of Equation 3-6 in lieu of Equation 3-5 are described in
Section 3.8.6 with respect to implications for epidemiology. If concentration measured at a central
site monitor is used to represent ambient concentration, then a, the ratio between personal exposure
to ambient PM and the ambient concentration of PM, can be defined as:
Equation 3-7
If the assumptions forming the basis for Equation 3-6 are valid, then a is the proportionality factor in
Equation 3-6:
Equation 3-8
a varies between 0 and 1. If a person's exposure occurs in a single microenvironment, the ambient
component of a microenvironmental PM concentration can be represented as the product of the
ambient concentration and F-^. Wallace et al. (2006, 089190) note that time-activity data and
corresponding estimates of Finf for each microenvironmental exposure are needed to compute an
individual's a with accuracy. If significant local sources and sinks are not captured by central site
monitors, then the ambient component of outdoor air must be estimated using dispersion models,
land use regression (LUR) models, receptor models, fine scale CTMs or some combination of these
techniques. Modeling methods are described in Section 3.8.3.
3.8.2. Personal and Microenvironmental Exposure Monitoring
The purpose of this section is to present new discoveries related to measuring
microenvironmental PM concentrations or personal exposure to PM. A review of over 100 personal
and microenvironmental PM exposure studies published since 2002 (see Table A-58 of Annex A)
reveals that the majority of the monitoring techniques in use were previously reviewed in the 2004
PM AQCD (U.S. EPA, 2004, 056905) for PM. Detailed descriptions of these methodologies are
provided in that document and therefore will not be repeated in this document. The following
sections will include only findings from 2002 or later regarding monitoring and modeling
methodologies in common use and significant advancements in understanding the capabilities and
limitations of these methods for assessment of PM exposure.
3.8.2.1. New Developments in Personal Exposure Monitoring Instrumentation
Current personal exposure sampling methodology consists largely of integrated filter sampling
using a cyclone or personal exposure monitor (PEM) to achieve a cut point at a desired particle size.
This method of sampling facilitates speciation work because the filters can be archived for chemical
and gravimetric analysis. Additionally, light scattering aerosol detection instruments, such as the
personal DataRam (pDR) and SidePak personal aerosol monitor have seen some use in personal
PM2.5, PMi0, and PMi monitoring (e.g., Lewne et al., 2006, 090556: Wallace et al., 2006, 088211).
Several researchers have noted that relative humidity causes overestimation of the particle mass in
December 2009 3-155
-------
light-scattering personal exposure monitors (e.g., Lowenthal et al., 1995, 045134; Ramachandran et
al., 2003, 195017); a correction factor has been applied to concentrations to address this issue. In the
DEARS, Williams et al. (2008, 191201) attempted to reduce humidity of the sample stream by
placing a drying column upstream of the pDR's detector. Although this addressed the humidity issue,
the drying column occasionally released particles and therefore caused artificial concentration peaks.
For this reason, Williams et al. (2008, 191201) determined that the humidity correction approach is
preferable.
One area of further development is in personal sampling of the thoracic and respirable particle
size distribution. Variations of the cascade impactor have been developed for personal sampling and
tested for use in field studies (Case et al., 2008, 155149; Lee et al., 2006, 098249; Singh et al., 2003,
156088). The model developed and tested by Lee et al. (2006, 098249) operates with a 1-um cut
point and therefore can characterize respirable particles well. The Case et al. (2008, 155149) two-
stage cascade impactor separated PM10_2.5 from PM2.5 for personal monitoring and sampled within ±
20% of a reference method. Hsiao et al. (2009, 191001) developed a mini-cyclone with a 1 urn or
0.3 um cut point for sampling accumulation mode and UFPs. The Personal Cascade Impactor
Sampler (PCIS) has the capability to sample down to a cut point of 250 nm (Singh et al., 2003,
156088). For PM2.5, the difference between the PCIS and the MOUDI cascade impactor was 11%,
while the difference between the PCIS and the SMPS-APS was only 2%. Differences between the
PCIS and the MOUDI for PM25 species compared with the MOUDI was generally higher: 11% for
SO42~, 22% for NO3~, 19% for EC, and 94% for OC. Mass was overestimated by 3%, 16%, and 31%
for PMi_o.5, PM0.5_o.25, and PM0.25, respectively, when compared with the SMPS-APS. Similarly, Case
et al. (2008, 155149) found a mass difference ranging from -11 to +10% for PMi0_2.5 with the
Personal Respirable Particulate Sampler (PRPS), and Lee et al. (2006, 098249) found a mass
difference of-6 to 0% for PM25 and -6 to -1% for PMi0 when comparing results from this device
with those from the PEM. Leith et al. (2007, 098241) redesigned the Wagner-Leith passive sampler
for measuring PMi0_2.5. In this work, the difference between a PMi0_2.5 FRM and the co-located
passive sampler was within 1 standard deviation of concentrations measured by the FRM samplers.
A number of personal PM monitors are under development as part of the National Institutes of
Health Genes, Environment, and Health Initiative (http://www.gei.nih.gov/exposurebiology/
program/sensor.asp). Funded projects for miniature personal monitors include a platform that records
real-time BC and PM concentrations and archives PM for further analysis, a badge containing a
sensor array that detects several compounds found in diesel PM, a micro-nephelometer recording
PM and endotoxin exposure, a complementary metal-oxide-semiconductor (CMOS) fitting in the
nose to measure allergen PM, and a micro-thermofluidic nanoparticle sensor. The mini-cyclone cited
above was designed to operate upwind of the micro-thermofluidic sensor (Hsiao et al., 2009,
191001). LeVine et al. (2009, 192091) and Schwartz et al. (2008, 192094) described use of the
CMOS technology for real-time DNA detection. Mulchandani et al. (2001, 191003) reviewed
amperometric biosensors used for organophosphate pesticide detection that are the basis of the diesel
detection badge; several additional articles have been published by this group that describe
applications of amperometric sensors.
3.8.2.2. New Developments in Microenvironmental Exposure Monitoring
Instrumentation
The majority of developments since the 2004 PM AQCD (U.S. EPA, 2004, 056905) regarding
microenvironmental PM characterization have involved real-time instrumentation in the UFP size
range. Because these methods are also used for ambient sampling, they are described in Section 3.4
and in Annex Table A-59.
New developments in microenvironmental sampling for exposure assessment have also
included construction, testing, and implementation of mobile environmental sampling laboratories
for PM mass, particle count, and composition, as well as other criteria pollutants (CO, SO2, NO2,
O3). These mobile laboratories typically contain a suite of real-time equipment with short sampling
intervals (e.g., 1-10 min), such as an SMPS with CPC, APS, laser photometers, and aethalometers
for aerosols; monitors for the gaseous criteria pollutants; a weather station for meteorological
variables; and a Global Positioning System (GPS) for position. Videotape or journal observations are
sometimes logged simultaneously to track local on-road sources of pollution. One key application of
mobile laboratory technology is assessment of the outdoor microscale environments and in-vehicle
microenvironments on roadways for determining exposure during on-road transportation (Pirjola et
December 2009 3-156
-------
al., 2004, 117564; Sabin et al, 2005, 087728; Weijers et al, 2004, 104186; Westerdahl et al, 2005,
086502). For instance, Sabin et al. (2005, 087728) used videotape records to determine whether BC
detected on a school bus was the result of local outdoor sources from other vehicles or
"self-pollution" from the school bus's own engine exhaust. Westerdahl et al. (2005, 086502) used the
GPS time series to determine that minima in the UFP time series corresponded to passage through
residential areas of Long Beach and Pasadena, in contrast to the pollution spikes observed along
highways. Studies have also shown that detection of PM from vehicle exhaust could be improved
through use of combined measurement results to improve statistical analysis (Ntziachristos and
Samaras, 2006, 116722).
3.8.3. Exposure Modeling
This section describes a variety of techniques used to model PM exposure. Many of these
methods are used in combination to link ambient PM levels in the atmosphere or source
characteristics to human exposure among individuals or sample populations. Recent developments in
exposure modeling are described in this subsection, and errors and uncertainties of these approaches
are described in Section 3.8.6.2.
3.8.3.1. Time-Weighted Microenvironmental Models
An individual's exposure is dictated by his or her activity patterns, as modeled by/ and/ in
Equation 3-5. Anumber of panel studies have tracked subject exposures using questionnaires, time-
activity diaries, or global positioning systems(e.g., Cohen et al., 2009, 190639; Elgethun et al., 2003,
190640; Johnson et al., 2000, 001660; Olson and Burke, 2006, 189951; Wallace et al., 2006,
089190). In many cases, the time-activity tracking is performed in conjunction with personal
exposure and/or indoor and outdoor PM concentration monitoring to estimate overall PM exposure.
Wu (2005, 086397) described a microenvironmental model of total personal exposure:
ET=fohC0+foaCa+flCl
Equation 3-9
where/,;, = fraction of time spent outdoors at home,/a = fraction of time spent outdoors away from
home, C0 = PM concentration outside the home, and C, = indoor PM concentration. In Equation 3-9,
ET can be calculated based on time-activity diary data and time-resolved PM concentration
measurements. ET can be expressed as a time-resolved value or cumulatively over a time period of
interest using this formulation. In this model, ambient and nonambient exposure cannot be separated
because it incorporates indoor concentrations that are a function of both ambient and nonambient
sources. Additionally, Equation 3-9 distinguishes ambient concentration measured at a monitor from
that measured immediately outside the home. Liu et al. (2003, 073841) found that this model
predicted elderly exposures adequately but was a poor predictor of PM2.5 exposure for asthmatic
children. Wu et al. (2005, 086397) point out that this may be due to lack of availability of the
children's time-activity data in school where children spend a substantial portion of their day. In a
study of school children's exposure patterns, DeCastro et al. (2007, 190996) computed odds ratios of
a panel subject's location within a given microenvironment using multivariate logistic models of the
indoor school, indoor home, and outdoor microenvironments. They found that (1) the city of
residence was a significant predictor of being indoors at school; (2) having an afterschool job was a
significant predictor of being indoors at home; and (3) age and having an afterschool job were
significant predictors of being outdoors. The results of the DeCastro et al. (2007, 190996) study were
designed to predict/ and/ in exposure modeling.
A second approach proposed by Wu et al. (2005, 086397) is similar in formulation to Equation
3-5 because it computes ambient PM exposure by considering the amount of outdoor PM infiltrated
indoors. This version also incorporates C0 and Ca:
December 2009 3-157
-------
Ea=fohC0+foaCa+ftFMC0
Equation 3-10
Equation 3-10 differs from Equation 3-5 because it accounts for concentrations immediately outside
the building rather than considering all outdoor exposures to be a function of that measured at a
community monitor. Factors influencing the contribution of Ea to ET may include sample population
characteristics, location of a site for microenvironmental monitoring, seasonal trends in PM
concentration, and regional differences affecting ambient concentration and infiltration.
Regression based approaches can also be incorporated into time-weighted microenvironmental
modeling. Chang et al. (2003, 053789) used data from the Scripted Activity Study and the Older
Adults Study, both conducted in Baltimore in 1998 and 1999, to compute total personal exposure
based on time-weighted microenvironmental exposures for each panel subject:
ET =
Equation 3-11
where ME = microenvironment (indoor or outdoor), /? = regression coefficient reflecting the
accuracy of the exposure estimate for a given microenvironment,^ = fraction of time performing an
activity k, and Ck = personal exposure while performing activity k. In this work, Chang et al. (2003,
053789) tested the models with hourly personal exposure data, hourly ambient concentration data,
and daily ambient concentration data. The study found that time-activity data improved estimates of
24-h PM2.5 exposure in comparison with using 24-h ambient PM2.5 data, but the use of hourly
ambient data was comparable to personal microenvironmental data in estimating exposure. When
using ambient concentration data, /? reflected infiltration for the indoor microenvironmental
estimates for a sample population. Using a similar activity-based exposure modeling approach for a
panel study in Seattle from 1999-2002, Allen et al. (2004, 190089) found that subjects' PM
exposures were best represented by modeling ambient and nonambient exposures separately because
ambient PM personal exposure was well correlated with PM concentration at central site monitors,
while nonambient PM exposure was not.
3.8.3.2. Stochastic Population Exposure Models
Population-based methods, such as the Air Pollution Exposure (APEX), Stochastic Human
Exposure and Dose Simulation (SHEDS), and EXPOLIS (exposure in polis, or cities) models,
involve stochastic treatment of the model input factors
(http://www.epa.gov/ttn/fera/human apex.html; (Burke et al., 2001, 014050; Kruize et al., 2003,
156661). These are described in detail in Annex 3.7 of the 2008 NOX ISA (U.S. EPA, 2008, 157073).
Stochastic models utilize distributions of pollutant-related and individual-level variables, such as
ambient and local PM concentration source contributions and breathing rate, respectively, to
compute the distribution of individual exposures across the modeled population. Using distributions
of input parameters in the model framework rather than point estimates allows the models to
explicitly incorporate uncertainty and variability into exposure estimates (Zidek et al., 2007,
190076). These models estimate time-weighted exposure for modeled individuals by summing
exposure in each microenvironment visited during the exposure period. The models also have the
capability to estimate received dose through a dosimetry model. The initial set of input data for
population exposure models is ambient air quality data, which may come from a monitoring network
or model estimates. Estimates of concentrations in a set of microenvironments are generated either
by mass balance methods or microenvironmental factors. Microenvironments modeled include
residential indoor microenvironments; other indoor microenvironments, such as schools, offices, and
public buildings; vehicles; and outdoor microenvironments. The sequence of microenvironments and
exertion levels during the exposure period is determined from characteristics of each modeled
individual. The APEX model does this by generating a profile for each simulated individual by
sampling from distributions of demographic variables such as age, gender, and employment;
physiological variables, such as height and weight; and situational variables, such as living in a
house with a gas stove or air conditioning. Activity patterns from a database such as Consolidated
Human Activity Database (CHAD) are assigned to the simulated individual using age, gender, and
December 2009 3-158
-------
biometric characteristics. Breathing rates are calculated for each activity based on exertion level, and
the corresponding received dose is then computed. For APEX, the PM dosimetry algorithm is based
on the International Commission on Radiological Protection's Human Respiratory Tract Model for
Radiological Protection (ICRP, 1994, 006988). and calculates the rate of mass deposition of PM in
the respiratory system (U.S. EPA, 2008, 191775). Summaries of individual- and population-level
metrics are produced, such as maximum exposure or dose, number of individuals exceeding a
specified exposure/dose threshold, and number of person-days at or above certain exposure levels.
The models also consider the non-ambient contribution to total exposure. Nonambient source terms
are added to the infiltration of ambient pollutants to calculate the total concentration in the
microenvironment. Output from model runs with and without nonambient sources can be compared
to estimate the ambient contribution to total exposure and dose.
Recent larger-scale human activity databases, such as those developed for CHAD or the
National Human Activity Pattern Survey (NHAPS), have been designed to characterize exposure
patterns among much larger population subsets than can be examined during individual panel studies
(Klepeis et al, 2001, 002437; McCurdy et al, 2000, 000782). CHAD consists of a consolidation of
human activity data obtained during several panel studies in which diary or retrospective activity
data were obtained, while NHAPS acquired sample population time-activity data through surveys
about human activity (Klepeis et al., 2001, 002437). The complex human activity patterns across the
population (all ages) for NHAPS are illustrated in Figure 3-93 (Klepeis et al., 2001, 002437). This
figure is presented to illustrate the diversity of daily activities among the entire population as well as
the proportion of time spent in each microenvironment. Different patterns would be anticipated when
breaking down activity patterns for subgroups, such as children or the elderly. With data for average
PM concentrations in each microenvironment, population exposures can be estimated from this
break-down of time-activity data.
B
O
S*
C3
^
0}
O
i-i
CO
O -—i (Nl i— i
-------
Stochastic and deterministic methods are often combined, as described below. Recently,
SHEDS has been linked with the Modeling Environment for Total Risk Studies (MENTOR) model
to expand population exposure assessment to individual risk assessment (Georgopoulos et al., 2005,
080269). In this formulation, CMAQ was used to predict initial concentrations at a coarse scale, and
then a spatiotemporal random field method (Vyas and Christakos, 1997, 156142) was applied to
interpolate the concentration to census tract scale in which exposure estimates are made. CHAD can
also be incorporated into MENTOR so that estimates of exposure are related to dose and metabolic
distributions to estimate risk of specific health impacts.
3.8.3.3. Dispersion Models
Dispersion models have been used both for direct estimation of exposure and as inputs for
stochastic modeling systems, as described above. Location-based exposures have been predicted
using models such as CALINE, AERMOD, CALPUFF, (all described in Section 3.6.2.3) or the
Operational Street Pollution Model (OSPM) for estimation of street-level PM pollution coupled with
infiltration models to represent indoor exposure to ambient levels (Gilliam et al., 2005, 056749;
Mensink et al., 2008, 155980: Wilson and Zawar-Reza, 2006, 088292). For instance, CALPUFF
was used to model transport and dispersion in lower Manhattan following the September 11, 2001
World Trade Center collapse to determine average location-based exposures (Gilliam et al., 2005,
056750). Wilson and Zawar-Reza (2006, 088292) used The Air Pollution Model (TAPM), which
integrated an emissions model with amesoscale meteorological driver, to assess PMi0 dispersion and
potential for exposure in Christchurch, New Zealand. Gulliver and Briggs (2005, 191079) used the
Atmospheric Dispersion Modeling System (ADMS) to model dispersion of "line-source" traffic
emissions in an urban environment. In a method similar to that employed by Georgopoulos et al.
(Georgopoulos et al., 2005, 080269) with SHEDS, Wu et al. (2005, 058570) used CALINE to predict
street-level concentrations of pollutants and input the results of that dispersion model into an
individual exposure model that accounts for infiltration of specific building characteristics. Wu et al.
(2005, 058570) employed CHAD to estimate the time-basis of exposures from the CALINE
predictions. With an individualized exposure approach, the model is deterministic. However,
population exposures can be estimated by performing repeated simulations using various housing
characteristics and then computing a posterior probability distribution function for exposure. Isakov
et al. (2009, 191192) developed a methodology to link a CTM (used to compute regional scale
spatiotemporally-varying concentration in an urban area) with stochastic population exposure
models to predict annual and seasonal variation in urban population exposure within urban
mi croenvironments.
3.8.3.4. Land Use Regression and CIS-Based Models
LUR models have also been developed to describe pollution levels as a function of source
characteristics (Briggs et al., 1997, 025950: Gilliland et al., 2005, 098820: Ryan and LeMasters,
2007, 156063). LUR is a regression derived from monitored concentration values as a function of
data from a combination of factors (e.g., land use designation, traffic counts, home heating usage,
point source strength, and population density). The regression is then computed for multiple
locations based on the independent variables at locations without monitors. At the census tract level,
a LUR is a multivariate description of pollution as a function of traffic, land use, and topographic
variables (Briggs et al., 1997, 025950). Originally, LUR was used for NO2 dispersion, but it was
adapted for PM2.5 exposure estimation by Brauer et al. (2003, 155702) for Stockholm, Sweden,
Munich, Germany, and throughout The Netherlands. This study found a measure of traffic density to
be the most significant variable predicting PM2.5 exposure. Ryan et al. (2008, 156064) reported on a
LUR model for childhood exposure to traffic-derived EC for the Cincinnati Allergy and Air Pollution
Study and also found traffic to be the most important determinant of diesel exhaust particle exposure.
In this case, wind direction was also factored into the model as a determinant of EC mixing. Like
deterministic dispersion models, LUR can be performed over wide areas to develop a posterior
probability distribution function of exposure at the urban scale. However, Hoek et al. (2008, 195851)
warn of several limitations of LUR, including distinguishing real associations between pollutants and
covariates from those of correlated copollutants, limitations in spatial resolution from monitor data,
applicability of the LUR model under changing temporal conditions, and introduction of
confounding factors when LUR is used in epidemiologic studies.
December 2009 3-160
-------
A GIS platform is typically used to organize the independent variable data and map the results.
The GIS software creates numerous lattice points for the regression of concentration as a function of
the covariates. For instance, Krewski et al. (2009, 191193) computed PM2.5 concentrations for the
New York City and Los Angeles metropolitan areas. For the Los Angeles analysis, the LUR was
applied at 18,000 points in the simulation domain, and an inverse distance weighting kriging method
was applied to interpolate the predicted concentration. In New York City, the LUR was applied at 49
monitors for a 3-yr model and 36 monitors for a model of winter 2000; kriging was employed only
for the purpose of visualizing the concentration between monitors. The models explained 69% and
66% of the variation in PM2.5 in Los Angeles and New York City, respectively.
GIS-based spatial smoothing models can be used to estimate PM concentration levels where
monitors are not located. Yanosky et al. (2008, 099467) described an approach to estimate
concentrations using a combination of reported AQS data and GIS-based and meteorological
covariates. Temporally stationary covariates included distance to nearest road for different PM size
fractions, urban land use, population density, point source emissions within 1 and 10 km buffers, and
elevation above sea level. Time-varying covariates included area source emissions, precipitation, and
wind speed. In this analysis, the GIS-based covariates were temporally stationary, while the
meteorological and PM monitored concentration inputs were time-varying. This approach was
applied to estimate PM2.5, PMi0_2.5, and PMi0 exposures for the Nurse's Health Study and provided
estimates of concentration at approximately 70,000 nodes with PM2.5 and/or PM10 data input from
more than 900 AQS sites with good validation of the PM25 and PMi0 models (Paciorek et al., 2009,
190090; Yanosky et al., 2008, 099467; Yanosky et al., 2009, 190114).
GIS-based methods can also be applied to integrate exposures over different
microenvironments. For example, Gulliver and Briggs (2005, 191079) described development of the
Space-Time Exposure Modeling System (STEMS) to model PMiq concentration. STEMS is a
multipronged model that links traffic emissions estimates, dispersion, background PM estimates, and
time-activity data within a GIS framework to create exposure estimates. Traffic emissions estimates
and meteorological parameters are input into the ADMS dispersion model, which along with
background PM measurements, are used to create hourly point estimates of PM concentration. Based
on the time-activity data, the concentration estimates were then used to calculate exposures to traffic-
related PMio along a commuting path while an individual is in transit. PMi0 was used by Gulliver
and Briggs (2005, 191079) because ADMS had not yet been validated for smaller PM size fractions.
In an analysis of the sensitivity of included variables, Gulliver and Briggs (2005, 191079) showed
the model to be most sensitive to fluctuations in local meteorology followed by sudden vehicle speed
reductions of 10 km/h. The STEMS model was primarily designed to model exposures during
transit, but the authors state that this technique can be applied to modeling other microenvironmental
exposures.
Source proximity is sometimes used as a covariate in GIS-based regression models. For
instance, Baxter et al. (2007, 092725) predicted indoor exposure to PM2 5, EC, and NOX based on
distance to roadways, indoor source characteristics, window opening, and ambient concentrations in
the Boston metropolitan area. In this effort, Baxter et al. examined a variety of factors estimated
using GIS including roadway density, roadway length, average daily traffic, and population density
to determine which variables were significant predictors. They found that point estimates of PM25
were largely influenced by regional ambient PM2 5 while EC estimates were more influenced by
local mobile sources. However, Baxter et al. (2008, 191194) found no association between distance
to a bridge toll booth station and indoor EC concentration in Detroit homes when studying the
impact of diesel emissions from traffic on the Ambassador Bridge as part of the Detroit Exposure
and Aerosol Research Study (DEARS). Being located downwind of the booth, however, was a
significant predictor of indoor EC concentration. Corburn (2007, 155738) tested two distinct
modeling approaches, the cumulative air toxics surface (CATS) and the U.S. EPA's National Air
Toxics Assessment (NATA) to determine how these approaches can yield estimates of human
exposure to diesel exhaust and 33 air toxics for environmental impact assessment. The CATS
approach included an exposure term incorporating source density and distance to source, and the
sources include traffic as well as bus depots and transfer stations, airports, and industrial point
sources. Corburn's results demonstrated that robust land use data can provide an approximation for
urban exposures, although he cautioned that such estimates should not supersede environmental
monitoring. In using these approaches, Huang and Batterman (2000, 156572) warn that geographic
divisions must be sufficiently small to avoid inter-zone variability in source and exposure
characteristics. Moreover, the HEI Report on Traffic Related Health Effects (2009, 191009)
December 2009 3-161
-------
discourages use of source proximity as a surrogate for traffic exposure in epidemiologic studies
because it is not specific to particular pollutants and can be subject to confounding factors such as
SES.
3.8.4. Exposure Assessment Studies
Table A-61 in Annex A lists exposure assessment studies performed in the U.S. by region of
the country with personal, microenvironmental, and ambient mass concentrations presented (note
that chemical speciation data, where available, are discussed below). The majority of urban-scale
studies focus on PM2.5 because PM2.5 concentrations are more homogeneously distributed. Studies of
microscale to neighborhood scale dispersion more commonly include data on UF and thoracic coarse
PM in addition to PM2.5 because they travel over shorter distances from the site of generation, as
described in Section 3.5. Some of these studies present the outdoor concentration measured outside
the test building, while others use ambient concentration obtained from a community site monitor.
As would be expected, there is considerable variability within and across regions of the country with
respect to indoor exposures and ambient concentrations. Furthermore, some regions are represented
by only one or two studies, while other regions have many studies. Most studies have been
conducted in only one or two metropolitan areas. Thus, the results presented may not be broadly
representative.
Results of these studies highlight the uncertainties surrounding various estimates of the
ambient contribution to personal exposure. This variation can be attributed to a number of factors,
including PM size distribution, scope and magnitude of microenvironmental sources, proximity to
microenvironmental sources, ambient concentrations of PM, percentages of time spent in various
microenvironments, the age and condition of indoor microenvironments, natural and urban
topography, and outdoor meteorology. Errors in exposure estimation are linked to the spatial scale of
concentration measurements because pollutant transport and dispersion varies over different spatial
scales as a function of the many factors mentioned in the previous sentence. Findings related to
identifying the ambient components of personal exposure and modes of PM infiltration indoors are
discussed in the subsequent subsections with respect to multiple spatial scales.
3.8.4.1. Micro-to-Neighborhood Scale Ambient PM Exposure
Near-Road Exposures
Sections 3.3 and 3.5 describe the physical and chemical composition of traffic emissions as
well as characterization of the plume away from roads. Table 3-24 contains data from recent studies
comparing outdoor personal exposure to fixed site monitors. Only studies where samples were
obtained outdoors and compared with a community-based ambient monitoring site were included
because indoor microenvironments have penetration losses that affect the comparability of the
results. Note that some of these studies included enclosed transportation microenvironments (e.g.,
cars, buses, subways), but all studies examined personal exposure in the outdoor microscale
environment. Also note that studies must be reviewed cautiously because most used different
instrumentation for personal, microenvironmental, and ambient measurements, and measurement
artifacts related to each instrument may differ. The Violante et al. (2006, 156140) study showed that
outdoor personal exposure to PMi0 was significantly higher than fixed community-based ambient
PMio measurements in downtown Bologna, Italy. Likewise, the Kaur et al. (2005, 088175). Kaur
et al. (2005, 086504). and Adams et al. (2001, 019350) studies showed PM2.5 measurements to be
significantly higher than fixed community-based ambient PM2 5 monitoring site measurements in
central London, U.K. Kinney et al. (2000, 001774) performed personal exposure monitoring on
study volunteers on streets in Manhattan and showed that PM2 5 concentrations were not significantly
different from ambient PM2 5 measurements; this is more consistent with the urban-scale
homogeneity in concentration of PM25.
Morwaska et al. (2008, 191006) stated that UFP number concentrations in the near-road
environment were roughly 18 times higher than in a non-urban background environment, while
measured concentrations in street canyons and tunnels were 27 and 64 times higher, respectively
December 2009 3-162
-------
than background. This suggests that trapping of sources in a semi-enclosed environment can lead to
higher UFP exposures. Additionally, fresh emission of short-lived UFPs would explain substantially
higher concentrations near the site of emission. By sampling UFP number concentrations at multiple
sites in Los Angeles, Moore et al. (2009, 191004) demonstrated five- to seven-fold differences
between concentrations measured directly next to a freeway and an oceanside site during morning
rush hour with substantial variability among sites throughout the day. When comparing sampling
campaign data for clear weather and rainy days next to the 1-710 freeway in Los Angeles,
Ntziachristos et al. (2007, 089164) found that particle number concentration obtained with a CPC
was 2.4 times higher in clear weather than when raining; particle surface area was 3.7 times larger in
clear weather; and, black carbon concentration was 1.7 times higher in clear weather. However,
SMPS data reported for rainy day particle number concentrations were almost 29 times higher in this
study. Likewise, Zhou and Levy (2007, 098633) noted in a meta-analysis of near-road studies that
the concentrations are generally elevated within 300-400 m of a roadway for EC and UFPs. Kinney
et al. (2000, 001774) showed EC to increase linearly with increasing traffic counts and large spatial
variations in two sites that had concentrations significantly higher than ambient measurements.
These observations suggest caution should be taken regarding the representativeness of community
averaged monitoring data for assessing exposures.
Particle chemistry is also an important consideration, because exposure may differ among PM
components. Farmer et al. (2003, 089017) found that exposure to particle-bound PAHs, including
benzo[a]pyrene, can be 2-3 times higher among those routinely exposed to outdoor traffic emissions
(e.g., police, bus drivers) compared with control subjects. Particle-bound PAH exposure can also
vary with vehicle operation. For example, Kinsey et al. (2007, 190073) estimated from continuous
idling and restart school bus operating conditions (without retrofitting) that over a 10-min period of
waiting at a bus stop, continuous idling resulted in exposure to 33% more particle-bound PAH than
in the case where the bus was restarted 2 min into the simulation and idled for 8 min. Continuous
idling produced approximately 34 times more particle-bound PAH than in another scenario where
the bus was off for 10 min then restarted.
December 2009 3-163
-------
Table 3-24. Examples of studies comparing near-road personal exposures with fixed site ambient
concentrations.
Reference
and Site
Ambient
monitors
Personal monitors m^™.
arJabTeT'' Ambientv- Personal Association
Primary Findings
Violante et al.
(2006,
1561401
Fixed PM10 and Active pump with PM10
benzene monitoring PEM, passive sample for
station (method not benzene desorbed and
Bologna, Italy specified). analyzed by GC-MS.
Localized traffic density
(vehicles/h);
Meteorology (wind
speed, wind direction,
visibility, relative
humidity).
Personal: 185.10+ 38.52 pg/m3
Fixed: 43.56 ± 24.10 pg/m3 (p<0.0001)
Fixed PMio correlated with
multivariate model of traffic and
meteorology, but not personal
PMi0; relationship between
benzene and PM10 not
explored.
Kaur et al.
(2005,
0865041
London, U.K.
Fixed TEOM for
PM2 5 and fixed CO
monitor at ambient
and curbside sites.
High flow personal
samplers for PM2 5,
P-Trak monitors for UFP,
LanganT15andT15v
for CO.
Exposures stratified by
mode of transport
(walk, cycle, bus, car,
taxi).
Average PM25 sampled by TEOM was 3
times lower than average personal PM25
sample, and 8 times lower than maximum
personal PM25 sample.
PM2 5 exposures during walking
significantly lower than during
car and taxi rides, UFP
exposures during walking
significantly lower than bus and
car rides, cycling exposures to
PM2 5 and UFP not significantly
different from those on bus,
car, or taxi.
Kaur et al.
(2005,
0881751
London, U.K.
Fixed TEOM for
PM2 5 and fixed CO
monitor at ambient
and curbside sites.
High flow personal
samplers for PM2 5
analyzed post-sample
for reflectance for EC,
P-Trak monitors for UFP,
LanganT15andT15v
for CO.
Volunteers walking at
set times and directions
along Marylebone Rd in
London.
Fixed vs. personal PM25: slope = 0.29,
R = 0.6; personal PM25 measurements
were >2 times background levels and
more than 15 pg/m greater than curbsidi
Median values: (ug/m3)
Adams et al.
London, U.K.
Kinney et al.
(2000,
0017741
New York City,
NY (Harlem)
Fixed TEOM for
PM2 5 and fixed CO
monitor at ambient
and curbside sites.
Ambient site filter in
greased impactor
with pump for
PM25; absorbance
testing on filter for
EC.
High flow personal
samplers for PM2 5.
Three high traffic sites
filter in greased impactor
with pump; absorbance
testing on filter for EC.
Exposures stratified by
mode of transport
(cycle, bus, car,
subway).
Localized traffic density
(vehicles/h).
Cycle
Bus
Car
Subway
Fixed
Curb
Mean values
Sitel
Site 2
SiteS
Ambient
Summer
345
390
37 7
247.2
15
24
: (ug/m3)
PM,,
45.7(10.1)
47.1 (16.4)
36.6(10.8)
38.7(10.9)
Winter
235
389
337
157.3
13
37
EC
6.2(1.9)
3.7 (0.6)
2.3 (0.9)
1.5(0.5)
Pedestrian exposures were
measurements. Results
a indicate that exposure declined
" up to 10% from curbside to
building edge within a street
canyon.
Exposures were 2.3-16.5 times
higher than ambient and
1.4-10.3 times higher than
curbside during summer.
During winter, only subway
exposures were appreciably
higher (4.3 times) than
curbside.
PM2 5 at high traffic sites was
not significantly higher than
ambient; EC was significantly
higher than ambient at 2 sites.
EC increased linearly with
traffic counts.
In-Vehicle and In-Transit Exposures
In-vehicle pollution has been identified in various studies as a source of exposure to PM2 5,
o and UFPs (Briggs et al., 2008, 156294; Diapouli et al., 2007, 156397; Fruin et al., 2008,
097183; Gomez-Perales et al., 2004, 054418; Gomez-Perales et al., 2007, 138816; Gulliver and
Briggs, 2004, 053238; Gulliver and Briggs, 2007, 155814; Rossner et al., 2008, 156927; Sabin et
al., 2005, 087728). Results from recent studies are provided in Table A-60 of Annex A. In many of
these studies, in-vehicle exposures are shown to be comparable to or less than that of walkers on the
same route. Typically, in-vehicle exposures were also higher than community-based ambient monitor
concentrations for TSP and PMio (Diapouli et al., 2008, 190119). Curbside measurements of UFPs
and PM2.5 obtained at a fixed site in the Kaur et al. (2005, 088175; 2005, 086504) studies were
generally lower than exposures during transit, including during walking and cycling. In contrast, the
Adams et al. (2001, 019350) study demonstrated that fixed site PM2.5 concentrations were higher
than curbside during the summer and lower than curbside during the winter. As particle size
decreased to the fine and UF range, less difference between in-vehicle and ambient concentrations
was observed for PM mass or count, with the exception of the Diapouli et al. (2008, 190119) study
where in-bus UFP concentrations were several times higher than indoor or outdoor residential and
school concentrations.
December 2009
3-164
-------
Fruin et al. (2008, 097183) and Westerdahl et al. (2005, 086502) observed that in-vehicle UFP
concentrations increased for freeways in comparison with arterial roads. They estimated that 36% of
exposure to UFPs occurred during a total daily commuting time of 1.5 h (6% of the day) in Los
Angeles; 22% of total exposure occurred during 0.5 h spent on freeways. Gong et al. (2009, 190124)
demonstrated that UFP deposition rate increased with decreasing particle size (down to ~30 nm) and
increasing surface area inside the vehicle, where deposited PM on the seats and dashboard can be
resuspended and then inhaled or ingested. UFP deposition rate also rose slightly with increased
number of passengers. Zhu et al. (2007, 179919) found that in-vehicle UFP counts were 85% lower
than outdoors when the fan was operating in recirculation mode. They estimated that a 1-h commute
(4% of the day) accounts for 10-50% of daily exposure to UFPs generated by traffic. Based on the
American Time Use Survey estimation of an average of 70.2 min spent in vehicles per person each
day (U.S. Bureau of Labor Statistics http://www.bis.gov/tus/), cumulative in-vehicle exposure can
become important.
In a study of PM2.5 exposure on school buses, Adar et al. (2008, 191200) found that PM2.5 on
school buses was 2 times higher than on-road levels and 4 times higher than central site
measurements. Sabin et al. (2005, 087728) demonstrated for school buses that emission control
technologies had a significant impact on in-bus concentrations of black carbon mass, and Hammond
et al. (2007, 190135) demonstrated significant reductions of particle number concentration measured
for 0.02-1 um particles when comparing buses using clean diesel or retrofits compared with non-
retrofitted buses. Although not tested here for other vehicle types with respect to PM, these findings
suggest that a portion of in-vehicle concentrations are due to self-pollution, defined by Behrentz et
al. (2004, 155682) as the fraction of a vehicle's own exhaust entering the vehicle microenvironment.
Behrentz et al. (2004, 155682) tested self-pollution with school buses using SF6 tracer gas and
demonstrated that 0.3% of in-vehicle air comes from self-pollution, and that this number was
roughly 10 times greater than in-vehicle concentrations related to self-pollution on newer buses. The
Behrentz et al. (2004, 155682) study also measured EC and particle-bound PAH and found that 25%
of the variability in EC concentration was related to self-pollution. Adar et al. (2008, 191200)
estimated that 35.5% of PM25 mass on school buses was from self-pollution. These findings are
important for exposure estimation when partitioning local and ambient sources of pollution during
transport in vehicles.
3.8.4.2. Ambient PM Exposure Estimates from Central Site Monitoring Data
The following paragraphs describe studies that estimate personal exposure to ambient PM
from central site monitoring data. As shown in Figure 3-93, the large majority of an individual's time
is spent indoors. Although calculation of infiltration and indoor personal exposure is an important
part of this assessment, such exposures are described with respect to central site monitors. Assessing
population-level exposure at the urban scale is particularly relevant for epidemiologic studies, which
typically provide information on the relationship between health effects and community-averaged,
rather than individual, exposure.
Indoor or other nonambient sources could significantly affect assessment of a person's total
exposure to many pollutants. For this reason, many studies use PM components to estimate
infiltration of ambient PM to indoor environments. Wilson et al. (2000, 010288) first proposed that
SO42~ could be used as a tracer of the ambient PM25 infiltration rate. Sarnat et al. (2002, 037056)
also noted that it is reasonable to assume that the size distribution of ambient SO4 ~ particles is
sufficiently similar to the size distribution of ambient PM2 5, and therefore that the ambient SO42~ to
personal SO42~ ratio is an acceptable surrogate for the ratio of the ambient PM25 exposure to the
ambient PM25 concentration. Sulfate has been used this way in several studies, including Ebelt et al.
(2005, 056907). Wallace and Williams (2005, 057485; 2006, 089190) and Wilson and Brauer (2006,
088933). For this method to be successful, indoor or other nonambient sources of the tracer must be
small compared to ambient sources over the period of sampling. Wilson and Brauer (2006, 088933)
noted that environmental tobacco smoke and tap water used in showers or humidifiers are indoor
sources of SO42~. Other concerns in using SO42~ as a tracer for PM2 5 arise because SO42~ tends to be
concentrated in the accumulation mode and thus it might not capture any coarse PM found in the
upper end of the PM2 5 distribution, which can include larger particles in the tail end of the coarse
mode (Wallace and Williams, 2005, 057485). Strand et al. (2007, 157018) suggested that Fe be used
as an additional tracer to correct for the infiltration of larger PM25 particles. Their study took place in
Denver, where indoor sources of Fe were small. However, there could be more substantial
December 2009 3-165
-------
contributions from tracking iron in soil indoors in other locations. The spatial variability of Fe is also
larger than that of PM2.5 across urban areas. Volatilization of nitrate or organic compounds after
infiltration of PM2.5 indoors could lead to bias in exposure estimates (Sarnat et al., 2006, 089166).
This could be a large problem in areas in which PM contains a large semi-volatile component.
Figure 3-94 shows total exposure to SO42~ as a function of measured ambient SO42~
concentration. Figure 3-95 shows estimated ambient exposure to PM2.5 as a function of measured
ambient PM2.5 concentration, where ambient personal exposure is calculated from the ambient
exposure factor for SO42~. Close agreement between these figures can be observed. Figure 3-96
shows total exposure to PM25 as a function of measured ambient PM25 concentration. However, the
total exposure to PM2 5 shows virtually no association with ambient PM2 5 because it contains
nonambient contributions to PM25.
LU
*
2345
Ambient Concentration, C50
-------
8-24
Figure 3-96.
0 5 '0 15 20 25 30
Ambient Concentration, C2 5 (pg/rn3)
Source: Reprinted with Permission of Nature Publishing Group from Wilson and Brauer (2006, 0889331
Total exposure to PM2.6 as a function of measured ambient PM2.6 concentration,
from the Vancouver study. Vancouver, British Columbia, April-September 1998,
with 16 non-smoking subjects aged 54-86 yr.
The estimated ambient exposure to PM2.5 is well correlated with measured ambient PM2.5
concentration with zero intercept, implying that nonambient sources were minor. This technique
works well in areas where SO4 " is a regional pollutant, because its spatial variability is small(Kim et
al., 2005, 083181; U.S. EPA, 2004, 056905). Wilson and Brauer (2006, 088933) reported that the
pooled Pearson correlation coefficient was 0.79 for personal ambient exposures (estimated by the
tracer element method) vs. ambient concentrations of PM2.5, and it was 0.001 for personal
nonambient PM2.5 exposure vs. ambient concentrations. Strand et al. (2006, 089203) conducted an
exposure study in Denver (2002-2004) for 6-12 year-old school children. Up to 10 personal exposure
samples were collected on each day, and ambient concentrations were measured simultaneously at a
fixed site located at the school. The daily average personal SO42" exposure was strongly associated
with ambient SO42" concentration (r = 0.96, 120 > N > 100). Koutrakis et al. (2005, 095800) reported
the median Spearman correlation coefficients between personal SO42" exposure and ambient SO42"
concentration were above 0.60 during both winter and summer in Boston and Baltimore (15 subjects
with 12 consecutive measurements during each season in both Boston and Baltimore). For another
Baltimore cohort (15 senior subjects with up to 23 consecutive measurements for each person),
Hopke et al. (2003, 095544) reported that the median Pearson correlation coefficient between
personal exposure to the SO42" factor and the ambient SO42" factor was 0.93 (ranging from 0.56 to
0.98 for different subjects), while the median Pearson correlation coefficients were 0.25 for the
crustal factor (ranging from -0.46 to 0.66) and 0.22 for a factor whose origin was not identified
(ranging from -0.19 to 0.88). The inferences drawn from using the SO42~ component of PM25 as an
indicator for personal exposure to ambient PM2 5 may apply in areas where SO42~ is a minor
component of PM25 or in the absence of significant nonambient sources of SO42~ (Sarnat et al., 2001,
019401).
Source apportionment techniques could also be used, in principle, to derive ambient personal
PM25 exposures. They would be especially useful in areas where the application of a tracer method
might be problematic. Hopke et al. (2003, 095544) noted that four outdoor factors (NH4NO3~ and
(NH4)2SO4, secondary SO42", OC, motor vehicle exhaust) would constitute an estimate of the
personal ambient PM25 concentration. However, the data used in this portion of the analysis were
obtained only with fixed monitors and did not include measurements made by PEMs. They also used
the Multilinear Engine to derive factors that were required to contribute jointly to central indoor and
outdoor, individual apartment, and PEM samples of a panel of residents. Hopke et al. (2003, 095544)
used PMF to derive source contributions to community, outdoor, and indoor PM exposures at a
retirement facility in Towson, MD. Hopke et al. (2003, 095544) found three sources: SO42~,
unknown (perhaps combustion related, according to the authors), and soil, jointly contributing 46%,
13%, and 4% of PM25 to the PEM samples, respectively. Further source resolution was not possible
because there was a lack of data for a number of components in the PEM samples. The largest and
most clearly identified contribution to personal exposure was from the SO42" factor. This study also
December 2009
3-167
-------
determined that a few minor indoor and personal activity sources contributed <10% of the ambient
SO42" source to personal exposures.
Wilson and Brauer (2006, 088933) presented an adaptation of the SO42~ method for estimating
exposure to PMi0_2.5. a is computed based on the SO42~ method as the ratio of exposure to ambient
SO42 (as measured by a personal monitor) to ambient SO42~ concentration. Then, knowing an
individual subjects' time diary and the penetration and loss properties of SO42~, the air exchange rate
for an individual location can be calculated from Equation 3-5 and Equation 3-7. Finally, the
penetration and loss rates of PMi0_2.5 from the PTEAM database (Ozkaynak et al., 1996, 073986) can
be input into the individual exposure model along with the individual activity pattern and residential
air exchange rate to compute the ambient PMi0_2.5 exposure factor and the ambient PMi0_2.5 exposure
if PM 10-2.5 concentration is measured; in the Wilson and Brauer (2006, 088933) paper, PMi0_2.5 was
estimated from ambient PMi0 and PM2 5 concentrations. Given that PMio-2.5 deposits more readily
and therefore disperses over a shorter distance than PM2.5, it is possible that use of ambient PM10_2.5
concentration may incur more error than in using this method for PM2 5. Ebelt et al. (2005, 056907)
observed in a Vancouver, Canada panel study that the correlation between ambient PMi0_2.5 exposure
and ambient PMi0 exposure (r = 0.72) was lower than the correlation between ambient PM25
exposure and ambient PMi0 exposure (r = 0.92). This is attributed to both a smaller Finf for PMi0_2.5
and PM25 comprising a greater fraction of the PMi0 for the Vancouver study. In this study, PMio_2.s
mass concentration was calculated from the difference between ambient PM10 and PM2 5 mass
concentration.
Wilson and Brauer (2006, 088933) state that their methodology for computing the ambient
exposure factor based on the PM25 SO4 ~ method can be applied to PM in the 0.1-0.5 (im size range.
Little SO42~ mass is found below 0.1 (im, so the SO42~ tracer method would not be applicable for
UFPs. Given the short atmospheric lifetime of UFPs resulting from particle growth and evaporation
processes, primary UFPs are most prevalent at microscale rather than at an urban spatial scale
(Sioutas et al., 2005, 088428). Moore et al. (2009, 191004) found substantial spatial, hourly, and
daily variability in UFP concentration in a saturation study of Los Angeles. Moore et al. (2009,
191004) and Harrison and Jones (2005, 191005) also found that UFPs and PM2 5 measurements were
poorly correlated at the monitoring sites.
3.8.4.3. Infiltration
Finf varies substantially given a vast array of conditions, and it can best be modeled
dynamically based on a distribution of air exchange and deposition or other UF, accumulation mode,
fine, and coarse PM loss rates rather than a single value (Bennett and Koutrakis, 2006, 089184;
Wallace et al., 2006, 089190). Given that air exchange rates within a building vary as a function of
ambient temperature and pressure, Fjnf is subject to seasonal and regional changes (Meng et al.,
2005, 058595; Sarnat et al., 2006, 089166; Wallace and Williams, 2005, 057485). These factors
make Finf a more accurate descriptor of infiltration than a simple I/O ratio because the I/O ratio also
includes contributions from indoor sources in addition to PM that infiltrates from outdoors. Wallace
et al. (2006, 089190) identified several significant factors affecting Fjnf, including window opening,
age of an indoor microenvironment, number of occupants, location on a dirt road, dryer usage, and
air conditioning usage. This term becomes even more complex when one considers transformation of
the size distribution and chemical composition of the PM through chemical reactions on the particle
surface, agglomeration, growth, and evaporation given that Finf depends on particle size (Keller and
Siegmann, 2001, 025881). Finf for PM is influenced by physical mechanisms, such as Brownian
diffusion, thermophoresis, and impaction, all of which are functions of particle size (Bennett and
Koutrakis, 2006, 089184: Tung et al., 1999, 049003). These differential effects are summarized
below. Recent studies on infiltration are summarized in Table A-64 of Annex A. Fjnf and I/O are
listed where available, although it is recognized that I/O is not as meaningful a descriptor but
provides an approximation of Finf.
A number of studies have examined the impact of season on PM infiltration. Season is
important because it affects the ventilation practices used (e.g., open windows, air conditioning or
heating use) and ambient temperature and humidity conditions influence the transport, dispersion,
and size distribution of the PM. Pandian et al. (1998, 090552) found that nationwide residential air
exchange rates vary by season as: summer > spring > winter > fall with summer air exchange
roughly 1.5-2 times greater than average air exchange rate for the entire year because the rates are
driven by home air conditioning and heating usage. Allen (2003, 053578) provided information on
December 2009 3-168
-------
the range and distribution of Finf for PM2.5 at 44 residences in Seattle. The mean F^ was calculated
using light scattering measurements in a recursive mass balance method with no species data. For
all sampling days, F^f (± SD) was 0.65 ± 0.21. Differences in infiltration were observed for the
heating season (0.53 ± 0.16), when windows would be expected to be closed, and for the non-heating
season (0.79 ± 0.18). Residences with open windows had a meanFinf of 0.69 vs. 0.58 for residences
with closed windows. The authors combined the light scattering results with indoor and outdoor
sulfur measurements to estimate that 79 ± 17% of indoor PM25 was generated outdoors. This study
provides important data on the distribution of residential Fjnf values and illustrates the magnitude of
the effect of season and window position on infiltration rates. Barn et al. (2008, 156252) and Baxter
et al. (2007, 092725) also noted that window opening was an important variable. Barn et al. (2008,
156252) foundFjnf of 0.61 ± 0.27 for 13 homes during summer and 0.27 ± 0.18 for 19 homes during
winter in Canada for PM2.5 from forest fires and wood smoke.
Likewise, location could impact residential ventilation practices and infiltration. Using the
SO42" method for estimating PM2.5 infiltration, Cohen et al. (2009, 190639) noted differences in
median infiltration among eight areas (including three comprising the Los Angeles region and two
comprising the New York City region). Indoor-outdoor SO42~ ratio was noted to be highest in New
York City (median: 0.85) and Los Angeles (median: 0.84) and lowest in St. Paul (median: 0.54).
Pandian et al. (1998, 090552) observed that residential air exchange rates vary by region as:
southwest > southeast > northeast > northwest, which reflects regional use of air conditioning. Sarnat
et al. (2006, 089166) noted differences in PM2 5 infiltration between coastal and inland residences,
although these differences were not statistically significant.
Differential Infiltration Related to PM Size
Differential infiltration as a function of particle size has been observed to occur. Infiltration
factors for particle diameters ranging from 20 nm to 10 um were measured using continuous SMPS-
APS monitoring in Boston by Long et al. (2001, 011526) during summer and fall for nighttime
periods, when personal activity patterns would be less likely to generate indoor PM. The maximum
infiltration factor was reported for particles between 80 and 500 nm to range from 0.8 to 1.0.
Summer values were uniformly higher than fall values, consistent with higher observed air exchange
rates. The infiltration factor decreased with size above 500 nm, reaching 0.1-0.2 for 6-10 um
particles. Particles smaller than 80 nm also were reported to have lower infiltration factors. This
demonstrates the size dependence of PM infiltration, which has been further studied by recent
investigators. Sarnat et al. (2006, 089166) examined infiltration as a function of particle size and
found that I/O varies by particle diameter, as measured by a SMPS-APS system. Figure 3-97
presents I/O values for size fractions ranging from 0.02 to 10 urn. The maximum infiltration was
observed around the accumulation mode (0.1-0.5 um), with I/O = 0.7-0.8. Reduced infiltration was
observed for coarse-mode PM (0.1-0.2 for Dp = 5-10 urn) and, to a lesser extent, UFPs (0.5-0.7 for
Dp = 0.02-0.1 um). This is consistent with increased removal mechanisms for those size fractions.
Deposition is caused by settling for coarse-mode particles. Deposition of UFP can occur by diffusion
leading to agglomeration into larger particles and subsequent settling, as well as losses to walls.
December 2009 3-169
-------
o
i
j
c
,,*
r'
/
. -4
t — {
1
\
N<
1
\
\
\
0.1
(,
X|
h— _.
\
\
V
H|-4
1 1
Particle Size (pm)
Source: Reprinted with Permission of Air & Waste Management Association from Sarnat et al. (2006, 0891661.
Figure 3-97. Finf as a function of particle size.
3.8.5. Multicomponent and Multipollutant PM Exposures
3.8.5.1. Exposure Issues Related to PM Composition
Annex A presents exposure studies that include chemical speciation data in Table A-62. Some
of these studies focused on SO42~, NO3~, or carbonaceous aerosols (EC, OC, particle-bound PAHs),
while others measured concentrations of trace elements from crustal (Ca, Fe, Mn, K, Al, S, Cl in
salt), mobile (Al, Ca, Fe, K, Mg, Na, Ba, Cr, Cu, Mn, Ni, Pb, S, Ti, V, and Zn), or industrial
(particle-bound Hg, Cl, V, Zn, Ti, Cu, Pb) sources. A number of source apportionment studies have
been performed over the last five years to determine the contribution of outdoor sources to indoor
and personal PM constituents.
Source apportionment studies by Kim et al. (2005, 083181). Hopke et al. (2003, 095544) and
Zhao et al. (2006, 156181) have shown that secondary SO4Z~ provides the largest ambient
contribution to personal and indoor exposures. These studies took place in Baltimore, MD and
Raleigh/Chapel Hill, NC. In a source apportionment study in Seattle, vegetative burning was the
most significant source of outdoor origin (Larson et al., 2004, 098145). Zhao et al. (2007, 156182)
performed a source apportionment study of personal exposure to PM2.s among residents in Denver
and also observed lower contributions from secondary SO42~ in comparison with motor vehicle
emissions and secondary NO3~. This suggests that personal exposure to SO42~ in parts of the West is
lower than in the Mid-Atlantic. These observations are consistent with the composition distribution
shown in Figure 3-17 and Figure 3-18. Viana et al. (2008, 156135) selected 4 sites of varying
population density to represent exposures of pregnant subjects in an early childhood epidemiologic
study. Viana et al. (2008, 156135) analyzed PM2.5 and PMi0 samples for several species along
urban-to-rural gradients centered in Valencia, Spain and found gradients for both size fractions in
anthropogenically-generated SO42~, OC, EC, NO3", Fe, and NH4+, but not in mineral species.
December 2009
3-170
-------
Combined, these findings suggest urban- and regional-scale variation in species composition can
influence exposure estimates. Personal PM exposure studies including source apportionment analysis
along with chemical speciation are presented in Annex A, Table A-63.
Source apportionment for carbonaceous aerosols is complicated by the fact that they can be
derived from indoor and outdoor combustion sources. Carbonaceous aerosols are difficult to trace to
specific indoor and outdoor sources because combustion is widespread. Sorensen et al. (2005,
089428). Ho et al. (2004, 056804). Larson et al. (2004, 098145). and Jansen et al. (2005, 082236) all
found that personal and microenvironmental exposure to total carbon or BC was lower than that
measured outdoors, while Sarnat et al. (2006, 089166) showed significant associations between
personal and ambient measurements of EC for measurements taken during the fall for low and high
ventilation conditions (slope = 0.66-0.73) and during the summer for high ventilation conditions
(slope = 0.41). Wu et al. (2006, 179950). Delfmo et al. (2006, 090745). Olson and Norris (2005,
156005) and Turpin et al. (2007, 157062) all demonstrated much higher levels of OC compared with
EC in personal samples, possibly due to indoor sources of OC from cooking and home heating. Reff
et al. (2007, 156045) and Meng et al. (2007, 194618) both reported findings from the Relationships
between Indoor, Outdoor, and Personal Air (RIOPA) study in Los Angeles, Houston, and Elizabeth,
NJ. Results from Reff et al. (2007, 156045) reveal significantly higher detection of aliphatic C-H
functional groups indoors and in personal samples compared with outdoors (Figure 3-98). This
information may help to distinguish carbonaceous compounds of indoor and outdoor origin in future
source apportionment studies of PM exposure. Little regional variation in the aliphatic, carbonyl, or
SO42~ groups tested were reported in this study. In Meng et al. (2007, 194618). indoor exposures
were shown to decrease for secondary formation aerosols including SO4 ~ but not NO3~ (not tested)
when compared with outdoor concentrations. In this study, indoor exposures to mechanically
generated aerosols decreased in comparison with outdoors (Figure 3-99).
Trace metal studies have shown variable results regarding personal exposure to ambient
constituents. For instance, Molnar et al. (2006, 156773) found that personal exposure was higher
than outdoor and ambient concentrations for mostly crustal Cl, K, Ca, Ti, Fe, and Cu. However,
Adgate et al. (2007, 156196) found that personal exposures were higher than ambient for Fe, Mg, K,
Zn, Cu, Pb, and Mn but lower than ambient for Al, Na, and Ti. Larson et al. (2004, 098145) found
that personal exposure to Ca and Cl were higher than concentrations measured at ambient (central
site) and residential outdoor monitors, lower for Fe, K, Mn, and As and the same for Al, Br, Cr, and
Cu. Source apportionment for trace metals can vary significantly among cities and over seasons. For
instance, in a Baltimore source apportionment study, exposure to Mn could be attributed nearly
equally to the Quebec wildfires, roadway wear, and soil, while Pb exposure was largely found to be
due to a local incinerator (Ogulei et al., 2006, 119973). In this case, the Quebec wildfires were a
transient episodic source, while roadway wear and incineration were continuous. However, in Larson
et al. (2004, 098145). Mn and Pb exposures in Seattle were largely attributable to mobile source and
stationary source emissions. For this reason, source composition behavior cannot be generalized for
characterizing exposures and resulting health effects across multiple locations or times.
December 2009 3-171
-------
~ 25
CO
| 20
"5
i 10 •
c
0)
§ 5-
O
0 J
Los Angeles Co., CA
1
I i i I
I I
25 ;
fT
IE 20 ;
^>
•f 15 :
o
1 10 ''•
"c
§ 5 ;
0
0 0 '
Elizabeth, NJ
T T 1
, i i ! I 1
^ (-^ & &
'/ ^ ^ /
co ,\o ~p ^y
^ 25 •
•^ 20 •
CD
^ 15 •
o
1 10 :
c
0)
O j-
c 5 :
0
O
0 •
Houston, TX
I ]
\^"
*y
^O- ^
,<#• o^
Source: Reprinted with Permission of Elsevier Ltd. from Reff et al. (2007,1560451
Figure 3-98. Apportionment of aliphatic carbon, carbonyl, and S042" components of outdoor,
indoor, and personal PM2.s samples, for Los Angeles (top), Elizabeth (center), and
Houston (bottom).
December 2009
3-172
-------
30 -
20
•in
ID
EJB^^™_
^
Mechanical Finf = 0.04 /
y = 0.04x + 0.51, R* = 0.51 /^
/
/ • AER: < 0.5 h''
^ V AER: 0.5 -1.5 h"'
/ D AER:>1.5h''
/
/
/
/
7 /
/
S
Zy'* '
offiBteafe-atJ ° V « V
0 5 10 15 20 25 30 3
-3fl ,____________^^
JU
^ ' 25 •
J=
50 20-
s3
C 15-
.2
•4— »
o3 10 •
'•4—*
1> 5 -
O
C L_
Primary FM = 0.51 /
y = 0.51x + 0.19, Rz = 0.61 //
/
/
/
S • AER: < 0.5 h'1
/' V AER: 0.5 - 1 .5 h *
/ D AER: > 1.5 h1
• s
a*Łf v
*aT^ J
aJEaH^^ v ^
^ n flHDiJIPfr1^
^ u HaJIwP*^ ' ' ' ' '
O 0 5 10 15 20 25 3
o 50
O
C 40 •
30 •
20
10 •
n i
f
Secondary Flnf = 0.78 /
y = 0.78x - 0.53, Rz = 0.75 /
/
S • AER: 1.5 h'1
D »^? ^
fiSflv _
^^^ * •
10
20
30
50
Outdoor Concentration (|ig/nr)
Source: Reprinted with Permission of ACS from Meng et al. (2007,194618).
Figure 3-99. Apportionment of infiltrated PM from mechanical generation (top), primary
combustion (center), and secondary combustion (bottom).
December 2009
3-173
-------
Differential Infiltration Related to PM Composition
A number of chemical factors influence the tendency for differential infiltration in PM.
Lunden et al. (2003, 156718) studied infiltration of BC and OC aerosols and found that F^f can vary
substantially as a function of gas transport properties with differing air exchange rates. This study
and Sarnat et al. (2006, 089166) also showed that BC aerosol infiltration is considerably higher than
infiltration of OC, and that carbonaceous aerosol infiltration differed substantially from NO3~ and
SO42~ aerosols under the same building air exchange conditions. These disparities are likely related
to differences in the particle size distribution of PM components, as described in Section 3.8.4.3. As
shown in Figure 3-100, the composition of indoor PM that has infiltrated from outdoors is different
from that of outdoor PM (Meng et al., 2007, 194618). In this case, the particles containing
photochemical products (primarily accumulation mode) have a higher infiltration rate than the larger
(primarily coarse mode) mechanically generated particles or the smaller primary combustion
particles (likely to consist mostly of UFPs in the nucleation or Aitken nuclei mode).
7-
5-
Photochemical
Secondary
•Accumulation Mode)
Primary
Combustion
Particles
~ Ultrafine)
Mechically
Generated
(~ Coarse Mode)
Out In
Out In Out In
Source: Reprinted with Permission of ACS from Meng et al. (2007,1946181.
Figure 3-100. Results of the positive matrix factorization model showing differences in the
mass of outdoor PM and PM that has infiltrated indoors based on source
category.
PM species enriched in the accumulation mode, such as SO42~, will infiltrate more efficiently
than components with larger size distributions, such as iron (Strand et al., 2007, 157018). Lunden et
al. (2008, 155949) also compared I/O ratios for PM25, total carbon, OC, and BC in an unoccupied
house and found the lowest ratio for PM2.5 (0.41 ± 0.2), the highest for BC (0.61 ± 0.2), and
intermediate values for total carbon (0.50 ± 0.2) and OC (0.47 ± 0.2). The authors attributed the
December 2009
3-174
-------
lower PM2.5 I/O ratio to indoor loss of NH4NO3 aerosol. The authors note that their BC I/O of 0.6 is
somewhat lower than BC ratios measured in occupied spaces (Polidori et al., 2007, 156877).
Conversely, indoor sources in occupied residences contribute to observed OC I/O ratios greater than
1 in other studies (Polidori et al., 2006, 156876; Sawant et al., 2004, 056798). Analytical results for
PM2.5 components from the Baxter et al. (2007, 092726) study found Fmf of 0.95 ± 0.07 for S and
0.60 ± 0.04 for V, the two components identified as having no indoor sources and which had I/O
ratios significantly less than 1 (Baxter et al., 2007, 092726). It is possible that association of V with
larger particles of lower penetration efficiency could contribute to a lower infiltration rate. Meng et
al. (2005, 058595) also noted that the lack of indoor sources of S and V result in much lower
variability in penetration and loss rates.
Volatilization of PM during infiltration can cause differences between the composition of
indoor and outdoor PM. NO3~, a prevalent PM component year-round in the western U.S. and during
winter throughout the mid-western and northeastern states, has a decreased Finf due to volatilization
of NO3~ indoors. Sarnat et al. (2006, 089166) calculated Fmf values forNO3~, PM2.5, and BC, and
found the values to increase in that order. NO3~ was low (median = 0.18, IQR = 0.12-0.33), while BC
was high (median = 0.84, IQR = 0.70-0.96); the intermediate value of PM2.5 (median = 0.48,
IQR = 0.39-0.57) reflected its composition as a mixture of those two components (among others).
Indoor volatilization of NO3~ enriches indoor ambient PM in other components, creating differences
in toxicity between indoor and outdoor ambient PM. The high infiltration of non-volatile BC creates
additional sorption sites for organics, including indoor-generated compounds. Meng et al. (2007,
194618) found that secondary formation accounts for 55% of indoor PM of outdoor origin, while
primary combustion accounts for 43%, and mechanical generation accounts for 2%. Meng et al.
(2007, 194618) noted that secondary formation processes often result in more accumulation mode
particles, so that diffusion losses are not as great as for primary combustion particles that are
composed primarily of nucleation and condensation size modes (Figure 3-100). Likewise, Polidori et
al. (2007, 156877) suggest that similarities in the EC and OC size distributions and infiltration
factors reflect low vapor pressure secondary organic aerosols in the OC component. Sioutas et al.
(2005, 088428) suggest that volatilization of UFPs while crossing the building envelope may impede
infiltration in this size range. Variations in the presence of outdoor PM indoors, and resulting
changes in removal behavior once indoors, relate to the species composition of PM.
3.8.5.2. Exposure to PM and Copollutants
Analysis of personal exposure to multipollutant mixtures is an area of growing research.
Several multipollutant studies involving UF, fine, and coarse PM are presented in Table A-65. Sarnat
et al. (2001, 019401) found significant associations between personal exposure to PM25 and ambient
concentrations of O3, NO2, CO (significant only for winter), and SO2 in a panel study conducted in
Baltimore. Personal exposures to PM2 5 and personal exposures to the gases were not correlated in
this study. This result may have arisen in part because personal exposures to the gases were often
beneath detection limits of the personal monitoring devices. Schwartz et al. (2007, 090220) also used
data from the Baltimore panel study to simulate distributions of personal exposures and ambient
concentrations of PM25, PM10, SO42~, NO2, and O3. They found that personal exposure to ambient
PM2 5 was significantly associated with ambient concentrations of PM2 5, NO2s and O3 (O3 in an
inverse relationship). They also reported that personal exposure to SO42~ was significantly positively
associated with ambient PM2 5 and O3 concentrations.
There is evidence that associations between ambient gases and personal exposure to PM2 5 of
ambient origin exist but are complex and vary by season and region. Seasonality of the associations
could be a result of seasonal variability in photochemistry, source generation, and building
ventilation. Sarnat et al. (2005, 087531) observed associations between personal exposure to total
PM2 5 and ambient concentrations of O3, NO2, and SO2 measured at community-based monitors for
groups of healthy senior citizens and school children in Boston during the summer. In this study,
significant associations between personal exposure to ambient PM2 5 and personal O3 exposures were
observed in summer and between personal PM2 5 and personal NO2 in winter and summer, unlike the
Baltimore study in which only summertime personal PM2 5 and personal NO2 were associated
(Sarnat et al., 2001, 019401). In their study of personal exposure to ambient air pollutants in
Steubenville, OH, Sarnat et al. (2006, 090489) found low but significant associations for ambient O3
with personal PM2 5, SO42~, and EC in the summer. Low but significant associations between ambient
SO2 and personal PM2 5, and between ambient NO2 and personal EC, were also observed. In the fall,
December 2009 3-175
-------
ambient O3 had a weak but significant association with personal EC, and SO2 had a weak but
significant association with personal SO42~. Ambient NO2 was also significantly associated with
personal PM25, SO42~, and EC with somewhat higher coefficient of determination (R2 = 0.25-0.49) in
the fall.
3.8.6. Implications of Exposure Assessment Issues for Interpretation of
Epidemiologic Studies
Environmental epidemiologic study designs vary by many factors, including study sample
size, measurement time interval, study duration, monitor type, and spatial distribution of the study
sample. A panel epidemiology study consists of a relatively small sample (typically tens) of study
participants followed over a period of days to months. Time-activity diary studies are examples of
panel studies (e.g., Cohen et al, 2009, 190639: Elgethun et al, 2003, 190640; Johnson et al., 2000,
001660; Olson and Burke, 2006, 189951). and a microenvironmental model might be applied to
represent exposure in this case. Community time-series studies may involve millions of people
whose exposure and health status is estimated over the course of a few years using a short
monitoring interval (hours to days). Because so many people are involved, community-averaged
concentration is typically used as a surrogate for exposure in community time-series studies.
Exposures and health effects are spatially aggregated over the time intervals of interest because they
are designed to examine health effects and their potential causes at the community level (e.g.,
Dominici et al., 2000, 005828; Peng et al., 2005, 087463). A longitudinal cohort epidemiology study
typically involves hundreds or thousands of subjects followed over several years or decades.
Concentrations are generally aggregated over time and by community to estimate exposures (e.g.,
Dockery et al., 1993, 044457; Krewski et al., 2000, 012281). The importance of exposure
misclassification varies with study design based on the spatial and temporal aspects of the design.
Other factors that could influence exposure estimates in PM epidemiologic studies include source
characteristics, particle size distribution, and particle composition. Potential issues that could
influence estimates of PM exposure include measurement, modeling, spatial variability, temporal
variability, use of surrogates for PM exposure, and compositional differences. These are described in
detail in the following sections.
3.8.6.1. Measurement Error
Measurement Error at Community-Based Ambient Monitors and Exposure
Assessment
Community-based ambient monitors are employed for time-series and longitudinal studies,
although they can be used for panel studies as well. Section 3.4 discusses potential errors in
measuring ambient PM in detail. Because there will likely be some random component to
instrumental measurement error, the correlation of the measured PM mass with the true PM mass is
expected to be <1. Sheppard et al. (2005, 079176) indicate that instrument error in the hourly or
daily average concentrations has "the effect of attenuating the estimate of a." Zeger et al. (2000,
001949) suggest that in order for this error to cause substantial bias in later estimation of a health
outcome, the measurement error must be strongly correlated with the measured concentrations.
Positive and negative artifacts resulting from sampling volatile PM may therefore lead to lack of
association with health endpoints in time-series and longitudinal studies. In multicity longitudinal
studies, where PM composition and associated artifacts may vary across cities, the cumulative
influence of such artifacts on exposure estimates is more difficult to predict.
Measurement Error for Personal Exposure Monitors
PEMs are primarily used in panel exposure studies to measure total exposure to PM (e.g.,
Cohen et al., 2009, 190639; Elgethun et al., 2003, 190640; Johnson et al., 2000, 001660; Olson and
Burke, 2006, 189951). PEMs are specialized monitors that, because people must carry them, have to
December 2009 3-176
-------
be small, light, quiet, and battery operated or passive. As a result, they may have lower face
velocities across the filter and lower pressure drops than ambient-based filter measurements of PM,
which typically sample at much higher flow rates, and consequently at much higher face velocities.
Light scattering measurements are biased when relative humidity is high and are sensitive to size
distribution (Lowenthal et al., 1995, 045134; Sioutas et al, 2000, 025223). Positive artifacts
resulting from adsorption of vapor-phase organic compounds and negative artifacts from evaporation
of semi-volatile PM also create challenges for interpreting personal exposure monitoring data (e.g.,
Pang et al., 2002, 030353). Olson and Norris (2005, 156005) attributed more OC particle mass
collection to face velocity differences when using PEMS compared with FRMs. Data quality of
PEMs is described in much greater detail in the 2004 PM AQCD (U.S. EPA, 2004, 056905). The
artifacts listed here could result in either negative or positive sampling bias.
3.8.6.2. Model-Related Errors
When models are used in lieu of or to supplement measurements of ambient PM exposure or
community-based ambient PM concentration, it is important to identify errors and uncertainties that
could affect estimates of PM-related health effects. Model-related errors are determined by four
factors: representativeness of the mathematical model, accuracy of model inputs, scale of model
resolution, and model sensitivity. If verification errors related to these four factors are minimized,
then the model can be evaluated against physical data to determine how well the model truly
captures a real situation (Roache, 1998, 156915). Detail of the model design and inputs can have
significant impact on validation, as observed in Meng et al.(2005, 058595) and Hering (2007,
155839). Meng et al. (2005, 058595) demonstrated how use of an increasingly more detailed
mathematical model decreases the variability of the results with respect to modeled indoor PM2.5
concentration of outdoor origin and to modeled infiltration factor. Hering (2007, 155839) compared
infiltration model results for PM-based EC, NO3~, and SO42~. Model inputs were from a central site
monitor only, central site monitor with air exchange data, and detailed inputs related to initial
outdoor (outside test building) and indoor concentrations. Use of more detailed inputs resulted in
significant reductions in error for indoor EC concentration, smaller improvements for indoor SO42~,
and negligible improvement in model results for indoor NO3~. This illustrates the impact of
differential infiltration discussed in Sections 3.8.4 and 3.8.5.
December 2009 3-177
-------
36 km CMAQ grid
12km CMAQ grid
4 km CMAQ grid
Census tract
centroids
in Philadelphia
county:
Tracts in each
4x4km
grid cells are shown
in different colors
4405
465 470
475
480 485 490 495 500
West-East distance (km)
505 510
515
Source: Reprinted with Permission of ACS from Isakov et al.(2007,195880).
Figure 3-101. Grid resolution of the CMAQ model in Philadelphia compared with distribution of
census tracts in which exposure assessment is performed.
For any spatial interpolation models, grid resolution is another source of error. Isakov et al.
(2007, 195880) linked CMAQ with the Hazardous Air Pollutant Exposure Model for exposure
assessment in Philadelphia. Their simulation was implemented on a 4 km nested grid within 12 km
and 36 km grids to bring the scale of their model from national to urban. However, the census tracts
in which Isakov et al. (2007, 195880) sought to describe exposure were distributed on a much finer
scale (Figure 3-101). They supplemented the CMAQ model with an Industrial Source Complex
Short Term (ISCST) dispersion model to resolve the subgrid scale behavior. If concentrations were
averaged across the cell in lieu of a more detailed subgrid representation, Isakov et al. (2007,
195880) found that exposures were overestimated by a factor of 2. Appel et al. (2008, 155660) noted
that their 36 km simulations provided a closer estimate of SO42~ aerosol concentration than did their
12 km nested simulation, which overestimated concentrations. Hogrefe et al. (2007, 156561) also
noted overestimation of the CMAQ model at the 12 km scale, where multiple point interpolation was
used to obtain subgrid estimates. Model convergence theory would suggest that the 36 km simulation
is not actually more accurate but coincidentally closer to the observed concentrations (Roache, 1998,
156915). It is possible that if secondary pollutants are more regionally dispersed, lower spatial
resolution would be required to attain a converged solution of the spatial concentration field.
However, higher spatial resolution in the simulation should produce very similar results if the
solution is convergent.
Use of geospatial statistical methods for grid interpolation, as performed in the
SHEDS/MENTOR simulation by Georgopoulos et al. (2005, 080269). provides another
methodology for grid interpolation. Similar to Isakov et al.(2007, 195880). Georgopoulos et al.
(2005, 080269) linked CMAQ with an exposure model for estimation of neighborhood-scale
exposures using a 4 km resolution grid nested within 12 km and 36 km grids. The authors found that
CMAQ underestimated PM2.5 concentration at many times during the simulation. Kim et al. (2009,
188446) compared results from an exposure model using six different levels of spatial resolution.
The model predicted PM2.5 at monitor locations as a function of the mean concentration and spatial
and random errors. The 6 levels of spatial resolution were simulated through assignment of the
spatial and random error terms and by defining the distance over which spatial errors are correlated.
December 2009
3-178
-------
Between monitors, Kim et al. (2009, 188446) compared results from assigning nearest monitor
values and from kriging. They found that model prediction error and bias in health effects estimates
decreased with increased spatial resolution of the error terms, as well as with kriging over a nearest
monitor scheme.
For GIS-based models designed to improve spatial resolution of exposure estimates, Yanosky
et al. (2008, 099467) described three sources from which they derived an estimate of total model
uncertainty: transient model components, stationary model components, and residual spatial and
temporal components of variance. When analyzing relative contributions to uncertainty, they found
that unexplained local spatial variability was the largest contributor. With model inputs from PMi0
monitors for this study, poor model performance and high uncertainty were observed where monitors
were sparsely located. High uncertainties were also calculated in a few select urban areas (New York
City, Detroit, Cleveland, and Pittsburgh) where concentrations tend to be higher, although the latter
may have been related to the Taylor series approximation used for the residual uncertainty term.
Spatial and temporal uncertainties were also reduced when temporal resolution was increased in the
model implementation.
In his review of various exposure assessment modeling techniques, Jerrett et al. (2005,
092864) reviewed source proximity and LUR for application to exposure assessment. The literature
contains mixed evidence of the association between health effects and source proximity (e.g.,
Langholz et al., 2002, 191771: Maheswaran and Elliott, 2003, 125271; Venn et al., 2000, 007895;
Venn et al., 2001, 023644). Jerrett et al. (2005, 092864) contend that source proximity modeling is
limited because other confounding covariates, such as SES, may be related to source proximity.
Additionally, subjects' time-activity patterns may vary from locations modeled through source
proximity. Wind direction and topography may bring PM plumes away from a site located even in
very close proximity to the source, so that high concentrations would be found at distances far
downwind of a source. Jerrett et al. (2005, 092864) state that LUR is an adaptable framework
allowing adaptation to localized conditions, but they caution that LUR is limited to fairly
homogeneous spatial regions. They point to Briggs' (2000, 191772) simulation of Amsterdam as an
example of LUR surfaces produced with little spatial variability.
LUR and kriging were both used in the ACS data reanalysis by Krewski et al. (2009, 191193)
to study mortality as a function of spatial variability in PM2 5 in New York and Los Angeles, as
described in Section 3.8.3.4. The LUR solution produced some observed overpredictions near
freeways. Kriged results were compared with LUR for both cities. For New York City, kriging
produced slightly attenuated mortality risk estimates, while for Los Angeles, kriging did not exhibit
as much spatial variability as LUR. The latter may be due to the fact that the monitoring network in
Los Angeles was not situated to capture spatial variability in PM2.5 concentration occurring near
areas of high traffic. Despite similarities in the LUR performance, health effects predictions were
quite different for New York City and Los Angeles, with increased hazard ratios of 1.56 and 1.39,
respectively using the same LUR covariates. Krewski et al. (2009, 191193) noted that in New York
City, the healthiest (and wealthiest) segment of the population lived in the most polluted areas, while
in Los Angeles, there was a strong association between pollution and mortality. This finding implies
that, because geographic regions may differ by multiple factors, such as building and power plant
fuel use, roadway design, traffic patterns, and building design, significant variables in an LUR
analysis may also differ by region.
3.8.6.3. Spatial Variability
For PM, spatial and temporal distribution as a function of particle size and composition also
plays a large role in the selection of an exposure model. For instance, use of Equation 3-6 might be
employed for a study of ambient PM2.5 exposure because spatial variability in PM2.5 concentration
can be low over urban to regional scales in comparison with more spatially variable PMi0_2.5 and
UFPs, as described in Section 3.8.4. Spatial issues leading to exposure misclassification are
discussed below.
In panel studies, exposure error will be introduced if the ambient PM concentration measured
at the central site monitor is used as an exposure surrogate and differs from the actual ambient PM
concentration outside a subject's residence and/or worksite. Filleul et al. (2006, 089862) computed
exposure based on varying contributions of community-based ambient monitors (deemed
background) and proximal monitors (to represent a receptor) in Le Havre, France for black smoke
measurements. They found that using a weighted mean with increasing weight for proximal monitors
December 2009 3-179
-------
resulted in non-significant but increased mean exposure estimates. Moore et al.'s (2009, 191004)
finding of high variability in UFPs across Los Angeles also suggests that exposure error would occur
from using one or a few UFP monitors. In an example using AQS data, PMi0 monitors in the
Chicago CSA are south of the most populated areas within Chicago (Figure A-8 in Annex A), where
intersampler correlations for urban scale PMi0 data for several monitor pairs are below 0.4 (Table A-
23 in Annex A). In another example from AQS data, PM2.5 and PMi0 monitor locations in the
Riverside CBSA are shown to correspond more closely with higher population density areas (Figures
A-25 and A-26 in Annex A), where urban scale intersampler correlations for both PM2 5 and PMi0 are
below 0.4 for several monitor pairs. For most cities, intersampler correlation is much higher for
PM2.5 than for PMi0. This is in accord with the findings of Sarnat et al. (2009, 180084) where, in an
Atlanta time-series study of the effect of spatial variation in concentration on epidemiologic
associations, spatially homogeneous PM2.5 and O3 were found to be consistently associated with
emergency room visits using any monitor within the study area, while associations were less
consistent across monitors for spatially heterogeneous CO and NO2 among the entire population
studied. Considering results reported in the literature along with inter-sampler correlations (reported
in Annex A, Section A.2 for PM2 5 and PMi0, and their corresponding monitor locations shown in
Annex A, Section A. 1), the magnitude of spatial exposure error likely depends on particle size as
well as monitor location, source location, and characteristics such as urban and natural topography
and meteorological trends.
Spatial variability among various studies further suggests that use of a single or small number
of ambient monitors introduces uncertainty in exposure assessment panel studies. Violante et al.
(2006, 156140) studied personal exposures to traffic of parking police in Bologna, Italy to determine
how personal exposure to outdoor PMi0 and benzene compares with that measured at a
community-based monitor. This study found that personal exposures to PMi0 were significantly
higher than at the community-based monitor, although the authors were not able to demonstrate
significant effects of meteorology or traffic on those exposures. Nerriere et al. (2007, 156801)
observed spatial heterogeneity of personal exposures to metals in PM25 and PMi0, with higher levels
found near high-traffic and industrial areas. In a Bayesian hierarchical model analysis of personal
exposure and ambient PM2 5 data from the pilot Baltimore Epidemiology-Exposure Panel Study of
16 subjects, McBride et al. (2007, 124058) showed that community monitors overestimated personal
exposures for the panel subjects, and that these results were not sensitive to model selection.
For community time-series epidemiology, the community-average concentration, not the
concentration at each fixed monitoring site, is the concentration variable of concern (Zeger et al.,
2000, 001949). Because variation in trends is of interest, bias in the central site monitor data will not
affect health effects estimates unless the central site monitor is not correlated with the community-
average concentration. The latter condition will cause the health effect estimate to be biased towards
the null (Sheppard et al., 2005, 079176). The correlation between the concentration at a central
community ambient monitor and the community-average concentration depends on homogeneity of
the spatial distribution and representativeness of the central-site monitor location. Kim et al. (2005,
083181) noted that spatial variability among PM species can add uncertainty to exposure estimates
in community time-series epidemiology studies exploring source contributions to health effects. The
monitoring site is selected to represent the community average of the PM characteristic (mass and/or
species) of interest. If the selected site is far from PM sources, then the average measurement may be
lower than actual ambient PM concentrations. Likewise, if the site is selected to measure a "hot
spot" or pollution from a nearby source, exposure estimates across the community could be skewed
upwards.
Intra-urban spatial heterogeneity could affect health effects estimates derived from community
time-series studies if a community is divided by urban or natural topographic features or by source
locations into several sub-communities that differ in the temporal pattern of pollution. Intra-urban
spatial heterogeneity is discussed in detail in Section 3.5. Community exposure may not be well-
represented when monitors cover large areas with several sub-communities having different sources
and topographies. This point is illustrated for Los Angeles in Figure 3-27 and Figure 3-37 where
intersampler correlation decreases with respect to distance moreso than for other cities shown. Using
zip code classified mortality data in a study of SES and acute cardiovascular mortality in Phoenix,
high risk ratios were computed when a small area near the monitoring site was studied (Mar et al.,
2003, 156731; Wilson et al., 2007, 157149). while use of larger-area county-wide data produced
non-significant associations (Moolgavkar, 2000, 010305; Smith et al., 2000, 010335). At least part of
the heterogeneity found between cities in multicity studies may be due to the use of a large
December 2009 3-180
-------
geographic area that is composed of several sub-communities that differ in the spatiotemporal
distribution of air pollutants. Note that when zip codes cover large areas (e.g., in western mountain
states) or when counties cover small areas (e.g., in the northeast), then assumptions change regarding
use of zip code- or county-level data for epidemiologic studies. For all metropolitan areas
investigated in this assessment, the PMi0 data have significantly more scatter than PM2.5. This
suggests that the uncertainty of the community average concentration would increase in the coarse
PM range. Metrics have been developed and used to compare the spatial variability of air pollutants
(Wongphatarakul et al., 1998, 049281). These metrics are useful in assessing the potential for
exposure error in the epidemiologic studies, especially when different monitors are used on different
days to construct city-wide averages.
Epidemiologic studies of long-term exposure rely on differences among communities in long-
term average ambient concentrations. If exposure errors are different in the different communities,
the differences in long-term ambient concentrations among communities may not represent the
differences in long-term average exposures (Dockery et al., 1993, 044457). Thus, in a regression of
health effects against average concentration as an indicator for average exposure, there could be a
different magnitude and direction of error in the exposure indicator for each spatial area. This could
bias the slope up or down. The following epidemiologic studies, described in detail in Section 7.6,
are cited here to illustrate the effect of spatial exposure error on health effects estimates. The Harvard
Six-City Study dealt with this issue by design, where the members of the cohort in each city were
located in a relatively small area near the monitor (Dockery et al., 1993, 044457). In the ACS study,
the spatial area was the Metropolitan Statistical Area (MSA) (Krewski et al., 2000, 012281); other
studies have used counties as the spatial area (Enstrom, 2005, 087356; Lipfert et al., 2000, 004087).
In a comparison of several of the larger long-term cohort studies, those using county level spatial
areas (Enstrom, 2005, 087356; Lipfert et al., 2000, 004087) sometimes did not find significant
associations, whereas those using MSAs (Pope et al., 1995, 045159; Pope et al., 2002, 024689) or
cities (Dockery et al., 1993, 044457) did find significant associations. Jerrett et al. (2005, 189405)
used smaller zip code areas within Los Angeles County and found effects that were both significant
and largest in magnitude compared to those reported for other long-term cohort studies. Krewski et
al. (2009, 191193) suggested that significant associations between cardiovascular health effects
estimates and PM2.s observed in Los Angeles but not in New York City were related to spatial
homogeneity of PM2.5 concentration in New York City. The Nurses' Health Study examined
associations of mortality with PMi0 and found higher and more significant associations when using
estimated concentrations at subjects' individual residences (Puett et al., 2008, 156891) in lieu of
county-level concentrations (Fuentes et al., 2006, 097647). These considerations suggest that studies
that include large U.S. counties as spatial areas and find no significant associations of health effects
with pollution cannot be considered definitive, because the likelihood of exposure error increases in
this situation. Reducing the exposure error by using concentrations based on residence address or
small zip code areas is associated with larger relative risk than those obtained with county-wide
averages of concentrations.
3.8.6.4. Temporal Variability
Temporal Correlation
Concentration time series analyzed for community time-series epidemiologic studies can
include those averaged over several monitors, a single monitor used as an estimate of the true
community average exposure, or a monitor used to represent nearby exposures. Within a city, lack of
correlation of relevant time series at various sites results in smoothing the exposure surrogate
concentration function over time and resulting loss of peak structure from the data series. Burnett
and Goldberg (2003, 042798) found that community time-series epidemiology results reflect actual
population dynamics only when five conditions are met: environmental covariates are fixed spatially
but vary temporally; the probability of the health effect estimate is small at any given time; each
member of the population has the same probability of the health effect estimate at any given time
after adjusting for risk factors; each member of the population is equally affected by environmental
covariates; and, if risk factors are averaged across members of the population, they will exhibit
smooth temporal variation. For this study, mortality was examined, but the temporal considerations
December 2009 3-181
-------
are generalizable to other health outcomes. Dominici et al. (2000, 005828) note that ensuring
correlation between ambient and community-average exposure time series is made difficult by
limitations in availability and duration of detailed ambient concentration and exposure time series
data and, as a result, is often a source of uncertainty. Sheppard et al. (2005, 079176) also add that the
health effect estimate can be biased by time-dependent error in a in a time-series study if the spatial
variation in PM concentration is not significant. The direction of bias is related to seasonal
correlation between a and Ca.
Seasonality
Community time-series studies can be designed to investigate seasonal effects by
incorporating seasonal interaction terms for the exposure surrogate and/or meteorology (e.g.,
Dominici et al., 2000, 005828). Studies from Section 6.5 are briefly mentioned here to illustrate how
seasonal exposures can influence health effect estimates. Bell et al. (2008, 156266) and Peng et al.
(2005, 087463) observed higher health effect estimates and stronger seasonal dependence in the
northeast than in the rest of the country for PM2.5 and PMi0, respectively. Peng et al. (2005, 087463)
stated that these results generated three hypotheses. First, the PM composition and resulting toxicity
might vary with season. Bell et al. (2008, 156266) showed seasonal differences between respiratory
and cardiovascular effect estimates that the authors hypothesized related to seasonal differences in
dominance of a given PM species. Second, Peng et al. (2005, 087463) suggested that less seasonality
in regions other than the northeast may reflect regional tendencies for spending more or less time
outdoors. Exposure estimates for time spent outdoors may be less subject to exposure error because
uncertainties related to infiltration are not a factor during that time. At the same time, air
conditioning usage, which is more common in the summer and in warm climates (Pandian et al.,
1998, 090552). has been associated with decreased association between PM2.5 and cardiovascular
morbidity (Bell et al., 2009, 191007). Third, infectious diseases are more prevalent during winter and
so may influence health outcomes. However, it would be expected that regions other than the
northeast would be affected by influenza in winter. Uncertainty in sources of seasonal bias may also
indicate other unknown factors.
Data Frequency
Most panel and many time-series studies examine the associations of health outcomes only
with exposure (or exposure surrogates) on the day of exposure (lag 0). Zanobetti et al. (2000,
004133) and Lokken et al. (2009, 186774) suggest that health effects may not occur until subsequent
days or be distributed over several days. When PM measurements are obtained every three or six
days, it is difficult to refine the study lag structure down to the day-level. In studies of the effects of
short-term PM2.5 exposure on cardiovascular and respiratory hospitalization in >200 urban U.S.
counties, Bell et al. (2008, 156266) and Dominici et al. (2006, 088398) worked with a combination
of air quality measurements obtained daily, those obtained every 3 days, and daily hospitalization
data. Time lags of 0, 1, and 2 days were applied, such that PM2.5 data obtained only on day 0 would
be applied as lag 0 for the corresponding day's hospitalization record, as lag 1 for the next day's
hospitalization, and as lag 2 for the following day. No lag 0 data would be available for day 1, and no
lag 0 or 1 data would be available for day 2 in this example. This analysis could be performed with
sufficient statistical power despite the reduction in days of PM data because a large number of
counties were analyzed. Single city studies using data obtained every three or six days could employ
the same approach but would lose statistical power compared with daily data because fewer data
points would exist for each lag. Likewise, Dominici et al. (2006, 088398) only applied distributed
lag analysis to the daily hospitalization data where daily PM2 5 concentration data were also
available.
December 2009 3-182
-------
3.8.6.5. Use of Surrogates for PM Exposure
Surrogates for Infiltration Tracers
In panel studies, a tracer can be used for PM that infiltrates indoors, such as sulfate as a
surrogate for PM2.5, as described in Section 3.8.4. For this method to be successful, indoor or other
nonambient sources of the tracer must be small compared to ambient sources over the period of
sampling. Wallace and Williams (2005, 057485) observed that, because SO42~ particles are typically
smaller than other PM contributing to measurable PM2.5 mass, Finf may be biased by using this term
to describe PM2.5 infiltration. Other concerns in using SO42~ as a tracer for PM2.5 arise because SO42~
tends to be concentrated in smaller particles and thus it might be a better tracer for fine mode
particles than for the coarse fraction at the upper tail of the PM25 particle size distribution.
Volatilization of ammonium nitrate or organic compounds after infiltration of PM25 indoors results
in these components being poor surrogates for ambient PM in exposure estimates (Lunden et al.,
2003, 081201).
Use of Ambient PM Concentration in Lieu of Ambient PM Exposure
Ambient PM concentration is often used as a surrogate for exposure to ambient PM in
epidemiologic studies. The ambient concentration may be based on measurements made just outside
the primary microenvironment, at the nearest community monitor, at a single community monitor, or
as the average of several community monitors. Based on the information presented in Section 3.8.4
related to urban-scale PM distribution, there is less exposure error for accumulation mode PM
because it has a more homogeneous spatial distribution and higher infiltration indoors, compared
with coarse or UFPs. If appropriate measurements are made, it is also possible to estimate the
ambient and nonambient components of total personal exposure and use four exposure surrogates in
panel epidemiologic studies: Ca, Er, Ea, and Ena (Ebelt et al., 2005, 056907; Koenig et al., 2005,
087384; Strand et al., 2006, 089203; Wilson and Brauer, 2006, 088933). Results from Wilson and
Brauer (2006, 088933) showed that exposure error is introduced by 1) using Ca instead of Ea and 2)
assuming Ea and Ena have the same effects on health outcomes. There was essentially no association
of the effect with ET or Ena. Wilson and Brauer (2006, 088933) noted that exposure to nonambient
PM will not affect the relationship between Ca and Ea, but "the difference between ambient
concentration and ambient exposure will bias the relative risk derived from epidemiologic studies."
Strand et al. (2006, 089203) also noted that inclusion of nonambient PM25 would not be expected to
change health effect estimates because ambient and nonambient PM2 5 calculations were not
correlated.
Zeger et al. (2000, 001949) pointed out that for community time-series epidemiology, it is the
correlation of the daily community-average personal exposure to the ambient concentration with
daily community-average concentration that is important, not the correlation of each individual's
daily exposure with the daily community-average concentration. Thus, the low correlation of
individual daily exposure with the daily community-average concentration, as frequently found in
pooled panel exposure studies, is not relevant to error in community time-series epidemiologic
analysis. Sheppard et al. (2005, 079176) also notes that an insufficient number of total personal
exposure samples used in a time-series design would introduce large classical measurement errors
related to high variability in Ena. Sheppard et al. (2005, 079176) further maintain that these errors
can be minimized by using the average concentration measured at community-based ambient
monitors. However, overestimation of the community-average exposure by substituting Ca for Ea
leads to underestimation of the effect estimate per unit mass of ambient PM. City-to-city variations
in the indoor air exchange rate, related to differences in climate or housing stock, will cause city-to-
city differences in the health effect endpoint estimate obtained from the study using Ca even if the
endpoint remained the same using the community-average exposure.
December 2009 3-183
-------
Relationship between PM and Copollutants
Uncertainties in the composition of multipollutant mixtures of gases and PM to which the
population is exposed can introduce uncertainties in health effects estimates. When copollutant
associations exist, as described in Section 3.8.5, the potential for one pollutant to act as a surrogate
for another pollutant or mix of pollutants introduces uncertainty into epidemiologic models (Sarnat
et al., 2001, 019401). For example, O3 may be an indicator of photochemical oxidation products
including organic PM. SO2 may be an indicator of Ni emissions from smelters, V from oil fired
power plants, or As, Se or Hg from coal-fired power plants. In another example, the HEI Report on
traffic-related health effects (2009, 191009) lists CO, NO2, PM2.5 and PMi0 mass, UFP count, EC,
benzene, and traffic metrics (e.g., count, fuel consumption) all as potential surrogates for traffic or
for the mix of all PM and gaseous pollutants in traffic because all of these pollutants are found in
mobile source emissions. Furthermore, in a multipollutant model, transfer of association can occur
by an increase in the slope of a confounding copollutant and a concurrent decrease in the slope of the
truly causal covariate. This can occur when copollutants are highly correlated with larger error for
the true copollutant, smaller error for the confounder, and correlation between the copollutant
measurement errors (U.S. EPA, 2004, 056905: Zeger et al., 2000, 001949; Zidek et al., 1996,
051879). For these reasons, this is an important area of uncertainty for interpretation of the
multipollutant models discussed in Chapter 6.
3.8.6.6. Compositional Differences
Differences between the composition of ambient PM and the ambient PM that has infiltrated
indoors may affect exposure estimates. Numerous differential infiltration studies related to indoor-
outdoor changes in size distribution and chemical composition are cited in Sections 3.8.4 and 3.8.5,
respectively. If differential infiltration results in differences in PM size distribution and chemical
composition between indoor-ambient PM and outdoor-ambient PM, then use of outdoor-ambient PM
could bias health effects estimates related to particular species. Baxter et al. (2007, 092726) showed
that V tends to have lower F^f, perhaps because metals exist more in the coarse range, while S has
Finf close to unity. Epidemiologic studies cited in Section 6.6 indicate that significant associations
between health effects estimates and PM trace metal exposures exist and may be modified by season.
Trace metal penetration efficiency estimates are thus relevant to those findings. Section 6.6 also
discusses significant associations between health effects endpoints and exposure to EC and OC in
PM. After initial emission, traffic-related PM is generally in the accumulation mode with volatile
components; accumulation mode PM tends to have the highest infiltration factors, but volatile
components may be lost during infiltration (Sarnat et al., 2006, 089166). If outdoor residential or
central-site measurements are used for an exposure surrogate, differences in indoor and outdoor PM
composition related to infiltration could introduce uncertainty into effects estimates.
Ebelt et al. (2005, 056907) illustrated that exposure error occurs when the PM on one or more
days is not representative of the normal community PM. Section 6.3.2.1 discusses this COPD panel
study of the association between respiratory and cardiovascular measures (i.e., lung function, blood
pressure, heart rate, HRV, and ectopic beats) and PM2.5, PMi0_2.5, and PMi0. In their analysis, one day
of dust from the Gobi Desert caused an increase in the concentration of fine and coarse PM. When
this day was deleted from the analysis, the associations of health effects estimates with PMi0 and
especially with PMi0_2.5 became larger and more significant. Similar peaks in PMi0 concentration
have been observed on the Iberian Peninsula as a result of high transport events carrying dust from
the Sahara Desert that could affect epidemiologic associations on given days (Artinano et al., 2001,
190099).
3.8.6.7. Conclusions
This section presents considerations for exposure assessment and the exposure
misclassification issues that can potentially affect health effects estimates. These issues can be
categorized into six areas: measurement, modeling, spatial variability, temporal variability, use of
surrogates for PM exposure, and compositional differences. Potential influences of each of these
sources on health effects estimates derived from panel, time-series, and longitudinal epidemiologic
studies are described above. Additionally, error sources often interact with each other and are driven
December 2009 3-184
-------
by particle size distribution. For example, fresh diesel-generated PM is characterized by UFPs that
dynamically grow and change in chemical composition over short time and spatial scales, and lack
of spatial and temporal resolution in measurements or models can result in misclassifying this
exposure (Moore et al., 2009, 191004). For this reason, conclusions regarding UFP exposure cannot
be drawn from PM2.5 concentration data, in part because PM2.5 concentration is more spatially
homogeneous across a city. In most circumstances, exposure error tends to bias a health effect
estimate downward (Sheppard et al., 2005, 079176: Zeger et al., 2000, 001949). Insufficient spatial
or temporal resolution to capture true variability and correlation of PM with copollutants are
examples of sources of uncertainty that could widen confidence intervals and so potentially reduce
the significance of health effects estimates.
3.9. Summary and Conclusions
3.9.1. Concentrations and Sources of Atmospheric PM
This section summarizes sources and concentrations of atmospheric PM. The following
summaries cover source characteristics from Section 3.3, measurement techniques from Section 3.4,
spacial and temporal variability and copollutant correlations from Section 3.5, source contributions
from Section 3.6 and policy relevant background concentrations from Section 3.7.
3.9.1.1. PM Source Characteristics
PM in the atmosphere contains both primary (i.e., emitted directly by sources) and secondary
components, which can be anthropogenic or natural in origin. Secondary components are produced
by the oxidation of precursor gases such as SO2 and NOX and reactions of acidic products with NH3
and organic compounds. Developments in the chemistry of formation of SOA indicate that oligomers
are likely a major component of OC in aerosol samples. Recent observations suggest that small, but
still significant quantities of SOA are formed from isoprene oxidation. Gasoline engines have been
found to emit a mix of OC, EC, and nucleation-mode heavy and large poly cyclic aromatic
hydrocarbons on which unspent fuel and trace metals condense, while diesel particles are composed
of a soot nucleus on which SO42~ and hydrocarbons condense. Data from standard emissions tests in
which there is insufficient dilution of fresh exhaust from combustion sources tend to overestimate
the primary component of organic aerosol at the expense of the semi-volatile components. These
semi-volatile components are precursors to secondary organic aerosol formation and their oxidation
results in more oxidized forms of SOA than previously considered, both in near source urban
environments and further downwind.
3.9.1.2. Measurement Techniques
The federal reference methods for PM2.5 and PM10 are based on criteria outlined in the CFR.
They are, however, subject to several limitations that should be kept in mind when using compliance
monitoring data for health outcome studies. FRM methods are subject to the loss of semi-volatile
species such as organic compounds and ammonium nitrate (especially in the West). Since FRM
gravimetric methods involve 24-h integrated filter samples, no information is available for variations
over shorter averaging times. However, methods have been developed to measure real-time PM2 5 or
PM10 mass concentrations (e.g., FDMS-TEOM). New FRMs and FEMs are available for PM10_2.5 and
various methods (dichotomous samplers, cascade impactors, and passive sampling techniques) are
under evaluation to improve PMi0_2.5 measurements. Techniques are available to characterize UFP
mass, surface area, and number concentrations. Continuous and semi-continuous measurement
techniques are also available for PM species, such as PILS for multiple ion analysis and AMS for
multiple component analysis. Advances have also been achieved in PM organic speciation (e.g., TD-
GC/MS).
December 2009 3-185
-------
3.9.1.3. Ambient PM Variability and Correlations
Advances in understanding the spatiotemporal distribution of PM mass and constituents have
recently been made, particularly with regard to PM2.5 mass and chemical composition and UFP
concentrations. Emphasis in this ISA was on the period 2005-2007 so that the most recent validated
EPA Air Quality System (AQS) data were used. Note, however, that a majority of U.S. counties were
not represented by AQS data since their population fell below the regulatory monitoring threshold
for PM. Moreover, monitors reporting to AQS were not uniformly distributed across the U.S. or
within counties, and conclusions drawn from AQS data may not apply equally to all parts of a
geographic region. Furthermore, biases can exist for some PM constituents (and hence total mass)
owing to volatilization losses of nitrates and other semi-volatile compounds, and, conversely, to
retention of particle-bound water by hygroscopic species. The degree of spatial variability in PM is
likely to be region-specific and strongly influenced by region-specific sources and meteorological
and topographic conditions.
Spatial Variability across the U.S.
County-scale, 24-h avg concentration data for PM2.5 between 2005-2007 showed considerable
variability across the U.S.. The highest reported 3-yr avg concentrations were for six counties within
the San Joaquin Valley and inland southern California, as well as Jefferson County, AL (containing
Birmingham) and Allegheny County, PA (containing Pittsburgh). The lowest reported annual average
PM2.5 concentrations were contained within 237 counties distributed throughout many western and
northern states as well as Florida and the Carolinas. Of the 15 individual CSAs/CBSAs selected for
detailed investigation based on their geographic distribution and importance in recent health effect
studies, the highest mean 24-h PM2.5 concentrations were reported for Riverside (17 (ig/m3),
Birmingham (16 (ig/m3) and Pittsburgh (16 (ig/m3); the lowest were reported for Denver (9 (ig/m3)
and Seattle (9 (ig/m3).
Since PM10_2.5 is not routinely measured and reported to AQS, co-located low-volume PM10
and PM2 5 measurements from the AQS network were used to investigate the spatial distribution in
PMio_2.5. Current data coverage (see Figure 3-10) and measurement errors limit the ability to draw
any meaningful conclusions regarding the large-scale spatial distribution of PMi0_2.5 in urban areas.
Only 6 of the 15 CSAs/CBSAs chosen for closer investigation had sufficient data for calculating
PMio_2.5. In general, in the eastern metropolitan areas including Atlanta, Boston, Chicago and New
York, most of the mass of PM10 was in the PM25 size fraction, with the highest ratio of PM25 to
PMio_25in Chicago (14 (ig/m3 PM25 5 (ig/m3 PMio_2s ratio = 2.8). In contrast, Denver (9 (ig/m3
PM2.5, 20 (ig/m3 PMio.2.5, ratio = 0.45) and Phoenix (10 (ig/m3 PM2.5, 22 (ig/m3 PM 10-2.5, ratio = 0.45)
contained most of PMi0 in the thoracic coarse mode.
Given the limited information available from AQS for PMi0_2.5 and the current National
Ambient Air Quality Standard for PMi0, analyses were performed on the more prevalent PMi0 data
acknowledging that PM10 incorporates both thoracic coarse and fine particles. The highest reported
3-yr avg PMi0 concentrations (>51 (ig/m3) occurred in two counties in southern California and five
counties in southern Arizona and central New Mexico. The lowest reported annual average PMi0
concentrations (< 20 ug/m3) were within 114 counties distributed fairly uniformly across the U.S. Of
the 15 CSAs/CBSAs investigated, the highest mean 24-h PMi0 concentrations was reported for
Phoenix (52 (ig/m3), considerably higher than the means for the other CSAs/CBSAs investigated.
The lowest was reported for Boston (17 (ig/m3) with New York, Philadelphia and Seattle only
slightly higher (19 (ig/m3).
Spatial variability in PM2 5 components obtained from the CSN varied considerably by
species. The highest annual average OC concentrations (>5 (ig/m3) were observed in the western and
southeastern U.S. Concentrations in the West peaked in the fall and winter, while concentrations in
the Southeast peaked anytime between spring and fall. Of the 15 CSAs/CBSAs investigated, OC was
the dominant PM25 component on an annual basis in the western cities, ranging from 34% of PM25
mass in Los Angeles to 58% in Seattle. EC exhibited less seasonal variability than OC and was
particularly stable in the eastern half of the U.S. Annual average EC concentrations greater than
1.5 (ig/m3 were present in Los Angeles, Pittsburgh, New York and El Paso. Concentrations of SO42~
were higher in the eastern U.S. resulting from higher SO2 emissions in the East compared with the
West. There is also considerable seasonal variability with higher SO42~ concentrations in the summer
months when the oxidation of SO2 proceeds at a faster rate than during the winter. Of the 15
December 2009 3-186
-------
CSAs/CBSAs selected, sulfate was the dominant PM2.5 component on an annual basis in the eastern
cities, ranging from 42% of PM2.5 mass in Chicago to 56% in Pittsburgh. NO3~ concentrations were
highest in California, with annual averages >4 (ig/m3 at many monitoring locations. There were also
elevated concentrations of NO3~ in the Midwest (>2 (ig/m3), with wintertime concentrations
exceeding 4 (ig/m3. In general, NO3~ was higher in the winter across the country, resulting from a
number of factors including: (1) lower temperatures which favor partitioning into particles; (2)
higher relative humidity, mainly in dry areas; (3) lower sulfate, allowing higher uptake of NO3~; and
(4) residential wood burning in specific areas of the U.S., especially in the Northwest. Exceptions
existed in Los Angeles and Riverside, where high NOjT readings appeared year-round. Crustal
material constituted a substantial fraction of PM2.5 year-round in Phoenix (28%) and Denver (16%),
and during the summer in Houston (26%).
Clearly there are variations in both PM2.5 mass and composition by city resulting from
numerous controlling variables (e.g., meteorology, the nature of sources, proximity to sources,
topography). These variables are frequently poorly characterized on a broad scale, making it
difficult to draw general conclusions regarding PM2 5 mass and composition across all cities within a
given geographic region.
Spatial Variability on the Urban and Neighborhood Scales
In general, PM2 5 has a longer atmospheric lifetime than PMi0_2.5 because larger particles have
a higher gravitational settling velocity. For PM2 5, most metropolitan areas exhibited high
correlations (generally >0.75) between monitoring sites out to a distance of 100 km. Notable
exceptions were Denver, Los Angeles and Riverside where correlations dropped below 0.75
somewhere between 20 and 50 km. Insufficient data were available in the 15 metropolitan areas to
perform similar analyses for PMi0_2.s using co-located, low volume FRM monitors. More abundant
PMio data, however, showed larger declines in inter-monitor correlations as a function of distance
relative to PM2 5. Atlanta, Boston, Denver, Los Angeles, New York City, Philadelphia, Phoenix,
Pittsburgh and Riverside all showed an average correlation of 0.75 at 40 km or greater monitor
separation while Birmingham, Chicago, Detroit, Houston and St. Louis had correlations that dropped
off much more quickly with distance (average correlation of 0.75 at 6 km or less monitor separation).
Furthermore, correlations between PMi0 concentrations exhibited substantially more scatter relative
to PM2 5. Shorter atmospheric lifetimes for PMi0 can result in local emission sources dominating
PMio annual average mass concentrations at particular monitors. Although the general understanding
of PM differential settling leads to an expectation of greater spatial heterogeneity in the PMio_25
fraction relative to the PM25 fraction in urban areas, deposition of particles as a function of size
depends strongly on local meteorological conditions, in particular on the degree of turbulence in the
mixing layer. Therefore, the findings from these 15 CSAs/CBSAs may not apply to all locations or at
all times.
Population density and associated building density are also important determinants of the
spatial distribution of PM concentrations. Inter-sampler correlations as a function of distance
between monitors obtained for sampler pairs located less than 4 km apart (i.e., on a neighborhood
scale) showed a shallower slope for PM25 than for PM10. The average correlation was 0.93 for PM25,
but it dropped to 0.70 for PMi0.
Few studies have performed direct comparisons of UFP measurements at multiple locations
within an urban area. A decrease in the number of UFPs was demonstrated with shifts from a
dominant mode at around 10 nm within 20 m of a freeway to a flattened dominant mode at around
50 nm at a distance of roughly 100-150 m. At the same time, accumulation mode particle number
concentration remained relatively constant to within -300 m from the freeway. These findings
suggest a high degree of spatial heterogeneity in UFPs compared with accumulation mode particles
on the urban scale.
3.9.1.4. Temporal Variability
A steady decrease in PM2 5 concentrations from 1999 (the beginning of nationwide monitoring
for PM2 5) to 2007 was observed in all 10 EPA Regions, with the 3-yr avg of the 98th percentile of
24-h PM25 concentrations dropping 10% over this time period. Similar trends in PMio concentrations
show a steady decline from 1988 to 2007 in all 10 EPA Regions.
December 2009 3-187
-------
Using hourly PM observations in the 15 metropolitan areas, diel variation showed peaks that
differ by PM size fraction and region. For PM2.5, a morning peak was observed starting at
approximately 6:00 a.m., corresponding with the start of morning rush hour. There was also an
evening PM2 5 concentration peak that was broader than the morning peak and extended into the
overnight period, likely reflecting a combination of evening rush hour and the concentration increase
caused by the usual collapse of the mixed layer after sundown. PM2.5 concentrations in Pittsburgh
remained elevated throughout the night, obscuring the morning peak. For PM1(j, all areas showed a
morning and afternoon peak in mean concentrations. The magnitude and duration of this peak varied
considerably by metropolitan area.
Studies indicate that UFPs in urban environments exhibit similar two-peaked diel patterns in
Los Angeles and the San Joaquin Valley as well as in Kawasaki City, Japan and Copenhagen,
Denmark. The afternoon peak in UFPs likely represents the combination of primary source
emissions such as evening rush-hour traffic and photochemical formation of secondary organic
aerosol and sulfate. Comparison between weekdays and Sundays as well as an urban street canyon
site and an urban background site in this figure suggest traffic is a major source of UFPs within a
street canyon during the morning rush hour. Any fluctuations or changes in the timing of the
individual daily peaks during the 3-yr period would result in a broadening of the distribution shown
in the diel plots.
3.9.1.5. Correlations between Copollutants
Correlations between PM size fractions and between PM and gaseous copollutants including
SO2, NO2, CO and O3 varied both seasonally and spatially between and within metropolitan areas.
On average, PM10 and PM2 5 were correlated with each other better than with the gaseous
copollutants. Correlations between PMi0 and PMi0_2.5 were greater in all locations than correlations
between PM2 5 and PMi0_2.5. Correlations between PMi0 and PMi0_2.5 were particularly high in
Denver and Phoenix (r > 0.88 in all seasons). There was relatively little seasonal variability in the
mean correlation between PM in both size fractions and SO2and NO2. CO, however, showed higher
correlations with PMi0 and PM2 5 on average in the winter compared with the other seasons. This
seasonality results in part because a larger fraction of PM is primary in origin during the winter. To
the extent that this primary component of PM is associated with common sources of NO2 and CO,
then higher correlations with these gaseous copollutants are to be expected. Increased atmospheric
stability in colder months would also reinforce these associations. The correlation between daily
maximum 8-h avg O3 and PM showed the highest degree of seasonal variability with positive
correlations on average in the spring, summer and fall, and negative correlations on average in the
winter. This situation arises as the result of seasonal differences in PM primary emissions and
photochemical production of secondary PM2 5 and O3
'3-
3.9.1.6. Source Contributions to PM
Results of receptor modeling calculations indicate that PM25 is produced mainly by
combustion of fossil fuel, either by stationary sources or by transportation. It is apparent that a
relatively small number of source categories, compared to the total number of chemical species that
typically are measured in ambient monitoring source receptor model studies, are needed to account
for the majority of the observed mass of PM in these studies. Trying to be more specific about
contributions from source categories could result in ambiguity. For example, quite different mobile
sources (e.g., trucks and locomotives) rely on diesel power and ancillary data is required to resolve
contributions from these sources. A compilation of study results shows that secondary sulfate
(mainly from EGUs), nitrate (from the oxidation of NOX emitted mainly from transportation and
EGUs), and primary mobile source categories constitute most of PM25 (and PMi0) in the East.
Fugitive dust, found mainly in the PMi0_2.5 size range, represents the largest source of ambient PMi0
in many locations in the western U.S. Quoted uncertainties in the source apportionment of
constituents in ambient aerosol samples typically range from 10 to 50%. A comparison of source
apportionment techniques indicated that the same major source categories of PM25 were consistently
identified by several independent groups working with the same data sets. Soil-, sulfate-, residual
oil-, and salt-associated mass were most clearly identified by the groups. Other sources with more
ambiguous signatures, such as vegetative burning and traffic-related emissions were less consistently
identified.
December 2009 3-188
-------
Spatial variability in source contributions across urban areas is an important consideration in
assessing the likelihood of exposure error in epidemiologic studies relating health endpoints to
sources. Concepts similar to those for using ambient concentrations as surrogates for personal
exposures apply here. Studies for PM2.5 indicate that intra-urban variability increases in the
following order: regional (e.g., secondary SO42~ from EGUs) < area (e.g., on-road mobile sources)
< point (e.g., stacks) sources. Only one study was available for PMi0_2.5, indicating a similar
ordering, but without a regional component (resulting from the short lifetime of PM10_2.5 compared to
transport times on the regional scale).
3.9.1.7. Policy-Relevant Background
The background concentrations of PM that are useful for risk and policy assessments
informing decisions about the NAAQS are referred to as policy-relevant background (PRB)
concentrations. PRB concentrations have historically been defined by EPA as those concentrations
that would occur in the U.S. in the absence of anthropogenic emissions in continental North America
defined here as the U.S., Canada, and Mexico. For this document, PRB concentrations include
contributions from natural sources everywhere in the world and from anthropogenic sources outside
continental North America. Background concentrations so defined facilitated separation of pollution
that can be controlled by U.S. regulations or through international agreements with neighboring
countries from those that were judged to be generally uncontrollable by the U.S. Over time
consideration of potential broader ranging international agreements may lead to alternative
determinations of which PM source contributions should be considered by EPA as part of PRB.
Contributions to PRB concentrations of PM include both primary and secondary natural and
anthropogenic components. For this document, PRB concentrations for the continental U.S. were
estimated using EPA's CMAQ modeling system, a deterministic CTM and with GEOS-Chem, a
global-scale model for CMAQ boundary conditions. PRB concentrations of PM2.5 were estimated to
be less than 1 (ig/m3 on an annual basis, with maximum daily average values in a range from 3.1 to
20 (ig/m3 and having a peak of 63 ug/m3 at the nine national park sites across the U.S. used to
evaluate model performance for this analysis. For further information on methods used in modeling
of PRB concentrations see Section 3.6, and for further information on the results of calculation of
PRB concentrations see Section 3.7.
3.9.2. Human Exposure
This section summarizes the findings from the recent exposure assessment literature, which
include the assessment of exposure to ambient PM, infiltration of ambient PM to indoor
environments, and source apportionment of PM exposure. This summary is intended to support the
interpretation of the findings from epidemiologic studies. For more detailed explication see
Section 3.8.
3.9.2.1. Characterizing Human Exposure
A number of techniques have been applied in the literature to model human exposure to PM.
Several studies have used time-weighted microenvironmental models to define total or ambient PM
exposure. Time-activity diaries or global positioning systems have been employed to capture the
time-basis for those models. Stochastic population exposure models, such as APEX and SHEDS, are
applied for PM exposure risk assessment among the population. Concentrations from chemistry
transport models have also been used to provide input to the stochastic exposure models at particular
locations. LUR models have been applied for individual exposures at the intra-urban scale to
examine exposure to pollution surrogates, such as traffic counts, land use, or topographic variables.
Source proximity and kriging have also been applied. GIS-based models have been used to model
exposure over large regions (e.g., for the Nurse's Health Study) using spatial smoothing models of
AQS data and incorporating GIS-based and meteorological covariates. GIS approaches have also
been used for intra-urban scale exposure studies. These methods all have their own uses and caveats.
LUR is an adaptable framework allowing adaptation to localized conditions but might best be
applied in relatively spatially homogeneous areas. In a comparison of LUR with kriging, kriging
produced slightly attenuated mortality risk estimates for New York City, while for Los Angeles,
December 2009 3-189
-------
kriging did not exhibit as much spatial variability as LUR. Source proximity modeling is relatively
simple to apply but is limited because other confounding covariates, such as socioeconomic status,
may be related to source proximity. Additionally, source proximity models do not incorporate time-
activity data.
New advancements in personal and microenvironmental monitoring techniques have been
reported. Personal monitoring developments include new models of cascade impactors and cyclones
to sample in the UF size range and miniature monitors for species detection. Additionally, new work
on microenvironmental modeling using mobile platforms and GPS technology has been reported.
The reader is referred to the 2004 PM AQCD (U.S. EPA, 2004, 056905) for descriptions of most
real-time and filter-based personal and microenvironmental PM monitors currently available.
3.9.2.2. Spatial Scales of PM Exposure Assessment
Assessing population-level exposure at the urban scale is particularly relevant for time-series
epidemiologic studies, which provide information on the relationship between health effects and
community-average exposure, rather than variations in individual exposure. The correlation between
the PM concentration measured at a central-site community ambient monitor and the true community
average concentration depends on the spatial distribution of the PM, location of the monitoring site
chosen to represent the community average, and division of the community by terrain features or
source locations into several sub-communities that differ in the temporal pattern of pollution.
Concentrations of SO42~ and some components of SOA measured at central-site monitors are
expected to be uniform in urban areas given the regional nature of their sources. However, this is not
true for primary components like EC whose sources are strongly spatially variable in urban areas.
Given that roughly 90% of an individual's day is spent indoors, assessment of exposure to infiltrated
ambient SO42~, whose formation and dispersion also occurs over urban-to-regional scales and whose
size distribution is in the accumulation mode, is commonly used to assess ambient PM2.5 exposure.
This technique has also been applied to assess PMi0_2.5 exposure but likely with more error than for
PM2.5 because PMi0_2.5 is more highly spatially variable than PM2 5. Source apportionment techniques
have also been applied to assess urban-scale PM2 5 exposures using community-based ambient
monitoring, outdoor, and indoor samples.
At micro-to-neighborhood scales, heterogeneity of sources and topography may cause more
variability in exposure. This is particularly true for PMi0_2.5 and for UFPs, both of which are more
highly spatially variable than PM2 5 Particle chemistry and source behavior also contribute to spatial
heterogeneity of PM concentration. Some studies, conducted mainly in Europe, have found personal
PM2 5 and PMi0 exposures for pedestrians in street canyons to be higher than ambient concentrations
measured by urban background ambient monitors. Likewise, microenvironmental UFP
concentrations were observed to be substantially higher in near-road environments, street canyons,
and tunnels when compared with other environments in urban areas. In-vehicle UFP exposures can
also be important. As a result, ambient monitors located at background, central urban, road side, or
near-residential sites might not reflect peak exposures to individuals who commute.
PM infiltration factors, F^f, depend on particle size, chemical composition, season, and region
of the country. Infiltration can best be modeled dynamically based on a distribution of air exchange
and deposition PM loss rates rather than being represented by a single value. There is significant
variability within and across regions of the country with respect to indoor exposures to ambient PM.
Infiltrated ambient PM concentrations depend in part on the ventilation properties of the building or
vehicle in which the person is exposed. Season is important to PM infiltration because it affects the
ventilation practices used, and ambient temperature and humidity conditions affect the transport,
dispersion, and size distribution of PM. Residential air exchange rates have been observed to be
higher in summer for regions with low air conditioning usage, and regional differences in air
exchange rates (Southwest < Southeast < Northeast < Northwest) also reflect ventilation practices.
Differential infiltration occurs as a function of PM size and composition. PM infiltration is largest
for accumulation mode particles, and decreases for UFPs lost to diffusion and for coarse PM lost
through inertial impaction mechanisms. Differential infiltration by size fraction can affect exposure
estimates if not properly characterized.
December 2009 3-190
-------
3.9.2.3. Multicomponent and Multipollutant PM Exposures
Emission inventories and source apportionment studies suggest that sources of PM exposure
vary by region. Comparison of studies performed in the eastern U.S. with studies performed in the
western U.S. suggest that the contribution of SO42~ to personal exposure is higher for the East
(16-46%) compared with the West (-4%) and that motor vehicle emissions and secondary NO3" are
larger sources of personal exposure for the West (-9%) as compared with the East (-4%). Results of
source apportionment studies of personal exposure to SO^ indicate that personal SO^ exposures
are mainly attributable to ambient sources. Source apportionment for OC and EC is difficult because
they originate from both indoor and outdoor sources. Exposure to OC of indoor and outdoor origin
can be distinguished by the presence of aliphatic C-H groups generated indoors, since outdoor
concentrations of aliphatic C-H are low. Trace metal studies have shown variable results regarding
personal exposure to ambient constituents with significant variation among cities and over seasons
that can be related to incinerator operation, fossil fuel combustion, biomass combustion (wildfires),
and presence of crustal materials in the built environment, among other sources. Differential
infiltration is also affected by variations in particle composition and volatility. For example EC
infiltrates more readily than OC. This can lead to outdoor-indoor differentials in PM toxicity.
A number of studies have examined whether gaseous copollutants could act as surrogates for
exposure to ambient PM. Several studies have concluded that ambient concentrations of O3, NO2,
and SO2 are associated with the ambient component of personal exposure to total PM2.5 as opposed
to the ambient component of personal exposures to the gases. However, in some studies this result
may have arisen in part because personal exposure to the gases was often beneath the detection
limits of the personal monitoring devices. Thus, the evidence that ambient gases can be considered
surrogates of PM2.5 exposure is mixed. It is likely that associations between ambient gases and
personal exposure to PM2 5 of ambient origin exist, but they are complex and vary by season and
location.
3.9.2.4. Implications for Epidemiologic Studies
The importance of exposure error varies with study design based on the spatial and temporal
aspects of the design. For PM epidemiology studies, source characteristics, particle size distribution,
and particle composition are also important factors in interpreting exposure error for an
epidemiology study. Potential sources of error that could influence estimates of PM exposure include
measurements, use of surrogates for PM exposure, modeling, spatial variability, temporal variability,
and compositional differences.
PM exposure estimates are subject to monitoring and modeling errors. Ambient and personal
exposure monitoring errors can bias health effects estimates if the error is strongly correlated with
the measurements of concentration. This can be an issue for sampling semi-volatile organic
compounds in PM, especially where PM exposures in cities with different PM composition are
compared. Ambient monitor height also affects estimates of exposure because PM concentration
varies as a function of height. Within a street canyon, changes in wind direction and speed cause
significant variability over a small distance. Wind tunnel studies have shown street canyon effects
exist for suburban settings as well as for heavily urbanized settings. Additionally, model-based
exposure estimates are subject to errors related to the spatial resolution of the modeling technique
and the measurement-based inputs used.
Variations in PM and its components could lead to errors in using ambient PM measures as
surrogates for PM exposure. PM2 5 concentrations are relatively well-correlated across monitors in
the urban areas examined. Correlation coefficients tend to be lower, and concentration differences
tend to be higher between PMi0 monitoring sites than between PM2 5 monitoring sites. Likewise,
studies have shown UFPs to be more spatially variable across urban areas. Even if PM25, PMi0_2.5,
and UFP concentrations measured at sites within an urban area are highly correlated, significant
differences in their concentrations can occur on any given day. The degree of urban-scale spatial
variability in PM concentrations varies across the country and with size fraction. Current information
suggests that UFPs, PMi0_2.5s and many PM components are more spatially variable than PM2 5. These
factors should be considered in using data obtained from monitoring networks to estimate
community-scale human exposure to ambient PM, and caution should be exercised in extrapolating
conclusions obtained from one urban area to another.
December 2009 3-191
-------
Community time-series epidemiologic studies use the average community PM concentration
as a surrogate for the average personal exposure to ambient PM. The resulting health effect risk
estimate, based on the average community ambient concentration, differs from the risk that would be
estimated if the average community ambient exposure were used in the epidemiologic study.
However, the risk estimate based on the ambient concentration gives the change in health effects
resulting from a change in ambient PM concentration and is, therefore, an appropriate measure for
risk assessment and risk management. Variations in ambient concentrations across a community,
variations in individual ambient exposures around the community average, and seasonal or daily
variation in the ambient exposure estimate may increase standard errors of PM health effects
estimates, making it more difficult to detect a true underlying association between the correct
exposure metric and the health outcome studied. Likewise, sampling time interval and lag time
selection both determine whether an epidemiologic model captures the phenomena of interest with
sufficient resolution. The use of the community average ambient PM2.5 concentration as a surrogate
for the community average personal exposure to ambient PM2.5 is not expected to change the
principal conclusions from PM2 5 epidemiologic studies that use community average health and
pollution data. Several recent studies support this by showing how the ambient component of
personal exposure to PM2.5 could be estimated using various tracer and source apportionment
techniques and that it is highly correlated with ambient concentrations of PM2 5. These studies also
show that the non-ambient component of personal exposure to PM25 is basically uncorrelated with
ambient PM2 5 concentrations. For long-term studies that use differences in long-term community
average ambient PM concentrations as an exposure metric, the effect of possible community-to-
community differences in the average ambient exposure factor or in the average non-ambient
exposure are less understood. For panel epidemiologic studies, the most appropriate exposure metric
may depend on the health outcome measured. However, sufficient information should be obtained to
enable determining the association of the health outcome with ambient concentration, ambient
exposure, non-ambient exposure, and total personal exposure.
Exposure error may occur if a measured PM component acts as a surrogate for another PM
constituent. Differences between composition of outdoor and indoor ambient PM may also cause
error in exposure assessment related either to differential losses of UF or coarse PM from diffusion,
evaporation of semi-volatile PM, or impaction. The resulting differences in PM size distribution and
chemical composition between indoor-ambient PM and outdoor-ambient PM are expected to cause
differences in toxicity that could affect health outcomes. Lack of information regarding these
relationships adds uncertainty to the health effects estimate.
December 2009 3-192
-------
Chapter 3 References
ACGIH. (2005). TLVs and BEIs: Based on the documentation of the threshold limit values for chemical substances and
physical agents and biological exposure indices. American Conference of Governmental Industrial Hygienists.
Cincinnati, Ohio. 156188
Adams HS; Nieuwenhuijsen MJ; Colvile RN; McMullen MAS; Khandelwal P. (2001). Fine particle (PM25) personal
exposure levels in transport microenvironments, London, UK. Sci Total Environ, 279: 29-44. 019350
Adar SD; Davey M; Sullivan JR; Compher M; Szpiro A; Liu L-J. (2008). Predicting airborne particle levels aboard
Washington State school buses . Atmos Environ, 42: 7590-7599. 191200
Adgate JL; Mongin SJ; Pratt GC; Zhang J; Field MP; Ramachandran G; Sexton K. (2007). Relationships between personal,
indoor, and outdoor exposures to trace elements in PM(2.5). Sci Total Environ, 386: 21-32. 156196
Al-Horr R; Samanta G; Dasgupta PK. (2003). A Continuous Analyzer for Soluble Anionic Constituents and Ammonium in
Atmospheric Particulate Matter. Environ Sci Technol, 37: 5711-5720. 153951
Albinet A; Leoz-Garziandia E; Budzinski H; Villenave E. (2007). Sampling precautions for the measurement of nitrated
polycyclic aromatic hydrocarbons in ambient air. Atmos Environ, 41: 4988-4994. 154426
Alexis A; Garcia A; Nystrom M; Rosenkranz K. (2001). The 2001 California almanac of emissions and air quality.
Retrieved 10-AUG-05, from http://www.arb.ca.gov/aqd/almanac/almanac01/almanac01.htm. 079886
Allen GA; Harrison D; Koutrakis P. (2001). A New Method for Continuous Measurement of Sulfate in the Ambient
Atmosphere. Presented at American Association for Aerosol Research, 20th Annual Conference. 156205
Allen R; Larson T; Sheppard L; Wallace L; Liu L-JS. (2003). Use of real-time light scattering data to estimate the
contribution of infiltrated and indoor-generated particles to indoor air. Environ Sci Technol, 37: 3484-3492. 053578
Allen R; Wallace L; Larson T; Sheppard L; Liu LJ. (2004). Estimated hourly personal exposures to ambient and
nonambient particulate matter among sensitive populations in Seattle, Washington. J Air Waste Manag Assoc, 54:
1197-1211. 190089
Andracchio A; Cavicchi C; Tonelli D; Zappoli S. (2002). A new approach for the fractionation of water-soluble organic
carbon in atmospheric aerosols and cloud drops. Atmos Environ, 36: 5097-5107. 155657
Andreae MO; Gelencser A. (2006). Black carbon or brown carbon? The nature of light-absorbing carbonaceous aerosols.
Atmos Chem Phys, 6: 3131-3148. 156215
Appel K; Bhave PV; Gilliland AB; Sarwar G; Roselle SJ. (2008). Evaluation of the community multiscale air quality
(CMAQ) model version 4.5: Sensitivities impacting model performance; Part IIa€"particulate matter. Atmos
Environ, 42: 6057-6066. 155660
Appel KW; Gilliland A; EderB. (2005). An operational evaluation of the 2005 release of models-3 CMAQ version 45.
National Oceanic and Atmospheric Administration-Air Resources. Washington DC. 089227
Arhami M; Kuhn T; Fine PM; Delfino RJ; Sioutas C. (2006). Effects of Sampling Artifacts and Operating Parameters on
the Performance of a Semicontinuous Particulate Elemental Carbon/Organic Carbon Monitor. Environ Sci Technol,
40: 945-954. 156224
Arnold JR; Dennis RL; Tonnesen GS. (2003). Diagnostic evaluation of numerical air quality models with specialized
ambient observations: testing the Community Multiscale Air Quality modeling system (CMAQ) at selected SOS 95
ground sites. Atmos Environ, 37: 1185-1198. 087579
Arnott WP; Moosmilller H; Sheridan PJ; Ogren JA; Raspet R; Slaton WV; Hand JL; Kreidenweis SM; Collett JL Jr. (2003).
Photoacoustic and filter-based ambient aerosol light absorption measurements: Instrument comparisons and the role
of relative humidity. J Geophys Res, 108: AAC 15-1. 037711
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009 3-193
-------
Artinano B; Querol X; Salvador P; Rodriguez S; Alonso DG; Alastuey A. (2001). Assessment of airborne particulate levels
in Spain in relation to the new EU-directive. Atmos Environ, 35: S43-S53. 190099
Auvray M; Bey I. (2005). Long-range transport to Europe: Seasonal variations and implications for the European ozone
budget. J Geophys Res, 110: D11303.1-D11303.22. 156237
Ayers GR (2004). Potential for simultaneous measurement of PM10, PM25 and PM1 for air quality monitoring purposes
using a single TEOM. Atmos Environ, 38: 3453-3458. 097440
Bae MS. (2007). Intercomparison of Real Time Ammonium Measurements at Urban and Rural Locations in New York.
Aerosol Sci Technol, 41: 329-341. 155669
Bae MS; Schauer JJ; Deminter JT; Turner JR. (2004). Hourly and daily patterns of particle-phase organic and elemental
carbon concentrations in the urban atmosphere. J Air Waste Manag Assoc, 54: 823-833. 156243
Bae MS; Schwab JJ; Zhang Q; Hogrefe O; Demerjian KL; Weimer S; Rhoads K; Orsini D; Venkatachari P; Hopke PK.
(2007). Interference of organic signals in highly time resolved nitrate measurements by low mass resolution aerosol
mass spectrometry. J Geophys Res, 112: D22305-D22305 (1-16). 156244
Baldauf R; Thoma E; Hays M; Shores R; Kinsey J; Gullett B; Kimbrough S; Isakov V; Long T; Snow R; Khlystov A;
Weinstein J; Chen FL; Seila R; Olson D; Gilmour I; Cho SH; Watkins N; Rowley P; Bang J. (2008). Traffic and
meteorological impacts on near-road air quality: Summary of methods and trends from the raleigh near-road study.
J Air Waste Manag Assoc, 58: 865-878. 190239
Baldauf R; Thoma E; Khlystov A; Isakov V; Bowker G; Long T; Snow R. (2008). Impacts of noise barriers on near-road air
quality. Atmos Environ, 42: 7502-7507. 191017
Baldauf R; Watkins N; Heist D; Bailey C; Rowley P; Shores R. (2009). Near-road air quality monitoring: Factors affecting
network design and interpretation of data. Air Qual. Atmos. Health , 2: 1-9. 191766
Baldauf RW; Gabele P; Crews W; SnowR; Cook JR. (2005). Criteria and air-toxic emissions from in-use automobiles in
the national low-emission vehicle program. J Air Waste Manag Assoc, 55: 1263-1268. 191184
Barn P; Larson T; Noullett M; Kennedy S; Copes R; Brauer M. (2008). Infiltration of forest fire and residential wood
smoke: an evaluation of air cleaner effectiveness. JExpo Sci Environ Epidemiol, 18: 503-511. 156252
Barreto RP; Albuquerque FC; Netto ADP (2007). Optimization of an improved analytical method for the determination of
1-nitropyrene in milligram diesel soot samples by high-performance liquid chromatography-mass spectrometry. J
Chromatogr, 1163: 219-227. 155676
Bauer H; Schueller E; Weinke G; Berger A; Hitzenberger R; Marr I; Puxbaum H. (2008). Significant contributions of
fungal spores to the organic carbon and to the aerosol mass balance of the urban atmospheric aerosol. Atmos
Environ, 42: 5542-5549. 189986
Baxter LK; Barzyk TM; Vette VF; Croghan C; Williams RW. (2008). Contributions of diesel truck emissions to indoor
elemental carbon concentrations in homes in proximity to Ambassador Bridge . Atmos Environ, 42: 9080-9086 .
191194
Baxter LK; Clougherty JE; Laden F; Levy JI. (2007). Predictors of concentrations of nitrogen dioxide, fine particulate
matter, and particle constituents inside of lower socioeconomic status urban homes. J Expo Sci Environ Epidemiol,
17: 433-444. 092726
Baxter LK; Clougherty JE; Paciorek CJ; Wright RJ; Levy JI. (2007). Predicting residential indoor concentrations of
nitrogen dioxide, fine particulate matter, and elemental carbon using questionnaire and geographic information
system based data. Atmos Environ, 41: 6561-6571. 092725
Behrentz E; Fitz DR; Pankratz DV; Sabin LD; Colome SD; Fruin SA; Winer AM. (2004). Measuring self-pollution in
school buses using a tracer gas technique. Atmos Environ, 38: 3735-3746. 155682
Bein KJ; Zhao Y; Wexler AS; Johnston MV. (2005). Speciation of size-resolved individual ultrafine particles in Pittsburgh,
Pennsylvania. J Geophys Res, 110: D07S05. 156265
Bell ML; Dominici F; Ebisu K; Zeger SL; Samet JM. (2007). Spatial and Temporal Variation in PM2. 5 Chemical
Composition in the United States for Health Effects Studies. Environ Health Perspect, 115: 989-995. 155683
Bell ML; Ebisu K; Peng RD; Dominici F. (2009). Adverse health effects of particulate air pollution: modification by air
conditioning. Epidemiology, 20: 682-686. 191007
December 2009 3-194
-------
Bell ML; Ebisu K; Peng RD; Walker J; Samet JM; Zeger SL; Dominic F. (2008). Seasonal and regional short-term effects
of fine particles on hospital admissions in 202 U.S. counties, 1999-2005. Am JEpidemiol, 168: 1301-1310. 156266
Bell ML; Kim JY; Dominici F. (2007). Potential confounding of particulate matter on the short-term association between
ozone and mortality inmultisite time-series studies. Environ Health Perspect, 115: 1591-1595. 093256
Benkovitz CM; Scholtz MT; Pacyna J; Tarrason L; Dignon J; Voldner EC; Spiro PA; Logan JA; Graedel TE. (1996). Global
gridded inventories of anthropogenic emissions of sulfur and nitrogen: North Atlantic Regional
Experiment(NARE). J Geophys Res, 101: 29239-29253. 156267
Bennett DH; Koutrakis P. (2006). Determining the infiltration of outdoor particles in the indoor environment using a
dynamic model. J Aerosol Sci, 37: 766-785. 089184
Binkowski F; Roselle S. (2003). Models-3 Community Multiscale Air Quality(CMAQ) model aerosol component 1. Model
description. J Geophys Res, 108: 4183. 191769
Binkowski FS; Arunachalam S; Adelman Z; Pinto JP. (2007). Examining photolysis rates with a prototype online
photolysis module in CMAQ. J Appl Meteor Climatol, 46: 1252-1256. 090563
Biswas S; Hu S; Verma V; Herner J; Robertson W; Ayala A; Sioutas C. (2008). Physical properties of particulate matter
(PM) from late model heavy-duty diesel vehicles operating with advanced PM and NOx emission control
technologies. Atmos Environ, 42: 5622-5634. 189969
Biswas S; Verma V; Schauer JJ; Sioutas C. (2009). Chemical speciation of PM emissions from heavy-duty diesel vehicles
equipped with diesel particulate filter (DPF) and. Atmos Environ, 43: 1917-1925. 189880
Blanchard P; Brook JR; Brazil P. (2002). Chemical characterization of the organic fraction of atmospheric aerosol at two
sites in Ontario, Canada. J Geophys Res, 107: ICC10.1-ICC10.8 . 047598
Blando JD; Turpin BJ. (2000). Secondary organic aerosol formation in cloud and fog droplets: a literature evaluation of
plausibility. Atmos Environ, 34: 1623-1632. 155692
Bond TC; Streets DG; Yarber KF; Nelson SM; Woo J-H; Klimont Z. (2004). A technology-based global inventory of black
and organic carbon emissions from combustion. J Geophys Res, 109: D14203. 056389
Borak J; Sirianni G; Cohen HJ; Chemerynski S; Wheeler R. (2003). Comparison of NIOSH 5040 Method versus
Aethalometer™ to Monitor Diesel Particulate in School Buses and at Work Sites. AIHA J, 64: 260-268. 156284
Borrego C; Tchepel O; Costa AM; Martins H; Ferreira J; Miranda AI. (2006). Traffic-related particulate air pollution
exposure in urban areas. Atmos Environ, 40: 7205-7214. 155697
Bortnick SM; Coutant BW; Eberly SI. (2002). Using continuous PM2.5 monitoring data to report an air quality index. J Air
Waste Manag Assoc, 52: 104-112. 156285
Brauer M; Hoek G; van Vliet P; Meliefste K; Fischer P; Gehring U; Heinrich J; Cyrys J; Bellander T; Lewne M;
Brunekreef B. (2003). Estimating long-term average particulate air pollution concentrations: application of traffic
indicators and geographic information systems. Epidemiology, 14: 228-239. 155702
Braun A. (2009). Two-process model for the atmospheric weathering, oxidation and ageing of diesel soot. Geophys Res
Lett, 36: L07810.189997
Briggs D. (2000). Exposure assessment. In Spatial Epidemiology: Methods and Applications (pp. 335-359). Oxford:
Oxford University Press. 191772
Briggs DJ; Collins S; Elliott P; Fischer P; Kingham S; Lebret E; Pryl K; Van Reeuwijk H; Smallbone K; Van Der Veen A.
(1997). Mapping urban air pollution using GIS: a regression-based approach. Int J Geogr Inform Sci, 11: 699-718.
025950
Briggs DJ; de Hoogh K; Morris C; Gulliver J. (2008). Effects of travel mode on exposures to particulate air pollution.
Environ Int, 34: 12-22. 156294
Brock CA; Hudson PK; Lovejoy ER; Sullivan A; Nowak JB; Huey LG; Cooper OR; Cziczo DJ; de Gouw J; Fehsenfeld
FC. (2004). Particle characteristics following cloud-modified transport from Asia to North America. J Geophys Res,
109: D23S26. 156295
Brook JR; Sirois A; Clarke JF. (1996). Comparison of dry deposition velocities for SO2, HNO3, and SO4(2-) estimated
with two inferential models. Water Air Soil Pollut, 87: 205-218. 024023
December 2009 3-195
-------
Brown AS; Yardley RE; Quincey PG; Butterfield DM. (2006). Studies of the effect of humidity and other factors on some
different filter materials used for gravimetric measurements of ambient particulate matter. Atmos Environ, 40:
4670-4678. 097665
Bruno P; Caselli M; de Gennaro G; Tutino M. (2007). Determination of poly cyclic aromatic hydrocarbons (PAHs) in
particulate matter collected with low volume samplers. Talanta, 72: 1357-1361. 155706
Burke JM; Zufall MJ; Ozkaynak H. (2001). A population exposure model for particulate matter: case study results for
PM25 in Philadelphia, PA. J Expo Sci Environ Epidemiol, 11: 470-489. 014050
Burnett RT; Goldberg MS. (2003). Size-fractionated particulate mass and daily mortality in eight Canadian cities. Health
Effects Institute. Boston, MA. 042798
Buser MD; Parnell Jr CB; Shaw BW; Lacey RE. (2007). Particulate matter sampler errors due to the interaction of particle
size and sampler performance characteristics: Ambient PM 10 samplers. Trans ASAE, 50: 229-240. 156311
Buser MD; Parnell Jr CB; Shaw BW; Lacey RE. (2007). Particulate matter sampler errors due to the interaction of particle
size and sampler performance characteristics: Ambient PM2. 5 samplers. Trans ASAE, 50: 241-254. 156310
Buser MD; Parnell Jr CB; Shaw BW; Lacey RE. (2007). Particulate matter sampler errors due to the interaction of particle
size and sampler performance characteristics: Background and Theory. Trans ASAE, 50: 221-228. 156312
Butler AJ; Andrew MS; Russell AG (2003). Daily sampling Of PM2. 5 in Atlanta: Results of the first year of the
Assessment of Spatial Aerosol Composition. J Geophys Res, 108: 8415. 156313
Buzorius G; Hameri K; Pekkanen J; Kulmala M. (1999). Spatial variation of aerosol number concentration in Helsinki city.
Atmos Environ, 33: 553-565. 081205
Bytnerowicz A; Miller PR; Olszyk DM. (1987). Dry deposition of nitrate, ammonium and sulfate to a Ceanothus
crassifolius canopy and surrogate surfaces. Atmos Environ, 21: 1749-1757. 036493
Byun D; Schere KL. (2006). Review of the governing equations, computational algorithms, and other components of the
models-3 community multiscale air quality (CMAQ) modeling system. Appl. Mech. Rev, 59: 51-77. 090560
ByunDW; Ching JKS (eds). (1999). Science algorithms of the EPA models-3 community multiscale air quality (CMAQ)
modeling system. U.S. Environmental Protection Agency, Office of Research and Development. Washington, DC.
EPA/600-R-99-030. http://www.epa.gov/asmdnerl/CMAO/CMAQscienceDoc.html. 156314
Cabada JC; Rees S; Takahama S; Khlystov A; Pandis SN; Davidson CI; Robinson AL. (2004). Mass size distributions and
size resolved chemical composition of fine particulate matter at the Pittsburgh supersite. Atmos Environ, 38: 3127-
3141. 148859
Cadle SH; Mulawa P; Groblicki P; Laroo C; Ragazzi RA; Nelson K; Gallagher G; Zielinska B. (2001). In-use light-duty
gasoline vehicle particulate matter emissions on the FTP, REP05, and UC cycles [final report]. 017192
Cadle SH; Mulawa PA; Hunsanger EC; Nelson K; Ragazzi RA; Barrett R; Gallagher GL; Lawson DR; Knapp KT; Snow R.
(1999). Composition of light-duty motor vehicle exhaust particulate matter in the Denver, Colorado area. Environ
Sci Technol, 33: 2328-2339. 007636
Carter WPL. (1990). A detailed mechanism for the gas-phase atmospheric reactions of organic compounds. Atmos Environ,
24:481-518.042893
Case MW; Williams R; Yeatts K; Chen F-L; Scott J; Svendsen E; Devlin RB. (2008). Evaluation of a direct personal coarse
particulate matter monitor. Atmos Environ, 42(19): 4446-4452. 155149
Cass GR. (1998). Organic molecular tracers for particulate air pollution sources. Trends Analyt Chem, 17: 356-366. 155716
Cass GR; Hughes LA; Bhave P; Kleeman MJ; Allen JO; Salmon LG (2000). The chemical composition of atmospheric
ultrafine particles. Proc Biol Sci, 358: 2581-2592. 020680
Chakrabarti B; Singh M; Sioutas C. (2004). Development of a Near-Continuous Monitor for Measurment of the Sub-150
nm PM Mass Concentration. Aerosol Sci Technol, 38: 239-252. 157426
Chang C-H; Meroney RN. (2003). Concentration and flow distributions in urban street canyons: wind tunnel and
computational data. J WindEng Ind Aerod, 91: 1141-1154. 090298
Chang CT; Tsai CJ. (2003). A model for the relative humidity effect on the readings of the PM10 beta-gauge monitor. J
Aerosol Sci, 34: 1685-1697. 155718
December 2009 3-196
-------
Chang L-T; Koutrakis P; Catalano PJ; Suh HH. (2003). Assessing the importance of different exposure metrics and time-
activity data to predict 24-h personal PM25 exposures. J Toxicol Environ Health A, 66: 1825-1846. 053789
Charron A; Harrison RM; Moorcroft S; Booker J. (2004). Quantitative interpretation of divergence between PM10 and
PM25 mass measurement by TEOM and gravimetric (Partisol) instruments. Atmos Environ, 38: 415-423. 053849
Chen FL; Williams R; Svendsen E; Yeatts K; Creason J; Scott J; Terrell D; Case M. (2007). Coarse particulate matter
concentrations from residential outdoor sites associated with the North Carolina Asthma and Children's
Environment Studies (NC-ACES). Atmos Environ, 41: 1200-1208. 147318
Chen LWA; Chow JC; Watson JG; MoosmSuller H; Arnott WP. (2004). Modeling reflectance and transmittance of quartz-
fiber filter samples containing elemental carbon particles: Implications for thermal/optical analysis. J Aerosol Sci,
35: 765-780. 199501
Chiapello I; Moulin C; Prospero JM. (2005). Understanding the long-term variability of African dust transport across the
Atlantic as recorded in both Barbados surface concentrations and large-scale Total Ozone Mapping Spectrometer
(TOMS) optical thickness. J Geophys Res, 110: D18S10. 156339
Ching J; Herwehe J; Swall J. (2006). On joint deterministic grid modeling and sub-grid variability conceptual framework
for model evaluation. Atmos Environ, 40: 4935-4945. 090300
Chow JC. (2007). The application of thermal methods for determining chemical composition of carbonaceous aerosols: A
review. J Environ Sci Health A Tox Hazard Subst Environ Eng, 42: 1521-1541. 157209
Chow JC; Doraiswamy P; Watson JG; Chen LW; Ho SS; Sodeman DA. (2008). Advances in integrated and continuous
measurements for particle mass and chemical composition. J Air Waste Manag Assoc, 58: 141-163. 156355
Chow JC; Watson JG; Chen LW; Chang MC; Robinson NF; Trimble D; Kohl S. (2007). The IMPROVE_A temperature
protocol for thermal/optical carbon analysis: maintaining consistency with a long-term database. J Air Waste Manag
Assoc, 57: 1014-1023. 156354
Chow JC; Watson JG; Chen LWA; Arnott WP; Moosmuller H; Fung K. (2004). Equivalence of elemental carbon by
thermal/optical reflectance and transmittance with different temperature protocols. Environ Sci Technol, 38: 4414-
4422. 156347
Chow JC; Watson JG; Louie PKK; Chen LWA; Sin D. (2005). Comparison of PM2.5 carbon measurement methods in
Hong Kong, China. Environ Pollut, 137: 334-344. 155728
Chow JC; Watson JG; Park K; Lowenthal DH; Robinson NF; Magliano KA. (2006). Comparison of particle light scattering
and fine particulate matter mass in central California. J Air Waste Manag Assoc, 56: 398-410. 099031
Chu D; Kaufman Y; Zibordi G; Chern J; Mao J; Li C; Holben B. (2003). Global monitoring of air pollution over land from
the Earth Observing System-Terra Moderate Resolution Imaging Spectroradiometer (MODIS). J Geophys Res, 108:
4661. 190049
Chung A; Herner JD; Kleeman MJ. (2001). Detection of alkaline ultrafine atmospheric particles at Bakersfield, California.
Environ Sci Technol, 35: 2184-2190. 017105
Claeys M; Graham B; Vas G; Wang W; Vermeylen R; Pashynska V; Cafmeyer J; Guyon P; Andreae MO; Artaxo P;
Maenhaut W. (2004). Formation of secondary organic aerosols through photooxidation of isoprene. Sci New York,
303: 1173-1176.058608
Clarke JF; Edgerton ES; Martin BE. (1997). Dry deposition calculations for the clean air status and trends network. Atmos
Environ, 31: 3667-3678. 025022
Clough WS. (1975). The deposition of particles on moss and grass surfaces. Atmos Environ, 9: 1113-1119. 070850
Cohen M; Adar S; Allen R; Avol E; Curl C; Gould T; Hardie D; Ho A; Kinney P; Larson T. (2009). Approach to estimating
participant pollutant exposures in the multi-ethnic study of atherosclerosis and air pollution (MESA Air). Environ
Sci Technol, 43: 4687-4693. 190639
Conny JM; Klinedinst DB; Wight SA; Paulsen JL. (2003). Optimizing thermal-optical methods for measuring atmospheric
elemental (black) carbon: A response surface study. Aerosol Sci Technol, 37: 703-723. 145948
Conny JM; Norris GA; Gould TR. (2009). Factorial-based response-surface modeling with confidence intervals for
optimizing thermal-optical transmission analysis of atmospheric black carbon. Anal Chim Acta, 635: 144-156.
191999
December 2009 3-197
-------
Cooke WF; Liousse C; Cachier H; Feichter J. (1999). Construction of a 1 ° x 1 ° fossil fuel emission data set for
carbonaceous aerosol and implementation and radiative impact in the ECHAM 4 model. J Geophys Res, 104:
22137-22162. 156365
Corburn J. (2007). Urban land use, air toxics and public health: Assessing hazardous exposures at the neighborhood scale.
Environ Impact Assess Rev, 27: 145-160. 155738
Crimmins BS; Baker JE. (2006). Improved GC/MS methods for measuring hourly PAH and nitro-PAH concentrations in
urban particulate matter. Atmos Environ, 40: 6764-6779. 097008
Crist KC; Liu B; Kim M; Deshpande SR; John K. (2008). Characterization of fine particulate matter in Ohio: Indoor,
outdoor, and personal exposures. Environ Res, 106: 62-71. 156372
Cyrys J; Heinrich J; Hoek G; Meliefste K; Lewne M; Gehring U; Bellander T; Fischer P; Van Vliet P; Brauer M;
Wichmann HE; Brunekreef B. (2003). Comparison between different traffic-related particle indicators: elemental
carbon (EC), PM25 mass, and absorbance. J Expo Sci Environ Epidemiol, 13: 134-143. 049634
Dasch JM. (1987). Measurement of dry deposition to surfaces in deciduous and pine canopies. Environ Pollut, 44: 261-277.
036496
Davidson CI; Wu Y-L. (1990). Dry deposition of particles and vapors. In Lindberg, S. E.; Page, A. L.; Norton, S. A.
(Ed.),Acidic precipitation: v. 3, sources, deposition, and canopy interactionsNew York, NY: Springer-Verlag.
036799
deCastro BR; Sax AN; Chillrud SN; Kinney PL; Spengler JD. (2007). Modeling time-location patterns of inner-city high
school students in New York and Los Angeles using a longitudinal approach with generalized estimating equations.
J Expo Sci Environ Epidemiol, 17: 233-247. 190996
Decesari S; Facchini MC; Fuzzi S; Tagliavini E. (2000). Characterization of water-soluble organic compounds in
atmospheric aerosol: Anew approach. J Geophys Res, 105: 1481-1489. 155748
Delfino RJ; Staimer N; Gillen D; Tjoa T; Sioutas C; Fung K; George SC; Kleinman MT. (2006). Personal and ambient air
pollution is associated with increased exhaled nitric oxide in children with asthma. Environ Health Perspect, 114:
1736-1743. 090745
Demokritou P; Lee S; Koutrakis P. (2004). Development and evaluation of a high loading PM2.5 speciation sampler.
Aerosol Sci Technol, 38: 111-119. 186901
Demokritou P; Lee SJ; Ferguson ST; Koutrakis P. (2004). A compact multistage (cascade) impactor for the characterization
of atmospheric' aerosols. J Aerosol Sci, 35: 281-299 . 190115
Dentener F; Stevenson D; Ellingsen K; Van Noije T; Schultz M; Amann M; Atherton C; Bell N; Bergmann D; Bey I;
Bouwman L; Butler T; Cofala J; Collins B; Drevet J; Doherty R; Eickhout B; Eskes H; Fiore A; Gauss M;
Hauglustaine D; Horowitz L; Isaksen ISA; Josse B; Lawrence M; Krol M; Lamarque JF; Montanaro V; Muller JF;
Peuch VH; Pitari G; Pyle J; Rast S; Rodriguez J; Sanderson M; Savage NH; Shindell D; Strahan S; Szopa S; Sudo
K; Van Dingenen R; Wild O; Zeng G. (2006). The global atmospheric environment for the next generation. Environ
Sci Technol, 40: 3586-3594. 088434
Derwent RG; Collins WJ; Johnson CE; Stevenson DS. (2001). Transient behaviour of tropospheric ozone precursors in a
global 3-D CTM and their indirect greenhouse effects. Clim Change, 49: 463-487. 047912
Diapouli E; Chaloulakou A; Spyrellis N. (2007). Levels of ultrafine particles in different microenvironments—implications
to children exposure. Sci Total Environ, 388: 128-136. 156397
Diapouli E; Grivas G; Chaloulakou A; Spyrellis N. (2008). PM10 and ultrafine particles counts in-vehicle and on-road in
the Athens area . Water Air Soil Pollut, 8: 89-97. 190119
Docherty KS; Wu W; Lim YB; Ziemann PJ. (2005). Contributions of organic peroxides to secondary aerosol formed from
reactions of monoterpenes with O3. Environ Sci Technol, 39: 4049-4059. 087613
Dockery DW; Pope CAIII; Xu X; Spengler JD; Ware JH; Fay ME; Ferris BG Jr; Speizer FE. (1993). An association
between air pollution and mortality in six US cities. N Engl J Med, 329: 1753-1759. 044457
Dodge MC. (1977). Combined use of modeling techniques and smog chamber data to derive ozone-precursor relationships.
Presented at International Conference on Photochemical Oxidant Pollution and Its Control, 09/12/1976-09/17/1976,
Research Triangle Park, NC. 038646
December 2009 3-198
-------
Dominici F; Peng RD; Bell ML; Pham L; McDermott A; Zeger SL; Samet JL. (2006). Fine particulate air pollution and
hospital admission for cardiovascular and respiratory diseases. JAMA, 295: 1127-1134. 088398
Dominici F; Zeger SL; Samet JM. (2000). A measurement error model for time-series studies of air pollution and mortality.
Biostatistics, 1: 157-175. 005828
Drewnick F; Jayne JT; Canagaratna M; Worsnop DR; DemerjianKL. (2004). Measurement of Ambient Aerosol
Composition During the PMTACS-NY 2001 Using an Aerosol Mass Spectrometer. Part II: Chemically Speciated
Mass Distributions. Aerosol Sci Technol, 38: 104-117. 155755
Drewnick F; Schwab J; Jayne J; Canagaratna M; Worsnop D; Demerjian K. (2004). Measurement of Ambient Aerosol
Composition During the PMTACS-NY 2001 Using an Aerosol Mass Spectrometer. Part I: Mass Concentrations.
Aerosol Sci Technol, 38: 92-103. 155754
Drewnick F; Schwab JJ; Hogrefe O; Peters S; Husain L; Diamond D; Weber R; Demerjian KL. (2003). Intercomparison
and evaluation of four semi-continuous PM2.5 sulfate instruments. Atmos Environ, 37: 3335-3350. 099160
Dutton SJ; Williams DE; Garcia JK; Vedal S; Hannigan MP. (2009). PM2.5 characterization for time series studies: organic
molecular marker speciation methods and observations from daily measurements in Denver. Atmos Environ, 43:
2018-2030. 194887
Ebelt ST; Wilson WE; Brauer M. (2005). Exposure to ambient and nonambient components of particulate matter: a
comparison of health effects. Epidemiology, 16: 396-405. 056907
Eder B; Kang D; Mathur R; Yu S; Schere K. (2006). An operational evaluation of the Eta-CMAQ air quality forecast
model. Atmos Environ, 40: 4894-4905. 142721
Eder B; Yu S. (2005). A performance evaluation of the 2004 release of Models-3 CMAQ. Atmos Environ, 40: 4811-4824.
089229
Edgerton ES; Casucclo GS; Saylor RD; Lersch TL; Hartsell BE; Jansen JJ; Hansen DA. (2009). Measurements of OC and
EC in coarse particulate matter in the southeastern United States. J Air Waste Manag Assoc, 59: 78-90. 180385
Edgerton ES; Hartsell BE; Saylor RD; Jansen JJ; Hansen DA; Hidy GM. (2005). The Southeastern Aerosol Research and
Characterization Study: Part II Filter-based measurements of fine and coarse particulate matter mass and
composition. J Air Waste Manag Assoc, 55: 1527-1542. 088686
Edney EO; Lewandowski M; Offenberg JH; Wang W; Claeys M; Kleindienst TE; Jaoui M. (2005). Formation of 2-methyl
tetrols and 2-methylglyeerie acid in secondary organic aerosol from laboratory irradiated isoprene/NOX/SO 2/air
mixtures and their detection in ambient PM2.5 samples collected in the eastern United States. Atmos Environ, 39:
5281-5289. 155760
Eiguren-Fernandez A; Miguel AH; Jaques PA; Sioutas C. (2003). Evaluation of a denuder-MOUDI-PUF sampling system
to measure the size distribution of semi-volatile polycyclic aromatic hydrocarbons in the atmosphere. Aerosol Sci
Technol, 37: 201-209. 142609
El-Zanan HS; Lowenthal DH; Zielinska B; Chow JC; Kumar N. (2005). Determination of the organic aerosol mass to
organic carbon ratio in IMPROVE samples. Chemosphere, 60: 485-496. 155764
Elgethun K; Fenske R; Yost M; Palcisko G (2003). Time-location analysis for exposure assessment studies of children
using a novel global positioning system instrument. Environ Health Perspect, 111: 115-122. 190640
Engel-Cox JA; Weber SA. (2007). Compilation and assessment of recent positive matrix factorization and UNMIX
receptor model studies on fine particulate matter source apportionment for the eastern United States. J Air Waste
Manag Assoc, 57: 1307-1316. 156419
Engewald W; Teske J; Efer J. (1999). Programmed temperature vaporisers-based large volume injection in capillary gas
chromatography. J Chromatogr, 842: 143-161. 155765
England GC; Watson JG; Chow JC; Zielinska B; Chang M. (2007). Dilution-Based Emissions Sampling from Stationary
Sources: Part 2- Gas-Fired Combustors Compared with Other Fuel-Fired Systems. J Air Waste Manag Assoc, 57:
79-93. 156421
England GC; Watson JG; Chow JC; Zielinska B; Oliver Chang MC; Loos KR; Hidv GM. (2007). Dilution-based emissions
sampling from stationary sources: Part 1-Compact sampler methodology and performance. J Air Waste Manag
Assoc, 57: 65-78. 156420
December 2009 3-199
-------
Enstrom JE. (2005). Fine particulate air pollution and total mortality among elderly Californians, 1973-2002. Inhal Toxicol,
17: 803-816. 087356
Erisman J-W; Vermetten AWM; Asman WAH; Waijers-IJpelaan A; Slanina J. (1988). Vertical distribution of gases and
aerosols: the behaviour of ammonia and related components in the lower atmosphere. Atmos Environ, 22: 1153-
1160.036510
Fairlie TD; Jacob DJ; Park RJ. (2007). The impact of transpacific transport of mineral dust in the United States. Atmos
Environ, 41: 1251-1266. 141923
Fan X; Brook JR; Mabury SA. (2003). Sampling atmospheric carbonaceous aerosols using an integrated organic gas and
particle sampler. Environ Sci Technol, 37: 3145-3151. 058628
Fan X; Lee PKH; Brook JR; Mabury SA. (2004). Improved Measurement of Seasonal and Diurnal Differences in the
Carbonaceous Components of Urban Particulate Matter Using a Denuder-Based Air Sampler. Aerosol Sci Technol,
38:63-69. 155770
Farmer PB; Singh R; Kaur B; Sram RJ; Binkova B; Kalina I; Popov TA; Garte S; Taioli E; Gabelova A; Cebulska-
Wasilewska A. (2003). Molecular epidemiology studies of carcinogenic environmental pollutants: effects of
poly cyclic aromatic hydrocarbons (PAHs) in environmental pollution on exogenous and oxidative DNA damage.
MutatRes, 544: 397-402. 089017
Feldpausch P; Fiebig M; Fritzsche L; Petzold A. (2006). Measurement of ultrafine aerosol size distributions by a
combination of diffusion screen separators and condensation particle counters. J Aerosol Sci, 37: 577-597. 155773
Filleul L; Zeghnoun A; Cassadou S; Declercq C; Eilstein D; Le Tertre A; Medina S; Pascal L; Prouvost H; Saviuc P;
QuenelP (2006). Influence of set-up conditions of exposure indicators on the estimate of short-term associations
between urban pollution and mortality. Sci Total Environ, 355: 90-97. 089862
Fine PM; Chakrabarti B; Krudysz M; Schauer JJ; Sioutas C. (2004). Diurnal Variations of Individual Organic Compound
Constituents of Ultrafine and Accumulation Mode Particulate Matter in the Los Angeles Basin. Environ Sci
Technol, 38: 1296-1304. 141283
Fine PM; Jaques PA; Hering SV; Sioutas C. (2003). Performance Evaluation and Use of a Continuous Monitor for
Measuring Size-Fractionated PM 2.5 Nitrate. Aerosol Sci Technol, 37: 342 - 354. 155775
Finlayson-Pitts BJ; Pitts JN Jr. (2000). Chemistry of the upper and lower atmosphere: theory, experiments and applications.
San Diego, CA: Academic Press. 055565
Fowler D; Cape JN; Unsworth MH. (1989). Deposition of atmospheric pollutants on forests. Proc Biol Sci, 324: 247-265.
002515
Fowler D; Duyzer JH; Baldocchi DD. (1991). Inputs of trace gases, particles and cloud droplets to terrestrial surfaces. Proc
Roy Soc Edinb B Biol, 97: 35-59. 046630
Frank NH. (2006). Retained nitrate, hydrated sulfates, and carbonaceous mass in federal reference method fine particulate
matter for six eastern U.S. cities. J Air Waste Manag Assoc, 56: 500-11. 098909
Fraser MP; Cass GR; Simoneit BRT. (1999). Particulate organic compounds emitted from motor vehicle exhaust and in the
urban atmosphere. Atmos Environ, 33: 2715-2724. 010819
Fraser MP; Yue Z W; Buzcu B. (2003). Source apportionment of fine particulate matter in Houston, TX, using organic
molecular markers. Atmos Environ, 37: 2117-2123. 042231
Fruin S; Westerdahl D; Sax T; Sioutas C; Fine PM. (2008). Measurements and predictors of on-road ultrafine particle
concentrations and associated pollutants in Los Angeles. Atmos Environ, 42: 207-219. 097183
Fuentes M; Raftery AE. (2005). Model evaluation and spatial interpolation by Bayesian combination of observations with
outputs from numerical models. Biometrics, 61: 36-45. 087580
Fuentes M; Song HR; Ghosh SK; Holland DM; Davis JM. (2006). Spatial association between speciated fine particles and
mortality. Biometrics, 62: 855-863. 097647
Gallagher MW; Choularton TW; Morse AP; Fowler D. (1988). Measurements of the size dependence of cloud droplet
deposition at a hill site. Q J Roy Meteorol Soc, 114: 1291 -1303. 046631
December 2009 3-200
-------
Gao S; Ng NL; Keywood M; Varutbangkul V; Bahrein! R; Nenes A; He J; Yoo KY; Beauchamp JL; Hodyss RP; Flagan
RC; Seinfeld JH. (2004). Particle phase acidity and oligomer formation in secondary organic aerosol. Environ Sci
Technol, 38: 6582-6589. 156460
Garner JHB; Pagano T; Cowling EB. (1989). An evaluation of the role of ozone, acid deposition, and other airborne
pollutants in the forests of eastern North America. Southeastern Forest Experiment Station. Asheville, NC. 042085
Garten CT Jr; Hanson PJ. (1990). Foliar retention of 15N-nitrate and 15N-ammonium by red maple (Acer rubrum) and
white oak (Quercus alba) leaves from simulated rain. Environ Exp Bot, 30: 333-342. 036803
Gaydos TM; Finder R; Koo B; Fahey KM; Yarwood G; Pandis SN. (2007). Development and application of a three-
dimensional aerosol chemical transport model, PMCAMx. Atmos Environ, 41: 2594-2611. 139738
Gaydos TM; Stanier CO; Pandis SN. (2005). Modeling of in situ ultrafine atmospheric particle formation in the eastern
United States. J Geophys Res, 110: D07S12. 191762
Gelencser A; Meszaros T; Blazso M; Kiss G; Krivacsy Z; Molnar A; Meszaros E. (2000). Structural Characterisation of
Organic Matter in Fine Tropospheric Aerosol by Pyrolysis-Gas Chromatography-Mass Spectrometry. J Atmos
Chem, 37: 173-183. 155785
Gelencser A; Varga Z. (2005). Evaluation of the atmospheric significance of multiphase reactions in atmospheric secondary
organic aerosol formation. Atmos Chem Phys, 5: 2823-2831. 156463
Geller ME); Ntziachristos L; Mamakos A; Samaras Z; Schmitz DA; Froines JR; Sioutas C. (2006). Physicochemical and
redox characteristics of particulate matter (PM) emitted from gasoline and diesel passenger cars. Atmos Environ,
40: 6988-7004. 139644
Generoso S; Bey I; Attie J-L; Breon F-M. (2007). A satellite- and model-based assessment of the 2003 Russian fires:
impact on the arctic region. J Geophys Res, 112: 5302. 155786
Georgopoulos PG; Wang S-W; Vyas VM; Sun Q; Burke J; Vedantham R; McCurdy T; Ozkaynak H. (2005). A source-to-
dose assessment of population exposures to fine PM and ozone in Philadelphia, PA, during a summer 1999 episode.
J Expo Sci Environ Epidemiol, 15: 439-457. 080269
Gery MW; Whitten GZ; Killus JP; Dodge MC. (1989). A photochemical kinetics mechanism for urban and regional scale
computer modeling. J Geophys Res, 94: 12,925-12,956. 043039
Giglio L; van der Werf GR; Randerson JT; Collatz GJ; Kasibhatla P. (2006). Global estimation of burned area using
MODIS active fire observations. Atmos Chem Phys, 6: 957-974. 156469
Gillette DA; Hanson KJ. (1989). Spatial and temporal variability of dust production caused by wind erosion in the United
States. J Geophys Res, 94: 2197-2206. 030212
Gilliam RC; Childs PP; Huber AH; Raman S. (2005). Metropolitan-scale transport and dispersion from the New York
World Trade Center following September 11, 2001 Part I: an evaluation of the CALMET meteorological model.
Pure Appl Geophys, 162: 1981-2003. 056749
Gilliam RC; Huber AH; Raman S. (2005). Metropolitan-scale transport and dispersion from the New York World Trade
Center following September 11, 2001 Part II: an application of the CALPUFF plume model. Pure Appl Geophys,
162: 2005-2028. 056750
Gilliland F; Avol E; Kinney P; Jerrett M; Dvonch T; Lurmann F; Buckley T; Breysse P; Keeler G; de Villiers T; McConnell
R. (2005). Air pollution exposure assessment for epidemiologic studies of pregnant women and children: lessons
learned from the Centers for Children's Environmental Health and Disease Prevention Research. Environ Health
Perspect, 113: 1447-54. 098820
Gomez-Perales JE; Colvile RN; Fernandez-Bremauntz AA; Gutierrez-Avedoy V; Paramo-Figueroa VH; Blanco-Jimenez S;
Bueno-Lopez E; Bernabe-Cabanillas R; Mandujano F; Hidalgo-Navarro M; Nieuwenhuijsen MJ. (2007). Bus,
minibus, metro inter-comparison of commuters' exposure to air pollution in Mexico City. Atmos Environ, 41: 890-
901. 138816
Gomez-Perales JE; Colvile RN; NieuwenhuijsenMJ; Fernandez-Bremauntz A; Gutierrez-Avedoy VJ; Paramo-Figueroa
VH; Blanco-Jimenez S; Bueno-Lopez E; Mandujano F; Bernabe-Cabanillas R. (2004). Commuters' exposure to
PM25, CO, and benzene in public transport in the metropolitan area of Mexico City. Atmos Environ, 38: 1219-
1229.054418
December 2009 3-201
-------
Gong L; Xu B; Zhu Y. (2009). Ultrafine particles deposition inside passenger vehicles . Aerosol Sci Technol, 43: 544-553.
190124
Goriaux M; Jourdain B; Temime B; Besombes JL; Marchand N; Albinet A; Leoz-Garziandia E; Wortham H. (2006). Field
comparison of particulate PAH measurements using a low-flow denuder device and conventional sampling systems.
Environ Sci Technol, 40: 6398-6404. 156484
Gouw Jd; Brock C; Atlas E; Bates T; Fehsenfeld F; Goldan P; Holloway J; Kuster W; Lerner B; Matthew B; Middlebrook
A; OnaschT; Peltier R; Quinn P; Senff C; Stohl A; Sullivan A; Trainer M; Warneke C; Weber R; Williams E.
(2008). Sources of particulate matter in the northeastern United States in summer: 1. Direct emissions and
secondary formation of organic matter in urban plumes. J Geophy s Res, 113: D08301. 191757
Graham B; Guyon P; Taylor PE; Artaxo P; Maenhaut W; Glovsky MM; Flagan RC; Andreae MO. (2003). Organic
compounds present in the natural Amazonian aerosol: Characterization by gas chromatography-mass spectrometry.
J Geophys Res, 108: 4766. 156489
Grell GA; Emeis S; Stockwell WR; Schoenemeyer T; Forkel R; Michalakes J; Knoche R; Seidl W. (2000). Application of a
multiscale, coupled MM5/chemistry model to the complex terrain of the VOTALP valley campaign. Atmos
Environ, 34: 1435-1453. 048047
Grover BD; Eatough NL; Eatough DJ; Chow JC; Watson JG; Ambs JL; Meyer MB; Hopke PK; Al-Horr R; Later DW;
Wilson WE. (2006). Measurement of Both Nonvolatile and Semi-Volatile Fractions of Fine Particulate Matter in
Fresno, CA. Aerosol Sci Technol, 40: 811-826. 138080
Grover BD; Kleinman M; Eatough NL; Eatough DJ; Gary RA; Hopke PK; Wilson WE. (2008). Measurement of fine
particulate matter nonvolatile and semi-volatile organic material with the Sunset Laboratory Carbon Aerosol
Monitor. J Air Waste Manag Assoc, 58: 72-77. 156502
Grover BD; Kleinman M; Eatough NL; Eatough DJ; Hopke PK; Long RW; Wilson WE; Meyer MB; Ambs JL. (2005).
Measurement of total PM25 mass (nonvolatile plus semivolatile) with the Filter Dynamic Measurement System
tapered element oscillating microbalance monitor. J Geophys Res, 110: D07S03. 090044
Gulliver J; Briggs DJ. (2004). Personal exposure to particulate air pollution in transport microenvironments. Atmos
Environ, 38: 1-8. 053238
Gulliver J; Briggs DJ. (2005). Time-space modeling of journey-time exposure to traffic-related air pollution using GIS.
Environ Res, 97: 10-25. 191079
Gulliver J; Briggs DJ. (2007). Journey-time exposure to particulate air pollution. Atmos Environ, 41: 7195-7207. 155814
Gustafsson O; Krusa M; Zencak Z; Sheesley RJ; Granat L; Engstrom E; Praveen PS; Rao PS; Leek C; Rodhe H. (2009).
Brown clouds over South Asia: biomass or fossil fuel combustion?. Science, 323: 495-498. 192000
Gutierrez-Daban A; Fernandez-Espinosa AJ; Ternero-Rodriguez M; Fernandez-Alvarez F. (2005). Particle-size distribution
of poly cyclic aromatic hydrocarbons in urban air in southern Spain. Anal Bioanal Chem, 381: 721-736. 155818
Hagler GSW; Baldauf RW; Thoma ED; Long TR; Snow RF; Kinsey JS; Oudejans L; Gullett BK. (2009). Ultrafine
particles near a major roadway in Raleigh, North Carolina: Downwind attenuation and correlation with traffic-
related pollutants. Atmos Environ, 43: 1229-1234. 191185
Hains JC; Chen L-WA; Taubman BF; Doddridge BG; Dickerson RR. (2007). A side-by-side comparison of filter-based
PM25 measurements at a suburban site: a closure study. Atmos Environ, 41: 6167-6184. 091039
Hammond DM; Lalor MM; Jones SL. (2007). In-vehicle measurement of particle number concentrations on school buses
equipped with diesel retrofits . Water Air Soil Pollut, 179: 217-225. 190135
Han Y; Cao J; Chow JC; Watson JG; An Z; Jin Z; Fung K; Liu S. (2007). Evaluation of the thermal/optical reflectance
method for discrimination between char-and soot-EC. Chemosphere, 69: 569-574. 155823
Hand JL; Kreidenweis SM. (2002). A New Method for Retrieving Particle Refractive Index and Effective Density from
Aerosol Size Distribution Data. Aerosol Sci Technol, 36: 1012-1026. 155824
Hansen K; Draaijers GPJ; Ivens WPMF; Gundersen P; van Leeuwen NFM. (1994). Concentration variations in rain and
canopy throughfall collected sequentially during individual rain events. Atmos Environ, 28: 3195-3205. 046634
Harrison D; Shik Park S; Ondov J; Buckley T; Roul Kim S; Jayanty RKM. (2004). Highly time resolved fine particle
nitrate measurements at the Baltimore Supersite. Atmos Environ, 38: 5321-5332. 136787
December 2009 3-202
-------
Harrison RM; Jones AM. (2005). Multisite study of particle number concentrations in urban air. Environ Sci Technol, 39:
6063-6070. 191005
Hasegawa S;Hirabayashi M;Kobayashi S;Moriguchi Y;Kondo Y;Tanabe K;Wakamatsu. (2005). Size Distribution and
Characterization of Ultrafine Particles in Roadside Atmosphere. J Environ Sci Health A Tox Hazard Subst Environ
Eng, 39: 2671-2690. 157355
Hays ME); Geron CD; Linna KJ; Smith ND; Schauer JJ. (2002). Speciation of gas-phase and fine particle emissions from
burning of foliar fuels. Environ Sci Technol, 36: 2281-2295. 026104
Hays MD; Lavrich RJ. (2007). Developments in direct thermal extraction gas chromatography-mass spectrometry of fine
aerosols. Trends Analyt Chem, 26: 88-102. 155831
Heald C; SpracklenD. (2009). Atmospheric budget of primary biological aerosol particles from fungal sources. Geophys
Res Lett, 36: L09806. 190014
HEI. (2009). Traffic-related air pollution: A critical review of the literature on emissions, exposure, and health effects.
Health Effects Institute. Boston. Special Report 17. 191009
Henderson R. (2005). Clean Air Scientific Advisory Committee (CASAC) review of the EPA staff recommendations
concerning a Potential Thoracic Coarse PM standard in the Review of the National Ambient Air Quality Standards
for Particulate Matter: Policy Assessment of Scientific and Technical Information. U.S. Environmental Protection
Agency. Washington, DC. 156537
Henderson R. (2009). Clean Air Scientific Advisory Committee (CASAC) letter the the Administrator: Consultation on
monitoring issues related to the NAAQS for particulate matter. U.S. Environmental Protection Agency .
Washington, DC. EPA-CASAC-09-006. 192001
Henry RC. (1997). History and fundamentals of multivariate air quality receptor models. Chemometr Intell Lab Syst, 37:
37-42. 020941
Hering S; Fine PM; Sioutas C; Jaques PA; Ambs JL; Hogrefe O; Demerjian KL. (2004). Field assessment of the dynamics
of particulate nitrate vaporization using differential TEOM® and automated nitrate monitors. Atmos Environ, 38:
5183-5192. 155837
Hering SV. (2007). Using Regional Data and Building Leakage to Assess Indoor Concentrations of Particles of Outdoor
Origin. Aerosol Sci Technol, 41: 639-654. 155839
Hering SV; Stolzenburg MR; Quant FR; Oberreit DR; Keady PB. (2005). A Laminar-Flow, Water-Based Condensation
Particle Counter (WCPC). Aerosol Sci Technol, 39: 659-672. 155838
Hermann M; Wehner B; Bischof O; Han HS; Krinke T; Liu W; Zerrath A; Wiedensohler A. (2007). Particle counting
efficiencies of newTSI condensation particle counters. JAerosol Sci, 38: 674-682. 155840
Herner JD; Aw J; Gao O; Chang DP; Kleeman MJ. (2005). Size and composition distribution of airborne particulate matter
in Northern California: I - Particulate mass, carbon, and water-soluble ions. J Air Waste Manag Assoc, 55: 30-51.
135983
Ho KF; Cao JJ; Harrison RM; Lee SC. (2004). Indoor/outdoor relationships of organic carbon (OC) and elemental carbon
(EC) in PM2.5 in roadside environment of Hong Kong. Atmos Environ, 38: 6327-6335. 056804
Hoek G; Beelen R; de Hoogh K; Vienneau D; Gulliver J; Fischer P; Briggs D. (2008). A review of land-use regression
models to assess spatial variation of outdoor air pollution. Atmos Environ, 42: 7561-7578. 195851
Hogrefe C; Hao W; Civerolo K; Ku JY; Sistla G; Gaza RS; Sedefian L; Schere K; Gilliland A; Mathur R. (2007). Daily
Simulation of Ozone and Fine Particulates over New York State: Findings and Challenges. J Appl Meteor Climatol,
46: 961-979. 156561
Hogrefe O; Schwab JJ; Drewnick F; Lala GG; Peters S; Demerjian KL; Rhoads K; Felton HD; Rattigan OV; Husain L.
(2004). Semicontinuous PM 2.5 sulfate and nitrate measurements at an urban and a rural location in New York:
PMTACS-NY Summer 2001 and 2002 Campaigns. J Air Waste Manag Assoc, 54: 1040-1060. 156560
Hopke PK; Ramadan Z; Paatero P; Norris GA; Landis MS; Williams RW; Lewis CW. (2003). Receptor modeling of
ambient and personal exposure samples: 1998 Baltimore Particulate Matter Epidemiology-Exposure Study. Atmos
Environ, 37: 3289-3302. 095544
December 2009 3-203
-------
Hosker RP Jr; Lindberg SE. (1982). Review: atmospheric deposition and plant assimilation of gases and particles. Atmos
Environ, 16: 889-910. 019118
Hsiao T-C; Chen D-R; Son SY. (2009). Development of mini-cyclones as the size-selective inlet of miniature particle
detectors. J Aerosol Sci, 40: 481-491. 191001
Hu S; Herner JD; Shafer M; Robertson W; Schauer JJ; Dwyer H; Collins J; Huai T; Ayala A. (2009). Metals emitted from
heavy-duty diesel vehicles equipped with advanced PM and NO"X emission controls. Atmos Environ, 43: 2950-
2959. 189886
Huang C; Chang C; Chang S; Tsai C; Shih T; Tang D. (2005). Use of porous foam as the substrate of an impactor for
respirable aerosol sampling. J Aerosol Sci, 36: 1373-1386. 186991
Huang L; Brook JR; Zhang W; Li SM; Graham L; Ernst D; Chivulescu A; Lu G. (2006). Stable isotope measurements of
carbon fractions (OC/EC) in airborne particulate: Anew dimension for source characterization and apportionment.
Atmos Environ, 40: 2690-2705. 097654
Huang YL; Batterman S. (2000). Selection and evaluation of air pollution exposure indicators based on geographic areas.
Sci Total Environ, 253: 127-144. 156572
Huebert BJ; Charlson RJ. (2000). Uncertainties in data on organic aerosols. Tellus B Chem Phys Meteorol, 52: 1249-1255.
156577
Huebert BJ; Luke WT; Delany AC; Brost RA. (1988). Measurements of concentrations and dry surface fluxes of
atmospheric nitrates in the presence of ammonia. J Geophys Res, 93: 7127-7136. 036569
Husar RB; Tratt DM; Schichtel BA; Falke SR; Li F; Jaffe D; Gasso S; Gill T; Laulainen NS; Lu F; Reheis MC; Chun Y;
Westphal D; Holben BN; Gueymard C; McKendry I; Kuring N; Feldman GC; McClain C; Frouin RJ; Merrill J;
DuBois D; Vignola F; Murayama T; Nickovic S; Wilson WE; Sassen K; Sugimoto N; Malm WC. (2001). Asian
dust events of April 1998. J Geophys Res, 106: 18,317-18,330. 024947
Hwang IJ; Hopke PK; Pinto JP (2008). Source apportionment and spatial distributions of coarse particles during the
regional air pollution study. Environ Sci Technol, 42: 3524-3530. 194533
Hwang K-W; Lee J-H; Jeong D-Y; Lee C-H; Bhatnagar A; Park J-M; Kim S-H. (2008). Observation of difference in the
size distribution of carbon and major inorganic compounds of atmospheric aerosols after the long-range transport
between the selected days of winter and summer. Atmos Environ, 42: 1057-1063. 134420
ICRP (1994). Human respiratory tract model for radiological protection: a report of a task group of the International
Commission on Radiological Protection. Ann ICRP, 24: 1-482. 006988
Isakov V; Irwin JS; Ching J. (2007). Using CMAQ for exposure modeling and characterizing the subgrid variability for
exposure estimates. J Appl Meteor Climatol, 46: 1354-1371. 195880
Isakov V; Touma JS; Burke J; Lobdell DT; Palma T; Rosenbaum A; Ozkaynak H. (2009). Combining regional- and local-
scale air quality models with exposure models for use in environmental health studies. J Air Waste Manag Assoc,
59: 461-472. 191192
Isakov V; Touma JS; Khlystov A. (2007). A method of assessing air toxics concentrations in urban areas using mobile
platform measurements. J Air Waste Manag Assoc, 57: 1286-1295. 156588
Jacob DJ. (1999). Introduction to atmospheric chemistry. New Jersey: Princeton University Press. 091122
Jacobson MZ. (2002). Atmospheric pollution: history, science, and regulation. New York: Cambridge University Press.
090667
Jacobson MZ; Kittelson DB; Watts WF. (2005). Enhanced coagulation due to evaporation and its effect on nanoparticle
evolution. Environ Sci Technol, 39: 9486-9492. 191187
Jaffe D; Price H; Parrish D; Goldstein A; Harris J. (2003). Increasing background ozone during spring on the west coast of
North America. Geophys Res Lett, 30: 1613. 052229
Jansen KL; Larson TV; Koenig JQ; Mar TF; Fields C; Stewart J; Lippmann M. (2005). Associations between health effects
and particulate matter and black carbon in subjects with respiratory disease. Environ Health Perspect, 113: 1741-
1746. 082236
Jaques PA; Ambs JL; Grant WL; Sioutas C. (2004). Field evaluation of the differential TEOM monitor for continuous PM
2.5 mass concentrations. Aerosol Sci Technol, 38: 49-59. 155878
December 2009 3-204
-------
Jeong CH; Hopke PK; Chalupa D; Utell M. (2004). Characteristics of nucleation and growth events of ultrafine
particlesmeasured in Rochester, NY. Environ Sci Technol, 38: 1933-1940. 180350
Jerrett M; Arain A; Kanaroglou P; Beckerman B; Potoglou D; Sahsuvaroglu T; Morrison J; Giovis C. (2005). A review and
evaluation of intraurban air pollution exposure models. J Expo Sci Environ Epidemiol, 15:185-204. 092864
Jerrett M; Burnett RT; Ma R; Pope 3rd CA; Krewski D; Newbold KB; Thurston G; Shi Y; Finkelstein N; Calle EE; Thun
MJ. (2005). Spatial analysis of air pollution and mortality in Los Angeles. Epidemiology, 16: 727-36. 189405
Jimenez JL; Jayne JT; Shi Q; Kolb CE; Worsnop DR; Yourshaw I; Seinfeld JH; Flagan RC; Zhang X; Smith KA. (2003).
Ambient aerosol sampling using the Aerodyne Aerosol Mass Spectrometer. J Geophys Res, 108: 8425. 156611
Johnson T; Long T; Ollison W. (2000). Prediction of hourly microenvironmental concentrations of fine particles based on
measurements obtained from the Baltimore scripted activity study. JExpo Sci Environ Epidemiol, 10: 403-411.
001660
Kahn R; Gaitley B; Martonchik J; Diner D; Crean K; Holben B. (2005). Multiangle Imaging Spectroradiometer (MISR)
global aerosol optical depth validation based on 2 years of coincident Aerosol Robotic Network (AERONET)
observations. J Geophys Res, 110: D10S04. 189961
Kalberer M; Paulsen D; Sax M; Steinbacher M; Dommen J; Prevot ASH; Fisseha R; Weingartner E; Frankevich V; Zenobi
R. (2004). Identification of Polymers as Major Components of Atmospheric Organic Aerosols. Science, 303: 1659-
1662. 156619
Kastner-Klein P; Plate EJ. (1999). Wind-tunnel study of concentration fields in street canyons. Atmos Environ, 33: 3973-
3979. 001961
Kaur S; Nieuwenhuijsen M; Colvile R. (2005). Personal exposure of street canyon intersection users to PM2.5, ultrafine
particle counts and carbon monoxide in central London, UK. Atmos Environ, 39: 3629-3641. 086504
Kaur S; Nieuwenhuijsen MJ; Colvile RN. (2005). Pedestrian exposure to air pollution along a major road in Central
London, UK. Atmos Environ, 39: 7307-7320. 088175
Keller A; SiegmannHC. (2001). The role of condensation and coagulation in aerosol monitoring. JExpo Sci Environ
Epidemiol, 11: 441-448. 025881
Kelly JM. (1988). Annual elemental input/output estimates for two forested watersheds in eastern Tennessee. J Environ
Qual, 17: 463-468. 037379
Kenny LC; Merrifield T; Mark D; Gussman R; Thorpe A. (2004). The Ddevelopment and Designation Testing of a New
USEPA-Approved Fine Particle Inlet: A Study of the USEPA Designation Process. Aerosol Sci Technol, 38: 15-22.
155895
Ketzel M; Wahlin P; Berkowicz R; Palmgren F. (2003). Particle and trace gas emission factors under urban driving
conditions in Copenhagen based on street and roof-level observations. Atmos Environ, 37: 2735-2749. 131251
Khlystov A; Stanier C; Pandis SN. (2004). An Algorithm for Combining Electrical Mobility and Aerodynamic Size
Distributions Data when Measuring Ambient Aerosol. Aerosol Sci Technol, 38: 229-238. 155897
Khlystov A; Stanier CO; Takahama S; Pandis SN. (2005). Water content of ambient aerosol during the Pittsburgh Air
Quality Study. J Geophys Res, 110: D07S10. 156635
Kidwell CB; Ondov JM. (2004). Elemental Analysis of Sub-Hourly Ambient Aerosol Collections. Aerosol Sci Technol, 38:
205-218. 155898
Kim E; Hopke PK; Pinto JP; Wilson WE. (2005). Spatial variability of fine particle mass, components, and source
contributions during the regional air pollution study in St Louis. Environ Sci Technol, 39: 4172-4179. 083181
Kim S Y; Sheppard L; Kim H. (2009). Health effects of long-term air pollution: influence of exposure prediction methods.
Epidemiology, 20: 442-50. 188446
Kinney PL; Aggarwal M; Northridge ME; Janssen NAH; Shepard P. (2000). Airborne concentrations of PM25 and diesel
exhaust particles on Harlem sidewalks: a community-based pilot study. Environ Health Perspect, 108: 213-218.
001774
Kinsey JS; Williams DC; Dong Y; Logan R. (2007). Characterization of fine particle and gaseous emissions during school
bus idling. Environ Sci Technol, 41: 4972-4979. 190073
December 2009 3-205
-------
Kiss G; Varga B; Galambos I; Ganszky I. (2002). Characterization of water-soluble organic matter isolated from
atmospheric fine aerosol. J Geophys Res, 107: 8339. 156646
KittelsonDB. (1998). Engines and nanoparticles: a review. J Aerosol Sci, 29: 575-588. 051098
Kittelson DB; Schauer JJ; Lawson DR; Watts WF; Johnson JR (2006). On-road and laboratory evaluation of combustion
aerosols-Part 2:. Summary of spark ignition engine results. J Aerosol Sci, 37: 931-949. 156648
Kittelson DB; Watts WF; Johnson JR (2006). On-road and laboratory evaluation of combustion aerosols—Part 1: Summary
of diesel engine results. J Aerosol Sci, 37: 913-930. 156649
Kleindienst TE; Edney EO; Lewandowski M; Offenberg JH; Jaoui M. (2006). Secondary Organic Carbon and Aerosol
Yields from the Irradiations of Isoprene and-Pinene in the Presence of NO x and SO 2. Environ Sci Technol, 40:
3807-3812. 156650
Klepeis NE; Nelson WC; OttWR; Robinson JPTsang AM; Switzer P; Behar JV; Hern SC; Engelmann WH. (2001). The
National Human Activity Pattern Survey (NHAPS): a resource for assessing exposure to environmental pollutants. J
Expo Sci Environ Epidemiol, 11: 231-252. 002437
Klouda GA; Filliben JJ; Parish HJ; Chow JC; Watson JG; Gary RA. (2005). Reference material 8785: Air particulate matter
on filter media. Aerosol Sci Technol, 39: 173-183. 130382
Koenig JQ; Mar TF; Allen RW; Jansen K; Lumley T; Sullivan JH; Trenga CA; Larson T; Liu LJ. (2005). Pulmonary effects
of indoor- and outdoor-generated particles in children with asthma. Environ Health Perspect, 113: 499-503. 087384
Kokhanovsky A; Breon F; Cacciari A; Carboni E; Diner D; Di Nicolantonio W; Grainger R; Grey W; Holler R; Lee K.
(2007). Aerosol remote sensing over land: a comparison of satellite retrievals using different algorithms and
instruments. Atmos Res, 85: 372-394. 190009
Koutrakis P; Suh HH; Sarnat JA; Brown KW; Coull BA; Schwartz J. (2005). Characterization of particulate and gas
exposures of sensitive subpopulations living in Baltimore and Boston. Health Effects Institute. Boston, MA. 131.
http://pubs.healtheffects.org/viewphp?id=91. 095800
Krewski D; Burnett RT; Goldberg MS; Hoover K; Siemiatycki J; Jerrett M; Abrahamowicz M; White WH. (2000).
Reanalysis of the Harvard Six Cities study and the American Cancer Society study of particulate air pollution and
mortality. Health Effects Institute. Cambridge, MA.http://pubs.healtheffects.org/view.php?id=6. 012281
Krewski D; Jerrett M; Burnett RT; Ma R; Hughes E; Shi Y; Turner MC; Pope AC III; Thurston G; Calle EE; Thun MJ.
(2009). Extended follow-up and spatial analysis of the American Cancer Society study linking particulate air
pollution and mortality. Health Effects Institute. Cambridge, MA. Report Nr. 140. 191193
Krieger UK; Rupp S; Hausammann E; Peter T. (2007). Simultaneous measurements of PM10 and PM1 using a single
TEOM. Aerosol Sci Technol, 41: 975-980. 129657
Kroll JH; Seinfeld JH. (2008). Chemistry of secondary organic aerosol: Formation and evolution of low-volatility organics
in the atmosphere. Atmos Environ, 42: 3593-3624. 155910
Krudysz MA; Froines JR; Fine PM; Sioutas C. (2008). Intra-community spatial variation of size-fractionated PM mass,
OC, EC, and trace elements in the Long Beach, CA area. Atmos Environ, 42: 5374-5389. 190064
Kruize H; Hanninen O; Breugelmans O; Lebret E; Jantunen M. (2003). Description and demonstration of the EXPOLIS
simulation model: two examples of modeling population exposure to particulate matter. J Expo Sci Environ
Epidemiol, 13: 87-99. 156661
Kuang C; McMurry PH; McCormick AV; Eisele FL. (2008). Dependence of nucleation rates on sulfuric acid vapor
concentration in diverse atmospheric locations. J Geophys Res, 113: D10209. 191196
KuhnT; Biswas S; Sioutas C. (2005). Diurnal and seasonal characteristics of particle volatility and chemical composition
in the vicinity of a light-duty vehicle freeway. Atmos Environ, 39: 7154-7166. 129448
Kuhns H; Knipping EM. (2005). Development of a United States-Mexico emissions inventory for the Big Bend Regional
Aerosol and Visibility Observational (BRAVO) Study. J Air Waste Manag Assoc, 55: 677-692. 156663
Kulmala M; Laaksonen A; Pirjola L. (1998). Parameterizations for sulfuric acid/water nucleation rates. J Geophys Res,
103: 8301-8307. 129411
December 2009 3-206
-------
Kulmala M; Mordas G; Petaja T; Gronholm T; Aalto PP; Vehkamaki H; Hienola AI; Herrmann E; Sipila M; Riipinen I.
(2007). The condensation particle counter battery (CPCB): A new tool to investigate the activation properties of
nanoparticles. J Aerosol Sci, 38: 289-304. 155911
Kulmala M; Riipinen I; Sipila M; Manninen HE; Petaja T; Junninen H; Maso MD; Mordas G; Mirme A; Vana M; Hirsikko
A; Laakso L; Harrison RM; Hanson I; Leung C; Lehtinen KE; Kerminen VM. (2007). Toward direct measurement
of atmospheric nucleation. Science, 318: 89-92. 097838
Kulmala M; Vehkamaki H; Petajda T; Dal Maso M; Lauri A; Kerminen V-M; Birmili W; McMurry PH. (2004). Formation
and growth rates of ultrafine atmospheric particles: a review of observations. J Aerosol Sci, 35: 143-176. 089159
Kuo Y; Huang S; Shih T; Chen C; Weng Y; Lin W. (2005). Development of a size-selective inlet-simulating ICRP lung
deposition fraction. Aerosol Sci Technol, 39: 437-443. 186997
Kurniawan A; Schmidt-Ott A. (2006). Monitoring the soot emissions of passing cars. Environ Sci Technol, 40: 1911-5.
098823
Lake DA; Tolocka MP; Johnston MV; Wexler AS. (2003). Mass Spectrometry of Individual Particles between 50 and 750
nm in Diameter at the Baltimore Supersite. Environ Sci Technol, 37: 3268-3274. 156669
Lake DA; Tolocka MP; Johnston MV; Wexler AS. (2004). The character of single particle sulfate in Baltimore. Atmos
Environ, 38: 5311-5320. 088411
Langholz B; Ebi K; Thomas D; Peters J; London S. (2002). Traffic density and the risk of childhood leukemia in a Los
Angeles case-control study. Ann Epidemiol, 12: 482-487. 191771
Larson T; Gould T; Simpson C; Liu LJ; Claiborn C; Lewtas J. (2004). Source apportionment of indoor, outdoor, and
personal PM2.5 in Seattle, Washington, using positive matrix factorization. J Air Waste Manag Assoc, 54: 1175-87.
098145
Lee HM. (2007). Fabrication of Reference Filter for Measurements of EC (Elemental Carbon) and OC (Organic Carbon) in
Aerosol Particles. Aerosol Sci Technol, 41: 284-294. 155926
Lee JH; Hopke PK; Holsen TM; Lee DW; AJaques P; Sioutas C; Ambs JL. (2005). Performance evaluation of continuous
PM2.5 mass concentration monitors. J Aerosol Sci, 36: 95-109. 155925
Lee JH; Hopke PK; Holsen TM; Polissar AV; Lee DW; Edgerton ES; Ondov JM; Allen G. (2005). Measurements of fine
particle mass concentrations using continuous and integrated monitors in Eastern US Cities. Aerosol Sci Technol,
39: 261-275. 128139
Lee SJ; Demokritou P; Koutrakis P; Delgado-Saborit JM. (2006). Development and evaluation of personal respirable
particulate sampler (PRPS). Atmos Environ, 40: 212-224. 098249
Leith D; Sommerlatt D; Boundy MG (2007). Passive sampler for PM10-2.5 aerosol. J Air Waste Manag Assoc, 57: 332-6.
098241
Levine PM; Gong P; Levicky R; Shepard KL. (2009). Real-time, multiplexed electrochemical DNA detection using an
active complementary metal-oxide-semiconductor biosensor array with integrated sensor electronics. Biosens
Bioelectron, 24: 1995-2001. 192091
Levy JI; Bennett DH; Melly SJ; Spengler JD. (2003). Influence of traffic patterns on particulate matter and poly cyclic
aromatic hydrocarbon concentrations inRoxbury, Massachusetts. JExpo Sci Environ Epidemiol, 13: 364-371.
052661
Lewne M; Nise G; Lind ML; Gustavsson P. (2006). Exposure to particles and nitrogen dioxide among taxi, bus and lorry
drivers. IntArchOccup Environ Health, 79: 220-226. 090556
Lim H-J; Turpin BJ; Edgerton E; Hering SV; Allen G; Maring H; Solomon P. (2003). Semi-continuous aerosol carbon
measurements: Comparison of Atlanta supersite measurements. J Geophys Res, 108: 8419. 037037
Lim MCH; Ayoko GA; Morawska L; Ristovski ZD; Jayaratne ER. (2007). The effects of fuel characteristics and engine
operating conditions on the elemental composition of emissions from heavy duty diesel buses. Fuel, 86: 1831-1839.
155931
Lindberg SE; Lovett GM. (1985). Field measurements of particle dry deposition rates to foliage and inert surfaces in a
forest canopy. Environ Sci Technol, 19: 238-244. 036530
December 2009 3-207
-------
Lipfert FW; Perry HM Jr; Miller JP; Baty JD; Wyzga RE; Carmody SE. (2000). The Washington University-EPPJ veterans'
cohort mortality study: preliminary results. Inhal Toxicol, 4: 41-73. 004087
Lippmann M. (2009). Semi-continuous speciation analyses for ambient air particulate matter: an urgent need for health
effects studies. J Expo Sci Environ Epidemiol, 19: 235-247. 190083
Lipsky E; Robinson A. (2006). Effects of dilution on fine particle mass and partitioning of semivolatile organics in diesel
exhaust and wood smoke. Environ Sci Technol, 40: 155-162. 189891
Little P; Wiffen RD. (1977). Emission and deposition of petrol engine exhaust Pb—I deposition of exhaust Pb to plant and
soil surfaces. Atmos Environ, 11: 437-447. 070869
Liu L-J; Box M; Kalman D; Kaufman J; Koenig J; Larson T; Lumley T; Sheppard L; Wallace L. (2003). Exposure
assessment of particulate matter for susceptible populations in Seattle. Environ Health Perspect, 11: 909-918.
073841
Liu XH; Hegg DA; Stoelinga MT. (2001). Numerical simulation of new particle formation over the northwest Atlantic
using the MM5 mesoscale model coupled with sulfur chemistry. J Geophys Res, 106: 9697-9715. 048201
Lokken PR; Wellenius GA; Coull BA; Burger MR; Schlaug G; Suh HH; Mittleman MA. (2009). Air Pollution and Risk of
Stroke: Underestimation of Effect Due to Misclassification of Time of Event Onset. Epidemiology, 20: 137-142.
186774
Long CM; Suh HH; Catalano PJ; Koutrakis P. (2001). Using time- and size-resolved particulate data to quantify indoor
penetration and deposition behavior. Environ Sci Technol, 35: 2089-2099. 011526
Long RW; McClenny WA. (2006). Laboratory and field evaluation of instrumentation for the semicontinuous
determination of particulate nitrate (and other water-soluble particulate components). J Air Waste Manag Assoc, 56:
294-305. 098214
Lovett GM. (1994). Atmospheric deposition of nutrients and pollutants in North America: an ecological perspective. Ecol
Appl, 4: 629-650. 024049
Lowenthal DH; Rogers CF; Saxena P; Watson JG; Chow JC. (1995). Sensitivity of estimated light extinction coefficients to
model assumptions and measurement errors. Atmos Environ, 29: 751-766. 045134
Lu R; Turco RP; Jacobson MZ. (1997). An integrated air pollution modeling system for urban and regional scales: 1
Structure and performance. J Geophys Res, 102: 6063-6079. 048202
Lu R; Turco RP; Jacobson MZ. (1997). An integrated air pollution modeling system for urban and regional scales: 2.
Simulations for SCAQS 1987 . J Geophys Res, 102: 6081-6098. 191768
Lunden MM; Kirchstetter TW; Thatcher TL; Hering SV; Brown NJ. (2008). Factors affecting the indoor concentrations of
carbonaceous aerosols of outdoor origin. Atmos Environ, 42: 5660-5671. 155949
Lunden MM; Revzan KL; Fischer ML; Thatcher TL; Littlejohn D; Hering SV; Brown NJ. (2003). The transformation of
outdoor ammonium nitrate aerosols in the indoor environment. Atmos Environ, 37: 5633-5644. 081201
Lunden MM; Thatcher TL; Hering SV; Brown NJ. (2003). The use of time-and chemically resolved particulate data to
characterize the infiltration of outdoor PM2.5 into a residence in the San Joaquin Valley. Environ Sci Pol, 37: 4724-
4732. 156718
Mader BT; Schauer JJ; Seinfeld JH; Flagan RC; Yu JZ; Yang H; Lim HJ; Turpin BJ; Deminter JT; Heidemann G. (2003).
Sampling methods used for the collection of particle-phase organic and elemental carbon during ACE-Asia. Atmos
Environ, 37: 1435-1449. 155955
Maheswaran R; Elliott P. (2003). Stroke mortality associated with living near main roads in England and wales: a
geographical study. Stroke, 34: 2776-2780. 125271
Mahrt L. (1998). Stratified atmospheric boundary layers and breakdown of models. Theor Comput Fluid Dynam, 11: 263-
279. 048210
Makar PA; Gravel S; Chirkov V; Strawbridge KB; Froude F; Arnold J; Brook J. (2006). Heat flux, urban properties, and
regional weather. Atmos Environ, 40: 2750-2766. 155959
Mar TF; Norris GA; Larson TV; Wilson WE; Koenig JQ. (2003). Air pollution and cardiovascular mortality in Phoenix,
1995-1997. 156731
December 2009 3-208
-------
Maricq MM. (2007). Chemical characterization of participate emissions from diesel engines: A review. J Aerosol Sci, 3 8:
1079-1118. 155973
Martin RV; Chance K; Jacob DJ; Kurosu TP; Spurr RJD; Bucsela E; Gleason IF; Palmer PI; Bey I; Fiore AM; Li Q;
Yantosca RM; Koelemeijer RBA. (2002). An improved retrieval of tropospheric nitrogen dioxide from GOME. J
Geophys Res, 107: 1-21. 089380
Massman WJ; Pederson J; Delany A; Grantz D; Denhartog G; Neumann HH; Oncley SP; Pearson R; Shaw RH. (1994). An
evaluation of the regional acid deposition model surface module for ozone uptake at 3 sites in the San-Joaquin
valley of California. J Geophys Res, 99: 8281-8294. 043681
Mathis U; Mohr M; Forss AM. (2005). Comprehensive particle characterization of modern gasoline and diesel passenger
cars at low ambient temperatures. Atmos Environ, 39: 107-117. 155970
Mathur R. (2008). Estimating the impact of the 2004 Alaskan forest fires on episodic particulate matter pollution over the
eastern United States through assimilation of satellite-derived aerosol optical depths in a regional air quality model.
J Geophys Res, 113:D17302. 156742
Matsumoto K; Hayano T; Uematsu M. (2003). Positive artifact in the measurement of particulate carbonaceous substances
using an ambient carbon particulate monitor. Atmos Environ, 37: 4713-4717. 124293
Matthias-Maser S; Bogs B; Jaenicke R. (2000). The size distribution of primary biological aerosol particles in cloud water
on the mountain Kleiner Feldberg/Taunus (FRG). Atmos Res, 54: 1-13. 155972
McBride SJ; Williams RW; Creason J. (2007). Bayesian hierarchical modeling of personal exposure to particulate matter.
Atmos Environ, 41: 6143-6155. 124058
McCurdy T; Glen G; Smith L; Lakkadi Y. (2000). The National Exposure Research Laboratory's consolidated human
activity database. JExpo Sci Environ Epidemiol, 10: 566-578. 000782
McDonald JD; Eide I; Seagrave J; Zielinska B; Whitney K; Lawson DR; Mauderly JL. (2004). Relationship between
composition and toxicity of motor vehicle emission samples. Environ Health Perspect, 112: 1527-1538. 087458
McKendry IG; Strawbridge KB; O?Neill NT; Macdonald AM; Liu PSK; Leaitch WR; Anlauf KG; Jaegle L; Fairlie TD;
Westphal DL. (2007). Trans-Pacific transport of Saharan dust to western North America: A case study. J Geophys
Res, 112:001103. 156748
McKenna DS; Konopka P; Grooss J-U; Gunther G; Muller R; Spang R; Offermann D; Orsolini Y. (2002). A new chemical
Lagrangian model of the stratosphere (CLaMS) 1. formulation of advection and mixing. J Geophys Res, 107: 4309.
053445
McMurry PH; Fink M; Sakurai H; Stolzenburg MR; Mauldin RL III; Smith J; Eisele F; Moore K; Sjostedt S; Tanner D;
Huey LG; Nowak JB; Edgerton E; Voisin D. (2005). A criterion for new particle formation in the sulfur-rich Atlanta
atmosphere. J Geophys Res, 110: 1-10. 191759
Mebust MR; Eder BK; Binkowski FS; Roselle SJ. (2003). Models-3 community multiscale air quality (CMAQ) model
aerosol component 2. Model evaluation. J Geophys Res, 108: 4184. 156749
Meng QY; Turpin BJ; Korn L; Weisel CP; Morandi M; Colome S; Zhang J; Stock T; Spektor D; Winer A; Zhang L; Lee
JH; Giovanetti R; Cui W; Kwon J; Alimokhtari S; Shendell D; Jones J; Farrar C; Maberti S. (2005). Influence of
ambient (outdoor) sources on residential indoor and personal PM2.5 concentrations: analyses of RIOPA data. J
Expo Sci Environ Epidemiol, 15: 17-28. 058595
Meng QY; Turpin BJ; Lee JH; Polidori A; Weisel CP; Morandi M; Colome S; Zhang JF; Stock T; Winer A. (2007). How
does infiltration behavior modify the composition of ambient PM2.5 in indoor spaces? An analysis of RIOPA data.
Environ Sci Tech, 41: 7315-7321. 194618
Mensink C; De Ridder K; Deutsch F; Lefebre F; Van de Vel K. (2008). Examples of scale interactions in local, urban, and
regional air quality modelling. Atmos Res, 89: 351-357. 155980
Mfula AM; Kukadia V; Griffiths RF; Hall DJ. (2005). Wind tunnel modelling of urban building exposure to outdoor
pollution. Atmos Environ, 39: 2737-2745. 123359
Middha P; Wexler A. (2006). Design of a Slot Nanoparticle Virtual Impactor. Aerosol Sci Technol, 40: 737-743. 155982
December 2009 3-209
-------
Middlebrook AM; Murphy DM; Lee S-H; Thomson DS; Prather KA; Wenzel RJ; Liu D-Y; Phares DJ; Rhoads KP; Wexler
AS; Johnston MV; Jimenez JL; Jayne JT; Worsnop DR; Yourshaw I; Seinfeld JH; Flagan RC. (2003). A comparison
of particle mass spectrometers during the 1999 Atlanta Supersites project. J Geophys Res, 108: 8424. 042932
Miller FJ; Gardner DE; Graham JA; Lee RE Jr; Wilson WE; Bachmann JD. (1979). Size considerations for establishing a
standard for inhalable particles. J Air Waste Manag Assoc, 29: 610-615. 070577
Miller SL; Anderson MJ; Daly EP; Milford JB. (2002). Source apportionment of exposures to volatile organic compounds I
Evaluation of receptor models using simulated exposure data. Atmos Environ, 36: 3629-3641. 030661
Misra C; Geller MD; Shah P; Sioutas C; Solomon PA. (2001). Development and evaluation of a continuous coarse (PM10-
PM25) particle monitor. J Air Waste Manag Assoc, 51: 1309-1317. 018998
Misra C; Geller MD; Sioutas C; Solomon PA. (2003). Development and evaluation of a PHI 0 impactor-inlet for a
continuous coarse particle monitor. Aerosol Sci Technol, 37: 271-281. 195001
Miyazaki Y; Kondo Y; Takegawa N; Komazaki Y; Fukuda M; Kawamura K; Mochida M; Okuzawa K; Weber RJ. (2006).
Time-resolved measurements of water-soluble organic carbon in Tokyo. J Geophys Res, 111: D23206. 156767
Molnar P; Johannesson S; Boman J; Barregard L; Sallsten G (2006). Personal exposures and indoor, residential outdoor,
and urban background levels of fine particle trace elements in the general population. J Environ Monit, 8: 543-551.
156773
Moolgavkar SH. (2000). Air pollution and hospital admissions for diseases of the circulatory system in three US
metropolitan areas. J Air Waste Manag Assoc, 50: 1199-1206. 010305
Moore K; Krudysz M; Pakbin P; Hudda N; Sioutas C. (2009). Intra-community variability in total particle number
concentrations in the San Pedro Harbor Area (Los Angeles, California). Aerosol Sci Technol, 43: 587-603. 191004
Moore KF; Ning Z; Ntziachristos L; Schauer JJ; Sioutas C. (2007). Daily variation in the properties of urban ultrafine
aerosol-Part I: Physical characterization and volatility. Atmos Environ, 41: 8633-8646. 122445
Morawska L; Keogh DU; Thomas SB; Mengersen K. (2007). Modality in ambient particle size distributions and its
potential as a basis for developing air quality regulation. Atmos Environ, 42: 1617-1628. 155990
Morawska L; Ristovski Z; Jayaratne ER; Keogh DU; Ling X. (2008). Ambient nano and ultrafine particles from motor
vehicle emissions: Characteristics, ambient processing and implications on human exposure . Atmos Environ, 42:
8113-8138. 191006
Morawska L; Thomas S; Jamriska M; Johnson G. (1999). The modality of particle size distributions of environmental
aerosols. Atmos Environ, 33: 4401-4411. 007609
Mulchandani A; Chen W; Mulchandani P; Rogers KR. (2001). Biosensors for direct determination of organophosphate
pesticides . Biosens Bioelectron, 16: 225-230. 191003
Muller K; Spindler G; Maenhaut W; Hitzenberger R; Wieprecht W; Baltensperger U; ten Brink H. (2004).
INTERCOMP2000, a campaign to assess the comparability of methods in use in Europe for measuring aerosol
composition. Atmos Environ, 38: 6459-6466. 097109
Murahashi T. (2003). Determination of mutagenic 3-nitrobenzanthrone in diesel exhaust particulate matter by three-
dimensional high-performance liquid chromatography. Analyst, 128: 42-5. 096539
Murphy BN; Pandis SN. (2009). Simulating the formation of semivolatile primary and secondary organic aerosol in a
regional chemical transport model. Environ Sci Technol, 43: 4722-4728. 190095
Nash DG; Tomas Baer T; Johnston MV. (2006). Aerosol mass spectrometry: an introductory review . Int J Mass Spectrom,
258:2-12. 199502
Nerriere E; Guegan H; Bordigoni B; Hautemaniere A; Momas I; Ladner J; Target A; Lameloise P; Delmas V; Personnaz
MB; Koutrakis P; Zmirou-Navier D. (2007). Spatial heterogeneity of personal exposure to airborne metals in
French urban areas. Sci Total Environ, 373: 49-56. 156801
Ng NL; Kroll JH; Chan AWH; Chhabra PS; Flagan RC; Seinfeld JH. (2007). Secondary organic aerosol formation from m-
xylene, toluene, and benzene. Atmos Chem Phys Discuss, 7: 4085-4126. 199528
Ning Z; Geller MD; Moore KF; Sheesley R; Schauer JJ; Sioutas C. (2007). Daily Variation in Chemical Characteristics of
Urban Ultrafine Aerosols and Inference of Their Sources. Environ Sci Technol, 41: 6000-6006. 156809
December 2009 3-210
-------
Ntziachristos L; Ning Z; Geller MD; Sioutas C. (2007). Particle concentration and characteristics near a major freeway
with heavy-duty diesel traffic. Environ Sci Technol, 41: 2223-2230. 089164
Ntziachristos L; Samaras Z. (2006). Combination of aerosol instrument data into reduced variables to study the consistency
of vehicle exhaust particle measurements. Atmos Environ, 40: 6032-6042. 116722
Odman MT; Boylan JW; Wilkinson JG; Russell AG; Mueller SF; Imhoff RE; Doty KG; Norris WB; McNider RT. (2002).
SAMI air quality modeling: final report. 092474
Offenberg JH; Lewandowski M; Edney EO; Kleindienst TE; Jaoui M. (2007). Investigation of a systematic offset in the
measurement of organic carbon with a semicontinuous analyzer. J Air Waste Manag Assoc, 57: 596-9. 098101
Ogulei D; Hopke PK; Zhou L; Patrick Pancras J; Nair N; Ondov JM. (2006). Source apportionment of Baltimore aerosol
from combined size distribution and chemical composition data. Atmos Environ, 40: 396-410. 119973
Olfert JS; Kulkarni P; Wang J. (2008). Measuring aerosol size distributions with the fast integrated mobility spectrometer. J
Aerosol Sci, 39: 940-956. 156004
Olivier JGJ; Bouwman AF; Berdowski JJM; Veldt C; Bloos JPJ; Visschedijk AJH; van der Maas CWM; Zandveld PYJ.
(1999). Sectoral emission inventories of greenhouse gases for 1990 on a per country basis as well as on 1 °x 1 °.
Environ Sci Pol, 2: 241-263. 156829
Olivier JGJ; Bouwman AF; Berdowski JJM; Veldt C; Bloos JPJ; Visschedijk AJH; Zandveld PYJ; Haverlag JL. (1996).
Description of EDGAR Version 2.0: A set of global emission inventories of greenhouse gases and ozone-depleting
substances for all anthropogenic and most natural sources on a per country basis and on 1 degree x 1 degree grid.
156828
Olson DA; Burke JM. (2006). Distributions of PM2.5 source strengths for cooking from the Research Triangle Park
particulate matter panel study. Environ Sci Technol, 40: 163-169. 189951
Olson DA; Mcdow SR. (2009). Near roadway concentrations of organic source markers. Atmos Environ, 43: 2862-2867.
191188
Olson DA; Norris GA. (2005). Sampling artifacts in measurement of elemental and organic carbon: Low-volume sampling
in indoor and outdoor environments. Atmos Environ, 39: 5437-5445. 156005
Orsini DA; Ma Y; Sullivan A; Sierau B; Baumann K; Weber RJ. (2003). Refinements to the particle-into-liquid sampler
(PILS) for ground and airborne measurements of water soluble aerosol composition. Atmos Environ, 37: 1243-
1259. 156008
Ott DK; Cyrs W; Peters TM. (2008). Passive measurement of coarse particulate matter, PM10-2.5. J Aerosol Sci, 39: 156-
167. 195004
Ott DK; Kumar N; Peters TM. (2008). Passive sampling to capture spatial variability in PM10-25. Atmos Environ, 42: 746-
756. 119394
Ozkaynak H; Xue J; Zhou H; Raizenne M. (1996). Associations between daily mortality and motor vehicle pollution in
Toronto, Canada. U.S. Environmental Protection Agency, Office of Reserch and Development, National Exposure
Research Laboratory. Research Triangle Park, NC. EPA/600/R-95/098. 073986
Paatero P; Tapper U. (1994). Positive matrix factorization: a non-negative factor model with optimal utilization of error
estimates of data values. Environmetrics, 5: 111-126. 086998
Paciorek CJ; Yanosky JD; Puett RC; Laden F; Suh HH. (2009). Practical large-scale spatio-temporal modeling of
particulate matter concentrations. Ann Appl Stat, 3: 370-397. 190090
Padro J. (1996). Summary of ozone dry deposition velocity measurements and model estimates over vineyard, cotton, grass
and deciduous forest in summer. Atmos Environ, 30: 2363-2369. 052446
Pakbin P;Ning Z;Schauer J;Sioutas C. (2009). Characterization of Particle Bound Organic Carbon from Diesel Vehicles
Equipped with Advanced Emission Control Technologies. Environ Sci Technol, 43: 4679-4686. 189893
Pandian MD; Behar JV; Ott WR; Wallace LA; Wilson AL; Colome S D. (1998). Correcting errors in the nationwide data
base of residential air exchange rates. J Expo Sci Environ Epidemiol, 8: 577-586. 090552
Pandis SN. (2004). Atmospheric aerosol processes. In McMurry PH; Shepard, M; Vickery JS (Ed.), Particulate Matter
Science for Policy Makers: ANARSTO AssessmentCambridge, UK: Cambridge University Press. 156838
December 2009 3-211
-------
Pang Y; Eatough NL; Modey WK; Eatough DJ. (2002). Evaluation of the RAMS continuous monitor for determination of
PM2.5 mass including semi-volatile material in Philadelphia, PA. J Air Waste Manag Assoc, 52: 563-572. 030353
Pang Y; Turpin B J; Gundel LA. (2006). On the Importance of Organic Oxygen for Understanding Organic Aerosol
Particles. Aerosol Sci Technol, 40: 128-133. 156012
Park J; Mitchell MJ; McHale PJ; Christopher SF; Myers TP (2003). Interactive effects of changing climate and
atmospheric deposition on N and S biogeochemistry in a forested watershed of the Adirondack Mountains, New
York State. Global Change Biol, 9: 1602-1619. 156842
Park K; Chow JC; Watson JG; Trimble DL; Doraiswamy P; Arnott WP; Stroud KR; Bowers K; Bode R; Petzold A; Hansen
AD. (2006). Comparison of continuous and filter-based carbon measurements at the Fresno supersite. J Air Waste
Manag Assoc, 56: 474-91. 098104
Park RJ; Stenchikov GL; Pickering; Dickerson RR; Allen DJ; Kondragunta S. (2001). Regional air pollution and its
radiative forcing: studies with a single column chemical and radiation transport model. J Geophys Res, 106:
28.751-28.770. 044169
Peltier RE; Hsu S-I; Lall R; Lippmann M. (2008). Residual oil combustion: A major source of airborne nickel in New York
City. J Expo Sci Environ Epidemiol, 19: 603-612. 197452
Peltier RE; Lippmann M. (2009). Residual oil combustion: 2. Distributions of airborne nickel and vanadium within New
York City. J Expo Sci Environ Epidemiol, TBD: 1-9. 197455
Peng RD; Dominici F; Pastor-Barriuso R; Zeger SL; Samet JM. (2005). Seasonal analyses of air pollution and mortality in
100 US cities. Am J Epidemiol, 161: 585-594. 087463
Petaja T; Mordas G; Manninen H; Aalto PP; Hmeri K; Kulmala M. (2006). Detection efficiency of a water-based TSI
condensation particle counter 3785. Aerosol Sci Technol, 40: 1090-1097. 156021
Peters K; Eiden R. (1992). Modelling the dry deposition velocity of aerosol particles to a spruce forest. Atmos Environ, 26:
2555-2564. 045277
Peters TM. (2006). Use of the Aerodynamic Particle Sizer to Measure Ambient PM 10-2. 5: The Coarse Fraction of PM 10.
J Air Waste Manag Assoc, 56: 411-416. 156860
Pettijohn FJ. (1957). Sedimentary rocks. New York: Harper & Brothers. 156862
Phares DJ; Rhoads KP; Johnston MV; Wexler AS. (2003). Size-resolved ultrafine particle composition analysis 2. Houston.
J Geophys Res, 108: 8420. 156866
Phuleria HC; Geller MD; Fine PM; Sioutas C. (2006). Size-Resolved emissions of organic tracers from light-and heavy-
duty vehicles measured in a California roadway tunnel. Environ Sci Technol, 40: 4109-4118. 156867
Phuleria HC; Sheesley RJ; Schauer JJ; Fine PM; Sioutas C. (2007). Roadside measurements of size-segregated particulate
organic compounds near gasoline and diesel-dominated freeways in Los Angeles, CA. Atmos Environ, 41: 4653-
4671.117816
Pierce JR; Adams PJ. (2009). Uncertainty in global CCN concentrations from uncertain aerosol nucleation and primary
emission rates. Atmos Chem Phys, 9: 1339-1356. 191189
Pirjola L; Parviainen H; Hussein T; Valli A; Haemeri K; Aaalto P; Virtanen A; Keskinen J; Pakkanen TA; Maekelae T;
Hillamo RE. (2004). 'Sniffer'; a novel tool for chasing vehicles and measuring traffic pollutants. Atmos Environ, 38:
3625-3635. 117564
Polidori A; Arhami M; Sioutas C; Delfino RJ; Allen R. (2007). Indoor/Outdoor relationships, trends, and carbonaceous
content of fine particulate matter in retirement homes of the Los Angeles Basin. J Air Waste Manag Assoc, 57: 366-
379. 156877
Polidori A; Turpin B; Meng QY; Lee JH; Weisel C; Morandi M; Colome S; Stock T; Winer A; Zhang J; Kwon J;
Alimokhtari S; Shendell D; Jones J; Farrar C; Maberti S. (2006). Fine organic particulate matter dominates indoor-
generated PM2.5 inRIOPAhomes. JExpo Sci Environ Epidemiol, 16: 321-331. 156876
Pope CAIII; Burnett RT; Thun MJ; Calle EE; Krewski D; Ito K; Thurston GD. (2002). Lung cancer, cardiopulmonary
mortality, and long-term exposure to fine particulate air pollution. JAMA, 287: 1132-1141. 024689
December 2009 3-212
-------
Pope CAIII; Thun MJ; Namboodiri MM; Dockery DW; Evans JS; Speizer FE; Heath CW Jr. (1995). Particulate air
pollution as a predictor of mortality in a prospective study of US adults. Am J Respir Grit Care Med, 151: 669-674.
045159
Poschl U. (2005). Atmospheric aerosols: composition, transformation, climate and health effects. Angew Chem Weinheim
Bergstr Ger, 44: 7520-7540. 156882
Pouliot G; Pace TG; Roy B; Pierce T; Mobley D. (2008). Development of a biomass burning emissions inventory by
combining satellite and ground-based information. J of Applied Remote Sensing, 2: 021501-021517. 156883
Price M; Bulpitt S; Meyer MB. (2003). A comparison of PM10 monitors at a Kerbside site in the northeast of England.
Atmos Environ, 37: 4425-4434. 098082
Puett RC; Schwartz J; Hart JE; Yanosky JD; Speizer FE; Suh H; Paciorek CJ; Neas LM; Laden F. (2008). Chronic
particulate exposure, mortality, and coronary heart disease in the nurses' health study. Am J Epidemiol, 168: 1161-
1168. 156891
Pun BK; Seigneur C; White W. (2003). Day-of-week behavior of atmospheric ozone in three US cities. J Air Waste Manag
Assoc, 53: 789-801. 047775
Qian S; Sakurai H; McMurry PH. (2007). Characteristics of regional nucleation events in urban East St Louis. Atmos
Environ, 41: 4119-4127. 116435
Qin X; Prather KA. (2006). Impact of biomass emissions on particle chemistry during the California Regional Particulate
Air Quality Study. Int J Mass Spectrom, 258: 142-150. 156895
Ramachandran G; Adgate JL; Pratt GC; Sexton K. (2003). Characterizing indoor and outdoor 15 minute average PM2.5
concentrations in urban neighborhoods. Aerosol Sci Technol, 37: 33-45. 195017
Rattigan OV; Hogrefe O; Felton HD; Schwab JJ; Roychowdhury UK; Husain L; Dutkiewicz VA; Demerjian KL. (2006).
Multi-year urban and rural semi-continuous PM25 sulfate and nitrate measurements in New York state: Evaluation
and comparison with filter based measurements. Atmos Environ, 40: 192-205. 115897
Rees SL; Robinson AL; Khlystov A; Stanier CO; Pandis SNSN. (2004). Mass balance closure and the Federal Reference
Method for PM2.5 in Pittsburgh, Pennsylvania. Atmos Environ, 38: 3305-3318. 097164
Reff A; Weisel CP; Zhang J; Morandi M; Stock T; Colome S; Winer A; Turpin BJ; Offenberg JH. (2007). A functional
group characterization of organic PM2.5 exposure: Results from the RIOPA study. Atmos Environ, 41: 4585-4598.
156045
Reithmeier C; Sausen R. (2002). ATTILA: atmospheric tracer transport in a Lagrangian model. Tellus Dyn Meteorol
Oceanogr, 54B: 278-299. 053447
Reponen T; Grinshpun SA; Trakumas S; Martuzevicius D; Wang Z-M; LeMasters G; Lockey JE; Biswas P. (2003).
Concentration gradient patterns of aerosol particles near interstate highways in the Greater Cincinnati airshed. J
Environ Monit, 5: 557-562. 088425
Rice J. (2004). Comparison of Integrated Filter and Automated Carbon Aerosol Measurements at Research Triangle Park,
North Carolina. Aerosol Sci Technol, 38: 23-36. 156049
Riddle SG; Jakober CA; Robert MA; Cahill TM; Charles MJ; Kleeman MJ. (2007). Large PAHs detected in fine particulate
matter emitted from light-duty gasoline vehicles. Atmos Environ, 41: 8658-8668. 115272
Rigby M; Toumi R. (2008). London air pollution climatology: Indirect evidence for urban boundary layer height and wind
speed enhancement. Atmos Environ, 42: 4932-4947. 156050
Roache PJ. (1998). Verification and validation in computational science and engineering. Socorro New Mexico: Hermosa
Publishing. 156915
Robinson AL; Donahue NM; Rogge WF. (2006). Photochemical oxidation and changes in molecular composition of
organic aerosol in the regional context. J Geophys Res, 111: D03302. 156918
Robinson AL; Donahue NM; Shrivastava MK; Weitkamp EA; Sage AM; Grieshop AP; Lane TE; Pierce JR; Pandis SN.
(2007). Rethinking organic aerosols: Semivolatile emissions and photochemical aging. Science, 315: 1259-1262.
156053
December 2009 3-213
-------
Rossner P; Svecova V; Milcova A; Lnenickova Z; Solansky N; Sram RJ. (2008). Seasonal variability of oxidative stress
markers in city bus drivers - Part I. Oxidative damage to DNA. Mutat Res Fund Mol Mech Mutagen, 642: 14-20.
156927
Rudich Y; Donahue NM; Mentel TF. (2007). Aging of Organic Aerosol: Bridging the Gap Between Laboratory and Field
Studies. Annu Rev Phys Chem, 58: 321. 156059
Russell A; Dennis R. (2000). NARSTO critical review of photochemical models and modeling. Atmos Environ, 34: 2283-
2324.035563
Russell M; Allen DT; Collins DR; Fraser MP. (2004). Daily, seasonal and spatial trends in PM25 mass and composition in
southeast Texas. Aerosol Sci Technol, 1: 14-26. 082453
Ryan PH; LeMasters GK. (2007). A Review of Land-use Regression Models for Characterizing Intraurban Air Pollution
Exposure. Inhal Toxicol, 19: 127. 156063
Ryan PH; LeMasters GK; Levin L; Burkle J; Biswas P; Hu S; Grinshpun S; Reponen T. (2008). A land-use regression
model for estimating microenvironmental diesel exposure given multiple addresses from birth through childhood.
Sci Total Environ, 404: 139-147. 156064
Ryno M; Rantanen L; Papaioannou E; Konstandopoulos AG; Koskentalo T; Savela K. (2006). Comparison of pressurized
fluid extraction, Soxhlet extraction and sonication for the determination of poly cyclic aromatic hydrocarbons in
urban air and diesel exhaust particulate matter. J Environ Monit, 8: 488-493. 156065
Saathoff H; Naumann KH; Schnaiter M; Schock W; Weingartner E; Baltensperger U; Kramer L; Bozoki Z; Poschl U;
Niessner R. (2003). Carbon mass determinations during the AIDA soot aerosol campaign 1999. J Aerosol Sci, 34:
1399-1420. 156066
Sabin LD; Kozawa K; Behrentz E; Winer AM; Fitz DR; Pankratz DV; Colome SD; Fruin SA. (2005). Analysis of real-time
variables affecting children's exposure to diesel-related pollutants during school bus commutes in Los Angeles.
Atmos Environ, 39: 5243-5254. 087728
Sadezky A; Muckenhuber H; Grothe H; Niessner R; Poschl U. (2005). Raman microspectroscopy of soot and related
carbonaceous materials: Spectral analysis and structural information. CarbonN Y, 43: 1731-1742. 097499
Sage AM; Weitkamp EA; Robinson AL; Donahue NM. (2008). Evolving mass spectra of the oxidized component of
organic aerosol: results from aerosol mass spectrometer analyses of aged diesel emissions. Atmos Chem Phys, 8:
1139-1152. 191758
Sahlodin AM; Sotudeh-Gharebagh R; Zhu Y. (2007). Modeling of dispersion near roadways based on the vehicle-induced
turbulence concept. Atmos Environ, 41: 92-102. 114058
Sakurai H; Tobias HJ; Park K; Zarling D; Docherty KS; Kittelson DB; McMurry PH; Ziemann PJ. (2003). On-line
measurements of diesel nanoparticle composition and volatility. Atmos Environ, 37: 1199-1210. 113924
Salminen K; Karlsson V. (2003). Comparability of low-volume PM10 sampler with β-attenuation monitor in
background air. Atmos Environ, 37: 3707-3712. 156070
Santarpia JL; Li RJ; Collins DR. (2004). Direct measurement of the hydration state of ambient aerosol populations. J
Geophys Res, 109: D18209. 156944
Sardar SB; Fine PM; Mayo PR; Sioutas C. (2005). Size-fractionated measurements of ambient ultrafine particle
chemicalcomposition in Los Angeles using the NanoMOUDI. Environ Sci Technol, 39: 932-944. 180086
Sardar SB; Solomon PA; Geller MD; Sioutas C. (2006). Development and evaluation of a high-volume dichotomous
sampler for chemical speciation of coarse and fine particles. J Aerosol Sci, 37: 1455-1466. 156071
Sarnat JA; Brown KW; Schwartz J; Coull BA; Koutrakis P. (2005). Ambient gas concentrations and personal particulate
matter exposures: implications for studying the health effects of particles. Epidemiology, 16: 385-395. 087531
Sarnat JA; Long CM; Koutrakis P; Coull BA; Schwartz J; Suh HH. (2002). Using sulfur as a tracer of outdoor fine
particulate matter. Environ Sci Technol, 36: 5305-5314. 037056
Sarnat JA; Schwartz J; Catalano PJ; Suh HH. (2001). Gaseous pollutants in particulate matter epidemiology: Confounders
or surrogates?. Environ Health Perspect, 109: 1053-1061. 019401
Sarnat SE; Coull BA; Ruiz PA; Koutrakis P; Suh HH. (2006). The influences of ambient particle composition and size on
particle infiltration in Los Angeles, CA residences. J Air Waste Manag Assoc, 56: 186-196. 089166
December 2009 3-214
-------
Sarnat SE; Klein M; Sarnat JA; Flanders WD; Waller LA; Mulholland JA; Russell AG; Tolbert PE. (2009). An examination
of exposure measurement error from air pollutant spatial variability in time-series studies. J Expo Sci Environ
Epidemiol, In Press: 1-12. 180084
Sarnat SE; Suh HH; Coull BA; Schwartz J; Stone PH; Gold DR. (2006). Ambient particulate air pollution and cardiac
arrhythmia in a panel of older adults in Steubenville, Ohio. Occup Environ Med, 63: 700-706. 090489
Sawant AA; Na K; Zhu X; Cocker K; Butt S; Song C; Cocker DR III. (2004). Characterization of PM25 and selected gas-
phase compounds at multiple indoor and outdoor sites in Mira Loma, California. Atmos Environ, 38: 6269-6278.
056798
Schauer JJ; Kleeman MJ; Cass GR; Simoneit BRT. (1999). Measurement of emissions from air pollution sources 2 C1
through C30 organic compounds from medium duty diesel trucks. Environ Sci Technol, 33: 1578-1587. 010582
Schauer JJ; Kleeman MJ; Cass GR; Simoneit BRT. (2002). Measurement of emissions from air pollution sources 5 Cl -
C32 organic compounds from gasoline-powered motor vehicles. Environ Sci Technol, 36: 1169-1180. 035332
Schauer JJ; Rogge WF; Hildemann LM; Mazurek MA; Cass GR. (1996). Source apportionment of airborne particulate
matter using organic compounds as tracers. Atmos Environ, 30: 3837-3855. 051162
Schnelle-Kreis J; Sklorz M; Peters A; Cyrys J; ZimmermannR. (2005). Analysis of particle-associated semi-volatile
aromatic and aliphatic hydrocarbons in urban particulate matter on a daily basis. Atmos Environ, 39: 7702-7714.
112944
Schwab JJ; Felton HD; Rattigan OV; Demerjian KL. (2006). New York State urban and rural measurements of continuous
PM2.5 mass by FDMS, TEOM, and BAM: Evaluations and Comparisons with the FRM . J Air Waste Manag
Assoc, 56: 372-83. 098449
Schwab JJ; Hogrefe O; Demerjian KL; Ambs JL. (2004). Laboratory characterization of modified tapered element
oscillating microbalance samplers. J Air Waste Manag Assoc, 54: 1254-63. 098450
Schwartz DE; Gong P; Shepard KL. (2008). Time-resolved Forster-resonance-energy-transfer DNA assay on an active
CMOS microarray. Biosens Bioelectron, 24: 383-390. 192094
Schwartz J; Sarnat JA; Coull BA; Wilson WE. (2007). Effects of exposure measurement error on particle matter
epidemiology: a simulation using data from a panel study in Baltimore, MD. J Expo Sci Environ Epidemiol, 17: S2-
S10. 090220
Seagrave JC; McDonald JD; Bedrick E; Edgerton ES; Gigliotti AP; Jansen JJ; Ke L; Naeher LP; Seilkop SK; Zheng M;
Mauderley JL. (2006). Lung toxicity of ambient particulate matter from southeastern US sites with different
contributing sources: relationships between composition and effects. Environ Health Perspect, 114: 1387-93.
091291
SeamanNL. (2000). Meteorological modeling for air quality assessments. Atmos Environ, 34: 2231-2259. 035562
Seinfeld JH; Pandis SN. (1998). Atmospheric chemistry and physics: from air pollution to climate change. New York: John
Wiley & Sons. 018352
Sheesley RJ; Schauer JJ; Meiritz M; Deminter JT; Bae M-S; Turner JR. (2007). Daily Variation in Particle-Phase Source
Tracers in an Urban Atmosphere. Aerosol Sci Technol, 41: 981-993. 112017
Shen S; Jaques PA; Zhu Y; Geller MD; Sioutas C. (2002). Evaluation of the SMPS-APS system as a continuous monitor
for measuring PM2. 5, PM10 and coarse (PM2. 5- 10) concentrations. Atmos Environ, 36: 3939-3950. 156086
Sheppard L; Slaughter JC; Schildcrout J; Liu L-JS; Lumley T. (2005). Exposure and measurement contributions to
estimates of acute air pollution effects. J Expo Sci Environ Epidemiol, 15: 366-376. 079176
Sheya SA; Glowacki C; Chang MC; Chow JC; Watson JG (2008). Hot filter/impinger and dilution sampling for fine
particulate matter characterization from ferrous metal casting processes. J Air Waste Manag Assoc, 58: 553-561.
156977
Shi Q; Sakurai H; McMurry PH. (2007). Climatology of Regional Nucleation Events in Urban East St. Louis. Atmos
Environ, 41: 4119-4127 . 191760
Shimpi S; Khalek I; Bougher T; Tennant C. (2009). Number count measurements of particulate trap-equipped diesel truck
exhaust from engines that meet 2007 US on-highway emission regulations. Presented at A&WMA's 102nd Annual
Conference and Exhibition, June 16-19, 2009, Pittsburgh, PA. 189888
December 2009 3-215
-------
Shrivastava MK; Subramanian R; Rogge WF; Robinson AL. (2007). Sources of organic aerosol: Positive matrix
factorization of molecular marker data and comparison of results from different source apportionment models.
Atmos Environ, 41: 9353-9369. 111594
Singh M; Misra C; Sioutas C. (2003). Field evaluation of a personal cascade impactor sampler (PCIS). Atmos Environ, 37:
4781-4793. 156088
Singh M; Phuleria HC; Bowers K; Sioutas C. (2006). Seasonal and spatial trends in particle number concentrations and
size distributions at the children's health study sites in Southern California. JExpo Sci Environ Epidemiol, 16: 3-18.
190136
Sioutas C; Delfino RJ; Singh M. (2005). Exposure assessment for atmospheric ultrafine particles (UFPs) and implications
in epidemiologic research. Environ Health Perspect, 113: 947-955. 088428
Sioutas C; Kim S; Chang M; Terrell LL; Gong H Jr. (2000). Field evaluation of a modified DataRAM MIE scattering
monitor for real-time PM25 mass concentration measurements. Atmos Environ, 34: 4829-4838. 025223
Slowik JG; Cross ES; Han J-H; Davidovits P; Onasch TB; Jayne JT; Williams LR; Canagaratna MR; Worsnop DR;
Chakrabarty RK; Moosm; uuml; Her H; Arnott WP; Schwarz JP; Gao R-S; Fahey DW; Kok GL; Petzold A. (2007).
An Inter-Comparison of Instruments Measuring Black Carbon Content of Soot Particles. Aerosol Sci Technol, 41:
295-314.096177
Smith JN; Dunn MJ; VanReken TM; lida K; Stolzenburg MR; McMurry PH; Huey LG (2008). Chemical composition of
atmospheric nanoparticles formed from nucleation in Tecamac, Mexico: evidence for an important role for organic
species in nanoparticle growth. Geophys Res Lett, 35: L04808. 199529
Smith JN; Moore KF; McMurry PH; Eisele FL. (2004). Atmospheric Measurements of Sub-20 run Diameter Particle
Chemical Composition by Thermal Desorption Chemical lonization Mass Spectrometry. Aerosol Sci Technol, 38:
100-110. 156090
Smith RL; Spitzner D; Kim Y; Fuentes M. (2000). Threshold dependence of mortality effects for fine and coarse particles
in Phoenix, Arizona. J Air Waste Manag Assoc, 50: 1367-1379. 010335
Smith WH; Staskawicz BJ. (1977). Removal of atmospheric particles by leaves and twigs of urban trees: some preliminary
observations and assessment of research needs. J Environ Manage, 1: 317-330. 046675
So ESP; Chan AT Y; Wong AYT. (2005). Large-eddy simulations of wind flow and pollutant dispersion in a street canyon.
Atmos Environ, 39: 3573-3582. 110746
Solomon P; Baumann K; Edgerton E; Tanner R; Eatough D; Modey W; Marin H; Savoie D; Natarajan S; Meyer MB.
(2003). Comparison of integrated samplers for mass and composition during the 1999 Atlanta supersites project. J
Geophys Res, 108: 8423. 156994
Solomon PA; Hopke PK. (2008). A Special Issue of JA&WMA Supporting Key Scientific and Policy-and Health-Relevant
Findings from EPA's Particulate Matter Supersites Program and Related Studies: An Integration and Synthesis of
Results. J Air Waste Manag Assoc, 58: 137. 156997
Solomon PA; Sioutas C. (2008). Continuous and semicontinuous monitoring techniques for particulate matter mass and
chemical components: a synthesis of findings from EPA's Particulate Matter Supersites Program and related studies.
J Air Waste Manag Assoc, 58: 164-195. 190139
Song CH; Carmichael GR. (2001). A three-dimensional modeling investigation of the evolution processes of dust and sea-
salt particles in east Asia. J Geophys Res, 106: 18,131-18,154.036064
Stanier CO; Khly stov AY; Chan WR; Mandiro M; Pandis SN. (2004). A Method for the In Situ Measurement of Fine
Aerosol Water Content of Ambient Aerosols: The Dry-Ambient Aerosol Size Spectrometer (DAASS). Aerosol Sci
Technol, 38: 215 - 228. 095955
Stein AF; Lamb D; Draxler RR. (2000). Incorporation of detailed chemistry into a three-dimensional Lagrangian-Eulerian
hybrid model: application to regional tropospheric ozone. Atmos Environ, 34: 4361-4372. 048341
Stevenson D; Dentener FJ; Schultz MG; Ellingsen K; Van Noije TPC; Wild O; Zeng G; Amann M; Atherton CS; Bell N;
Bergmann DJ; Bey I; Butler T; Cofala J; Collins WJ; Derwent RG; Doherty RM; Drevet J; Eskes HJ; Fiore AM;
Gauss M; Hauglustaine DA; Horowitz LW; Isaksen ISA; Krol MC; Lamarque J-F; Lawrence MG; Montanaro V;
Muller J-F; Pitari G; Prather MJ; Pyle JA; Rast S; Rodriguez JM; Sanderson MG. (2006). Multimodel ensemble
simulations of present-day and near-future tropospheric ozone. J Geophys Res, 111: D08301. 089222
December 2009 3-216
-------
Stockwell WR; Middleton P; Chang JS; Tang X. (1990). The second generation Regional Acid Deposition Model chemical
mechanism for regional air quality modeling. J Geophys Res, 95: 16,343-16,367. 043095
Stolzenburg MR; Dutcher DD; Kirby BW; Hering SV. (2003). Automated Measurement of the Size and Concentration of
Airborne Particulate Nitrate. Aerosol Sci Technol, 37: 537-546. 156102
Stolzenburg MR; Hering SV. (2000). Method for the automated measurement of fine particle nitrate in the atmosphere.
Environ Sci Technol, 34: 907-914. 013289
Stracquadanio M; Bergamini D; Massaroli E; Trombini C. (2005). Field evaluation of a passive sampler of poly cyclic
aromatic hydrocarbons (PAHs) in an urban atmosphere (Bologna, Italy). J Environ Monit, 7: 910-915. 156104
Strand M; Hopke PK; Zhao W; Vedal S; Gelfand E; Rabinovitch N. (2007). A study of health effect estimates using
competing methods to model personal exposures to ambient PM2.5. J Expo Sci Environ Epidemiol, 17: 549-558.
157018
Strand M; Vedal S; Rodes C; Dutton SJ; Gelfand EW; Rabinovitch N. (2006). Estimating effects of ambient PM2.5
exposure on health using PM2.5 component measurements and regression calibration. J Expo Sci Environ
Epidemiol, 16: 30-38. 089203
Streets DG; Zhang Q; Wang L; He K; Hao J; Wu Y; Tang Y; Carmichael GR. (2006). Revisiting China?s CO emissions
after the Transport and Chemical Evolution over the Pacific (TRACE-P) mission: Synthesis of inventories,
atmospheric modeling, and observations. J Geophys Res, 111: Dl4306. 157019
Subramanian R; Khlystov A; Robinson A. (2006). Effect of Peak Inert-Mode Temperature on Elemental Carbon Measured
Using Thermal-Optical Analysis. Aerosol Sci Technol, 40: 763-780. 156107
Subramanian R; Khlystov AY; Cabada JC; Robinson AL. (2004). Positive and negative artifacts in particulate organic
carbon measurements with denuded and undenuded sampler configurations. Aerosol Sci Technol, 1: 27-48. 081203
Sullivan AP; Peltier RE; Brock CA; de Gouw JA; Holloway JS; Warneke C; Wollny AG; Weber RJ. (2006). Airborne
measurements of carbonaceous aerosol soluble in water over northeastern United States: Method development and
an investigation into water-soluble organic carbon sources. J Geophys Res, 111: 1-14. 157030
Sullivan AP; Weber RJ. (2006). Chemical characterization of the ambient organic aerosol soluble in water: 1. Isolation of
hydrophobic and hydrophilic fractions with a XAD-8 resin. J Geophys Res, 111: D05314. 157031
Sullivan AP; Weber RJ; Clements AL; Turner JR; Bae MS; Schauer JJ. (2004). A method for on-line measurement of
water-soluble organic carbon in ambient aerosol particles: Results from an urban site. Geophys Res Lett, 31:
LI3105. L57029
Sullivan JH; Hubbard R; Liu SL; Shepherd K; Trenga CA; Koenig JQ; Chandler WL; Kaufman JD. (2007). A community
study of the effect of particulate matter on blood measures of inflammation and thrombosis in an elderly population.
Environ Health Perspect, 6:3. 100083
Swartz E; Stockburger L; Gundel LA. (2003). Recovery of Semivolatile Organic Compounds during Sample Preparation:
Implications for Characterization of Airborne Particulate Matter. Environ Sci Technol, 37: 597-605. 157035
Swietlik D; Faust M. (1984). Foliar nutrition of fruit crops. Hortic Rev (Am Soc Hortic Sci), 6: 287-355. 046678
S0rensen M; Loft S; Andersen HV; Raaschou-Nielsen O; Skovgaard LT; Knudsen LE; Nielsen IV; Hertel O. (2005).
Personal exposure to PM25, black smoke and NO2 in Copenhagen: relationship to bedroom and outdoor
concentrations covering seasonal variation. JExpo Sci Environ Epidemiol, 15: 413-422. 089428
Taha G; Box GP; Cohen DD; Stelcer E. (2007). Black Carbon Measurement using Laser Integrating Plate Method. Aerosol
Sci Technol, 41: 266 - 276. 096277
Taylor DA. (2002). Dust in the wind. Environ Health Perspect, 110: A80-A87. 025693
Taylor GE Jr; Hanson PJ; Baldocchi DD. (1988). Pollutant deposition to individual leaves and plant canopies: sites of
regulation and relationship to injury. In Heck, W. W.; Taylor, O. C.; Tingey, D. T. (Ed.),Assessment of crop loss
from air pollutants (pp. 227-257). New York, NY: Elsevier Applied Science. 019289
Temime-Roussel B; Monod A; Massiani C; Wortham H. (2004). Evaluation of an annular denuder for atmospheric PAH
partitioning studies—2: evaluation of mass and number particle losses. Atmos Environ, 38: 1925-1932. 098530
December 2009 3-217
-------
Temime-Roussel B; Monod A; Massiani C; Wortham H. (2004). Evaluation of an annular denuder tubes for atmospheric
PAH partitioning studies—1: evaluation of the trapping efficiency of gaseous PAHS. Atmos Environ, 38: 1913-
1924.098521
ten Brink H; Hoek G; Khlystov A. (2005). An approach to monitor the fraction of elemental carbon in the ultrafine aerosol.
Atmos Environ, 39: 6255-6259. 156115
ten Brink H; Maenhaut W; Hitzenberger R; Gnauk T; Spindler G; Even A; Chi X; Bauer H; Puxbaum H; Putaud J-P; Tursic
J; Berner A. (2004). INTERCOMP2000: the comparability of methods in use in Europe for measuring the carbon
content of aerosol. Atmos Environ, 38: 6507-6519. 097110
Tesche TW; Morris R; Tonnesen G; McNally D; Boylan J; Brewer P. (2006). CMAQ/CAMx annual 2002 performance
evaluation over the eastern US. Atmos Environ, 40: 4906-4919. 157050
Thornburg J; Rodesa CE; Lawless PA; Williams R. (2009). Spatial and Temporal Variability of Outdoor Coarse Particulate
Matter Mass Concentrations Measured with a New Coarse Particle Sampler During the Detroit Exposure and
Aerosol Research Study . Atmos Environ, 43: 4251-4258. 190999
Thurston GD; Spengler JD. (1985). A quantitative assessment of source contributions to inhalable particulate matter
pollution in metropolitan Boston. Atmos Environ, 19: 9-25. 056074
Tobias HJ; Kooiman PM; Docherty KS; Ziemann PJ. (2000). Real-Time Chemical Analysis of Organic Aerosols Using a
Thermal Desorption Particle Beam Mass Spectrometer. Aerosol Sci Technol, 33: 170-190. 156121
Tobias HJ; Ziemann PJ. (1999). Compound Identification in Organic Aerosols Using Temperature-Programmed Thermal
Desorption Particle Beam Mass Spectrometry. Anal Chem, 71: 3428-3435. 157053
Tolocka M; Jang PM; Ginter JM; Cox FJ; Kamens RM; Johnston MV. (2004). Formation of oligomers in secondary
organic aerosol. Environ Sci Technol, 38: 1428-1434. 087578
Trijonis JC; Malm WC; Pitchford M; White WH; Charlson R. (1990). Acidic deposition: State of science and technology.
Report 24. Visibility: Existing and historical conditions-causes and effects. Final report. 157058
Tsai CJ; Chang CT; Huang CH. (2006). Direct field observation of the relative humidity effect on the beta-gauge readings.
J Air Waste Manag Assoc, 56: 834-40. 098312
Tsai YI; Kuo SC. (2006). Development of diffuse reflectance infrared Fourier transform spectroscopy for the rapid
characterization of aerosols. Atmos Environ, 40: 1781-1793. 156127
TSI. (2005). Data Merge Software Module for merging and fitting of SMPAand APS Data Files, P/N 1930074, User's
Guide, Model 390069, Revision A. St. Paul, MN. A. St Paul, MN: TSI, Incorporated 157196
Tuch TM; Herbarth O; Franck U; Peters A; Wehner B; Wiedensohler A; Heintzenberg J. (2006). Weak correlation of
ultrafine aerosol particle concentrations< 800 nm between two sites within one city. J Expo Sci Environ Epidemiol,
16:486-491. 157060
Tung TCW; Chao CYH; Burnett J. (1999). A methodology to investigate the particulate penetration coefficient through
building shell. Atmos Environ, 33: 881-893. 049003
Turpin BJ; Lim H-J. (2001). Species contributions to PM25 mass concentrations: revisiting common assumptions for
estimating organic mass. Aerosol Sci Technol, 35: 602-610. 017093
Turpin BJ; Weisel CP; Morandi M; Colome S; Stock T; Eisenreich S; Buckley B. (2007). Relationships of Indoor, Outdoor,
and Personal Air (RIOPA): Part II. Analyses of concentrations of particulate matter species. 157062
U.S. EPA. (1982). Air quality criteria for particulate matter and sulfur oxides, Vol 1, 2, 3. U.S. Environmental Protection
Agency, Environmental Criteria and Assessment Office. Washington, D.C.. EPA 600/8-82-029a; EPA 600/8-82-
029b; EPA 600/8-82-029c. http://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=3000188Z.txt;
http ://nepis. epa. gov/Exe/ZyPURL. cgi?Dockey=3 00018EV.txt;
http://nepis.epa.gov/Exe/ZyPURL.cgi?Dockev=300053KV.txt. 017610
U.S. EPA. (1996). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research Triangle
Park, NC. EPA/600/P-95/001aF-cF. 079380
U.S. EPA. (1998). SLAMS/NAMS/PAMS network review guidance. 093211
U.S. EPA. (2004). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research Triangle
Park, NC. EPA/600/P-99/002aF-bF. 056905
December 2009 3-218
-------
U.S. EPA. (2005). Review of the national ambient air quality standards for participate matter: Policy assessment of
scientific and technical information OAQPS staff paper. U.S. Environmental Protection Agency. Washington, DC.
EPA/452/R-05-005a. http://www.epa.gov/ttn/naaqs/standards/pm/data/pmstaffpaper_20051221 .pdf. 090209
U.S. EPA. (2006). 2002 National Emissions Inventory Data and Documentation. Retrieved 02-DEC-09, from
http ://www. epa. gov/ttnchie 1 /net/2002inventory.html. 157070
U.S. EPA. (2006). Air quality criteria for ozone and related photochemical oxidants. EPA. DC. 088089
U. S. EPA. (2006). Provisional Assessment of Recent Studies on Health Effects of Particulate Matter Exposure. U. S.
Environmental Protection Agency. Research Triangle Park, NC. 157071
U.S. EPA. (2008). Analysis of particulate matter emissions from light-duty gasoline vehicles in Kansas City. U.S.
Environmental Protection Agency. Washington, D.C.. EPA420-R-08-010. 191767
U.S. EPA. (2008). Integrated Science Assessment for Oxides of Nitrogen - Health Criteria. U.S. Environmental Protection
Agency. Research Triangle Park, NC. EPA/600/R-08/071. 157073
U.S. EPA. (2008). Integrated science assessment for oxides of nitrogen and sulfur: Ecological criteria. U.S. Environmental
Protection Agency. Research Triangle Park, NC. EPA/600/R-08/082F. 157074
U.S. EPA. (2008). Integrated Science Assessment for Sulfur Oxides - Health Criteria. U.S. Environmental Protection
Agency. Research Triangle Park, NC. EPA/600/R-08/047F.
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=l 98843. 157075
U.S. EPA. (2008). National Air Quality Status and Trends Through 2007. U.S. Environmental Protection Agency. Research
Triangle Park. EPA-454/R-08-006 . 191190
U.S. EPA. (2008). Total Risk Integrated Methodology (TRIM) Air Pollutants Exposure Model Documentation
(TRIM.Expo/APEX, Version 4.3). Volume 1: User's Guide. U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards. Research Triangle Park, NC. EPA-452/B-08-001b. 191775
U.S. EPA. (2008). U.S. EPA's 2008 Report on the Environment (Final Report). U.S. Environmental Protection Agency. DC.
157076
U. S. EPA. (2009). Risk and Exposure Assessment for Review of the Secondary National Ambient Air Quality Standards for
Oxides of Nitrogen and Oxides of Sulfur Second Draft. U.S. Environmental Protection Agency. Washington, D.C..
EPA-452/P-09-004a. http ://www.epa.gov/ttnnaaqs/standards/no2so2sec/cr_rea.html. 191774
U.S. EPA. (2009). Verified Retrofit Technologies. Retrieved 02-DEC-09, from http://www.epa.gov/otaq/retrofit/index.htm.
189885
UNCEC. (2007). Hemispheric Transport of Air Pollution 2007. New York: United Nations Publications. 157078
Unsworth MH; Wilshaw JC. (1989). Wet, occult and dry deposition of pollutants on forests. Agr Forest Meteorol, 47: 221-
238. 046682
VanCuren RA; Cahill TA. (2002). Asian aerosols in North America: Frequency and concentration of fine dust. J Geophys
Res, 107: 4804. 157087
Van AalstRM. (1982). Dry deposition of NOx. In Schneider, T.; Grant, L. (Ed.),Air pollution by nitrogen
oxidesAmsterdam, The Netherlands: Elsevier Scientific Publishing Company. 036481
van der Werf GR; Randerson JT; Giglio L; Collatz GJ; Kasibhatla PS; Arellano Jr AF. (2006). Interannual variability in
global biomass burning emissions from 1997 to 2004. Atmos Chem Phys, 6: 3423-3441. 157084
Vega E; Reyes E; Wellens A; Sanchez G; Chow JC; Watson JG (2003). Comparison of continuous and filter based mass
measurements in Mexico City. Atmos Environ, 37: 2783-2793. 105974
Venkataraman C; Lyons JM; Friedlander SK. (1994). Size distributions of poly cyclic aromatic hydrocarbons and elemental
carbon 1 Sampling, measurement methods, and source characterization. Environ Sci Technol, 28: 555-562. 002475
Venn A; Lewis S; Cooper M; Hubbard R; Hill I; Boddy R; Bell M; Britton J. (2000). Local road traffic activity and the
prevalence, severity and persistence of wheeze in school children: combined cross sectional and longitudinal study.
Occup Environ Med, 57: 152-158. 007895
Venn AJ; Lewis SA; Cooper M; Hubbard R; Britton J. (2001). Living near a main road and the risk of wheezing illness in
children. Am J Respir Grit Care Med, 164: 2177-2180. 023644
December 2009 3-219
-------
Veranth JM; Moss TA; Chow JC; Labban R; Nichols WK; Walton JC; Walton JG; Yost GS. (2006). Correlation of in vitro
cytokine responses with the chemical composition of soil-derived particulate matter. Environ Health Perspect, 114:
341-349. Q87479
Viana M; Chi X; Maenhaut W; Cafmeyer J; Querol X; Alastuey A; Mikuska P; Vecera Z. (2006). Influence of Sampling
Artefacts on Measured PM, OC, and EC Levels in Carbonaceous Aerosols in an Urban Area. Aerosol Sci Technol,
40: 107-117. 179987
Viana M; Querol X; Alastuey A; Ballester F; Llop S; Esplugues A; FernAjndez-Patier R; GarcA-a dos Santos S; Herce
MD. (2008). Characterising exposure to PM aerosols for an epidemiological study. Atmos Environ, 42: 1552-1568.
156135
Violante FS; Barbieri A; Curti S; Sanguinetti G; Graziosi F; Mattioli S. (2006). Urban atmospheric pollution: Personal
exposure versus fixed monitoring station measurements. Chemosphere, 64: 1722-1729. 156140
Virkkula A; Makela T; Hillamo R; Yli-Tuomi T; Hirsikko A; Hameri K; Koponen IK. (2007). A simple procedure for
correcting loading effects of aethalometer data. J Air Waste Manag Assoc, 57: 1214-1222. 157098
Voisin D; Smith JN; Sakurai H; McMurry PH; Eisele FL. (2003). Thermal Desorption Chemical lonization Mass
Spectrometer for Ultrafine Particle Chemical Composition. Aerosol Sci Technol, 37: 471-475. 156141
Volckens J; Olson DA; Hays MD. (2008). Carbonaceous species emitted from handheld two-stroke engines. Atmos
Environ, 42: 1239-1248. 105465
Vyas VM; Christakos G. (1997). Spatiotemporal analysis and mapping of sulfate deposition data over Eastern USA. Atmos
Environ, 31: 3623-3633. 156142
Wagner J; LeithD. (2001). Passive aerosol sampler. Parti: Principle of operation. Aerosol Sci Technol, 34: 186-192.
190153
Wagner J; LeithD. (2001). Passive aerosol sampler. Part II: Wind tunnel experiments. Aerosol Sci Technol, 34: 193-201.
190154
Wallace L. (2000). Real-time monitoring of particles, PAH, and CO in an occupied townhouse. J Occup Environ Hyg, 15:
39-47. 000803
Wallace L; Williams R. (2005). Use of personal-indoor-outdoor sulfur concentrations to estimate the infiltration factor and
outdoor exposure factor for individual homes and persons. Environ Sci Technol, 39: 1707-1714. 057485
Wallace L; Williams R; Rea A; Croghan C. (2006). Continuous weeklong measurements of personal exposures and indoor
concentrations of fine particles for 37 health-impaired North Carolina residents for up to four seasons. Atmos
Environ, 40: 399-414. 088211
Wallace L; Williams R; Suggs J; Jones P. (2006). Estimating contributions of outdoor fine particles to indoor
concentrations and personal exposures: effects of household characteristics and personal activities. U.S.
Environmental Protection Agency; National Exposure Research Laboratory. Research Triangle Park, NC.
EPA/600/R-06/023. 089190
Wang J; Christopher SA; Nair US; Reid JS; Prins EM; Szykman J; Hand JL. (2006). Mesoscale modeling of Central
American smoke transport to the United States: 1. "Top-down" assessment of emission strength and diurnal
variation impacts. J Geophys Res, 111: DOS SI 7. 157109
Wang LH; Milford JB; Carter WPL. (2000). Reactivity estimates for aromatic compounds Part 2 uncertainty in incremental
reactivities. Atmos Environ, 34: 4349-4360. 048365
Wang LH; Milford JB; Carter WPL. (2000). Reactivity estimates for aromatic compounds Part I: uncertainty in chamber-
derived parameters. Atmos Environ, 34: 4337-4348. 048357
Watson JG; Chen LW; Chow JC; Doraiswamy P; Lowenthal DH. (2008). Source apportionment: findings from the U.S.
Supersites Program. J Air Waste Manag Assoc, 58: 265-288. 157128
Watson JG; Chow JC. (2002). Comparison and evaluation of in situ and filter carbon measurements at the Fresno Supersite.
J Geophys Res, 107: 8341. 037873
Watson JG; Chow JC; Chen LWA. (2005). Summary of organic and elemental carbon/black carbon analysis methods and
intercomparisons. Aerosol Air Qual Res, 5: 69-102. 157125
December 2009 3-220
-------
Watson JG; Fujita EM; Chow JC; Zielinska B; Richards LW; Neff W; Dietrich D. (1998). Northern front range air quality
study Final report. 012257
Watson JG; Robinson NF; Chow JC; Henry RC; Kim BM; Pace TG; Meyer EL; Nguyen Q. (1990). The USEPA/DRI
chemical mass balance receptor model, CMB 70. Environ Software, 5: 38-49. 004848
Weber RJ; Orsini D; Daun Y; Lee Y-N; Kotz PJ; Brechtel F. (2001). A particle-into-liquid collector for rapid measurement
of aerosol bulk chemical composition. Aerosol Sci Technol, 35: 718-727. 024640
Weber RJ; Orsini D; Duan Y; Baumann K; Kiang CS; Chameides W; Lee YN; Brechtel F; Klotz P; Jongejan P. (2003).
Intercomparison of near real time monitors of PM2. 5 nitrate and sulfate at the US Environmental Protection
Agency Atlanta Supersite. J Geophys Res, 108: 8421. 157129
Weijers EP; Khlystov AY; Kos GPA; Erisman JW (2004). Variability of particulate matter concentrations along roads and
motorways determined by a moving measurement unit. Atmos Environ, 38: 2993-3002. 104186
Weingartner E; Saathoff H; Schnaiter M; Streit N; Bitnar B; Baltensperger U. (2003). Absorption of light by soot particles:
determination of the absorption coefficient by means of aethalometers. J Aerosol Sci, 34: 1445-1463. 156149
Wenzel RJ; Liu DY; Edgerton ES; Prather KA. (2003). Aerosol time-of-flight mass spectrometry during the Atlanta
Supersite Experiment: 2. Scaling procedures. J Geophys Res, 108: 8426. 157139
Wesely ML; Eastman JA; Stedman DH; Yalvac ED. (1982). An eddy-correlation measurement of NO2 flux to vegetation
and comparison to O3 flux. Atmos Environ, 16: 815-820. 036564
Wesely ML; Hicks BB. (2000). Areview of the current status of knowledge on dry deposition. Atmos Environ, 34: 2261-
2282. 025018
Westerdahl D; Fruin S; Sax T; Fine PM; Sioutas C. (2005). Mobile platform measurements of ultrafine particles and
associated pollutant concentrations on freeways and residential streets in Los Angeles. Atmos Environ, 39: 3597-
3610. 086502
Whitby KT. (1978). The physical characteristics of sulfur aerosols. Atmos Environ, 12: 135-159. 071181
Williams B; Goldstein A; Kreisberg N; Hering S. (2006). An In-Situ Instrument for Speciated Organic Composition of
Atmospheric Aerosols: Thermal Desorption Aerosol GC/MS-FID (TAG). Aerosol Sci Technol, 40: 627-638.
156157
Williams R; Rea A; Vette A; Croghan C; Whitaker D; Stevens C; McDow S; Fortmann R; Sheldon L; Wilson H; Thornburg
J; Phillips M; Lawless P; Rodes C; Daughtrey H. (2008). The design and field implementation of the Detroit
exposure and aerosol research study. J Expo Sci Environ Epidemiol, 19: 643-659. 191201
Williams TC; Shaddix CR; Jensen KA; Suo-Anttila JM. (2006). Measurements of the Dimensionless Extinction Coefficient
of Soot Within Laminar Diffusion Flames. Combust Flame, 50: 1616-1630. 157148
Wilson JG; Zawar-Reza P. (2006). Intraurban-scale dispersion modelling of particulate matter concentrations: applications
for exposure estimates in cohort studies. Atmos Environ, 40: 1053-1063. 088292
Wilson WE; Brauer M. (2006). Estimation of ambient and non-ambient components of particulate matter exposure from a
personal monitoring panel study. J Expo Sci Environ Epidemiol, 16: 264-274. 088933
Wilson WE; Grover BD; Long RW; Eatough NL; Eatough DJ. (2006). The measurement of fine particulate semivolatile
material in urban aerosols. J Air Waste Manag Assoc, 56: 384-387. 091142
Wilson WE; Mage DT; Grant LD. (2000). Estimating separately personal exposure to ambient and nonambient particulate
matter for epidemiology and risk assessment: why and how. J Air Waste Manag Assoc, 50: 1167-1183. 010288
Wilson WE; Mar TF; Koenig JQ. (2007). Influence of exposure error and effect modification by socioeconomic status on
the association of acute cardiovascular mortality with particulate matter in Phoenix. J Expo Sci Environ Epidemiol,
17: S11-S19. 157149
Wilson WE; Stanek J; Han HS; Johnson T; Sakurai H; Pui DY; Turner J; Chen DR; Duthie S. (2007). Use of the electrical
aerosol detector as an indicator of the surface area of fine particles deposited in the lung. J Air Waste Manag Assoc,
57:211-20.098398
Wilson WE; Suh HH. (1997). Fine particles and coarse particles: concentration relationships relevant to epidemiologic
studies. J Air Waste Manag Assoc, 47: 1238-1249. 077408
December 2009 3-221
-------
Winkler PM; Steiner G; Vrtala A; Vehkamaki H; Noppel M; Lehtinen KEJ; Reischl GP; Wagner PE; Kulmala M. (2008).
Heterogeneous Nucleation Experiments Bridging the Scale from Molecular Ion Clusters to Nanoparticles. Science,
319: 1374-1377. 156160
Womiloju TO; Miller JD; Mayer PM; Brook JR. (2003). Methods to determine the biological composition of particulate
matter collected from outdoor air. Atmos Environ, 37: 4335-4344. 179954
Wongphatarakul V; Friedlander SK; Pinto JP (1998). Acomparative study of PM2.5 ambient aerosol chemical databases.
Environ Sci Technol, 32: 3926-3934. 049281
Woo K-S; Chen D-R; Pui DYH; McMurry PH. (2001). Measurement of Atlanta aerosol size distributions: observations of
ultrafine particle events. Aerosol Sci Technol, 34: 75-87. 011702
Wu C-F; Delfino RJ; Floro JN; Quintana;. (2005). Exposure assessment and modeling of particulate matter for asthmatic
children using personal nephelometers. Atmos Environ, 39: 3457-3469. 086397
Wu CF; Delfino RJ; Floro JN; Samimi BS; Quintana PJ; Kleinman MT; Liu LJ. (2005). Evaluation and quality control of
personal nephelometers in indoor, outdoor and personal environments. J Expo Sci Environ Epidemiol, 15: 99-110.
157155
Wu CF; Jimenez J; Claiborn C; Gould T; Simpson CD; Larson T; Liu LJS. (2006). Agricultural burning smoke in Eastern
Washington: Part II. Exposure assessment. Atmos Environ, 40: 5379-5392. 179950
Wu J; Houston D; Lurmann F; Ong P; Winer A. (2009). Exposure of PM2. 5 and EC from diesel and gasoline vehicles in
communities near the Ports of Los Angeles and Long Beach, California. Atmos Environ, 43: 1962-1971. 191773
Wu J; Lurmann F; Winer A; Lu R; Turco R;. (2005). Development of an individual exposure model for application to the
Southern California children's health study. Atmos Environ, 39: 259-273. 058570
Wu S; Mickley L; Jacob D; Rind D; Streets D. (2008). Effects of 2000-2050 changes in climate and emissions on global
tropospheric ozone and the policy-relevant background surface ozone in the United States. J Geophys Res, 113:
Dl 8312. 190039
Xia X; Hopke PK. (2006). Seasonal Variation of 2-Methyltetrols in Ambient Air Samples. Environ Sci Technol, 40: 6934-
6937. 179947
Xiaomin X; Zhen H; Jiasong W (2006). The impact of urban street layout on local atmospheric environment. Build
Environ, 41: 1352-1363. 156165
Yang H; Li Q; Yu JZ. (2003). Comparison of two methods for the determination of water-soluble organic carbon in
atmospheric particles. Atmos Environ, 37: 865-870. 156167
Yanosky JD; Paciorek CJ; Schwartz J; Laden F; Puett R; Suh HH. (2008). Spatio-temporal modeling of chronic PM10
exposure for the Nurses' Health Study. Atmos Environ, 42(18): 4047-4062. 099467
Yanosky JD; Paciorek CJ; SuhHH. (2009). Predicting chronic fine and coarse particulate exposures using spatiotemporal
models for the northeastern and midwestern United States. Environ Health Perspect, 117: 522-529. 190114
Yi SM; Ambs JL; Patashnick H; Rupprecht G; Hopke PK. (2004). Particle Collection Characteristics of a Prototype
Electrostatic Precipitator (ESP) for a Differential TEOM System. Aerosol Sci Technol, 38: 46-51. 156169
Yu H; Remer LA; Chin M; Bian H; Kleidman RG; Diehl T. (2008). A satellite-based assessment of transpacific transport of
pollution aerosol. J Geophy s Res, 113: 1-15. 157168
Yu JZ; Yang H; Zhang H; Lau AKH. (2004). Size distributions of water-soluble organic carbon in ambient aerosols and its
size-resolved thermal characteristics. Atmos Environ, 38: 1061-1071. 156172
Zanis P; Trickl T; Stohl A; Wernli H; Cooper O; Zerefos C; Gaeggeler H; Schnabel C; Tobler L; Kubik PW; Priller A;
Scheel HE; Kanter HJ; Cristofanelli P; Forster C; James P; Gerasopoulos E; Delcloo A; Papayannis A; Claude H.
(2003). Forecast, observation and modelling of a deep stratospheric intrusion event over Europe. Atmos Chem
Phys, 3:763-777. 053423
Zanobetti A; Wand MP; Schwartz J; Ryan LM. (2000). Generalized additive distributed lag models: quantifying mortality
displacement. Biostatistics, 1: 279-292. 004133
Zeger SL; Thomas D; Dominici F; Samet JM; Schwartz J; Dockery D; Cohen A. (2000). Exposure measurement error in
time-series studies of air pollution: concepts and consequences. Environ Health Perspect, 108: 419-426. 001949
December 2009 3-222
-------
Zeng Y. (2006). A comprehensive particulate matter monitoring system and dosimetry-based ambient participate matter
standards. J Air Waste Manag Assoc, 56: 518-29. 098375
Zhang K; Wexler A. (2008). Modeling urban and regional aerosols—Development of the UCD Aerosol Module and
implementation in CMAQ model. Atmos Environ, 42: 3166-3178 . 191770
Zhang Q; Jimenez JL; Canagaratna MR; Allan JD; Coe HL Ulbrich I; Alfarra MR; Takami A; Middlebrook AM; Sun YL;
Dzepina K; Dunlea E; Docherty K; DeCarlo PF; Salcedo D; Onasch T; Jayne JR; Miyoshi T; Shimono A;
Hatakeyama S; Takegawa N; Kondo Y; Schneider J; Drewnick F; Borrmann S; Weimer S; Demerjian K; Williams
P; Bower K; Bahreini R; Cottrell L; Griffin RJ; Rautiainen J; Sun JR; Zhang YM; Worsnop DR. (2007). Ubiquity
and dominance of oxygenated species in organic aerosols in anthropogenically-influenced Northern Hemisphere
midlatitudes. Geophys Res Lett, 34: LI3801.189998
Zhang Q; Jimenez JL; Canagaratna MR; Jayne JT; Worsnop DR. (2005). Time- and size-resolved chemical composition of
submicron particles in Pittsburgh: Implications for aerosol sources and processes. J Geophys Res, 110: 1-19.
157185
Zhang X; Zhuang G; Guo J; Yin K; Zhang P. (2007). Characterization of aerosol over the Northern South China Sea during
two cruises in 2003. Atmos Environ, 41: 7821-7836. 101119
Zhao W; Hopke PK; Norris G; Williams R; Paatero P. (2006). Source apportionment and analysis on ambient and personal
exposure samples with a combined receptor model and an adaptive blank estimation strategy. Atmos Environ, 40:
3788-3801. 156181
Zhao W; Rabinovitch N; Hopke PK; Gelfand EW (2007). Use of an expanded receptor model for personal exposure
analysis in schoolchildren with asthma. Atmos Environ, 41: 4084-4096. 156182
Zheng M; Cass GR; Schauer JJ; Edgerton ES. (2002). Source apportionment of PM2.5 in the southeastern United States
using solvent-extractable organic compounds as tracers. Environ Sci Technol, 36: 2361-2371. 026100
Zheng M; Ke L; Edgerton ES; Schauer JJ; Dong M; Russell AG (2006). Spatial distribution of carbonaceous aerosol in the
southeastern United States using molecular markers and carbon isotope data. J Geophys Res, 111: DlOS06. 157189
Zhou Y; Levy JI. (2007). Factors influencing the spatial extent of mobile source air pollution impacts: a meta-analysis.
BMC Public Health, 7: 89. 098633
Zhu K; Zhang J; Lioy PJ. (2007). Evaluation and comparison of continuous fine particulate matter monitors for
measurement of ambient aerosols. J Air Waste Manag Assoc, 57: 1499-506. 098367
Zhu Y; Eiguren-Fernandez A; Hinds WC; Miguel AH. (2007). In-Cabin Commuter Exposure to Ultrafine Particles on Los
Angeles Freeways. Environ Sci Technol, 41: 2138-2145. 179919
Zhu Y; Hinds WC; Kim S; Shen S; Sioutas C. (2002). Study of ultrafine particles near a major highway with heavy-duty
diesel traffic. Atmos Environ, 36: 4323-4335. 041553
Zhu Y; Hinds WC; Shen S; Sioutas C. (2004). Seasonal trends of concentration and size distribution of ultrafine particles
near major highways in Los Angeles. Aerosol Sci Technol, 38: 5-13. 156184
Zhu Y; Kuhn T; Mayo P; Hinds WC. (2005). Comparison of Daytime and Nighttime Concentration Profiles and Size
Distributions of Ultrafine Particles near a Major Highway. Environ Sci Technol, 39: 2531-2536. 157191
Zidek JV; Shaddick G; Meloche J; Chatfield C; White R. (2007). A framework for predicting personal exposures to
environmental hazards. Environ Ecol Stat, 14: 411-431. 190076
Zidek JV; Wong H; Le ND; Burnett R. (1996). Causality, measurement error and multicollinearity in epidemiology.
Environmetrics, 7: 441-451.051879
December 2009 3-223
-------
Chapter 4. Dos i me try
4.1. Introduction
Particle dosimetry refers to the characterization of deposition, translocation, clearance, and
retention of particles and their constituents within the respiratory tract and extrapulmonary tissues.
This chapter summarizes basic concepts presented in dosimetry chapters of the 1996 and 2004 PM
AQCDs (U.S. EPA, 1996, 079380: U.S. EPA, 2004, 056905). and updates the state of the science
based upon new literature appearing since publication of these PM AQCDS. Although the basic
understanding of the mechanisms governing deposition and clearance of inhaled particles has not
changed, there is significant additional information on the role of certain biological determinants
such as gender, age and lung disease on deposition and clearance. Additionally, new studies have
further characterized the retention and translocation of ultrafine particles (UFPs; also commonly
referred to as nanoparticles) following deposition in the respiratory tract.
The dose from inhaled particles deposited and retained in the respiratory tract is governed by a
number of factors. These include exposure concentration and duration, activity and ventilatory
parameters, and particle properties (e.g., particle size, hygroscopicity, and solubility in airway fluids
and cellular components). The basic characteristics of particles as they relate to deposition and
retention, as well as anatomical and physiological factors influencing particle deposition and
retention, were discussed in depth in Chapter 10 of 1996 PM AQCD and updated in Chapter 6 of the
2004 PM AQCD. Species differences between humans and rats in particle exposures, deposition
patterns, and pulmonary retention were also reviewed in Brown et al. (2005, 089308). The current
review of PM dosimetry focuses mainly on issues that may affect the susceptibility of an individual
to adverse effects as well as issues that affect our ability to extrapolate findings between studies
(e.g., in vitro to in vivo) and between species. Other than a brief overview in this introductory
section, the disposition (i.e., deposition, absorption, distribution, metabolism, and elimination) of
fibers and unique nano-objects (viz., dots, hollow spheres, rods, fibers, tubes) is not reviewed herein.
Substantial exposures to fibers and unique nano-objects generally occur in the occupational settings
rather than the ambient environment.
The deposition by interception of micro-sized fibers was briefly discussed in the 1996 and
2004 PM AQCD, but fiber retention in the respiratory tract was not addressed. Airborne fibers
(length/diameter ratio > 3), can exceed 150 um in length and appear to be relatively stable in air.
This is because their aerodynamic size is determined predominantly by their diameter, not their
length. Fibers longer than 10 um can deposit by interception and when aligned with the direction of
airflow may penetrate deep into the respiratory tract. Once deposited, macrophage mediated
clearance is the primary mechanism of removing micro-sized particles from the pulmonary region.
The length of fibers can, however, affect their phagocytosis and clearance. For example, fibers of
>17 um in length are too long to be fully engulfed by rat alveolar macrophages and can protrude
from macrophages (i.e., macrophage frustration) (Zeidler-Erdely et al., 2006, 190967). Further
discussion of the fiber disposition in the respiratory tract is beyond the scope of this chapter.
The term "ultrafine particle" has traditionally been used by the aerosol research and
occupational and environmental health communities to describe airborne particles or other laboratory
generated aerosols used in toxicological studies that are <100 nm in size (based on physical size,
diffusivity, or electrical mobility). Generally consistent with the definition of an UFP, the
International Organization for Standardization (ISO) recently defined a nanoparticle as an object
with all 3 external dimensions in the nanoscale, i.e., from approximately 1 and 100 nm (ISO, 2008,
190066). The ISO also defined a nano-object as a material with one or more external dimensions in
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009 4-1
-------
the nanoscale. The terms, nanoparticle and UFP, have been used rather synonymously in the recent
literature. However, the terms nanoparticle and nano-object are more commonly associated with
engineered materials that are created for consumer products and industrial applications. With the
current interest in nanotechnologies, many nano-objects have been created by manipulating materials
at the atomic or molecular scale for the purpose of forming new materials, structures, and devices
that exploit the unique physical and chemical properties associated with their nanoscale.
Toxicological studies are becoming available that evaluate in vivo translocation and heath effects
unique of nano-obj ects (viz., dots, hollow spheres, rods, tubes). The in vivo disposition of these
unique nano-objects is not, however, necessarily relevant to the behavior of UF aerosols in the urban
environment that are created by combustion sources and photochemical formation of secondary
organic aerosols. Therefore, the disposition of unique nano-objects (viz., dots, hollow spheres, rods,
fibers, tubes) is not considered in this chapter.
4.1.1. Size Characterization of Inhaled Particles
Particle size is a major determinant of the fraction of inhaled particles depositing in and
cleared from various regions of the respiratory tract. The distribution of particle sizes in an aerosol is
typically described by the lognormal distribution (i.e., the situation in which the logarithms of
particle diameter are distributed normally). The geometric mean is the median of the distribution,
and the variability around the median is the geometric standard deviation (GSD or og) and is given
by:
GSD = o- = ii% =^o%
"50% "16%
Equation 4-1
where: di6%, d50%, d84% are the particle diameters associated with the 16th, 50th (i.e., the median), and
the 84th percentiles from the cumulative frequency distribution of particle sizes. By definition, GSD
must be greater than one. The particle size associated with any percentile of the distribution, di, is
given by:
d = ^o% ff*(P)
Equation 4-2
where: z(P) is the normal standard deviate for a given probability. In most cases, the aerosols to
which people are naturally exposed are poly disperse. By contrast, most experimental studies of
particle deposition and clearance in the lung use monodisperse particles (GSD <1.15). Ambient
aerosols may also be composed of multiple size modes, each mode should be described by its
specific median diameter and GSD.
Aerosol size distributions may be measured and described in various ways. When a
distribution is described by counting particles, the median is called the count median diameter
(CMD). On the other hand, the median of a distribution based on particle mass in an aerosol is the
mass median diameter (MMD). Impaction and sedimentation of particles in the respiratory tract
depend on a particle's aerodynamic diameter (dae), which is the size of a sphere of unit density that
has the same terminal settling velocity as the particle of interest. The size distribution is frequently
described in terms of dae as the mass median aerodynamic diameter (MMAD), which is the median
of the distribution of mass with respect to aerodynamic equivalent diameter. Alternative descriptions
should be used for particles with actual physical sizes below ~ 0.5 urn because, for these sized
particles, aerodynamic properties become less important and diffusion becomes ever more important.
For these smaller particles, their physical diameter or CMD are typically used since diffusivity is not
a function of particle density. For small irregular shaped particles and aggregates, the diameter of a
spherical particle that has the same diffusion coefficient in air as the particle in question is
appropriate.
December 2009 4-2
-------
4.1.2. Structure of the Respiratory Tract
The basic structure of the human respiratory tract is illustrated in Figure 4-1. In the literature,
the terms extrathoracic (ET) region and upper airways are used synonymously. The term lower
airways is used to refer to the intrathoracic airways, i.e., the combination of the tracheobronchial
(TB) region which is the conducting airways and the alveolar region which is the functional part or
parenchyma of the lung. A recent review of interspecies similarities and differences in the structure
and function of the respiratory tract is provided by Phalen et al. (2008, 156865). Although the
structure varies, the illustrated anatomic regions are common to all mammalian species with the
exception of the respiratory bronchioles. Respiratory bronchioles, the transition region between
ciliated and fully alveolated airways, are found in humans, dogs, ferrets, cats, and monkeys.
Respiratory bronchioles are absent in rats and mice and abbreviated in hamsters, guinea pigs, oxen,
sheep, and pigs. The branching structure of the ciliated bronchi and bronchioles also differs between
species from being a rather symmetric and dichotomous branching network of airways in humans to
a more monopodial branching network in other mammals.
Extrathoracic
Region
Posterior
Nasal Passage
fNasal Part
Pharynx-^ -. .....
' [ Oral Part
Trachea
Tracheobronchial
Region
Main Bronchi
Bronchi
Bronchioles
Bronchiolar Region
Alveolar Interstitial
Alveolar
Region
Bronchioles
Terminal Bronchioles
Respiratory Bronchioles
Alveolar Duct +
Alveoli
Source: Based on ICRP (1994,;
Figure 4-1.
Diagrammatic representation of respiratory tract regions in humans. Structures
are anterior nasal passages, ETi; oral airway and posterior nasal passages, ET2;
bronchial airways, BB; bronchioles, bb; and alveolar interstitial, Al.
Another species difference relevant to particle dosimetry is the route of breathing. For
instance, rodents are obligate nose breathers, whereas most humans are oronasal breathers who
breathe through the nose when at rest and increasingly through the mouth with increasing activity
December 2009
4-3
-------
level. There is inter-individual variability in the route by which people breathe. Most people, 87%
(26 of 30) in the Niinimaa et al. (1981, 071758) study, breathed through their nose until an activity
level was reached where they switched to oronasal breathing. Thirteen percent (4 of 30) of the
subjects, however, were oronasal breathers even at rest. These two subject groups are commonly
referred to in the literature (e.g. ICRP, 1994, 006988) as "normal augmenters" and "mouth
breathers," respectively. In contrast to healthy subjects, Chadha et al. (1987, 037365) found that the
majority (11 of 12) of patients with asthma or allergic rhinitis breathe oronasally even at rest.
a.
Bronchus
Bronchiolus
Aveolus
b.
Air
Liquid
Tissue
Air
Tr 11 n
Liquid
Tissue
Air
Air
Source: Panel (a) reprinted with permission from McGraw Hill (Fishman and Elias, 1980,156436)
Figure 4-2. Structure of lower airways with progression from the large airways to the
alveolus. Panel (a) illustrates basic airway anatomy. Structures are epithelial
cells, EP; basement membrane, BM; smooth muscle cells, SM; and
fibrocartilaginous coat, FC. Panel (b) illustrates the relative amounts of liquid,
tissue, and blood with distal progression.
December 2009
4-4
-------
The site of particle deposition within the respiratory tract has implications related to lung
retention and surface dose of particles as well as potential systemic distribution of particles or their
constituents. Figure 4-2 illustrates the progressive change in airway anatomy with distal progression
into the lower respiratory tract. In the bronchi there is a thick liquid lining and mucociliary clearance
rapidly moves deposited particles toward the mouth. In general, in the bronchi, only highly soluble
materials moving from the air into the liquid layer will have systemic access via the blood. With
distal progression, the protective liquid lining diminishes and clearance rates slow. Soluble
compounds and some poorly soluble UFPs may cross the air-liquid interface to enter the tissues and
the blood especially in the alveolar region.
4.2. Particle Deposition
Inhaled particles may be either exhaled or deposited in the ET, TB, or alveolar region. A
particle becomes deposited when it moves from the airway lumen to the wall of an airway. The
deposition of particles in the respiratory tract depends primarily on inhaled particle size, route of
breathing (nasal or oronasal), tidal volume (VT), breathing frequency (f), and respiratory tract
morphology. The distinction between air passing through the nose versus the mouth is important
since the nasal passages more effectively remove inhaled particles than the oral passage. Respiratory
tract morphology, which affects particle transport and deposition, varies between species, the size of
an animal or human, and health status.
The fraction of inhaled aerosol becoming deposited in the human respiratory tract has been
measured experimentally. Studies, using light scattering or particle counting techniques to quantify
the amount of aerosol in inspired and expired breaths, have characterized total particle deposition for
varied breathing conditions and particle sizes. The vast majority of in vivo data on the regional
particle deposition has been obtained by scintigraphic methods where external monitors are used to
measure gamma emissions from radiolabeled particles. These scintigraphic data have shown highly
variable regional deposition with sites of highly localized deposition or "hot spots" in the obstructed
lung relative to the healthy lung. Even in the healthy lung, "hot spots" occur in the region of airway
bifurcations. Mathematical models aid in predicting the mixed effects of particle size, breathing
conditions, and lung volume on total and regional deposition. Experimentally, however, there is
considerable inter-individual variability in total and regional deposition even when inhaled particle
size and breathing conditions are strictly controlled. Section 4.2.4 on Biological Factors Modulating
Deposition provides more detailed information on factors affecting deposition among individuals.
In order to potentially become deposited in the respiratory tract, particles must first be inhaled.
The inspirable particulate mass fraction of an aerosol is that fraction of the ambient airborne particles
that can enter the uppermost respiratory tract compartment, i.e., the head (Soderholm, 1985,
156992). The American Conference of Governmental Industrial Hygienists (ACGIH) and the
International Commission on Radiological Protection (ICRP) have established inhalability criteria
for humans (ACGIH, 2005, 156188: ICRP, 1994, 006988). These criteria are indifferent to route of
breathing and assume random orientation with respect to wind direction. They are based on
experimental inhalability data for dae < 100 um at wind speeds of between 1 and 8 m/s. For the
ACGIH criterion, inhalability is 97% for an dae = 1 um, 87% for an dae = 5 um, 77% for an
dae = 10 um, and plateaus at 50% for dae above -40 um. The ICRP criterion, which also plateaus at
50% for very large dae, does not become of real importance until an dae = 5 um where inhalability is
97%. Dai et al. (2006, 156377) reported slightly lower nasal particle inhalability in humans during
moderate exercise than rest (e.g., 89.2 versus 98.1% for 13 um particles, respectively). Nasal particle
inhalability is similar between an adult and 7-year-old child (Hsu and Swift, 1999, 155855).
Inhalability into the mouth from calm air in humans also becomes important for dae >10 um
(Anthony and Flynn, 2006, 155659; Brown, 2005, 156299). Unlike the inhalability from high wind
speeds which plateaus at 50% for dae greater than ~40 um, particle inhalability from calm air
continues to decrease toward zero with increasing dae.
Inhalability data in laboratory animals, such as rats, are only available for breathing from
relatively calm air (velocity < 0.3 rn/s). For nasal breathing, inhalability becomes an important
consideration for dae of above 1 um in rodents and 10 um in humans (Menache et al., 1995, 006533).
The inhalability of particles having dae of 2.5, 5, and 10 um is 80, 65, and 44% in rats, respectively,
whereas it only decreases to 96% for an dae of 10 um in humans during nasal breathing (Menache et
December 2009 4-5
-------
al., 1995, 006533). Asgharian et al. (2003, 153068) suggested that an even more rapid decrease in
inhalability with increasing dae may occur in rats. Inhalability is a particularly important
consideration for rodent exposures. Section 4.2.3 provides additional discussion of interspecies
patterns of particle deposition.
4.2.1. Mechanisms of Deposition
Particle deposition in the lung is predominantly governed by diffusion, impaction, and
sedimentation. Most discussion herein focuses on these three dominant mechanisms of deposition.
Simple interception, which is an important mechanism of fiber deposition, is not discussed in this
chapter. Electrostatic and thermophoretic forces as mechanisms of deposition have not been
thoroughly evaluated and receive limited discussion. Some generalizations with regard to deposition
by these mechanisms follows, but should not be viewed as absolute rules. Both experimental studies
and mathematical models have demonstrated that breathing patterns can dramatically alter regional
and total deposition for all sized particles. The combined processes of aerodynamic and diffusive (or
thermodynamic) deposition are important for particles in the range of 0.1 um to 1 um. Aerodynamic
processes predominate above and thermodynamic processes predominate below this range.
Diffusive deposition, by the process of Brownian diffusion, is the primary mechanism of
deposition for particles having physical diameters of less than 0.1 um. For particles having physical
diameters of roughly between 0.05 and 0.1 urn, diffusive deposition occurs mainly in the small distal
bronchioles and the pulmonary region of the lung. However, with further decreases in particle
diameter below -0.05 um, increases in particle diffusivity shift more deposition proximally to the
bronchi and ET regions.
Governed by inertial or aerodynamic properties, impaction and sedimentation increase with
dae. When a particle has sufficient inertia, it is unable to follow changes in flow direction and strikes
a surface thus depositing by the process of impaction. Impaction occurs predominantly at
bifurcations in the proximal airways, where linear velocities and secondary eddies are at their
highest. Sedimentation, caused by the gravitational settling of a particle, is most important in the
distal airways and pulmonary region of the lung. In these regions, residence time is the greatest and
the distances that a particle must travel to reach the wall of an airway are minimal.
The electrical charge on some particles may result in an enhanced deposition over what would
be expected based on size alone. With an estimated charge of 10-50 negative ions per 0.5 um
particle, Scheuch et al. (1990, 006948) found deposition in humans (VT = 500 mL, f = 15 min"1) to
increase from 13.4% (no charge) to 17.8% (charged). This increase in deposition is thought to result
from image charges induced on the surface of the airway by charged particles. Yu (1985, 006963)
estimated a charge threshold level above which deposition fractions would be increased of about 12,
30, and 54% for 0.3, 0.6, and 1.0 um diameter particles, respectively. Electrostatic deposition is
generally considered negligible for particles below 0.01 um because so few of these particles carry a
charge at Boltzmann equilibrium. This mechanism is also thought to be a minor contributor to
overall particle deposition, but it may be important in some laboratory studies due to specific aerosol
generation techniques such as nebulization. Laboratory methods such as passage of aerosols through
a Kr-85 charge neutralizer prior to inhalation are commonly used to mitigate this effect.
The National Radiological Protection Board (NRPB) recently evaluated the potential for
corona discharges from high voltage power lines to charge particles and enhance particulate doses
(NRPB, 2004, 156815). They concluded that electrostatic effects would be the most important for
particles in the size range from about 0.1-1 um, where deposition may theoretically increase by a
factor of three to ten. However, given that only a small fraction of ambient particles would pass
through the corona to become charged, the small range of relevant particle sizes (0.1-1 um), and the
subsequent required transport of charged particles to expose individuals; the NRPB concluded that
effects, if any, of electric fields on particle deposition in the human respiratory tract would likely be
minimal.
Thermophoretic forces on particles occur due to temperature differences between respired air
and respiratory tract surfaces. Temperature gradients of around 20°C are thought to produce
sufficient thermophoretic force to oppose diffusive and electrostatic deposition during inspiration
and to perhaps augment deposition by these mechanisms during expiration (Jeffers, 2005, 156608).
Thermophoresis is only relevant in the extrathoracic and large bronchi airways and reduces to zero
as the temperature gradient decreases deeper in the lung. Theoretical analysis of thermophoresis has
been done for smooth walled tubes and is important over distances that are several orders of
December 2009 4-6
-------
magnitude smaller than the diameter of the trachea. The alteration of the flow patterns by airway
surface features such as cartilaginous rings may affect particle transport and deposition over far
greater distances than thermophoretic force.
4.2.2. Deposition Patterns
Knowledge of sites where particles of different sizes deposit in the respiratory tract and the
amount of deposition therein is necessary for understanding and interpreting the health effects
associated with exposure to particles. Particles deposited in the various respiratory tract regions are
subjected to large differences in clearance mechanisms and pathways and, consequently, retention
times. Deposition patterns in the human respiratory tract were described in considerable detail in
dosimetry chapters of prior PM AQCD (U.S. EPA, 1996, 079380: U.S. EPA, 2004, 056905): as such,
they are only briefly described here.
Predicted total and regional deposition for an adult male during rest and light exercise are
illustrated in Figure 4-3 and Figure 4-4, respectively. Note that a large proportion of inhaled coarse
particles in the 3-6 um (dae) range can reach and deposit in the lower respiratory tract, particularly
the TB airways. Although these figures were provided in Chapter 6 of the 2004 PM AQCD, they are
reproduced here to illustrate changes in deposition as a function of particle size and breathing
conditions. The predictions were based on two publicly available particle deposition models, the
ICRP (1994, 006988) and the Multi-Path Particle Dosimetry model (MPPD; Version 1.0, ©2002).
The ICRP (1994, 006988) model was implemented by Lung Dose Evaluation Program (LUDEP;
Version 2.07, June 2000). The MPPD1 model was developed by the CUT Centers for Health
Research with support from the Dutch National Institute of Public Health and the Environment.
0.1 1
Diameter, |jm
0.1 1
Diameter, |jm
Figure 4-3. Comparison of total and regional deposition results from the ICRP and MPPD
models for a resting breathing pattern (VT = 625 ml, f = 12 min"1) and corrected
for particle inhalability. Regions are extrathoracic, ET; tracheobronchial, TB; and
alveolar, A. Panels a-b are for nose breathing; panels c-d are for mouth breathing.
1 For more information about this model, the reader is referred to: http://www.ara.com/products/mppd capabilities.htm.
December 2009
4-7
-------
1.0
Diameter,
Diameter, \im
Figure 4-4. Comparison of total and regional deposition results from the ICRP and MPPD
models for a light exercise breathing pattern (VT = 1250 ml, f = 20 min"1) and
corrected for particle inhalability. Regions are extrathoracic, ET;
tracheobronchial, TB; and alveolar, A. Panels a-b are for nose breathing; panels
c-d are for mouth breathing.
4.2.2.1. Total Respiratory Tract Deposition
The efficiency of deposition in the respiratory tract may generally be described as a
"U-shaped" curve on a plot of deposition efficiency versus the log of particle diameter. Total
deposition shows a minimum for particle diameters in the range of 0.1 to 1.0 um, where particles are
small enough to have minimal sedimentation or impaction and sufficiently large so as to have
minimal diffusive deposition. Total deposition does not decrease to zero for any sized particle, in
part, because of mixing between particle laden tidal air and residual lung air. The particles mixed
into residual air remain in the lung following a breath and are removed on subsequent breaths or
gradually deposited. Total deposition approaches 100% for particles of roughly 0.01 um (physical
diameter) due to diffusive deposition and for particles of around 10 um (dae) due to the efficiency of
sedimentation and impaction.
Total human lung deposition, as a function of particle size, is depicted in Figure 4-5. These
experimental data were obtained by using monodisperse spherical test particles in healthy adults
during controlled breathing on a mouthpiece. Despite the control of inhaled particle size and
breathing conditions, this figure illustrates variability in deposition efficiencies due to inter-
individual differences in lung size and anatomical variability in airway dimensions and branching
patterns.
December 2009
4-8
-------
c
o
c
o
o
Q.
0.04
0.06
0.08 0.10 1 3
Particle Diameter (|im)
* = p < 0.05
Male
Female
Vt = 500 ml
Q = 250 mUs
Source: Data from Kim and Hu (1998, 0860661 and Kim and Jaques (2000, 0128111.
Figure 4-5.
Total lung deposition measured in healthy adults (UF, 11 M, 11 F, 31 ± 4 yr; fine
and coarse, 11 M, 11 F, 25 ± 4 yr) during controlled breathing on a mouthpiece.
Deposition calculated from aerosol bolus measurements between 50 and 500 ml
into a breath with 50 ml increments. Illustrated data are means and standard
errors. Asterisk indicates significantly greater total deposition in females versus
males.
4.2.2.2. Extrathoracic Region
The first line of defense for protecting the lower respiratory tract from inhaled particles is the
nose and mouth. Particle deposition in the ET region, especially the nasal passages, reduces the
amount available for deposition in the TB and alveolar regions. Recent data have become available,
but are largely derived from computational fluid dynamics (CFD) modeling and experimental
measurements in casts. As most of these studies do not substantially improve our understanding of
deposition in the ET region they are not reviewed here.
For particles >1 um dae, deposition efficiency in the oral and nasal passages has been generally
described as a function of an impaction parameter (Stokes number) with the addition of a flow
regime parameter (Reynolds number) for the oral passages (Finlay and Martin, 2008, 155776; Grgic
et al, 2004, 155810; Kelly et al, 2005, 155894; Schroeter et al., 2006, 156076). For an adult male,
the CFD simulations of Schroeter et al. (2006, 156076) predicted nasal deposition of 10 um dae
particles was 90%, and 100% for a VE of 7.5 L/min (rest) and 15 L/min (light activity), respectively.
Thus, relatively few large coarse particles will pass through the nasal passages into the lungs. Since
the nasal passages are more efficient at removing inhaled particles than the oral passage, an
individual's mode of breathing (i.e., oral versus nasal) influences the quantity of particles penetrating
to the lung.
In limited studies, it has been shown that children tend to have more oral breathing both at rest
and during exercise and also displayed more variability than adults (Becquemin et al., 1999, 155679;
Bennett et al., 2008, 156269; James et al., 1997, 042422). In contrast to adults, there is little data on
the uptake of particles for oral or nasal breathing in children. Theoretical calculations by Xu and Yu
(1986, 072697) predict enhanced deposition of particles (>2 um) in the head region for children
when compared to adults. Studies of fine particle deposition in physical models of the nose, scaled to
adult versus children sizes, predict that deposition efficiency in the nose is a function of pressure
drop across the nose (Phalen et al., 1989, 156023). Consequently, these model analyses suggest that,
when properly scaled physiological flows are used in the calculation of nasal deposition, children,
who have higher nasal resistance than adults, should have higher nasal deposition compared to
adults. Surprisingly, the few studies reporting measures of nasal deposition in children, found lower
nasal deposition efficiencies for fine particles (1-3 um dae) as compared to adults, despite their higher
nasal resistances (Becquemin et al., 1991, 009187; Bennett et al., 2008, 156269). These findings of
lesser nasal versus oral breathing and less efficient nasal deposition suggest that children's lower
respiratory tract (i.e., the TB and alveolar regions) may receive a higher dose of ambient PM
December 2009
4-9
-------
compared to adults. Normalized to lung surface area, the dose rate to the lower airways of children
versus adults is increased further because children breathe at higher minute ventilations relative to
their lung volumes (see Section 4.2.4.2 on age as a factor modulating deposition).
4.2.2.3. Tracheobronchial and Alveolar Region
Inhaled particles passing the ET region enter and may become deposited in the lungs. For any
given particle size, the pattern of particle deposition influences clearance by partitioning deposited
material among lung regions. Deposition in the tracheobronchial airways and alveolar region cannot
be directly measured in vivo. Much of the available deposition data for the TB and alveolar regions
have been obtained from experiments with radioactively labeled, poorly soluble particles (U.S. EPA,
1996, 079380) or by use of aerosol bolus techniques (U.S. EPA, 2004, 056905). In general, the
ability of these experimental data to define specific sites of particle deposition is limited to
anatomically large regions of the respiratory tract such as the head, larynx, bronchi, bronchioles, and
alveolar region. Mathematical modeling can provide more refined predictions of deposition sites.
Comparisons of the modeling results obtained with two publicly available models were provided in
Figure 4-3 and Figure 4-4. Highly localized sites of deposition within the bronchi are described in
Section 4.2.2.4. Both experimental and modeling techniques are based on many assumptions that
may be relatively good for the healthy lung but not for the diseased lung. For discussion of these
issues, the reader is referred to Sections 4.2.4.4 and 4.2.4.5.
4.2.2.4. Localized Deposition Sites
From a toxicological perspective, it is important to realize that not all epithelial cells in an
airway will receive the same dose of deposited particles. Localized deposition in the vicinity of
airway bifurcations has been analyzed using experimental and mathematical modeling techniques. In
the 1996 PM AQCD, experimental data were available illustrating the peak deposition of coarse
particles (3, 5, and 7 um dae) in daughter airways during inspiration and the parent airway during
expiration, but always near the carinal ridge (Kim and Iglesias, 1989, 078539; Kim et al., 1989,
078538). In the 2004 PM AQCD, mathematical models predicted distinct "hot spots" of deposition
in the vicinity of the carinal ridge for both coarse (10 um) and UF (0.01 um) particles (Heistracher
and Hofmann, 1997, 047514; Hofmann et al., 1996, 047515). In a model of lung generations 4-5
during inspiration, hot spots occurred at the carinal ridge for 10 um dae particles due to inertial
impaction and for 0.01 um particles due to secondary flow patterns formed at the bifurcation. During
expiration, preferential sites of deposition for both particle sizes occurred 1) approaching the
juncture of daughter airways on the walls forming and across the lumen from the carinal ridge; and
2) the top and bottom (visualizing the Y-shaped geometry laying horizontal) of the parent airway
downstream of the bifurcation.
Recent studies further support these findings (Balashazy et al., 2003, 155671; Farkas and
Balashazy, 2008, 157358; Farkas et al., 2006, 155771; Isaacs et al., 2006, 155861). Most of these
studies quantified localized deposition in terms of an enhancement factor. Typically, the
enhancement factor is the ratio of the deposition in a pre-specified surface area (e.g., 100 x 100 um
which corresponds to -10 x 10 epithelial cells) to the average deposition density for the whole
airway geometry. These enhancement factors are very sensitive to the size of the surface considered
(Balashazy et al., 1999, 043201). The studies by Farkas et al. (2006, 155771) and Farkas and
Balashazy (2008, 157358) investigated the phenomena of localized deposition down to 0.001 um
particles. The deposition of 0.001 um was rather uniform, however, the deposition pattern became
increasingly less uniform with increasing particle size. These studies indicate that, for particles
greater than -0.01 um, some cells located near the carinal ridge of bronchial bifurcations may
receive hundreds to thousands times the average dose (particles per unit surface area) of the parent
and daughter airways. Furthermore, the inertial impaction of particles > 1 um dae at the carinal ridge
of large bronchi will increase with increasing inspiratory flows. In a comparison of constricted
versus healthy airways, Farkas et al. (2006, 155771) also reported that the overall deposition
efficiency of 10 um dae particles at bifurcations downstream of a constriction may be increased by
18 times. Given these considerations, Phalen and Oldham (2006, 156024) noted that substantial
doses of particles (> 1 um dae) may be justified for in vitro studies using tracheobronchial epithelial
cell cultures.
December 2009 4-10
-------
4.2.3. Interspecies Patterns of Deposition
The primary purpose of this document is to assess the health effects of particles in humans. As
such, human dosimetry studies have been stressed in this chapter. Such studies avoid the
uncertainties associated with the extrapolation of dosimetric data from laboratory animals to humans.
However, animal models have been and continue to be used in evaluating PM health effects because
of ethical considerations regarding the types of studies that can be performed with human subjects.
Thus, there is a considerable need to understand dosimetry in animals and dosimetric differences
between animals and humans. Limited new data are becoming available. Similar deposition
efficiencies have been reported in nasal casts of human and rhesus monkey for 1-10 um dae for
inspiratory flows mimicking resting breathing patterns (Kelly et al, 2005, 155894). Oldham and
Robinson (2007, 156003) recently provided morphological data and predicted particle deposition in
an asthma mouse model.
Interspecies similarities and differences in deposition were described in detail in the last two
PM AQCDs (U.S. EPA, 1996, 079380; U.S. EPA, 2004, 056905). It was concluded that the general
pattern of total particle deposition efficiency was similar between laboratory animals and humans:
deposition increases on both sides of a minimum that occurs for particles of 0.2-1 um. There are,
however, marked interspecies differences in uptake into the respiratory tract and regional deposition.
For instance, the nasal inhalability of 10 urn dae particles is predicted to be 96% in humans, whereas
it is only 44% in rats (Menache et al., 1995, 006533). In most laboratory animal species (rat, mouse,
hamster, guinea pig, and dogs), deposition in the ET region is near 100% for particles >5 um dae
(Raabe et al., 1988, 001439). indicating greater efficiency than that seen in humans. Detailed
presentation of dosimetric difference between rats and humans are available elsewhere (Brown et al.,
2005, 089308: Jarabek et al., 2005, 056756).
Brown et al. (2005, 089308) conducted a thorough evaluation of extrapolations between rats
and humans in relation to PM exposures. One of many factors they considered was the choice of a
dose metric appropriate for comparison between species. For example, deposited mass may be an
appropriate PM indicator for health effects associated with soluble PM constituents. For health
effects associated with insoluble PM, the particle number, surface area, or mass may be appropriate
indicators. Given interspecies differences in deposition patterns and clearance rates, the question of
retained versus deposited dose was also discussed. It was concluded that for acute effects, the
incremental dose may be the appropriate type of dose metric. For chronic effects, long-term burden
may be more appropriate. For various dose metrics, estimates of particle concentration and exposure
duration required for a rat to receive the same dose as received by a human were obtained with
consideration of activity levels, mode of breathing, and particle size distributions. It was noted that
high PM exposures over the period of months can lead to particle overload in rats (see
Section 4.3.4.4). Exposure regimes were derived as a function of particle size and exposure duration
that should avoid overwhelming macrophage mediated clearance achieving particle overload in rats
(see Table 12 in Brown et al., 2005, 089308). The dosimetric calculations indicated that to achieve
nominally similar acute doses per surface area in rats, relative to humans undergoing moderate to
high exertion, PM exposure concentrations for rats would need to be somewhat higher than for
humans. Since particle clearance from the lungs of rats is faster than humans, much higher exposure
concentrations are required for the rat to simulate retained burdens of humans. Illustrating the
complexity of such analyses, in some cases, rats were found to require lower exposures than humans
to have comparable doses (generally when considering a scenario of humans at rest).
4.2.4. Biological Factors Modulating Deposition
Evaluation of factors affecting particle deposition is important to help understand potentially
susceptible subpopulations. Differences in biological response following pollutant exposure may be
caused by dosimetry differences as well as by differences in innate sensitivity. The effects of
different biological factors on deposition were discussed in the 2004 PM AQCD (U.S. EPA, 2004,
056905) and are summarized briefly here.
December 2009 4-11
-------
4.2.4.1. Physical Activity
The activity level of an individual is well recognized to affect their minute ventilation and
route of breathing. Changes in minute ventilation during exercise are accomplished by increasing
both VT and f (Table 4-1). Humans are oronasal breathers tending to breathe through the nose when
at rest and increasingly through the mouth with increasing activity level. There is considerable inter-
individual variability in both the route by which people breathe and the way breathing pattern
changes occur.
Table 4-1. Breathing patterns with activity level in adult human male.
Activity
Breaths/min
Tidal volume, mL
Minute ventilation, L/min
Awake
Rest3
12
625
7.5
Slow
Walk3
16
813
13
Light
Exertion
19
1000
19
Moderate
Exertion 3
28
1429
40
Heavy
Exertion
26
1923
50
Sources: 'Winter-Sorkina and Cassee (2002, 043670): bICRP (1994, 006988)
Individuals typically breathe through their nose while at rest, switching to oronasal breathing
as ventilation increases (Bennett et al., 2003, 191977; Niinimaa et al, 1981, 071758). The role of the
nose in filtering particles is diminished as airflow is diverted from the nose to the mouth during
exercise, bringing more particles to the lower respiratory tract. A recent study in adults (Bennett et
al., 2003, 191977) found that nasal ventilation during exercise varied as a function of both race and
gender. African-Americans possessed a greater nasal contribution to breathing during exercise than
Caucasians. At similar exercise efforts (i.e., normalized to a % maximum work capacity) the females
also had a greater nasal contribution to breathing during exercise than males.
In addition, when individuals increase their ventilation with activity the total number of
particles inhaled per unit time (i.e., exposure rate) increases, but the fractional deposition of particles
in each breath also changes with breathing pattern. Figure 4-3 and Figure 4-4 illustrate predicted
deposition fractions in the respiratory tract during rest versus light exercise, respectively. During
exercise, both VT and f increase. Fractional deposition for all particles increases with increased VT.
Increasing the f, however, decreases the fractional deposition of fine and UFPs due to decreased time
for gravitational and diffusive deposition. For particles larger than a dae of roughly 3 um, increasing f
can increase the deposition fraction due to increased impaction in the extrathoracic and TB airways.
Thus, it should be expected that the change in deposition fraction with activity will vary among
individuals depending on the relative influences of these two variables (i.e., VT and f) in a given
subject and the particle size to which they are exposed. Experimentally, the lung deposition fractions
of fine particles during moderate exercise and mouth breathing are unchanged between rest and
exercise (Bennett et al., 1985, 190034: Morgan et al., 1984, 190035). Kim (2000, 013112) evaluated
differences in deposition of 1, 3, and 5 um (MMAD) particles under varying breathing patterns
(simulating breathing conditions of sleep, resting, and mild exercise). Total lung deposition increased
with increasing VT at a given flow rate and with increasing flow rate at a given breathing period.
These experimental studies suggest that the total deposited dose rate (i.e., deposition per unit time)
of particles will generally increase in direct proportion to the increase in minute ventilation
associated with exercise.
The changes in ventilation, i.e., breathing pattern and flow rate, may also alter the regional
deposition of particles. Coarse particle deposition increases in the TB and ET regions during exercise
due to the increased flow rates and associated impaction. A rapid-shallow breathing pattern during
exercise may result in more bronchial airway versus alveolar deposition, while a slow-deep pattern
will shift deposition to deeper lung regions (Valberg et al., 1982, 190019). Bennett et al. (1985,
190034) showed for 2.6 urn particles that moderate exercise shifted deposition from the lung
periphery towards ET and larger, bronchial airways. Similarly, Morgan et al. (1984, 190035) showed
that even for fine particles (0.7 um) TB deposition was enhanced with exercise. This shift in
deposition toward the bronchial airways results in a much greater dose per unit surface area of tissue
in those regions. Morgan et al. (1984, 190035) also found that the apical-to-basal distribution of fine
December 2009 4-12
-------
particles increased with exercise, i.e., a shift towards increased deposition in the lung apices. This
shift may be less likely for larger particles, however, whose deposition in large airway bifurcations
may preclude their transport to these more apical regions (Bennett et al., 1985, 190034).
4.2.4.2. Age
Airway structure and respiratory conditions vary with age, and these variations may alter the
amount and site of particle deposition in the respiratory tract. It was concluded in the 2004 PM
AQCD (U.S. EPA, 2004, 056905) that significant differences between adults and children had been
predicted by mathematical models and observed in experimental studies. Studies generally indicated
that ET and TB deposition was greater in children and that children received greater doses of
particles per lung surface area than adults. Deposition studies in the elderly are still quite limited.
A few studies have attempted to measure oronasal breathing in children as compared to adults
(Becquemin et al., 1999, 155679; Bennett et al., 2008, 156269; James et al., 1997, 042422). This is
important since particles deposit with greater efficiency in the nose relative to the mouth, thereby
affecting exposure of the lower respiratory tract. James et al. (1997, 042422) found that children (age
7-16 yr, n = 10) displayed more variability than adults with respect to their oronasal pattern of
breathing with exercise. However, it was not possible to predict the pattern of the partitioning of
ventilation during exercise based on age, gender, or nasal airway resistance. Further, in a limited
number of children (age 8-16 yr, n = 10), Becquemin et al. (1999, 155679) found that the children
tended to display more oral breathing both at rest and during exercise than the adults. The highest
oral fractions were also found in the youngest children. None of these studies, however, was able to
show a relationship between nasal resistance and the relative contribution of nasal breathing in
children. Bennett et al. (2008, 156269) made preliminary measurements of the relative contributions
of oral versus nasal breathing at rest and during incrementally graded submaximal exercise on the
cycle ergometer for children (age 6-10 yr, n = 12) and adults (age 18-27 yr, n = 11). There was a
trend for children to have a lesser nasal contribution to breathing at rest and during exercise, but the
differences from adults were not statistically significant.
Breathing patterns are well recognized to change with increasing age, i.e., VT increase and
respiratory rates decrease (Tabachnik et al., 1981, 157036; Tobin et al., 1983, 156122). Bennett and
Zeman (1998, 076182) measured the deposition fraction of inhaled, fine particles (2 um dae) in
children (age 7-14 yr, n=16) and adults (age 19-35 yr, n=12) as they breathed the aerosol with their
natural, resting breathing pattern. Among the children, variation in deposition fractions, measured by
photometry at the mouth, was highly dependent on intersubject variation in VT. On the other hand,
they found no difference in deposition fractions between children versus adults for these fine
particles. This finding and the modeling predictions (Hofmann et al., 1989, 006922) are explained in
part by the smaller VT and faster breathing rate of children relative to adults for natural breathing
conditions. Bennett et al. (2008, 156269) also recently reported measures of fine particle (1 and 2 um
dae) deposition at ventilation rates typical of light exercise in children (age 6-10 yr, n=12) and adults
(age 18-27 yr, n=ll) and showed that, like with resting breathing, deposition fractions were
predicted by breathing pattern and did not differ or tended to be less in children compared to adults.
On the other hand, because children breathe at higher minute ventilations relative to their lung
volumes, the rate of deposition of fine particles normalized to lung surface area may be greater in
children versus adults (Bennett and Zeman, 1998, 076182).
Bennett and Zeman (2004, 155686) expanded their measures of fine particle deposition during
resting breathing to a larger group of healthy children (6-13 yr; 20 boys, 16 girls) and found again
that the variation in total deposition was best predicted by VT (r = 0.79, p < 0.001). But both VT and
resting minute ventilation increased with both height and body mass index (BMI) of the children.
Interestingly, these data suggest that for a given height and age, children with higher BMI have
larger minute ventilations and VT at rest than those with lower BMI. These differences in breathing
patterns as a function of BMI translated into increased deposition of fine particles in the heaviest
children. The rate of deposition (i.e., particles depositing per unit time) in the overweight children
was 2.8 times that of the leanest children (p < 0.02). Among all children, the rate of deposition was
significantly correlated with BMI (r = 0.46, p < 0.004). Some of the increase in deposition fractions
of heavier children may be due to their elevated VT, which was well correlated with BMI (r = 0.72,
p< 0.001).
In 62 healthy adults with normal lung function aged 18-80 yr, Bennett et al. (1996, 083284)
showed there was no effect of age on the whole lung deposition fractions of 2-um particles under
December 2009 4-13
-------
natural breathing conditions. Across all subjects, the deposition fractions were found to be
independent of age, depending on breathing period (r = 0.58, p < 0.001) and airway resistance
(r = 0.46, p < 0.001). In the same adults breathing with a fixed pattern (360 mL VT, 3.4 sec breathing
period), there was a mild decrease in deposition with increasing age, which could be attributed to
increased peripheral airspace dimensions in the elderly.
4.2.4.3. Gender
Males and females differ in body size, conductive airway size, and ventilatory parameters;
therefore, gender differences in deposition might be expected. In some of the controlled studies,
however, the men and women were constrained to breathe at the same VT and f Since women are
generally smaller than men, the increased minute ventilation of women compared to their normal
ventilation could affect deposition patterns. This may help explain why gender related effects on
deposition have been observed in some studies.
Kim and Hu (1998, 086066) assessed the regional deposition patterns of 1-, 3-, and 5-um
MMAD particles in healthy adult males and females using controlled breathing. The total fractional
deposition in the lungs was similar for both genders with the 1-um particle size, but was greater in
women for the 3- and 5-um particles. Deposition also appeared to be more localized in the lungs of
females compared to those of males. Kim and Jaques (2000, 012811) measured deposition in healthy
adults using sizes in the UF mode (0.04-0.1 um). Total fractional lung deposition was greater in
females than in males for 0.04- and 0.06-um particles. The region of peak fractional deposition was
shifted closer to the mouth and peak height was slightly greater for women than for men for all
exposure conditions. The total lung deposition data from these studies are illustrated in Figure 4-5.
These differences were generally attributed to the smaller size of the upper airways, particularly of
the laryngeal structure, in females.
In another study (Bennett et al, 1996, 083284). the total respiratory tract deposition of 2-um
particles was examined in adult males and females aged 18-80 yr who breathed with a normal resting
pattern. There was a tendency for greater deposition fractions in females compared to males.
However, since males had greater minute ventilation, the deposition rate (i.e., deposition per unit
time) was greater in males than in females. More recently, Bennett and Zeman (2004, 155686) found
no difference in the deposition of 2-um particles in boys versus girls aged 6-13 yr (n = 36).
4.2.4.4. Anatomical Variability
Anatomical variability, even in the absence of respiratory disease, can affect deposition
throughout the respiratory tract. The ET region is the first exposed to inhaled particles and, therefore,
deposition within this region would reduce the amount of particles available for deposition in the
lungs. Variations in relative deposition within the ET region will, therefore, propagate through the
rest of the respiratory tract, creating differences in calculated doses among individuals.
The influence of variations in nasal airway geometry on particle deposition has been
investigated. Cheng et al. (1996, 047520) examined nasal airway deposition in healthy adults using
particles ranging in size from 0.004 to 0.15 um and at 2 constant inspiratory flow rates, 167 and
333 mL/s. Inter-individual variability in deposition was correlated with the wide variation of nasal
dimensions; in that, greater surface area, smaller cross-sectional area, and increasing complexity of
airway shape were all associated with enhanced deposition. Bennett and Zeman (2005, 155687) have
also shown that nasal anatomy influences the efficiency of particle uptake in the noses of adults. For
light exercise breathing conditions in adults, their study demonstrated that nasal deposition
efficiencies for both 1- and 2-um monodisperse particles were significantly less in African
Americans versus Caucasians. The lesser nasal efficiencies in African-Americans were associated
with both lower nasal resistance and less elliptical nostrils compared to Caucasians.
Within the lungs, the branching structure of the airways may also differ between individuals.
Zhao et al. (2009, 157187) recently examined the bronchial anatomy of the left lung in patients
(132 M, 84 W; mean age 47 yr) that underwent conventional thoracic computed tomography scans
for various reasons. At the level of the segmental bronchus in the upper and lower lobes, a
bifurcation occurred in the majority of patients. A trifurcation, however, was observed in 23% of the
upper and 18% of the lower lobes. Other more unusual findings were also reported such as four
bronchi arising from the left upper lobe bronchus. As described in Section 4.2.2.4, deposition can be
highly localized near the carinal ridge of bifurcations. The effect of a bifurcation versus other
December 2009 4-14
-------
branching patterns on airflow patterns and particle deposition has not been described in the literature.
Martonen et al. (1994, 000847) showed that a wide blunt carinal ridge shape dramatically affected
the flow stream lines relative to a narrower and more rounded ridge shape. Specifically, there were
high flow velocities across the entire area of the blunt carinal ridge versus a smoother division of the
airstream in the case of the narrow rounded ridge shape. The implication may be that localized
particle deposition on the carinal ridge would increase with ridge width. A similar situation might be
expected for a trifurcation versus a bifurcation. These differences in branching patterns provide a
clear example of anatomical variability among individuals that might affect both air flow patterns
and sites of particle deposition.
4.2.4.5. Respiratory Tract Disease
The presence of respiratory tract disease can affect airway structure and ventilatory
parameters, thus altering deposition compared to that occurring in healthy individuals. The effect of
airway diseases on deposition has been studied extensively, as described in the 1996 and 2004 PM
AQCD (U.S. EPA, 1996, 079380: U.S. EPA, 2004, 056905). Studies described therein showed that
people with chronic obstructive pulmonary disease (COPD) had very heterogeneous deposition
patterns and differences in regional deposition compared to healthy individuals. People with
obstructive pulmonary diseases tended to have greater deposition in the TB region than did healthy
people. Furthermore, there tended to be an inverse relationship between bronchoconstriction and the
extent of deposition in the alveolar region, whereas total respiratory tract deposition generally
increased with increasing degrees of airway obstruction.
The vast majority of deposition studies in individuals with respiratory disease have been
performed during controlled breathing, i.e., all subjects breathed with the same VT and f However,
although resting VT is similar or elevated in people with COPD compared to healthy individuals, the
former tend to breathe at a faster rate, resulting in higher than normal tidal peak flow and resting
minute ventilation. Thus, given that breathing patterns differ between healthy and obstructed
individuals, particle deposition data for controlled breathing may not be appropriate for estimating
respiratory doses from ambient PM exposures.
Bennett et al. (1997, 078839) measured the fractional deposition of insoluble 2-um particles in
moderate-to-severe COPD patients (n = 13; mean age 62 yr) and healthy older adults (n = 11; mean
age 67 yr) during natural resting breathing. COPD patients had about a 50% greater deposition
fraction and a 50% increase in resting minute ventilation relative to the healthy adults. As a result,
the patients had an average deposition rate of about 2.5 times that of healthy adults. Similar to
previously reviewed studies (U.S. EPA, 1996, 079380: U.S. EPA, 2004, 056905). these investigators
observed an increase in deposition with an increase in airway resistance, suggesting that deposition
increased with the severity of airway disease.
Brown et al. (2002, 043216) measured the deposition of an UF aerosol (CMD = 0.033 um)
during natural resting breathing in 10 patients with moderate-to-severe COPD (mean age 61 yr) and
9 healthy adults (mean age 53 yr). The COPD group consisted of 7 patients with chronic bronchitis
and 3 patients with emphysema. The total deposition fraction in the bronchitic patients (0.67) was
significantly (p < 0.02) greater than in either the patients with emphysema (0.48) or the healthy
subjects (0.54). Minute ventilation increased with disease severity (healthy, 5.8 L/min; chronic
bronchitic, 6.9 L/min; emphysema, 11 L/min). Relative to the healthy subjects, the average dose rate
was significantly (p < 0.05) increased by 1.5 times in the COPD patients, whereas the average
deposition fraction only tended to be increased by 1.1 times. These data further demonstrate the need
to consider dose rates (which depend on minute ventilation) rather than just deposition fractions
when evaluating the effect of respiratory disease on particle deposition and dose.
December 2009 4-15
-------
0.8
.2 0.6
0.4
Q
0.2
0.0
Hygroscopic
Hydrophobic
0.03 0.1 0.4
Inhaled particle diameter (|jm)
Source:DatafromTu and Knutson (1984, 0728701.
Figure 4-6. Total deposition of hygroscopic sodium chloride and hydrophobic
aluminosilicate aerosols during oral breathing (VT= 1.0 L; f = 15 min"1).
4.2.4.6. Hygroscopicity of Aerosols
Experimental and modeling studies of hygroscopic aerosol growth and deposition in the lung
were extensively discussed in Section 10.4.3.1 of the 1996 PM AQCD (U.S. EPA, 1996, 079380).
Hygroscopic ambient aerosols include sulfates, nitrates, some organics, and aerosols laden with
sodium or potassium. The high relative humidity in the lungs contributes to rapid growth of
hygroscopic particles and dramatically alters the deposition characteristics of ambient hygroscopic
aerosols relative to nonhygroscopic aerosols. Nonhygroscopic particles in the range of 0.3 um have
minimal intrinsic mobility and low total deposition in the lungs. However, a 0.3 um salt particle
(dry) will grow in vivo to nearly 2 um and deposit to a far greater extent (Anselm et al., 1990,
156217). The hygroscopic growth of particles in the respiratory tract decreases diffusive deposition
and increases aerodynamic deposition as illustrated in Figure 4-6.
4.2.5. Summary
Particle deposition in the respiratory tract occurs predominantly by diffusion, impaction, and
sedimentation. Deposition is minimal for particle diameters in the range of 0.1 to 1.0 um, where
particles are small enough to have minimal sedimentation or impaction and sufficiently large so as to
have minimal diffusive deposition. In humans, total respiratory tract deposition approaches 100% for
particles of roughly 0.01 um (physical diameter) due to diffusive deposition and for particles of
around 10 um dae due to the efficiency of sedimentation and impaction.
The first line of defense for protecting the lower respiratory tract from inhaled particles is the
nose and mouth. Nasal deposition approaches 100% in the average human for 10 um dae particles.
Experimental studies show lower nasal particle deposition in children than adults. Relative to adults,
children also tend to breathe more through their mouth which is less efficient for removing inhaled
particles than the nose. These findings suggest that the lower respiratory tract of children may
receive a higher dose of ambient PM compared to adults. Since children breathe at higher minute
ventilations relative to their lung volumes, the rate of particle deposition normalized to lung surface
area may be further increased relative to adults.
People with COPD generally have greater total deposition and more heterogeneous deposition
patterns compared to healthy individuals. The observed increase in deposition correlates with
increases in airway resistance, suggesting that deposition increases with the severity of airway
disease. COPD patients also have an increased resting minute ventilation relative to the healthy
adults. This demonstrates the need to consider dose rates (which depend on minute ventilation)
December 2009
4-16
-------
rather than just deposition fractions when evaluating the effect of respiratory disease on particle
deposition and dose.
Modeling studies indicate that, for particles greater than -0.01 um, some cells located near the
carinal ridge of bronchial bifurcations may receive hundreds to thousands times the average dose
(particles per unit surface area) of the parent and daughter airways. The inertial impaction of
particles > 1 um dae at the carinal ridge of large bronchi increases with increasing inspiratory flows.
Airway constriction can further augment the overall deposition efficiency of coarse particles at
downstream bifurcations. These findings suggest that substantial doses of particles (> 1 urn dae) may
be justified for in vitro studies using tracheobronchial epithelial cell cultures.
Our ability to extrapolate between species has not generally changed since the 2004 PM
AQCD (U.S. EPA, 2004, 056905). However, some considerations related to coarse particles warrant
comment. The inhalability of particles having dae of 2.5, 5, and 10 um is 80, 65, and 44% in rats,
respectively, whereas it remains near 100% for a dae of 10 um in humans. In most laboratory animal
species (rat, mouse, hamster, guinea pig, and dogs), deposition in the extrathoracic region is near
100% for particles greater than 5 um dae. By contrast, in humans nasal deposition approaches 100%
for 10 um dae. Oronasal breathing versus obligate nasal breathing further contributes to greater
penetration of coarse particles into the lower respiratory tract of humans than rodents.
4.3. Clearance of Poorly Soluble Particles
This section discusses the clearance and translocation of poorly soluble particles that have
deposited in the respiratory tract. The term "clearance" is used here to refer to the processes by
which deposited particles are removed by mucociliary action or phagocytosis from the respiratory
tract. "Translocation" is used here mainly to refer to the movement of free particles across cell
membranes and to extrapulmonary sites. In the literature, translocation may also refer to the extra-
and intracellular dissolution of particles and the subsequent transfer of dissociated material to the
blood through extra- and intracellular fluids and across the various cell membranes and lung tissues.
The clearance and distribution of soluble particles and soluble constituents of particles are discussed
in Section 4.4.
A basic overview of biological mechanisms and clearance pathways from various regions of
the respiratory tract are presented in the following sections. Then regional kinetics of particle
clearance are addressed. Subsequently, an update on interspecies patterns and rates of particle
clearance is provided. The translocation of UFPs is also discussed. Finally, information on biological
factors that may modulate clearance is presented.
4.3.1. Clearance Mechanisms and Kinetics
For any given particle size, the deposition pattern of poorly soluble particles influences
clearance by partitioning deposited material between lung regions. Tracheobronchial clearance of
poorly soluble particles in humans, with some exceptions, is thought (in general) to be complete
within 24-48 h through the action of the mucociliary escalator. Clearance of poorly soluble particles
from the alveolar region is a much slower process which may continue from months to years.
4.3.1.1. Extrathoracic Region
Particles deposited in either the nasal or oral passages are cleared by several mechanisms.
Particles depositing in the mouth may generally be assumed to be swallowed or removed by
expectoration. Particles deposited in the posterior portions of the nasal passages are moved via
mucociliary transport towards the nasopharynx and swallowed. Mucus flow in the most anterior
portion of the nasal passages is forward, toward the vestibular region where removal occurs by
sneezing, wiping, or nose blowing.
December 2009 4-17
-------
4.3.1.2. Tracheobronchial Region
Mucociliary clearance in the TB region has generally been considered to be a rapid process
that is relatively complete by 24-48 h post-inhalation in humans. Mucociliary clearance is frequently
modeled as a series of "escalators" moving material proximally from one generation to the next. As
such, the removal rate of particles from an airway generation increases with increasing tracheal
mucus velocity. Assuming continuity in the amount of mucus between airway generations, mucus
velocities decrease and transit times within an airway generation increase with distal progression.
Although clearance from the TB region is generally rapid, experimental evidence discussed in the
1996 and 2004 PM AQCD (U.S. EPA, 1996, 079380: U.S. EPA, 2004, 056905) showed that a
fraction of material deposited in the TB region is retained much longer.
The slow-cleared TB fraction (i.e., the fraction of particles deposited in the TB region that are
subject to slow clearance) was thought to increase with decreasing particle size. For instance, Roth et
al. (1993, 156928) showed approximately 93% retention of UFPs (30 nm median diameter) thought
to be deposited in the TB region at 24 h post-inhalation. The slow phase clearance of these UFPs
continued with an estimated half-time (ti/2) of around 40 days. Using a technique to target inhaled
particles (monodisperse 4.2 urn MMAD) to the conducting airways, Moller et al. (2004, 155987)
observed that 49 ± 9% of particles cleared rapidly (ti/2 of 3.0 ±1.6 h), whereas the remaining fraction
cleared considerably slower (ti/2 of 109 ± 78 days). The ICRP (1994, 006988) human respiratory
tract model assumes particles < 2.5 urn (physical diameter) to have a slow-cleared TB fraction of
50%. The slow-cleared fraction assumed by the ICRP (1994, 006988) decreases with increasing
particle size to <1% for 9 urn particles. Considering the UF data of Roth et al. (1993, 156928) in
addition to data considered by the ICRP (1994, 006988). Bailey et al. (1995, 190057) estimated a
slow-cleared TB fraction of 75% for UFPs. At that time, they (Bailey et al. 1995) also estimated the
slow-cleared fraction to decrease with increasing particle size to 0% for particles > 6 urn. Recent
experimental evidence from the same group (Smith et al., 2008, 190037) showed no difference in TB
clearance among humans for particles with geometric sizes of 1.2 versus 5 urn, but the same dae (5
urn) so as to deposit similarly in the TB airways. For at least micron-sized particles, these recent
findings do not support the particle size dependence of a slow-cleared TB fraction. As discussed
further below, much of the apparent slow-cleared TB fraction may be accounted for by differences in
deposition patterns, i.e., greater deposition in the alveolar region than expected.
A portion of the slow cleared fraction from the TB region appears to be associated with small
bronchioles. For large particles (dae = 6.2 urn) inhaled at a very slow rate to theoretically deposit
mainly in small ciliated airways, 50% had cleared by 24 h post-inhalation. Of the remaining
particles, 20% cleared with ati/2 of 2.0 days and 80% with ati/2 of 50 days (Falk et al., 1997,
086080). Using the same techniques, Svartengren et al. (2005, 157034) also reported the existence of
long-term clearance in humans from the small airways. It should be noted that the clearance rates for
the slow-cleared TB fraction still exceeds the clearance rate of the alveolar region in humans.
Kreyling et al. (1999, 039175) targeted inhaled particle (dae = 2.2 and 2.5 urn) deposition to the TB
airways of adult beagle dogs and subsequently quantified particle retention using scintigraphic and
morphometric analyses. Despite the use of shallow aerosol bolus inhalation to a volumetric lung
depth of less than the anatomic dead space, 2.5-25% of inhaled particles deposited in alveoli. At 24
and 96 h post-inhalation, more than 50% of the retained particles were in alveoli. However, 40% of
particles present at 24 and 96 h were localized to small TB airways of between 0.3 and 1 mm in
diameter. Collectively, these studies suggest that although mucociliary clearance is fast and effective
in healthy large airways, it is less effective and sites of longer retention exist in the smaller TB
airways.
The underlying sites and mechanisms of long-term TB retention in the smaller airways remain
largely unknown. Several factors may contribute to the existence or experimental artifact of slow
clearance from the smaller TB airways. Even when inhaled to very shallow lung volumes, some
particles reach the alveolar region (Kreyling et al., 1999, 039175). Therefore, experiments utilizing
bolus techniques to target inhaled particle deposition to the TB airways may have had some
deposition in the alveolar region. This may occur due to variability in path length and the number of
generations to the alveoli (Asgharian et al., 2001, 017025) and/or differences in regional ventilation
(Brown and Bennett, 2004, 190032). Nonetheless, the experimentally measured clearance rates
measured for the slow cleared TB fraction are faster than that of the alveolar region in both humans
and canines. Thus, although experimental artifacts likely occur, they do not discount the existence of
a slow cleared TB fraction. To some extent, it is possible that the slow cleared TB fraction may be
due to bronchioles that do not have a continuous ciliated epithelium as in the larger bronchi. Neither
December 2009 4-18
-------
path length, ventilation distribution, nor a discontinuous ciliated epithelium explains an apparently
slow cleared TB fraction with decreasing particle size below 0.1 urn. As discussed in Section 4.3.3
on Particle Translocation, UFPs cross cell membranes by mechanisms different from larger (~1 urn)
particles. Based on that body of literature, particles smaller than a micron may enter epithelial cells
resulting in their prolonged retention, particularly in the bronchioles where the residence time is
longer and distances necessary to reach the epithelium are shorter compared to that in the bronchi.
4.3.1.3. Alveolar Region
The primary alveolar clearance mechanism is macrophage phagocytosis and migration to
terminal bronchioles where the cells are cleared by the mucociliary escalator. Alveolar macrophages
originate from bone marrow, circulate briefly as monocytes in the blood, and then become
pulmonary interstitial macrophages before migrating to the luminal surfaces. Under normal
conditions, a small fraction of ingested particles may also be cleared through the lymphatic system.
This may occur by transepithelial migration of alveolar macrophage following particle ingestion or
free particle translocation with subsequent uptake by interstitial macrophages. Snipes et al. (1997,
156092) have also demonstrated the importance of neutrophil phagocytosis in clearance of particles
from the alveolar region. Rates of alveolar clearance of poorly soluble particles vary between species
and are briefly discussed in Section 4.3.2. The translocation of particles from their site of deposition
is discussed in Section 4.3.3.
The efficiency of macrophage phagocytosis is thought to be greatest for particles between 1.5
and 3 urn (Oberdorster, 1988, 006857). The decreased efficiency of alveolar macrophage for
engulfing UFPs increases the time available for these particles to be taken up by epithelial cells and
moved into the interstitium (Ferin et al., 1992, 044401). Consistent with this supposition
(i.e., translocation increases with time), an increase in titanium dioxide (TiO2) particle transport to
lymph nodes has been reported following inhalation of a cytotoxin to macrophages (Greenspan et al.,
1988, 045031). Interestingly, the long-term clearance kinetics of the poorly soluble UF (15-20 nm
CMD) iridium (Ir) particles were found to be similar to the kinetics reported in the literature for
micrometer-sized particles (Semmler et al., 2004, 055641; Semmler-Behnke et al., 2007, 156080).
Semmler-Behnke et al. (2007, 156080) concluded that UF Ir particles are less phagocytized by
alveolar macrophage than larger particles, but are effectively removed from the airway surface into
the interstitium. Particles are then engulfed by interstitial macrophages which then migrate to the
airway lumen and are removed by mucociliary clearance to the larynx. The major role of
macrophage-mediated clearance was supported by lavage of relatively few free particles versus
predominantly phagocytized particles at time-points of up to 6 mo. It is also possible that some free
particles as well as particle-laden macrophages were carried from interstitial sites via the lymph flow
to bronchial and bronchiolar sites, including bronchial-associated lymphatic tissue, where they were
excreted again into the airway lumen.
4.3.2. Interspecies Patterns of Clearance and Retention
There are differences between species in both the rates of particle clearance from the lung and
manner in which particles are retained in the lung. For instance, based on models of mucociliary
clearance from un-diseased airways, >95% of particles deposited in the tracheobronchial airways of
rats are predicted to be cleared by 5 h post deposition, whereas it takes nearly 40 h for comparable
clearance in humans (Hofmann and Asgharian, 2003, 055579). As noted in Section 4.3.1.2,
however, there is considerable evidence that a sizeable fraction of particles deposited at the
bronchiolar level of the ciliated airways in humans (as well as canines) are cleared at a far slower
rate. The slow cleared TB fraction appears to increase with decreasing particle size.
December 2009 4-19
-------
1.00
-------
4.3.3.1. Alveolar Region
Numerous studies have examined the translocation of UFPs from their site of deposition in the
lung. Traditionally viewed as a relatively inert particle type, UF TiO2 has received the most study. At
the time the 2004 PM AQCD was released, there were conflicting results regarding the rate and
magnitude of UF carbon translocation from the human lung. Since that time, it has become well-
established that the transport of UF carbon particles from the human lung is far slower than that of
soluble materials. However, it has also been shown in animal studies (primarily of rats) that UFPs
cross cell membranes by mechanisms different from larger (~1 um) particles and that a small
fraction of these particles enter capillaries and may distribute systemically. Described in brief below,
details of selected new studies investigating the disposition of poorly soluble particles are provided
in Annex B.
There has been some contention regarding ability of UF carbon particles to rapidly diffuse
from the lungs into the systemic circulation. Based on their study of 5 healthy volunteers, Nemmar et
al. (2002, 024914) suggested that UF carbon particles (<100 nm) pass rapidly into the systemic
circulation. However, Brown et al. (2002, 043216) found that the majority of UF carbon particles
(CMD, 33 ± 2 nm) were still in the lungs of healthy human adult volunteers (n = 9; aged 40-67 yr)
and COPD patients (n = 10; 45-70 yr) at 24 h post-inhalation. Brown et al. (2002, 043216) and
Burch (2002, 056754) contended that the findings reported by Nemmar et al. (2002, 024914) were
consistent with soluble pertechnetate clearance, but not insoluble UF carbon particles. Highly
soluble in normal saline, pertechnetate clears rapidly from the lung with a half-time of-10 min and
accumulates most notably in the bladder, stomach, thyroid, and salivary glands. Three recent studies
have confirmed that the majority (>95%) of UF carbon particles deposited in the lungs of human
volunteers are retained at 24 h post-inhalation (Mills et al., 2006, 088770; Wiebert et al., 2006,
157146; Wiebert et al., 2006, 156154). Wiebert et al. (2006, 157146) modified their aerosol
generation system to reduce leaching of the 99mTc radiolabel from carbon particles. Except for a
small amount of radiotracer leaching from particles (1.0 ± 0.6% of initially deposited activity in
urine by 24 h), these investigators found negligible radiolabel and associated particle clearance from
the lungs by 70 h. The available data show that there is not a rapid or significant amount of UF
carbon particle migration into circulation (Brown et al., 2002, 043216; Burch, 2002, 056754; Mills et
al., 2006, 088770; Moller et al., 2008, 156771: Wiebert et al., 2006, 157146: Wiebert et al., 2006,
156154).
Although human studies show that the vast majority of UF carbon particles are retained in the
lungs until at least 24 h post-inhalation, both in vitro and in vivo studies support the rapid (< 1 h)
translocation of free UF TiO2 particles across pulmonary cell membranes (Churg et al., 1998,
085815: Ferin et al., 1992, 044401: Geiser et al., 2005, 087362). Peculiar to TiO2 aerosols, there is
evidence that particle aggregates may disassociate once deposited in the lungs. This disassociation
makes inhaled aggregate size the determinant of deposition amount and site, but primary particle size
the determinant of subsequent clearance (Bermudez et al., 2002, 055578: Ferin et al., 1992, 044401:
Takenaka et al., 1986, 046210). Following disaggregation, the UF TiO2 particles are cleared more
slowly and cause a greater inflammatory response (neutrophil influx) than fine TiO2 particles
(Bermudez et al., 2002, 055578: Ferin et al., 1992, 044401: Oberdorster et al., 1994, 046203:
Oberdorster et al., 1994, 056285; Oberdorster et al., 2000, 039014). The differences in inflammatory
effects and possibly lymph burdens between fine and UF TiO2 in many studies appear related to lung
burden in terms of particle surface area and not particle mass or number (Oberdorster, 1996, 039852:
Oberdorster et al., 1992, 045110: Oberdorster et al., 2000, 039014: Tran et al., 2000, 013071). More
recently, others have noted that particle surface area is not an appropriate metric across all particle
types (Warheit et al., 2006, 088436). Surface characteristics such as roughness can also affect protein
binding and potentially clearance kinetics, with smoother TiO2 surfaces being more hydrophobic
(Sousa et al., 2004, 089866).
Geiser et al. (2005, 087362) conducted a detailed examination of the disposition of inhaled UF
TiO2 in 20 healthy adult rats. They found that distributions of particles among lung tissue
compartments appeared to follow the volume fraction of the tissues and did not significantly differ
between 1 and 24 h post-inhalation. Averaging 1- and 24-h data, 79.3 ± 7.6% of particles were on the
luminal side of the airway surfaces, 4.6 ± 2.6% were in epithelial or endothelial cells, 4.8 ± 4.5%
were in connective tissues, and 11.3 ± 3.9% were within capillaries. Particles within cells were not
membrane bound. It is not clear why the fraction of particles identified in compartments such as the
capillaries did not differ between 1 and 24 h post-inhalation. These findings were consistent with the
smaller study of 5 rats by Kapp et al. (2004, 156624) who reported identifying TiO2 aggregates in a
December 2009 4-21
-------
type II pneumocyte; a capillary close to the endothelial cells; and within the surface-lining layer
close to the alveolar epithelium immediately following a 1-h exposure. These studies effectively
demonstrate that some inhaled UF TiO2 particles, once deposited on the pulmonary surfaces, can
rapidly (< 1 h) translocate beyond the epithelium and potentially into the vasculature.
Extrapulmonary translocation has also been described for poorly soluble UF gold and Ir
particles. In male Wistar-Kyoto rats exposed to UF gold particles (5-8 nm), Takenaka et al. (2006,
156110) reported a low but significant fraction (0.03 to 0.06% of lung concentration) of gold in the
blood from 1 to 7 days post-inhalation. Semmler et al. (2004, 055641) also found small but
detectable amounts of poorly soluble Ir particle (15 and 20 nm CMD) translocation from the lungs of
male Wistar-Kyoto rats to secondary target organs like the liver, spleen, brain, and kidneys. Each of
these organs contained about 0.2% of deposited Ir. The peak levels in these organs were found 7 days
post-inhalation. The translocated particles were largely cleared from extrapulmonary organs by 20
days and Ir levels were near background at 60 days post-inhalation. Particles may have been
distributed systemically via the gastrointestinal tract. Immediately after the 6-h inhalation exposure,
18 ± 5% of the deposited Ir particles had already cleared into the gastrointestinal tract. After 3 wk,
31 ± 5% of the deposited particles were retained in the lung. By 2 and 6 mo post-inhalation, lung
retention was 17 ± 3 and 7 ± 1%, respectively. The particles appeared to be cleared predominantly
from the peripheral lung via the mucociliary escalator into the GI tract and were found in feces.
A few recent studies have characterized differences in the behavior of fine and UF particles in
vitro. Geiser et al. (2005, 087362) found that both UF and fine (0.025 (im gold, 0.078 (im TiO2, and
0.2 (im TiO2) particles cross cellular membranes by non-endocytic (i.e., involving vesicle formation)
mechanisms such as adhesive interactions and diffusion, whereas the phagocytosis of larger 1 (im
TiO2 particles is ligand-receptor mediated. Edetsberger et al. (2005, 155759) found that UFPs
(0.020 um polystyrene) translocated into cells by first measurement (~1 min after particle
application). Intracellular agglomerates of 88-117 nm were seen by 15-20 min and of 253-675 nm by
50-60 min after particle application. These intracellular aggregates were thought to result from
particle incorporation into endosomes or similar structures since Genistein or Cytochalasin treatment
generally blocked aggregate formation. Interestingly, particles did not translocate into dead cells,
rather they attached to the outside of the cell membrane. Amine- or carboxyl-modified surfaces (46
nm polystyrene) did not affect translocation across cultures of human bronchial epithelial cells with
about 6% regardless of the surface characteristics (Geys et al., 2006, 155789).
4.3.3.2. Olfactory Region
Numerous studies have demonstrated the translocation of soluble and poorly soluble particles
from the olfactory mucosa via the axons to the olfactory bulb of the brain. The vast majority of these
studies were conducted in rodents. However, DeLorenzo (1970, 156391) observed the rapid (within
30-60 min) movement of 50 nm silver-coated colloidal gold particles instilled on the olfactory
mucosa into the olfactory bulb of squirrel monkeys. The specifics of this and other key studies that
have investigated the translocation of particles to the olfactory bulb are provided in Annex B.
Two recent studies reported the movement of UFPs deposited in the olfactory region of the
nose along the olfactory nerve and into the olfactory bulb of the brain in rats. Oberdorster et al.
(2004, 055639) exposed rats to UF carbon particles (36 nm CMD, 1.7 og) containing 13C in a whole-
body chamber for 6 h. The distribution of1 C was followed for 7 days postexposure. There was a
significant increase in 13C in the olfactory bulb on Day 1 with persistent and continued increase
through Day 7. Elder et al. (2006, 089253) exposed rats to manganese (Mn) oxide (~30 nm
equivalent sphere with 3-8 nm primary particles) via whole-body inhalation exposure for 12 days (6
h/day, 5 days/wk) with both nares open or Mn oxide for 2 days (6 h/day) with right nostril blocked.
After the 12 days exposure via both nostrils, Mn in the olfactory bulb increased 3.5-fold. After the
2-day exposure with the right nostril blocked, Mn was found mainly in the left olfactory bulb (2.4-
fold increase). These studies suggest the neuronal uptake and translocation of UFPs without particle
dissolution and in the absence of mucosal injury.
Elder et al. (2006, 089253) also addressed the issue of whether solubilization of particles was
requisite for translocation along the olfactory nerve and into the brain. Similar amounts of soluble
manganese chloride (MnCl2) and poorly soluble Mn oxide were instilled onto the left naris of
anesthetized rats. At 24 h post-instillation, similar amounts of Mn were found in the left olfactory
bulb of rats instilled with MnCl2 (8.2 ± 3.6% of instilled) and Mn oxide (8.2 ± 0.7% of instilled). If
solubilization were required for translocation, then a lower amount of Mn oxide than MnCl2 should
December 2009 4-22
-------
have reached the olfactory bulb. Following 14 consecutive days of aerosol exposure, Dorman et al.
(2001, 055433) demonstrated that more soluble Mn sulfate reaches the olfactory bulb and striatum of
rat brains than the poorly soluble form of Mn tetroxide. Nonetheless, the Mn levels were statistically
increased in both the olfactory bulb and striatum following exposure to Mn tetroxide relative to
filtered air. In a subsequent 13-wk exposure study, Dorman et al. (2004, 155752) also demonstrated
that more soluble manganese sulfate (MnSO4) reached the olfactory bulb than was observed for the
less soluble Mn form (hureaulite). Both the soluble and less soluble forms of Mn resulted in
statistically increased levels of Mn in the olfactory bulb relative to air exposed controls. The soluble
MnSO4 was also observed to reach the striatum and cerebellum. In addition, Yu et al. (2003, 156171)
demonstrated increased Mn levels in the brains of rats exposed to welding-fumes for 60 days,
however, the role of transport via the blood is less clear in this study.
The translocation of zinc (Zn) and TiO2 to the olfactory bulb has also been reported in the
literature. Persson et al. (2003, 051846) observed the translocation of Zn to the olfactory bulbs
following instillation in both rats and freshwater pike. Wang et al. (2007, 156146) reported the
translocation of both fine (155 nm) and UF (21 and 71 nm) TiO2 particles in mice. Interestingly, a
qualitative analysis of the data showed that more of the fine TiO2 than UF TiO2 reached the olfactory
bulb. Wang et al. (2007, 156146) suggested that a strong hydrophilic character and propensity for
aggregation reduced the translocation of the UF TiO2.
The importance of particle translocation to the brain is not yet understood. Translocation via
the axon to the olfactory bulb has been observed for numerous compounds of varying composition,
particle size, and solubility. Although the rate of translocation is rapid, perhaps less than an hour, the
magnitude of transport remains poorly characterized. With regard to the magnitude of transport,
Elder et al. (2006, 089253) found that as much as 8% of both soluble and insoluble forms of Mn
were translocated to the olfactory bulb in rats following intranasal instillation. It is also still unclear
to what extent translocation to the olfactory bulb and other brain regions may vary between species.
The olfactory mucosa covers approximately 50% of the nasal epithelium in rodents versus only
about 5% in primates (Aschner et al., 2005, 155663). Additionally, a greater portion of inhaled air
passes through the olfactory region of rats relative to primates (Kimbell, 2006, 155902). These
differences may predispose rats, more so than humans, to deposition of particles in the olfactory
region with subsequent particle translocation to the olfactory bulb.
4.3.4. Factors Modulating Clearance
4.3.4.1. Age
It was previously concluded that there appeared to be no clear evidence for any age-related
differences in clearance from the lung or total respiratory tract, either from child to adult, or young
adult to elderly (U.S. EPA, 1996, 079380; U.S. EPA, 2004, 056905). Studies showed either no
change or some slowing in mucus clearance with age after maturity. Although some differences in
alveolar macrophage function were reported between mature and senescent mice, no age-related
decline in macrophage function had been observed in humans. A comprehensive review of the recent
and older literature supports a decrease in mucociliary clearance with increasing age beyond
adulthood in humans and animals. Limited animal data also suggest macrophage-mediated alveolar
clearance may also decrease with age.
Studies addressing the effects of age on respiratory tract clearance are provided in Annex B.
Ho et al. (2001, 156549) demonstrated that nasal mucociliary clearance rates were about 40% lower
in old (age >40-90 yr) versus young (age 11-40 yr) men and women. Tracheal mucus velocities in
elderly (or aged) humans and beagle dogs are about 50% that of young adults (Goodman et al., 1978,
071130; Whaley et al., 1987, 156153). Several human studies have demonstrated decreasing rates of
mucociliary particle clearance from the large and small bronchial airways with increasing age
(Puchelle et al., 1979, 006863; Svartengren et al., 2005, 157034; Vastag et al., 1985, 157088). Linear
fits to the data show that rapid clearance (within 1 h) from large bronchi and prolonged clearance
(between 1-21 days) from the small bronchioles in an 80-year-old is only about 50% of that in a
20-year-old (Svartengren et al., 2005, 157034; Vastag et al., 1985, 157088). One study reported that
alveolar particle clearance rates decreased by nearly 40% in old versus young rats (Muhle et al.,
1990, 006853). Another study has reported that older rats have an increased susceptibility to
December 2009 4-23
-------
pulmonary infection due to altered alveolar macrophage function and slowed bacterial clearance
(Antonini et al., 2001, 156219). Although data are somewhat limited, they consistently show a
depression of clearance throughout the respiratory tract with increasing age from young adulthood in
humans and laboratory animals.
4.3.4.2. Gender
Gender was not found to affect clearance rates in prior reviews (U.S. EPA, 1996, 079380;
U.S. EPA, 2004, 056905). Studies not included in those reviews also show that human males and
females have similar nasal mucus clearance rates (Ho et al., 2001, 156549). tracheal mucus
velocities (Yeates et al., 1981, 095391). and large bronchial airway clearance rates (Vastag et al.,
1985. 157088).
4.3.4.3. Respiratory Tract Disease
At the time of the last two reviews (U.S. EPA, 1996, 079380: U.S. EPA, 2004, 056905). it was
well recognized that obstructive airways disease may influence both the site of initial deposition and
the rate of mucociliary clearance from the airways. When deposition patterns are matched,
mucociliary clearance rates are reduced in patients with COPD relative to healthy controls. The
effects of acute bacterial/viral infections and cough on mucociliary clearance were briefly
summarized in Section 10.4.2.5 (EPA, 1996, 079380) and Section 6.3.4.4 (EPA, 2004, 056905) of
past reviews. While cough is generally a reaction to some inhaled stimulus, in some cases, especially
respiratory disease, it can also serve to clear the upper bronchial airways of deposited substances by
dislodging mucus from the airway surface. One of the difficulties in assessing effects on infection on
mucociliary clearance is that spontaneous coughing increases during acute infections. Cough has
been shown to supplement mucociliary clearance of secretions, especially in patients with
obstructive lung disease and primary ciliary dyskinesia.
Using a bolus technique to target specific lung regions, Moller et al. (2008, 156771) examined
particle clearance from the ciliated airways and alveolar region of healthy subjects, smokers, and
patients with COPD. Airway retention after 1.5 hours was significantly lower in healthy subjects
(89 ± 6%) than smokers (97 ± 3%) or COPD patients (96 ± 6%). At 24 and 48 h, retention remained
significantly higher in COPD patients (86 ± 6% and 82 ± 6%, respectively) than healthy subjects
(75 ± 10% and 70 ± 9%, respectively). However, these findings are confounded by the more central
pattern of deposition in the healthy subjects than in the smokers and COPD patients. Alveolar
retention of particles was similar between the groups at 48 h post-inhalation.
The effect of asthma on lung clearance of particles may depend on disease status. Lay et al.
(2009, 190060) found significantly (p < 0.01) more rapid particle (0.22 (im) mucociliary clearance
over a 2-h period post-inhalation in mild asthmatics than in healthy volunteers. Although the pattern
of deposition tended to be more central in the asthmatics, there was not a statistically significant
difference from healthy controls. In vivo uptake by airway macrophages in mild asthmatics was also
enhanced relative to healthy volunteers (p < 0.01). In an ex vivo study, airway macrophages from
individuals with more severe asthma had impaired phagocytic capacity relative to less severely
affected asthmatics and healthy volunteers (Alexis et al., 2001, 190013). Lay et al. (2009, 190060)
concluded that enhanced uptake and processing of particulate antigens could contribute to the
pathogenesis and progression of allergic airways disease in asthmatics and may contribute to an
increased risk of exacerbations with parti culate exposure.
Chen et al. (2006, 147267) investigated the effect of endotoxin on the disposition of particles.
Healthy rats and those pretreated with endotoxin (12 h before particle instillation) were instilled with
UF (56.4 nm) or fine (202 nm) particles. In healthy rats, there were no marked differences in lung
retention or systemic distribution between the UF and fine particles. In healthy animals, UFPs were
primarily retained in lungs (72 ± 10% at 0.5-2 h; 65 ± 1% at 1 day; 62 ± 5% at 5 days). Particles
were also detected in the blood (2 ± 1% at 0.5-2 h; 0.1 ± 0.1% at 5 days) and liver (3 ± 2% at 0.5-2
h; 1 ± 0.1% at 5 days) of the healthy animals. At 1 day post-instillation, about 13% of the particles
were excreted in the urine or feces of the healthy animals. In rats pretreated with endotoxin, by 2 h
post-instillation, the UFPs accessed the blood (5 versus 2%) and liver (11 versus 4%) to a
significantly greater extent than fine particles. The endotoxin-treated rats also had significantly
greater amounts of UFPs in the blood (5% versus 2%) and liver (11% versus 3%) relative to the
December 2009 4-24
-------
healthy control rats. This study demonstrates that acute pulmonary inflammation caused by
endotoxin increases the migration of UFPs into systemic circulation.
Adamson and Preditis (1995, 189982) investigated the possibility that particle deposition into
an already injured lung might affect particle retention and enhance the toxicity of "inert" particles.
Bleomycin was instilled into mice to induce epithelial necrosis and subsequent pulmonary fibrosis.
Instilled 3 days following bleomycin treatment, while epithelial permeability was compromised,
carbon black particles in treated mice were translocated to the interstitium and showed increased
pulmonary retention relative to untreated mice. When instilled 4 wk post-bleomycin treatment, after
epithelial integrity was restored, carbon black particle retention was similar between treated and
untreated mice with minimal translocation to the interstitium. The instillation of carbon particles did
not appear to increase lung injury in the bleomycin treated mice at either time point. This study
shows that integrity of the epithelium affects particle retention and translocation into interstitial
tissues.
4.3.4.4. Particle Overload
Unlike other laboratory animals, rats appear susceptible to "particle overload" effects due to
impaired macrophage-mediated alveolar clearance. Numerous reviews have discussed this
phenomenon and the difficulties it poses for the extrapolation of chronic effects in rats to humans
(International, 2000, 002892; Miller, 2000, 011822; Morrow, 1994, 006850; Oberdorster, 1995,
046596; Oberdorster, 2002, 021111). Large mammals have slow pulmonary particle clearance and
retain particles in interstitial tissues under normal conditions, whereas rats have rapid pulmonary
clearance and retain particles in alveolar macrophages (Snipes, 1996, 076041). With chronic high
doses of PM there is a shift in the pattern of dust accumulation and response from that observed at
lower doses in rat lungs (Snipes, 1996, 076041; Vincent and Donaldson, 1990, 002462). Rats
chronically exposed to high concentrations of insoluble particles experience a reduction in their
alveolar clearance rates and an accumulation of interstitial particle burden (Bermudez et al, 2002,
055578; Bermudez et al., 2004, 056707; Ferin et al., 1992, 044401; Oberdorster et al., 1994, 046203;
Oberdorster et al., 1994, 056285; Warheit et al., 1997, 086055). With continued exposure, some rats
eventually develop pulmonary fibrosis and both benign and malignant tumors(Lee et al., 1985,
067628; Lee et al., 1986, 067629; Warheit et al., 1997, 086055). Oberdorster (1996, 039852; 2002,
021111) proposed that high-dose effects observed in rats may be associated with two thresholds. The
first threshold is the pulmonary dose that results in a reduction in macrophage-mediated clearance.
The second threshold, occurring at a higher dose than the first, is the dose at which antioxidant
defenses are overwhelmed and pulmonary tumors develop. Intrapulmonary tumors following TiO2
exposures are exclusive to rats and are not found in mice or hamsters (Mauderly, 1997, 084631).
Moreover, Lee et al. (1985, 067628) noted that the squamous cell carcinomas observed with
prolonged high concentration TiO2 exposures developed from the alveolar lining cells adjacent to the
alveolar ducts, whereas squamous cell carcinomas in humans which are generally linked with
cigarette smoking are thought to arise from basal cells of the bronchial epithelium. Quoting Lee et al.
(1986, 067629). "Since the lung tumors were a unique type of experimentally induced tumor under
exaggerated exposure conditions and have not usually been seen in man or animals, their relevance
to man in questionable."
4.3.5. Summary
For any given particle size, the pattern of poorly soluble particle deposition influences
clearance by partitioning deposited material between regions of the respiratory tract. Particles
depositing in the mouth may generally be assumed to be swallowed or removed by expectoration.
Particles deposited in the posterior portions of the nasal passages or the TB airways are moved via
mucociliary transport towards the nasopharynx and swallowed. Although clearance from the TB
region is generally rapid, there appears to be fraction of material deposited in the TB region of
humans that is retained much longer. The underlying sites and mechanisms of long-term TB
retention are not known. The primary alveolar clearance mechanism is macrophage phagocytosis and
migration to terminal bronchioles where the cells are cleared by the mucociliary escalator. Clearance
from both the TB and alveolar region is more rapid in rodents than humans. Mucociliary and
macrophage-mediated clearance decreases with age beyond adulthood.
December 2009 4-25
-------
Human data show that there is not a rapid or significant amount of UF carbon particle
migration into circulation. However, both in vitro and in vivo animal studies support the rapid
(< 1 h) translocation of free UF TiO2 particles across pulmonary cell membranes. Extrapulmonary
translocation has also been described in rats for poorly soluble UF gold and Ir particles. A low, but
statistically significant, fraction (0.03-0.06% of lung concentration) of UF gold particles has been
observed in the blood of rats from 1 to 7 days post-inhalation. The translocation in detectable
amounts (<1% of deposited material) of poorly soluble Ir particles (15 and 20 nm CMD) from the
lungs of rats to secondary target organs like the liver, spleen, brain, and kidneys has also been
reported. However, the systemic distribution of particles may have occurred via normal clearance
from the lungs to the gastrointestinal tract.
Although the importance of particle translocation to the brain is not yet understood,
translocation from the olfactory mucosa via the axon to the olfactory bulb has been reported in
primates, rodents, and freshwater pike for numerous compounds of varying composition, particle
size, and solubility. The rate of translocation is rapid, perhaps less than an hour. In rats, as much as
8% of material may become translocated to the olfactory bulb following intranasal instillation. It is
unclear to what extent translocation to the olfactory bulb and other brain regions may vary between
species. Interspecies differences may predispose rats, more so than humans, to the deposition of
particles in the olfactory region with subsequent translocation to the olfactory bulb.
4.4. Clearance of Soluble Materials
Soluble particles and soluble constituents of particles may be absorbed through the epithelium
and distributed systemically or retained in the lung. The rate of dissolution depends on a number of
factors, including particle surface area and chemical structure. Some dissolved materials bind to
proteins or other components in the airway surface liquid layer. In the ciliated airways, solutes are
cleared by mucociliary transport and diffuse into underlying tissues and the blood. In the alveolar
regions, the thin barrier between the air and blood allows for rapid transport of solutes into the blood.
The movement of soluble materials depends on the site of deposition in the lung, the rate of material
dissolution from particles, and the molecular weight of the solute. The rate of soluble material
clearance from the lungs depends on epithelial permeability which may be affected by age,
respiratory disease, and concurrent exposures. While enhanced clearance of insoluble particles acts
to reduce dose to airway tissue, increased transport of soluble matter into the blood stream may
enhance effects on extra-pulmonary organs.
4.4.1. Clearance Mechanisms and Kinetics
The rate of absorption across the epithelium for materials that dissolve in the airway or
alveolar lining fluid is fairly rapid (minutes to hours) and is a function of their molecular size and
their water or lipid solubility (Enna and Schanker, 1972, 155767; Huchon et al, 1987, 024923;
Oberdorster, 1988, 006857; Schanker et al., 1986, 005100). Huchon et al. (1987, 024923) studied the
clearance of a variety of aerosolized solutes from the lungs of dogs. Solute clearance was inversely
related to molecular weight. Negligible clearance of the largest molecular weight solute (transferrin
mol wt -76,000 daltons) in their study was found over a 30-min observation period. At the other
extreme, free pertechnetate (mol wt -163 daltons) had a clearance rate of 6% per min. Clearance of
hydrophilic solutes is diffusion limited by pore sizes associated with intercellular tight junctions
(estimated at 0.6-1.5 nm). Absorption of lipophilic compounds that pass easily through cell
membranes is perfusion limited and thus generally occurs very rapidly. However, if lipophilic
materials are adsorbed onto insoluble particles their retention in the lung may be prolonged (Creasia
et al., 1976, 059713). In addition to diffusion through intercellular junctions, transcellular transport
of large solutes by pinocytosis into epithelial cells has also been observed (Chinard, 1980, 156341).
A portion of poorly soluble particles may become dissolved with subsequent solute clearance.
More rapid dissolution of poorly soluble nano- or UF particles relative to micro-sized particles
occurs due to an increasing surface-to-volume ratio with decreasing particle size. Kreyling et al.
(2002, 037332) examined the dissolution of poorly soluble UF Ir particle agglomerates (15-80 nm
CMD) composed of 5 nm primary particles. After 7 days, <1% of the particles were dissolved in
buffered saline, whereas 6% dissolved in 1 N hydrochloric acid after 1 day. Thus, the high surface-
December 2009 4-26
-------
to-volume ratio of UFPs should not be misconstrued to imply rapid dissolution of poorly soluble
particles following deposition in the respiratory tract. However, poorly soluble particles that become
phagocytosed may slowly dissolve in the acidic (pH of 4.3-5.3) environment of the phagolysosome
to be released in their solubilized form from the cell and potential move across the epithelium into
the bloodstream. The dissolution rate is inversely related to particle size and directly related to
specific surface area (Kreyling and Scheuch, 2000, 056281) and facilitated by the acidic
environment of the macrophage (Kreyling, 1992, 067243).
There is considerable evidence that soluble particles depositing in the bronchial airways are
cleared by dual mechanisms (Bennett and Ilowite, 1989, 000835; Lay et al., 2003, 155920; Matsui
et al., 1998, 040405; Sakagami et al., 2002, 156936; Wagner and Foster, 2001, 156143). The relative
contribution of their removal by transepithelial absorption versus mucociliary clearance is likely a
function of both the molecular size and water or lipid solubility of the material (Enna and Schanker,
1972, 155767; Huchon et al., 1987, 024923; Oberdorster, 1988, 006857; Sakagami et al., 2002,
156936). Furthermore, the rate of mucociliary transport for soluble particles may be less than that of
insoluble particles (Lay et al., 2003, 155920). Consequently, non-permeating hydrophilic solutes
may remain in contact with the airway epithelium for a longer period than insoluble particles. This
may be due to diffusion of a greater portion of the solute into the periciliary sol layer which may be
transported less efficiently than the mucus layer during mucociliary clearance. Bronchial blood flow
has also been shown to modulate airway retention of soluble particles (Wagner and Foster, 2001,
156143). i.e., decreasing blood flow increases airway retention of soluble particles.
As an example of how transport of soluble components of PM may clear the lung by
transepithelial absorption, Wallenborn et al. (2007, 156144) measured elemental content of lungs,
plasma, heart, and liver of healthy male WKY rats (12-15 wk old) 4 or 24 h following a single
intratracheal (IT) instillation of saline or 8.33 mg/kg of oil combustion PM containing a variety of
transition metals with differing water and acid solubility. Metals with high water solubility and
relatively high concentration in oil combustion PM were increased in extrapulmonary organs.
Elements with low water or acid solubility, like silicon and aluminum, were not detected in
extrapulmonary tissues despite decreased levels in the lung suggesting they cleared the lung
primarily by mucociliary clearance. Thus, PM-associated metals deposited in the lung may be
released into systemic circulation at different rates depending on their water/acid solubility.
The amount and type of water soluble or leachable metals associated with PM varies with
location and by source. Furthermore some metals such as Zn, Cu, and Fe are essential to body
function while others such as vanadium and nickel are nonessential metals. Consequently the body
and the lung have different ways of dealing with excesses in inhaled soluble metals associated with
PM. Bioavailability, and potentially the toxicity, of leachable metals may be altered by protein
binding within the lung and blood as well as the affinities of these binding sites. For example Zn is
tightly regulated by a variety of metal binding proteins, including metallothionein and a family of Zn
specific transporters. In the plasma, Zn binds to many proteins, including a2-macroglobulin and
albumin. Zn is an example of a common abundant water soluble metal in ambient air that may
contribute to increased respiratory and cardiovascular disease risk associated with PM exposure.
Wallenborn et al. (2009, 191172) recently showed that soluble Zn sulfate (in the form of °Zn, a rare
isotope of Zn) introduced into the lungs by instillation not only reaches, but accumulates in
extrapulmonary organs, including the heart and liver, following pulmonary exposure. However, the
retention of greater than 50% of Zn in the lung at 4 h post-instillation suggested that the
transepithelial absorption of soluble Zn was indeed slowed by binding to proteins in the lungs. While
it could not be ascertained if 70Zn measured in the heart was replacing endogenous Zn pools, any
accumulation in cardiac Zn levels could lead to mitochondrial dysfunction and ion channel
disruption, possibly explaining adverse cardiac effects from inhalation of Zn-rich PM. Effects of Zn
instillation on epithelial integrity were not evaluated.
4.4.2. Factors Modulating Clearance
A number of studies have evaluated the epith
99mTc-diethylenetriaminepentaacetic acid (99mTc-Dl
0.57 nm). These studies are the basis for much of the discussion in this section.
A number of studies have evaluated the epithelial permeability by measuring the clearance of
99mTc-diethylenetriaminepentaacetic acid (99mTc-DTPA), a small hydrophilic solute (492 daltons,
December 2009 4-27
-------
4.4.2.1. Age
In humans, the clearance of water-soluble particles (99mTc-DTPA) from the alveolar epithelium
generally slows with increasing age (Braga et al., 1996, 156289; Pigorini et al., 1988, 156027).
However, Tankersley et al. (2003, 096363) recently showed enhanced permeability of soluble
particles (99mTc-DTPA) in terminally senescent mice just before death, suggesting that a
disintegration of the epithelial barrier may be a feature of lung homeostatic loss during this period of
terminal senescence.
4.4.2.2. Physical Activity
The transepithelial transport rates of soluble particles, 99mTc-DTPA, have also been found to
increase during exercise (Hanel et al., 2003, 155826; Lorino et al., 1989, 155946; Meignan et al.,
1986, 156752). This enhancement was linked to increases in VT associated with exercise (Lorino et
al., 1989, 155946). Regionally, this effect was dominated by increased apical lung clearance and
attributed to an increase in apical blood flow (Meignan et al., 1986, 156752). The increased
permeability with exercise appears to resolve to baseline after a short period post exercise, i.e.,
within a couple hours (Hanel et al., 2003, 155826).
4.4.2.3. Disease
Because the integrity of the epithelial surface lining of the lungs may be damaged from lung
disease, particles (either insoluble or soluble) may gain greater access to the interstitium, lymph, and
blood stream. Damage to the epithelial barrier is most likely to acutely affect transepithelial transport
rates of soluble particles. From bronchial biopsies, Laitinen et al. (1985, 037521) found various
degrees of epithelial damage, from loosening of tight junctions to complete denudation of the airway
epithelium, in asthmatics. Consistent with these findings, Ilowite et al. (1989, 156584) found that
asthmatics had increased permeability of the bronchial mucosa to the hydrophilic solute
99mTc-DTPA. On the other hand, a more recent study in a sheep model showed that the presence of
bronchial edema could slow the uptake of soluble DTPA into the blood and enhanced retention in the
airways, likely within the expanded interstitial barrier (Foster and Wagner, 2001, 155778). Both a
leaky epithelial barrier and expanded interstitial barrier associated with asthma may result in
enhanced exposure of submucosal immune and smooth muscle cells to xenobiotic substances.
Alveolar epithelial permeability was also shown to be affected by the presence of lung
inflammation. The most common finding has been a clear increase in alveolar permeability induced
by cigarette smoking (Jones et al., 1980, 155883). This effect appears to be dependent on the recent
cigarette smoke exposure as indexed by carboxyhaemoglobin (Jones et al., 1983, 155884) and is
rapidly reversible within a week of smoking cessation (Mason et al., 1983, 013169). In fact, Huchon
et al. (1984, 156576) demonstrated that COPD patients who have stopped smoking have normal
clearance of 9ymTc-DTPA.
In general, increased alveolar permeability to 99mTc-DTPA has been found to be associated
with any lung syndrome characterized by pulmonary edema. While the trans-alveolar transport of a
small solute like DTPA is very sensitive to even mild acute lung injury (such as that associated with
even mild cigarette smoking), increased transport rates of larger molecules (>100K daltons) across
the alveolar epithelium require more severe damage like that seen in adult respiratory distress
syndrome (ARDS) (Braude et al., 1986, 155701; Peterson et al., 1989, 024922). Interstitial lung
disease and pulmonary fibrosis are also characterized by increased alveolar permeability (Antoniou
et al., 2006, 156220; Bodolay et al., 2005, 156280; Watanabe et al., 2007, 157115). Interestingly,
these recent studies have also shown that the increased permeability in these patients could be
corrected with immunosuppressive/steroid treatments (Bodolay et al., 2005, 156280; Watanabe et al.,
2007, 157115). Furthermore, studies of DTPA clearance in bleomycin injured dogs, a model of
pulmonary fibrosis, suggest that the enhanced permeability is associated with the initial acute phase
of the lung damage, with clearance rates returning to normal as chronic fibrosis developed over time
(Sugaetal., 2003,157024).
Finally, as evidence of lung complications associated with non-insulin dependent diabetes
(type 2), Lin et al. (2002, 155932) found impairment of alveolar integrity as shown by increased
transport rates of both hydrophilic and lipophilic solutes from the lungs in these patients. By
contrast, a number of other studies have found epithelial permeability reduced, i.e., slower transport
December 2009 4-28
-------
rates, in diabetes (Caner et al., 1994, 156320; Mousa et al, 2000, 156786; Ozsahin et al., 2006,
156833) that may be related to disease duration and metabolic control (Ozsahin et al., 2006,
156833) . These findings are consistent with thickening of alveolar basement membrane detected in
autopsies of diabetes patients (Weynand et al., 1999, 157140).
4.4.2.4. Concurrent Exposures
The integrity of the alveolar epithelium may be disrupted by co-pollutants such that soluble
components of inhaled particles can more easily enter the interstitium and blood stream. Like active
cigarette smoking discussed previously, Beadsmoore et al. (2007, 156259) showed clearance half-
times in healthy passive smokers to be shorter compared with healthy non-smokers but still longer
than in healthy smokers. These findings show a progressive increase in epithelial permeability with
exposure to cigarette smoke. Similarly, acute exposure of humans to 0.4 ppm ozone for 2 h with
intermittent exercise has been shown to alter epithelial integrity and increase clearance of soluble
hydrophilic particles from the alveolar surfaces of the lung (Kehrl et al., 1987, 040824). This effect
persists to at least 24 h post-exposure even following lower exposure levels (0.24 ppm average for
130 min) of ozone (Foster and Stetkiewicz, 1996, 079920). Similarly, 0.8 ppm O3 exposure for 2 h
in rats caused increased permeability to macromolecules at all levels of the respiratory tract (Bhalla
et al., 1986, 040407) that persisted in the alveolar region beyond 24 h post-exposure. Cohen et al.
(1997, 009213) may have best illustrated the competing effects of mucociliary and trans epithelial
transport by showing that coexposure to ozone affected the retention of inhaled chromium in rats
differently depending on its solubility. In its soluble potassium chromate form, ozone decreased the
retention of chromium, but when chromium was inhaled as insoluble barium chromate, its retention
in the lung was increased by ozone coexposure. Similarly, a study that showed decreased clearance
of insoluble cesium oxide particles following influenza infection also showed a virus-induced
enhancement of clearance for a soluble cesium chloride (Lundgren et al., 1978, 155950). Chang et
al. (2005, 097776) also recently showed that UF carbon black acts through a reactive oxygen species
(ROS) dependent pathway to increase epithelial permeability in mice.
Chronic exposure to other particulate or gaseous pollutants does not always led to increased
epithelial permeability. Studying subjects with a variety of occupational exposures, Kaya et al.
(2006, 156632) showed that nonsmoking welders actually have decreased epithelial permeability
relative to nonsmoking control subjects, and occupational exposure of painters to isocyanates has no
effect on bronchoalveolar epithelial permeability (Kaya et al., 2003, 156631).
4.4.3. Summary
The healthy airway and alveolar epithelium is generally impermeable to very large insoluble
macromolecules and particles. Water and acid soluble particles may more rapidly move through the
epithelium as they dissolve on the airway surface or within the phagolysomes of macrophages. The
presence of airway inflammation in a variety of airway diseases (e.g., asthma, fibrosis, ARDS,
pulmonary edema, inflammation from smoking) alters epithelial integrity to allow more rapid
movement of these solutes into the bloodstream. While diabetics are another group recently shown to
have increased susceptibility to particulate air pollution (Zanobetti and Schwartz, 2002, 034821). it
is unclear whether transport of soluble particles across the epithelium is affected in these patients. In
general, it appears that coexposure to irritant pollutants results in a disruption of epithelial integrity
and macrophage function which, on the one hand, retards mucociliary and alveolar clearance, but
also allows for a more rapid movement of soluble constituents across the epithelial surface into the
interstitium and blood stream. Alterations in epithelial permeability by disease, pollutant exposure,
or infection may partially explain increased susceptibility to PM associated with these co-conditions.
December 2009 4-29
-------
Chapter 4 References
ACGIH. (2005). TLVs and BEIs: Based on the documentation of the threshold limit values for chemical substances and
physical agents and biological exposure indices. American Conference of Governmental Industrial Hygienists.
Cincinnati, Ohio. 156188
Adamson I; Prieditis H. (1995). Response of mouse lung to carbon deposition during injury and repair. Environ Health
Perspect, 103: 72-76. 189982
Alexis N; Soukup J; Nierkens S; Becker S. (2001). Association between airway hyperreactivity and bronchial macrophage
dysfunction in individuals with mild asthma. Am J Physiol Lung Cell Mol Physiol, 280: L369-L375. 190013
Anselm A; Heibel T; Gebhart J; Perron GA. (1990). In vivo studies of growth factors of sodium chloride particles in the
human respiratory tract. J Aerosol Sci, 21: S427-430. 156217
Anthony TR; FlynnMR. (2006). Computational fluid dynamics investigation of particle inhalability. J Aerosol Sci, 37:
750-765. 155659
Antonini JM; Roberts JR; Clarke RW; Yang HM; Barger MW; Ma JYC; Weissman DN. (2001). Effect of age on respiratory
defense mechanisms: pulmonary bacterial clearance in Fischer 344 rats after intratracheal instillation of Listeria
monocytogenes. Chest, 120: 240-249. 156219
Antoniou KM; Malagari K; Tzanakis N; Perisinakis K; Symvoulakis EK; Karkavitsas N; Siafakas NM; Bouros D. (2006).
Clearance of technetium-99m-DTPA and HRCT findings in the evaluation of patients with Idiopathic Pulmonary
Fibrosis. BMC Pulm Med, 6: 4. 156220
Aschner M; Erikson KM; Dorman DC. (2005). Manganese Dosimetry: Species Differences and Implications for
Neurotoxicity. Grit Rev Toxicol, 35: 1-32. 155663
Asgharain B; Kelly JT; Tewksbury EW. (2003). Respiratory deposition and inhalability of monodisperse aerosols in Long-
Evans rats. Toxicol Sci, 71: 104-111. 153068
Asgharian B; Hofmann W; Miller F J. (2001). Mucociliary clearance of insoluble particles from the tracheobronchial
airways of the human lung. J Aerosol Sci, 32: 817-832. 017025
Bailey M; Dorrian M; Birchall A. (1995). Implications of airway retention for radiation doses from inhaled radionuclides. J
Aerosol Med, 8: 373-390. 190057
Balashazy I; Hofmann W; Heistracher T. (1999). Computation of local enhancement factors for the quantification of
particle deposition patterns in airway bifurcations. J Aerosol Sci, 30: 185-203. 043201
Balashazy I; Hofmann W; Heistracher T. (2003). Local particle deposition patterns may play a key role in the development
of lung cancer. J Appl Physiol, 94: 1719-1725. 155671
Beadsmoore C; Cheow HK; Szczepura K; Ruparelia P; Peters AM. (2007). Healthy passive cigarette smokers have
increased pulmonary alveolar permeability. Nucl Med Comm, 28: 15-11. 156259
Becquemin MH; Swift DL; Bouchikhi A; Roy M; Teillac A. (1991). Particle deposition and resistance in the noses of adults
and children. Eur Respir J, 4: 694-702. 009187
Becquemin MM; Bertholon JF; Bouchikhi A; Malarbet JL; Roy M. (1999). Oronasal Ventilation Partitioning in Adults and
Children: Effect on Aerosol Deposition in Airways. Radiat Prot Dosimetry, 81: 221-228. 155679
Bennett W; Messina M; Smaldone G (1985). Effect of exercise on deposition and subsequent retention of inhaled particles.
J Appl Physiol, 59: 1046-1054. 190034
Bennett W; Zeman K; Jarabek A. (2003). Nasal contribution to breathing with exercise: effect of race and gender. J Appl
Physiol, 95: 497-503. 191977
Bennett WD; Ilowite JS. (1989). Dual pathway clearance of 99mTc-DTPA from the bronchial mucosa. Am Rev Respir Dis,
139: 1132-1138.000835
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009 4-30
-------
Bennett WD; Zeman KL. (1998). Deposition of fine particles in children spontaneously breathing at rest. Inhal Toxicol, 10:
831-842. 076182
Bennett WD; Zeman KL. (2004). Effect of body size on breathing pattern and fine-particle deposition in children. J Appl
Physiol, 97: 821-826. 155686
Bennett WD; Zeman KL. (2005). Effect of Race on Fine Particle Deposition for Oral and Nasal Breathing. Inhal Toxicol,
17: 641-648. 155687
Bennett WD; Zeman KL; Jarabek AM. (2008). Nasal contribution to breathing and fine particle deposition in children
versus adults. J Toxicol Environ Health A, 71: 227-237. 156269
Bennett WD; Zeman KL; Kim C. (1996). Variability of fine particle deposition in healthy adults: effect of age and gender.
Am J Respir Grit Care Med, 153: 1641-1647. 083284
Bennett WD; Zeman KL; Kim C; Mascarella J. (1997). Enhanced deposition of fine particles in COPD patients
spontaneously breathing at rest. Inhal Toxicol, 9: 1-14. 078839
Bermudez E; Mangum JB; Asgharian B; Wong BA; Reverdy EE; Janszen DB; Hext PM; Warheit DB; Everitt JI. (2002).
Long-term pulmonary responses of three laboratory rodent species to subchronic inhalation of pigmentary titanium
dioxide particles. Toxicol Sci, 70: 86-97. 055578
Bermudez E; Mangum JB; Wong BA; Asgharian B; Hext PM; Warheit DB; Everitt JI. (2004). Pulmonary responses of
mice, rats, and hamsters to subchronic inhalation of ultrafine titanium dioxide particles. Toxicol Sci, 77: 347-357.
056707
Bhalla DK; Mannix RC; Kleinman MT; Crocker TT. (1986). Relative permeability of nasal, tracheal, and bronchoalveolar
mucosa to macromolecules in rats exposed to ozone. J Toxicol Environ Health, 17: 269-283. 040407
Bodolay E; Szekanecz Z; Devenyi K; Galuska L; Csipo I; Vegh J; Garai I; Szegedi G. (2005). Evaluation of interstitial lung
disease in mixed connective tissue disease (MCTD). Rheumatology, 44: 656-661. 156280
Braga F; Mango JC; Souza JF; Ferrioli E; De Andrade J; lazigi N. (1996). Age-related reduction in 99Tcm-DTPA alveolar-
capillary clearance in normal humans. Nucl Med Comm, 17: 971-974. 156289
Braude S; Nolop KB; Hughes JMB; Barnes PJ; Royston D. (1986). Comparison of lung vascular and epithelial
permeability indices in the adult respiratory distress syndrome. Am Rev Respir Dis, 133: 1002-1005. 155701
Brown J; Bennett W (2004). Deposition of coarse particles in cystic fibrosis: model predictions versus experimental
results. J Aerosol Med, 17: 239-248. 190032
Brown JS. (2005). Particle inhalability at low wind speeds. Inhal Toxicol, 17: 831-837. 156299
Brown JS; Wilson WE; Grant LD. (2005). Dosimetric comparisons of particle deposition and retention in rats and humans.
Inhal Toxicol, 17: 355-385. 089308
Brown JS; Zeman KL; Bennett WD. (2002). Ultrafine particle deposition and clearance in the healthy and obstructed lung.
Am J Respir Grit Care Med, 166: 1240-1247. 043216
Burch WM. (2002). Comment on "Passage of inhaled particles into the blood circulation in humans". Circulation, 106:
el41-e!42. 056754
Caner B; Ugur O; Bayraktar M; Ulutuncel N; Mentes T; Telatar F; Bekdik C. (1994). Impaired lung epithelial permeability
in diabetics detected by technetium-99m-DTPA aerosol scintigraphy. J Nucl Med, 35: 204-206. 156320
Chadha TS; Birch S; Sackner MA. (1987). Oronasal distribution of ventilation during exercise in normal subjects and
patients with asthma and rhinitis. Chest, 92: 1037-1041. 037365
Chang C-C; Chiu H-F; Wu Y-S; Li Y-C; Tsai M-L; Shen C-K; Yang C-Y (2005). The induction of vascular endothelial
growth factor by ultrafine carbon black contributes to the increase of alveolar-capillary permeability. Environ
Health Perspect, 113: 454-460. 097776
Chen J; Tan M; Nemmar A; Song W; Dong M; Zhang G; Li Y (2006). Quantification of extrapulmonary translocation of
intratracheal-instilled particles in vivo in rats: effect of lipopolysaccharide. Toxicology, 222: 195-201. 147267
Cheng K-H; Cheng Y-S; Yeh H-C; Guilmette RA; Simpson SQ; Yang Y-H; Swift DL. (1996). In vivo measurements of
nasal airway dimensions and ultrafine aerosol deposition in the human nasal and oral airways. J Aerosol Sci, 27:
785-801.047520
December 2009 4-31
-------
Chinard FP. (1980). The alveolar-capillary barrier: some data and speculations. Microvasc Res, 19: 1-17. 156341
Churg A; Stevens B; Wright JL. (1998). Comparison of the uptake of fine and ultrafine TiO2 in a tracheal explant system.
Am J Physiol, 274: L81-L86. 085815
Cohen MD; Zelikoff JT; Chen L-C; Schlesinger RB. (1997). Pulmonary retention and distribution of inhaled chromium:
effects of particle solubility and coexposure to ozone. Inhal Toxicol, 9: 843-865. 009213
Creasia DA; Poggenburg JK Jr; Nettesheim P. (1976). Elution of benzo[a]pyrene from carbon particles in the respiratory
tract of mice. J Toxicol Environ Health, 1: 967-975. 059713
Dai YT; Juang YJ; Wu Y; Breysse PN; Hsu DJ. (2006). In vivo measurements of inhalability of ultralarge aerosol particles
in calm air by humans. J Aerosol Sci, 37: 967-973. 156377
DeLorenzo AJD. (1970). The olfactory neuron and the blood-brain barrier. In Taste and Smell in Vertebrates (pp. 151-175).
London: Churchill Livingstone. 156391
Dorman DC; McManus BE; Parkinson CU; Manuel CA; McElveen AM; Everitt JI. (2004). Nasal Toxicity of Manganese
Sulfate and Manganese Phosphate in Young Male Rats Following Subchronic (13-Week) Inhalation Exposure. Inhal
Toxicol, 16: 481-488. 155752
Dorman DC; Struve MF; James RA; Marshall MW; Parkinson CU; Wong BA. (2001). Influence of particle solubility on
the delivery of inhaled manganese to the rat brain: manganese sulfate and manganese tetroxide pharmacokinetics
following repeated (14-day) exposure. Toxicol Appl Pharmacol, 170: 79-87. 055433
Edetsberger M; Gaubitzer E; Valic E; Waigmann E; Kohler G. (2005). Detection of nanometer-sized particles in living cells
using modern fluorescence fluctuation methods. Biochem Biophys Res Commun, 332: 109-116. 155759
Elder A; Gelein R; Silva V; Feikert T; Opanashuk L; Carter J; Potter R; Maynard A; Ito Y; Finkelstein J; Oberdorster G.
(2006). Translocation of inhaled ultrafine manganese oxide particles to the central nervous system. Environ Health
Perspect, 114: 1172-1178. 089253
Enna SJ; Schanker LS. (1972). Absorption of drugs from the rat lung. Am J Physiol, 223: 1227-1231. 155767
Falk R; Philipson K; Svartengren M; Jarvis N; Bailey M; Camner P. (1997). Clearance of particles from small ciliated
airways. Exp Lung Res, 23: 495-515. 086080
Farkas A; Balashazy I. (2008). Quantification of particle deposition in asymmetrical tracheobronchial model geometry.
Comput Biol Med, 38: 508-18. 157358
Farkas A; Balashazy I; SzQcs K. (2006). Characterization of Regional and Local Deposition of Inhaled Aerosol Drugs in
the Respiratory System by Computational Fluid and Particle Dynamics Methods. J Aerosol Med, 19: 329-343.
155771
Ferin J; Oberdorster G; Penney DP. (1992). Pulmonary retention of ultrafine and fine particles in rats. Am J Respir Cell
Mol Biol, 6: 535-542. 044401
Finlay WH; Martin AR. (2008). Recent advances in predictive understanding of respiratory tract deposition. J Aerosol Med
Pulm Drug Deliv, 21: 189-206. 155776
Fishman AP; Elias JA. (1980). Fishman's Pulmonary diseases and disorders. New York: McGraw-Hill. 156436
Foster WM; Stetkiewicz PT. (1996). Regional clearance of solute from the respiratory epithelia: 18—20 h postexposure to
ozone. J Appl Physiol, 81: 1143-1149. 079920
Foster WM; Wagner EM. (2001). Bronchial edema alters 99mTc-DTPA clearance from the airway surface in sheep. JAppl
Physiol, 91: 2567-2573. 155778
Geiser M; Rothen-Rutishauser B; Kapp N; Schurch S; Kreyling W; Schulz H; Semmler M; Im Hof V; Heyder J; Gehr P.
(2005). Ultrafine particles cross cellular membranes by nonphagocytic mechanisms in lungs and in cultured cells.
Environ Health Perspect, 113: 1555-1560. 087362
Geys J; Coenegrachts L; Vercammen J; Engelborghs Y; Nemmar A; Nemery B; Hoet PHM. (2006). In vitro study of the
pulmonary translocation of nanoparticles A preliminary study. Toxicol Lett, 160: 218-226. 155789
Goodman RM; Yergin BM; Landa JF; Golinvaux MH; Sackner MA. (1978). Relationship of smoking history and
pulmonary function tests to tracheal mucous velocity in nonsmokers, young smokers, ex-smokers, and patients with
chronic bronchitis. Am Rev Respir Dis, 117: 205-214. 071130
December 2009 4-32
-------
Greenspan BJ; Morrow PE; Ferin J. (1988). Effects of aerosol exposures to cadmium chloride on the clearance of titanium
dioxide from the lungs of rats. Exp Lung Res, 14: 491-499. 045031
Grgic B; Finlay WH; Burnell PKP; Heenan AF. (2004). In vitro intersubject and intrasubject deposition measurements in
realistic mouth-throat geometries. J Aerosol Sci, 35: 1025-1040. 155810
Hanel B; Law I; Mortensen J. (2003). Maximal rowing has an acute effect on the blood-gas barrier in elite athletes. J Appl
Physiol, 95: 1076-1082. 155826
Heistracher T; Hofmann W. (1997). Flow and deposition patterns in successive airway bifurcations. Presented at In: Cherry,
N.; Ogden, T., eds. Inhaled Particles VIII: proceedings of an international symposium on inhaled particles organised
by the British Occupational Hygiene Society; August 1996; Cambridge, UK. Ann. Occup. Hyg. 41(suppl): 537-
542. 047514
Ho JC; Chan KN; Hu WH; Lam WK; Zheng L; Tipoe GL; Sun J; Leung R; Tsang KW. (2001). The effect of aging on nasal
mucociliary clearance, beat frequency, and ultrastructure of respiratory cilia. Am J Respir Crit Care Med, 163: 983-
988. 156549
Hofmann W; Asgharian B. (2003). The effect of lung structure on mucociliary clearance and particle retention in human
and rat lungs. Toxicol Sci, 73: 448-456. 055579
Hofmann W; Balashazy I; Heistracher T; Koblinger L. (1996). The significance of particle deposition patterns in bronchial
airway bifurcations for extrapolation modeling. Aerosol Sci Technol, 25: 305-327. 047515
Hofmann W; Martonen TB; Graham RC. (1989). Predicted deposition of nonhygroscopic aerosols in the human lung as a
function of subject age. J Aerosol Med, 2: 49-68. 006922
Hsu DJ; Swift DL. (1999). The measurements of human inhalability of ultralarge aerosols in calm air using mannikins. J
Aerosol Sci, 30: 1331-1343. 155855
Huchon GJ; Montgomery AB; Lipavsky A; Hoeffel JM; Murray JF. (1987). Respiratory clearance of aerosolized
radioactive solutes of varying molecular weight. J Nucl Med, 28: 894-902. 024923
Huchon GJ; Russell JA; Barritault LG; Lipavsky A; Murray JF. (1984). Chronic air-flow limitation does not increase
respiratory epithelial permeability assessed by aerosolized solute, but smoking does. Am Rev Respir Dis, 130: 457-
460. 156576
ICRP (1994). Human respiratory tract model for radiological protection: a report of a task group of the International
Commission on Radiological Protection. Ann ICRP, 24: 1-482. 006988
Ilowite JS; Bennett WD; Sheetz MS; Groth ML; Nierman DM. (1989). Permeability of the bronchial mucosa to 99mTc-
DTPAin asthma. Am Rev Respir Dis, 139: 1139-1143. 156584
International Life Sciences Institute. (2000). The relevance of the rat lung response to particle overload for human risk
assessment: A workshop consensus report. Inhal Toxicol, 12: 1-17. 002892
Isaacs KK; Schlesinger RB; Martonen TB. (2006). Three-dimensional computational fluid dynamics simulations of particle
deposition in the tracheobronchial tree. J Aerosol Med, 19: 344-352. 155861
ISO. (2008). Nanotechnologies — Terminology and definitions for nano-objects — Nanoparticle, nanofibre and nanoplate.
International Organization for Standardization. Geneva. 190066
James DS; Stidley CA; Lambert WE; Chick TW; Mermier CM; Samet JM. (1997). Oronasal distribution of ventilation at
different ages. Arch Environ Occup Health, 52: 118-123. 042422
Jarabek AM; Asgharian B; Miller FJ. (2005). Dosimetric adjustments for interspecies extrapolation of inhaled poorly
soluble particles (PSP). Inhal Toxicol, 17: 317-334. 056756
Jeffers DE. (2005). Relative magnitudes of the effects of electrostatic image and thermophoretic forces on particles in the
respiratory tract. Radiat Prot Dosimetry, 113: 189-194. 156608
Jones JG; Minty BD; Lawler P; Hulands G; Crawley JC; Veall N. (1980). Increased alveolar epithelial permeability in
cigarette smokers. Lancet, 1: 66-68. 155883
Jones JG; Minty BD; Royston D; Royston JP (1983). Carboxyhaemoglobin and pulmonary epithelial permeability in man.
Thorax, 38: 129-133. 155884
December 2009 4-33
-------
Kapp N; Kreyling W; Schulz H; Im Hof V; Gehr P; Semmler M; Geiser M. (2004). Electron energy loss spectroscopy for
analysis of inhaled ultrafine particles in rat lungs. Microsc Res Tech, 63: 298-305. 156624
Kaya E; Fidan F; Unlu M; Sezer M; Tetik L; Acar M. (2006). Evaluation of alveolar clearance by Tc-99m DTPA
radioaerosol inhalation scintigraphy in welders. Ann Nucl Med, 20: 503-510. 156632
Kaya M; Salan A; Tabakoglu E; Aydogdu N; Berkarda S. (2003). The bronchoalveolar epithelial permeability in house
painters as determined by Tc-99m DTPA aerosol scintigraphy. Ann Nucl Med, 17: 305-308. 156631
Kehrl HR; Vincent LM; Kowalsky RJ; HorstmanDH; O'Neil JJ; McCartney WH; Bromberg PA. (1987). Ozone exposure
increases respiratory epithelial permeability in humans. Am J Respir Crit Care Med, 135: 1124-1128. 040824
Kelly JT; Asgharian B; Wong BA. (2005). Inertial Particle Deposition in a Monkey Nasal Mold Compared with that in
Human Nasal Replicas. Inhal Toxicol, 17: 823-830. 155894
Kim CS. (2000). Methods of calculating lung delivery and deposition of aerosol particles. Respir Care, 45: 695-711.
013112
Kim CS; Hu SC. (1998). Regional deposition of inhaled particles in human lungs: comparison between men and women. J
ApplPhysiol, 84: 1834-1844. 086066
Kim CS; Iglesias AJ. (1989). Deposition of inhaled particles in bifurcating airway models: I inspiratory deposition. J
Aerosol Med, 2: 1-14. 078539
Kim CS; Iglesias AJ; Garcia L. (1989). Deposition of inhaled particles in bifurcating airway models: II expiratory
deposition. J Aerosol Med, 2: 15-27. 078538
Kim CS; Jaques PA. (2000). Respiratory dose of inhaled ultrafine particles in healthy adults. Philos Transact A Math Phys
Eng Sci, 358: 2693-2705. 012811
Kimbell JS. (2006). Nasal Dosimetry of Inhaled Gases and Particles: Where Do Inhaled Agents Go in the Nose?. Toxicol
Pathol, 34: 270-273. 155902
Kreyling WG (1992). Intracellular particle dissolution in alveolar macrophages. Environ Health Perspect, 97: 121-126.
067243
Kreyling WG; Blanchard JD; Godleski JJ; Haeussermann S; Heyder J; Hutzler P; Schulz H; Sweeney TD; Takenaka S;
Ziesenis A. (1999). Anatomic localization of 24- and 96-h particle retention in canine airways. J Appl Physiol, 87:
269-284. 039175
Kreyling WG; Scheuch G (2000). Clearance of particles deposited in the lungs. In Gehr, P.; Heyder, J. (Ed.),Particle-lung
interactions (pp. x). New York, NY: Marcel Dekker, Inc. 056281
Kreyling WG; Semmler M; Erbe F; Mayer P; Takenaka S; Schulz H; Oberdorster G; Ziesenis A. (2002). Translocation of
ultrafine insoluble iridium particles from lung epithelium to extrapulmonary organs is size dependent but very low.
J Toxicol Environ Health A, 65: 1513-1530. 037332
Laitinen LA; Heino M; Laitinen A; Kava T; Haahtela T. (1985). Damage of the airway epithelium and bronchial reactivity
in patients with asthma. Am Rev Respir Dis, 131: 599-606. 037521
Lay J; Alexis N; Zeman K; Peden D; Bennett W. (2009). In vivo uptake of inhaled particles by airway phagocytes is
enhanced in patients with mild asthma compared with normal volunteers. Thorax, 64: 313-320. 190060
Lay JC; Stang MR; Fisher PE; Yankaskas JR; Bennett WD. (2003). Airway Retention of Materials of Different Solubility
following Local Intrabronchial Deposition in Dogs. JAerosolMed, 16: 153-166. 155920
Lee KP; Henry NW III; Trochimowicz HJ; Reinhardt CF. (1986). Pulmonary response to impaired lung clearance in rats
following excessive TiO2 dust deposition. Environ Res, 41: 144-167. 067629
Lee KP; Trochimowicz HJ; Reinhardt CF. (1985). Transmigration of titanium dioxide (TiO2) particles in rats after
inhalation exposure. Exp Mol Pathol, 42: 331-343. 067628
Lin CC; Chang CT; Li TC; Kao A. (2002). Objective Evidence of Impairment of Alveolar Integrity in Patients with Non-
Insulin-Dependent Diabetes Mellitus Using Radionuclide Inhalation Lung Scan. Lung, 180: 181-186. 155932
Lorino AM; Meignan M; Bouissou P; Atlan G. (1989). Effects of sustained exercise on pulmonary clearance of aerosolized
99mTc-DTPA. J Appl Physiol, 67: 2055-2059. 155946
December 2009 4-34
-------
Lundgren DL; Hahn FF; Grain CR; Sanchez A. (1978). Effect of influenza virus infection on the pulmonary retention of
inhaled 144Ce and subsequent survival of mice. Health Phys, 34: 557-567. 155950
Martonen TB; Yang Y; Xue ZQ. (1994). Effects of carinal ridge shapes on lung airstreams. Aerosol Sci Technol, 21: 119-
136.000847
Mason GR; Uszler JM; Effros RM; Reid E. (1983). Rapidly reversible alterations of pulmonary epithelial permeability
induced by smoking. Chest, 83: 6-11.013169
Matsui H; Randell SH; Peretti SW; Davis CW; Boucher RC. (1998). Coordinated clearance of periciliary liquid and mucus
from airway surfaces. J Clin Invest, 102: 1125-1131. 040405
Mauderly JL. (1997). Relevance of particle-induced rat lung tumors for assessing lung carcinogenic hazard and human
lung cancer risk. Environ Health Perspect, 105 (Suppl 5): 1337-1346. 084631
Meignan M; Rosso J; Leveau J; Katz A; Cinotti L; Madelaine G; Galle P. (1986). Exercise increases the lung clearance of
inhaled technetium-99m DTPA. J Nucl Med, 27: 274-280. 156752
Menache MG; Miller FJ; Raabe OG. (1995). Particle inhalability curves for humans and small laboratory animals. Ann
Occup Hyg, 39: 317-328. 006533
Miller FJ. (2000). Dosimetry of particles in laboratory animals and humans in relationship to issues surrounding lung
overload and human health risk assessment: a critical review. Inhal Toxicol, 12: 19-57. 011822
Mills NL; Amin N; Robinson SD; Anand A; Davies J; Patel D; de la Fuente JM; Cassee FR; Boon NA; Macnee W; Millar
AM; Donaldson K; Newby DE. (2006). Do inhaled carbon nanoparticles translocate directly into the circulation in
humans?. Am J Respir Grit Care Med, 173: 426-431. 088770
Moller W; Felten K; Sommerer K; Scheuch G; Meyer G; Meyer P; Haussinger K; Kreyling WG. (2008). Deposition,
retention, and translocation of ultrafine particles from the central airways and lung periphery. Am J Respir Crit Care
Med, 177: 426-432. 156771
Moller W; Haussinger K; Winkler-Heil R; Stahlhofen W; Meyer T; Hofmann W; Heyder J. (2004). Mucociliary and long-
term particle clearance in the airways of healthy nonsmoker subjects. J Appl Physiol, 97: 2200-2206. 155987
Morgan W; Ahmad D; Chamberlain M; Clague H; Pearson M; Vinitski S. (1984). The effect of exercise on the deposition
of an inhaled aerosol. Respir Physiol, 56: 327-338. 190035
Morrow PE. (1994). Mechanisms and significance of "particle overload". In Mohr U (Ed.),Toxic and Carcinogenic Effects
of Solid Particles in the Respiratory Tract (pp. 17-25). Washington, DC: ILSI Press . 006850
Mousa K; Onadeko BO; Mustafa HT; Mohamed M; Nabilla A; Omar A; Al-Bunni A; Elgazzar A. (2000). Technetium
99mTc-DTPA clearance in the evaluation of pulmonary involvement in patients with diabetes mellitus. Respir Med,
94: 1053-1056. 156786
Muhle H; Creutzenberg O; BellmannB; HeinrichU; Mermelstein R. (1990). Dust overloading of lungs: investigations of
various materials, species differences, and irreversibility of effects. J Aerosol Med, 1: S111-S128. 006853
Nemmar A; Hoet PHM; Vanquickenborne B; Dinsdale D; Thomeer M; Hoylaerts MF; Vanbilloen H; Mortelmans L;
Nemery B. (2002). Passage of inhaled particles into the blood circulation in humans. Circulation, 105: 411-414.
024914
Niinimaa V; Cole P; Mintz S; Shephard RJ. (1981). Oronasal distribution of respiratory airflow. Respir Physiol Neurobiol,
43: 69-75. 071758
Nikula KJ; Vallyathan V; Green FH Y; Hahn FF. (2001). Influence of exposure concentration or dose on the distribution of
particulate material in rat and human lungs. Environ Health Perspect, 109: 311-318. 016641
NRPB. (2004). Particle deposition in the vicinity of power lines and possible effects on health: Report of an independent
advisory group on non-ionising radiation and its ad hoc group on corona ions. National Radiological Protection
Board. Oxfordshire, England. 156815
Oberdorster G. (1988). Lung clearance of inhaled insoluble and soluble particles. J Aerosol Med, 1: 289-330. 006857
Oberdorster G. (1995). Lung particle overload: implications for occupational exposures to particles. Regul Toxicol
Pharmacol, 27: 123-135. 046596
December 2009 4-35
-------
Oberdorster G. (1996). Significance of particle parameters in the evaluation of exposure-dose-response relationships of
inhaled particles. Inhal Toxicol, 8 Supplement: 73-89. 039852
Oberdorster G. (2002). Toxicokinetics and effects of fibrous and nonfibrous particles. Inhal Toxicol, 14: 29-56. 021111
Oberdorster G; Ferin J; GeleinR; Soderholm SC; Finkelstein J. (1992). Role of the alveolar macrophage in lung injury:
studies with ultrafine particles. Environ Health Perspect, 97: 193-199. 045110
Oberdorster G; Ferin J; Lehnert BE. (1994). Correlation between particle size, in vivo particle persistence, and lung injury.
Environ Health Perspect, 102 : 173-179. 046203
Oberdorster G; Ferin J; Soderholm S; Gelein R; Cox C; Baggs R; Morrow PE. (1994). Increased pulmonary toxicity of
inhaled ultrafine particles: due to lung overload alone?. Ann Occup Hyg, 38: 295-302. 056285
Oberdorster G; Finkelstein JN; Johnston C; Gelein R; Cox C; Baggs R; Elder ACP (2000). Acute pulmonary effects of
ultrafine particles in rats and mice. Health Effects Institute. Boston. Report Number 96. 039014
Oberdorster G; Sharp Z; Atudorei V; Elder A; Gelein R; Kreyling W; Cox C. (2004). Translocation of inhaled ultrafine
particles to the brain. Inhal Toxicol, 16: 437-445. 055639
Oldham MJ; Robinson RJ. (2007). Predicted tracheobronchial and pulmonary deposition in a murine asthma model. Anat
Rec (Hoboken), 290: 1309-1314. 156003
Ozsahin K; Tugrul A; Mert S; Yilksel M; Tugrul G. (2006). Evaluation of pulmonary alveolo-capillary permeability in Type
2 diabetes mellitus Using technetium 99mTc-DTPA aerosol scintigraphy and carbon monoxide diffusion capacity. J
Diabetes Complications, 20: 205-209. 156833
Persson E; Henriksson J; Tallkvist J; Rouleau C; Tjalve H. (2003). Transport and subcellular distribution of intranasally
administered zinc in the olfactory system of rats and pikes. Toxicology, 191: 97-108. 051846
Peterson BT; Dickerson KD; James HL; Miller EJ; McLarty JW; Holiday DB. (1989). Comparison of three tracers for
detecting lung epithelial injury in anesthetized sheep. J Appl Physiol, 66: 2374-2383. 024922
Phalen RF; Oldham MJ. (2006). Aerosol dosimetry considerations. Clin Occup Environ Med, 5: 773-784. 156024
Phalen RF; Oldham MJ; Mautz WJ. (1989). Aerosol Deposition in the Nose As a Function of Body Size. Health Phys, 57:
299-305. 156023
Phalen RF; Oldham MJ; Wolff RK. (2008). The relevance of animal models for aerosol studies. J Aerosol Med Pulm Drug
Deliv, 21: 113-124. 156865
Pigorini F; Maini CL; Pau F; Giosue S. (1988). The influence of age on the pulmonary clearance of 99Tcm-DTPA
radioaerosol. Nucl Med Comm, 9: 965-971. 156027
Puchelle E; Zahm J-M; Bertrand A. (1979). Influence of age on bronchial mucociliary transport. Scand J Respir Dis, 60:
307-313.006863
Raabe OG; Al-Bayati MA; Teague SV; Rasolt A. (1988). Regional deposition of inhaled monodisperse, coarse, and fine
aerosol particles in small laboratory animals. In Inhaled particles VI: Proceedings of an international symposium
and workshop on lung dosimetry (pp. 53-63). Cambridge, U.K.: PergamonPress. 001439
RothC; ScheuchG; Stahlhofen W. (1993). Clearance of the human lungs for ultrafine particles. J Aerosol Sci, 24: S95-S96
.156928
Sakagami M; Byron PR; Venitz J; Rypacek F. (2002). Solute disposition in the rat lung in vivo and in vitro: determining
regional absorption kinetics in the presence of mucociliary escalator. J Pharmacol Sci, 91: 594-604. 156936
Schanker LS; Mitchell EW; Brown RA Jr. (1986). Species comparison of drug absorption from the lung after aerosol
inhalation or intratracheal injection. Drug Metab Dispos, 14: 79-88. 005100
Scheuch G; Gebhart J; Roth C. (1990). Uptake of electrical charges in the human respiratory tract during exposure to air
loaded with negative ions. J Aerosol Sci, 21: S439-S442. 006948
Schroeter JD; Kimbell JS; Asgharian B. (2006). Analysis of particle deposition in the turbinate and olfactory regions using
a human nasal computational fluid dynamics model. J Aerosol Med Pulm Drug Deliv, 19: 301-313. 156076
Semmler M; Seitz J; Erbe F; Mayer P; Heyder J; Oberdorster G; Kreyling WG (2004). Long-term clearance kinetics of
inhaled ultrafine insoluble iridium particles from the rat lung, including transient translocation into secondary
organs. Inhal Toxicol, 16: 453-459. 055641
December 2009 4-36
-------
Semmler-Behnke M; Takenaka S; Fertsch S; Wenk A; Seitz J; Mayer P; Oberdorster G; Kreyling WG. (2007). Efficient
elimination of inhaled nanoparticles from the alveolar region: evidence for interstitial uptake and subsequent
reentrainment onto airways epithelium. Environ Health Perspect, 115: 728-733. 156080
Smith JR; Bailey MR; Etherington G; Shutt AL; Youngman MJ. (2008). Effect of particle size on slow particle clearance
from the bronchial tree. Exp Lung Res, 34: 287-312. 190037
Snipes MB. (1996). Current information on lung overload in nonrodent mammals: contrast with rats. Inhal Toxicol, 8: 91-
109. 076041
Snipes MB; Harkema JR; Hotchkiss JA; Bice DE. (1997). Neutrophil Involvement in the Retention and Clearance of Dust
Intratracheally Instilled into the LUNGS of F344/N Rats. Exp Lung Res, 23: 65-84. 156092
Soderholm SC. (1985). Size-selective sampling criterial for inspirable mass fraction. In Particle size-selective sampling in
the workplace : report of the ACGIH Technical Committee on Air Sampling Procedures (pp. 27-32). Cincinnati,
Ohio: American Conference of Governmental Industrial Hygienists. 156992
Sousa SR; Moradas-Ferreira P; Saramango B; Viseu ML; Barbosa MA. (2004). Human serum albumin adsorption on TiO2
from single protein solutions and from plasma. Langmuir, 20: 9745-9754. 089866
Suga K; Yuan Y; Ogasawara N; Tsukuda T; Matsunaga N. (2003). Altered Clearance of Gadolinium
Diethylenetriaminepentaacetic Acid Aerosol from Bleomycin-injured Dog Lungs: Initial Observations. Am J Respir
Grit Care Med, 167: 1704-1710. 157024
Svartengren M; Falk R; Philipson K. (2005). Long-term clearance from small airways decreases with age. Eur Respir J, 26:
609-615. 157034
Tabachnik E; Muller N; Toye B; Levison H. (1981). Measurement of ventilation in children using the respiratory inductive
plethysmograph. J Pediatr, 99: 895-899. 157036
Takenaka S; Dornhofer-Takenaka H; Muhle H. (1986). Alveolar distribution of fly ash and of titanium dioxide after long-
term inhalation by Wistarrats. JAerosol Sci, 17: 361-364. 046210
Takenaka S; Karg E; Kreyling W; Lentner B; Moller W; Behnke-Semmler M; Jennen L; Walch A; Michalke B; Schramel P.
(2006). Distribution Pattern of Inhaled Ultrafine Gold Particles in the Rat Lung. Inhal Toxicol, 18: 73 3-740. 156110
Tankersley CG; Shank JA; Flanders SE; Soutiere SE; Rabold R; Mitzner W; Wagner EM. (2003). Changes in lung
permeability and lung mechanics accompany homeostatic instability in senescent mice. J Appl Physiol, 95: 1681-
1687. 096363
TobinMJ; ChadhaTS; Jenouri G; Birch SJ; GazerogluHB; SacknerMA. (1983). Breathing patterns. 1. Normal subjects.
Chest, 84: 202-205. 156122
Tran CL; Buchanan D; Cullen RT; Searl A; Jones AD; Donaldson K. (2000). Inhalation of poorly soluble particles II
Influence of particle surface area on inflammation and clearance. Inhal Toxicol, 12: 1113-1126. 013071
Tu KW; Knutson EO. (1984). Total deposition of ultrafine hydrophobic and hygroscopic aerosols in the human respiratory
system. Aerosol Sci Technol, 3: 453-465. 072870
U.S. EPA. (1996). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research Triangle
Park, NC. EPA/600/P-95/001aF-cF. 079380
U.S. EPA. (2004). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research Triangle
Park, NC. EPA/600/P-99/002aF-bF. 056905
Valberg P; Brain J; Sneddon S; LeMott S. (1982). Breathing patterns influence aerosol deposition sites in excised dog
lungs. J Appl Physiol, 53: 824-837. 190019
Vastag E; Matthys H; Kohler D; Gronbeck L; Daikeler G. (1985). Mucociliary clearance and airways obstruction in
smokers, ex-smokers and normal subjects who never smoked. Eur J Respir Dis, 139: 93-100. 157088
Vincent JH; Donaldson K. (1990). A dosimetric approach for relating the biological response of the lung to the
accumulation of inhaled mineral dust. Br J Ind Med, 47: 302-307. 002462
Wagner EM; Foster WM. (2001). Interdependence of bronchial circulation and clearance of 99mTc-DTPA from the airway
surface. J Appl Physiol, 90: 1275-1281. 156143
December 2009 4-37
-------
Wallenborn JG; Kovalcik KD; McGee JK; Landis MS; Kodavanti UP. (2009). Systemic translocation of (70)zinc: kinetics
following intratracheal instillation in rats. Toxicol Appl Pharmacol, 234: 25-32. 191172
Wallenborn JG; McGee John K; Schladweiler Mette C; Ledbetter Allen D; Kodavanti Urmila P. (2007). Systemic
translocation of particulate matter-associated metals following a single intratracheal instillation in rats. J Toxicol
Sci, 98: 231-239. 156144
Wang JX; Chen CY; Yu HW; Sun J; Li B; Li YF; Gao YX; He W; Huang YY; Chai ZF. (2007). Distribution of TiO 2
particles in the olfactory bulb of mice after nasal inhalation using microbeam SRXRF mapping techniques. Journal
of Radioanal Chem, 272: 527-531. 156146
Warheit DB; Hansen JF; Yuen IS; Kelly DP; Snajdr SI; Hartsky MA. (1997). Inhalation of high concentrations of low
toxicity dusts in rats results in impaired pulmonary clearance mechanisms and persistent inflammation. Toxicol
Appl Pharmacol, 145: 10-22. 086055
Warheit DB; Webb TR; Sayes CM; Colvin VL; Reed KL. (2006). Pulmonary instillation studies with nanoscale TiO2 rods
and dots in rats: Toxicity is not dependent upon particle size and surface area. Toxicol Sci, 91: 227-236. 088436
Watanabe N; Tanada S; Sasaki Y (2007). Pulmonary clearance of aerosolized 99mTc-DTPA in sarcoidosis I patients. Q J
Nucl Med Mol Imaging, 51: 82-90. 157115
Weynand B; Jonckheere A; Frans A; Rahier J; Saint-Luc CU. (1999). Diabetes mellitus Induces a Thickening of the
Pulmonary Basal Lamina. Respiration, 66: 14-19. 157140
Whaley SL; Muggenburg BA; Seiler FA; Wolff RK. (1987). Effect of aging on tracheal mucociliary clearance in beagle
dogs. JApplPhysiol,62: 1331-1334. 156153
Wiebert P; Sanchez-Crespo A; Falk R; Philipson K; Lundin A; Larsson S; Moller W; Kreyling W; Svartengren M. (2006).
No Significant Translocation of Inhaled 35-nm Carbon Particles to the Circulation in Humans. Inhal Toxicol, 18:
741-747. 156154
Wiebert P; Sanchez-Crespo A; Seitz J; Falk R; Philipson K; Kreyling WG; Moller W; Sommerer K; Larsson S; Svartengren
M. (2006). Negligible clearance of ultrafine particles retained in healthy and affected human lungs. Eur Respir J,
28: 286-290. 157146
Winter-Sorkina Rde; Cassee FR. (2002). From concentration to dose: factors influencing airborne particulate matter
deposition in humans and rats. National Institute for Public Health and the Environment. Bilthoven, Netherlands.
650010031/2002. www.rivm.nl/bibliotheek/rapporten/650010031.html. 043670
Xu GB; YuCP. (1986). Effects of age on deposition of inhaled aerosols in the human lung. Aerosol Sci Technol, 5: 349-
357. 072697
YeatesDB; Gerrity TR; Garrard CS. (1981). Particle deposition and clearance in the bronchial tree. Ann BiomedEng, 9:
577-592. 095391
Yu CP (1985). Theories of electrostatic lung deposition of inhaled aerosols. Ann Occup Hyg, 29: 219-227. 006963
Yu IJ; Park JD; Park ES; Song KS; Han KT; Han JH; Chung YH; Choi BS; Chung KH; Cho MH. (2003). Manganese
distribution in brains of Sprague-Dawley rats after 60 days of stainless steel welding-fume exposure.
Neurotoxicology, 24: 777-785. 156171
Zanobetti A; Schwartz J. (2002). Cardiovascular damage by airborne particles: are diabetics more susceptible?.
Epidemiology, 13: 588-592. 034821
Zeidler-Erdely PC; Calhoun WJ; Ameredes BT; Clark MP; Deye GJ; Baron P; Jones W; Blake T; Castranova V. (2006). In
vitro cytotoxicity of Manville Code 100 glass fibers: Effect of fiber length on human alveolar macrophages. Part
Fibre Toxicol, 28: 3:5. 190967
Zhao X; Ju Y; Liu C; Li J; Huang M; Sun J; Wang T. (2009). Bronchial anatomy of left lung: a study of multi-detector row
CT. Surg Rad Anat, 31: 85-91. 157187
December 2009 4-38
-------
Chapter 5. Possible Pathways/
Modes of Action
The mechanisms underlying pulmonary effects of inhaled PM have been well-studied in the
laboratory and there is general agreement regarding the key roles played by cellular injury and
inflammation. These pathways are initiated following the deposition of inhaled particles on
respiratory tract surfaces. Since most of these studies were conducted at concentrations of PM much
higher than ambient levels, there is some question regarding the relevance of these responses and
mechanisms to ambient exposures.
Interestingly, inhaled PM may also affect the cardiovascular, hematopoietic and other systems.
Mechanisms underlying these extra-pulmonary effects are incompletely understood. However,
pulmonary inflammation can lead to systemic inflammation and pulmonary reflexes can activate the
autonomic nervous system (ANS). These latter responses may mediate cardiovascular and other
systemic effects, as will be discussed below. In addition, it has been proposed that PM or soluble
components of PM reach the circulation by translocating across the epithelial and endothelial
barriers of the respiratory tract. In this way, PM or its components may interact directly with cells in
the vasculature and blood and be transported to the heart and other organs. At this time, evidence
clearly supports the translocation of small solutes following inhalation exposures and the
translocation of soluble components of PM following some high dose exposures involving
intratracheal (IT) instillation. However, there is insufficient evidence to support translocation of
appreciable amounts of intact particles following inhalation exposures at lower concentrations
(Section 4.3.3.1). Future studies will be required to resolve these issues.
The following sections discuss biological pathways which comprise proposed modes of action
for the pulmonary and extra-pulmonary effects of inhaled PM. Overall themes are emphasized and
supportive evidence from new in vitro and in vivo animal studies is cited. The characterization of
evidence here is for PM in general, since most of the potential pathways or modes of action do not
appear to be specific to a particular size class of PM. However, characteristics of ultrafine particles
(UFPs) may allow for unique modes of action or effects disproportionate to their mass, as will be
described below. Recent studies suggest an enhanced potential of this size class of PM to cause
adverse effects; however evidence supporting this hypothesis is limited. Finally, a compilation of
results from new inhalation studies which are relevant to ambient PM exposures and which confirm
and extend these proposed mechanisms is found at the end of this chapter. Detailed descriptions of
these key new studies are found in Chapters 6 and 7.
5.1. Pulmonary Effects
5.1.1. Reactive Oxygen Species
A great deal of research interest has focused on the role of reactive oxygen species (ROS) in
the initiation of pulmonary injury and inflammation following exposure to PM. Numerous studies
have demonstrated PM oxidative potential in in vitro and in vivo assay systems (Ayres et al., 2008,
155666: Cho et al., 2005, 087937: Shi et al., 2003, 088248: Tao et al., 2003, 156111). Both redox
active surface components, such as metals and organic species, and the surface characteristics of
crystal structures have been shown to contribute to oxidative potential (Jiang et al., 2008, 156609:
Tao et al., 2003, 156111: Warheit et al., 2007, 090482). In this way, PM may be a direct source of
ROS in the respiratory tract (Figure 5-1).
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009 5-1
-------
Redox Active
Surface Components
Metals, Organics
Surface Characteristics
of Crystal Structures
PM Oxidative
Potential
Cell-free assay or
oxidation of components
in a cellular system
Figure 5-1. PM oxidative potential.
PM may also act as an indirect source of ROS in the respiratory tract by stimulating cells to
produce ROS (Ayres et al, 2008, 155666: Tao et al., 2003, 156111) (Figure 5-2). This may explain
the observation that the oxidative potential of isolated PM does not always correlate with cellular or
tissue oxidative stress induced by PM exposure. Exposure to PM increases intracellular production
of ROS by a variety of mechanisms. For example, PM interaction with cell surfaces results in
stimulation of NADPH oxidase in macrophages (i.e., the respiratory burst) (Dostert et al., 2008,
155753) and in epithelial cells (Amara et al., 2007, 156212; Becher et al., 2007, 097125; Tamaoki et
al., 2004, 157040). Absorption of PM soluble components (e.g., PAH, transition metals) by
respiratory tract cells can occur (Penn et al., 2005, 088257) and be followed by microsomal
transformation of PAHs to quinones or by redox cycling of transition metals with production of
intracellular ROS (Molinelli et al., 2002, 035347; Xia et al., 2004, 087486). Disruption of
intracellular iron homeostasis with the subsequent generation of ROS has also been demonstrated
following PM exposure (Ohio and Cohen, 2005, 088272). In some cases, mitochondria serve as the
source of ROS in response to PM (Huang et al., 2003, 156573; Risom and Loft, 2005, 089070;
Soberanes et al., 2006, 156991; Soberanes et al., 2009, 190483). Furthermore, PM interaction with
cells can lead to the induction of nitric oxide synthase (Becher et al., 2007, 097125; Lindbom et al.,
2007, 155934; Xiao et al., 2005, 156164; Zhao et al., 2006, 100996) and the production of nitric
oxide and other reactive nitrogen species (RNS).
December 2009
5-2
-------
Iron Sequestration
Redox Cycling
Altered iron homeostasis
NADPH
Oxidase
rf-
Binding to Cell Surfaces
Phagocytosis
Cytoskeletal Interactions
Mitochondria!
Electron Transport
Nitric Oxide
Synthase
Cellular Sources
ofROS/RNS
Soluble metals
Microsomal
Metabolism
(PAH/Quinones)
ROS/RNS Assay
Oxidation of Cellular Components
Lipid Peroxidation, Nitrotyrosine
HO-1 Induction
Figure 5-2 PM stimulates pulmonary cells to produce ROS/RNS.
Although all size fractions of PM may contribute to oxidative and nitrosative stress, UFPs may
contribute disproportionately to their mass due to their large surface/volume ratio. The relative
enrichment ofredox active surface components, such as metals and organics, per unit mass may
translate to a relatively greater oxidative potential of UFPs compared with larger particles with
similar surface components. In addition, the greater surface per unit volume could potentially deliver
relatively more adsorbed soluble components to cells. These components may undergo intracellular
redox cycling following cellular uptake. Furthermore, per unit mass, UFPs may have more
opportunity to interact with cell surfaces due to their greater surface area and their greater particle
number compared with larger PM. These interactions with cell surfaces can lead to ROS generation,
as described above. Recent studies have also demonstrated that UFPs have the capacity to cross
cellular membranes by non-endocytotic mechanisms involving adhesive interactions and diffusion
(Geiser et al., 2005, 087362). as described in Chapter 4. This may allow UFPs to interact with or
penetrate intracellular organelles.
In general, high levels of intracellular ROS/RNS can lead to irreversible protein modifications,
loss of cellular membrane integrity, DNA damage and cellular toxicity. Lower levels of ROS/RNS
may cause reversible protein modifications that trigger intracellular signaling pathways and/or
adaptive responses. Thus PM-dependent generation of ROS may be responsible for a continuum of
responses from cell signaling to cellular injury.
5.1.2. Activation of Cell Signaling Pathways
Activation of cell signaling pathways by ROS/RNS has received increasing attention by
numerous investigators over the years. An early example was provided by Kaul and Forman (1996,
155892) who demonstrated that respiratory burst-derived H2O2 activates the transcription factor
December 2009
5-3
-------
nuclear factor kappa-light-chain-enhancer of activated B cells (NF-KB). Numerous studies since then
have demonstrated that PM, which serves as both a direct and indirect source of ROS/RNS, activates
cell signaling pathways by this mechanism.
PM also has the potential to activate cell signaling by mechanisms that are independent of
ROS/RNS. For example, PM delivers water-soluble components, such as endotoxin and zinc (Zn), to
cell surfaces. Endotoxin binds to toll-like receptors on alveolar macrophages and other cells,
resulting in the upregulation of cytokines (Becker et al., 2002, 052419). Zn, a transition metal which
does not redox cycle, inhibits protein tyrosine phophatases in airway epithelial cells resulting in a
cascade of cell signaling events (Tal et al., 2006, 108588). Similarly, PM-mediated delivery of lipid
soluble components such as PAH results in binding and activation of the arylhydrocarbon receptor
(AhR). AhR is a transcription factor responsible for the upregulation of CYP1A1, a cytochrome
oxidase involved in PAH biotransformation to metabolites capable of forming DNA adducts or
eliciting oxidative stress responses (Rouse et al., 2008, 156930). In addition, interaction of PM with
cell surfaces may activate cell signaling by perturbation of the cytoskeleton, adherence,
internalization, or receptor-mediated pathways.
Recent studies involving PM exposures have focused on intracellular pathways involving
protein kinases, such as mitogen-activated protein kinase (MAPK) (Bayram et al., 2006, 088439;
Lee et al., 2005, 156682: Roberts et al., 2003, 156051: Soberanes et al., 2009, 190483): AKT (Ahsan
et al., 2005, 156200): src (Cao et al., 2007, 156322) and epidermal growth factor receptor (Blanchet
et al., 2004, 087982: Cao et al., 2007, 156322: Tamaoki et al., 2004, 157040). as well as ras
(Tamaoki et al., 2004, 157040). toll-like receptors (Becker et al., 2002, 052419: Becker et al., 2005,
088590). protein tyrosine phosphatases (Tal et al., 2006, 108588). phospholipases A2 (Lee et al.,
2003, 156678). calcium (Agopyan et al., 2003, 155649: Brown et al., 2004, 155705: Brown et al.,
2004, 088663: Geng et al., 2005, 096689: 2006, 097026: Sakamoto et al., 2007, 096282). caspases
(Soberanes et al., 2006, 156991: Zhang et al., 2007, 156179). poly (ADP-ribose) polymerase family
member 1 (PARP-1) (Zhang et al., 2007, 156179) and histone acetylation (Gilmour et al., 2003,
096959). The transcription factors regulated by these pathways, including NF-KB (Bayram et al.,
2006, 088439: Lee et al., 2005, 156682: Takizawa et al., 2003, 157039). activator protein 1 (AP-1)
(Donaldson et al., 2003, 156408). signal transducers and activators of transcription protein (STAT)
(Cao et al., 2007, 156322). antioxidant response element (ARE) (Li and Nel, 2006, 156694). and
AhR (Rouse et al., 2008, 156930) have also been studied following PM exposures. Activation of
these intracellular pathways and transcription factors leads to the upregulation of genes responsible
for inflammatory, immune and acute phase responses as well as genes responsible for antioxidant
defense and xenobiotic metabolism.
5.1.3. Pulmonary Inflammation
Following PM exposure, transcription factor activation in macrophages and epithelial cells
stimulates the synthesis and release of soluble mediators involved in inflammatory and immune
responses including cytokines, chemokines, proteases and eicosanoids (Figure 5-3). These
substances play a role in recruiting inflammatory cells such as neutrophils, monocytes, mast cells
and eosinophils to the lung. Interactions between macrophages and epithelial cells enhance these
responses (Tao and Kobzik, 2002, 157044).
December 2009 5-4
-------
Cell Surface
Interactions
Endotoxin
Zinc
PAH
• * *
Cell Signaling/
Transcription
Factor Activation
),
/
Cytokines
Chemokines
Proteases
Mediators
Cell Injury
Influx of Leukocytes
Pulmonary Inflammation
Figure 5-3. PM activates cell signaling pathways leading to pulmonary inflammation.
Inflammatory cells can serve as a source of extracellular ROS which, along with soluble
mediators derived from the inflammatory cells, amplify the inflammatory response. Unchecked
inflammation may cause cellular and tissue injury through the generation of excess amounts of ROS
and soluble mediators. In some cases the oxidative potential of PM is well-correlated with the degree
of inflammation (Dick et al., 2003, 036605). suggesting that the inflammation is a direct
consequence of PM-generated ROS. However, in other cases the oxidative potential of PM is not
well-correlated with the degree of inflammation (Beck-Speier et al., 2005, 156262). suggesting that
the inflammation is a consequence of the ROS-independent mechanisms by which PM activates
intracellular signaling pathways. Particle surface area has been identified as a key determinant of the
extent of inflammation in the case of low-toxicity, low-solubility particles (Donaldson et al., 2008,
December 2009
5-5
-------
190217). Moreover, UFPs may cause inflammation disproportionately to their mass compared with
PM of larger sizes given their large surface/volume ratio compared with other PM fractions.
PM exposure often results in neutrophilic inflammation in laboratory studies (Tao et al., 2003,
156111). Neutrophilic inflammation is also associated with acute lung injury in humans as well as
chronic lung diseases such as COPD and certain forms of asthma (Barnes, 2007, 191139; Cowburn
et al., 2008, 191142). Circulating neutrophils respond to chemotactic factors in the lung such as
leukotriene B4 and IL-8 (Barnes, 2007, 191139). They migrate into the lung parenchyma across the
pulmonary capillary network and into the airways from the bronchial circulation (Cowburn et al.,
2008, 191142). As a consequence of priming by inflammatory mediators or contact with
extracellular matrix components, neutrophils become hyperresponsive to activating signals and
insensitive to chemotactic signals (Cowburn et al., 2008, 191142). Activation results in neutrophil
degranulation, respiratory burst responses and soluble mediator release (Cowburn et al., 2008,
191142). Neutrophils eventually undergo apoptosis and are phagocytized by inflammatory
macrophages (Cowburn et al., 2008, 191142). This is accompanied by the release of anti-
inflammatory mediators such as IL-10 and transforming growth factor-|3 (TGF- |3) (Cowburn et al.,
2008, 191142). These steps are key to the resolution of inflammation and prevent unregulated release
of toxic neutrophil products such as neutrophil elastase (Cowburn et al., 2008, 191142). Thus,
circumstances leading to the decreased ingestion of apoptotic neutrophils by macrophages may lead
to tissue injury. Impairment of macrophage function may serve as an important mechanism by which
PM contributes to disease.
5.1.4. Respiratory Tract Barrier Function
Epithelial injury can lead to an increase in permeability of the airway epithelial and
alveolar-capillary barriers (Braude et al., 1986, 155701). Enhanced transport of soluble and possibly
of insoluble PM components into the circulation may occur under these conditions. Increased
epithelial permeability is also associated with enhanced immune responses to proteins, including
allergens, on the epithelial surface, presumably due to the greater availability of antigens to
underlying immune cells (Wan et al., 1999, 191903). Furthermore, endothelial injury can
compromise the integrity of the alveolar-capillary barrier resulting in transvascular fluid and solute
flux (Braude et al., 1986, 155701). Soluble mediators derived from inflammatory and lung cells
(Chang et al., 2005, 097776) and peptides released by some nerve cells (Widdicombe and Lee,
2001, 019049) can increase the permeability of the alveolar-capillary barrier and result in alveolar
edema. Compromised barrier function in the airways may lead to airway edema. Edema occurring
secondarily to nerve cell stimulation is one component of the process termed neurogenic
inflammation.
Given the small size of UFPs, modest changes in epithelial permeability may particularly
affect the disposition of this fraction. Enhanced translocation to interstitial compartments or to the
circulation may be important sequelae. A recent study described in Section 4.3.4.3 demonstrated
greater translocation of UFPs compared with PM^sinto the circulation of rodents treated with
endotoxin to induce acute lung injury prior to IT instillation of PM (Chen et al., 2006, 147267).
Furthermore, epithelial injury in another model resulted in greater translocation of UFPs into the
interstitial compartment (Adamson and Prieditis, 1995, 189982).
5.1.5. Antioxidant Defenses and Adaptive Responses
Antioxidant defenses and adaptive responses are important modulators of oxidative stress and
other cellular stresses resulting from PM exposure. Antioxidants are present in the epithelial lining
fluid in all regions of the respiratory tract. In addition, they are present in cells of the lung
parenchyma and inflammatory cells found in airways and alveoli. Some antioxidants act directly
against oxidant species (e.g., glutathione, ascorbate, superoxide dismutase) while others act
indirectly (e.g., gamma-glutamylcysteine synthetase [yGCS], glutathione reductase). Furthermore,
some antioxidants (e.g., Phase 2 enzymes heme oxygenase-l[HO-l], NADPH quinone
oxidoreductase 1 [NQO1], glutathione-S-transferase [GST]) are inducible via activation of the
nuclear factor (erythroid-derived 2)- related factor 2 (Nrf2)-ARE pathway, which occurs as an
adaptive response to stress (Cho et al., 2006, 156345; Li and Nel, 2006, 156694). Antioxidants play
December 2009 5-6
-------
an important role in reducing the oxidative potential of those PM species that directly generate ROS.
They also inhibit responses due to generation of intracellular ROS.
Recently a three-tier response to oxidative stress was proposed (Li and Nel, 2006, 156694). In
this scheme, mild oxidative stress enhances antioxidant defenses by upregulating Phase 2 and other
antioxidant enzymes (Tier 1). Further increase in oxidative stress induces inflammation (Tier 2) and
cell death (Tier 3). Experimental evidence is supportive of this scheme. Numerous studies have
demonstrated that enhancement of lung and cellular antioxidant defenses inhibits inflammation,
cytotoxicity and other responses following exposure to PM (Ahsan et al, 2005, 156200; Bachoual et
al., 2007, 155667; Bayram et al., 2006, 088439; Chang et al., 2005, 097776; Imrich et al., 2007,
155859; Koike and Kobayashi, 2005, 088303; Koike et al., 2004, 058555; Li et al., 2007, 155929;
Ramage and Guy, 2004, 055640; Rhoden et al., 2004, 087969; Steerenberg et al., 2004, 087981;
Takizawa et al., 2003, 157039; Tao et al., 2003, 156111; Upadhyay et al., 2003, 097370; Wan and
Diaz-Sanchez, 2006, 097399; Wan and Diaz-Sanchez, 2007, 156145; Yin et al., 2004, 087983).
Cellular and tissue exposure to xenobiotics carried by PM can lead to induction of Phase 1 and
Phase 2 detoxifying enzymes following the activation of cell signaling pathways and transcription
factors AhR and ARE, respectively (Rengasamy et al., 2003, 156907; Rouse et al., 2008, 156930;
Zhao et al., 2006, 100996).
AHR and
Airway
Remodeling
Allergic Asthma
And Other
Allergic
Disorders
Impaired
Host Defense
and Infections
Progression of
Pre-existing
Lung
Disease
DMA Damage
and
Lung
Cancer
Death or Hospitalization for Asthma, Pneumonia, COPD and Lung Cancer
Figure 5-4. Potential pathways for the effects of PM on the respiratory system.
December 2009
5-7
-------
5.1.6. Pulmonary Function
PM exposure may alter pulmonary function by a variety of different mechanisms (Figure 5-4).
In the short-term, airway hyperresponsiveness (AHR) may ensue due to the influence of
inflammatory mediators. In the long-term, morphological changes may occur, in some cases leading
to mucus hypersecretion and airway remodeling. Activation of irritant receptors and stimulation of
the ANS in the respiratory tract is another mechanism by which PM exposure may alter pulmonary
function (Section 5.4).
5.1.7. Allergic Disorders
PM exposure sometimes leads to the development of allergic immune responses (Figure 5-4).
These responses are predominately mediated by T helper 2 cells (Th2). Thl responses, characterized
by IFN-y and classical macrophage activation, are inflammatory; in excess they can lead to tissue
damage. Alternatively, Th2 responses are associated with allergy and asthma and are characterized
by IL-4, IL-5, IL-13, influx of eosinophils, B-lymphocyte production of IgE, and alternative
macrophage activation. PM exposure can also lead to the exacerbation of airway allergic responses,
such as antigen-specific IgE production and AHR.
Due to soluble mediators and immune cell trafficking, pulmonary exposure may result in
systemic immune alterations. Not only do macrophages ingest PM, but they are also antigen
presenting cells whose level of activation dictates costimulation and thus subsequent T cell
responses. These cells are highly mobile and can transport PM to other sites such as lymph nodes.
Dendritic cells (DC) also play a key role as antigen presenting cells and in modulating T and B cell
activity. A cell culture model of the human epithelial airway wall was used to demonstrate that DC
extended processes between epithelial cells through the tight junctions, collected particles in the
lumenal space and transported them across the epithelium (Blank et al., 2007, 096521). DC also
transmigrated through the epithelium to take up particles on the epithelial surface (Blank et al., 2007,
096521). Furthermore, DC interacted with particle-loaded macrophages on top of the epithelium and
with other DC within or beneath the epithelium to transfer particles (Blank et al., 2007, 096521). In
vitro studies also demonstrated that the adjuvant activity of diesel exhaust particles (DEPs) involved
stimulation of immature monocyte-derived dendritic cells (iMDDC) to undergo maturation in
response to an altered airway epithelial cell-derived microenvironment (Bleck et al., 2006, 096560).
Additionally, DEP directly influenced the profile of cytokines secreted by DC and caused a
predisposition toward Th2-mediated or allergic responses (Chan et al., 2006, 097468). Thus PM can
negatively affect both innate immunity through effects on macrophage pathogen handling
(Section 5.1.8) as well as adaptive immunity by altering macrophage or DC antigen presenting
activity and subsequent T cell responses.
Moreover, recent studies have demonstrated that ambient PM can act as an adjuvant for
allergic sensitization with UFPs having a greater effect than fine particles (de Haar et al., 2006,
144746; Li et al., 2009, 190457). This has been attributed to higher oxidative potential of the UFPs
compared with the same mass of particles of larger size (Li et al., 2009, 190457). although the larger
surface area and particle number per unit mass as well as the propensity of UFPs for trans-epithelial
movement may also contribute to this effect.
5.1.8. Impaired Lung Defense Mechanisms
PM exposure may impair lung defense mechanisms and result in frequent or persistent
infections (Figure 5-4). Potential targets include mucociliary transport, surfactant function and
pathogen clearance. Pathogen clearance is dependent on the integrity of macrophages and their
migration, phagocytosis and respiratory burst functions. PM-mediated cytotoxicity of macrophages
with the concomitant release of lysosomal contents may affect pathogen clearance and cause damage
to nearby cells and tissues. IT instillation and cell culture experiments have demonstrated
PM-dependent impairment of lung defense mechanisms (Jaspers et al., 2005, 088115; Kaan and
Hegele, 2003, 095753: Long et al., 2005, 087454: Moller et al., 2005, 156770: Monn et al., 2003,
052418: Roberts et al., 2007, 097623: Yin et al., 2004, 087983).
December 2009 5-8
-------
5.1.9. Resolution of Inflammation/Progression or Exacerbation of
Disease
Resolution of pulmonary inflammation and injury has been demonstrated in many
experimental models using higher than ambient concentrations of PM. Factors contributing to this
complex process are likely to include the uptake and clearance of PM by macrophages; the retention
of PM in parenchymal cells and tissues; the balance of pro/anti-inflammatory soluble mediators,
oxidants/antioxidants and proteases/anti-proteases; and the presence of pre-existing disease. These
factors may also influence the resolution of pulmonary responses to ambient PM exposures (Figure
5-4). The long-term consequences of prolonged inflammation are not likely to be beneficial and may
lead to remodeling of the respiratory tract and to the progression or exacerbation of disease.
5.1.9.1. Factors Affecting the Retention of PM
Clearance of poorly soluble particles from ET, TB and alveolar regions is extensively
discussed in Section 4.3. While clearance from ET and TB regions generally occurs over hours to
days, clearance from the alveolar compartment is much slower, occurring over months to years
depending on the species. Phagocytosis by alveolar macrophages and transport by the mucociliary
escalator is the primary mechanism of clearance from the alveolar compartment although neutrophil
phagocytosis also plays a role (Snipes et al, 1997, 156092). Pre-existing disease can alter the extent
and localization of PM deposition as discussed in Section 4.2.4.5. In addition, mechanisms of
clearance may be altered in cases of pre-existing disease as discussed in Section 4.3.4.3. While mild
asthma was associated with enhanced mucociliary clearance, acute lung injury was associated with
enhanced particle translocation to the circulation and the interstitial compartment. Whether retained
particles are localized in alveolar macrophages or parenchymal tissue also differs according to
species (Snipes, 1996, 076041).
UFPs may have a special propensity for retention given the decreased efficiency of alveolar
macrophages for phagocytosis of this size particle (Oberdorster, 1988, 006857) and the
demonstration that UFPs can readily cross cellular membranes (Geiser et al., 2005, 087362). Some
studies suggest that UFPs are taken up by epithelial cells and move to the interstitum where they are
cleared by other pathways (Semmler-Behnke et al., 2007, 156080); however clearance mechanisms
are not entirely understood.
Enhanced deposition of particles in "hot spots" may influence retention. For example,
deposition in the centriacinar or proximal alveolar region, where clearance is slow, may result in
accumulated particle dose in this region and the potential for prolonged inflammation at the site
leading to the development of pulmonary fibrosis or emphysema (Donaldson et al., 2008, 190217). A
recent study suggests an important role for retained particles in the progression of disease.
Complexation of endogenous iron by retained particles resulted in retained particles growing larger
over time. The authors suggested that redox cycling of complexed iron may be responsible for
disease progression (Ohio and Cohen, 2005, 088272: Ohio et al., 2004, 155790).
5.1.9.2. Factors Affecting the Balance of Pro/Anti-Inflammatory Mediators,
Oxidants/Anti-Oxidants and Proteases/Anti-Proteases
Inflammation can be enhanced by pro-inflammatory mediators or dampened by anti-
inflammatory mediators. Production of anti-inflammatory mediators normally occurs at several steps
of the inflammation pathway, such as the release of IL-10 and TGF-(3 during phagocytosis of
apoptotic neutrophils by macrophages (Cowburn et al., 2008, 191142). Dsyregulation of the
inflammatory process may prevent the resolution of inflammation. PM exposure may result in the
production of pro-inflammatory mediators as well as decrease the production of anti-inflammatory
mediators by impairing macrophage function.
An unfavorable balance of oxidants to antioxidants in the lung is associated with inflammatory
lung diseases including asthma and COPD (Rahman et al., 2006, 191165). PM is likely to contribute
to an unfavorable balance through its oxidative potential and capacity to promote cellular production
of ROS. Exacerbations of asthma and COPD resulting from bacterial and viral infections are also
associated with increased oxidative stress (Barnes, 2007, 191139). Conversely, antioxidants may
reduce neutrophilic inflammation associated with oxidative stress (Barnes, 2007, 191139).
December 2009 5-9
-------
Protease/anti-protease balance has long been tied to the pathogenesis of emphysema and other
forms of COPD (Owen, 2008, 191162). Key steps include the release of proteinases by inflammatory
cells which degrade the extracellular matrix components of alveolar walls. Destruction of alveolar
walls and airspace enlargement ensues. Endogenous anti-protease defenses in the lung modulate this
response but may be insufficient to prevent it during prolonged inflammation. Proteases also play a
role in pathologies which lead to small airway fibrosis (Owen, 2008, 191162). Oxidative stress has
been linked to both activation of proteases and inactivation of anti-proteases (Owen, 2008, 191162).
PM may contribute to an unfavorable protease/anti-protease balance through the generation of ROS.
Although there are numerous inflammatory cell-derived proteases and lung anti-proteases,
many recent studies have focused on matrix metalloproteinases (MMPs). MMPs are a family of Zn-
containing enzymes normally found in an inactive pro-enzyme form. Activation involves proteolytic
cleavage or oxidation of the "cysteine switch" (Pardo and Selman, 2005, 191163). Inhibitors include
tissue inhibitors of metalloproteinases (TIMPs) (Pardo and Selman, 2005, 191163). In particular,
MMP-1 is well-studied and found to play an important role in physiological processes such as
development and wound repair as well as in diseases such as pulmonary emphysema, fibrosis,
asthma and bronchial carcinoma (Li et al., 2009, 190424; Pardo and Selman, 2005, 191163). In
addition to its activity in degrading collagenase, MMP-1 also acts on non-matrix substrates and cell
surface molecules suggesting that it may influence cell signaling (Pardo and Selman, 2005, 191163).
MMP-2 and MMP-9 are thought to be involved in the pathogenesis of disease through their
gelatinase activity. Interestingly, recent in vitro studies have demonstrated upregulation of MMP-1
by hydrogen peroxide, cigarette smoke and DEPs (Amara et al., 2007, 156212; Li et al., 2009,
190424; Mercer et al., 2004, 191180). and up-regulation of MMP-12 following instillation of PM
collected from the Paris subway (Bachoual et al., 2007, 155667). These considerations suggest that
particulate air pollution may act via MMP to mediate progression or exacerbation of lung disease.
5.1.9.3. Pre-Existing Disease
In addition to its effects on deposition, retention and clearance of PM described above, pre-
existing disease may also alter the balance of the aforementioned factors. For example, acute
exacerbations of COPD are characterized by a rapid influx of neutrophils into the airways (Owen,
2008, 191162). However, clearance of apoptotic neutrophils by macrophages is impaired in COPD
leading to greater release of neutrophil-derived inflammatory mediators, oxidants and proteases
(Owen, 2008, 191162). Thus, exacerbation of disease may occur as a result of unchecked
inflammation.
5.1.10. Pulmonary DMA Damage
Pulmonary DNA damage can occur primarily or secondarily to PM exposure. Primary effects
include oxidative DNA injury or DNA adduct formation due directly to PM while secondary effects
occur due to PM-mediated inflammation (De Kok et al., 2005, 088656; Gabelova et al., 2007,
156457; Gallagher et al., 2003, 140171; Schins and Knaapen, 2007, 156074). These responses may
lead to chromosomal aberrations or DNA strand breaks. PM effects on cell cycle arrest, proliferation,
apoptosis, and DNA repair mechanisms may also influence the genotoxic, mutagenic or carcinogenic
potential of DNA damage as reviewed by Schins et al. (2007, 156074).
5.1.11. Epigenetic Changes
Epigenetic mechanisms regulate the transcription of genes without altering the nucleotide
sequence of DNA. These mechanisms generally involve DNA methylation and histone
modifications, leading to alterations which may have long-term consequences or are heritable (Jones
and Baylin, 2007, 191153; Keverne and Curley, 2008, 191154). DNA methylation and histone
modifications, which include methylation, acetylation, phosphorylation, ubiquitylation and
sumoylation, are known to be linked (Hitchler and Domann, 2007, 191151; Jones and Baylin, 2007,
191153). Numerous studies have identified epigenetic processes in the control of cancer (Foley et al.,
2009, 191144; Gopalakrishnan et al., 2008, 191147; Jones and Baylin, 2007, 191153; Valinluck et
al., 2004, 191170). embryonic development (Foley et al., 2009, 191144; Gopalakrishnan et al., 2008,
December 2009 5-10
-------
191147; Keverne and Curley, 2008, 191154) and inflammation and other immune system functions
(Adcock et al., 2007, 191178).
Epigenetic modifications resulting in decreased expression of tumor suppressor genes and
increased expression of transforming genes have been observed in human tumors (Valinluck et al.,
2004, 191170). In general, transcription repression is associated with DNA methylation in promoter
regions of genes. Cytosine methylation in CpG dinucleotides has emerged as an important, heritable
epigenetic modification which can result in chromatin remodeling and decreased gene expression
(Valinluck et al., 2004, 191170). Global changes in DNA methylation are also seen in cancer and
hypomethylation is associated with genomic instability (Gopalakrishnan et al., 2008, 191147).
Embryonic development is characterized by several phases of epigenetic modifications. DNA
methylation is very dynamic following fertilization, with demethylation and re-methylation of egg
and sperm genomes occurring immediately (Foley et al., 2009, 191144). Imprinted genes, however,
retain the methylation profile of the parent of origin (Foley et al., 2009, 191144). Epigenetic changes
accumulated through a life course may be passed from parent to offspring in the germline (i.e.,
germline transmission of epimutation) if they survive the epigenetic remodeling that occurs during
gametogenesis and early embryogenesis (Foley et al., 2009, 191144). Early development is
characterized by the process of cell differentiation, which produces different cell types and involves
the selective activation of some sets of genes and the silencing of others in a temporal pattern (Foley
et al., 2009, 191144; Gopalakrishnan et al., 2008, 191147). DNA methylation is postulated to provide
a basis for cell differentiation (Gopalakrishnan et al., 2008, 191147).
In the lung, histone acetylation and methylation have been linked to inflammatory gene
expression, T cell differentiation, and the regulation of macrophage function following pathogen
challenge (Adcock et al., 2007, 191178). Furthermore, altered patterns of methylation and
acetylation have been reported in inflammatory diseases (Adcock et al., 2007, 191178). Reduced
expression and activity of histone deacetylase have been demonstrated in lung and inflammatory
cells in COPD and asthma (Adcock et al., 2007, 191178; Barnes, 2007, 191139). Consequently,
histone deacetylase has been identified as a potential therapeutic target for epigenetic therapy
(Adcock et al., 2007, 191178; Jones and Baylin, 2007, 191153).
Epigenetic mechanisms have been identified as potential targets for gene-environment
interactions and recent studies have demonstrated that diet, cigarette smoking, endocrine disrupters,
heavy metals and bacterial infection can alter the epigenetic profile in animals and humans (Foley et
al., 2009, 191144). A role for PM in promoting epigenetic changes has been proposed and new
studies, discussed in later chapters, provide some evidence for this pathway (Baccarelli et al., 2009,
192155; Liu et al., 2008, 156709; Reed et al., 2008, 156903; Tarantini et al., 2009, 192010; Tarantini
et al., 2009, 192153; Yauk et al., 2008, 157164).
Early life exposures may be especially important in this regard since periods of rapid cell
division and epigenetic remodeling are likely to occur at this time (Foley et al., 2009, 191144;
Keverne and Curley, 2008, 191154; Wright and Baccarelli, 2007, 191173). This may provide a basis
for fetal origins of adult disease.
It has been suggested that DNA methylation is regulated by oxygen gradients and redox status
(Hitchler and Domann, 2007, 191151). While this is of particular importance during development
where oxygen gradients and redox status are linked to cellular differentiation, these processes are
also important for cell signaling during all stages of life. A common metabolic precursor for both
methylation reactions and glutathione availability (involved in redox status) is methionine (Hitchler
and Domann, 2007, 191151). Methionine availability regulates the cell's ability to generate
S-adenosyl methionine which is directly involved in DNA and histone methylation and the cell's
ability to generate homocysteine/cysteine which is involved in glutathione biosynthesis (Hitchler
and Domann, 2007, 191151). Furthermore, the folate cycle is a key determinant of methionine
bioavailability (Hitchler and Domann, 2007, 191151). In this way, cellular intermediary metabolism
is linked to epigenetic processes, with oxidative stress necessitating a metabolic shift resulting in
decreased DNA methylation and increased glutathione production.
5.1.12. Lung Development
Lung development is a multi-step process which begins in embryogenesis and continues to
adult life (Pinkerton and Joad, 2006, 091237). This allows for a long period of potential
vulnerability to environmental and other stressors. Furthermore, enzymatic systems responsible for
detoxification of xenobiotic compounds are not fully developed until the postnatal period (Pinkerton
December 2009 5-11
-------
and Joad, 2006, 091237). Disruption of cell signaling during development could affect cellular
differentiation, branching morphogenesis and overall lung growth, possibly leading to life-long
consequences. Although very little is known about the effects of maternal exposure to PM on the
fetus or the effects of exposure during childhood, recent animal studies demonstrate respiratory and
immune system effects of perinatal exposure to sidestream cigarette smoke (Pinkerton and Joad,
2006, 091237: Wang and Pinkerton, 2007, 1799751
5.2. Systemic Inflammation
Pulmonary inflammation resulting from PM exposure may trigger systemic inflammation
through the action of cytokines and other soluble mediators which leave the lung and enter the
circulation (Figure 5-5). Epithelial permeability may exert an important influence on this process
(Section 5.1.4). Cytokines released by alveolar macrophages can stimulate bone marrow production
of leukocytes resulting in an increased number of total and immature leukocytes in the circulation
(Van and Hogg, 2002, 088111: Van Eeden et al, 2001, 019018). They also can activate neutrophils
and promote their sequestration in microvascular beds (Van Eeden et al., 2001, 019018). The time
course of these responses varies according to the acute or chronic nature of the PM exposure
(van Eeden et al., 2005, 157086).
Systemic inflammation is seen under conditions of mild pulmonary inflammation - and
sometimes under conditions of no measurable pulmonary inflammation - following PM exposure.
The time-dependent nature of pulmonary and systemic inflammatory responses may in part explain
these findings since biomarkers of inflammation are frequently measured only at one time point.
Furthermore, chronic exposures may lead to adaptive responses. In general, systemic inflammation is
associated with changes in circulating white blood cells, the acute phase response, pro-coagulation
effects, endothelial dysfunction and the development of atherosclerosis (Figure 5-5). Adverse effects
on the cardiovascular and cerebrovascular systems such as thrombosis, plaque rupture, MI and stroke
may result. Systemic inflammation may affect other organ systems such as the liver or the CNS.
One recent study demonstrated that alveolar macrophage-derived IL-6 mediated
pro-coagulation effects in mice exposed by IT instillation to PMi0 (Mutlu et al., 2007, 121441). This
study provides a clear link between lung cytokines and systemic responses in one model system.
Whether this mechanism or others account for the majority of extra-pulmonary effects following
inhalation of PM at concentrations relevant to ambient exposures is not yet known.
December 2009 5-12
-------
Pulmonary
Oxidative Stress
and Inflammation
Direct Effects ?
Translocation/Absorption
of Soluble Components
Pulmonary Reflexes
Other Reflexes ?
Autonomic
Nervous System
Liver
Acute
Phase
Response
Altered Sympathetic/
Parasympathetic
Tone
Atherosclerosis
Plaque Destabihzation
Or Rupture
Death or
Hospitalization
for Throm bo-
Em bolic Disease
Death or Hospitalization
for Stroke
Death or Hospitalization for Coronary
Heart Disease or Congestive Heart Failure
Figure 5-5. Potential pathways for the effects of PM on the cardiovascular system.
5.2.1. Endothelial Dysfunction and Altered Vasoreactivity
The lumenal surface of blood vessels is lined by endothelial cells which, in addition to
providing a barrier function, are key regulators of vascular homeostasis. Endothelial cells synthesize
and release vasodilators such as nitric oxide (NO) and prostacyclin and vasoconstrictors such as
endothelin (ET), which act on neighboring smooth muscle cells. ET also stimulates endothelial NO
synthesis through a feedback mechanism. Inhalation of high concentrations of PM has been reported
to increase ET levels in the circulation (Thomson et al, 2005, 087554). ET has also been proposed to
play a role in hypoxia-induced MI (Caligiuri et al., 1999, 156318). However, the role of ET in
mediating cardiovascular effects following ambient PM exposures is unclear (Section 6.2.4.3).
December 2009
5-13
-------
Endothelial dysfunction can lead to or follow endothelial activation under conditions of
systemic inflammation and/or vascular oxidative stress. Both systemic inflammation and vascular
oxidative stress have been associated with PM exposure (Nurkiewicz et al., 2006, 088611;
Van Eeden et al., 2001, 019018). Cytokines activate endothelial cells and upregulate endothelial cell
adhesion molecules. They also promote the sequestration of neutrophils in microvascular beds.
Neutrophil sequestration is sometimes associated with the deposition of myeloperoxidase (MPO) on
endothelial cell surfaces (Nurkiewicz et al., 2006, 088611). ROS-derived from neutrophils, MPO,
other adhered inflammatory cells and/or other sources can perturb the balance of vasodilator and
vasoconstrictor species produced by endothelial cells. Oxidative stress can result in decreased
synthesis of NO due to limitation of the redox-sensitive essential cofactor tetrahydrobiopterin and in
decreased bioavailability of NO due to reaction with superoxide. Prostacyclin synthesis is also
decreased by oxidative stress. Importantly, these processes can affect vasoreactivity such that blood
vessels may be unable to respond to vasoconstrictor stimuli with compensatory vasodilation.
Loss of NO and prostacyclin synthesis due to PM-dependent vascular oxidative stress may
have other consequences since both exert negative influences on platelet and neutrophil activation.
While endothelial surfaces normally are anti-thrombotic, endothelial dysfunction can contribute to
thrombus formation. Furthermore, inflammation and oxidative stress associated with endothelial
dysfunction can contribute to the development or progression of atherosclerosis (van Eeden et al.,
2005, 157086).
5.2.2. Activation of Coagulation and Acute Phase Response
The primary function of the coagulation cascade is to stop the loss of blood after vascular
injury by forming a fibrin clot. However in some cases, activation of coagulation can promote
intravascular thrombosis (Karoly et al., 2007, 155890). It has been proposed that air pollution-
associated PM can activate clotting pathways and enhance the likelihood of an obstructive cardiac
ischemic event (e.g., MI) or cerebral event (e.g., stroke) (Seaton et al., 1995, 045721).
Coagulation is regulated by intrinsic and extrinsic pathways. The intrinsic pathway occurs
following activation of Factor XII and does not require the addition of an exogenous agent
(Mackman, 2005, 156722). On the other hand, the extrinsic pathway is an inducible signaling
cascade that can be activated by tissue factor (TF) produced in response to inflammation or
endothelial injury (Karoly et al., 2007, 155890).
In general, platelets, red blood cells (RBCs) and endothelial cells are effector cells for
inducing a pro-coagulant state in the vasculature. Circulating factors may enhance coagulation or
promote activation of platelets. Cytokines formed during tissue damage and inflammation lead to TF
induction. TF is the initiating stimulus for coagulation following vascular injury or plaque erosion.
Complexes of TF:Factor Vila form on endothelial cell surfaces and play a key role in thrombin
generation by initiating the extrinsic blood coagulation pathway (Gilmour et al., 2005, 087410).
Thrombin generates fibrin from fibrinogen and amplifies the intrinsic pathway (Karoly et al., 2007,
155890). TF and thrombin also have pro-inflammatory actions independent of coagulation functions
(Chu, 2005, 155730); thus activation of coagulation may lead to or potentiate inflammation.
Endothelial cell-derived von Willebrand factor also contributes to coagulation.
The fibrinolytic system opposes these processes by facilitating the removal of a clot. The
fibrinolytic pathway is regulated by the ratio of tissue plasminogen activator (tPA) and plasminogen
activator inhibitor (PAI). Furthermore, the endothelial cell surface has anti-thrombotic properties due
to the expression of tissue factor pathway inhibitor (TFPI) and thrombomodulin (Mackman, 2005,
156722).
Inhibition of the fibrinolytic pathway, along with increased plasma viscosity and increased
concentrations of plasma fibrinogen and Factor VII, contributes to a pro-thrombotic state (Gilmour et
al., 2005, 087410). In acute lung injury, vascular cells have enhanced pro-coagulant activity and
impaired fibrinolytic activity (Gilmour et al., 2005, 087410). In arterial atherosclerosis, TF
expression is increased within plaques. As a result, spontaneous plaque rupture may trigger
intravascular clotting (Karoly et al., 2007, 155890).
Acute phase responses also play a role in hemostasis by exerting pro-coagulant effects.
Cytokines such as IL-6 stimulate the liver to produce acute phase proteins including C-reactive
protein (CRP), fibrinogen and antiproteases (van Eeden et al., 2005, 157086). To date, there is
limited evidence supporting a role for ambient PM in stimulating acute phase responses (Ruckerl et
al., 2007, 156931 and reviewed therein) (also Section 6.2.7 and 6.2.8 in this ISA).
December 2009 5-14
-------
5.2.3. Atherosclerosis
Atherosclerosis is a chronic progressive disease which contributes greatly to cardiovascular
morbidity (Libby, 2002, 192009). Mainly a disease of the large arteries, it is characterized by the
accumulation of lipid and fibrous tissue in atheromas, or swellings of the vessel wall (Libby, 2002,
192009). Although a strong link is known to exist between hypercholesterolemia and atherogenesis,
there is growing appreciation of the key role played by inflammation in the initiation and progression
of atherosclerosis (Libby, 2002, 192009). Furthermore, inflammation has the potential to promote
thrombosis which can complicate this disorder and lead to MI and stroke (Libby, 2002, 192009). As
discussed above, PM exposure is associated with systemic inflammation, potentially contributing to
the development of atherosclerosis.
Atheroma formation in experimental animals fed a high fat diet begins with the accumulation
of modified lipoprotein particles in the arterial intima, as reviewed by Libby (2002, 192009). The
modification of lipoprotein particles often involves oxidation. As discussed above, PM exposure is
associated with oxidative stress, suggesting a potential role for PM in the modification of lipoprotein
particles. Endothelial dysfunction may also be key to these early events (Halvorsen et al., 2008,
191149). Recent studies, which are discussed in later chapters, demonstrate PM-dependent
endothelial dysfunction (Nurkiewicz et al., 2009, 191961). As described by Libby (2002, 192009).
oxidative stress leads to lipid modification and uptake by endothelial cells and initiates an
inflammatory response by activating NF-KB. As a result, cell adhesion molecules such as vascular
cell adhesion molecule-1 (VCAM-1) are upregulated and expressed in endothelial cells. Pro-
inflammatory cytokines associated with systemic inflammation may also contribute to adhesion
molecule upregulation. Subsequently, monocytes and T cell lymphocytes adhere to the activated
endothelium, then migrate into the tunica intima directed by chemokines such as monocyte
chemoattractant protein-1 (MCP-1) and IL-8. Monocytes undergo transformation to tissue
macrophages and later to foam cells. As a part of this transformation, monocyte/macrophages
express scavenger receptors and bind to and internalize the modified lipoprotein particles. These
cells secrete growth factors and cytokines, produce ROS, replicate within the lesion and contribute to
further lesion progression. Macrophage colony-stimulating factor (M-CSF) and granulocyte-
macrophage colony-stimulating factor (GM-CSF) are thought to be involved in these latter steps
which eventually lead to the formation of a fatty streak. T cell lymphocytes also become activated in
the atheroma and secrete pro- or anti-inflammatory cytokines. Degranulation of mast cells found in
the atheroma may also contribute to the lesion.
Atheroma progression involves the proliferation of smooth muscle cells, which is stimulated
by macrophage-derived growth factors (Libby, 2002, 192009). This results in smooth muscle
accumulation in the lesion, the elaboration of extracellular matrix material and the formation of a
more bulky lesion which can occlude the arterial lumen. Matrix proteins contribute to the evolution
of the lesion from a fatty streak to a fibrous plaque. Mechanisms which trigger plaque disruption,
including endothelial erosions and plaque rupture, can result in thrombosis as well as in further
expansion of the lesion. Resident T cells, mast cells and circulating platelets may also play a role in
destabilizing plaques (Halvorsen et al., 2008, 191149). It is thought that most atheromatous lesions
progress in a discontinuous manner due to cycles of disruption, expansion and repair.
A major factor regulating plaque disruption is the thickness of the fibrous cap, with more
stable plaques characterized by a thick fibrous cap (Libby, 2002, 192009). Collagen in the fibrous
cap can be degraded by proteases, especially MMPs. Inflammation in the intima reduces collagen
production by smooth muscle cells and promotes the expression and activation of MMPs. ROS may
mediate MMP upregulation (Lund et al., 2009, 180257). Both macrophages and smooth muscle cells
produce MMPs in the lesion areas (Halvorsen et al., 2008, 191149): fully differentiated macrophages
selectively upregulate certain MMPs with more destructive potential (Newby, 2008, 191161).
Rupture of the fibrous cap allows pro-thrombotic factors within the plaque (e.g., TF) to come into
contact with coagulation factors in the blood possibly resulting in the formation of an occlusive
blood clot. However in some cases, blood fibrinolytic mechanisms minimize clot formation and
repair processes ensue. The result is a more fibrous plaque and/or an expanded lesion. The proposed
role of PM in activating coagulation pathways, which was discussed above, may influence the
outcome of plaque disruption.
In this manner oxidative stress, inflammation and pro-coagulant activity in the blood are
involved in the initiation and progression of atheromatous lesions as well as plaque disruption and
occlusive blood clot formation. PM exposure may contribute to these pathways and in fact, studies
December 2009 5-15
-------
described in later chapters demonstrate PM-dependent effects on atherosclerosis progression (Araujo
et al, 2008, 156222; Chen and Nadziejko, 2005, 087219; Sun et al., 2005, 087952; Ying et al,
2009, 190111).
5.3. Activation of the Autonomic Nervous System by
Pulmonary Reflexes
Chemosensitive receptors, including rapidly adapting receptors (RARs) and sensory C-fiber
receptors, are found at all levels of the respiratory tract and are sensitive to irritant particles as well
as to irritant gases (Alarie, 1973, 070967; Coleridge and Coleridge, 1994, 156362; Widdicombe,
2006, 155519). Activation of trigeminal afferents in the nose causes CNS reflexes resulting in
decreases in respiratory rate through a lengthened expiratory phase, closure of the glottis, closure of
the nares with increased nasal airflow resistance and effects on the cardiovascular system such as
bradycardia, peripheral vasoconstriction and a rise in systolic arterial blood pressure (Alarie, 1973,
070967). Sneezing, rhinorrhea and vasodilation with subsequent nasal vascular congestion are also
nasal reflex responses involving the trigeminal nerve (Sarin et al., 2006, 191166). Activation of vagal
afferents in the tracheobronchial and alveolar regions of the respiratory tract causes CNS reflexes
resulting in bronchoconstriction, mucus secretion, mucosal vasodilation, cough, apnea followed by
rapid shallow breathing and effects on the cardiovascular system such as bradycardia and
hypotension or hypertension (Coleridge and Coleridge, 1994, 156362; Widdicombe, 2003, 157145;
2006, 155519; Widdicombe and Lee, 2001, 019049). Some evidence suggests that cardiovascular
responses may be mediated primarily by the C-fiber receptors (Coleridge and Coleridge, 1994,
156362) and that irritants in the lower respiratory tract cause more pronounced cardiovascular
responses than irritants in the upper respiratory tract (Widdicombe and Lee, 2001, 019049).
Early experiments demonstrated that sectioning of the trigeminal nerve abrogated irritant
effects on respiratory rate, heart rate and systolic arterial blood pressure (Alarie, 1973, 070967).
These nasal reflexes were attributed to the ophthalmic branch of the trigeminal nerve since they were
identical to reflex responses following diving or immersion of the face in water (Alarie, 1973,
070967). Early experiments also demonstrated that non-nasal reflexes were mediated by cholinergic
parasympathetic pathways involving the vagus nerve and inhibited by atropine (Grunstein et al.,
1977, 071445; Nadel et al., 1965, 014846). More recent experiments have shown that noncholinergic
mechanisms may also be involved. For example, stimulation of C-fiber receptors can activate local
axon reflexes. These local axon pathways are responsible for secretion of neuropeptides and the
development of neurogenic inflammation (Widdicombe and Lee, 2001, 019049). It has been
proposed that, in some cases, neurogenic pulmonary responses can switch from their normally
protective function to one that perpetuates pulmonary inflammation (Wong et al., 2003, 097707).
Differences in respiratory tract innervation between rodents and humans suggest that C-fiber
mediated neurogenic inflammation may be more important in rodents than in humans (Groneberg et
al., 2004, 138134; Widdicombe, 2003, 157145; Widdicombe and Lee, 2001, 019049) However, the
role of neurogenic inflammation in mediating pulmonary responses in humans is an active area of
investigation.
VR1 receptors represent a subset of neuropeptide and acid-sensitive irritant receptors which
belong to the transient receptor potential (TRP) family. They are located on the sensory C-fibers
which lie underneath and between lung epithelial cells and on immune and non-immune airway
cells. Some investigators have focused on the role played by these receptors in mediating
inflammation following exposure to PM (Veronesi and Oortgiesen, 2001, 015977). Exposure of
bronchial epithelial cells and neurons to PM in vitro has been shown to result in an immediate
increase in intracellular calcium followed by the release of neuropeptides and inflammatory
cytokines (Veronesi et al., 1999, 048764; Veronesi et al., 2000, 017062). In one study, this response
was found to be due to an intrinsic property of the particle core and was not metal-dependent
(Oortgiesen et al., 2000, 013998). while in another study electrostatic charge was found to activate
VR1 receptors (Veronesi et al., 2003, 094384). PM-mediated activation of VR1 receptors which
results in increases in intracellular calcium and apoptosis in epithelial cells has also been
demonstrated (Agopyan et al., 2003, 155649). New studies, discussed in later chapters, provide
evidence for the involvement of TRP VI irritant receptor involvement in PM-dependent responses
(Ghelfi et al., 2008, 156468; Rhoden et al., 2005, 087878).
December 2009 5-16
-------
Recently it has been proposed that pulmonary reflex responses may be modulated by CNS
plasticity (Bonham et al, 2006, 191140; Sekizawa et al., 2008, 191167). Plasticity is a property of
neurons or synapses which allows change in response to previous events. In the case of the
respiratory tract, visceral afferent inputs to the CNS are integrated primarily in the nucleus tractus
soltarius (NTS) region of the brain (Bonham et al., 2006, 191140). Inputs from local networks,
higher brain regions and circulating mediators contribute to the reflex output (Bonham et al., 2006,
191140). This integration allows for plasticity in that repeated or prolonged exposure to a particular
stimuli may lead to altered reflex responses to the same or subsequent stimuli. An exaggerated reflex
response was recently observed in guinea pigs exposed to ETS for an extended period of time
(Sekizawa et al., 2008, 191167). It is not known whether CNS plasticity influences responses
following acute and chronic exposure to ambient PM, but it is a mechanism that may possibly
explain hyperresponsiveness and/or adaptation of reflex-related responses.
At this time, it is not clear how activation of the ANS by pulmonary reflexes contributes to the
kinds of altered conduction and/or repolarization properties of the heart which may be linked to
arrhythmias (Figure 5-5). Pulmonary reflexes, as they are currently understood, initially lead to
increases in parasympathetic tone. However, decreased heart rate variability appears to be reflective
of decreased parasympathetic tone and/or increased sympathetic tone. A PM-dependent sympathetic
stress response mediated by cytokines has been postulated (Godleski et al., 2000, 000738). but there
is little new information to support this mechanism. Thus, activation of the autonomic nervous
system by mechanisms other than pulmonary reflexes seems likely in response to PM. Very little is
known about putative alternative mechanisms leading to sympathoexcitatory responses although one
study demonstrated a role for the olfactory bulb-NTS pathway in regulating cardiovascular functions
following smoke exposure (Moffitt et al., 2002, 191160). In addition, sympathoexcitatory responses
occur during myocardial ischemia, mediated by the release of adenosine or the production of ROS
by the myocardium (Longhurst et al., 2001, 191158). Hence, PM-dependent effects leading to
myocardial ischemia may stimulate sympathoexcitatory responses. Furthermore, the effects of
pre-existing alterations in the ANS due to disease processes (e.g., increased sympathetic tone
observed in cardiac diseases) on PM responses are not understood. Possibly, integration of neural
signals resulting from pulmonary and cardiac reflexes at the level of the NTS may have an influence
on ANS responses to PM. Further investigation will be required to clarify these mechanisms.
5.4. Translocation of UFPs or Soluble PM Components
UFPs can translocate across cell membranes by non-endocytotic mechanisms involving
adhesive interactions and diffusion (Geiser et al., 2005, 087362). as described in Section 4.3.3.1. In
this study, there was no measurable loss of PM from the lung over 24 h despite the rapid
translocation of inhaled UFPs into alveolar epithelial cells and capillary endothelial cells. In another
study, UFPs were localized in macrophage mitochondria as demonstrated by electron microscopy (Li
et al., 2003, 042082). Other studies found extrapulmonary translocation of poorly soluble UFPs, but
the process was slow and resulted in only a small amount leaving the lung. It is possible that in these
studies PM gained access to the circulation after initial transport to the lymph nodes or the
gastrointestinal system. Hence, there is limited evidence to date that UFPs or other PM size
fractions access the circulation by traversing the epithelial barrier of the respiratory tract.
However, soluble components from all size fractions of PM have the potential to translocate
across the airway epithelium into the bronchial circulation or across the alveolar epithelium into the
systemic circulation as depicted in Figure 5-5. Absorption across nasal epithelium may also occur
(Ilium, 2006, 191205). Factors affecting this process include the rate of dissolution of the solute
from the particle and the molecular weight of the solute (Section 4.4). More rapid dissolution of
soluble components may occur in the case of UFPs due to the higher surface/volume ratios compared
with larger particles.
Several interesting studies investigated the translocation of water-soluble metals from the lung.
Gilmour et al. (2006, 156472) demonstrated the rapid appearance of Zn in the plasma of rats
following IT instillation of zinc sulfate (ZnSO4). Similarly, Wallenborn et al. (2007, 156144)
demonstrated the rapid appearance of water-soluble metals in the blood, heart and liver following IT
instillation of oil combustion PM in rats. Using a more sensitive technique, these same investigators
demonstrated the accumulation of 70Zn, a rare isotope of Zn, in blood, heart and liver following IT
December 2009 5-17
-------
instillation of ZnSO4 (Wallenborn et al., 2009, 191172). In three other studies, soluble Zn and Cu
were associated with cardiac effects following IT instillation of rats with different forms of Zn- and
Cu-containing PM (Gilmour et al., 2006, 088489; Gottipolu et al., 2008, 191148; Kodavanti et al.,
2008, 155907). These results suggest the possibility that PM-derived soluble Zn and Cu translocated
across the alveolar-capillary barrier into the circulation and exerted effects on the heart. However, in
the two studies in which barrier function was measured it was found to be compromised (Gilmour et
al., 2006, 088489; Gottipolu et al., 2008, 191148) suggesting an acute lung injury response to IT
instillation of high concentrations of metal. Acute lung injury is not likely to occur in healthy
individuals exposed to PM at concentrations relevant to ambient levels. Cardiac effects were also
observed following subchronic inhalation exposure to low concentrations of aerosolized ZnSO4
(Wallenborn et al., 2008, 191171). Since it was not possible to measure extra-pulmonary Zn in this
study, it remains unclear whether cardiac effects were a direct effect of translocated Zn or an indirect
effect of exposure to Zn-containing PM. Nonetheless, translocation of soluble components derived
from inhaled PM remains a viable hypothesis to explain some extra-pulmonary effects.
Epithelial permeability is a key determinant of translocation and is discussed in detail in
Section 4.4.2. In brief, a number of studies have measured clearance of 99mTc-DTPA as an index of
alveolar epithelial membrane integrity and permeability of alveolar-capillary barrier (Braude et al.,
1986, 155701). Endothelial integrity also contributes to the alveolar-capillary barrier and is measured
by transvascular protein flux but is not discussed here (Braude et al., 1986, 155701). In laboratory
animals, increased alveolar permeability was shown in terminally senescent mice (Tankersley et al.,
2003, 096363). In human volunteers, epithelial permeability was transiently increased following 3 h
of moderate exercise but not following 24-h exposure to particle-rich urban air (Brauner et al., 2009,
190244). A previous study found that the exercise-induced increase in epithelial permeability was
transient and suggested that it was due to increased ventilation and elevated vascular pressure which
altered the properties of tight junctions (Hanel et al., 2003, 155826). Smokers (Jones et al., 1983,
155884) and individuals with acute respiratory distress syndrome (Braude et al., 1986, 155701) or
interstitial lung disease (Rinderknecht et al., 1980, 191965) also exhibited increased alveolar
epithelial permeability. The changes in smokers were reversible upon cessation of smoking.
Increased airway epithelial permeability was found in asthmatics when 99mTc-DTPA clearance was
used to measure the permeability of the bronchial mucosa (Ilowite et al., 1989, 156584). These
studies demonstrate that epithelial permeability is increased following moderate exercise and in lung
syndromes associated with inflammation and suggest that compromised epithelial barrier functions
in the lung may contribute to PM-mediated effects.
Interaction of circulating PM or soluble PM components with vascular endothelial cells,
platelets, and other leukocytes is a potential mechanism underlying the cardiovascular and systemic
effects of inhaled PM. A role for PM-derived ROS and/or cellular-derived ROS has been proposed.
Furthermore, soluble metals that do not redox-cycle may activate cell signaling pathways without the
generation of ROS. In this way, PM may promote adverse cardiovascular effects such as endothelial
dysfunction, atherosclerosis and thrombosis. Circulating PM or soluble PM components also have
the potential to impact other organ systems. However, convincing evidence that this occurs to an
appreciable extent in healthy individuals following inhalation of PM at concentrations relevant to
ambient exposures is lacking.
5.5. Disease of the Cardiovascular and Other Organ
Systems
As discussed above, deposition of PM in the lung may lead not only to pulmonary disease but
also to diseases of other systems (Figure 5-5). In the cardiovascular system, myocardial ischemia and
MI may occur as a result of the above proposed effects of PM on atherosclerosis, plaque instability,
thrombosis, plaque rupture and/or altered vasoreactivity of coronary vessels. Myocardial ischemia
and MI may alter the conduction and depolarization properties of the heart and lead to arrhythmic
events. In addition, thrombosis may lead to stroke and/or thromboembolic disease. Many of these
processes may be interlinked and responses to ambient PM exposures may involve multiple
mechanisms simultaneously with some variability depending on PM composition. Furthermore, it is
not clear at this time whether PM initiates cardiovascular disease or whether it perturbs existing
disease.
December 2009 5-18
-------
In addition, recent studies which are discussed in later chapters have demonstrated PM-
dependent effects on the CNS (Campbell et al., 2005, 087217; Kleinman et al, 2008, 190074;
Sirivelu et al., 2006, 111151; Veronesi et al., 2005, 087481; Win-Shwe et al., 2008, 190146). At this
time, it is not known whether this is a direct or indirect consequence of PM exposure. Translocation
of soluble and poorly soluble particles from the olfactory mucosa via the axons to the olfactory bulb
of the brain has been proposed as a possible mechanism by which PM or its components may
directly access the CNS. Evidence for this pathway is discussed in Section 4.3.3.2. Alternative
mechanisms proposed for PM-mediated CNS effects involve systemic inflammation and autonomic
responses. These are new and intriguing possibilities which warrant further investigation.
PM-dependent effects on the reproductive system, reproductive outcomes and perinatal
development have also been identified and are discussed in a later chapter. Mechanisms involved in
these responses have not been determined. However, it seems possible that systemic inflammation
and/or oxidative stress may play a role. Developmental windows of susceptibility may also be an
important consideration. Furthermore, it has been hypothesized that oxygen gradients and redox
status are key to cell differentiation and epigenetic processes occurring during development
(Section 5.1.11) (Hitchler and Domann, 2007, 191151).
5.6. Acute and Chronic Responses
In general, repeated acute responses may lead to cumulative effects which manifest as chronic
disease. Several examples relevant to the modes of action discussed in this chapter are that of
allergic responses, atherosclerosis and lung development. Allergic responses require repeated
exposures to antigen over time. Co-exposure to an adjuvant, possibly DEP or ultrafme concentrated
ambient particles (CAPs), can enhance this response. Furthermore, the presence of oxidative stress,
as may occur in response to PM, can contribute to allergic responses. Allergic sensitization often
underlies allergic asthma, characterized by inflammation and AHR. In this way, repeated or chronic
exposures involving multifactorial responses (immune system activation, oxidative stress,
inflammation) can lead to irreversible outcomes. Similarly, the development of atherosclerosis
involves inflammation and remodeling of the blood vessel wall. Factors contributing to this process
include systemic inflammation, endothelial dysfunction, oxidative stress and high levels of
circulating lipids. PM exposure is associated with three out of four of these processes. The role of
PM in initiating, promoting or complicating this disease or its outcomes has yet to be determined.
Critical windows of susceptibility during development also provide an opportunity for repeated
exposures to injurious agents to lead to irreversible changes in organ structure and function. The
extended period of postnatal lung development in humans and other species heightens this
vulnerability. PM may serve as such an injurious agent as has been demonstrated previously for
hyperoxia (Randell et al., 1990, 191956).
Furthermore, adverse outcomes may be precipitated by acute events superimposed on chronic
disease states. In the case of allergic asthma, acute PM exposure may provoke asthmatic responses
through oxidative stress and inflammatory pathways. Additionally, PM can act as a carrier of
aeroallergens and other biological materials which can potentially trigger asthma attacks. Similarly,
PM exposure may provoke inflammatory or thrombotic responses leading to rupture of an
atherosclerotic plaque which subsequently results in acute MI. In this way, the outcome of an acute
exposure to PM may be drastically worsened by the underlying chronic disease.
5.7. Results of New Inhalation Studies which Contribute
to Modes of Action
Prior to this review, much of the evidence for the proposed modes of action was obtained from
animal studies involving IT instillation or inhalation of high concentrations of PM and from cell
culture experiments. In many cases, the types of PM used were of questionable relevance to ambient
exposures (i.e., high concentrations of ROFA, metals and ambient PM collected on filters). Since
then, many inhalation studies have been conducted using CAPs, combustion-derived PM, urban air
December 2009 5-19
-------
and carbon black, generally using concentrations of PM lower than 1 mg/m3. Much of this research
has been conducted in animal models of disease. These key new studies, described in detail in
Chapters 6 and 7, add to the understanding of modes of action which are relevant to ambient PM
exposure. A compilation of pertinent results is found below.
• Altered lung function including changes in respiratory frequency and AHR following
short-term exposures to CAPs and combustion-derived PM (Section 6.3.2.3)
• Mild pulmonary inflammation in response to short-term exposures to CAPs, urban air,
combustion-derived PM and carbon black (Section 6.3.3.3)
• Mild pulmonary injury in response to short-term exposure to CAPs and combustion-
derived PM (Section 6.3.5.3)
• Inhibition of cell proliferation in the proximal alveolar region of neonatal animals
following short-term exposure to iron-soot (Section 6.3.5.3)
• Pulmonary oxidative stress in response to short-term exposure to CAPs, urban air,
combustion-derived PM, carbon black and iron-soot; pulmonary nitrosative stress in
response to titanium dioxide (TiO2) (Section 6.3.4.2)
• Antioxidant intervention which ameliorates PM effects on oxidative stress, allergic
responses, and AHR (Sections 6.3.4.2 )
• Allergic sensitization and exacerbation of allergic responses in response to CAPs and
combustion-derived PM (Section 6.3.6.3)
• Altered methylation of promoter regions of IFN-y and IL-4 genes suggestive of pro-
allergic Th2 gene activation following short-term exposure to combustion-derived PM in
an allergy model (Section 6.3.6.3)
• Increased susceptibility to respiratory infection following exposure to combustion-
derived PM (Section 6.3.7.2)
• Effects on nasal epithelial mucosubstances, airway morphology and airway
mucosubstances following chronic exposure to urban air-derived PM and woodsmoke
(Section 7.3.5.1)
• Worsening of papain-induced emphysema following chronic exposure to urban air-
derived PM (Section 7.3.5.1)
• Effects on lung development following chronic exposure to urban air-derived PM
(Sections 7.3.2.2 and 7.3.5.1)
• Prolonged exposure to CAPs and combustion-derived PM leading sometimes to mild
pulmonary inflammation, oxidative stress and injury and sometimes to loss of
inflammatory, oxidative stress and AHR responses which were observed after short-term
exposures (Sections 7.3.2.2, 7.3.3.2, 7.3.4.1, 7.3.5.1 and 7.3.6.2)
• Hypermethylation of lung DNA following chronic exposure to combustion-derived PM
(Section 7.3.5.1)
• A role for TRPV1 irritant receptors in activating local axon and CNS reflexes following
short-term exposure to CAPs and combustion-derived PM (Section 6.2.9.3)
December 2009 5-20
-------
A role for TRPV1 irritant receptors in mediating lung and heart oxidative stress through
increased parasympathetic and sympathetic activity in response to CAPs
(Sections 6.2.9.3 and 6.3.4.2)
Altered heart rate variability in response to CAPs, combustion-derived PM and carbon
black (Section 6.2.1.3)
Arrhymthmic events in response to CAPs and combustion-derived PM (Section 6.2.2.2)
Altered cardiac contractility following short-term exposure to CAPs and carbon black
(Section 6.2.6.1)
Enhanced myocardial ischemia following short-term exposure to CAPs (Section 6.2.3.3)
Endothelial dysfunction and altered vascular reactivity following short-term exposure to
CAPs, combustion-derived PM and TiO2 (Section 6.2.4.3)
Increases in blood pressure following short-term exposure to CAPs and carbon black
(Section 6.2.5.3)
Changes in blood leukocyte counts following short-term exposure to CAPs and carbon
black (Section 6.2.7.3)
Increased levels of blood coagulation factors following short-term exposure to CAPs and
on-road highway aerosols (Section 6.2.8.3)
Systemic and cardiovascular oxidative stress in response to short-term exposure to CAPs,
road dust and combustion-derived PM (Section 6.2.9.3)
Progression of atherosclerosis, induction of TF in aortic plaques, vascular oxidative
stress and altered vasomotor function following long-term exposure to CAPs in a
susceptible animal model (Section 7.2.1.2)
Vascular remodeling following chronic exposure to urban air-derived PM
(Section 7.2.1.2).
Enhanced angiotensin II-induced hypertension accompanied by vascular oxidative stress
and altered vasoreactivity in response to chronic exposure to CAPs (Section 7.2.5.2)
Exaggerated insulin resistance, visceral adiposity and systemic inflammation in response
to chronic exposure to CAPs and a high-fat diet (Section 7.2.3.1)
CNS responses following short- and long-term exposures to CAPs and combustion-
derived PM (Section 6.4.3)
Effects on the reproductive system, reproductive outcomes and developmental outcomes
following chronic exposure to urban-air derived PM (Section 7.4.2)
DNA adducts in nose, lung and liver following chronic exposure to urban air
(Section 7.5.2.1)
Germ line mutations, DNA strand breaks and global hypermethylation in sperm
following chronic exposure to urban air-derived PM (Section 7.5.3)
December 2009 5-21
-------
5.8. Gaps in Knowledge
The new studies highlighted in Section 5.7 confirm and extend findings from older studies.
However, this increasing body of evidence does not provide a complete picture of the biological
pathways involved in mediating PM effects. For example, a lack of information regarding the time-
dependence of many responses makes it difficult to understand the underlying biological
mechanisms. Existing gaps in knowledge include:
• The spatial distribution of retained particles in the lung and its impact
• The deposition, uptake and clearance of UFPs in the lung
• Effects of ambient PM exposures on epithelial barrier function in the lung
• Time dependence of responses
• The putative modulation of neural reflexes by pre-existing disease or other factors
• The putative role of neural reflexes besides those involving pulmonary irritant receptors
• The putative role of ET in altering vasomotor tone following PM exposure
• The putative translocation of PM or soluble components across the epithelial barrier of
the lung into the circulation
• The putative translocation of PM from olfactory epithelium to the olfactory bulb and
other brain regions
Additional studies will be required to clarify the biological mechanisms underlying the health
effects of PM.
December 2009 5-22
-------
Chapter 5 References
Adamson I; Prieditis H. (1995). Response of mouse lung to carbon deposition during injury and repair. Environ Health
Perspect, 103: 72-76. 189982
Adcock IM; Tsaprouni L; Bhavsar P; Ito K. (2007). Epigenetic regulation of airway inflammation. Curr Opin Immunol, 19:
694-700. 191178
AgopyanN; Bhatti T; Yu S; Simon SA. (2003). Vanilloid receptor activation by 2-and 10-|im particles induces responses
leading to apoptosis in human airway epithelial cells. Toxicol Appl Pharmacol, 192: 21-35. 155649
Ahsan MK; Nakamura H; Tanito M; Yamada K; Utsumi H; Yodoi J. (2005). Thioredoxin-1 suppresses lung injury and
apoptosis induced by diesel exhaust particles (DEP) by scavenging reactive oxygen species and by inhibiting DEP-
induced downregulation of Akt. Free Radic Biol Med, 39: 1549-1559. 156200
Alarie Y (1973). Sensory irritation by airborne chemicals. Grit Rev Toxicol, 2: 299-363. 070967
Amara N; Bachoual R; Desmard M; Golda S; Guichard C; Lanone S; Aubier M; Ogier-Denis E; Boczkowski J. (2007).
Diesel exhaust particles induce matrix metalloprotease-1 in human lung epithelial cells via a NADP(H)
oxidase/NOX4 redox-dependent mechanism. Am J Physiol Lung Cell Mol Physiol, 293: LI70-181. 156212
Araujo JA; Barajas B; Kleinman M; Wang X; Bennett BJ; Gong KW; Navab M; Harkema J; Sioutas C; Lusis AJ; Nel AE.
(2008). Ambient particulate pollutants in the ultrafine range promote early atherosclerosis and systemic oxidative
stress. Circ Res, 102: 589-596. 156222
Ayres JG; Borm P; Cassee FR; Castranova V; Donaldson K; Ghio A; Harrison RM; Hider R; Kelly F; Kooter IM; Marano
F; Maynard RL; Mudway I;NelA; SioutasC; Smith S; Baeza-SquibanA; Cho A; Duggan S; Froines J. (2008).
Evaluating the Toxicity of Airborne Particulate Matter and Nanoparticles by Measuring Oxidative Stress Potential -
A Workshop Report and Consensus Statement. Inhal Toxicol, 20: 75 - 99. 155666
Baccarelli A; Wright RO; Bollati V; Tarantini L; Litonjua AA; Suh HH; Zanobetti A; Sparrow D; Vokonas PS; Schwartz J.
(2009). Rapid DNA Methylation Changes after Exposure to Traffic Particles . Am J Respir Grit Care Med, 179:
572-578. 192155
Bachoual R; Boczkowski J; Goven D; Amara N; Tabet L; On D; Lecon-Malas V; Aubier M; Lanone S. (2007). Biological
effects of particles from the Paris subway system. Chem Res Toxicol, 20: 1426-1433. 155667
Barnes PJ. (2007). New molecular targets for the treatment of neutrophilic diseases. J Clin Immunol, 119: 1055-62. 191139
Bayram H; Ito K; Issa R; Ito M; Sukkar M; Chung KF. (2006). Regulation of human lung epithelial cell numbers by diesel
exhaust particles. Eur Respir J, 27: 705-713. 088439
Becher R; Bucht A; Ovrevik J; Hongslo JK; Dahlman HJ; Samuelsen JT; Schwarze PE. (2007). Involvement of NADPH
oxidase and iNOS in rodent pulmonary cytokine responses to urban air and mineral particles. Inhal Toxicol, 19:
645-655. 097125
Beck-Speier I; Dayal N; Karg E; Maier KL; Schumann G; Schulz H; Semmler M; Takenaka S; Stettmaier K; Bors W; Ghio
A; Samet JM; Heyder J. (2005). Oxidative stress and lipid mediators induced in alveolar macrophages by ultrafine
particles. Free Radic Biol Med, 38: 1080-1092. 156262
Becker S; Fenton MJ; Soukup JM. (2002). Involvement of microbial components and toll-like receptors 2 and 4 in cytokine
responses to air pollution particles. Am J Respir Cell Mol Biol, 27: 611-618. 052419
Becker S; Mundandhara S; Devlin RB; Madden M. (2005). Regulation of cytokine production in human alveolar
macrophages and airway epithelial cells in response to ambient air pollution particles: further mechanistic studies.
Toxicol Appl Pharmacol, 207: 269-275. 088590
Blanche! S; Ramgolam K; Baulig A; Marano F; Baeza-SquibanA. (2004). Fine particulate matter induces amphiregulin
secretion by bronchial epithelial cells. Am J Respir Cell Mol Biol, 30: 421-427. 087982
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009 5-23
-------
Blank F; Rothen-Rutishauser B; Gehr P. (2007). Dendritic cells and macrophages form a transepithelial network against
foreign particulate antigens. Am J Respir Cell Mol Biol, 36: 669-77. 096521
Bleck B; Tse Doris B; Jaspers I; de Lafaille Maria AC; Reibman J. (2006). Diesel exhaust particle-exposed human
bronchial epithelial cells induce dendritic cell maturation. J Immunol, 176: 7431-7437. 096560
Bonham AC; Chen CY; Sekizawa S; Joad JP (2006). Plasticity in the nucleus tractus solitarius and its influence on lung
and airway reflexes. J Appl Physiol, 101: 322-7. 191140
Braude S; Nolop KB; Hughes JMB; Barnes PJ; Royston D. (1986). Comparison of lung vascular and epithelial
permeability indices in the adult respiratory distress syndrome. Am Rev Respir Dis, 133: 1002-1005. 155701
Brauner EV; Mortensen J; Moller P; Bernard A; Vinzents P; Wahlin P; Glasius M; Loft S. (2009). Effects of ambient air
particulate exposure on blood-gas barrier permeability and lung function. Inhal Toxicol, 21: 38-47. 190244
Brown DM; Donaldson K; Borm PJ; Schins RP; Dehnhardt M; Gilmour P; Jimenez LA; Stone V. (2004). Calcium and
ROS-mediated activation of transcription factors and TNF-alpha cytokine gene expression in macrophages exposed
to ultrafine particles. Am J Physiol, 286: L344-L353. 155705
Brown DM; Donaldson K; Stone V. (2004). Effects of PM10 in human peripheral blood monocytes and J774 macrophages.
Respir Res, 5: 29. 088663
Caligiuri G; Levy B; Pernow J; Thoren P; Hansson GK. (1999). Myocardial infarction mediated by endothelin receptor
signaling in hypercholesterolemic mice. PNAS, 96: 6920-6924. 156318
Campbell A; Oldham M; Becaria A; Bondy SC; Meacher D; Sioutas C; Misra C; Mendez LB; Kleinman M. (2005).
Particulate matter in polluted air may increase biomarkers of inflammation in mouse brain. Neurotoxicology, 26:
133-140.087217
Cao D; Tal TL; Graves LM; Gilmour I; Linak W; Reed W; Bromberg PA; Samet JM. (2007). Diesel exhaust particulate-
induced activation of Stat3 requires activities of EGFR and Src in airway epithelial cells. Am J Physiol Lung Cell
Mol Physiol, 292: L422-L429. 156322
Chan RC-F; Wang M; Li N; Yanagawa Y; Onoe K; Lee JJ; Nel AE. (2006). Pro-oxidative diesel exhaust particle chemicals
inhibit LPS-induced dendritic cell responses involved in T-helper differentiation. J Allergy Clin Immunol, 118: 455-
465. 097468
Chang C-C; Chiu H-F; Wu Y-S; Li Y-C; Tsai M-L; Shen C-K; Yang C-Y (2005). The induction of vascular endothelial
growth factor by ultrafine carbon black contributes to the increase of alveolar-capillary permeability. Environ
Health Perspect, 113: 454-460. 097776
Chen J; Tan M; Nemmar A; Song W; Dong M; Zhang G; Li Y. (2006). Quantification of extrapulmonary translocation of
intratracheal-instilled particles in vivo in rats: effect of lipopolysaccharide. Toxicology, 222: 195-201. 147267
Chen LC; Nadziejko C. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice V CAPs
exacerbate aortic plaque development in hyperlipidemic mice. Inhal Toxicol, 17: 217-224. 087219
Cho AK; Sioutas C; Miguel AH; Kumagai Y; Schmitz DA; Singh M; Eiguren-Fernandez A; Froines JR. (2005). Redox
activity of airborne particulate matter at different sites in the Los Angeles Basin. Environ Res, 99: 40-47. 087937
Cho H-Y; Reddy SP; Kleeberger SR. (2006). Nrf2 defends the lung from oxidative stress. Antioxid Redox Signal, 8: 76-87.
156345
ChuAJ. (2005). Tissue factor mediates inflammation. Arch Biochem Biophys, 440: 123-132. 155730
Coleridge HM; Coleridge JCG (1994). Pulmonary Reflexes: Neural Mechanisms of Pulmonary Defense. AnnuRev
Physiol, 56: 69-91. 156362
Cowburn AS; Condliffe AM; Farahi N; Summers C; Chilvers ER. (2008). Advances in neutrophil biology: clinical
implications. Chest, 134: 606-12. 191142
de Haar C; Hassing I; Bol M; Bleumink R; Pieters R. (2006). Ultrafine but not fine particulate matter causes airway
inflammation and allergic airway sensitization to co-administered antigen in mice. Clin Exp Allergy, 36: 1469-
1479. 144746
De Kok TM; Hogervorst JG; Briede JJ; Van Herwijnen MH; Maas LM; Moonen EJ; Driece HA; Kleinjans JC. (2005).
Genotoxicity and physicochemical characteristics of traffic-related ambient particulate matter. Environ Mol
Mutagen, 46: 71-80. 088656
December 2009 5-24
-------
Dick CAJ; Brown DM; Donaldson K; Stone V. (2003). The role of free radicals in the toxic and inflammatory effects of
four different ultrafine particle types. Inhal Toxicol, 15: 39-52. 036605
Donaldson K; Borm PJA; Oberdorster G; Pinkerton KE; Stone V; Tran CL. (2008). Concordance between in vitro and in
vivo dosimetry in the proinflammatory effects of low-toxicity, low-solubility particles: the key role of the proximal
alveolar region. Inhal Toxicol, 20: 53-62. 190217
Donaldson K; Stone V; Borm PJA; Jimenez LA; Gilmour PS; Schins RPF; Knaapen AM; Rahman I; Faux SP; Brown DM;
MacNee W. (2003). Oxidative stress and calcium signaling in the adverse effects of environmental particles
(PM10). Free Radic Biol Med, 34: 1369-1382. 156408
Dostert C; Petrilli V; Van Bruggen R; Steele C; Mossman BT; Tschopp J. (2008). Innate Immune Activation Through Nalp3
Inflammasome Sensing of Asbestos and Silica. Science, 320: 674-677. 155753
Foley DL; Craig JM; Morley R; Olsson CA; Dwyer T; Smith K; Saffery R. (2009). Prospects for epigenetic epidemiology.
Am J Epidemiol, 169: 389-400. 191144
Gabelova A; Valovicova Z; Labaj J; Bacova G; Binkova B; Farmer Peter B. (2007). Assessment of oxidative DNA damage
formation by organic complex mixtures from airborne particles PM(10). Mutat Res, 620: 135-144. 156457
Gallagher J; Sams R; Inmon J; GeleinR; Elder A; Oberdorster G; Prahalad AK. (2003). Formation of 8-oxo-7,8-dihydro-2'-
deoxy guano sine in rat lung DNA following subchronic inhalation of carbon black. Toxicol Appl Pharmacol, 190:
224-231. 140171
Geiser M; Rothen-Rutishauser B; Kapp N; Schurch S; Kreyling W; Schulz H; Semmler M; Im Hof V; Heyder J; Gehr P.
(2005). Ultrafine particles cross cellular membranes by nonphagocytic mechanisms in lungs and in cultured cells.
Environ Health Perspect, 113: 1555-1560. 087362
Geng H; Meng Z; Zhang Q. (2005). Effects of blowing sand fine particles on plasma membrane permeability and fluidity,
and intracellular calcium levels of rat alveolar macrophages. Toxicol Lett, 157: 129-137. 096689
Geng H; Meng Z; Zhang Q. (2006). In vitro responses of rat alveolar macrophages to particle suspensions and water-
soluble components of dust storm PM(2.5). Toxicol In Vitro, 20: 575-584. 097026
Ghelfi E; Rhoden CR; Wellenius GA; Lawrence J; Gonzalez-Flecha B. (2008). Cardiac oxidative stress and
electrophysiological changes in rats exposed to concentrated air particles are mediated by TRP-dependent
pulmonary reflexes. Toxicol Sci, 102: 328-336. 156468
Ghio AJ; Churg A; Roggli VL. (2004). Ferruginous bodies: Implications in the mechanism of fiber and particle toxicity.
Toxicol Pathol, 32: 643-649. 155790
Ghio AJ; Cohen MD. (2005). Disruption of iron homeostasis as a mechanism of biologic effect by ambient air pollution
particles. Inhal Toxicol, 17: 709-716. 088272
Gilmour PS; Morrison ER; Vickers MA; Ford I; Ludlam CA; Greaves M; Donaldson K; MacNee W. (2005). The
procoagulant potential of environmental particles (PM10). Occup Environ Med, 62: 164-171. 087410
Gilmour PS; Nyska A; Schladweiler MC; McGee JK; Wallenbom JG; Richards JH; Kodavanti UP. (2006). Cardiovascular
and blood coagulative effects of pulmonary zinc exposure. Toxicol Appl Pharmacol, 211: 41-52. 088489
Gilmour PS; Rahman I; Donaldson K; MacNee W. (2003). Histone acetylation regulates epithelial IL-8 release mediated by
oxidative stress from environmental particles. Am J Physiol Lung Cell Mol Physiol, 284: L533-L540. 096959
Gilmour PS; Schladweiler MC; Nyska A; McGee JK; Thomas R; Jaskot RH; Schmid J; Kodavanti UP. (2006). Systemic
imbalance of essential metals and cardiac gene expression in rats following acute pulmonary zinc exposure. J
Toxicol Environ Health A, 69: 2011-2032. 156472
Godleski JJ; Verrier RL; Koutrakis P; Catalano P; Coull B; Reinisch U; Lovett EG; Lawrence J; Murthy GG; Wolfson JM;
Clarke RW; Nearing BD; Killingsworth C. (2000). Mechanisms of morbidity and mortality from exposure to
ambient air particles. Res Rep Health Eff Inst, 91: 5-88; discussion 89-103. 000738
Gopalakrishnan S; Van Emburgh BO; Robertson KD. (2008). DNA methylation in development and human disease. Mutat
Res, 647: 30-38. 191147
Gottipolu RR; Landa ER; Schladweiler MC; McGee JK; Ledbetter AD; Richards JH; Wallenborn GJ; Kodavanti UP.
(2008). Cardiopulmonary responses of intratracheally instilled tire particles and constituent metal components.
Inhal Toxicol, 20: 473-84. 191148
December 2009 5-25
-------
Groneberg DA; Quarcoo D; Frossard N; Fischer A. (2004). Neurogenic mechanisms in bronchial inflammatory diseases.
Allergy, 59: 1139-52. 138134
Grunstein MM; Hazucha M; Sorli J; Milic-Emili J. (1977). Effect of SO2 on control of breathing in anesthetized cats. J
ApplPhysiol, 43: 844-851.071445
Halvorsen B; Otterdal K; Dahl TB; Skjelland M; Gullestad L; Oie E; Aukrust P. (2008). Atherosclerotic plaque stability-
what determines the fate of a plaque?. Prog Cardiovasc Dis, 51: 183-94. 191149
Hanel B; Law I; Mortensen J. (2003). Maximal rowing has an acute effect on the blood-gas barrier in elite athletes. J Appl
Physiol, 95: 1076-1082. 155826
Hitchler MJ; Domann FE. (2007). An epigenetic perspective on the free radical theory of development. Free Radic Biol
Med, 43: 1023-1036. 191151
Huang Y-CT; Soukup J; Harder S; Becker S. (2003). Mitochondrial oxidant production by a pollutant dust and NO-
mediated apoptosis in human alveolar macrophage. Am J Physiol Lung Cell Mol Physiol, 284: C24-C32. 156573
Ilium L. (2006). Nasal clearance in health and disease. J Aerosol Med, 19: 92-99. 191205
Ilowite JS; Bennett WD; Sheetz MS; Groth ML; Nierman DM. (1989). Permeability of the bronchial mucosa to 99mTc-
DTPAin asthma. Am Rev Respir Dis, 139: 1139-1143. 156584
Imrich A; Ning Y; Lawrence J; Coull B; Gitin E; Knutson M; Kobzik L. (2007). Alveolar macrophage cytokine response to
air pollution particles: oxidant mechanisms. Toxicol Appl Pharmacol, 218: 256-264. 155859
Jaspers I; Ciencewicki JM; Zhang W; Brighton LE; Carson JL; Beck MA; Madden MC. (2005). Diesel exhaust enhances
influenza virus infections in respiratory epithelial cells. Toxicol Sci, 85: 990-1002. 088115
Jiang J; Oberdrster G; Elder A; Gelein R; Mercer P; Biswas P. (2008). Does nanoparticle activity depend upon size and
crystal phase?. Nanotoxicology, 2: 33-42. 156609
Jones JG; Minty BD; Royston D; Royston JP (1983). Carboxyhaemoglobin and pulmonary epithelial permeability in man.
Thorax, 38: 129-133. 155884
Jones PA; Baylin SB. (2007). The epigenomics of cancer. Cell, 128: 683-92. 191153
Kaan PM; Hegele RG (2003). Interaction between respiratory syncytial virus and particulate matter in guinea pig alveolar
macrophages. Am J Respir Cell Mol Biol, 28: 697-704. 095753
Karoly ED; Li Z; Dailey LA; Hyseni X; Huang YCT. (2007). Up-regulation of Tissue Factor in Human Pulmonary Artery
Endothelial Cells after Ultrafine Particle Exposure. Environ Health Perspect, 115: 535-540. 155890
Kaul N; Forman HJ. (1996). Activation of NF?B by the respiratory burst of macrophages. Free Radic Biol Med, 21: 401-
405. 155892
Keverne EB; Curley JP. (2008). Epigenetics, brain evolution and behaviour. Front Neuroendocrinol, 29: 398-412. 191154
Kleinman MT; Araujo JA; Nel A; Sioutas C; Campbell A; Cong PQ; Li H; Bondy SC. (2008). Inhaled ultrafine particulate
matter affects CNS inflammatory processes and may act via MAP kinase signaling pathways. Toxicol Lett, 178:
127-130. 190074
Kodavanti UP; Schladweiler MC; Gilmour PS; Wallenborn JG; Mandavilli BS; Ledbetter AD; Christiani DC; Runge MS;
Karoly ED; Costa DL; Peddada S; Jaskot R; Richards JH; Thomas R; Madamanchi NR; Nyska A. (2008). The role
of particulate matter-associated zinc in cardiac injury in rats. Environ Health Perspect, 116: 13-20. 155907
Koike E; Hirano S; Furuyama A; Kobayashi T. (2004). cDNA microarray analysis of rat alveolar epithelial cells following
exposure to organic extract of diesel exhaust particles. Toxicol Appl Pharmacol, 201: 178-185. 058555
Koike E; Kobayashi T. (2005). Organic extract of diesel exhaust particles stimulates expression of la and costimulatory
molecules associated with antigen presentation in rat peripheral blood monocytes but not in alveolar macrophages.
Toxicol Appl Pharmacol, 209: 277-285. 088303
Lee CC; Cheng YW; Kang JJ. (2005). Motorcycle exhaust particles induce IL-8 production through NF-kappaB activation
in human airway epithelial cells. J Toxicol Environ Health A, 68: 1537-1555. 156682
Lee SS; Woo CH; Chang JD; Kim JH. (2003). Roles of Rac and cytosolic phospholipase A2 in the intracellular signalling
in response to titanium particles. Cell Signal, 15: 339-345. 156678
December 2009 5-26
-------
Li J; Ohio AJ; Cho SH; Brinckerhoff CE; Simon SA; Liedtke W. (2009). Diesel exhaust particles activate the matrix-
metalloproteinase-1 gene in human bronchial epithelia in a beta-arrestin-dependent manner via activation of RAS.
Environ Health Perspect, 117: 400-409. 190424
Li N; Nel AE. (2006). Role of the Nrf2-mediated signaling pathway as a negative regulator of inflammation: Implications
for the impact of particulate pollutants on asthma. Antioxid Redox Signal, 8: 88-98. 156694
Li N; Sioutas C; Cho A; Schmitz D; Misra C; Sempf J; Wang M; Oberley T; Froines J; Nel A. (2003). Ultrafine particulate
pollutants induce oxidative stress and mitochondrial damage. Environ Health Perspect, 111: 455-460. 042082
Li N; Wang M; Bramble LA; Schmitz DA; Schauer JJ; Sioutas C; Harkema JR; Nel AE. (2009). The adjuvant effect of
ambient particulate matter is closely reflected by the particulate oxidant potential. Environ Health Perspect, 117:
1116-1123. 190457
Li Y-J; Kawada T; Matsumoto A; Azuma A; Kudoh S; Takizawa H; Sugawara I. (2007). Airway inflammatory responses to
oxidative stress induced by low-dose diesel exhaust particle exposure differ between mouse strains. Exp Lung Res,
33: 227-244. 155929
Libby P. (2002). Inflammation in atherosclerosis. Nature, 420: 868-874 . 192009
Lindbom J; Gustafsson M; Blomqvist G; Dahl A; Gudmundsson A; Swietlicki E; Ljungman AG (2007). Wear particles
generated from studded tires and pavement induces inflammatory reactions in mouse macrophage cells. Chem Res
Toxicol, 20: 937-946. 155934
Liu J; Ballaney M; Al-Alem U; Quan C; Jin X; Perera F; Chen LC; Miller RL. (2008). Combined Inhaled Diesel Exhaust
Particles and Allergen Exposure Alter Methylation of T Helper Genes and IgE Production In Vivo. Toxicol Sci, 102:
76-81. 156709
Long JF; Waldman WJ; Kristovich R; Williams M; Knight D; Dutta PK. (2005). Comparison of ultrastructural cytotoxic
effects of carbon and carbon/iron particulates on human monocyte-derived macrophages. Environ Health Perspect,
113: 170-174. 087454
Longhurst JC; Tjen-a-looi SC; Fu LW. (2001). Cardiac sympathetic afferent activation provoked by myocardial ischemia
and reperfusion. Mechanisms and reflexes. Ann N Y Acad Sci, 940: 74-95. 191158
Lund AK; Lucero J; Lucas S; Madden MC; McDonald JD; Seagrave JC; Knuckles TL; Campen MJ. (2009). Vehicular
emissions induce vascular MMP-9 expression and activity associated with endothelin-1 mediated pathways.
Arterioscler Thromb Vase Biol, 29: 511-517. 180257
Mackman N. (2005). Tissue-Specific Hemostasis in Mice. Arterioscler Thromb Vase Biol, 25: 2273-2281. 156722
Mercer BA; Kolesnikova N; Sonett J; D'Armiento J. (2004). Extracellular regulated kinase/mitogen activated protein
kinase is up-regulated in pulmonary emphysema and mediates matrix metalloproteinase-1 induction by cigarette
smoke. J Biol Chem, 279: 17690-17696. 191180
Moffitt JA; Grippo AJ; Holmes PV; Johnson AK. (2002). Olfactory bulbectomy attenuates cardiovascular
sympathoexcitatory reflexes in rats. Am J Physiol Heart Circ Physiol, 283: H2575-H2583. 191160
Molinelli AR; Madden MC; McGee JK; Stonehuerner JG; Ghio AJ. (2002). Effect of metal removal on the toxicity of
airborne particulate matter from the Utah Valley. Inhal Toxicol, 14: 1069-1086. 035347
Moller W; Brown DM; Kreyling WG; Stone V (2005). Ultrafine particles cause cytoskeletal dysfunctions in macrophages:
role of intracellular calcium. Part Fibre Toxicol, 2: 7. 156770
Monn C; Naef R; Roller T. (2003). Reactions of macrophages exposed to particles <10 "micron"m. Environ Res, 91: 35-44.
052418
Mutlu GM; Green D; Bellmeyer A; Baker CM; Burgess Z; Rajamannan N; Christman JW; Foiles N; Kamp DW; Ghio AJ;
Chandel NS; Dean DA; Sznajder JI; Budinger GR. (2007). Ambient particulate matter accelerates coagulation via
an IL-6-dependent pathway. JClin Invest, 117: 2952-2961. 121441
Nadel JA; Salem H; Tamplin B; Tokiwa Y. (1965). Mechanism of bronchoconstriction during inhalation of sulfur dioxide. J
ApplPhysiol, 20: 164-167. 014846
Newby AC. (2008). Metalloproteinase expression in monocytes and macrophages and its relationship to atherosclerotic
plaque instability. Arterioscler Thromb Vase Biol, 28: 2108-2114. 191161
December 2009 5-27
-------
Nurkiewicz TR; Porter DW; Barger M; Millecchia L; Rao KMK; Marvar PJ; Hubbs AF; Castranova V; Boegehold MA.
(2006). Systemic microvascular dysfunction and inflammation after pulmonary particulate matter exposure.
Environ Health Perspect, 114: 412-419. 088611
Nurkiewicz TR; Porter DW; Hubbs AF; Stone S; Chen BT; Frazer DG; Boegehold MA; Castranova V. (2009). Pulmonary
nanoparticle exposure disrupts systemic microvascular nitric oxide signaling. Toxicol Sci, 110: 191-203. 191961
Oberdorster G. (1988). Lung clearance of inhaled insoluble and soluble particles. J Aerosol Med, 1: 289-330. 006857
Oortgiesen M; Veronesi B; Eichenbaum G; Kiser PF; Simon SA. (2000). Residual oil fly ash and charged polymers activate
epithelial cells and nociceptive sensory neurons. Am J Physiol, 278: L683-L695. 013998
Owen CA. (2008). Roles for proteinases in the pathogenesis of chronic obstructive pulmonary disease. Int J of COPD, 3:
253-268. 191162
Pardo A; Selman M. (2005). MMP-1: the elder of the family. Int J Biochem Cell Biol, 37: 283-8. 191163
Perm A; Murphy G; Barker S; Henk W; Perm L. (2005). Combustion-derived ultrafine particles transport organic toxicants
to target respiratory cells. Environ Health Perspect, 113: 956-963. 088257
Pinkerton KE; Joad JP. (2006). Influence of air pollution on respiratory health during perinatal development. Clin Exp
Pharmacol Physiol, 33: 269-272. 091237
Rahman I; Biswas SK; Kode A. (2006). Oxidant and antioxidant balance in the airways and airway diseases. Eur J
Pharmacol, 533: 222-39. 191165
Ramage L; Guy K. (2004). Expression of C-reactive protein and heat-shock protein-70 in the lung epithelial cell line A549,
in response to PM10 exposure. Inhal Toxicol, 16: 447-452. 055640
Randell SH; Mercer RR; Young SL. (1990). Neonatal hyperoxia alters the pulmonary alveolar and capillary structure of 40-
day-oldrats. Am JPathol, 136: 1259-1266. 191956
Reed MD; Barrett EG; Campen MJ; Divine KK; Gigliotti AP; McDonald JD; Seagrave JC; Mauderly JL; Seilkop SK;
Swenberg JA. (2008). Health effects of subchronic inhalation exposure to gasoline engine exhaust. Inhal Toxicol,
20: 1125-1143. 156903
Rengasamy A; Barger MW; Kane E; Ma JKH; Castranova V; Ma JYC. (2003). Diesel exhaust particle-induced alterations
of pulmonary phase I and phase II enzymes of rats. J Toxicol Environ Health A,66: 153-167. 156907
Rhoden CR; Lawrence J; Godleski JJ; Gonzalez-Flecha B. (2004). N-acetylcysteine prevents lung inflammation after
short-term inhalation exposure to concentrated ambient particles. Toxicol Sci, 79: 296-303. 087969
Rhoden CR; Wellenius GA; Ghelfi E; Lawrence J; Gonzalez-Flecha B. (2005). PM-induced cardiac oxidative stress and
dysfunction are mediated by autonomic stimulation. Biochim Biophys Acta, 1725: 305-313. 087878
Rinderknecht J; Shapiro L; Krauthammer M; Taplin G; Wasserman K; Uszler JM; Effros RM. (1980). Accelerated
clearance of small solutes from the lungs in interstitial lung disease. Am Rev Respir Dis, 121: 105-117. 191965
Risom LMoller P; Loft S. (2005). Oxidative stress-induced DNA damage by particulate air pollution. Mutat Res, 592: 119-
137.089070
Roberts ES; Richards JH; Jaskot R; Dreher KL. (2003). Oxidative stress mediates air pollution particle-induced acute lung
injury and molecular pathology. Inhal Toxicol, 15: 1327-1346. 156051
Roberts JR; Young S-H; Castranova V; Antonini JM. (2007). Soluble metals in residual oil fly ash alter innate and adaptive
pulmonary immune responses to bacterial infection in rats. Toxicol Appl Pharmacol, 221: 306-319. 097623
Rouse RL; Murphy G; Boudreaux MJ; Paulsen DB; Perm AL. (2008). Soot nanoparticles promote biotransformation,
oxidative stress, and inflammation in murine lungs. Am J Respir Cell Mol Biol, 39: 198-207. 156930
Ruckerl R; Greven S; Ljungman P; Aalto P; Antoniades C; Bellander T; Berglind N; Chrysohoou C; Forastiere F;
JacqueminB; vonKlot S; Koenig W; Kuchenhoff H; Lanki T; Pekkanen J; Perucci CA; Schneider A; Sunyer J;
Peters A. (2007). Air pollution and inflammation (interleukin-6, C-reactive protein, fibrinogen) in myocardial
infarction survivors. Environ Health Perspect, 115: 1072-1080. 156931
Sakamoto N; Hayashi S; Gosselink J; Ishii H; Ishimatsu Y; Mukae H; Hogg JC; van Eeden SF. (2007). Calcium dependent
and independent cytokine synthesis by air pollution particle-exposed human bronchial epithelial cells. Toxicol Appl
Pharmacol, 225: 134-141.096282
December 2009 5-28
-------
Sarin S; UndemB; Sanico A; Togias A. (2006). The role of the nervous system in rhinitis. J Clin Immunol, 118: 999-1016.
191166
Schins RPF; Knaapen AM. (2007). Genotoxicity of poorly soluble particles. Inhal Toxicol, 19 Suppl 1: 189-198. 156074
Seaton A; MacNee W; Donaldson K; Godden D. (1995). Particulate air pollution and acute health effects. Lancet, 8943:
176-178. 045721
Sekizawa S; Chen CY; Bechtold AG; Tabor JM; Brie JM; Pinkerton KE; Joad JP; Bonham AC. (2008). Extended
secondhand tobacco smoke exposure induces plasticity in nucleus tractus solitarius second-order lung afferent
neurons in young guinea pigs. Eur J Neurosci, 28: 771-81. 191167
Semmler-Behnke M; Takenaka S; Fertsch S; Wenk A; Seitz J; Mayer P; Oberdorster G; Kreyling WG (2007). Efficient
elimination of inhaled nanoparticles from the alveolar region: evidence for interstitial uptake and subsequent
reentrainment onto airways epithelium. Environ Health Perspect, 115: 728-733. 156080
Shi T; Knaapen AM; Begerow J; Birmili W; Borm PJ; Schins RP (2003). Temporal variation of hydroxyl radical
generation and 8-hydroxy-2'-deoxyguanosine formation by coarse and fine particulate matter. Occup Environ Med,
60: 315-321.088248
Sirivelu MP; MohanKumar SMJ; Wagner JG; Harkema JR; MohanKumar PS. (2006). Activation of the stress axis and
neurochemical alterations in specific brain areas by concentrated ambient particle exposure with concomitant
allergic airway disease. Environ Health Perspect, 114: 870-874. 111151
Snipes MB. (1996). Current information on lung overload in nonrodent mammals: contrast with rats. Inhal Toxicol, 8: 91-
109. 076041
Snipes MB; Harkema JR; Hotchkiss JA; Bice DE. (1997). Neutrophil Involvement in the Retention and Clearance of Dust
Intratracheally Instilled into the LUNGS of F344/N Rats. Exp Lung Res, 23: 65-84. 156092
Soberanes S; Panduri V; Mutlu GM; Ghio A; Bundinger GRS; Kamp DW. (2006). p53 mediates particulate matter-induced
alveolar epithelial cell mitochondria-regulated apoptosis. Am J Respir Crit Care Med, 174: 1229-1238. 156991
Soberanes S; Urich D; Baker CM; Burgess Z; Chiarella SE; Bell EL; Ghio AJ; De Vizcaya-Ruiz A; Liu J; Ridge KM;
Kamp DW; Chandel NS; Schumacker PT; Mutlu GM; Budinger GRS. (2009). Mitochondrial complex Ill-generated
oxidants activate ASK1 and JNK to induce alveolar epithelial cell death following exposure to particulate matter air
pollution. J Biol Chem, 284: 2176-2186. 190483
Steerenberg P; Withagen C; Dalen W; Dormans J; Loveren H. (2004). Adjuvant Activity of Ambient Particulate Matter in
Macrophage Activity-Suppressed, N-Acetylcysteine-Treated, iNOS- and IL-4-Deficient Mice. Inhal Toxicol, 16:
835-843. 087981
Sun Q; Wang A; Jin X; Natanzon A; Duquaine D; Brook RD; Aguinaldo JG; Fayad ZA; Fuster V; Lippmann M; Chen LC;
Rajagopalan S. (2005). Long-term air pollution exposure and acceleration of atherosclerosis and vascular
inflammation in an animal model. JAMA, 294: 3003-3010. 087952
Takizawa H; Abe S; Okazaki H; Kohyama T; Sugawara I; Saito Y; Ohtoshi T; Kawasaki S; Desaki M; Nakahara K;
Yamamoto K; Matsushima K; Tanaka M; Sagai M; Kudoh S. (2003). Diesel exhaust particles upregulate eotaxin
gene expression in human bronchial epithelial cells via nuclear factor-kappa B-dependent pathway. Am J Physiol
Lung Cell Mol Physiol, 284: L1055-L1062. 157039
Tal TL; Graves LM; Silbajoris R; Bromberg PA; Wu W; Samet JM. (2006). Inhibition of protein tyrosine phosphatase
activity mediates epidermal growth factor receptor signaling in human airway epithelial cells exposed to Zn2+.
Toxicol Appl Pharmacol, 214: 16-23. 108588
Tamaoki J; Isono K; Takeyama K; Tagaya E; Nakata J; Nagai A. (2004). Ultrafine carbon black particles stimulate
proliferation of human airway epithelium via EGF receptor-mediated signaling pathway. Am J Physiol Lung Cell
MolPhysiol, 287: L1127-L1133. 157040
Tankersley CG; Shank JA; Flanders SE; Soutiere SE; Rabold R; Mitzner W; Wagner EM. (2003). Changes in lung
permeability and lung mechanics accompany homeostatic instability in senescent mice. J Appl Physiol, 95: 1681-
1687. 096363
Tao F; Gonzalez-Flecha B; Kobzik L. (2003). Reactive oxygen species in pulmonary inflammation by ambient particulates.
Free Radic Biol Med, 35: 327-340. 156111
December 2009 5-29
-------
Tao F; Kobzik L. (2002). Lung macrophage-epithelial cell interactions amplify particle-mediated cytokine release. Am J
Respir Cell Mol Biol, 26: 499-505. 157044
Tarantini L; Bonzini M; Apostoli P; Pegoraro V; Bollati V; Marinelli B; Cantone L; Rizzo G; Hou L; Schwartz J; Bertazzi
PA; Baccarelli A. (2009). Effects of particulate matter on genomic DNA methylation content and iNOS promoter
methylation. Environ Health Perspect, 117: 217-222. 192010
Tarantini L; Bonzini M; Apostoli P; Pegoraro V; Bollati V; Marinelli B; Cantone L; Rizzo G; Hou L; Schwartz J; Bertazzi
PA; Baccarelli A. (2009). Errata to Effects of particulate matter on genomic DNA methylation content and iNOS
promoter methylation . Environ Health Perspect, 117: A143. 192153
Thomson E; Kumarathasan P; Goegan P; Aubin RA; Vincent R. (2005). Differential regulation of the lung endothelin
system by urban particulate matter and ozone. Toxicol Sci, 88: 103-113. 087554
Upadhyay D; Panduri V; Ghio A; Kamp DW. (2003). Particulate matter induces alveolar epithelial cell DNA damage and
apoptosis: role of free radicals and the mitochondria. Am J Respir Cell Mol Biol, 29: 180-187. 097370
Valinluck V; Tsai HH; Rogstad DK; Burdzy A; Bird A; Sowers LC. (2004). Oxidative damage to methyl-CpG sequences
inhibits the binding of the methyl-CpG binding domain (MBD) of methyl-CpG binding protein 2 (MeCP2). Nucleic
Acids Res, 32: 4100-8. 191170
Van Eeden SF; Hogg JC. (2002). Systemic inflammatory response induced by particulate matter air pollution: the
importance of bone-marrow stimulation. J Toxicol Environ Health A, 65: 65:1597-1613. 088111
Van Eeden SF; Tan WC; Suwa T; Mukae H; Terashima T; Fujii T; Qui D; Vincent R; Hogg JC. (2001). Cytokines involved
in the systemic inflammatory response induced by exposure to particulate matter air pollutants (PM10). Am J
Respir Grit Care Med, 164: 826-830. 019018
van Eeden SF; Yeung A; Quinlam K; Hogg JC. (2005). Systemic response to ambient particulate matter: relevance to
chronic obstructive pulmonary disease. Proc Am Thorac Soc, 2: 61-67. 157086
Veronesi B; Makwana O; Pooler M; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles
VII Degeneration of dopaminergic neurons inApo E-/- mice. Inhal Toxicol, 17: 235-241. 087481
Veronesi B; Oortgiesen M. (2001). Neurogenic inflammation and particulate matter (PM) air pollutants. Neurotoxicology,
22: 795-810. 015977
Veronesi B; Oortgiesen M; Carter JD; Devlin RB. (1999). Particulate matter initiates inflammatory cytokine release by
activation of capsaicin and acid receptors in a human bronchial epithelial cell line. Toxicol Appl Pharmacol, 154:
106-115.048764
Veronesi B; Oortgiesen M; Roy J; Carter JD; Simon SA; Gavett SH. (2000). Vanilloid (capsaicin) receptors influence
inflammatory sensitivity in response to particulate matter. Toxicol Appl Pharmacol, 169: 66-76. 017062
Veronesi B; Wei G; Zeng JQ; Oortgiesen M. (2003). Electrostatic charge activates inflammatory vanilloid (VR1) receptors.
Neurotoxicology, 24: 463-473. 094384
Wallenborn JG; Evansky P; Shannahan JH; Vallanat B; Ledbetter AD; Schladweiler MC; Richards JH; Gottipolu RR;
Nyska A; Kodavanti UP. (2008). Subchronic inhalation of zinc sulfate induces cardiac changes in healthy rats.
Toxicol Appl Pharmacol, 232: 69-77. 191171
Wallenborn JG; Kovalcik KD; McGee JK; Landis MS; Kodavanti UP. (2009). Systemic translocation of (70)zinc: kinetics
following intratracheal instillation in rats. Toxicol Appl Pharmacol, 234: 25-32. 191172
Wallenborn JG; McGee John K; Schladweiler Mette C; Ledbetter Allen D; Kodavanti Urmila P. (2007). Systemic
translocation of particulate matter-associated metals following a single intratracheal instillation in rats. J Toxicol
Sci, 98: 231-239. 156144
Wan H; Winton HL; Soeller C; Tovey ER; Gruenert DC; Thompson PJ; Stewart GA; Taylor GW; Garrod DR; Cannell MB;
Robinson C. (1999). Der p 1 facilitates transepithelial allergen delivery by disruption of tight junctions. J Clin
Invest, 104: 123-133. 191903
Wan J; Diaz-Sanchez D. (2006). Phase II enzymes induction blocks the enhanced IgE production in B cells by diesel
exhaust particles. JImmunol, 177: 3477-3483. 097399
Wan J; Diaz-Sanchez D. (2007). Antioxidant enzyme induction: a new protective approach against the adverse effects of
diesel exhaust particles. Inhal Toxicol, 19 Suppl 1: 177-182. 156145
December 2009 5-30
-------
Wang L; Pinkerton KE. (2007). Air pollutant effects on fetal and early postnatal development. Birth Defects Res C Embryo
Today, 81: 144-154. 179975
Warheit DB; Webb TR; Colvin VL; Reed KL; Sayes CM. (2007). Pulmonary bioassay studies with nanoscale and fine-
quartz particles in rats: toxicity is not dependent upon particle size but on surface characteristics. Toxicol Sci, 95:
270-280. 090482
Widdicombe J. (2006). Reflexes from the lungs and airways: historical perspective. J Appl Physiol, 101: 628. 155519
Widdicombe J; Lee L-Y. (2001). Airway reflexes, autonomic function, and cardiovascular responses. Environ Health
Perspect, 4: 579-584. 019049
Widdicombe JG. (2003). Overview of neural pathways in allergy and asthma. Pulm Pharmacol Ther, 16: 23-30. 157145
Win-Shwe TT; Yamamoto S; Fujitani Y; Hirano S; Fujimaki H. (2008). Spatial learning and memory function-related gene
expression in the hippocampus of mouse exposed to nanoparticle-rich diesel exhaust. Neurotoxicology, 29: 940-
947. 190146
Wong SS; Sun NN; Keith I; Kweon C-B; Foster DE; Schauer James J; Witten ML. (2003). Tachykinin substance P
signaling involved in diesel exhaust-induced bronchopulmonary neurogenic inflammation in rats. Arch Toxicol, 77:
638-650. 097707
Wright RO; Baccarelli A. (2007). Metals and neurotoxicology. J Nutr, 137: 2809-2813. 191173
Xia T; Korge P; Weiss JN; Li N; Venkatesen MI; Sioutas C; Nel A. (2004). Quinones and aromatic chemical compounds in
particulate matter induce mitochondrial dysfunction: implications for ultrafine particle toxicity. Environ Health
Perspect, 112: 1347-1358. 087486
Xiao GG; Nel AE; Loo JA. (2005). Nitrotyrosine-modified proteins and oxidative stress induced by diesel exhaust
particles. Electrophoresis, 26: 280-292. 156164
Yauk C; Polyzos A; Rowan-Carroll A; Somers CM; Godschalk RW; Van Schooten FJ; Berndt ML; Pogribny IP; Koturbash
I; Williams A; Douglas GR; Kovalchuk O. (2008). Germ-line mutations, DNA damage, and global
hypermethylation in mice exposed to particulate air pollution in an urban/industrial location. PNAS, 105: 605-610.
157164
Yin XJ; Ma JYC; Antonini JM; Castranova V; Ma JKH. (2004). Roles of reactive oxygen species and heme oxygenase-1 in
modulation of alveolar macrophage-mediated pulmonary immune responses to listeria monocytogenes by diesel
exhaust particles. Toxicol Sci, 82: 143-153. 087983
Ying Z; Kampfrath T; Thurston G; Farrar B; Lippmann M; Wang A; Sun Q; Chen LC; Rajagopalan S. (2009). Ambient
particulates alter vascular function through induction of reactive oxygen and nitrogen species. Toxicol Sci, 1: 1-36.
190111
Zhang J; Ghio AJ; Chang W; Kamdar O; Rosen GD; Upadhyay D. (2007). Bim mediates mitochondria-regulated
particulate matter-induced apoptosis in alveolar epithelial cells. FEES Lett, 581: 4148-4152. 156179
Zhao H; Barger MW; Ma JKH; Castranova V; Ma JYC. (2006). Cooperation of the inducible nitric oxide synthase and
cytochrome P450 1 Al in mediating lung inflammation and mutagenicity induced by diesel exhaust particles.
Environ Health Perspect, 114: 1253-1258. 100996
December 2009 5-31
-------
Chapter 6. Integrated Health Effects
of Short-Term PM Exposure
6.1. Introduction
This chapter reviews, summarizes, and integrates the evidence of relationships between short-
term exposures to PM and a variety of health-related outcomes and endpoints. Cardiovascular and
respiratory health effects of short-term exposure to various size fractions and sources of PM have
been examined in numerous epidemiologic, controlled human exposure and toxicological studies. In
addition, there is a large body of literature evaluating the relationship between mortality and short-
term exposure to PM. The association between PM exposure and central nervous system function
has also been assessed, although far fewer studies are available. The research approaches used to
evaluate health effects of PM exposure are described in Section 1.5 along with advantages and
limitations of the various study types. Chapter 5 provides an overview of the potential
pathophysiological pathways and modes of action underlying the PM-induced health effects
observed in animal and human studies. Evidence from the scientific literature of specific
cardiovascular and systemic effects, respiratory effects, and central nervous system (CNS) effects
associated with exposure to PM are presented in Sections 6.2, 6.3, and 6.4, respectively. Evidence of
associations between short-term exposure to PM and mortality are described in Section 6.5. The
chapter concludes with an evaluation of PM-induced health effects attributable to specific
constituents or sources (Section 6.6). More detailed descriptions of each study evaluated for this
assessment are presented in Annexes C, D, E, and F.
Findings for cardiovascular and respiratory effects are presented by specific endpoint or
measure of effect, leading from more subtle health outcome measures (e.g., heart rate variability
[HRV]) to the more severe, such as hospitalization and mortality for cardiovascular disease.
Conclusions from the 2004 PM AQCD (U.S. EPA, 2004, 056905) are briefly summarized at the
beginning of each section, and the evaluation of evidence from recent studies builds upon what was
available during the previous review. For each health outcome, results are summarized for studies
from the specific scientific discipline, i.e., epidemiologic, controlled human exposure, and
toxicological studies. The sections conclude with summaries of the evidence on the various health
outcomes and integration of the findings that leads to conclusions regarding causality based upon the
framework described in Chapter 1. Determination of causality is made for the overall health effect
category, such as cardiovascular effects, with coherence, consistency and biological plausibility
being based upon the evidence from across disciplines and also across the suite of related health
outcomes ranging from the more subtle health outcomes to cause-specific mortality. In the summary
sections for cardiovascular and respiratory effects and all-cause mortality, the evidence is
summarized and independent conclusions drawn for relationships with PM2.5, PMi0_2.5, and ultrafine
particles (UFPs) (Sections 6.2.12, 6.3.10, and 6.5.3, respectively). Evidence of central nervous
system effects is also divided by scientific discipline; however, the lack of data does not allow for
informative summaries of effect by PM metric in discussing CNS effects (Section 6.4.4).
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009 6-1
-------
6.2. Cardiovascular and Systemic Effects
6.2.1. Heart Rate and Heart Rate Variability
Heart rate (HR), HRV, and BP are all regulated, in part, by the sympathetic and
parasympathetic nervous systems. Changes in one or more may increase the risk of cardiovascular
events (e.g., arrhythmias, MI, etc.). Decreases in HRV have been associated with cardiovascular
mortality/morbidity in older adults and those with significant heart disease (TFESC, 1996, 003061).
In addition, decreased HRV may precede some clinically important arrhythmias, such as atrial
fibrillation, as well as sudden cardiac death, in high risk populations (Chen and Tan, 2007, 197461;
Sandercock and Brodie, 2006, 197465; Thong and Raitt, 2007, 197462).
HRV is measured using electrocardiograms (ECG) and can be analyzed in the time domain
(e.g., standard deviation of all NN intervals [SDNN], square root of the mean squared successive NN
interval differences [rMSSD]), and/or the frequency domain measured by power spectral analysis
(e.g., high frequency [HF], low frequency [LF], ratio of LF to HF [LF/HF]). SDNN generally
reflects the overall modulation of HR by the autonomic nervous system (ANS), whereas rMSSD and
frequency variations in HR generally reflect parasympathetic activity. Thus, rMSSD is generally well
correlated with HF, which also reflects the parasympathetic modulation of HR. LF is predominately
determined by both sympathetic parasympathetic tone and increased LF/HF indicates
sympathoexcitation, which correlates with decreased overall HRV (SDNN, rMSSD). Thus LF/HF is
thought to estimate the ratio of sympathetic influences on HR to parasympathetic influences.
While HRV is commonly described as being a reflection of vagal and adrenergic input to the
heart, there is clearly a more complex phenomenon reflected in HRV parameters. Rowan et al. (2007,
191911) provide a review of HRV and its use and interpretation with respect to air pollution studies.
To summarize, HRV indices are excellent measures of extrapulmonary effects from inhaled
pollutants, but the characterization of the acute, reversible responses to air pollution as being either
parasympathetic or sympathetic in origin, much less predictive of some adverse outcomes such as
ventricular arrhythmia, is relatively unsupported by the clinical literature. This is consistent with the
2004 PM AQCD (U.S. EPA, 2004, 056905) which stated that there is inherent variability in the
minute-to-minute spectral measurements, but long-term HRV measures demonstrate excellent
day-to-day reproducibility.
The 2004 PM AQCD (U.S. EPA, 2004, 056905) presented limited evidence of PM-induced
changes in HRV. However, findings from epidemiologic, controlled human exposure and
toxicological studies demonstrated both decreases and increases in HRV following PM exposure.
Recent epidemiologic studies have demonstrated a more consistent decrease in HRV (SDNN and
rMSSD), which is supported by several controlled human exposure studies published since 2003. In
these studies, decreases in HRV were observed among healthy adults following short-term exposures
to PM2.5 and PMi0_2.5 CAPs. It is interesting to note that these effects were not observed in adults
with asthma or COPD. The effect of PM on HRV observed in animal toxicological studies continues
to vary greatly, which may be due in part to strain differences in baseline HRV.
6.2.1.1. Epidemiologic Studies
The 2004 PM AQCD (U.S. EPA, 2004, 056905) reviewed several studies of PM exposure and
HR or HRV and described mixed findings across studies. Several additional studies have
investigated the association between acute changes in multiple HRV parameters and ambient air
pollutant concentrations in the U.S., Canada, Europe, Mexico, and Asia. Features and results of these
studies are presented in Table 6-1, and are summarized below.
In a multicity study, Liao and colleagues (2004, 056590) used data from the fourth cohort
evaluation of the Atherosclerosis Risk in Communities (ARIC) Study (1996-1998). The 6,784
subjects were 45-64 yr of age and lived in Washington County, MD, Forsyth County, NC, or the
suburbs of Minneapolis, MN. Linear regression models were used to examine the change in HRV
associated with PMi0, O3, SO2, CO, and NO2 concentrations in the 1-3 days prior to ECG
measurement. Among all subjects, each 11.5 (ig/m3 increase in mean daily PMi0 concentration 1 day
before the ECG measurement was associated with a 0.06 ms2 decrease in log-transformed HF (95%
December 2009 6-2
-------
CI: -0.10 to -0.02) and a 1.03 ms decrease in SDNN (95% CI: -1.64 to -0.42). A smaller
non-significant decrease was also observed for log transformed LF. This reduction in cardiac
autonomic control was larger among hypertensive subjects, suggesting that this group may be
susceptible to the effects of PM.
In a study of randomly selected participants in the Women's Health Initiative (WHI), a
multicity U.S. study, Whitsel et al. (2009, 191980) found decreases in rMMSD and SDNN in
association with PM10 concentration. The associations were stronger among participants with
diabetes. For example, in subjects with impaired fasting glucose, the reduction in rMSSD was 8.3%
(-13.9, -2.4) among those with high levels of insulin and 0.6% (-2.1, 1.6) among those with low
levels of insulin. Similar results were observed comparing high and low levels of insulin resistance.
Timonen et al. (2006, 088747) conducted a multicity panel study of elderly subjects with
stable coronary heart disease who lived in 3 European cities (Amsterdam, the Netherlands; Erfurt,
Germany; or Helsinki, Finland). They collected ECGs biweekly for six months in each subject. This
analysis, done as part of the ULTRA Study, examined changes in HRV (resting, paced breathing,
supine, and 5-min beat-to-beat NN intervals) associated with changes in fixed monitor particulate
concentrations (PM2.5, PMi0_2.5) with an emphasis on counts of UFPs (0.01-0.1 (im particles) and
accumulation mode particles (ACP; 0.1-1.0 (im particles). Mixed models were first fit to estimate the
change in HRV associated with PM (UFP, ACP, PM2.5, and PMi0_2.s) concentrations on the same and
previous 4 days in each city. In pooled analyses, the most consistent results identified were for
LF/HF (Table 6-1). Estimates for PM2.5, however, differed across cities. PM2.5 was associated with
decreased HF power and increased LF/HF in Helsinki, increased HF power and decreased LF/HF in
Erfurt, and not associated with any HRV metric in Amsterdam. In a subsequent analysis, de Hartog
et al. (2009, 191904) investigated whether exposure misclassification, effect modification by
medication use, or particle composition differences across the three cities could explain the result
observed. These authors found that PM2.5 apportioned from traffic, long-range transported PM2.5 and
outdoor PM2 5 were associated with reduced HRV most strongly among those not taking beta-
blockers (Table 6-1). Indoor and personal PM25 were not associated with decreased HRV in this
study. Therefore, the authors concluded that effect modification by medication use and particle
composition differences across the three cities may, in part, explain the heterogeneous PM2 5 findings
in the previous analysis.
The association between HRV and short-term increases in PM2 5, PM10_2.5, PM10, other size
fractions and components was also examined in single-city studies conducted in the U.S. or Canada
(Table 6-1). Among U.S. and Canadian cities, increases in PM25 were generally associated with
decreased SDNN and/or decreased HF power but not in all studies. However, studies also reported
increased SDNN associated with PM25 concentrations (Riediker et al., 2004, 056992; Wheeler et
al., 2006, 088453). In addition, Yeatts et al. (2007, 091266) reported increased rMSSD and HF
power with increased PM25 concentrations as well as SDANN5 (standard deviation of the average of
normal to normal intervals in all 5-min intervals in a 24-h period), and SDNN24HR (standard
deviation of the average of all normal to normal intervals in a 24-h period).
Other size fractions (e.g. coarse PM and UFPs) were also associated with decreases in HRV
metrics in several single-city studies conducted in the U.S. or Canada. Lipsett et al. (2006, 088753)
reported significantly decreased SDNN associated with increases in 2- and 6-h mean PMi0 and
PM10_2.5 concentrations. Yeatts et al. (2007, 091266) reported decreased rMSSD, SDNN24HR,
SDANN5, ASDNN5 (mean of the standard deviation in all 5-min segments of a 24-h recording),
proportion of NN intervals <50 m apart (pNN50) (7 min and 24 h), and HF power associated with
increased PMi0_25 concentration. Of those studies examining HRV associations with particle counts
(Adar et al., 2007, 001458: Park et al., 2005, 057331). only Adar et al. (2007, 001458) found clear
evidence of such effects (e.g., decreased SDNN, LF, HF). Decreased HRV was also associated with
increases in ambient mean SO42~ concentration (Luttmann-Gibson et al., 2006, 089794). ambient
mean BC concentration (Park et al., 2005, 057331; Schwartz et al., 2005, 074317). and traffic
generated particles/pollution (Adar et al., 2007, 001458; Riediker et al., 2004, 056992) in these
single-city studies.
Studies in Asia, Europe, and Mexico have also reported decreases in one or several HRV
metrics (Table 6-1) associated with increases in PM25 concentration or other size fractions. However,
a study conducted in Scotland reported no PM-HRV associations (Barclay et al., 2009, 179935).
Riojas-Rodriguez et al. (2006, 156913) reported significantly decreased LF and HF power associated
with each 1 ppm increase in CO concentration, but only small non-significant decreases associated
with PM2.5.
December 2009 6-3
-------
Summary of Epidemiologic Studies of Heart Rate and HRV
HRV studies investigated lagged pollutant concentrations from 2 h-5 days before ECG
measurement, reporting effects associated with mean pollutant concentrations lagged as short as
1-2 h, and more consistently with lags of 24-48 h. Taken together, these international and
U.S./Canadian studies show decreases in HRV associated with PM25 in most studies that use SDNN,
rMSSD or HF power. The effects of PMi0_2.5, UFPs, and components were evaluated in fewer studies
but associations with decreased HRV (e.g., both time and frequency measures) were observed. PMi0
studies also found evidence for PM-induced alterations in HRV, however, it is difficult to determine
which size fraction of PMi0 (e.g., PMi0_2.5, PM25 or UFPs) imparts the effects observed. As a result,
PM10 studies provide supportive evidence for the overall effect of PM on HRV, but not for a specific
size fraction. The proportion of studies reporting decreases in HRV may be inflated by publication
bias (i.e., studies showing little or no effects are not submitted for publication).
HRV Studies Investigating Specific Mechanisms
Panel studies investigating PM-HRV associations have also been useful in investigating
potential mechanistic pathways by which PM may elicit a cardiovascular response. A series of
analyses using data from the Normative Aging Study, a cohort of older men living in the Boston
metropolitan area, has also provided mechanistic insights into the PM-HRV association (Baccarelli
et al, 2008, 191959: Chahine et al, 2007, 156327: Park et al, 2005, 057331: Park et al, 2006,
091245: Park et al., 2008, 156845: Schwartz et al., 2005, 086296).
Park et al. (2005, 057331) studied the association between short-term increases in ambient air
pollution and changes in HRV using males enrolled in the Normative Aging Study. Using linear
regression models, the association between HRV metrics and PM2.5, O3, NO2, SO2, CO, BC, and
particle number count (PNC) moving averages (ma) in the previous 4, 24, and 48 h were examined.
The modifying effects of hypertension, diabetes, ischemic heart disease (IHD), and use of
hypertensive medications were also estimated. Of the pollutants examined, only PM25 and O3 were
associated with reductions in HRV, and each pollutant's effect appeared independent of the other.
Each 8 (ig/m3 increase in mean PM25 concentration in the previous 48 h was associated with a 20.8%
decrease in HF power (95% CI: -34.2 to -4.6), with larger effects among subjects with hypertension,
IHD, and diabetes. The authors state that since BC concentrations were also associated with adverse
changes in HRV, this suggests that traffic pollution may be partially responsible for the HRV
changes.
Schwartz et al. (2005, 086296) examined the hypothesis that adverse changes in HRV due to
PM2 5 are mediated by an oxidative stress response among participants in the Normative Aging
Study. They examined whether the change in HF power associated with each 10 (ig/m3 increase in
48-h mean PM25 was modified by the presence or absence of the allele for glutathione S-transferase
Ml (GSTM1), use of statins, obesity, high neutrophil counts, higher blood pressure (BP), and/or
older age. In subjects without the GSTM1 allele and its protection against oxidative stress, each
10 (ig/m3 increase in 48-h mean PM25 concentration was associated with a 34% decrease in HF
power (95% CI: -52 to -9). There was no association among those with at least one copy of the
allele. Obesity and high neutrophil counts also worsened the effect of PM on HRV regardless of
allele.
Park et al. (2006, 091245) investigated whether transition metals may be responsible for
cardiorespiratory effects that are observed in association with PM2 5. Again using the Normative
Aging Study cohort, they investigated whether subjects with two hemochromatosis (HFE)
polymorphisms associated with increased iron uptake had a smaller decrease in HF power associated
with PM than those subjects without either variant. Each 10 (ig/m3 increase in 48-h mean PM25 was
associated with a 31.7% decrease in HF (95% CI: -48.1 to -10.3) among subjects without either
polymorphism, but not among those with the 2 protective HFE alleles.
Chahine et al. (2007, 156327) reported a 10.5% reduction in SDNN (95% CI: -18.2 to -2.2)
associated with each 10 (ig/m3 increase in the mean 48-h PM25 concentration among Normative
Aging Study participants without the GSTM1 allele, but only a 2.0% SDNN decrease
(95% CI: -11.3, 8.3) in those with the allele. This supports the PM-HF power findings of Schwartz
et al. (2005, 086296). Further, subjects with the long repeat polymorphism in the HO-1 promoter had
a greater decline in SDNN associated with each 10 (ig/m3 increase in the mean 48-h PM25
December 2009 6-4
-------
concentration (-8.5% [95% CI: -14.8 to -1.8) than those with the short repeat polymorphism in HO-1
(7.4 % increase [95% CI: -8.7 to 26.2). Again, this suggests that PM-HRV changes are mediated, in
part, by oxidative stress.
Baccarelli et al. (2008, 191959) investigated whether the PM2.5-HRV association was modified
by dietary intakes of methyl nutrients (folate, vitamins B6 and B12, and methionine) and related
gene polymorphisms thought to either confer increased or decreased risk of CVD among men
enrolled in the Normative Aging Study. Each 10 (ig/m3 increase in PM2 5 in the previous 48 h was
associated with -8.8% (95% CI: -16.7 to -0.2) and -11.8% (95% CI: -20.8 to -1.8) decreases in
SDNN, among those with CC/TT genotypes of the C677T methylenetetrahydrofolate reductase
(MTHFR) polymorphism, and the CC genotype of the C1420T cytoplasmic serine
nydroxymethyltransferase (cSHMT) polymorphism, respectively. There were no changes among
those with CC MTHFR and CC/TT cSHMT. Further, there were similar HRV reductions in those
subjects with lower intakes of B6, B12, and/or methionine, but no decreases in those with high
intakes. Thus these genetic and nutritional variations in the methionine cycle may modify the PM-
HRV association.
Finally, among those Normative Aging Study subjects with high chronic lead exposure as
measured using X-ray fluorescence of the tibia, each 7 (ig/m3 increase in mean PM2.5 concentration
in the previous 48 h was associated with a 22% decrease in HF power (95% CI: -37.4 to -1.7) (Park
et al., 2008, 093027). Decreases in HF HRV were also associated with each 2.5 (ig/m3 increase in
mean SO42~ concentration in the previous 48 h (22% decrease [95% CI: -40.4 to 1.6). The authors
suggest that these findings are consistent with an oxidative stress response. Although this series of
studies suggest a role of oxidative stress and perhaps methyl nutrients and related polymorphisms in
these short-term associations of PM2.5 with HRV, replication by other investigators in other cities and
in other populations will aid interpretations of these findings.
Using data from a randomized controlled trial in Mexico City, Romieu et al. (2005, 086297)
investigated whether omega-3 fatty acids in fish oil supplements would mitigate the adverse effects
of acute PM exposure on HRV. Residents of a Mexico City nursing home were randomized to either
2 g/day offish oil or 2 g/day of soy oil. They used random-effects regression models to estimate the
change in HRV associated with mean PM2 5 concentration in the pre-supplementation and
supplementation phases. In the group receiving the fish oil supplement, each 8 (ig/m3 increase in 24-
h mean total PM25 exposure (weighted average of indoor and outdoor PM25 based on time activity
diaries) was associated with a 54% reduction (95% CI: -72 to -24) in log transformed HF power in
the pre-supplementation phase. However, in the supplementation phase of the trial, each 8 (ig/m3
increase in 24-h mean total PM2 5 concentration was associated with only a 7% reduction in log
transformed HF power (95% CI: -20 to 7). Decreases in other HRV parameters associated with PM25
were also muted in the supplementation phase. In the group receiving the soy oil supplement, the
reduction in HF power was also smaller in magnitude during the supplementation phase. However,
among those receiving the soy oil supplement, the differences between the pre-supplementation
PM2 5-HF change and the supplementation PM2 5-HF change were smaller compared to those
receiving the fish oil, and were not statistically significant. Romieu et al. (2008, 156922)also report
that omega-3 polyunsaturated fatty acids appear to modulate the adverse effect of PM25 based on
measured biomarkers of oxidative response (Section 6.2.9.1).
Summary of HRV Studies Investigating Specific Mechanisms
In summary, several analyses of data from the Normative Aging Study have provided evidence
that effect of PM25 on HRV is modulated by genetic polymorphisms related to oxidative stress
(Chahine et al., 2007, 156327; Park et al., 2006, 091245; Schwartz et al., 2005, 086296) or dietary
methyl nutrients or related genetic polymorphisms (Baccarelli et al., 2008, 191959). In addition,
preexisting conditions such as diabetes, IHD, and hypertension (Park et al., 2005, 057331; Whitsel
et al., 2009, 191980). beta-blocker use (Folino et al., 2009, 191902; Park et al., 2005, 057331).
chronic lead exposure (Park et al., 2008, 093027) and omega-3 fatty acid (Romieu et al., 2005,
086297) are reported to modulate the effect of PM25 on HRV.
December 2009 6-5
-------
Table 6-1.
Characteristics of epidemiologic studies investigating associations between PM and
changes in HRV.
PM Type, Exposure
Lag
Study Subjects
Ambient Concentration Recording CRMM i c HF,
(ug/m3)* Length SDNN LF rMSSD
LF/HF
MUL77C/rYSrUD/ES
Liao et al. (2004,
056590)
Whitsel et al.
(2009, 1919801
Timonen et al.
(2006, 0887471
De Hartog et al.
(2009, 1919041
PM,o, 24-h, lag 1-day
PM10, 24-h, 3-d avg
within 5 days preceding
exam
UFP, lags 0-2 days
AC, lags 0-2 days
PM25, lags 0-2 days
PM10.25, 2-day lag
24 h PM25 outdoor, PM25
traffic, long-range
transported PM25
N=6784 (mean age = 62 yrs),
ARIC study: MD, NC, MN
N=4295 randomly selected
participants in the WHI Trial
Stable IHD patients (65+ yr)
Amsterdam, Netherlands
(N=37)
Erfurt, Germany (N=47)
- Helsinki, Finland (N=47)
Stable IHD patients (65+)
Amsterdam, Netherlands
(N=37)
Erfurt, Germany (N=47)
Helsinki, Finland (N=47)
(Effects strongest among those
NOT taking beta-blockers)
24.3 5-min [ I I
28 visit 1
27 visit 2 10 second [ I
27 visit 3
Amsterdam:
17,300 particles/cm3
Erfurt:
21, 100 particles/cm3 i T
Helsinki:
17,000 particles/cm3
Amsterdam:
21 00 particles/cm3 5-min
Erfurt: (Pooled
1800 particles/cm3 estimates ^ 1
,,,.,. during paced
Helslnkl: 3 breathing
1400 particles/cm presented to
Amsterdam: 20.0 Inerl9n"
Erfurt: 23.1 [ t
Helsinki: 12.7
Amsterdam: 15.3
Erfurt: 3.7
Helsinki: 6.7
Median Outdoor:
Amsterdam: 16.7
5min I I
Erfurt: 16.3
Helsinki: 10.6
1
1
1
1
U.S. AND CANADIAN STUDIES
Park etal. (2005,
0573311
Riediker et al.
(2004, 056992)
Schwartz et al.
(2005, 0743171
Yeatts et al. (2007,
0912661
PM25, 48-h avg
PNC, 48-h avg
BC, 48-h avg
In-vehicle PM25 (mass)
9-h avg
BC, 24-h
PM25, 24-h
Secondary PM
(estimated), 1-h
PM,o.2.5, 24-h
PM25,24-h
N=497 men (mean age = 73
yr), Normative Aging Study
Boston, MA
N=9 healthy state police
N=28 older adults (61 -89 yr),
- 1 2 wk follow-up, Boston, MA
N=12 adult asthmatics, Chapel
Hill, NC
24-h: 11. 4
98th: 30.58
24-h: 28, 942 (13, 527) 4-min
particles/cm3
24-h: 0.92 [ [ [
9-h in-vehicle: 23 10-min t ~» t
24-h Median: 1.0 [ I
24-h Median: 10 23.mjn 1 |
1-h Median: -1.7 [ I
24-h: 5.3 [ [ [
24-h: 12.5 t 1 t
t
i
r
i
r
r
r
December 2009
6-6
-------
Wheeler et al.
(2006, 0884531
Dales 2004 (2004,
099036)
Luttmann-Gibson
et al. (2006,
0897941
Adaretal. (2007,
0014581
Pope et al. (2004,
055238)
Sullivan et al.
(2005, 1094181
Lipsett et al. (2006,
0887531
Ebelt et al. (2005,
0569071
Baccarelli et al.
(2008, 1919591
Fan et al. (2008,
1919791
PM Type, Exposure
Lag
PM25, 4-h avg
PM25, 4-h avg
EC, 4-h avg
EC, 4-h avg
PM25, 24-h avg
(personal)
PM25, lag 1-day
Sulfate, lag 1-day
Nonsulfate PM, lag 1-day
EC, lag 1-day
PM25, 24-h avg
BC, 24-h avg
PNC fine
PNC course
PM25(FRM), 24-h, lag
1-day
PM25, 1,2, 24-h avg
PM,0,
PM,0-2.5
PM25
PM10, 24 h
PM,0-2.5
PM25, 24-h
PM25 Sulfate, 24-h
outdoor
PM25, 48 h
PM2 5 personal, 1 h
Study Subjects
N=18 COPD, Atlanta, GA
N=1 2 Ml, Atlanta, GA
N= 18 COPD, Atlanta, GA
N=1 2 Ml, Atlanta, GA
N=36 IHD patients, Toronto,
Canada
N=32 (65+ yr)
Steubenville, OH
N=44 (60+ yr), diesel bus riders
St. Louis, MO
N=88 (65+ yr; 250 p-days),
Utah Valley
N=21 (65+ yr) with CVD,
Seattle WA
N=13 (65+ yr) w/out CVD,
Seattle WA
N=19IHD (65+yr), 12wkfu,
Coachella Valley, CA
N=16COPD, Vancouver,
Canada
N=549 Normative Aging Study
and residents of Boston
metropolitan area
N=11 crossing guards in New
Jersey
Ambient Concentration
(ug/m3)*
4-h: 17.8
• 4-h: 2.3
24-h personal: 19.9
24-h: 19.7
24-h: 6.9
24-h: 10.0
24-h: 1.1
24-h: 10.17
98th: 22. 43
330 ng/m3
42 particles/cm3
0.02 particles/cm3
23.7
• Median:10.7
31.0 and 46.1
None given
14 and 23. 2
17
5.6
11.4
98th: 23
2.0
Geometric mean (95%
confidence interval)
10.5(10.0,10.9)
Only change in 1-h PM25
reported
Morning shift: 35.2
Afternoon shift: 24.1
Recording CRMM i c HF,
Length SDNN LF rMSSD
t t t
on-min
r
i
Not described -> -> ->
1 1 1
1 1 1
r i
i i i
5-min III
1 1 1
r r r
24-h 1 |
5-min III
domain; 2-h, 1 1 ->
24-h Time
domain lit
1 1
r
24-h i i
i
7min I
24 h |
LF/HF
r
i
->
r
r
r
i
INTERNATIONAL STUDIES
Chan et al. (2004,
0873981
Chuang et al.
(2005, 0879891
NCo.02-1, 1-4 h
PMi.o-o.3, 1-4 h
PM2.5-i.o, 1-4 h
PM,o.2.5, 1-4 h
N=9 adults (1 9-29 yr) with lung
function impairment, Taipei,
Taiwan
N=10 adults (42-79 yr) with
lung function impairment,
Taipei, Taiwan
N=16, Patients with IHD/
hypertension , Taipei, Taiwan
23,407(19,836)
particles/cm
25,529 (20,783)
particles/cm3
37.2
12.6
14.0
1 1 1
1 1 1
5-min III
1 1 1
1 1 1
i
i
r
r
r
December 2009
6-7
-------
Holguin et al.
(2003, 0573261
Romieu et al.
(2005, 0862971
Riojas-Rodriguez
et al. (2006,
156913)
Barclay et al.
(2009, 1799351
Cardenas et al.
(2008, 1919001
Folino et al. (2009,
1919021
Min et al. (2008,
1919011
PM Type, Exposure
Lag
PM,,-o.3, 1-4 h
PM2.5-i.o, 1-4 h
PM,0-2.5, 1-4 h
PM25, 24-h
P M2 .5, 24-h (outdoor and
indoor)
Personal PM25
PM10, daily
PNC, daily
Estimated PM25 and PNC
PM25-outdoor
PM25-indoor
PM10,24h
PM25,24h
PM025,24h
PM,o, 12 h
Study Subjects
N=10 IHD, Taipei, Taiwan
N=21 without hypertension
(60-96 yr), Mexico City
N=13with hypertension (60-88
yr), Mexico City
N=50 nursing home residents
65+ yr, Mexico City
N=30 IHD patients, Mexico City
N=132, stable coronary heart
failure
Aberdeen, Scotland
N=52 (31 women, 21 men; 20-
40 yr), southeast of Mexico City
N=39 (36 male, 3 female;
mean age = 60 yr)
Padua, Italy
N=1349 (596 males; mean age
= 44 yr), Korea
Ambient Concentration
(ug/m3)*
26.8
10.9
16.4
•37.2
Outdoor: 19.4
Indoor: 18.3
Geometric mean: 46.8
Range of daily means: 7.4
to 68
Median PM25 outdoor:
28.3 pg/m3
Median PM25 indoor: 10.8
PM10
Summer: 46.4
Winter:73.0
Spring: 38.3
PM25
Summer: 33.9
Winter: 62.1
Spring: 30.8
PM0.25
Summer: 17.6
Winter: 30.5
Spring: 18.8
1-havg:33.2
Recording CRMM i c HF,
Length SDNN LF rMSSD
1 1 1
1 1 1
1 1 1
1 1
1 1
6-min
(Indoor PM25,
pre-supplement II 1
phase
presented)
5-min I I
24 h ->
15min I I
24 h |
5min III
LF/HF
->
1
r
r
r
i
Notes: Increases (|), decreases (J.) and no effects (-*) in HRV associated with PM concentration are indicated. Statistical significance was not necessary to categorize an effect as an increase or decrease.
For time domain measures moving average lags up to 24-h were explored. For frequency domain measures lags of 2-h, 4-h and 24-h were explored.
** All concentrations are means measured in u/m3, unless otherwise noted.
6.2.1.2. Controlled Human Exposure Studies
The 2004 PM AQCD (U.S. EPA, 2004, 056905) cited one study in which HRV indicators of
parasympathetic activity increased relative to filtered air control following a 2-h exposure with
intermittent exercise to PM2.5 CAPs (avg concentration 174 ug/m3) in both healthy and asthmatic
volunteers (Gong et al., 2003, 042106). This effect was observed immediately following exposure
and at 1 day post-exposure, but not at 4 h post-exposure. Although not statistically significant, HRV
(total power) increased following exposure to filtered air and decreased following exposure to CAPs.
More recent controlled human exposure studies are described below.
CAPs
Two new studies have evaluated the effect of PM2 5 CAPs (2-h exposures to concentrations of
20-200 ug/m3) on HRV in elderly subjects (Devlin et al., 2003, 087348: Gong et al., 2004, 087964).
In both studies, subjects experienced significant decreases in HRV following exposure to CAPs
relative to filtered air exposures. Interestingly, Gong et al. (2004, 087964) found that decreases in
HRV were more pronounced in healthy older adults than in those with COPD. In another study,
December 2009
6-8
-------
healthy and asthmatic adults were exposed to PM10_2.5 CAPs (avg concentration 157 ug/m3) for 2 h
with intermittent exercise (Gong et al., 2004, 055628). HRV was not affected immediately following
the exposure, but decreased in both groups at 4 and 22 h after the end of the exposure, with greater
responses observed in non-asthmatics. In a recent study among healthy adults exposed for 2 h with
intermittent exercise to PMi0_25 CAPs (avg concentration 89 ug/m3, MMAD 3.59 um, Chapel Hill,
NC), Graff et al. (2009, 191981) observed a significant decrease in overall HRV (SDNN) at 20 h
post-exposure, although no other measures of HRV were affected. Using a similar study design, the
same laboratory also evaluated the effect of ultrafme CAPs (avg concentration 49.8 ug/m3, <0.16 um
in diameter) on various HRV parameters (Samet et al., 2009, 191913) Relative to filtered air, both
HF and LF power increased 18 h following exposure to UF CAPs (36-42% increase per 105
particles/cm3). Exposure to UF CAPs, expressed as mass concentration, was not associated with
changes in HF power, and time domain parameters of HRV did not differ between CAPs and filtered
air in the 24 h following exposure. Gong et al. (2008, 156483) also recently evaluated changes in
HRV following controlled human exposures to UF CAPs and reported a small and transient decrease
in LF power (p < 0.05) among healthy (n = 17) and asthmatic (n = 14) adults 4 h after the completion
of a 2-h exposure with intermittent exercise in Los Angeles (avg concentration 100 ug/m3, avg
PNC 145,000/cm3). No other measure of HRV was shown to be significantly affected by exposure to
UF CAPs. In one of the largest studies of controlled human exposures to CAPs conducted to date,
Fakhri et al. (2009, 191914) evaluated changes in HRV among 50 adult volunteers during 2-h
exposures to PM2.5 CAPs (127 ug/m3) and O3 (114 ppb), alone and in combination. Neither exposure
to CAPs nor O3 resulted in any significant changes in HRV relative to filtered air. However, trends
were observed suggesting a negative concentration-response relationship between CAPs
concentration and SDNN, rMSSD, HF power and LF power when subjects were concomitantly
exposed to O3.
Diesel Exhaust
In a double-blind, crossover, controlled-exposure study, Peretz et al. (2008, 156855) exposed
three healthy adult volunteers and 13 adults with metabolic syndrome while at rest to filtered air and
two levels of diluted DE (PM2.5 concentrations of 100 and 200 ug/m3) in 2-h sessions. HRV
parameters were assessed prior to exposure, as well as at 1, 3, 6 and 22 h following the start of
exposure, and included both time domain (SDNN and rMSSD) and frequency domain parameters
(HF power, LF power, and the LF/HF ratio). In an analysis including all 16 subjects, the authors
observed an increase in HF power and a decrease in LF/HF 3 h after the start of exposure to
200 ug/m3 relative to filtered air. Although these changes were statistically significant (p < 0.05) the
effects were not consistent among the study subjects. No other significant effect of DE on HRV was
observed at either concentration or time point. The authors attributed the lack of consistent effects to
the small and non-homogeneous population and the timing of measurement. There was no difference
in either baseline or diesel-induced changes in HRV parameters between normal individuals and
patients with metabolic syndrome, although the number of normal individuals was quite small. It is
unclear if patients with metabolic syndrome were taking any medications.
Model Particles
Several additional recent controlled human exposure studies have evaluated the effect of
laboratory generated particles on HRV in healthy and health-compromised individuals. In a random
order crossover controlled human exposure study, Routledge et al. (2006, 088674) examined the
effects of UF elemental carbon (EC) particles (50 ug/m3) alone and in combination with 200 ppb
SO2 on HRV among 20 healthy older adults (age 56-75 yr), as well as 20 older adults with coronary
artery disease (age 52-74 yr). Five minute recordings of HRV data were obtained prior to and
immediately following the 1-h exposure, as well as 3 h post-exposure. In healthy subjects, exposure
to EC particles resulted in small increases in RR-interval, SDNN, rMSSD, and LF power
immediately following exposure compared to filtered air control. At 3 h post-exposure, there were no
significant differences in HRV measures between EC particle and filtered air exposures. Conversely,
SO2-induced decreases in HRV were observed at 3 h, but not immediately following exposure.
Concomitant exposure to EC particles and SO2 followed a pattern similar to that observed with SO2
December 2009 6-9
-------
alone, but did not reach statistical significance. Subjects with coronary artery disease did not
experience any significant changes in HRV following exposure to either pollutant. The authors
postulated that this lack of effect may be due to differences in medication between the two groups, as
70% of subjects with stable angina reported using (3 blockers, which are known to increase cardiac
vagal control. The lack of any significant effects on HRV following exposure to EC particles is an
important finding, as it provides evidence to suggest that the health effects observed following
exposure to PM may be due to particle constituents other than carbon, or to reactive species found on
the surface of the particle. These findings are in agreement with those of Zareba et al. (2009,
190101) who reported small and variable changes in HRV among a group of healthy adults
following exposure to UF EC. While exposure both at rest and during exercise to 10 ug/m3 UF EC
resulted in an increase in time domain parameters (rMSSD and SNDD), no such effect was observed
following exposure to a higher concentration of UF EC (25 ug/m3) in the same subjects. A recent
pilot study reported no effect of exposure to EC and ammonium nitrate particles (250-300 ug/m3) on
HRV parameters in five adults with allergic asthma (Power et al., 2008, 191982). However, when
the exposure occurred concomitantly with O3 (0.2 ppm), subjects were observed to experience
significant changes in both time and frequency HRV parameters. These observations should be
considered very preliminary as the study was limited by a small sample size (n = 5) and did not
evaluate the effect of exposure to O3 without particles. However, these findings are in agreement
with the previously described study of CAPs and O3 conducted by Fakhri et al. (2009, 191914). In
addition to the studies of laboratory generated carbon described above, Beckett et al. (2005, 156261)
used ZnO as a model particle and exposed twelve resting, healthy adults for 2 h to filtered air and
500 ug/m3 in the ultrafine (40.4 ± 2.7 nm) and fine (291.2 ± 20.2 nm) modes. Neither ultrafine nor
fine ZnO produced a significant change in any time or frequency domain parameter of HRV.
Summary of Controlled Human Exposure Study Findings for Heart Rate Variability
The results of several new controlled human exposure studies provide limited evidence to
suggest that acute exposure to near ambient levels of PM may be associated with small changes in
HRV. Changes in HRV parameters, however, are variable with some showing increased
parasympathetic activity relative to sympathetic activity and others showing the opposite. Although a
direct comparison between younger and older adults has not been made, PM exposure appears to
result in a decrease in HRV more consistently in healthy older adults (Devlin et al., 2003, 087348;
Gong et al.. 2004. 087964).
6.2.1.3. lexicological Studies
Toxicological studies that examined HR and HRV are presented in the 2004 PM AQCD
(U.S. EPA, 2004, 056905) and overall demonstrated differing responses, which were collectively
characterized as providing limited evidence for PM-related cardiovascular effects. The studies
described that reported HR or HRV effects following PM exposure were conducted with a variety of
particle types (CAPs, diesel, ROFA, metals), exposure methods (inhalation and IT instillation), and
doses (100-3,000 ug/m3 for inhalation; up to 8.3 mg/kg for IT instillation).
CAPs
Two groups of SH rats exposed to CAPs in Tuxedo, NY for 4 h (single-day mean PM2.5
concentrations 80 and 66 ug/m3; February 2001 and May 2001, respectively) demonstrated
decreased HR when exposure groups were combined that returned to baseline values when exposure
ceased (Nadziejko et al., 2002, 087460). Fine or UF H2SO4 exposure (mean concentration 225 and
468 ug/m3) did not induce any HR effects. Another study demonstrated a trend toward increased HR
in WKY rats following a 1- or 4-day PM2.5 CAPs exposure in Yokohama City, Japan (4.5 h/day; May
2004, November 2004, and September 2005), but the correlation between change in HR and
cumulative PM mass collected was not significant (Ito et al., 2008, 096823). Increased HR was
observed in SH rats exposed to PM2.5 CAPs for two 5-h periods during the spring (mean mass
concentration 202 ug/m3) in a suburb of Taipei, Taiwan (Chang et al., 2004, 055637). The response
December 2009 6-10
-------
was less prominent in the summer (mean mass concentration 141 ug/m3), despite the number
concentrations being similar for the two seasons (2.30><105 and 2.78x10 particles/cm3, respectively).
For HRV, decreased SDNN was observed in SH rats exposed to PM2.5 CAPs (mean mass
concentration 202 ug/m3; mean number concentration 2.30><105 particles/cm3) for two 5-h periods
separated by 24 h (Chang et al., 2005, 088662). Each of the four animals served as their own control
and the estimated mean PM effects for the SDNN decreases during exposure were 85-60% of
baseline. CAPs effects on rMSSD were less remarkable. In a study of Tuxedo, NY PM2.5 CAPs, no
acute changes in rMSSD or SDNN were observed in either ApoE~'~ or C57 mice when the 48-h time
period postexposure was evaluated (6 h/dayx5 day/wk; mean mass concentration over 5-mo period
110 ug/m3) (Chen and Hwang, 2005, 087218).
Diesel Exhaust
Anselme et al. (2007, 097084) used a MI model of congestive heart failure (CHF) where the
left anterior descending coronary artery of WKY rats was occluded to induce ischemia. After 3 mo of
recovery, rats were exposed to diesel emissions for 3 h (PM concentration 500 ug/m3; mass mobility
diameter 85 nm; NO2 1.1 ppm; CO 4.3 ppm) and decreases in rMSSD were observed during the first
2 h of the exposure, which returned to baseline values for the last hour of exposure. Healthy rats also
demonstrated decreased rMSSD when measured over the entire exposure period.
Model Particles
In WKY rats exposed to UF carbon particles (mass concentration 180 ug/m3; mean number
concentration 1.6x107 particles/cm3) for 24 h, HR increased and SDNN decreased during particle
inhalation (Harder et al., 2005, 087371). These measures returned to baseline values during the
recovery period. This study provides evidence that ultrafine carbon exerts its effects through changes
in ANS mediation, as the HR and HRV responses occurred quickly after exposure started and
pulmonary inflammation was only observed at the 24-h time point (and not at 4 h). SH rats exposed
to ultrafine carbon particles under the same conditions (mass concentration 172 ug/m3; mean number
concentration 9.0xl06 particles/cm3) demonstrated similar responses, albeit not until recovery days 2
and 3 (Upadhyay et al., 2008, 159345).
A model of premature senescence has been developed by Tankersley et al. (2003, 053919).
using aged AKR mice whose body weight abruptly declines ~5 wk prior to death and is accompanied
by deficiencies in other vital physiological function including HR and temperature regulation. When
exposed to carbon black ([CB]; mean concentration 160 ug/m3; 3 h/dayx3 day), terminal senescent
mice responded with robust cardiovascular effects, including bradycardia and increased rMSSD and
SDNN (Tankersley et al., 2004, 094378). SDNN and LF/HF were also increased in healthy
senescent mice exposed to CB. These studies indicate that HR regulatory mechanisms are altered in
susceptible mice exposed to PM (sympathetic and parasympathetic changes in healthy senescent
mice and increased parasympathetic influence in terminally senescent mice), which may translate
into lowered homeostatic competence in these animals. Results from the near-terminal group should
be interpreted with caution, as only three mice were in this group.
Subsequent research with a similar exposure protocol (mean CB concentration 159 ug/m3)
used C57BL/6J and C3H/HeJ mice to determine whether an acute PM challenge can modify HR
regulation in two mice strains with differing baseline HR (Tankersley et al., 2007, 097910). There
were no CB-specific effects on HR or HRV in C3H/HeJ compared to C57BL/6J mice (average HR
-80 bpm lower than C3H/HeJ at baseline). Administration of a sympathetic antagonist (propanolol)
to C57BL/6J mice prior to CB exposure resulted in elevated HR and decreased rMSSD compared to
air during the last 2 h of exposure, indicating withdrawal of parasympathetic tone. There may be
differences in regional particle deposition based on strain-specific breathing patterns that may affect
HR and HRV responses. However, this study revealed that inherent autonomic tone, which is
genetically varied between these mouse strains, may affect cardiovascular responses following PM
exposure. In extrapolating these results to humans, individual variation in genetic factors likely plays
some role in PM-induced adjustments in HR control via the ANS.
A recent study in mice (C3H/HeJ, C57BL/6J, and C3H/HeOuJ) examined the effects of a 2-h
O3 (mean concentration 0.584 ppm) pretreatment followed by a 3-h exposure to CB (mean
December 2009 6-11
-------
concentration 536 ug/m3) on HR and HRV measures (Hamade et al., 2008, 156515) HR decreased
to the greatest extent during O3 pre-exposure for all strains that were then exposed to CB. The
percent change in SDNN and rMSSD were increased in C3H mice during O3 pre-exposure and CB
exposure compared to the filtered air group; however, these HRV parameters gradually decreased
over the duration of the experiment and appeared to be O3 dependent. Together, these findings
indicate that increases in parasympathetic tone and/or decreases in sympathetic input may explain
the observed bradycardia. In a subset of all mice pre-exposed to O3, rMSSD remained significantly
elevated during the CB exposure compared to filtered air. The results from this study confirm what
was observed in Tankersley et al. (2007, 097910) in that genetic determinants affect HR regulation in
mice with exposure to air pollutants.
Summary of lexicological Study Findings for Heart Rate and Heart Rate Variability
Both increases and decreases in HR have been observed in rats or mice following PM
exposure. Fine or UF H2SO4 did not result in HR changes in SH rats. Similarly, decreased SDNN
was reported for UF CAPs exposure and lowered rMSSD was observed with diesel exposure. In
near-terminal senescent mice, HRV responses were robust following CB exposure and represented
increased parasympathetic influence. Strain differences in baseline HR and HRV likely contribute to
PM responses. HRV changes with preexposure to O3 and CB appeared to be O3 dependent, although
rMSSD remained elevated during PM exposure.
Source Apportionment and PM Components
An additional analysis of CAPs data (Chen and Hwang, 2005, 087218; Hwang et al., 2005,
087957) was conducted to link short-term HR and HRV effects to major PM source categories using
source apportionment methodology (Lippmann et al., 2005, 087453).
The source categories were: (1) regional secondary SO4Z~ comprised of high S, Si, and OC
(mean 63.41 ug/m3); (2) resuspended soil characterized by high concentrations of Ca, Fe, Al, and Si
(mean concentration 5.88 ug/m3); (3) fly ash emissions from power plants burning residual oil in the
eastern U.S. and containing high levels of V, Ni, and Se (mean concentration 1.53 ug/m3); and (4)
motor vehicle traffic and other unknown sources (34.92 ug/m3) (Lippmann et al., 2005, 087453).
Exposures occurred from 9:00 a.m. to 3:00 p.m., 5 days/wk for 5 mo. PM2.5 mass was associated
with a daily interquartile change of -4.1 beat/min HR during exposure in ApoE"7" mice1 and a similar
magnitude of effect was observed with resuspended soil (-4.5 beat/min). Resuspended soil was also
associated with a HR increase in the afternoon post-exposure (2.6 beat/min); the secondary SO42~
factor was linked to lowered HR in the same period (-2.5 beat/min). A 6.2% increase in rMSSD
collected in the afternoon post-exposure was associated with the residual oil factor, compared to a
5.6% and 2.4% decrease in rMSSD at night for secondary SO42~ and PM2 5 mass, respectively.
Resuspended soil was associated with a 4.3% increase in rMSSD the night following CAPs
exposure. The residual oil and secondary SO42~ categories showed similar statistically significant
parameter estimates for SDNN as rMSSD.
Recent studies of ECG alterations in mice have indicated a role for PM-associated Ni in
driving the cardiovascular effects. Lippman et al. (2006, 091165) presented a posthoc analysis of
daily variations in PM2.5 CAPs (mean concentration: 85.6 ug/m3; 7/21/ 2004-1/12/2005; Tuxedo,
NY) and changes in cardiac dynamics in ApoE"7" mice. On the 14 days that the exposed mice had
unusually elevated HR, Ni, Cr, and Fe comprised 12.4% of the PM mass, compared to only 1.5% on
Atherosclerosis and related pathways have been studied primarily in the Apolipoprotein E (ApoE) knockout mouse. Developed by
Nobuyo Maeda's group in 1992 (Piedrahita et al., 1992, 156868; Zhang et al., 1992, 157180). the ApoE-/- mouse and related models have
become the workhorse of atherosclerosis research over the past 15 years. The ApoE molecule is involved in the clearance of fats and
cholesterol. When ApoE (or the LDL receptor) is deleted from the genome, mice develop severely elevated lipid and cholesterol profiles;
ApoE"'" mice on a high-fat ("Western") diet exhibit cholesterol levels exceeding 1000 mgdL (normal is —150 mgdL) (Huber et al., 1999,
156575; Moore et al., 2005, 156780). As a result, the lipid uptake into the vasculature is increased and the atherosclerotic process is
dramatically hastened. Furthermore, the LDLs in ApoE" mice are highly susceptible to oxidation (Hayek et al., 1994, 156527). which may
be a crucial event in the air pollution-mediated vascular changes. However it should be noted that this model is primarily one of peripheral
vascular disease rather than coronary artery disease.
December 2009 6-12
-------
the other 89 days. Back trajectory analyses indicated high-altitude winds from the northwest that did
not traverse population centers and industrial areas except the Sudbury Ni smelter in Ontario,
Canada. On the 14 days that high HR was observed, the HR elevation lasted for two days, but only
the current day CAPs concentration was statistically significant. SDNN decreases were statistically
significant for all three lags (0, 1,2 days). The GAM regression analysis showed that only Ni
produced a statistically significant effect for HR and SDNN.
6.2.2. Arrhythmia
Epidemiologic and toxicological studies presented in the 2004 PM AQCD (U.S. EPA, 2004,
056905) provided some evidence of arrhythmia following exposure to PM. However, a positive
association between PM and ventricular arrhythmias among patients with implantable cardioverter
defibrillators was only observed in one study conducted in Boston, MA, while toxicological studies
reported arrhythmogenesis in rodents following exposure to ROFA, DE, or metals. Recent
epidemiologic studies have confirmed the findings of PM-induced ventricular arrhythmias in Boston,
MA, and have also reported increases in ectopic beats in studies conducted in the Midwest and
Pacific Northwest regions of the U.S. In addition, two studies from Germany have demonstrated
positive associations between traffic and combustion particles and changes in repolarization
parameters among patients with IHD. Findings of recent toxicological studies are mixed, with both
demonstrated decreases and increases in frequency of arrhythmia following exposure to CAPs.
6.2.2.1. Epidemiologic Studies
Studies of Arrhythmias Using Implantable Cardioverter Defibrillators
One study reviewed in the 2004 PM AQCD assessed the effect of short-term fluctuations in
PM2.5 on ventricular arrhythmias and several recent studies examining this relationship have been
conducted. Ventricular ectopy and arrhythmia include ventricular premature beats (VPBs),
ventricular tachycardia (VT), and ventricular fibrillation (VF). VPBs are spontaneous beats
originating from either the right or left ventricles. VT refers to three or more VPBs in succession at a
rate of 100 beats per minute or greater, while VF is characterized by rapid and disorganized
ventricular electrical activation incapable of generating an organized mechanical contraction or
cardiac output. AF is the most common type of arrhythmia. In this condition, ectopic electrical
impulses arising in the atria or pulmonary veins, i.e., outside their normal anatomic origin (the
sinoatrial node), can result in atrioventricular dilatation, dysfunction, and/or thromboembolism.
Despite being common, clinical and subclinical forms of AF are associated with reduced functional
status and quality of life. Moreover, the arrhythmia accounts for a large proportion of ischemic
stroke (Laupacis et al., 1994, 190901; Prystowsky et al., 1996, 156031) and is a strong risk factor
for CHF (Roy et al., 2009, 190902). contributing to both cardiovascular disease (CVD) and all-
cause mortality (Kannel et al., 1983, 156623).
Ventricular arrhythmia is commonly associated with myocardial infarction, heart failure,
cardiomyopathy, and other forms of structural (e.g., valvular) heart disease. Pathophysiologic
mechanisms underlying this established cause of sudden cardiac death include activators and
facilitators of arrhythmia, such as electrolyte abnormalities, modulation of the ANS, membrane
channels, gap junctions, oxidant stress, myocardial stretch and ischemia.
Previously, Peters et al. (2000, 011347) conducted a pilot study in Boston, MA to examine the
association between short-term changes in ambient air pollutant concentrations and increased risk of
ventricular arrhythmias, among a cohort of patients with implantable cardioverter defibrillators
(ICD). ICDs continuously monitor cardiac rhythm and upon detection of an abnormal rhythm (i.e.,
rapid HR), they can be programmed to deliver pacing and/or shock therapy to restore normal sinus
rhythm. Those abnormal rhythms that are most severe or rapid are assumed to be due to VT or VF
(i.e., life-threatening arrhythmias), and are thus treated with electric shock. These ICD devices also
store information on each abnormal rhythm detected, including the date, time, and therapy given.
Thus, using the date and time of those arrhythmias resulting in electric shock, Peters et al. (2000,
011347) reported an increased risk of ICD shock associated with mean NO2 concentration in the
December 2009 6-13
-------
previous two days. Among subjects with frequent events (10 or more during 3 yr of follow-up) an
increased risk of ICD shock was also associated with interquartile range increases in CO, NO2,
PM2.5, and BC in the previous 2 days. Several studies were conducted to confirm these findings. The
study characteristics, as well as the reported effect estimates and 95% CI associated with each PM
metric, are shown in Table 6-2.
Dockery et al. (2005, 078995: 2005, 090743) conducted a follow-up study of ICD patients
living in eastern Massachusetts and followed subjects for a longer period of time (up to 7 yr). They
were the first to review the ECG, classify each ICD-detected arrhythmia (e.g., ventricular
arrhythmia, VF, atrial tachycardia, sinus tachycardia, etc.), and include only ventricular arrhythmias
(VF or VT; excluding supraventricular arrhythmias). In single-pollutant models using generalized
estimating equations, increased risks of confirmed ventricular arrhythmias were associated with IQR
increases in every pollutant (PM2.5, BC, SO42~, NO2, SO2, O3, and PNC). Among those with a prior
ventricular arrhythmia in the past three days, interquartile range increases in 2-calendar-day mean
PM2 5, NO2, SO2, CO, O3, SO42~, and BC concentrations were all associated with significant and
markedly higher risks of ventricular arrhythmia than among those without a prior arrhythmia. The
pollutants associated with increased risk of ventricular arrhythmia implicate traffic pollution.
Rich et al. (2005, 079620) conducted a case-crossover analysis of these same data to
investigate moving average pollutant concentrations lagged <48 h. They reported an increased risk
of ventricular arrhythmia associated with mean PM25 and O3 concentrations in the 24 h before the
arrhythmia. Each pollutant effect appeared independent in two pollutant models. In single-pollutant
models, NO2 and SO2 were associated with increased risk, but when included in two pollutant
models with PM2 5, only PM2 5 remained associated with increased risk. They did not, however, find
evidence of a more acute arrhythmic response to pollution (i.e., larger risk estimates associated with
moving averages <24 h before arrhythmia detection). In an ancillary case-crossover analysis of data
from the Boston ICD study, Rich et al. (2006, 088427) identified 91 confirmed episodes of
paroxysmal AF among 29 subjects. In single pollutant models, they reported a significantly increased
risk of AF associated with mean O3 and PM25 concentrations in the hour before the arrhythmia and
BC concentration in the 24 h before the arrhythmia.
Rich et al. (2006, 089814) conducted another case-crossover study in the St. Louis, MO
metropolitan area. Using the same methods as in Boston, they reported increased risk of ventricular
arrhythmia associated with mean SO2 concentration in the 24 h before the arrhythmia, but not PM2 5
(in single-pollutant models). Again, they found no evidence of an arrhythmic response with moving
average pollutant concentrations <24 h before the arrhythmia.
In Vancouver, Canada, Vedal et al. (2004, 055630) did not find increased risk of ICD shocks
associated with increases in any pollutant concentration (PMi0, O3, SO2, NO2, and CO). Secondary
analyses among those subjects with two or more discharges per year, and analyses stratified by
season were also null for PMi0, although an association with SO2 (lag 2 days) was observed. A case
crossover analysis of these same data examining additional particle pollutant concentrations
available for a shorter time frame (e.g., PM2 5, SO42~, EC, and OC) also found no increased risk of
ICD shock associated with any pollutant (Rich et al., 2004, 055631).
The largest ICD study to date examined the risk of ventricular arrhythmias associated with
increases in the daily concentration of numerous PM and gaseous pollutants in Atlanta, GA (Metzger
et al., 2007, 092856) (see Table 6-2 for specific pollutants evaluated). Similar to Vedal et al. (2004,
055630). they did not find significant or consistently increased risk of a ventricular arrhythmia
associated with any IQR increase in mean daily PM or gaseous pollutant concentration at any lag
examined.
Ljungman et al. (2008, 180266) conducted a similar study, using case-crossover methods, on
ICD patients in Gothenburg and Stockholm, Sweden. They investigated the triggering of confirmed
ventricular arrhythmias by ambient PMi0 and NO2 concentrations, and reported increased relative
odds of ventricular arrhythmia associated with each 10 (ig/m3 increase in the 2-h ma PMi0
concentration (OR = 1.22 [95% CI: 1.00-1.51]), with a smaller non-significant risk associated with
each 10.3 ug/m3 increase in the 24-h ma PMi0 concentration (OR = 1.23 [95% CI: 0.87-1.73]). The
NO2 and PM2 5 effect estimates were much smaller and not statistically significant. Effect estimates
were larger for events occurring near the air pollution monitors in Gothenburg (compared to
Stockholm).
Albert et al. (2007, 156201). although not investigating associations with ambient pollution,
conducted a case-crossover study of the association between ventricular arrhythmia and traffic
exposure in the hours before the arrhythmia. They reported an increased risk of ventricular
December 2009 6-14
-------
arrhythmia associated with traffic exposure or driving in the previous hour. They hypothesized that
this increased risk was due to either a stress response from being in a car in heavy traffic, or from
traffic-generated air pollution, or a combination of both.
Table 6-2.
Reference
Dockery et al.
(2005, 078995;
2005, 0907431
Eastern MA
Rich et al. (2005,
0796201
Eastern MA
Rich et al. (2006,
0898141
St. Louis metro
area
Vedal et al.
(2004, 0556301
Vancouver, BC,
Canada
Ljungman et al.
(2008, 1802661
Gothenburg and
Stockholm,
Sweden
Rich et al. (2004,
0556311
Vancouver, BC,
Canada
Epidemiologic studies of ventricular arrhythmia and ambient PM concentration, in
patients with implantable cardioverterdefibrillators.
Outcome and
Sample Size
N=670 days with >1
confirmed ventricular
arrhythmias among
n=84 subjects
N=798 confirmed
ventricular
arrhythmias among
n=84 subjects
N=139 confirmed
ventricular
arrhythmias among
n=56 subjects
N=257 days with >1
ICD shock among
n=50 subjects
N=114 ventricular
arrhythmias among
73 subjects. 211 total
subjects were
followed.
N=77 to 98 days with
> 1 ICD shock
among n=34
subjects
Study Design and r*nnnii,,tnnt* PM
Analytic Method c°P°llutants Metric
Generalized estimating
equations Ng2] cc, Sg2]
Lags Evaluated: 2 calendar ^3
day means
Time-stratified case--
crossover study. NQ CQ SQ
Conditional logistic regres- 0 ' ' '
sion. Lags evaluated: 3, 6, 3
24, 48-h ma
Time-stratified
case-crossover study.
Conditional logistic N°2> co> s°2>
regression. Lags 3
Evaluated: 6, 12, 24, 48-h
ma
Generalized estimating
equations NC,2] co, S02,
Lags Evaluated: 0, 1,2, 3 °3
daily ma
Conditional logistic
regression NC,2
Lags evaluated: 2 h, 24 h
Am bi-directional
case-crossover study.
Conditional logistic
regression N02, CO, S02,
03
Lags Evaluated: 0, 1, 2,
and 3 day ma
PM25
BC
Sulfate
PNC
PM25
BC
PM25
EC
Organic
Carbon
PM10
PM,0,
PM25
PM25
PM10
EC
Organic
Carbon
Sulfate
Ambient
Concentration
Daily Median:
10.3|jg/m3
Daily Median:
0.98 pg/m3
Daily Median:
2.55 pg/m3
Daily Median:
29,300 particles/cm3
Daily Median:
9.8 pg/m
Daily Median:
0.94 pg/m3
Daily Median:
16.2|jg/m3
Daily Median:
0.6 pg/m3
Daily Median:
4.0 pg/m
Daily Median:
11.6|jg/m3
Median Gothenburg
2h:18.95|jg/m3
24 h: 19.92 pg/m3
Stockholm
2 h: 14.62 pg/m3
24 h: 15.23 pg/m3
Median
Stockholm pg/m3
2 h: 9.17
24 h: 9.49 pg/m3
Daily Mean:
8.2 pg/m3
Daily Mean:
13. 3 pg/m3
Daily Mean:
0.8 pg/m3
Daily Mean:
4.5 pg/m3
Daily Mean:
1.3 pg/m3
Lag and its
Increment
Units
2 day
6.9 pg/m3
2 day
0.74 pg/m
2 day
2.04 pg/m3
2 day
19,120
particles/cm3
24-h ma
7.8 pg/m3
24-h ma
0.83 pg/m3
24-h ma
9.7 pg/m3
24-h ma
0.5 pg/m3
24-h ma
2.3 pg/m3
Lag Day 0
5.6 pg/m3
2-hma:14.16
pg/m3
24-h ma: 11:49
pg/m3
2-h ma: 6.69
pg/m3
24-h ma: 5.27
pg/m3
Lag Day 0
5.2 pg/m3
Lag Day 0
7.4 pg/m3
Lag Day 0
0.4 pg/m3
Lag Day 0
2.2 pg/m3
Lag Day 0
0.9 pg/m
OR
1.08
1.11
1.05
1.14
1.19
0.93
0.95
1.18
1.08
1.00*
2h:
1.31
24 h:
1.24
2h:
1.23
24 h:
1.28
LOT
0.9t
1.1T
1.1T
0.9t
95%
Confidence
Interval
0.96,
0.95,
0.92,
0.87,
1.02,
0.74,
0.72,
0.93,
0.81,
0.82,
1.00,
0.87,
0.84,
0.90,
0.9,1
0.5,1
0.9,1
0.9, 1
0.7, 1
1.22
1.28
1.20
1.50
1.38
1.18
1.27
1.50
1.43
1.19*
1.72
1.76
1.80
1.84
•1T
•5T
•3T
•3T
•2T
December 2009
6-15
-------
"-— r?Ł TagES
Generalized estimating
Metzgeretal. N=6287 confirmed equations
(2007, 092856) ventricular
arrhythmias among Lags Evaluated:
Atlanta, GA n=51 8 subjects
0, 1, and 2 day ma
<*— • A
PM25
PM10
PM,0-2.5
NO* «>. SO* PM25EC
PM25 OC
PM25
S042"
PM25 water
soluble
elements
Ambient
Concentration
Daily Median:
16.2|jg/m3
Daily Median:
26.4 pg/m3
Daily Median:
8.7 pg/m
Daily Median:
1.4|jg/m3
Daily Median:
3.9 pg/m3
Daily Median:
4.1 pg/m3
Daily Median:
0.022 pg/m3
Units
24-h ma
10|jg/m3
24-h ma
10|jg/m3
24-h ma
5 pg/m3
24-h ma
1 pg/m3
24-h ma
2 pg/m3
24-h ma
5 pg/m3
24-h ma
0.03 pg/m3
OR
1.00
1.00
1.03
1.01
1.01
0.99
0.95
95%
Confidence
Interval
0.95, 1.0
0.97, 1.03
1.00, 1.07
0.98, 105
0.98, 1.03
0.93, 1.06
0.90, 1.00
Estimated from Figure 3 Vedal et al. (2004, O556301.t Estimated from Figure 3 Rich et al. (2004, 05563H
Summary of Epidemiologic Studies of Arrhythmias using ICDs
Since 2004, only two studies (in Boston and Sweden), reported adverse associations of PM2.s,
other size fractions and components with ICD-detected ventricular arrhythmias (Dockery et al.,
2005, 078995: Dockery et al., 2005, 090743; Ljungman et al., 2008, 180266; Rich et al., 2005,
079620). Studies of ICD-detected ventricular arrhythmias conducted elsewhere did not report
associations (Dusek et al., 2006, 155756; Metzger et al., 2007, 092856; Rich et al., 2004, 055631;
Vedal et al., 2004, 055630) nor was an association observed in a study of PMi0 and ICD shock in
Vancouver, Canada (Vedal et al., 2004, 055630). A range in exposure lags was evaluated in the
Boston study (3 h-3 days) (Dockery et al., 2005, 078995; Dockery et al., 2005, 090743; Rich et al.,
2005, 079620) and Sweden study (2 h and 24 h) (Ljungman et al., 2008, 180266). Reasons for the
inconsistent findings may include differing degrees of exposure misclassification within each study
or city due to differences in PM composition and pollutant mixes (e.g., less transition metals and
sulfates in the Pacific Northwest than the Northeast U.S.), and differences in the size of study areas
(Boston: within 40 km of PM2.5 monitoring site; Vancouver: Lower Mainland of British Columbia 90
km east of Vancouver). In addition, Rich et al. (2005, 079620) reported that use of the mean
pollutant concentration from the specific 24 h before the arrhythmia rather than just the day of the
arrhythmia, resulted in less exposure misclassification and less bias towards the null, possibly
explaining the lack of association when using just the day of ICD discharge and daily PM
concentrations.
Ectopy Studies Using ECG Measurements
A few panel studies have used ECG recordings to evaluate associations between ectopic beats
(ventricular or supraventricular) and mean PM concentrations in the previous hours and/or days
(Berger et al., 2006, 098702; Ebelt et al., 2005, 056907; Liao et al., 2009, 199519; Sarnat et al.,
2006, 090489).
Ectopic beats are defined as heart beats that originate at a location in the heart outside of the
sinus node. They are the most common disturbance in heart rhythm. Ectopic beats are usually
benign, and may present with or without symptoms, such as palpitations or dizziness. Such beats can
arise in the atria, AV node, conduction system or ventricles. When the origin is in the atria the beat is
called an atrial or supraventricular ectopic beat. When such a beat occurs earlier than expected it is
referred to as a premature supraventricular or atrial premature beat. Likewise, when the origin is in
December 2009
6-16
-------
the ventricle the beat is defined as a ventricular ectopic beat, or when early a premature ventricular
beat. When three or more occur ectopic beats occur in succession, this is called a non-sustained run
of either supraventricular (atrial) or ventricular origin. When the rate of the run is greater than 100
beats per minute it is defined as a tachycardia. Sustained VT are the arrhythmias investigated in the
ICD studies described above.
Using data from the WHI done in 59 U.S. exam sites in 24 cities, Liao et al. (2009, 199519)
estimated mean PM2.5 and PM^ concentrations at the addresses of 57,422 study subjects undergoing
ECG monitoring. They then estimated the risks of ventricular and supraventricular ectopy during that
10-s ECG recording associated with increases in mean PMi0 and PM2.5 concentrations on the same
day and previous 2 days, as well as over the previous 30 days. Mean PM2.5 and PMi0 concentrations
during the study period were 13.8 and 27.5 (ig/m3, respectively. Using a 2-stage random effects
model, they reported that among smoking subjects, each 10 (ig/m3 increase in PM2.5 concentration on
lag day 1 was associated with a significantly increased risk of ventricular ectopy (OR = 2.0 [95% CI:
1.32-3.3]). Similarly, each 10 (ig/m3 increase in lag 1 PMi0 concentration was associated with an
increased risk of ventricular ectopy (OR = 1.32 [95% CI: 1.07-1.65]). The lag day 2 PM2.5 risk
estimate was similar in size, but not statistically significant. There were no associations between
PMio, PM2 5 and supraventricular ectopy among smokers or non-smokers, and no association with
any PM metric and ventricular ectopy among non-smokers.
Sarnat et al. (2006, 090489) conducted a panel study among 32 nonsmoking older adults
residing in Steubenville, OH. In this study, the median daily PM2 5, SO42~, and EC concentrations
were 17.7, 5.7, and 1.0 (ig/m3, respectively They used logistic regression models to examine lagged
effects of 1- to 10-day ma concentrations of PM25, SO42~, EC, O3, NO2, and SO2. Supraventricular
ectopy and ventricular ectopy were measured using Holter monitors during a 30-minute protocol of
alternating rest in the supine position, standing, walking and paced breathing. In single-pollutant
models, each 10.0 (ig/m increase in 5-day mean PM25 concentration was associated with increased
risk of supraventricular ectopy (OR = 1.42 [95% CI: 0.99-2.04]), but not ventricular ectopy
(OR = 1.02 [95% CI: 0.63-1.65]). Similarly, increased risk of supraventricular ectopy, but not
ventricular ectopy, was associated with each interquartile range increase in 5-day mean SO42~ and O3
concentration.
Ebelt et al. (2005, 056907) conducted a repeated measures panel study of 16 patients with
COPD in Vancouver, British Columbia. Their goal was to evaluate the relative impact of ambient
and non-ambient exposures to PM25, PMi0, and PMi0_2.5 on several health measures. Subjects wore
an ambulatory ECG monitor for 24 h to record heart rhythm data and ascertain supraventricular
ectopic beats. The mean PM25 concentration during this study was 11.4 ug/m3. Using mixed models
with random subjects effects to investigate only same-day PM concentrations, an increase in
supraventricular ectopic beats was associated with same day ambient exposures to each PM size
fraction.
Berger and colleagues (2006, 098702) conducted a panel study of 57 men with coronary heart
disease living in Erfurt, Germany. Using 24-h ECG measurements made once every 4 wk, they
studied associations between runs of supraventricular and ventricular tachycardia and lagged
concentrations of PM2.5, UFP (0.01-0.1 (im), ACP (0.1-1.0 (im), SO2, NO2, CO, and NO. Using
GAMs, as well as Poisson and linear regression models, they reported increases in supraventricular
tachycardia and the number of runs of ventricular tachycardia associated with 5-day mean PM25,
UFP counts, and ACP counts. They found these associations at all lags evaluated (during ECG
recording, 0-23 h before, 24-47 h before, 48-71 h before, 72-95 h before, and 5-day mean), but the
largest effect estimates were generally associated with the 24- to 47-h mean and the 5-day mean.
Summary of Ectopy Studies Using ECG Measurements
Four studies of ectopic beats and runs of supraventricular and ventricular tachycardia, captured
using ECG measurements, all report at least one positive association. Further, they report findings in
regions other than Boston and Sweden (i.e., Midwest U.S., Pacific Northwest, 24 U.S. cities, and
Erfurt, Germany). A range of lags and/or moving averages were investigated (0-30 days) with the
strongest effects observed for either the 5-day mean, same day, or 1-day lagged PM concentrations.
Taken together, these ICD studies and ectopy studies provide evidence of an arrhythmic response to
PM, although further study is needed to understand the variable ICD study findings.
December 2009 6-17
-------
ECG Abnormalities Associated with the Modulation of Repolarization
No reported investigations of the relationship of PM concentration and ECG abnormalities
indicating arrhythmia were conducted prior to 2002 and thus were not included in the 2004 PM
AQCD (U.S. EPA, 2004, 056905). Abnormalities in the myocardial substrate, myocardial
vulnerability, and resulting repolarization abnormalities are believed to be key factors contributing to
the development of arrhythmogenic conditions such as those discussed above. These abnormalities
include ECG measures of repolarization such as QT duration (time for depolarization and
repolarization of the ventricles), T-wave complexity (a measure of repolarization morphology), and
T-wave amplitude (height of the T-wave). Abnormalities in repolarization may also identify subjects
potentially at risk of more serious events such as sudden cardiac death (Atiga et al., 1998, 156231;
Berger et al., 1997, 155688; Chevalier et al., 2003, 156338; Okin et al., 2000, 156002; Zabel et al.,
1998, 156176). Recent studies of changes in these measures following acute increases in air
pollution are described below.
Two studies conducted in Erfurt, Germany, (Henneberger et al., 2005, 087960; Yue et al.,
2007, 097968) examined the association between measures of repolarization (QT duration, T-wave
complexity, T-wave amplitude, T-wave amplitude variability) and particulate air pollution.
Henneberger et al. (2005, 087960) conducted a panel study of 56 males with IHD. Each subject was
measured every 2 wk for 6 mo. During the study, the median daily PM2.5, EC, and OC concentrations
were 14.9, 1.8, and 1.4 ug/m3,respectively. The median count of UFP was 11,444 particles/cm3,
while the median count of ACP (0.1-1.0 (im) was 1,238 particles/cm3. They examined the change in
these ECG parameters associated with the mean pollutant (UFP, ACP, PM2.5, OC, and EC)
concentrations 0-5, 6-11, 12-17, 18-23, and 0-23 h before, and 2-5 days before the ECG
measurement. Significant decreases in T-wave amplitude were associated with PM2.5 mass, UFP, and
ACP. Each 16.4 (ig/m3 increase in the mean PM2.5 concentration in the previous 5 h was associated
with a 6.46 (iV decrease in T-wave amplitude (95% CI: -10.88 to -2.04). Each 0.7 (ig/m3 increase in
the mean OC concentration in the previous 5 h was associated with a 4.15 ms increase in QT interval
(95% CI: 0.22-8.09). There was a similar sized effect for 24-h mean OC concentration. Significant
increases in the variability of T-wave complexity were also associated with acute increases in EC and
OC concentration.
Yue et al. (2007, 097968) then used positive matrix factorization to identify 5 sources of
ambient PM (airborne soil, local traffic-related UFP, combustion-generated aerosols, diesel traffic-
related particles, and secondary aerosols). Using similar statistical models, they examined the
association between these same repolarization changes and incremental increases in the mean
concentration of each particle source in the 24 h before the ECG measurement. They also examined
associations with CRP and vWF concentrations in the blood. Both UFP from local traffic and diesel
particles from traffic had the strongest associations with repolarization parameters.
Summary of Epidemiologic Studies of ECG Abnormalities Associated with the
Modulation of Repolarization
These two analyses demonstrate associations between PM pollution and repolarization
changes, at lags of 5 h to 2 days. Moreover, the findings from the Yue et al. (2007, 097968) study
demonstrate a potential role of traffic particles/pollution.
6.2.2.2. lexicological Studies
The ECG of animal research models frequently exhibit different characteristics than that of
humans. Mice and rats are notable in this regard, as they do not have an isoelectric ST-segment
typical of larger species, likely owing to their rapid heart rates (-600 and -350 bpm, respectively)
and repolarizing currents. However, the ultimate function of the pumping heart is conserved and
reflected by the ECG in a remarkably consistent manner across species. Thus, atrial depolarization
causes an electrical inflection represented by the P-wave, ventricular depolarization elicits the QRS
complex, and the T-wave represents repolarization of the ventricles.
December 2009 6-18
-------
The earliest indication that there may be cardiovascular system effects of PM came from ECG
studies in susceptible animal models (rats with pulmonary hypertension and dogs with coronary
occlusion), which were summarized in the 2004 PM AQCD (U.S. EPA, 2004, 056905). However, a
study of dogs exposed to ROFA did not demonstrate ECG changes, perhaps due to differences in
disease state, as these were the oldest dogs in the colony with signs of preexisting, naturally
occurring heart disease (Muggenburg et al., 2000, 189163). Much of the research conducted since
the release of the last PM AQCD has been focused on exploring susceptibility or varying exposure
methodologies, with little new evidence into the mechanisms for ECG changes of inhaled PM.
CAPS
Wellenius et al. (2004, 087874) used a susceptible model that was previously shown to
produce significant results with exposures to ROFA (Wellenius et al., 2002, 025405) to examine
ECG-related PM2.5 effects. Using an anesthetized model of post-infarction myocardium sensitivity,
Wellenius and colleagues tested the effects of Boston, MA CAPs on the induction of spontaneous
arrhythmias in SD rats (1 h; mean mass concentration 523.11 ug/m3; range of mass concentration
60.3-2202 ug/m3). Decreased (67.1%) VPB frequency was observed during the post-exposure period
in rats with a high number of pre-exposure VPB. No interaction was observed with coexposure to
CO (35 ppm). CAPs number concentration or the mass concentration of any single element did not
predict VPB frequency. In a follow-up publication, a decreased number of supraventricular ectopic
beats (SVEB) was reported with CAPs (mean mass concentration 645.7 ug/m3) (Wellenius et al.,
2006, 156152). Furthermore, an increase in CAPs number concentration of 1,000 particles/cm3 was
associated with a 3.3% decrease in SVEB frequency. The findings of decreased ventricular
arrhythmia differ from those observed following ROFA exposure in the same animal model in that
an increased frequency of premature ventricular complexes was observed with ROFA, albeit the
ROFA exposure concentration was >3,000 ug/m3 (Wellenius et al., 2002, 025405). It is difficult to
directly compare the results of these studies due to differences in exposure concentrations and
particle type, but collectively they may suggest an important role for the soluble components of PM,
including transition metals, as only ROFA induced increases in ventricular arrhythmia occurrence.
In older rats (Fisher 344; -18 months) exposed to PM2.5 CAPs in Tuxedo, NY (4 h; mean
concentration 180 ug/m3; August 2000), the frequency of delayed beats was greater than in rats
exposed to air (Nadziejko et al., 2004, 055632). The majority of these beats were characterized as
pauses (a delay of 2.5 times the adjacent interbeat intervals) rather than premature beats. When the
same animals were exposed to generated UF carbon particles (single-day concentrations 500 and
1280 ug/m3) or SO2 (1.2 ppm), no significant differences were observed in arrhythmia frequency
between air controls and exposed animals. The authors also report using the same protocol for young
WKY rats (concentration 215 ug/m3) and very few arrhythmias were observed, thus precluding
statistical analysis. The results of this study indicate (1) involvement of the sino-atrial node, as the
observed arrhythmias were mostly of a delayed nature; and (2) particle size and PM25 constituents
may play a role in these effects.
Diesel and Gasoline Exhaust
Anselme and colleagues (2007, 097084) exposed rats with and without induced CHF to DE for
3 h (PM concentration 500 ug/m3; mass mobility diameter 85 nm; NO2 1.1 ppm; CO 4.3 ppm).
While no dramatic change was noted in HR, prominent increases in the incidence of VPB were
observed in CHF rats, which lasted at least 4-5 h after exposure ceased. The duration of VPB
attributable to diesel exposure in CHF rats lasted much longer than the rMSSD change (>5 h
post-exposure), indicating that the HRV response was not driving the increased arrhythmia
incidence. It is interesting to contrast the work of Anselme with the studies by Wellenius et al. (2002,
025405; 2004, 087874; 2006, 156152). as the arrhythmia incidence in the acute infarction model was
greatest with ROFA, while the CHF model demonstrated sensitivity to DE exposure. However,
several differences in the research designs preclude strong comparisons.
Using ApoE"7" mice on a high-fat diet as a model of pre-existing coronary insufficiency
(Caligiuri et al., 1999, 156318). Campen and colleagues studied the impact of inhaled diesel and
gasoline exhaust and road dust (6 h/day><3 day) on ECG morphology (Campen et al., 2005, 083977;
December 2009 6-19
-------
2006, 096879). Moreover, a high efficiency particle filter was used to compare the whole exhaust
with an atmosphere containing only the gaseous components. For gasoline exhaust, the
PM-containing atmosphere (PM mean concentration 61 ug/m3; PNMD 15 nm; NOX mean
concentration 18.8 ppm; CO mean concentration 80 ppm) induced T-wave morphological alterations,
while the PM-filtered atmosphere did not (Campen et al., 2006, 096879). Resuspended road dust
(PM25), at up to 3500 ug/m3 had no impact on ECG. For DE (PM mean concentration 512, 770, or
3,634 ug/m3; MMD 100 nm, CMD 80 nm; NOX mean concentration 19, 105, 102 ppm for low whole
exhaust, high PM filtered, and high whole exhaust, respectively), dramatic bradycardia, decreased
T-wave area, and arrhythmia (atrioventricular-node block and VPB) were only observed in mice
exposed to high filtered and high whole exhaust (Campen et al., 2005, 083977). These effects
remained after filtration of PM, suggesting that the gaseous components of the whole DE drove the
cardiovascular findings. The diesel- and gasoline-induced ECG changes contrast, in that the gasoline
exhaust required particles to induce T-wave changes, whereas the DE did not require PM to cause an
effect on ECG. However, the differing responses could be attributable to higher PM concentrations in
the whole DE.
Summary of lexicological Study Findings for ECG Abnormalities
The above toxicological studies demonstrate mixed results for arrhythmias, which may be
somewhat attributable to the different disease models used. Wellenius et al. (2004, 087874; 2006,
156152) showed decreased frequency of VPB and SVEB following PM2.5 CAPs exposure in rats
with induced MI (>12 h prior to exposure). One study reported increased frequency of premature
beats in older rats exposed to CAPs, which were not observed with UF carbon particles (Nadziejko
et al., 2004, 055632). Rats with a MI model of CHF (3-mo recovery) had increased incidence of
VPB with DE exposure (Anselme et al., 2007, 097084). As for ECG morphology changes, T-wave
alterations were reported for gasoline exhaust that were absent when the PM was filtered (Campen
et al., 2006, 096879). However, for DE, increased atrioventricular-node block, VPB, and decreased
T-wave area were observed with whole exhaust and remained after filtration of PM, indicating that
the gases were responsible for the effects (Campen et al., 2005, 083977).
6.2.3. Ischemia
Although no evidence from epidemiologic or controlled human exposure studies of PM-
induced myocardial ischemia was included in the 2004 PM AQCD (U.S. EPA, 2004, 056905). one
toxicological study was cited that observed ST-segment changes in dogs following a 3-day exposure
to CAPs. In epidemiologic studies published since the 2004 PM AQCD (U.S. EPA, 2004, 056905).
associations have been demonstrated between PM and ST-segment depression, and one new
controlled human exposure study reported significant increases in exercise-induced ST-segment
depression among men with prior MI following a controlled exposure to DE. Results from recent
toxicological studies confirm the findings presented in the 2004 PM AQCD (U.S. EPA, 2004,
056905) and provide coherence and biological plausibility for the effects observed in epidemiologic
and controlled human exposure studies.
6.2.3.1. Epidemiologic Studies
ECG Changes Suggestive of Increased Ischemia
The ST-segment duration is typically in the range of 0.08-0.12 s (80-120 ms). The direction of
the ST change is influenced by the extent of the acute myocardial injury. If the ischemia or infarction
is transmural, i.e., penetrates the entire thickness of the ventricular wall, it usually causes ST-
segment elevation, while ischemia confined primarily to the ventricular endocardium often causes
ST-segment depression. Clinical ischemia is typically defined to include a downsloping ST segment
depression of >0.1 mV (ECG voltages are calibrated such that 1 mV equals 10 mm in the vertical
direction). The studies described below evaluate a range of ECG changes suggestive of increased
December 2009 6-20
-------
ischemia including subclinical ST segment depressions (e.g. less than 0.1 mV or 1 mm) in relation to
ambient PM concentration.
In a large study of the WHI Trial, Zhang et al. (2009, 191970) examined the change and risk of
subclinical ST-segment abnormalities, T-wave abnormalities, and T-wave amplitude associated with
ambient PM25 concentrations on the same and previous 6 days. Using logistic regression, each 10
(ig/m3 increase in the mean PM2 5, on lag days 0-2, was associated with a 4% (95% CI: -3 to 10)
increase in the relative odds of a ST-segment abnormality, and a 5% (95% CI: 0-9) increase in the
relative odds of a T-wave abnormality.
Gold et al. (2005, 087558) studied 24 elderly residents of Boston, MA (aged 61-88 yr) residing
at or near an apartment complex that was -0.5 km from an air pollution monitoring station. A
protocol of continuous Holter monitoring including 5 min of rest, 5 min of standing, 5 min of
outdoor exercise, 5 min of rest, and then 20 cycles of paced breathing was done up to 12 times for
each subject (n = 269 ECG measurements for analysis). From these ECG measurements, they
identified occurrences of ST-segment depression and examined whether mean BC, CO, and PM2 5
concentrations in the previous 5 and 12 h were associated with ST-segment depression. The median
5-h and 12-h mean BC concentrations were 1.28 and 1.14 (ig/m3, respectively (PM2.5 concentrations
are in Table 6-3). The mean BC concentrations in the 5 and 12 h before testing predicted ST-segment
depression in most portions of the protocol. However, these effects were strongest in the
post-exercise periods. For example, during the post-exercise rest period, each 10th-90th percentile
increase (1.59 (ig/m3) in the mean 5-h BC concentration was associated with a -0.11 mm ST-segment
depression (95% CI: -0.18 to -0.05). In two pollutant models, CO did not appear to confound this
association. PM2 5 was not associated with ST-segment depression in this study. These findings
suggest traffic-generated particulate pollution may be associated with ST-segment depression.
Previously, Pekkanen et al. (2002, 035050) conducted a panel study of 45 subjects with stable
coronary heart disease living in Helsinki, Finland. Each subject had biweekly sub-maximal exercise
testing for 6 mo (n = 342 exercise tests with 72 exercise-induced ST-segment depressions). The
median daily count of ACP (ACP: 0.1-1.0 (im) was 1,200 particles/cm3 (PM25 concentrations are
found in Table 6-3). They examined the risk of ST-segment depression associated with mean
pollutant concentrations (UFP, ACP, PMb PM2.5, PMi0_2.5, NO2, CO) in the previous 24 h, and the 3
previous lagged 24-h periods. Each 7.9 (ig/m3 increase in mean PM25 concentration, lagged 2 days,
was associated with significantly increased risk of ST-segment depression >0.1 mV (OR: 2.84 [95%
CI: 1.42-5.66]). Each 760 particles/cm3 increase in the count of ACP, lagged 2 days, was also
associated with significantly increased risk of ST-segment depression >0.1 mV (OR: 3.29 [95%
CI: 1.57-6.92]). Similarly sized increased risks of ST-segment depression were also found for other
particulate pollutants, including PMi0_2.5, PMi, and UFP counts.
This same research group, then conducted a principal components analysis to identify five
PM25 sources (crustal, long range transport, oil combustion, salt, and local traffic) (Lanki et al.,
2006, 088412). Using similar statistical models, each 1 (ig/m3 increase in "local traffic" particle
concentration, lagged 2 days, was associated with increased risk of ST-segment depression (OR: 1.53
[95% CI: 1.19-1.97]). Similarly, each 1 (ig/m3 increase in "long-range transport" particle
concentration was also associated with increased risk of ST-segment depression (OR: 1.11 [95%
CI: 1.02-1.20]). No significant associations for other sources were reported for any lag time.
In Boston, Chuang et al. (2008, 155731) studied 48 patients with a prior percutaneous
intervention following MI, acute coronary syndrome (ACS) without MI, or stable coronary artery
disease without ACS,. Each patient had a 24-h ECG measurement up to four times during study
follow-up. Using logistic regression, they estimated the risk of ST-segment depression of >0.1 mm,
during 30-min segments, associated with increases in the mean PM2 5, BC, CO, NO2, O3, and SO2
concentration in the previous 24 h. Each 6.93 (ig/m3 increase in mean PM25 concentration was
associated with a significantly increased risk of ST-segment depression (OR = 1.50 [95% CI:
1.19-1.89]). Using linear additive models to estimate the change in ST level associated with the same
PM2.5 change, they observed a significant -0.031 mm change (95% CI: -0.042 to -0.019). In single
pollutant models, risk estimates were of similar magnitude and statistically significant for BC, NO2,
and SO2. In two pollutant models, however, PM25 risk estimates were reduced to 1.0 in all models
with BC, NO2, and SO2. In contrast, the risk estimates for BC, NO2, and SO2 remained elevated and
statistically significant when modeled with PM2 5.
In a panel study of 14 Helsinki resident, non-smoking, elderly subjects with coronary artery
disease, Lanki et al. (2008, 191984) used logistic regression to report that each 10 (ig/m3 increase in
personal PM2 5 concentration in the previous hour was associated with a significantly increased risk
December 2009 6-21
-------
of ST-segment depression (OR = 3.26 [95% CI: 1.07-9.98]). In addition, each 10 ug/m3 increase in
outdoor mean PM2 5 concentration in the previous 4 h was also associated with an increased risk (OR
= 2.47 [95% CI: 1.05-5.85]). Last, the risk estimates for all time lags examined (1, 4, 8, 12, and 22 or
24 h) for all PM size fractions were increased, but none other than those described above were
statistically significant.
Summary of Epidemiologic Study Findings for Ischemia
These studies demonstrate associations between PM2.5 pollution and ST-segment depression at
lags of 1 h-2 days. Moreover, these findings demonstrate a potential role for traffic (Chuang et al.,
2008, 155731: Gold et al., 2005, 087558) and long-range transported PM2.5 (Lanki et al., 2006,
089788). Mean and upper percentile concentrations reported in these studies are found in Table 6-3.
Table 6-3. PM Concentrations reported in epidemiologic studies ECG changes suggestive of
ischemia.
Author
Location
Mean Concentration (ug/m3)
Upper Percentile
Concentrations (ug/m3
PM2.5
Zhang etal. (2009,191970)
PMiO-2.5
Mulitcity, US:WHI Clinical Trial NR
NR
Pekkanen et al. (2002, 035050) Helsinki, Finland
Gold et al. (2005. 087558) Boston, MA
Chuang et al. (2008, 155731) Boston, MA
Lanki et al. (2008, 191984) Helsinki, Finland
24-havg: 10.6 (median)
5-havg:9.5 (median)
12-havg:9.8 (median)
12-havg:9.91 (median)
24-havg: 9.20 (median)
Personal
1-havg:11.5(median)
4-havg: 10.1 (median)
22-havg:9.3 (median)
Outdoor
24-havg: 12.5
75th: 16.0
Max: 39.8
5-havg 90th: 25.6 Max: 41.0
12-havg 90th: 25.9 Max: 35.6
12-havg75th:13.18
24-havg max: 40. 38
Personal
1-havg 75th: 17.2; Max: 746.3
4-havg 75th: 15.7; Max 189.6
22-havg 75th: 13.2; Max 52.9
Outdoor
24-havg 75th: 17.7; Max: 30.5
Pekkanen et al. (2002, 035050)
Helsinki, Finland
24-havg:4.8 (median)
75th:8.5
Max: 37.0
6.2.3.2. Controlled Human Exposure Studies
Diesel Exhaust
Among a group of 20 men with prior MI, Mills et al. (2007, 091206) found that DE
(300 ug/m3 particle concentration, median particle diameter 54 nm) significantly increased
exercise-induced ischemic burden during exposure, calculated as the product of exercise duration
and change in ST-segment amplitude. The mechanism by which DE induced the exacerbation of
ischemic burden remains unclear, and appears to be unrelated to impaired vasodilation. However, the
December 2009
6-22
-------
authors suggest that this discrepancy may be due to the timing of the vascular assessment, as
measures of blood-flow were taken 5 h after the observed increase in ischemic burden. Although it is
reasonable to assume that the observed increase in ST-segment depression during exercise represents
an increased magnitude of ischemia, it is important to note that there are other potential explanations
for the ST change. For example, it is possible that the ST-segment depression could be secondary to
heterogeneity of electrophysiological responses of particle exposure on the myocardium that is
enhanced by the metabolic and ionic conditions associated with ischemia or increased HR. It is also
important to note that the effects observed in this study cannot be conclusively attributed to the
particles per se, as subjects were also exposed relatively high levels of NO (3.45 ppm), NO2
(1.01 ppm), CO (2.9 ppm), and total hydrocarbons (2.8 ppm).
6.2.3.3. lexicological Studies
CAPs
A study that examined ECG changes in dogs (female; retired mongrel breeder dogs) following
PM2.5 CAPs exposure in Boston, MA (mean mass concentration 345 ug/m3; 9/2000-3/2001) and left
anterior descending coronary artery occlusion as an indicator of myocardial ischemia reported
changes in ST-segment (Wellenius et al., 2003, 055691). The experimental protocol was a 6-h
exposure to CAPs via tracheostomy, followed by a preconditioning occlusion (5 min), rest interval
(20 min), and the experimental occlusion (5 min). Increased ST-segment elevation was observed
following PM2.5 during the experimental occlusion period compared to filtered air. Furthermore,
peak ST-segment elevation attributable to CAPs was reported with the experimental occlusion,
which remained elevated 24 h post-exposure. Ventricular arrhythmias were rarely observed during
occlusion and when observed, were unrelated to CAPs exposure. The results from this study support
those previously observed (Godleski et al., 2000, 000738) and provides greater support that
enhanced myocardial ischemia occurs relatively quickly (within hours) following PM exposure.
The Wellenius et al. (2003, 055691) study also attempted to link ST-segment changes with
four CAPs elements (Si, Ni, S, and BC) as tracers of PM2.5 sources in Boston. In the multivariate
regression analyses, peak ST-segment elevation and integrated ST-segment change were significantly
associated with only the mass concentration of Si (Si mean concentration 8.17 ug/m3; Si
concentration 2.31-13.93 ug/m3). In the univariate regression analyses, Pb also demonstrated a
significant association for both ST-segment measures, although the p-value was greater than that
observed with Si.
A recent study in dogs (female mixed-breed canines) evaluated myocardial blood flow during
myocardial ischemia following 5-h PM2 5 Boston CAPs exposures (daily mean mass concentration
94.1-1556.8 ug/m3; particle number concentration 3-69.3><103 particles/cm3; BC concentration
1.3-32.0 ug/m ) (Bartoli et al., 2009, 179904). Similar methods were used for the coronary
occlusion and exposure method as Wellenius et al. (2003, 055691). Immediately following exposure,
microspheres were injected (15 urn diameter) into the left atrium after 3 min of ischemia during the
second occlusion. Post-mortem analysis of cardiac tissue and blood samples allowed for
quantification of microspheres. CAPs-exposed dogs had decreased total myocardial blood flow and
increased coronary vascular resistance during occlusion that was greatest in tissue within or near the
ischemic zone. The rate-pressure product (product of HR and SBP) during occlusion was unchanged
in animals exposed to CAPs, indicating that cardiac metabolic demand was not altered. The
multilevel linear mixed models demonstrated that myocardial blood flow and coronary vascular
resistance during occlusion were inversely and significantly associated with CAPs mass
concentration, particle number concentration, and BC concentration, with the strongest effects
observed with particle number concentration. The results of this study provide evidence that
exacerbation of myocardial ischemia following PM exposure is due to reduced myocardial blood
flow, perhaps via dysfunctional collateral vessels.
December 2009 6-23
-------
Intratracheal Instillation
Cozzi et al. (2006, 091380) exposed ICR mice to UF PM (100 (ig IT instillation), followed by
ischemia/reperfusion injury to the left anterior coronary artery 24 h later. The area-at-risk (the region
of tissue perfused by the left anterior descending coronary artery) and the infarct size were measured
2 h following reperfusion, and while the area-at-risk was not affected by PM exposure, the infarct
size was nearly doubled in mice who received UF PM. Increases in infarct size were associated with
increased myocardial neutrophil density in the infarct zone and lipid peroxidation in the
myocardium.
Summary of lexicological Study Findings for Ischemia
The studies described above provide evidence that PM can induce greater myocardial
responses following ischemic events, as demonstrated by, enhanced ischemia, decreased myocardial
blood flow and increased coronary vascular resistance, and increased infarct size.
6.2.4. Vasomotor Function
The most noteworthy new cardiovascular-related revelation in the past six years with regards
to PM exposure is that the systemic vasculature may be a target organ. The vasculature of all tissues
is lined with endothelial cells that will naturally encounter any systemically absorbed toxin. The
endothelium (1) maintains barrier integrity to ensure fluid compartmentalization; (2) communicates
dilatory and constrictive stimuli to vascular smooth muscle cells; and (3) recruits inflammatory cells
to injured regions. Smooth muscle cells lie within the layer of endothelium and are crucial to the
regulation of blood flow and pressure. In states of injury and disease, both cell types can exhibit
dysfunction and even pathological responses.
Endothelial dysfunction is a factor in many diseases and may contribute to the origin and/or
exacerbation of perfusion-limited diseases, such as MI or IHD, as well as hypertension. Endothelial
dysfunction is also a characteristic feature of early and advanced atherosclerosis. A primary outcome
of endothelial dysfunction is impaired vasodilatation, frequently due to uncoupling of NOS. It is this
uncoupling that appears central to impaired vasodilation and thus endothelial dysfunction.
One controlled human exposure study cited in the 2004 PM AQCD (U.S. EPA, 2004, 056905)
reported a decrease in bronchial artery diameter (BAD) among healthy adults following exposure to
CAPs in combination with O3. Conclusions based on this finding were limited due to the
concomitant exposure to O3 as well as a lack of published results from epidemiologic and
toxicological studies. Recent controlled human exposure studies have provided support to the
findings described in the 2004 PM AQCD (U.S. EPA, 2004, 056905). with changes in vasomotor
function observed following controlled exposures to DE and EC particles. In addition, epidemiologic
studies have observed associations between PM and decreases in BAD and flow mediated dilatation
(FMD) in healthy adults and diabetics. These findings are further supported by a large body of new
toxicological evidence of impaired vasodilation following exposure to PM.
6.2.4.1. Epidemiologic Studies
O'Neill et al. (2005, 088423) examined the association between 2 measures of vascular
reactivity, non-endothelium dependent nitroglycerin mediated reactivity and endothelium-dependent
flow-mediated reactivity, and ambient mean particulate pollutant concentration (PM2.5, SO42~, BC,
PNC) on the same and previous few days. They studied a panel of 270 subjects with diabetes or at
risk for diabetes, who lived in the greater Boston metropolitan area. Using linear regression models,
the change in vascular reactivity associated with moving average pollutant concentrations across the
same and previous 5 days was estimated. Interquartile range (values not reported) increases in mean
PM2.5 concentration, BC concentration, and PNC over the previous 6 days were associated with
decreased vascular reactivity among diabetics, but not among subjects at risk for diabetes. For SO42~,
the mean concentration on lag day 0, lag day 1, and the 3-day, 4-day, and 5-day ma all were
associated with similarly sized reductions in both metrics of vascular reactivity. Among diabetics,
each interquartile range increase in the mean SO42~ concentration over the previous 6 days was
December 2009 6-24
-------
associated with a 5.4% decrease in nitroglycerin-mediated reactivity (95% CI: -10.5 to -0.1) and
flow-mediated reactivity (-10.7% [95% CI: -17.3 to -3.5]). Also among diabetics, each interquartile
range increase in the mean PM2.5 concentration over the previous 6 days was associated with a 7.6%
decrease in nitroglycerin-mediated reactivity (95% CI: -12.8 to -2.1) and anon-significant 7.6%
decrease in flow-mediated reactivity (95% CI: -14.9 to 0.4). Each interquartile range increase in the
mean BC concentration over the previous 6 days was associated with a 12.6% decrease in flow
mediated reactivity (95% CI: -21.7 to -2.4), but not nitroglycerin-mediated reactivity. PNC was
associated with non-significant decreases in both measures. Effect estimates were larger for type 2
diabetics than type 1 diabetics.
Dales et al. (2007, 155743) conducted a panel study of 39 healthy volunteers who sat at 1 of 2
bus stops in Ottawa, Canada for 2 h. FMD of the brachial artery was measured immediately after the
bus stop exposure, but not before. They examined the association between FMD and 2-h mean
PM2.5, PMi, NO2, and traffic density at the bus stop (vehicles/h). The authors report that each
30 ug/m3 increase in 2-h mean PM2.5 concentration was associated with a significant 0.48%
reduction in FMD. This represented a 5% relative change in the maximum ability to dilate.
This same research group conducted a panel study of 25 type 1 or 2 diabetic subjects living in
Windsor, Ontario (aged 18-65 yr) (Liu et al., 2007, 156705). For each subject, personal PMi0
concentrations were measured for 24 h before measurements of BAD, FMD, and other biomarkers.
Each 10 ug/m3 increase in personal 24-h mean PM10 concentration was associated with a 0.20%
increase in end-diastolic FMD (95% CI: 0.04-0.36) and a 0.38% increase in end-systolic FMD (95%
CI: 0.03-0.73), but decreases in end-diastolic basal diameter (-2.52 urn [95% CI: -8.93 to 3.89]) and
end-systolic basal diameter (-9.02 urn [95% CI: -16.04 to -2.00]).
Rundell et al. (2007, 156060) examined the change in FMD associated with high and low PMi
(0.02-1.0 urn) pollution in a panel of 16 young intercollegiate athletes (mean age = 20.5±2.4 yr) in
Scranton, PA, who were non-smokers, non-asthmatics, and free of cardiovascular disease (Rundell
et al., 2007, 156060). Each subject had FMD of the brachial artery measured 10-20 min before and
20-30 min after each of two 30-min exercise tests (85-90% of maximal HR). The exercise tests were
done outside either on an inner campus location free of automobile and truck traffic (low PMi;
mean = 5,309±1,942 particles/cm3) or on a soccer field adjacent to a major highway (high PMi;
mean = 143,501±58,565 particles/cm3). The order of the exercise test locations was chosen
randomly. Using paired t-tests for analysis, they reported FMD was impaired after high PM:
exposure (pre-exercise: 6.8±3.58%; post-exercise: 0.30±2.74%), but not low PMi exposure
(pre-exercise: 6.6±4.04%; post-exercise: 4.89±4.42%). Further, they found basal brachial artery
vasoconstriction (4%; pre-exercise BAD: 4.66±0.61 mm; post-exercise BAD: 4.47±0.63 mm) after
the 'high PM/ exposure, but not the 'low PMi' exposure (-0.3% pre-exercise BAD: 4.66±0.63 mm;
post-exercise BAD: 4.68±0.61 mm).
In a prospective panel study of 22 type 2 diabetics (aged 61 ± 8 yr), Schneider et al. (2008,
191985) examined the change in FMD, BAD, small artery elasticity index, larger artery elasticity
index, and systemic vascular resistance associated with ambient PM2 5 as measured in Chapel Hill,
NC (November 2004-December 2005). Using additive mixed models with a random subject effect,
each 10 ug/m3 increase in PM25 in the previous 24 h was associated with a decrease in FMD (-
17.3% [95% CI: -34.6 to 0.0]). Similarly, each 10 ug/m3 increases in PM25 was associated with a
decrease in small artery elasticity index lagged 1 day (-15.1% [95% CI: -29.3 to -0.9]), and lagged 3
days (-25.4% [95% CI: -45.4 to -5.3]). Significant decreases in larger artery elasticity index and
increases in systemic vascular resistance lagged 2 and 4 days were also reported. Further, effects
were greatest among those with high BMI, high glycosylated hemoglobin Ale, low adiponectin, or
the null GSTM1 polymorphism. However, high myeloperoxidase (MPO) levels were associated with
greater PM2 5 effects on these measures.
In a similar study done in Paris, France, Briet (2007, 093049) similarly reported that each
increase in PM25 was associated with a -0.32% decrease in FMD (95% CI: -1.10 to 0.46). Significant
FMD reductions were associated with increased SO2, NO2, and CO concentrations. Each 1 standard
deviation increase (units not given) in PM25 in the previous 2 wk was associated with a 15.68%
(95% CI: 7.11-23.30) increase in small artery reactive hyperemia. Each 1 standard deviation increase
(units not given) in PMi0 in the previous 2 wk was associated with a 15.91% (95% CI: 7.74-24.0)
increase in small artery reactive hyperemia.
December 2009 6-25
-------
Summary of Epidemiologic Study Findings for Vasomotor Function
Vasomotor function has been evaluated using several metrics in the studies described above,
including FMD, small artery elasticity index, larger artery index, systemic vascular resistance, BAD,
end diastolic basal diameter, and nitroglycerin-mediated reactivity. The most common measures
evaluated were BAD, a measure of the relatively static, anatomic/physiological baseline vasomotor
function, and FMD, the dynamic measure of post- minus pre-occlusion BAD. Each study
demonstrated an acute association between these measures of vascular function and ambient PM25
concentrations (Briet et al., 2007, 093049: Dales et al., 2007, 155743: Liu et al., 2007, 156705:
O'Neill et al., 2005, 088423: Rundell et al., 2007, 156060: Schneider et al., 2008, 191985). An
association with PM10 was observed in a study conducted in Windsor Ontario (Liu et al., 2007,
156705). Three studies evaluated effects on diabetics (Liu et al., 2007, 156705: O'Neill et al., 2005,
088423: Schneider et al., 2008, 191985). and three evaluated PM-related changes in vasomotor
function on young healthy subjects (Briet et al., 2007, 093049: Dales et al., 2007, 155743: Rundell
et al., 2007, 156060). Only two studies investigated multiple lags (lag days 0 to 6) (O'Neill et al.,
2005, 088423: Schneider et al., 2008, 191985). with one reporting the strongest association with the
6-day mean PM concentration (O'Neill et al., 2005, 088423). and the other with lag day 0. In other
studies, responses were observed in as short as 30 min after the exposure (Rundell et al., 2007,
156060). The Rundell et al. (2007, 156060) findings are consistent with other studies showing an
adverse response to ambient particulate pollution emitted from vehicular traffic (Adar et al., 2007,
098635: Adar et al., 2007, 001458: Riediker et al., 2004, 056992: Riediker et al., 2004, 091261).
Mean and upper percentile concentrations reported in these studies are found in Table 6-4.
Table 6-4. PM concentrations reported in epidemiologic studies of vasomotor function.
Author
Location
Mean Concentration (ug/m3)
Upper Percentile
Concentrations (ug/m )
PM2.5
Briet (2007, 093049)
Dales (2007, 155743)
O'Neill (2005, 088423)
Schneider (2008. 191 985)
Paris, France
Ottawa, Canada (bus stops)
Boston, MA
Chapel Hill, NC
NR
Bus stop 1 : 40
Bus stop 2: 10
11.5
13.6
NR
NR
Range: 1.1 -20.0
NR
PM10
Briet (2007, 093049)
Liu (2007. 156705)
Paris, France
Windsor, Ontario
NR
24h (personal): 25.5
NR
5th to 95th: 9.8 -133
6.2.4.2. Controlled Human Exposure Studies
Some evidence of a PM-induced increase in brachial artery vasoconstriction is presented in the
2004 PM AQCD (U.S. EPA, 2004, 056905). Brook et al. (2002, 024987) exposed 24 healthy adults
to PM2.5 CAPs (150 ug/m3) along with 120 ppb O3 for a period of 2 h. A significant decrease in BAD
was observed immediately following exposure compared with filtered air control. No significant
changes were observed in either endothelial-dependent or endothelial-independent vasomotor
function, as determined by FMD and nitroglycerin-mediated dilatation, respectively. As described
below, many more recent studies have evaluated the effects of various types of particles on
vasomotor function following controlled exposures among healthy and health-compromised
individuals.
December 2009
6-26
-------
CAPS
A subsequent analysis of the CAPs constituents from the Brook et al. (2002, 024987) study
revealed a significant negative association between the post-exposure change in BAD and both the
OC and EC concentrations of CAPs (Urch et al., 2004, 055629). However, the observed vasomotor
effects cannot conclusively be attributed to PM2.5, as subjects were exposed concurrently to PM2.5
and O3. Mills et al. (2008, 156766) evaluated the effect of fine and UF CAPs on vasomotor function
in a group of 12 males with stable coronary heart disease (average age 59 yr), as well as in 12
healthy males (average age 54 yr). Relative to filtered air exposure, exposure to PM (average
concentration 190 ug/m3) did not significantly affect vascular function in either group. The authors
attributed the lack of response in endothelial function to the composition of the CAPs used in the
study, which were low in combustion-derived particles and consisted largely of sea salt.
Urban Traffic Particles
The effect of exposure to urban traffic particles on vasomotor function has recently been
evaluated among a group of adult volunteers (Brauner et al., 2008, 191966). In this study, healthy
young adults (average age 27 yr) exposed for 24 h to urban traffic particles (average PM2.5
concentration 10.5 ug/m ) were not observed to experience any change in microvascular function
after 6 or 24 h of exposure relative to filtered air.
Diesel Exhaust
Mills et al. (2005, 095757) exposed 30 healthy men (20-38 yr) to both diluted DE (300 ug/m3)
and filtered air control for 1 h with intermittent exercise. Half of the subjects underwent vascular
assessments at 6-8 h following exposure to DE or filtered air, while in the other 15 subjects, vascular
assessments were performed at 2-4 h post-exposure. DE attenuated forearm blood flow increase
induced by bradykinin, acetylcholine (ACh), and sodium nitroprusside (SNP) infusion measured 2
and 6 h after exposure. The authors postulated that the effect of DE on vasomotor function may be
the result of reduced NO bioavailability in the vasculature stemming from oxidative stress induced
by the nanoparticulate fraction of DE. A DE-induced decrease in the release of tPA was also
observed at 6 h post-exposure, which may provide additional mechanistic evidence supporting the
observed association between air pollution and MI. As presented in Tornqvist et al. (2007, 091279).
changes in vascular function were also evaluated 24 h following exposure in 15 of the 30 subjects.
Compared with filtered air, exposure to DE significantly reduced endothelium-dependent (ACh)
vasodilation at 24 h post exposure. Bradykinin-induced vasodilation was marginally attenuated by
DE, while no effects of diesel on endothelium-independent vasodilation (SNP) were observed.
Although the release of tPA was not affected by DE 24 h following exposure, the authors suggest
that the persistent association between diesel exposure and vasomotor function observed in this study
provides supporting mechanistic evidence of increases in cardiovascular events occurring 24 h after
a peak in PM concentration.
To further investigate the effects of DE on vasomotor function, Mills et al. (2007, 091206)
exposed 20 men (avg age 60 yr) with previous MI on two separate occasions to dilute DE
(300 ug/m3; mean particle size 54 nm) or filtered air for 1 h with intermittent exercise. Contrary to
previous findings in younger, healthy adults (Mills et al., 2005, 095757). DE was found not to affect
vasomotor function in peripheral resistance vessels at 6 h post-exposure as measured by
endothelium-dependent (ACh) and endothelium-independent (SNP) vasodilation (forearm blood
flow). However, vascular assessments were not performed at 2 h post-exposure in this study. The
same laboratory evaluated the effect of exposure to DE with slightly higher particle concentrations
(330 ug/m3, particle number 1.26xl06/cm ) on arterial stiffness among healthy adults (Lundback et
al., 2009, 191967). Using radial artery pulse wave analysis, significant increases in augmentation
pressure and augmentation index, as well as a significant reduction in the time to wave reflection
were observed 10 and 20 min following exposure to DE relative to filtered air. This finding of a DE-
induced reduction in arterial compliance provides additional evidence to suggest that exposure to
particles may adversely affect vasomotor function.
December 2009 6-27
-------
Peretz et al. (2008, 156854) exposed both healthy adults (n = 10) and adults with metabolic
syndrome (n = 17) for 2 h to filtered air and two concentrations of diluted DE (PM2.5 concentrations
of 100 and 200 ug/m3). Compared with filtered air, DE at 200 ug/m3 elicited a statistically significant
decrease in BAD (0.11 mm [95% CI: 0.02-0.18 mm]) immediately following exposure. A smaller
DE-induced decrease in BAD (0.05 mm) was observed following exposure to 100 ug/m3. Although
this latter decrease was not statistically significant, the average decrease was approximately 50% of
the decrease at the higher particle concentration, which provides suggestive evidence of a linear
concentration response in this range of concentrations. Exposure to DE was not shown to
significantly affect endothelium-dependent FMD. Plasma levels of endothelin-1 (ET-1) were
observed to increase relative to filtered air exposure approximately 1 h after exposure to 200 ug/m3
DE (p = 0.01). Samples collected following the 100 ug/m3 exposure session were not assayed for
ET-1. The results of this study provide evidence of an acute endothelial response and arterial
vasoconstriction resulting from short-term exposure to DE. DE-induced changes in vasoconstriction
and ET-1 release were more pronounced in the healthy subjects than in the subjects with metabolic
syndrome. The authors postulated that subjects with metabolic syndrome may have stiffer vessels
that are not as responsive to vasoconstrictor stimuli. In a study utilizing a similar exposure protocol,
Lund et al. (2009, 180257) observed a significant increase in ET-1 in healthy adults following a 2-h
exposure to DE with a particle concentration of 100 ug/m3.
In the previously described studies by Mills et al. (2005, 095757; 2007, 091206). Peretz et al.
(2008, 156854). Tornqvist et al. (2007, 091279) and Lund et al. (2009, 180257). subjects were
exposed to DE, which, in addition to PM, includes DE gases such as NOX, CO, and hydrocarbons.
Therefore, it is possible that the observed effects may be due in part to exposure to non-particle
components of DE. While the majority of these DE exposures have contained relatively high levels
of gaseous emissions including NO2 concentrations >2 ppm, the concentrations of these gases were
much lower in the Peretz et al. (2008, 156854) study (NO2 concentrations ~ 20 ppb) which used a
newer diesel engine (2002 Cummins B-series) operating under load at 75% of rated capacity. In this
study, an apparent linear concentration response relationship was observed between increasing DE
exposure and decreases in BAD at particle concentrations between 100 and 200 ug/m3.
Gasoline Emissions
Rundell and Caviston (2008, 191986) exposed 15 college athletes to particles generated using
a 2.5 hp gasoline engine, as well as a clean air control during 6-min periods of maximal exercise on a
cycle ergometer. Subjects were exposed twice under each condition, with the two clean air exposures
occurring first, separated by 3 days. The 2 exposures to gasoline emissions were also separated by
3 days, with the first exposure occurring 7 days after the second clean air exposure. During
exposures to gasoline emissions, average PNC of PM <1.0 urn were reported as 336,730 and
396,200 particles/cm3 during the first and second exposures, respectively, with an average CO
concentration of 6.3 ppm. There were no differences observed in total work done (kJ) over the 6-min
exercise periods between the two clean air exposures or between the clean air exposures and the first
exposure to gasoline exhaust. However, the second gasoline exhaust exposure was demonstrated to
significantly decrease work accumulated over the 6-min exercise period compared with either of the
other exposure conditions. The results of this study provide limited evidence to suggest that a very
short term exposure to gasoline emissions may affect exercise performance in healthy adults. The
authors speculated that the observed effect of exposure on work accumulated during maximal
exercise could be due to vasoconstriction and decrease in blood flow in the skeletal muscle
microcirculation. However, the effect of exposure on vasoreactivity was not explicitly assessed.
Model Particles
The results of a recent study by Shah et al. (2008, 156970) provides evidence that exposure to
UF EC particles (50 ug/m3) without coexposure to organics, metals, or gaseous copollutants may
alter vasomotor function in healthy adults. In this study, venous occlusion plethysmography was
used to measure reactive hyperemia of the forearm prior to exposure, immediately following
exposure, and 3.5 h, 21 h, and 45 h following a 2-h exposure with intermittent exercise. Peak
December 2009 6-28
-------
forearm blood flow was observed to increase after exposure to filtered air, but not following
exposure to UF EC at 3.5 h post-exposure (p = 0.03).
Summary of Controlled Human Exposure Study Findings for Vasomotor Function
Taken together, the two studies by Mills et al. (2005, 095757; 2007, 091206) along with the
studies by Peretz et al. (2008, 156854). Lund et al. (2009, 180257) and Tornqvist et al. (2007,
091279) suggest that, in healthy subjects, DE exposure inhibits endothelium-dependent and
endothelium-independent vasodilation acutely (within 2-6 h), and that the suppression of
endothelium-dependent vasodilation may remain up to 24 h following exposure. In patients with
coronary artery disease, vasodilator function does not appear to be affected 6-8 h following
exposure; however, vascular assessments were not performed at earlier time points. In addition, the
use of medications in these patients may have blunted the response to PM. The findings of Shah
et al. (2008, 156970) suggest that UFP carbon core may be sufficient to produce small changes in
systemic vascular function, but the mechanisms remain obscure. The authors demonstrated a
decrease in nitrate levels following exposure to UF EC; however, venous nitrite level, which more
closely reflects NO production, was unchanged. Exposure to urban traffic particles was not
demonstrated to alter vasomotor function among healthy adults.
6.2.4.3. lexicological Studies
Vascular dysfunction is a function of altered production of vasoconstrictors and vasodilators.
In the 2004 PM AQCD (U.S. EPA, 2004, 056905). studies examining ET as an activator of
vasoconstriction were limited to those conducted by Bouthiller et al. (1998, 087110) and Vincent
et al. (2001, 021184). in which increased plasma ET levels were observed in rats exposed to high
concentrations (40 or 5 mg/m3) of resuspended Ottawa (EHC-93) or diesel PM, respectively. The
authors postulated that PM altered vasoconstriction via elevated ET. No studies were cited in the
2004 PM AQCD (U.S. EPA, 2004, 056905) that looked at direct measures of vasoreactivity.
As this area is newly emerging, some studies are included below that utilize IT exposure or
high concentrations; the studies that exposed vessels directly to particles ex vivo are included in
Annex D only, as their relevance is questionable. There is clearly a need for more toxicological
research examining the relationship between vascular measurements and PM exposures using
ambient particles at lower concentrations. Furthermore, no new studies have advanced the
knowledge in regards to ET as a biomarker of PM-induced vasoconstriction since the last PM
review.
CAPS
SD rats were exposed to PM2.5 CAPs (5 h/day><3 days; daily mean mass concentration
73.5-733 ug/m3; Boston, MA; 3/1997-6/1998) then the pulmonary arterial vasculature was evaluated
(Batalha et al., 2002, 088109). Some animals were repeatedly exposed to SO2 (5 h/day><5
days/wk><6 wk) to induce chronic bronchitis. Morphometric measurements indicated that the
pulmonary artery lumen-to-wall (L/W) ratio (an indicator of arterial narrowing) was decreased for
the both CAPs groups compared to the normal/air group. Furthermore, decreased L/W ratio in
CAPs-exposed animals (regardless of pre-treatment) was significantly associated with particle mass
and composition when the mean concentrations from the second and third exposure days were used
in a univariate linear regression. These results indicate a change in vascular tone following acute
exposure to PM. Univariate analyses were conducted that regressed log L/W on differential exposure
concentrations of tracer elements determined using principal components analysis (Batalha et al.,
2002, 088109). For CAPs exposure (regardless of pretreatment), CAPs mass, Si, Pb, SO42~, EC, and
OC were all negatively correlated with L/W ratio. Si and SO42~ were negatively correlated with L/W
ratio in normal rats and Si and OC were negatively correlated with L/W ratio in bronchitic rats.
When a multivariate analysis was conducted using normal and bronchitic animals, only the
association with Si remained significant. V was not associated with L/W ratio in any analysis.
December 2009 6-29
-------
Diesel Exhaust
The venous circulation plays a prominent role in heart failure exacerbation (Gehlbach and
Geppert, 2004, 155784). In heart failure, patients are often volume overloaded and are subsequently
placed on diuretics to alleviate symptoms of pulmonary congestion and chest pain. Knuckles et al.
(2008, 191987) hypothesized that if veins constrict in a manner similar to arteries, then patients with
severe CHF may have temporary shunting of fluid to the pulmonary circulation, which may elicit
signs and symptoms of CHF. Using mesenteric vessels from mice (C57BL/6) exposed to DE
(350 ug/m x4 h; MMD 100 nm, CMD 80 nm), the authors reported a significant enhancement of
ET-1-induced vasoconstriction in veins with much weaker responses in arteries. In an ex vivo
experiment, venous constriction was blocked by the arginine analog, L-NAME, which eliminates the
feedback NOS activation via endothelial ETB receptors; this is indicative of impaired or uncoupled
eNOS. The authors hypothesized that volatile organic compounds might be responsible these effects,
but no significant effects were observed for acetaldehyde, formaldehyde, acetone, hexadecane, or
pristane.
Model Particles
A study by Nurkiewicz et al. (2008, 156816) compared the arteriole dilation responses in the
spinotrapezius muscle with inhalation exposure to fine or UF TiO2 (1 urn and 21 nm, respectively;
mean mass concentration 3-16 and 1.5-12 mg/m3, respectively) for durations of 4-12 h in SD rats.
Both size fractions of TiO2 induced impaired dilation with a NO-dependent Ca2+ ionophore in a
dose-dependent manner. When fine and UF TiO2 were compared at similar mass doses, the systemic
microvascular dysfunction was greater with the UFPs. Furthermore, three exposures of differing
durations and concentrations that produced equal calculated pulmonary deposition of UF TiO2
(30 ug) demonstrated similar dilation responses, indicating that impairment is dependent upon the
timexconcentration product. No effects on dilation were observed with a dose of 4 ug UF TiO2
(1.5 mg/m3 for 4 h) or 8 ug fine TiO2 (3 mg/m3 for 4 h).
In a follow-up study, Nurkiewicz et al. (2009, 191961) examined the effect of pulmonary fine
and UF TiO2 exposure on endogenous microvasculature NO production in SD rats. The exposure
concentrations and durations were selected to produce -50% impairment of microvascular reactivity
(67 and 10 ug for fine1 and UF2 TiO2, respectively). Similar to the study above (Nurkiewicz et al.,
2008, 156816). impaired endothelium-dependent arteriolar dilation was observed 24 h post-exposure
with infusion of a Ca2+ ionophore. Earlier studies that used residual oil fly ash (ROFA) or TiO2 via
IT instillation reported similar findings, regardless of particle type (Nurkiewicz et al., 2004, 087968;
Nurkiewicz et al., 2006, 088611). There was no difference in arteriolar dilation between sham and
TiO2 exposed groups with direct administration of the NO donor SNP to the exterior arteriolar wall
and this response was consistent with that observed following ROFA administered intratracheally
(Nurkiewicz et al., 2004, 087968). The lack of response to SNP indicates that vascular smooth
muscle sensitivity to NO is not altered after particle exposure. The amount of ROS in the
microvascular wall was increased following exposure to either TiO2 size. Local ROS may consume
endothelial-derived NO and generate peroxynitrite radicals, as microvascular nitrotyrosine (NT)
formation (the end product of peroxynitrite reactions) was demonstrated after TiO2 exposure. NO
production was compromised in a dose-dependent manner following particle exposure (8-90 ug for
fine and 4-38 ug for UF TiO2), and was partially restored with agents for radical scavenging or
enzyme inhibition for NADPH oxidase and MPO.
Intratracheal Instillation
Nurkiewicz et al. (2004, 087968: 2006, 088611) have shown impairment of
endothelium-dependent dilation in the systemic microvasculature of SD rats following ROFA or
TiO2 exposure (0.1 or 0.25 mg/rat). NO-independent arteriolar dilation was also impaired by ROFA,
1 Produced by a 300-min exposure to 16 mg/m3 of fine TiO2
2 Produced by a 240-min exposure to 6 mg/m3 of ultrafine TiO2
December 2009 6-30
-------
but arteriole adrenergic sensitivity to phenylephrine (PHE) was not affected by 0.25 mg ROFA,
indicating that contractile activity was unchanged. In addition, increased venular leukocyte rolling
and adhesion in the spinotrapezius muscle was also observed following ROFA exposure (Nurkiewicz
etal.. 2004. 0879681
Further characterization of the leukocyte adherence and "rolling" effects for both ROFA and
TiO2 were indicative of an activated endothelium (Nurkiewicz et al., 2006, 088611). Vascular
deposition of MPO was observed in the spinotrapezius muscle 24 h post-exposure and the authors
suggested that the adherent leukocytes may have deposited the MPO to be taken up by endothelial
cells (Nurkiewicz et al., 2006, 088611). However, this is in contrast to another study (Cozzi et al.,
2006, 091380) that did not find changes in blood neutrophil MPO release in ICR mice exposed to
UF PM (100 ug from Chapel Hill, NC; assessed 24 h post-exposure), although this finding may be a
reflection of differing protocols. Increased oxidative stress in the arteriolar wall was also reported
with exposure to 0.25 mg ROFA. TiO2 and ROFA induced varying degrees of pulmonary
inflammation in these animals, but elicited very similar vascular effects, indicating that the vascular
responses may be due to PM presence in the lung rather than its physiochemical properties or
intrinsic pulmonary toxicity.
PM10
Tamagawa et al. (2008, 191988) reported reduced ACh-stimulated relaxation in carotid arteries
from rabbits (New Zealand White) exposed to PMi0 (EHC-93) via intrapharyngeal instillation for 5
days or 4 wk (total doses 8 and 16 mg/kg, respectively). Endothelium-dependentNO-mediated
vasorelaxation correlated with increased serum IL-6 levels in the acute study and during wk 1 and 2
of the 4-wk exposure, which may indicate a role for systemic inflammation in the response. Maximal
SNP-induced dilation was not affected by PM exposure, indicating that the dilatory response was not
acting via endothelium-independent NO-mediated mechanisms. This finding is consistent with that
by Nurkiewicz et al. (2004, 087968) and suggests that the arteriolar smooth muscle is not involved in
the PM-impaired dilatation response.
Vasoreactivity of aortic rings was measured in SH rats following exposure to 10 mg/kg PMi0
(EHC-93), with an increase in ACh-induced vasorelaxation observed (Bagate et al., 2004, 087945).
This endothelium-dependent response was greatest at 4 h and was still present at 24 h. Similarly,
vasorelaxation induced by SNP 4-h post-PM exposure was enhanced. The vasorelaxation response
was attenuated after denudation of the aortic rings, suggesting that the effect was endothelium
dependent. The findings of enhanced dilation with PM exposure contrast with those reported by
Nurkiewicz et al. (2004, 087968; 2006, 088611). Tamagawa et al. (2008, 191988). and Cozzi et al.
(2006, 091380) and may be attributable to differences in PM type, animal species, or disease models.
The authors attribute their findings to the SH rat as a well-documented model of sympathetic
hyperactivity (increased affinity of aortic smooth muscle a-adrenergic receptors) that demonstrates
upregulation of NO formation and/or release (Safar et al., 2001, 156068). No change in
vasoconstriction was observed with PM with PHE or potassium chloride.
Consistent with the impaired vasodilatory responses observed in the microvasculature and
aortic rings following PM exposure, Courtois et al. (2008, 156369) demonstrated less relaxation to
ACh in intrapulmonary arteries of Wistar rats exposed to a high dose (5 mg) of ambient PM
(SRM1648). This response was only observed 12 h after PM exposure and not at shorter (6 h) or
longer (24 or 72 h) time points. Fine TiO2 did not alter ACh-induced relaxation.
Ultrafine PM
Cozzi et al. (2006, 091380) used ICR mice to examine the effects of UF PM exposure (100 ug
collected from Chapel Hill, NC) on vascular reactivity following PM exposure and
ischemia/reperfusion injury. Aortic rings were evaluated for their contractile and dilatory responses
24 h post-exposure and following the ischemia/reperfusion protocol. Maximum ACh-induced
relaxation was impaired in UF PM-exposed vessels, as well as a rightward shift in sensitivity to
ACh. There was no difference in constriction to PHE between aortic rings from control and
PM-exposed mice. The reduced ACh-induced relaxation may be important for reperfusion of critical
vascular beds following occlusion, potentially leading to a greater area of infarction (as in this
study). A new study in dogs supports the results observed in the above study and provides evidence
of reduced myocardial blood flow following PM exposure (Bartoli et al., 2009, 179904). and is
discussed in more detail in Section 6.2.3.3.
December 2009 6-31
-------
Summary of Toxicological Study Findings for Vasoreactivity
The toxicological findings with respect to vascular reactivity are generally in agreement and
demonstrate impaired dilation following PM exposure that is likely endothelium dependent. These
effects have been demonstrated in varying vessels (right spinotrapezius muscle, carotid arteries, and
aortic rings) and in response to different PM types (ROFA, TiO2, EHC-93, UF ambient PM). The
work by Nurkiewicz et al. (2004, 087968; 2006, 088611; 2008, 156816; 2009, 191961) supports a
role for increased ROS and RNS production in the microvascular wall that leads to altered NO
bioavailability and dysfunction following particle exposure. Only one study showed enhanced
dilation with PM exposure, but the authors attributed the conflicting results to the SH rat. No
constriction changes in response to PHE were observed following PM exposure. The responses
observed in the pulmonary circulation after PM exposure include pulmonary vasoconstriction,
decreased L/W ratio, and impaired vasodilation in intrapulmonary arteries. These results are
consistent and indicate altered vascular tone. Enhancement of vasoconstriction in mesenteric veins
following DE is the first study of its kind to report on venous circulatory effects.
Endothelin
In addition to studies that look at vascular reactivity, three recent studies have examined
plasma ET levels following exposure to vehicle emissions and a few studies examined the mRNA
expression of ET-1 and ET receptors in the hearts of rodents following PM exposure.
CAPs
The upregulation of mRNA expressions of ET-1 and the ETA receptor in WKY rats exposed to
CAPs (1 or 4 days; 4.5 h/day; mean mass concentration range 1,000-1,900 ug/m3; Yokohoma City,
Japan) was correlated with increasing PM cumulative mass collected on chamber filters (Ito et al.,
2008, 096823). Furthermore, relative cardiac mRNA expressions of ET-1 and ETA receptor were
significantly correlated with CYP1B1 and HO-1 expression, indicating a possible relationship
between ET-1 metabolism and oxidative stress.
Another plasma mediator of vasomotor tone is asymmetric dimethylarginine (ADMA), which
is an endogenous inhibitor of NOS that is associated with impaired vascular function and increased
cardiovascular events. Dvonch et al. (2004, 055741) assessed levels of ADMA in Brown Norway
rats 24 h following a 3-day PM2.5 CAPs exposure in southwest Detroit (8 h/day; July 2002). CAPs
(mean mass concentration 354 ug/m3) resulted in increased plasma ADMA compared to air controls,
although the levels reported were well below the 2 uM range associated with increased CVD risk in
humans in chronic studies. Therefore, the preliminary results identified anew potential biomarker of
vascular tone that had not previously been used in air pollution toxicological studies.
Traffic-Related Particles
A study of old rats (21 mo; F344) exposed to on-road highway aerosols (number concentration
range 0.95-3.13><105 particles/cm3; Interstate 90 between Rochester and Buffalo, NY) for 6 h
demonstrated decreased plasma ET-2 (18 h post-exposure) and unchanged levels of ET-1 and ET-3
(Elder et al., 2004, 0873541
Gasoline Exhaust
In contrast to the study above, circulating levels of ET-1 (measured 18 h post-exposure) were
elevated in animals exposed to gasoline exhaust and filtration of particles did not reduce this effect
(study details in Section 6.2.2.2) (Campen et al., 2006, 096879). The results of Campen et al. (2006,
096879) are consistent with those observed by Bouthillier et al. (1998, 087110) following a very
high exposure to EHC-93, but it is difficult to attribute the effects to PM alone, as Campen et al.
(2006, 096879) showed that the gaseous components of the gasoline mixture were required for the
ET-1 increase.
Aorta ET-1 mRNA expression was increased with a 7-day gasoline exhaust exposure (60
ug/m3) in ApoE"" mice, but was not changed following a single-day exposure (Lund et al., 2009,
180257). The expression and activity of MMP-2 and -9 and oxidative stress in aortas of exposed
December 2009 6-32
-------
mice were also elevated. The ET-1 and MMP-9 mRNA expressions were attenuated with the addition
of an ETA receptor antagonist (but not a radical scavenger), indicating that ET-1 may mediate the
expression of MMP-9 through the ETA receptor.
Model Particles
Another study examined the effects of UF carbon particles (mass concentration 172 ug/m3;
mean number concentration 9.0xl06 particles/cm3) and there was no difference in ET-1, ETA or ETB
receptor mRNA expression between air- and particle-exposed SH rats 1 or 3 days post-exposure
(Upadhyay et al, 2008, 159345). In lung homogenates, ET-1, ETA and ETB receptor mRNA
expressions were elevated 3 days after exposure to UF carbon particles (Upadhyay et al., 2008,
159345).
Summary of lexicological Study Findings for Endothelin
The ET responses were mixed, with one study demonstrating ET-1 increases after exposure to
gasoline emissions that were particle independent and another reported decreased ET-2, but no
change in ET-1 or ET-3 with on-road highway exposure. Elevated levels of ET-1 and ETA receptor
mRNA expression were noted in hearts of rats exposed to CAPs, but not in rats exposed to UF
carbon particles. However, ET-1, ETA and ETB receptor mRNA expressions were increased in lung
homogenates of rats following UF carbon exposure. The ETA receptor was found to be involved in
the ET-1 and MMP-9 responses in the aortas of mice exposed to gasoline exhaust. A relatively novel
marker, ADMA, was used to evaluate vasomotor tone in rats and was found to be elevated following
exposure to CAPs, although the results are preliminary and have not been confirmed.
6.2.5. Blood Pressure
One of the potential outcomes of air pollution-mediated alterations in vascular tone is its
impact on variable BP or hypertension. BP is tightly regulated by autonomic (central and local),
cardiac, renal, and regional vascular homeostatic mechanisms with changes in arterial tone being
countered by changes in cardiac contractility, HR, or fluid volume. The evidence of PM-induced
changes in BP presented in the 2004 PM AQCD (U.S. EPA, 2004, 056905) is limited and
inconsistent. Recent epidemiologic, controlled human exposure, and toxicological studies have
similarly reported conflicting results regarding the effect of PM on BP. However, the majority of
these studies have evaluated changes in BP at some point following exposure to PM. Significant
increases in DBP have been observed in controlled human exposure studies that evaluated BP during
exposure (concomitant exposure to CAPs and O3). In addition, evidence from toxicological studies
suggests that the effect of PM on BP may be modified by health status, as PM-induced increases in
BP have been more consistently observed in SH rats.
6.2.5.1. Epidemiologic Studies
Increased BP was associated with PM concentration in two of three studies reviewed in the
2004 PM AQCD (U.S. EPA, 2004, 056905). Increases in left ventricular BP (systolic and diastolic)
are well established risk factors for cardiovascular mortality/morbidity (Welin et al., 1993, 156151).
Changes in HR and BP both reflect changes in autonomic tone, and have been examined following
short-term increases in PM pollution in several recent studies.
Ibald-Mulli et al. (2004, 087415) examined associations between BP and ambient PM2.5
concentrations, UFP counts, and ACP counts in a multicity panel study (Amsterdam, the
Netherlands; Helsinki, Finland; Erfurt, Germany) of 131 adults with coronary heart disease.
Although based on the same ULTRA Study (Timonen et al., 2006, 088747) with study methods as
described previously in Section 6.2.1.1, the study period was different. They investigated changes in
BP (SBP and DBP) associated with mean PM2.5, UFP, and ACP concentration/counts (lag days 0, 1,
and 2, as well as the 5-day mean) in each city and then generated a pooled estimate across the cities.
The median PM2.5 concentration for each city is provided in Table 6-5. Pooled analyses across all 3
cities showed small, but statistically significant decreases in SBP and DBP associated with various
single day lagged concentrations/counts of each particulate pollutant. Each 10 (ig/m3 increase in the
December 2009 6-33
-------
mean PM25 concentration over the previous 5 days was associated with a 0.36 mmHg decrease in
SBP (95% CI: -0.99 to 0.27) and a 0.39 mmHg decrease in DBF (95% CI: -0.75 to -0.03). Each
10,000 particles/cm3 increase in UFP was associated with a 0.72 mmHg decrease in SBP (95%
CI: -1.92 to 0.49), and a 0.70 mmHg decrease in DBF (95% CI: -1.38 to -0.02). Each
1,000 particles/cm3 increase in 5-day avg ACP was associated with a 1.11 mmHg decrease in SBP
(95% CI: -2.12 to -0.09) and a 0.95 mmHg decrease in DBP (95% CI: -1.53 to -0.37). The authors
concluded that these findings do not support previous findings of an increase in BP associated with
increases in particulate pollutant concentrations.
Single-city studies examining the association between BP and particulate air pollution have
been done in several U.S. and Canadian cities. Dales et al. (2007, 155743) conducted a panel study
of 39 healthy volunteers who sat outside at two different bus stops for 2-h in Ottawa, Canada. The
median PM2.5 concentrations measured at the bus stops during each 2-h exposure session were 40
and 10 (ig/m . Post-exposure SBP and DBP were not associated with the mean PM2.5 concentration
measured at the bus stops during the 2-h exposure session. The change in BP from pre- to
post-exposure was not evaluated, as health measurements were only made after the 2-h exposure
sessions.
Jansen et al. (2005, 082236) studied changes in BP among 16 older subjects (aged 60-86 yr)
with asthma or COPD in Seattle, Washington, associated with indoor, outdoor, and personal PMi0,
PM2.5, and BC concentrations on 12 consecutive days. The study authors reported that no
associations were observed between BP and daily mean PMi0, PM2.5, or BC concentrations.
Zanobetti et al. (2004, 087489) examined the association between BP (SBP, DBP, and mean
arterial BP) and mean PM25 concentrations in the previous 24, 48, 72, 96, and 120 h in 62 elderly,
cardiac rehabilitation patients in Boston, MA (Zanobetti et al., 2004, 087489). Each 10.4 (ig/m3
increase in mean PM25 concentration in the previous 120 h was associated with significant increases
in resting DBP (2.82 mmHg [95% CI: 1.26-4.41]), SBP (2.68 mmHg [95% CI: 0.04-5.38]), and
mean arterial BP (2.76 mmHg [95% CI: 1.07-4.48]).
Mar et al. (2005, 087566) studied this same PM25-BP association in 88 subjects aged >57 yr in
Seattle, WA. Among healthy subjects taking medications (bronchodilators, inhaled corticosteroids,
anti-hypertensives, |3-blockers, calcium channel blockers, and/or cardiac glycosides), each 10 (ig/m3
increase in mean outdoor PM2 5 concentration on the same day as the BP measurement was made
was associated with small increases in SBP and DBP. However, among all subjects, each 10 (ig/m3
increase in same day mean PM2 5 concentration was associated with non-significant decreases in SBP
(-0.81 mmHg [95% CI: -2.34 to 0.73]) and DBP (-0.46 mmHg [95% CI: -1.49 to 0.57]).
As described earlier, Ebelt et al. (2005, 056907) conducted a repeated measures panel study of
16 patients with COPD in the summer of 1998 in Vancouver, British Columbia to evaluate the
relative impact of ambient and non-ambient exposures to PM25, PMi0, and PMi0_2.5 on multiple
health outcomes including ectopy and BP. Using the same analytic methods, pollutant
concentrations, and lags, they reported decreased SBP associated with same day ambient exposures
to each PM size fraction.
Two similar studies were done in Incheon, South Korea (Choi et al., 2007, 093196) and
Taipei, Taiwan (Chuang et al., 2005, 156356). Choi et al. (2007, 093196) reported significantly
increased SBP and DBP associated with the mean PMi0 concentration over the same and previous 2
days in the warm season only (July to September). Chuang et al. (2005, 156356) reported significant
increases in SBP and DBP associated with the mean UFP count (0.01-0.1 (im particles) 1-3 h before
the BP measurement.
Summary of Epidemiologic Studies of Blood Pressure
These studies (Choi et al., 2007, 093196; Chuang et al., 2005, 156356; Dales et al., 2007,
155743; Ibald-Mulli et al., 2004, 087415; Mar et al., 2005, 087566; Zanobetti et al., 2004, 087489)
are not entirely consistent with regard to their BP-PM associations. Most have reported increases in
SBP and DBP associated with increases in either PM25, PMi0, or UFP (Choi et al., 2007, 093196;
Chuang et al., 2005,156356; Mar et al.. 2005. 087566; Zanobetti et al.. 2004. 087489). However.
two studies reported small decreases in BP associated with multiple particulate pollutants (Ibald-
Mulli et al., 2004, 087415; Mar et al., 2005, 087566). Dales et al. (2007, 155743) reported no
change in BP associated with a 2-h exposure to bus stop PM2 5 and Jansen at al. (2005, 082236)
reported null findings among older adults in Seattle, WA. Exposure lags ranging from 1-3 h (Chuang
December 2009 6-34
-------
et al, 2005, 156356). to the same day(Ebelt et al., 2005, 056907; Mar et al., 2005, 087566). to the
mean across the previous 5 days (Zanobetti et al., 2004, 087489) were reported as having the
strongest associations with BP. Mean and upper percentile concentrations for PM from these studies
are presented in Table 6-5.
Table 6-5. Mean PM concentrations reported in epidemiologic studies of blood pressure.
Author
Location
Mean Concentration (ug/m
Upper Percentile
Concentrations (ug/m3
PM2.5
Dales (2007,155743)
Ottawa, Canada (bus stops)
Bus stop 1:40
Bus stop 2:10
NR
Ebelt (2005.056907)
Ebelt (2005,056907)
Vancouver, Canada
Ambient (measured): 11.4
Personal (estimated): 7.9
Personal (measured): 18.5
Ambient (measured) range:
4.2-28.7
Personal (estimated) range:
0.9-21.3
Personal (measured) range:
2.2-90.9
Amsterdam, Netherlands
Ibald-Mulli (2004, 087415) Erfurt, Germany
Helsinki, Finland
Jansen (2005, 082236) Seattle, WA
Mar (2005, 087566) Seattle, WA
Zanobetti (2004, 087489) Boston, MA
20
23.1
12.7
10.47
Healthy: Personal- 9.3
Indoor- 7.4
Outdoor- 9
CVD: Personal- 10.8 Indoor- 9.5
Outdoor- 12.6
COPD: Personal- 10.5
Indoor- 8.5
Outdoor- 9.2
Median: 8.8
50th: 16. 9
75th: 23.9
Max: 82.2
50th: 16. 3
75th: 27.4
Max: 118.1
50th: 10. 6
75th: 16
Max: 39.8
NR
NR
90th: 17. 6
PMiO-2.5
Vancouver, Canada
Ambient (calculated): 5.6
Personal (estimated): 2.4
Ambient (calculated) range: -1.2 to
11.9
Personal (estimated) range: -0.4 to
7.2
PMu
Choi (2007, 093196)
Incheon, South Korea
July-Sept: 42.1
Oct.-Dec: 53.5
July-Sept.: 75%: 52.2
Max: 136.7
Oct.-Dec.: 75%: 64.5
Max: 209.6
December 2009
6-35
-------
Author Location
Chuang (2005, 156356) Taipei, Taiwan
Ebelt (2005, 056907) Vancouver, Canada
Jansen (2005. 082236) Seattle, WA
Mar (2005. 087566) Seattle, Washington
Mean Concentration (ug/m3)
54.1
Ambient (calculated): 17
Personal (estimated): 10.3
13.47
Healthy: 14.5
CVD:18
COPD:14.3
Upper Percentile
Concentrations (ug/m3)
Range: 10.3-1 39.8
Ambient (calculated) range: 7-36
Personal (estimated) range:
1.5-23.8
NR
NR
Right Ventricular Pressure
Several recent studies, summarized in the section on hospital admissions and emergency
department (ED) visits for CVD causes, have reported increased risk of hospital admissions for CHF
associated with increased PM concentration on the same day (Wellenius et al., 2005, 087483; 2006,
088748). As a possible mechanism for these reported associations, Rich et al. (2008, 156910)
hypothesized that these hospital admissions for decompensation of heart failure would be preceded
by more subtle increases in pulmonary arterial (PA) and right ventricular (RV) diastolic pressures.
They used passively monitored PA and RV pressures on 5,807 person-days, among 11 subjects
implanted with the Chronicle Implantable Hemodynamic Monitor [Medtronic, Inc. Medtronic,
MN]). Using a two-stage modeling process, they examined the change in daily mean right heart
pressures associated with mean PM2.5 concentration on the same and previous 6 days. Each
11.62 ug/m3 increase in same day mean PM2.5 concentration was associated with small, but
statistically significant increases in estimated PA diastolic pressure (0.19 mmHg [95% CI: 0.05-
0.33]) and RV diastolic pressure (0.23 mmHg [95% CI: 0.11-0.34]). These effects were not
attenuated when controlling for all lags simultaneously. Thus, PM induced right heart pressure
increases may mark another potential pathway between PM exposure and incidence of
cardiovascular events, but further studies on this same hypothesis are needed for confirmation.
Wellenius et al. (2007, 092830) conducted a panel study of 28 subjects living in the greater
Boston metropolitan area, each with chronic stable heart failure and impaired systolic function. They
hypothesized that circulating levels of B-type natriuretic peptide (BNP), measured in whole blood at
0, 6, and 12 wk, were associated with acute changes in ambient air pollution, as a possible
mechanistic explanation for the observed association between hospital admissions for CHF and
ambient PM concentration (Wellenius et al., 2005, 087483; 2006, 088748). During the study the
mean PM2.5 concentration was 10.9 ug/m3, while the mean BC concentration was 0.73 ug/m . Using
linear mixed models, they reported no association between any pollutant (PM2.5, CO, SO2, NO2, O3,
and BC) and BNP at any lag (e.g., each 10 ug/m3 increase in mean daily PM25 concentration [0.8%
increase in BNP (95% CI: -16.4 to 21.5)]) (Wellenius et al., 2007, 092830). However, BNP the
active peptide has a very short half-life and might not be the best biomarker for such a study. Thus
the absence of a correlation between PM and BNP may not suggest that PM does not have an impact
on RV or LV function in individuals with impaired cardiac mechanics.
6.2.5.2. Controlled Human Exposure Studies
Only one controlled human exposure study cited in the 2004 PM AQCD (U.S. EPA, 2004,
056905) reported any PM-induced changes in BP. Gong et al. (2003, 042106) found that exposure to
PM2.5 (174 ug/m3) decreased SBP in asthmatics, but increased SBP in healthy subjects. Among
healthy adults, BP was not affected following 2-h exposures to 200 ug/m3 diesel PM (Nightingale et
al., 2000 011659). 150 ug/m3 PM25 CAPs with 120 ppb O3 (Brook et al., 2002, 024987). or
10 ug/m UF carbon particles (Frampton, 2001, 019051). The effect of PM on BP has been further
investigated in several recent controlled human exposure studies, which are described below.
December 2009
6-36
-------
CAPS
One recent study demonstrated a significant increase (9.3%) in DBF among healthy adults
immediately prior to the end of a 2-h exposure to 150 ug/m3 PM25 CAPs in combination with
120 ppb O3 (Urch et al., 2005, 081080). The authors also found that the magnitude of change in BP
was significantly associated with PM2.5 carbon content, but not total PM2.5 mass. It was postulated
that the disparity between these findings and those of a similar study by the same group (Brook et
al., 2002, 024987) could be due to differences in experimental methods. The Brook et al. (2002,
024987) study measured post-exposure BP approximately 10 min following exposure, while the
study by Urch et al. (2005, 081080) measured BP during exposure. In a follow up study that
evaluated changes in BP during a 2-h exposure to PM2 5 CAPs, Fakhri et al. (2009, 191914) reported
a significant increase in DBP with exposure to CAPs with, but not without, coexposure to O3.
Diesel Exhaust
Several recent studies have assessed BP changes following a 1-h exposure to DE with a
particle concentration of 300 ug/m3. Mills et al. (2005, 095757) evaluated changes in BP 2 h
following exposure to DE and found a 6 mmHg increase in DBP of marginal statistical significance
(p = 0.08) compared to filtered air control. In this same group of subjects, Tornqvist et al. (2007,
091279) did not observe any such changes in BP 24 h following DE exposure. At lower particle
concentrations in diluted DE (100-200 ug/m3 PM2.5), Peretz et al. (2008, 156854) did not observe
any changes in systolic or DBP in either healthy adults or adults with metabolic syndrome
immediately following a 2-h exposure. Further, although Lundback et al. (2009, 191967) reported an
increase in arterial stiffness following exposure to DE with a particle concentration of 330 ug/m3
among healthy young adults, no changes in systolic or diastolic BP were observed during or
following exposure relative to filtered air.
Model Particles
Routledge et al. (2006, 088674) did not observe any changes in BP among healthy older adults
and older adults with stable angina following a 1-h exposure to UF EC (50 ug/m3), with or without
coexposure to 200 ppb SO2. Similarly, neither Shah et al. (2008, 156970). nor Beckett et al. (2005,
156261) reported any changes in BP among healthy adults following exposure to UF EC (50 ug/m3)
or ZnO (500 ug/m3 fine and ultrafine), respectively.
Summary of Controlled Human Exposure Study Findings for BP
The findings of these new studies do not provide convincing evidence of an association
between PM exposure and an increase in BP; however, they do suggest that there is a need for
additional investigations of PM-induced changes in BP at various time points following exposure.
6.2.5.3. lexicological Studies
In healthy animal models, little evidence exists for significant BP changes following inhalation
exposure to environmentally-relevant concentrations of PM. Only one animal toxicological study is
mentioned in the 2004 PM AQCD (U.S. EPA, 2004, 056905) that examined BP with PM exposure
and no effect was observed (Vincent et al., 2001, 021184).
CAPs
In a recent study of dogs, exposure to PM2 5 CAPs from Boston (mean mass concentration
358.1 ug/m3; mass concentration 94.1-1557 ug/m3) for 5 h resulted in increased SBP (2.7 mmHg),
DBP (4.1 mmHg), mean arterial pressure (3.7 mmHg), and lowered pulse pressure (1.7 mmHg)
when measured upstream of the femoral artery (Bartoli et al., 2009, 156256). Administration of an
December 2009 6-37
-------
a-adrenergic antagonist (prazosin) prior to CAPs attenuated the BP responses. These findings
indicate that CAPs exposure may have activated a-adrenergic receptors and increased peripheral
vascular resistance. Baroreflex sensitivity was measured immediately before and after exposure
during a transient elevation of arterial pressure that was induced by PHE; increased baroreflex
sensitivity was observed in subgroup of dogs exposed to CAPs, which is consistent with an
upregulation of vagal reflexes.
Chang et al. (2004, 055637) noted slight increases in SH rat BP (5-10 mmHg) when exposed
to PM2.5 CAPs (mean mass concentration 202 ug/m3) during spring months. However, during
summer months, when the CAPs exposure level was less (140 ug/m3), this effect was not observed.
It was unclear, therefore, whether the effects were seasonal or dose-related. In a preliminary study of
SH rats exposed to CAPs during a dust storm event, mean BP was elevated the third and fourth hour
of a 6-h exposure, although interpretation of this finding is difficult due to few animals in the
exposure group (n = 2) (Chang et al., 2007, 155719). In another study, the increased change in mean
BP measured using the tail cuff method following CAPs exposure weakly correlated with PM mass
accumulated on chamber filters over the entire exposure duration (Section 6.2.4.3 for details) (Ito et
al., 2008, 096823). Furthermore, ETA receptor mRNA expression in cardiac tissue was positively
correlated with the change in mean BP.
Model Particles
In WKY rats, 24-h exposure to UF carbon particles (mass concentration 180 ug/m3; mean
number concentration 1.6x107 particles/cm3) did not alter mean BP during exposure or the recovery
periods (Harder et al., 2005, 087371). SH rats exposed to UF carbon particles for 24 h (mass
concentration 172 ug/m3; mean number concentration (9.0xl06 particles/cm3) resulted in elevated
mean BP (by 6 mmHg) on the first and second days of recovery following exposure that was
attributable to increases in both SBP and DBP (Upadhyay et al., 2008, 159345). Increased plasma
renin concentrations were observed in CB-exposed rats on the first and second days of recovery,
although renin activity and angiotensin (Ang) I and II concentrations were not affected by particle
exposure.
Summary of lexicological Study Findings for Blood Pressure
Limited toxicological evidence provides support for elevated BP in dogs or compromised rats
with CAPs, UF CAPs, CAPs during a dust storm event, or UF carbon particle exposure. However,
most of the CAPs studies were conducted outside of the U.S.
6.2.6. Cardiac Contractility
The 2004 PM AQCD (U.S. EPA, 2004, 056905) did not include any toxicological studies that
evaluated cardiac contractility either directly or indirectly following exposure to PM. Two recent
animal toxicological studies have demonstrated reductions in cardiac fractional shortening,
diminished ejection shortening, or changes in the QA interval following PM exposure. The results of
these studies provide some evidence of PM-induced changes in cardiac contractility in animal
models.
6.2.6.1. Toxicological Studies
The strength of the contracting heart is reflected by its contractility. In heart failure,
contractility wanes significantly and the heart cannot compensate during periods of increased
physical activity. Measuring true contractility in a whole animal is difficult, requiring extensive
surgical instrumentation and monitoring.
December 2009 6-38
-------
CAPS
Using radiotelemetry to indirectly measure cardiac contractility through the QA interval, SH
rats were repeatedly and alternately exposed to UF CAPs in Taiwan on separate days in spring or
summer (details provided in Section 6.2.5.3) (Chang et al., 2004, 055637). The QA interval was
calculated as the time duration between the Q wave in the ECG and point A (upstroke in aortic
pressure) in the pressure trace and is not as reliable as other measures, such as echocardiography.
During the spring exposure, QA interval decreased by 1.6 ms (as demonstrated by fixed effects in
linear mixed-effects modeling), which indicates an increase in cardiac contractility. There were no
changes in QA interval observed for the summer months, which may be attributable to lower UF PM
concentrations (mean mass concentration 140 ug/m3) or differing PM compositions.
Model Particles
A recent study using old (18-28-mo) mice (C57BL/6, C3H/HeJ, and B6C3F1) demonstrated
significant reductions in cardiac fractional shortening (due to increased left ventricular end-diastolic
and end-systolic diameters) following a 4-day (3 h/day) exposure to CB (PM2.5 mean concentration
401 ug/m ; PMi0 mean concentration 553 ug/m3) using echocardiography (Tankersley et al., 2008,
157043). Hemodynamic measurements of diminished ejection fraction and maximum change in
pressure over time further supported lowered myocardial contractility. Furthermore, increased right
ventricular pressure associated with elevated right atrial and pulmonary vascular pressures and
resistance, was indicative of pulmonary vasoconstriction in CB-exposed mice. Heart tissue and
isolated cardiomyocytes from exposed animals demonstrated enhanced ROS that was partially
attributable to NO S3-uncoupling and elevated MMP-2 and MMP-9 levels, which may implicate
myocardial remodeling. The combined results from this study suggest that cellular mechanisms
involving NOS-uncoupled ROS generation likely mediate PM-induced cardiac effects. Furthermore,
mRNA expression for atrial and brain natriuretic peptides was increased in hearts from exposed mice
compared to control, which is consistent with pulmonary congestion. There were no reported
strain-related differences in any response.
Intratracheal Instillation
Similar to the responses observed by Tankersley et al. (2008, 157043). decreases in fractional
shortening and increases in left ventricular end diastolic diameter measured by echocardiography
were also reported for SD rats at 24 h post-IT exposure to DE particles (250 ug) (Yan et al., 2008,
098625). A subset of rats received isoproterenol to induce myocardial injury prior to IT instillation of
DE particles and these animals demonstrated lowered fractional shortening at baseline, which was
decreased to an even greater extent with DE particle exposure; left ventricular end diastolic diameter
was not affected by DE particles in these rats.
Summary of lexicological Study Findings for Cardiac Contractility
The studies above provide some evidence that cardiac contractility may be altered
immediately following PM exposure in animal models. Results from the Tanksersley (2008, 157043)
and Yan (2008, 098625) studies provide the strongest support for PM-induced contractility changes
with inhalation exposure, as echocardiography and hemodynamic measurements are well-established
for examining cardiac function.
6.2.7. Systemic Inflammation
The evidence presented in the 2004 PM AQCD (U.S. EPA, 2004, 056905) of increases in
markers of systemic inflammation associated with PM was limited and not sufficient to formulate a
definitive conclusion. Recent controlled human exposure and toxicological studies continue to
provide mixed results for an effect of PM on markers of systemic inflammation including cytokine
December 2009 6-39
-------
levels, C-reactive protein (CRP), and white blood cell (WBC) count. While results from recent
epidemiologic studies have also been inconsistent across studies, there is some evidence to suggest
that PM levels may have a greater effect on inflammatory markers among populations with
preexisting diseases.
6.2.7.1. Epidemiologic Studies
Several studies reviewed in the 2004 PM AQCD (U.S. EPA, 2004, 056905) investigated the
association of short-term fluctuations in PM concentration with markers of inflammation. These
studies were found to offer limited support for mechanistic explanations of the associations between
PM concentration and heart disease outcomes. Recent studies, published since 2002, are reviewed
below. CRP was measured in multiple studies, allowing the consistency of findings across
epidemiologic studies to be evaluated. Several other markers were examined in only a few studies, in
relation to a wide range PM size fractions and components. These markers included IL-6, TNF-a,
vascular cell adhesion molecule-1 (VCAM-1), intercellular adhesion molecule-1 (ICAM-1), soluble
CD40 ligand (sCD40L), WBCs, and soluble adhesion molecules (sP-selectin and e-selectin).
Diez-Roux et al. (2006, 156400) examined whether CRP increased in response to changes in
the mean ambient PM2.5 concentrations in the prior day, prior 2 days, prior week, prior 30 days, and
prior 60 days among participants in the Multi-Ethnic Study of Atherosclerosis (MESA) cohort.
Subjects (n = 5,634) lived in either Baltimore City or County, MD, Chicago, IL, Forsyth County,
NC, Los Angeles County, CA, Northern Manhattan and the Bronx, NY, or St. Paul, MN. The authors
report finding no evidence of a short-term effect of PM2.5 on CRP in their population-based sample.
Of the five exposure measures examined, only the 30-day and 60-day mean exposures showed
positive associations with PM2.5 (3% [95% CI: -2 to 10] and 4% [95% CI: -3 to 11] per 10 (ig/m3,
respectively).
Ruckerl et al. (2007, 156931) conducted a multicity longitudinal study to examine whether
changes in markers of inflammation were associated with short-term increases in particulate
concentrations (PMi0, PM2.5, PNC) and gaseous pollutant (NO2, SO2, CO, O3). Study subjects were
MI survivors (n= 1,003) living in either Athens, Greece; Augsburg, Germany; Barcelona, Spain;
Helsinki, Finland; Rome, Italy; or Stockholm, Sweden. Repeated measurements of IL-6 and CRP
were made during the study. Fibrinogen was also measured in this study and results are discussed in
Section 6.2.8.1. The mean city-specific pollutant concentrations during the study are shown below in
Table 6-6. In pooled analyses, each interquartile range (not provided) increase in PNC in the 12-17 h
before the health measurement was associated with a 2.7% increase in the geometric mean IL-6
levels (95% CI: 1.0-4.6). None of the pollutants, at any lag, were associated with CRP levels in these
subjects. There did not appear to be effect modification of these results by smoking, diabetes, or
heart failure. Ljungman et al. (2009, 191983) studied the modification of the IL-6 association with
several PM size fractions (PMi0, PM2 5, PNC) by three IL-6 SNPs, one fibrinogen a chain (FGA)
single-nucleotide polymorphism (SNP) and one fibrinogen (3 chain (FGB) SNP The associations of
PM25 and PMi0 with plasma level of IL-6 were stronger among those with the homozygous minor
allele genotype of FGB rs!800790 and among those homozygous for the major allele genotype of
IL-6 rs2069832. Gene-environment interactions were most pronounced for CO. Modification the
PNC-IL-6 association by genotype was not apparent in these data, nor was modification of the PM-
IL-6 associations by FBA.
Single-city studies of systemic inflammation have also been conducted in the U.S. and
Canada. Delfmo et al. (2008, 156390) measured CRP, IL-6, TNF-a, sP-selectin, sVCAM-1 and
sIC AM-1 in blood during a period of 12 wk. Associations of these markers with average PM
concentration (PM0.25, PM0.25_2.5, PM10_2.?, PNC, EC, OC, BC, primary OC, secondary OC) 24 h to 9
days prior to the blood draw were examined. Subjects included residents of two downtown Los
Angeles nursing homes who were >65 yr old with a history of coronary artery disease. Both 24-
h avg and multiday average concentrations of PM0.25, EC, primary OC, BC, PNC and gaseous
pollutants were associated with CRP, IL-6 and sP-selectin.
Pope et al. (2004, 055238) conducted a panel study of 88 non-smoking, elderly subjects
residing in the Salt Lake City, Ogden, and Provo metropolitan area of Utah. Each 100 (ig/m3 increase
in same day mean PM25 concentration was associated with a 0.81 mg/dL increase in CRP (95%
CI: 0.48-1.14), but not WBCs. However, when excluding 1 influential subject, each 100 ug/m3
increase in same day mean PM25 concentration was associated with only a 0.19 mg/dL increase in
December 2009 6-40
-------
CRP (95% CI: -0.01 to 0.39). Several markers of coagulation were examined in this study and are
discussed in Section 6.2.8.1.
Zeka et al. (2006, 157177) studied 710 elderly members of the VA Normative Aging Study to
examine changes in CRP, sediment rate and WBCs with acute changes in PM concentrations in the
previous 48 h, 1 wk, and 4 wk. Results for fibrinogen are discussed in Section 6.2.8.1. They did not
find consistent or significant associations with any pollutant and CRP or WBC count. Sediment rate
was significantly increased with PNC, BC and PM2 5 concentration averaged over the previous 4 wk
period. Modification of these PM effects by obesity, GSTM1 genotype and statin use was suggested
in this study.
O'Neill et al. (2007, 091362) conducted a cross-sectional study of 92 Boston residents with
type 2 diabetes, to examine the association between plasma levels of 1C AM-1, VCAM-1 and PM
concentrations. Results for markers of coagulation measured in this study are discussed in
Section 6.2.8.1. PM2.5, BC, and SO42" concentrations were measured 0.5 km from the patient exam
site. For all moving averages examined (1-6 days), increases in mean PM25 and BC concentration
were associated with increased ICAM-1 and VCAM-1 concentrations. Each 7.6 (ig/m3 increase in
the mean PM25 concentration over the previous 6 days was associated with a 11.76 ng/mL increase
in VCAM-1 (95% CI: 3.48-20.70), and each 0.6 (ig/m3 increase in the mean BC concentration over
the previous 6 days was associated with a 27.51 ng/mL increase in VCAM-1 (95% CI: 11.96-45.21).
There were no consistent associations between mean SO42" concentration and any marker at any lag.
Sullivan et al. (2007, 100083) conducted a panel study of 47 subjects (aged >55 yr) either with
COPD (n = 23) or without COPD (n = 24) in Seattle, WA. They examined the association between
levels of CRP and mean daily PM2 5 concentration. Most values for IL-6 and TNF-a were below the
limit of detection, so these cytokines were not included in the analyses. Results for fibrinogen and
D-dimer are discussed in Section 6.2.8.1. They did not find any associations between 24-h mean
PM2 5 concentrations and levels of CRP in individuals with or without COPD.
In the study by Liu et al. (2006, 192002; 2007, 156705). conducted in Toronto, Ontario,
neither CRP (0.11 (ig/mL [95% CI: -0.03 to 0.25]) nor TNF-a (0.03 pg/mL [95% CI: -0.07 to 0.13])
was associated with personal exposure to PMi0 (24-h averaging time).
Similarly, there was no association with IL-6. However, significant positive associations with
markers of oxidative stress, FMD and BP were found and are discussed in Sections 6.2.9.1, 6.2.4.1,
and 6.2.5.1, respectively.
In the St. Louis Bus Study, each 5.4 (ig/m3 increase in the mean PM2 5 concentration over the
previous week was associated with 5.5% increase in WBCs (95% CI: 0.10-11) (Dubowsky et al.,
2006, 088750). Each 6.1 (ig/m3 increase in the mean PM2 5 concentration over the previous 5 days
was associated with a 14% increase in CRP among all subjects (95% CI: -5.4 to 37), but an 81%
increase in CRP (95% CI: 21-172) among subjects with diabetes, obesity, and/or hypertension.
Associations between PM2 5 and IL-6 were only observed among those with diabetes, obesity, and/or
or hypertension. In another study of in-vehicle PM2 5, each 10 (ig/m3 increase during a work-shift
was associated with decreased lymphocytes, increased mean corpuscular volume, neutrophils, and
CRP over the next 10-14 h among 9 healthy North Carolina state troopers (Riediker et al., 2004,
056992). Associations of roadside and ambient PM25 with systemic inflammatory markers were
weaker and non-significant in this population.
International studies of the effect of air pollution on markers of inflammation have been
conducted with mixed results. Two studies conducted among 57 male patients with coronary heart
disease in Erfurt, Germany, found associations of UFP, ACP and PMi0 with CRP (Ruckerl et al.,
2006, 088754) and UFP and ACP with sCD40L, a marker for platelet activation (Ruckerl et al.,
2007, 156931). In a large cross-sectional study of healthy subjects in Tel Aviv, Steinvil et al. (2008,
188893) examined biological markers of inflammation (CRP and WBCs) collected as part of routine
health examinations for 3,659 individuals. Associations with air pollutants (including PM10)
measured at local monitoring sites for the day of the examination and up to 7 days prior were
examined. No significant associations were found between pollutant levels and indications of
enhanced inflammation. By contrast, PMi0, PM25, SO42~ and nitrate (3-day avg concentrations) were
associated with increases in hs-CRP in healthy students in Taiwan (Chuang et al., 2007, 091063).
PMio, PM25 and PM0.2s were not associated with CRP in a study of MI patients in Italy, although
associations with autonomic dysregulation and more severe arrhythmias were observed (Folino et
al., 2009, 191902). Kelishadi et al. (2009, 191960) reports that CRP, as well as markers of insulin
resistance and oxidative stress (discussed in Section 6.2.9.1), were associated with PMi0 in a cross-
December 2009 6-41
-------
sectional study of a population-based sample of children 10-18 yr old in Iran (mean PM
concentration 122.08 (ig/m3).
10
Summary of Epidemiologic Study Findings for Systemic Inflammation
The most commonly measured marker of inflammation in the studies reviewed was CRP. CRP
was not consistently associated with short-term PM concentrations (PM2.5, PMi0, SO^, EC, OC,
PNC). A multicity study of MI survivors in Europe (Ruckerl et al., 2007, 156931) failed to provide
evidence of an effect of PM (e.g., PMi0, PM2.5, PNC) on CRP and no effect was observed by Diez-
Roux et al. (2006, 156400) in a population-based study when concentrations were averaged over
periods less than 30 days. Several other markers of inflammation have been examined in relation to
several PM size fractions and components, but the number of studies examining the same
marker/PM metric combination is too few to allow results to be compared across epidemiologic
studies. Mean and upper percentile concentrations for those epidemiologic studies that evaluated
systemic inflammation are included in Table 6-6.
Table 6-6. PM concentrations reported in epidemiologic studies of inflammation, hemostasis,
thrombosis, coagulation factors and oxidative stress.
Author
Location
Mean Concentration (ug/m
Upper Percentile Concentrations
(ug/m3)
PM2.5
Chuang (2007, 091063)
Diez-Roux (2006, 156400)
Dubowsky (2006, 088750)
Folino (2009, 191902)
O'Neill (2007, 091362)
Park (2008, 156845)
Peters (2009, 191992)
Taipei, Taiwan
Chicago, IL
Baltimore, MD
Forsyth County, NC
Los Angeles, CA
New York City, NY
St. Paul, MN
St. Louis (bus stops)
Padua, Italy
Boston, MA
Boston, MA
Helsinki, Finland
Stockholm, Sweden
Augsburg, Germany
Rome, Italy
Barcelona, Spain
1-dayavg:31.8
2-dayavg:36.4
3-dayavg:36.5
Prior day (median): 14.3
Prior 2 days (median): 14.4
Prior 7 days (median): 15.24
Prior 30 days (median): 15.69
Prior 60 days (median): 15.9
16
Summer: 33.9
Winter: 62.1
Spring: 30.8
11.4
12
Helsinki: 8.2
Stockholm: 8.8
Augsburg: 17.4
Rome: 24.5
Barcelona: 24.2
Total: 16.4
1-day avg (range): 16.2-50.1
2-day avg (range): 15-53.4
3-day avg (range): 12. 7-59. 5
Prior day (75th): 20.9
Prior 2 days (75th): 20.35
Prior 7 days (75th): 19.7
Prior 30 days (75th): 19.22
Prior 60 days (75th): 19.08
75th: 22
100th: 28
NR
Range: 0.07-33.7
Range: 2-62
Helsinki (range): 1-28
Stockholm (range): 0-27
Augsburg (range): 6-39
Rome (range): 4-95
Barcelona (range): 3-95
Total (range): 0-95
December 2009
6-42
-------
Author
Pope (2004, 055238)
Riediker (2004, 056992)
Ruckerl (2007, 156931)
Ssrensen (2003, 157000)
Sullivan (2007, 100083)
Zeka (2006. 157177)
Location
Salt Lake City, Ogden, Provo Utah
North Carolina State Troopers
Helsinki, Finland
Stockholm, Sweden
Augsburg, Germany
Rome, Italy
Barcelona, Spain
Athens, Greece
Copenhagen, Denmark
Seattle, WA
Boston, MA
Mean Concentration (ug/m3)
FRM-Filled:23.7
Not filled: 25.8
TEOM:18.9
RAMS/PC-BOSS: 26. 5
Light Scatter: 24.1
Mass: 23
Ambient: 32.3
Roadside: 32.1
8.2(19.4)
8.8(19.1)
17.4(29.3)
24.5(54.1)
24.2 (64.7)
23 (46)
Personal (median): 16.1
Urban background (median): 9.2
Outdoor (median): 7.7
Indoor (median): 7.7
48h (median): 9.39
Upper Percentile Concentrations
(ug/m3)
FRM-Filled (range): 1.7-74
Not filled (range): 1.7-74
TEOM (range): 2.2-61 .5
RAMS/PC-BOSS (range): 5.6-72.4
Light Scatter (range): 4.5-54.4
Mass (range): 7.1-38.7
Ambient (range): 9.9-68.9
Roadside (range): 8.9-62.2
NR
NR
NR
NR
NR
NR
Personal (025-075): 10-24.5
Urban background (025-075): 5.3-14.8
Outdoor: 75th- 11. 5
90th- 19.9
Max- 33.9
Indoor: 75th- 12.1
90th- 16
Max- 81 .4
75th: 14.57
90th: 21 .48
PMiO-2.5
Delfino (2008, 156390)
Peters (2009, 191992)
Los Angeles, CA
Helsinki, Finland
Stockholm, Sweden
Augsburg, Germany
Rome, Italy
Barcelona, Spain
Outdoor: 10.04 (4.07)
Indoor: 4.1 2 (4.76)
Helsinki: 8.9
Stockholm: 9
Augsburg: 15.8
Rome: 16.8
Barcelona: 16.5
Total: 13.3
Outdoor (range): 1.76-22.38
Indoor (range): 0.1 2-37.63
Helsinki (range): 1-38
Stockholm (range): 0-40
Augsburg (range) :-1 to 35
Rome (range): -33 to 65
Barcelona (range): 1-102
Total (range): -33 to 102
PM10
Baccarelli (2007. 090733)
Baccarelli (2007. 091310)
Chuang (2007, 091063)
Lombardia Region, Italy
Lombardia Region, Italy
Taipei, Taiwan
Sep-Nov(median):51.2
Dec-Feb (median): 68.5
Mar-May (median): 64.1
Jun-Aug (median): 44.3
Median: 34.1
1-dayavg:49.2
2-dayavg:55.3
3-day avg: 54.9
Sep-Nov (max): 148.9
Dec-Feb (max): 238.3
Mar-May (max): 158.5
Jun-Aug (max): 94.7
Maximum: 390
1 -day avg (range): 29. 5-83. 4
2-day avg (range): 25. 5-85.1
3-day avg (range): 22.2-87.2
December 2009
6-43
-------
Author
Location
Mean Concentration (ug/m
Upper Percentile Concentrations
(ug/m3)
Folino (2009, 191902) Padua, Italy
Kelishadi (2009, 191960) Isfahan, Iran
Summer: 46.4
Winter: 73
Spring: 38.3
122.08
NR
75th: 153
100th: 191
Washington County, MD
Forsyth County, NC
Minneapolis, MN (suburbs)
29.9
Q4:47.3
Windsor, Ontario, Canada
Personal (median):
0-24 h before clinical visit: 25.5
0-6 h before clinical visit: 15.3
7-12 h before clinical visit: 17
13-18 h before clinical visit: 28.5
19-24 h before clinical visit: 30.5
Personal (5th to 95th):
0-24 h before clinical visit: 9.8-133
0-6 h before clinical visit: 5.3-83.2
7-12 h before clinical visit: 7.1-186.3
13-18 h before clinical visit: 11.4-167
19-24 h before clinical visit: 10.1-148.2
Peters (2009, 191992)
Ruckerl (2007, 156931)
Steinvil (2008, 188893)
Helsinki, Finland
Stockholm, Sweden
Augsburg, Germany
Rome, Italy
Barcelona, Spain
Helsinki, Finland
Stockholm, Sweden
Augsburg, Germany
Rome, Italy
Barcelona, Spain
Athens, Greece
Tel Aviv, Israel
Helsinki: 17.1
Stockholm: 17.8
Augsburg: 33.1
Rome: 42.1
Barcelona: 40.7
Total: 30.3
17.1
17.8
33.1
42.1
40.7
38.5
64.5
Helsinki (range): 4-53
Stockholm (range): 0-57
Augsburg (range): 7-71
Rome (range): 15-91
Barcelona (range): 6-194
Total (range): 0-1 94
NR
NR
NR
NR
NR
NR
75th: 60.7
6.2.7.2. Controlled Human Exposure Studies
Several controlled human exposure studies were included in the 2004 PM AQCD (U.S. EPA,
2004, 056905) which evaluated markers of systemic inflammation following exposure to PM. Salvi
et al. (1999, 058637) exposed 15 healthy volunteers (21-28 yr) for 1 h to DE (300 ug/m3 particle
concentration) and observed a significant increase in neutrophils in peripheral blood 6 h post-
exposure compared with filtered air control. However, Ohio et al. (2003, 087363) reported no
changes in plasma cytokine levels (e.g., IL-6 and TNF-a), WBC count, or CRP 0 or 24 h following a
2-h exposure to PM2.5 CAPs (120 ug/m3). Gong et al. (2003, 042106) did not observe any effect of
PM2.5 CAPs (174 ug/m3) on serum amyloid A, while Frampton (2001, 019051) reported no change in
leukocyte activation following exposure to a low concentration (10 ug/m ) of UF carbon. The results
of studies published since the completion of the 2004 PM AQCD (U.S. EPA, 2004, 056905) are
discussed below.
December 2009
6-44
-------
CAPS
Several controlled human exposure studies have reported no change in plasma CRP levels
0-24 h after exposure to UF (avg concentration 50-100 ug/m3), PM25 (avg concentration 190 ug/m3),
or PM10_25 (avg concentration 89 ug/m3) CAPs (Gong et al., 2008, 156483; Graff et al, 2009,
191981; Mills et al., 2008, 156766; Samet et al., 2009, 191913). In a study of exposures to PM2.5
CAPs (200 ug/m3), Gong et al. (2004, 087964) observed increased peripheral basophils 4 h
following a 2-h exposure in a group of healthy older adults, which provides limited evidence of a
CAPs-induced systemic inflammatory response.
Urban Traffic Particles
In a recent investigation of controlled exposures (24 h) to urban traffic particles, Brauner et al.
(2008, 191966) observed no effect of PM concentration (avg PM2.5 concentration 10.5 ug/m3) on
markers of inflammation including CRP, IL-6 and TNF- a in peripheral venous blood.
Diesel Exhaust
Recent controlled human exposure studies have observed no effect of DE on plasma CRP
concentrations or peripheral blood cell counts (Blomberg et al., 2005, 191991; Carlsten et al., 2007,
155714; Mills et al., 2005, 095757; Mills et al., 2007, 091206; Tornqvist et al., 2007, 091279).
Mills et al. (2005, 095757) found no effect of DE (300 ug/mj) on serum IL-6 or TNF-a among
healthy adult volunteers 6 h after exposure. However, as reported by Tornqvist et al. (2007, 091279).
a significant increase in these cytokines was observed 24 h after exposure. Although the
physiological significance of this finding is unclear, this study does provide evidence of a mild
systemic inflammatory response induced by exposure to DE. In an effort to better understand the
inflammatory response of exposure to PM, Peretz et al. (2007, 156853) conducted a pilot study in
which gene expression in peripheral blood mononuclear cells (PBMCs) of healthy human volunteers
was evaluated following a 2-h controlled exposure to DE (200 ug/m3 PM2 5). Adequate RNA samples
for microarray analysis from both pre- and 4 h post-exposure to filtered air and DE were available in
4 of the 11 subjects enrolled. The authors found differential expression of 10 genes involved in the
inflammatory response when comparing DE exposure (8 upregulated, 2 downregulated) to filtered
air. Two participants had paired samples from 20 h post-exposure which were adequate for analysis.
At this time point, DE was associated with 4 differentially expressed genes (1 upregulated, 3
downregulated). However, this study is limited by a small sample size with limited statistical power.
Wood Smoke
Barregard et al. (2006, 091381) recently reported an increase in serum amyloid A at 0, 3, and
20 h following a 4-h exposure to wood smoke (PM2 5 concentrations of 240-280 ug/m3) among a
group of 13 healthy adults (20-56 yr).
Model Particles
Frampton et al. (2006, 088665) evaluated the effect of varying concentrations (10-50 ug/m3) of
UF EC on blood leukocyte expression of adhesion molecules in healthy and asthmatic adults.
Healthy subjects (n = 40) were exposed for 2 h to filtered air and UF EC under three separate
protocols: 10 ug/m3 at rest (n = 12), 10 and 25 ug/m3 with intermittent exercise (n = 12), and
50 ug/m3 with intermittent exercise (n = 16). Asthmatics (n = 16) were exposed at a single
concentration (10 ug/m3) for 2 h with intermittent exercise. Leukocyte expression of surface markers
were quantified using flow cytometry on peripheral venous blood samples collected prior to and
immediately following exposure, as well as at 3.5 and 21 h post-exposure. Among healthy resting
adults, UF EC exposure at a concentration of 10 ug/m3 had no effect on blood leukocytes. The
expression of adhesion molecules CD54 and CD 18 on monocytes, and CD 18 on PMNs was shown
December 2009 6-45
-------
to decrease with UF EC exposure in healthy exercising adults. In exercising asthmatics, expression
of CD lib on monocytes and eosinophils, as well as CD54 on PMNs were reduced following
exposure to UF EC. In both asthmatics and healthy adults, a UF EC-induced decrease in eosinophils
and basophils was observed 0-21 h following exposure. Although the clinical significance of these
findings is unclear, the authors concluded that their findings of UF EC-induced changes in leukocyte
distribution and expression were consistent with increased retention of leukocytes in the pulmonary
vasculature, which may be due to an increase in pulmonary vasoconstriction. Other studies have
reported no changes in plasma cytokine levels, peripheral blood counts, or CRP following exposure
to ZnO or UF EC (Beckett et al., 2005, 156261; Routledge et al., 2006, 088674).
Summary of Controlled Human Exposure Study Findings for Systemic
Inflammation
New studies involving controlled exposures to various particle types have provided limited
and inconsistent evidence of a PM-induced increase in markers of systemic inflammation.
6.2.7.3. lexicological Studies
There has been limited evidence that enhanced hematopoiesis may occur in animals exposed
to PM. Two studies in the 2004 PM AQCD (U.S. EPA, 2004, 056905) provided support for this
effect, with one study measured stimulated release of PMNs from bone marrow and another
examined peripheral blood PMN and blood cell counts; however, one study did not find associations
between CAPs and peripheral blood counts. Thus, it was concluded that consistent evidence of
PM-induced hematopoiesis remained to be demonstrated. However, in a study of humans exposed to
biomass burning during the 1997 Southeast Asian smoke-haze episodes, PMi0 demonstrated the best
relationship with blood PMN band cell counts expressed as a percentage of total PMN at lag 0 and 1,
indicating a relatively quick response (Tan et al., 2000, 002304).
CAPs
A 2-day CAPs study employing SH rats did not report increased WBCs 18-20 h post-exposure
(Kodavanti et al., 2005, 087946). A study utilizing fine and/or UF CAPs demonstrated decreased
WBCs in SH rats 18 h after a 2-day (6 h/day) exposure (Kooter et al., 2006, 097547). The decrease
was largely attributable to lowered neutrophils in the fine CAPs-exposed rats and reduced
lymphocytes in the fine+UF CAPs-exposed animals.
Model Particles
In a study of fine and UF CB particles (WKY rats; 7 h; mean mass concentration 1,400 and
1,660 ug/m3 for fine and UF CB, respectively; mean number concentration 3.8><103 and 5.2><104
particles/cm3, respectively), only UF CB induced elevated blood leukocytes at 0 and 48 h
post-exposure compared to the control rats and no effect was observed at 16 h (Gilmour et al., 2004,
054175). In another study of SH rats exposed to UF carbon particles for 24 h (mass concentration
172 ug/m3; mean number concentration 9.0xl06 particles/cm3), the percent neutrophils and
lymphocytes were increased on the first recovery day, but not the third day (Upadhyay et al., 2008,
159345); CRP was unchanged. In another study, blood neutrophils were decreased in SH rats
exposed to UF CB for 6 h and no effects were observed in old F344 rats (Elder et al., 2004,
055642). Plasma IL-6 levels were unchanged (Elder et al., 2004, 055642).
Coal Fly Ash
Smith et al. (2006, 110864) examined the hematology parameters in SD rats following a 3-day
inhalation exposure (4 h/day) to coal fly ash (mean mass concentration 1,400 ug/m3) and reported
increased blood neutrophils and reduced blood lymphocytes at 36 h but not 18 h post-exposure.
December 2009 6-46
-------
Intratracheal Instillation
Elevated systemic IL-6 and TNF-a levels were observed following PMi0 instillation in mice
(details provided in Section 6.2.8.3) (Mutlu et al., 2007, 121441). IL-6 was decreased with PM
exposure in macrophage-depleted mice, indicating that some of the IL-6 release originated from
macrophages. For mice (male C57B1/6J) exposed to PMi0_2.5 derived from coal fly ash (200 ug),
increased plasma IL-6 levels were only observed in animals that also received 100 ug of LPS
(Finnerty et al., 2007, 156434) and this response was not observed with LPS alone, indicating a role
for PM 10.2.5.
Summary of lexicological Study Findings for Systemic Inflammation
Overall, these studies provide evidence of time-dependent responses of systemic inflammation
induced by PM exposure. Alterations in WBCs have been reported generally as elevations
immediately (0 h) or <36 h post-exposure and no change or reductions are noted from 18-24 h.
6.2.8. Hemostasis, Thrombosis and Coagulation Factors
The 2004 PM AQCD (U.S. EPA, 2004, 056905) presented limited and inconsistent evidence
from epidemiologic, controlled human exposure, and toxicological studies of PM-induced changes in
blood coagulation markers. The body of scientific literature investigating hemostatic effects of PM
has grown significantly since the publication of the 2004 PM AQCD (U.S. EPA, 2004, 056905). with
a limited number of epidemiologic studies demonstrating consistent increases in von Willebrand
factor (vWf) associated with PM and less consistent associations with fibrinogen. Recent controlled
human exposure and toxicological studies have also observed changes in blood coagulation markers
(e.g., fibrinogen, vWf, factor VII, t-PA) following exposure to PM. However, the findings of these
studies are somewhat inconsistent, which may be due in part to differences in the post-exposure
timing of the assessment.
6.2.8.1. Epidemiologic Studies
Several studies investigating the association of short-term fluctuations in PM concentration
with markers of coagulation (e.g., blood viscosity and fibrinogen) were included in the 2004 PM
AQCD (U.S. EPA, 2004, 056905). These preliminary studies offered limited support for mechanistic
explanations of the associations of PM concentration with heart disease outcomes. New studies,
published since 2002, are reviewed in this section. Only vWF and fibrinogen were measured in
enough comparable studies to allow the consistency of findings to be evaluated across epidemiologic
studies. Other markers of coagulation studied included D-dimer, prothombin time, Factor VII/VIII
and tPA.
Liao et al. (2005, 088677) used a cross-sectional study to examine the association between
short-term increases in air pollutant concentrations (mean PMi0, NO2, CO, SO2, and O3 over the
previous 3 days) and several plasma hemostatic markers (fibrinogen, factor VIII-C, vWF, albumin).
Study subjects were middle aged participants in the ARIC (Atherosclerosis Risk in Communities)
study (n = 10,208), and were residents of Washington County, MD, Forsyth County, NC, selected
suburbs of Minneapolis, MN, or Jackson, MS. Each 12.8 ug/m3 increase in the mean PMi0
concentration 1 day before the health measurements were made was associated with a 3.93%
increase in vWF (95% CI: 0.40-7.46) among diabetics, but not among non-diabetics (-0.54% [95%
CI: -1.68 to 0.60]). Each 12.8 ug/m3 increase in the mean PMi0 concentration 1 day before the health
measurements were made was also associated with a 0.006 g/dL decrease in serum albumin
(95% CI: -0.012 to 0.000) among those with cardiovascular disease (CVD), but not among those
without CVD (0.029 g/dL increase [95% CI: -0.004 to 0.062]). The mean CO concentration on the
previous day was also associated with a significant decrease in serum albumin. The authors reported
significant curvilinear associations between PMi0 and factor VIII-C, which may indicate a threshold
effect. Similar curvilinear associations were observed between O3 with fibrinogen, and vWF, and
SO2 with factor VIII-C, WBC, and serum albumin (Liao et al., 2005, 088677). No significant
associations with fibrinogen and PM10 or gaseous pollutants were observed.
December 2009 6-47
-------
In the European multicity study described in Section 6.2.7.1, Ruckerl et al. (2007, 156931)
found that each 13.5 (ig/m3 increase in the mean PMi0 concentration over the previous 5 days was
associated with a 0.6% increase in the arithmetic mean fibrinogen level (95% CI: 0.1-1.1). Further
these investigators found that promoter polymorphisms within FGA and FGB modified the
association of 5-day avg PMi0 concentration with plasma fibrinogen levels (Peters et al., 2009,
191992). This association was 8-fold higher among those homozygous for the minor allele genotype
of FGB rs!800790 compared with those homozygous for the major allele.
Several smaller studies have been conducted in the U.S. and Canada. Delfino et al. (2008,
156390) measured fibrinogen and D-dimer in blood of subjects who resided at two downtown Los
Angeles nursing homes. As described in Section 6.2.7.1, measurements were made over a period of
12 wk and subjects were >65 yr old with a history of coronary artery disease. These markers were
not associated with the broad array PM metrics studied (e.g., PM0.25, PM0.25-2.5, PMi0_2.5, EC, OC,
primary OC, BC). In the study of 92 Boston residents with type 2 diabetes described previously,
O'Neill et al. (2007, 091362) found that increases in mean PM2 5 and BC concentration were
associated with vWF concentrations for all moving averages examined (1-6 days). Reidiker et al.
(2004, 056992) reported that in-vehicle PM2.5 was associated with increased vWF over the next
10-14 h among nine police troopers. Sullivan et al. (2007, 100083) did not observe associations with
fibrinogen, or D-dimer in individuals with or without COPD. Red blood cells (RBCs), platelets, nor
blood viscosity were associated with PM2 5 concentration in a panel study of 88 non-smoking elderly
subjects residing in the Salt Lake City, Ogden and Provo metropolitan area of Utah (Pope et al.,
2004, 055238). Although Zeka et al. (2006, 157177) did not observe an association with CRP in the
analysis of the Normative Aging Study population in Boston (Section 6.3.7.1), increased fibrinogen
level was associated with increases in the number of particles/cm3 over the previous 48 h and 1 wk,
and an incremental increase in BC concentration over the previous 4 wk. There were no consistent
findings for lagged PM2.5 or sulfates (Zeka et al., 2006, 157177).
Several studies of coagulation markers were conducted outside the U.S. and Canada. In a
study of healthy individuals in Taiwan, associations were observed for PM25, PMi0, nitrate, and
SO4 ~ concentrations with fibrinogen and plasminogen activator fibrinogen inhibitor-1 (PAI-1)
(Chuang et al., 2007, 091063). In a large cross-sectional study of healthy subjects in Tel-Aviv,
Steinvil et al. (2008, 188893) examined fibrinogen collected as part of routine health examinations
for 3,659 individuals. No significant associations were found between pollutant levels (lagged
1-7 days) and fibrinogen. Finally, Baccarelli and colleagues reported associations between PMi0 and
prothrombin time among normal subjects (Baccarelli et al., 2007, 090733).
Summary of Epidemiologic Study Findings for Hemostasis, Thrombosis and
Coagulation
The most commonly measured markers of coagulation in the studies reviewed were fibrinogen
and vWF. Associations of PM10 (Liao et al., 2005, 088677) and PM2.5 (O'Neill et al., 2007, 091362;
Riediker et al., 2004, 056992) with increased vWF were observed across the limited number of
studies examining this association among both diabetics and healthy state troopers state troopers
(Liao et al., 2005, 088677: Riediker et al., 2004, 056992). Results for fibrinogen were not
consistent across epidemiologic studies. Positive associations with fibrinogen were reported in older
adults residing in Boston (Zeka et al., 2006, 157177) and in the multicity European study of MI
survivors. Liao et al. (2005, 088677) in a population based multicity study and Sullivan et al. (2007,
100083) did not observe associations of PMip or PM25 with fibrinogen. Several other markers have
been examined (e.g., D-dimer, prothrombin time), but not in adequate numbers of studies to allow
comparisons across epidemiologic studies. Mean and upper percentile concentrations of the studies
discussed in this section are listed in Table 6-6.
6.2.8.2. Controlled Human Exposure Studies
In two separate studies conducted by Ohio and colleagues, controlled exposures (2 h) to fine
CAPs (Chapel Hill, NC) at concentrations between 15 and 350 ug/m3 were shown to increase blood
fibrinogen 18-24 h following exposure among healthy adults (Ohio et al., 2000, 012140; Ohio et
al., 2003, 087363). Increases in blood fibrinogen or factor VII would suggest an increase in blood
coagulability, which could result in an increased risk of coronary thrombosis. However, a similar
December 2009 6-48
-------
study conducted in Los Angeles observed a PM2.5 CAPs-induced decrease in factor VII blood levels
in healthy subjects and found no association between PM2.5 CAPs and blood fibrinogen among
healthy and asthmatic volunteers (Gong et al., 2003, 042106). Since the publication of the 2004 PM
AQCD (U.S. EPA, 2004, 056905). several new controlled human exposure studies have evaluated
the effects of PM on blood coagulation markers.
CAPs
Two studies of controlled human exposures to Los Angeles CAPs among older adults with
COPD (PM2.5 CAPs) and adults with and without asthma (UF CAPs) reported no significant
association between exposure and blood coagulation markers at 0, 4, or 22 h post-exposure (Gong et
al., 2004, 087964: 2008, 156483). Graff et al. (2009, 191981) observed a decrease in the
concentration of D-dimer of marginal statistical significance in healthy adults (11.3% decrease per
10 ug/m3, p = 0.07) following exposure to PMi0_2.5 CAPs (89 ug/m3). At 20 h post-exposure, levels of
tPA in plasma were shown to decrease by 32.9% from baseline per 10 ug/m3 increase in CAPs
concentration. No other markers of hemostasis or thrombosis were affected by exposure to PMi0_2.5
CAPs. However, in a similar study from the same laboratory, Samet et al. (2009, 191913) reported a
statistically significant increase in D-dimer immediately following, as well as 18 h, after a 2-h
exposure to UF CAPs (49.8 ug/m3; 120,662 particles/cm3) in a group of healthy adults (18-35 yr).
Plasma concentrations of PAI-1 were also reported to increase 18 h after exposure to UF CAPs,
although this increase was not statistically significant (p = 0.1). No changes in fibrinogen, tPA, vWF,
plasminogen, or factor VII were observed. The finding of an increase in D-dimer following exposure
to UF CAPs provides potentially important information in elucidating the relationship between
elevated concentrations of PM and cardiovascular morbidity and mortality observed in
epidemiologic studies. Whereas many coagulation markers provide evidence of an increased
potential to form clots (e.g., an increase in fibrinogen or a decrease in tPA), D-dimer is a degradation
product of a clot that has formed.
Urban Traffic Particles
In a study of controlled 24-h exposures to urban traffic particles (avg PM2.5 concentration
10.5 ug/m3) among 29 healthy adults, Brauner et al. (2008, 191966) did not observe any particle-
induced change in plasma fibrinogen, factor VII, or platelet count after 6 or 24 h of exposure.
Similarly, Larsson et al. (2007, 091375) observed no change in PAI-1 or fibrinogen in peripheral
blood of healthy adult volunteers 14 h after a 2-h exposure to road tunnel traffic with a PM2.5
concentration of 46-81 ug/m3.
Diesel Exhaust
Mills and colleagues have recently demonstrated a significant effect of DE (particle
concentration 300 ug/m3) on fibrinolytic function both in healthy men (n = 30) and in men with
coronary heart disease (n = 20) (Mills et al., 2005, 095757; 2007, 091206). In both groups of
volunteers, bradykinin-induced release of tPA was observed to decrease 6 h following exposure to
DE compared to filtered air exposure. The same laboratory did not observe an attenuation of tPA
release 24 h after a 1-h exposure to DE (300 ug/m3) in a group of health adults (Tornqvist et al.,
2007, 091279). or observe any change in markers of hemostasis or thrombosis 6 or 24 h following
DE exposure at the same particle concentration among a group of older adults with COPD
(Blomberg et al., 2005, 191991). Carlsten et al. (2007, 155714) conducted a similar study involving
exposure of healthy adults to DE with a PM2.5 concentration of 200 ug/m3. Although the authors
observed an increase in D-dimer, vWF, and platelet count 6 h following exposure to DE, these
increases did not reach statistical significance. In a subsequent study with a similar study design, the
same laboratory found no effect of a 2-h exposure to DE (100 and 200 ug/m3 PM25) on
prothrombotic markers in a group (n = 16) of adults with metabolic syndrome (Carlsten et al., 2008,
156323). The authors postulated that the lack of significant findings could be due to a relatively
small sample size. In addition, Carlsten et al. (2007, 155714; 2008, 156323) exposed subjects at rest
December 2009 6-49
-------
while Mills et al. (2005, 095757) exposed subjects to a higher concentration (300 ug/m3) with
intermittent exercise. A more recent study of DE which exposed healthy adults to a slightly higher
particle concentration (330 ug/m3) evaluated the effect of DE on thrombus formation using an ex
vivo perfusion chamber (Lucking et al., 2008, 191993). Thrombus formation, as well as in vivo
platelet activation, was observed to significantly increase 2 h following exposure to DE relative to
filtered air, thus providing some evidence of a potential physiological mechanism which may explain
in part the associations between PM and cardiovascular events observed in epidemiologic studies.
Wood Smoke
Barregard et al. (2006, 091381) recently evaluated the effect of wood smoke on markers of
coagulation, inflammation, and lipid peroxidation. Subjects (n = 13) were healthy males and females
(20-56 yr) and were exposed for 4 h to PM2.s concentrations of 240-280 ug/m3. The authors reported
an increase in the ratio of factor VIII/vWF, which is an indicator of an increased risk of venous
thromboembolism, at 0, 3, and 20 h following exposure to wood smoke.
Model Particles
Routledge et al. (2006, 088674) did not observe any changes in fibrinogen or D-dimer
following a 1-h exposure to UF carbon among a group of resting healthy older adults and older
adults with stable angina. Similarly, Beckett et al. (2005, 156261) found no changes in hemostatic
markers (e.g., factor VII, fibrinogen, and vWF) following exposure to UF and fine ZnO (500 ug/m3).
Summary of Controlled Human Exposure Study Findings for Hemostasis,
Thrombosis and Coagulation
Taken together, these new studies have provided some additional evidence that short-term
exposure to PM at near ambient levels may have small, yet statistically significant effects on
hemostatic markers in healthy subjects or patients with coronary artery disease.
6.2.8.3. lexicological Studies
In general, the limited toxicological studies reviewed in the 2004 PM AQCD (U.S. EPA, 2004,
056905) reported positive and negative findings for plasma fibrinogen levels or other factors
involved in the coagulation cascade. Rats exposed to New York City CAPs did not have any
exposure-related effects on any measured coagulation markers (Nadziejko et al., 2002, 050587).
whereas rats exposed to a high concentration of ROFA demonstrated increased plasma fibrinogen
(Kodavanti et al., 2002, 025236).
CAPs
APM2.5 CAPs exposure conducted for 2 days (4 h/day; mean mass concentration
144-2,758 ug/m3; 8-10/2001; RTP, NC) in SH rats induced plasma fibrinogen increases (measured
18-20 h post-exposure) in 5 of 7 separate studies (Kodavanti et al., 2005, 087946). Fibrinogen was
not different from the air control group on the two days with the highest CAPs concentrations (1,129
and 2,758 ug/m3), indicating that the response was likely not attributable to mass alone.
In SH rats exposed to PM2.5 CAPs for 6 h in one of three locations in the Netherlands (mean
mass concentration range 270-2,400; 335-3,720; and 655-3,660 ug/m3), plasma fibrinogen was
increased 48 h post-exposure when all CAP-exposed animals were combined in the analysis (Cassee
et al., 2005, 087962). In WKY rats pre-exposed to O3 (8 h; 1,600 ug/m3) and CAPs for 6 h, increases
in RBCs, hemoglobin, and hematocrit were observed 2 days after CAPs exposure. For SH rats
exposed to CAPs only, decreased mean corpuscular hemoglobin concentration were reported.
December 2009 6-50
-------
A similar study conducted by the same group (Kooter et al., 2006, 097547) reported no
changes in plasma fibrinogen measured 18 h after a 2-day exposure (6 h/day) to PM2.5 or PM2.5+UF
CAPs (mean mass concentration range 399.0-1,067.5 and 269.0-555.8 ug/m , respectively; 1/2003-
4/2004). However, elevated vWF was observed in SH rats exposed to the highest concentration of
PM2.5 CAPs. Decreases in mean corpuscular volume (MCV), and elevations in mean platelet volume
(MPV) and mean platelet component (MPC) were reported in SH rats 18 h following a 2-day
exposure to PM2.5+UF CAPs in a freeway tunnel.
Traffic-Related Particles
Plasma fibrinogen levels were elevated 18 h following a single 6-h exposure to on-road
highway aerosols when groups of rats pretreated with saline or influenza virus were combined
(i.e., there was a significant effect of particles) (Elder et al., 2004, 087354).
Model Particles
The coagulation effects of inhaled UF CB at a concentration of 150 ug/m3 (number count not
provided) for 6 h were evaluated 24 h post-exposure in two aged rat models (11-14 mo SH and 23
mo F344), some of which received LPS via intraperitoneal injection prior to particle exposure (Elder
et al., 2004, 055642). LPS has been shown to induce the expression of molecules involved in
coagulation, inflammation, oxidative stress, and the acute-phase response. In those animals only
exposed to CB, SH rats demonstrated increased thrombin-anti-thrombin complexes (TAT) and
decreased fibrinogen. For F344 rats, TAT complexes and fibrinogen were elevated only in those that
received LPS and CB. Whole-blood viscosity was not altered in either rat strain with particle
exposure.
In another study of SH rats exposed to UF carbon particles for 24 h (mass concentration 172
ug/m3; mean number concentration 9.0xl06 particles/cm ), the number of RBCs and platelets and
hematocrit percent, were unchanged 1 and 3 days following exposure (Upadhyay et al., 2008,
159345). Fibrinogen levels were similar in both air and UF carbon-exposed groups. However,
mRNA expression of PAI-1 and TF in lung homogenates (but not in heart) was increased on
recovery day 3 after exposure. A study of similar design that employed SH rats did not report any
effect on plasma fibrinogen 4 or 24 h following UF carbon exposure (mass concentration 180 ug/m3;
mean number concentration 1.6><107 particles/cm3) (Harder et al., 2005, 087371). Similarly, clotting
factor Vila and thrombomodulin, PAI-1, and tPA mRNA expression were not affected by UF carbon
exposure at 24 h post-exposure.
Coal Fly Ash
One study that employed coal fly ash (mean mass concentration 1,400 ug/m3; 4 h/day><3 days)
demonstrated increases in hematocrit and MCV in SD rats at 36 h but not 16 h post-exposure (Smith
etal, 2006, 110864).
Intratracheal Instillation
Mutlu et al. (2007, 121441) used a PM10 sample collected from Dusseldorf, Germany, in mice
(C57BL/6) with and without the gene coding for IL-6. The authors report using a moderate IT
instillation dose (10 ug/mouse; roughly equivalent to 400-500 ug/kg); the PM sample had previously
been characterized as having significant Fe, Ni, and V content (Upadhyay et al., 2003, 097370). In
C57BL/6 mice, the Dusseldorf PM shortened bleeding (32%), prothrombin (13%), and activated
partial thromboplastin (16%) times and increased platelet count, fibrinogen, and Factors II, VIII, and
X activities 24 h following exposure. The authors further demonstrated accelerated coagulation by a
reduction in the left carotid artery occlusion time (experimentally-derived by direct application of
FeCl3). Additional experiments demonstrated that IL-6"7" or macrophage-depleted mice showed
dramatically attenuated effects of PMi0 on hemostatic indices, thrombin generation, and occlusion
December 2009 6-51
-------
time. In IL-6"7" mice, there was no change in total cell counts or differentials in BALF compared to
the wild-type mice, despite the lack of IL-6. In contrast, the model of macrophage depletion had
reduced levels of macrophages and IL-6 in BALF, following PM exposure. These studies suggest
that instillation of Dusseldorf PMi0 activates clotting through an alveolar macrophage-dependent
release of IL-6; however, other factors may also be involved in the prothrombotic response
(i.e., activation of neutrophils, other inflammatory cells, or alterations in the levels of other
cytokines).
In a study employing PMi0_2.5 collected from six European locations with contrasting traffic
profiles, fibrinogen increases were observed in SH rats exposed to 10 mg/kg via IT instillation at 24
h post-exposure and similar responses were observed with PM2.5 (Gerlofs-Nijland et al, 2007,
097840). PMio_2.5 and PM2.5 samples from Prague or Barcelona administered intratracheally to SH
rats (7 mg/kg) resulted in elevated plasma fibrinogen levels 24 h post-exposure compared to rats
instilled with water (Gerlofs-Nijland et al., 2009, 190353). No changes were observed in vWF for
whole particle suspensions, but Barcelona PMi0_2.5 organic extract induced greater levels of vWF
than Barcelona PMio-2.5-
Summary of lexicological Study Findings for Hemostasis, Thrombosis and
Coagulation
Increases in coagulation and thrombotic markers were observed in some studies of rats or mice
exposed to PM. Plasma TAT complexes were increased in CB-exposed SH rats and shortened
bleeding, prothrombin, and activated partial thromboplastin times were observed in mice exposed
via IT instillation to PMi0. Furthermore, the latter study also reported increased levels of Factors II,
VIII, and X activities in mice. Another study demonstrated increased vWF in response to PM2.5
CAPs. As for plasma fibrinogen, these studies provide some evidence that increased levels are
observed 18-48 h post-exposure to PM, although one study reported no change and another reported
a decrease in this biomarker. Alterations in platelet measurements have also been observed with PM
exposure, including increased platelet number, mean platelet volume, and mean platelet component.
The toxicological results of RBC-related measurements are limited and inconsistent following PM
exposure, which may be attributable to different exposure protocols, time of analysis, or rat strain.
6.2.9. Systemic and Cardiovascular Oxidative Stress
Very little information on systemic oxidative stress associated with PM was available for
inclusion in the 2004 PM AQCD (U.S. EPA, 2004, 056905). However, recent epidemiologic studies
have provided consistent evidence of PM-induced increases in markers of systemic oxidative stress
including plasma thiobarbituric acid reactive substances (TEARS), CuZn-super oxide dismutase
(SOD), 8-oxo-7-hydrodeoxyguanosine (8-oxodG), and total homocysteine. This is supported by a
limited number of controlled human exposure studies that observed PM-induced increases in free-
radical mediated lipid peroxidation, as well as upregulation of the DNA repair gene hOGGl. In
addition, recent toxicological studies have demonstrated an increase in cardiovascular oxidative
stress following PM exposure in rats.
6.2.9.1. Epidemiologic Studies
No studies of markers of oxidative stress were reviewed in the 2004 PM AQCD (U.S. EPA,
2004, 056905). Since 2002, numerous studies have examined whether short-term increases in mean
PM concentrations are associated with changes in systemic markers of oxidative stress.
In an analysis of the randomized trial of omega-3 fatty acid supplementation in Mexico City
nursing home residents described previously (Section 6.2.1.1), Romieu et al. (2008, 156922)
investigated the effect of this intervention on markers of systemic oxidative stress (Cu/Zn SOD
activity, LPO in plasma and GSH in plasma). A significant decrease of Cu/Zn SOD was associated
with a 10 ug/m3 increase of PM2.5 in both groups (Fish oil: (3 = -0.17 [SE = 0.05], p = 0.002; Soy oil:
(3 = -0.06 [SE = 0.02], p <0.001). A decrease in GSH was associated with a 10 ug/m3 increase in
PM2.5 in the fish oil group ((3 = -0.09 [SE = 0.04], p = 0.017).
December 2009 6-52
-------
Two studies evaluated plasma homocysteine levels in relation to PM. Baccarelli et al. (2007,
091310) investigated fasting and post-methionine load total homocysteine (tHcy) among 1,213
normal subjects in Lombardia, Italy. Plasma homocysteine is a risk factor for CVD and a marker for
oxidative stress. Among smokers, average PMi0 level during the 24 h preceding the measurement
was associated with 6.3% (95% CI: 1.3-11.6) and 4.9% (95% CI: 0.5-9.6) increases in fasting and
post-methionine load tHcy, respectively. No associations were observed among non-smokers. Park
et al. (2008, 156845) investigated the association of BC, OC, SO42~ and PM2.5 with tHcy among 960
male participants of the Normative Aging Study. Effect modification by folate and vitamins B6 and
B12 was also examined. BC and OC were associated with increases in tHcy and associations were
more pronounced in those with lower plasma folate and vitamin B12.
In smaller studies with 25-50 healthy or diseased participants, several markers of oxidative
stress have been associated with PM size fractions or components. These associations include
TEARS with 24-h PMi0 (Liu et al., 2006, 192002); Cu/Zn-SOD with several PM metrics (e.g., UF,
PMio-zs, EC, OC, BC and PNC) (Delfino et al., 2008, 156390): PM2.5, BC, V and Cr with plasma
proteins (Sorensen et al., 2003, 157000); DNA damage assessed by 8-oxodG in lymphocytes
(Sorensen et al., 2003, 157000). and 8-OHdG with sulfates (Chuang et al., 2007, 091063). In
addition, a cross-sectional study of children (10-18 yr) in Iran showed an association of PMi0 with
oxidized LDL (oxLDL), malondialdehyde (MDA) and conjugated diene (CDE) (Kelishadi et al.,
2009, 191960).
Summary of Epidemiologic Study Findings for Systemic and Cardiovascular
Oxidative Stress
Oxidative stress responses measured by one or more markers (plasma tHcy, CuZn-SOD,
TEARS, 8-oxodG, oxLDL and MDA) have been consistently observed (Baccarelli et al., 2007,
091310: Chuang et al., 2007, 091063: Delfino et al., 2008, 156390: Kelishadi et al., 2009, 191960:
Liu et al., 2007, 156705: Romieu et al., 2008, 156922: Sorensen et al., 2003, 157000). In addition,
a series of analyses examining the modification the PM-HRV association by genetic polymorphisms
related to oxidative stress has provided insight into the possible mechanisms of CVD observed in
association with PM concentrations (Section 6.2.1.1). Mean and upper percentile concentrations of
the epidemiologic studies of systemic oxidative stress are included in Table 6-6.
6.2.9.2. Controlled Human Exposure Studies
Urban Traffic Particles
Brauner et al. (2007, 091152) recently investigated the effect of urban traffic particles on
oxidative stress-induced damage to DNA. Healthy adults (20-40 yr) were exposed to low
concentrations of urban traffic particles as well as filtered air for periods of 24 h, with and without
two 90-min periods of exercise. Exposures took place in an exposure chamber above a busy road
with high traffic density in Copenhagen. Non-filtered air was pumped into the chamber from above
the street, with avg PM25 and PMi0_2.5 mass concentrations of 9.7 ug/m3 and 12.6 ug/m3,
respectively. The UF/PM2.5 (6-700 nm) particle number concentration was continuously monitored
throughout the exposure (avg PNC 10,067 particles/cm3). The PM2.5 fraction was rich in sulfur, V,
Cr, Fe, and Cu. PBMCs were isolated from blood samples collected at 6 and 24 h. DNA damage, as
measured by strand breaks (SB) and formamidopyrimidine-DNA glycosylase (FPG) sites, was
evaluated using the Comet assay. The activity and mRNA levels of the DNA repair enzyme
7,8-dihydro-8-oxoguanine-DNA glycosylase (OGG1) were also measured. The authors observed
increased levels of DNA strand breaks and FPG sites following 6 and 24 h of exposure to PM. Using
a mixed-effects regression model, the particle concentration at the 57 nm mode was found to be the
major contributor of these measures of DNA damage. The results of this study suggest that short-
term (6-24 h) exposure to ambient levels of UFPs cause systemic oxidative stress resulting in
damage to DNA.
December 2009 6-53
-------
Diesel Exhaust
Tornqvist et al. (2007, 091279) reported an increase in plasma antioxidant capacity in a group
of healthy volunteers 24 h after a 1-h exposure to DE with a particle concentration of 300 ug/m3. The
investigators suggested that systemic oxidative stress occurring following exposure may have caused
this up-regulation in antioxidant defense. Peretz et al. (2007, 156853) observed some significant
differences in expression of genes involved in oxidative stress pathways between exposure to DE
(200 ug/m3 PM25) and filtered air. However, the conclusions of this investigation are limited by a
small number of subjects (n = 4).
Wood Smoke
In a controlled human exposure study of controlled exposure to wood smoke, Barregard et al.
(2006, 091381) found an increase in urinary excretion of free 8-iso-prostaglandin2a among healthy
adults (n = 9) approximately 20 h following a 4-h exposure to PM2 5 (mass concentration of
240-280 ug/m3). This finding provides evidence of a PM-induced increase in free-radical mediated
lipid peroxidation. From the same study, Danielsen et al. (2008, 156382) reported an increase in the
mRNA levels of the DNA repair gene hOGGl in peripheral mononuclear cells 20 h after exposure to
wood smoke relative to filtered air.
Summary of Controlled Human Exposure Study Findings for Systemic and
Cardiovascular Oxidative Stress
Based on the results of these studies, it appears that exposure to PM at or near ambient levels
may increase systemic oxidative stress in human subjects.
6.2.9.3. Toxicological Studies
Very little information was available for inclusion in the 2004 PM AQCD (U.S. EPA, 2004,
056905) on oxidative stress in the cardiovascular system. A few new studies have evaluated ROS in
blood or the heart following PM exposure. Some studies have used chemiluminescence (CL), which
is measured using the decay of excited states of molecular oxygen, and may also be prone to artifact.
CAPS
Gurgueira et al. (2002, 036535) measured oxidative stress in SD rats immediately following a
5-h CAPs exposure (PM2.5 mean mass concentration 99.6-957.5 ug/m3; Boston, MA; 7/2000-2/2001)
and reported increased in situ CL in hearts of CAPs-exposed animals. CL evaluated after 1- and 3-h
CAPs exposure did not demonstrate changes from the filtered air group, although a 5-h exposure
resulted in increased CL in hearts. When animals were allowed to recover for 24 h, oxidative stress
returned to control values. To compare potential particle-induced differences in CL, rats were
exposed to ROFA (1.7 mg/m3 for 30 min) or CB (170 ug/m3 for 5 h) and only the ROFA-treated
animals exhibited increased CL in cardiac tissue. Additionally, levels of antioxidant enzymes in the
heart (Cu/Zn-SOD and MnSOD) were increased in CAPs-exposed rats. Individual PM component
concentrations were linked to CL levels in rat heart tissue using separate univariate linear regression
models, with total PM mass, Al, Si, Ti, and Fe having p-values < 0.007 (Gurgueira et al., 2002,
036535). The highest R2 value in the regression analyses was for Al (0.67) and its concentration
ranged from 0.000 to 8.938 ug/m3.
Recently, Rhoden et al. (2005, 087878) tested the role of the ANS in driving CAPs-induced
cardiac oxidative stress in heart tissues of SD rats. At PM2.5 mass concentrations of 700 ug/m3
(Boston, MA), pretreatment with an antioxidant, a Pi-receptor antagonist, or a muscarinic receptor
antagonist attenuated the CL and TEARS effects observed in the heart following a 5-h PM2.5
exposure. The wet/dry ratio (edema) of cardiac tissue also returned to control values in animals
treated with the antioxidant prior to CAPs. These combined results indicate involvement of both the
December 2009 6-54
-------
sympathetic and parasympathetic pathways in the cardiac oxidative stress response observed
following PM exposure.
More recently, a type of irritant receptor, the transient receptor potential vanilloid receptor 1
(TRPV1), was identified as central to the inhaled CAPS-mediated induction of cardiac tissue CL and
TEARS in SD rats (Ghelfi et al., 2008, 156468). In these studies (PM2.5 mean mass concentration
218 ug/m3; Boston, MA), capsazapine (aTRPVl inhibitor) abrogated cardiac CL, TEARS, edema,
and QT-interval shortening when measured at the end of the 5-h exposure. These studies provide
some evidence that the ANS may be involved in producing cardiac oxidative stress following
exposure to CAPs. Furthermore, this response could be acting, at least in part, via TRPV receptors.
In WKY rats exposed to PM2.5 CAPs in Japan, relative mRNA expression of HO-1 was
increased in cardiac tissue and was also significantly correlated with the cumulative mass of PM
collected on chamber filters throughout the exposure (Ito et al., 2008, 096823).
Road Dust
A composite of PM2.5 road dust samples obtained from New York City, Los Angeles, and
Atlanta induced cardiac ROS as measured by CL in the low exposure group (306 ug/m3) and TEARS
in the high exposure group (954 ug/m3); thus, the CL and TEARS methods provided different results
for the various source types (Seagrave et al., 2008, 191990).
Gasoline and Diesel Exhaust
Gasoline exhaust exposure also resulted in increased ROS (measured by TEARS) in aortas of
ApoE"7" mice, as discussed in Section 6.2.4.3 (Lund et al., 2009, 180257). Similarly, a 6-h exposure
to gasoline exhaust (PM mass concentration 60 ug/m3, CMD 15-20 nm; MMD 150 nm; CO
concentration 104 ppm, NO concentration 16.7 ppm, NO2 concentration 1.1 ppm, SO2 concentration
1.0 ppm) in SD rats demonstrated increased CL in the heart, but no change in TEARS and the CL
response was not duplicated when the particles were filtered (Seagrave et al., 2008, 191990).
Increased lipid peroxides in the serum of male SH rats exposed to gasoline exhaust (PM mass
concentration 59.1 ug/m3; NO concentration 18.4 ppm; NO2 concentration 0.9 ppm; CO
concentration 107.3 ppm; SO2 concentration 0.62 ppm) was observed following a 1-wk exposure to
gasoline exhaust and this effect was attenuated with particle filtration (Reed et al., 2008, 156903).
An IT instillation study of diesel particles in mice demonstrated increased myocardial MPO activity
12 and 24 h post-exposure to the residual particle component that remained after extraction with
dichloromethane (Yokota et al., 2008, 190109).
Model Particles
Other studies previously presented also demonstrated ROS (via CL) and NT expression (via
ELISA) in the left ventricle with CB exposure (Tankersley et al., 2008, 157043) and oxidative stress
in the systemic microvasculature following TiO2 inhalation (Nurkiewicz et al., 2009, 191961) or
ROFA IT instillation exposure (Nurkiewicz et al., 2006, 088611). Decreased HO-1 mRNA
expression in hearts of SH rats exposed to UF carbon particles was observed 3 days following
exposure (Upadhyay et al., 2008, 159345) and there was a trend toward increased HO-1 mRNA
expression 1 day post-exposure.
Summary of lexicological Study Findings for Systemic and Cardiovascular
Oxidative Stress
When considered together, the above studies provide evidence that PM exposure results in
oxidative stress as measured in cardiac tissue by CL, TEARS, HO-1 mRNA expression, and NT
expression. However, the PM concentration/dose and method of ROS measurement could also affect
the response. Cardiac oxidative stress may have resulted from PM stimulation of the ANS, although
these studies have only been conducted in one laboratory. Multiple studies from two different
December 2009 6-55
-------
laboratories provide support for vascular oxidative stress as demonstrated in aortas following
gasoline exhaust exposure and in the microvasculature after TiO2 inhalation or ROFA IT exposure.
6.2.10. Hospital Admissions and Emergency Department Visits
The 1996 PM AQCD (U.S. EPA, 1996, 079380) considered just two time-series studies
regarding the association between daily variations in PM levels and the risk of CVD morbidity as
measured by the number of daily hospitalizations with primary discharge diagnoses related to CVD
(Burnett et al., 1995, 077226; Schwartz and Morris, 1995, 046186). In contrast, the 2004 PM
AQCD (U.S. EPA, 2004, 056905) reviewed more than 25 publications relating PM and risk of CVD
hospitalizations. Results from a handful of larger multicity studies were emphasized, with the
greatest emphasis placed on findings from the U.S. National Morbidity, Mortality, and Air Pollution
Study (NMMAPS) (Samet et al., 2000, 010269) and a subsequent reanalysis (Zanobetti and
Schwartz, 2003, 157174). The NMMAPS study evaluated the effect of daily changes in ambient PM
levels on total CVD hospitalizations among elderly Medicare beneficiaries in 14 U.S. cities and
found a ~1% excess risk per 10 (ig/m3 increase in PMi0. The 2004 PM AQCD concluded that these
results, along with those of the other single- and multicity studies reviewed "generally appear to
confirm likely excess risk of CVD-related hospital admissions for U.S. cities in the range of
[0.6-1.7% per 10 (ig/m3] PM10, especially among the elderly" (U.S. EPA, 2004, 056905). The 2004
PM AQCD (U.S. EPA, 2004, 056905) also concluded that there was some evidence from single-city
studies suggesting an excess risk specifically for hospitalizations related to IHD and heart failure.
Furthermore, the 2004 PM AQCD (U.S. EPA, 2004, 056905) found that "insufficient data exist from
the time-series CVD admissions studies [... ] to provide clear guidance as to which ambient PM
components, defined on the basis of size or composition, determine ambient PM CVD effect
potency" (U.S. EPA, 2004, 056905). The key studies reviewed in the 2004 PM AQCD (U.S. EPA,
2004, 056905) on this topic included those by Burnett and colleagues (1997, 084194; 1999, 017269).
Lippman and colleagues (2000, 011938). Ito (2003, 042856). and Peters et al. (2001, 016546).
Recent large studies conducted in the U.S., Europe, and Australia and New Zealand have
confirmed these findings for PMi0, and have also observed consistent associations between PM2.5
and cardiovascular hospitalizations. However, findings from single-city studies have demonstrated
regional heterogeneity in effect estimates. It is apparent from these recent studies that the observed
increases in cardiovascular hospitalizations are largely due to admissions for IHD and CHF rather
than CBVDs (such as stroke). The new literature on hospitalizations and ED visits for cardiovascular
causes published since 2002 is reviewed in the following sections. First, the specific CVD outcomes
captured using ICD codes from hospital admissions databases are discussed. Second, the methods
used in the large and multicity studies are described. For each outcome considered, evidence from
large/multicity studies is emphasized and results from U.S. and Canadian single-city studies are also
discussed. Although the single-city studies may lack statistical power needed to evaluate interactions
and detect some of the subtle effects of air pollution, they inform the interpretation of the
heterogeneous effect estimates that have been observed across North America.
Cardiovascular Disease ICD Codes
When the 2004 PM AQCD (U.S. EPA, 2004, 056905) was written, few studies had evaluated
the link between ambient PM and specific CVD outcomes such as CHF, IHD or ischemic stroke. In
contrast, the majority of recent studies have focused on specific CVD outcomes. This trend is
justified by the fact that the short-term exposure effects of PM may be very different for different
cardiovascular outcomes. For example, given the current putative biological pathways involved in
the acute response to PM exposure, there is no a priori reason why short-term fluctuations in PM
levels would have similar effects on the risk of acute MI, chronic atherosclerosis of the coronary
arteries, and hemorrhagic stroke.
Almost all of the published time-series studies of cardiovascular hospitalizations and ED visits
identified cases based on administrative discharge diagnosis codes as defined by the International
Classification of Disease 9th revision (ICD-9) or 10th revision (ICD-10) (NCHS, 2007, 157194). A
complicating factor in interpreting the results of these studies is the lack of consistency in both
defining specific health outcomes and in the nomenclature used.
December 2009 6-56
-------
Table 6-7. Description of ICD-9 and ICD-10 codes for diseases of the circulatory system.
Description
All Cardiovascular Disease
IHD
Acute Ml
Diseases Of Pulmonary Circulation
CHF
Arrhythmia
CBVD
Ischemic Stroke And Transient Ischemic Attack (TIA)
Hemorrhagic Stroke
Peripheral Vascular Disease (PVD)
ICD-9 Codes
390-459
410-414
410
415-417
428
427
430-438
430-432
433-435
440-448
ICD-10 Codes
IOO-I99
I20-I25
121
I26-I28
ISO
I47, I48, I49
I60-I69
I63
I60-I62
I70-I79
Table 6-7 shows major groups of diagnostic codes used in air pollution studies for diseases of
the circulatory system. The codes ICD-9: 390-459 are frequently used to identify all CVD morbidity.
Note that this definition of CVD includes diseases of the heart and coronary circulation, CBVD, and
peripheral vascular disease. In contrast, the term cardiac disease specifically excludes diseases not
involving the heart or coronary circulation. While this distinction is conceptually straightforward, the
implementation of the definition of cardiac disease in terms of ICD-9 or ICD-10 codes varies among
authors. Even greater heterogeneity can be found among studies in the implementation of definitions
related to CBVD.
Design and Methods of Large and Multicity Hospital Admission and ED Visit
Studies
Recently, multiple research groups in the U.S., Europe, and Australia have created large
datasets to evaluate specific CVD and respiratory endpoints using more detailed and relevant
measures of PM concentration. In the U.S., the MCAPS analyses of Dominici et al. (2006, 088398).
Bell et al. (2008, 156266) and Peng et al. (2008, 156850) are large, comprehensive and informative
studies based on Medicare hospitalization data. Likewise, the Atlanta-based SOPHIA study (Metzger
et al., 2004, 044222; Peel et al., 2005, 056305; Tolbert et al., 2007, 090316) is the largest and most
comprehensive study of U.S. cardiovascular and respiratory ED visits. In Europe, the APHEA
initiative (Le et al., 2002, 023746; Le et al., 2003, 042820) the more recent HEAPSS study (Von et
al., 2005, 088070). and the French PSAS program (Host et al., 2008, 155852; Larrieu et al., 2007,
093031) are similarly noteworthy for their large sample size, geographic diversity, and consideration
of specific CVD and/or respiratory endpoints. These studies contain adequate data to examine
interactions by season and region; the effects of different size fractions, components and sources of
PM; or the effect of PM on susceptible populations. The following section provides a detailed review
of the study design and methods used by each of the large studies. A discussion of the results of each
study can be found later in Section 6.2.10.
MCAPS; Medicare Air Pollution Study
Dominici et al. (2006, 088398) created a database of daily time-series of hospital admission
rates (1999-2002) for a range of cardiovascular and respiratory outcomes among Medicare
beneficiaries aged >65 yr, ambient PM2.5 levels, and meteorological variables for 204 U.S. urban
counties. The specific CVD outcomes considered were: CBVD (ICD-9: 430-438), peripheral
vascular disease (440-448), IHD (410-414, 429), heart rhythm disturbances (426, 427), and CHF
December 2009
6-57
-------
(428). Injuries (800-849) were evaluated as a control outcome. Gaseous and other particulate
pollutant size fractions were not considered.
Data on PM2.5 were obtained from the AQS database of the U.S. EPA. Within each county,
associations between cause-specific hospitalization rates and same-day PM2.5 levels were evaluated
using Poisson regression models controlling for long-term temporal trends and meteorologic
conditions with natural cubic splines. County-specific results were subsequently averaged using
Bayesian hierarchical models. In addition to evaluating single-day lags, 3-day distributed lag models
(lags 0, 1, and 2 days) were also considered in a subset of 90 U.S. counties with daily PM2.5 data
available during the study time period.
Subsequently, Peng et al. (2008, 156850) and Bell et al. (2008, 156266) extended the database
of daily time-series of hospital admissions, PM25, and other covariates for 202 U.S. counties through
2005. Importantly, Peng et al. (2008, 156850) added data on PMi0.2.5 to this database for 108 U.S.
counties with one or more co-located PM25 and PM10 monitors. Analyses with PM10_25 were carried
out using similar methods to those of Dominici et al. (2006, 088398). Peng et al. (2008, 156850)
evaluated the robustness of PM25 associations to adjustment for PMi0_2.5 (Peng et al., 2008, 156850).
Gaseous pollutants were not considered in these analyses.
SOPHIA: Study of Particulates and Health in Atlanta
SOPHIA investigators (Metzger et al.. 2004. 044222: Peel et al., 2005, 056305; Tolbert et
al., 2000, 010320) compiled data on 4,407,535 ED visits between 1993 and 2000 to 31 hospitals in
the Atlanta metropolitan statistical area (20 counties). Specific cardiovascular outcomes considered
were: IHD (ICD-9: 410-414), acute MI (410), cardiac dysrhythmias (427), cardiac arrest (427.5),
CHF (428), peripheral vascular and CBVD (433-437, 440, 443-444, 451-453), atherosclerosis (440),
and stroke (436). Finger wounds (883.0) were evaluated as a control outcome.
The air quality data included measurements of criteria pollutants (PM and gaseous pollutants)
for the entire study period, as well as detailed measurements of mass concentrations for PM25 and
PM10_2.5 and several physical and chemical characteristics of PM25 for the final 25 mo of the study
using data from the ARIES monitoring station. Rates of ED visits for specific causes were assessed
in relation to the 3-day ma (lags 0-2 days) of daily measures of air pollutants using Poisson
generalized linear models (GLMs) controlling for long-term temporal trends and meteorologic
conditions with cubic splines. Tolbert et al. (2007, 090316) published interim results of this study in
relation to both cardiovascular and respiratory disease visits, Metzger et al. (2004, 044222)
published the main results for CVD visits, and Peel et al. (2005, 056305) published the main results
for respiratory conditions. An analysis of co-morbid conditions that may make individuals more
susceptible to PM-related cardiovascular risk was carried out by Peel et al. (2007, 090442). Tolbert
et al. (2007, 090316) extended the available data through 2002 and compared results from single and
multipollutant models, while Sarnat et al. (2008, 097972) evaluated the risk of ED visits for
cardiovascular and respiratory diseases in relation to specific sources of ambient PM using the
extended dataset.
APHEA andAPHEA-2: Air Pollution and Health: a European Approach
APHEA-2 investigators compiled daily data on cardiovascular (Le et al., 2002, 023746; 2003,
042820) and respiratory (Atkinson et al., 2001, 021959; 2003, 042797) disease hospital admissions
in the following 8 European locations: Barcelona, Birmingham, London, Milan, the Netherlands
(considered a "city" for this study, due to its small size and dense population), Paris, Rome, and
Stockholm. (The publications on respiratory diseases were reviewed in the 2004 PM AQCD). The
specific CVD outcomes considered in each city were: cardiac diseases (ICD-9: 390-429), IHD
(410-413) and CBVDs (430-438). Routine registers in all cities provided daily data on
hospitalizations. Only emergency hospitalizations were considered, except in Milan, Paris, and
Rome where only general admissions data were available.
Ambient PMi0 levels were available in all cities except Paris (PMi3 used), and Milan and
Rome (TSP used). Data on gaseous pollutants (NO2, SO2, CO, and O3) were also available in most
cities. Five of the eight cities provided data on black smoke (BS). The length of the available
time-series varied by city but generally spanned from the early to mid-1990s.
Within each city, associations between cause-specific hospitalization rates and same-day PM2 5
levels were evaluated using Poisson GAMs controlling for long-term temporal trends and
meteorologic conditions. City-specific results were subsequently averaged using standard
December 2009 6-58
-------
meta-analytic methods. The original analyses (Atkinson et al., 2001, 021959; Le et al., 2002,
023746) were carried out using general additive models (GAM) and LOESS smoothers. Following
reports of problems associated with using the default convergence criteria in the standard S-plus
GAM procedure (Dominici et al., 2002, 030458). study authors reanalyzed the data on cardiac
admissions using GAMs and stricter convergence criteria, and GLMs with natural splines and
penalized splines (Atkinson et al., 2003, 042797: Le et al., 2003, 042820). The authors found that
the results of the original analyses were insensitive to the choice of convergence criteria and that the
use of GLMs with penalized splines yielded very similar results.
HEAPSS: Health Effects of Air Pollution among Susceptible Subpopulations
HEAPSS investigators collected data on patients hospitalized for a first MI in five European
cities between 1992 and 2000. Patients were identified from MI registers in Augsburg and
Barcelona, and from hospital discharge registers in Helsinki, Rome and Stockholm. Data on daily
levels of PMio, were measured at central monitoring sites in each city. Particle number concentration
was measured for a year in each city and then modeled retrospectively for the whole study period.
Associations of outcomes with gaseous criteria pollutants were also evaluated.
Von Klot et al. (2005, 088070) identified 22,006 survivors of a first MI in the five participating
European cities and collected data on subsequent first cardiac re-hospitalizations between 1992 and
2001. Readmissions of interest were those with primary diagnoses of acute MI, angina pectoris, or
cardiac disease (which additionally includes dysrhythmias and CHF). Within each city, associations
between cause-specific hospitalization rates and same-day levels of PMi0 were evaluated using
Poisson GAMs controlling for long-term temporal trends and meteorologic conditions using
penalized splines. City-specific results were combined using standard meta-analytic methods.
Subsequently, Lanki et al. (2006, 089788) used HEAPSS data from 26,854 patients to evaluate the
association between daily PMio and particle number concentrations and the risk of hospitalization for
first MI.
PSAS: The French National Program on Air Pollution Health Effects
Larrieu et al. (2007, 093031) evaluated the association between PM10 and the risk of
hospitalization in eight French cities between 1998 and 2003. The cities examined were: Bordeaux,
Le Havre, Lille, Lyon, Marseille, Paris, Rouen and Toulouse. The specific CVD outcomes
considered in each city included: total CVD (ICD-10: 100-199), cardiac disease (100-152), IHD
(120-125) and stroke (160-164, G45-G46). The available data did not differentiate between emergency
and non-emergency hospitalizations. Daily mean PMio and NO2 levels as well as 8-h max O3 levels
were obtained from a network of monitors in each city.
Within each city, associations between cause-specific hospitalization rates and 2-day ma (lag
0-1 days) levels of PMio were evaluated using Poisson GAMs controlling for long-term temporal
trends and meteorologic conditions using penalized splines. City-specific results were combined
using standard meta-analytic methods. Host et al. (2008, 155852) used a subset of these data (6
cities, 2000-2003) to compare the effects of PM2.5 and PMi0_2.5 on the risk of cardiovascular and
respiratory admissions. CVD outcomes assessed in this analysis were all CVD (ICD-10 100-199),
cardiac disease (100-152) and IHD (120-125). PM2.5 levels were obtained from the same network of
background monitors described above. PMi0_2.s was calculated by subtracting PM2 5 levels from PMio
levels. Gaseous pollutants and hospital admissions for stroke were not considered in this analysis.
Multicity Studies in Australia and New Zealand
Barnett et al. (2006, 089770) collected data on daily CVD emergency hospital admissions
among older adults and pollution data between 1998 and 2001 in five Australian cities (Brisbane,
Canberra, Melbourne, Perth, Sydney) and two cities in New Zealand (Auckland, Christchurch). In
2001, these cities covered 53% of the Australian population and 44% of the New Zealand
population. The specific outcomes considered in each city were: all circulatory diseases (ICD-9
390-429, ICD-10 100-199 with exclusions); CHF (ICD-9 428, ICD-10 150); arrhythmia (ICD-9 427
ICD-10 146-49); cardiac disease (ICD-9 390-429, ICD-10 100-152,197.0, 197.1,198.1); IHD (ICD-9
410-413, ICD-10 120-24,125.2); acute MI (ICD-9 410, ICD-10 121-22); and stroke (ICD-9 430-438,
ICD-10 160-66,167,168,169, G45-46 with exclusions).
December 2009 6-59
-------
Air pollutants considered were 24-h avg PM10, 24-h avg PM2.5, BSP and gaseous pollutants.
Within each city, associations between cause-specific hospitalization rates and 2-day ma (lags
0-1 days) of PMio were evaluated using the time-stratified case-crossover approach which controls
for long-term and seasonal time trends by design rather than analytically. City-specific results were
combined using random effects meta-analytic methods.
EMECAS: Spanish Multicentric Study on the Relation between Air Pollution and Health
Ballester et al. (2006, 088746) collected data on daily cardiovascular emergency hospital
admission and air pollution data between approximately 1995 and 1999 in 14 cities in Spain. The
specific outcomes considered in each city were: total CVD (ICD-9: 390-459) and heart diseases
(410-414, 427, 428). Air pollutants considered were PMio, TSP, BS, SO2, NO2 (24-h avg), CO and
O3 (8-h max).
Within each city, associations between cause-specific hospitalization rates and daily levels of
each pollutant metric were evaluated using Poisson GAMs with strict convergence criteria. In all
models, pollutants were entered as linear continuous variables and included control for confounding
by meteorological variables, influenza rates, long-term time trends, and unusual events. The authors
considered both distributed lag models (lags 0-3 days) and the 2-day ma of pollution (lags 0-1 days).
City-specific results were combined using standard meta-analytic methods.
6.2.10.1. All Cardiovascular Disease
The 2004 PM AQCD (U.S. EPA, 2004, 056905) incorporated the results of a large number of
time-series studies in the U.S. and elsewhere relating ambient PM levels and risk of hospitalization
for CVD. The 2004 PM AQCD (U.S. EPA, 2004, 056905) noted that the strongest evidence for this
association came from the NMMAPS study (Samet et al., 2000, 010269) and the subsequent
reanalysis by Zanobetti and Schwartz (2003, 157174).
Since then, the U.S. MCAPS study evaluated the association between PM2.5 and risk of CVD
hospitalization in 202 U.S. counties between 1999 and 2005 and found a 0.8% (95% posterior
interval (PI): 0.6-1.0) increase in risk per 10 (ig/m3 increase in PM25 on the same day (Bell et al.,
2008, 156266: Peng et al., 2008, 156850). In 108 U.S. counties with co-located PMi0 and PMi0.2.5
monitors, Peng et al. found a 0.4% (95% PI, 0.1- 0.7, lag 0) increase in risk per 10 (ig/m3 PMi0_2.5
and no associations at lags of 1 and 2 days (Peng et al., 2008, 156850). In a two-pollutant model
adjusted for PM25, the association between PMi0_25 and CVD hospitalization lost precision (0.3%
[95% PI: -0.1 to 0.6, lag 0]). Bell et al. (2008, 156266) found evidence of substantial and
statistically significant variability in the effects of PM25 on cardiovascular hospitalizations by season
and region, with the highest national average estimates occurring in the winter and the highest
regional estimates in the northeastern U.S. (1.08% [95% PI: 0.79-1.37, lag 0, per 10 (ig/m3 increase
in PM2.5]). Estimates for the nation (1.49% [95% PI: 1.09-1.89, lag 0]) and northeast (2.01% [95%
PI: 1.39-2.63, lag 0]) were highest in the winter.
Bell et al. (2009, 191997) and Peng et al. (2009, 191998) used data from the MCAPS study
and the EPA's Speciation Trends Network (STN) to identify the components of PM2 5 that are most
strongly associated with hospitalizations for CVD. Peng et al. (2009, 191998) focused on the
components that make up the majority of PM25 mass (SO42~, NO3~, Si, EC, OC, Na+ and NH4+) and
found that in multipollutant models, only EC and OC were significantly associated with risk of
hospitalization for CVD. Bell et al. (2009, 191997) used data from 20 PM2 5 components and found
that EC, Ni, and V were most positively and significantly associated with the risk of cardiovascular
hospitalizations. These results suggest that the observed associations between PM2 5 and CVD
hospitalizations may be primarily due to particles from oil combustion and traffic.
Additional evidence is provided by several large multicity studies conducted outside of the
U.S. The European APHEA2 study (Le et al., 2002, 023746) looked at admissions for CVD among
those aged >65 and found a 0.7% (95% CI: 0.4-1.0, lag 0-1 day avg) increase in risk per 10 (ig/m3
PMio. The Spanish EMECAS study (Ballester et al., 2006, 088746) looked at admissions for CVD
and found a 0.9% (95% CI: 0.4-1.5, lag 0-1 day avg) increase in risk per 10 (ig/m3 PMi0. The French
PSAS program looked at CVD hospitalizations among the elderly and found a 1.9% (95% CI:
0.9-3.0, lag 0-1 day avg) increase in risk with a 10 ug/m3 increase in PM25 and a 1.1% (95% CI: 0.5-
1.7) increase in risk with PM10 (Host et al., 2008, 155852; Larrieu et al., 2007, 093031).
Non-significant increases in CVD hospital admissions association with PMi0_2.5 were reported (1.0%
December 2009 6-60
-------
[95% CI: -1.0 to 3.0]) (Host et al., 2008, 155852). In multiple cities across New Zealand and
Australia, Barnett et al. (2006, 089770) found a 1.3% (95% CI: 0.6-2.0, lag 0-1 day avg) increase in
risk per 10 (ig/m3 increase in PM25.
The Atlanta-based SOPHIA study found a 3.3% (95% CI: 1.0-5.6, lag 0-2 day avg) and a 0.9%
(95% CI: -0.2 to 1.9, lag 0-2 day avg) increase in risk with a 10 (ig/m3 increase in PM2.5 and PMi0,
respectively (Metzger et al., 2004, 044222). In a more recent analysis from this study with an
additional four years of data, ED visits for CVD were not significantly associated with PM10 or
PM25, but were significantly associated with total carbon (1.6% [95% CI: 0.5-2.6, per IQR
increase]), EC (1.5% [95% CI: 0.5-2.5, per IQR increase]) and OC (1.5% [95% CI: 0.5-2.6, per IQR
increase]) components of PM25 (2007, 090316). A weak non-significant association PMi0_25 was
observed in these data (Tolbert et al., 2007, 090316) More recently, Sarnat et al. (2008, 097972)
used multiple source-apportionment methods to evaluate the association between all CVD ED visits
and specific PM2 5 sources and found consistent positive associations with sources related to motor
vehicles and biomass combustion. These results were insensitive to the source-apportionment
technique used. It is noteworthy that other traffic-related gaseous pollutants were associated with
CVD ED visits in the SOPHIA study (Metzger et al., 2004, 044222).
Using meta-regression techniques and the reported association between PMi0 and CVD
hospitalizations from the 14 cities included in the NMMAPS analysis, Janssen et al. (2002, 016743)
examined whether the between-city variability in relative risk estimates were related to the local
contribution of a number of PM sources. The authors found that in multivariate analyses PMi0
coefficients increased significantly with increasing percentage of PMi0 emissions from highway
vehicles/diesels and oil combustion.
A small number of additional single-city studies have been published showing positive
associations between hospital admissions and ambient PM in Copenhagen, Denmark (Andersen et
al., 2007, 093201). weak nonsignificant associations in Spokane, WA (Schreuder et al., 2006,
097959; Slaughter et al., 2005, 073854), and no associations in two small counties in Idaho (Ulirsch
et al., 2007, 091332). Schreuder et al. (2006, 097959) performed a source apportionment analysis
using seven years of daily speciation data from the same residential monitor in Spokane, WA used by
Slaughter et al. (2005, 073854). These authors related daily levels of four sources (wood smoke, an
As-rich source, motor vehicle emissions, and airborne soil) to the excess risk of cardiovascular ED
visits. During the heating season, the only notable association for CVD-related ED visits was with
wood smoke, while in the non-heating season the only notable association was with airborne soil.
While neither of these associations reached statistical significance, the study likely lacked the
statistical power to find effects of the expected magnitude. In fact, it is doubtful that studies
conducted outside of large metropolitan areas have sufficient statistical power to detect associations
of the expected magnitude. Delfino et al. (2009, 191994) evaluated the effects of the 2003 California
wildfires and observed a slightly larger excess risk of total CVD admissions during the wildfire
period compared to the period prior to the wildfire, although excess risk estimates were generally
weak and non-significant.
Studies in several cities in Australia have investigated the association of CVD admissions with
PM concentration and sources. A study from Sydney, Australia found a 1.8% (95% CI: 0.4-3.2) and
0.3% (95% CI: -0.8 to 1.4) excess risk per 10 (ig/m increase in the 2-day ma (lags 0-1 days) in
PM2.5 and PM10, respectively (Jalaludin et al., 2006, 189416). Johnston et al. (2007, 155882) and
Hanigan et al. (2008, 156518) studied the association between PMio and cardiovascular and
respiratory hospitalizations in Darwin, Australia, where the predominant source of PM is from
biomass combustion. The authors found little or no evidence of an association between PMi0 and
CVD hospital admissions in the general population.
Crustal material has also been investigated in an effort to explain associations of PM
concentration with CVD admissions. Studies of a dust storm in the Gobi desert that transported PM
across the Pacific Ocean reaching the western U.S. in the spring of 1998 have been conducted. An
analysis of the health impacts of this event on the population of British Columbia's (Canada) Lower
Fraser Valley found no excess risk of cardiac or respiratory hospital admissions despite hourly PMi0
levels >100 (ig/m3 (Bennett et al., 2006, 088061). On the other hand, a number of studies in Asia
and eastern Europe have reported associations between CVD hospital admissions and dust storm
events. Middleton et al. (2008, 156760) found that dust storms in Cyprus were associated with a
4.7% (95% CI: 0.7-9.0) and 10.4% (95% CI: -4.7 to 27.9) increase in risk of hospitalization for all
causes and CVD, respectively. Chan et al. (2008, 093297) studied the effects of Asian dust storms on
cardiovascular hospital admissions in Taipei, Taiwan and also found significant adverse effects
December 2009 6-61
-------
during 39 Asian dust events with high PM10 levels (daily PM10 >90 (ig/m3). Bell et al. (2008,
091268) analyzed these data independently and concluded that Asian dust storms were positively
associated with risk of hospitalization for IHD.
Study
Bell etal. (2008,156266)
Hostetal. (2008, 155852)
Barnett etal. (2006. 089770)
Metzger et al. (2004, 044222)
Tolbert etal. (2007,090316)
Slaughter etal. (2005. 073854)
Delfino etal. (2009, 191994)
Location
202 Counties, US
6 Cities, France
7 Cities, Australia/NZ
Atlanta, GA
Atlanta, GA
Spokane, WA
CA
Lag
0
0
0
0
0-1
0-1
0-2
0-2
1
0-1
Covariates Effect Estimate (95% Cl)
65+ Overall
65+ NE
65+, Winter
65+ NE Winter
65+
65+
All Ages
All Ages _
All Ages
*. PM2.5
-*-
U —
All Ages, Wildfires J_i —
Peng etal. (2008. 156850)
Hostetal. (2008, 155852)
Tolbert etal. (2007, 09031 6)
108 Counties, US
6 Cities, France
Atlanta, GA
0
1
2
0-1
0-2
65+ [*- PMio-2.5
65+ -f
65+ -•-
All Ages
Zanobetti & Schwartz (2003, 043119)
Metzger et al. (2004, 044222)
Tolbert etal. (2007. 09031 6)
Ballester etal. (2006, 088746)
Ulirsch etal. (2007.091332)
Slaughter et al. (2005, 073854)
LeTertre etal. (2002, 023746)
10 Cities, US
Atlanta, GA
Atlanta, GA
14 Cities, Spain
Pocatello, Chubbuck, ID
Spokane, WA
8 Cities, Europe
0-1
0-2
0-2
0-1
0
1
0-1
65+
All ages
All Ages
65+
•
1 * PMio
.T
50-r « 1
All dgeb *
65+
+
iii iii
-0-4-20 2 4 6
Excess Risk Estimate
Figure 6-1. Excess risk estimates per 10 ug/m3 increase in 24-h avg PM2.s, PMio. 2.5, and
concentration for CVD ED visits and HAs. Studies represented in the
figure include all multicity studies, as well as single-city studies conducted in the
U.S. or Canada.
The effect estimates from multicity studies and single-city studies conducted in the U.S. and
Canada are included in Figure 6-1. Information on PM concentrations during the relevant study
period is presented in Table 6-8. In summary, large studies from the U.S., Europe, and Australia/New
Zealand published since the 2004 PM AQCD (U.S. EPA, 2004, 056905) provide support for an
association between short-term increases in ambient levels of PM2.5 and PMio and increased risk of
hospitalization for total CVD. The evidence for an association of CVD hospitalization with PMi0_2.5
is relatively limited. Peng et al. (2008, 156850) reported that their PM10_2.5 estimate was not robust to
adjustment for PM2.5 and estimates from the other studies are imprecise. The average excess risk
among the U.S. elderly is likely in the range of 0.5-1.0% per 10 (ig/m3 increase in PM25, although
substantial variability by region of the country and season has been demonstrated. An excess risk of
ED visits for CVD of a similar magnitude appears likely. The excess risk of CVD hospitalization
may be somewhat greater in Europe and Australia/New Zealand than in the U.S. Sources including
wood burning, oil burning, traffic and crustal material have been associated with increases in
cardiovascular hospitalization or ED visits, but the best evidence suggests that in the U.S., oil
combustion, wood burning, and traffic are likely the sources of PM25 most strongly associated with
cardiovascular hospitalizations or ED visits.
December 2009
6-62
-------
Table 6-8. Characterization of ambient
admission and ED visits for
Pollutant
Study
PM concentrations in epidemiologic studies of hospital
cardiovascular diseases.
Location
Mean Concentration
(ug/m3)
Upper Percentlle
Concentration (ug/m3)
PM2.5
Barnett et al. (2006, 089770)
Bell et al. (2008, 1562661
Burnett etal. (1999, 017269)
Dominic! et al. (2006, 088398)
Delfino et al. (2009, 191994)
Host et al. (2008, 1558521
Ito et al. (2003, 0428561: Lippman (2000, 0119381
Lisabeth et al. (2008, 1559391
Metzger etal. (2004,044222)
Pope et al. (2006, 0912461
Slaughter etal. (2005, 0738541
Sullivan et al. (2005, 0508541
Symons et al. (2006, 091258)
Tolbertetal. (2007, 090316)
Villenueve et al. (2006, 0901911
Zanobetti and Schwartz (2005, 0880691
7 cities in Australia
202 counties in the U.S.
Toronto Canada
204 counties in the U.S.
6 counties CA
6 cities in France
Detroit, Ml
Atlanta, GA
Wasatch Front, Utah
Spokane, WA
King County, WA
Baltimore, MD
Atlanta, GA
Edmonton, Canada
Boston, MA
8.1-11.0
12.92
18
13.4
18.4-32.7
13.8-18.8
18
7
17.8
10.1-11.3
NR
12.8
16
17.1
8.5
11.1 (median)
NR
34.16
95th: 34.0, Max: 90
NR
45.3-76.1 (wildfire period)
95th: 25-33
98th: 55.2
75th: 10
90th: 32.3
98th: 39.8
Max: 82-1 44
90th: 20.2
90th 27.3, Max: 147
Max: 69.2
98th: 38.7
75th: 11
95th: 26.31
98th: 55.2
PM10-2.5
Burnett etal. (1999, 0172691
Host et al. (2008, 1558521
Ito et al. (2003, 042856): Lippman (2000, 011938)
Le Tertre et al. (2002, 023746)
Metzger etal. (2004, 0442221
Peng et al. (2008, 1568501
Peters etal. (2001,0165461
Slaughter etal. (2005, 073854)
Tolbertetal. (2007, 090316)
Toronto, Canada
6 cities in France
Detroit, Ml
8 cities in Europe
Atlanta, GA
204 cities in the U.S.
Boston, MA
Spokane, WA
Atlanta, GA
12.2
7-11
13
NR
9.1
9.8 (Median)
7.4
NR
9
Max: 68
95th: 12.5-21.0
Max: 50
NR
90th: 16.2
75th: 15.0
95th: 15.2
NR
Max: 50.3
PM10
Ballester et al. (2006, 088746)
Barnett et al. (2006, 0897701
Burnett et al. (1999, 0172691
Ito et al. (2003, 042856): Lippman (2000, 011938)
Jalaludin et al. (2006, 1894161
Larrieu et al. (2007, 093031)
Le Tertre et al. (2002, 023746)
14 cities in Spain
7 cities in Australia and New
Zealand
Toronto, Canada
Detroit, Ml
Sydney, Australia
8 cities in France
8 cities in Europe
32.8-43.2
16.5-20.6
30.2
31
16.8
21.0-28.9
Range: 15.5-55.7
90th: 50.3-62.6
NR
95th: 56.0
NR
75th: 19.9
Max: 103.9
NR
Range 75th: 19. 9-66
December 2009
6-63
-------
Pollutant Study
Linn et al. (2000, 0028391
Metzgeretal. (2004, 044222)
Morris etal. (1998, 0249241
Peters etal. (2001,0165461
Schwartz et al. (1995, 0461861
Slaughter etal. (2005, 073854)
Tolbertetal. (2007, 090316)
Ulirsch et al. (2007, 0913321*
Wfellenius et al. (2005, 0874831
Wfellenius et al. (2005, 0886851
Wfellenius et al. (2006, 088748)
Zanobetti and Schwartz (2005, 088069)
Location
Los Angeles, California
Atlanta, GA
Chicago, Illinois
Boston, MA
Detroit, Ml
Spokane, WA
Atlanta, GA
2 cities in southeast Idaho
Pittsburgh, PA
9 cities in the U.S.
7 cities in the U.S.
Boston, MA
Mean Concentration
(ug/m3)
45
26.3
41
19.4
48
NR
26.6
24.2/23.2
31.1
28.4 (median)
28.3 (median)
28.4 (median)
Upper Percentlle
Concentration (ug/m3)
78 (summer) -132 (fall)
90th: 44.7
75th: 51
Max: 117
95th: 37.0
90th: 82
90th: 41. 9
Max: 98.4
90th: 40.7/37.4
95th: 70.5
90th: 57.9
90th: 57
90th: 53.6
*Results presented separately for 2 separate time series
6.2.10.2. Cardiac Diseases
Cardiac disease represents a subset of CVD which specifically excludes hospitalizations for
CBVD, peripheral vascular disease, and other circulatory diseases not involving the heart or
coronary circulation. Only a small number of studies published since the 2004 PM AQCD
(U.S. EPA, 2004, 056905) have evaluated the association between ambient PM and hospitalizations
for cardiac diseases, as most investigators have focused instead on more narrowly defined outcomes.
The French PSAS program found a 2.4% (95% CI: 1.2-3.7, lag 0-1) and 1.5% (95% CI:
0.5-2.2, lag 0-1) excess risk among the elderly per 10 (ig/m3 increase in PM25 and PMi0,
respectively (Host et al., 2007, 155851: Larrieu et al., 2007, 093031). Host et al. (2008, 155852)
also found a positive less precise association with PMi0_2.5, (excess relative risk per 10 ug/m3: 1.6%
[95% CI: -0.8 to 4.1]). The European HEAPSS study looked at cardiac readmissions among
survivors of a first MI and found a 2.1% (95% CI: 0.4-3.9, lag 0) excess risk per 10 (ig/m3 increase
in PM10 (Von et al., 2005, 088070). A 1.9% (95% CI: 1.0-2.7, lag 0-1) excess risk per 10 ug/m3
increase in PM2.5 was observed in several cities in Australia and New Zealand (Barnett et al., 2006,
089770). Single-city studies of hospital admissions from Kaohsiung and Taipei, Taiwan, and an ED
visit study from Sydney, Australia also reported statistically significant positive associations(Chang
et al., 2005, 080086: Jalaludin et al., 2006, 189416: Yang et al., 2004, 094376). On the other hand,
Slaughter et al. (2005, 073854) found no association between either PM2.5 or PMi0 and risk of
cardiac hospitalization in Spokane, Washington.
In summary, although relatively few studies have focused on all cardiac diseases, large studies
from Europe and Australia/New Zealand published since the 2004 PM AQCD (U.S. EPA, 2004,
056905) report positive associations between short-term increases in ambient levels of PM2.5,
PM10_2.5, and PM10and increased risk of hospitalization for cardiac disease. The results from small
single-city studies are less consistent. The excess risk for cardiac hospitalizations may be somewhat
larger than for total CVD hospitalizations.
6.2.10.3. Ischemic Heart Disease
IHD represents a subset of all cardiac disease hospitalizations and typically includes acute MI
(ICD 9: 410), other acute and subacute forms of IHD (411), old MI (412), angina pectoris (413), and
other forms of chronic IHD (414). Some authors term this category coronary heart disease. Published
studies evaluating IHD as a single outcome are considered first, followed by consideration of studies
looking at acute MI, a specific form of IHD.
In one of the first studies to evaluate IHD, Schwartz and Morris (1995, 046186) reported a
0.6% (95% CI: 0.2-1.0) excess risk of hospitalization for IHD per 10 (ig/m3 increase in mean PMi0
December 2009
6-64
-------
levels over the previous two days among elderly Medicare beneficiaries living in Detroit between
1986 and 1989. As reviewed in the 2004 PM AQCD (U.S. EPA, 2004, 056905). similar associations
were subsequently observed in many single-city studies including: London, England (Atkinson et
al., 1999, 007882). Toronto, Canada (Burnett et al., 1999, 017269). and Seoul, Korea (Lee et al.,
2003, 095552). Studies in Hong Kong (Wong et al., 1999, 009172: Wong et al., 2002, 023232).
Birmingham, England (Anderson et al., 2001, 017033). and London, England (Wong et al., 2002,
023232) yielded positive point estimates of a similar magnitude, but did not reach statistical
significance.
The positive associations between short-term changes in PM and IHD hospitalizations
observed in the early single-city studies have been confirmed in several large multicity studies. The
U.S. MCAPS study (Dominici et al., 2006, 088398) found a 0.4% (95% CI: 0.0-0.8) excess risk of
hospitalization for IHD per 10 ug/m increase in PM25 two days earlier. The European APHEA-2
study (Le et al., 2002, 023746) considered PM10 and found a 0.8% (95% CI: 0.3-1.2, lag 0-1) excess
risk among those aged >65 yr. Among the elderly in 5 cities in Australia and New Zealand (Barnett
et al., 2006, 089770) there was a 4.3% (95% CI: 1.9-6.4, lag 0-1) excess risk per 10 ug/m3 increase
in PM25. Among the elderly in several French cities there was a 4.5% (95% CI: 2.3-6.8, lag 0-1),
6.4% (95% CI: 1.6-11.4, lag 0-1) and 2.9% (95% CI: 1.5-4.3, lag 0-1) excess risk per 10 ug/m3
increase in PM25 PMi0_25 (Host et al., 2008, 155852). and PMi0, respectively (Larrieu et al., 2007,
093031).
With regard to ED visits, the Atlanta-based SOPHIA study (Metzger et al., 2004, 044222)
found positive associations with PM25 and PMi0 (ranging from 1.1 to 2.3%), but the effect estimates
did not reach statistical significance. Similarly, associations of EC and OC with IHD were increased
but not significant. In 6 cities across Canada, Szyszkowicz (2009, 191996) observed a 2.4% (95%
CI: 1.2-3.6) and 1.4% (95% CI: 0.7-2.0) excess risk of ED visits for angina per 10 ug/m3 increase in
same-day PM25 and PM10, respectively. Although excess risks were generally weak and non-
significant, Delfmo et al. (2009, 191994) observed a slightly larger excess risk of IHD during
wildfires compared to the pre-wildfire period. In Sydney, Australia, Jalaludin et al. (2006, 189416)
found a 2.6% (95% CI: 0.1-5.2) and 0.8% (95% CI: -1.2 to 2.8) excess risk of ED visits for IHD per
10 ug/m3 increase in 2-day ma of PM25 and PMi0, respectively. A recent study in Helsinki, Finland,
found no evidence of an association of IHD hospital admissions with UFP, ACP, PM25, PMi0_25, or
source-specific PM2.5 (Halonen et al., 2009, 180379).
To explore this link further, Pope et al. (2006, 091246) used data from an ongoing registry of
patients undergoing coronary angiography at a single referral center in Salt Lake City, UT, between
1994-2004. The authors found a 4.8% (95% CI: 1.0-8.8, lag 0) excess risk of acute MI or unstable
angina per 10 ug/m3 increase in PM25 among 4,818 patients. These results were robust to changes in
the definition of the outcome. The results of this study are particularly noteworthy given the high
specificity of the outcome definition.
In summary, large studies from the U.S., Europe, and Australia/New Zealand published since
the 2004 PM AQCD (U.S. EPA, 2004, 056905) provide support for an association between
short-term increases in ambient levels of PMi0 and PM25 and increased risk of hospitalization or ED
visits for ischemic heart diseases. Although estimates are less precise for PMi0_2.5, most results from
single pollutant models provide evidence of a positive association between PMi0_2.5 and IHD.
Moreover, Host et al. (2008, 155852) found that the effect estimates for the association of PM25 and
PMio_2.5 with IHD were very similar when scaled to the IQR of each metric. Estimates of the excess
risk vary considerably between studies, but as was the case for total CVD hospitalizations, the excess
risk appears to somewhat greater in Europe and Australia/New Zealand. Results from multicity
studies and U.S. and Canadian single-city studies are presented in Figure 6-2.
December 2009 6-65
-------
Study
Location
Lag Age
Effect Estimate (95% Cl)
ISCHEMIC HEART DISEASE
Ito (2003, 042856)
Popeetal (2006 091246)
Hostetal. (2007, 155851)
Metzgeretal (2004 044222)*
Barnettetal (2006 089770)
Dominici et al. (2006, 088398)
Barnettetal. (2006.089770)
Hostetal. (2007. 155851)
Burnett etal. (1999. 01 7269)
Delfinoetal. (2009. 191994)
Mpt7npr pt ai or\r\A 044999^
Burnett etal. (1999. 01 7269)
Ito (2003. 042856)
Le Tertre et al. (2002, 023746)
Metzger et al. (2004, 044222)*
Larrieu etal. (2007. 09X31)
Burnett etal. (1999.017269)
Le Tertre etal. (2002. 023746)
Jalaludin etal. (2006. 189416)*
Larrieu etal. (2007, 09X31)
MYOCARDIAL INFARCTION
Peters etal (2001 016546)
Sullivan et al (2005 050854)
Zanobetti & Schwartz (2006, 090195)
pptprcptai nnm mfi^di^
Linn etal. (2000.002839)
Zanobetti & Schwartz (2005, 088069)
Detroit, Ml
Utah Valley UT
6 Cities, France
Atlanta GA
Australia/NZ
204 Counties, US
Australia/NZ
6 Cities, France
Toronto, Can
6 Counties, CA
(Wildfires)
Atlanta flA
Toronto, Can
Detroit, Ml
8 Cities, Europe
Atlanta, GA
8 Cities, France
Toronto, Can
8 Cities, Europe
Sydney, Australia
8 Cities, France
Boston MA
King County WA
Boston, MA
Los Angeles, CA
21 Cities, US
1
o
0-1
0-3
0-1
0
1
2
0-2 DL
0-1
0-1
0,1
0,1
n ?
0
1
0-1
0-2
0-1
0-1
0-1
0-1
0-1
2h
24 h
1 h
2h
4h
24 h
0
9 h
94 h
0
9 h
94 h
0
65+ -
All
All J
All
15-64
65+
65+ i
65+
65+ n
65+
65+
All
All -
All f •
All
65+
<65
All
All
All
65+
65+ —i
65+
61 6 Mean
61 6 Mean
21-98
9 1-98
0 1-98
9 1-98
65+
R1 R Mmnf
>x
fi1 fi Mpan
R1 R Mpan
65+
-. PM2.5
1 — • —
t-
•-
»-
•-
-• —
*
-•-
-t— PMio
••
-• —
-•-
h-»
PM2.5
*
*
_
*
•
PMlO-2.5
^ PMio
*
•»
i r
-6
o
n i r
6
\ r
12
18
~ i i
24 28
* ED Visits
DL Distributed Lag
Figure 6-2.
Excess Risk (%)
Excess risk estimates per 10 ug/m increase in 24-h avg (unless otherwise noted)
PM2.6, PMio-2.5, and PMio concentration for Ml and IHD ED visits and HAs. Studies
represented in the figure include all multi-city studies as well as single-city
studies conducted in U.S. or Canada.
December 2009
6-66
-------
6.2.10.4. Acute Myocardial Infarction
Because even IHD refers to a heterogeneous collection of diseases and syndromes, several
authors have evaluated the association between short-term fluctuations in ambient PM and acute MI,
a specific form of IHD.
In 2001, Peters et al. (2001, 016546) published their study evaluating the effects of PM on the
risk of MI among 772 Boston-area participants in the Determinants of MI Onset Study. The authors
found that a 10 ug/m3 increase in the 2-h or 24-h avg levels of PM25 was associated with a 17%
(95% CI: 4-32) and 27% (95% CI: 6-53) excess risk of MI, respectively. An imprecise, non-
significant association between PMi0_2.5 and onset of MI was observed in Boston. In contrast, a study
among 5793 patients in King County, WAthat used similar methods, found no association with PM25
with lag times of 1, 2, 4, or 24 h (Sullivan et al., 2005, 050854). Among 852 hospitalized patients in
Augsburg, Germany, Peters et al. (2005, 087759) also found no association between PM2 5 and MI
risk within this time frame, although they did find a positive and statistically significant association
with time spent in traffic (Peters et al., 2004, 087464).
These three studies are particularly important because in each one: (1) cases were
prospectively identified based on clinical criteria rather than retrospectively based on discharge
diagnoses; and (2) time of MI symptom onset was used for exposure assessment rather than date of
hospital admission. Whether the discrepant results among these studies are due to regional
differences in population characteristics and/or air pollution sources remains unclear. The King
County study suggests that differences in statistical approaches are unlikely to account for the
discrepant results (Sullivan et al., 2005, 050854). Analyses from the U.S. MCAPS study suggest
that substantial heterogeneity of effects are to be expected across regions of the country (Bell et al.,
2008, 156266)
Several studies have assessed the association between acute exposure to ambient PM and MI
using administrative databases. In the U.S., MI was not one of the specific endpoints evaluated in the
MCAPS study (Dominici et al., 2006, 088398) or in the Atlanta-based SOPHIA study of ED visits
(Metzger et al., 2004, 044222). However, Zanobetti and Schwartz (2005, 088069) found a 0.7%
(95% CI: 0.3-1.0) excess risk of MI per 10 ug/m3 increase in same-day PM10 among elderly
Medicare beneficiaries in 21 cities. Subsequently, the same authors found that among elderly
Medicare beneficiaries living in the Boston metropolitan region, a 10 ug/m3 increase in PM25 was
associated with a 4.9% (95% CI: 1.1-8.2) excess risk on the same day (Zanobetti and Schwartz,
2006, 090195).
This body of evidence may implicate traffic-related pollution generally as a risk factor for MI.
In the study described above, Peters et al. (2001, 016546) found positive associations between risk of
hospitalization for MI and potential markers of traffic-related pollution measured at a central monitor
including BC, CO and NO2. However, none of these associations were statistically significant in
models adjusting for season, meteorological variables, and day of week. Zanobetti and Schwartz
(2006, 090195) examined the association between traffic-related pollution and risk of hospitalization
for MI among Medicare beneficiaries in the Boston area and found that MI risk was positively and
significantly associated with measures of PM2 5, BC, NO2, and CO, but not with levels of
non-traffic-related PM25. Peters et al. (2004, 087464) interviewed 691 subjects with MI who
survived at least 24-h after the event and found a strong positive association between self-reported
exposure to traffic and the onset of MI within 1 h (OR: 2.9 [95% CI: 2.2-3.8]). The association was
somewhat stronger among subjects traveling by bicycle or public transportation in the hour prior to
the event. Of note, however, this study did not directly measure traffic-related pollution.
Similar studies with administrative databases have been conducted in Europe, Australia, and
New Zealand. Barnett et al. (2006, 089770) observed that in five cities in Australia and New
Zealand, a 10 ug/m3 increase in PM25 was associated with a 7.3% (95% CI: 3.5-11.4, lag 0-1 day)
excess risk. In Rome, D'Ippoliti et al. (2003, 074311) carried out a case-crossover study and found a
statistically significant positive association between TSP and the risk of hospitalization for MI. In
contrast, the HEAPSS study found no evidence of an association between PMi0 and risk of
hospitalization for a first MI in five European cities (Lanki et al., 2006, 089788). although there is
some indication that among survivors of a first MI, risk of re-hospitalization for MI may be related
to transient elevations in PMi0 (Von et al., 2005, 088070).
In summary, large studies from the U.S., Europe, and Australia/New Zealand published since
the 2004 PM AQCD (U.S. EPA, 2004, 056905) provide support for an association between
short-term increases in ambient levels of PM25 and PMi0and increased risk of hospitalization for MI.
Some of the heterogeneity of results is likely explained by regional differences in pollution sources,
December 2009 6-67
-------
components, and measurement error. One study of the effect of 2- and 24-h avg PM10_2.5
concentration on admissions for MI produced effect estimates that were positive, but imprecise
(Peters et al., 2001, 016546). These results need to be interpreted together with those studies
evaluating hospitalization for IHD since Mis make up the majority of hospitalizations for IHD. U.S.
studies of MI are included in Figure 6-2.
6.2.10.5. Congestive Heart Failure
Perhaps the first suggestion of an association between ambient PM and hospitalization for
CHF was provided by the study of Schwartz and Morris (1995, 046186). These authors reported that
among elderly Medicare beneficiaries living in Detroit between 1986-1989, a 10 ug/m3 increase in
mean PM10 levels over the previous two days was associated with a 1.0% (95% CI: 0.4-1.6) increase
in risk of hospitalization for CHF. As reviewed in the 2004 PM AQCD (U.S. EPA, 2004, 056905).
using similar approaches, statistically significant positive associations with PM2.5 or PMi0 were
subsequently reported in single-city studies looking at hospitalizations for CHF in Toronto (Burnett
et al., 1999, 017269). Hong Kong (Wong et al., 1999, 009172). and Detroit (Ito, 2003, 042856). but
not Los Angeles (Linn et al., 2000, 002839) or Denver (Koken et al., 2003, 049466). Burnett et al.
(1999, 017269) reports a significantly increased risk of CHF hospitalization with PM10_25 while
Metzger et al. (2004, 044222) and Ito et al. found (2003, 042856) less precise associations.
Subsequent multicity studies support the presence of a positive association between PM
concentration and CHF hospitalization. In the U.S., the MCAPS study found a 1.3% (95%: 0.8-1.8)
excess risk per 10 ug/m3 increase in same-day PM25 (Dominici et al., 2006, 088398). In addition,
Wellenius et al. (2006, 088748) reported a 0.7% (95% CI: 0.4-1.1) excess risk of hospitalization for
CHF per 10 ug/m3 increase in same-day PM10 among elderly Medicare beneficiaries in seven cities.
In Australia and New Zealand, Barnett et al. (2006, 089770) found a 9.8% (95% CI: 4.8-14.8, lag 0-1
day) and 4.6% (95% CI: 2.8-6.3, lag 0-1 days) excess risk of hospitalization for CHF associated with
a 10 ug/m3 increase in PM25 and PMi0, respectively. Results from more recent single-city studies in
Pittsburgh (Wellenius et al., 2005, 087483). Utah's Wasatch Front (Pope et al., 2008, 191969).
Kaohsiung, Taiwan (Lee et al., 2007, 093271) and Taipei, Taiwan (Yang, 2008, 157160) have also
reported positive associations between PM and CHF hospital admissions. In addition, Yang et al.
(2009, 190341) found that hospitalizations for CHF were elevated during or immediately following
54 Asian dust storm events (while single day lags 0-3 were evaluated, maximum excess risk
occurred at lag 1: 11.4% [95% CI: -0.7 to 25.0]). Delfino et al. (2009, 191994) observed a slightly
larger excess risk of total CHF during wildfires occurring in California compared to the period
before the wildfires.
While most studies suggest an association at very short lags (0-1 days), the study by Pope
et al. (2008, 191969) failed to find such short term associations and instead suggested that PM2 5
levels averaged over the past 2-3 wk may be more important. Pope et al. (2008, 191969) observed a
13.1% (95% CI: 1.3-26.2) increase in CHF hospitalization per 10 ug/m3 increase in PM25 (imputed
values used in analysis). Whether findings at longer lags in this population represent true cumulative
effects of PM or are due to misclassification of symptom onset times remains to be determined.
Findings from the Atlanta-based SOPHIA study (Metzger et al., 2004, 044222) also support
the presence of a positive association between PM and CHF ED visits. Specifically, the SOPHIA
study found a 5.5% (95% CI: 0.6-10.5, lag 0-2 days) excess risk of ED visits for CHF per 10 ug/m3
increase in the 3-day ma of PM25. Positive associations were also observed for CHF and EC and OC
components of PM25. No associations were observed with PMi0and a weak, imprecise increase was
observed in association with PMi0_2.5.
Only one published study has attempted to evaluate the effects of ambient particles on CHF
symptom exacerbation using data which was not derived from administrative databases. Symons
et al. (2006, 091258) interviewed 135 patients with prevalent CHF hospitalized for symptom
exacerbation in Baltimore, MD. The authors found a 7.4% (95% CI: -7.5 to 24.2) excess risk of
hospitalization per 10 ug/m3 increase in PM25 two days prior to symptom onset. This finding did not
reach statistical significance and may be attributable to the lack of statistical power needed to find an
effect of the expected magnitude.
In summary, large studies from the U.S., Europe, and Australia/New Zealand published since
the 2004 PM AQCD (U.S. EPA, 2004, 056905) provide support for an association between
short-term increases in ambient levels of PM25 and PMi0 and increased risk of hospitalization and
ED visits for CHF. Although the number of studies is fewer (and only Metzger et al., 2004, 044222
December 2009 6-68
-------
is new since the 2005 AQCD), elevated risks of hospitalization or ED visits for CHF in association
with PMio_2.5 have been observed. The excess risks associated with CHF hospitalizations and ED
visits are consistently greater than those observed for other CVD endpoints. The results of multicity
studies and U.S. and Canadian single-city studies are summarized in Figure 6-3.
Study
Burnett etal. (1999.017269)
Ito (2003, 042856)
Metzger et al. (2004, 044222)*
Barnett etal. (2006.089770)
Symons et al. (2006, 091258)
Dominici et al. (2006, 088398)
Barnett etal. (2006.089770)
Delfino etal. (2009. 191994)
Pope etal. (2008, 191969)
Burnett etal. (1999.017269)
Ito (2003, 042856)
Metzger et al. (2004, 044222)*
Burnett etal. (1999.017269)
Linn etal. (2000,002839)
Ito etal. (2003.042856)
Metzger et al. (2004, 044222)*
Barnett etal. (2006. 089770)
Schwartz & Morris (1995, 046186)
Morris & Naumova (1998, 086857)
Wellenius etal. (2005. 087483)
Wellenius etal. (2006. 088748)
Barnett etal. (2006.089770)
Location
Toronto, Canada
Detroit, Ml
Atlanta, GA
Australia/NZ
Baltimore, MD
204 U.S counties
Australia/NZ
CA wildfires
Utah
Toronto, Canada
Detroit, Ml
Atlanta, GA
Toronto, Canada
Los Angeles, CA
Detroit, Ml
Atlanta, GA
Australia/NZ
Detroit, Ml
Chicago, IL
Pittsburgh, PA
7 Cities, U.S.
Australia/NZ
Lag
0-2
1
0-2
0-1
2
0
0-1
0-1
14dDL
0-2
0
0-2
0-2
0
0
0-2
0-1
0-1
0
0
0
0-1
Age Effect Estimate (95% Cl)
All Ages
65+
All Ages
1c; GA
AIIAg°s
65+"
65+
All Anno
All Ages
All Ages
R^I
All Anr-
AIIAges
>30
65+
1^ R4
65+
65+
65+
65+
65+
1 1 i 1 I
* PM2.5
— *— PMio
-•-
-»
I I I i I I I I
-8 -6 -4-20
6 8
* ED Visits
Figure 6-3.
Excess Risk Estimate
Excess risk estimates per 10 ug/m3 increase in 24-h avg PM2.6, PMi0.2.5, and
concentration for CHF ED visits and HAs. Studies represented in the
figure include all multicity studies as well as single-city studies conducted in the
U.S. and Canada.
6.2.10.6. Cardiac Arrhythmias
A number of studies based on administrative databases have sought to evaluate the association
between short-term fluctuations in ambient PM levels and the risk of hospitalization for cardiac
arrhythmias (also known as dysrhythmias). Typically in these studies a primary discharge diagnosis
of ICD-9 427 has been used to identify hospitalized patients. However, ICD-9 427 includes a
heterogeneous group of arrhythmias including paroxysmal ventricular or supraventricular
tachycardia, atrial fibrillation and flutter, ventricular fibrillation and flutter, cardiac arrest, premature
beats, and sinoarterial node dysfunction. One study in the Netherlands found that the positive
predictive value of ICD-9 codes related to ventricular arrhythmias and sudden cardiac death was
82% (De et al., 2005, 155746). The positive predictive value of other codes related to cardiac
arrhythmias is unknown, but likely to be lower.
The results from early studies of arrhythmia-related hospitalizations have been inconsistent,
with positive findings in Toronto (Burnett et al., 1999, 017269) and null findings in Detroit
(Schwartz and Morris, 1995, 046186). Los Angeles (Linn et al., 2000, 002839). and Denver (Koken
et al., 2003, 049466). The U.S. MCAPS study found a statistically significant 0.6% (95% CI:
0.0-1.2) excess risk of hospitalization for the combined outcome of cardiac arrhythmias and
conduction disorders (ICD-9: 426, 427) per 10 (ig/m3 increase in same-day PM2.5 (Dominici et al.,
2006, 088398). A multicity study in Australia and New Zealand found no evidence of an association
between arrhythmia hospitalizations and either PM2.5 or PMio (Barnett et al., 2006, 089770). A study
in Helsinki, Finland, found no evidence of an association between either PM2.s or PMi0_2.5 and risk of
hospitalization for arrhythmias (Halonen et al., 2009, 180379). although there was an association
with smaller particles (0.03-0.1 (im).
December 2009
6-69
-------
With regard to ED visits, the Atlanta-based SOPHIA study found no evidence of an
association between any measure of ambient PM and the rate of ED visits for cardiac arrhythmias
(Metzger et al, 2004, 044222). However, in Sao Paulo, Brazil, Santos et al. (2008, 192004) found a
3.0% (95% CI: 0.5-5.4) excess risk of ED visits for arrhythmias per 10 ug/m increase in PMio on
the same day.
In summary, the current evidence does not support the presence of a consistent association
between short-term increases in ambient levels of PM25, PM10_2.5, or PM10and increased risk of
hospitalization for cardiac arrhythmias. However, it should be noted that studies of hospital
admissions or ED visits are ill-suited to the study of cardiac arrhythmias since most arrhythmias do
not lead to hospitalization. Studies in patients with implanted defibrillators, human panel studies
with ambulatory ECG recordings, and animal toxicological studies provide a more appropriate
setting for evaluating this endpoint. Results of these studies are described in Section 6.2.2.
6.2.10.7. Cerebrovascular Disease
Time-series studies evaluating the hypothesis that short-term increases in ambient PM2.5 or
PMio levels are associated with increased risk of hospitalization for CBVD have been inconsistent,
with few studies reporting positive associations (Chan et al., 2006, 090193; Dominici et al., 2006,
088398; Metzger et al., 2004, 044222; Wordley et al., 1997, 082745). and several studies reporting
null or negative associations (Anderson et al., 2001, 017033; Barnett et al., 2006, 089770; Burnett
et al., 1999, 017269; Halonen et al., 2009, 180379; Jalaludin et al., 2006, 189416; Larrieu et al.,
2007, 093031; Le et al., 2002, 023746; Peel et al., 2007, 090442; Villeneuve et al., 2006, 090191;
Wong et al., 1999, 009172).
The U.S. MCAPS study found a 0.8% (95% CI: 0.3-1.4) excess risk of hospitalization for
CBVD per 10 ug/m3 increase in same-day PM25 (Dominici et al., 2006, 088398). The association
showed regional variability with the strongest associations observed in the eastern U.S. The
Atlanta-based SOPHIA study found a 5.0% (95% CI: 0.8-9.3, lag 0-2 days) excess risk of ED visits
for cerebrovascular and peripheral vascular disease combined (excluding hemorrhagic strokes) per
10
10 ug/nf increase in PM2 5 and a 2.0% (95% CI: -0.1 to 4.3, lag 0-2 days) excess risk for PM
(Metzger et al., 2004, 044222). Delfino et al. (2009, 191994) observed a weak association between
excess risk of CBVD admissions before and during a wildfire occurring in California and slightly
higher risks after the wildfire period.
Large multicity studies conducted outside of North America have failed to observe an
association between PM and CBVD hospitalizations. The APHEA study found no excess risk (0.0%
[95% CI: -0.3 to 0.3]) of hospitalization for CBVD per 10 ug/m3 increase in the 2-day ma of PMio in
8 European cities (Le et al., 2002, 023746). Investigators from the French PSAS program reported a
0.8% (95% CI: -0.9 to 2.5, lag 0-1 days) excess risk per 10 ug/m3 increase in PMio among patients
aged >65 yr and a 0.2% (95% CI: -1.6 to 1.9, lag 0-1 days) excess risk among all patients (Larrieu et
al., 2007, 093031). Although neither estimate was statistically significant, the estimated excess risk
among the elderly is very similar to that observed in the U.S. MCAPS study. Barnett et al. (2006,
089770) examined this hypothesis in New Zealand and Australia and reported no association.
All of the above studies have identified cases of CBVD based on ICD-9 or ICD-10 codes
(most commonly ICD-9 430-438). However, the range of ICD codes commonly used in these studies
includes ischemic strokes, hemorrhagic strokes, transient ischemic attacks (TIAs) and several poorly
defined forms of acute neurological events (e.g., seizures from a vascular cause) (Table 6-7). It is
plausible that ambient PM has different effects on each of these disparate outcomes.
Ischemic Strokes and Transient Ischemic Attacks
An increasing number of studies have specifically evaluated the association between PMio and
PM25 and the risk of ischemic stroke (Chan et al., 2006, 090193; Henrotin et al., 2007, 093270;
Linn et al., 2000, 002839; Lisabeth et al., 2008, 155939; Low et al., 2006, 090441; Szyszkowicz,
2008, 192128; Tsai et al., 2003, 080133; Villeneuve et al., 2006, 090191; Wellenius et al., 2005,
087483). Linn et al. (2000, 002839) found a 1.3% (95% CI: 1.0-1.6 per 10 ug/m3, PM10 lag 0) excess
risk of hospitalization for ischemic stroke in the Los Angeles metropolitan area. Wellenius et al.
(2005, 087483) reported a statistically significant 0.4% (95% CI: 0.0-0.9) excess risk per 10 ug/m3
increase in same-day PMio among elderly Medicare beneficiaries in nine U.S. cities. Low et al.
December 2009 6-70
-------
reported an absolute increase of 0.08 (95% CI; 0.002-0.16) ischemic stroke
hospitalizations per 10 ug/m3 increase in PMi0 in New York City. In Kaohsiung, Taiwan, Tsai et al.
(2003, 080133) found a 5.9% (95% CI: 4.3-7.4, lag 0-2 days) excess risk of hospitalization for
ischemic stroke per 10 ug/m3 increase in PMi0 after excluding days with mean daily temperature
<20°C. Meanwhile, in Taipei, Taiwan, Chan et al. (2006, 090193) found a 3.0% (95% CI: -0.8 to 6.6,
lag 3) and 1.6% (95% CI: -0.8 to 3.9, lag 3) excess risk per 10 ug/m3 increase in PM25and PMi0,
respectively. Villeneuve et al. (2006, 090191) and Szyszkowicz et al. (2008, 192128) found no
association between either PM2.5 or PMi0 and ED visits for acute ischemic stroke in Edmonton,
Canada.
Two recent studies are particularly noteworthy given the high specificity of the outcome
definition. Henrotin et al. (2007, 093270) used data on 1432 confirmed cases of ischemic stroke
from the French Dijon Stroke Register and found 0.9% (95% CI: -7.0 to 9.4) excess risk of ischemic
stroke per 10 ug/m increase in PM10 on the same day and a 1.1% (95% CI: -0.2 to 9.4) excess risk
on the previous day (lag 1 day). Lisabeth et al. (2008, 155939) used data on 2,350 confirmed cases
of ischemic stroke and 1,158 cases of TIA from the Brain Attack Surveillance in Corpus Christi
Project (BASIC), a population-based stroke surveillance project designed to capture all strokes in
Nueces County, Texas. The authors found a 6.0% (95% CI: -0.8 to 13.2) and 6.0% (95% CI: -1.8 to
14.4) excess risk of ischemic stroke/TIA per 10 ug/m3 increase in PM2.5 on the previous day and the
same day, respectively.
Only the study by Villeneuve et al. (2006, 090191) specifically evaluated the association
between ambient PM and the risk of TIAs. This study failed to find any evidence of an association
with either PM2 5 or PMi0.
A limitation of all of these studies is that they have assessed exposure based on the date of
hospital admission or ED presentation rather than the date and time of stroke symptom onset. It has
been shown that this can bias health effect estimates towards the null by up to 60% (Lokken et al.,
2009, 186774). Therefore, if there is a causal link between PM and the risk of stroke, it is likely that
the existing studies underestimate the true effects. Moreover, most of these studies have evaluated
only very short-term effects (lags of 0-2 days) and none have considered lags longer than 5 days. It is
possible that the lag structure of the association between PM and stroke differs from that of other
CVDs and it might even differ by stroke type.
Hemorrhagic Strokes
Most of the studies in the preceding section also evaluated the association between ambient
PM and the risk of hemorrhagic stroke (Chan et al., 2006, 090193: Henrotin et al., 2007, 093270:
Tsai et al., 2003, 080133: Villeneuve et al., 2006, 090191: Wellenius et al., 2005, 087483). In
Kaohsiung, Taiwan, Tsai et al. (2003, 080133) noted a 6.7% (95% CI: 4.2-9.4, lag 0-2 days) excess
risk of hospitalization for hemorrhagic stroke per 10 ug/m3 increase in PMi0, after excluding days
where the mean temperature was <20°C. However, in the U.S., Wellenius et al. (2005, 088685)
failed to find any association between ambient PMi0 levels and risk of hemorrhagic stroke among
Medicare beneficiaries in nine U.S. cities. Similarly, Villeneuve et al. (2006, 090191) found no
evidence of an association between ED visits for hemorrhagic stroke and either PM25 or PM10 levels
in Edmonton, Canada. Henrotin et al. (2007, 093270) found no evidence of an association between
risk of hospitalization and PMi0 levels in Dijon, France, and Chan et al. (2006, 090193) found no
evidence of an association between risk of hospitalization and either PM25 or PMi0levels in Taipei,
Taiwan.
In summary, large studies from the U.S., Europe, and Australia/New Zealand published since
the 2004 PM AQCD (U.S. EPA, 2004, 056905) provide inconsistent support for an association
between short-term increases in ambient levels of PM25 and PMi0 and risk of hospitalization and ED
visits for CBVD (Figure 6-4). Studies of PMi0_2.5 and CBVD or stroke have not been conducted. The
heterogeneity in results is likely partly attributed to: (1) differences in the sensitivity and specificity
of the various outcome definitions used in the studies; (2) lag structures between PM exposure and
stroke onset which may vary by stroke type and patient characteristics; and (3) exposure
misclassification due to the use of hospital admission date rather than stroke onset time, which may
vary by region, population characteristics, and stroke type. Effect estimates from multicity studies
and single-city U.S. and Canadian studies are included in Figure 6-4.
December 2009 6-71
-------
6.2.10.8. Peripheral Vascular Disease
In the U.S., the large MCAPS study Dominici et al. (2006, 088398) evaluated the association
between mean daily PM25 levels and the risk of hospitalization among elderly Medicare
beneficiaries in 204 urban counties and found that a 10 ug/m3 increase in PM2.5 was not significantly
associated with risk of hospitalization for peripheral vascular disease 0-2 days later. An earlier study
in Toronto (Burnett et al., 1999, 017269) found a negative association with PM25 (point estimate
and confidence intervals not reported), a positive statistically significant association with PMi0_25
(2.2% [95% CI: 0.1-4.3]), and a positive non-significant association with PM10 (0.5% [95% CI: -0.5
to 1.6]). The Atlanta-based SOPHIA study (Metzger et al., 2004, 044222) of ED visits grouped
visits for PVD with those for CBVD, making interpretation of these results challenging.
Study
Dominici et al. (2006, 088398)
Metzger et al. (2004, 044222)*
Delfinoetal. (2009. 191994)
Metzger et al. (2004, 044222)*
LeTertreetal. (2003, 042820)
Larrieuetal. (2007, 093031)
Villeneuve et al* (2006, 090191)
Welleniusetal. (2005. 088685)
Lisabethetal. (2008. 155939)
Villeneuve et al. (2006, 090191)*
Linn etal. (2000.002839)
Villeneuve et al. (2006, 090191)*
Wellenius et al. (2005, 088685)
Villeneuve et al. (2006, 090191)'
Age Location Lag
65+ 204 counties, U.S. 0
Northeast, U.S. 0
Southeast, U.S. 0
Midwest, U.S. 0
South, U.S. 0
All Ages Atlanta, GA 0-2
All Ages 6 counties, CA 0-1
All Ages Atlanta, GA 0-2
8 Cities, Europe 0-1
65+ 8 Cities, France 0-1
All Ages 8 Cities, France 0-1
65+ Edmonton, Canada 0-2
0-2
0-2
0-2
65+ 9 Cities, U.S. 0
All Ages Nueces County, TX 1
0
65+ Edmonton, Canada 0-2
0-2
0-2
0-2
>X Los Angeles, CA 0
65+ Edmonton, Canada 0-2
0-2
65+ 9 Cities, U.S. 0
65+ Edmonton, Canada 0-2
0-2
Outcome Effect Estimate (95% CI)
CBVD
CBVD
CBVD -I
CBVD
CBVD
CBVD/PVD
CBVD, Wildfire
CBVD/PVD
CBVD 4
CBVD -I
CBVD —I
IS roo\ c •
TIA Cool < *
IS
IS/TIA
IS/TIA
IS Cool •
iq \Afarm I
IS
H° CYinl f
HS 1
HS Cool <
HS Warm
-* PM2.5
L-»
*•
_,_
. PMio
\
-t—
t_
PM"r
* PMio
••
PM2.5
*
— PMio
_ '
i i i \ \ i i i t i i i \ i i i
-12 -8 -4 0 4 8 12 16 20
* ED Visits
Figure 6-4.
Excess Risk Estimate
Excess risk estimates per 10 ug/m3 increase in 24-h avg PM2.s and
concentration for CBVD ED visits and HAs. Studies represented in the
figure include all multicity studies as well as single-city studies conducted in the
U.S. and Canada.
In summary, there is insufficient published data to determine whether or not there may be an
association between short-term increases in ambient levels ofs PM2 5, PM10_2.5s or PM10 and increased
risk of hospitalization and ED visits for PVD.
6.2.10.9. Copollutant Models
Relatively few studies have evaluated the effects of PM25 and PM10_2.5 on the risk of hospital
admissions and ED visits in the context of two-pollutant models. Generally, results for health effects
of both size fractions are similar even after controlling for SO2 or O3 levels (Figure 6-5). However,
December 2009
6-72
-------
controlling for NO2 or CO has yielded mixed results. Among the large multicity studies, the Atlanta-
based SOPHIA study found that the association between PM2.5 (total carbon) and risk of
cardiovascular ED visits was somewhat attenuated in two-pollutant models additionally controlling
for either CO or NO2 (Tolbert et al, 2007, 090316). Barnett et al. (2006, 089770) found that the
associations they observed between PM2.5 and cardiac hospitalizations in Australia and New Zealand
were attenuated after control for 24-h NO2, but not after control for CO.
Only a few studies have attempted to evaluate the effects of one PM size fraction while
controlling for another PM size fraction. The large U.S. MCAPS study evaluating the effects of
PMio_2 5 on cardiovascular hospital admissions lost precision after controlling for PM2 5, but did not
consider gaseous pollutants (Peng et al., 2008, 156850). Andersen et al. (2008, 189651) found that
associations between both PMi0 and PM2 5 and cardiovascular hospitalizations in Copenhagen were
not attenuated by control for particle number concentration.
A number of studies have also evaluated PMi0 effects in the context of two-pollutant models
with inconsistent results. The multicity Spanish EMECAS study (Ballester et al., 2006, 088746)
found that the statistically significant positive associations observed between PMi0 and cardiac
hospitalizations were robust to control for other pollutants in two-pollutant models. Jalaludin et al.
(2006, 189416) found that the effects of PMi0 as well as PM25 on cardiovascular ED visits in Sydney
Australia were attenuated by additional control for either NO2 or CO. Wellenius et al. (2005,
087483) found that the PM10-related risk of hospitalization for CHF in Allegheny County, PA, was
attenuated in two-pollutant models controlling for either CO or NO2. In contrast, Chang et al. (2005,
080086) examined CVD hospitalizations in Taipei and found attenuation of PMi0 effects by control
for NO2 or CO, but only during warm days. In Kaohsiung, Taiwan, Tsai et al. (2003, 080133) found
that the association between PMi0 and ischemic stroke hospitalizations was not materially attenuated
in two-pollutant models controlling for either NO2 or CO.
The inconsistent findings after controlling for gaseous pollutants or other size fractions are
likely due to differences in the correlation structure among pollutants, as well as differing degrees of
exposure measurement error.
December 2009 6-73
-------
Study
Tolbertetal. (2007.090316)
Barnettetal (2006 089770)
Villeneuve et al (2006 090191)
Burnett etal (1997 084194)
Ito (2003 042856)
Moolgavkar (2003, 051316)
Jalaludin etal. (2006. 189416)
Peng etal. (2008. 156850)
Andersen etal (2008 189651)
Ito (2003 042856)
Burnett etal (1997 084194)
Peng etal. (2008. 156850)
Outcome
CVD
Heart Disease
TIE
Heart Disease
CHF
IHD
CVD
CVD
CVD
CVD
IHD
CHF
CVD
Pollutant
PM25 —I
PM25TC
PM25TC+N02
PM25
PM25+CO
PM-- <
PH-r
PH-- <
PM'r
pu"4.Mn
pM^ioV
pu'J+rn'
PM25
pM,,+n,
PM25+S02
PM25+N02
PM25+CO
PM-C
PM- -"•«-!-
PM25+N02
PM25
PM25+CO J
PM25
PM25+N02
PM25+N02
PM25+N02
PM25
PM25+PM10-25
py,.
UFP(Number)
PMm-_
PMnn-ic+Oo
PM10-25+N02
PM1° 25
PM,,, ,,+NO, i
PMnn-oc
PM |0 2r+CO
PMnrzs
PMlo-15+PM2.5
I I I
-6 -4 -2 C
Effect Estimate (95% Cl)
PM2 5 Adjusted for Gases and Other Size Fractions
u —
^
_
^
^
*
_, —
__- —
X
_
PMm ; Adjusted for Gases and Other Size Fractions
_
^
*
I I I I I I I I
) 2 4 6 8 10 12 14 16
Excess Risk (%)
Figure 6-5. Excess risk estimates per 10 ug/m increase in 24-h avg PM2.6, and PMi0.2.s for
cardiovascular disease ED visits or HAs, adjusted for co-pollutants.
December 2009
6-74
-------
6.2.10.10. Concentration Response
The concentration-response relationship has been extensively analyzed primarily through
studies that examined the relationship between PM and mortality. These studies, which have focused
on short- and long-term exposures to PM have consistently found no evidence for deviations from
linearity or a safe threshold (Daniels et al., 2004, 087343; Samoli et al., 2005, 087436; Schwartz,
2004, 078998; Schwartz et al., 2008, 156963) (Sections 6.5.2.7 and 7.1.4). Although on a more
limited basis, studies that have examined PM effects on cardiovascular hospital admissions and ED
visits have also analyzed the PM concentration-response relationship, and contributed to the overall
body of evidence which suggests a log-linear, no-threshold PM concentration-response relationship.
The evaluation of cardiovascular hospital admission and ED visit studies in 2004 PM AQCD
(U.S. EPA, 2004, 056905) found no evidence for a threshold in the dose-response relationship
between short-term exposure to PMi0and IHD hospital admissions (Schwartz and Morris, 1995,
046186). An evaluation of recent single- and multicity studies of hospital admission and ED visits
for CVD further supports this finding.
Ballester et al. (2006, 088746) examined the linearity of the relationship between air
pollutants (including PM10) and cardiovascular hospital admissions in 14 Spanish cities within the
EMECAM project. In this exploratory analysis, the authors examined the models used when
pollutants were added in either a linear or non-linear way (i.e., with a spline smoothing function) to
the model. Although the study does not present the results for each of the pollutants evaluated
individually, overall Ballester et al. (2006, 088746) found that the shape of the pollutant-
cardiovascular hospital admission relationship was most compatible with a linear curve. Wellenius
et al. (2005, 087483) conducted a similar analysis when examining the relationship between PM10
and CHF hospital admissions among Medicare beneficiaries. The authors examined the assumption
of linearity using fractional polynomials and linear splines. The results of both approaches
contributed to Wellenius et al. (2005, 087483) concluding that the assumption of linearity between
the log relative risk of cardiovascular hospital admissions and PM concentration was reasonable.
Unlike the aforementioned studies that examined the linearity in the concentration-response
curve as part of the model selection process (i.e., to determine the most appropriate model to use to
examine the relationship between PM and cardiovascular hospital admissions and ED visits),
Zanobetti and Schwartz (2005, 088069) conducted an extensive analysis of the shape of the
concentration-response curve and the potential presence of a threshold when examining the
association between PMi0 and MI hospital admissions among older adults in 21 U.S. cities. The
authors examined the concentration-response curve by fitting a piecewise linear spline with slope
changes at 20 and 50 ug/m3. This approach resulted in an almost linear concentration-response
relationship between PMi0 and MI hospital admissions with a steeper slope occurring below
50 ug/m3 (Figure 6-6). Additionally, Zanobetti and Schwartz (2005, 088069) found no evidence for a
threshold.
December 2009 6-75
-------
7.0
i6-"
•- 5.0
05
Cfl
| 4.0
o
.= 3.0
8 2.0
<2 1,0
0.0
0
10
I
20
I
30
1
40
I
50
80 70 80
Source: Zanobetti and Schwartz (2005, 0880691.
Figure 6-6. Combined random-effect estimate of the concentration-response relationship
between Ml emergency hospital admissions and PMi0, computed by fitting a
piecewise linear spline, with slope changes at 20 ug/m3 and 50 ug/m3.
Overall, the limited evidence from the studies that examined the concentration-response
relationship between PM and cardiovascular hospital admissions and ED visits supports a no-
threshold, log-linear model, which is consistent with the observations made in studies that examined
the PM-mortality relationship (Section 6.5.2.7).
6.2.10.11. Out of Hospital Cardiac Arrest
One study of out of hospital cardiac death conducted in Seattle, WA (Checkoway et al., 2000,
015527). which reported no association with PM was included in the 2004 PM AQCD (U.S. EPA,
2004, 056905). In the U.S., the survival rate of sudden cardiac arrest is less than 5%. In addition, as
discussed in Section 6.5, Zeka et al. (2006, 088749) found that the estimated mortality risk due to
short-term exposure to PM10 was much higher for out-of-hospital cardiovascular deaths than for
in-hospital cardiovascular deaths. The analysis of studies that examine the association between PM
and cardiac arrest could provide evidence for an important link between the morbidity and mortality
effects attributed to PM.
Sullivan et al. (2003, 043156) examined the association between the incidence of primary
cardiac arrest and daily measures of PM2.5 (measured by nephelometry) using a case-crossover
analysis of 1,206 Washington State out-of-hospital cardiac arrests (1985-1994) among persons with
(n = 774) and without (n = 432) clinically recognized heart disease. The authors examined PM
associations at 0- through 2-day lags using the time-stratified referent sampling scheme (i.e., the
same day of the week and month of the same year). The estimated relative risk for a 13.8-ug/m3
increase in 1-day lag PM2.5 (nephelometry: IQR = 0.54 10"1 km"1 bsp) was 0.94 (95% CI: 0.88-1.02),
or 0.96 (95% CI: 0.91-1.0) per 10 ug/m3 increase. Similar estimates were reported for 0- or 2-day
lags. The presence or absence of clinically recognized heart disease did not alter the result. This
finding is consistent with the previous study of cardiac arrest in Seattle (Levy et al., 2001, 017171)
that reported no PM association. It is also consistent with the Sullivan et al. (2005, 050854) analysis
of PM and onset of MI, and the Sullivan et al. (2007, 100083) analysis of PM and blood markers of
inflammation in the elderly population, both of which were conducted in Seattle. Note also that the
analysis of the NMMAPS data for the years 1987-1994 also found no PMi0 association for all-cause
mortality in Seattle. Overall, the results of studies conducted in Seattle consistently found no
association between PM and cardiovascular outcomes or all-cause mortality.
December 2009
6-76
-------
Rosenthal et al. (2008, 156925) examined associations between PM2.5 and out-of-hospital
cardiac arrests in Indianapolis, Indiana for the years 2002-2006 using a case-crossover design with
time-stratified referent sampling. Using all the cases (n = 1,374), they found no associations between
PM2.5 and cardiac arrest in any of the 0- through 3-day lags or multiday averages thereof (e.g., for
0-day lag, OR =1.02 [CI: 0.94-1.11] per 10 (ig/m3 increase in PM2.5). However, for cardiac arrests
witnessed by bystanders (n = 511), they found a significant association with PM25 exposure (by
TEOM, corrected with FRM measurements) during the hour of the arrest (OR =1.12 [CI: 1.01-1.25]
per 10 (ig/m3 increase in PM25), and even larger risk estimates for older adults (age 60-75) or those
that presented with asystole. There have been very few PM studies that used hourly PM
measurements, and further studies are needed to confirm associations at such time scales.
In Rome, Forastiere et al. (2005, 086323) examined associations between air pollution (PNC,
PMio, CO, NO2, and O3) and out-of-hospital coronary deaths (n = 5,144) for the study period of
1998-2000. A case-crossover design with the time-stratified referent sampling was used to examine
the pollution indices at lag 0- through 3 days and the average of 0-1 lags. They found associations
between deaths and PNC (lag 0 and 0-1), PMi0 (lag 0, 1, and 0-1), and CO (lag 0 and 0-1) but not
with NO2 or O3. The risk estimate for 0-day lag PMi0 was 1.59% (CI: 0.03-3.18) per 10 ug/m3
increase. The older adults (65-74 and >75 age groups) showed higher risk estimates than the younger
(35-64) age group. Because PNC is considered to be associated with UFPs, and CO was also
associated with out-of-hospital cardiac arrests, combustion sources were implicated.
In summary, only a few studies have examined out-of-hospital cardiac arrest or deaths. The
two studies from Seattle, WA consistently found no association (also consistent with other cardiac
effects and mortality studies conducted in that locale); a study in Indianapolis, IN found an
association with hourly PM2 5 but not daily PM2 5; and a study in Rome found an association with
PMio, but also with PNC and CO. Because multicity mortality studies examining this association
found heterogeneity in PM risk estimates across regions, future studies of out-of-hospital cardiac
arrest will need to consider location and the air pollution mixture during their design. Mean and
upper percentile concentrations are found in Table 6-9.
Table 6-9. PM concentrations reported in studies of out-of-hospital cardiac arrest.
Author
Location
Mean Concentration (ug/m3
PM2.5
Sullivan et al. (2003, 043156)
Washington State
Nephelometry:0.71 x10"1 km"1 bsp Maximum: 5.99 x 10"1 km"1 bsp
Rosenthal et al. (2008,156925) Indianapolis, Indiana
NR
NR
PM,,
Sullivan et al. (2003, 043156)
Washington State
28.05
89.83
Zeka et al. (2006, 088749)
Range in Means: 15.9 (Honolulu) - MR
37.5 (Cleveland) INK
Forastiere etal. (2005,086323) Rome, Italy
52.1
75th: 65.7
6.2.11. Cardiovascular Mortality
An evaluation of studies that examined the association between short-term exposure to PM2 5
and PMio_2.5 and mortality provides additional evidence for PM-related cardiovascular health effects.
Although the primary analysis in the majority of mortality studies evaluated consists of an
examination of the relationship between PM2 5 or PMi0_2.5 and all-cause (nonaccidental) mortality,
some studies have examined associations with cause-specific mortality including cardiovascular-
related mortality.
December 2009
6-77
-------
Multicity mortality studies that examined the PM2 5-cardiovascular mortality relationship on a
national scale (Franklin et al. (2007, 091257) - 27 U.S. cities; Franklin et al. (2008, 155779) - 25
U.S. cities; and Zanobetti and Schwartz (2009, 188462) - 112 U.S. cities) have found consistent
positive associations between short-term exposure to PM2 5 and cardiovascular mortality ranging
from 0.47 to 0.85% per 10(ig/m3 at lag 0-1 (Section 6.5). The associations observed on a national
scale are consistent with those presented by Ostro et al. (2006, 087991) in a study that examined the
PM2 ^-mortality relationship in nine California counties (0.6% [95% CI: 0-1.1] per 10 (ig/m3). Of the
multicity studies evaluated, one examined single day lags and found evidence for slightly larger
effects at lag 1 compared to the average of lag days 0 and 1 for cardiovascular mortality (94%
[95% CI: -0.14 to 2.02] per 10 ug/m3) (Franklin et al., 2007, 091257). Although the overall effect
estimates reported in the multicity studies evaluated are consistently positive, it should be noted that
a large degree of variability exists between cities when examining city-specific effect estimates
potentially due to differences between cities and regional differences in PM2 5 composition (Figure
6-24). Only a limited number of studies that examined the PM2 5-mortality relationship have
conducted analyses of potential confounders, such as gaseous copollutants, and none examined the
effect of copollutants on PM25 cardiovascular mortality risk estimates. Although the recently
evaluated multicity studies did not extensively examine whether PM2 5 mortality risk estimates are
confounded by gaseous copollutants, evidence from the limited number of single-city studies
evaluated in the 2004 PM AQCD (U.S. EPA, 2004, 056905) suggests that gaseous copollutants do
not confound the PM2 5-cardiovascular mortality association. This is further supported by studies that
examined the PMi0-mortality relationship in both the 2004 PM AQCD (U.S. EPA, 2004, 056905)
and this review. The evidence from epidemiologic, controlled human exposure, and toxicological
studies that examined the association between short-term exposure to PM2 5 and cardiovascular
morbidity provide coherence and biological plausibility for the cardiovascular mortality effects
observed. Overall, the cardiovascular mortality PM2 5 effects were similar to those reported for all-
cause (nonaccidental) mortality (Section 6.5), and are consistent with the effect estimates observed
in the single- and multicity studies evaluated in the 2004 PM AQCD (U.S. EPA, 2004, 056905).
Zanobetti and Schwartz (2009, 188462) also examined PMi0_2.5 mortality associations in 47
U.S. cities and found evidence for cardiovascular mortality effects (0.32% [95% CI: 0.00-0.64] per
10 (ig/m at lag 0-1) similar to those reported for all-cause (nonaccidental) mortality (0.46% [95% CI:
0.21-0.67] per 10 (ig/m). In addition, Zanobetti and Schwartz (2009, 188462) reported seasonal (i.e.,
larger in spring and summer) and regional differences in PMi0_2.5 cardiovascular mortality risk
estimates. A few single-city studies evaluated also reported associations, albeit somewhat larger than
the multicity study, between PMi0_25 and cardiovascular mortality in Phoenix, AZ (Wilson et al.,
2007, 157149) (3.4-6.6% at lag 1) and Vancouver, Canada (Villeneuve et al., 2003, 055051) (5.4%
at lag 0). The difference in the PMi0_2.5 risk estimates observed between the multi- and single-city
studies could be due to a variety of factors including differences between cities and compositional
differences in PMi0_2.5 across regions (Figure 6-29). Only a small number of studies have examined
potential confounding by gaseous copollutants or the influence of model specification on PMi0_2.s
mortality risk estimates, but the effects are relatively consistent with those studies evaluated in the
2004 PM AQCD (U.S. EPA, 2004, 056905).
6.2.12. Summary and Causal Determinations
6.2.12.1. PM2.5
Several studies cited in the 2004 AQCD reported positive associations between short-term
PM2 5 concentrations and hospital admissions or ED visits for CVD, although few were statistically
significant. In addition, U.S. and Canadian-based studies (both multi- and single-city) that examined
the PM2 ^-mortality relationship reported associations for cardiovascular mortality consistent with
those observed for all-cause (nonaccidental) mortality and relatively stronger than those for
respiratory mortality. Significant associations were also observed between MI and short-term PM2 5
concentrations (averaged over 2 or 24 h), as well as decreased HRV in association with PM2 5.
Several controlled human exposure and animal toxicological studies demonstrated HRV effects from
exposure to PM2 5 CAPs, as well as changes in blood coagulation markers. However, the effects in
these studies were variable. Arrhythmogenesis was reported for toxicological studies and generally
December 2009 6-78
-------
these results were observed in animal models of disease (SH rat, MI, pulmonary hypertension)
exposed to combustion-derived PM2.5 (i.e., ROFA, DE, metals). One study demonstrated significant
vasoconstriction in healthy adults following controlled exposures to CAPs, although this response
could not be conclusively attributed to the particles as subjects were concomitantly exposed to
relatively high levels of O3. The results reported for systemic inflammation in toxicological studies
were mixed.
A large body of evidence from studies of the effect of PM2.5 on hospital admissions and ED
visits for CVD has been published since the 2004 PM AQCD. Associations with PM2 5 are
consistently positive with the majority of studies reporting increases in hospital admissions or ED
visits ranging from 0.5 to 3.4% per 10 ug/m3 increase in PM25 (Section 6.2.10). The largest U.S-
based multicity study, MCAPs, reported excess risks in the range of approximately 0.7% with the
largest excess risks in the Northeast (1.08%) and in the winter (1.49%), providing evidence of
regional and seasonal heterogeneity (Bell et al., 2008, 156266; Dominici et al, 2006, 088398).
Weak or null findings for PM2 5 have been observed in two single-city studies both conducted in
Washington state (Slaughter et al., 2005, 073854; Sullivan et al., 2007, 100083) and may be
explained by this heterogeneity. Weak associations were also reported in Atlanta for PM2 5 and CVD
ED visits, with PM2 5 traffic components being more strongly associated with CVD ED visits than
other components (Tolbert et al., 2007, 090316). Multicity studies conducted outside the U.S. and
Canada have shown positive associations with PM2 5. Studies of specific CVD outcomes indicate that
IHD and CHF may be driving the observed associations (Sections 6.2.10.3 and 6.2.10.5,
respectively). Although estimates from studies of cerebrovascular diseases are less precise and
consistent, ischemic diseases appear to be more strongly associated with PM2 5 compared to
hemorrhagic stroke (Section 6.2.10.7). The available evidence suggests that these effects occur at
very short lags (0-1 days), although effects at longer lags have rarely been evaluated. Overall, the
results of these studies provide support for associations between short-term PM25 exposure and
increased risk of cardiovascular hospital admissions in areas with mean concentrations ranging from
7 to 18 ug/m3.
Epidemiologic studies that examined the association between PM2 5 and mortality provide
additional evidence for PM25-related cardiovascular effects (Section 6.2.11). The multicity studies
evaluated found consistent, precise positive associations between short-term exposure to PM2 5 and
cardiovascular mortality ranging from 0.47 to 0.85% at mean 24-h avg PM25 concentrations above
13 ug/m3. These associations were reported at short lags (0-1 days), which is consistent with the
associations observed in the hospital admission and ED visit studies discussed above. Although only
a limited number of studies examined potential confounders of the PM2 5-cardiovascular mortality
relationship, the studies evaluated in both this review and the 2004 PM AQCD (U.S. EPA, 2004,
056905) support an association between short-term exposure to PM2 5 and cardiovascular mortality.
Recent studies that apportion ambient PM2 5 into sources and components suggest that
cardiovascular hospital admissions associated with PM2 5 may be attributable to traffic-related
pollution and, in some cases, biomass burning (Section 6.2.10). Further supporting evidence is
provided by studies that have used PMi0 collection filters (median diameter generally <2.5 um) to
identify combustion- or traffic-related sources associated with cardiovascular hospital admissions.
Metals have also been implicated in these effects (Bell et al., 2009, 191997). A limited number of
older publications have reported that particle acidity of PM25 is not more strongly associated with
CVD hospitalizations or ED visits than other PM metrics.
Changes in various measures of cardiovascular function have been demonstrated by multiple
independent laboratories following controlled human exposures to different types of PM2 5. The most
consistent effect is changes in vasomotor function, which has been demonstrated following exposure
to CAPs and DE. The majority of the new evidence of particle-induced changes in vasomotor
function comes from studies of exposures to DE (Section 6.2.4.2). None of these studies have
evaluated the effects of DE with and without a particle trap. Therefore, the changes in vasomotor
function cannot be conclusively attributed to the particles in DE as subjects are also concomitantly
exposed to relatively high levels of NO2, NO, CO, and hydrocarbons. However, it is important to
note that a study by Peretz et al. (2008, 156854) used a newer diesel engine with lower gaseous
emissions and reported significant DE-induced decreases in BAD. In addition, increasing the particle
exposure concentration from 100 to 200 ug/m3, without proportional increases in NO, NO2, or CO,
resulted in an approximate 100% increase in response. An additional consideration is that, while
fresh DE used in these studies contains relatively high concentrations of PM25, the MMAD is
typically < 100 nm, which makes it difficult to determine whether the observed effects are due to
December 2009 6-79
-------
PM2.5 or, more specifically, due to the UF fraction. Further evidence of a particle effect on vasomotor
function is provided by significant changes in BAD demonstrated in healthy adults following
controlled exposure to CAPs with O3 (Brook et al., 2002, 024987). These findings are consistent
with epidemiologic studies of various measures of vasomotor function (e.g., FMD and BAD were
the most common), which have demonstrated an association with short-term PM2.5 concentration in
healthy and diabetic populations (Section 6.2.4.1). A limited number of epidemiologic studies
examined multiple lags and the strongest associations were with either the 6-day mean concentration
(O'Neill et al., 2005, 088423) or the concurrent day (Schneider et al., 2008, 191985).
The toxicological findings with respect to vascular reactivity are generally in agreement and
demonstrate impaired dilation following PM2 5 exposure that is likely endothelium dependent
(Section 6.2.4.3). These effects have been demonstrated in varying vessels and in response to
different PM2.5 types, albeit using IT instillation exposure in most studies. Further support is
provided by IT instillation studies of ambient PM10 that also demonstrate impaired vasodilation and a
PM25 CAPs study that reported decreased L/W ratio of the pulmonary artery. An inhalation study of
Boston PM2 5 CAPs reported increases in coronary vascular resistance during ischemia, which
indicated a possible role for PM-induced coronary vasoconstriction. The mechanism behind impaired
dilation following PM exposure may include increased ROS and RNS production in the
microvascular wall that leads to altered NO bioavailability and endothelial dysfunction. Despite the
limited number of inhalation studies conducted with concentrations near ambient levels, the
toxicological studies collectively provide coherence and biological plausibility for the myocardial
ischemia observed in controlled human exposure and epidemiologic studies.
Consistent with the observed effects on vasomotor function, one recent controlled human
exposure study reported an increase in exercise-induced ST-segment depression (a potential indicator
of ischemia) during exposure to DE in a group of subjects with prior MI (Mills et al., 2007,
091206). In addition, toxicological studies from Boston that employed CAPs provide further
evidence for PM2 5 effects on ischemia, with changes in ST-segment and decreases in total
myocardial blood flow reported (Section 6.2.3.3). These findings from toxicological and controlled
human exposure studies provide coherence and biological plausibility for the associations observed
in epidemiologic studies, particularly those of increases in hospital admissions and ED visits for
IHD. Several epidemiology studies have reported associations between short-term PM2 5
concentration (including traffic sources or components such as BC) and ST-segment depression or
abnormality (Section 6.2.3.1).
Toxicological studies provide biological plausibility for the PM2 5 associations with CHF
hospital admissions by demonstrating increased right ventricular pressure and diminished cardiac
contractility in rodents exposed to CB and DE (Section 6.2.6.1). Similarly, increased coronary
vascular resistance was observed following PM2 5 CAPs exposure in dogs with
experimentally-induced ischemia. Further, a recent epidemiology study reported small but
statistically significant decreases in passively monitored diastolic pressure and right ventricular
diastolic pressure (Rich et al., 2008, 156910).
In addition to the effects of PM on vasomotor response, there is a growing body of evidence
that demonstrates changes in markers of systemic oxidative stress following controlled human
exposures to DE, wood smoke, and urban traffic particles. However, these effects may be driven in
part by the UF fraction of PM25. Toxicological studies provide evidence of increased cardiovascular
ROS following PM2 5 exposure to CAPs, road dust, CB, and TiO2, as well as increased systemic ROS
in rats exposed to gasoline exhaust (Section 6.2.9.3). Epidemiologic studies of markers of oxidative
stress (e.g., tHcy, CuZn-SOD, TEARS, 8-oxodG, oxLDL and MDA) are consistent with these
toxicological findings (Section 6.2.9.1).
A few epidemiologic studies of ventricular arrhythmias recorded on ICDs that were conducted
in Boston and Sweden (Table 6-2) found associations with short-term PM2 5 concentration (also BC
and sulfate). While Canadian and U.S. studies conducted outside of Boston did not find positive
associations between PM2 5 and ICD recorded ventricular arrhythmias, several such studies observed
associations with ectopic beats and runs of supraventricular or ventricular tachycardias (Section
6.2.2.1). Toxicological studies also provide limited evidence of arrhythmia, mainly in susceptible
animal models (i.e., older rats, rats with CHF) (Section 6.2.2.2).
Most epidemiologic studies of HRV have reported decreases in SDNN, LF, HF, and rMSSD
(Section 6.2.1.1). While there are also a significant number of controlled human exposure studies
reporting PM-induced changes in HRV, these changes are often variable and difficult to interpret
(Section 6.2.1.2). Similarly, HRV increases and decreases have been observed in animal
December 2009 6-80
-------
toxicological studies that employed CAPs or CB (Section 6.2.1.3). In a study in mice, resuspended
soil, secondary sulfate, residual oil, and motor vehicle/other sources, as well as Ni were implicated in
HRV effects (Lippmann et al., 2006, 091165). Further, cardiac oxidative stress has been implicated
as a consequence of ANS stimulation in response to CAPs. Modification of the PM-HRV association
by genetic polymorphisms related to oxidative stress has been observed in a series of analyses of the
population enrolled in the Normative Aging Study. Changes in HRV measures (whether increased or
decreased) are likely to be more useful as indicators of PM exposure rather than predictive of some
adverse outcome. Furthermore, the HRV result may be reflecting a fundamental response of an
individual that is determined in part by a number of factors including age and pre-existing
conditions.
Although not consistently observed across studies, some investigators have reported
PM2.5-induced changes in BP, blood coagulation markers, and markers of systemic inflammation in
controlled human exposure studies (Sections 6.2.5.2, 6.2.8.2, and 6.2.9.2, respectively). Findings
from epidemiologic studies, which are largely cross-sectional and measure a wide array markers of
inflammation and coagulation, are not consistent; however, a limited number of recent studies of
gene-environment interactions offer insight into potential individual susceptibility to these effects
(Ljungman et al., 2009, 191983: Peters et al., 2009, 191992). Similarly, toxicological studies
demonstrate mixed results for systemic inflammatory markers and generally indicate relatively little
change at 16-20 h post-exposure (Section 6.2.7.3). Increases in BP have been observed in
toxicological studies (Section 6.2.5.3), with the strongest evidence coming from dogs exposed to
PM2.5 CAPs. For blood coagulation parameters, the most commonly reported change in animal
toxicological studies is elevated plasma fibrinogen levels following PM2 5 exposure, but this
response is not consistently observed (Section 6.2.8.3).
In summary, associations of hospital admissions or ED visits with PM2.5 for CVD
(predominantly IHD and CHF) are consistently positive with the majority of studies reporting
increases ranging from 0.5 to 3.4% per 10 ug/m increase in PM25. Seasonal and regional variation
observed in the large multicity study of Medicare recipients is consistent with null findings reported
in several single city studies conducted in the Western U.S. The results from the hospital admission
and ED visit studies are supported by the associations observed between PM2 5 and cardiovascular
mortality, which also provide additional evidence for regional and seasonal variability in PM2 5 risk
estimates. Changes in various measures of cardiovascular function that may explain these
epidemiologic findings have been demonstrated by multiple independent laboratories following
controlled human exposures to different types of PM25. The most consistent PM25 effect is for
vasomotor function, which has been demonstrated following exposure to CAPs and DE.
Toxicological studies finding reduced myocardial blood flow during ischemia and altered vascular
reactivity provide coherence and biological plausibility for the myocardial ischemia that has been
observed in both controlled human exposure and epidemiologic studies. Further, PM2 5 effects on ST-
segment depression have been observed across disciplines. In addition to ischemia, PM2 5 may act
through several other pathways. Plausible biological mechanisms (e.g., increased right ventricular
pressure and diminished cardiac contractility) for the associations of PM25 with CHF have also been
proposed based on toxicological findings. There is a growing body of evidence from controlled
human exposure, toxicological and epidemiologic studies demonstrating changes in markers of
systemic oxidative stress with PM2 5 exposure. Inconsistent effects of PM on BP, blood coagulation
markers and markers of systemic inflammation have been reported across the disciplines. Together,
the collective evidence is sufficient to conclude that a causal relationship exists between
short-term PM25 exposures and cardiovascular effects
6.2.12.2. PM10.2.5
There was little evidence in the 2004 AQCD regarding PMi0_2.5 cardiovascular health effects.
Two single-city epidemiologic studies found positive associations of PMi0_25 with cardiovascular
hospital admissions in Toronto (Burnett et al., 1999, 017269) and Detroit, MI (Ito, 2003, 042856;
Lippmann, 2000, 024579) and the effect estimates were of the same general magnitude as for PMi0
and PM2 5. Both studies reported positive associations and estimates appeared robust to adjustment
for gaseous copollutants in two-pollutant models. An imprecise, non-significant association between
PMio_2.5 and onset of MI was observed in Boston (Peters et al., 2001, 016546). No controlled human
exposure or toxicological studies of PMi0_2.5 were presented in the 2004 AQCD.
December 2009 6-81
-------
Several recent epidemiologic studies of the effect of ambient PM10_2.5 concentration on hospital
admissions or ED visits for CVD were conducted (Section 6.2.10). In a study of Medicare patients in
108 U.S. counties, Peng et al. (2008, 156850) reported a significant association between PMi0_2.5 and
CVD hospitalizations in their single pollutant model. In a study of six French cities, Host et al.
(2008, 155852) reported a significant increase in IHD hospital admissions in association with PMi0_
25. In contrast, associations of cardiovascular outcomes with PMi0_25 were weak for CHF and null
for IHD in the Atlanta-based SOPHIA study (Metzger et al., 2004, 044222). Results from single-city
studies are generally positive, but effect sizes are heterogeneous and estimates are imprecise
(Section 6.2.10). Crustal material from a dust storm in the Gobi desert that was largely coarse PM
(generally indicated using PMi0) was associated with hospitalizations for CVD, including IHD and
CHF in most studies (Section 6.2.10). Mean PMi0_2.5 concentrations in the hospital admission and
ED visit studies ranged from 7.4-13 ug/m3. A few epidemiologic studies that examined the
association between short-term exposure to PM10_2.5 and cardiovascular mortality (Section 6.2.11)
provide supporting evidence for the hospital admission and ED visit studies at similar 24-h avg
PMio_2.5 concentrations (i.e., 6.1-16.4 ug/m3). Amulticity study reported risk estimates for
cardiovascular mortality of similar magnitude to those for all-cause (nonaccidental) mortality
(Zanobetti and Schwartz, 2009, 188462). However, the single-city studies evaluated (Villeneuve et
al., 2003, 055051; Wilson et al., 2007, 157149) reported substantially larger effect estimates, but this
could be due to differences between cities and compositional differences in PM10_2.5 across regions.
Of note is the lack of analyses within the studies evaluated that examined potential confounders of
the PMio_2.5-cardiovascular mortality relationship.
The U.S. study of Medicare patients (Peng et al., 2008, 156850) and the multicity study that
examined the association between PMi0_2.5 and mortality (Zanobetti and Schwartz, 2009, 188462)
were the only studies to adjust PMi0_2.5 for PM2.5. Peng, et al. (2008, 156850) found that the PMi0_2.5
association with CVD hospitalizations remained, but diminished slightly after adjustment for PM2.5.
These results are consistent with those reported by Zanobetti and Schwartz (2009, 188462). which
found PMio_2.5-cardiovascular mortality risk estimates remained relatively robust to the inclusion of
PM2.5 in the model. Because of the greater spatial heterogeneity of PMi0_2.5, exposure measurement
error is more likely to bias health effect estimates towards the null for epidemiologic studies of
PMio_2.5 versus PMi0 or PM2.5, making it more difficult to detect an effect of the coarse size fraction.
In addition, models that include both PM10_2.5 and PM2.5 may suffer from instability due to
collinearity. Further, the lag structure of PMi0_2.5 effects on risk of cardiovascular hospital admissions
and ED visits, as well as mortality, has not been examined in detail.
Several epidemiologic studies of cardiovascular endpoints including HRV, BP, ventricular
arrhythmia, and ECG changes indicating ectopy or ischemia were conducted since publication of the
2004 PM AQCD. Supraventricular ectopy and ST-segment depression were associated with PMi0_2.5
(Section 6.2.3.1), and the only study to examine the effect of PM10_25 on BP reported a decrease in
SBP (Ebelt et al., 2005, 056907) (Section 6.2.5.1). HRV findings were mixed across the
epidemiologic studies (Section 6.2.1.1). A limited number of studies have evaluated the effect of
controlled exposures to PMi0_2.5 CAPs on cardiovascular endpoints in human subjects. These studies
have provided some evidence of decreases in HRV (SDNN) and tPA concentration among healthy
adults approximately 20 hours following exposure (Section 6.2.1.2). However, it is important to note
that no other measures of HRV (e.g., LF, HF, or LF/HF), nor other hemostatic or thrombotic markers
(e.g., fibrinogen) were significantly affected by particle exposure in these studies.
There are very few toxicological studies that examined the effect of exposure to PMi0_2.5 on
cardiovascular endpoints or biomarkers in animals. The few studies that evaluated cardiovascular
responses were comparative studies of various size fractions, and only blood or plasma parameters
were measured (Sections 6.2.7.3 and 6.2.8.3). These studies used IT instillation methodologies, as
there are challenges to exposing rodents via inhalation to PM 10.2,5, due to near 100% deposition in
the ET region for particles >5 urn (Raabe et al., 1988, 001439) and only 44% nasal inhalability of a
10 urn particle in the rat (Menache et al., 1995, 006533). These studies also employed relatively
high doses of PMi0_2.s. Despite these shortcomings, increased plasma fibrinogen was observed and
the response was similar to that observed with PM2.5. At this time, evidence of biological plausibility
for cardiovascular morbidity effects following PMi0_2.5 exposure is sparse, due to the small number
of studies, few endpoints examined, and the limitations related to the interpretation of IT instillation
exposures.
In summary, several epidemiologic studies report associations with cardiovascular endpoints
including IHD hospitalizations, Supraventricular ectopy, and changes in HRV. Further, dust storm
December 2009 6-82
-------
events resulting in high concentrations of crustal material are linked to increases in cardiovascular
disease hospital admissions or ED visits for cardiovascular diseases. A large proportion of inhaled
coarse particles in the 3-6 urn (dae) range can reach and deposit in the lower respiratory tract,
particularly the TB airways (Figures 4-3 and 4-4). The few toxicological and controlled human
exposure studies examining the effects of PMi0_2.5 provide limited evidence of cardiovascular effects
and biological plausibility to support the epidemiologic findings. Therefore the available evidence is
suggestive of a causal relationship between PM,,, exposures and cardiovascular
effects
6.2.12.3. UFPs
There was very little evidence available in the 2004 PM AQCD (U.S. EPA, 2004, 056905) on
the cardiovascular effects of UFPs. Findings from one study presented in the 2004 PM AQCD
(U.S. EPA, 2004, 056905) of controlled exposures to UF EC suggested no particle-related effects on
various cardiovascular endpoints including blood coagulation, HRV, and systemic inflammation. No
epidemiologic studies of short-term UFP concentration and cardiovascular endpoints were included
in the 2004 AQCD and there were no relevant toxicological studies reviewed in the 2004 PM AQCD
(U.S. EPA, 2004, 056905) that exposed animals to UFPs. A small number of new epidemiologic
studies, as well as several controlled human exposure and toxicological studies have been conducted
in recent years, but substantial uncertainties remain as to the cardiovascular effects of UFPs. For a
given mass, the enormous number and large surface area of UFPs highlight the importance of
considering the size of the particle in assessing response. For example, UFPs with a diameter of
20 nm, when inhaled at the same mass concentration, have a number concentration that is
approximately six orders of magnitude higher than for a 2.5-um diameter particle. Particle surface
area is also greatly increased with UFPs. Many studies suggest that the surface of particles or
substances released from the surface (e.g., transition metals, organics) interact with biological
substrates, and that surface-associated free radicals or free radical-generating systems may be
responsible for toxicity, resulting in greater toxicity of UFPs per particle surface area than larger
particles. Additionally, smaller particles may have greater potential to cross cell membranes and
epithelial barriers.
Controlled human exposure studies are increasingly being utilized to evaluate the effect of
UFPs on cardiovascular function. While the number of studies of exposure to UFPs is still limited,
there is a relatively large body of evidence from exposure to fresh DE, which is typically dominated
by UFPs. As described under the summary for PM2.s, studies of controlled exposures to DE (100-300
ug/m3) have consistently demonstrated effects on vasomotor function among adult volunteers
(Section 6.2.4.2). In addition, exposure to UF EC (50 ug/m3, 10.8xl06 particles/cm3) was recently
shown to attenuate FMD (Shah et al., 2008, 156970). Changes in vasomotor function have been
observed in animal toxicological studies of UFPs, although very few studies have been conducted
(Section 6.2.4.3). Inhaled UF TiO2 impaired arteriolar dilation when compared to fine TiO2 at similar
mass doses (Nurkiewicz et al., 2008, 156816). This response may have been due to ROS in the
microvascular wall, which may have led to consumption of endothelial-derived NO and generation
of peroxynitrite radicals. Support for an UFP effect on altered vascular reactivity is also provided by
studies of DE and IT instillation exposure to ambient PM. The response to DE did not appear to be
due to VOCs. One epidemiologic study showed that PNC was associated with a nonsignificant
decrease in flow- and nitroglycerine-mediated reactivity as measures of vasomotor function in
diabetics living in Boston (O'Neill et al., 2005, 088423).
New studies have reported increases in markers of systemic oxidative stress in humans
following controlled exposures to different types of PM consisting of relatively high concentrations
of UFPs from sources including wood smoke, urban traffic particles, and DE (Section 6.2.9.2).
Increased cardiac oxidative stress has been observed in mice and rats following gasoline exhaust
exposure and it appeared the effect was particle-dependent (Section 6.2.9.3).
The associations between UFPs and HRV measures in epidemiologic studies include increases
and decreases (Section 6.2.1.1), providing some evidence for an effect. Exposure to UF CAPs has
been observed to alter parameters of HRV in controlled human exposure studies, although this effect
has been variable between studies (Section 6.2.1.2). Alterations in HR, HRV, and BP were reported
in rats exposed to <200 ug/m3 UF CB (<1.6xl07 particles/cm3) (Sections 6.2.1.3 and 6.2.5.3). The
effects of UFPs on BP have been mixed in epidemiologic studies (Section 6.2.5.1).
December 2009 6-83
-------
There is some evidence of changes in markers of blood coagulation in humans following
controlled exposure to UF CAPs, as well as wood smoke and DE; however, these effects have not
been consistently observed across studies (Section 6.2.8.2). Toxicological studies demonstrate mixed
results for systemic inflammation and blood coagulation as well (Sections 6.2.7.3 and 6.2.8.3).
Few time-series studies of CVD hospital admissions have evaluated UFPs. The SOPHIA study
found no association between any outcome studied (all CVD, dysrhythmia, CHF, IHD, peripheral
vascular and cerebrovascular disease) and 24-h mean levels of UFP (Metzger et al. 2004). The
median UF particle count in Atlanta during the study period was 25,900 particles/cm3. UFP were not
associated with CVD hospitalizations in the elderly in Copenhagen, Denmark, but were associated
with cardiac readmission or fatal MI in the European HEAPSS study (Section 6.2.10). In the
Copenhagen study, the mean count of particles with a 100 nm mean diameter was 0.68*104
particles/cm3, whereas the PNC range was approximately 1.2-7.6*104 particles/cm3 in HEAPSS
study. Spatial variation in UFP concentration, which diminishes within a short distance from the
roadway, may introduce exposure measurement error, making it more difficult to observe an
association if one exists.
A limited number of epidemiologic studies have evaluated subclinical cardiovascular measures
and a number of these were conducted in Boston. UFPs have been linked to ICD-recorded
arrhythmias in Boston and supraventricular ectopic beats in Erfurt, Germany (Section 6.2.2.1). One
study reported no UFP association with ectopy (Barclay et al., 2009, 179935). ST-segment
depression in subjects with stable coronary heart disease was associated with UFPs in Helsinki
(Section 6.2.3.1). The limited number of studies that examine this size fraction makes it difficult to
draw conclusions about these cardiovascular measures.
In summary, there is a relatively large body of evidence from controlled human exposure
studies of fresh DE, which is typically dominated by UFPs, demonstrating effects of UFP on the
cardiovascular system. In addition, cardiovascular effects have been demonstrated by a limited
number of laboratories in response to UF CB, urban traffic particles and CAPs. Responses include
altered vasomotor function, increased systemic oxidative stress and altered HRV parameters. Studies
using UF CAPs, as well as wood smoke and DE, provide some evidence of changes in markers of
blood coagulation, but findings are not consistent. Toxicological studies conducted with UF TiO2,
CB, and DE demonstrate changes in vasomotor function as well as in HRV. Effects on systemic
inflammation and blood coagulation are less consistent. PM-dependent cardiac oxidative stress was
noted following exposure to gasoline exhaust. The few epidemiologic studies of UFPs conducted do
not provide strong support for an association of UFPs with effects on the cardiovascular system.
Based on the above findings, the evidence is suggestive of a causal relationship between
ultrafine PM exposure and cardiovascular effects.
6.3. Respiratory Effects
6.3.1. Respiratory Symptoms and Medication Use
The 2004 PM AQCD (U.S. EPA, 2004, 056905) presented evidence from epidemiologic
studies of increases in respiratory symptoms associated with PM, although this was not supported by
the findings of a limited number of controlled human exposure studies. Recent epidemiologic studies
have provided evidence of an increase in respiratory symptoms and medication use associated with
PM among asthmatic children, with less evidence of an effect in asthmatic adults. The lack of an
observed effect of PM exposure on respiratory symptoms in controlled human exposure studies does
not necessarily contradict these findings, as very few studies of controlled exposures to PM have
been conducted among groups of asthmatic or healthy children.
6.3.1.1. Epidemiologic Studies
The 2004 PM AQCD (U.S. EPA, 2004, 056905) concluded that the effects of PMi0 on
respiratory symptoms in asthmatics tended to be positive, although they were somewhat less
consistent than PMi0 effects on lung function. Most studies showed increases in cough, phlegm,
December 2009 6-84
-------
difficulty breathing, and bronchodilator use, although these increases were generally not statistically
significant for PMi0. The results from one study of respiratory symptoms and PMi0_2.5 (Schwartz and
Neas, 2000, 007625) found a statistically significant association with cough with PMi0_2.5. The
results of two studies examining respiratory symptoms and PM25 revealed slightly larger effects for
PM2.5 than for PMi0.
Asthmatic Children
Two large, longitudinal studies in urban areas of the U.S. investigated the effects of ambient
PM on respiratory symptoms and/or asthma medication use with similar analytic techniques (i.e.,
multistaged modeling and generalized estimating equations [GEE]): the Childhood Asthma
Management Program (CAMP) (Schildcrout et al, 2006, 089812) and the National Cooperative
Inner-City Asthma Study (NCICAS) (Mortimer et al., 2002, 030281). A number of smaller panel
studies conducted in the U.S. evaluated the effects of ambient PM concentrations on respiratory
symptoms and medication use among asthmatic children (Delfmo et al., 2002, 093740; 2003,
090941: 2003, 050460: Gent et al., 2003, 052885: 2009, 180399: 2006, 088031: Slaughter et al.,
2003, 086294).
In the CAMP study, the association between ambient air pollution and asthma exacerbations in
children (n = 990) from eight North American cities was investigated (Schildcrout et al., 2006,
089812). In contrast to several past studies (Delfmo et al., 1996, 080788: 1998, 051406). no
associations were observed between PMi0 and asthma exacerbations or medication use. PMi0
concentrations were measured on less than 50% of study days in all cities except Seattle and
Albuquerque. While PMi0 effects were not observed for the entire panel of children, they were
observed in recent reports on the children participating at the Seattle center (Slaughter et al., 2003,
086294: Yu et al., 2000, 013254). In a smaller panel study of asthmatic children (n = 133) enrolled
in the CAMP study, daily particle concentrations averaged over three central sites in Seattle was used
as the exposure metric (Slaughter et al., 2003, 086294). Children were followed for 2 months, on
average. Daily health outcomes included both a 3-category measure of asthma severity based on
symptom duration and frequency, and inhaled albuterol use. In single-pollutant models, an increased
risk of asthma severity was associated with a 10 (ig/m3 increase in lag 1 PM25 (OR 1.20 [95% CI:
1.05-1.37]) and with a 10 ug/m3 increase in lag 0 PM10 (OR 1.12 [95% CI: 1.05-1.22]). In
copollutant models with CO, the associations remained (OR for PM2.5 1.16 [95% CI: 1.03-1.30]; OR
for PMio 1.11 [95% CI: 1.03-1.19]). Associations between inhaler use and PM were positive in
single-pollutant models (RR lag 1 PM2.5 1.08 [95% CI: 1.01-1.15]; RR lag 0 PM10 1.05 [95% CI:
1.00-1.09]), but attenuated and no longer statistically significant in copollutant models.
The eight cities included in the NCICAS (Mortimer et al., 2002, 030281) were all in the East
or Midwest: New York City (Bronx, E. Harlem), Baltimore, Washington DC, Cleveland, Detroit, St.
Louis, and Chicago. In this study, 864 asthmatic children, aged 4-9 yr, were followed daily for four
2-wk periods over the course of nine months. Morning and evening asthma symptoms (analyzed as
none vs. any) and peak flow were recorded. For the three urban areas with air quality data, each
10 (ig/m3 increase in the mean of the previous 2 days (lag 1-2) PMio, increased the risk for morning
asthma symptoms (OR 1.12 [95% CI: 1.00-1.26]). This effect was robust to the inclusion of O3 (OR
1.12 [95% CI: 0.98-1.27]). In a related study, O'Connor et al. (2008, 156818) examined the
relationship between short-term fluctuations in outdoor air pollutant concentrations and changes in
pulmonary function and respiratory symptoms among children with asthma in seven U.S. inner-city
communities. PM2 5 concentration was not statistically associated with respiratory symptoms in this
study.
December 2009 6-85
-------
Study
Mar etal. (2004.057309)
Rodriguez etal. (2007, 092842)
Ranzi etal. (2004. 089500)
Aekplakorn et al. (2003, 089908)
Gent etal. (2009. 180399)
Mar etal. (2004,057309)
Rodriguez etal. (2007, 092842)
Gent etal. (2009. 180399)
OConnor et al. (2008, 156818)
Mar etal. (2004. 057309)
Rodriguez etal. (2007, 092842)
Gent etal. (2009. 180399)
Mar etal. (2004, 057309)
Aekplakorn et al. (2003, 089908)
Mar etal. (2004, 057309)
Gent etal. (2009. 180399)
Rabinovitch et al .(2006, 088031)
Slaughter etal. (2003. 086294)
Rabinovitch et al. (2004, 096753)
Slaughter et al. (2003, 086294)
Mar etal. (2004. 057309)
Aekplakorn et al. (2003, 089908)
Mar etal. (2004. 057309)
Aekplakorn et al. (2003, 089908)
Mar etal. (2004.057309)
Mar etal. (2004.057309)
Aekplakorn et al. (2003, 089908)
Just etal. (2002. 035429)
Jalaludin et al. (2004, 056595)
Mar etal. (2004. 057309)
Jalaludin etal. (2004. 056595)
Andersen et al. (2008, 096150)
Mar etal. (2004.057309)
Delfino etal. (2003. 050460)
Mortimer et al. (2002, 030281)
Rabinovitch et al. (2004, 096753)
Mar etal. (2004,057309)
Delfino etal. (2002. 093740)
Schildcroutetal. (2006, 089812)
Aekplakorn et al. (2003, 089908)
Just etal. (2002.035429)
Mar etal. (2004,057309)
Rabinovitch et al. (2004, 096753)
Jalaludin et al. (2004, 056595)
Slaughter et al. (2003, 086294)
Schildcrout et al. (2006, 089812)
Rabinovitch et al. (2004, 096753)
Just etal (2002 035429)
Slaughter etal. (2003. 086294)
ORs and 95% CIs standardized to increments of 10 (jg/m3
Location
Spokane, WA
Perth, Australia
Emilia-Romagna, Italy
Thailand
New Haven, CT
Spokane, WA
Perth, Australia
New Haven, CT
Multicity, US
Spokane, WA
Perth, Australia
New Haven, CT
Spokane, WA
Thailand
Spokane, WA
New Haven, CT
Denver, CO
Seattle, WA
Denver, CO
Seattle, WA
Spokane, WA
Thailand
Spokane, WA
Thailand
Spokane, WA
Spokane, WA
Thailand
Paris, France
Australia
Spokane, WA
Australia
Copenhagen, Den
Spokane, WA
California
Multicity, US
Denver, CO
Spokane, WA
CA, 1-hMax
CA, 8-h Max
CA 24-h Avg
US and Can
Thailand
Paris, France
Spokane, WA
Denver, CO
Australia
Seattle, WA
US and Can
Denver CO
Paris France
Seattle, WA
Lag
0
0
0-5
0-3
0
0-2
0
0-5
0-2
0-2
0-4
0
0
0-5
0-2
0
0
0
0
0-2
1
1
0-2
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0-4
0-5
0
0-5
2-4
0
0
0
1-2
0-3
0-3
0
0-2
0-2
0-2
0-2
0
0
0-4
0
0-3
0-5
0
0-2
0-3
0-4
0
Endpoint
Phlegm
+ Runny Nose
Cough
Wheeze
+Shortness of Breath
Chest Tightness
Any Symptoms
+ URS
+ LRS
+ LRS
Med Use
Asthma Exacerbation
Phlegm
+ Runny Nose
Cough
Wheeze
+Shortness of Breath
Any Symptoms
+ URS
+ LRS
+ LRS
Phlegm
+ Runny Nose
Cough
Wheeze
+Shortness of Breath
Any Symptom
+ Score >1
+ Current Day
+ Previous Night
+ URS
+ LRS
+ LRS
Med U^e <Ł
s
^
Asthma Exacerbation
0.1 0.6
Effect Estimate (95% Cl)
t
0
i-«-
-0-
i A
*
—0—
— '-t —
-0!—
|
t
i — •
t«-
+0 —
i •
-*-
'»
-!0-
+
— •
i
i — • —
-t-0 —
i 0
-i — •
— 10 —
-L0 —
— 10
1 ^
—10
I
-p-0
|0-
-10-
1-0-
10-
J 0
. 1
ft
1
| *
|_ft_
1 ft
1 ft
1 Ł
^
V
'ft
1
1
-0-
*-
*
«,
i .
1 j
1.1 1.6 2.1
Odds Ratio
PM2.5
PMl 0-2.5
PMio
^ *s^
•w
r
2.6
Figure 6-7. Respiratory symptoms and/or medication use among asthmatic children
following acute exposure to PM.
December 2009
6-86
-------
Table 6-10. Characterization of ambient PM concentrations from epidemiologic studies of
respiratory morbidity and short-term exposures in asthmatic children and adults. All
concentrations are for the 24-h avg unless otherwise noted.
Study
Location
Mean Concentration
(ug/m3)
Upper Percentile
Concentrations (ug/m3)
PM2.5
Adamkiewicz et al. (2004, 0879251
Adaretal. (2007, 0014581
Aekplakorn et al. (2003, 089908)
Allen et al. (2008, 1562081
Barraza-Villarreal et al. (2008, 1562541
Bourotte et al. (2007, 150040)
de Hartog et al. (2003, 001061)
Delfino et al. (2006, 090745)
DeMeo et al. (2004, 0873461
Ebelt et al. (2005, 0569071
Ferdinands et al. (2008, 156433)
Fischer et al. (2007, 156435)
Gent et al. (2003, 0528851
Gent et al. (2009, 180399)
Girardot et al. (2006, 088271)
Hogervorst et al. (2006, 156559)
Hong et al. (2007, 0913471
Jansen et al. (2005, 0822361
Johnston et al. (2006, 091386)
Koenig et al. (2003, 156653)
Lagorio et al. (2006, 0898001
Lee et al. (2007, 0930421
Lewis etal. (2004, 097498)
Liu et al. (2009, 1920031
Mar etal. (2004, 0573091
Mar etal. (2005, 0887591
McCreanor et al. (2007, 092841)
Moshammer et al. (2006, 090771)
Murata et al. (2007, 1891591
O'Connor etal. (2008, 1568181
Steubenville, OH
St. Louis, MO
North Thailand
Seattle, WA
Mexico City
Sao Paulo, Brazil
Multicity, Europe
Southern CA
Boston, MA
Vancouver, Canada
Atlanta, GA
The Netherlands
CTSMA
New Haven, CT
Smoky Mountains
The Netherlands
Incheon City, Korea
Seattle, WA
Darwin, Australia
Seattle, WA
Rome, Italy
Seoul, South Korea
Detroit, Ml
Wndsor, Ontario
Spokane, WA
Seattle, WA
London, England
Linz, Austria
Tokyo, Japan
Multicity, U.S.
20.43
10.13
11.2
8-h max: 28.9
11.9
12.8-23.4
3.9-6.9
10.8
11.4
27.2
56
13.1
17.0
13.9
19.0
20.27
14.0
11.1
13.3
27.2
51.15
15.7-17.5
7.1
8.1-11.0
5-26
1-h avg: 11. 9-28.3
8-h avg: 15.70
39.0
14
75th: 23
98th: 51. 79
Max: 51. 79
98th: 22.43
Max: 23.24
Max: 24.8-26.3
Max: 40.38
Max: 102.8
Max: 26.6
Max: 39.8-118.1
Max: 8.8-11. 6
NR
98th: 23
Max: 28.7
Max: 34.7
75th: 187
60th: 12.1
80th: 19.0
NR
Max: 38.4
NR
Max: 36.28
Max: 44
Max: 36.5
Max: 40.4
Max: 100
75th: 87.54
Max: 92.71
Max: 56.1
95th: 19.0
98th: 19.0.
NR
NR
1-h max: 55. 9-76.1
Max 24-h avg: 76. 39
Max 1-h avg: 120
Max: 35
December 2009
6-87
-------
Study
Peled et al. (2005, 1560151
Penttinen et al. (2006, 0879881
Rabinovitch et al. (2004, 0967531
Rabinovitch et al. (2006, 0880311
Ranzi et al. (2004, 0895001
Rodriguez et al. (2007, 0928421
Slaughter etal. (2003,086294)
Strand et al. (2006, 089203)
Timonen et al. (2004, 087915)
Trenga et al. (2006, 1552091
von Klot et al. (2002, 0347061
Ward et al. (2002, 025839)
Location
Multicity, Israel
Helsinki, Finland
Denver, CO
Denver, CO
Emilia-Romagna, Italy
Perth, Australia
Seattle, WA
Denver, CO
Multicity, Europe
Seattle, WA
Erfurt, Germany
Birmingham and Sandwell, U.K.
Mean Concentration
(ug/m3)
23.9-29.2
8.37
10.8
10.8
Urban: 53.07
Rural: 29.11
1-havg:20.8
24-h avg: 8.5
7.3a
12.7
12.7-23.1
8.6-9.6a
30. 3b
12.3-12.7
Upper Percentile
Concentrations (ug/m3)
NR
75th: 11.15
Max: 33.53
98th: 29.3
Max: 53.5
98th: 23.4
NR
Max 1-h avg: 93.4
Max 24-h avg: 39. 4
75th: 11. 3
Max: 32.3
Max: 39.8-118.1
75th: 13.1-14.8
Max: 40. 4-41. 5
75th: 41. 3b
Max: 133.8b
Max: 28-37
PMiO-2.5
Aekplakorn et al. (2003, 089908)
Bourotte et al. (2007, 150040)
Ebelt et al. (2005, 056907)
Lagorio et al. (2006, 0898001
Mar etal. (2004, 0573091
von Klot et al. (2002, 0347061
North Thailand
Sao Paulo, Brazil
Vancouver, Canada
Rome, Italy
Spokane, WA
Erfurt, Germany
NR
21.7
5.6
15.6
8.7-13.5
10.3
NR
Max: 62.0
Max: 11.9
Max: 39.6
NR
75th: 14.6
Max: 64.3
PUn
Aekplakorn et al. (2003, 089908)
Andersen et al. (2008, 096150)
Boezen et al. (2005, 087396)
de Hartog et al. (2003, 0010611
Delfino et al. (2002, 0937401
Delfino et al. (2003, 050460)
Delfino et al. (2004, 0568971
Delfino et al. (2006, 0907451
Desqueyroux et al. (2002, 0260521
Ebelt et al. (2005, 056907)
Hong et al. (2007, 091347)
Jalaludin et al. (2004, 0565951
Jansen et al. (2005, 0822361
North Thailand
Copenhagen, Denmark
The Netherlands
Multicity, Europe
Alpine, CA
Los Angeles, CA
Alpine, CA
Southern CA
Paris, France
Vancouver, Canada
Incheon City, Korea
Sydney, Australia
Seattle, WA
31.9-37.5
25.1
26.6-44.1
19.6-36.5
20
59.9
29.7
35.7-70.8
23-28
17
35.3
22.8
18.0
Max: 113.3-153.3
75th: 30.2
Max: 89.9-242.2
Max: 67.4-112.0
90th: 32
Max: 42
90th: 86/0/Max: 126
90th: 40.9
Max: 50.7
Max: 105.5-154.1
Max: 63-84
Max: 36
Max: 124.87
75th: 122.8
Max: 51
December 2009
6-88
-------
Study
Johnston et al. (2006, 0913861
Just et al. (2002, 035429)
Lagorio et al. (2006, 089800)
Laurent et al. (2008, 1566721
Lee et al. (2007, 0930421
Maretal. (2004, 057309)
Mortimer etal. (2002, 030281)
Moshammer et al. (2006, 0907711
Odajima et al. (2008, 1920051
Peacock etal. (2003, 0420261
Peled et al. (2005, 156015)
Preutthipan et al. (2004, 055598)
Rabinovitch et al. (2004, 0967531
Segala et al. (2004, 0904491
Schildcrout et al. (2006, 0898121
Slaughter etal.(2003, 086294)
Steinvil et al. (2008, 1888931
von Klot et al. (2002, 0347061
Location
Darwin, Australia
Paris, France
Rome, Italy
Strasbourg, France
Seoul, South Korea
Spokane, WA
Multicity, U.S.
Linz, Austria
Fukuoka, Japan
Southern England
Multicity, Israel
Bangkok, Thailand
Denver, CO
Paris, France
Multicity, U.S.
Seattle, WA
Tel Aviv, Israel
Erfurt, Germany
Mean Concentration
(ug/m3)
20
23.5
42.8
20.8
71.40
16.8-24.5
53
8-h avg: 24.85
3-havg: 32.6-41. 5
21.2
31.0-67.1
111.0
28.1
24.2
17.7-32.43
21. Oa
64.5
45.4
Upper Percentile
Concentrations (ug/m3)
Max: 43.3
Max: 44.0
Max: 123
Max: 106.3
75th: 87.54
Max: 148.34
NR
NR
Max 24-h: 76.39
Max 3-havg: 126.0-191. 3
Max: 87.9
NR
Max: 201
Max: 102.0
Max: 97.4
75th: 26.2-42.7
90th: 32.5-53.9
75th: 29.3
75th: 60.7
75th: 59.7
Max: 172.4
3Median
Includes UFP, for complete information on number concentration from this study, please see corresponding table in Annex E.
Mar et al. (2004, 057309) studied asthmatic children (n = 9) in Spokane, WA. Increases in 0-,
1- or 2-day lags of each of the PM size classes studied were associated with cough. When all lower
respiratory tract symptoms (wheeze, cough, shortness of breath, sputum production) were grouped
together, positive associations were reported for each 10 (ig/m3 increase in same-day PMi0 (OR 1.07
[95% CI: 1.00-1.14]), or lag 0 or lag 1 PM25 (OR 1.18 [95% CI: 1.00-1.38]; OR 1.21 [95% CI:
1.00-1.46], respectively), and 10 (ig/m3 increase in lag 0 and lag 1 PM10 (OR 1.21 [95% CI:
1.01-1.44]; OR 1.25 [95% CI: 1.01-1.55], respectively). No associations were reported for PM10_2.5
and grouped lower respiratory tract symptoms (Mar et al., 2004, 057309).
Gent et al. (2003, 052885) reported on daily symptom and medication use during one summer
for 271 asthmatic children living in southern New England. In single-pollutant models for users of
maintenance medication (n = 130), PM2.5 >19 (ig/m3 lagged by 1 day was associated with a 10-25%
increase in risk of symptoms compared to PM25 <6.9 (ig/m3: OR for persistent cough 1.12 (95% CI:
1.02-1.24); OR for chest tightness 1.21 (95% CI: 1.00-1.46); OR for shortness of breath 1.26
(95% CI: 1.02-1.54). Effects were attenuated in models including O3 (OR for persistent cough 1.00
[95% CI: 0.88-1.15]; OR for chest tightness 0.91 [95% CI: 0.71-1.17]; OR for shortness of breath
1.20 [95% CI: 0.94-1.52]). No statistical associations between ambient particle exposure and
respiratory health were found for asthmatic children not on maintenance medication.
Annual PM2 5 levels at monitoring sites in New Haven, CT exceed the annual standard of
15 (ig/m3. Gent et al. (2009, 180399) conducted a study here to examine the associations between
daily exposure to PM2 5 components and sources identified through source apportionment, and daily
symptoms and medication use in asthmatic children. Asthmatic children (n = 149) aged 4-12 yr were
enrolled in the study between 2000 and 2003. Factor analysis was used to identify six sources of
PM2 5 (motor vehicle, road dust, sulfur, biomass burning, oil, and sea salt). Total PM2 5 was not
associated with any symptoms or medication use; however trace elements originating from motor
vehicle, road dust, biomass burning and oil sources were associated with symptoms and/or
December 2009 6-89
-------
medication use. For example, an increased risk of wheeze, shortness of breath, chest tightness or
short-acting inhaler use was associated with increasing EC mass concentration. Risks remain in
models that include all six PM2.5 sources as well as NO2, which may be considered a marker for
traffic. NO2 was found to be an independent risk factor for increased wheeze.
Two panel studies were conducted over the course of three winters at a school in Denver
(Rabinovitch et al., 2004, 096753: 2006, 088031). In the first report, approximately 86 different
children contributed data on asthma symptoms and medication use over three consecutive winters
(Rabinovitch et al., 2004, 096753). The exposure metric was the 3-day average concentration of
PM2 5 measured at a site located next to the school for the first two winters and from a central site
located 4.8 km (3 miles) away for the third. A strong correlation was observed during the first two
winters between PM2 5 values measured locally and at a downtown monitoring station (Pearson
product-moment correlation = 0.93) and between PMi0 values measured locally and at a downtown
monitoring station (correlation = 0.84). Therefore, in year 3, all ambient data were collected from
nearby community monitoring stations. No statistical associations were found between asthma
symptoms or medication use and PM. Rabinovitch et al. (2006, 088031) enrolled a panel of 73
children and evaluated associations with morning maximum PM2 5 measured at the central site. PM
measurements were available hourly from two co-located monitors, an FRM and a TEOM monitor.
Each 10 (ig/m3 increase in morning maximum 1-h PM25 concentration was associated with an
increased likelihood of rescue medication use (OR for FRM 1.02 [95% CI: 1.01-1.03]; OR for
TEOM 1.03 [95% CI: 1.00-1.6]). Interestingly, the association between inhaler use and particle
exposure was not evident when the 24-h avg PM2 5 was used in the model.
Two smaller panel studies enrolling asthmatic children conducted by Delfino et al. (2002,
093740: 2003, 050460) in southern California examined the health effects of different averaging
times for PMi0 (1-h, 8-h, 24-h) (Delfino et al., 2002, 093740). and 24-h avg of two PMio
components (EC and OC) (Delfino et al., 2003, 050460). In the first study, 22 children living in a
"lower" pollution area were followed daily for two months in spring. In contrast with Gent et al.
(2003, 052885). positive statistical associations with asthma symptoms (measured on a 6-point
severity scale) were found only for the children not taking anti-inflammatory medication. For these
12 children, in single-pollutant models each 10 (ig/m3 increase in lag 0 1-h max PMio nearly doubled
the risk of clinically meaningful symptoms (i.e., an asthma symptom score >3) (OR 1.14 [95% CI:
1.04-1.24]) and each 10 ug/m3 increase in 3-day avg 24-h PM10 increased the risk by 1.25 (95% CI:
1.06-1.48). No statistical associations were found between exposure to ambient particles and
symptoms in the ten children who were taking anti-inflammatory medication. No multipollutant
models were reported. The second study enrolled 22 asthmatic children living in an area of higher
pollution. For children living in this community, each 10 (ig/m3 increase in lag 0, 24-h PMio was
associated with an increased risk of asthma symptom score >1: OR 1.10, (95% CI: 1.03-1.19)
(Delfino et al., 2003, 050460). The correlation among PM10, EC and OC was substantial: 0.80
between PMio and either EC or OC, and 0.94 between EC and OC. Associations between EC or OC
and asthma symptoms were very similar to those for PMi0: each 3 (ig/m3 increase in lag 0, 24-h EC
or 5 (ig/m3 increase in lag 0, 24-h OC was associated with an increased risk of asthma symptoms
(OR 1.85 [95% CI: 1.11-3.08] or OR 1.88 [95% CI: 1.12-3.17], respectively) (Delfino et al., 2003,
050460).
The association between incident wheezing symptoms and air pollution was assessed in the
Copenhagen Prospective Study of Asthma in Children among a birth cohort of 205 children in
Copenhagen, Denmark. In addition to PMio and other gaseous air pollutants, the study examined
UFP concentrations collected from a central background monitoring station. This is the only study
identified that examined the association between UFPs and respiratory symptoms in children. There
were strong adverse effects for PMio and UFPs, as well as for NO2, NOX, and CO for wheezing
symptoms in infants which attenuated after the age of 1 yr (lag 2-4 PM10 OR 1.21 (95% CI 0.99-
1.48); lag 2-4 UFP OR 1.92 (95% CI: 0.98-3.76)). These associations remained in copollutant
models including NO2, NOX and CO.
Studies from Australia (Rodriguez et al., 2007, 092842). Europe (Andersen et al., 2008,
096150: Laurent et al., 2008, 156672: Laurent et al., 2009, 192129: Ranzi et al., 2004, 089500).
and Asia (Aekplakorn et al., 2003, 089908) provide additional evidence of an association between
ambient PM and respiratory symptoms and/or medication use among asthmatic children. Two studies
(Jalaludin et al., 2004, 056595; Just et al., 2002, 035429) found no association between ambient PM
levels and these health endpoints.
December 2009 6-90
-------
Asthmatic Adults
Since the 2004 PM AQCD (U.S. EPA, 2004, 056905). one U.S. and several European studies
have investigated the effects of ambient PM levels on respiratory symptoms and medication use
among asthmatic adults. The respiratory symptom and medication use results from these studies are
summarized by particle size and displayed in Table 6-10 and Figure 6-8. Relatively few studies
examined these effects in healthy adults, and they did not identify a relationship between ambient
PM levels and respiratory symptoms or medication use. These studies of healthy adults are
summarized in Annex E, but will not be described in detail in this section.
Mar et al. (2004, 057309) studied asthmatic adults (n = 16) in Spokane, WA over a 3-yr time
period. No associations were found between PM and respiratory symptoms among the adults.
Several panel studies conducted in Europe have examined effects of daily exposures to air
pollution on adults with asthma, including studies in the Pollution Effects on Asthmatic Children in
Europe (PEACE) study (Boezen et al., 2005, 087396). Exposure and Risk Assessment for Fine and
UFPs in Ambient Air (ULTRA) study (De et al., 2003, 001061). in Germany (Von et al., 2002,
034706). and in Paris (Desqueyroux et al., 2002, 026052: 2004, 090449). Boezen et al. (2005,
087396) enrolled 327 elderly adults in the Netherlands to examine the role of airway
hyperresponsiveness (AHR) and IgE levels in susceptibility to air pollution. For subjects with both
AHR (defined as > 20% FEVi decline at < 2 mg cumulative methacholine [Mch]) and high total IgE
(>20 kU/L), each 10 (ig/m3 increase in lag 2 PMi0 concentration was associated with an increased
risk of upper respiratory symptoms (URS) among males (OR 1.06 [95% CI: 1.02-1.10]), and at lag 0
with increased cough among females (OR 1.04 [95% CI: 1.00-1.08]). Each 10 (ig/m3 increase in BS
at lag 0, lag 1, and the 5-day mean was associated with URS and cough among males. The strongest
association in both cases was for the 5-day mean (OR for URS 1.43 [95% CI: 1.20-1.69]; OR for
cough 1.16 [95% CI: 1.05-1.29]). The authors suggest that the sex differences observed may be
explained by differential daily exposure to traffic exhaust experienced by men compared to women
(Boezen et al.. 2005. 087396).
As part of the multicenter ULTRA study, de Hartog et al. (2003, 001061) enrolled 131 older
adults with coronary artery disease in three cities (Amsterdam, Erfurt [Germany], and Helsinki).
Pooling data from all 3 cities, associations were observed between PM2.5 and shortness of breath and
phlegm: each 10 ug/m3 increase in the 5-day avg PM25 was associated with an increased risk of
symptoms (OR for shortness of breath 1.12 [95% CI: 1.02-1.24]; OR for phlegm 1.16 [95% CI:
1.03-1.32]). Unlike fine particles, UFPs were not consistently associated with symptoms.
In a study that took place in Erfurt, Germany, von Klot et al. (2002, 034706) examined daily,
winter time exposure to ambient PM10_2.5, PM2.5_0.oi and PM0.i_0.oi and respiratory health effects in 53
adult asthmatics. The authors examined associations between wheeze, use of inhaled, short-acting
p2-agonists or inhaled corticosteroids and exposure to particles in single and multipollutant models.
Particle exposure metrics examined included same-day, 5-day and 14-day average concentrations.
No effects were observed for wheeze and exposure to PMi0_2.5 for any averaging time. The strongest
association between wheeze and exposure to UFPs was for a 14-day avg: each 7,700 increase in the
NCooi-o i increased the risk of wheeze by 27% (OR 1.27 [95% CI: 1.13-1.43]). The effect was
attenuated in copollutant models that also included PM25 001 (OR 1.12 [95% CI: 1.01-1.24]), NO2
(OR 1.12 [95% CI: 0.99-1.26]), CO (OR 1.05 [95% CI: 0.92-1.19]) or SO2 (OR 1.14 [95% CI:
1.04-1.26]). The correlations between UFPs and two gaseous pollutants, NO2 and CO, were high:
0.66 for each.
December 2009 6-91
-------
Study
Location Lag Endpoint
Effect Estimate (95% Cl)
de Hartog et al. (2003,001061) Netherlands
Mar et al. (2004,057309) Spokane, WA
de Hartog et al. (2003,001061) Netherlands
Mar et al. (2004,057309) Spokane, WA
Johnston et al. (2006,091386) Australia
Mar et al. (2004,057309) Spokane, WA
Mar et al. (2004,057309) Spokane, WA
von Klotetal. (2002.034706) Germany
Mar et al. (2004,057309) Spokane, WA
von Klotetal. (2002.034706) Germany
Mar et al. (2004,057309) Spokane, WA
Mar et al. (2004,057309) Spokane, WA
Boezen et al. (2005,087396) Netherlands
Segala et al. (2004,090449) Paris, France
Mar et al. (2004,057309) Spokane, WA
Desqueyroux et al. (2002,026052) Paris, France
Johnston et al. (2006,091386) Australia
Mar et al. (2004,057309) Spokane, WA
Boezen et al. (2005,087396) Netherlands
Mar et al. (2004,057309) Spokane, WA
0-4 Phlegm
0
0 + Runny Nose
0 Cough
0-4 Wheeze
0 Shortness of breath
0
0-5 Asthma Symptoms
0
0 +LRS
0 Phlegm
0 + Runny Nose
0-4 Wheeze
0
0 Shortness of breath
0 Cough
0 Any Symptom
0-4 Medication Use
0 Asthma Symptoms
0 Phlegm
0 + Runny Nose
0-4 Cough
0
0
0 Wheeze
0 Shortness of breath
3-5 Asthma Symptoms
0-5
0
0-4 + URS
0 +LRS
PMZ5
1 •
PM«
PM«
0.50 0.75 1.00 1.25
Odds Ratio
1.50
1.75
Figure 6-8. Respiratory symptoms and/or medication use among asthmatic adults following
acute exposure to particles. Summary of studies using 24-h avg of PMio, PM2.5,
PMio-2.6- ORs and 95% CIs were standardized to increments of 10 ug/m3.
In the same study, no association was found between exposure to PMi0_2.5,, PM2 5, or UFPs and
use of short-acting inhalers, though there was an association with maintenance medication. Increased
likelihood of maintenance medication was significantly associated with PM of all sizes and all
averaging times (same-day, 5- and 14-day avg) and gaseous copollutants in single or copollutant
models. The strongest effects were seen for 14-day avg of PMi0_25 (for each 10 ug/m3 increase OR
1.43 [95% CI: 1.28-1.60]), PM25.001 (for each 20 ug/m3 increase OR 1.54 [95% CI: 1.43-1.66]),
NCo.oi-o.1 (for each 7,700 increase OR 1.45 [95% CI: 1.29-1.63]). For PM2.5.0.0i, effects were
unchanged in copollutant models, including a model with UFPs. The authors conclude that this is
evidence for independent effects of PM2.5 and UFPs (Von et al., 2002, 034706).
In Paris, Segala et al. (2004, 090449) recruited 78 adults from an otolaryngology clinic and
followed them for three months. Both PM10 and BS (which were highly correlated [r =.88]) were
associated with cough: OR 1.24 (95% CI: 1.01-1.52) for a 10 ug/m increase in mean 0-4 day PMi0
and OR 1.18 (95% CI: 1.02-1.39) for a 10 ug/m3 increase in BS.
Also in Paris, 60 severe asthmatics were followed for 13 months and the relationship between
daily air quality (including 24-h PMi0 as measured at the site nearest to the subject's home) and
asthma attack (defined as the need to increase rescue medication use and one or more positive signs
on clinical exam of wheezing, expiratory brake, thoracic distention, hypertension with tachycardia,
polypnea) were examined with GEE models (Desqueyroux et al., 2002, 026052). Each 10 ug/m3
increase in PMi0 increased the risk of asthma attack, but only after lags of 3-5 days. The strongest
effect was seen for the mean lag of days 3-5 (OR 1.21 [95% CI: 1.04-1.40]). Effect sizes were larger
among patients not on regular oral steroid therapy: for PMi0 lag 3-5 (OR 1.41 [95% CI: 1.15-1.73]).
This effect persisted in copollutant models for winter time levels of PMi0 and SO2 (OR 1.51 [95%
December 2009
6-92
-------
CI: 1.20-1.90]) or NO2 (OR 1.43 [95% CI: 1.16-1.76]), but not in summer time models with O3 (OR
1.09 [95% CI: 0.71-1.67]).
Copollutant Models
A limited number of respiratory symptoms studies reported results of copollutant models.
Generally, the associations between respiratory symptoms and PM were robust to the inclusion of
copollutants (Figure 6-9), though Desqueyroux et al. (2002, 026052) indicate the effects of PM may
be potentiated by NO2 and SO2 during the winter months. Gent et al. (2003, 052885) also reported
the results of copollutant models, though the categorical exposure groups used in the analysis did not
allow these results to be included in Figure 6-9. As reported above, the investigators found that
effects were attenuated in models including O3.
Study
Outcome
Pollutant
Effect Estimate (95% CI)
Slaughter et al. (2003,086294)
Aekplakorn et al. (2003,089909)
Asthma Severity
Cough
Aekplakorn et al. (2003,089909) Cough
PM25
PM25+CO
PM25
PM2.5+S02
PM1(>2.5
PM10.2.5+S02
Slaughter et al. (2003,086294)
Mortimer et al. (2002,030281)
Aekplakorn et al. (2003,089909)
Andersen et al. (2008,096150)
Asthma Severity PMm
PMm+CO
AM Asthma Symptoms PM10
PM10+03
Cough PM10
PM10+S02
Wheeze PM10
PM10+N02
PMm+CO
Andersen et al. (2008,096150) Wheeze
Desqueyroux et al. (2002,026052) Asthma Attack
UFP
UFP+ PM10
UFP+ N02
UFP+CO
PM10
PM10+N02
PM10+03
PM10+S02
I
0.5
PM2.5 Asthmatic Children
PMi0-2.5 Asthmatic Children
PMio Asthmatic Children
UFP Asthmatic Children
PMio Asthmatic Adults
\ I I
1.0 1.5 2.0
Odds Ratio
2,5
Figure 6-9. Respiratory symptoms following acute exposure to particles and additional
criteria pollutants. Circles represent single pollutant effect estimates and
squares represent copollutant effect estimates.
6.3.1.2. Controlled Human Exposure Studies
CAPs
Neither new controlled human exposure studies nor studies cited in the 2004 PM AQCD
(U.S. EPA, 2004, 056905) have found significant effects of CAPs on respiratory symptoms among
healthy or asthmatic adults, or among older adults with COPD (Gong et al., 2000, 155799; 2003,
042106; 2004, 087964; 2004, 055628; 2005, 087921; 2008, 156483; Petrovic et al., 2000, 004638).
December 2009
6-93
-------
Urban Traffic Particles
One new study reported an increase in respiratory symptoms (upper and lower airways) among
healthy volunteers (19-59 yr) during a 2-h exposure to road tunnel traffic (PM25 concentration 46-
81 ug/m3) (Larsson et al., 2007, 091375). However, information on specific respiratory symptoms
(e.g., throat irritation, wheeze or chest tightness) was not provided. In addition, this study only
evaluated respiratory symptoms pre- versus post-exposure, and did not compare response with a
filtered air control exposure.
Diesel Exhaust
Respiratory symptoms including mild nose and throat irritation have been reported following
controlled exposure to DE; however, other symptoms such as cough, wheeze and chest tightness
have not been observed (Mudway et al., 2004, 180208).
Model Particles
Pietropaoli et al. (2004, 156025) found no association between exposure to UF carbon
particles and respiratory symptoms in healthy adults at concentrations between 10 and 50 ug/m3, or
asthmatics at a concentration of 10 ug/m3. Beckett et al. (2005, 156261) exposed healthy subjects to
UF and fine ZnO (500 ug/m3) and observed no difference in respiratory symptoms compared to
filtered air control 24 h following exposure. In a study evaluating respiratory effects of exposure to
ammonium bisulfate or aerosolized H2SO4 (200 and 2,000 ug/m ) among healthy and asthmatic
adults, Tunnicliffe et al. (2003, 088744) observed no change in respiratory symptoms with either
particle type or concentration relative to filtered air. This finding is in agreement with many similar
older studies which have generally reported no increase in respiratory symptoms following exposure
to acid aerosols at concentrations <1,000 ug/m3 (U.S. EPA, 1996, 079380; 2004, 056905).
Summary of Controlled Human Exposure Study Findings for Respiratory
Symptoms
These new studies confirm previous reports that have found no association between PM
exposure and respiratory symptoms.
6.3.2. Pulmonary Function
Epidemiologic studies cited in the 2004 PM AQCD (U.S. EPA, 2004, 056905) observed small
decrements in pulmonary function associated with both PM25 and PMi0 (U.S. EPA, 2004, 056905).
The majority of controlled human exposure studies reported no effect of PM on pulmonary function,
while the results from toxicological studies were mixed, with some evidence of changes in tidal
volume and respiratory rate following exposure to CAPs. Epidemiologic studies published since the
2004 PM AQCD (U.S. EPA, 2004, 056905) have reported an association between PM2.5
concentration and decrements in forced expiratory volume in one second (FEVi), particularly among
asthmatic children. These findings are coherent with recent toxicological evidence of AHR
following CAPs exposure. Results from recent controlled human exposure studies have been
inconsistent, with some studies demonstrating small decreases in arterial oxygen saturation, FEVi or
maximal mid-expiratory flow following exposure to CAPs or EC. It is interesting to note that these
effects appear to be more pronounced among healthy adults than adults with asthma or COPD. A
number of recent animal toxicological studies demonstrated alterations in respiratory frequency
following short-term exposure to CAPs.
December 2009 6-94
-------
6.3.2.1. Epidemiologic Studies
The 2004 PM AQCD (U.S. EPA, 2004, 056905) concluded that both PM2.5 and PMi0 appeared
to affect lung function in asthmatics. A limited number of studies evaluated UFPs and found them to
be associated with a decrease in peak expiratory flow (PEF). Few analyses were able to clearly
distinguish the effects of PM2.5 and PMi0 from other pollutants. Results for PMi0 PEF analyses in
non-asthmatic studies were inconsistent, with fewer studies reporting strong associations.
Asthmatic Children
Several recent panel studies have been conducted in the U.S. examining the association of
exposure to ambient PM and lung function in asthmatic children (Allen et al., 2008, 156208 in
Seattle; Lewis et al., 2003, 088413 in Southern California; 2004, 097498: Lewis et al., 2005,
081079 in Detroit; O'Connor et al., 2008, 156818: Rabinovitch et al., 2004, 096753 in Denver).
Mean concentration data from these studies are summarized in Table 6-10. In the Inner-City Asthma
Study (ICAS), FEVi and PEF tidal were statistically related to the 5-day avg of PM2.5 but not to the
1-day avg concentration (O'Connor et al., 2008, 156818). The risk of experiencing a percent-
predicted FEVi more than 10% below personal best was related to the 5-day avg concentration of
PM25 (1.14 [95% CI: 1.01-1.29]). The risk of experiencing a percent-predicted PEF rate more than
10% below personal best was related to PM2.5 (1.18 [95% CI: 1.03-1.35]). This effect remained
robust in copollutant models with O3 and NO2 for the FEVi effect, but not the PEF rate effect.
The Denver study (Rabinovitch et al., 2004, 096753), described in Section 6.3.1.1, also
examined daily FEVi and PEF in 86 asthmatic children over the course of three winters (some
subjects participated in more than one winter). Lung function measurements were performed under
supervision daily at the elementary school where all subjects attended, and without supervision every
evening and on nonschool days. As described above, the authors chose to use a 3-day moving
average of 24-h PM25 or PMi0 as the exposure metric. No statistical associations were observed
between morning or afternoon FEVi or PEF and particle exposure. The same group of researchers
(Strand et al., 2006, 089203) used regression calibration to estimate personal exposures to ambient
PM2.5 and found that a 10 (ig/m3 increase in PM2.5 was associated with a 2.2% (95% CI: 0.0-4.3)
decrease in FEVi at a 1-day lag as compared with the estimate of a 1.0% decrease in FEVi using
ambient PM25 concentrations from fixed monitors. These results underscore the effects of exposure
error on epidemiologic study results; the effect estimate using an estimate of personal exposure to
ambient PM2 5 was twice that for central site PM2 5.
From winter 2001 to the spring of 2002, the same number (n = 86) of primary school-age
asthmatic children participated in six 2-wk seasonal assessments of lung function in Detroit (Lewis
et al., 2005, 081079). Using a protocol similar to that used in Rabinovitch et al. (2004, 096753).
morning lung function measurements (FEVi, PEF) were self-administered at school under
supervision by research staff. Evening and weekend measurements were recorded by subjects at
home, without supervision from research staff. Community-level exposure was assessed using
monitors placed on a school rooftop of both of the communities. Most of the subjects (82 of 86)
lived within 5 km of their respective community monitors. In single-pollutant models using GEE and
only among children reporting the use of maintenance medication (corticosteroids), each 10 (ig/m3
increase in lag 2 PMi0 was associated with a decrease in the lowest daily percent predicted FEVi (a
reduction of 1.15%, [95% CI: -2.1 to -0.25]). Among children reporting presence of URI on the day
of lung function measurement, increases in the average of lag 3-5 of either PM25 or PMi0 resulted in
a decrease in the lowest daily FEVi (for a 10 (ig/m3 increase in PM25 the reduction was 2.24% [95%
CI; -4.4 to -0.25]; and for a 10 ug/m3 increase in PMi0 the reduction was 2.4% [95% CI: -4.5 to
-0.3]). In copollutant models that included one particle pollutant and O3, and among children using
maintenance medication, lag 3-5 PM25 continued to be associated with lowest daily FEVi as well as
diurnal FEVi variability: each 10 (ig/m3 increase was associated with a 2.23% decrease in FEVi
(95% CI: -3.92 to -0.57) and a 2.22% increase in FEVi variability (95% CI: 1.0 to 3.50). Increases in
lag 1 or lag 2 of PMi0 were associated with FEVi and FEVi diurnal variability in copollutant models.
The strongest association was with lag 2 for diurnal variability (for each 10 (ig/m3 increase
variability increased by 7.0% [95% CI: 4.2-9.6). It is unclear what role the lack of supervision during
the evening and weekend measures may have had on these diurnal results.
Two panel studies in southern California examined the association of PM exposure on lung
function in asthmatic children (Delfino et al., 2003, 050460: 2004, 056897). In Delfino et al. (2003,
December 2009 6-95
-------
050460). described above, no association between exposure to particles and PEF was found for 22
Hispanic, asthmatic children living in an area of relatively high pollution. In Delfino et al. (2004,
056897) 19 asthmatic children, aged 9-17 yr, were followed for 2 weeks and daily, self-administered
FEVi measurements were taken. Particle exposures studied included central-site PMi0 in addition to
personal PM (in the range of 0.1-10 (im range, with the highest response in the fine PM range), and
home stationary measurements of both PM2.5 and PMi0. The authors report inverse associations
between percent expected FEVi and PM indicators. The strongest association for exposure to
personal PM was for a 5-day moving average of 12-h daytime PM: for each 10 (ig/m3 increase, FEVi
decreased by 7.1% (95% CI: -9.9 to -2.9). Effects for all stationary sites (inside and outside of
residence, central site) for PM2.5 were on the order of 1-2% reductions in FEVi, with the strongest
associations for the 5-day moving average (presented in figures only). Likewise for PMi0 measured
at stationary sites, the strongest effects were for the 5-day moving average and ranged from
approximately 3.8% reduction associated with indoor monitors to about 1.5% for both the outdoor
and central site monitors (presented in figures only). A helpful comparison among all 24-h measures
is given for 10 (ig/m3 increases in personal PM and PM2.5 associated with decreases in percent
predicted FEVi: an increase of 10 (ig/m3 personal PM is associated with a decrease in FEVi of 3.0%
(95% CI: -5.6 to -0.5); 10 (ig/m3 increase in indoor PM with 2.4% decrease (95% CI: -4.2 to -0.6);
10 (ig/m3 increase in outdoor PM with 1.5% decrease (95% CI: -3.4 to 0.1); 10 ug/m3 increase in
central site PM with 0.9% decrease (95% CI: -2.6 to 0.5).
Trenga et al. (2006, 155209) reported associations among personal, residential, and central site
PM2.5 and lung function in 17 asthmatic children in Seattle. The only statistical association with
decline in FEVi was with indoor measurements of PM25: each 10 (ig/m3 increase in lag 1 indoor
PM2.5 was associated with a decline in FEVi of 64.8 mL (95% CI: -111.3 to 18.3) (a 3.4% decline
from the mean of 1.9 L). Indoor PM25 (lag 1) was also associated with declines in PEF (by 9.2 L/min
[95% CI: -17.5 to -0.9], a 3.6% decline from the 254 L/min avg) and in maximal mid-expiratory
flow (MMEF) for the six subjects not taking anti-inflammatory medication (by 12.6 L/min [95% CI:
-20.7 to -4.6], a 13.7% decline from the 92 L/min avg). Personal PM25 (lag 1) was only statistically
associated with PEF for the six subjects not on anti-inflammatory medication: each 10 (ig/m3
increase resulted in a 10.5 L/min ([95% CI: -18.7 to -2.3], a 4.5% decline from the 233 L/min avg)
reduction in PEF. Anti-inflammatory medication use attenuated associations with PM2 5.
Also in Seattle, Allen et al. (2008, 156208) evaluated the effect of different PM2 5 exposure
metrics in relation to lung function among children in wood smoke-impacted areas. The authors
found that the ambient-generated component of PM25 exposure was associated with decrements in
lung function only among children not using inhaled corticosteroids, whereas no association was
reported with the nonambient exposure component. All of the ambient concentrations were
associated with decrements in both PEF and maximal expiratory flow (MEF). There were no
associations between any exposure metrics and forced vital capacity (FVC). The authors suggest that
lung function may be especially sensitive to the combustion-generated component of ambient PM2 5,
whereas airway inflammation may be more closely related to some other source.
In a longitudinal study, Liu et al. (2009, 192003) examined the association between acute
increases in ambient air pollutants and pulmonary function among children (ages 9-14 yr) with
asthma. FEVi and FEF25_75o/0 exhibited a consistent trend of negative associations with PM25 across
lag days 0, 1, 0-1, and 0-2, with the strongest effects for FEF25.75o/0 on lag day 0 (-1.12% [95% CI:
-2.06 to -0.18]) and lag days 0-1 (-1.18% [95% CI: -2.24 to -0.12]). Copollutant models including
O3, SO2 or NO2 did not result in marked changes in the PM2 5 risk estimates for FEVi or FEF25_75o/0.
Moshammer and Neuberger (2003, 041956) used a novel technique for assessing exposure to
PM in a study they conducted in Austria. They employed a diffusion charging particle sensor (model
LQ 1-DC, Matter Engineering AG, Wohlen, Switzerland) and a photoelectric aerosol sensor (model
PAS 2000 CE, EcoChem Analytics, League City, TX) to relate the spirometry scores of Upper
Austrian children, aged 7-10 yr, to particle surface area and particle-bound PAH concentration,
respectively. Details on these methods for measuring surface area and PAH can be found in Shi et al.
(2001, 078292) and Burtscher (2005, 155710). respectively. By measuring the surface area
distribution, it was possible to understand potential for contact area with respiratory tract cells. The
authors found that acute decrements of pulmonary function (FVC, FEVi, MEF50) were related to the
active surface of particles after adjustment for PMi0. For short-term lung impairments, this indicates
that active particle surface is a better index of exposure than PM mass.
A number of additional panel studies conducted outside of the U.S. and Canada also examined
lung function using more traditional exposure metrics. Several European and Asian studies reported
December 2009 6-96
-------
associations with PM measurements and decrements in pulmonary function (FEVi, FVC, FEF, MEF,
PEF rate) (Hogervorst et al., 2006, 156559; Hong et al., 2007, 091347; Moshammer et al, 2006,
090771; Odajima et al., 2008, 192005; Peacock et al., 2003, 042026; Peled et al., 2005, 156015).
Others found little evidence for a relationship between PM and daily changes in PEF after correction
for the confounding effects of weather, trends in the data, and autocorrelation (Fischer et al., 2002,
025731; Holguin et al., 2007, 099000; Just et al., 2002, 035429; Preutthipan et al., 2004, 055598;
Ranzi et al., 2004, 089500; Ward, 2003, 157111).
Adults
Trenga et al. (2006, 155209) examined personal, residential, and central site monitoring of
particles and the relationship with lung function in Seattle. In models controlling for gaseous
copollutants (CO, NO2), adults, regardless of COPD status, experienced a decline in FEVi associated
only with measurements of PM2.5 at the central site: each 10 ug/m3 increase in lag 0 PM2.5 was
associated with a 35.3 mL (95% CI: -70 to -1.0) decrease in FEVi. This represents a 2.2% decline in
mean FEVi (mean 1.6 L during the study). Results for personal, indoor and outdoor measures of
PM2.5 were inconsistent. No statistical associations were reported with outdoor PMi0_2.5.
Girardot et al. (2006, 088271) assessed the effects of PM25 on the pulmonary function of adult
day hikers in the Great Smoky Mountains National Park. Hikers performed spirometry both before
their hike and when they returned from their hike. The authors reported no statistically significant
responses in pulmonary function with an average of five hours of outdoor exercise at ambient PM25
levels that were below the current NAAQS. Specifically, post-hike percentage changes in FVC,
FEVi, FEVi/FVC, FEF25_75, and PEF were not associated with PM25 exposure.
Ebelt et al. (2005, 056907) developed an approach to separately estimate exposures to PM of
ambient and non-ambient origin based on a mass balance model. These exposures were linked with
respiratory and cardiovascular health endpoints for 16 patients with COPD in Vancouver, Canada
(mean age 74 yr). Effect estimates for estimated ambient exposure were generally equal to or larger
than those for the respective ambient concentration levels for post-FEV and AFEVi, and were
statistically significant for all AFEVi comparisons (estimated from figure).
Several studies outside of the U.S. and Canada examined the relationship between PM
concentrations and lung function and all reported a decrease in lung function in adults (FEVi, FVC,
PEFR) associated with PM exposure (Boezen et al., 2005, 087396; Bourotte et al., 2007, 150040;
Lagorio et al., 2006, 089800; Lee et al., 2007, 093042; McCreanor et al., 2007, 092841; Penttinen
etal, 2006, 087988).
Measures of Oxygen Saturation
Oxygen saturation measures the percentage of hemoglobin binding sites in the bloodstream
occupied by oxygen. DeMeo et al. (2004, 087346) estimated the change in oxygen saturation and
mean PM25 concentration in the previous 24 h in a panel of elderly subjects. They used the same
panel of elderly Boston residents (n = 28) and study protocol and analytic methods (12 wk of
repeated oxygen saturation measurements) as Gold et al. (2005, 087558) and Schwartz et al. (2005,
074317) in studies of ST-segment depression and HRV, respectively. At each clinic visit, subjects had
5 min each of rest, standing, post-exercise rest, and 20 cycles of paced breathing. The median PM25
concentration during the study period was 10.0 (ig/m3 (Schwartz et al., 2005, 074317). Each
10 (ig/m3 increase in the mean PM25 concentration in the previous 6 h was associated with a 0.15%
decrease in oxygen saturation (95% CI: -0.22 to 0.0) during the baseline rest period. Each 10 (ig/m3
increase in mean 6-h PM2 5 concentration was also associated with a decline in oxygen saturation
during the post-exercise period (-0.15% [95% CI: -0.22 to 0.0]), and post-exercise paced breathing
period (-0.07% [95% CI: -0.22 to 0.0]), but not during the exercise period. The authors suggest that
these oxygen saturation reductions may result from pulmonary vascular and inflammatory changes.
In a similar study, Goldberg et al. (2008, 180380) examined the association between oxygen
saturation, pulse rate, and ambient PM25, NO2, and SO2 concentrations in a panel of 31 subjects in
Montreal, with NYHA Class II or III heart failure who were aged 50-85 yr. Although each 10 (ig/m3
increase in PM25 on lag day 0 was associated with a -0.119 (95% CI = -0.196 to -0.042) change in
oxygen saturation in unadjusted models, once adjusted for temperature and barometric pressure, the
estimated change was smaller and no longer significant (-0.077 [95% CI = -0.160 to 0.007). Only
December 2009 6-97
-------
SO2 was significantly associated with reduced oxygen saturation in copollutant models. None of the
pollutants examined, including PM2.5, were associated with a change in pulse rate.
6.3.2.2. Controlled Human Exposure Studies
As with respiratory symptoms, there is little evidence from controlled human exposure studies
of PM-induced changes in pulmonary function. One study cited in the 2004 PM AQCD (U.S. EPA,
2004, 056905) noted a significant decrement in thoracic gas volume in healthy adults following a 2-h
exposure to PM2.5 CAPs (92 ug/m3); however, no significant changes were observed in spirometric
measurements, diffusing capacity (DLCO), total lung capacity, or airways resistance (Petrovic et al.,
2000, 004638). Other studies found no significant changes in pulmonary function in healthy adults
following exposure to inhaled iron oxide particles (Lay et al., 2001, 020613) or UF EC (Frampton,
2001, 019051). or in healthy and asthmatic adults following exposure to CAPs (Ohio et al., 2000,
012140; Gong et al., 2000, 155799; 2003, 087365). Rudell et al. (1996, 056577) reported a
significant increase in specific airways resistance following exposure to DE, an effect that was not
attenuated by reducing the particle number by 46% (2.6xlO6particles/cm3 compared with 1.4xl06
particles/cm ) using a particle trap. The particle trap did not affect the concentrations of other
measured diesel emissions including NO2, NO, CO, or total hydrocarbons. As described below, more
recent controlled human exposure studies provide limited and inconsistent evidence of changes in
lung function following exposure to particles from various sources.
CAPs
Among a group of healthy and asthmatic adults exposed to UFPs (Los Angeles, mean
concentration 100 ug/m3), Gong et al. (2008, 156483) observed small, yet statistically significant
decrements in arterial oxygen saturation immediately following exposure, 4 h post-exposure, and
22 h post-exposure (0.5% mean decrease relative to filtered air across all time points, p < 0.05). A
statistically significant decrease in FEVi was also observed, but only at 22 h post-exposure (2%
decrease relative to filtered air, p < 0.05). The responses demonstrated in this study were not affected
by health status. No such effects were observed in a similar study conducted in Chapel Hill, NC
which exposed healthy adults to a lower concentration of UF CAPs (49.8 ug/m3) (Samet et al.,
2009, 191913). In addition, two studies evaluating effects of exposure to PMi0_2.5 CAPs (average
concentration 89-157 ug/m3) on lung function observed no changes in spirometric measurements,
DLCO or arterial oxygen saturation 0-22 h post-exposure in asthmatic or healthy adults (Gong et al.,
2004, 055628: Graff et al., 2009, 191981). While Gong et al. (2004, 087964) did not observe a
significant association between exposure to PM25 CAPs and spirometry in older subjects (60-80 yr),
the investigators did report a decrease in oxygen saturation immediately following CAPs exposure.
This effect was observed more consistently in healthy older adults than in older adults with COPD.
These findings were confirmed by a subsequent study conducted by the same laboratory (Gong et
al., 2005, 087921). The authors also observed a small decrease in MMEF following a 2-h exposure
to PM2 5 CAPs (200 ug/m3) which was more pronounced in healthy subjects.
Urban Traffic Particles
Neither short-term exposure to relatively high levels of urban traffic particles nor longer
exposures to lower concentrations of urban particles have been observed to alter pulmonary function
in controlled exposures among healthy adults. Larsson et al. (2007, 091375) exposed 16 adults for
2 h to PM25 concentrations of 46-81 ug/m3 in a room adjacent to a busy road tunnel, with
concomitant exposure to NO2(0.12 ppm), NO (0.71 ppm), and CO (5 ppm). Although respiratory
effects in this study were not compared to filtered air control, no difference in lung function was
observed 14 h after exposure to traffic particles relative to lung function measured on a day
following typical activities that did not include transit though a road tunnel. In a study of 24-h
exposures to urban traffic particles (PM25 9.7 ug/m3), no change in lung function was reported at
2.5 h after the start of exposure relative to filtered air (Brauner et al., 2009, 190244).
December 2009 6-98
-------
Diesel Exhaust
Mudway et al. (2004 180208) exposed 25 healthy adults to DE with an average particle
concentration of 100 ug/m and observed mild bronchoconstriction (airways resistance) immediately
following exposure relative to filtered air. No changes were observed in FEVi or FVC following DE
exposure in these subjects, or in a group of 15 asthmatics exposed using the same protocol (Mudway
et al., 2004, 180208; Stenfors et al., 2004, 157009).
Model Particles
Pietropaoli et al. (2004, 156025) observed a reduction in MMEF and DLCO in healthy adults
21 h after a 2-h exposure to UF carbon particles (50 ug/m3). This reduction in DLCO may reflect a
PM-induced vasoconstrictive effect on the pulmonary vasculature. Tunnicliffe et al. (2003, 088744)
did not observe any significant change in lung function following exposure to ammonium bisulfate
or aerosolized H2SO4 (200 and 2,000 ug/m3) in healthy or asthmatic adults, which is consistent with
findings of the majority of studies of controlled exposures to acid aerosols presented in the last two
PM AQCDs (U.S. EPA, 1996, 079380; 2004, 056905).
Summary of Controlled Human Exposure Study Findings for Pulmonary Function
Taken together, the majority of controlled human exposure studies do not provide evidence of
PM-induced changes in pulmonary function; however, some investigators have observed slight
decreases in DLCO, MMEF, FEVi, oxygen saturation, or increases in airways resistance following
exposure to CAPs, DE, or UF EC.
6.3.2.3. lexicological Studies
The 2004 PM AQCD (U.S. EPA, 2004, 056905) included three animal toxicological studies
which measured pulmonary function following multiday short-term inhalation exposure to CAPs. A
decreased respiratory rate was noted in the one study involving dogs. Increased tidal volume was
observed in one study involving rats while no changes were observed in the other rat study. AHR
was found in four studies of mice, healthy rats or SH rats exposed to ROFA by IT instillation or
inhalation. Studies conducted since the last review are discussed below.
CAPs
SH rats exposed to Tuxedo, NY CAPs via nose-only inhalation for 4 h (mean concentration
73 ug/m3; single-day concentrations 80 and 66 ug/m3; 2/2001 and 5/2001, respectively) had a
statistically significant decreased respiratory rate compared with air-exposed controls (Nadziejko et
al., 2002, 087460). This measure was obtained from BP fluctuations using radiotelemetry. The
decrease in respiratory rate of 25-30 breaths/min was an immediate response to CAPs, beginning
shortly after the exposure began and ceasing with the end of exposure. It was accompanied by a
decrease in HR (Section 6.2.1.3). Rats were also exposed to fine (MMAD 160 nm; 49-299 ug/m3)
and UF H2SO4(MMAD 50-75 nm; 140-750 ug/m3) (Nadziejko et al., 2002, 087460) because H2SO4
aerosols have the potential to activate irritant receptors. Irritant receptors, found at all levels of the
respiratory tract, include rapidly-adapting receptors and sensory C-fiber receptors (Alarie, 1973,
070967; Bernardi et al., 2001, 019040; Coleridge and Coleridge, 1994, 156362; Widdicombe, 2003,
157145; Widdicombe, 2006, 155519). Activation of trigeminal afferents in the nose causes CNS
reflexes resulting in decreases in respiratory rate through a lengthened expiratory phase, closure of
the glottis, closure of the nares with increased nasal airflow resistance and effects on the
cardiovascular system such as bradycardia, peripheral vasoconstriction and a rise in systolic arterial
blood pressure. Sneezing, rhinorrhea and vasodilation with subsequent nasal vascular congestion are
also nasal reflex responses involving the trigeminal nerve (Sarin et al., 2006, 191166). Activation of
vagal afferents in the tracheobronchial and alveolar regions of the respiratory tract causes CNS
reflexes resulting in bronchoconstriction, mucus secretion, mucosal vasodilation, cough, and apnea
December 2009 6-99
-------
followed by rapid shallow breathing. Besides effects on the respiratory system, effects on the
cardiovascular system can also occur including bradycardia and hypotension or hypertension. Fine
H2SO4 induced an overall decrease in respiratory rate, with UF H2SO4 resulting in elevated
respiratory rate compared to control (Nadziejko et al., 2002, 087460). The authors suggested that
both CAPs and fine H2SO4 aerosols activated sensory irritant receptors in the upper airways,
resulting in a decreased respiratory rate. The response to UF H2SO4 aerosols differed from the other
responses and was thought to be due to deposition of UFPs deeper into the lung with the subsequent
activation of pulmonary irritant receptors which trigger an increase in respiratory rate. Since irritant
receptors in nasal, tracheobronchial and alveolar regions act via trigeminal- and vagal-mediated
pathways, this study indicates a role for neural reflexes in respiratory responses to CAPs.
Kodavanti et al. (2005, 087946) measured respiratory frequency 1 day after a 2-day exposure
of SH and WKY rats to CAPs from RTP, NC (mean mass concentration range 144-2,758 (ig/m3;
<2.5 (im in size; 8/27-10/24/2001) for 4 h/day. Increases in inspiratory and expiratory times were
seen in SH, but not WKY rats, exposed to CAPs compared with filtered air controls.
Effects of CAPs on pulmonary function were also investigated in a rat model of pulmonary
hypertension using SD rats pre-treated with monocrotaline (Lei et al., 2004, 087999). In this study,
rats were exposed to CAPs from an urban high traffic area in Taiwan (mean mass concentration
371 (ig/m3) for 6 h/day on three consecutive days and pulmonary function was evaluated 5 h
post-exposure using whole-body plethysmography. A statistically significant decrease in respiratory
frequency and an increase in tidal volume were observed following CAPs exposure, along with an
increase in airway responsiveness (measured as Penh) following Mch challenge.
In many animal studies changes in ventilatory patterns are assessed using whole body
plethysmography, for which measurements are reported as enhanced pause (Penh). Some
investigators report increased Penh as an indicator of AHR, but these are inconsistently correlated
and many investigators consider Penh solely an indicator of altered ventilatory timing in the absence
of other measurements to confirm AHR. Therefore use of the terms AHR or airway responsiveness
has been limited to instances in which the terminology has been similarly applied by the study
investigators.
Diesel Exhaust
Li et al. (2007, 155929) exposed BALB/c and C57BL/6 mice to clean air or to low dose DE
(containing 100 (ig/m3 particles) for 7 h/day and 5 days/wk for 1, 4 and 8 wk. Average gas
concentrations were reported to be 3.5 ppm CO, 2.2 ppm NO2, and <0.01 ppm SO2. AHR was
evaluated by whole-body plethysmography at day 0 and after 1, 4 and 8 wk of exposure. Exposure to
DE for 1 wk resulted in an increased sensitivity of airways to Mch, measured as Penh, in C57BL/6
but not BALB/c, mice. Other short-term responses of this study are discussed in Sections 6.3.3.3 and
6.3.4.2.
McQueen et al. (2007, 096266) investigated the role of vagally-mediated pathways in
respiratory responses to PM. Respiratory minute volume (RMV) was increased in anesthetized
Wistar rats 6 h after treatment with 500 (ig DE particles (SRM2975) by IT instillation. This response
was blocked by severing the vagus nerve or pretreatment with atropine. The absence of a respiratory
response with vagotomy or atropine indicated that the increase in RMV following DE particle
exposure involved a neural reflex acting via vagal afferents. No statistically significant changes in
mean BP, HR or HRV were observed in response to DE particles in this study. Vagally-mediated
inflammatory responses to DEP were also observed in this study and are discussed in
Section 6.3.3.3.
Model Particles
In a study by Last et al. (2004, 097334). BALB/c mice were exposed to 250 (ig/m3
laboratory-generated iron-soot (size range 80-110 nm; about 200 (ig/m3 as soot) for 4 h/day and
3 days/wk for 2 wk. Pulmonary function was measured by whole-body plethysmography after
challenge with Mch. No AHR, as measured by Penh, was observed following 2-wk exposure to
iron-soot. Other findings of this study are reported in Sections 6.3.3.3 and 6.3.5.3.
December 2009 6-100
-------
Summary of Toxicological Study Findings for Pulmonary Function
Several recent studies demonstrated alterations in respiratory frequency and in airway
responsiveness following short-term exposure to CAPs and DE. Two studies provide evidence for
the involvement of irritant receptors and vagally-mediated neural reflexes in mediating changes in
respiratory functions.
6.3.3. Pulmonary Inflammation
The discussion of the effects of PM on pulmonary inflammation in the 2004 PM AQCD
(U.S. EPA, 2004, 056905) was limited by a relative lack of information from controlled human
exposure and toxicological studies. Although no epidemiologic studies of pulmonary inflammation
were described in the 2004 PM AQCD (U.S. EPA, 2004, 056905). several recent studies have
observed a positive association between PM concentration and exhaled NO (eNO). New controlled
human exposure and toxicological studies have also generally observed an increase in markers of
inflammation in the pulmonary compartment following exposure to PM.
6.3.3.1. Epidemiologic Studies
No epidemiologic studies of pulmonary inflammation were described in the 2004 PM AQCD
(U.S. EPA, 2004, 056905).
Exhaled Nitric Oxide -Asthmatic Children
Exhaled NO, a biomarker for airway inflammation, was the outcome studied in panels of
asthmatic children in southern California (Wu et al, 2006, 157156) and Seattle (Allen et al., 2008,
156208: Koenig et al., 2003, 156653: 2005, 087384: Mar et al., 2005, 088759). Mean concentration
data from these studies are summarized in Table 6-10. Delfino et al. (2006, 157156) followed 45
asthmatic children for ten days with offline fractional eNO and examined the associations with
exposures to personal PM2.5 and 24-h PM2.5, EC and OC as well as ambient PM2.5, EC and OC. The
strongest associations were between eNO and 2-day avg pollutant concentrations: for a 10 (ig/m3
increase in personal PM2.5, eNO increased by 0.46 ppb (95% CI: 0.04-0.79); for 0.6 (ig/m3 personal
EC, eNO increased by 0.7 ppb (95% CI: 0.3-1.1). An association with exposure to ambient PM25
was only statistically significant in 19 subjects taking inhaled corticosteroids: for each 10 (ig/m
increase in PM2.5, eNO increased by 0.77 ppb (95% CI: 0.07-1.47).
In a panel of 19 asthmatic children in Seattle, effects were observed only among the ten
non-users of inhaled corticosteroids. For each 10 (ig/m3 increase in personal, outdoor, indoor, or
central site PM25, eNO increased from 3.82 ppb (associated with central site, 95% CI: 1.22-6.43) to
4.48 ppb (with personal PM2.5, 95% CI: 1.02-7.93) (Koenig et al., 2003, 156653). Further analysis
examining the association between eNO and outdoor and indoor-generated particles suggested that
eNO was associated more strongly with ambient particles, but only for non-users of medication: each
10 (ig/m3 increase in estimated ambient PM25 results in an increase in eNO of 4.98 ppb (95% CI:
0.28-9.69) (Koenig et al., 2005, 087384).
Also in Seattle, WA, Mar et al. (2005, 088759) examined the association between eNO and
ambient PM25 concentration among children (aged 6-13 yr) recruited from an asthma/allergy clinic.
Fractional exhaled nitric oxide (FeNO) was associated with hourly averages of PM25 up to 10-12 h
after exposure. Each 10 (ig/m3 increase in 1-h mean PM25 concentration was associated with a 6.99
ppb increase in eNO (95% CI: 3.43-10.55) among children not taking inhaled corticosteroids, but
associated with only a 0.77 ppb decrease in eNO (95% CI: -4.58 to 3.04) among those taking inhaled
corticosteroids.
Allen et al. (2008, 156208). in a reanalysis of data from Koenig et al. (2005, 087384).
evaluated the effect of different PM2 5 exposure metrics in relation to airway inflammation among
children in wood smoke-impacted areas of Seattle. The authors found that for the nine non-users of
inhaled corticosteroids, the ambient-generated component of PM2 5 exposure was associated with
respiratory responses, both airway inflammation and decrements in lung function, whereas the non-
ambient PM2 5 exposure component was not. They did note, however, different relationships for
December 2009 6-101
-------
airway inflammation and decrements in lung function, with the former significantly associated with
total personal PM2 5, personal light-absorbing carbon (LAC), and ambient generated personal PM2.5
and the latter related to ambient PM2 5 and its combustion markers. The different results between
FeNO and lung function were not unexpected; epidemiologic data show that airway inflammation
indicated by FeNO does not correlate strongly with either respiratory symptoms or lung function
(Smith and Taylor, 2005, 192176). The authors conclude that lung function decrements may be
associated with the combustion-generated component of ambient PM2.5, whereas airway
inflammation may be related to some other component of the ambient PM25 mixture.
In a longitudinal study, Liu et al. (2009, 192003) examined the association between acute
increases in ambient air pollutants and FeNO among children (ages 9-14 yr) with asthma. FeNO had
a trend of positive associations with PM25, with the strongest association on lag day 0 (3.12% [95%
CI: -2.12 to 8.82]). Copollutant models including O3, SO2 or NO2 did not result in marked changes in
the PM2 5 risk estimates for FeNO.
A few studies outside of the U.S. examined eNO in relation to PM exposure among children.
Fischer et al. (2002, 025731) and Murata et al. (2007, 189159) found a statistical association
between increases in PM and increases in the percent of eNO. Holguin et al. (2007, 099000) found
no association between exposure to PM and eNO. However, they did see statistical associations
between increases in eNO for the 95 asthmatic subjects and measures of road density of roads 50-
and 75-m from the home.
Exhaled Nitric Oxide -Adults
Three recent panel studies examined the effects of particle exposure on eNO measured in older
adults (Adamkiewicz et al., 2004, 087925 in Steubenville, OH; Adar et al.. 2007. 001458; Jansen et
al., 2005, 082236 in Seattle). Mean concentration data from these studies are characterized in Table
6-10. Breath samples were collected weekly for 12 weeks from a group of 29 elderly adults in
Steubenville, OH (Adamkiewicz et al., 2004, 087925). In single-pollutant models, each 10 (ig/m3
increase in 24-h ambient PM2.5 increased eNO by 0.82 ppb (95% CI: 0.19-1.45), a change of 15%
compared to mean eNO (9.9 ppb). Effects were essentially unchanged in copollutant models that
included ambient and/or indoor NO. The effect estimates for the seven COPD subjects were higher
than for normal subjects (2.20 vs. 0.45 ppb, p = 0.03) (Adamkiewicz et al., 2004, 087925).
In the Seattle panel of older adults (aged 60-86 yr), seven subjects were asthmatic and nine
had a diagnosis of COPD (five with asthma and four without) (Jansen et al., 2005, 082236). Exhaled
NO was measured daily for 12 days, along with personal, indoor, outdoor and central site PMi0,
PM2 5 and BC. The strongest associations between 24-h avg PM and eNO were found for the
asthmatic subjects: 10 (ig/m3 increases in outdoor levels (measured outside the subjects' homes) of
PM25 or PMio were associated with increases in eNO of 4.23 ppb (95% CI: 1.33-7.13), an increase
of 22% above the group mean of 19.2 ppb, and 5.87 ppb (95% CI: 2.87-8.88), an increase of 31%,
respectively. BC measured indoors, outdoors or personally was also associated with increases in
eNO (of 3.97, 2.32, and 1.20 ppb, respectively) (Jansen et al., 2005, 082236).
Adar et al. (2007, 001458) conducted a panel study of 44 non-smoking senior citizens residing
in St. Louis, MO. As part of the study, subjects were taken on group trips to a theater performance,
Omni movie, outdoor band concert, and a Mississippi River boat cruise. Subjects were driven to and
from each event aboard a diesel bus. Before and after each bus trip, eNO was measured on each
subject. Two carts containing continuous air pollution monitors were used to measure group-level
micro-environmental exposures to PM25, BC, and size-specific particle counts (0.3-2.5 (im and
2.5-10 (im) on the day of each trip. Each 10 ug/m3 increase in 24-h mean PM2 * concentration was
associated with a 36% increase in eNO pre-trip (95% CI: 5-71). Each 10 (ig/m increase in
micro-environmental PM25 concentration (i.e., during the bus ride) was associated with a 27%
increase in eNO post-trip (95% CI: 17-38).
These studies all demonstrated an association between increased levels of eNO and increases
in PM in the previous 4-24 h. Further, three studies demonstrated effects in elderly populations
(Adamkiewicz et al.. 2004. 087925: Adar et al.. 2007. 001458: Jansen et al., 2005, 082236) while
four others reported a similar acute increase in eNO among children (Delfino et al., 2006, 090745;
Koenig et al., 2003, 156653; 2005, 087384; 2005, 088999).
Outside of the U.S., one study examined eNO in a panel of 60 adult asthmatic subjects in
London. McCreanor et al. (2007, 092841) reported that 1 (ig/m3 increase in personal exposure to EC
December 2009 6-102
-------
was associated with increases of approximately 1.75-2.25% in eNO (results were presented
graphically only) for up to 22 h post-exposure.
Other Biomarkers of Pulmonary Inflammation and Oxidative Stress
Other biomarkers of respiratory distress that have been examined in recent panel studies
include urinary leukotriene E4 (LTE4) in asthmatic children (Rabinovitch et al., 2006, 088031); two
oxidative stress markers: TEARS and 8-isoprostane in asthmatic children (Liu et al., 2009, 192003)
and breath acidification in adolescent athletes (Ferdinands et al., 2008, 156433). Mean concentration
data from these studies are characterized in Table 6-10.
In Rabinovitch et al. (2006, 088031). LTE4, an asthma-related biological mediator, was used to
study the response to short-term particle exposure. In the second winter of their 2-yr study of
asthmatic children (described above in Section 6.3.1.1), urine samples were collected at
approximately the same time of day from 57 subjects for eight consecutive days. Controlling for
days with URI symptoms, each 10 (ig/m3 increase in morning maximum PM25 (measured by
TEOM), was associated with an increase in LTE4 levels by 5.1% (95% CI: 1.6-8.7). No statistically
significant effects were observed on the same day or up to 3 days later based on 24-h averaged
concentrations from the TEOM monitor or from the FRM central site monitor.
In a longitudinal study conducted in Windsor, Ontario, Liu et al. (2009, 192003) examined the
association between acute increases in ambient air pollutants and TEARS and 8-isoprostane among
children (ages 9-14 yr) with asthma. TEARS, but not 8-isoprostane, was positively associated with
PM2.5 (percent change in TEARS 40.6% [95% CI: 11.8-81.3], lag 0-2 days). The association with
TEARS persisted for at least three days. Adverse changes in pulmonary function (Section 6.3.2.1)
were consistent with those of TEARS in response to PM2.5 with a similar lag structure, suggesting a
coherent outcome for small airway function and oxidative stress.
The effects of vigorous outdoor exercise during peak smog season in Atlanta, GA on breath
pH, a biomarker of airway inflammation, in adolescent athletes (n = 16, mean age = 14.9 yr) were
examined by Ferdinands et al. (2008, 156433). Median pre-exercise breath pH was 7.58 (range
4.39-8.09) and median post-exercise breath pH was 7.68 (range 3.78-8.17). The authors observed no
significant association between ambient PM and post-exercise breath pH. However both pre- and
post-exercise breath pH were strikingly low in these athletes when compared to 14 relatively
sedentary healthy adults and to published values of breath pH in healthy subjects. The authors
speculate that repetitive vigorous exercise may induce airway acidification.
Effect of Measurement Location on Studies of Pulmonary Function and
Inflammation
A number of studies examining exposure to PM2.5 and pulmonary function and inflammation
have compared the results of exposure assessment based on concentrations recorded from personal,
indoor, outdoor, and/or ambient monitors (Allen et al., 2008, 156208; Delfmo et al., 2004, 056897;
Delfmo et al., 2006, 090745; Koenig et al., 2005, 087384; Trenga et al., 2006, 155209). Two
investigations evaluated PM2 5 concentrations from indoor, outdoor, personal and central site
monitors and the relationship with FEVi. Delfmo et al. (2004, 056897) reported that personal
exposure estimates showed a stronger association with FEVi than any of the stationary exposures,
and that indoor exposure estimates were associated with a stronger effect than either outdoor or
central site exposure estimates. However, Trenga et al. (2006, 155209) reported the largest declines
in FEVi associated with central site exposure estimates, though the most consistent association with
declines in FEVi came from the exposure estimates measured by indoor monitors. Delfmo et al.
(2006, 090745) used personal and ambient exposure estimates in a study of FeNO among asthmatic
children and found that the personal exposure estimates were more robust than the ambient exposure
estimates. Two studies conducted in Seattle, WA partitioned personal exposure to PM2 5 into its
ambient-generated and indoor-generated components. Koenig et al. (2005, 087384) reported that
ambient-generated PM2 5 was consistently associated with an increase in FeNO, while the indoor-
generated component of PM25 was less strongly associated with FeNO. This could reflect the
difference in composition of indoor-generated PM25 as compared to ambient-generated PM25
Similarly, Allen et al. (2008, 156208) found that FeNO was associated with the ambient-generated
December 2009 6-103
-------
component of personal PM2.5 exposure, but not with ambient PM2.5 concentrations measured by
central site monitors. Overall, these studies provide a unique perspective on how measurement
location influences the findings of epidemiologic studies. This small group of studies indicates that
effects are associated with all types of PM measurement, suggesting health effects of both ambient-
generated and indoor-generated particles. It is likely that variability in season, meteorology,
topography, geography, behavior and exposure patterns contribute to the observed differences.
6.3.3.2. Controlled Human Exposure Studies
Studies of controlled human exposures presented in the 2004 PM AQCD (U.S. EPA, 2004,
056905) provided evidence of pulmonary inflammation induced by exposure to PM. Lay et al.
(1998, 007683) found that instillation of iron oxide particles (2.6 urn) produced an increase in
alveolar macrophages and neutrophils in bronchoalveolar lavage fluid (BALF) collected 24 h
post-instillation. Ohio and Devlin (2001, 017122) evaluated the inflammatory response following
bronchial instillation of particles extracted from filters collected in the Utah Valley both prior to and
after the closure of an area steel mill. Subjects who underwent pulmonary instillation of particles
(500 ug) collected while the steel mill was operating (n = 16) had significantly higher levels of
neutrophils 24 h post-instillation compared with either saline instillation or with subjects (n = 8) who
were instilled with the same mass of PM collected during the mill's closure. This finding indicates
that metals may be an important PM component for this health outcome. In an inhalation study of
exposure to PM2.5 CAPs (23-311 ug/m3) from Chapel Hill, NC, Ohio et al. (2000, 012140) observed
an increase in airway and alveolar neutrophils 18 h after the 2-h exposure. A similar finding was
reported by Rudell et al. (1999, 001964) following exposure to DE among healthy adults. In this
study, reducing the particle number from 2.6x106particles/cm3 to 1.3x10 particles/cm3 while
maintaining the concentration of gaseous diesel emissions was not observed to attenuate the
response. One study of controlled exposures to UF EC among healthy adults did not report particle-
related effects on eNO (Frampton, 2001, 019051). As summarized below, several recent studies of
controlled exposures have provided some additional evidence of pulmonary inflammation associated
with PM.
CAPS
A series of exposures to UF, PM2.5, and PMi0_2.5 CAPs from Los Angeles with average particle
concentrations between 100 and 200 ug/m3 have not been shown to have a significant effect on
markers of airway inflammation in healthy or health-compromised adults (Gong et al., 2004,
087964; 2004, 055628; 2005, 087921; 2008, 156483). However, two recent studies conducted in
Chapel Hill, NC reported significant increases in percent PMNs and concentration of IL-8 in BALF
among healthy adults 18-20 h following controlled exposures to PMi0_25 (89 ug/m3) and UF (49.8
ug/m3) CAPs, respectively (Graff et al., 2009, 191981; Samet et al., 2009, 191913). As discussed
above, the same laboratory previously reported a mild inflammatory response in the lower
respiratory tract following exposure to PM25 CAPs (Ghio et al., 2000, 012140). In a follow-up
analysis, Huang et al. (2003, 087377) found the increase in BALF neutrophils demonstrated by Ghio
et al. (2000, 012140) to be positively associated with the Fe, Se, and SO4 ~ content of the particles.
Alexis et al. (2006, 154323) recently evaluated the effect of PMi0_2.5 on markers of airway
inflammation, specifically focusing on the impact of biological components of PMi0_2.s. Healthy men
and women (n = 9) between the ages of 18 and 35 inhaled nebulized saline (0.9%) as well as
aerosolized PMi0_2.5 collected from ambient air. Subjects were exposed to PM10_2.s on two separate
occasions, once using PMi0_2.5 that had been heated to inactivate biological material and once using
non-heated PMi0_2.5. Approximately 0.65 mg PMi0_2.5 was deposited in the respiratory tract of
subjects during the exposures. Markers of inflammation and immune function were analyzed in
induced sputum collected 2-3 h after inhalation of saline or PMi0_2.5. Both heated and non-heated
PMio_2.5 were observed to increase the neutrophil response compared with saline. Exposure to
non-heated PM10_2.5 was found to increase levels of monocytes, eotaxin, macrophage TNF-a mRNA,
and was also associated with an upregulation of macrophage cell surface markers. No such effects
were observed following exposure to biologically inactive PMi0_2.s. These results suggest that while
PMio_2.5-induction of neutrophil response is not dependent on biological components, heat sensitive
December 2009 6-104
-------
components of PM10_2.5 (e.g., endotoxin) may be responsible for PM-induced alveolar macrophage
activation.
Traffic Particles
Larsson et al. (2007, 091375) exposed 16 healthy adults to air pollution in a road tunnel for 2 h
during the afternoon rush hour in Stockholm, Sweden. The median PM2.5 and PMi0 concentrations
during the road tunnel exposures were 64 ug/m3 and 176 ug/m3, respectively. Bronchial biopsies
were obtained and bronchoscopy and BAL were performed 14 h after the exposure. The results were
compared with a control exposure which consisted of exposure to urban air during normal activity.
The authors reported significant BALF increases in percentage of lymphocytes, total cell number,
and alveolar macrophages following exposure to road tunnel exposure versus control. These results
provide evidence of a significant association between exposure to road tunnel air pollution and
airway inflammation. However, unlike other controlled exposure studies, the control exposure was
not a true clean air control, but only a lower exposure group with no characterization of personal
exposure. In addition, it is not possible to separate out the contributions of each air pollutant,
including PM, on the observed inflammatory response.
Diesel Exhaust
In a recent study evaluating the effect of DE exposure on markers of airway inflammation,
Behndig et al. (2006, 088286) exposed healthy adults (n = 15) for 2 h with intermittent exercise to
filtered air or DE with a reported PM10 concentration of 100 ug/m3. Eighteen hours after exposure to
DE, the authors found significant increases in neutrophil and mast cell numbers in bronchial tissue,
as well as significant increases in neutrophil numbers and IL-8 in BALF compared with filtered air
control. Similarly, Stenfors et al. (2004, 157009) observed an increase in pulmonary inflammation
(e.g., airways neutrophilia and an increase in IL-8 in BALF) among healthy adults 6 h following
exposure to DE (PMi0 average concentration 108 ug/m3). It is interesting to note, however, that no
such inflammatory effects were observed in a group of mild asthmatic subjects in the same study.
The DE-induced neutrophil response in the airways of healthy subjects observed in these two studies
(Behndig et al., 2006, 088286; Stenfors et al., 2004, 157009) is qualitatively consistent with the
findings of Ohio et al. (2000, 012140) who exposed healthy subjects to Chapel Hill PM2.5 CAPs. In a
group of healthy volunteers, Bosson et al. (2007, 156286) demonstrated that exposure to O3 (2 h at
0.2 ppm) may enhance the airway inflammatory response of DE relative to clean air (1-h exposure to
300 ug/m3). Exposure to O3 was conducted 5 h after exposure to DE, and resulted in an increase in
the percentage of neutrophils in induced sputum collected 18 h after exposure to O3. In a subsequent
study using a similar protocol at the same concentrations, prior exposure to DE was shown to
increase the inflammatory effects of O3 exposure, demonstrated as an increase in neutrophil and
macrophage numbers in bronchial wash (Bosson et al., 2008, 196659).
Wood Smoke
Barregard et al. (2008, 155675) examined the effect of a short-term exposure (4 h) to wood
smoke (240-280 ug/m3) on markers of pulmonary inflammation in a group of healthy adults.
Exposure to wood smoke increased alveolar NO compared to filtered air (2.0 ppb versus 1.3 ppb) 3 h
after exposure. Although these results provide some evidence of a PM-induced increase in
pulmonary inflammation, the physiological significance of the relatively small increase in alveolar
NO is unclear.
Model Particles
Pietropaoli et al. (2004, 156025) observed a lack of airway inflammatory response 21 h after
exposure to UF EC particles (10-50 ug/m3) among healthy and asthmatic adults. The same laboratory
reported no effect of exposure to UF or fine ZnO (500 ug/m3) on total or differential sputum cell
December 2009 6-105
-------
counts 24 h after exposure in a group of healthy adults (Beckett et al., 2005, 156261). Tunnicliffe
et al. (2003, 088744) measured levels of eNO as a marker of airway inflammation following 1-h
controlled exposures to ammonium bisulfate or aerosolized H2SO4 (200 and 2,000 ug/m3) in a group
of healthy and asthmatic adults. While exposure to ammonium bisulfate increased the concentration
of eNO immediately following exposure in asthmatics, no such effect was observed in healthy
adults, or in either healthy or asthmatic adults following exposure to aerosolized H2SO4.
Instillation
Schaumann et al. (2004, 087966) investigated the inflammatory response of human subjects
instilled with PM2.5 (100 ug) collected from two different cities in Germany, Hettstedt and Zerbst.
Although endobronchial instillation of PM from both cities were shown to induce airway
inflammation, instillation of PM from the more industrial area (Hettstedt) resulted in greater influxes
of BALF monocytes compared to PM collected from Zerbst. The authors postulated that the
difference in response between PM from the two cities may be due to the higher concentration of
transition metals observed in the samples collected from Hettstedt. Another study reported no change
in inflammatory markers in nasal lavage fluid 4 and 96 h following intranasal instillation of DEP
(300 ug/nostril) in asthmatics and healthy adults (Kongerud et al., 2006, 156656). Pre-exposure of
DEP to O3 was not shown to have any effect on the response. Although not a cross-over design, these
findings suggest that exposure to DEP without the gaseous component of DE may have little effect
on inflammatory responses in human subjects.
Summary of Controlled Human Exposure Study Findings for Pulmonary
Inflammation
These new studies strengthen the evidence of PM-induced pulmonary inflammation; however,
the response appears to vary significantly depending on the source and composition of the particles.
6.3.3.3. lexicological Studies
The 2004 PM AQCD (U.S. EPA, 2004, 056905) discussed numerous studies investigating
pulmonary inflammation in response to CAPs, ROFA, DEPs, metals and acid aerosols. A wide
variety of responses was reported depending on the type of PM and route of administration. In
general, IT instillation exposure to fly ash and metal PM resulted in notable pulmonary
inflammation. In contrast, inhalation of sulfates and acid aerosols had minimal, if any, effect on
pulmonary inflammation. More recent animal toxicological studies using CAPs, DE and other
relevant PM types are summarized below.
CAPs
The 2004 PM AQCD (U.S. EPA, 2004, 056905) found that exposure to PM2.5 CAPs at
concentrations of 100-1,000 ug/m3 for 1-6 h/day and 1-3 days generally resulted in minimal to mild
inflammation in rats and dogs. Somewhat enhanced inflammation was observed in a model of
chronic bronchitis. Since the last review, numerous studies have investigated inflammatory responses
to PM2 5 and UF CAPs in both healthy and compromised animal models.
In one study of healthy animals, SD rats were exposed to CAPs for 4 h/day on 3 consecutive
days in Fresno, CA, in fall 2000 and winter 2001 (PM25 mean mass concentration 190-847 ug/m3)
(Smith et al., 2003, 042107). The particle concentrator used in these studies was capable of
enhancing the concentration of UF as well as fine particles. Immediately after exposure on the third
day, BALF was collected and analyzed for total cells and neutrophils. Statistically significant
increases were observed in numbers of neutrophils during the first week of the fall exposure period
and in numbers of total cells, neutrophils and macrophages during the first week of the winter
exposure period. CAPs concentrations were >800 ug/m during both of those weeks.
Two studies were conducted using CAPs in Boston. In a study by Godleski et al. (2002,
156478), healthy SD rats were exposed for 5 h/day for 3 consecutive days to CAPs ranging in
December 2009 6-106
-------
concentration from 73.5-733.0 (ig/m3. BALF and lung tissue were collected for analysis 1 day later.
Neutrophilic inflammation was indicated by a statistically significant increase in percent neutrophils
in BALF. Microarray analysis of RNA from lung tissue and BALF cells demonstrated increased gene
expression of pro-inflammatory mediators, markers of vascular activation and enzymes involved in
organic chemical detoxification. This study overlapped in part with previously described studies by
Saldiva et al. (2002, 025988) and Batalha et al. (2002, 088109) (Section 6.2.4.3). In another study
(Rhoden et al., 2004, 087969). healthy SD rats were exposed for 5 h to CAPs (mean mass
concentration 1228 fig/m3; June 20-August 16, 2002). A statistically significant increase in BALF
neutrophils was observed 24 h following CAPs exposure. Histological analysis confirmed the influx
of inflammatory cells (Section 6.3.5.3). Inflammation was accompanied by injury which is discussed
in Section 6.3.5.3.
Kodavanti et al. (2005, 087946) reported two sets of studies involving PM2.5 CAPs exposure
during fall months in RTP, NC. In the first study, SH rats were exposed to filtered air or CAPs (mean
mass concentration range 1,138-1,765 (ig/m3; <2.5 (im) for 4 h and analyzed 1-3 h later. No increase
in BALF inflammatory cells or other measured parameter was observed. In the second study, SH and
WKY rats were exposed to filtered air or CAPs (mean mass concentration range 144-2,758 (ig/m3;
<2.5 (im) for 4 h/day on 2 consecutive days and analyzed 1 day afterward. Differences in baseline
parameters were noted for the two rat strains since SH rats had greater numbers of BALF neutrophils
than WKY rats. Following the 2-day CAPs exposure, increased BALF neutrophils were observed in
the WKY rats but not in the SH rats compared with filtered air controls. Inflammation was not
accompanied by increases in BALF markers of injury (Section 6.3.5.3).
Two CAPs studies involving SH rats were conducted in the Netherlands. In the first, SH rats
were exposed by nose-only inhalation to CAPs (ranging in concentration from 270-3,660 (ig/m3 and
in size from 0.15-2.5 (im) from three different sites in the Netherlands (suburban, industrial and
near-freeway) for 6 h (Cassee et al., 2005, 087962). Increased numbers of neutrophils were
observed in BALF 2 days post-exposure compared to air controls. When CAPs exposure was used as
a binary term, the relationship between CAPs concentration and number of PMN in BALF was
statistically significant. In contrast, Kooter et al. (2006, 097547) reported no changes in markers of
pulmonary inflammation measured 18 h after a 2-day exposure (6 h/day) of SH rats to PM2.5 or
PM2.5+UFP CAPs from sites in the Netherlands (mean mass concentration range 399-3613 and
269-556 ug/m3, respectively; PM2.5 CAPs site in Bilthoven and PM2 5+UF CAPs site in freeway
tunnel in Hendrik-Ido-Ambacht).
Pulmonary inflammation was investigated in two studies using a rat model of pulmonary
hypertension (i.e., SD rats pre-treated with monocrotaline). In the first study, rats were exposed to
PM25 CAPs from an urban high traffic area in Taiwan (mean mass concentration of 371 (ig/m3) (Lei
et al., 2004, 087999) for 6 h/day on 3 consecutive days and BALF was collected 2 days later. A
statistically significant increase in total cells and neutrophils was observed in BALF. Levels of TNF-
a and IL-6 in the BALF were not altered by CAPs exposure. In the second study, rats were exposed
to PM25 CAPs (mean mass concentration 315.6 and 684.5 ug/m3 for 6 and 4.5 h, respectively;
Chung-Li area, Taiwan) during a dust storm event occurring March 18-19, 2002 (Lei et al., 2004,
087884). Only one animal served as control during the 6-h exposure (from 2100-300 on the first
exposure day) so results from that one animal were combined with that of three control animals from
the 4.5-h exposure (from 300-730) on the second exposure day. A statistically significant increase in
total cells and neutrophils in BALF occurred in both CAPs-exposed groups. In addition, increases in
BALF IL-6 and markers of injury (Section 6.3.5.3) were observed as a function of CAPs exposure.
In summary, pulmonary inflammation was noted in all three studies involving multiday
exposure of healthy rats to CAPs from different locations. No pulmonary inflammation was seen in
one study of SH rats exposed to CAPs for 4 h and analyzed 1-3 h later. In studies involving multiday
exposure of SH rats, one demonstrated pulmonary inflammation while two did not. In the rat
monocrotaline model of pulmonary hypertension, both single-day and multiday exposures to CAPs
resulted in mild pulmonary inflammation.
On-Road Exposures
In a study by Elder et al. (2004, 087354) old rats (21 mo) were exposed to on-road highway
aerosols (particle concentration range 0.95-3.13xl05 particles/cm3; mass concentration estimated to
be 37-106 (ig/m3; Interstate 90 between Rochester and Buffalo, NY) for 6 h on one or three
December 2009 6-107
-------
consecutive days. No increase in BALF inflammatory cells was observed 18 h post-exposure in any
of the treatment groups.
Urban Air
To evaluate inflammatory responses to ambient particles from vehicles, Wistar rats were
exposed to ambient urban air from a high traffic site (concentration range 22-225 (ig/m3 PMi0; Porto
Alegre, Brazil) or to the same air which was filtered to remove the PM (Pereira et al, 2007,
156019). Concentrations of gases were not reported. Compared with controls exposed to filtered
urban air, a significant increase in total number of BALF cells was observed 24 h following the 20 h
continuous exposure, but not following the 6 h of exposure to unfiltered urban air.
Diesel Exhaust
The 2004 PM AQCD (U.S. EPA, 2004, 056905) summarized findings of the 2002 EPA Diesel
Document regarding the health effects of DE. Short-term inhalation exposure to low levels of DE
results in the accumulation of diesel PM in lung tissue, pulmonary inflammation and alveolar
macrophage aggregation and accumulation near the terminal bronchioles. More recent studies are
summarized below.
Pulmonary inflammatory responses were investigated in C57BL/6 mice exposed to DE
7 h/day for 6 consecutive days (Harrod et al., 2003, 097046). Compared with controls, inflammatory
cell counts in BALF were increased in mice exposed to the higher concentration of DE (1,000 (ig/m
PM) but not in mice exposed to the lower concentration of DE (30 (ig/m3 PM). Concentrations of
gases present in the higher dose DE were reported to be 43 ppm NOX, 20 ppm CO and 364 ppb SO2.
In a second study evaluating DE effects on BALF inflammatory cells, no increases in numbers
of neutrophils, lymphocytes or eosinophils were observed in BALB/c mice exposed by inhalation to
500 or 2,000 (ig/iri DE particles for 4 h/day on 5 consecutive days (Stevens et al., 2008, 1570101
Concentrations of gases reported in this study were 4.2 ppm CO, 9.2 ppm NO, 1.1 ppm NO2, and
0.2 ppm SO2 for the higher concentration of DE. Transcriptional microarray analysis demonstrated
upregulation of chemokine and inflammatory cytokine genes, as well as genes involved in growth
and differentiation pathways, in response to the higher concentration of DE. No gene expression
results were reported for the lower concentration of DE. Sensitization and challenge with ovalbumin
(OVA) significantly altered these findings (Section 6.3.6.2). These results demonstrate that changes
in gene expression can occur in the absence of measurable pulmonary inflammation or injury
markers (Section 6.3.5.3).
Li et al. (2007, 155929) exposed mice to clean air or to low dose DE (100 (ig/m3 PM) for
7 h/day and 5 days/wk for 1, 4 and 8 wk as described in Section 6.3.2.3. Analysis of BALF and
histology of lung tissues was carried out at day 0 and after 1, 4 and 8 wk of exposure. Total numbers
of cells and macrophages in BALF were significantly increased in C57BL/6 mice, but not in
BALB/c mice, after 1-wk exposure to DE compared with 0 day controls. Neutrophils and
lymphocytes were increased after 1-wk exposure to DE in both strains compared with 0 day controls.
Differences in BALF cytokines were also noted between the two strains after 1-wk exposure to DE.
No changes were observed by histological analysis. Pulmonary function and oxidative responses
were also evaluated (Sections 6.3.2.3 and 6.3.4.2). Long-term exposure responses are discussed in
Sections 7.3.2.2, 7.3.3.2 and 7.3.4.1.
Healthy F344 rats and A/J mice were exposed to DE containing 30, 100, 300 and 1,000 (ig/m3
PM by whole body inhalation for 6 h/day, 7 days/wk for either 1 wk or 6 months in a study by Reed
et al. (2004, 055625). Concentrations of gases were reported to be from 2.0-45.3 ppm NO,
0.2-4.0 ppmNO2, 1.5-29.8 ppm CO and 8-365 ppb for SO2 in these exposures. One week of
exposure resulted in no measurable effects on pulmonary inflammation. Long-term exposure
responses are discussed in Section 7.3.3.2.
In a study by Wong et al. (2003, 097707). also reported by Witten et al. (2005, 087485).
F344/NH rats were exposed nose-only to filtered room air or to DE at concentrations of 35.3 (ig/m3
and 669.3 (ig/m3 PM (particle size range 7.2-294.3 nm) for 4 h/day and 5 days/wk for 3 wk. Gases
associated with the high dose exposure were reported to be 3.59 ppm NO, 3.69 ppm NOX, 0.1 ppm
NO2, 2.95 ppm CO, 518.96 ppm CO2 and 0.031 ppm total hydrocarbon. The focus of this study was
December 2009 6-108
-------
on the possible role of neurogenic inflammation in mediating responses to DE. Neurogenic
inflammation is characterized by both the influx of inflammatory cells and plasma extravasation into
the lungs following the release of neuropeptides from bronchopulmonary C-fibers. Pulmonary
inflammation was evaluated by histological analysis of lung tissue at the end of the 3-wk exposure
period. Following high, but not low, concentration exposure to DE, a large number of alveolar
macrophages was found in the lungs. Small black particles, presumably DE particles, were found in
the cytoplasm of these alveolar macrophages. Perivascular cuffing consisting of mononuclear cells
was also observed in high dose-exposed animals. Influx of neutrophils or eosinophils was not seen,
although mast cell number was increased in high-dose exposed animals. Pulmonary plasma
extravasation was measured by the 99mTechnecium-albumin technique and found to be
dose-dependently increased in the bronchi and lung parenchyma. Alveolar edema was also observed
by histology in high concentration-exposed animals. A significant decrease in substance P content in
lung tissue was reported in DE-exposed rats. These responses initially suggested that DE resulted in
stimulation of C-fibers and activation of a local axon reflex resulting in the repeated release of the
stored neuropeptide substance P. Subsequent experiments were conducted using capsaicin
pretreatment, which inhibits neurogenic inflammation by activating C-fibers and causing the
depletion of neuropeptide stores. Pretreatment with capsaicin was found to reduce the influx of
inflammatory cells, but not plasma extravasation, in response to DE. Hence, DE is unlikely to act
through bronchopulmonary C-fibers to cause neurogenic edema in this model, although there may be
a different role for bronchopulmonary C-fibers in mediating the inflammatory cell influx.
Stimulation of bronchopulmonary C-fibers can result in activation of both local and CNS
reflexes through vagal parasympathetic pathways. McQueen et al. (2007, 096266) investigated the
role of vagally-mediated pathways in acute inflammatory responses to DE particles. A statistically
significant increase in BALF neutrophils was observed 6 h after IT instillation treatment of
anesthetized Wistar rats with 500 ug DE particles (SRM2975). This response was blocked by
severing the vagus nerve or pretreatment with atropine (McQueen et al., 2007, 096266). Similarly,
atropine treatment blocked the increase in BALF neutrophils seen 6 h after DE particle exposure in
conscious Wistar rats. These results provide evidence for the involvement of a pulmonary vagal
reflex in the inflammatory response to DE particles.
In summary, several studies demonstrate that short-term inhalation exposure to DE
(100-1,000 ug/m PM) causes pulmonary inflammation in rodents. No attempt was made in these
studies to determine whether the responses were due to PM components or to gaseous components.
However, PM from DE was found to be capable of inducing an inflammatory response, as
demonstrated by the one IT instillation study described above. Evidence was presented suggesting
that DEP may act through bronchopulmonary C-fibers to stimulate pulmonary inflammation.
Gasoline Emissions and Road Dust
Healthy male Swiss mice were exposed to gasoline exhaust (635 ug/m3 PM and associated
gases) or filtered air for 15 min/day for 7, 14, and 21 days (Sureshkumar et al., 2005, 088306).
BALF was collected for analysis 1 h after the last exposure. Histological analysis was also carried
out at 7, 14, and 21 days. The number of leukocytes in BALF was increased after exposure to
gasoline exhaust, but this increase did not achieve statistical significance. However, levels of the
pro-inflammatory cytokines TNF-a and IL-6 were significantly increased in BALF following 14 and
21 days of exposure. Furthermore, inflammatory cell infiltrate in the peribronchiolar and alveolar
regions were observed by histology. Evidence of lung injury was also found (Section 6.3.5.3). In this
study, BALF analysis of inflammatory cells was a less sensitive indicator of pulmonary
inflammation than BALF analysis of cytokines and histological analysis of lung tissue. Results of
this study cannot entirely be attributed to the presence of PM in the gasoline exhaust since
0.11 mg/m3 SOX, 0.49 mg/m3 of NOX and 18.7 ppm of CO were also present during exposure.
Using ApoE"7" mice on a high-fat diet, Campen et al. (2006, 096879) studied the impact of
inhaled gasoline emissions and road dust (6 h/day><3 day) on pulmonary inflammation. For gasoline
emissions, the PM-containing atmosphere (PM mean concentration 61 ug/m3; NOX mean
concentration 18.8 ppm; CO mean concentration 80 ppm) failed to increase numbers of
inflammatory cells in BALF collected 18 h after the last exposure. However, a statistically
significant increase in total cells and macrophages was observed in response to resuspended road
dust (PM25) at 3,500 ug/m3, but not at 500 ug/m3.
December 2009 6-109
-------
Model Particles
In a study by Elder et al. (2004, 055642). pulmonary inflammation was investigated in two
compromised, aged animal models (11-14 mo old SH and 23 mo old F344) exposed by inhalation to
UF CB (count median diameter = 36 nm) at a relevant concentration (150 ug/m3). No changes in
BALF cells were seen 24 h post-exposure in either model.
An increase in BALF neutrophils was observed at 24 h, but not at 4 h, in WKY rats exposed to
UF carbon particles (median particle size 38 nm; mass concentration 180 ug/m3; mean number
concentration 1.6xl07 particles/cm3) for up to 24 h (Harder et al., 2005, 087371). Changes in HR
and HRV demonstrated in this study (Section 6.2.1.3) occurred much more rapidly than the
inflammatory response.
No evidence of pulmonary inflammation was found by analysis of BALF or histology one or
three days following 24-h exposure of SH rats to UF carbon particles under similar conditions
(median particle size 31 nm; mass concentration 172 ug/m3; mean number concentration 9.0xl06
particles/cm3) (Upadhyay et al., 2008, 159345). However increased expression of HO-1, ET-1, ETA
and ETB, tPA and, plasminogen activator-1 was found in lung tissue three days following exposure.
In a study by Gilmour et al. (2004, 054175), adult Wistar rats were exposed for 7 h to fine and
UF CB particles (mean mass concentration 1,400 and 1,660 ug/m3 for fine and UF CB, respectively;
mean number concentration 3.8><103 and 5.2><104 particles/cm , respectively; count median
aerodynamic diameter 114 nm and 268 nm, respectively). Both treatments resulted in increased
BALF neutrophils 16 h post-exposure, with the UFPs having the greater response. UFPs also
increased total BALF leukocytes and macrophage inflammatory protein-2 (MIP-2) mRNA in BALF
cells. Although these exposures may not be relevant to ambient exposures, this study demonstrated
the greater propensity of UF CB particles to cause a pro-inflammatory response compared with fine
CB particles.
In a study by Last et al. (2004, 097334). mice were exposed to 250 ug/m3 laboratory-generated
iron-soot over a 2-wk period as described in Section 6.3.2.3. BALF was collected 1-h after the last
exposure and analyzed for total cells. No increase in total cell number was observed following
iron-soot exposure. Other findings of this study are described in Sections 6.3.2.3 and 6.3.5.3.
Pinkerton et al. (2008, 190471) exposed young adult male SD rats to filtered air, iron, soot or
iron-soot for 6 h/day for 3 days. The iron particles were mainly less than 100 nm aerodynamic
diameter, while the soot particles were initially 20-40 nm in diameter but formed clusters of 100-
200 nm in diameter. The size-distribution of iron-soot particles was bimodal over 10-250 nm and
averaged 70-80 nm in diameter. Rats were exposed to 45, 57 and 90 ug/m3 iron or to 250 ug/m3 soot
alone or in combination with 45 ug/m3 iron. Increased levels of the pro-inflammatory cytokine IL-1J3
were observed in lung tissue of rats exposed for 6 h/day for 3 days to 90 ug/m3, but not 57 ug/m3,
iron. No change in BALF inflammatory cells was observed after exposure to 57 ug/m3 or 90 ug/m3
iron. Exposures to 250 ug/m3 soot in combination with 45 ug/m3 iron also resulted in increased
levels of lung IL-1(3 and activation of the transcription factor NF-KB. Levels of lung IL-1(3 were
increased in neonatal rats exposed to 250 ug/m3 soot in combination with 100, but not 30, ug/m3
iron. Other endpoints of this study are described in Section 6.3.4.2.
Summary of lexicological Study Findings for Pulmonary Inflammation
New studies involving short-term exposures to CAPs and urban air strengthen the evidence of
PM-induced pulmonary inflammation. In addition, several studies demonstrated pulmonary
inflammation in response to diesel and gasoline exhaust; however it is not known whether PM or
gaseous components of the exhaust were responsible for these effects. Mixed results were obtained
in studies using model particles such as CB and iron-soot.
6.3.4. Pulmonary Oxidative Responses
The results of a small number of controlled human exposure and toxicological studies
presented in the 2004 PM AQCD (U.S. EPA, 2004, 056905) provided some initial evidence of an
association between exposure to PM and pulmonary oxidative stress. Recent controlled human
December 2009 6-110
-------
exposure studies have provided support for previous findings of an increase in markers of pulmonary
oxidative stress following exposure to DE, and one new study has observed a similar effect
following controlled exposure to wood smoke. New findings from toxicological studies provide
further evidence that oxidative species are involved in PM-mediated effects. No epidemiologic
studies have evaluated the association between PM concentration and pulmonary oxidative response.
6.3.4.1. Controlled Human Exposure Studies
Two studies cited in the 2004 PM AQCD (U.S. EPA, 2004, 056905) observed effects on
markers of airway oxidative response in healthy adults following controlled exposures to fresh DE or
resuspended DE particles (Blomberg et al., 1998, 051246: Nightingale et al, 2000, 011659).
Several recent studies are described below which have further evaluated the oxidative response
following exposure to particles in human volunteers.
Diesel Exhaust
Pourazar et al. (2005, 088305) exposed 15 adults (11 males and four females) for 1 h to air or
DE (PMio concentration 300 ug/m ) in a controlled cross-over study. Bronchoscopy with airway
biopsy was performed 6 h after exposure. The expression of NF-KB, AP-1 (c-jun and c-fos), p38, and
JNK in bronchial epithelium was quantified using immunohistochemical staining. DE was observed
to significantly increase nuclear translocation of NF-KB, AP-1, phosphorylated p38, and
phosphorylated JNK; however, the findings of this study require confirmation with more quantitative
methods such as Western blot analysis. The observed activation of redox-sensitive transcription
factors by DE may result in the induction of pro-inflammatory cytokines. There is some evidence to
suggest that this bronchial response to DE is mediated through the epidermal growth factor receptor
signaling pathway (Pourazar et al., 2008, 156884). Behndig et al. (2006, 088286) evaluated the
upregulation of endogenous antioxidant defenses following exposure to DE (100 ug/m3 PMi0) in a
group of 15 healthy adults. Increases in urate and reduced GSH were observed in alveolar lavage,
but not bronchial wash, 18 h after exposure. In a study utilizing the same exposure protocol,
Mudway et al. (2004, 180208) observed an increase in GSH and ascorbate in nasal lavage fluid 6 h
following exposure to DE in a group of 25 healthy adults.
Wood Smoke
Barregard et al. (2008, 155675) observed a significant increase in malondialdehyde levels in
breath condensate of healthy volunteers (n = 13) immediately following and 20 h after a 4-h
exposure to wood smoke (240-280 ug/m PM).
Endobronchial Instillation
Schaumann et al. (2004, 087966) demonstrated an increased oxidant radical generation of
BALF cells following endobronchial instillation of urban particles compared with instillation of
particles collected in a rural area. The authors suggested that this difference was likely due to the
greater concentration of transition metals found in the urban particles.
Summary of Controlled Human Exposure Study Findings for Pulmonary Oxidative
Responses
Taken together, these studies suggest that short-term exposure to PM at near ambient levels
may produce mild oxidative stress in the lung. Limited data suggest that proximal and distal lung
regions may be subject to different degrees of oxidative stress during exposures to different pollutant
particles.
December 2009 6-111
-------
6.3.4.2. Toxicological Studies
The 2004 PM AQCD (U.S. EPA, 2004, 056905) reported one study which provided evidence
that ROS were involved in PM-mediated responses. This particular study used pre-treatment with the
antioxidant DMTU to block the neutrophilic response to ROFA. More recently, several studies
evaluated the effects of PM exposure on pulmonary oxidative stress. Oxidative stress can be directly
determined by measuring ROS or oxidation products of lipids and proteins. An indirect assay
involves measurement of the enzyme HO-1 or of the antioxidant enzymes SOD or catalase, all of
which can be induced by oxidative stress. Antioxidant interventions which inhibit or prevent
responses are a further indirect measure of oxidative stress playing a role in the pathway of interest.
CAPS
Gurgueira et al. (2002, 036535) measured oxidative stress as in situ CL. Immediately
following a 5-h PM2.5 CAPs exposure (mean mass concentration range 99.6-957.5 ug/m3; Boston,
MA) increased CL was observed in lungs of CAPs-exposed SD rats. CL evaluated after CAPs
exposure durations of 3 h was also increased but did not achieve statistical significance compared to
the filtered air group. When animals were allowed to recover for 24 h following the 5-h CAPs
exposure, CL levels returned to control values. Interestingly, a decrease in lung CL was observed in
rats breathing filtered air for three days compared with rats breathing room air for the same duration.
To compare potential particle-induced differences in in situ CL, rats were exposed to ROFA (1.7
mg/m3 for 30 min) or CB (170 ug/m3 for 5 h). Only the ROFA-treated animals exhibited increased
CL in lung tissue. Additionally, levels of antioxidant enzymes in the lung (MnSOD and catalase)
were increased in CAPs-exposed rats. A CAPs-associated increase in CL was also seen in the heart
(Section 6.2.9.3), but not the liver.
In a similar study, Rhoden et al. (2004, 087969) exposed SD rats for 5 h to PM2.5 CAPs from
Boston (mean mass concentration 1,228 ug/m3) or to filtered air. Significant increases in TEARS
and protein carbonyl content (a measure of protein oxidation) were observed 24 h post-exposure to
CAPs. Pretreatment with the thiol antioxidant NAC (50 mg/kg i.p.) 1-h prior to exposure prevented
not only the lipid and protein oxidation observed in response to CAPs, but also the increase in BALF
neutrophils and pulmonary edema in this model (Sections 6.3.3.3 and 6.3.5.3). Results of this study
demonstrate the key role played by oxidative stress in these CAPs-mediated effects.
A later study by Rhoden et al. (2008, 190475) investigated the role of superoxide in mediating
pulmonary inflammation following exposure to ambient air particles. In this study, adult SD rats
were exposed by IT instillation to 1 mg of SRM1649. Two hours prior to exposure, half of the rats
were pretreated with the membrane-permeable SOD mimetic MnTBAP (10 mg/kg, i.p.). MnTBAP
abrogated the inflammatory response, measured by increased BALF inflammatory cells, and the
increase in lung superoxide, measured by CL, observed 4 h following exposure to urban air particles.
Kooter et al. (2006, 097547) reported an increase in HO-1 in BALF and lung tissue measured
18 h after a 2-day exposure (6 h/day) of SH rats to PM2.5 or PM2.5+UF CAPs (mean mass
concentration range 399-3613 and 269-556 ug/m3, respectively; PM25 CAPs site in Bilthoven and
PM2 5+UF site in freeway tunnel in Hendrik-Ido-Ambacht, the Netherlands). This occurred in the
absence of any measurable pulmonary inflammation (Section 6.3.3.3).
Urban Air
To evaluate oxidative stress responses to ambient particles from vehicles, Wistar rats were
exposed to ambient urban air from a high traffic site (concentration range 22-225 ug/m3 PMi0; Porto
Alegre, Brazil) or to the same air which was filtered to remove the PM (Pereira et al., 2007,
156019). Several exposure regimens were carried out: 6- and 20-h continuous exposures or to
intermittent exposures of 5 h/day for four consecutive days. A significant increase in lipid
peroxidation (measured as malondialdehyde) was seen in lung tissue immediately following the 20-h
continuous exposure, but not following the 6-h exposure or the intermittent exposures.
Inflammation-related endpoints are described in Section 6.3.3.3.
December 2009 6-112
-------
Diesel Exhaust
Li et al. (2007, 155929) exposed mice to clean air or to low dose DE (100 (ig/m3 PM) for
7 h/day and 5 days/wk for 1, 4 and 8 wk as described in Section 6.3.2.3. HO-1 mRNA and protein
were increased in lung tissues of both mouse strains after 1 wk of DE exposure. In addition, AHR
and changes in BALF cells and cytokines were observed (Sections 6.3.2.3 and 6.3.3.3). Pretreatment
with the thiol antioxidant NAC (320 mg/kg, i.p.) on days 1-5 of DE exposure greatly attenuated the
AHR and inflammatory response seen after 1 wk of DE exposure. Long-term responses are
discussed in Sections 7.3.2.2, 7.3.3.2 and 7.3.4.1.
A study by Whitekus et al. (2002, 157142) investigated the adjuvant effects of DE particles in
an allergic animal model and is discussed in detail below (Section 6.3.6.3). Intervention with the
thiol antioxidants bucillamine and NAC inhibited the increases in allergen-specific IgE and IgGi as
well as the increases in protein carbonyl and lipid hydroperoxides in the lung following DE particle
exposure.
Gasoline Exhaust
Pulmonary oxidative stress was evaluated by measurement of CL and TEARS following
exposure of SD rats to gasoline engine exhaust (Seagrave et al., 2008, 191990). Animals were
exposed for 6 h in a nose-only inhalation exposure system. PM mass concentration was reported to
be 60 ug/m3; count median diameter 20 nm; mass median diameter 150 nm; while the concentrations
of gaseous copollutants were 104 ppm CO, 16.7 ppmNO, 1.1 ppmNO2 and 1.0 ppm SO2. A
statistically significant increase in lung CL was observed without a concomitant increase in lung
TEARS. Discordant results were also observed for road dust exposures in the heart (Section 6.2.9.3).
The discrepancy between oxidative stress indicators suggests that the responses may follow different
time courses. Furthermore, no CL was seen when the gasoline exhaust was filtered to remove the
particulate fraction.
Model Particles
Increased expression of HO-1 was observed in lung tissue three days following 24-h exposure
of SH rats to UF carbon particles (median particle size 31 nm; mass concentration 172 ug/m3; mean
number concentration 9.0x106 particles/cm3) despite no evidence of pulmonary inflammation
(Section 6.3.3.3) (Upadhyay et al., 2008, 159345)
In a study conducted by Pinkerton et al. (2008, 190471). young adult male SD rats were
exposed to filtered air, soot, iron or iron-soot for 6 h/day for three days as described in Section
6.3.3.3. A statistically significant decrease in total antioxidant power and a statistically significant
increase in glutathione-S-transferase activity were observed in lung tissue from rats exposed to 90
(ig/m3 iron. This high concentration iron exposure also resulted in increased levels of ferritin protein
in lung tissue, indicating the presence of free iron which has the potential to redox cycle and cause
oxidative stress. Lung tissue total antioxidant power was decreased and glutathione redox ratio was
increased by the combined exposure to 250 (ig/m3 soot and 45 (ig/m3 iron. The iron-soot exposure
also increased oxidized glutathione in BALF and lung tissue. These results demonstrate that co-
exposure to soot enhanced iron-mediated oxidative stress. Furthermore, co-exposure to soot and iron
resulted in increased expression of cytochrome P450 isozymes CYP1A1 and CYP2E1 in lung tissue,
an effect not observed in response to either agent alone. Inflammation-related endpoints observed in
this study are described in Section 6.3.3.3.
In a parallel study, Pinkerton et al. (2008, 190471) exposed neonatal male SD rats to iron-soot
or filtered air 6 h/day for three days during the second and fourth week of life. Both 30 (ig/m3 and
100 (ig/m3 iron in combination with 250 (ig/m3 soot resulted in increased BALF oxidized
glutathione, glutathione redox ratio and glutathione-S-transferase activity and decreased total
antioxidant power. The higher concentration exposure resulted in increased ferritin expression in
lung tissue. Effects on cellular proliferation in specific regions of the lung were also noted as
described in Section 6.3.5.3.
Nurkiewicz et al. (2009, 191961) exposed SD rats to fine (count median diameter 710 nm) and
UF (count median diameter 100 nm) TiO2 particles via aerosol inhalation at concentrations of 1.5-16
December 2009 6-113
-------
mg/m3 for 240-720 min. These exposures were chosen in order to produce deposition of 4-90 ug/rat,
which was demonstrated in a previous study to result in different degrees of impaired microvascular
function (Nurkiewicz et al., 2008, 156816). Histological analysis of lung tissue did not find any
significant inflammation, although particle accumulation in alveolar macrophages and a frequent
association of alveolar macrophage with the alveolar wall was observed 24 h following exposure
(Nurkiewicz et al., 2008, 156816). Although the main focus of the more recent study was on effects
of TiO2 on NO production and microvascular reactivity in the spinotrapezius muscle
(Section 6.2.4.3), the presence of nitrotyrosine was determined in both lung tissue and spinotrapezius
muscle as a measure of peroxynitrite formation. Peroxynitrite formation occurs mainly as a result of
the rapid reaction of NO with superoxide and suggests an increase in local superoxide production.
The area of lung tissue containing nitrotyrosine immunoreactivity increased three-fold 24 h
following exposure to 10 ug UF TiO2. Nitrotyrosine immunoreactivity was localized in
inflammatory cells found in the alveolar region of the lung.
Summary of lexicological Study Findings for Pulmonary Oxidative Responses
New studies involving short-term exposure to CAPs, urban air, diesel and gasoline exhaust,
and model particles such as CB, iron-soot and TIO2 consistently demonstrate pulmonary oxidative
responses. Furthermore, antioxidant treatment ameliorated effects observed in response to CAPs,
DE and DE particles.
6.3.5. Pulmonary Injury
The 2004 PM AQCD (U.S. EPA, 2004, 056905) presented evidence from several toxicological
studies of small PM-induced increases in markers of pulmonary injury including thickening of
alveolar walls and increases in BALF protein. These findings are consistent with the results of recent
toxicological and controlled human exposure studies demonstrating mild pulmonary injury
accompanying inflammatory responses to CAPs and wood smoke. One recent epidemiologic study
has also observed a positive association between PM and urinary concentrations of lung Clara cell
protein.
6.3.5.1. Epidemiologic Studies
One epidemiologic study examined biomarkers of pulmonary injury. The mean concentration
data from this study are characterized in Table 6-10. Timonen et al. (2004, 087915) enrolled subjects
with coronary heart disease in Amsterdam (n = 37), Erfurt, Germany (n = 47) and Helsinki (n = 47)
to study daily variation in PM and urinary concentrations of lung Clara cell protein (CC16). No
associations were seen between the PNC of the smallest particles (NC0.oi-o.i) and CC16. Significant
associations with NC0.i-i and PM25 (which were strongly correlated with each other [r = 0.8]) were
seen only for Helsinki subjects: same day, lag 3 and 5-day mean NC0.i-i increases of 1000
particles/cm3 were associated with increases in In (CC16/creatinine) of 15.5% (95% CI: 0.001-30.9),
17.4% (95% CI: 3.4-31.4), and 43.2% (95% CI: 17.4-69.0), respectively. Similar associations were
seen for 10 (ig/m3 increases in PM25: lag 0 and 5-day mean PM25 were associated with increases in
In (CC16/creatinine) of 23.3% (95% CI: 6.3-40.3) and 38.8% (95% CI: 15.8-61.8), respectively.
6.3.5.2. Controlled Human Exposure Studies
No studies of controlled human exposures presented in the 2004 PM AQCD (U.S. EPA, 2004,
056905) specifically examined the effect of PM on pulmonary injury. However, several recent
studies have evaluated changes in markers of injury and increased alveolar permeability following
exposures to various types of particles.
December 2009 6-114
-------
Urban Traffic Particles
Brauner et al.(2009 190244) evaluated the effect of exposure to urban traffic particles (24-h
exposure, PM2.5 9.7 ug/m ) on the integrity of the alveolar epithelial membrane in a group of 29
healthy adults, with and without exercise. Following 2.5 h of exposure, alveolar epithelial
permeability was assessed by measuring the pulmonary clearance of 99mTc-DTPA, which was
administered as an aerosol during 3 min of tidal breathing. While pulmonary clearance of
99mTc-DTPA was observed to increase following exercise, there was no significant difference in
clearance between exposure to urban traffic particles and filtered air. In addition, PM exposure was
not observed to affect the level of CC16 in plasma or urine at 6 or 24 h after the start of exposure.
Diesel Exhaust
Relative to filtered air, exposure for 1 h to DE (300 ug/m3 PM) was not observed to affect the
plasma CC16 concentration at 6 or 24 h post exposure in a group of 15 former smokers with COPD
(Blomberg et al., 2005, 191991).
Wood Smoke
In a study examining the respiratory effects of wood smoke, Barregard et al. (2008, 155675)
exposed two groups of healthy adults in separate 4-h sessions to wood smoke with median particle
concentrations of 243 and 279 ug/m3. At 20 h post-exposure, the mean serum CC16 concentration
was significantly higher after exposure to wood smoke when compared with filtered air. However,
when the analysis was stratified by exposure session, a statistically significant effect of wood smoke
on serum CC16 was observed in the subjects in session 1 but not those in session 2. It is interesting
to note that while the mean particle concentration was only slightly higher in session 1, the mean
particle number in session 1 was almost 90% higher than the particle number in session 2, with
geometric mean particle diameters of 42 and 112 nm, respectively.
Summary of Controlled Human Exposure Study Findings for Pulmonary Injury
The findings from these studies provide limited evidence to suggest that exposures to particles
may increase markers of pulmonary injury in healthy adults.
6.3.5.3. lexicological Studies
The 2004 PM AQCD (U.S. EPA, 2004, 056905) reported mild increases in BALF protein, a
marker of pulmonary injury, in several studies involving inhalation exposure to CAPs. In addition,
histological analysis demonstrated that the bronchoalveolar junction was the site of the greatest
inflammation following CAPs exposure. Low level exposure to DE was associated with Type 2 cell
proliferation and thickening of alveolar walls near alveolar macrophages according to the 2002 EPA
Diesel Document (U.S. EPA, 2002, 042866). In addition, IT instillation of fly ash and
metal-containing PM generally caused pulmonary injury as measured by increases in BALF protein,
LDH and albumin. Proliferation of bronchiolar epithelium was also noted. More recent studies of
BALF markers of pulmonary injury and histological analysis of lung tissue are summarized below.
BALF Markers of Pulmonary Injury and Increased Permeability
CAPs
Kodavanti et al. (2005, 087946) exposed SH and WKY rats to filtered air or PM2.5 CAPs from
RTP, NC as described in Section 6.3.3.3. Differences in baseline parameters were noted for the two
rat strains since SH rats had greater levels of protein and lower levels of LDH, NAG, ascorbate and
December 2009 6-115
-------
uric acid in the BALF compared with WKY rats. One day after the 2-day CAPs exposure, increased
levels of GGT were observed in BALF (a marker of epithelial injury) of SH rats, but not WKY rats,
compared with filtered air controls. Injury was not accompanied by inflammation (Section 6.3.3.3).
In a study by Cassee et al. (2005, 087962). SH rats were exposed for 6 h by nose-only
inhalation to CAPs from three different sites in the Netherlands as described in Section 6.3.3.3. The
pulmonary injury marker CC16 was increased in BALF two days following CAPs exposure.
Inflammation was also observed (Section 6.3.3.3).
Gurgueira et al. (2002, 036535) exposed SD rats to Boston, MA CAPs as described in
Section 6.3.4.2 and reported a small but statistically significant increase in lung wet/dry ratios after 3
and 5 h of exposure, indicating the presence of mild edema. This response was accompanied by
increased oxidative stress as measured by in situ CL (Section 6.3.4.2). In a similar study, Rhoden
et al. (2004, 087969) reported an increase in lung wet/dry ratio in rats 24 h following a 5-h exposure
to Boston CAPs which was diminished by pre-treatment of the antioxidant NAC (Section 6.3.4.2).
Pulmonary injury was investigated in two studies using a rat model of pulmonary hypertension
(SD rats pre-treated with monocrotaline) which is described in greater detail in Section 6.3.3.3 (Lei
et al., 2004, 087999). Significant increases in BALF LDH and protein were observed in response to
CAPs. Pulmonary inflammation was observed in both of these studies (Section 6.3.3.3).
Diesel Exhaust
In a study evaluating the effects of DE, no changes were observed in BALF protein and LDH
in mice exposed by inhalation to concentrations of 50 and 2000 (ig/m3 DE particles for 4 h/day on 5
consecutive days as described in Section 6.3.3.3 (Stevens et al., 2008, 157010). Changes in gene
expression were observed in the higher exposure group. This study demonstrates that changes in
gene expression can occur in the absence of measurable markers of injury or pulmonary
inflammation.
In a study by Wong et al. (2003, 097707). also reported by Witten et al. (2005, 087485). rats
were exposed nose-only to filtered room air or to DE over a 3-wk period. This study, focusing on
neurogenic inflammation, is described in greater detail in Section 6.3.3.3. Pulmonary plasma
extravasation was measured by the 99mTechnecium-albumin technique and found to be
dose-dependently increased in the bronchi and lung. Pretreatment with capsaicin, which inhibits
neurogenic inflammation by activating C-fibers and causing the depletion of neuropeptide stores, did
not reduce plasma extravasation following DE exposure. Hence, DE is unlikely to act through
bronchopulmonary C-fibers to cause neurogenic edema in this model. Inflammatory responses
measured in this study are discussed in Section 6.3.3.3.
Gasoline Exhaust
Healthy male Swiss mice were exposed to gasoline exhaust (635 ug/m3 PM and associated
gases) or filtered air for 15 min/day for 7, 14, and 21 days as described in Section 6.3.3.3
(Sureshkumar et al., 2005, 088306). BALF was collected for analysis 1-h after the last exposure.
Statistically significant increases in BALF markers of lung injury, alkaline phosphatase,
gamma-glutamyl transferase and LDH, were observed at all time points studied. Alveolar edema was
noted following 14 and 21 days of exposure. Other findings of this study, including inflammation
and histopathological changes, are discussed in Section 6.3.3.3 and below.
Histopathology
CAPs
Histopathological changes were demonstrated in rats exposed for 5 h to Boston CAPs as
described in Section 6.3.3.3 (Rhoden et al., 2004, 087969). Slight bronchiolar inflammation and
thickened vessels at the bronchiole were observed 24 h post-exposure, consistent with the influx of
polymorphonuclear leukocytes observed in BALF (Section 6.3.3.3).
December 2009 6-116
-------
Diesel Exhaust
In a study by Wong et al. (2003, 097707). also reported by Witten et al. (2005, 087485). rats
were exposed nose-only to filtered room air or to DE over a 3-wk period. This study, focusing on
neurogenic inflammation, is described in greater detail in Section 6.3.3.3. Pulmonary inflammation
was evaluated by histological analysis of lung tissue. Following high, but not low,
concentration-exposure to DE, a large number of alveolar macrophages was found in the lungs.
Small black particles, presumably DE particles, were found in the cytoplasm of these alveolar
macrophages. Perivascular cuffing consisting of mononuclear cells was also observed in the high
exposure animals. Influx of neutrophils or eosinophils was not seen although mast cell number was
increased. Other indices of injury demonstrated in this study are described above.
Gasoline Exhaust
Another study, which is described in greater detail in Section 6.3.3.3, demonstrated
histopathological responses to gasoline exhaust in mice exposed to gasoline exhaust or filtered air
for 15 min/day for 7, 14, and 21 days (Sureshkumar et al., 2005, 088306). Histological observations
showed inflammatory cell infiltrate in the peribronchiolar and alveolar region, alveolar edema and
thickened alveolar septa at 14 and 21 days post-exposure. Levels of pro-inflammatory cytokines and
marker enzymes of lung damage were also increased in BALF. The numbers of inflammatory cells in
BALF was increased but not significantly, demonstrating that BALF analysis of inflammatory cells
was a less sensitive indicator of pulmonary inflammation in this study than histological analysis.
Other indices of injury found in this study are described above.
Model Particles
In a study investigating the effects of iron-soot, mice were exposed to 250 (ig/m3
laboratory-generated iron-soot as described in Sections 6.3.2.3 and 6.3.3.3 (Last et al., 2004,
097334). Analysis of airway collagen content was conducted by histology and by biochemical
analysis of microdissected airways. No increases in airway collagen content were found by either
method in mice exposed to iron-soot for two weeks. Furthermore, no goblet cells were observed in
airways of air or iron-soot exposed animals. Other findings of this study are described in Sections
6.3.2.3 and 6.3.3.3.
One study demonstrating histopathological responses to PM in neonatal rats was reported by
Pinkerton et al. (2004, 087465). Rat pups (10 days old) were exposed to soot and iron particles
(mean mass concentration of 243 (ig/m ; iron concentration 96 (ig/m3; size range 10-50 nm) for
6 h/day on 3 consecutive days. Cell proliferation in different lung regions was evaluated following
bromodeoxyuridine injection 2 h prior to necropsy. The rate of cell proliferation in the proximal
alveolar region (immediately beyond the terminal bronchioles) was significantly reduced in iron-soot
exposed animals compared to controls. This was a region-specific response since the rate of cell
proliferation was not altered in the terminal bronchioles or the general lung parenchyma. However
alveolar septation, the process by which alveoli are formed during development, and alveolar growth
were not altered by iron-soot exposure. Decreased cell viability and increased LDH was also noted
in BALF of neonatal rats (Pinkerton et al., 2008, 190471). The authors suggest the possibility of
greater susceptibility to air pollution during the critical postnatal lung development period which
occurs in animals and humans and that neonatal exposure to PM may contribute to impaired lung
growth seen in children.
Summary of lexicological Study Findings for Pulmonary Injury
New studies involving short-term exposure to CAPs and diesel and gasoline exhaust
demonstrate mild pulmonary injury, including enhanced BALF markers of injury, pulmonary edema
and histopathology. In general, injury responses were accompanied by inflammatory responses. In
addition, altered cellular proliferation in the proximal alveolar region was observed in neonatal rats
exposed to iron-soot, suggesting the possibility of greater susceptibility to PM during postnatal lung
development.
December 2009 6-117
-------
Relative Toxicity of PM Size Fractions
Ambient PM Studies
A recently undertaken multinational project entitled "Chemical and biological characterization
of ambient thoracic coarse (PMi0_2.5), fine (PM2.5_o.2), and UFPs (PM0.2) for human health risk
assessment in Europe" (PAMCHAR) takes a systematic approach to expanding the present
knowledge about the physiochemical and toxicological effects of these three PM size fractions. Six
European cities were selected that represented contrasting ambient PM profiles: Helsinki, Duisburg,
Prague, Amsterdam, Barcelona, and Athens. For PM collected at all sites, PMi0_2.5 induced the
greatest pulmonary effects in C57BL/6J mice IT instilled with 1, 3, or 10 mg/kg of particles (Happo
et al., 2007, 096630). Dose-response relationships in BALF parameters measured 24 h post-IT
instillation exposure, including cell number and protein, were observed for all sites following
PMio_2.5, and neutrophils were the predominant cell type present in the BALF (Happo et al., 2007,
096630). Prague PMi0_2.s exposure resulted in decreased macrophages in BALF at 12 h, and
Amsterdam, Barcelona, and Athens PMi0_2.5 induced lymphoplasmacytic cells in BALF (Happo et
al., 2007, 096630). No inflammatory responses were observed for UFPs measured 12-h after
exposure. Protein was elevated for PMi0_2.5 for all locations with the 10 mg/kg dose; Athens UFPs
induced protein release only at the two lowest doses 12 h post-exposure. For TNF-a and IL-6, the
greatest response was observed with PMi0_2.5 4 h following exposure (Happo et al., 2007, 096630).
Exposure to UFPs from Duisburg resulted in elevated TNF-a for the 1 and 3 mg/kg doses. Only the
Helsinki sample appeared to induce the same level of IL-6 release for PMi0_2.5 and PM0.2 at 10
mg/kg, albeit the collection times differed. In vitro TNF-a and IL-6 responses did not always reflect
in vivo effects (Table 6-11), as the Duisburg PMi0_2.5 sample was the most potent in vivo compared
to the other sites and elicited much lower cytokine release compared to other cities (except Helsinki)
in vitro (Happo et al., 2007, 096630; Jalava et al., 2006, 155872; Jalava et al., 2008, 098968).
Helsinki PM was collected in the spring and generally had the lowest in vivo and in vitro activity for
PMio_2.5 compared to the other cities (Happo et al., 2007, 096630; Jalava et al., 2006, 155872;
Jalava et al., 2008, 098968). Spring-time samples were collected because episodes of resuspended
road dust occur frequently during this season (Pennanen et al., 2007, 155357). There was a high
correlation between EC content in PM2 5 and PM10_2.5, indicating that traffic impacted both size
fractions (Sillanpaa et al., 2005, 156980). Duisburg PM collected in fall had the greatest amounts of
Mn and Zn compared to PM samples from other locations (Pennanen et al., 2007, 155357). Metals
industries in Duisburg are likely contributors to the observed PM metals concentrations. For the
Prague winter PM samples, the As content was higher than at any other location (Pennanen et al.,
2007, 155357). Prague also had the highest PAH levels in all three size fractions, possibly
attributable to stable atmosphere conditions and incomplete combustion of coal and biomass in
residential heating (Pennanen et al., 2007, 155357). High levels of ammonium and nitrate in PM
samples from Amsterdam suggest traffic as a large source of air pollution (Pennanen et al., 2007,
155357). Approximately one-third of PMi0_2.5 mass from Amsterdam was comprised of sea salt
(Sillanpaa et al., 2005, 156980). double that of any other city. In Barcelona and Athens, high
calcium or Ca2+ contents in spring and summer PM25 and PMi0_2.5 are indicative of resuspended
soil-derived particles (Pennanen et al., 2007, 155357).
December 2009 6-118
-------
Table 6-11. PAMCHAR PMi0.2.6 inflammation results with ambient PM.
City and Season
Helsinki spring
Duisburg fall
Prague winter
Amsterdam winter
Barcelona spring
Athens summer
BALF protein
+10
+10
+10
+10
+10
+10
BALF TNF-a
+10
+10
[+310]
+10
+10
[+310]
In Vivo3 (mg/kg)
BALF IL-6
+10
+10
+10
+10
[+310]
[+310]
BALF KC
[+310]
+10
[+310]
+10
+10
[+310]
BALF PMN
+10
+10
+10
+10
+10
+10
In
BALF AM TNF-a
+150,300
+150,300
+10 +150,300
+150
+150,300
+150,300
Vitro" (ug/mL)
IL-6
+150,300
+150,300
+150,300
+150,300
+150,300
+150,300
MIP-2
+150,300
+300
+150,300
+150,300
+150,300
+150,300
"Source: Happo et al. (2007, 096630): 2 cell lines used for in vitro study were RAW264.7
bSource: Jalava et al. (2006,155872): + indicates increased response and numbers that follow indicate at which dose the response was observed
Schins et al. (2004, 054173) employed PM from two cities in Germany, Duisburg and Borken,
in another study. In contrast to the PAMCHAR study where animals were administered PM
suspended in pathogen-free water (Happo et al., 2007, 096630). animals received PM via IT
instillation suspended in saline at a dose of 320 ug (Schins et al., 2004, 054173). In female Wistar
rats, neutrophils in BALF were significantly elevated for PMi0_2.5 from Duisburg and Borken (Table
6-12), albeit the percent of neutrophils with the PMi0_2.5 from Borken was nearly double that of
Duisburg. The responses with PM2.5 were much smaller. When these PM10_2.5 particles were
introduced into whole blood to determine overall inflammogenic capacity, IL-8 and TNF-a were
released in greater quantities than in response to PM2.5. Furthermore, PMi0_2.5 from Borken induced
higher cytokine responses than Duisburg PMi0_2.5.
An in vivo study involving SH rats was conducted using PMio_2.5 and PM2 5 from six different
European locations with varying traffic densities (3 or 10 mg/kg IT instillation; UFPs were not
collected) (Gerlofs-Nijland et al., 2007, 097840). It was reported that PM10_2.5 generally induced
greater responses than PM25. IT instillation of PMi0_2.s from a location with high traffic influence in
Munich, Germany, demonstrated the greatest response in terms of LDH activity, protein, total cells,
neutrophils, and lymphocytes in BALF 24 h post-exposure. PMi0_2.5 collected from a low traffic site
in Munich induced the greatest cytokine response for TNF-a and MIP-2. Some correlations were
observed between PMi0_2 5 components (Ba and Cu) and BALF parameters, but were largely driven
by one location (Gerlofs-Nijland et al., 2007, 097840).
Table 6-12. Other ambient PM - in vivo PMi0-2.6 studies - BALF results, 18-24 h post-IT exposure.
Location
Germany, Borken; rural Feb-May 2000
Germany, Duisburg; heavy industry
Feb-May 2000
USA, Seattle, WA
Feb-March 2004
USA, Salt Lake City, UT
Apr-May 2004
USA, South Bronx, NY
Dec 2003-Jan 2004
USA, Sterling Forest, NY
Dec 2003-Jan 2004
Endotoxin Dose Cell n/tnUno
(-Values) (mg/kg) Differentials °yiOKme
6.6 EU/mg 0.58-0.91 t*%PMN t TNF-a
5.0 EU/mg 0.58-0.91 t % PMN t MIP-2
6.0 EU/mg 1.25,5.0
6.3 EU/mg 1.25,5.0
2.8 EU/mg 1.25,5.0 fPMN t MIP-2
2.9 EU/mg 1.25,5.0
Injury
s Biomarkers e erence
Schins et al. (2004, 054173)
Schins et al. (2004, 0541731
Gilmour, et al. (2007, 0964331
t protein Gilmour, et al. (2007, 0964331
Gilmour, et al. (2007, 0964331
Gilmour, et al. (2007, 0964331
December 2009
6-119
-------
Location
Endotoxin Dose Cell
(~ Values) (mg/kg) Differentials
<***"»• Bio'Sers
Reference
USA, RTP, NC
Oct-Nov1996
Germany, Munich Ost Bahnhof; high
traffic A
Aug 2002
Netherlands, Hendrik-ldo-Ambacht;
high traffic
Sept 2002
Italy, Rome; high traffic
Apr 2002
Netherlands, Dordrecht; moderate
traffic
Apr 2002
Germany, Munich Grosshadern
Hospital; low traffic
Jun-Jul 2002
Sweden, Lycksele; low traffic
Feb-March 2002
0.96EU/mg 0.5,2.5,5.0 tt PMN
tf total cells
ttAM
2.9EU/mg 3,10
tt*PMN
tt* Lymph
tt total cells
tt*AM
3,10
6.5 EU/mg tt PMN
tt Lymph
t total cells
ttAM
1.5 EU/mg 3,10
ttPMN
tt Lymph
tt total cells
tAM
0.6 EU/mg 3,10
ttPMN
t Lymph
t total cells
ttAM
2.9 EU/mg 3,10
ttPMN
tt Lymph
tt total cells
tAM
0.9 EU/mg 3,10
ttPMN
t Lymph
tlL-6
tt MIP-2
tt TNF-a
t MIP-2
tt TNF-a
tt MIP-2
tt TNF-a
tt* MIP-2
tt* TNF-a
tt*LDH
t* protein
ttLDH
t protein
ttLDH
ttLDH
t protein
tt*LDH
t protein
ttLDH
t protein
Dick (2003, 088776)
Gerlofs-Nijland, et al.
Gerlofs-Nijland, et al.
Gerlofs-Nijland, et al.
Gerlofs-Nijland, et al.
Gerlofs-Nijland, et al.
Gerlofs-Nijland, et al.
(2007, 0978401
(2007, 0978401
(2007, 0978401
(2007, 0978401
(2007, 0978401
(2007, 0978401
For Gerlofs-Nijland study, composition data were averaged across seasons. | significant only at highest dose.
tt Significant at lowest and highest dose.
* Greatest potency for that endpoint and study. Gilmour et al. (2007, 096433)exposure was via aspiration.
A more recent study by these investigators (Gerlofs-Nijland et al., 2009, 190353) compared
responses to PM from three different European cities based on size fraction and content of metals
and PAH. SH rats were IT instilled with 7 mg/kg PM, and markers of toxicity and inflammation were
measured in BALF 24 h later. Blood markers of coagulation were also measured and are described in
Section 6.2.8.3. In the first part of the study, both PM2.5 and PMi0_2.5 from Duisburg were found to
have dramatic effects on inflammatory cell influx and activation as well as on the injury markers
LDH, protein and albumin in the BALF. The antioxidant species uric acid was increased in BALF
from rats exposed to both size fractions and was interpreted as an adaptive response to oxidative
stress. Statistical analysis demonstrated that PMi0_2.5 was more potent in eliciting these responses
than PM2.s. In the second part of the study, responses to metal-rich PM from Duisburg and metal-
poor PM from Prague were determined. A statistically significant greater enhancement of BALF
markers of inflammation and injury was observed for the Duisburg PM compared with the Prague
PM. Furthermore, responses to PAH-rich PMi0_2.5 from Prague and PAH-poor PMi0_2.5 from
Barcelona were determined. PMi0_2.5 from Prague was found to have statistically significant greater
effects compared with PMi0_2.5 from Barcelona. However, organic extracts of these PMi0_2.5 fractions
had very little capacity to produce inflammation or toxicity in this model. These findings suggest an
important role for specific components associated with PMi0_2.5 in mediating the pro-inflammatory
effects.
December 2009
6-120
-------
In another study investigating specific components of PM10_2.5, BALB/c mice were IT instilled
with 25 and 50 ug PMio-2.5 from a rural area of the San Joaquin Valley, California (Wegesser and
Last, 2008, 190506). Inflammatory cell influx into BALF began at 6 h and peaked at 24 h following
IT instillation with 50 ug PM, with the increase in neutrophils preceding the increase in
macrophages. Pro-inflammatory effects were found to be mainly due to insoluble components of
PM. Furthermore, heat-treatment, which was capable of inactivating endotoxin, had no effect on
inflammation. Numbers of neutrophils in the BALF were found to correlate with the content of MIP-
2, a known neutrophil chemoattractant released from macrophages and epithelial cells. Taken
together, these results demonstrate that the pro-inflammatory effect of this PMi0_2.5 was associated
with insoluble components and not with endotoxin.
In an in vivo study that employed ambient PM collected in fall 1996 from RTP, NC,
neutrophilic influx was observed in BALF of female GDI mice 18 h post-IT instillation (10, 50 or
100 ug) of coarse PM (3.5-20 um), although a dose-response relationship was not evident (Dick et
al., 2003, 088776). Mice were also exposed to fine (1.7-3.5 um) and fine/ultrafme (<1.7 um) PM
fractions. Only the two highest doses of PM for the smaller size fractions induced elevated
neutrophils. Significant responses in albumin and TNF-a were only observed for the fine PM
(1.7-3.5 um) exposure group. Total protein, LDH and NAG responses were unchanged from control
levels for all PM size fractions. Levels of IL-6 were elevated in mice exposed to 100 ug for coarse,
fine, and fine/ultrafme (<1.7 um) PM. When dimethylthiourea (DMTU) was administered
intravenously prior to exposure, the neutrophil response was attenuated in all groups to levels below
control.
Another study compared PMi0_25, PM25, and UFPs collected in Seattle, WA, Salt Lake City,
UT, South Bronx, NY, and Sterling Forest, NY (Gilmour et al., 2007, 096433). In female BALB/c
mice, the 100 ug dose of PMi0_2.5 (approximately 5 mg/kg) from Salt Lake City induced a significant
increase in protein in BALF, and the level released was almost as high as that observed after LPS
exposure. PMi0_2.5 from the South Bronx resulted in dose-related increases in neutrophil number and
MIP-2 levels in BALF. In contrast, no effects were observed with PMi0_2.5 from Sterling Forest. The
greatest amount of LPS was observed in the Salt Lake City and Seattle PMi0_2.s samples. There was a
less discernable pattern of response with fine and UFPs.
Coal Fly Ash
Coal fly ash of differing size fractions and composition was administered to female GDI mice
via oropharyngeal aspiration (25 or 100 ug) to assess lung inflammation and injury 18 h following
exposure (Gilmour et al., 2004, 057420). Montana (low-sulfur subbituminous; 0.83% sulfur, 11.72%
ash content) or western Kentucky (high-sulfur bituminous; 3.11% sulfur, 8.07% ash content) coal
was combusted using a laboratory-scale down-fired furnace. Interestingly, no significant effects on
BALF neutrophils, TNF-a, MIP-2, albumin, total protein, LDH activity, or NAG activity were
observed 18 h post-exposure to PMi0_2.5 from either coal fly ash. However, the UF fraction (PM0.2)
of combusted Montana coal induced greater numbers of neutrophils than PM10_2.5 or PM25 at both
doses. TNF-a was only elevated in animals exposed to 100 ug of the Montana UFPs; MIP-2 was also
increased at both doses. The PM2 5 western Kentucky coal fly ash caused increased BALF
neutrophils, MIP-2, albumin, and protein (Gilmour et al., 2004, 057420).
In a similar study employing Montana subbituminous coal fly ash particles >2.5 um,
C57BL/6J mice were IT instilled with PM alone or PM+LPS and BALF was obtained 18 h
post-exposure (Finnerty et al., 2007, 156434). TNF-a and IL-6 in lung homogenates were only
elevated in the animals exposed to PM+100 ug LPS, although it appeared that there was a
greater-than additive effect. Total cells and cell differentials were not measured.
Summary of lexicological Study Findings for Relative Toxicity of PM Size Fractions
Biomarkers of injury and inflammation were measured in in vivo and in vitro studies
comparing the toxicity of different size fractions of ambient PM from various locations. Responses
were measured in BALF from rodents following IT instillation or aspiration of PM. In general, the
PMio_2.5 size fraction was more potent than PM2 5 or UFPs and endotoxin levels did not appear
responsible. In one study, rural PMi0_2.5 from Germany induced a greater inflammatory and cytokine
response than PM10_2.5 from an industrial location. In contrast, PM10_2.5 from Sterling Forest, NY did
not lead to any change in BALF inflammation or injury markers. A study that employed coal fly ash
December 2009 6-121
-------
indicated that the UF fraction was the most inflammogenic. All of these studies were conducted
using high doses of PM (0.58-10 mg/kg) and it is unclear if similar effects would be observed at
lower doses.
6.3.6. Allergic Responses
A large number of toxicological and controlled human exposure studies cited in the 2004 PM
AQCD (U.S. EPA, 2004, 056905) reported an exacerbation of existing allergic airway disease
following exposure to laboratory-generated and ambient particles. In addition, numerous studies
have demonstrated that PM can alter the immune response to challenge with specific antigens and
suggest that PM may act as an adjuvant to promote allergic sensitization. Recent toxicological
studies have provided evidence of enhanced allergic responses and allergic sensitization following
exposure to CAPs and DE that is consistent with the findings presented in the 2004 PM AQCD. PM
can enhance allergic responses by facilitating delivery of allergenic material and promoting
subsequent immune reactivity. The initiation or exacerbation of allergic responses has important
implications for allergic asthma, the most common form of asthma. Additionally, PM has been
shown to alter ventilatory measures in non-allergic animal models, suggesting a possible role in
other forms of asthma.
6.3.6.1. Epidemiologic Studies
Allergy contributes to a number of respiratory morbidity outcomes, including asthma.
However, relatively few epidemiologic studies of PM have specifically examined indicators of
allergy. The 2004 PM AQCD (U.S. EPA, 2004, 056905) presented one study (Hajat et al., 2001,
016693) showing an association between doctor visits for allergic rhinitis and PM10 among children
in London. This association was strongest at a lag of 3 or 4 days. Similar results were obtained in a
new study by Tecer et al. (2008, 180030), which found significant associations between PM2.5,
PMio_2.5, and PMi0 with hospital admissions for allergic rhinitis in Turkish children, particularly at
lag day 4. While exacerbation of allergic symptoms may occur relatively rapidly, repeated or longer
exposures may be required for allergic sensitization to develop; a number of studies associating long-
term exposure to PM with specific indicators of allergic sensitization are described in Chapter 7.
6.3.6.2. Controlled Human Exposure Studies
Exacerbation of Allergic Responses
Diesel Exhaust and Diesel Exhaust Particles
Exposure to DE particles was shown to increase the allergic response among atopic
individuals in several controlled human exposure studies cited in the 2004 PM AQCD (U.S. EPA,
2004, 056905). Nordenhall et al. (2001, 025185) found that exposure to DE significantly decreased
the concentration of Mch required to induce a 20% decrease in FEVi in a group of atopic asthmatics
24 h post-exposure. In addition, Diaz-Sanchez et al. (1997, 051247) demonstrated an increase in
allergen-specific IgE following exposure via intranasal spray to ragweed plus DE particles (0.3 mg)
relative to ragweed allergen alone. Decreases in IFN-y and IL-2, as well as increases in IL-4, IL-5,
IL-6, IL-10, and IL-13 were also observed when ragweed allergen was administered with DE
particles. It should be noted that the DE particles used in this study were collected during a cold start
of a light-duty Isuzu diesel engine, and thus contained relatively high levels of incomplete
combustion materials and semi-volatiles organics (e.g., PAHs). One new study using the same source
of DE particles (Bastain et al., 2003, 098690) also observed an increase in IL-4 and allergen specific
IgE, as well as a decrease in IFN-y following intranasal administration of ragweed allergen with DE
particles (0.3 mg) in atopic adults. The protocol was repeated in this study for all subjects, and the
enhancement of allergic response by coexposure to DE particles was observed to be highly
reproducible within individuals. In addition, Gilliland et al. (2004, 156471) demonstrated that GST
December 2009 6-122
-------
polymorphisms may alter the adjuvant effects of DE particles on allergic response, with individuals
with GSTM1 null or GSTP1 1105 wild type genotypes showing the largest effects.
Allergic Sensitization
Diesel Exhaust Particles
One controlled human exposure study has demonstrated that de novo sensitization to a
neoantigen can be induced by exposure to DE particles. In this study, Diaz-Sanchez et al. (1999,
011346) dosed 25 atopic adults intranasally with 1 mg keyhole limpet hemocyanin (KLH), followed
by two biweekly challenges with 100 ug KLH. In 15 of the 25 subjects, cold-start DE particles (0.3
mg) were administered intranasally 24 h prior to each KLH exposure, while in the other ten subjects,
no DE particles were administered. No KLH-specific IgE was observed in the nasal lavage fluid of
any of the subjects exposed to KLH without exposure to DE particles. However, KLH-specific IgE
was present in the nasal lavage fluid of 9 out of 15 subjects 28-32 days after the initial KLH
immunization when exposures were preceded by administration of DE particles.
CAPs
Increased levels of eotaxin, a marker of allergic activation, were observed in healthy adult
volunteers after inhalation of nebulized ambient Chapel Hill PMi0_2.5 (Alexis et al., 2006, 154323).
This particular effect was found to be due to endotoxin, based on its elimination by heat-inactivation;
study details are provided in Section 6.3.3.2.
6.3.6.3. lexicological Studies
Exacerbation of Allergic Responses
Increased use of actual ambient air particle mixes in toxicological studies since the 2004 CD
has greatly expanded evidence relevant to assessing these and other immunotoxic effects. A number
of studies have also included ambient-level concentrations, although many still include relatively
high doses of questionable relevance compared to the doses inhaled by humans. Recent dosimetric
models reveal that a small fraction of epithelial cells located at the carinal ridges of airway
bifurcations can receive massive doses that may be even a few hundred times higher than the
average dose for the whole airway (Chapter 4). These areas, coincidentally, are locations of bronchus
associated lymphoid tissues (BALT) which are sites at which interaction of T and B lymphocytes
with antigen presenting cells (APC) occurs. Hence the deposited particles are in near-ideal proximity
to immunologically active tissues. Doses used for assessing PM immunotoxicity should be viewed
with this perspective. In many animal studies, changes in ventilatory patterns are assessed using
whole body plethysmography, for which measurements are reported as enhanced pause (Penh).
Some investigators report increased Penh as an indicator of AHR, but these are inconsistently
correlated and many investigators consider Penh solely an indicator of altered ventilatory timing in
the absence of other measurements to confirm AHR. Therefore use of the terms AHR or airway
responsiveness has been limited to instances in which the terminology has been similarly applied by
the study investigators.
CAPs
Existing allergic sensitization confers susceptibility to the effects of PM in rodent models. For
example, studies in allergic rats (Harkema et al., 2004, 056842; Morishita et al., 2004, 087979)
suggest that allergic sensitization enhances the retention of PM in the airways. Recovery of
anthropogenic trace elements (La, V, Mn, S) from lung tissue was greater for Detroit PM2.5 CAPs
exposed OVA sensitized/challenged BN rats than for air exposed or non-allergic CAPs exposed
controls (24 h post-exposure for 4 or 5 consecutive 10-h days during July or September; time
weighted avg mass concentration of 676 ± 288 or 313 ± 119 (ig/m3, respectively) (Harkema et al.,
December 2009 6-123
-------
2004, 056842). Interestingly, despite lower avg mass concentration, increases in these elements were
observed in September, when the avg number concentration of UFPs was nearly double that of July
(10,879 ± 5,126 vs. 5,753 ± 2,566 particles/cm3). September CAPs was associated with eosinophil
influx and BALF protein content, as well as significantly increased airway mucosubstances, and the
authors speculated that the high concentration of UFPs facilitated particle penetration into the
alveolar region of the lungs. IT instillation of fractionated insoluble PM2.5 collected from this period
resulted in a mild pulmonary neutrophilic inflammation in healthy BN rats, but no differential effects
were obtained after IT instillation of total, soluble, or insoluble PM2.5 in allergic rats.
Research has also been conducted to determine the effect of proximity to the roadway on
exacerbation of existing allergic disease. OVA-allergic BALB/c mice were exposed to PM2.5 or UF
(<0.15 um) CAPs, (avg total concentration 400 ug/m3) for five 4-h days a week over 2 wk at 50 or
150 m downwind of a heavily trafficked road (Kleinman et al., 2005, 087880). Markers of allergy
(serum OVA-specific IgE and IgGl, lung IL-5 and eosinophils) were significantly higher in mice
exposed to CAPs (PM2.5 or UF) than in air-exposed mice after OVA challenge. IL-5, IgGl, and
eosinophils were higher in mice closer to the roadway (50 m) than in mice 150 m downwind. The
authors suggest that the enhanced responses closer to the roadway may reflect a greater proportion of
UFPs in this vicinity, given that the concentrations of sub-25-nm particles decrease rapidly with
distance from the roadway and the PM2 5 CAPs closer to the roadway contained a greater number of
particles for a similar mass, a portion of which were UF. Animal-to-animal variability among the
biomarkers tested made it necessary to combine values from two exposures spanning two years for
statistical power (determined prior to the start of the experiment). A subsequent publication
(Kleinman et al., 2007, 097082) included a third exposure regimen as well as compositional
analysis. PM2 5 CAPs mass concentration was intentionally adjusted to an avg concentration of
approximately 400 ug/m3, ranging from 163 to 500 ug/m3, with an estimated particle number of
2.1*105 particles/cm at 50 m and 1.6xl05 particles/cm3 at 150 m. UFPs ranged from 146 to
430 ug/m3, with particle counts of 4.9 ± 1.4xl05particles/cm3 at 50 m, and 4.4 ± 2.1xl05
particles/cm3 at 150 m. Analysis of results from the three exposures indicated that OVA-sensitized
mice exposed 50 m downwind of the roadway exhibited increased levels of IL-5 and IgGi compared
to mice exposed 150 m downwind or exposed to air. No markers of allergy-related responses were
observed in the 150 m exposure groups, and very little difference was seen between PM25 and UF
CAPs responses, perhaps because PM2 5 contained 20-32% UF components. The strongest
associations between component concentrations and biological markers of allergy (IL-5 and IgGl)
were with EC and OC. These studies demonstrate that CAPs can enhance allergic responses, and that
proximity to a source may be an important factor.
In a BN rat model for allergic asthma (Heidenfelder et al., 2009, 190026). thirteen 8-h days of
exposure to Grand Rapids, MI PM2 5 CAPs alone did not result in differential gene expression or
indicators of asthmatic pathology in the lung, but the combination of CAPs and OVA resulted in
differential expression of genes predominantly related to inflammation and airway remodeling, along
with significant increases in IgE, mucin, and total protein in BALF. Consistent with these changes in
gene expression and BALF markers, OVA with CAPs also induced a more severe allergic
bronchopneumonia (distribution and severity of bronchiolitis and alveolitis) and increased mucus
cell metaplasia/hyperplasia and mucosubstances, indicating exacerbation of allergic or asthmatic
disease. CAPs was collected in July and characterized as having an average mass of 493±391, OC
244±144, EC 10±4, SO42~ 79±131 (13 day avg was only about 10% of the CAPs, but a spike
occurred during the first week), nitrate 39±67 ammonium 39±59, and urban dust (estimated from
Fe, Al, Ca, and Si) 18±6 (mean ± SD in ug/m3).
Diesel Exhaust Particles
Resuspended DE particles influences airway responses in mice with existing allergic
sensitization. A single 5-h nose-only exposure to 870 ug/m3 aerosolized filter-collected DE particles
(PM2 5) increased Mch-induced increases in ventilatory timing (Penh) in OVA sensitized/challenged
C57BL/6J mice (Farraj et al., 2006, 088469). Intranasal pretreatment with an antibody against the
pan neurotrophin receptor p75 attenuated the DE particle-induced increase in airflow obstruction,
indicating a role for neurotrophins. Neurotrophins are expressed by various structural, nerve and
immune cells within the respiratory tract and are linked to the etiology of asthma in both humans and
animal models. DE particles alone in unsensitized mice caused a significant increase in lung
macrophages; this response was also inhibited by anti-p75, which may suggest mediation of
macrophage influx by neurotrophin or alternatively may reflect anti-p75 dependent depletion of
December 2009 6-124
-------
macrophages due to expression of the p75 receptor. Aside from increased macrophages, the single
exposure to DE particles had little effect on other markers of airway inflammation. In a similar
subsequent study, these authors demonstrate neurotrophin-mediated DE particle-induced airflow
obstruction in OVA sensitized and challenged BALB/c mice (Farraj et al., 2006, 141730). in this
case using a higher 2000 ug/m3 single 5-h exposure to aerosolized filter-collected PM2.5. Differences
between whole body plethysmography and tracheal ventilation measurements indicated that airflow
obstruction may have originated in the nasal passages. Again, very few indices of inflammation were
increased; however, similar neurotrophin-dependent increases in lung macrophages were observed
after DE particle exposure alone, and BALF IL-4 protein levels were increased 5-fold in sensitized,
challenged, DE particle-exposed mice. This neurotrophin-dependent IL-4 response was not evident
in the first study, and may be related to the higher dose used in the second study or the characteristic
allergic/Th2 bias of the BALB/c strain. Airflow obstruction in the absence of airway inflammation in
OVA-sensitized animals seen in both studies by Farraj et al. (2006, 088469; 2006, 141730) may
reflect DE particle-induced acute enhancement of neurogenic as opposed to immunologic
inflammation.
Diesel Exhaust
Exposure to relatively low doses of DE has been shown to exacerbate asthmatic responses in
OVA sensitized/challenged BALB/c mice (Matsumoto et al., 2006, 098017). Mice were intranasally
challenged one day prior to chamber exposure to DE (100 ug/m3 PM; CO, 3.5 ppm; NO2, 2.2 ppm;
SO2 <0.01 ppm) for 1 day or 1, 4, or 8 wk (7h/day, 5 days/wk, endpoints 12-h post-DE exposure).
Results from the 8 wk study are described in Section 7.3.6.2. It should be noted that control mice
were left in a clean room as opposed to undergoing chamber exposure to filtered air. Significant
AHR upon Mch challenge was observed after 1 and 4 wk of exposure, and airway sensitivity
(provocative concentration of Mch causing a 200% increase in Penh) was significantly increased
after 1 wk of exposure but not 4 wk. DE had no effect on total cells in BALF, but transiently
increased expression of IL-4, IL-5, and IL-13 after 1 day of exposure, MDC after 1 wk, and
RANTES after 2 and 3 wk. Eotaxin, TARC, and MCP-1 were elevated without statistical
significance after short-term (1 day or wk) exposure. Statistical power may have been lacking due to
few animals in the exposure group (n=3). Protein levels of IL-4 and RANTES were significantly
elevated after one day of DE exposure. DE had no effect on OVA challenge-induced peribronchial
inflammatory or mucin positive cells. Therefore DE-induced AHR was observed in the absence of
neutrophilic inflammation, similar to the responses described for aerosolized or nebulized DE
particles by Farraj et al. (2006, 088469; 2006, 141730) and Hao et al. (2003, 096565).
Gasoline Exhaust
Acute exposure to fresh gasoline engine exhaust PM does not appear to exacerbate allergic
responses (Day et al., 2008, 190204). BALB/c mice were exposed to whole exhaust diluted 1:10
(H), 1:15 (M), or 1:90 (L), filtered exhaust at the 1:10 (HF), or clean air for 6 h/day over three days.
Analytes for the high (H) and high filtered (HF) concentrations were: PM mass (ug/m3) 59.1±28.3
(H) and 2.3±2.6 (HF), PM number (particles/cm3) 5.0xl05 and l.lxlO4; CO (mg/m3) 102.8±33.0 and
99.5±1.6; NO (nig/m3) 18.4±2.8 and 17.2±1.9; NO2 (mg/m3) 1.4±0.3 and 1.7±0.2; SO2 (ug/m3)
1366.8±56.0 and 1051.1±43.0; NH3 (ug/m3) 1957.7±8.1 and 1241.5±6.1; NMHC (mg/m3) 15.9 and
25.9. Particles represented only 0.04% of the total exposure mass and particle size in the H exposure
ranged from 5.5 to 150 nm with the majority between 5-20 nm (MMD 150 nm) (McDonald et al.,
2008, 191978). Although particles were filtered out, it should be noted that NMHC (non-methane
volatile organics) increased by 62%. Mice were exposed with or without prior sensitization to OVA,
after one aerosol challenge and with or without secondary challenge. Acute gasoline engine exhaust
exposure had variable effects on inflammatory and allergic markers depending on the exposure
protocol, but there were no statistically significant differences between the H and HF exposure
results, suggesting that the PM fraction of gasoline engine exhaust does not appear to contribute
significantly to observed health effects.
Hardwood Smoke
One study indicated that hardwood smoke exposure only minimally exacerbated indices of
allergic airway inflammation in an OVA-sensitized BALB/c mouse model and did not alter Thl/Th2
December 2009 6-125
-------
cytokine levels (Barrett et al., 2006, 155677). Trend analysis indicated increasing BALF eosinophils
with increasing dose of hardwood smoke, becoming significantly elevated at 300 (ig/m3 (CO,
1.6±0.3 ppm; total vapor hydrocarbon, 0.6±0.2 ppm; NOX, below limit of quantitation, PM MMAD
0.35±2.0 (im), and increasing, but not significantly, OVA-specific IgE levels with hardwood smoke
up to 1,000 ug/m3.
Model Particles
Exposures to an aerosol of soot and iron oxide generated from ethylene (0.235 mg/m3 PM2.5)
were conducted to test whether the sequence of exposure to OVA aerosol challenge and PM affected
the observed response of OVA sensitized BALB/c mice (Last et al., 2004, 097334). Though called
PM2.5, the authors characterized the PM material as UF, 80-110 nm, with the iron oxide crystals often
spatially segregated from the soot (200 (ig/m3 soot, remainder iron oxide, CO <0.8 ppm, NOX
<0.4 ppm, PAH below detection). Mice were exposed to PM via chamber inhalation for 2 wk
(4h/day, 3 days/wk) before or after 4 wk of OVA inhalation, or simultaneously to PM and OVA for
6 wk. Among endpoints (BALF cells, Penh, airway collagen, and goblet cells) only goblet cell
counts were significantly increased with PM exposure in any combination with OVA. There was a
trend toward increased Penh responses with exposure to PM alone or with OVA, particularly when
PM exposure immediately preceded Mch challenge (after or during OVA challenge). Results from
this study are difficult to interpret due to the varied elapsed times between cessation of PM or OVA
treatment and endpoint determination. The mild responses to PM may be related to the
intraperitoneal sensitization protocol used, reputed to generate a highly allergic mouse in which any
additive effects of PM may be obscured by maximal responses to antigen challenge (Deurloo et al.,
2001, 156396; Hao et al.. 2003. 096565).
Res/cfua/0;7F/y/\s/7
Arantes-Costa and colleagues (2008, 187137) estimated that 60 (ig of ROFA would be inhaled
by a mouse during one day of exposure to Sao Paulo air. This dose, given intranasally every other
day for 4 days, increased AHR in both nonsensitized and OVA sensitized/challenged BALB/c mice
upon Mch challenge 2 days after the last exposure. ROFA had no significant impact on eosinophil or
macrophage numbers in the lung, nor did it increase the chronic lung inflammation or thickening
induced by OVA. In many studies, particular effects such as airway obstruction are only evident
when allergic sensitization precedes exposure, but this study and a few others demonstrate allergen-
independent AHR after exposure to PM including CAPs (Lei et al., 2004, 087999) and DE or DE
particles (Hao et al., 2003, 096565; Li et al., 2007,155929).
Allergy in Pregnancy or Early Life
Pregnancy or in utero exposure may confer susceptibility to PM-induced asthmatic responses.
Exposure of pregnant BALB/c mice to aerosolized ROFA leachate by inhalation or to DE particles
intranasally increased asthma susceptibility in their offspring (Fedulov et al., 2008, 097482;
Hamada et al., 2007, 091235). The offspring from dams exposed for 30 min to 50 mg/mL ROFA 1,
3, or 5 days prior to delivery responded to OVA immunization and aerosol challenge with AHR and
increased antigen-specific IgE and IgGl antibodies. AHR was also observed in the offspring of dams
intranasally instilled with 50 (ig of DE particles or TiO2, or 250 (ig CB, indicating that the same
effect could be demonstrated using relatively "inert" particles. Pregnant mice were particularly
sensitive to exposure to DE particles or TiO2 particles, and genetic analysis indicated differential
expression of 80 genes in response to TiO2 on the pregnant background. Thus pregnancy may
enhance responses to PM, and exposure to even relatively inert particles may result in offspring
predisposed to asthma.
Allergic Sensitization
A large number of in vivo animal studies and in vitro studies have demonstrated that particles
can alter the immune response to challenge with specific antigens and suggest that PM acts as an
adjuvant to promote allergic sensitization. This phenomenon was introduced in the 2002 Diesel
December 2009 6-126
-------
Document, and has been noted in multiple animal and human studies by the 2004 PM AQCD
(U.S. EPA, 2004, 056905). Adjuvants enhance the immune response to antigens through various
means, including chemoattraction, cytokines, or enhanced antigen presentation and costimulation,
and may act on a number of cell types. Importantly, adjuvants may be major contributors to the
development of inappropriate immune responses. These immune responses, mediated by T helper
cells, fall along a continuum from T helper type 1 (Thl) to T helper type 2 (Th2). Thl responses,
characterized by IFN-y, are inflammatory and in excess can lead to tissue damage. Alternatively, Th2
responses are characterized by IL-4, IL-5, IL-13, eosinophils, and IgE, and are associated with
allergy and asthma. Autoimmune diseases may be driven by Thl, Th2, or mixed responses, but
allergic diseases are predominantly Th2 mediated, and many of the immunologic effects observed
for PM fall into the Th2 category.
It has been suggested that the capacity of particles to enhance allergic sensitization is
associated more strongly with particle number and surface area than particle mass, and several
studies comparing size fractions of the same material show greater adjuvant activity for an
equivalent mass dose of smaller particles (de Haar et al., 2006, 144746; Inoue et al., 2005, 088625;
Nygaard et al., 2004, 058558). This is particularly true of inert or homogeneous materials, such as
carbon, polystyrene, and TiO2, which vary little in composition with size fraction. Studies using
CAPs have also observed that adjuvancy and allergic exacerbation are more strongly associated with
the UF fraction, possibly due to greater oxidative potential (Kleinman et al., 2005, 087880;
Kleinman et al., 2007, 097082; Li et al., 2009, 190457). In some studies of ambient PM, however,
PMio_25 or PMio have demonstrated equal or greater adjuvancy compared to PM25 (Nygaard et al.,
2004, 058558; Steerenberg et al., 2004, 096024; Steerenberg et al., 2005, 088649). More inhalation
studies to compare size fractions are needed in order to elucidate the role of particle size in
mediating adjuvancy, but this may prove difficult given the influence of composition, e.g.,
combustion related materials (Steerenberg et al., 2006, 088249) and metal content (Gavett et al.,
2003, 053153). which differs among various size fractions and sources.
CAPs
As little as 0.1 (ig of UF Los Angeles CAPs administered intranasally with OVA was able to
significantly boost allergic antibody responses in BALB/c mice (Li et al., 2009, 190457). A
comparison of UFPs (aerodynamic diameter <0.15 um) with a mix of sub-2.5 urn particles
(PM2.5/UFP) collected 200 m from a major freeway delivered intranasally five times over the course
of nine days showed that UFP but not PM2.5/UFP were associated with significant adjuvant effects.
0.5 ug of UFP with OVA (but not alone) led to an increase in BALF eosinophils, allergic cytokines,
inflammatory mediators, and serum OVA-specific IgE/IgGl, as well as allergic tissue inflammation
in the upper and lower airways. Adjuvant effects of UFP were observed with two independently
collected samples (1/2007 and 9/2006) and could not be replicated by administering the same
amount of endotoxin measured in the particles, indicating that the effects were not unique to the
sampling period nor mediated by contaminating endotoxin. UFP had a greater OC and PAH content
than PM2.5/UFP, and induced greater oxidative stress in vitro. Partial blocking of the adjuvant effects
by antioxidant administration implicates redox potential as a key factor in mediating these effects.
The authors suggest that the lack of adjuvancy for UF carbon particles (being mostly EC) is due to a
lack of redox cycling compounds, but this was not tested. In contrast, UF (30-50 nm) CB particles
have demonstrated intranasal adjuvant activity in other studies (de et al., 2005, 097872) when
administered with OVA over three consecutive days. A 200-ug dose increased serum OVA-specific
IgE, local lymph node dendritic cells and OVA-specific Th2 lymphocytes in the lung draining lymph
nodes and lung, as well as post-challenge airway eosinophilia. Doses as low as 20 ug were able to
activate adoptively transferred OVA-specific T cells.
Diesel Exhaust Particles
Resuspended DE particles have been shown to enhance OVA-specific IgGl and IgE in
BALB/c mice exposed via inhalation to doses as low as 200 and 600 ug/m3, respectively (Whitekus
et al., 2002, 157142). Mice were exposed to DE particles (200, 600 and 2,000 ug/m3) for 1 h daily
for 10 days prior to aerosol OVA challenge. Compared with responses to OVA alone, antibody levels
were increased by all OVA+DE particle exposures. Statistical significance was reached for IgGl at
all DE particle exposure levels, whereas OVA specific IgE was significantly increased at the 600 and
2,000 ug/m3 doses and total IgE was significantly elevated at 2,000 ug/m3. Although strong adjuvant
December 2009 6-127
-------
effects were observed, no general markers of inflammation such as eosinophils, IL-5, GM-CSF,
mucin, morphological changes, or eosinophilic major basic protein (MBP) deposition in the airways
were observed in exposed mice. In vitro experiments using the RAW 264.7 macrophage-like cell line
indicated a DE particle-induced lipid peroxidation and protein oxidation, which could be inhibited
by a variety of antioxidants. Also observed was a decrease in the GSH:GSSG ratio and an increase in
HO-1 expression, both of which were inhibited only by the thiol antioxidants NAC and BUG. These
same thiol antioxidants were able to completely block DE particle-related increases in IgE and IgGl,
as well as lipid peroxides and oxidized proteins recovered from BALF at the 2,000 (ig/m3 dose. Thus
solid correlations between in vivo and in vitro antioxidant activities were found, and the reversal of
adjuvant effects by antioxidants in vivo clearly indicates a link between oxidative stress and DE
particle adjuvancy. However, the intranasal adjuvant activity of Ottawa, Canada, dust (EHC-93) in
the same strain of mice was not inhibited by NAC pretreatment (Steerenberg et al., 2004, 087981).
suggesting that disparate pathways may be utilized by different materials to exert immune
stimulation.
Diesel Exhaust
DE inhalation during allergen exposure has been shown to augment IgE production and alter
methylation of T helper genes in BALB/c mice (Liu et al., 2008, 156709) Animals were exposed to
DE (1280 (ig/m3 PM) over a 3-wk period, 5 h per day, concurrent with periodic intranasal
sensitization to the common fungus Aspergillus fumigatus. Gas concentrations were not reported.
Total IgE and BALF eosinophils were elevated with A. fumigatus sensitization and further increased
by concomitant DE exposure. Greater methylation of the IFN-y promoter was observed following
DE and A. fumigatus exposure (but not DE alone) compared to A. fumigatus alone, indicating that
combined DE and allergen exposure might induce methylation and thus suppress expression of Thl
genes. Furthermore, hypomethylation of the IL-4 promoter was detected after exposure to A.
fumigatus and DE compared with exposure to A. fumigatus or DE alone, suggesting pro-allergic Th2
gene activation upon combined exposure to allergen and DE. The changes in methylation status of
these genes were associated with alterations in IgE levels in individual animals, indicating that
modifications at the genetic level could result in predicted downstream effects. This study shows for
the first time that DE exposure can exert pro-allergic in vivo effects on the mouse immune system at
the epigenetic level.
A toxicogenomic approach to investigate early response mechanisms of DE adjuvancy was
taken by Stevens et al. (2008, 157010). BALB/c mice were chamber exposed to filtered air, 500 or
2,000 (ig/m3 PM in DE for 4 h/day over 5 consecutive days and intranasally exposed to OVA on each
of the first 3 days. In the low (500 (ig/m3) vs. high (2,000 (ig/m3) DE exposures, CO, NO, NO2, and
SO2 were <0.1 versus 4.3, <2.5 vs. 9.2, <0.25 vs. 1.1 and <0.06 vs. 0.2 ppm; particle number median
diameters were 80 and 86 nm, and volume median diameters were 184 and 195 nm, respectively.
Lung tissues were assessed for alterations in global gene expression (n = 4) 4 h after the last DE
exposure on day 4. Mice were intranasally challenged with OVA or saline on day 18 and then with
OVA on day 28. Post-challenge results demonstrated mild adjuvancy with antigen and DE exposure
as evidenced by significant increases in eosinophils, neutrophils, lymphocytes, and IL-6 in the
BALF. Antibody responses were not significantly affected by DE exposure, although a slight
increase in IgE after high concentration exposure was observed. DE alone only increased
neutrophils, indicating the need for combined exposure to DE and antigen in the development of
allergic outcomes. Comparison of low DE /OVA vs. air/OVA resulted in no significant changes in
gene sets associated with this treatment. Comparison of the high DE/OVA versus air/OVA, however,
showed significant changes in 23 gene sets, including neutrophil homing and other chemokines,
inflammatory cytokines, numerous interleukins and TNF subtypes, and growth/differentiation
pathways.
Summary of lexicological Study Findings for Allergic Responses
Studies conducted since the last review confirm and extend the 2004 PM AQCD's (U.S. EPA,
2004, 056905) finding that PM can modulate immune reactivity in both humans and animals to
promote allergic sensitization and exacerbate allergic responses. Numerous forms of PM, including
inert materials, have been shown to function as adjuvants, and although toxicological studies of
relatively homogeneous materials demonstrate greater adjuvancy for smaller particles, some analyses
December 2009 6-128
-------
of ambient PM do not. Recent toxicological studies comparing size fractions of well-characterized
ambient PM for adjuvant activity in a direct, controlled fashion via inhalation exposure suggest a
role for oxidative potential, but thus far the relative contributions of size and composition are not
entirely clear. Although epidemiologic studies examining specific allergic outcomes and short-term
exposure PM are relatively rare, the available studies, conducted primarily in Europe, positively
associate various PM size fractions with allergic rhinitis. Similar findings from a number of long
term studies are described in Chapter 7.
6.3.7. Host Defense
The normal and very important role of respiratory immune defense is the detection and/or
destruction of pathogens that enter the lung via inhalation and removal of damaged, transformed
(cancerous), or infected cells. Innate immune defenses of the respiratory tract include mucociliary
clearance, release of toxic antimicrobial proteins into airway surface liquid, and activation of
alveolar macrophages. The innate immune system is the earliest responder to irritation or infection,
initiating the normal inflammatory response including the majority of detrimental inflammatory
processes discussed. Activated macrophages and epithelial cells release cytokines and chemokines
that can bring into play the adaptive immune system, which in turn can produce long-lasting
pathogen-specific immune responses critical for resolving and preventing infections.
6.3.7.1. Epidemiologic Studies
Collectively, results from multicity studies of hospital admissions and ED visits for respiratory
infection as well as single-city studies conducted in the U.S. and Canada (summarized in Figure
6-14) show a positive association between PM and respiratory infections. Lag structure was not
investigated in most studies and effects have been observed in association with current day
concentration (Zanobetti and Schwartz, 2006, 090195) as well as with concentrations modeled using
a 14-day distributed lag function (Peel et al., 2005, 056305). Of studies examining multiple lag
times, associations with increasing lag times were observed (Dominici et al., 2006, 088398; Peel et
al., 2005, 056305; Peng et al., 2008, 156850). Although no significant positive associations were
reported, Slaughter et al. (2005, 073854) observed a trend of increasing association with increasing
lag for acute respiratory infection ED visits with PMb PM2.5, PMi0 and PMi0_2.5. This delay in the
onset of disease may reflect the time necessary for an infection to become established and
symptomatic. The majority of toxicological evidence, described below and in the 2004 PM AQCD
(U.S. EPA, 2004, 056905). suggests that PM impairs innate immunity, the first line of defense in
preventing infection.
6.3.7.2. Toxicological Studies
Several toxicological studies were cited in the 2004 PM AQCD (U.S. EPA, 2004, 056905) that
demonstrated increased susceptibility to infectious agents following exposure to PM. A limited
number of new studies have evaluated the effect of PM on host defense in rodents. Two recent
studies have observed an increase in susceptibility to influenza infection and respiratory syncytial
virus in mice. However, one new study found that wood smoke had no effect on bacterial clearance
in rodents.
Bacterial Infection
Several studies included in the 2004 PM AQCD (U.S. EPA, 2004, 056905) demonstrated
increased susceptibility to infectious agents following exposure to various forms of PM. CAPs
exposed aged rats demonstrated increased S. pneumoniae burdens when a 24-h exposure (65 (ig/m3)
followed infection (Zelikoff et al., 2003, 039009). In another study, IT instillation exposure to
ROFA was found to affect bacterial clearance (Antonini et al., 2002, 035342). Examinations of
mechanisms related to PM interference with host defenses have demonstrated impaired mucociliary
clearance and modified macrophage phagocytosis and chemotaxis. Prolonged exposure to inhaled
particles at sufficiently high concentrations can lead to diminished clearance of PM from the alveolar
December 2009 6-129
-------
region of the lung, resulting in the accumulation of retained particles and an accompanying chronic
alveolar inflammation. Diminished clearance of PM may also increase susceptibility to pulmonary
infection by impeding clearance of pathogens. Impaired phagocytosis by alveolar macrophages may
contribute to a decrease in the lung's capacity to deal with increased particle loads (as occurs during
high-pollution episodes) or infections and affect the local and systemic responses through the release
of biologically active compounds (cytokines, ROS, NO, isoprostanes).
Diesel Exhaust
Since the last review, several additional studies have reported impairment of pathogen
clearance following exposure to various sources of PM. All levels of DE (30, 100, 300 or 1,000
ug/m3) decreased lung bacterial clearance in C57BL/6 mice exposed for 1 wk (7 days/wk, 6 h/day)
prior to infection with Pseudomonas aeruginosa (Harrod et al., 2005, 088144). This effect appeared
concentration dependent up to 100 ug/m3 and was not enhanced at higher concentrations. Lung
inflammation was not induced by DE in the absence of infection, but infection-induced inflammation
was exacerbated by DE at all concentrations without apparent concentration dependency. Measures
of histopathology in infected animals were increased by DE exposure in a concentration-dependent
manner, peaking at 100 ug/m3 and leveling off or decreasing with higher concentrations. Particle
deposition was readily apparent in the lungs after exposure to the lowest concentration of 30 ug/m3.
A loss of ciliated cells was observed at 30 ug/m3 and 100 ug/m3 in large airways and in small
airways at the higher concentration. Alterations in Clara cell morphology and function were
observed at both concentrations as well. Concentrations of gases were reported to be 2.0-45.3 ppm
NO, 0.2-4.0 ppm NO2, 1.5-29.8 ppm CO and 8-365 ppb for SO2 (McDonald et al., 2004, 055644).
PM mass median diameter was -100-150 nm at all exposure levels (>90% below 1 urn in
aerodynamic diameter), with lower exposure concentrations having a slightly smaller size
distribution (Reed et al., 2004, 055625).
Gasoline Exhaust
In a study by Reed et al. (2008, 156903). short or long-term exposure to fresh gasoline exhaust
(6h/day, 7day/wk for 1 wk or 6 mo) did not affect clearance of P. aeruginosa from the lungs of
C57BL/6 mice. Atmospheric characterizations are described above for the Day et al. (2008, 190204)
and McDonald et al. (2008, 191978) studies in Section 6.3.6.3.
Hardwood Smoke
Similar to gasoline exhaust, hardwood smoke does not appear to have significant impact on
pathogen clearance. C57BL/6 mice were exposed to 30-1,000 ug/m3 hardwood smoke by whole-
body inhalation for 1 wk and 6 months (Reed et al., 2006, 156043). Long-term responses are
discussed in Sections 7.3.3.2 and 7.3.7.2. Concentrations of gases ranged from 229.0-14,887.6
mg/m3 for CO, 54.9-139.3 ug/m3 for ammonia, and 177.6-3,455.0 ug/m3 for nonmethane volatile
organic carbon in these exposures. Bacterial clearance of instilled P. aeruginosa was unaffected by
hardwood smoke.
Intratracheal Instillation
Studies demonstrate that ROFA impairs host defenses and that soluble metals are important
contributors. Antonini et al. (2004, 097199) compared sources of ROFA in SD rats. Precipitator
ROFA induced an inflammatory response and diminished pulmonary clearance of L. monocytogenes
while air heater ROFA had no effect on lung bacterial clearance at the same IT dose of 1 mg/lOOg
body weight. Precipitator ROFA generated a metal-dependent hydroxyl radical suggesting that
differences in metal composition were a determinant of the immunotoxicity of ROFA. Subsequent
studies using soluble extracts of ROFA with or without a chelating agent confirmed that soluble
metals were responsible for weakening defenses against bacterial infection and impairing both innate
and adaptive lung immune responses (Roberts et al., 2004, 196994; Roberts et al., 2007, 097623)
ROFA has also been shown to result in ciliated cell loss in BALB/c mice after intranasal
administration of 60 ug every other day for 4 days (Arantes-Costa et al., 2008, 187137).
December 2009 6-130
-------
Viral Infection
Diesel Exhaust
Viral respiratory infections in early life are associated with increased incidence of childhood
asthma and other pulmonary diseases. DE exposure can enhance the progression of influenza
infection. BALB/c mice that were chamber exposed to DE 4 h/day for 5 days and subsequently IT
instilled with influenza A/Bangkok/1/79 virus had increased susceptibility to influenza infection
(Ciencewicki et al, 2007, 096557). Exposures to two concentrations of DE were conducted:
500 ng/m3 (0.9 ppm CO, O.25 ppmNO2, <2.5 ppmNO, and 0.06 ppm SO2) and 2,000 (ig/m3
(5.4 ppm CO, 1.13 ppmNO2, 10.8 ppm NO, and 0.32 ppm SO2). Responses were greater for animals
exposed to 500 (ig/m DE than to 2,000 (ig/m3, and were associated with a significant increase in
IL-6 protein and mRNA expression and IFN-J3 expression. The authors present the possibility that
damage to the epithelium at the higher exposure prevented viral infection and replication. After
exposure to 500 (ig/m3 DE alone or prior to infection, decreased expression of surfactant proteins
(SP) A and D was observed. These proteins are part of the IFN-independent defense against
influenza.
Similarly, Harrod et al. (2003, 097046) demonstrated decreased SP-A expression in the lungs
following DE exposure and linked it to increased susceptibility to respiratory syncytial virus (RSV),
the most common cause of respiratory infection in young children. C57BL/6 mice, a relatively
RSV-resistant strain, were exposed via inhalation to DE at a concentration of 30 or 1,000 (ig/m3 PM
6h/day for 7 consecutive days prior to intratracheal viral inoculation. Gaseous copollutants ranged
from 2.0-43.3 ppm for NOX (~ 90% NO), 0.94-29.0 ppm CO, and 8.3-364.9 ppb SO2. Exposure to
30 (ig/m3 DE did not induce a statistically significant increase in BALF cell numbers compared to
air-treated, infected animals. However, distinct consolidated inflammatory infiltrates were observed
in the peribronchial regions of RSV-infected animals exposed to this concentration, along with
alterations in Clara cell morphology, decreased CCSP production by these cells, and occasional
regional myofibril layer thickening. These changes were more pronounced in RSV-infected animals
exposed to 1000 (ig/m3, and the higher concentration also resulted in significant increases in
inflammatory cells, predominantly macrophages, in both uninfected and infected mice compared to
air-exposed controls. Both doses elicited significant levels of TNF-a and IFN-y in the lungs of
infected animals, but decreased levels of SP-A. Consistent with this study's finding of decreased
SP-A and increased viral gene and inflammatory cytokine expression after DE exposure, SP-A"7"
mice demonstrate decreased clearance of RSV concordant with increased lung inflammation (Levine
et al., 1999, 156687). Thus, DE may enhance susceptibility to respiratory viral infections by
reducing the expression and production of SP (Ciencewicki et al., 2007, 096557; Harrod et al.,
2003, 097046). although the contribution of gaseous copollutants, in some instances concentrated
1,000 times, should be considered for both studies. SP are also essential for clearance of other
pathogens, including group B Streptococcus (GBS), Haemophilus influenzcte, and P. aeruginosa
(LeVine and Whitsett, 2001, 155928).
A reduction in host defense molecules and an increase in viral entry sites was observed by
Gowdy et al. (2008, 097226) after BALB/c mice were exposed to HEPA filtered room air or DE at
0.5 or 2.0 mg/m3 for 4hr/day for one or five consecutive days [O2 (%) 21.0±0.10 or 20.7±0.09, CO
(ppm) 1.7±0.15 or 5.4±0.07, NOX (ppm) 2.0±0.36 or 7.4±0.61, SO2 (ppm) 0.0±0.0 or 0.4±0.3,
number median (nm) 96.2±2 or 97±2, volume median (nm) 238±2 or 249±2, OC/EC (wt ratio)
0.4±0.04 or 0.4±0.07 for the 0.5 or 2.0 mg/m3 exposures, respectively]. One of the more notable
features of this study was the observation that effects of extended exposure to the lower
concentration (0.5 mg/m3 for 5 days) tended to persist beyond 18 h post-exposure. Exposure to DE
significantly increased BALF neutrophils in the higher exposure group, and this response persisted
beyond 18 h only after the five day exposure. An increase in ICAM-1 expression (a viral entry site)
was observed in both exposure groups, and was persistent in the lower concentration group after a
5-day exposure. Persistently elevated expression of pro-inflammatory cytokines IL-6 and TNF-a and
pro-allergic cytokine IL-13 was observed after five days of low concentration exposure. Non-
statistically significant effects of either concentration or exposure regimen included increased IFN-y
and MIP-2. Host defense molecules CCSP, SP-A and SP-D were decreased after either exposure
regimen, persisting beyond 18 h in the low concentration group.
December 2009 6-131
-------
Taken together, these data suggest that exposure to DE can weaken host defenses, in some
cases persistently. A role for PM in these responses is supported by studies demonstrating changes in
host defense molecules and viral entry sites in vitro consistent with those observed in vivo. In lung
epithelial cells, DE particles increased the mRNA expression of 1C AM-1, LDL and
platelet-activating factor (PAF) receptors, which can act as receptors for viruses or bacteria (Ito et
al., 2006, 096648). DE particles may therefore enhance the susceptibility to infection by the
upregulation of bacterial and viral invasion sites in the lungs. Expression of the (3-defensin-2 gene,
which is one antimicrobial mechanism of host defense in the airway, was significantly inhibited by V
and not Ni or Fe in airway epithelial cells incubated with aqueous leachate of ROFA (Klein-Patel et
al., 2006, 097092).
Immunosuppressive Effects of PM
Diesel Exhaust
DE may affect systemic immunity. Decreased thymus weight was observed in female F344
rats exposed to 300 ug/m3 DE for 1 wk by Reed et al. (2004, 055625). Concentrations of gases for
this PM concentration were reported to be approximately 16.1 ppm for NO, 0.8 ppm for NO2,
9.8 ppm for CO, and 115 ppb for SO2. Long-term responses are discussed in Section 7.3.8.
Summary of lexicological Study Findings for Host Defense
Toxicological studies demonstrate that short-term inhalation exposures to CAPs and DE, but
not gasoline exhaust or wood smoke, can increase susceptibility to infection by bacterial and viral
pathogens. While gaseous copollutants may be contributing factors, a role for particles is
demonstrated by studies utilizing IT instillation exposure and in vitro studies of PM where
biomarkers parallel those observed in vivo. Although ethical considerations limit controlled exposure
studies in humans, epidemiologic evidence reflects an association between most PM size fractions
and hospital admissions for respiratory infections. Importantly, toxicological studies demonstrate
impaired host defense against the etiological agents of influenza, pneumonia (S. pneumoniae), and
bronchiolitis (RSV), which are commonly reported respiratory morbidities associated with PM.
6.3.8. Respiratory ED Visits, Hospital Admissions and Physician Visits
The epidemiologic evidence presented in the 2004 PM AQCD (U.S. EPA, 2004, 056905)
linking short-term increases in PM concentration with respiratory hospitalizations and ED visits was
consistent across studies. Recent investigations provide further support for this relationship, with
larger effect estimates observed among children and older adults. However, effect estimates are
clearly heterogeneous, with evidence of both regional and seasonal differences at play.
Excess risk estimates for hospitalizations or ED visits for all respiratory diseases combined,
reported in studies reviewed in the 2004 PM AQCD (U.S. EPA, 2004, 056905) fell within the range
of approximately 1-4% per 10 ug/m3 increase in PMi0. On average, excess risks for asthma were
higher than excess risks for COPD and pneumonia. Associations with PM2.5 (including PMi) and
PM10_2.5 were also reported in the limited body of evidence reviewed in the 2004 AQCD. Excess risk
estimates fell within the range of approximately 2.0-6.0% per 10 ug/m3 increases in PM2.5 or PMi0_2.5
for all respiratory diseases combined as well as COPD admissions. Larger estimates were reported
for asthma admissions. Many of the associations of respiratory admissions and ED visits with
short-term PM2 5 concentration were statistically significant. The associations with PMi0_2.5 were less
precise with fewer reaching statistical significance (U.S. EPA, 2004, 056905). Finally, several
studies reviewed in the 2004 AQCD reported associations of PM with outpatient physician visits,
suggesting that the population impacted by short-term increases in PM is not restricted to those
admitted to the hospital or seeking medical attention through an ED.
December 2009 6-132
-------
Table 6-13. Description of ICD-9 and ICD-10 codes for diseases of the respiratory system.
Description
ICD 9 Codes
ICD 10 Codes
Diseases of the Respiratory System 460-519
JOO-J99
Asthma
493
J45
COPD and allied conditions
490-496 (asthma, chronic bronchitis, emphysema,
bronchiectasis, extrinsic allergic alveolitis)
Chronic lower respiratory diseases
J40-J47 (bronchitis, emphysema, other COPD, asthma, status
asthmaticus, bronchiectasis)
Acute Respiratory Infections
460-466 (common cold, sinusitis, pharyngitis, tonsillitis,
laryngitis & tracheitis, bronchitis & bronchiolitis)
Acute Upper Respiratory Infections
JOO-J06 (common cold, sinusitis, pharyngitis, tonsillitis,
laryngitis & tracheitis, croup & epiglottitis)
Acute bronchitis and bronchiolitis
466
J20-J22
Allergic Rhinitis
477
J30.1
Pneumonia
480-486
J13-J18
Wheezing
786.09
Hospital admissions or ED visits for respiratory diseases and ambient concentrations of PM
have been the subject of more than 90 peer-reviewed research publications since 2002 (Annex E).
Included among these new publications are several large single-city and multicity studies. These new
studies complement those reviewed in the 2004 AQCD by examining the effect of several PM size
fractions and components on increasingly specific disease endpoints, as well as evaluating the
presence of effect modification by factors such as season and region.
Specific design and methodological considerations of the large and multicity studies included
in this review were discussed previously (Section 6.2.10). Like the CVD endpoints discussed, the
respiratory endpoints examined in these studies were heterogeneous and approaches to selecting
cases for inclusion in the studies were varied. ICD codes commonly used in hospital admission and
ED visits studies for diseases of the respiratory system are found in Table 6-13.
6.3.8.1. All Respiratory Diseases
Findings from new studies of PM and respiratory hospitalization and ED visits among children
are summarized in Figure 6-10. Results from new studies of adults are summarized in Figure 6-11.
Information on the PM concentrations during the relevant study periods is found in Table 6-14.
Children
Barnett et al. (2005, 087394) used a case-crossover design to study respiratory hospital
admissions (ICD-9 460-519) of children (age groups 0, 1-4, and 5-14 yr) in seven cities in Australia
and New Zealand from 1998 to 2001. All respiratory diseases (ICD 10 JOO-J99) except Mendelson's
Syndrome, post-procedural disorders, asphyxia and certain other symptoms (ICD 10 codes
J95.4-J95.9, R09.1, R09.8) were included in the study. In addition, scheduled admissions and
transfers from other hospitals were excluded. Using an a priori lag (0- to 1-day avg), increases in
respiratory hospital admissions of 2.0% (95% CI: -0.13 to 4.3) among infants <1 yr old, 2.3%
(95% CI: 1.9-7.3) among children 1-4 yr old and 2.5% (95% CI: 0.1-5.1) among children 5-14 yr old
per 10 ug/m3 increase in 24-h avg PM10 were observed. Increases of 6.4% (95% CI: 2.7-10.3) among
infants <1 yr and 4.5% (95% CI: 1.9-7.3) among children 1-4 yr per 10 ug/m3 increase in PM2.5 were
observed.
Ostro et al. (2009, 191971) studied the effect of PM2.5 and components on respiratory disease
(ICD9 460-519) hospitalizations among children <19 yr from 2000 to 2003 in six counties in
California. The nine components examined (EC, OC, nitrates, sulfates, Cu, Fe, K, Si and Zn), were
chosen because they made up relatively large proportion of PM2.5, had a signal to noise ratio >2, or
December 2009
6-133
-------
the majority of their values were greater than the level of detection. Single day lags of 0-3 days were
evaluated. The largest risks were observed at lag 3 days for PM25 (2.8% [95%CI: 1.2-4.3] per 10
(ig/m3), EC (5.4% [95% CI: 0.8-10.3] per IQR) and Fe (4.7% [95% CI: 2.2-7.2] per IQR increase).
Although not as great, positive associations were also observed for OC, SO42~, nitrate, Cu and Zn.
In a study of PM2.5 from wildfires in California during 2003, Delfino et al. (2009, 191994)
evaluated conducted stratified analyses comparing PM2.5 associations pre-, post- and during the
wildfires. Four age groups (0-4, 5-19, 20-64 and >65 yr) were considered in these analyses. Authors
found increased respiratory disease admissions in the periods before (2.6% [95%CI: -5.4 to 11.3])
and during (2.7% [95%CI: -1.6 to 7.6]) the wildfires among children 5-19 yr old, but not after the
wildfire period. Among younger children (0-4 yr), hospital admissions were increased during fire
periods (4.5% [95% CI: 1-8.2]), but not before or after the wildfire period. Estimated zip code level
PM2.5 concentrations were 90 (ig/m3 and 75 (ig/m3 during heavy and light smoke conditions,
respectively, compared to 20 (ig/m3 during non-fire periods.
In the study of six cities in France described previously (PSAS), investigators report a change
of 0.4% (95%CI: -1.2 to 2) per 10 ug/m3 increases in PM25 for all respiratory diseases combined
(ICD-10: JOO-J99) among children from 0-14 yr old (Host et al., 2008, 155852). The same study
reported a larger increase associated with PMi0_2.5 of 6.2% (95% CI: 0.4-12.3, 0-1 day avg) per
10 ug/m3 increase among children. A relatively large effect for PMio_2.s (31% [95% CI: -4.7 to 80])
was also observed in a single-city study of children <3 yr in Vancouver (Yang et al., 2004, 087488).
The non-significant PM2 5 effect estimates were not presented in the publication. Luginaah et al.
(2005, 057327) did not observe significant increases in respiratory hospitalizations with increasing
PMio concentrations among male or female children in Ontario Canada, while Ulirsch et al. (2007,
091332) reported increased admissions for respiratory hospitalizations, ED and urgent care visits
combined among children <17 yr in association with PMi0.
As shown in Figure 6-10, studies of respiratory hospitalizations or ED visits reported
increased risks to children in association with all size fractions. However, increased risk among boys
was not observed in Ontario (Luginaah et al., 2005, 057327). Estimates are imprecise and it is not
clear if associations with PM2 5, PMi0_2.5, or both are driving associations observed with PMi0.
Study
Location
Lag Age
Effect Estimate (95% CI)
Barnettetal. (2005.087394) Australia/NZ 0-1
Host et al. (2008, 1 55852) 6 Cities France 0-1
Ostro etal. (2009,191971) 6 Counties CA 3
Delfino etal. (2009. 190254) 6 Counties, CA 0-1
(wildfires)
Host et al. (2008, 1 55852) 6 Cities France 0-1
Yang etal. (2004, 087488) Vancouver
Barnettetal. (2005.087394) Au/NZ 0-1
0-1
0-1
Ulirsch et al. (2007, 091332)* 2 Cities, SE Idaho 0
Luginaah et al. (2005. 057327) Windsor, Can 1
2
3
1
2
3
"Hospital admissions, ED and urgent care visits combined
1-4
C -I /
0-19
0-4
5-19
14
<3 (31% (-4.7 to 80)^
1-4 J
5-14
0-17
0-1 4 Females
D-1 4 Females
0-1 4 Males •
014 Male" 4 »
_ "
_
PMio
_
^
i i I i i i i
-4 -2 Q 2 4 6 8
Excess Risk (%)
Figure 6-10. Excess risk estimates per 10 ug/m3 24-h avg PM2.5, PM10.2.5, and PM10
concentration for ED visits and HAs for respiratory diseases in children. Studies
represented in the figure include all multicity studies as well as single-city
studies conducted in the U.S. or Canada.
December 2009
6-134
-------
Adults and All Ages Combined
In the study of four million ED visits from 31 hospitals in Atlanta described previously,
SOPHIA investigators reported an excess risk of 1.3% (95% CI: 0.4-2.1, lag 0-2) per 10 ug/m3
increase in 24-h avg PM10 for ED visits for respiratory causes combined (ICD-9: 460-466, 477,
480-486, 491-493, 496, 786.09) among all ages during January 1993-August 2000 (Peel et al., 2005,
056305). PM2.5, PMio_2.5, UF number count and PM2 5 components (SO42~, acidity, EC, OC, and an
index of water-soluble transition metals) were available for inclusion in analyses beginning August
1, 1998. Excess risks of 1.6% (95% CI: -0.003 to 3.5) per 10 ug/m3 increase in 24-h avg PM2.5 and
0.6% (95% CI: -3.6 to 5.1) per 10 ug/m3 increase in PMio-2.5 were reported. Weaker, less precise
associations with components were reported and no increase with UF PNC was observed.
Analyses with four additional years of data were conducted and more recently reported by
SOPHIA investigators (Tolbert et al., 2007, 090316). Single-pollutant results are included in Figure
6-11. The effect of PMi0 remained with the additional years of data, while the effect of PM2.5 was
diminished and a decrease in ED visits with PMi0_2.5 was observed. The association of PMi0 with
respiratory disease ED visits was robust to adjustment for O3, CO and NO2. In another recent
analysis using SOPHIA data from 1998 through 2002 to compare source apportionment methods,
Sarnat et al. (2008, 097972) reported that PM2 5 from mobile sources, PM2 5 from biomass burning
and SO42~-rich secondary PM2 5 were associated with respiratory ED visits and associations were
robust to the choice of the method. Excess risks were statistically significant, ranging from
approximately 2-4%, depending on the method.
In a French multicity study, larger increases were observed in association with 24-h avg
PMio_2.5 concentration compared to PM2 5 concentration among adults as well as children. Among
adults 15-64 yr, investigators reported increases in respiratory hospitalizations of 0.8% (95%CI:
-0.7 to 2.3) and 2.6% (95%CI: -0.5 to 5.8) per 10 ug/m3 for PM25 and PM10_25, respectively (lag
0-1 days) (Host et al., 2008, 155852).
In a study of respiratory hospital admission and ED visits (ICD-9 Codes 460-519) among all
ages conducted in Spokane, Washington, no associations were observed with any size fraction of PM
considered (e.g., PMi, PM25, PM10_2.5, PM10) (Slaughter et al., 2005, 073854). Furthermore, several
of the same investigators conducted a source apportionment analysis using daily PM25 filter samples
from the same residential monitor in Spokane (Schreuder et al., 2006, 097959). In this investigation,
PM2 5 from vegetative burning in the previous day (lag 1) was associated with respiratory hospital
admissions (2.3% [95% CI: 0.9-3.8] per interquartile range increase in the source marker). In a study
of PM25 from wildfires in California during 2003, associations with respiratory hospitalizations were
generally stronger relative to associations in the periods before and after the fires (Delfino et al.,
2009, 191994). Among adults 20-64 yr, an increase of 2.4% (95% CI: 0.5-4.4 per 10 ug/m3) was
reported during the wildfire period compared to 0.9% (95%CI: -0.1 to 1.8 per 10 ug/m ) for all
periods combined (pre-, post- and during wildfires).
Luginaah et al. (2005, 057327) examined respiratory hospital admissions in relation to PMi0
concentration across strata for age and gender and compared time series to case-crossover
approaches. The results for all ages combined, which were relatively precise, stratified by gender and
all lags are presented in Figure 6-11; the largest estimates for PMi0 were for adult males (15-64 yr
old). Fung et al. (2005, 093262) did not report evidence of an association between respiratory
admissions and 24-h PMi0 concentration among adults <65 yr, in a study in Ontario, Canada, while
Ulirsch et al. (2007, 091332) reported a significant positive association among all ages and adults
(18-64 yr) in two Southeast Idaho cities for hospitalizations, ED and urgent care visits combined.
This estimate was robust to adjustment for gaseous pollutants.
Older Adults
Among older adults, MCAPS investigators observed largely null findings for PM25 and
respiratory hospitalizations (ICD-9: 490-492, 464-466, 480-487) for the U.S. as a whole, but
reported heterogeneity in effect estimates across the country that were explained by regional and
seasonal factors (Bell et al., 2008, 156266). The nationwide excess risk of respiratory admissions
with PM2.5 was 0.22% (95% PI: -0.12 to 0.56, lag 0) (Bell et al., 2008, 156266). The largest increase
was observed during the winter in the Northeast (1.76% [95% PI: 0.60-2.93], lag 0). Significant
increases in respiratory admissions were also observed at lag 2. In an analysis of PMi0_2.5, MCAPS
December 2009 6-135
-------
investigators observed small imprecise increases in respiratory admissions with 24-h PM10 25
concentration (0.33% [95% PI: -0.21 to 0.86, per 10 ug/m3, lag 0]) (Peng et al, 2008, 156850).
which decreased after adjustment for PM25 (0.26% [95% PI: -0.32 to 0.84 per 10 ug/mj lag 0]).
Associations with PM2.5 increased (0.7% [95% PI: 0-1.5, lag 0]) or persisted (0.6% [95% PI: -0.2 to
1.25, lag 2]), after adjustment for PMi0_2.5.
Two recent MCAPS analyses evaluate the effect of PM25 components on respiratory hospital
admissions. Bell et al. (2009, 191997) analyzed a subset of MCAPS data restricted to 106 counties
with data available for both long-term average concentrations of PM25 components (Bell et al.,
2007, 155683) and PM25 total mass (1999-2005). The components evaluated included 20 chemicals
with demonstrated toxicity or that contribute a large proportion of PM25 mass (Al, NH4+, As, Ca, Cl,
Cu, EC, OCM, Fe, Pb, Mg, Ni, NO3", K, Si, Na+, Ti, V, Zn). Increases in effect estimates of 511%
(95% PI: 80.7-941) for EC, 223% (95% PI: 36.9-410) for Ni and 392% (95% CI: 46.3-738) for V per
IQR increases in county-specific component fraction were observed. Associations were somewhat
reduced and non-significant in two-pollutant models. When Queens or New York County were
excluded, the association of V with hospital admissions lost significance. Associations were also
diminished when alternative lag structures were considered.
Peng et al. (2009, 191998) linked data on hospital admissions for respiratory causes among
older adults from 2000-2006 to daily air levels from the STN in 119 counties in which both sets of
data were available. Chemical constituents evaluated were SO42~, nitrate, Si, EC, OCM, sodium and
ammonium ions. Single-day lags of 0-2 days were considered. These investigators found a 0.82%
increase (95% PI: 0.22-1.44) per IQR increase in same day OCM. After adjustment for the other
components, a 1.01% (95% PI: 0.04-1.98, lag 0) increase in respiratory admissions per IQR increase
OCM was observed.
French PSAS investigators reported a non-significant increase in hospitalizations for
respiratory diseases (ICD-10 JOO-J99) with 24-h avg PMi0_2.5 among older adults. PM25 estimates
were also not significant (Host et al., 2008, 155852). Adjusted estimates from two-pollutant models
were not presented. Positive associations of first hospitalization, overall hospitalizations and
readmission for respiratory diseases and PMi0_2 5 were also reported among older adults in Vancouver
(Chen et al., 2005, 087555). PMi0.2.5 was associated with an increase of 15% (95% CI: 4.8-22.8) in
overall admissions per 10 ug/m3. Increases associated with PMi0_2.s were larger for readmissions
compared to overall admissions. The association for PM25 with overall admissions was 5.1% (95%
CI: -4.9 to 13) and the association with readmissions was not larger. In this study, effect estimates for
PMio_2.5 and PMi0lost precision, but were robust to adjustment for gaseous pollutants, while the
estimate for PM2 5 was null after adjustment for gaseous pollutants. In Vancouver, Fung et al. (2006,
089789) report increased admissions of 1.8% (95% CI: -2.5 to 5.8) per 10 ug/m increase in PM2
2.5
and 3.8% (95% CI: 0-7.6) per 10 ug/m increase in PMi0_2.5 (lag 0-1 day avg) among adults >65 yr.
In a multicity Australian study, Simpson et al. (2005, 087438) examined the association
between PM25 measured by nephelometry and respiratory hospital admissions (ICD-9 460-519)
among older adults (>65 yr) and reported significant associations (1.055 [95% CI: 1.008-1.1045], lag
0-1 day avg) from a meta-analysis combining effect estimates from all cities. Results from three
statistical models were considered, including standard GAM, which produced similar results.
Delfino et al. (2009, 191994) reported that PM2 5 from wildfire in California was associated
with respiratory hospital admissions among older adults (3% 95% CI: 1.1-4.9 per 10 ug/m3). In two
analyses of data collected in Copenhagen, Denmark between 1999 and 2004, several size fractions
including UF and accumulation mode (Andersen et al., 2008, 189651) and PMi0 sources (Andersen
et al., 2007, 093201) were investigated in relation to respiratory hospitalizations (J41-42, J43,
J44-46) among adults >65 yr of age. Of the size fractions examined (NC total, NC median diameter
of 12 nm [NCai2], NCa23, NCa57, NCai00, NCa2i2, PMi0, PM2.5) NCa2i2, typically aged secondary
long-range transported, NCa57 and PM10 were significantly associated with respiratory
hospitalizations (Andersen et al., 2008, 189651). PMi0 sources including biomass combustion,
secondary inorganic compounds, oil combustion, and crustal were associated with respiratory
hospitalizations (excess risks ranged from 3.5% to 5.4% per interquartile range, respectively)
(Andersen et al., 2007, 093201). PMi0 associations were diminished somewhat in two-pollutant
models (Andersen et al., 2007, 093201: 2008, 189651): the authors note that it was difficult to
separate the effects of PM10 and NCa2i2, which were highly correlated in these data. PM2 5 was not
associated with respiratory hospitalizations in these data.
Results from other single-city studies offer somewhat consistent evidence for the effect of
on respiratory admissions among older age groups. Ulirsch et al. (2007, 091332) found
December 2009 6-136
-------
increases in hospitalizations, ED and urgent care visits combined among this age group in two cities
of Southeast Idaho. Two studies in Vancouver report increased admissions for respiratory causes
with the largest effects observed for a 3-day ma (0-2 days) (Chen et al., 2005, 087555; Fung et al.,
2006, 089789). Fung et al. (2005, 093262) observed non-significant increases in admissions with
PMio among older adults in Ontario, Canada, while another study conducted in Ontario (Luginaah et
al., 2005, 057327) did not provide compelling evidence for an effect that was robust to method
selection, although some increases among males were observed. Finally, a study of hospital
admissions for cardiopulmonary conditions combined among older adults (>65 yr) in Allegheny
County, PA found a positive association with PMi0 at lag 0 (Arena et al., 2006, 088631).
Effect estimates for adults (and combined age groups) as well as older adults are found in
Figure 6-11. Effects observed in single-city studies are generally imprecise but most studies report
positive associations. Regional and seasonal variation was observed with the largest effect estimate
reported by Bell et al. (2008, 156266) in the Northeast during the winter. Although the number of
studies examining components or sources was limited, EC, OC, Ni, V, and PM2.5 from mobile
sources were associated with increased respiratory admissions. Several additional studies conducted
outside the U.S. and Canada reported positive associations of respiratory hospitalizations with PMi0
for different age groups and lags (Bedeschi et al., 2007, 090712: Chen et al., 2005, 087555: Chen
et al., 2006, 087947: Hanigan et al., 2008, 156518: Lai and Cheng, 2008, 180301: Larrieu et al.,
2009, 180294; Middleton et al., 2008, 156760; Oftedal et al., 2003, 055623), PM25 (Hinwood et
al., 2006, 088976; Neuberger et al., 2004, 093249; Vigotti et al., 2007, 090711). BS (Bartzokas et
al., 2004, 093252; Tecer et al., 2008, 180030) and with PM10.2.5 (Tecer et al., 2008, 180030). Other
studies reported no associations with PMi0 (Vegni and Ros, 2004, 087448) or TSP (Llorca et al.,
2005, 087825).
6.3.8.2. Asthma
Results from multicity studies of hospital admissions and ED visits for asthma as well as
single-city studies conducted in the U.S. and Canada are summarized in Figure 6-12. Studies
reviewed in the 2004 AQCD are included for continuity. Concentrations of PM for the relevant study
period are found in Table 6-14.
Children
SOPHIA investigators (Peel et al., 2005, 056305) reported that, of the PM mass indicators
examined, the largest effect estimate observed using the a priori lag (0- to 2-day avg) was the
association of PMi0 with pediatric (2-18 yr) asthma ED visits (1.6% [95% CI: -0.2 to 3.4]). ED visits
for both asthma (ICD-9: 493) and wheezing (ICD-9: 786.09) were included in their study. New York
State DOH (2006, 090132) conducted a study comparing effect estimates for ED visits for asthma
and 24-h PM25 and 1-h PM25 across two communities in New York City (the Bronx and Manhattan).
No associations with 24-h PM25 were reported for either borough for age categories 0-4 or 5-18 yr.
Non-significant increases with 1-h maximum PM2 5 were reported for the Bronx. Asthma hospital
admissions (ICD-10 J45, J46, J44.8) in children <14 yr were examined in the Australia/New Zealand
multicity study (Barnett et al., 2005, 087394). In this study, associations for asthma hospital
admissions with PM2 5 and PMio were increased but imprecise.
Lin et al. (2002, 026067) used both time series and case-crossover approaches to investigate
the influence of PM on asthma hospitalization in children, 6-12 yr old, in Toronto from 1981 to
1993. These authors report relatively small differences in results obtained through bi-directional case
crossover and time series approaches, but indicate that unidirectional case-crossover methods may
overestimate the relative risks. Single- to 7-day avg lags were investigated and estimates appeared to
increase and then level off at the longer lags (0- to 2-day and 0- to 5-day lags are shown in Figure
6-12). Effect estimates for PM2 5 are not easily distinguished from the null, but PMi0_2.5 is
significantly associated with asthma admissions among boys and among girls. These associations
were imprecise, but robust to adjustment for gaseous pollutants, among all children combined.
December 2009 6-137
-------
Study
Location Lag Age Gender
Effect Estimate (95% Cl)
ADULTS OR ALL AGES COMBINED
Peel etal. (2005.056305)*
Tolbertetal. (2007,090316)*
Slaughter et al. (2005, 073854)*
Host etal. (2008, 155852)
Delfino etal. (2009. 191994)
Peel etal. (2005.056305)*
Tolbertetal. (2007, 09031 6)*
Slaughter et al. (2005, 073854)*
Host etal. (2008, 155852)
Peel etal. (2005.056305)*
Tolbertetal. (2007, 09031 6)*
Luginaah et al. (2005, 057327)
Slaughter et al. (2005, 073854)*
Ulirsch etal. (2007. 091 332)**
Fung etal. (2005.093262)
Atlanta GA 0-2 All
Atlanta, GA 0-2 All -i
Spokane WA 1 All
6 Cities, France 0-1 15-64 — J
California Wildfires 0-1 20-64
Atlanta GA 0 ° All
Atlanta GA 0-° All *
Spokane, WA 1 All — i
6 Cities France 0-1 15-64 — i
Atlanta, GA 0-2 All
Atlanta, GA 0-2 All
Windsor Can 1 All Female i
? All Female ,......J
3 All Female
1 All Male
0 All MalP
3 All Male
Spokane, WA 1 All _i
2 All _j
3 All _i
Idaho 0 All
0 18-64
Ontarin,r.an n
-------
Although Ostro et al. (2009, 191971) presented estimates for all respiratory diseases
combined, these authors note that PM2.5 and its components were associated with asthma
hospitalizations among the children in six counties of Los Angeles studied. Delfino et al. (2009,
191994) examined the association of PM25 before, during, and after wildfires in California with
asthma hospitalizations among age and gender subgroups. Associations were observed for children
0-4 yr among children during the wildfire period (8.3% [95% CI: 2.1-14.9] per 10 ug/m3), but not
before or after the wildfire period. For older children, 5-19 yr, non-significant increases in asthma
hospitalizations were found before the wildfire period, but not during or after the fires.
Hirshon et al. (2008, 180375) studied hospital admissions and ED visits by children 0-17 yr
old in Baltimore, MD from June 2002-November 2002, in relation to Zn as a component in PM2.5.
Single day lags from 0-2 days were tested with the highest estimates observed for the previous day.
A 23% (95% CI: 7-41) increase in admissions was observed comparing medium (8.63-20.76 ng/m3)
concentrations on the previous day to low concentrations (<8.63 ng/m ) on the previous day.
Previous day high concentration (>20.76 ng/m3) was associated with an increase in admissions of
16% (95% CI: -3 to 39) compared to previous day low concentration. Zinc associations were robust
to adjustment for EC, CO, NO2, Ni, and Cr. However, evidence of effect modification by EC and
NO2 at lags 1 and 2 was observed.
Mohr et al. (2008, 180215) used measurements of EC, O3, SO2, and total NOX from the EPA
supersite in St. Louis for June 2001-May 2003, to examine the association of EC, temperature and
season with asthma ED visits among children 2-17 yr old. The association of EC with asthma ED
visits varied by age, season and weekday versus weekend. The largest associations were observed for
2-5 yr olds during the fall weekends (3% [95% CI: 1-5] per 0.1 ug/m3) and 11-17 yr olds during
winter weekdays (3% [95% CI: 0-5] per 0.1 ug/m3) and summer weekends (9% [95% CI: 2-17] per
0.1 ug/m3). Investigators also report that temperature modified the effect of EC after adjusting for
gaseous copollutants, such that the association of ED visits with EC increased with increasing
temperature during the summer and increased with decreasing temperature during the winter.
Authors attribute the temperature modification to time-activity patterns among this age group.
Sinclair and Tolsma (2004, 088696) investigated respiratory ambulatory care visits using
ARIES data in Atlanta, GA (also used by SOPHIA investigators) and health insurance records. These
authors evaluated three 3-day ma lags (0-2, 2-5 and 6-8 days) and reported relative risks, with no
confidence intervals, for significant results only (not included in Figure 6-12). For childhood asthma
outpatient visits, OHC, PMi0_2.5, PMi0, EC and OC were significantly associated with ambulatory
care visits at lags 0-2 or 2-5 days.
A study in Anchorage used medical records to examine effects of particle exposure on
pediatric asthma outpatient visits, inpatient visits and prescriptions for short-acting inhalers
(Chimonas and Gessner, 2007, 093261). Authors examined Medicaid claims for asthma-related and
lower respiratory infection visits among children less than 20 yr of age for 5 yr (approximately
25,000 children were enrolled in Medicaid each year between 1999 and 2003). Citing work done in
the mid-1980's, the authors describe their city's particles as arising primarily from natural, geologic
sources (PMi0), and to a lesser extent from local automotive emissions (PM2 5) (Pritchett and
Cooper, 1985, 156886). Using GEE in a time-series analysis of daily and weekly effects of particle
exposure on health outcomes, the authors found that each 10 ug/m3 increase in 24-h avg PMi0 was
associated with a 0.6% increase (95% CI: 0.1-1.3) in outpatient visits for asthma. The same increase
in weekly PMi0 concentration resulted in a 2.1% increase (95% CI: 0.4-3.8) in asthma visits, after
adjustment for gaseous pollutants. No meaningful associations were observed for PM25.
In Copenhagen, Denmark, Anderson et al. (2007, 093201) found an association between PMi0
attributed to vehicle emissions and asthma hospitalizations among children 5-18 yr (5.4% 95% CI:
0.57-22.9 per 10 ug/m3, 0- to 5-day avg). In an analysis of size distribution and number
concentration, accumulation mode particles were most strongly associated with asthma admissions
(8% [95% CI: 0-17] per 495 particles/cm3, lag 0-5). (Andersen et al., 2008, 189651). In Helsinki,
Halonen et al. (2008, 189507) examined the association of various size fractions of PM (e.g., Aitken,
accumulation mode, PM2 5, PMi0_2.5) with ED visits for asthma among children <15 yr. These authors
evaluated lags 0-5 and noted a different lag structure depending on age with children experiencing
greater effects at lags 3-5 days compared to adults at lag 0. Aitken, accumulation mode particles and
traffic-related PM were significantly and most strongly associated with asthma visits among
children, while no association with PMi0_2.5 was observed in this age group.
December 2009 6-139
-------
Sti
c
•o
.c
15
.0
o
and All Ages C
M
•a
<
idy
Rarnott ot al HOOfi nfiQ77rT\
Linetal (2002 026067)
nplfinnptal f9nDQ 1Q1QQ41
093261)
NYS DOH (2006 090132) *
Linetal (2002 026067)
Barnettetal (2006 089770)
Peel etal. (2005, 056305)*
Linetal (2002 026067)
Chirnonas & Gessner (2007
093261)
Peel etal. (2005, 056305)*
Ito etal. (2007.091262)*
^lannhtprptal (TTfK flTWiai*
nplfinnptal f9nDQ 1Q1QQ41
Sheppard et al. (2003, 042826)
MY^ now f9nnfi noniTO*
pppiptai nnnR rww^*
Slaughter et al. (2005, 073854)*
Sheppard et al. (2003, 042826)
NYS DOH (2006 090132)*
Peel etal. (2005. 056305)*
Jaffe etal. (2003. 041 957)*
Slaughter et al. (2005, 073854)*
Sheppard etal. (2003. 042826)
Location
Bronx NY
Manhattan NY
Australia/NZ
Atlanta, GA
Atlanta, GA
New York, NY
(wildfire)
Seattle, WA
RrnnY MY
Manhattan MY
Atlanta I^A
Spokane, WA
Seattle, WA
Bronx NY
Manhattan NY
Atlanta, GA
3 Cities, OH
Spokane, WA
Seattle, WA
Lag
n 1
0-1
0-2
n ^
n 9
0-5
n 1
0-1
o
o
0-4
0-4
0-4
0-4
0-2
0-5
0-2
n ^
0-1
0-1
0-2
0-2
0-5
0-2
0-5
o
0
0-2
0-1
0-1
0-1
1
2
•t
n 1
0-1
0
n 4
n 4
n 9
1
2
3
0
0-4
0-4
0-2
0-1 SDL
3
3
2
3
1
2
3
0
Covariates, Age
1 A
5 14
Boy 6 1°
Pirl- R 1°
Girl" 61°
n.4
5 19 i
Outpatient 019 •
0 1
518 *
Q /\ <; t
5 18 <«
Boys 6-12 —I
Boys 6-12
Girls 6-12
Rirk fi 19
1-4 -J
5-1/1
2-18 J
Boys 61°
Boys 61°
Girls 612
Girls 6-1°
Inpatient 019
Outpatient, 0-19
All Ages 1
All Year, All
Cool Season, All
nil
All
All
9n ftd J
65+
<65
All
All
All •
All J
All — I
All «i
<65 —I
All
All
All J
All
All cities, 5-34 — i
Columbus, 5-34
Columbus 5-34 '
All J
All — I
All i
<65
Effect Estimate (95% Cl)
• PMl r
||
—
L,
f PMl'-r
9 v.
* PMm
L_»
L*.
^ PM2.5
—
^
^
—
PM-imr
L_.
l»
I
I .
[ .
,
L*_ PMio
L-»
I
,_ —
L_,_
L»—
k—
* ED Visits
DL Distributed Lag
Figure 6-12.
i I I i i i I i i i I i i i I i i
-8-40 4 8 12 16 20 24
Excess Risk (%)
Excess risk estimates per 10 ug/m increase in 24-h avg PM2.6, PMi0.2.5, and
for asthma ED visits and HAs. Studies represented in the figure include all
multicity studies as well as single-city studies conducted in the U.S. or Canada
Adults and All Ages Combined
Results from the Atlanta SOPHIA study based on the a priori models examining a 3-day ma
(lag 0-2 days) revealed no statistically significant associations with asthma (ICD-9 493, 786.09)
December 2009
6-140
-------
among all ages for any of the PM metrics studied (e.g., PM2.5, PM10_2.5, PM10, UF PNC, PM
components) (Peel et al., 2005, 056305). However, the 14-day unconstrained distributed lag model
produced an excess risk of 9.9% (95% CI: 6.5-13.5 per 10 ug/m3 PMi0). The authors note that
associations of PM25 and OC with asthma tended to be stronger during the warmer months. Sinclair
and Tolsma (2004, 088696) report a significant association between adult outpatient visits for asthma
and UFPs, but not other PM size fractions (not included in Figure 6-12 because only significant
results were presented).
Jaffe et al. (2003, 041957) examined the effects of ambient pollutants (PMi0, O3, NO2 and
SO2) during the summer months (June through August) on the daily number of ED visits for asthma
among Medicaid recipients aged 5-34 yr from 1991 to 1996 in Cincinnati, Columbus, and Cleveland.
Lags 1 to 3 were tested and only statistically significant lags were presented. For all cities combined,
the overall effect estimate for 24-h avg PM10 was 1.0% (95% CI: -1.44 to 3.54 per 10 ug/m3
increase). The effect estimate for Cleveland was the only significantly elevated estimate (2.3%
[95% CI: 0.0-4.9] per 10 ug/m3 increase) when the cities were examined independently. The authors
reported results from analyses indicating a possible concentration response for O3, but no consistent
effects for PMi0.
In New York City, Ito et al. (2007, 156594) examined numbers of ED visits for asthma among
all ages (ICD-9 493) in relation to pollution levels from 1999 to 2002; several weather models were
evaluated. Although the association with NO2 was the strongest, PM2 5 was significantly associated
with asthma ED visits in each weather model (strongest during the warm months) and remained
significant after adjustment for O3, NO2, CO and SO2. Slaughter et al. (2005, 073854) reported no
associations with ED visits or hospitalizations for asthma, among all ages, in Spokane, Washington
for the PM size fractions studied (PMb PM2 5, PMi0, PMi0_2.5). An association with CO, which the
authors attribute to combustion related pollution in general, was observed. The effect of 24-h avg
and 1-h max PM25, PM10_2.5, EC and OC on ED visits for asthma among all ages combined,
comparing two communities in New York City was investigated (ATSDR, 2006, 090132). In the
Bronx, an increase in visits of 3.1% (95% CI: 0.6-6.2 per 10 (ig/m3) was observed in relation to 24-h
avg PM25. For PMi0_2.5, an increase of 2.7% (95% CI: 0.0-5.4) was observed in the Bronx. Smaller,
less precise estimates were observed for Manhattan. Increased asthma visits were observed with OC,
EC and total metals. In the Bronx, the association of 1-h max PM25 with ED visits was larger than
the association with 24-h PM2 5 when standardized to the mean concentration for both communities
and was generally robust to adjustment for copollutants.
Delfmo et al. (2009, 191994) examined the association of PM25 before, during and after
wildfires in California with asthma hospitalizations among age and gender subgroups. The increase
among older adults >65 yr of 10% (95% CI: 3-17.8 per 10 (ig/m3) was larger than the increase
among adults 20-64 yr of 4.1% (95% CI: -0.5 to 9 per 10 (ig/m3). For older adults, the association
was stronger during the wildfire period compared to the pre-wildfire period and did not diminish
during the post-wildfire period.
Effect estimates from studies of hospital admissions and ED visits for asthma are summarized
in Figure 6-12. Associations with PM25 concentration among children are imprecise and not
consistently positive across different age groups and lags. Findings from two studies of PMi0_2.s
(Sinclair and Tolsma, 2004, 088696). as well as PMi0 studies both show positive associations,
although estimates lack precision. Among adults and adults and children combined, associations of
asthma hospital admissions and ED visits with PM25 concentration were observed in most studies.
Positive, non-significant associations of PMio_25 concentration with asthma admissions and ED visits
were observed in some studies of adults. Again, PMi0 estimates are more consistently positive and
precise compared to other size fractions. Associations were observed with several PM2 5 components
(e.g., EC, OC and Zn) and sources (e.g., traffic, wildfires). Many factors (e.g., the underlying
distribution of individual sensitivity and severity, medication use and other personal behaviors) can
influence the lag time observed in observational studies (Forastiere et al., 2008, 186937). Excess
risk estimates for asthma were generally sensitive to choice of lag and increase with longer or
cumulative lags times. Most additional single-city studies conducted in Europe, South America and
Asia, have investigated the associations of asthma hospitalizations, ED visits or doctor visits and
most have reported evidence of an association with TSP (Arbex et al., 2007, 091637; Migliaretti
and Cavallo, 2004, 087425; 2005, 088689). PM10 (Bell et al., 2008, 156266; Bell et al., 2008,
091268; Chardon et al., 2007, 091308; Chen et al., 2006, 087947; Erbas et al., 2005, 073849;
Galan et al., 2003, 087408; Jalaludin et al., 2004, 056595; Kim et al., 2007, 092837; Ko et al.,
2007. 091639; Kuo et al.. 2002. 036310; Lee et al.. 2002. 034826; Lee et al., 2006, 090176) and
December 2009 6-141
-------
PM2.5 (Chardon et al., 2007, 091308; Ko et al, 2007, 091639)) (Ko et al., 2007, 196606) while a
few have not shown an association with PMi0 (Larrieu et al., 2009, 180294; Masjedi et al., 2003,
052100; Tsai et al., 2006, 089768; Yang and Chen, 2007, 092847; Yang et al., 2007, 092848).
6.3.8.3. Chronic Obstructive Pulmonary Disease
Results from multicity studies of hospital admissions and ED visits for COPD as well as
single-city studies conducted in the U.S. and Canada are summarized in Figure 6-13. Studies
reviewed in the AQCD are included in the figure for continuity. Concentrations of PM for the
relevant study period are found in Table 6-14.
In a study of Medicare recipients in 204 U.S. counties, Dominici et al. (2006, 088398)
reported an overall increase of about 1% in COPD hospitalizations (ICD-9 490-492) associated with
24-h avg PM2.5, with the largest effects at lags 0 and 1. In this study effect estimates were
heterogeneous across the U.S. with a significant increase of about 4% observed in the Southeast at
lag 0. In another study using Medicare data in 36 U.S. cities (1986-1999) short-term exposure to
PMio was associated with an increase in COPD hospital admissions (ICD-9 490-496, excluding 493)
of 1.47% (95% CI: 0.93-2.01, lag 1) during the warm season (Medina-Ramon et al., 2006, 087721).
A smaller effect was observed during the cold season.
In Atlanta, SOPHIA investigators reported a comparably sized effect estimate for COPD
(ICD-9 491, 492, 496) and 24-h avg PM25 (1.5% [95% CI: -3.1 to 6.3], 0- to 2-day avg]). The
association of PMi0 with COPD reported by Peel et al. (2005, 056305) was 1.8% (95% CI: -0.6 to
4.3). No associations were observed for PMi0_2.5, UF or PM2.5 components. Slaughter et al. (2005,
073854) reported no associations between any size fraction of PM in Spokane, Washington (PM25,
PM10_25, PM10) and COPD (ICD-9 491, 492, 494, 496). In contrast, Chen et al. (2004, 087262)
reported increases in COPD admissions (ICD-9 490-492, 494, 496) for PM25 (17.1% [95% CI:
4.6-31.0], 0- to 2-day avg), PM10_25 (10.0% [95% CI: -1.2 to 22.8, 0- to 2-day avg]), and PM10
(16.5% [95% CI: 6.88-27.02], 0- to 2-day avg]). However, the estimates for PM metrics were
diminished after adjustment for NO2.
Delfmo et al. (2009, 191994) examined the association of PM25 from the wildfires of 2003 in
California with COPD hospitalizations among age and gender subgroups. Among older adults (>65
years), associations were similar across pre-, post- and wildfire periods with none reaching
significance. The increase for all periods combined in this age group was 1.9% (95% CI: -0.6 to 4.4,
per 10 (ig/m3). Michaud et al. (2004, 188530) reported an association for asthma and COPD ED
visits combined with PMi (lag 1) in Hilo, Hawaii in a study designed to investigate the effect of
volcanic fog.
Halonen et al. (2008, 189507) conducted a study of ED visits for COPD and asthma combined
(J41, J44-J46) among adults 15-64 yr and older adults >65 yr. These authors examined the effects of
Aitken mode particles, accumulation mode particles, PM2 5 and PMi0_2.5 as well as several sources of
PM2 5 (traffic, long range transported particles, road dust and coal/oil combustion). Concentrations,
lagged from 0-5 days, were examined and the largest effects among older adults were observed in
association with concurrent day PM2 5, PM10_2.5, accumulation mode particles, NO2, and CO
concentrations. The PM25 association was diminished with adjustment for UFPs, NO2and CO. A
similar diminishment was observed when PMi0_2.5 was adjusted for PM2 5, NO2 and CO. However,
traffic related particles and long range transported particles (e.g., accumulation mode particles such
as carbon compounds, sulfates and nitrates from central Europe and Russia) were associated with
COPD and asthma among older adults. This same research group conducted additional analyses of
hospital admissions using the same PM metrics focusing on older adults (>65 yr) (Halonen et al.,
2009, 180379). The PM2 5 results and lag structure were similar to the earlier ED visit study. The
strongest effect was for accumulation mode particles with COPD/asthma admissions. Traffic related
PM2 5 was associated with COPD/asthma admissions at lag 1 while no effect was observed with
concurrent day concentration. Long range transported particles and road dust were also associated
with admissions for asthma and COPD.
With the exception of one study conducted in Spokane Washington (Slaughter et al., 2005,
073854). associations have been consistently observed for PM2 5 and PMio with COPD in multicity
and single-city studies conducted in the U.S. and Canada. Associations with PMi0_2.5 are fewer and
less consistent. A study that examined seven single-day lags in association with pooled COPD and
asthma ED visits in Finland reported that PM2 5, PMi0_2.5, traffic sources as well as gaseous pollutants
December 2009 6-142
-------
had a more immediate effect in older adults (lags 0 and 1) compared to children experiencing asthma
(3- to 5-day lags) (Halonen et al., 2008, 189507). Larger estimates at shorter lags were not observed
consistently across other studies. Most single-city studies conducted outside of the U.S. or Canada
focused on PMi0 (Chiu et al., 2008, 191989: Hapcioglu et al., 2006, 093263: Ko et al., 2007,
091639: Ko et al.. 2007. 092844: Martins et al.. 2002. 035059: Masjedi et al.. 2003. 052100:
Sauerzapf et al., 2009, 180082: Yang and Chen, 2007, 092847).
Study
Location
Lag Age/Climate
Effect Estimate (95% Cl)
Dominici et al. (2006, 088398)
Chen (2004 087262)
Delfinoetal. (2009, 191994)
Ito (2003 042856)
Moolgavkar (2003, 051316)
Slaughter et al (2005 073854)*
Chen (2004 087262)
Ito (2003 042856)
Zanobetti & Schwartz (2003, 043119)
Medina-Ramon (2006, 087721)
Peel (2005. 056305)*
Slaughter et al (2005 073854)*
Chen (2004 087262)
Ito (2003 042856)
Moolgavkar (2003, 051316)
* ED Visits
204 Counties, US
6 Counties, CA
Detroit Ml
Los Angeles, CA
Atlanta ft A
Vancouver Can
Detroit Ml
14 Cities, US
36 Cities, US
Atlanta, GA
Spokane WA
Detroit Ml
Los Angeles, CA
Cook County, I L
0 '-*•
1 '-•-
2 -"•—
0-2 DL «-•-
1 * '
3 *
1 65+ '
0 i
0-1 avg 65* -J — »
0 6S4- b
0 65+ «-«_
1 '•
i « i
3 ' •
1 1
0 1 ,
T. 1
^ =wn i
0 ' »
0-1 avg i -*-
0 Warm U-
1 Warm i _»..
0 Cold -*-
1 Cold Jo.
0 1 3 DL '
1 k
2 .------^----^
1
?
^
0-2 avg
0 •
o u-
0 L«,
1
1 1 1 1 1 1 1 1 I
-12 -8-40 4
Excess Risk (%)
A
PM2.5
—
10-2.5
^
PMio
*
i 1 I 1 I 1 I I 1
8 12 16 20 24
for COPD ED visits and HAs among older adults (65+ yr, unless other age group
is noted). Studies represented in the figure include all multicity studies as well as
single-city studies conducted in the U.S. or Canada.
6.3.8.4. Pneumonia and Respiratory Infections
Results from multicity studies of hospital admissions and ED visits for respiratory infection as
well as single-city studies conducted in the U.S. and Canada are summarized in Figure 6-14. The
figure includes studies of respiratory infection reviewed in the 2004 AQCD. Concentrations of PM
for the relevant study period are found in Table 6-14.
December 2009
6-143
-------
Children
In the study of seven cities in Australia and New Zealand, associations of PM25 with
pneumonia and acute bronchitis (ICD-10 J12-J17, J18.0, J18.1, J18.8, J18.9, J20, J21) were observed
among infants <1 yr old (4.54% [95% CI: 0.00-9.20]) and children 1-4 yr old (6.44% [95% CI:
0.26-12.85]) (Barnett et al, 2005, 087394). Although quantitative results were only presented for all
respiratory diseases combined, Ostro et al. (2009, 191971) examined several specific respiratory
diseases including acute bronchitis and pneumonia. They reported that PM2.5 and its components
were more strongly associated with these endpoints compared to other respiratory diseases. Delfino
et al. (2009, 191994) reports imprecise increases in admissions among children during wildfire
periods for acute bronchitis and bronchiolitis, as well as pneumonia.
Inpatient and outpatient visits for lower respiratory tract infections among children in
Anchorage, Alaska, were not associated with PM25 or PMi0 (Chimonas and Gessner, 2007, 093261).
Lin et al. (2005, 087828) observed associations of respiratory infections (ICD-9 464, 466, 480-487)
with PM 10-2.5 and PMi0 that persisted after adjustment for gaseous pollutants among subjects <15 yr
old living in Toronto. Analyses were stratified by gender and both single and multiple day lags were
examined (4- and 6-day avg were presented). The largest significant effect estimates were for
PMio_2.5. The size of the PM2.5 estimate varied by gender and was sensitive to the choice of lag. PM2.5
results were not generally robust to adjustment for gases.
All Ages and Older Adults
SOPHIA investigators examined ED visits for upper respiratory tract infections (URI) (ICD-9
460-466, 477) and pneumonia (ICD-9 480-486) among all ages. An excess risk of 1.4% (95% CI:
0.4-2.5 per 10 ug/m3, lag 0- to 2-day avg) for PMi0 was associated with URI visits. With the
exception of a small increase in risk for OC of 2.8% (95% CI: 0.4-5.3 per 2 ug/m3, 0- to 2-day avg)
with pneumonia visits, Peel et al. (2005, 056305) reported no association with other PM size
fractions or components evaluated. However, Sinclair and Tolsma (2004, 088696). who also used
ARIES data in their analysis, reported significant associations with outpatient visits for LRI. These
associations were generally observed for 3- to 5-day ma lags, in association with PMi0_2.5, PMi0, EC,
OC, and PM2 5 water soluble metals (not pictured in figure because only significant lags were
reported). No associations with pneumonia for any size fractions were observed among all ages in a
study conducted in Spokane, Washington (effect estimates were not reported) (Slaughter et al.,
2005, 073854).
French PSAS investigators examined the effect of PM25 and PMi0_25 on hospital admissions
for respiratory infection (ICD-10: J10-22) among all ages. Increases of 2.5% (95% CI: 0.1-4.8) and
4.4% (95%CI: 0.9-8.0) per 10 ug/m3 were observed in association with PM2.5 and PMi0_2.5,
respectively (Host et al., 2008, 155852). In a multicity study of older adults (>65 yr) Medina-Ramon
et al. (2006, 087721) examined hospital admissions for pneumonia (ICD-9 480-487) in 36 U.S. cities
in relation to 24-h avg PM10 concentration. An increase in pneumonia admissions of 0.84% (95% CI:
0.50-1.19 per 10 ug/m3, lag 0) was reported by these investigators during the warm season. Cold
season associations were weaker (0.30% [95% CI: 0.07-0.53] per 10 ug/m3, lag 0) as were lag 1
associations. Dominici et al. (2006, 088398) investigated hospital admissions for all respiratory
infections including pneumonia (ICD-9 464-466, 480-487) among older adults in 204 urban U.S.
counties in relation to PM2 5 and reported a significant increased risk only at lag 2. Heterogeneity in
effect estimates was observed across the U.S. with the largest associations reported for the South and
Southeast.
In Boston, excess risks of pneumonia hospitalization in association with PM25, BC, and CO
were observed among older adults (Zanobetti and Schwartz, 2006, 090195). A measure of
non-traffic PM, e.g., the residuals from the regression of PM25 on BC, was not associated with
pneumonia hospitalization in these data. In a California study (Delfino et al., 2009, 190254). effect
estimates were of similar magnitude for pneumonia admissions associated with PM2 5 from wildfires
among all ages combined and older adults (2.8% [95% CI: 0.7-5.0] per 10 ug/m3, all ages
combined). The PM2 5 association with acute bronchitis and bronchiolitis admissions during the
wildfire period for all age groups showed an approximately 10% increase (9.6% 95% CI: 1.8-17.9,
per 10 ug/m3). The increase was not larger during the wildfire period compared to the pre-fire period
for either outcome.
December 2009 6-144
-------
In a study of four cities in Australia, statistically significant associations of pneumonia and
acute bronchitis with particles measured by nephelometry (but not PM2 5 mass) and NO2 were
observed among older adults (Simpson et al, 2005, 087438). Halonen et al. (2009, 180379)
examined pneumonia among older adults (ICD10 J12-J15) in their most recent analysis. Associations
of PM2.5 (5.0% [95% CI: 1.0-9.3] per 10 (ig/m3, lag 5-day mean), as well as accumulation mode
particles, with pneumonia admissions were observed.
Although the body of literature is small, several studies of children reported associations of
PM2.5, PMio_2.5 and PMi0 with respiratory infections but the outcomes studied are heterogeneous and
effect estimates are imprecise. Studies of adults show a similar pattern of increased risk for each of
these size fractions. Several other single-city studies conducted outside the U.S. and Canada reported
associations for PMi0 (Cheng et al.. 2007. 093034; Hwang and Chan, 2002, 023222; Nascimento et
al., 2006, 093247) and PM2.5 (Hinwood et al., 2006, 088976) with hospitalization or ED visits for
respiratory infections.
Study
Location
Lag Age Outcome
Effect Estimate (95% CI)
CHILDREN
Barnettetal. (2006.089770)
7 Cities, Australia/NZ
Chimonas & Gessner (2007,093261) Anchorage, AK
Delfino et al. (2009,191994) 6 Counties, S. CA
Lin etal. (2005.087828)
Lin etal. (2005.087828)
Barnettetal. (2006.089770)
Ontario, Can
Ontario, Can
7 Cities, Australia/NZ
Chimonas & Gessner (2007,093261) Anchorage, AK
Lin etal. (2005.087828) Ontario, Can
ALL AGES COMBINED OR OLDER ADULTS
Dominici et al. (2006,088398) 204 Counties, US
0-1
0-1
0
0
0-1
0-1
0-1
0-3
0-3
0-1
0-1
0
0-1
0-3
Dominici et al. (2006,088398)
Zanobetti & Schwartz (2006,09019
Peel etal. (2005,056305)*
Host etal. (2008.155852)
Delfino etal. (2009.191994)
Ito (2003.042856)
Peel etal. (2005.056305)*
90 Counties, US
Boston, MA
Atlanta, GA
6 Cities, France
6 Counties, S.CA
Detroit, Ml
Atlanta, GA
Host et al. (2008,155852) 6 Cities, France
Ito (2003,042856) Detroit, Ml
Medina-Ramon et al. (2006,087721) 36 Cities, US
Peel etal. (2005.056305)*
Zanobetti (2003.043119)
Ito (2003.042856)
Atlanta, GA
14 Cities, US
Detroit, Ml
1-4
0-19
0-19
0-4
0-4
5-19
1-4
0-19
0-19
Pneumonia
Pneumonia
LRI
LRI
0
1
2
0-2 DL
0
0-1
0-3
0-3
0-1
0-1
0-1
0
0-3
0-3
0-1
0
0
1
0
1
0-3
0-1 SDL
0-3
0-1 SDL
0-1
0
65+
65+
65+
65+
65+
65+
All
All
All
All
All
65+
All
All
All
65+
65+
65+
65+
65+
All
All
All
All
65+
65+
RTI
RTI
RTI
RTI
Pneumonia
Pneumonia
URI
Pneumonia
RTI
Pneumonia
Bronchiolitis
Pneumonia
Pneumonia
URI
RTI
Pneumonia
Pneumonia
Pneumonia
Pneumonia
Pneumonia
URI
URI
Pneumonia
Pneumonia
Pneumonia
Pneumonia
PM2.5
Bronchitis, Bronchiolitis
Pneumonia —
Pneumonia
RTI
RTI
Pneumonia, Acute Bronchitis
Pneumonia, Acute Bronchitis
LRI
LRI
RTI
i Outpatient
i— Inpatient
-Wildfire
-Wildfire
Wildfire
- Outpatient
•—Inpatient
PM1025
PM10
PM2.
-Wildfire
-Wildfire
'•Warm
i*Warm
i»Cold
PM1(>
PM,<
*ED Visits
DL = Distributed Lag
Figure 6-14.
-10
20 30
Excess Risk (%)
Excess risks estimates per 10 ug/m increase in 24-h avg PM2.6, PMi0.2.5, and PMi0
for respiratory infection ED visits* and HAs. Studies represented in the
figure include all multicity studies as well as single-city studies conducted in the
U.S.
December 2009
6-145
-------
Table 6-14. PM concentrations in epidemiologic studies of respiratory diseases.
Study
Location
Mean Concentration
(ug/m3)
Upper Percentile
concentrations (ug/m )
PMts
Andersen et al. (2007, 093201)
Barnett et al. (2005, 087394)
Bell et al. (2008, 156266)
Chardonetal. (2007, 091308)
Chen et al. (2004, 087262; 2005, 087555)
Chimonas and Gessner (2007, 093261)
Delfino et al. (2009, 191994)
Dominici et al. (2006, 088398)
Funq et al. (2006, 089789)
Halonen et al. (2008, 189507)
Host et al. (2008, 155852)
Itoetal. (2007, 091262)
Lin et al. (2002, 026067)
Lin et al. (2005, 087828)
Moolqavkar (2003, 051316)
New York State DOH (2006, 090132)
Peel et al. (2005, 056305)
Sinclair and Tolsma (2004, 088696)
Sheppard et al. (2003, 042826)
Slaughter etal. (2005, 073854)
Tolbertetal. (2007, 090316)
Yanq et al. (2004, 087488)
Zanobetti and Schwartz (2006, 090195)
Copenhagen, Denmark
7 Cities Australia, NZ
202 U.S. counties
Paris, France
Vancouver, Canada
Anchoraae, AK
6 counties, CA
204 U.S. counties
Vancouver, Canada
Helsinki, Finland
6 Cities France
New York, NY
Toronto Canada
Ontario, Canada
LosAnaeles, CA
Bronx/Manhattan
Atlanta, GA
Atlanta, GA
Seattle, WA
Spokane, WA
Atlanta, GA
Vancouver, Canada
112 U.S. cities
10
8.1-11
12.92
14.7
7.7
6.1
18.4-32.7
13.4
7.72
NR; Median = 9.5
13.8-18.8
All vr: 15.1
17.99
9.59
22 (median)
15.0/16.7
19.2
17.62
16.7
NR
17.1
7.7
11.1 (Median)
99th: 28
Max: 29.3-122.8
98th: 34. 16
75th: 18.2
Max: 32
Max: 69.8
45.3-76.1 (mean durina wildfire period)
75th: 15.2
Max: 32
Max: 69.5
95th: 25.0-33.0
All vr: 95th: 32
Max: 89.59
Max: 73
Max: 86
NR
90th: 32.3; 98th: 39.8
NR
98th: 46.6
Max: 20.2 (usina 90% of concentrations)
90th: 28.8; 98th: 38.7
Max: 32.0
95th: 26.31
PMw-u
Chen et al. (2004, 087262; 2005, 087555)
Funa et al. (2006, 089789)
Halonen et al. (2008, 189507)
Hostetal. (2008, 155852)
Lin et al. (2002, 026067)
Lin et al. (2005, 087828)
New York State DOH
Peel et al. (2005, 056305)
Pena et al. (2008, 156850)
Sinclair and Tolsma (2004, 088696)
Sheppard et al. (2003, 042826)
Slaughter etal. (2005, 073854)
Tolbertetal. (2007, 090316)
Yana et al. (2004, 087488)
Vancouver, Canada
Vancouver, Canada
Helsinki, Finland
6 Cities France
Toronto, Canada
Ontario, Canada
Bronx/Manhattan
Atlanta, GA
108 U.S. counties
Atlanta, GA
Seattle, WA
Spokane, WA
Atlanta, GA
Vancouver, Canada
5.6
5.6
NR; Median: 9.9
7.0-11.0
12.17
10.86
7.69/7.10
9.7
NR; Median: 9.8
9.67
16.2
NR
9
7.7
Max: 24.6
Max: 27.07
Max: 101.4
95th: 12.5-21.0
Max: 68.00
Max: 45
NR
90th: 16.2
75th: 15.0
NR
Max: 88
NR
90th: 15.1: Max: 50.3
Max: 24.6
PM10
Andersen et al. (2007, 093201)
Barnett et al. (2005, 087394)
Chardonetal. (2007, 091308)
Chen et al. (2004, 087262; 2005, 087555)
Chimonas and Gessner (2007, 093261)
Funa et al. (2005, 093262)
Funa et al. (2006, 089789)
Gordian and Choudhurv (2003, 054842)
Jaffeetal. (2003, 041957)
Copenhaaen, Denmark
7 Cities, Australia, NZ
Paris, France
Vancouver, Canada
Anchoraae, AK
Ontario, Canada
Vancouver, Canada
Anchoraae, AK
Cincinnati, OH
25/24
16.5-20.6
23
13.3
27.6
38
13.3
36.11
43
75th: 30 /99th: 72
Max: 50.2-156.3
Max: 97.3
Max: 52.2
Max: 421
Max: 248
Max: 52. 17
Max: 210.0
Max: 90
December 2009
6-146
-------
Study
Location
Mean Concentration
(ug/m3)
Upper Percentile
concentrations (ug/m
Jalaludin et al (2004 056595)
Lin et al. (2002, 026067)
Lin et al. (2005, 087828)
Luainaah et al. (2005, 057327)
Medina-Ramon et al. (2006, 087721)
Moolaavkar (2003, 051316)
Moolaavkar (2003, 051316)
Peel et al. (2005, 056305)
Sinclair and Tolsma (2004, 088696)
Slaughter etal. (2005, 073854)
Tolbertetal. (2007, 090316)
Ulirschetal. (2007, 091332)
Yana et al. (2004, 087488)
Zanobetti (2003, 043119);Sametetal. (2000, 010269)
Svdnev Australia
Toronto, Canada
Ontario, Canada
Ontario, Canada
36 U.S. Cities
LosAnaeles, CA
Cook Countv, IL
Atlanta, GA
Atlanta, GA
Spokane, WA
Atlanta, GA
Idaho
Vancouver, Canada
14 U.S. Cities
228
30.16
20.41
50.6
15.9-44.0
22 (median)
35 (median)
27.9
29.03
NR
26.6
23.2
13.3
24.4-45.3
May 44 9
Max: 116.20
Max: 73
Max: 349
NR
Max: 86
Max: 365
Max: 44.7
NR
Max: 41.9 (using 90% of concentrations)
90th: 42.8
Max: 183.0
Max: 52.2
Max 94.8-605.8
UFP
Andersen et al. (2008,1896511
Copenhagen, Denmark Mean particles/cm : 6847
99th: 19,895 particles/cm3
Halonen et al. (2008,1895071
NR: Median particles/cm3: 8,203 Max: 50,990 particles/cm3
6.3.8.5. Copollutant Models
Some studies have investigated potential confounding by copollutants through the application
of multipollutant models (Figure 6-15). Several Canadian studies of respiratory hospital admissions
reported larger effects for PMi0 2 5 compared to PM2 5 that were robust to adjustment for gaseous
pollutants (Chen et al., 2005, 087555; Lin et al., 2002, 026067; Yang et al., 2004, 087488). The
COPD associations between PM2.5 and PMi0_2.5 reported by Chen et al. (2004, 087262) remained
positive but were diminished slightly after adjustment for NO2. The associations reported by Ito et al.
(2003, 042856) of PM25 and PMi0_2.5 with pneumonia hospital admissions remained after adjustment
for gases, while the association of PM10_25 with COPD admissions was not robust to adjustment for
O3. Associations reported by Burnett et al. (1997, 084194). Moolgavkar et al. (2003, 042864) and
Delfino et al. (1998, 093624) were not consistently robust to adjustment for gaseous copollutants. In
the MCAPS study, the effect of PM2.5 was robust to adjustment for PMi0_2.5, while the PMi0_2.5 effect
on respiratory admissions was diminished after adjustment for PM25 (Peng et al., 2008, 156850).
Effect estimates for PMi0 were robust to adjustment for gases in several recent studies (Andersen et
al., 2007, 093201; Tolbert et al., 2007, 090316; Ulirsch et al., 2007, 091332).
Multiple pollutant analyses for other size fractions and components have been conducted in a
some additional studies. PMi0 associations with respiratory disease did not change in models also
containing total PNC, nor did the association of ACP diminish after adjustment for UFP
concentration(Andersen et al., 2008, 189651). Peng et al. (2009, 191998) reports an OCM effect that
was robust to adjustment for other components while the associations with Ni, V, and EC were
somewhat diminished in models containing multiple components.
Inconsistency across these study findings is likely due to differences in the correlation
structure among pollutants as well as differing degrees of exposure measurement error.
December 2009
6-147
-------
Study
Peng etal. (2008, 1568501
Lin etal. (2005, 0878281
Chen etal. (2005, 0875551
Chen etal. (2004, 0872621
Ito (2003, 0428561
Moolgavkar (2003, 0428641
Sheppard (2003, 0428261
Lin etal. (2002. 0260671
Delfinoetal. (1998, 0936241
Burnett etal. (1997,0841941
Thurstonetal. (1994, 0439211
Peng etal. (2008, 1568501
Lin etal. (2005,0878281
Chen etal. (2005, 0875551
Chen etal. (2004, 0872621
Yang etal. (2004,0874881
Ito (2003, 0428561
Lin etal. (2002, 0260671
Burnett etal. (1997,0841941
Outcome
Respiratory Disease
Respiratory Infection
Respiratory Disease
COPD
Pneumonia
COPD
Asthma
Asthma, Boys
Asthma, Girls
Asthma, Children
Respiratory Disease
Respiratory Disease
Respiratory Disease
Respiratory Infection
Respiratory Disease
COPD
Respiratory Disease
COPD
Pneumonia
Asthma, Boys
Asthma, Girls
Asthma, Children
Respiratory Disease
Pollutant Effect Estimate (95% Cl)
P M2 5 | PM2 5 Adjusted for Gases and Other Size Fractions
PM2.5+PM10.2.5 r*-
PM.T .
pM.-j-rn Qn. |\p. n. ,, ,
PM-- , •
pM.-*rru.n,4.|\|n.4.Qn. , .
PM-- i . V
PM--+PM10-- i •
PM-+rr> ,
PM--+n. .
PM-rJ-SO- ' •
PMls ' •
PM-.4-n, • .
PM25+S02 «
PM25+N02 .
PM25+CO , .
PM25 ,-^
PM25+N02 _*_
PM2.5 , ——
PM25+CO , _^_
PM-- < . i
PM- : ,
piuu+rn so, MO, n, < ,
PM- ' .
PM"+<~), ^
PM2.5 -+-
PM25+03 .
PM25+N02 .
PM25+S02 !_._
PM25 .
PM,, .
PM,,+n, , ,
PM1lM5 Adjusted for Gases and other Size Fractions
PM1M.5+PM2.5 *
PM • V/
PH +rn cr> Mn n ' •
PM +rn+n +MH +°n ' •
PM ' • V7
rMm.^ , • ^/
PM +SO i •
pMr._+QO so- [\JQ- 0- '/ • >
PMr - -+SO- •
pjyi^^+lvjO^ «
p|y||C^r+QQ . 9
PM,--- •
PM|^25+S02 •
p[y||C^2r+[\j02 t
pM,-,rJ.pQ -
PMiri-r • ' *JI
p|y|r2r+rn gn2 |\|n2 n. ,
PMW25+03 .
PM1M5+N02 — i— «
PMW25+S02 .
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 // 1
-8-40 4 8 12 16 20 24 28 32 36 60
Excess Risk Estimate
Figure 6-15.
Excess risk estimates per 10 ug/m3 increase in 24-h avg PM2.6, and PMi0.2.sfor
respiratory disease ED visits or HAs, adjusted for co-pollutants.
December 2009
6-148
-------
6.3.9. Respiratory Mortality
An evaluation of studies that examined the association between short-term exposure to PM2.5
and PMio_2.5 and mortality provides additional evidence for PM-related respiratory health effects.
Although the primary analysis in the majority of mortality studies evaluated consists of an
examination of the relationship between PM2 5 or PMio_2.s and all-cause (nonaccidental) mortality,
some studies have examined associations with cause-specific mortality including respiratory-related
mortality.
Multicity mortality studies that examine the PM-respiratory mortality relationship on a
national scale - Franklin et al. (2007, 091257): 27 U.S. cities and Zanobetti and Schwartz (2009,
188462): 112 U.S. cities - have found consistent positive associations between short-term exposure
to PM2.5 and respiratory mortality of approximately 1.68% per 10 (ig/m3 at lag 0-1 (Section 6.5). The
associations observed on a national scale are consistent with those presented by Ostro et al. (2006,
087991) in a study that examined the PM2.5-mortality relationship in nine California counties (2.2%
[95% CI: 0.6-3.9] per 10 (ig/m3). An evaluation of studies that examined additional lag structures of
associations found smaller respiratory mortality effect estimates when using the average of lag days
1 and 2 (1.01% [95% CI: -0.03 to 2.05] per 10 (ig/m3) (Franklin et al., 2008, 155779). and
associations consistent with those observed at lag 0-1 when examining single-day lags, specifically
lag 1 (1.78% [95% CI: 0.2-3.36]). Although the overall effect estimates reported in the multicity
studies evaluated are consistently positive, it should be noted that a large degree of variability exists
between cities when examining city-specific effect estimates potentially due to differences between
cities and regional differences in PM2.5 composition (Figure 6-25). Only a limited number of studies
that examined the PM2.5-mortality relationship have conducted analyses of potential confounders,
such as gaseous copollutants, and none examined the effect of copollutants on PM2 5 respiratory
mortality risk estimates. Although the recently evaluated multicity studies did not extensively
examine whether PM2 5 mortality risk estimates are confounded by gaseous pollutants, evidence
from the limited number of single-city studies evaluated in the 2004 PM AQCD (U.S. EPA, 2004,
056905) suggest that gaseous copollutants do not confound the PM2 5-respiratory mortality
association. This is further supported by studies that examined the PMi0-mortality relationship in
both the 2004 PM AQCD (U.S. EPA, 2004, 056905) and this review. Overall, the respiratory PM2.5
effects observed in the new studies evaluated were larger, but less precise than those reported for all-
cause (nonaccidental) mortality (Section 6.5), and are consistent with the effect estimates observed
in the single- and multicity studies evaluated in the 2004 PM AQCD (U.S. EPA, 2004, 056905).
Zanobetti and Schwartz (2009, 188462) also examined PMi0_2.5 mortality associations in 47
U.S. cities and found evidence for respiratory mortality effects (1.16% [95% CI: 0.43-1.89] per
10 (ig/m3 at lag 0-1), which are somewhat larger than those reported for all-cause (nonaccidental)
mortality (0.46% [95% CI: 0.21-0.671] per 10 ug/m3). In addition, Zanobetti and Schwartz (2009,
188462) reported seasonal (i.e., larger in spring) and regional differences in PMi0_2.s respiratory
mortality risk estimates. However, single-city studies conducted in Atlanta, GA (Klemm et al., 2004,
056585) and Vancouver, Canada ((Villeneuve et al., 2003, 055051) reported no associations between
short-term exposure to PMi0_2.s and respiratory mortality. The difference in the results observed
between the multi- and single-city studies could be due to a variety of factors including differences
between cities and compositional differences in PMi0_2.5 across regions (Figure 6-30). Only a small
number of studies have examined potential confounding by gaseous copollutants or the influence of
model specification on PMi0_2 5 mortality risk estimates, but the effects are relatively consistent with
those studies evaluated in the 2004 PM AQCD (U.S. EPA, 2004, 056905).
6.3.10. Summary and Causal Determinations
6.3.10.1. PM2.5
Several studies of the effect of PM25 on hospital admissions for respiratory diseases reviewed
in the 2004 AQCD (U.S. EPA, 2004, 056905) reported positive associations for several diseases. The
2004 AQCD (U.S. EPA, 2004, 056905) presented limited epidemiologic evidence of PM25 being
associated with respiratory symptoms (including cough, phlegm, difficulty breathing, and
bronchodilator use); observations for PM2 5 were positive, with slightly larger effects for PM2 5 than
December 2009 6-149
-------
for PM10. In addition, mortality studies reported relatively higher PM2.5 risk estimates for respiratory-
related mortality compared to all-cause (nonaccidental) mortality. Controlled human exposure
studies did not provide support for effects of CAPs on respiratory symptoms. Small decrements in
peak flow for both PM25 and PMi0 in asthmatics and nonasthmatics were reported in epidemiologic
studies included in the 2004 PM AQCD (U.S. EPA, 2004, 056905). whereas controlled human
exposure and animal toxicological studies reported few or no effects on pulmonary function with
inhalation of CAPs. In addition, the 2004 PM AQCD (U.S. EPA, 2004, 056905) presented a number
of controlled human exposure and toxicological studies that reported mild pulmonary inflammation
following exposure to PM2 5 CAPs and DE or DE particles, as well as ROFA or other
metal-containing PM in animals. The 2004 PM AQCD (U.S. EPA, 2004, 056905) described
controlled human exposure studies showing increases in allergic responses among previously
sensitized atopic subjects after short-term exposure to DE particles. These observations were
supported by many toxicological studies that added to existing evidence demonstrating that various
types of PM could promote allergic disease and exacerbate allergic asthma in animal models.
Toxicological studies also indicated that PM2.5 increased susceptibility to respiratory infection.
Overall, in recent studies PM2 5 effects on respiratory hospitalizations and ED visits have been
consistently observed. Most effect estimates were in the range of-1-4% and were observed in areas
with mean 24-h PM25 concentrations between 6.1 and 22 (ig/m3. Further, recent studies have focused
on increasingly specific disease endpoints such as asthma, COPD, and respiratory infection. The
strongest recent evidence of an association comes from large multicity studies of COPD, respiratory
tract infection, and all respiratory diseases among Medicare recipients (>65 yr) (Bell et al., 2008,
156266; Dominici et al., 2006, 088398). Studies of children have also found evidence of an effect of
PM2 5 on hospitalization for all respiratory diseases, including asthma and respiratory infection.
However, many of these effect estimates are imprecise, their magnitude and statistical significance
are sensitive to choice of lag, and some null associations were observed. Although the association of
PM25 with pediatric asthma was not examined specifically, it is noteworthy that one of the strongest
associations observed in the Atlanta-based SOPHIA study was between PMi0 and pediatric asthma
visits; PM25 makes up a large proportion of PMi0 in Atlanta (Peel et al., 2005, 056305). Positive
associations between PM25 (or PMi0) and hospital admissions for respiratory infection ( Figure 6-14)
are supported by animal toxicological studies which add to previous findings of increased
susceptibility to infection following exposure to PM2 5. These include studies demonstrating reduced
clearance of bacteria (Pseudomonas, Listeria) or enhanced pathogenesis of viruses (influenza, RSV)
after exposure to DE or ROFA.
Epidemiologic studies that examined the association between PM2 5 and mortality provide
additional evidence for PM25-related respiratory effects (Section 6.3.9). The multicity studies
evaluated found consistent, precise positive associations between short-term exposure to PM2 5 and
respiratory mortality ranging from 1.67 to 2.20% increases at mean 24-h PM25 avg concentrations
above 13 (ig/m3. Although only a limited number of studies examined potential confounders of the
PM2 5-respiratory mortality relationship, the studies evaluated in both this review and the 2004 PM
AQCD (U.S. EPA, 2004, 056905) support an association between short-term exposure to PM25 and
respiratory mortality.
Epidemiologic studies of asthmatic children have observed increases in respiratory symptoms
and asthma medication use associated with higher PM2 5 or PM10 concentrations. Associations with
respiratory symptoms and medication use are less consistent among asthmatic adults, and there is no
evidence to suggest an association between respiratory symptoms with PM2 5 among healthy
individuals. In addition, respiratory symptoms have not been reported following controlled
exposures to PM2 5 among healthy or health-compromised adults (Section 6.3.1.2).
Although more recent epidemiologic studies of pulmonary function and PM25 have yielded
somewhat inconsistent results, the majority of studies have found an association between PM25
concentration and FEVi, PEF, and/or MMEF. In asthmatic children, a 10 ug/m3 increase in PM25 is
associated with a decrease in FEVi ranging from 1-3.4% (Section 6.3.2.1). A limited number of
controlled human exposure studies have reported small decreases in arterial oxygen saturation and
MMEF following exposure to PM2 5 CAPs with more pronounced effects observed in healthy adults
than in asthmatics or older adults with COPD (Section 6.3.2.2). In toxicological studies, changes in
pulmonary function have been observed in healthy and compromised rodents after inhalation
exposures to CAPs from a variety of locations or DE. A role for the PM fraction of DE is supported
by altered pulmonary function in healthy rats after IT instillation of DE particles (Section 6.3.2.3).
December 2009 6-150
-------
Several lines of evidence suggest that PM2.5 promotes and exacerbates allergic disease, which
often underlies asthma (Section 6.3.6). Although epidemiologic studies examining specific allergic
outcomes and short-term exposure to PM are relatively rare, the available studies, conducted
primarily in Europe, positively associate PM2.5 and PMio with allergic rhinitis or hay fever and skin
prick reactivity to allergens. Short-term exposure to DE particles in controlled human exposure
studies has been shown to increase the allergic response among previously sensitized atopic subjects,
as well as induce de novo sensitization to an antigen. Toxicological studies continue to provide
evidence that PM2.5, in the form of CAPs, resuspended DE particles, or DE, but not wood smoke,
spurs and intensifies allergic responses in rodents. Proposed mechanisms for these effects include
mediation by neurotrophins and oxidative stress, and one study demonstrated that effects were
mediated at the epigenetic level (Liu et al., 2008, 156709).
A large body of evidence, primarily from toxicological studies, indicates that various forms of
PM induce oxidative stress, pulmonary injury, and inflammation. Notably, CAPs from a variety of
locations induce inflammatory responses in rodent models, although this generally requires multiday
exposures. The toxicology findings are consistent with several recent epidemiologic studies of PM2.5
and the inflammatory marker eNO, which reported statistically significant, positive effect estimates
with some inconsistency in the lag times and use of medication. In asthmatic children, a 10 ug/m3
increase in PM25 is associated with an increase in eNO ranging from 0.46 to 6.99 ppb. Several new
controlled human exposure studies report traffic or DE-induced increases in markers of inflammation
(e.g., neutrophils and IL-8) in BALF from healthy adults. Recent studies have provided additional
evidence in support of a pulmonary oxidative response to DE in humans, including induction of
redox-sensitive transcription factors and increased urate and GSH concentrations in nasal lavage. In
addition, exposure to wood smoke has recently been demonstrated to increase the levels of eNO and
malondialdehyde in breath condensate of healthy adults (Barregard et al., 2008, 155675).
Preliminary findings indicate little to no pulmonary injury in humans following controlled exposures
to PM25 urban traffic particles or DE, in contrast to a number of toxicological studies demonstrating
injury with CAPs or DE (Sections 6.3.5.2 and 6.3.5.3, respectively).
Recent studies have reported associations of hospital admissions, ED or urgent care visits for
several respiratory diseases with PM2 5 components and sources including Ni, V, OC and EC, wood
smoke and traffic emissions, in studies of both children and adults. Delfmo et al. (2003, 090941;
2006, 090745) found positive associations between EC and OC components of PM and asthma
symptoms and between EC and eNO. Particle composition and/or source also appears to heavily
influence the increase in markers of pulmonary inflammation demonstrated in studies of controlled
human exposures to PM2 5. For example, whereas exposures to PM2 5 CAPs from Chapel Hill, NC
have been shown to increase BALF neutrophils in healthy adults, no such effects have been observed
in similar studies conducted in Los Angeles. In addition, differential inflammatory responses have
been observed following bronchial instillation of particles collected at different times or from
different areas (Section 6.3.3.2). One new study found that the increased airway neutrophils
previously observed by Ohio et al. (2000, 012140) in human volunteers after Chapel Hill CAPs
exposure could be largely attributed to the content of sulfate, Fe, and Se in the soluble fraction
(Huang etal. 2003. 0873771
In summary, new evidence of ED visits and hospital admissions builds upon the positive and
statistically significant evidence presented in the 2004 PM AQCD to support a consistent association
with ambient concentrations of PM25 Most effect estimates with respiratory hospitalizations and ED
visits were in the range of-1-4% and were observed in areas with mean 24-h PM25 concentrations
between 6.1 and 22 ug/m3. The evidence for PM2 5-induced respiratory effects is strengthened by
similar hospital admissions and ED visit associations for PMi0, along with the consistent positive
associations observed between PM2 5 and respiratory mortality in multicity studies. Panel studies also
indicate associations with PM2 5 and respiratory symptoms, pulmonary function, and pulmonary
inflammation among asthmatic children. Further support for these observations is provided by recent
controlled human exposure studies in adults demonstrating increased markers of pulmonary
inflammation following DE and other traffic-related exposures, oxidative responses to DE and wood
smoke, and exacerbations of allergic responses and allergic sensitization following exposure to DE
particles. Although not consistent across studies, some controlled human exposure studies have
reported small decrements in various measures of pulmonary function following exposures to PM25.
Numerous toxicological studies demonstrating a wide range of responses provide biological
plausibility for the associations between PM2 5 and respiratory morbidity observed in epidemiologic
studies. Altered pulmonary function, mild pulmonary inflammation and injury, oxidative responses,
December 2009 6-151
-------
AHR in allergic and non-allergic animals, exacerbations of allergic responses and increased
susceptibility to infections were observed in a large number of studies involving exposure to CAPs,
DE, other traffic-related PM, and wood smoke. The evidence for an effect of PM2.5 on respiratory
outcomes is somewhat restricted by limited coherence between some of the findings from
epidemiologic and controlled human exposure studies for the specific health outcomes reported and
the sub-populations in which those health outcomes occur. For instance, although there is evidence
for respiratory symptoms among asthmatic children in epidemiologic panel studies, the studies of
hospital admissions and ED visits provide more evidence for effects from COPD and respiratory
infections than for asthma. Additionally, controlled human exposure studies report greater effects in
healthy adults when compared to asthmatics or those suffering from COPD. Finally, there is limited
information which could explain the relationship between the clinical and subclinical respiratory
outcomes observed and the magnitude of the PM2 5-respiratory mortality associations reported.
Therefore, the evidence is sufficient to conclude that a C3USal relationship JS likely to exist
between short-term PM25 exposures and respiratory effects.
6.3.10.2. PM10.2.5
The 2004 PM AQCD (U.S. EPA, 2004, 056905) presented the results from several
epidemiologic studies of respiratory symptoms and PMi0_2.5, which provided limited evidence for
cough and effects on morning PEF. Toxicology data for PMi0_2.5 were extremely limited, and there
were no controlled human exposure studies presented in the 2004 PM AQCD (U.S. EPA, 2004,
056905) that evaluated the effect of PM10_2.5 on respiratory symptoms, pulmonary function, or
inflammation. Epidemiologic studies of the effect of PMi0_2.s on hospitalizations or ED visits for
respiratory diseases (i.e., pneumonia, COPD and respiratory diseases combined) reviewed in the
2004 AQCD (U.S. EPA, 2004, 056905) reported positive associations. Additionally, the few
mortality studies that examined cause-specific mortality suggested somewhat larger risk estimates
for respiratory mortality compared to all-cause (nonaccidental) mortality.
Several new studies report associations between PMi0_2.s and respiratory hospitalizations with
the most consistent evidence among children (Figure 6-10 through Figure 6-14), however, effect
estimates are imprecise. Although a number of studies provide evidence of respiratory effects in
older adults, a recent analysis of MCAPS data reports that weak associations of PMi0_2.5 with
respiratory hospitalizations are further diminished after adjustment for PM2 5. It is not clear that
PMio_2.5 estimates across all populations and regions are confounded by PM2 5. An examination of
PM10_2.5 mortality associations on a national scale found a strong association between PM10_2.5 and
respiratory mortality, but this association varied when examining city-specific risk estimates
(Zanobetti and Schwartz, 2009, 188462). The regional variability in PMi0_2.5 mortality risk estimates
is further confirmed by the negative associations reported in the single-city studies evaluated.
However, there is greater spatial heterogeneity in PMi0_2.5 compared to PM2 5 and consequently
greater potential for exposure measurement error in epidemiologic studies relying on central site
monitors. This exposure measurement error may bias effect estimates toward the null and could
explain some of the regional variability in the observed associations between PMi0_2.5 and respiratory
morbidity and mortality.
Mar et al. (2004, 057309) provide evidence for an association with increased respiratory
symptoms in asthmatic children, but not asthmatic adults. Consistent with this, controlled human
exposures to PMi0_2.5 have not been observed to affect lung function or respiratory symptoms in
healthy or asthmatic adults. However, increases in markers of pulmonary inflammation have been
demonstrated in healthy volunteers. In these studies, an increase in neutrophils in BALF or induced
sputum was observed, with additional evidence of alveolar macrophage activation associated with
biological components of PMi0_2.5 (i.e., endotoxin). Toxicological studies using inhalation exposures
are still lacking, but pulmonary injury and inflammation have been observed in animals after IT
instillation exposure and both rural and urban PMi0_2.5 have induced these responses. In some cases,
PM10_2.5 from urban air was more potent than PM25 (Section 6.3.3.3). PM10_2.5 respiratory effects may
be due to components other than endotoxin (Wegesser and Last, 2008, 190506).
Overall, the most compelling new evidence comes from a number of recent epidemiology
studies conducted in Canada and France showing significant associations between respiratory ED
visits or hospitalization and short-term exposure to PMi0_2.5. Effects have been observed in areas
where the mean 24-h avg PMi0_2.5 concentrations ranged from 7.4 to 13.0 ug/m3. The strongest
relationships were observed among children, whereas studies of adults and older adults show less
December 2009 6-152
-------
consistent evidence of an association. While controlled human exposure studies have not observed
an effect on lung function or respiratory symptoms in healthy or asthmatic adults in response to
exposure to PMi0_2.5, healthy volunteers have exhibited increases in markers of pulmonary
inflammation. Toxicological studies using inhalation exposures are still lacking, but pulmonary
injury has been observed in animals after IT instillation exposure to both rural and urban PMi0_2.5,
which may not be entirely attributed to endotoxin. Overall, epidemiologic studies, along with the
limited number of controlled human exposure and toxicological studies that examined PMi0_2.5 and
respiratory outcomes, provide evidence that is Suggestive Of 3 C3USal relationship between
short-term PM1025 exposures and respiratory effects.
6.3.10.3. UFPs
The 2004 PM AQCD (U.S. EPA, 2004, 056905) included a few epidemiologic and controlled
human exposure studies that examined the effect of UFPs on respiratory morbidity. Collectively
these studies provided limited evidence of an association between UFPs and respiratory symptoms,
medication use, inflammation, and decreased pulmonary function. Evidence from toxicological
studies presented in the 2004 AQCD, although limited, suggested that exposure via inhalation to
high concentrations of UF TiO2 may increase pulmonary inflammation in healthy rodents. Since the
publication of the 2004 AQCD there has been an increased focus among the scientific community on
gaining a better understanding of the potential health effects associated with exposure to UFPs
(U.S. EPA, 2004, 056905). A number of recent controlled human exposure and toxicological studies
have evaluated respiratory responses following exposures to UF CAPs, model particles, and fresh
diesel or gasoline exhaust. While DE contains both PM2.5 and UFPs, the MMAD is typically
< 100 nm, and therefore the results of these studies may be used to support findings from studies
utilizing other sources of UFP.
UFPs were associated with incident wheezing symptoms among infants (<1 yr) in a study
conducted in Copenhagen, Denmark, where the mean UFP number concentration was 8,092
particles/cm3, though this association did not persist for children between ages 1-3 yr (Andersen et
al., 2008, 096150). Recent epidemiologic studies conducted in Copenhagen, Denmark and Helsinki,
Finland, reported associations between UFPs and hospital admissions or ED visits for respiratory
diseases, including childhood asthma and pneumonia in adults (Andersen et al., 2008, 189651;
Halonen et al., 2008, 189507). The median UFP number concentrations in Copenhagen and Helsinki
were 6,243 particles/cm3 and 8,203 particles/cm3, respectively. Associations between UFP and ED
visits for respiratory diseases were not observed in the Atlanta-based SOPHIA study, where the mean
UFP number concentration was 38,000 particles/cm3.
A single recent epidemiologic study has examined associations between UFP and pulmonary
function, and observed that asthmatic adults exhibited decreased lung function after exposure to
diesel traffic pollution in London (McCreanor et al., 2007, 092841). Two new controlled human
exposure studies have reported small decreases in pulmonary function among healthy adults
approximately following exposure to Los Angeles UF CAPs or UF EC (Gong et al., 2008, 156483;
Pietropaoli et al., 2004, 156025). Exposures to lower concentrations of UF CAPs from Chapel Hill,
NC did not result in any changes in pulmonary function (Samet et al., 2009, 191913). However,
while Gong et al. (2008, 156483) did not observe any effect of exposure to UF CAPs on markers of
pulmonary inflammation, Samet et al. (2009, 191913) reported an UF CAPs-induced increase in IL-
8 in BALF at 18 hours post-exposure. A limited number of controlled human exposure studies have
also demonstrated increases in the pulmonary inflammatory response following exposure to UF and
PM2.5 from DE, which may be enhanced by exposure to O3 (Section 6.3.3.2).
Altered pulmonary function and inflammation have also been observed in toxicological studies
of DE and UF model particles (Sections 6.3.2.3 and 6.3.3.3). In one rat model, pulmonary
inflammation was observed after exposure to UF CB at concentrations as low as 180 (ig/m3 (Harder
et al., 2005, 087371). However, inflammatory responses vary considerably depending on the animal
model, dose, test material, and exposure duration. In cases where pulmonary inflammation was not
observed, oxidative stress was often evident (Section 6.3.4.2). Oxidative stress is a major mechanism
by which PM may exert effects (Chapter 5), and some toxicological studies suggest that UFPs are
more potent than PM2.5, possibly due to a higher proportion of pro-oxidative OC and PAH content
and greater surface area with which to deliver these components.
The relationship between exposure to UFP and pulmonary injury has not been widely
examined. No association with pulmonary injury biomarkers was found for UFP in a European
December 2009 6-153
-------
multicity epidemiologic study (Timonen et al., 2004, 087915). In controlled human exposure
studies, UFP from wood smoke resulted in significantly increased markers of injury in healthy
adults, but this effect was not evident in COPD sufferers exposed to DE (Section 6.3.5.2). Exposure
of neonatal rats to UF iron-soot particles resulted in a significantly reduced rate of cell proliferation
in the proximal alveolar region, which suggests that postnatal lung development may be susceptible
to air pollution, consistent with impaired lung function growth observed in children (Pinkerton et
al., 2004, 087465). In contrast, no histopathological responses were evident in adult mice exposed to
UF iron-soot particles (Last et al., 2004, 097334). Some toxicological studies have reported
pulmonary injury after inhalation of DE or gasoline exhaust (Section 6.3.5.3). In studies that
evaluated ambient PM size fractions from a variety of European and U.S. cities for relative toxicity
in rodents following IT instillation exposure, UFPs were generally less injurious than the larger size
fractions. However, the UF fraction of Montana coal fly ash induced greater injury and inflammation
than the PMio_2.5 fraction (Gilmour et al., 2004, 057420).
In rodent studies, UF CAPs appeared to be more potent than PM2.5 CAPs in inducing and
exacerbating allergic responses (Section 6.3.6.3). In addition to CAPs, UF CB or iron-soot particles,
but not particles from fresh gasoline exhaust, have been shown to induce or exacerbate allergic
responses in mice. Bacterial clearance appears unaffected by hardwood smoke or gasoline engine
exhaust. However, host defenses are impaired by DE, which has been shown to reduce bacterial
clearance, impair defenses against viral infection, and reduce thymus weight, indicating systemic
immunosuppression.
Several toxicological studies demonstrated oxidative, inflammatory, and allergic responses
following exposure to a number of different UFP types, including model particles (i.e., CB, iron-soot
particles), CAPs, and DE. Although the respiratory effects of controlled exposures to UFPs have not
been extensively examined in humans, two controlled human exposure studies have observed small
UFP-induced decreases in pulmonary function; however, no increases in respiratory symptoms have
been reported. In a limited number of studies, markers of pulmonary inflammation were increased
following controlled human exposures to UFP, which has been most consistently observed in studies
using fresh DE. In both controlled human exposure and animal toxicological studies using fresh DE,
the relative contributions of gaseous copollutants to the observed effects remain unresolved.
However, similar effects are reported using resuspended DE particles, and although not UFPs, these
particles can be assumed to have similar composition. A limited number of epidemiologic studies
have provided some evidence of an association between short-term exposure to UFPs and respiratory
symptoms, as well as asthma hospitalizations. However, the interpretation of these findings is
difficult due to the spatial variability of UFPs. Thus, the current collective evidence is SUgQBStJVB
of a causal relationship between short-term UFP exposure and respiratory effects.
6.4. Central Nervous System Effects
While evidence of an effect of PM on the CNS was not presented in the 2004 PM AQCD
(U.S. EPA, 2004, 056905). a limited number of recent epidemiologic, controlled human exposure
and toxicological studies provide some evidence that exposure to PM may be associated with
changes in neurological function. The majority of studies included in this section are of short-term
exposure, however, there are also a few studies of long-term exposure. As CNS effects of PM are a
newly emerging area, and since there are so few studies, all studies that evaluate CNS responses are
included in this section.
6.4.1. Epidemiologic Studies
Chen and Schwartz (2009, 179945) used extant data on CNS function from the Third National
Health and Nutrition Examination Survey (NHANES III) to characterize the association between
cognitive function in adults (ages 20-59 yr) and exposure to ambient air pollution. Three
computerized neurobehavioral tests were used: a simple reaction time test (SRTT), a basic measure
of visuomotor speed; a symbol digit substitution test (SDST) on coding ability; and a serial digit
learning test (SDLT) on attention and short-term memory. The authors used annual PMi0
concentrations to approximate the long-term exposure to ambient air pollution prior to the
December 2009 6-154
-------
NHANES-III examination. Increased PM10 levels were associated with reduced performance in all
three neurobehavioral tests, and were particularly strong for SDST and SDLT scores in models
adjusted for age and sex. However, after additional adjustment for race/ethnicity or SES, the
magnitudes of these associations were greatly diminished and largely null. It is possible that the
observed associations disappeared after adjustment for race/ethnicity and SES due to the potential
confounding by residential segregation of ethnic minorities and poorer people in areas with high
levels of ambient PMi0 concentrations.
Two additional epidemiologic studies evaluated the effect of ambient PM on the CNS
(Calderon-Garciduefias et al., 2008, 156317; Suglia et al, 2008, 157027). These studies examined
long-term exposure to non-specific PM indicators and are detailed in Annex E.
6.4.2. Controlled Human Exposure Studies
In a recent controlled human exposure study, Cruts et al. (2008, 156374) exposed 10 healthy
males (18-39 yr) to filtered air and dilute DE (300 ug/m3 PM) for 1 h using a randomized crossover
study design. Changes in brain activity were measured during and following exposure using
quantitative electroencephalography (QEEG). Exposure to DE was observed to significantly increase
the median power frequency (MPF) in the frontal cortex during exposure, as well as in the hour
following the completion of the exposure. While this study does provide some evidence of an acute
cortical stress response to DE, it is important to note that the QEEG findings are very nonspecific,
and could have been caused by factors other than diesel PM such as DE gases (e.g., CO, NO and
NO2) or the odor of the DE.
6.4.3. lexicological Studies
Evidence is mounting that the CNS may be a critical target of PM and that adverse health
effects may result from PM exposure. Whether these health effects are a direct or indirect effect of
PM has not yet been established. One hypothesis suggests that UFPs which deposit onto nasal
olfactory epithelium enter the CNS by axonal olfactory transport to the olfactory bulb and lead to a
cascade of effects involving inflammatory cytokines and ROS. An increased potential for
neurodegenerative processes may ensue. Evidence for translocation of UFPs to the olfactory bulb via
olfactory neurons is discussed in Chapter 4, but its relevance to CNS health effects is unknown.
Another hypothesis suggests that brain inflammation occurs secondarily to PM-mediated systemic
inflammation. Finally, it has been suggested that PM-stimulation of the ANS via respiratory tract
receptors results in inflammatory or other effects in the CNS. This is an emerging field with many
unknowns.
6.4.3.1. Urban Air
Calderon-Garciduenas et al. (2003, 156316) conducted a long-term observational study in
mongrel dogs from Mexico City and Tlaxcala. DNA damage and inflammation in the brain and
respiratory tract were evaluated in dogs living in Mexico City (exposed group) and dogs living in
Tlaxcala (control group). These cities are similar in altitude but differ in air pollutant levels.
Measurements of air pollutant levels were presented only for Mexico City, the more polluted city.
Statistically significant greater levels of apurinic/apyrimidinic sites (an indicator of DNA damage)
were observed in the olfactory bulbs and hippocampus of Mexico City dogs compared with controls.
These differences were not seen in other brain regions examined or in nasal respiratory epithelium.
In addition, Mexico City dogs demonstrated greater histopathological changes in the respiratory and
olfactory epithelium of the nasal cavity compared with controls. Immunohistochemical staining of
brain tissue from the Mexico City dogs demonstrated greater immunoreactivity for NF-KB, iNOS,
cyclooxygenase-2, glial fibrillatory acidic protein (GFAP), ApoE, amyloid precursor product and
p-amyloid compared with controls. These results are indicative of inflammation and stress protein
responses. This study has several limitations given that the dogs were of mixed breeds and of
variable ages and that there was no standardization of exposures or diets. However results suggest a
possible relationship between air pollution and brain inflammation.
December 2009 6-155
-------
6.4.3.2. CAPs
Several new inhalation studies have provided evidence of CNS effects due to ambient PM
exposures. In one study, Campbell et al. (2005, 087217) exposed OVA-sensitized BALB/c mice to
filtered air or near-highway Los Angeles CAPs (a 20-fold concentration of PM2.5+UFPs or UFPs
only; mean exposure concentration UFPs 282.5 ug/m3 and PM2.5 441.7 ug/m3) for 4 h/day and 5
days/wk over a 2-wk period. The animals were subsequently challenged with OVA to elicit an
allergic response in the lungs; brain tissue was obtained one day later. Exposure to CAPs, but not
filtered air, resulted in activation of the immune-related transcription factor NF-KB and upregulation
of the cytokines TNF-a, and IL-la in the brain, demonstrating pro-inflammatory responses that
could contribute to neurodegenerative disease. While this study demonstrates CAPs effects in an
allergic animal model, it is not known whether these responses also occur in non-allergic animals.
In a second study, control or OVA-sensitized and challenged Brown Norway rats were exposed
for 8 h to filtered air or PM2.5 CAPs (500 ug/m3) in Grand Rapids, MI (Sirivelu et al., 2006,
111151). Brain tissue was obtained 1 day later. CAPs exposure resulted in brain region-specific
modulation of neurotransmitters. In animals which were not pretreated with OVA, statistically
significant increases in norepinephrine were observed in the paraventricular nucleus and olfactory
bulb of CAPs-exposed rats compared with filtered air controls. In animals which were pretreated
with OVA, a statistically significant increase in dopamine was observed in the medial preoptic area
in CAPs-exposed rats compared with controls. Furthermore, exposure to CAPs resulted in a
statistically significant increase in serum corticosterone. These data suggest that the
hypothalamo-pituitary-adrenal axis (i.e., stress axis) may be activated by PM exposure, causing
aggravation of allergic airway disease. The authors discuss the possible role of the olfactory bulb in
mediating neuroendocrine control of autonomic activities involved in respiratory and cardiovascular
functions; however these relationships require clarification.
Pro-inflammatory responses were examined in a subchronic CAPs study involving normal
(C57BL/6J) and ApoE"" mice (Kleinman et al., 2008, 190074). Mice were exposed to filtered air or
to two concentrations of UF CAPs from a near-highway area of central Los Angeles (average of 30.4
and 114.2 ug/m3) for 5 h/day and 3 days/wk over a 6-wk period. Brain tissue was harvested one day
after the last exposure and cortical samples prepared. CAPs exposure resulted in activation of
transcription factors, with a dose-dependent increase observed for AP-1 and an increase in NF-KB
observed at the higher concentration. Increased levels of GFAP (representing activation of
astrocytes) and phosphorylated JNK (representing MAP kinase activation) were observed at the
lower but not higher concentration of CAPs. No changes were observed in levels of or activation of
the other MAP kinases p38 and ERK or of 1KB. These findings provide evidence that inhalation of
CAPs can lead to activation of cell signaling pathways involved in upregulation of pro-inflammatory
cytokine genes in the cortical region of the mouse brain.
In another study utilizing normal (C57BL/6) and ApoE"" mice, brain histopathology was
examined following a 4-month chronic exposure to PM2.5 CAPs from Tuxedo, NY (March, April or
May through September 2003) (Veronesi et al., 2005, 087481). The average PM25 exposure
concentration was 110 ug/m3. CAPs exposure resulted in a statistically significant decrease in
dopaminergic neurons, measured by tyrosine hydroxylase immunoreactivity, in the substantia nigra
of ApoE"7" mice but not in control mice. This population of neurons is targeted in neurodegenerative
diseases such as Parkinson's. Furthermore, a statistically significant increase in GFAP
immunoreactivity, a marker for astrocytes, was observed in the nucleus compacta of CAPs-exposed
ApoE"7" mice compared to air-exposed ApoE"7" mice. These results suggest that the ApoE"7" mice, a
genetic model involving increased oxidative stress, are susceptible to PM-induced
neurodegeneration. Evidence for brain oxidative stress has also been found in normal animals
following IT instillation of high concentrations of PM25 from Taiyuan, China (Liu and Meng, 2005,
088650) and of gasoline exhaust (Che et al., 2007, 096460) and following chronic exposure to
ROFA by intranasal instillation (Zanchi et al., 2008, 157173).
6.4.3.3. Diesel Exhaust
A recent study tested the effects of DE inhalation on spatial learning and memory function-
related gene expression in the hippocampus (Win-Shwe et al., 2008, 190146). Male BALB/c mice
were exposed to DE (148.86 ug/m3 PM) for 5 h/day and 5 day/wk over a 4-wk period. Particle size
was 26.21±1.50 nm and PNC was 1.92xl06 ± 6.18><104 particles/m3. Concentrations of gases were
December 2009 6-156
-------
3.27 ppm CO, 0.01 ppm SO2, 0.53 ppmNO?, 0.98 ppm NO and 0.07 ppm CO2. Half of the animals
were injected i.p. once per week with lipoteichoic acid (LTA), a bacterial cell wall component used
to induce systemic inflammation. The ability of the mice to perform spatial learning tasks was
examined the day after the final exposure to DE and on two subsequent days. Impaired acquisition of
spatial learning was observed in DE-exposed mice on the first day and on all three days in DE-
exposed mice that had also been treated with LTA. LTA by itself had no effect. Since the NMD A (a
type of neurotransmitter) receptors in the hippocampus play an important role in spatial learning
ability, mice were sacrificed and total RNA from hippocampus was extracted and analyzed for
expression of NMD A receptor subunits. DE exposure resulted in a statistically significant increase in
the expression of one subunit while the combined exposure to DE and LTA resulted in statistically
significant increases in the expression of three subunits compared with controls. The expression of
pro-inflammatory cytokines was also examined in the hippocampus. DE exposure resulted in a
statistically significant increase in TNF-a mRNA, while LTA exposure resulted in a statistically
significant increase IL-1(3 mRNA compared with controls. Neither exposure altered the expression of
HO-1. These results demonstrated that subchronic exposure to UF-rich DE resulted in impaired
spatial learning and altered expression of hippocampal genes involved in memory function and
inflammation. These responses were modulated by systemic inflammation.
6.4.3.4. Summary of lexicological Study Findings of CMS Effects
In summary, PM may produce adverse effects in the CNS by direct or indirect mechanisms
which are at present incompletely understood. Two recent short-term PM2.5 CAPs inhalation studies
demonstrated pro-inflammatory responses in the brain and brain region-specific modulation of
neurotransmitters and suggest the involvement of neuroimmunological pathways. One recent chronic
PM2.5 CAPs inhalation study demonstrated loss of dopaminergic neurons in the substantia nigra and
suggested that oxidative stress contributes to neurodegeneration. Veronesi et al. (2005, 087481) have
noted that the brain is very vulnerable to the oxidative stress induced by PM due to the brain's high
energy demands, low levels of endogenous free radical scavengers, and high content of lipids and
proteins. PM-mediated upregulation of inflammatory cytokines and mediators may also contribute to
neurodegeneration. In fact, a recent subchronic study involving UF CAPs demonstrated the
activation of cell signaling pathways associated with upregulation of pro-inflammatory cytokines in
brain cortical regions. Furthermore, a subchronic study involving UF-rich DE demonstrated impaired
spatial learning and altered expression of pro-inflammatory and neurotransmitter receptor genes in
the hippocampus. Further investigations are required to delineate mechanisms involved in these
responses.
6.4.4. Summary and Causal Determination
Recent animal toxicological studies involving acute or chronic CAPs exposure have
demonstrated pro-inflammatory responses in the brain, brain region-specific modulation of
neurotransmitters and loss of dopaminergic neurons in the substantia nigra (Campbell et al., 2005,
087217: Kleinman et al., 2008, 190074: Sirivelu et al., 2006, 111151: Veronesi et al., 2005,
087481). However, the mechanisms underlying these effects need to be delineated. A single
controlled human exposure study provides some evidence of an acute cortical stress response to DE,
though these findings are nonspecific and could have been caused by DE gases rather than DE
particles (Cruts et al., 2008, 156374). Similar consideration is warranted for the single animal
toxicological study involving DE which demonstrated impaired spatial learning and altered
expression of pro-inflammatory and neurotransmitter genes in the hippocampus following
subchronic exposure (Win-Shwe et al., 2008, 190146). The single epidemiology study that
examined CNS outcomes did not find associations between long-term exposure to PMi0 and
cognitive function in adults after adjustment for race/ethnicity or SES (Chen and Schwartz, 2009,
179945). Though the effect of ambient air pollution on CNS outcomes has recently begun to draw
more attention, the evidence for a PM-induced CNS effect is limited. While most available studies
have evaluated the effects of fine particle exposures, there is insufficient evidence to draw
conclusions regarding effects of specific PM size fractions. Overall, the evidence JS inadequate to
determine if a causal relationship exists between short-term exposures to PM2.5, PM™ 25,
or UFPs and CNS effects.
December 2009 6-157
-------
6.5. Mortality
The relationship between short-term exposure to PM and mortality has been extensively
addressed in previous PM assessments (U.S. EPA, 1982, 017610: 1996, 079380: 2004, 056905). A
positive association between PM concentration and mortality was consistently demonstrated across
studies cited in the 2004 PM AQCD (U.S. EPA, 2004, 056905): these results are summarized below
in Section 6.5.1. Numerous studies have been published since the previous review, including a
number of multicity analyses and many single-city studies. The current body of evidence examines
the association between short-term exposure to PM of various size fractions (i.e., PMi0, PMi0_2.5,
PM2.5, and UFPs) and mortality through the use of time-series and/or case-crossover studies. Both
study designs aim to disentangle the PM-mortality effect through either complex modeling (i.e.,
time-series) or matching strategies (i.e., case-crossover). Overall, the results of the more recent
studies build upon the conclusions from the previous review, showing consistent positive
associations between mortality and short-term exposure to PM2.5 and PMi0_2.5.
Section 6.5.2 reviews and summarizes the results of recent studies that examined mortality
associations with the four PM size classes listed above. Each section integrates the results of recent
studies with those available in previous PM reviews. This assessment first focuses on multicity
studies that examined mortality associations with PMi0 because this is an important body of
literature that provides information on potential effect modifiers, potential confounding by
copollutants, evaluation of concentration-response relationships, and the influence of different
modeling approaches on the PM-mortality relationship (Section 6.5.2.1). ThePMio studies have
provided the most data among the PM indices thus far; therefore this evaluation begins with the
consideration of those findings as they relate to the general association between PM and mortality. It
is difficult to interpret the extent to which these studies inform an evaluation of the effects of PM2.5
or PMio_2.5, since data are combined from multiple cities with different PM composition.
Interpretations of the PM size fraction that contributes the most to the PMi0 effects observed are
provided when appropriate in the following review. The multicity studies that examine the
association between PMi0 and mortality also offer new evidence on regional and seasonal differences
in effect estimates, building upon observations made in the 2004 PM AQCD (U.S. EPA, 2004,
056905).
Recent study findings on associations with PM25!PMio_2.5, and UFPs are evaluated in Sections
6.5.2.2, 6.5.2.3, and 6.5.2.4, respectively. For PM2.5, the focus of the assessment remains on multicity
study findings; however, for PMi0_2.5 and UFPs, some additional emphasis is placed on single-city
studies, due to the relative sparseness of peer-reviewed literature on these size fractions. Some
studies have also evaluated relationships between mortality and specific components and sources of
PM, and the results are summarized in Sections 6.5.2.4 and 6.5.2.5. Finally, Section 6.5.2.6 assesses
evidence on the concentration-response relationship between short-term PM exposure and mortality.
6.5.1. Summary of Findings from 2004 PM AQCD
The 2004 PM AQCD (U.S. EPA, 2004, 056905) found strong evidence that PMi0 and PM2.5, or
one or more PM2 5 components, acting alone and/or in combination with gaseous copollutants, are
associated with total (nonaccidental) mortality and various cause-specific mortality outcomes. For
PMio, several multicity studies in the U.S., Canada, and Europe provided strong support for this
conclusion, reporting associations with total mortality highlighted by effect estimates ranging from
-0.2 to 0.7% (per 10 ug/m3 increase in PMi0) (U.S. EPA, 2004, 056905). Numerous studies also
reported PMi0 associations with cause-specific mortality, specifically cardiovascular- and
respiratory-related mortality. For PM25, the strength of the evidence varied across categories of
cause-specific mortality, with relatively stronger evidence for associations with cardiovascular
compared to respiratory mortality. The resulting effect estimates reported from the U.S.- and
Canadian-based studies (both multi- and single-city) analyzed for these two categories ranged from
1.2 to 2.7% for cardiovascular-related mortality and 0.8 to 2.7% for respiratory-related mortality, per
10 ug/m3 increase in PM2.5 (U.S. EPA, 2004, 056905). In regards to PMi0.2.5, the PM AQCD found a
limited body of evidence that was suggestive of associations between short-term exposure to ambient
PMio_2.5 and various mortality outcomes (e.g., 0.08-2.4% increase in total [nonaccidental] mortality
per 10 ug/m3 increase in PMi0_2.5). The positive effect estimates obtained from studies that analyzed
the association between PMi0_2.5 and mortality resulted in the conclusion that PMi0_2.5, or some
December 2009 6-158
-------
constituent component(s) (including those on the surface) of PM10_2.5, may contribute, in certain
circumstances, to increased human health risks.
Some additional studies examined the association between specific PM2.5 chemical
components and mortality. These studies observed associations for SO42~, NO3", and CoH, but not
crustal particles. The strength of the association for each component varied from city to city
(U.S. EPA, 2004, 056905). Source-oriented analyses were also conducted to identify specific
source-types associated with mortality. These studies implicate PM2.5 from anthropogenic origin,
such as motor vehicle emissions, coal combustion, oil burning, and vegetative burning, as being
important in contributing to increased mortality (U.S. EPA, 2004, 056905).
6.5.2. Associations of Mortality and Short-Term Exposure to PM
The recent literature examines the association between short-term exposure to various PM size
fractions (i.e., PM10, PM10_2.5, PM25, UFPs, or species [e.g., OC, EC, transition metals, etc.]) and
mortality. This ISA, similar to previous AQCDs, focuses more heavily on multicity studies, and
especially those conducted in the U.S. and Canada (Table 6-15). By using this approach it is possible
to: (1) obtain a more representative sample of or insight into the PM-mortality relationship observed
across the U.S.; (2) analyze the association between mortality and short-term exposure to PM at or
near ambient conditions observed in the U.S.; (3) examine the potential heterogeneity in effect
estimates between cities and regions; and (4) analyze the confounders and/or effect modifiers that
may explain the PM-mortality relationship in the U.S. Although this section focuses on mortality
outcomes in response to short-term exposure to PM, it does not evaluate studies that examine the
association between PM and infant mortality. These studies are evaluated in Section 7.5,,although it
is possible that short- and long-term in utero exposures may contribute to infant mortality. In
addition, the exposure windows of interest for this unique health outcome can be difficult to
characterize and may span both short- and long-term exposure periods.
Table 6-15. Overview of U.S. and Canadian multicity PM studies of mortality analyzed in the 2004 PM
AQCD and the PM ISAb.
Study
Location
Mean Concentration 98th; 99th Upper Percentile:
(ug/m ) Percentiles (ug/m ) Concentrations (ug/m )
PM10
Dominici et al. (2003, 1564071'
Burnett and Goldberg (2003, 042798)3
Peng et al. (2005, 087463)
Dominici et al. (2007, 097361)'
Wfelty and Zeger (2005, 0874841 '
Bell et al. (2009, 1910071
Burnett et al. (2004, 0862471
Samoli et al. (2008, 1884551
Schwartz (2004, 0789981
Schwartz (2004, 0535061
Zeka et al. (2005, 088068)
Zeka et al. (2006, 088749)
90 U.S. cities
8 Canadian cities
100 U.S. cities
100 U.S. cities
100 U.S. cities
84 U.S. urban
communities
12 Canadian cities
12 Canadian cities
90 U.S. cities8
22 European cities
14 U.S. cities
14 U.S. cities
20 U.S. cities
20 U.S. cities
15.3-53.2
25.9
13-49
13-49
13-49
NR
NR
NR
23-36d
23-36d
15-37.5
15.9-37.5
NR
95th: 54; Maximum: 121
50th: 27.1; 75th:
Maximum: 48.7
50th: 27.1; 75th:
Maximum: 48.7
50th: 27.1; 75th:
Maximum: 48.7
NR
NR
NR
75th: 31 -57
75th: 31 -57
NR
NR
32.0
32.0
32.0
December 2009
6-159
-------
Study
, .. Mean Concentration 98th; 99th Upper Percentile:
Locatlon (ug/m3) Percentiles (ug/m3) Concentrations (ug/m3)
PM2.5
Burnett and Goldberg (2003, 042798)3
Dominici et al. (2007, 0973611
Zanobetti and Schwartz (2009, 1884621
Franklin et al. (2007, 091257)
Franklin et al. (2008, 097426)9
Ostro et al. (2006, 0879911
Ostro et al. (2007, 0913541
Burnett et al. (2004, 0862471
8 Canadian cities
96 U.S. cities
112 U.S. cities
27 U.S. cities
25 U.S. cities
9 California counties
6 California counties
12 Canadian cities
13.3 38.9; 45.4
NR
13.2 34.3; 38.6
15.6 45.8; 54.7
14.8 43.0; 50.9
19.9 68.2; 82.0
18.4 61.2; 70.1
12.8 38.0; 45.0
95th: 32; Maximum: 86
NR
Maximum: 57.4
Maximum: 239
Maximum: 239.2
95th: 61. 3; Maximum: 160.0
Maximum: 116.1
Maximum: 86.0
PM10-2.5
Burnett and Goldberg (2003, 04279813
Zanobetti and Schwartz (2009, 1884621
Burnett et al. (2004, 086247)
Villeneuve et al. (2003, 055051)
Klemmetal. (2004, 0565851
Slaughter etal. (2005,073854)
Wilson et al. (2007, 1571491
Kettunen et al. (2007, 0912421
Perez et al. (2008, 1560201
8 Canadian cities
47 U.S. cities
12 Canadian cities
Vancouver, Canada
Atlanta, Georgia
Spokane, Washington
Phoenix, Arizona
Helsinki, Finland
Barcelona, Spain
12.6
11.8 40.2; 47.2
11.4
6.1
9.7 20.7
NR
NR
Cold season: 6.7d
Warm season: 8.4d
Saharan Dust Days: 16.4
Non-Saharan Dust Days:
14.9
95th: 30; Maximum: 99
Maximum: 88.3
Maximum: 151
90th: 13.0; Maximum: 72.0
50th: 9.34; 75th: 11.94
Maximum: 25.17
NR
NR
Cold season: 50th: 6.7
75th: 12.5; Maximum: 101.4
Warm season: 50th: 8.4
75th: 11. 8; Maximum: 42.0
Saharan Dust Days
50th: 14.8; 75th: 21. 8
Maximum: 36.7
Non-Saharan Dust Days
50th: 12.6; 75th: 18.9
Maximum: 93.1
3 Multicity studies examined in the 2004 PM AQCD (U.S. EPA, 2004, 0
b Because only two multicity study was identified that examined PM10_2 5, single-city and international studies that examined PM10_2 5 were analyzed in this ISA and are included in this table.
cThe majority of multicity studies examined in the PM ISA provide the mean PM concentration of each individual city, not an overall PM concentration across all cities. As a result, the range of PM
concentrations for a particular study are presented, which represents the lowest and highest mean PM concentrations reported across cities, if an overall mean is not provided within the study.
a Median PM concentration.
eThe study included 90 U.S. cities in the 1-day lag analysis, but only 15 U.S. cities in the analysis of the average of lag days 0-1.
f The concentrations reported for these studies were estimated from Peng et al. (2005, 087463) because they used the same number of cities and years of data from NMMAPS.
9 This study did not present an overall mean 24-h avg PM2 5 concentration across all cities for each season. The range of mean 24-h avg concentrations reported in this table for each season represents the
lowest mean 24-h avg PM2 5 concentration and the highest 24-h avg PM2 5 concentration reported across all cities included in the study.
6.5.2.1. PM10
The majority of studies that examined the association between short-term exposure to PM and
mortality focused on effects attributed to PMi0. Although these studies do not characterize the
compositional differences in PMi0 across the cities examined in each of the studies evaluated, they
can provide an underlying basis for the overall pattern of associations observed when examining the
relationship between PMi0_2.5 and PM2.5 and mortality. The studies evaluated in this review analyzed
the PMio-mortality relationship through either a time-series or case-crossover design.1
1 Schwartz (1981, 078988) used a case-crossover study design, but also conducted a time-series analysis to validate the results obtained
using the case-crossover approach.
December 2009
6-160
-------
Time-Series Analyses
Mortality associations with short-term exposure to PMi0 in the U.S. have been examined in
several updated time-series analyses of the NMMAPS. In the previous NMMAPS analysis
(Dominici et al., 2003, 156407: Samet et al., 2000, 005809; Samet et al., 2000, 010269) of the
1987-1994 data, which was reviewed in the 2004 PM AQCD (U.S. EPA, 2004, 056905). the
strongest association was found for nonaccidental mortality for 1-day lag, with a combined estimate
across 90 cities of 0.21% (95% PI: 0.09-0.33) per 10 ug/m increase in PMi0. The association was
found to be robust to the inclusion of other gaseous copollutants in the regression models, but the
investigators found heterogeneity across regions, with the strongest associations in northeastern
cities. In the new updated analyses, the investigators examined additional issues surrounding the
association between PM and mortality including: seasonal effect modification; change in risk
estimates over time; sensitivity of results to alternative weather models; and effect modification by
air conditioning use. The NMMAPS data has also been used to examine the PM concentration-
response relationship using PMi0 data from 20 cities (Section 6.5.2.7). A few multicity studies
conducted in Canada and Europe provide additional information, which further clarifies and supports
the association between PM and mortality presented in the NMMAPS analyses.
Seasonal Analyses of PM10-Mortality Associations
Using the updated NMMAPS data, which consisted of 100 U.S. cities for the period
1987-2000, Peng et al. (2005, 087463) examined the effect of season on PMi0-mortality associations.
In their first stage regression model, for each city, the PMi0 effect was modeled to have a sinusoidal
shape that completes a cycle in a year, but was constrained to be periodic across years using
sine/cosine terms. The authors also considered a model that consisted of PMio-season interactions
using season indicators. Both of these models also included covariates that were used in their earlier
NMMAPS analyses. In the second stage model, the seasonal patterns of PMi0 mortality coefficients
were estimated for seven geographic regions and on average for the entire U.S. Peng et al. (2005,
087463) found for 1-day lag, at the national level, season specific increases in nonaccidental
mortality per 10 (ig/m3 increase in PM10 of: 0.15% (95% PI: -0.08 to 0.39), 0.14% (95% PI: -0.14 to
0.42), 0.36% (95% PI: 0.11-0.61), and 0.14% (95% PI: -0.06 to 0.34) for winter, spring, summer, and
fall, respectively. The corresponding all-season estimate was 0.19% (95% PI: 0.10-0.28). After the
inclusion of SO2, OB, or NO2 in the model with PMi0 in a subset of cities (i.e., 45 cities) for which
data existed, PMi0 risk estimates remained fairly robust. An analysis by geographic region found a
strong seasonal pattern in the Northeast. Figure 6-16 presents the estimated seasonal pattern of PMi0
risk estimates by region from Peng et al. (2005, 087463). which includes a sensitivity analysis aimed
to determine the appropriate number of degrees of freedom for temporal adjustment. It is clear from
Figure 6-16 that the Northeast has the strongest association with PMi0 and mortality, which peaks in
the summer and is robust to the extent of temporal adjustment. The industrial Midwest also shows
the summer peak, but with smaller risk estimates. Other regions have either no seasonal pattern
(Southeast) or a suggestion of a spring peak that appears to be sensitive to the extent of temporal
adjustment. On a nationwide basis, the PMi0 risk estimates appear to peak between spring and
summer. Overall, this study identified an effect modifier that may be useful in identifying the
specific chemical component(s) of PM that are related to specific regions and times of the year.
Change in PM10-Mortality Associations over Time
Dominici et al. (2007, 097361) conducted an analysis of the extended NMMAPS data set (i.e.,
1987-2000) to examine if short-term PMi0-mortality risk estimates changed during the course of the
study period. The investigators estimated the average PMi0 mortality risk coefficient for 1-day lag,
using essentially the same model specification as in their 2003 analysis, separately for three time
periods (i.e., 1987-1994, 1995-2000, and 1987-2000) the "eastern U.S." (62 counties), the "western
U.S." (38 counties), and all 100 U.S. counties. To produce national and regional estimates, two-stage
hierarchical models were used as in the previous NMMAPS studies. As shown in Table 6-16, the
authors found a continuation of the PMi0-mortality association in the nationwide data for the entire
study period. A comparison of the relative risk estimates for 1987-1994 vs. 1995-2000 suggests weak
evidence (not a statistically significant difference) that short-term effects declined. Most of the
decline in the national estimate appears to be attributable to the eastern U.S. counties. However, the
decline in the risk estimate for all-cause mortality in the eastern U.S. appears to be
December 2009 6-161
-------
disproportionately influenced by the reduction in the risk estimate for the "other" mortality category
(i.e., all-cause minus cardio-respiratory category, which may be 40-50% of all-cause deaths in U.S.
cities). Likewise, the apparent increase in the risk estimate for all-cause mortality in the western U.S.
appears to be affected by the increase in the risk estimate for the "other" mortality category. Because
the study does not clearly identify the specific cause(s) in the "other" mortality category that are
affected by PM, interpreting the reduction in risk estimates for all-cause mortality requires caution.
In contrast, the apparent reductions (-23%) in PMi0 risk estimates for cardio-respiratory deaths were
more comparable between the two regions.
Industrial Midwest
c 1-5-
c
Southeast
1.0-
0-5
0.0 —
i i i
100 200 300
Northeast
Southwest
Northwest
100 200 300
Day in year
i i i
100 200 300
Southern California
i i r
100 200 300
Source: Reprinted with Permission of Oxford University Press from Peng et al. (2005, 0874631
Figure 6-16. National and regional estimates of smooth seasonal effects for PM10 at a 1-day
lag and their sensitivity to the degrees of freedom assigned to the smooth
function of time in the updated NMMAPS data 1987-2000. Note: The degrees of
freedom chosen were 3 df (short-dashed line), 5 df (dotted line), 7 df (solid line), 9
df (dotted-and-dashed line), and 11 df (long-dashed line) per year of data.
In addition, the investigators estimated time-varying PMi0 mortality risk as a linear function of
calendar time for the period 1987-2000, producing the percentage rate change in the PM10 risk
estimate with a change in time of 1 yr. The estimated rate of decline in slope for all-cause mortality
and the combination of cardiovascular and respiratory mortality were -0.012 (95% PI: -0.037 to
0.014) and -0.016 (95% PI: -0.058 to 0.027), respectively. The authors also estimated a PM2.5
mortality risk for the period 1999-2000 (discussed in Section 6.5.2.2.).
December 2009
6-162
-------
Table 6-16. NMMAPS national and regional percentage increase in all-cause, cardio-respiratory, and
other-cause mortality associated with a 10 ug/m3 increase in PMio at lag 1 day for the
periods 1987-1994,1995-2000, and 1987-2000.
1987-1994
95% PI
1996-2000
95% PI
1987-2000
95% PI
ALL CAUSE
East
West
National
0.29
0.12
0.21
0.12, 0.46
-0.07, 0.30
0.10, 0.32
0.13
0.18
0.18
-0.19, 0.44
-0.07, 0.44
0.00, 0.35
0.25
0.12
0.19
0.11,0.39
-0.02, 026
0.10,0.28
CARDIORESPIRATORY
East
West
National
0.39
0.17
0.28
0.16, 0.63
-0.07, 0.40
0.14, 0.43
0.30
0.13
0.21
-0.13, 0.73
-0.23, 0.50
-0.03, 0.44
0.34
0.14
0.24
0.15,0.54
-0.05, 0.33
0.13, 0.36
OTHER
East
West
National
0.21
0.09
0.15
-0.03, 0.44
-0.21,0.38
-0.02, 0.32
0.00
0.23
0.17
-0.49, 0.50
-0.15, 0.62
-0.07, 0.41
0.15
0.11
0.15
-0.09, 0.39
-0.10, 0.33
0.00, 0.29
Source: Reprinted with Permission of HEI from Dominici et al. (2007, 097361)
The objective of the Dominici et al. (2007, 097361) study described above was motivated by
accountability research, the idea of measuring the impact of policy interventions. However, unlike
the intervention studies conducted in Hong Kong (Hedley et al., 2002, 040284) and Dublin, Ireland
(Clancy et al., 2002, 035270) that were reviewed in the 2004 PM AQCD (U.S. EPA, 2004, 056905).
this study was not designed to estimate a reduction in mortality in response to a sudden change in air
pollution. In fact, the figure of observed trend in PMio levels presented in the Dominici et al. (2007,
097361) study indicates that the decline in PMio levels during the study period was very gradual,
with much of the decline appearing in the first few years (median values of ~33 (ig/m3 in 1987 to
~25 (ig/m3 in 1992, then down to ~23 (ig/m3 in 2000). A flaw in the use of the time-series study
design for this type of analysis is that it adjusts for long-term trends, and, therefore, does not
estimate the change in mortality in response to the gradual change in PM10. The apparent change,
though weak, in the PMio risk estimates may also reflect a potential change in the composition of
PMio (i.e., PMio_2.5 or PM2.5). The study listed a number of PMio-related air pollution control
programs that were implemented between 1987 and 2000. Some of these programs, such as the Acid
Rain Control Program, did result in major reductions in emissions, and, therefore, could have
contributed to the results observed, but the analytic approach used in the study does not allow for a
systematic analysis of the effect of air pollution policies on the risk of mortality.
Sensitivity of PM-Mortality Associations to Alternative Weather Models
To examine the sensitivity of PM10-mortality risk estimates to alternative weather models that
consider longer lags, Welty and Zeger (2005, 087484) analyzed the updated NMMAPS 100 U.S.
cities data. All of the previous NMMAPS analyses only considered temperature and dew point up to
3-day lags. In this analysis, the authors considered various forms of a constrained distributed lag
model: (1) containing a step function of temperature with steps at lag 0, 2, 7 and extended to
14 days; (2) similar to (1) but with time-varying coefficients to change over season and study period;
and, (3) containing a smooth function to account for non-linearity in the temperature-mortality
relationship. With the combination of degrees of freedom for temporal trends and the number of
distributed lags, more than 20 models were applied to each of the 3 lag days (0, 1, and 2) of PMi0.
These city-specific risk estimates were then combined across the 100 cities in the second stage
Bayesian model. The combined PMio risk estimates were generally consistent within the lag. In
particular, the risk estimates for nonaccidental mortality for lag 1 day ranged between 0.15% and
December 2009
6-163
-------
0.25% per 10 (ig/m3 increase in PM10, and were always statistically significant regardless of the
model used. In addition, the range of these point estimates across the models was found to be much
narrower than the regression posterior intervals. Thus, the PMi0 risk estimates at lag 1 day were
robust to alternative temperature models that considered temperature effects lasting up to a 2-week
period.
In summary, the above three analyses of the updated NMMAPS data provided useful
information on PM-mortality risks, resulting in the following conclusions: (1) estimated PMi0
mortality risk is particularly high in the northeast and in the summer; (2) there remains an overall
PMio-mortality association in the 1987-2000 time period as well as the 1995-2000 time period; (3)
there is a weak indication that PMio-mortality risk estimates are declining; and (4) PMio-mortality
risk estimates were not sensitive to alternative temperature models.
Effect Modification of PM10-Mortality Associations by Air Conditioning Use
It has been hypothesized that air conditioning (AC) use reduces an individual's exposure to
PM and subsequently modifies the PM-mortality association. Bell et al. (2009, 191007) investigated
the role of AC use on the relationship between PMi0 and all-cause mortality using the NMMAPS
PMio risk estimates from 84 U.S. urban communities from 1987-2000.1 Bayesian hierarchical
modeling was used to examine if AC prevalence (i.e., fraction of households with central or any AC)
explained city-to-city variation in PMi0 risk estimates. The authors calculated yearly, summer-only,
and winter-only effect estimates stratified by housing stock that had either central AC or any AC,
which includes window units. Risk estimates for lag 1 (previous day) were used in the analysis
because this lag showed the strongest association with mortality in the original NMMAPS analyses.
Community-specific AC prevalence was calculated from national survey U.S. Census American
Housing Survey (AHS) data, which is available every two years. The investigators computed percent
change in PMi0 effect estimates per an additional 20% of the population acquiring AC.
The AC variables were not strongly correlated with socio-economic variables (poverty rate,
unemployment, and education) from the U.S. Census (correlation ranged from -0.27 to 0.29). Bell
et al. (2009, 191007) found that communities with higher AC prevalence had lower PMi0 mortality
risk estimates for all-cause mortality (-30.4% [95% PI: -80.4 to 19.6] per an additional 20% of the
population acquiring any AC; -39.0% [95% PI: -81.4 to 3.3] for central AC), but results were not
statistically significant. When restricting the analysis to the summer months and focusing on the 45
cities with summer-peaking PMi0 concentrations, the authors reported positive, non-significant risk
estimates (29.9% [95% PI: -84.0 to 144] per an additional 20% of the population acquiring any AC; -
2.0% [95% PI: -60.3 to 64.3] for central AC). A similar analysis was conducted for winter months
using data from six cities with winter peaking PMio concentrations, but the confidence bands were
too wide (due to the small sample size) for meaningful interpretation.
Although the estimated reductions in PMio all-cause mortality risks from AC use reported in
the Bell et al. (2009, 191007) study were not statistically significant, their large magnitude suggests
that AC use may reduce an individual's exposure to PM. Given the expected additional increase in
AC use in the future, and the results from recent multicity studies, which have reported stronger PM-
mortality associations during the warm season, AC use may play a larger role in determining an
individuals exposure to PM. Studies that have examined the effect of AC use on the PM2 5-mortality
association have reported similar results. For example, Franklin et al. (2007, 091257) (discussed in
detail in Section 6.5.2.2) found that AC use non-significantly modified PM2.5 mortality risk
estimates, but the result was suggestive of higher PM2.5 effects in cities with lower AC use,
especially in cities with summer-peaking PM2.5 concentrations. Overall, further investigation is
needed to fully understand the relationship between AC use and mortality attributed to short-term
exposure to PM.
PM10-Mortality Associations in Canada and Europe
Burnett et al. (2004, 086247) examined the association between mortality and various air
pollutants in 12 Canadian cities, and reported that the most consistent association was found for
NO2. For this analysis, PM was measured every 6th day for the majority of the study period, and the
PMio concentrations used in the study represent the sum of the PM25 and PMi0_2.5, which were
directly measured by dichotomous samplers. The authors found that the simultaneous inclusion of
1 This study also examined risk estimates for cardiovascular and respiratory hospital admissions in older adults (> 65).
December 2009 6-164
-------
NO2 and PM10 in a model, on those days with PM data, greatly reduced the PM10 association with
nonaccidental mortality, from 0.47% (95% CI: 0.04-0.89) to 0.07% (95% CI: -0.44 to 0.58) per
10 (ig/m3 increase. The previous Canadian multicity analysis (Burnett and Goldberg, 2003, 042798).
a re-analysis of Burnett et al. (2000, 010273) reviewed in the 2004 PM AQCD (U.S. EPA, 2004,
056905). did not consider gaseous pollutants. Thus, PMi0 risk estimates in the Canadian data appear
to be more sensitive to NO2 than those estimates reported in U.S. studies.
The association between PM10 and mortality in Europe was also reviewed in the 2004 PM
AQCD (U.S. EPA, 2004, 056905) through Katsouyanni et al. (2003, 042807). which presented
results from the APHEA-2 study, a multicity study that examined PMi0 effects on total mortality in
29 European cities. Analitis et al. (2006, 088177) published a brief report on effect estimates for
cardiovascular and respiratory deaths also based on the 29 European cities, within the APHEA2
study. They reported for the average of 0- and 1-day lags, PMi0 risk estimates per 10 (ig/m3 of 0.76%
(95% CI: 0.47-1.05) for cardiovascular deaths and 0.71% (95% CI: 0.22-1.20) for respiratory deaths
in random effects models.
Comparison of PM-Mortality Associations in Europe, Canada, and the U.S.
The APHENA study (Samoli et al., 2008, 188455) was a collaborative effort by the APHEA,
NMMAPS, and the Canadian multicity study investigators to evaluate the coherence of PMi0
mortality risk estimates across locations and possible effect modifiers of the PM-mortality
relationship using a common protocol. To adjust for temporal trends, Samoli et al. (2008, 188455)
used 3, 8, and 12 degrees of freedom (df) with natural splines and penalized splines, as well as the
minimization of the sum of the absolute values of the partial auto-correlation function (PACF). The
investigators also included a smooth function of temperature on the same day of death and the day
before death. The study reported risk estimates for a 1-day lag (from all three data sets), the average
of lag day 0 and 1 (all but for the Canadian data because PM data was collected every 6th day), and
an unconstrained distributed lag model using lags of 0, 1, and 2 days (all but for the Canadian data).
The second-stage regression included: (a) the average pollution level and mix in each city; (b) air
pollution exposure characterization (e.g., number of monitors, density of monitors); (c) the health
status of the population (e.g., cardio-respiratory deaths as a percentage of total mortality, crude
mortality rate, etc.); and (d) climatic conditions (e.g., mean and variance of temperature). In addition,
unemployment rate was examined for 14 European cities and all U.S. cities. Effect modification
patterns were examined only for cities with complete time-series data and using the average of lags 0
and 1 day, resulting in the exclusion of the Canadian data.
Generally, the risk estimates from Europe and the U.S. were similar, but those from Canada
were substantially higher.1 For example, the percent excess risks per 10 ug/m3 increase in PMi0 for
all ages using 8 df/yr and penalized splines were 0.84% (95% CI: 0.30-1.40), 0.33% (95% CI: 0.22-
0.44), and 0.29% (95% CI: 0.18-0.40) for the Canadian, European, and U.S. data, respectively. Note
that the risk estimate for the 90 U.S. cities is slightly larger than that reported in the original
NMMAPS study (0.21%, using natural splines, and more temperature variables). In the all ages
model, the average of lag days 0 and 1, and the distributed lag model with lags 0, 1, and 2 did not
result in larger risk estimates compared to those for a 1 day lag. In copollutant models, PMi0 risk
estimates did not change when controlling for O3. Figure 6-17 shows the risk estimates from the
three data sets for alternative extent of temporal smoothing and smoothing methods. The Canadian
data appear less sensitive to the extent of temporal smoothing or smoothing methods (Panel A of
Figure 6-17). When stratifying by age the risk estimates for the older age group (> 75 yr) were
consistently larger than those for the younger age group (<75 yr) (e.g., 0.47% vs. 0.12% for the U.S.
data) for all the three data sets. Although the study did not quantitatively present the results from the
effect modification analyses, some evidence of effect modification across the study regions was
observed. The investigators reported that, in the European data, higher levels of NO2 and a larger
NO2/PMio ratio were associated with greater PMi0 risk estimates, and that while this pattern was also
present in the U.S. data, it was less pronounced. Additionally, in the U.S. data, smaller PM10 risk
estimates were observed among older adults in cities with higher O3 levels. Effect modification by
temperature was also observed, but only in the European data.
1 The risk estimate reported for the 12 Canadian cities examined in the APHENA study is higher than that reported by Burnett et al. (2004,
086247). This is because the APHENA study did not use the 12 cities data from Burnett et al. (2004, 086247). but instead used a
composite of the data from three previous studies conducted by the same group by the same group (Burnett and Goldberg, 2003,
042798; 1998, 029505; Burnett etal., 2000, 010273).
December 2009 6-165
-------
5fl
ee
Oi
'Z i.o
c
!B
g 0.5
Q_
0.0
-0.5
• PS
ANS
3 8 12
Canada (n= 12)
3 8 12
Europe (n = 22)
df/year
3 8 12
USA(n = 90)
2-5
CO
2 1.5
u
'Z i-O
Ł
0,0
-0.5
3 8 12PACF
Europe (n = 22)
3 8 12PACF
USA(n=15)
i i i i
3 8 12PACF
Europe-USA
df/year
Source: Samoli et al. (2008, 188455)
Figure 6-17. Percent increase in the daily number of deaths, for all ages, associated with a
10-ug/m3 increase in PMi0: lag 1 (A) and lags 0 and 1 (B) for all three centers.
PACF indicates df based on minimization of PACF.
In this study, the underlying basis for the larger PMi0 risk estimates (by twofold) in the
Canadian data compared to the European and U.S. data could not be identified, even when consistent
statistical methods were applied across each of the data sets. Because the effect modification of PMi0
risk estimates were not examined in the Canadian data, the potential influence of air pollution type or
mixture could not be ruled out as a potential source of heterogeneity across the three data sets. It
should be noted that both the original U.S. and European studies reported regional heterogeneity in
PM risk estimates, and the U.S. data also demonstrated seasonal heterogeneity. In both of these cases
the specific characteristics associated with the regions that contributed to the heterogeneity observed
were not identified. Thus, further investigation is needed to identify factors that influence the
heterogeneity in PM risk estimates observed between different countries and across regions.
Case-Crossover Analyses
Since the 2004 PM AQCD (U.S. EPA, 2004, 056905) investigators have used the
case-crossover study design more frequently as an alternative to time-series analyses to examine the
association between short-term exposure to PM and mortality. This study design allows for the
control of seasonal variation, time trends, and slow time varying confounders without the use of
complex models. However, similar to any study design, biases can be introduced into the study
depending on the control (i.e., referent) period selected (Janes et al., 2005, 087535). The multicity
case-crossover analyses discussed below match cases (i.e., days in which a death occurred) to
controls (i.e., days in which a death did not occur), to control for (1) seasonal patterns and gaseous
pollutants; or (2) temperature. In addition, the studies attempted to examine the heterogeneity of
effect estimates through the analysis of individual-level and city-specific effect modification.
Controlling for Temperature
Schwartz (2004, 078998) investigated the PMi0-mortality association in 14 U.S. cities for the
years 1986-1993 (some cities started in later years because of PMi0 data availability) using a
case-crossover study design. Note that in this analysis, four more cities (Boulder, CO; Cincinnati,
OH; Columbus, OH; and Provo-Orem, UT) were added to the cities Schwartz (2003, 042800)
previously analyzed using a time-series study design. These cities were chosen for this analysis
because they collected daily PMi0 data, unlike most U.S. cities, which only monitor PMi0 every six
days. Lag 1-day PMi0 risk estimates were computed using several methods. Model 1 (i.e., the main
model) and Model 2 were constructed from a case-crossover analysis with bidirectional control days
December 2009
6-166
-------
(7-15 days before and after the case). Model 1 obtained city-specific estimates in the first stage
analysis, followed by a second stage random-effects model to obtain a combined estimate. Model 2
is the same as Model 1, but consisted of a single stage model, which included data from all 14 cities.
Models 3 and 4 were also constructed from a case-crossover analysis, but used time-stratified control
days (i.e., matched on season and temperature within the same degree in Celsius). Model 3 obtained
single-city estimates in the first stage analysis, followed by a second stage random-effects model to
obtain combined estimates. Model 4 used the same approach as Model 3, but consisted of a single
stage model including data from all 14 cities. The final model, Model 5 consisted of a two-stage
Poisson time-series model, which produced city-specific estimates in the first stage, and combined
estimates across cities in the second stage. In the main model the estimated excess risk for
nonaccidental mortality was 0.36% (95% CI: 0.22-0.50) per 10 (ig/m3 increase in PMi0. The other
models yielded a similar magnitude of effect estimates, ranging from 0.32% (Model 2) to 0.53%
(Model 4). Thus, the methods used to select control days and adjust for weather in the case-crossover
design did not result in major differences in effect estimates, and in addition, were comparable to the
estimates obtained from the time-series analysis, 0.40% (Model 5).
Controlling for Gaseous Pollutants
In a subsequent analysis, Schwartz (2004, 053506) analyzed the same 14 cities data described
above, using a case-crossover design, to investigate the potential confounding effects of gaseous
pollutants. For each case day, control days were selected from all other days of the same month of
the same year. In addition, control days were selected if they had gaseous pollutant concentrations
within: 1 ppb, 1 ppb, 2 ppb, or 0.03 ppm for SO2, NO2, 1-h max O3, and CO, respectively, of the case
day. Unlike the study described above (Schwartz, 2004, 078998) in this analysis, the excess risk was
estimated for the average of 0- and 1-day lag PMi0 (rather than 1-day lag). In addition, apparent
temperature (a composite index of temperature and humidity) was used rather than temperature and
humidity individually. The case-crossover analysis was conducted in each city, and a combined
estimate was computed in a second-stage random effects model. The number of cities analyzed
varied across pollutants depending on the availability of monitors. The study reported PMi0 risk
estimates for nonaccidental mortality of 0.81% (95% CI: 0.47-1.15), 0.78% (95% CI: 0.42-1.15),
0.45% (95% CI: 0.12-0.78), and 0.53% (95% CI: 0.04-1.02) per 10 (ig/m3 increase, for the analysis
matched by SO2 (10 cities), NO2 (8 cities), O3 (13 cities), and CO (13 cities), respectively.
Schwartz (2004, 053506) only presented PMi0 risk estimates matched by gaseous pollutants,
therefore, it is unclear in this analysis how matching by gaseous pollutants affected (i.e., reduced or
increased) unmatched PMi0 risk estimates. The estimates reported were computed using the average
of 0- and 1-day lagged PMi0 and, therefore, cannot be directly compared to the 1-day lag PMi0 risk
estimates obtained in the Schwartz (2004, 078998) 14-city study described above. The estimates
reported in the case-crossover analysis that controlled for gaseous pollutants (Schwartz, 2004,
053506) are generally larger than those obtained in the analysis that controlled for temperature
(Schwartz, 2004, 078998). which was expected since the Schwartz (2004, 053506) analysis used
2-day avg PMi0. However, the estimates reported in Schwartz (2004, 053506) are comparable to the
average of 0- and 1-day lagged PMi0 risk estimate for nonaccidental mortality (0.55% [95% CI:
0.39-0.70]) per 10 (ig/m3 increase from the 10-city study (Schwartz, 2003, 042800). which was
reviewed in the 2004 PM AQCD (U.S. EPA, 2004, 056905). Overall, Schwartz (2004, 053506)
provided an alternative method to assess the influence of gaseous copollutants. The results suggest
that PMio is significantly associated with all-cause mortality after controlling for each of the gaseous
copollutants.
City-Level Effect Modification
Zeka et al. (2005, 088068) expanded the 14 cities analyses conducted by Schwartz (2004,
078998: 2004, 053506) to 20 cities, added more years of data (1989-2000), and investigated PM10
effects on total and cause-specific mortality using a case-crossover design. Individual 0-, 1-, and
2-day lags as well as an unconstrained distributed lag model with 0, 1, and 2 lag days were
examined. For each case day, control days were defined as every third day in the same month of the
same year, to eliminate serial correlation. The authors also investigated potential effect modifiers in
the second stage regression using city-specific variables including percent using AC, population
density, standardized mortality rates, the proportion of elderly in each city, daily minimum apparent
December 2009 6-167
-------
temperature in summer, daily maximum apparent temperature in winter, and the estimated
percentage of primary PMi0 from traffic sources.
o
0>
0,8
I °4
-P
o °-4
Ł
•E 0.2
o
-0.2
-0.4
1000 3000
(count/mile )
25% 75%
25% 75% 25% 75%
Population
density
Variance of
summer AT
Mean of
winter AT
% Primary
PMIO from
traffic
City effect modifiers
Source: Reprinted with Permssion of BMJ Group from Zeka et al. (2005, 088068)
Figure 6-18. Effect modification by city characteristics in 20 U.S. cities. Note: The two
estimates and their Cl for each of the modifying factors represent the percentage
increase in mortality for a 10 ug/m3 increase in PMio, for the 25th percentile, and
75th percentile of the modifier distribution across the 20 cities.
The investigators found that, for all-cause (nonaccidental) mortality, lag 1 day showed the
largest risk estimate (0.35% [95% CI: 0.21-0.49] per 10 (ig/m3) among the individual lags.
Respiratory mortality exhibited associations at lag 0, 1, and 2 days (0.34%, 0.52%, and 0.51%,
respectively), whereas cardiovascular mortality was most strongly associated with PMi0 at lag day 2
(0.37%). The sum of the distributed lag risk estimates (e.g., 0.45% [95% CI: 0.25-0.65] for all-cause
mortality) was generally larger than those for single-day lag estimates. The excess risk estimates for
single-day lags for specific respiratory and cardiovascular causes had generally wider confidence
intervals due to their smaller daily mortality counts, but some of the categories showed markedly
larger estimates when included in the combined distributed lag model (e.g., pneumonia 1.24% [95%
CI: 0.46-2.02]). As shown in Figure 6-18, Zeka et al. (2005, 088068) also found evidence indicative
of several PM10 effect modifiers including higher population density and the estimated percentage of
primary PMio from traffic. When 25th versus 75th percentiles of these city-specific variables were
evaluated, the estimated percent increase in mortality attributed to PMio appears to contrast
substantially (e.g., 0.09% vs. 0.52% for variance of summer time apparent temperature).
The effect modifiers investigated by Zeka et al. (2005, 088068) consisted of city-specific
variables. Some of these variables are ecological in nature, and therefore, interpreting the meaning of
"effect modification" requires some caution. As the investigators pointed out, the population density
and the estimated percentage of primary PMio from traffic were correlated in this data set (r = 0.65)1.
These variables may also be a surrogate for another or composite aspects of "urban" characteristics.
1 The correlation coefficient was calculated based on the numbers provided in Table 1 of Zeka et al. (2005,
December 2009
6-168
-------
Thus, the apparent effect modification by traffic-related PM10 needs further investigation.
Interestingly, the percent of homes with central AC was not a significant effect modifier of PMi0 risk
estimates, which questions the impact of reduced building ventilation rates on PM exposure. Overall,
this study presented PMi0 risk estimates that are consistent with those found in other analyses, but
also provided new information on the risk estimated for broad and specific respiratory and
cardiovascular mortality designations, along with possible effect modifying city-specific
characteristics.
Individual-level Effect Modification
In an additional analysis, Zeka et al. (2006, 088749) examined individual-level, instead of
city-specific, effect modification of PMi0-mortality associations in the 20 U.S. cities described above
using the same case-crossover design. City-specific estimates were obtained in the first stage model,
followed by a second stage model which estimated the overall effects across all cities. Figure 6-19
shows PMio excess risks by four of the individual characteristics examined in the study (i.e., gender,
race, age group, and education). It should be noted that the lag and averaging of days for the
associations reported varied across the outcomes: all-cause and heart disease deaths used the average
of lag 1 and 2 days; respiratory deaths used the average of lag 0 through 2 days; MI deaths used lag
0 day; and stroke deaths used lag 1 day. PMi0 risk estimates do not appear to differ by gender or by
race. However, significant differences were found for the youngest vs. oldest age groups for
all-cause and heart disease mortality. For all-cause mortality, the level of education appeared to be
inversely related to the PMi0 risk estimates (i.e., greater risk for lower education level), but this
observation was not statistically significant. The study also examined effect modification by location
of death ("out-of-hospital" versus "in-hospital") and season (Figure 6-20). The "out-of-hospital"
deaths showed larger PMi0 risk estimates than were found for "in-hospital deaths" with a significant
difference per 10 ug/m3 for all-cause (0.71% versus 0.22%) and heart disease (0.93% versus 0.15%)
deaths. Stroke deaths also showed a significant difference (0.87% vs. 0.06%, not shown in Figure
6-20).
December 2009 6-169
-------
Gender
tn
d
Male ,
Female
Race
0 White
• Black
0)
tr
Age group
1° <6b
\& 6b-/b
0)
CC
Education
Ł•
a
in
9
0 Low
Medium
High
cc
cc
Figure 6-19. Percent excess risk in mortality (all-cause [nonaccidental] and cause-specific)
per 10 ug/m3increase in PMioby individual-level characteristics. The risk
estimates and 95% confidence intervals were plotted using numerical results
from tables in Zeka et al. (2006, 088749). The estimates with * next to them are
significantly higher than the lowest estimate in the group.
December 2009
6-170
-------
Location of death
Season
0 In-hospital
Out-of-hospitall
ac
111
in
CD
o
><
LU
in
CD
in
CD
f
Winter
Summer
Spring/Fall
31
cc
so
oj
I
2 -L
'EL
:/3
Q~
Figure 6-20. Percent excess risk in mortality (all-cause [nonaccidental] and cause-specific)
per 10 ug/m3increase in PMi0 by location of death and by season. The risk
estimates and 95% confidence intervals were plotted using numerical results
from tables in Zeka et al. (2006, 088749). The estimates with * next to them are
significantly higher than the lowest estimate in the group.
Overall, Zeka et al. (2006, 088749) showed a consistent pattern of effect modification by
contributing causes of death (i.e., pneumonia, stroke, heart failure, and diabetes) on PMi0 risk
estimates for primary causes of death (Figure 6-21; not all results for contributing cause are shown).
However, because the contributing causes of death counts were relatively small, as reflected by the
wide confidence intervals in Figure 6-21, most of the differences observed did not achieve statistical
significance.
December 2009
6-171
-------
5
.fr 4
1
o 3-
E
,E
a 2 -
U)
i
8 1-
o
* n
i °
s
„» .1-
K h
4
>
> 1
1
l>
^
4
1*
*
4
1
I
>
"2 *
//
// /,
4?
All cause
Ml
Stroke
Figure 6-21.
Respiratory
Primary cause of death
Source: Adatped with Permission of Oxford University Press from Zeka et al. (2006, 088749)
Percent increase in mortality (all-cause [nonaccidental] and cause-specific) per
10 ug/m3 increase in PMi0 by contributing causes of death. The estimates with *
(added to the original figure) indicates a significant difference.
In addition, when examining the other effect modifiers, the results that show no difference in
risk estimates between gender or race for all-cause and cardiovascular deaths are important,
given the relatively narrow confidence bands of these estimates. The effect modification by the
location of death has been reported previously in smaller studies, but the large contrast found for
all-cause and cardiovascular mortality in this large multicity analysis is noteworthy. The elevated
PMio risks reported by Zeka et al. (2006, 088749) for all-cause, heart disease (and stroke)
"out-of-hospital" deaths are also consistent with the hypothesis of acute PMio effects on "sudden
deaths" brought on by systemic inflammation or dysregulation of the ANS. The finding regarding the
seasonal effect modification, though significant only for respiratory deaths, is somewhat in contrast
with the Peng et al. (2005, 087463) analysis of the extended NMMAPS data, which observed the
greatest effects during the summer season. The apparent inconsistency may be due to the difference
in geographic coverage (i.e., 20 versus 100 cities) or methodology (i.e., case-crossover with referent
days in the same month of the same year vs. time-series analysis with adjustment for temporal trend
in the regression model).
Summary of PM10 Risk Estimates
Overall, the recent studies continue to show an association between short-term exposure to PM
and mortality. Although these studies do not examine mortality effects attributed to PM size fractions
that compose PMio, the regional, seasonal, and effect modification analyses conducted contribute to
the evidence for the PM2.5 and PM10_2.5 associations presented in Sections 6.5.2.2and 6.5.2.3,
respectively. Of the PMio studies evaluated, depending on the lag/averaging time and the number of
cities included, the estimates for all-cause (nonaccidental) mortality for all ages ranged from 0.12%
(Dominici et al., 2007, 097361) to 0.84% (Samoli et al., 2008, 188455) per 10 (ig/m3 increase in
PMio, regardless of the study design used (i.e., time-series vs. case crossover). Although this range of
PM mortality risk estimates is smaller than those reported for PMi0_2.5 and PM2.5they do support the
December 2009
6-172
-------
association between PM and mortality. The majority of studies examined present estimates for either
a lag of 1 day or a 2-day avg (lag 0-1), both of which have been found to be strongly associated with
the risk of death (Schwartz, 2004, 078998; 2004, 053506). The use of a distributed lag model (using
lag 0, 1, and 2 days) was found to result in slightly larger (by -30%) estimates compared to those for
single-day lags in the 20 cities study (Zeka et al., 2005, 088068). but when using the 15 cities data
from NMMAPS analyzed in the APHENA study (Samoli et al., 2008, 188455). the 1-day lag
combined risk estimate was larger than the distributed lag (lag, 0, 1, and 2 days) estimate. Overall, an
examination of the PMi0 risk estimates stratified by cause-specific mortality and age, for all U.S.-
and Canadian-based studies, further supports the findings of the multicity studies discussed in the
2004 PM AQCD (U.S. EPA, 2004, 056905) (i.e., consistent positive associations between short-term
exposure to PMi0 and mortality) and this ISA, however, it must be noted that a large degree of
variability exists between cities when examining city-specific risk estimates.
The variability in PMi0 mortality risk estimates reported within and between multicity studies
may be due to the difference in the cities analyzed and the potential regional differences in PM
composition. The NMMAPS studies have found that geographic regions and seasons are the two
most important factors that determine the variability in risk estimates, with estimates being larger in
the eastern U.S. and during the summer. These findings were fairly consistent across studies, but
Zeka et al. (2006, 088749) observed the strongest association during the transition period (spring and
fall); however, this may be due to the difference in geographic coverage or the difference in the
model specification used compared to Peng et al. (2005, 087463).
Finally, examination of potential confounders showed that the size of PMi0 risk estimates are
fairly robust to the inclusion of gaseous copollutants in models (Peng et al., 2005, 087463) or by
matching days with similar gaseous pollutant concentrations (Schwartz, 2004, 053506). These
findings further confirmed that PMi0 risk estimates are not, at least in a straightforward manner,
confounded by gaseous copollutants.
December 2009 6-173
-------
Study
Peng et al. (2005, 087463). NMMAPS, Lag 1
Dominici et al. (2007, 097361) NMMAPS, Lag 1
Welty and Zeger (2005, 087484) NMMAPS, Lag 1
Schwartz (2004, 078998) 14 U.S. cities, Lag 1
Schwartz (2004, 053506) 14 U.S. cities, Lag 0-1
Zeka et al. (2005, 088068) 20 U.S. cities
LagO
Lag1
Lag 2
Sum of distributed lag 0-2
Zeka et al. (2006, 088749) 20 U.S. cities, Lag 1-2
Effect Estimate (95% Cl)
All seasons
Summer
Fill
Nationwide, 1987-1994
1995-2000
1987-2000
East, 1987-1994
1995°000
1987-2000
West 1987-1994 '
1 995-2000 I
1 OR? 9nnn j
Bidirectional, 2-stage
Bidirectional, 1 -stage
Matched by temperature, 2-stage
Matched by temperature, 1 -stage
Time-series
Matched by CO (13 cities)
Matched by 03 (13 cities)
Matched by N02 (8 cities)
Matched by S02 (10 cities)
Winter
Spring/Fall
In-hospital
Out-of-hospital
Burnett et al. (2004, 086247) 12 Canadian cities, Lag 1
Samoli et al. (2008, 188455) APHENA, Lag 1
Canada (12 cities)
U.S. (90 cities)
i nun
II I I I I I I I I
-0.4 0.0 0.4 0.8 1.2 1.6
% Increase
Figure 6-22. Summary of percent increase in all-cause (nonaccidental) mortality from recent
multicity studies per 10 ug/m3 increase in PMio. The number after the study
location indicates lag/average used for PMi0 (e.g., "01" indicates the average of
lag 0 and 1 days). For Welty and Zeger (2005, 087484), the vertical lines represent
point estimates for 23 different weather models, and the horizontal band spans
the 95% posterior intervals of these point estimates.
6.5.2.2. PM2.5
A nationwide monitoring system for PM2.5 was not established until 1999. This in conjunction
with the unavailability of nationwide mortality data from the National Center of Health Statistics
(NCHS) starting in 20011, has contributed to the relatively small literature base that has examined
1 In 2008 the EPA facilitated the availability of the mortality data for EPA-funded researchers, which should eventually increase the
literature base of studies that examine the association between short-term exposure to PM2.5 and mortality.
December 2009
6-174
-------
the association between short-term exposure to PM2.5 and mortality. To date, the studies that have
been conducted examined national (i.e., in multiple cities across the country) or regional (i.e., in one
location of the country) PM2.5 associations with mortality.
PM2.s - Mortality Associations on a National Scale
The NMMAPS study conducted by Dominici et al. (2007, 097361) (described in
Section 6.5.2.1), also conducted a national analysis of PM2 5-mortality associations using the same
methodology and data for 1999-2000. The PM25 risk estimates at lag 1 day were 0.29% (95%PI:
0.01-0.57) and 0.38% (95%PI: -0.07 to 0.82) per 10 (ig/m3 increase for all-cause and
cardio-respiratory mortality, respectively. The authors also conducted a sensitivity analysis of the
risk estimates based on the extent of adjustment for temporal trends in the model, changing the
degrees of freedom (df) of temporal adjustment from 1 to 20/yr (the main result used 7 df/yr). In
comparison to the PMi0 results, the PM2 5 risk estimates appeared more sensitive to the extent of
temporal adjustment between 5 and 10 df/yr, but this may be in part due to the much smaller sample
size used for the PM2 5 analysis (i.e., mortality counts from 1999-2000) compared to the PMi0
analysis (i.e., mortality counts from 1987-2000).
Franklin et al. (2007, 091257) analyzed 27 cities across the U.S. that had PM25 monitoring and
daily mortality data for at least two years of a 6-yr period, 1997-2002. The mortality data up to year
2000 were obtained from the NCHS, while the 2001-2002 data were obtained from six states (CA,
MI, MN, PA, TX, and WA), resulting in 12 out of the 27 cities having data up to 2002. The start year
for each city included in the study was set at 1999, except for Milwaukee, WI (1997) and Boston,
MA (1998), which is due to PM25 data availability in these two cities. In the case-crossover analysis
in each city, control days for each death were chosen to be every 3rd-day within the same month and
year that death occurred in order to reduce autocorrelation. The first stage regression examined the
interaction of effects with age and gender, while the second stage random effects model combined
city-specific PM2 5 risk estimates and examined possible effect modifiers using city-specific
characteristics (e.g., prevalence of central AC and geographic region). For all of the mortality
categories, the estimates for lag 1 day showed the largest estimates. The combined estimates at lag 1
day were: 1.2% (95%CI: 0.29-2.1), 0.94% (95%CI: -0.14 to 2.0), 1.8% (95%CI: 0.20-3.4), and 1.0%
(95%CI: 0.02-2.0) for all-cause, cardiovascular, respiratory, and stroke deaths, respectively, per
10 (ig/m3. When examining the city-specific risk estimates most of the cities with negative estimates
were also those with a high prevalence of central AC (Dallas, 89%; Houston, 84%; Las Vegas, 93%;
Birmingham, 77%). It is unclear why these cities exhibit negative (and significant) risk estimates
rather than null effects.
In the analysis of effect modifiers, Franklin et al. (2007, 091257) found that individuals
> 75 yr showed significantly higher PM2 5 risk estimates than those individuals < 75 yr. The
estimated effects were also found to vary by geographic location with larger estimates in the East
than in the West, which are consistent with the regional pattern found in the NMMAPS PMi0 risk
estimates. In addition, a higher prevalence of central AC was associated with decreased PM25 risk
estimates when comparing the lower (25th percentile) versus the higher (75th percentile) AC use rates,
especially in the cities where PM2 5 concentrations peak in the summer. Finally, the risk estimates
were not found to be different between communities with PM25 concentrations < 15 vs. >15 (ig/m3.
The risk estimates for each effect modifier are presented in Figure 6-25. Note the wide confidence
intervals associated with each of the risk estimates, specifically for Franklin et al. (2007, 091257)
and Ostro et al. (2006, 087991). which suggests low statistical power for testing the differences
between effect modifiers.
Franklin et al. (2008, 097426) analyzed 25 cities that had PM2 5 monitoring and daily mortality
data between the years 2000-2005 (with the study period varying from city to city). The choice of the
25 communities was based on the availability of PM25 mass concentrations and daily mortality
records for at least four years, along with PM2 5 speciation data for at least 2 years between 2000 and
2005. Similar to Franklin et al. (2007, 091257). all-cause, cardiovascular, respiratory, and stroke
deaths were examined; however, of the 25 cities included in the study, only 15 overlap with the 27
cities analyzed in Franklin et al. (2007, 091257). The authors obtained mortality data from the
NCHS and various state health departments (CA, MA, MI, MN, MO, OH, PA, TX, and WA).
Although the main objective of the study was to examine the role of PM25 chemical species in the
second stage analysis (Section 6.5.2.5), the first stage analysis conducted a time-series regression of
December 2009 6-175
-------
mortality on PM2.5. In addition, the first stage regression performed a seasonal analysis in order to
take advantage of seasonal variation in PM2.5 chemical species across cities and to possibly explain
the city-to-city variation in PM2.5 mortality risk estimates. From this analysis a strong seasonal
pattern was observed with the greatest effects occurring in the spring and summer seasons (Figure
6-25).
Overall, the risk estimates for all-cause, cardiovascular, and respiratory deaths reported by
Franklin et al. (2008, 097426) are comparable to those presented in the 27 cities study (2007,
091257). as shown in Figure 6-26. When comparing the 2007 and 2008 studies conducted by
Franklin et al. (2007, 091257; 2008, 097426). although only 15 cities overlap between the two
studies and each study was designed differently (i.e., time-series vs. case-crossover), the magnitude
of the PM2.5 risk estimates reported were similar for the same averaging time, and both studies
reported a regional pattern (East > West) similar to that found in the NMMAPS studies previously
discussed.
Zanobetti and Schwartz (2009, 188462) conducted a multicity time-series study to examine
associations between PM25 and mortality in 112 U.S. cities. The cities included in this analysis
encompass the majority of cities included in the Franklin et al. (2007, 091257; 2008, 097426)
analyses. In this analysis a city represents a single county; however, 14 of the cities represent a
composite of multiple counties. In addition to examining PM25, the investigators also analyzed
PM10_2.5; these results are discussed in Section 6.5.2.3. Zanobetti and Schwartz (2009, 188462)
analyzed PM2 5 associations with all-cause, cardiovascular disease (CVD), MI, stroke, and
respiratory mortality for the years 1999-2005. To be included in the analysis, each of the cities
selected had to have at least 265 days of PM2 5 data per year and at least 300 days of mortality data
per year. The authors conducted a city- and season-specific Poisson regression to estimate excess
risk for PM25 lagged 0- and 1-days, adjusting for smooth functions (natural cubic splines) of days
(1.5 df per season), the same-day and previous day temperature (3 df each), and day-of-week. The
city specific estimates were then combined using a random effects model. Based on the assumption
that climate affects PM exposures (e.g., ventilation and particle characteristics), the investigators
combined city-specific estimates into six regions based on the Koppen climate classification scheme
(e.g., "Mediterranean climates" for CA, OR, WA, etc.).
The overall combined excess risk estimates were: 0.98 % (95% CI: 0.75, 1.22) for all-cause;
0.85 % (95% CI: 0.46-1.24) for CVD, 1.18 % (95% CI: 0.48-1.89) for MI; 1.78 % (95% CI: 0.96-
2.62) for stroke, and 1.68 % (95% CI: 1.04-2.33) for respiratory mortality for a 10 ug/m3 increase in
PM25 at lag 0-1. When the risk estimates were combined by season, the spring estimates were the
largest for all-cause and for all of the cause-specific mortality outcomes examined. For example, the
risk estimate for all-cause mortality for the spring was 2.57% (95% CI: 1.96-3.19) with the estimates
for the other seasons ranging from 0.25% to 0.95%. When examining cities that had both PM2 5 and
PM10_2.5 data (i.e., 47 cities), the addition of PM10_2.5 in the model did not alter the PM25 estimates
substantially, only decreasing slightly from 0.94% in a single pollutant model to 0.77% in a
copollutant model with PMi0_2.5. When the risk estimates were combined by climatic regions, the
estimated PM25risk for all-cause mortality were similar (all above 1% per 10 ug/m3 increase) for all
the regions except for the "Mediterranean" region (0.5%) which includes cities in CA, OR and WA,
though the estimates in that region were significantly heterogeneous (Figure 6-24).
December 2009 6-176
-------
Climatic Region
All-cause
CVD
Respiratory
Humid subtropical and maritime
V\ferm summer continental
Hot summer continental
Dry
Dry, continental
Mediterranean
-2024
% Increase
-2024
% Increase
-2
024
% Increase
Source: Data from Zanobetti and Schwartz (2009,188462).
Figure 6-23.
Percent increase in all-cause (nonaccidental) and cause-specific mortality per
10 ug/m3 increase in the average of 0- and 1-day lagged PIVb.s, combined by
climatic regions.
December 2009
6-177
-------
The PM2.5 risk estimate for all-cause mortality reported by Zanobetti and Schwartz (2009,
188462) for 112 cities (0.98% per 10 (ig/m3 increase in the average of 0- and 1-day lags) is generally
consistent with that reported by Franklin et al. (2007, 091257) for 27 cities (0.82% [0.02-1.63]) and
Franklin et al. (2008, 097426) for 25 cities (0.74% [95% CI: 0.41-1.07]) using the same 0- and 1-
day avg exposure time. The seasonal pattern (i.e., higher risk estimates in the spring) found in this
study is also consistent with the result from Franklin et al. (2008, 097426). Figure 6-23 highlights the
risk estimates for all-cause, CVD, and respiratory morality combined by region. The regional
division based on climatic types used in this study makes it difficult to directly compare the regional
pattern of results from previous studies. However, an examination of empirical Bayes-adjusted effect
estimates for each of the cities included in the analysis further confirms the heterogeneity observed
between some cities and regions of the country (Figure 6-24). It is noteworthy that, unlike
NMMAPS, which focused on PMi0 and indicated larger risk estimates in the northeast, Zanobetti
and Schwartz (2009, 188462) found that the all-cause mortality risk estimates were fairly uniform
across the climatic regions, except for the "Mediterranean" region.
-2.5 -1.5 -0.5 0.5 1.5 2.5 3.5 4.5 -2.5 -1.5 -0.5 0.5 1.5 2.5 3.5 4.5 ~2'5 "1-5 -0.5 0.5 1.5 2.5 3.5 4.5
% Increase in Total Mortality % Increase In Cardiovascular Mortality % Increase In Respiratory
Mortality
Figure 6-24. Empirical Bayes-adjusted city-specific percent increase in total (nonaccidental),
cardiovascular, and respiratory mortality per 10 ug/m3 increase in the average of
0-and 1-day lagged PM2.6 by decreasing mean 24-h avg PM2.s concentrations.
Based on estimates calculated from Zanobetti and Schwartz (2009,188462) using
the approach specified in Le Tertre et al. (2005, 087560).
December 2009 6-178
-------
Key to Figure 6-24
City
Rubidoux, CA
Bakersfield, CA
Los Angeles, CA
Fresno, CA
Atlanta, GA
Steubenville, OH
Cincinnati, OH
Birmingham, AL
Middletown, OH
Indianapolis, IN
Cleveland, OH
Dayton, OH
Columbus, OH
Detroit, Ml
Akron, OH
Louisville, KY
Chicago, IL
Pittsburgh, PA
Harrisburg, PA
Baltimore, MD
Youngstown, OH
Knoxville, TN
Gary, IN
Charlotte, NC
Warren, OH
Washington, DC
Wilmington, DE
Carlisle, PA
Mean
24.7
21.7
19.7
18.7
17.6
17.1
17.1
16.5
16.5
16.4
16.3
16.3
16.2
16.2
16.0
15.9
15.8
15.7
15.6
15.6
15.6
15.5
15.5
15.3
15.2
15.2
15.1
15.1
98th
68.0
80.3
51.1
64.9
38.2
41.4
39.9
38.8
38.4
38.2
40.5
38.3
38.3
41.0
39.0
38.0
39.1
43.1
40.2
38.8
38.1
32.9
37.5
32.7
37.4
37.2
37.6
40.0
City
Taylors, SC
Toledo, OH
Anaheim, CA
New York, NY
Washington, PA
Wnston, NC
Elizabeth, NJ
Philadelphia, PA
St. Louis, MO
Allentown, PA
Richmond, VA
Spartanburg, SC
Durham, NC
Little Rock, AR
Easton, PA
Raleigh, NC
Greensboro, NC
Mercer, PA
Annandale, VA
Nashville, TN
Dumbarton, VA
Columbia, SC
Milwaukee, Wl
New Haven, CT
Grand Rapids, Ml
El Cajon, CA
Gettysburg, PA
State College, PA
Mean
15.0
14.9
14.9
14.7
14.7
14.7
14.6
14.6
14.5
14.4
14.3
14.2
14.2
14.2
14.2
14.1
14.1
14.1
14.0
13.9
13.8
13.7
13.7
13.6
13.6
13.5
13.4
13.4
98th
32.2
36.6
44.1
38.1
37.0
34.1
38.2
36.6
33.7
38.9
33.0
31.4
32.9
31.8
39.7
31.8
31.0
36.4
34.6
31.0
31.9
30.7
36.3
36.8
36.4
34.9
36.5
38.5
City
Waukesha, Wl
Baton Rouge, LA
Memphis, TN
Erie, PA
Dallas, TX
Houston, TX
Chesapeake, VA
Wlkes-Barre, PA
Norfolk, VA
Sacramento, CA
Springfield, MA
New Orleans, LA
Ft. Worth, TX
Pensacola, FL
Davenport, IA
Avondale, LA
Boston, MA
Holland, Ml
Charleston, SC
Tampa, FL
Tulsa, OK
Kansas, MO
Scranton, PA
Hartford, CT
Minneapolis, MN
Worcester, MA
Salt Lake, UT
Providence, Rl
Mean
13.4
13.4
13.3
12.9
12.8
12.8
12.8
12.8
12.7
12.6
12.5
12.5
12.4
12.3
12.3
12.3
12.3
12.1
12.1
12.1
12.1
12.0
11.9
11.8
11.6
11.5
11.5
11.5
98th
35.3
30.1
32.4
36.1
28.7
27.5
29.8
32.5
29.6
45.0
35.1
29.0
27.7
31.2
32.1
28.6
30.2
35.0
27.9
25.8
32.3
28.6
33.0
33.5
31.6
30.2
52.4
30.5
City
Phoenix, AZ
Tacoma, WA
Port Arthur, TX
Cedar Rapids, IA
Dodge, Wl
Oklahoma, OK
Des Moines, IA
Jacksonville, FL
Omaha, NE
Denver, CO
Pinellas, FL
Austin, TX
Orlando, FL
Klamath, OR
Seattle, WA
Medford, OR
Bath, NY
Provo, UT
Miami, FL
El Paso, TX
Spokane, WA
San Antonio, TX
Portland, OR
Davie, FL
Eugene, OR
Palm Beach, FL
Bend, OR
Albuquerque, NM
Mean
11.4
11.4
11.1
11.0
10.9
10.8
10.5
10.5
10.5
10.5
10.4
10.4
10.3
10.2
10.1
10.0
9.6
9.5
9.4
9.0
8.9
8.9
8.9
8.4
8.1
7.8
7.7
6.6
98th
30.7
38.1
25.7
31.0
32.9
26.1
27.9
25.3
28.0
26.4
23.1
24.5
24.3
40.7
27.9
37.3
29.3
38.5
20.5
24.4
30.6
21.9
25.4
19.1
29.9
18.4
23.5
17.9
Note: The top effect estimate in the figures represents the overall effect estimate for that mortality outcome across all cities. The remaining effect estimates are ordered by the highest (i.e., Rubidoux, CA) to lowest (i.e.,
Albuquerque, NM) mean 24-h PM2 5 concentrations across the cities examined. In the key the cities are reported in this order, which represents the policy relevant concentrations for the annual standard, but the policy
relevant PM25 concentrations for the daily standard (i.e., 98th percentile of the 24-h average) are also listed for each city (from Zanobetti and Schwartz (2009,188462))
PM2.s-Mortality Associations on a Regional Scale: California
Ostro et al. (2006, 087991) examined associations between PM2.5 and daily mortality in nine
heavily populated California counties (Contra Costa, Fresno, Kern, Los Angeles, Orange, Riverside,
Sacramento, San Diego, and Santa Clara) using data from 1999 through 2002. The authors used a
two-stage model to examine all-cause, respiratory, cardiovascular, ischemic heart disease, and
diabetes mortality individually and by potential effect modifier (i.e., age, gender, race, ethnicity, and
education level). The a priori exposure periods examined included the average of 0- and 1-day lags
(lag 0-1) and the 2-day lag (lag 2). The authors selected these non-overlapping lags (i.e., rather than
selecting lag 1 as the single-day lag) because previous studies have reported stronger associations at
lags of 1 or 2 days or with cumulative exposure over three days. It is unclear why the investigators
chose these non-overlapping lags (i.e., single-day lag of 2 instead of 1) even though they state they
based the selection of their lag days on results presented in previous studies, which found the
strongest association for PM lagged 1 or 2 days. Using the average of 0- and 1-day lags Ostro et al.
December 2009
6-179
-------
(2006, 087991) reported combined estimates of: 0.6% (95% CI: 0.2-1.0), 0.6% (95% CI: 0.0-1.1),
0.3% (95% CI: -0.5 to 1.0), 2.2% (95% CI: 0.6-3.9), and 2.4% (95% CI: 0.6-4.2) for all-cause,
cardiovascular, ischemic heart disease, respiratory, and diabetes deaths, respectively, per 10 (ig/m3.
The authors also conducted a sensitivity analysis of risk estimates based on the extent of temporal
adjustment, which showed monotonic reductions for all of the death categories examined when 4, 8,
and 12 degrees of freedom per year were used.
Five of the nine counties examined in the Ostro et al. (2006, 087991) analysis contain cities
that are among the 27 cities examined in the Franklin et al. (2007, 091257) analysis for the same
period, 1999-2002. While the lags used were different between these two studies, both presented
PM2.5 risk estimates in individual cities or counties (graphically in the Franklin et al. study (2007,
091257); in a table in the Ostro et al. study (2006, 087991)). which allowed for a cursory evaluation
of consistency between the two analyses. In Franklin et al. (2007, 091257). PM2.5 risk estimates at
lag 1 day for the cities Los Angeles and Riverside were slightly negative, whereas Fresno,
Sacramento, and San Diego showed positive values above 1% per 10 (ig/m3 increase in PM25. The
2-day lag result presented in Ostro et al. (2006, 087991) is qualitatively consistent, with Los Angeles
and Riverside, both of which show slightly negative estimates, while the other 3 locations all show
positive, but somewhat smaller estimates, than those reported by Franklin et al. (2007, 091257). The
estimates for the average of 0- and 1-day lags for these five counties in Ostro et al. (2006, 087991).
which contain cities examined in Franklin et al. (2007, 091257). were all positive. Thus, these two
PM2.5 studies showed some consistencies in risk estimates even though they used different lag
periods and a different definition for the study areas of interest (i.e., counties vs. cities). The risk
estimates for Franklin et al. (2007, 091257) and Ostro et al. (2006, 087991). stratified by various
effect modifiers (e.g., gender, race, etc.), are summarized in Figure 6-25. Of note is the contrast in
the results presented for the effect modification analysis for "in-hospital" versus "out-of-hospital"
deaths for Ostro et al. (2006, 087991). which differs from the results presented in the PM10 study
conducted by Zeka et al. (2006, 088749). Ostro et al. (2006, 087991) observed comparable risk
estimates for "in-hospital" vs. "out-of-hospital" deaths, whereas Zeka et al. (2006, 088749) observed
a large difference between the two in the 20 cities study discussed earlier. This difference in effects
observed between the two studies is more than likely due to the compositional differences in PMi0 in
the cities examined in Zeka et al. (2006, 088749) (i.e., PMi0 more or less dominated by PM25 and the
subsequent composition of PM25).
December 2009 6-180
-------
Study
Effect Modifier
Effect Estimate (95% Cl)
Ostro et al. (2006, 0879911, 9 CA counties, Lag 0-1
All-cause
Age > 65
Male
Female
White
Black
Hispanic
In-hospital
Out-of-hospital
> High School
< High School
Franklin et al. (2007, 0912571, 27 U.S. cities, Lag 1
All-cause
Age > 75
Age < 75
Male
Female
East
West
PM25>15|jg/m3
PM25<15|jg/m3
25th percentile air conditioning
75th percentile air conditioning
Summer peaking PM2 5 cities: 25th percentile air conditioning
75th percentile air conditioning
Franklin et al. (2008, 0974261, 25 U.S. cities, Lag 0-1
All-cause
Wnter
Spring
Summer
Fall
West
East & Central
Zanobetti and Schwartz (2009,1884621,112 U.S. cities, Lag 0-1
All-cause
Wnter
Spring
Summer
Autumn
-2.0
0.0
% Increase
i
1.0
I
2.0
3.0
I
4.0
Figure 6-25. Summary of percent increase in all-cause (nonaccidental) mortality per 10 ug/m3
increase in PM2.s by various effect modifiers.
December 2009
6-181
-------
PM2.s-Mortality Associations in Canada
An analysis of multiple pollutants, including PM2.5, in 12 Canadian cities found the most
consistent associations for NO2 (Burnett et al, 2004, 086247). In this analysis, PM2.5 was only
measured every 6th day in much of the study period, and the simultaneous inclusion of NO2 and
PM2.5 in a model on the days when PM2 5 data were available eliminated the PM2 5 association (from
0.60% to -0.10% per 10 (ig/m3 increase in PM25). However, the investigators noted that during the
later study period of 1998-2000 when daily TEOM PM25 data were available for 11 of the 12 cities,
a simultaneous inclusion of NO2 and PM25 resulted in considerable reduction of the NO2 risk
estimate, while the PM2.5 risk estimate was only slightly reduced from 1.1% to 0.98% (95% CI: -0.16
to 2.14). Thus, the relative importance of NO2 and PM25 on mortality effect estimates has not been
resolved when using the Canadian data sets.
Summary of PM2.5 Risk Estimates
The risk estimates for all-cause mortality for all ages ranged from 0.29% Dominici et al.
(2007, 097361) to 1.21% Franklin et al. (2007, 091257) per 10 (ig/m3 increase in PM2.5 (Figure
6-26). An examination of cause-specific risk estimates found that PM2 5 risk estimates for
cardiovascular deaths are similar to those for all-cause deaths (0.30-1.03%), while the effect
estimates for respiratory deaths were consistently larger (1.01-2.2%), albeit with larger confidence
intervals, than those for all-cause or cardiovascular deaths using the same lag/averaging indices.
Figure 6-27 summarizes the PM25 risk estimates for all U.S.- and Canadian-based studies by
cause-specific mortality.
An examination of lag structure observed results similar to those reported for PMi0 with most
studies reporting either single day lags or two-day avg lags with the strongest effects observed on lag
1 or lag 0-1. In addition, seasonal patterns of PM25 risk estimates were found to be similar to those
reported for PMip, with the warmer season showing the strongest association. An evaluation of
regional associations found that in most cases the eastern U.S. had the highest PM25 mortality risk
estimates, but this was dependent on the geographic designations made in the study. When grouping
cities by climatic regions, similar PM2 5 mortality risk estimates were observed across the country
except in the Mediterranean region, which included CA, OR, and WA.
Of the studies evaluated, only Burnett et al. (2004, 086247). a Canadian multicity study,
analyzed gaseous pollutants and found mixed results, with possible confounding of PM25 risk
estimates by NO2. Although the recently evaluated U.S.-based multicity studies did not analyze
potential confounding of PM25 risk estimates by gaseous pollutants, evidence from single-city
studies evaluated in the 2004 PM AQCD (U.S. EPA, 2004, 056905) suggest that gaseous
copollutants do not confound the PM2 5-mortality association, which is further supported by studies
that examined the PMio-mortality relationship.
December 2009 6-182
-------
Study
Cause-Specific Mortality
Effect Estimate (95% Cl)
Dominici et al. (2007, 0991351, NMMAPS
Franklin et al. (2007, 0912571, 27 U.S. cities
Franklin et al. (2008, 0974261, 25 U.S. cities
All-cause, lag 1
Cardio-respiratory, lag 1
All-cause, lag 1
Cardiovascular, lag 1
Respiratory, lag 1
All-cause, lag 0-1
Cardiovascular, lag 0-1
Respiratory, lag 0-1
All-cause, lag 0-1
Cardiovascular, lag 0-1
Respiratory, lag 1-2
Zanobetti and Schwartz (2009, 1884621,112 U.S. cities
Ostro et al. (2006, 0879911, 9 CA counties
All-cause, lag 0-1
Cardiovascular, lag 0-1
Respiratory, lag 0-1
All-cause, lag 0-1
Cardiovascular, lag 0-1
Respiratory, lag 0-1
-0.5 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
% Increase
Figure 6-26. Summary of percent increase in all-cause (nonaccidental) and cause-specific
mortality per 10 ug/m3 increase in PM2.5from recent multicity studies.
December 2009
6-183
-------
Study
Location
Burnett and Goldberg (2003, 042798)* 8 Cities, Canada
Klemm and Mason (2003, 042801)* 6 Cities, U.S.
Moolgavkar (2003, 051316)* Los Angeles, CA
Ito (2003, 042856)* Detroit, Ml
Fairley (2003, 042850)* Santa Clara County, CA
Tsai et al. (2000, 006251)* Newark, NJ
Elizabeth, NJ
Camden, NJ
Chock et al. (2000, 010407)* Pittsburgh, PA
Dominici et al. (2007, 097361) 100 Cities, U.S.
Zanobetti and Schwartz (2009, 188462) 112 Cities, U.S.
Franklin et al. (2007, 091257) 27 Cities, U.S.
25 Cities, U.S.
Burnett et al. (2004, 086247) 1 2 Cities, Canada
Ostro et al. (2006, 087991) 9 Counties, CA
Slaughter et al. (2005, 073854) Spokane, WA
Klemm et al. (2004, 056585) Atlanta, GA
Villeneuve et al. (2003, 055051) Vancouver, Canada
Tsai et al. (2000, 006251)* Newark, NJ
Elizabeth, NJ
Camden, NJ
Dominici et al. (2007, 097361) 100 Cities, U.S.
Klemm and Mason (2003, 042801)* 6 Cities, U.S.
Ostro et al. (1 995, 0791 97)* Southern CA
Lipfert et al. (2000, 004088)* Philadelphia, PA
Moolgavkar (2003, 051316)* Los Angeles, CA
Ito (2003, 042856)* Detroit, Ml
Mar et al. (2003, 042841)* Phoenix, AZ
Fairley (2003, 042850)* Santa Clara County, CA
Zanobetti and Schwartz (2009, 188462) 112 Cities, U.S.
Franklin et al. (2007, 091257) 27 Cities, U.S.
Franklin et al. (2008, 097426) 25 Cities, U.S.
Ostro et al. (2007, 091 354) 9 Counties, CA
Ostro et al. (2006, 087991) 9 Counties, CA
Holloman et al. (2004, 087375) 7 Counties, NC
Wilson et al. (2007, 1 571 49) Phoenix, AZ
Villeneuve et al. (2003, 055051) Vancouver, Canada
Klemm and Mason (2003, 042801)* 6 Cities, U.S.
Ostro et al. (1 995, 0791 97)* Southern California
Moolgavkar (2003, 051 31 6)* Los Angeles, CA
Ito (2003, 042856)* Detroit, Ml
Fairley (2003, 042850)* Santa Clara County, CA
Zanobetti and Schwartz (2009, 188462) 112 Cities, U.S.
Franklin et al. (2007, 091257) 27 Cities, U.S.
Franklin et al. (2008, 097426) 25 Cities, U.S.
Ostro et al. (2006, 087991) 9 Counties, CA
Villeneuve et al. (2003, 055051) Vancouver, Canada
"Studies represent the collective
evidence from the 2004 PM AQCD (2004, 056905).
Lag Age Effect Estimate (95% Cl)
1
0-1
1
3 ->
0
0
0
1
0-1
1
0-1
1
0-1
0-1 65+
0
0
1
0-1
0 -i
1
1 -i
1
0-1
1 J
0-1
3
0-1
0 J
0-1
0-1
1-2
0-1
— •— Nonaccidental
-•—
i— •
—
*
«•
.0.
-•-
*
— •— Cardio-respiratory
•«-
'— • — Cardiovascular
«-
r-»
i — •
r»-
n -„._„,
— • —
*
— •
I I I I I I I I I I I
-5-3-1 1 3 5 7 9 11 13 15
% Increase
Figure 6-27. Summary of percent increase in all-cause (nonaccidental) and cause-specific
mortality per 10 ug/m3 increase in PM2.6 for all U.S.- and Canadian-based studies.
The three vertical lines for the Wilson et al. (2007,157149) estimate represent the
central, middle, and outer Phoenix estimates.
6.5.2.3. Thoracic Coarse Particles (PM10.2.5)
In the 2004 PM AQCD (U.S. EPA, 2004, 056905). a limited number of studies, mostly single-
city analyses, were evaluated that examined thoracic coarse (PMi0_2.5) PM for its association with
mortality. Of these studies a small number examined both PM25 and PMi0_2.5 effects, and found some
evidence for PMi0_2.5 effects of the same magnitude as PM25. However, multiple limitations in these
studies were identified including measurement and exposure issues for PM10_2.5 and the correlation
between PM2 5 and PMi0_2.5. These limitations increased the uncertainty surrounding the
concentrations at which PMi0_2.5-mortality associations are observed.
December 2009
6-184
-------
A thorough analysis of PM10_2.5 mortality associations requires information on the speciation of
PMio_2.5. This is because, while a large percent of the composition of coarse particles may consist of
crustal materials by mass, depending on available sources, the surface chemical characteristics of
PMio_2.5 may also vary from city to city. Thus, without information on the chemical speciation of
PMio_2.5, the apparent variability in observed associations between PMi0_2.5 and mortality across cities
is difficult to characterize. Although this type of information is not available in the current literature,
the relative importance of the associations observed between PMi0_2.5 and mortality in the following
studies is of interest.
PM10.2.5 Concentrations Estimated Using the Difference Method
The Zanobetti and Schwartz (2009, 188462) multicity analysis, described for PM2.5
section (Section 6.5.2.2), also examined the association between computed PMi0_2.5 and all-causes,
cardiovascular disease (CVD), MI, stroke, and respiratory mortality for the years 1999-2005. Of the
112 cities included in the PM2.5 analysis only 47 cities had both PM2.s and PMi0 data available.
PMio_2.5 was estimated in these cities by differencing the county wide averages of PMi0 and PM2.5. In
addition to examining the association between PMi0_2.5 and mortality for the average of lags 0 and
1 day, the investigators also considered a distributed lag of 0-3 days. The risk estimates for PM10_2.5
were presented for both a single pollutant model and a copollutant model with PM2.5, and were also
combined by season and climatic regions as was done in the PM2.5 analysis.
The study found a significant association between the computed PMiq_2.5 and all-cause, CVD,
stroke, and respiratory mortality. The combined estimate for the 47 cities using the average of 0- and
1-day lag PMio.2.5 for all-cause mortality was 0.46% (95% CI: 0.21-0.71) per 10 (ig/m3 increase. The
estimate obtained using the distributed lag model was smaller (0.31% [95% CI: 0.00-0.63]). The
seasonal analysis showed larger risk estimates in the spring for all-cause (1.01%) and respiratory
mortality (2.56%) (i.e., the same pattern observed in the PM2.5 analysis); however, for CVD
mortality, the estimates for spring (0.95%) and summer (1.00%) were comparable. When the risk
estimates were combined by climatic region (Figure 6-28), for all-cause mortality, the "dry,
continental" region (which included Salt Lake City, Provo, and Denver, all of which had relatively
high estimated PMi0_2.5 concentrations) showed the largest risk estimate (1.11% [95% CI:
0.11-2.11]), but the "dry" region (which included Phoenix and Albuquerque, the two cities with high
PMio_2.5 concentrations) and the "Mediterranean" region (which included cities in CA, OR, and WA)
did not show positive associations. The other three regions (i.e., "hot summer, continental," "warm
summer, continental," and "humid, subtropical and maritime"), which included cities that correspond
to the mid-west, southeast, and northeast geographic regions as defined in previous NMMAPS
analyses, all showed significantly positive associations. Similar regional patterns of associations
were found for CVD and respiratory mortality, which are further confirmed when examining the
empirical Bayes-adjusted city-specific estimates in Figure 6-29. The regional pattern of associations
for MI and stroke are less clear, because of the wider confidence intervals due to the smaller number
of deaths in these specific categories. The lack of a PMi0_2.5-mortality association in the "dry" region
reported in this study is in contrast to the results from three studies that analyzed Phoenix data and
found associations, as reviewed in the 2004 PM AQCD (U.S. EPA, 2004, 056905). and Wilson et al.
(2007, 157149) (discussed below).
Although the results from this analysis are informative because it is the first multicity U.S.-
based study that examined the association between short-term exposure to PMi0_2.5 and mortality on a
large scale, some limitations do exist. Specifically, it is not clear how the computed PMi0_2.5
measurements used by Zanobetti and Schwartz (2009, 188462) compare with the PMi0_2.5
concentrations obtained by directly measuring PM10_2.5 using a dichotomous sampler, or the PM10_2.5
concentrations computed using the difference of PMi0 and PM2.5 measured at co-located samplers.
Additional studies evaluated the association between short-term exposure to PMi0_2.5 and
mortality using PMi0_2.5 concentrations estimated by subtracting PMi0 from PM2.5 concentrations at
co-located monitors. Although PMi0_2.5 concentrations estimated using this approach are not ideal,
the results from these studies are informative in evaluating the PMi0_2.5 mortality association.
December 2009 6-185
-------
Climatic Region
All-cause
CVD
Respiratory
Humid subtropical and maritime
V\ferm summer continental
Hot summer continental
Dry
Dry continental
Mediterranean
i i i I i i i i
-3-2-10 1 2 3 4
I I T
-3-2-101234 -3-2-101234
% Increase
% Increase
% Increase
Source: Data from Zanobetti and Schwartz (2009,188462).
Figure 6-28. Percent increase in all-cause (nonaccidental) and cause-specific mortality per
10 ug/m3 increase in the average of 0- and 1-day lagged PMio-2.5, combined by
climatic regions.
December 2009
6-186
-------
0.5 1.0 1.5 2.0 2.5 3.0 3.5 -2.5
05 10 15 2.0 2.5 3.0 3.5
0.5 1.0 1.5 2.0 2.5 3.0 3.5
% Increase in Total Mortality
% Increase In Cardiovascular Mortality % Increase In Respiratory Mortality
Figure 6-29. Empirical Bayes-adjusted city-specific percent increase in total (nonaccidental),
cardiovascular, and respiratory mortality per 10 ug/m3 increase in the average of
0- and 1-day lagged PM10.2.s by decreasing 98th percentile of mean 24-h avg
PM10.2.6 concentrations. Based on estimates calculated from Zanobetti and
Schwartz (2009,188462) using the approach specified in Le Tertre et al. (2005,
087560).
December 2009
6-187
-------
Key for Figure 6-29
City
El Paso, TX
St. Louis, MO
Phoenix, AZ
Detroit, Ml
Gary, IN
Omaha, NE
Albuquerque, NM
New Haven, CT
Bakersfield, CA
Des Moines, IA
Denver, CO
Salt Lake, UT
98th
105.1
81.9
80.1
77.5
71.3
65.6
64.3
58.4
55.9
55.0
53.8
52.6
Mean
25.4
15.2
33.3
17.3
6.9
24.7
22.9
11.9
16.1
16.2
18.1
19.2
City
Cleveland, OH
Davenport, IA
Birmingham, AL
Provo, UT
Chicago, IL
Easton, PA
Steubenville, OH
Columbia, SC
Los Angeles, CA
Spokane, WA
Columbus, OH
Pittsburgh, PA
98th
51.2
49.9
49.6
49.3
46.1
43.9
43.5
42.9
42.5
41.8
40.0
32.0
Mean
15.2
15.3
14.2
18.2
12.4
12.0
12.1
8.4
13.5
13.8
11.2
9.4
City
Sacramento, CA
Tampa, FL
Toledo, OH
Washington, PA
Allentown, PA
Atlanta, GA
Davie, FL
Taylors, SC
Memphis, TN
Seattle, WA
Baltimore, MD
Cincinnati, OH
98th
31.5
29.1
28.8
27.8
27.8
27.4
25.5
25.4
24.3
23.7
23.5
23.3
Mean
10.2
12.9
7.6
6.5
4.5
8.6
9.4
8.0
9.3
9.0
8.9
7.8
City
Louisville, KY
Wlkes-Barre, PA
New York, NY
Wlmington, DE
Raleigh, NC
Scranton, PA
Harrisburg, PA
Akron, OH
Charleston, SC
Wnston, NC
Erie, PA
98th
23.3
22.2
22.0
21.8
20.9
19.2
18.6
17.7
17.6
16.5
14.9
Mean
8.3
6.2
6.4
7.0
6.9
6.1
5.4
5.3
6.6
7.4
3.1
Note: The top effect estimate in the figures represents the overall effect estimate for that mortality outcome across all cities. The remaining effect estimates are ordered by the highest (i.e., El Paso, TX) to
lowest (i.e., Erie, PA) 98th percentile of the mean 24-h PMW7 5 concentrations across the cities examined, which is the policy relevant concentration for the daily standard [from Zanobetti and Schwartz (2009,
18846211.
Slaughter et al. (2005, 073854) examined the association of various PM size fractions (PMi,
PM2.5, PMio, PMio_2.5) and CO with ED visits, HAs, and mortality in Spokane, WA for the period
1995-2001. Although the authors did not report mortality risk estimates for PM10_2.5, they did not find
an association between any PM size fraction (or CO) and mortality or cardiac HAs at lags of
0-3 days.
Wilson et al. (2007, 157149) examined the association between size-fractionated PM (PM2.5
and PMio_2.5) and cardiovascular mortality in Phoenix for the study period 1995-1997, using
mortality data aggregated for three geographic regions: "Central Phoenix," "Middle Ring," and
"Outer Phoenix," which were constructed as a composite of zip codes of residence in order to
compare population size among the three areas. The authors reported apparently different patterns of
associations between PM2.5 and PMi0_2.s in terms of the size of the risk estimate across the three areas
and temporal patterns of associations. In the "Middle Ring" where PMi0_2.5 showed the strongest
association, the estimated risk per 10 (ig/m3 increase for a 1 day lag was 3.4% (95% CI: 1.0-5.8).
The estimated risk for PM2.5 found for "Central Phoenix" was 6.6% (95% CI: 1.1-12.5) for lag 1.
The authors speculated that the apparent difference in estimated risks across the areas might be due
to the lower SES in "Central Phoenix" or the lower exposure error, but the relatively wide
confidence bands of these estimates make it difficult to establish such relationships (Section 8.1.7 for
a detailed discussion on SES and susceptibility to PM exposure).
Kettunen et al. (2007, 091242) analyzed UFPs, PM2.5, PMi0, PMi0_2.5, and gaseous pollutants
for their associations with stroke mortality in Helsinki during the study period of 1998-2004. The
authors did not observe an association between air pollution and mortality for the whole year or cold
season, but they did find associations for PM2.5 (13.3% [95% CI: 2.3-25.5] per 10 ug/m3), PMi0, and
CO during the warm season, most strongly at lag 1 day. An association was also observed for
PMio_2.5 during the warm season (7.8% [95% CI: -7.4 to 25.5] per 10 (ig/m3 at lag 1 day); however, it
was weaker than PM2 5.
The Perez et al. (2008, 156020) analysis tested the hypothesis that outbreaks of Saharan dust
exacerbate the effects of PM25 and PM10_2.5 on daily mortality. Changes of effects between Saharan
and non-Saharan dust days were assessed using a time-stratified case-crossover design involving
24,850 deaths between March 2003 and December 2004 in Barcelona, Spain. Saharan dust days
were identified from back-trajectory and satellite images. Chemical speciation, but not an analysis
for microbes or fungi, was conducted approximately once a week during the study period. On
Saharan dust days, mean concentrations were 1.2 times higher for PM25 (29.9 (ig/m3) and 1.1 times
higher for PM10_25 (16.4 (ig/m3) than on non-Saharan dust days. During Saharan dust days (90 days
out of 602), the PM 10-2.5 risk estimate was 8.4% (95% CI: 1.5-15.8) per 10 (ig/m3 increase at lag 1
day, compared with 1.4% (95% CI: -0.8 to 3.4) during non-Saharan dust days. In contrast, there was
not an additional increased risk of daily mortality for PM25 during Saharan dust days (5.0%
December 2009
6-188
-------
[95% CI: 0.5-9.7]) compared with non-Saharan dust days (3.5% [95% CI: 1.6-5.5]). However,
differences in chemical composition (i.e., PM25 was primarily composed of nonmineral carbon and
secondary aerosols; whereas PMi0_2.5 was dominated by crustal elements) did not explain these
observations. Note also when examining all days combined, both size fractions were associated with
mortality, but the PM2.5 association was found to be stronger.
PM10.2.5 Concentrations Directly Measured
In Burnett et al. (2004, 086247). which analyzed the association of multiple pollutants with
mortality in 12 Canadian cities, described previously; the authors also examined PMi0_2.5. In this
study the authors collected PMi0_2.5 using dichotomous samplers with an every-6th-day schedule.
When both NO2 and PM10 2 5 were included in the regression model, the PM10 2 5 effect estimate was
reduced from 0.65% (95% CI: -0.10 to 1.4) to 0.31% (95% CI: -0.49 to 1.1) per 10 (ig/m3 increase in
1-day lag PMi0_2.5. These risk estimates are similar to those reported for PM25, which were also
reduced upon the inclusion of NO2 in the two-pollutant model, but to a greater extent, from 0.60%
(95% CI: -0.03 to 1.2) to -0.1% (95% CI: -0.86 to 0.67).
Villeneuve et al. (2003, 055051) analyzed the association between PM2.5, PMi0_2.5, TSP, PMi0,
SO42~, and gaseous copollutants in Vancouver, Canada, using a cohort of approximately 550,000
whose vital status was ascertained between 1986 and 1999. In this study PM25 and PMi0_2.5 were
directly measured using dichotomous samplers. The authors examined the association of each air
pollutant with all-cause, cardiovascular, and respiratory mortality, but only observed significant
results for cardiovascular mortality at lag 0 for both PMi0_2.5 and PM2 5. They found that PMi0_2.5
(5.4% [95% CI: 1.1-9.8] per 10 (ig/m3), was more strongly associated with cardiovascular mortality
than PM2.5 (4.8% [95% CI: -1.9 to 12.0] per 10 ug/m3).
Klemm et al. (2004, 056585) analyzed various components of PM and gaseous pollutants for
their associations with mortality in Fulton and DeKalb Counties, Georgia for the 2-yr period,
1998-2000. PMio_2.5 concentrations were obtained from the ARIES database, which directly
measured PMi0_2.5 using dichotomous samplers. In this analysis the authors adjusted for temporal
trend using quarterly, monthly, and biweekly knots, and reported estimates for all-cause, circulatory,
respiratory, cancer, and other causes mortality for each scenario. Overall, PM2 5 was, more strongly
associated with mortality than PMi0_2.5. For example, using the average of 0- and 1-day lags, the risk
estimates for PM25 and PMi0_25 in the monthly knots model for all-cause mortality, ages > 65 yr
were 5.6% (95% CI: 1.9-9.5) and 6.4% (95% CI: -0.5 to 14.1) per 10 (ig/m3 increase, respectively.1
Summary of PM10.2.5 Risk Estimates
The results from newly available studies that examined the association between short-term
exposure to PMi0_2.5 primarily consisted of single-city studies. Collectively these studies found
consistent, positive associations, with the precision of each association varying by study location.
The evidence from those single-city studies conducted in the U.S. and Canada in combination with
the multicity studies evaluated (i.e., in the U.S. and Canada), provide evidence for PMi0_2.5 effects.
However, the various methods used to estimate exposure to PMi0_2.5 (e.g., direct measurement of
PMio_2.5 using dichotomous samplers, calculating the difference between PMi0 and PM2 5
concentrations) in the studies evaluated add uncertainty to the associations observed. Specifically, a
new U.S. multicity study (Zanobetti and Schwartz, 2009, 188462) estimated PMi0_2.5 by calculating
the difference between the county-average PMi0 and PM2 5 concentrations. Although there are
limitations in the method used by Zanobetti and Schwartz (2009, 188462) associations between
PMio_2.5 and mortality were observed throughout multiple regions of the country. However, some of
the findings of this new multicity study (e.g., no associations in "dry" region where PMi0_2.5 levels
are high) are not consistent with the findings of the PMi0_2.5 studies evaluated in the 2004 PM AQCD
(U.S. EPA, 2004, 056905). and suggest that the coarse fraction is associated with mortality in areas
of the U.S. where PMi0_2.5 levels are not high. Limitations also exist in the PMi0_2.5 associations
reported due to the small number of PM10_2.5 studies that have investigated confounding by gaseous
1 The monthly knot model was selected for comparison because, overall, PM2.5 showed the strongest association with all-cause mortality
among the 15 air pollution indices examined when using this model.
December 2009 6-189
-------
copollutants or the influence of model specification on PM10_2.5 risk estimates. Additionally, more
data is needed to characterize the chemical and biological components that may modify the potential
toxicity of PMio_2.5. Figure 6-30 summarizes the PMi0_2.5 risk estimates for all U.S.-, Canadian-, and
international-based studies by cause-specific mortality.
Study
Location
Lag Age**
Effect Estimate (95% Cl)
Klemm et al. (2003,042801)*
Burnett and Goldberg (2003,042798)*
Ito (2003.042856)*
Fairley (2003,042850)*
Chock etal. (2000.010407)*
Zanobetti and Schwartz (2009,188462)
Burnett etal. (2004.086247)
Klemm etal. (2004.056585)
Villeneuve et al. (2003,055051)
Lipfertetal. (2000,004088)*
Ito (2003.042856)*
Mar etal. (2003.042841)*
Fairley (2003,042850)*
Ostro etal. (2003.042824)*
Zanobetti and Schwartz (2009,188462)
Wilson etal. (2007.157149)
Villeneuve et al. (2003,055051)
Ito (2003.042856)*
Fairley (2003,042850)*
Zanobetti and Schwartz (2009,188462)
Villeneuve et al. (2003,055051)
6 Cities, U.S. 0-1
8 Cities, Canada 1
Detroit, Ml 1
Santa Clara County, CA 0
Pittsburgh, PA 0
0
47 Cities, U.S. 0-1
12 Cities, Canada 1
Atlanta, GA 0-1
Vancouver, Can 0
Philadelphia, PA 1
Detroit, Ml 1
Phoenix, AZ 0
Santa Clara County, CA 0
Coachella Valley, CA 0
47 Cities, U.S. 0-1
Phoenix, AZ 0-5
Vancouver, Can 0
Detroit, Ml 2
Santa Clara County, CA 0
47 Cities, U.S. 0-1
Vancouver, Canada 0
<75
75+
65+
65+
65+
Nonaccidental
Cardiovascular
H—h
Respiratory
-• $
* Slides represent the collective evidence from the 2004 PM AQCD (2004, 0569051.
**lf age not specified, study included all ages.
-5 -3-11 3
% Increase
9 11 13 15
Figure 6-30. Summary of percent increase in total (nonaccidental) and cause-specific
mortality per 10 ug/m3 increase in PMi0-2.5 for all U.S.-, Canadian-, and
international-based studies. The three vertical lines for the Wilson et al. (2007,
157149) estimate represent the central, middle, and outer Phoenix estimates.
6.5.2.4. Ultrafine Particles
The 2004 PM AQCD (U.S. EPA, 2004, 056905) reviewed Wichmann et al.'s (reanalyzed by
Stolzel et al., 2003, 042842: 2000, 013912) study of fine and ultrafine particles (UFPs) (diameter:
0.01-0.1 (im) in Erfurt, Germany, for the study period 1995-1998. Stolzel et al. (2007, 091374)
extended the study period to include the years 1995-2001 and updated the analysis. Number
concentrations (NC) for four size ranges of UFPs (0.01-0.1, 0.01-0.03, 0.03-0.05, and 0.05-0.1 (im)
as well as mass concentration (MC) for three size ranges (0.01-2.5, 0.1-0.5, and 10 (im) were
analyzed. The authors found associations with UFP NC and all-cause as well as cardio-respiratory
mortality, each for a 4 day lag. The risk estimates associated with a 9,748/cm3 increase in UFP NC
was 2.9% (95% CI: 0.3-5.5) for all-cause mortality and 3.1% (95% CI: 0.3-6.0) for
cardio-respiratory mortality. The UFP-mortality association, and the lag structure of association, is
consistent with the results from their earlier analysis, but the PM2.5 association found in the previous
study was not observed in the updated analysis. Both UFP and PM2.5 concentrations were higher
during the cold season in this locale.
Breitner et al. (2009, 188439) analyzed UFP data from Erfurt, Germany, over a 10.5-yr period
(October 1991-March 2002) after the German unification, when air quality improved. In this analysis
associations between all-cause mortality and UFPs and PM2.5 were analyzed from September 1995 to
March 2002, while PMi0, NO2 and CO was analyzed for the whole study duration. The exposure
time window / averages used in this study were different from those used by Stolzel (2003, 042842)
and Stolzel et al. (2007, 091374). Breitner et al. (2009, 188439) investigated the cumulative effect of
air pollution on mortality at lags 0-5 and 0-14, using (a) a semiparametric Poisson regression model;
and (b) a third degree polynomial distributed lag (PDL) model. The authors estimated the mortality
risk for the entire study period as well as specific time periods to examine the effect of declining air
pollution levels on the air pollution-mortality association. Of the air pollutants examined, UFPs were
December 2009
6-190
-------
found to be most consistently associated with mortality. NO2 and CO were also found to be
significantly associated with mortality using the 15-day PDL and 15-day avg models, respectively.
PM2.5 and PMi0 also showed positive, but much weaker associations with mortality. In this data set,
UFPs were only moderately correlated with PM25 (r = 0.48) and PMi0 (r = 0.57). Of the pollutants
examined, NO2 showed the strongest (but overall a moderate) correlation with UFPs (r = 0.62).
When the risk estimates were compared between the two latter time periods of the study (September
1995-February 1998; and March 1998-March 2002), the estimates obtained using the 6-day avg for
these pollutants generally declined. For example, the all-cause mortality risk estimates associated
with a 8,439/cmyincrease in UFP NC was 5.5% (95% CI: 1.1-10.5) for the earlier period and -1.1%
(95% CI: -6.8 to 4.9) for the later period. However, such patterns were less clear when using 15-
day avg pollutant concentrations. In summary, UFPs appear to be the pollutant most consistently
associated with mortality in Erfurt, Germany, but combined with the results for NO2 and CO, these
associations may implicate the role of local combustion sources on the mortality association
observed.
Kettunen et al.'s (2007, 091242) study in Helsinki also examined the relationship between
UFPs and stroke mortality. As described earlier, PM2 5, PMi0, and CO was associated with stroke
mortality only during the warm season. The association with UFPs was borderline non-significant
(8.5% [95% CI: -1.2 to!9.1] per 4,979/cm3 increase in UFPs at lag 1 day), but its lag structure of
association and the magnitude of the effect estimate per interquartile-range are similar to those for
PM2 5. Note that the UFP NC levels in Helsinki (median equals 8,986/cm during the cold season and
7,587/cm3 during the warm season) are lower than those in Erfurt (mean = 13,549/cm3), but clearly
higher in the cold season.
Summary of UFP Risk Estimates
Only a few new studies, all of them conducted in Europe, examined and reported associations
between UFPs and mortality. In Erfurt, UFPs showed the strongest associations with mortality
among all of the PM indices, but its lag structure of association is either unique with the strongest
association at lag 4 days in Stolzel et al. (2007, 091374). not consistent with the lag structure of
associations found in other mortality studies, or the time-windows examined are longer (0-5 and
0-14 days) ((Breitner et al., 2009, 188439). making it difficult to compare whether the associations
observed are consistent with those reported in other studies. In Helsinki, the association between
UFPs and stroke mortality was weaker than that for PM2 5, but its lag structure of association was
similar to that for PM2 5 (strongest at lag 1 day). However, Kettunen et al. (2007, 091242) only
examined lags 0-3 days. Overall, the results of these studies should be viewed with caution because
UFPs were consistently found to be correlated with gaseous pollutants derived from local
combustion sources, and one or more of the gaseous pollutants were also found to be associated with
mortality. Clearly, more research is needed to further investigate the role of UFPs on PM-mortality
associations.
6.5.2.5. Chemical Components of PM
A few recent studies have examined the association between mortality and components of
PM25. This endeavor has been undertaken by some investigators through the use of the newly
available PM2 5 chemical speciation network data. The PM2 5 chemical speciation network consists of
more than 250 monitors that have been collecting over 40 chemical species since 2000; however,
most sites started collecting data in 2001. One caveat to the new network is that because the
sampling frequencies of the monitors are either every third day or every sixth day, there have not
been, generally, a sufficient number of days to examine associations with mortality in single cities.
To circumvent this issue, some investigators (Bell et al., 2009, 191997; Dominici et al., 2007,
099135: Franklin et al., 2008, 097426: Lippmann et al., 2006, 091165) have used the PM2.5
chemical species data in a second stage regression to explain the heterogeneity in PMi0 or PM2 5
mortality risk estimates across cities. However, it should be noted these studies assume that the
relative contributions of PM2 5 have remained the same over time. There have also been some studies
that directly analyzed speciated PM25 data (e.g., Klemm et al., 2004, 056585; Ostro et al., 2007,
091354).
December 2009 6-191
-------
Explaining the Heterogeneity of PM10 Risk Estimates Using PM2.5 Chemical
Speciation Data
Lippmann et al. (2006, 091165). in addition to their primary analysis1, investigated the
consistency of the associations between specific elements and health outcomes by examining the
heterogeneity of published 1-day lagged NMMAPS PMi0 mortality risk estimates for 1987-1994
across cities as a function of the average PM2.5 chemical components across cities. They matched
PM2.5 chemical species in 60 out of 90 cities. Lippmann et al. (2006, 091165) noted that the
concentrations of the 16 chemical species examined averaged over the years 2000-2003 were highly
skewed across cities. They therefore regressed PMi0 risk estimates on each of the PM2.5 components,
raw and log-transformed, with weights based on the standard error of the PMi0 risk estimates. The
log-transformed values yielded better predictive power, and the authors subsequently presented the
results with log-transformed values. As shown in Figure 6-31, the 16 PM2.5 species showed varying
extent of predictive power in explaining the PM10 risk estimates across 60 cities, with Ni and V
being the best predictors.
EC
Cu
OC
PM
0
"
-0.5 0.0 0.5 1.0
Percent per 10-jit(/ni3 increase in PM10
Source: Lippmann et al. (2006, 091165)
Figure 6-31. Percent increase in PMio risk estimates (point estimates and 95% CIs) associated
with a 5th-95th percentile increase in PM2.s and PM2.5 chemical components. The
PM2.6 chemical components were log-transformed in the regression. The PMio
risk estimates were for 60 NMMAP cities for 1987-1994.
Dominici et al. (2007, 099135) examined the influence of Ni and V on the updated NMMAPS
PMio mortality risk estimates for 1987-2000, using 72 counties in which Ni and V data were
collected. A Bayesian hierarchical model was used to estimate the role of Ni and V on the
heterogeneity of PMio risk estimates. While they found both Ni and V to be significant predictors of
variation in PMi0 mortality risk estimates across cities, they also noted that this result was sensitive
to the inclusion of the New York City data. Lippmann et al. (2006, 091165) and Dominici et al.
(2007, 099135) both reported that the Ni levels in New York City are particularly high (-10 times the
national average). Figure 6-32 shows the result of the sensitivity analysis for Ni. Note that the Ni in
this result was not log-transformed, as clearly reflected in the change in the width of confidence
bands when the New York data were removed (i.e., a skewed distribution produces narrow bands).
1 The main focus of the study was to examine the role of PM2.5 chemical components in a mouse model of atherosclerosis (ApoE ')
exposed to concentrated fine PM (CAPs) in Tuxedo, NY.
December 2009
6-192
-------
Dominici et al. (2007, 099135) further noted that they reached "the same conclusion" when
log-transformed data were used in the analysis, but the results were not presented.
No communities removed
•a
0)
o
=
OJ
.&
o
u
I
o
New York removed
—i
-10
-20 -10 0 10 20 30
Percent increase in PM10 risk estimates per IQR Ni
Source: Reprinted with Permission of Oxford University Press from Dominici et al. (2007, 099135)
Figure 6-32. Sensitivity of the percent increase in PM™ risk estimates (point estimates and
95% CIs) associated with an interquartile increase in Ni. The Ni concentration
was not log-transformed in this regression model. The PM10 risk estimates were
for 72 NMMAP cities for 1987-2000. The top estimate is achieved by including
data for all the 69 communities. The other estimates are calculated by excluding
one of the 69 communities at a time.
Bell et al. (2009, 191997) presented a supplemental analysis similar to both Lippmann et al.
(2006, 091165) and Dominici et al. (2007, 099135) in their examination of whether the variation in
PM2 5 risks for cardiovascular and respiratory hospital admissions is due to differences in PM2.5
chemical composition. The authors used the 100 U.S. cities included in the Peng et al. (2005,
087463) analysis and PMi0 data for the years 1987-2000 along with PM25 chemical component data
for 2000-2005. Using a Bayesian hierarchical model, Bell et al. (2009, 191997) found that PMi0
relative risks for total mortality were greater in counties and during seasons with higher PM2.5 Ni
concentrations. However, in a sensitivity analysis when selectively removing cities from the overall
estimate, the significant association between the PMi0 mortality risk estimate and the PM2 5 Ni
fraction was diminished upon removing New York city from the analysis, which is consistent with
the results presented by Dominici et al. (2007, 099135).
Explaining the Heterogeneity of PM2.5 Risk Estimates Using PM2.5 Chemical
Speciation Data
The first stage of the Franklin et al. (2008, 097426) 25 cities study, described previously,
focused on a time-series regression of mortality on PM2 5 by season. In the second stage random
effects meta regression, the PM25 mortality risk estimates (25 cities><4 seasons = 100 estimates) were
regressed on the ratio of mean seasonal PM25 species to the total PM25 mass. The authors included
those species that had at least 25% of the reported concentrations above the minimum detection
limit, which resulted in 18 species being included in the analysis. Their rationale for using species
proportions as effect modifiers, according to the investigators, was that "in the first stage of the
analysis the mortality risk was estimated per unit of the total PM25 mass, which encompassed all
December 2009
6-193
-------
measured species, and therefore it would not be meaningful to use the species concentrations directly
as the effect modifier" (Franklin et al., 2008, 097426). In the second stage regression model,
Franklin et al. (2008, 097426) also included a quadratic function of seasonally averaged temperature
to capture the inverted U-shape relationship between PM2.5 penetration and temperature. They found
that the fitted relationship between PM2.5 risk estimates across cities and seasonally averaged
temperature substantiates the use of temperature as a surrogate for ventilation (Franklin et al., 2008,
097426). Table 6-17 shows the resulting effect modification by PM25 species. Al, As, Ni, Si, and
SO4 ~ were found to be significant effect modifiers of PM2.5 risk estimates, and simultaneously
including Al, Ni, and SO4 ~ together, or Al, Ni, and As together further increased explanatory power.
Of all the species examined, Al and Ni explained the most residual heterogeneity. Franklin et al.
(2008, 097426) also examined the effect of demographic variables on PM2 5 risk estimates and found
that only median household income was significantly associated with mortality.
Table 6-17. Effect modification of composition on the estimated percent increase in mortality with a
10 [jg/rn3 increase in PM2.6.
Cause
Nonaccidental
Univariate
Nonaccidental
Multivariate
(1)
Nonaccidental
Multivariate
(2)
p-value for effect % increase in nonaccidental mortality per 10 ug/m3 Heteroaeneitv
Species modification by species to increase in PIV^.s for an interquartile increase in exnlained f°/ v
PM2.s mass proportion species to PM2.s mass proportion* expiamea (/<>)
Al
As
Br
Cr
EC
Fe
K
Mn
Na+
Ni
N03
NrV
OC
Pb
Si
S042"
V
Zn
Al
Ni
S042"
Al
Ni
As
<0.001
0.02
0.11
0.12
0.79
0.43
0.10
0.42
0.22
0.01
0.07
0.84
0.59
0.31
0.03
0.01
0.28
0.19
<0.001
0.01
<0.001
<0.001
0.01
<0.001
0.58
0.55
0.38
0.33
0.06
0.12
0.41
0.14
0.20
0.37
-0.49
0.04
-0.02
0.17
0.41
0.51
0.30
0.23
0.79
0.34
0.75
0.61
0.35
0.58
45
35
5
16
0
3
28
10
14
41
28
3
4
11
25
33
3
15
100
100
*Adjusted for temperature
Includes heterogeneity explained by temperature
Source: Reprinted with Permission of Lippincott Williams & Wilkins from Franklin et al. (2008, 09742'
Although Lippmann et al. (2006, 091165) used NMMAPS PM19 risk estimates and Franklin
et al. (2008, 097426) used PM2 5 risk estimates to examine effect modification due to various PM
December 2009
6-194
-------
species, 14 out of the 18 species analyzed in these two studies overlap (Figure 6-31 and Table 6-17).
Both studies found that Ni explained the heterogeneity in PM risk estimates. Note that New York
City was not included in the 25 cities examined in Franklin et al. (2008, 097426) and, thus, could not
influence the result. Sulfate positively, but not significantly, explained the PMi0 risk estimates in the
Lippmann et al. (2006, 091165) analysis. However, SO42~ was a significant predictor of PM2.5 risk
estimates in the Franklin et al. (2008, 097426) analysis. Al and Si were negative (i.e., less than the
average PM10 risk estimates across cities), though not significant predictors in the Lippmann et al.
(2006, 091165) analysis. Unlike the Franklin et al. (2008, 097426) analysis, arsenic (As) showed no
association with mortality in the Lippmann et al. (2006, 091165) analysis. The source of these
differences may be due to the difference in geographic coverage, PM size (PM2 5 may represent more
secondary aerosols than PMi0), or the difference in the analytical methods used in each study.
Specifically, the analytical approach used by Franklin et al. (2008, 097426) does have an advantage
of delineating seasonal variations in PM components and the associated potential seasonal mortality
effects.
In light of the results presented in speciation studies it must be noted that second stage
analyses that use PM chemical species as effect modifiers have some limitations. Unlike analyses
that directly examine the associations between chemical species and mortality, if an effect
modification is observed it may be confounded if the variations of the mean levels of the chemical
species examined are correlated with other demographic factors that vary across cities. Thus, more
concrete conclusions could be formulated if direct associations are found between mortality and PM
chemical components in time-series analyses.
Association between PM2.s Chemical Components and Mortality
Ostro et al. (2007, 091354) examined associations between PM2.5 chemical components and
mortality in six California counties (Fresno, Kern, Riverside, Sacramento, San Diego, and Santa
Clara), which had at least 180 days of speciation data for the years 2000-2003. The study examined
all-cause, cardiovascular, and respiratory mortality for individual lags of 19 specific PM25 chemical
components. The second stage random-effects model combined risk estimates at each lag across
cities. The number of available days for chemical species data ranged from 243 (San Diego County)
to 395 (Sacramento County). The authors found an association between mortality, especially
cardiovascular mortality, and several chemical components. For example, cardiovascular mortality
was associated with EC, OC, nitrate, Fe, K, and Ti at various lags.
Even though this was a multicity study, the relatively small number of available days and the
every-third-day (or every-sixth-day) sampling frequency for PM2 5 chemical species made it difficult
to interpret the results of the lag structure of associations observed for the chemical species. To
evaluate the impact of non-daily sampling frequency, Ostro et al. (2007, 091354) examined both the
PM2 5 series that coincides with the speciation sampling days (for the initial six counties
[i.e., PM25c]) and PM25 data that was available on all days for an extended set of counties (the initial
six counties plus Contra Costa, Los Angeles, and Orange Counties [i.e., PM25ext]). Figure 6-33 shows
the association between all-cause mortality and selected PM2 5 chemical species as well as for PM2 5c
and PM2 5ext. Note the wide confidence bands for the risk estimates for each PM2 5 chemical species
and PM25c, apparently reflecting the low statistical power of the data. The lag structure of
associations is more clearly defined for PM2 5ext, and appears to be different from that for PM2 5.
December 2009 6-195
-------
300
se 2.00
Ł
Ł
&
ja 100
''fc*
in
OS
u
Ł 0
Q
QJ
°- -too
-2.00
-
(
1
i
4
*
| J
1
M i :
:
* '•
»
j _
«
•
•K"
1
•
If !
4* *
f !
1 !
J" (
.„, . * *
lr T*» t»
r o I i T
«| *f. ** 1 *1
-. -(; : ;- -' "T :T
* <> j i ~ j. - j.
i
0123 0123 9123 0123 0123 0123 fl 1 2 3 0123 0123
PM?Sl, EC OC
HO, SO, Cu K Zn PM;!1^,
Species and lag day
Source: Ostro et al. (2007, 091354)
Figure 6-33.
Percent excess risk (Cl) of total (nonaccidental) mortality per IQR of
concentrations. Note: PM2.6 has the same sampling days as chemical species.
PM2.6 has all available PM2.6ext data for nine counties. * p < 0.10; ** p < 0.05
Ostro et al. (2008, 097971) used the speciation data from the six counties analyzed in their
2007 analysis, described above, in an additional analysis to examine effect modification of
cardiovascular mortality effects, which showed the strongest association in the 2007 analysis,
attributed to PM2.5 and 13 chemical components by socio-economic and demographic factors. The
results of the analysis were combined using random effects meta-analysis. The investigators tested
statistical differences in risk estimates between strata using a t-test, and reported that, for many of the
PM2.5 chemical species; there were significantly higher effect estimates among those with lower
educational attainment and Hispanics. While these patterns were apparent in their results table,
interpretation of the results is not straightforward because the table only presented the most
significant (and positive) lags, and they were often different between the strata (e.g., the most
frequent significant lag for the Hispanic group was 1 day, while it was 2 or 3 days for the White
group). As the investigators pointed out, the every -third-day sampling frequency of the speciation
data also complicates the interpretation of the results for different lags.
Overall, the two studies by Ostro et al. (2007, 091354) were the first attempt to directly
analyze associations between the newly available chemical speciation data and mortality. While
suggestive associations between several chemical species and mortality were reported, a longer
length of observations is needed to more clearly determine the associations.
6.5.2.6. Source-Apportioned PM Analyses
Chemically speciated PM data allow for the source apportionment of PM. The idea of using
source-apportioned PM for health effects analyses is appealing because, if such
source-apportionment could be reliably conducted, it would allow for an evaluation of PM2.5 mass
concentrations by source types. However, the uncertainties associated with source-apportionment
methods have not been well characterized.
To address this issue, in 2003, several groups of EPA-funded researchers organized a
workshop and independently conducted source apportionment on two sets of data: Phoenix, AZ, and
Washington, DC, compared the results (Hopke et al., 2006, 088390). and then conducted time-series
December 2009
6-196
-------
mortality regression analyses using each group's source-apportioned data (Ito et al., 2006, 088391;
Mar et al., 2006, 086143; Thurston et al., 2005, 097949). The various research groups generally
identified the same major source types, each with similar elemental compositions. Inter-group
correlation analyses indicated that soil-, SO42~-, residual oil-, and salt-associated mass concentrations
were most unambiguously identified by various methods, whereas vegetative burning and traffic
were less consistent. Aggregate source-specific mortality relative risk (RR) estimate confidence
intervals overlapped each other, but the SO42~-related PM2 5 component was most consistently
significant across analyses in these cities.
The results from the source-apportionment workshop quantitatively characterized the
uncertainties associated with the factor analysis-based methods, but they also raised new issues. The
mortality analyses conducted in Phoenix, AZ, and Washington, DC, both found that different
source-types showed varying lag structure of associations with mortality. For example, Figure 6-34
shows cardiovascular mortality risk estimates for three of the PM2.5 sources from the Phoenix, AZ,
analysis (Mar et al., 2006, 086143). The strongest associations for "traffic" PM2.5 was found for lag
1-day, while for "secondary SO4 ~" PM25, it was lag 0, with a monotonic decline towards longer
lags. These results are consistent with those in the 2004 PM AQCD (U.S. EPA, 2004, 056905). in
which associations were reported with combustion-related PM2 5, but not crustal source PM2 5. It is
conceivable that PM from different source types produces different lagged effects, but it is also
likely that different PM species have varying lagged correlations with the covariates in the health
effects regression models (e.g., temperature, day-of-week) resulting in apparent differences in lagged
associations with mortality. Thus, interpretation of these source-apportioned PM health effect
estimates remains challenging.
1.4-
1.?
1.0
0.8
1.4-
1.2
1.0
0.8
Traffic
1.4-
1.2
1.0
0.8
Secondary sulfate
ABCDEFGH
ABCOtf-GH
A B C D t F G
Source: Reprinted with Permission of Nature Publishing Group from Mar et al. (2006, 0861431
Figure 6-34. Relative risk and Cl of cardiovascular mortality associated with estimated PM2.s
source contributions. Y-axis: relative risk per 5th-to-95th percentile increment of
estimated PM2.s source contribution. X-axis: the alphabet denotes investigator/
method; lagged PM2.6 source contribution for lag 0 through 5 days, left to right,
are shown for each investigator/method.
6.5.2.7. Investigation of Concentration-Response Relationship
The results from large multicity studies reviewed in the 2004 PM AQCD (U.S. EPA, 2004,
056905) suggested that strong evidence did not exist for a clear threshold for PM mortality effects.
However, as discussed in the 2004 PM AQCD (U.S. EPA, 2004, 056905). there are several
challenges in determining and interpreting the shape of PM-mortality concentration-response
functions and the presence of a threshold, including: (1) limited range of available concentration
levels (i.e., sparse data at the low and high end); (2) heterogeneity of susceptible populations; and (3)
December 2009
6-197
-------
the influence of measurement error. Regardless of these limitations, studies have continued to
investigate the PM-mortality concentration-response relationship.
Daniels et al. (2004, 087343) evaluated three concentration-response models: (1) log-linear
models (i.e., the most commonly used approach, from which the majority of risk estimates are
derived); (2) spline models that allow data to fit possibly non-linear relationship; and (3) threshold
models, using PMi0 data in 20 cities from the 1987-1994 NMMAPS data. They reported that the
spline model, combined across the cities, showed a linear relation without indicating a threshold for
the relative risks of death for all-causes and for cardiovascular-respiratory causes in relation to PMi0,
but "the other cause" deaths (i.e., all cause minus cardiovascular-respiratory) showed an apparent
threshold at around 50 ug/m3 PMi0, as shown in Figure 6-35. For all-cause and cardio-respiratory
deaths, based on the Akaike's Information Criterion (AIC), a log-linear model without threshold was
preferred to the threshold model and to the spline model.
The HEI review committee commented that interpretation of these results required caution,
because (1) the measurement error could obscure any threshold; (2) the city-specific concentration-
response curves exhibited a variety of shapes; and (3) the use of AIC to choose among the models
might not be appropriate due to the fact it was not designed to assess scientific theories of etiology.
Note, however, that there has been no etiologically credible reason suggested thus far to choose one
model over others for aggregate outcomes. Thus, at least statistically, the result of Daniels et al.
(2004, 087343) suggests that the log-linear model is appropriate in describing the relationship
between PMi0 and mortality.
Total Mortality CVDRESP Mortality Other Cause Mortality
o
Ł
"es
CC
PM (MQ/m3)
Source: Reprinted with Permission of HEI from Daniels et al. (2004, 087343)
Figure 6-35. Concentration-response curves (spline model) for all-cause, cardiovascular,
respiratory and other cause mortality from the 20 NMMAPS cities. Estimates are
posterior means under Bayesian random effects model. Solid line is mean lag,
triangles are lag 0 (current day), and squares are lag 1 (previous day).
The Schwartz (2004, 078998) analysis of PMi0 and mortality in 14 U.S. cities, described in
Section 6.5.2.1, also examined the shape of the concentration-response relationship by including
indicator variables for days when concentrations were between 15 and 25 (ig/m3, between 25 and
34 (ig/m3, between 35 and 44 (ig/m3, and 45 (ig/m3 and above. In the model, days with
concentrations below 15 (ig/m3 served as the reference level. This model was fit using the single
stage method, combining strata across all cities in the case-crossover design. Figure 6-36 shows the
resulting relationship, which does not provide sufficient evidence to suggest that a threshold exists.
The authors did not examine city -to-city variation in the concentration-response relationship in this
study.
December 2009 6-198
-------
2.5
_« 2.0
O
ID
T)
increase in
O Oi
i
u 0.5
a_
0.0
~
—
A *
A
- A | III
10 20 30 40 50 60
PM10 (ng/rn3)
Source: Reprinted with Permission of BMJ Group from Schwartz (2004, 0789981
Figure 6-36. Percent increase in the risk of death on days with PM10 concentrations in the
ranges of 15-24, 25-34, 35-44, and 45 ug/m and greater, compared to a reference
of days when concentrations were below 15 ug/m3. Risk is plotted against the
mean PM10 concentration within each category.
Samoli et al. (2005, 087436) investigated the concentration-response relationship between
PM10 and mortality in 22 European cities (and BS in 15 of the cities) participating in the APHEA
project. In nine of the 22 cities, PMi0 levels were estimated using a regression model relating co-
located PMi0 to BS or TSP They used regression spline models with two knots (30 and 50 (ig/m3)
and then combined the individual city estimates of the splines across cities. The investigators
concluded that the association between PM and mortality in these cities could be adequately
estimated using the log-linear model. However, in an ancillary analysis of the concentration-response
curves for the largest cities in each of the three distinct geographic areas (western, southern, and
eastern European cities): London, England; Athens, Greece; and Cracow, Poland, Samoli et al.
(2005, 087436) observed a difference in the shape of the concentration-response curve across cities.
Thus, while the combined curves (Figure 6-37) appear to support no-threshold relationships between
PMio and mortality, the heterogeneity of the shapes across cities makes it difficult to interpret the
biological relevance of the shape of the combined curves.
December 2009
6-199
-------
*2
s
0
07
15-
10-
,••
/
I
Total
Cardiovascular
Respiratory
. . j j j- S ffjfjjit^rfyi
o 5-
-5-
0
50
100
PM10 (^g/m
150
3)
200
Source: Samoli et al. (2005, 0874361
Figure 6-37. Combined concentration-response curves (spline model) for all-cause,
cardiovascular, and respiratory mortality from the 22 APHEA cities.
The results from the three multicity studies discussed above support no-threshold log-linear
models, but issues such as the possible influence of exposure error and heterogeneity of shapes
across cities remain to be resolved. Also, given the pattern of seasonal and regional differences in
PM risk estimates depicted in recent multicity study results (e.g., Peng et al., 2005, 087463). the
very concept of a concentration-response relationship estimated across cities and for all-year data
may not be very informative.
6.5.3. Summary and Causal Determinations
6.5.3.1. PM2.5
The 2004 PM AQCD (U.S. EPA, 2004, 056905) found that the strength of evidence from U.S.-
and Canadian-based studies (both multi- and single-city) for PM2.5 mortality associations varied
across outcomes, with relatively stronger evidence for associations with cardiovascular compared to
respiratory causes. The resulting effect estimates reported for these two endpoints ranged from 1.2 to
2.7% for cardiovascular-related mortality and 0.8 to 2.7% for respiratory-related mortality, per
10 (ig/m3 increase in PM2.5 (U.S. EPA, 2004, 056905).
In the current review, PM2.5 risk estimates were found to be consistently positive, and slightly
larger than those reported for PM10 for all-cause, and respiratory- and cardiovascular-related
mortality. The risk estimates for all-cause (nonaccidental) mortality ranged from 0.29% (Dominici
et al., 2007, 097361) to 1.21% (Franklin et al., 2007, 091257) per 10 (ig/m3 increase in PM2.5. These
associations were consistently observed at lag 1 and lag 0-1, which have been confirmed through
extensive analyses in PMi0-mortality studies. Cardiovascular-related mortality risk estimates were
found to be similar to those for all-cause mortality; whereas, the risk estimates for respiratory-related
mortality were consistently larger: 1.01% (Franklin et al., 2007, 091257) to 2.2% (Ostro et al.,
2006, 087991) using the same lag (i.e., lag 1 and lag 0-1) and averaging indices. The studies
evaluated that examined the relationship between short-term exposure to PM2 5 and cardiovascular
December 2009
6-200
-------
effects (section 6.2) provide coherence and biological plausibility for PM2.5-induced cardiovascular
mortality, which represents the largest component of total (nonaccidental) mortality (~ 35%)
(American Heart Association, 2009, 198920). However, as noted in section 6.3, there is limited
coherence between some of the respiratory morbidity findings from epidemiologic and controlled
human exposure studies for the specific health outcomes reported and the subpopulations in which
those health outcomes occur, complicating the interpretation of the PM25 respiratory mortality
effects observed.
Regional and seasonal patterns in PM2 5 risk estimates were observed with results similar to
those presented for PMio (Dominici et al.. 2007. 097361; Peng et al., 2005, 087463; Zeka etal,
2006, 088749). with the greatest effects occurring in the eastern U.S. (Franklin et al., 2007, 091257;
Franklin et al., 2008, 097426) and during the spring (Franklin et al.. 2007. 091257; Zanobetti and
Schwartz, 2009, 188462). Of the studies evaluated only Burnett et al. (2004, 086247). a Canadian
multicity study, analyzed gaseous pollutants and found mixed results, with possible confounding of
PM2.5 risk estimates by NO2. Although the recently evaluated U.S.-based multicity studies did not
analyze potential confounding of PM25 risk estimates by gaseous pollutants, evidence from single-
city studies evaluated in the 2004 PM AQCD (U.S. EPA, 2004, 056905) suggest that gaseous
copollutants do not confound the PM2 5-mortality association, which is further supported by studies
that examined the PMi0-mortality relationship. An examination of effect modifiers (e.g.,
demographic and socioeconomic factors), specifically AC use which is sometimes used as a
surrogate for decreased pollutant penetration indoors, has suggested that PM2 5 risk estimates
increase as the percent of the population with access to AC decreases (Franklin et al., 2007, 091257;
2008, 097426). Collectively, the epidemiologic evidence is sufficient to conclude that 3 C3US3I
relationship exists between short-term exposure to PM and mortality.
6.5.3.2. PM10.2.5
The 2004 PM AQCD (U.S. EPA, 2004, 056905) found a limited body of evidence that was
suggestive of associations between short-term exposure to ambient PM10_2.5 and various mortality
outcomes (e.g., 0.08 to 2.4% increase in total [nonaccidental] mortality per 10 ug/m3 increase in
PMio_2.5). As a result, the AQCD concluded that PMi0_2.5, or some constituent component(s)
(including those on the surface) of PMi0_2.5, may contribute, in certain circumstances, to increased
human health risks.
The majority of studies evaluated in this review that examined the relationship between
PMio_2.5 and mortality reported consistent positive associations in areas with mean 24-h avg
concentrations ranging from 6.1-16.4 ug/m3. However, uncertainty surrounds the PMi0_2.5
associations reported due to the different methods used to estimate PMi0_2.5 concentrations across
studies (e.g., direct measurement of PMi0_2.5 using dichotomous samplers, calculating the difference
between PMi0 and PM2 5 concentrations).
Anew study of 47 U.S. cities (Zanobetti and Schwartz, 2009, 188462). which estimated
PM10_2.5 by calculating the difference between the county-average PM10 and PM2 5, found
associations between PMi0_2.5 and mortality across the U.S., including regions where PMi0_2.5 levels
are not high. In addition, one well conducted multicity Canadian study (Burnett et al., 2004,
086247) provided evidence for an association between short-term exposure to PMi0_2.5 and mortality.
However, unlike PM25 very few of the PMi0_2.5 studies have investigated confounding by gaseous
copollutants or the influence of model specification on PMi0_2.5 risk estimates. Zanobetti and
Schwartz (2009, 188462) did provide preliminary evidence for greater effects occurring during the
warmer months (i.e., spring and summer), which is consistent with the results from PMi0-mortality
studies (Peng et al., 2005, 087463; Zeka et al., 2006, 088749). Overall, although more data is
needed to: adequately characterize the chemical and biological components that may modify the
potential toxicity of PMi0_2.5 and compare the different methods used to estimate exposure, consistent
positive associations between short-term exposure to PMi0_2.5 and mortality were observed in the
U.S. and Canadian-based multicity studies evaluated, as well as the single-city studies conducted in
these locations. Therefore, the epidemiologic evidence JS Suggestive Of 3 C3US3I relationship
between short-term exposure to PM1025 and mortality.
December 2009 6-201
-------
6.5.3.3. UFPs
Limited evidence was available during the review of the 2004 PM AQCD (U.S. EPA, 2004,
056905) regarding the potential association between UFPs and mortality. The lone study evaluated
was conducted in Germany and provided some evidence for an association, but this association was
reduced upon the inclusion of gaseous pollutants in a two-pollutant model.
Only a few new studies, all of them from Europe, were identified during this review, which
examined the association between short-term exposure to UFPs and mortality. Inconsistencies were
observed in the lag structure of association reported by each study in terms of both the lag day with
the greatest association and the number of lag days considered in the study. Overall the studies
consistently found that UFPs were correlated with gaseous pollutants derived from local combustion
sources and that one or more of the gaseous pollutants were also associated with mortality. The
limited number of studies available and the discrepancy in results between studies further confirms
the need for additional data to examine the UFP-mortality relationship. In conclusion, the
epidemioiogic evidence is inadequate to infer a causal association between short-term
exposure to UFPs and mortality
6.6. Attribution of Ambient PM Health Effects to Specific
Constituents or Sources
From a mechanistic perspective, it is highly plausible that the chemical composition of PM
would be a better predictor of health effects than particle size. The observed geographical gradients
in a number of PM2.5 constituents (e.g., EC, OC, nitrate, and SO42^ and regional heterogeneity in
PM-related health effects reported in epidemioiogic studies are consistent with this hypothesis.
Recent studies in epidemiology, controlled human exposure, and toxicology have begun using
information on ambient PM composition, and apportionment of constituents into sources, in an
attempt to identify those with links to health outcomes and endpoints.
This section focuses on short-term exposure studies that (1) assessed health effects for ambient
PM sources or components; and (2) used quantitative methods to relate those sources and
components to health effects. Epidemioiogic, controlled human exposure, and toxicological studies
that took into consideration a large set of PM constituents (typically minerals, metals, EC, OC, and
ions such as SO42~) and aimed to segregate which constituents or groups of constituents may be
responsible for the PM-related health effects observed are included. Most of these studies were
reviewed earlier in this chapter and evaluated the relationship between specific chemical constituents
derived from ambient PM and health effects. However, there were many studies presented earlier, as
well as others only included in the Annexes, which only selected one or a small number of PM
constituents a priori. Several controlled human exposure and toxicological studies likewise used a
single compound found in PM rather than ambient PM. Additionally, studies that presented ambient
PM composition and health data without systematically and explicitly investigating relationships are
not included in this section. The few epidemioiogic studies of long-term exposure that examined
potential relationships between composition and sources of PM with mortality are discussed in
Section 7.6.2.
Prior to the 2004 PM AQCD (U.S. EPA, 2004, 056905). only a handful of epidemioiogic
studies had attempted to relate specific constituents or sources of ambient PM to health outcomes
without selecting constituents a priori. In this review, approximately 40 new epidemioiogic,
controlled human exposure, and toxicological studies explore the health effects attributed to
chemical constituents and sources of ambient PM. The following summary (Section 6.6.3) provides
a synthesis of the findings, including discussions on the coherence and consistency of the results.
6.6.1. Evaluation Approach
Relating a large number of ambient PM constituents with a large number of health outcomes
presents difficulties that are related to both the nature of PM and methods of quantitative analysis.
First, the number of constituents that comprise PM is not only large, but the correlations between
them can be high. Reducing the correlation between constituents has been accomplished in most of
December 2009 6-202
-------
the recent studies through various forms of factor analysis, which limits the correlations between
constituents by grouping the most highly correlated ambient PM constituents into less correlated
groups or factors. Some studies identify the resulting groups or factors with named sources of
ambient PM, but many do not draw explicit links between factors and actual sources. The methods
used in estimating source contributions to ambient PM are reviewed in Section 3.6.1.
Most studies reviewed herein, regardless of discipline, were based on data for between 7 and
20 ambient PM constituents, with EC, OC, SO4s and NO3 most commonly measured. Most studies
first reduced the number of ambient PM constituents by grouping them with various factorization or
source apportionment techniques and the majority labeled the constituent groupings according to
their presumed source. A separate analysis was then used to examine the relationship between the
grouped PM constituents and various health effects. A few performed these two steps simultaneously
using Partial Least Squares (PLS) procedures or Structural Equation Modeling (SEM). A small
number of controlled human exposure and toxicological studies did not apply any kind of grouping
to the ambient PM speciation data.
There are some differences in the type of PM constituent data used in epidemiologic,
controlled human exposure and toxicological studies. In epidemiologic studies, ambient PM
speciation data is obtained from atmospheric monitors; for controlled human exposure and
toxicological studies, the technique used in the experimental exposure determines the type of PM
data. Thus, all epidemiologic studies relied on monitor data, while all of the controlled human
exposure and the majority of the toxicological studies used CAPs (and analyzed the concentrations
of constituents therein). The remaining toxicological studies used ambient PM samples collected on
filters at various U.S. sites. Further details on the studies included can be found in Appendix F.
Some important limitations in interpreting these studies together is that few, if any of the
results are easily comparable, due to: (1) differences in the sets of ambient PM constituents that
make up each of the factors; (2) the subjectivity involved in labeling factors as sources; (3) the
numerous potential health effects examined in these studies, including definitive outcomes (e.g.,
HAs) as well as physiological alterations (e.g., increased inflammatory response); and (4) the various
statistical methods and analytical approaches used in the studies. There are no well-established,
objective methods for conducting the various forms of factor analysis and source apportionment,
leaving much of the model operation and factor assignment open to judgment by the individual
investigator. For example, the Al/Si factor identified in one study may differ from the Al/Ca/Fe/Si
factor from another study, and the "Resuspended Soil" factor from a third study. After factorization
or apportionment of the ambient PM data, the methods used for analyzing the potential association
between ambient PM constituents or sources and health effects also varied. Except for the studies
that used PLS or SEM, controlled human exposure and toxicological studies all used univariate
mixed model regression for every identified PM factor or source. A number of toxicological studies
followed the univariate step with multivariate regression for all factors. Epidemiologic studies
generally related short-term exposure to sources with health outcomes through various forms of
Poisson regression.
6.6.2. Findings
The results that follow are organized by discipline, with epidemiologic studies followed by
controlled human exposure and toxicological studies. This section ends with a summary table, Table
6-18. Table 6-18 is broken out by PM2.5 sources, and includes those epidemiologic, controlled human
exposure, and in vivo toxicological studies that either grouped ambient PM2.5 constituents or used
tracers for each source. The table does not include all factors or sources examined in the studies
listed: those factors or sources for which no association with effects was found not included.
6.6.2.1. Epidemiologic Studies
Results from the 2004 PM AQCD
Three epidemiologic studies that examined the association between PM constituents or sources
and specific health effects were evaluated in the 2004 PM AQCD (U.S. EPA, 2004, 056905). Of
December 2009 6-203
-------
these studies, one study associated daily mortality with a mobile sources PM factor in Knoxville, TN
and St. Louis, MO and coal in Boston, MA, while the crustal factor was not found to be significant
for any of the six cities studied (Laden et al, 2000, 012102; Schwartz, 2003, 042811). Another
study demonstrated an association between a regional SO4 ~ factor and total mortality at lag 0 in
Phoenix and factors for regional SO42~, motor vehicles, and vegetative burning with cardiovascular
mortality at lags of 0, 1, and 3, respectively (Mar et al., 2000, 001760: 2003, 042841). Negative
associations were observed between total mortality and regional SO42~ at lag 3, along with local SO2
and soil factors (Mar et al., 2000, 001760; 2003, 042841). Finally, Tsai et al. (2000, 006251)
identified significant associations between PMi5-derived industrial sources and total daily deaths in
Newark and Camden, NJ; SO42~ was also linked to cardiopulmonary deaths in both locations. Total
mortality and cardiopulmonary deaths were also significantly associated with PM from oil burning in
Camden (2000, 006251).
Comparative Analyses of Source Apportionment Methods
Hopke et al. (2006, 088390) conducted a comparative analysis of source apportionment
techniques used by investigators at multiple institutions, and subsequently used in epidemiologic
analyses (Ito et al., 2006, 088391; Mar et al., 2006, 086143). An overarching conclusion of this set
of analyses, reported in Thurston et al. (2005, 097949). is that variation in the source apportionment
methods was not a major source of uncertainty in the epidemiologic effect estimates. In the primary
analyses, mortality was associated with secondary SO4 ~ in both Phoenix and Washington D.C.,
although lag times differed (0 and 3, respectively). The SO42~ effect was stronger for total mortality
in Washington D.C. and for cardiovascular mortality in Phoenix (Ito et al., 2006, 088391; Mar et
al., 2006, 086143). In addition, Ito et al. (2006, 088391) found some evidence for associations with
primary coal and traffic with total mortality in Washington D.C. (Ito et al., 2006, 088391) while
copper smelter, traffic, and sea salt were associated with cardiovascular mortality in Phoenix at
various lag times (Mar et al., 2006, 086143). In contrast to Phoenix, sea salt and traffic were not
associated with mortality in Washington D.C. (Ito et al., 2006, 088391). but in both locations no
associations were observed between biomass/wood combustion and mortality (Ito et al., 2006,
088391; Mar et al., 2006, 086143). In an additional study that compared three source apportionment
methods in Atlanta-PMF, modified CMB, and a single-species tracer approach-found that the
epidemiologic results were robust to the choice of analytic method (Sarnat et al., 2008, 097972).
There were consistent associations between ED visits for cardiovascular diseases with PM2 5 from
mobile sources (gasoline and diesel) and biomass combustion (primarily prescribed forest burning
and residential wood combustion), whereas PM2 5 from secondary SO42~ was associated with
respiratory disease ED visits (Sarnat et al., 2008, 097972). Sarnat et al. (2008, 097972) also found
that the primary power plant PM2 5 source identified by the CMB approach was negatively associated
with respiratory ED visits while no association was found for PM2 5 from soil and secondary
nitrates/ammonium nitrate. In these studies, effect estimates based on the different source
apportionment methods were generally in close agreement.
Source Apportionment Studies
A study that examined associations with mortality in Santiago, Chile, identified a motor
vehicle source of PM25 as having the greatest association with total and cardiac mortality at lag 1
(Cakmak et al., 2009, 191995). There was effect modification by age, with the total mortality
relative risks associated with PM2 5 from motor vehicles being greatest for those >85 yr. Soil and
combustion sources were also associated with cardiac mortality. Risk estimates for respiratory
mortality were the greatest for the motor vehicle source, with combustion and soil source factors also
demonstrating positive associations for lag 1 (Cakmak et al., 2009, 191995).
An epidemiologic study that evaluated respiratory ED visits was conducted in Spokane, WA
and used tracers as indicators of ambient PM25 sources (Schreuder et al., 2006, 097959). In this
study, only PM2 5 from vegetative burning (total carbon) was associated with increased respiratory
ED visits for lag 1, while PM25 indicators for motor vehicles (Zn) and soil (Si) were not associated
with cardiac hospital or respiratory ED visits. Andersen et al. (2007, 093201) conducted a source
apportionment analysis to identify the sources of ambient PMi0 associated with cardiovascular and
December 2009 6-204
-------
respiratory hospital admissions in older adults and children (ages 5-18) in Copenhagen, including
two-pollutant models with various sources of PMi0. Andersen et al. (2007, 093201) found that
secondary and crustal sources of PMi0 were associated with cardiovascular hospital admissions;
biomass sources were associated with respiratory hospital admissions; and vehicle sources were
associated with asthma hospital admissions.
Several panel epidemiologic studies have examined the association between PM sources and
physiological alterations in cardiovascular function. Lanki et al. (2006, 089788) reported positive
associations between PM2.5 from local traffic (measured as absorbance, which is correlated with EC
content) and long-range transported PM2.5 with ST-segment depression in elderly adults in a study
conducted in Helsinki, Finland. Positive associations with ST-segment depression were also reported
with PM2.5 from crustal and salt sources, but these associations were not statistically significant. In
an additional study, Yue et al. (2007, 097968) found that adult males with coronary artery disease in
Erfurt, Germany, demonstrated changes in repolarization parameters associated with traffic-related
PM2.5, with increased vWF linked to traffic and combustion-generated particles, although the source
apportionment was based solely on particle size distribution. In addition, elevated CRP levels were
associated with all sources of PM25 (soil, local traffic, secondary aerosols from local fuel
combustion, diesel, and secondary aerosols from multiple sources) (Yue et al., 2007, 097968).
Reidiker et al. (2004, 056992). in a study of young male highway patrol officers, found that the most
significant effects (HRV, supraventricular ectopic beats, hematological markers, vWF) were
associated with a speed-change factor for PM2 5 (2004, 056992). In addition, the authors observed an
association between crustal factor and cardiovascular effects, but no health-related associations with
steel wear or gasoline PM2 5 source factors.
Two recent studies have examined the associations between ambient PM2 5 sources and
respiratory symptoms and lung function. Positive associations with PM2 5 motor vehicle and road
dust sources were reported for respiratory symptoms and inhaler use in asthmatic children in New
Haven, CT, and negative associations with wheeze or inhaler use for biomass burning at lag 0-2
(Gent et al., 2009, 180399). These positive effects for motor vehicle and road dust sources were
robust to the inclusion of a gaseous copollutant (NO2, CO, SO2, or O3) in the regression model.
Penttinen et al. (2006, 087988) in a study consisting of asthmatic adults living in Helsinki, Finland,
found that decrements in PEF were associated with ambient PM2 5 soil, long-range transport, and
local combustion sources at lags from 0-5 days. In addition, negative associations with asthma
symptoms and medication use were reported for PM2 5 from sea salt and long-range transport sources
(Penttinen et al., 2006, 087988).
PM Constituent Studies
Some studies considered large sets of ambient PM constituents and attempted to identify
which were associated with various health effects, but without grouping them into factors, or
identifying sources. The majority of these studies focused on health effects associated with short-
term exposure to PM2 5. Peng et al. (2009, 191998) examined the association between PM2 5
constituents (i.e., EC, OC, SO42", NO3", Si, Na, NH4+) and cardiovascular and respiratory hospital
admissions in 119 U.S. cities. When including each constituent in a multipollutant model, they found
that EC and OC were robust to the inclusion of the other constituents at lag 0 for cardiovascular and
respiratory hospital admissions, respectively. Although this study did not include analyses to identify
sources of the constituents examined, EC and OC are often attributed to motor vehicle emissions,
particularly diesel engines, and wood burning (Peng et al., 2009, 191998). Ostro et al. (2007,
091354; 2008, 097971) conducted two studies in six California counties to examine the association
between ambient PM constituents and mortality. In the 2007 analysis, Ostro et al. (2006, 087991)
found associations between Cu and all-cause mortality; EC, K, and Zn and CVD mortality; and Cu
and Ti and respiratory mortality at lags ranging from 0 to 3 days. Associations during the summer
were only observed between K for both CVD and respiratory mortality; and Al, Cl, Cu, Pb, Ti, and
Zn and respiratory mortality. Overall, the most consistent associations were observed during the cool
season. In a subsequent analysis, Ostro et al. (2008, 097971) examined the association between
ambient PM constituents and cardiovascular mortality in potentially susceptible subpopulations. The
authors found positive associations between EC, OC, NO3", SO42", K, Cu, Fe, and Zn and
cardiovascular mortality. These associations were higher in individuals with lower educational
December 2009 6-205
-------
attainment and of Hispanic ethnicity. In addition, similar to the 2007 analysis, associations were
observed at lags ranging from 0 to 3 days.
Evaluation of Effect Modification by PM Constituents
Several studies have conducted secondary analyses to examine whether the variation in
associations between PM2.5 and morbidity and mortality or PMi0 and mortality reflects differences in
PM2.5 constituents. An assumption in these types of analyses, especially when examining the effects
on PMio mortality risk estimates, is that the relative contributions of PM2.5 have remained the same
over time; these studies used PMi0 data for years prior to 2000, while PM2.5 speciation data has only
been routinely collected since about 2000. Bell et al. (2009, 191997) found statistically significant
associations between the county average concentrations of V, Ni, and EC (106 counties) and effect
estimates for both cardiovascular and respiratory hospital admissions with short-term exposure to
PM25. In this analysis the ambient PM25 constituents that comprised the majority of PM25 total mass
in the study locations were NH4+, EC, OC, NO3", and SO42". Bell et al. (2009, 191997) also
conducted a similar analysis for PMio-mortality risk estimates and found that only Ni increased the
risk estimate. However, in a sensitivity analysis, when selectively dropping out the communities
examined one at a time, removing New York City diminished the Ni association. Both Lippmann
et al. (2006, 091165) and Dominici et al. (2007, 099135) conducted similar analyses, albeit using a
smaller subset of cities and/or different years of PMi0 data. In both studies, Ni and V were found to
modify the PMi0-mortality risk estimates. Similar to Bell et al. (2009, 191997). Dominici et al.
(2007, 099135) also found that excluding New York City as part of a sensitivity analysis resulted in a
diminished association with Ni and V. In an additional study, Franklin et al. (2008, 097426)
examined the potential modification of the PM2 5-mortality relationship by PM constituents in 25
U.S. cities. In a second-stage analysis using the species-to-PM25 mass proportion of multiple
constituents, the authors found that Al, As, Ni, Si, and SO42" significantly modified the association
between PM2 5 and nonaccidental mortality.
6.6.2.2. Controlled Human Exposure Studies
A few controlled human exposure studies employed PCA, although not all linked groupings of
PM constituents to the measured physiological parameters. Huang et al. (2003, 087377)
demonstrated associations between increased fibrinogen and Cu/Zn/V and increased BALF
neutrophils and Fe/Se/SO4 in young, healthy adults exposed to RTP, NC CAPs; however, only water-
soluble constituents were analyzed. In the other study that examined physiological cardiovascular
effects, Fe and EC were associated with changes in ST-segment, while SO42~ was associated with
decreased SBP in asthmatic and healthy human volunteers exposed to Los Angeles CAPs (2003,
087377). In Gong et al. (2003, 087365) the majority of the PM was in the thoracic coarse fraction. In
the other study that used Los Angeles CAPs, the only observed association was between SO42~
content and decreased lung function (FEVi and FVC) in elderly volunteers with and without COPD
(Gong et al., 2005, 087921). Two additional controlled human exposure studies that did not perform
grouping and employed Toronto CAPs plus O3 demonstrated increased DBP and increased brachial
artery vasoconstriction associated with carbon content (Urch et al., 2004, 055629; 2005, 081080).
6.6.2.3. lexicological Studies
The only toxicological in vivo study that characterized PM sources corresponding to identified
sources was conducted in Tuxedo, NY, over a 5-mo period. This study reported that all sources
(regional SO42~, resuspended soil, residual oil, traffic and other unknown sources) were linked to HR
or HRV changes in mice at one time or another during or after daily exposure (Lippmann et al.,
2005, 087453). In a simultaneous in vitro study using CAPs from the same location, NF-KB in
BEAS-2B cells were correlated with the oil combustion factor (r = 0.289 and 0.302 for V and Ni,
respectively) (Maciejczyk and Chen, 2005, 087456). The other in vitro toxicological study (Duvall
et al., 2008, 097969) that named sources employed samples from 5 U.S. cities and found a good fit
for the regression model with increased IL-8 release in primary human airway epithelium cells and
coal combustion (R2= 0.79), secondary nitrate (R2= 0.63), and mobile sources (R2= 0.39). In
addition, soil (R2= 0.48), residual oil combustion (R2= 0.38), and wood combustion (R2= 0.33) were
December 2009 6-206
-------
associated with COX-2 effects; whereas, secondary SO42 (R2= 0.51) was correlated with HO-1.
Wood combustion and soil were negatively associated with HO-1.
Several toxicological studies employed Boston CAPs and identified at least four groupings of
ambient PM2 5 constituents (V/Ni, S, Al/Si, and Br/Pb), but they named sources only partially and
tentatively (Batalha et al., 2002, 088109: Clarke et al, 2000, 011806: Godleski et al., 2002,
156478: Nikolov et al., 2008, 156808: Saldiva et al., 2002, 025988: Wellenius et al., 2003,
055691). When examining cardiovascular effects these studies reported that Si was associated with
changes in the ST-segment of dogs (Wellenius et al., 2003, 055691) and decreased L/W ratio in rat
pulmonary arteries (Batalha et al., 2002, 088109) in multivariate analyses. In addition, blood
hematological results were associated with V/Ni, Al/Si, Na/Cl, and S in dogs (Clarke et al., 2000,
011806). An examination of respiratory effects in the latter study found that V/Ni and Br/Pb were
associated with increased inflammation in BALF for only the third day of exposure (Clarke et al.,
2000, 011806). Decreased respiratory rate and increased airway irritation (Penh) in dogs were
associated with road dust (Al) and motor vehicles (OC), respectively (Nikolov et al., 2008, 156808).
Individual PM2 5 constituents associated with elevated neutrophils in BALF were Br, EC, OC, Pb,
and SO42~ (Godleski et al., 2002, 156478). which is consistent with the findings (Br, EC, OC, Pb, V,
and Cl) of Saldiva et al. (2002, 0259881
The two toxicological studies that used PLS methodologies identified PM25 constituents
linked to respiratory parameters. Seagrave et al. (2006, 091291) demonstrated associations between
cytotoxic responses and a gasoline plus nitrates source factor (OC, Pb, hopanes/steranes, nitrate, and
As) along with inflammatory responses and a gasoline plus diesel source factor (including major
metal oxides) in rats exposed via IT instillation. In the other study, Veranth et al. (2006, 087479)
collected loose surface soil from 28 sites in the Western U.S. and exposed BEAS-2B cells to PM2.5.
OCi, OC3, OC2, EC2, Br, ECi, and Ni correlated with IL-8 release, decreased IL-6 release, and
decreased viability at low and high doses (10 and 80 ug/cm2, respectively).
Table 6-18. Study-specific PM2.6 factor/source categories associated with health effects.
Source Category Location
Health Effects
Time study*1 sPecies Reference
CRUSTAUSOIUROAD DUST
Al, Si, Fe
Not provided
Al, Ca, Fe, Si
Al, Si, Ca, K, Fe
Al, Si, Ca, K, Fe
Al, Si
Al, Si, Ca
Al, Si, Ti, Fe
Al, Si, Ca, Fe
Al, Si
Phoenix, AZ
Washington,
D.C.
Santiago, Chile
Helsinki, Finland
Los Angeles, CA
Boston, MA
Boston, MA
VMe County,
NC
Tuxedo, NY
Boston, MA
nega^ssociationwith
tCV mortality
tCV mortality
t respiratory mortality
ST-segment depression
I ST-segment voltage
ST-segment change
I lumen/wall ratio
t uric acid
t mean cycle length
|HR
tHR
tSDNN, tRMSSD
t blood PMN %
1 blood lymphocytes %
tWBC
Lag 2 E
Lag 4 E
Lag1 E
Lag3 E
2 days post-exposure H
Following exposure T
24 h post-exposure T
Lag 15 h E
During exposure
Afternoon post-exposur
e. Night post-exposure
Following exposure T
Human
Human
Human
Human
Human
Dog
Rat
Human
Mouse
Dog
Mar etal. (2000, 0017601
ltoetal.(2006, 0883911
Cakmak etal. (2009, 1919951
Lanki et al. (2006, 089788)
Gong et al. (2003, 0421061
Wellenius etal. (2003, 0556911
Batalha et al. (2002, 088109)
Riedikeretal. (2004, 0569921
Lippmann et al. (2005, 0874531
Clarke etal. (2000, 0118061
Si, Fe, Al, Ca, Ba, Ti New Haven, CT
t respiratory symptoms and
inhaler use
Lag 0-2
Human Gent et al. (2009,1803991
December 2009
6-207
-------
Source Category
Si, Al, Ca, Fe, Mn
Al
Location
Helsinki, Finland
Boston, MA
Health Effects
I mean PEF
I airway irritation (penh)
Time
Lag3
During exposure
Type of
Study1
E
T
Species
Human
Dog
Reference
Penttinen et al. (2006, 087988)
Nikolov et al. (2008, 156808)
SALT
Not provided
Na, Cl
Na, Cl
Na, Cl
Na, Cl
SECONDARY S042'
S
Not provided
Not provided
S, K, Zn, Pb
S042"
S, Si, OC
S
S042", NH4*, OC
S, K, Zn, PM mass
S042" (+N02)
Phoenix, AZ
Helsinki, Finland
Boston, MA
Helsinki, Finland
Boston, MA
tCV mortality
ftotal mortality
negative association with
total mortality
ST-segment depression
t blood lymphocyte %
Negatively associated with
bronchodilator use and
corticosteroid use
t lung PMN density
Lag5
LagO
Lag3
Following exposure
Lag 0-5 avg
24 h post-exposure
E
E
T
E
T
Human
Human
Dog
Human
Rat
Mar etal. (2006,086143)
Lanki et al. (2006, 089788)
Clarke etal.(2000, 011806)
Penttinen et al. (2006, 0879881
Saldiva et al. (2002, 025988)
/LONG-RANGE TRANSPORT
Phoenix, AZ
Washington,
D.C.
Phoenix, AZ
Helsinki, Finland
Los Angeles, CA
Tuxedo, NY
Boston, MA
Atlanta, GA
Helsinki, Finland
Los Angeles, CA
t total mortality
negative association with
total mortality
t total mortality
tCV mortality
ST-segment depression
|SBP
|HR
|SDNN, IRMSSD
|RBC
t hemoglobin
t respiratory ED visits
1 mean PEF. Negative
association with asthma
symptom prevalence
|FEV,
|FVC
LagO
Lag5
LagS
LagO
Lag 2
4 h post-exposure
Afternoon post-
exposure
Night post-exposure
Following exposure
LagO
Lag1
LagS
Following exposure
E
E
E
E
H
T
T
E
E
H
Human
Human
Human
Human
Human
Mouse
Dog
Human
Human
Human
Mar etal. (2000, 0017601
Ito et al. (2006, 0883911
Mar etal. (2006, 0861431
Lanki et al. (2006, 089788)
Gong et al. (2003, 042106)
Lippmann et al. (2005, 0874531
Clarke et al. (2000, 0118061
Sarnat et al. (2008, 0979721
Penttinen et al. (2006, 0879881
Gong et al. (2005, 0879211
TRAFFIC
Pb, Br, Cu
Not provided
Mn, Fe, Zn, Pb, OC,
EC, CO, N02
CO, N02, EC, OC
Gasoline (OC, N03",
NH/)
Diesel (EC, OC, N03")
NOX, EC, ultrafine
count
Harvard Six
Cities
Phoenix, AZ
Phoenix, AZ
Santiago, Chile
Atlanta, GA
Atlanta, GA
Helsinki, Finland
t total mortality
tCV mortality
t CV mortality
tCV mortality
t respiratory mortality
tCVD ED visits
tCVD ED visits
ST-segment depression
Lag 0-1
Lag1
Lag1
Lag1
LagO
LagO
Lag 2
E
E
E
E
E
E
E
Human
Human
Human
Human
Human
Human
Human
Laden et al. (2000, 0121021
Mar etal. (2006, 086143)
Mar etal. (2000, 0017601
Cakmak etal. (2009, 1919951
Sarnat et al. (2008, 0979721
Sarnat et al. (2008, 0979721
Lanki et al. (2006, 089788)
December 2009
6-208
-------
Source Category
Speed-change factor
(Cu, S, aldehydes)
Motor vehicle/other
(Br, Pb, Se, Zn, N0r)
EC, Zn, Pb, Cu, Se
Local combustion
(NOX, ultrafine PM, Cu,
Zn, Mn, Fe)
Gasoline+ secondary
nitrate*
Gasoline+diesel*
Location
Vteke County,
NC
Tuxedo, NY
New Haven, CT
Helsinki, Finland
Birmingham, AL;
Atlanta, GA;
Pensacola, FL;
Centreville, AL
Birmingham, AL;
Atlanta, GA;
Pensacola, FL;
Centreville, AL
Health Effects
t blood urea nitrogen
t mean red cell volume
t blood PMN %
1 blood lymphocytes %
t von Willebrand factor
(vWF)
I protein C
t mean cycle length
tSDNN
t PNN50
t supraventricular ectopic
beats
| RMSSD
t respiratory symptoms
I mean PEF
cytotoxic responses
(potency)
inflammatory responses
(potency)
Time
Lag 15 h
Afternoon post-
exposure
Lag 0-2
Lag 0-5 avg
24 h post-exposure
24 h post-exposure
Study*1 sPecies Reference
E Human Riediker et al. (2004, 056992)
T Mouse Lippmann et al. (2005, 087453)
E Human Gent et al. (2009, 1803991
E Human Penttinen et al. (2006, 0879881
T Rat Seagrave et al. (2006, 0912911
T Rat Seagrave et al. (2006, 0912911
OIL COMBUSTION
V, Ni
V, Ni, Se
Ni
V, Ni
Boston, MA
Tuxedo, NY
Boston, MA
Boston, MA
t blood PMN %
1 blood lymphocytes %
tBALFAM%
|SDNN
| RMSSD
I respiratory rate
t lung PMN density
Following exposure
Following exposure
24 h post-exposure
Afternoon post-
exposure
During exposure
24 h post-exposure
T Dog Clarke et al. (2000, 0118061
T Mouse Lippmann et al. (2005, 0874531
T Dog Nikolov et al. (2008, 1568081
T Rat Saldiva et al. (2002, 0259881
COAL COMBUSTION
Se, S042"
Not provided
Harvard Six
Cities
Washington,
D.C.
t total mortality
t total mortality
Lag 0-1
Lag3
E Human Laden et al. (2000, 0121021
E Human Itoetal. (2006, 0883911
OTHER METALS
Cu smelter (not
provided)
Incinerator
Metal processing
(S042", Fe, NH4 , EC,
OC)
Phoenix, AZ
Washington,
D.C.
Atlanta, GA
tCV mortality
ftotal mortality
Negative association with
total and CV mortality
tCVD ED visits
LagO
LagO
LagO
E Human Mar etal. (2006, 0861431
E Human Ito et al. (2006, 0883911
E Human Sarnat et al. (2008, 0979721
December 2009
6-209
-------
Source Category Location
Health Effects Time study1' sPecies
Reference
r'CU' Santiago, Chile
t respiratory mortality
Lag1
Human Cakmak et al. (2009, 1919951
WOODSMOKE/ VEGETATIVE BURNING
OC, K
OC, EC, K, NH4*
Total C
Phoenix, AZ
Atlanta, GA
Spokane, WA
t CV mortality
tCVD ED visits
t respiratory ED visits
Lag3
LagO
Lag1
E
E
E
Human
Human
Human
Mar etal. (2000, 0017601
Sarnat et al. (2008, 097972)
Schreuderetal. (2006, 097959)
UNNAMED FACTORS
Zn-Cu-V
Fe-Se-S042~
Br, Cl, Pb
Br, Pb
Br, Pb
Chapel Hill, NC
Chapel Hill, NC
Santiago, Chile
Boston, MA
Boston, MA
t blood fibrinogen
t BALF PMN
tCV mortality
t respiratory mortality
t BALF PMN %
t lung PMN density
18 h post-exposure
18 h post-exposure
Lag1
24 h post-exposure
24 h post-exposure
H
H
E
T
T
Human
Human
Human
Dog
Rat
Huang et al. (2003, 087377)
Huang et al. (2003, 087377)
Cakmak etal. (2009, 1919951
Clarke etal. (2000, 0118061
Saldiva et al. (2002, 025988)
Constituents not provided.
1 E = Epidemiologic study; H = Controlled human exposure study; T = Toxicological study
An in vitro toxicological study that employed Chapel Hill PM10 used PCA but did not name
specific PM sources (Becker et al., 2005, 088590). In this study, the release of IL-6 from human
alveolar macrophages and IL-8 from normal human bronchial epithelial cells was associated with a
PMio factor comprised of Cr, Al, Si, Ti, Fe, and Cu. No statistically significant effects were observed
for a second PMi0 factor (Zn, As, V, Ni, Pb, and Se).
Those toxicological studies that did not apply groupings to the ambient PM2.5 speciation data
demonstrated a variety of results. Two Boston CAPs studies demonstrated lung oxidative stress
correlated with a number of individual PM2 5 constituents including, Mn, Zn, Fe, Cu, and Ca
(Gurgueira et al., 2002, 036535) and Al, Si, Fe, K, Pb, and Cu (Rhoden et al., 2004, 087969) in rats
using univariate regression.
The remaining toxicological study that did not use ambient PM constituent groupings reported
a correlation between Zn and plasma fibrinogen in SH rats when constituents were normalized per
unit mass of CAPs (Kodavanti et al., 2002, 035344).
6.6.3. Summary by Health Effects
Recent epidemiologic, toxicological, and controlled human exposure studies have evaluated
the health effects associated with ambient PM constituents and sources, using a variety of
quantitative methods applied to a broad set of PM constituents, rather than selecting a few
constituents a priori. As shown in Table 6-18, numerous ambient PM2.5 source categories have been
associated with health effects, including factors for PM from crustal and soil, traffic, secondary
SO42~, power plants, and oil combustion sources. There is some evidence for trends and patterns that
link particular ambient PM constituents or sources with specific health outcomes, but there is
insufficient evidence to determine whether these patterns are consistent or robust.
For cardiovascular effects, multiple outcomes have been linked to a PM crustal/soil/road dust
source, including cardiovascular mortality in Washington D.C. (Ito et al., 2006, 088391)and
Santiago, Chile, (Cakmak et al., 2009, 191995) and ST-segment changes in Helsinki (Lanki et al.,
2006, 089788). Los Angeles (Gong et al., 2003, 042106). and Boston (Wellenius et al., 2003,
055691). Interestingly, the ST-segment changes have been observed in an epidemiologic panel study,
a controlled human exposure study, and a toxicological study, although the majority of the CAPs in
the controlled human exposure study was PMi0_2.5. Further support for a crustal/soil/road dust source
associated with cardiovascular health effects comes from a PM10 source apportionment study in
Copenhagen that reported increased cardiovascular hospital admissions (Andersen et al., 2007,
093201).
December 2009
6-210
-------
PM2.5 traffic and wood smoke/vegetative burning sources have also been linked to
cardiovascular effects. Cardiovascular mortality in Phoenix (Mar et al., 2000, 001760; 2006,
086143) and Santiago, Chile, (Cakmak et al., 2009, 191995) was associated with traffic at lag 1.
Gasoline and diesel sources were associated with ED visits in Atlanta for cardiovascular disease at
lag 0 (Sarnat et al., 2008, 097972). Cardiovascular mortality in Phoenix (Mar et al., 2000, 001760)
and ED visits in Atlanta (Sarnat et al., 2008, 097972) were associated with wood smoke/vegetative
burning.
Studies that only examined the effects of individual PM2.5 constituents linked EC to
cardiovascular hospital admissions in a multicity analysis (Peng et al., 2009, 191998) and
cardiovascular mortality in California (Ostro et al., 2007, 091354; 2008, 097971).
These studies suggest that cardiovascular effects may be associated with PM2.5 from motor
vehicle emissions, wood or biomass burning, and PM (both PM2.5 and PMi0_2.5) from crustal or road
dust sources. In addition there are many studies that observed associations between other sources
(i.e., salt, secondary SO4 71ong-range transport, other metals) and cardiovascular effects, but at this
time, there does not appear to be a consistent trend or pattern of effects for those factors.
There is less consistency in observed associations between PM sources and respiratory health
effects, which may be partially due to the fact that fewer studies have been conducted that evaluated
respiratory-related outcomes and measures. However, there is some evidence for associations with
secondary SO42~ PM2 5. Sarnat et al. (2008, 097972) found an increase in respiratory ED visits in
Atlanta that was associated with a PM2 5 secondary SO42~ factor. Decrements in lung function in
Helsinki (Lanki et al., 2006, 089788) and Los Angeles (Gong et al., 2005, 087921)in asthmatic and
healthy adults, respectively, were also linked to this factor. Health effects relating to the
crustal/soil/road dust and traffic sources of PM included increased respiratory symptoms in
asthmatic children (Gent et al., 2009, 180399) and decreased PEF in asthmatic adults (Penttinen et
al., 2006, 087988). Inconsistent results were also observed in those PM25 studies that use individual
constituents to examine associations with respiratory morbidity and mortality, although Cu, Pb, OC,
and Zn were related to respiratory health effects in two or more studies.
A few studies have identified PM2 5 sources associated with total mortality. These studies
found an association between mortality and a PM2 5 coal combustion factor (Laden et al., 2000,
012102). while others linked mortality to a secondary SO42~/long-range transport PM2 5 source (Ito
et al., 2006, 088391: Mar et al., 2006, 086143).
Recent studies have evaluated whether the variation in associations between PM2 5 and
morbidity and mortality or PM10 and mortality reflects differences in PM25 constituents (Bell et al.,
2009, 191997; Dominici et al., 2007, 099135: Lippmann et al., 2006, 091165). In three studies (Bell
et al., 2009, 191997: Dominici et al., 2007, 099135: Lippmann et al., 2006, 091165) PMi0-mortality
effect estimates were greater in areas with a higher proportion of Ni in PM2 5, but the overall PMiq-
mortality association was diminished when New York City was excluded in a sensitivity analysis in
two of the studies. V was also found to modify PMio-mortality effect estimates as well as those for
PM2 5 with respiratory and cardiovascular hospital admissions (Bell et al., 2009, 191997). When
examining the effect of species-to-PM2 5 mass proportion on PM2 5-mortality effect estimates Ni was
found to modify the association along with Al, As, Si, and SO42", but not V (Franklin et al., 2008,
097426).
6.6.4. Conclusion
Recent studies show that source apportionment methods have the potential to add useful
insights into which sources and/or PM constituents may contribute to different health effects. Of
particular interest are several epidemiologic studies that compared source apportionment methods
and the associated results. One set of studies compared epidemiologic associations with PM25 source
factors using several methods - PCA, PMF, and UNMIX - independently analyzed by separate
research groups (Hopke et al., 2006, 088390: Ito et al., 2006, 088391: Mar et al., 2006, 086143:
Thurston et al., 2005, 097949). Schreuder et al. (2006, 097959) compared UPM and two versions of
UNMIX to derive tracers and Sarnat et al. (2008, 097972) compared PMF, modified CMB, and a
single-species tracer approach. In all analyses, epidemiologic results based on the different methods
were generally in close agreement. The variation in risk estimates for daily mortality between source
categories was significantly larger than the variation between research groups (Ito et al., 2006,
088391: Mar et al., 2006, 086143: Thurston et al., 2005, 097949). Additionally, the variation in risk
estimates based on the source apportionment model used had a much smaller effect than the
December 2009 6-211
-------
variation caused by the different source constituents. Further, the most strongly associated source
types were consistent across all groups. This supports the general validity of such approaches,
though greater integration of results would be possible if the methods employed for grouping PM
constituents were more consistent across studies and disciplines. Further research would aid
understanding of the contribution of different factors, sources, or source tracers of PM to health
effects by increasing the number of locations where similar health endpoints or outcomes are
examined.
Overall, the results displayed in Table 6-18 indicate that many constituents of PM can be
linked with differing health effects and the evidence is not yet sufficient to allow differentiation of
those constituents or sources that are more closely related to specific health outcomes. These
findings are consistent with the conclusions of the 2004 PM AQCD (U.S. EPA, 2004, 056905). that a
number of source types, including motor vehicle emissions, coal combustion, oil burning, and
vegetative burning, are associated with health effects. Although the crustal factor of fine particles
was not associated with mortality in the 2004 PM AQCD (U.S. EPA, 2004, 056905). recent studies
have suggested that PM (both PM2.s and PMi0_2.5) from crustal, soil or road dust sources or PM
tracers linked to these sources are associated with cardiovascular effects. In addition, secondary
SO42~ PM2.5 has been associated with both cardiovascular and respiratory effects.
December 2009 6-212
-------
Chapter 6 References
Adamkiewicz G; Ebelt S; Syring M; Slater J; Speizer FE; Schwartz J; Suh H; Gold DR. (2004). Association between air
pollution exposure and exhaled nitric oxide in an elderly population. Thorax, 59: 204-209. 087925
Adar SD; Adamkiewicz G; Gold DR; Schwartz J; Coull BA; Suh H. (2007). Ambient and microenvironmental particles and
exhaled nitric oxide before and after a group bus trip. Environ Health Perspect, 115: 507-512. 098635
Adar SD; Gold DR; Coull BA; Schwartz J; Stone PH; Suh H. (2007). Focused exposures to airborne traffic particles and
heart rate variability in the elderly. Epidemiology, 18: 95-103. 001458
Aekplakorn W; Loomis D; Vichit-Vadakan N; Shy C; Plungchuchon S. (2003). Acute effects of SO2 and particles from a
power plant on respiratory symptoms of children, Thailand. Southeast Asian J Trap Med Public Health, 34: 906-
914. 089908
Alarie Y. (1973). Sensory irritation by airborne chemicals. Grit Rev Toxicol, 2: 299-363. 070967
Albert CM; Rosenthal L; Calkins H; Steinberg JS; Ruskin JN; Wang P; Muller JE; Mittleman MA. (2007). Driving and
Implantable Cardioverter-Defibrillator Shocks for Ventricular Arrhythmias: Results From the TOVA Study. J Am
Coll Cardiol, 50: 2233-2240. 156201
Alexis NE; Lay JC; Zeman K; Bennett WE; Peden DB; Soukup JM; Devlin RB; Becker S. (2006). Biological material on
inhaled coarse fraction particulate matter activates airway phagocytes in vivo in healthy volunteers. J Allergy Clin
Immunol, 117: 1396-1403. 154323
Allen RW; Mar T; Koenig J; Liu LJ; Gould T; Simpson C; Larson T. (2008). Changes in lung function and airway
inflammation among asthmatic children residing in a woodsmoke-impacted urban area. Inhal Toxicol, 20: 423-433.
156208
American Heart Association. (2009). Cardiovascular disease statistics. Retrieved 17-NOV-09, from
http://www.am ericanheart.org/presenter.jhtml?identifier=4478. 198920
Analitis A; Katsouyanni K; Dimakopoulou K; Samoli E; Nikoloulopoulos AK; Petasakis Y; Touloumi G; Schwartz J;
Anderson HR; Cambra K; Forastiere F; Zmirou D; Vonk JM; Clancy L; Kriz B; Bobvos J; Pekkanen J. (2006).
Short-term effects of ambient particles on cardiovascular and respiratory mortality. Epidemiology, 17: 230-233.
088177
Andersen ZJ; Loft S; Ketzel M; Stage M; Scheike T; Mette MN; Bisgaard H. (2008). Ambient Air Pollution Triggers
Wheezing Symptoms in Infants. Thorax, 63: 710-716. 096150
Andersen ZJ; Wahlin P; Raaschou-Nielsen O; Ketzel M; Scheike T; Loft S. (2008). Size distribution and total number
concentration of ultrafine and accumulation mode particles and hospital admissions in children and the elderly in
Copenhagen, Denmark. Occup Environ Med, 65: 458-66. 189651
Andersen ZJ; Wahlin P; Raaschou-Nielsen O; Scheike T; Loft S. (2007). Ambient particle source apportionment and daily
hospital admissions among children and elderly in Copenhagen. J Expo Sci Environ Epidemiol, 17: 625-636.
093201
Anderson HR; Bremner SA; Atkinson RW; Harrison RM; Walters S. (2001). Particulate matter and daily mortality and
hospital admissions in the west midlands conurbation of the United Kingdom: associations with fine and coarse
particles, black smoke and sulphate. Occup Environ Med, 58: 504-510. 017033
Anselme F; Loriot S; Henry J-P; Dionnet F; Napoleoni J-G; Thnillez C; Morin J-P (2007). Inhalation of diluted diesel
engine emission impacts heart rate variability and arrhythmia occurrence in a rat model of chronic ischemic heart
failure. Arch Toxicol, 81: 299-307. 097084
Antonini JM; Roberts JR; Jernigan MR; Yang H-M; Ma JYC; Clarke RW. (2002). Residual oil fly ash increases the
susceptibility to infection and severely damages the lungs after pulmonary challenge with a bacterial pathogen.
Toxicol Sci, 70: 110-119. 035342
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009 6-213
-------
Antonini JM; Taylor MD; Leonard SS; Lawryk NJ; Shi X; Clarke RW; Roberts JR. (2004). Metal composition and
solubility determine lung toxicity induced by residual oil fly ash collected from different sites within a power plant.
Mol Cell Biochem, 255: 257-265. 097199
Arantes-Costa F; Lopes F; Toledo A; Magliarelli-Filho P; Moriya H; Carvalho-Oliveira R; Mauad T; Saldiva P; Martins M.
(2008). Effects of residual oil fly ash (ROFA) in mice with chronic allergic pulmonary inflammation. Toxicol
Pathol, 36: 680-686. 187137
Arbex MA; Martins LC; De Oliveira RC; Pereira LA; Arbex FF; Cancado JE; Saldiva PH; Braga AL. (2007). Air pollution
from biomass burning and asthma hospital admissions in a sugar cane plantation area in Brazil. J Epidemiol
Community Health, 61: 395-400. 091637
Arena VC; Mazumdar S; Zborowski JV; Talbott EO; He S; Chuang YH; Schwerha JJ. (2006). A retrospective investigation
of PM10 in ambient air and cardiopulmonary hospital admissions in Allegheny County, Pennsylvania: 1995-2000. J
Occup Environ Med, 48: 38-47. 088631
Atiga WL; Calkins H; Lawrence JH; Tomaselli GF; Smith JM; Berger RD. (1998). Beat-to-beat repolarization lability
identifies patients at risk for sudden cardiac death. J Cardiovasc Electrophysiol, 9: 899-908. 156231
Atkinson RW; Anderson HR; Sunyer J; Ayres J; Baccini M; Vonk JM; Boumghar A; Forastiere F; Forsberg B; Touloumi G;
Schwartz J; Katsouyanni K. (2001). Acute effects of particulate air pollution on respiratory admissions: results from
APHEA 2 project. Am JRespir Grit Care Med, 164: 1860-1866.021959
Atkinson RW; Anderson HR; Sunyer J; Ayres J; Baccini M; Vonk JM; Boumghar A; Forastiere F; Forsberg B; Touloumi G;
Schwartz J; Katsouyanni K. (2003). Acute effects of particulate air pollution on respiratory admissions. Health
Effects Institute. Cambridge, MA. 042797
Atkinson RW; Bremner SA; Anderson HR; Strachan DP; Bland JM; Ponce de Leon A. (1999). Short-term associations
between emergency hospital admissions for respiratory and cardiovascular disease and outdoor air pollution in
London. Arch Environ Occup Health, 54: 398-411. 007882
ATSDR. (2006). A study of ambient air contaminants and asthma in New York City: Part A and B. Agency for Toxic
Substances and Disease Registry; Public Health Service; U.S. Department of Health and Human Services. Atlanta,
GA.http://permanent.access.gpo.gov/lps88357/ASTHMA_BRONX_FINAL_REPORT.pdf. 090L32
Baccarelli A; Cassano P; Litonjua A; Park S; Suh H; Sparrow D; Vokonas P; Schwartz J. (2008). Cardiac autonomic
dysfunction: effects from particulate air pollution and protection by dietary methyl nutrients and metabolic
polymorphisms. Circulation, 117: 1802. 191959
Baccarelli A; Zanobetti A; Martinelli I; Grillo P; Hou L; Giacomini S; Bonzini M; Lanzani G; Mannucci PM; Bertazzi PA;
Schwartz J. (2007). Effects of exposure to air pollution on blood coagulation. J Thromb Haemost, 5: 252-260.
090733
Baccarelli A; Zanobetti A; Martinelli I; Grillo P; Hou L; Lanzani G; Mannucci PM; Bertazzi PA; Schwartz J. (2007). Air
pollution, smoking, and plasma homocysteine. Environ Health Perspect, 115: 176-181. 091310
Bagate K; Meiring JJ; Gerlofs-Nijland ME; Vincent R; Cassee FR; Borm PJA. (2004). Vascular effects of ambient
particulate matter instillation in spontaneous hypertensive rats. Toxicol Appl Pharmacol, 197: 29-39. 087945
Ballester F; Rodriguez P; Iniguez C; Saez M; Daponte A; Galan I; Taracido M; Arribas F; Bellido J; Cirarda FB; Canada A;
Guillen JJ; Guillen-Grima F; Lopez E; Perez-Hoyos S; Lertxundi A; Toro S. (2006). Air pollution and
cardiovascular admisisons association in Spain: results within the EMECAS project. J Epidemiol Community
Health, 60: 328-336. 088746
Barclay JL; Miller BG; Dick S; Dennekamp M; Ford I; Hillis GS; Ayres JG; Seaton A. (2009). A panel study of air
pollution in subjects with heart failure: negative results in treated patients. Occup Environ Med, 66: 325-334.
179935
Barnett AG; Williams GM; Schwartz J; Best TL; Neller AH; Petroeschevsky AL; Simpson RW. (2006). The effects of air
pollution on hospitalizations for cardiovascular disease in elderly people in Australian and New Zealand cities.
Environ Health Perspect, 114: 1018-1023. 089770
Barnett AG; Williams GM; Schwartz J; Neller AH; Best TL; Petroeschevsky AL; Simpson RW. (2005). Air pollution and
child respiratory health: a case-crossover study in Australia and New Zealand. Am J Respir Crit Care Med, 171:
1272-1278. 087394
December 2009 6-214
-------
Barraza-Villarreal A; Sunyer J; Hernandez-Cadena L; Escamilla-Nunez MC; Sienra-Monge JJ; Ramirez-Aguilar M;
Cortez-Lugo M; Holguin F; Diaz-Sanchez D; Olin AC; Romieu I. (2008). Air pollution, airway inflammation, and
lung function in a cohort study of Mexico City schoolchildren. Environ Health Perspect, 116: 832-838. 156254
BarregardL; SallstenG; AnderssonL; Almstrand AC; GustafsonP; AnderssonM; OlinAC. (2008). Experimental exposure
to wood smoke: effects on airway inflammation and oxidative stress. Occup Environ Med, 65: 319-324. 155675
Barregard L; Sallsten G; Gustafson P; Andersson L; Johansson L; Basu S; Stigendal L. (2006). Experimental exposure to
wood-smoke particles in health humans: effects on markers of inflammation, coagulation, and lipid peroxidation.
Inhal Toxicol, 18: 845-853. 091381
Barrett EG; Henson RD; Seilkop SK; McDonald JD; Reed MD. (2006). Effects of hardwood smoke exposure on allergic
airway inflammation in mice. Inhal Toxicol, 18: 33-43. 155677
Bartoli CR; Wellenius GA; Coull BA; Akiyama I; Diaz EA; Lawrence J; Okabe K; Verrier RL; Godleski JJ. (2009).
Concentrated ambient particles alter myocardial blood flow during acute ischemia in conscious canines. Environ
Health Perspect, 117: 333-337. 179904
Bartoli CR; Wellenius GA; Diaz EA; Lawrence J; Coull BA; Akiyama I; Lee LM; Okabe K; Verrier RL; Godleski JJ.
(2009). Mechanisms of Inhaled Fine Particulate Air Pollution-Induced Arterial Blood Pressure Changes. Environ
Health Perspect, 117: 361-366. 156256
Bartzokas A; Kassomenos P; Petrakis M; Celessides C. (2004). The effect of meteorological and pollution parameters on
the frequency of hospital admissions for cardiovascular and respiratory problems in Athens. Indoor Built Environ,
13:271-275.093252
BastainTM; Gilliland FD; Li Y-F; Saxon A; Diaz-Sanchez D. (2003). Intraindividual reproducibility of nasal allergic
responses to diesel exhaust particles indicates a susceptible phenotype. Clinlmmunol, 109: 130-136. 098690
Batalha JR; Saldiva P H; Clarke RW; Coull BA; Stearns RC; Lawrence J; Murthy GG; Koutrakis P; Godleski JJ. (2002).
Concentrated ambient air particles induce vasoconstriction of small pulmonary arteries in rats. Environ Health
Perspect, 110: 1191-1197. 088109
Becker S; Mundandhara S; Devlin RB; Madden M. (2005). Regulation of cytokine production in human alveolar
macrophages and airway epithelial cells in response to ambient air pollution particles: further mechanistic studies.
Toxicol Appl Pharmacol, 207: 269-275. 088590
Beckett WS; Chalupa DF; Pauly-Brown A; Speers DM; Stewart JC; Frampton MW; Utell MJ; Huang LS; Cox C; Zareba
W; Oberdorster G. (2005). Comparing inhaled ultrafine versus fine zinc oxide particles in healthy adults - A human
inhalation study. Am J Respir Grit Care Med, 171: 1129-1135. 156261
Bedeschi E; Campari C; Candela S; Collini G; Caranci N; Frasca G; Galassi C; Francesca G; Vigotti MA. (2007). Urban air
pollution and respiratory emergency visits at pediatric unit, Reggio Emilia, Italy. J Toxicol Environ Health A, 70:
261-265. 090712
Behndig AF; Mudway IS; Brown JL; Stenfors N Helleday R Duggan ST; Wilson SJ; Boman C Cassee FR; Frew AJ; Kelly
FJ; Sandstrom T Blomberg A. (2006). Airway antioxidant and inflammatory responses to diesel exhaust exposure in
healthy humans. Eur Respir J, 27: 359-365. 088286
Bell M; Ebisu K; Peng R; Samet J; Dominici F. (2009). Hospital Admissions and Chemical Composition of Fine Particle
Air Pollution. Am J Respir Grit Care Med, 179: 1115-1120. 191997
Bell ML; Dominici F; Ebisu K; Zeger SL; Samet JM. (2007). Spatial and Temporal Variation in PM2. 5 Chemical
Composition in the United States for Health Effects Studies. Environ Health Perspect, 115: 989-995. 155683
Bell ML; Ebisu K; Peng RD; Dominici F. (2009). Adverse health effects of particulate air pollution: modification by air
conditioning. Epidemiology, 20: 682-686. 191007
Bell ML; Ebisu K; Peng RD; Walker J; Samet JM; Zeger SL; Dominic F. (2008). Seasonal and regional short-term effects
of fine particles on hospital admissions in 202 U.S. counties, 1999-2005. Am JEpidemiol, 168: 1301-1310. 156266
Bell ML; Levy JK; Lin Z. (2008). The effect of sandstorms and air pollution on cause-specific hospital admissions in
Taipei, Taiwan. Occup Environ Med, 65: 104-111. 091268
Bennett CM; McKendry IG; Kelly S; Denike K; Koch T. (2006). Impact of the 1998 Gobi dust event on hospital
admissions in the Lower Fraser Valley, British Columbia. Sci Total Environ, 366: 918-925. 088061
December 2009 6-215
-------
Berger A; Zareba W; Schneider A; Ruckerl R; Ibald-Mulli A; Cyrys J; WichmannHE; Peters A. (2006). Runs of ventricular
and supraventricular tachycardia triggered by air pollution in patients with coronary heart disease. J Occup Environ
Med, 48: 1149-58.098702
Berger RD; Kasper EK; Baughman KL; Marban E; Calkins H; Tomaselli GF. (1997). Beat-to-beat QT interval variability:
novel evidence for repolarization lability in ischemic and nonischemic dilated cardiomyopathy. Circulation, 96:
1557-1565. 155688
Bernardi L; Passino C; Wilmerding V; Dallam GM; Parker DL; Robergs RA; Appenzeller O. (2001). Breathing patterns
and cardiovascular autonomic modulation during hypoxia induced by simulated altitude. J Hypertens, 19: 947-958.
019040
Blomberg A; Sainsbury C; Rudell B; FrewAJ; Holgate ST; Sandstrom T; Kelly FJ. (1998). Nasal cavity lining fluid
ascorbic acid concentration increases in healthy human volunteers following short term exposure to diesel exhaust.
Free Radic Res, 28: 59-67. 051246
Blomberg A; Tornqvist H; Desmyter L; Deneys V; Hermans C. (2005). Exposure to diesel exhaust nanoparticles does not
induce blood hypercoagulability in an at-risk population. J Thromb Haemost, 3: 2103-2105. 191991
Boezen HM; Vonk JM; Van Der Zee SC; Gerritsen J; Hoek G; Brunekreef B; Schouten JP; Postma DS. (2005).
Susceptibility to air pollution in elderly males and females. Eur Respir J, 25: 1018-1024. 087396
Bosson J; Barath S; Pourazar J; Behndig AF; Sandstrom T; Blomberg A; Adelroth E. (2008). Diesel exhaust exposure
enhances the ozone-induced airway inflammation in healthy humans. Eur Respir J, 31: 1234-40. 196659
Bosson J; Pourazar J; Forsberg B; Adelroth E; Sandstrom T; Blomberg A. (2007). Ozone enhances the airway inflammation
initiated by diesel exhaust. Respir Med, 101: 1140-1146. 156286
Bourotte C; Curi-Amarante A-P; Forti M-C; APereira LA; Braga AL; Lotufo PA. (2007). Association between ionic
composition of fine and coarse aerosol soluble fraction and peak expiratory flow of asthmatic patients in Sao Paulo
city (Brazil). Atmos Environ, 41: 2036-2048. 150040
Bouthillier L; Vincent R; Goegan P; Adamson IYR; Bjarnason S; Stewart M; Guenette J; Potvin M; Kumarathasan P.
(1998). Acute effects of inhaled urban particles and ozone: lung morphology, macrophage activity, and plasma
endothelin-1. Am JPathol, 153: 1873-1884. 087110
Brauner EV; Forchhammer L; Moller P; Simonsen J; Glasius M; Wahlin P; Raaschou-Nielsen O; Loft S. (2007). Exposure
to ultrafine particles from ambient air and oxidative stress-induced DNA damage. Environ Health Perspect, 115:
1177-82.091152
Brauner EV; Mortensen J; Moller P; Bernard A; Vinzents P; Wahlin P; Glasius M; Loft S. (2009). Effects of ambient air
particulate exposure on blood-gas barrier permeability and lung function. Inhal Toxicol, 21: 38-47. 190244
Brauner EV; M011er P; Barregard L; Dragsted LO; Glasius M; Wahlin P; Vinzents P; Raaschou-Nielsen O; Loft S. (2008).
Exposure to ambient concentrations of particulate air pollution does not influence vascular function or
inflammatory pathways in young healthy individuals. Part Fibre Toxicol, 5:13. 191966
Breitner S; Stolzel M; Cyrys J; Pitz M; Wolke G; Kreyling W; Kuchenhoff H; Heinrich J; Wichmann H; Peters A. (2009).
Short-Term Mortality Rates during a Decade of Improved Air Quality in Erfurt, Germany. Environ Health Perspect,
117:448-454. 188439
Briet M; Collin C; Laurent S; Tan A; Azizi M; Agharazii M; Jeunemaitre X; Alhenc-Gelas F; Boutouyrie P. (2007).
Endothelial function and chronic exposure to air pollution in normal male subjects. Hypertension, 50: 970-976.
093049
Brook RD; Brook JR; Urch B; Vincent R; Rajagopalan S; Silverman F. (2002). Inhalation of fine particulate air pollution
and ozone causes acute arterial vasoconstriction in healthy adults. Circulation, 105: 1534-1536. 024987
Brown TP; Rushton L; Mugglestone MA; Meechan DF. (2003). Health effects of a sulphur dioxide air pollution episode. J
Publ Health Med, 25: 369-371. 089909
Burnett RT; Brook J; Dann T; Delocla C; Philips O; Cakmak S; Vincent R; Goldberg MS; Krewski D. (2000). Association
between particulate- and gas-phase components of urban air pollution and daily mortality in eight Canadian cities.
Inhal Toxicol, 12: 15-39. 010273
Burnett RT; Cakmak S; Brook JR. (1998). The effect of the urban ambient air pollution mix on daily mortality rates in 11
Canadian cities. Can J Public Health, 89: 152-156. 029505
December 2009 6-216
-------
Burnett RT; Cakmak S; Brook JR; Krewski D. (1997). The role of participate size and chemistry in the association between
summertime ambient air pollution and hospitalization for cardiorespiratory diseases. Environ Health Perspect, 105:
614-620. 084194
Burnett RT; Dales R; Krewski D; Vincent R; Dann T; Brook JR. (1995). Associations between ambient particulate sulfate
and admissions to Ontario hospitals for cardiac and respiratory diseases. Am J Epidemiol, 142: 15-22. 077226
Burnett RT; Goldberg MS. (2003). Size-fractionated particulate mass and daily mortality in eight Canadian cities. Health
Effects Institute. Boston, MA. 042798
Burnett RT; Smith-Doiron M; Stieb D; Cakmak S; Brook JR. (1999). Effects of particulate and gaseous air pollution on
cardiorespiratory hospitalizations. Arch Environ Occup Health, 54: 130-139. 017269
Burnett RT; Stieb D; Brook JR; Cakmak S; Dales R; Raizenne M; Vincent R; Dann T. (2004). Associations between short-
term changes in nitrogen dioxide and mortality in Canadian cities. Arch Environ Occup Health, 59: 228-236.
086247
Burtscher H. (2005). Physical characterization of particulate emissions from diesel engines: a review. J Aerosol Sci, 36:
896-932. 155710
Cakmak S; Dales RE; Vida CB. (2009). Components of particulate air pollution and mortality in Chile. Int J Occup Environ
Health, 15: 152-158. 191995
Calderon-Garciduenas L; Maronpot RR; Torres-Jardon R; Henriquez-Roldan C; Schoonhoven R; Acuna-Ayala H;
Villarreal-Calderon A; Nakamura J; Fernando R; Reed W; Azzarelli B; Swenberg JA. (2003). DNA damage in nasal
and brain tissues of canines exposed to air pollutants is associated with evidence of chronic brain inflammation and
neurodegeneration. Toxicol Pathol, 31: 524-538. 156316
Calderon-Garciduenas L; Mora-Tiscareno A; Ontiveros E; Gomez-Garza G; Barragan-Mejia G; Broadway J; Chapman S;
Valencia-Salazar G; Jewells V; Maronpot RR; Henriquez-Roldan C; Perez-Guille B; Torres-Jardon R; Herrit L;
Brooks D; Osnaya-Brizuela N; Monroy M. (2008). Air pollution, cognitive deficits and brain abnormalities: A pilot
study with children and dogs. Brain Cognit, 68: 117-127. 156317
Caligiuri G; Levy B; Pernow J; Thoren P; Hansson GK. (1999). Myocardial infarction mediated by endothelin receptor
signaling in hypercholesterolemic mice. PNAS, 96: 6920-6924. 156318
Campbell A; Oldham M; Becaria A; Bondy SC; Meacher D; Sioutas C; Misra C; Mendez LB; Kleinman M. (2005).
Particulate matter in polluted air may increase biomarkers of inflammation in mouse brain. Neurotoxicology, 26:
133-140.087217
Campen MJ; Babu NS; Helms GA; Pett S; Wernly J; Mehran R; McDonald JD. (2005). Nonparticulate components of
diesel exhaust promote constriction in coronary arteries from ApoE -/- mice. Toxicol Sci, 88: 95-102. 083977
Campen MJ; McDonald JD; Reed MD; Seagrave J. (2006). Fresh gasoline emissions, not paved road dust, alter cardiac
repolarization in ApoE-/- mice. Cardiovasc Toxicol, 6: 199-210. 096879
Cardenas M; Vallejo M; Romano-Riquer P; Ruiz-Velasco S; Ferreira-Vidal AD; Hermosillo AG (2008). Personal exposure
to PM2.5 air pollution and heart rate variability in subjects with positive or negative head-up tilt test. Environ Res,
108: 1-6. 191900
Carlsten C; Kaufman JD; Trenga CA; Allen J; Peretz A; Sullivan JH. (2008). Thrombotic markers in metabolic syndrome
subjects exposed to diesel exhaust. Inhal Toxicol, 20: 917-921. 156323
Carlsten C; Kaufman Joel D; Peretz A; Trenga Carol A; Sheppard L; Sullivan Jeffrey H. (2007). Coagulation markers in
healthy human subjects exposed to diesel exhaust. Thromb Res Suppl, 120: 849-855. 155714
Cassee FR; Boere AJF; Fokkens PHB; Leseman DLAC; Sioutas C; Kooter IM; Dormans JAMA. (2005). Inhalation of
concentrated particulate matter produces pulmonary inflammation and systemic biological effects in compromised
rats. J Toxicol Environ Health A, 68: 773-796. 087962
Chahine T; Baccarelli A; Litonjua A; Wright RO; Suh H; Gold DR; Sparrow D; Vokonas P; Schwartz J. (2007). Particulate
air pollution, oxidative stress genes, and heart rate variability in an elderly cohort. Environ Health Perspect, 115:
1617-1622. 156327
Chan C-C; Chuang K-J; Chien L-C; Chen W-J; Chang W-T. (2006). Urban air pollution and emergency admissions for
cerebrovascular diseases in Taipei, Taiwan. Eur Heart J, 27: 1238-1244. 090193
December 2009 6-217
-------
Chan CC; Chuang KJ; Chen WJ; Chang WT; Lee CT; Peng CM. (2008). Increasing cardiopulmonary emergency visits by
long-range transported Asian dust storms in Taiwan. Environ Res, 106: 393-400. 093297
Chan CC; Chuang KJ; Shiao GM; Lin LY. (2004). Personal exposure to submicrometer particles and heart rate variability
in human subjects. Environ Health Perspect, 112: 1063-1067. 087398
Chang C-C; Hwang J-S; Chan C-C; Wang P-Y; Cheng T-J. (2007). Effects of concentrated ambient particles on heart rate,
blood pressure, and cardiac contractility in spontaneously hypertensive rats during a dust storm event. Inhal
Toxicol, 19: 973-978. 155719
Chang C-C; Hwang J-S; Chan C-C; Wang P-Y; Hu T-H; Cheng T-J. (2004). Effects of concentrated ambient particles on
heart rate, blood pressure, and cardiac contractility in spontaneously hypertensive rats. Inhal Toxicol, 16: 421-429.
055637
Chang C-C; Tsai S-S; Ho S-C; Yang C-Y (2005). Air pollution and hospital admissions for cardiovascular disease in
Taipei, Taiwan. EnvironRes, 98: 114-119. 080086
Chang CC; Hwang JS; Chan CC; Wang PY; Hu TH; Cheng TJ. (2005). Effects of concentrated ambient particles on heart
rate variability in spontaneously hypertensive rats. J Occup Health, 47: 471-480. 088662
Chardon B; Lefranc A; Granados D; Gremy I. (2007). Air pollution and doctors' house calls for respiratory diseases in the
greater Paris area (2000-3). Occup Environ Med, 64: 320-324. 091308
Che W; Zhang Z; Zhang H; Wu M; Liang Y; Liu F; Shu Y; Li N. (2007). Compositions and oxidative damage of
condensate, particulate and semivolatile organic compounds from gasoline exhausts. Environ Toxicol Pharmacol,
24: 11-18.096460
Checkoway H; Levy D; Sheppard L; Kaufman J; Koenig J; Siscovick D. (2000). A case-crossover analysis of fine
particulate matter air pollution and out-of-hospital sudden cardiac arrest. Health Effects Institute. Cambridge, MA.
015527
Chen C-H; Xirasagar S; Lin H-C. (2006). Seasonality in adult asthma admissions, air pollutant levels, and climate: a
population-based study. J Asthma, 43: 287-292. 087947
Chen J-C; Schwartz J. (2009). Neurobehavioral effects of ambient air pollution on cognitive performance in US adults.
Neurotoxicology, 30: 231-239. 179945
Chen LC; Hwang JS. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice IV
Characterization of acute and chronic effects of ambient air fine particulate matter exposures on heart-rate
variability. Inhal Toxicol, 17: 209-216. 087218
Chen PS; Tan AY. (2007). Autonomic nerve activity and atrial fibrillation. Heart Rhythm, 4: S61-S64. 197461
Chen Y; Yang Q; Krewski D; Burnett RT; Shi Y; McGrail KM. (2005). The effect of coarse ambient particulate matter on
first, second, and overall hospital admissions for respiratory disease among the elderly. Inhal Toxicol, 17: 649-655.
087555
Chen Y; Yang Q; Krewski D; Shi Y; Burnett RT; McGrail K. (2004). Influence of relatively low level of particulate air
pollution on hospitalization for COPD in elderly people. Inhal Toxicol, 16: 21-25. 087262
Cheng M-F; Tsai S-S; Wu T-N; Chen P-S; Yang C-Y. (2007). Air pollution and hospital admissions for pneumonia in a
tropical city: Kaohsiung, Taiwan. J Toxicol Environ Health A, 70: 2021 -2026. 093034
Chevalier P; Burri H; Adeleine P; Kirkorian G; Lopez M; Leizorovicz A; Andre-Fouet X; Chapon P; Rubel P; Touboul P.
(2003). QT dynamicity and sudden death after myocardial infarction: results of a long-term follow-up study. J
Cardiovasc Electrophysiol, 14: 227-233. 156338
Chimonas MA; Gessner BD. (2007). Airborne particulate matter from primarily geologic, non-industrial sources at levels
below National Ambient Air Quality Standards is associated with outpatient visits for asthma and quick-relief
medication prescriptions among children less than 20 years old enrolled in Medicaid in Anchorage, Alaska. Environ
Res, 103:397-404. 093261
Chiu H; Tiao M; Ho S; Kuo H; Wu T; Yang C. (2008). Effects of Asian Dust Storm events on hospital admissions for
chronic obstructive pulmonary disease in Taipei, Taiwan. Inhal Toxicol, 20: 777-781. 191989
Chock DP; Winkler SL; Chen C. (2000). A study of the association between daily mortality and ambient air pollutant
concentrations in Pittsburgh, Pennsylvania. J Air Waste Manag Assoc, 50: 1481-1500. 010407
December 2009 6-218
-------
Choi JH; Xu QS; Park SY; Kim JH; Hwang SS; Lee KH; Lee HJ; Hong YC. (2007). Seasonal variation of effect of air
pollution on blood pressure. J Epidemiol Community Health, 61: 314-318. 093196
Chuang K-J; Chan C-C; Chen N-T; Su T-C; Lin L-Y (2005). Effects of particle size fractions on reducing heart rate
variability in cardiac and hypertensive patients. Environ Health Perspect, 113: 1693-1697. 087989
Chuang K-J; Chan C-C; Su T-C; Lee C-T; Tang C-S. (2007). The effect of urban air pollution on inflammation, oxidative
stress, coagulation, and autonomic dysfunction in young adults. Am J Respir Grit Care Med, 176: 370-376. 091063
Chuang KJ; Chan CC; Shiao GM; Su TC. (2005). Associations between submicrometer particles exposures and blood
pressure and heart rate in patients with lung function impairments. J Occup Environ Med, 47: 1093-1098. 156356
Chuang KJ; Coull BA; Zanobetti A; Suh H; Schwartz J; Stone PH; Litonjua A; Speizer FE; Gold DR. (2008). Particulate
Air Pollution as a Risk Factor for ST-Segment Depression in Patients With Coronary Artery Disease. Circulation,
118: 1314-1320. 155731
Ciencewicki J; Gowdy K; Krantz QT; Linak WP; Brighton L; Gilmour MI; Jaspers I. (2007). Diesel exhaust enhanced
susceptibility to influenza infection is associated with decreased surfactant protein expression. Inhal Toxicol, 19:
1121-1133.096557
Clancy L; Goodman P; Sinclair H; Dockery DW. (2002). Effect of air pollution control on death rates in Dublin, Ireland: an
intervention study. Lancet, 360: 1210-1214. 035270
Clarke RW; Catalano P; Coull B; Koutrakis P; Krishna Murthy GG; Rice T; Godleski JJ. (2000). Age-related responses in
rats to concentrated urban air particles (CAPs). Presented at In: Phalen, R. F, ed. Inhalation toxicology:
proceedings of the third colloquium on particulate air pollution and human health (first special issue); June, 1999;
Durham, NC. Inhalation Toxicol. 12(suppl. 1): 73-84. 011806
Coleridge HM; Coleridge JCG. (1994). Pulmonary Reflexes: Neural Mechanisms of Pulmonary Defense. AnnuRev
Physiol, 56: 69-91. 156362
Courtois A; Andujar P; Ladeiro Y; Baudrimont I; Delannoy E; Leblais V; Begueret H; Galland MAB; Brochard P; Marano
F. (2008). Impairment of NO-Dependent Relaxation in Intralobar Pulmonary Arteries: Comparison of Urban
Particulate Matter and Manufactured Nanoparticles. Environ Health Perspect, 116: 1294-1299. 156369
Cozzi E; Hazarika S; Stallings HW; Cascio WE; Devlin RB; Lust RM; Wingard CJ; Van Scott MR. (2006). Ultrafine
particulate matter exposure augments ischemia-reperfusion injury in mice. Am J Physiol, 291: H894-H903. 091380
Cruts B; van Etten L; Tornqvist H; Blomberg A; Sandstrom T; Mills NL; Borm PJ. (2008). Exposure to diesel exhaust
induces changes in EEG in human volunteers. Part Fibre Toxicol, 5: 4. 156374
DTppoliti D; Forastiere F; Ancona C; Agabiti N; Fusco D; Michelozzi P; Perucci CA. (2003). Air pollution and myocardial
infarction in Rome: a case-crossover analysis. Epidemiology, 14: 528-535. 074311
Dales R. (2004). Ambient carbon monoxide may influence heart rate variability in subjects with coronary artery disease. J
Occup Environ Med, 46: 1217-1221.099036
Dales R; Liu L; Szyszkowicz M; Dalipaj M; Willey J; Kulka R; Ruddy TD. (2007). Particulate air pollution and vascular
reactivity: the bus stop study. Int Arch Occup Environ Health, 81: 159-164. 155743
Daniels MJ; Dominici F; Zeger SL; Samet JM. (2004). The national morbidity, mortality, and air pollution study Part III:
PM10 concentration-response curves and thresholds for the 20 largest US cities. Health Effects Institute.
Cambridge, MA. 087343
Danielsen PH; Brauner EV; Barregard L; Sallsten G; Wallin M; Olinski R; Rozalski R; Moller P; Loft S. (2008).
Oxidatively damaged DNA and its repair after experimental exposure to wood smoke in healthy humans. Mutat Res
Fund Mol Mech Mutagen, 642: 37-42. 156382
Day KC; Reed MD; McDonald JD; Keilkop SK; Barrett EG. (2008). Effects of gasoline engine emissions on preexisting
allergic airway responses in mice. Inhal Toxicol, 20: 1145-1155. 190204
De Bruin ML; van Hemel NM; Leufkens HG; Hoes AW. (2005). Hospital discharge diagnoses of ventricular arrhythmias
and cardiac arrest were useful for epidemiologic research. J Clin Epidemiol, 58: 1325-1329. 155746
de Haar C; Hassing I; Bol M; Bleumink R; Pieters R. (2005). Ultrafine carbon black particles cause early airway
inflammation and have adjuvant activity in a mouse allergic airway disease model. Toxicol Sci, 87: 409-418.
097872
December 2009 6-219
-------
De Hartog JJ; Hoek G; Peters A; Timonen KL; Ibald-Mulli A; Brunekreef B; Heinrich J; Tiittanen P; Van Wijnen JH;
Kreyling W; Kulmala M; Pekkanen J. (2003). Effects of fine and ultrafine particles on cardiorespiratory symptoms
in elderly subjects with coronary heart disease: the ULTRA study. Am J Epidemiol, 157: 613-623. 001061
Delfino R; Brummel S; Wu J; Stern H; Ostro B; Lipsett M; Winer A; Street D; Zhang L; Tjoa T. (2009). The relationship of
respiratory and cardiovascular hospital admissions to the southern California wildfires of 2003. Occup Environ
Med, 66: 189. 191994
Delfino RJ; Chang J; Wu J; Ren C; Tjoa T; Nickerson B; Cooper D; Gillen DL. (2009). Repeated hospital encounters for
asthma in children and exposure to traffic-related air pollution near the home. Ann Allergy Asthma Immunol, 102:
138-44. 190254
Delfino RJ; Coate BD; Zeiger RS; Seltzer JM; Street DH; Koutrakis P. (1996). Daily asthma severity in relation to personal
ozone exposure and outdoor fungal spores. Am J Respir Grit Care Med, 154: 633-641. 080788
Delfino RJ; Gone H; Linn WS; Pellizzari ED; Hu Y. (2003). Asthma symptoms in Hispanic children and daily ambient
exposures to toxic and criteria air pollutants. Environ Health Perspect, 111: 647-656. 050460
Delfino RJ; Gong H; Linn WS; Hu Y; Pellizzari ED. (2003). Respiratory symptoms and peak expiratory flow in children
with asthma in relation to volatile organic compounds in exhaled breath and ambient air. J Expo Sci Environ
Epidemiol, 13: 348-363. 090941
Delfino RJ; Murphy-Moulton AM; Becklake MR. (1998). Emergency room visits for respiratory illnesses among the
elderly in Montreal: association with low level ozone exposure. Environ Res, 76: 67-77. 093624
Delfino RJ; Quintana PJE; Floro J; Gastanaga VM; Samimi BS; Kleinman MT; Liu L-JS; Bufalino C; Wu C-F; McLaren
CE. (2004). Association of FEV1 in asthmatic children with personal and microenvironmental exposure to airborne
particulate matter. Environ Health Perspect, 112: 932-941. 056897
Delfino RJ; Staimer N; Gillen D; Tjoa T; Sioutas C; Fung K; George SC; Kleinman MT. (2006). Personal and ambient air
pollution is associated with increased exhaled nitric oxide in children with asthma. Environ Health Perspect, 114:
1736-1743. 090745
Delfino RJ; Staimer N; Tjoa T; Polidori A; Arhami M; Gillen DL; Kleinman MT; Vaziri ND; Longhurst J; Zaldivar F;
Sioutas C. (2008). Circulating biomarkers of inflammation, antioxidant activity, and platelet activation are
associated with primary combustion aerosols in subjects with coronary artery disease. Environ Health Perspect,
116:898-906. 156390
Delfino RJ; Zeiger RS; Seltzer JM; Street DH. (1998). Symptoms in pediatric asthmatics and air pollution: differences in
effects by symptom severity, anti-inflammatory medication use and particulate averaging time. Environ Health
Perspect, 106: 751-761.051406
Delfino RJ; Zeiger RS; Seltzer JM; Street DH; McLaren CE. (2002). Association of asthma symptoms with peak
particulate air pollution and effect modification by anti-inflammatory medication use. Environ Health Perspect,
110: A607-A617. 093740
DeMeo DL; Zanobetti A; Litonjua AA; Coull BA; Schwartz J; Gold DR. (2004). Ambient air pollution and oxygen
saturation. Am J Respir Grit Care Med, 170: 383-387. 087346
Desqueyroux H; Pujet J-C; Prosper M; Squinazi F; Momas I. (2002). Short-term effects of low-level air pollution on
respiratory health of adults suffering from moderate to severe asthma. Environ Res, 89: 29-37. 026052
Deurloo DT; van Esch BC; Hofstra CL; Nijkamp FP; van Oosterhout AJ. (2001). CTLA4-IgG reverses asthma
manifestations in a mild but not in a more "severe" ongoing murine model. Am J Respir Cell Mol Biol, 25: 751-
760. 156396
Devlin RB; Ghio AJ; Kehrl H; Sanders G; Cascio W. (2003). Elderly humans exposed to concentrated air pollution
particles have decreased heart rate variability. Eur Respir J, 40: 76S-80S. 087348
de Haar C; Hassing I; Bol M; Bleumink R; Pieters R. (2006). Ultrafine but not fine particulate matter causes airway
inflammation and allergic airway sensitization to co-administered antigen in mice. Clin Exp Allergy, 36: 1469-
1479. 144746
December 2009 6-220
-------
de Hartog JJ; Lanki T; Timonen KL; Hoek G; Janssen NA; Ibald-Mulli A; Peters A; Heinrich J; Tarkiainen TH; van
Grieken R; van Wijnen JH; Brunekreef B; Pekkanen J. (2009). Associations between PM2.5 and heart rate
variability are modified by particle composition and beta-blocker use in patients with coronary heart disease.
Environ Health Perspect, 117:105-111. 191904
Diaz-Sanchez D; Garcia MP; Wang M; Jyrala M; Saxon A. (1999). Nasal challenge with diesel exhaust particles can induce
sensitization to a neoallergen in the human mucosa. J Allergy Clin Immunol, 104: 1183-1188. 011346
Diaz-Sanchez D; Tsien A; Fleming J; Saxon A. (1997). Combined diesel exhaust particulate and ragweed allergen
challenge markedly enhances human in vivo nasal ragweed-specific IgE and skews cytokine production to a T
helper cell 2-type pattern. J Immunol, 158: 2406-2413. 051247
Dick CA; Singh P; Daniels M; Evansky P; Becker S; Gilmour MI. (2003). Murine pulmonary inflammatory responses
following instillation of size-fractionated ambient particulate matter. J Toxicol Environ Health A, 66: 2193-2207.
088776
Diez Roux AV; Auchincloss AH; Astor B; Barr RG; Cushman M; Dvonch T; Jacobs DR Jr; Kaufman J; Lin X; Samson Px.
(2006). Recent exposure to particulate matter and C-reactive protein concentration in the multi-ethnic study of
atherosclerosis x. Am J Epidemiol, 164: 437-448. 156400
Dockery DW; Luttmann-Gibson H; Rich DQ; Link MS; Mittleman MA; Gold DR; Koutrakis P; Schwartz JD; Verrier RL.
(2005). Association of air pollution with increased incidence of ventricular tachyarrhythmias recorded by implanted
cardioverter defibrillators. Environ Health Perspect, 113: 670-674. 078995
Dockery DW; Luttmann-Gibson H; Rich DQ; Link MS; Schwartz JD; Gold DR; Koutrakis P; Verrier RL; Mittleman MA.
(2005). Particulate air pollution and nonfatal cardiac events Part II Association of air pollution with confirmed
arrhythmias recorded by implanted defibrillators. Health Effects Institute. Cambridge, MA. 090743
Dominici F; McDermott A; Daniels M; Zeger SL; Samet J. (2003). Revised Analyses of Time-Series Studies of Air
Pollution and Health: Mortality Among Residents of 90 Cities. Health Effects Institute. Boston, MA. 156407
Dominici F; McDermott A; Zeger SL; Samet JM. (2002). On the use of generalized additive models in time-series studies
of air pollution and health. Am J Epidemiol, 156: 193-203. 030458
Dominici F; Peng RD; Bell ML; Pham L; McDermott A; Zeger SL; Samet JL. (2006). Fine particulate air pollution and
hospital admission for cardiovascular and respiratory diseases. JAMA, 295: 1127-1134. 088398
Dominici F; Peng RD; Ebisu K; Zeger SL; Samet JM; Bell ML. (2007). Does the effect of PM10 on mortality depend on
PM nickel and vanadium content? A reanaly sis of the NMMAPS data. Environ Health Perspect, 115: 1701-1703.
099135
Dominici F; Peng RD; Zeger SL; White RH; Samet JM. (2007). Particulate air pollution and mortality in the United States:
did the risks change from 1987 to 2000?. Am J Epidemiol, 166: 880-888. 097361
Dubowsky SD; Suh H; Schwartz J; Coull BA; Gold DR. (2006). Diabetes, obesity, and hypertension may enhance
associations between air pollution and markers of systemic inflammation. Environ Health Perspect, 114: 992-998.
088750
Dusek R; Frank GP; Hildebrandt L; Curtius J; Schneider J; Walter S; Chand D; Drewnick F; Kings S; Jung D; Borrmann S;
Andreae MO. (2006). Size matters more than chemistry for cloud-nucleating ability of aerosol particles. Science,
312: 1375-1378. 155756
Duvall RM; Norris GA; Dailey LA; Burke JM; McGee JK; Gilmour MI; Gordon T; Devlin RB. (2008). Source
apportionment of particulate matter in the US and associations with lung inflammatory markers. Inhal Toxicol, 20:
671-683. 097969
Dvonch JT; Brook RD; Keeler GJ; Rajagopalan S; D'Alecy LG; Marsik FJ; Morishita M; Yip FY; Brook JR; Timm EJ;
Wagner JG; Harkema JR. (2004). Effects of concentrated fine ambient particles on rat plasma levels of asymmetric
dimethylarginine. Inhal Toxicol, 16: 473-480. 055741
Ebelt ST; Wilson WE; Brauer M. (2005). Exposure to ambient and nonambient components of particulate matter: a
comparison of health effects. Epidemiology, 16: 396-405. 056907
Elder A; Gelein R; Finkelstein J; Phipps R; Frampton M; Utell M; Kittelson DB; Watts WF; Hopke P; Jeong CH; Kim E;
Liu W; Zhao W; Zhuo L; Vincent R; Kumarathasan P; Oberdorster G. (2004). On-road exposure to highway
aerosols 2 Exposures of aged, compromised rats. Inhal Toxicol, 16 Suppl 1: 41-53. 087354
December 2009 6-221
-------
Elder ACP; Gelein R; Azadniv M; Frampton M; Finkelstein J; Oberdorster G. (2004). Systemic effects of inhaled ultrafine
particles in two compromised, aged rat strains. Inhal Toxicol, 16: 461-471. 055642
Erbas B; Kelly A-M; Physick B; Code C; Edwards M. (2005). Air pollution and childhood asthma emergency hospital
admissions: estimating intra-city regional variations. Int J Environ Health Res, 15: 11-20. 073849
Fairley D. (2003). Mortality and air pollution for Santa Clara County, California, 1989-1996, In: Revised analyses of time-
series studies of air pollution and health. Special report. Health Effects Institute. Boston,
MA.http://wwwhealtheffects.org/Pubs/TimeSeries.pdf. 042850
Fakhri AA; Ilic LM; Wellenius GA; Urch B; Silverman F; Gold DR; Mittleman MA. (2009). Autonomic effects of
controlled fine particulate exposure in young healthy adults: Effect modification by ozone. Environ Health
Perspect, 117: 1287-1292. 191914
Fan Z; Meng Q; Weisel C; Laumbach R; Ohman-Strickland P; Shalat S; Hernandez M; Black K. (2008). Acute exposure to
elevated PM (2.5) generated by traffic and cardiopulmonary health effects in healthy older adults. J Expo Sci
Environ Epidemiol, 19: 525-533. 191979
Farraj AK; Haykal-Coates N; Ledbetter AD; Evansky PA; Gavett SH. (2006). Inhibition of pan neurotrophin receptor p75
attenuates diesel particulate-induced enhancement of allergic airway responses in C57/B16J mice. Inhal Toxicol,
18:483-491.088469
Farraj AK; Haykal-Coates N; Ledbetter AD; Evansky PA; Gavett SH. (2006). Neurotrophin mediation of allergic airways
responses to inhaled diesel particles in mice. Toxicol Sci, 94: 183-192. 141730
Fedulov AV; Leme A; Yang Z; Dahl M; Lim R; Mariani TJ; Kobzik L. (2008). Pulmonary exposure to particles during
pregnancy causes increased neonatal asthma susceptibility. Am J Respir Cell Mol Biol, 38: 57-67. 097482
Ferdinands JM; Crawford CA; Greenwald R; Van Sickle D; Hunter E; Teague WG (2008). Breath acidification in
adolescent runners exposed to atmospheric pollution: a prospective, repeated measures observational study. Environ
Health Global Access Sci Source, 7: 10. 156433
Finnerty K; Choi J-E; Lau A; Davis-Gorman G; Diven C; Seaver N; Linak William P; Witten M; McDonagh Paul F.
(2007). Instillation of coarse ash particulate matter and lipopolysaccharide produces a systemic inflammatory
response in mice. J Toxicol Environ Health A, 70: 1957-1966. 156434
Fischer PH; Steerenberg PA; Snelder JD; Van Loveren H; Van Amsterdam JGC. (2002). Association between exhaled nitric
oxide, ambient air pollution and respiratory health in school children. Int Arch Occup Environ Health, 75: 348-353.
025731
Fischer SL; Koshland CP (2007). Daily and peak 1 h indoor air pollution and driving factors in a rural Chinese village.
Environ Sci Technol, 41: 3121-3126. 156435
Folino AF; Scapellato ML; Canova C; Maestrelli P; Bertorelli G; Simonato L; Iliceto S; Lotti M. (2009). Individual
exposure to particulate matter and the short-term arrhythmic and autonomic profiles in patients with myocardial
infarction. Eur Heart J, 30: 1614-1620. 191902
Forastiere F; Stafoggia M; Berti G; Bisanti L; Cernigliaro A; Chiusolo M; Mallone S; Miglio R; Pandolfi P; Rognoni M;
Serinelli M; Tessari R; Vigotti M; Perucci C. (2008). Particulate Matter and Daily Mortality: A Case-Crossover
Analysis of Individual Effect Modifiers. Epidemiology, 19: 571-580. 186937
Forastiere F; Stafoggia M; Picciotto S; Bellander T; DTppoliti D; Lanki T; Von Klot S; Nyberg F; Paatero P; Peters A;
Pekkanen J; Sunyer J; Perucci CA. (2005). A case-crossover analysis of out-of-hospital coronary deaths and air
pollution in Rome, Italy. Am J Respir Grit Care Med, 172: 1549-1555. 086323
Frampton MW. (2001). Systemic and cardiovascular effects of airway injury and inflammation: ultrafine particle exposure
in humans. Environ Health Perspect, 109: 529-532. 019051
Frampton MW; Stewart JC; Oberdorster G; Morrow PE; Chalupa D; Pietropaoli AP; Frasier LM; Speers DM; Cox C;
Huang LS; Utell MJ. (2006). Inhalation of ultrafine particles alters blood leukocyte expression of adhesion
molecules in humans. Environ Health Perspect, 114: 51-58. 088665
Franklin M; Koutrakis P; Schwartz J. (2008). The role of particle composition on the association between PM2.5 and
mortality. Epidemiology, 19: 680-689. 155779
Franklin M; Koutrakis P; Schwartz J . (2008). The role of particle composition on the association between PM2.5 and
mortality. Epidemiology, 19: 680-689. 097426
December 2009 6-222
-------
Franklin M; Zeka A; Schwartz J. (2007). Association between PM2.5 and all-cause and specific-cause mortality in 27 US
communities. JExpo Sci Environ Epidemiol, 17: 279-287. 091257
Fung KY; Khan S; Krewski D; Chen Y. (2006). Association between air pollution and multiple respiratory hospitalizations
among the elderly in Vancouver, Canada. Inhal Toxicol, 18: 1005-1011. 089789
Fung KY; Luginaah IKMG; Webster G. (2005). Air pollution and daily hospitalization rates for cardiovascular and
respiratory diseases in London, Ontario. Int J Environ Stud, 62: 677-685. 093262
Galan I; Tobias A; Banegas JR; Aranguez E. (2003). Short-term effects of air pollution on daily asthma emergency room
admissions. Eur Respir J, 22: 802-808. 087408
Gavett SH; Haykal-Coates N; Copeland L B; Heinrich J; Gilmour MI. (2003). Metal composition of ambient PM2.5
influences severity of allergic airways disease in mice. Environ Health Perspect, 111: 1471-1477. 053153
Gehlbach BK; Geppert E. (2004). The pulmonary manifestations of left heart failure. Chest, 125: 669-682. 155784
Gent JF; Koutrakis P; Belanger K; Triche E; Holford TR; Bracken MB; Leaderer BP (2009). Symptoms and medication
use in children with asthma and traffic-related sources of fine particle pollution. Environ Health Perspect, 117:
1168-1174. 180399
Gent JF; Triche EW; Holford TR; Belanger K; Bracken MB; Beckett WS; Leaderer BP. (2003). Association of low-level
ozone and fine particles with respiratory symptoms in children with asthma. JAMA, 290: 1859-1867. 052885
Gerlofs-Nijland ME; Rummelhard M; Boere AJF; Leseman DLAC; Duffin R; Schins RPF; Borm PJA; Sillanpaa M;
Salonen RO; Cassee FR. (2009). Particle induced toxicity in relation to transition metal and polycyclic aromatic
hydrocarbon contents . Environ Sci Technol, 43: 4729-4736. 190353
Gerlofs-Nijland ME; Dormans JA; Bloemen HJ; Leseman DL; John A; Boere F; Kelly FJ; Mudway IS; Jimenez AA;
Donaldson K; Guastadisegni C; Janssen NA; Brunekreef B; Sandstrom T; van Bree L; Cassee FR. (2007). Toxicity
of coarse and fine particulate matter from sites with contrasting traffic profiles. Inhal Toxicol, 19: 1055-1069.
097840
Ghelfi E; Rhoden CR; Wellenius GA; Lawrence J; Gonzalez-Flecha B. (2008). Cardiac oxidative stress and
electrophysiological changes in rats exposed to concentrated air particles are mediated by TRP-dependent
pulmonary reflexes. Toxicol Sci, 102: 328-336. 156468
Ghio AJ; Devlin RB. (2001). Inflammatory lung injury after bronchial instillation of air pollution particles. Am J Respir
Grit Care Med, 164: 704-708. 017122
Ghio AJ; Hall A; Bassett MA; Cascio WE; Devlin RB. (2003). Exposure to concentrated ambient air particles alters
hematologic indices in humans. Inhal Toxicol, 15: 1465-1478. 087363
Ghio AJ; Kim C; Devlin RB. (2000). Concentrated ambient air particles induce mild pulmonary inflammation in healthy
human volunteers. Am J Respir Grit Care Med, 162: 981-988. 012140
Gilliland FD; Li YF; Saxon A; Diaz-Sanchez D. (2004). Effect of glutathione-S-transferase Ml and PI genotypes on
xenobiotic enhancement of allergic responses: randomised, placebo-controlled crossover study. Lancet, 363: 119-
125. 156471
Gilmour MI; McGee J; Duvall Rachelle M; Dailey L; Daniels M; Boykin E; Cho S-H; Doerfler D; Gordon T; Devlin
Robert B. (2007). Comparative toxicity of size-fractionated airborne particulate matter obtained from different
cities in the United States. Inhal Toxicol, 19 Suppl 1: 7-16. 096433
Gilmour MI; O'Connor S; Dick CAJ; Miller CA; Linak WP. (2004). Differential pulmonary inflammation and in vitro
cytotoxicity of size-fractionated fly ash particles from pulverized coal combustion. J Air Waste Manag Assoc, 54:
286-295. 057420
Gilmour PS; Ziesenis A; Morrison ER; Vickers MA; Drost EM; Ford I; Karg E; Mossa C; Schroeppel A; Perron GA;
Heyder J; Greaves M; MacNee W; Donaldson K. (2004). Pulmonary and systemic effects of short-term inhalation
exposure to ultrafine carbon black particles. Toxicol Appl Pharmacol, 195: 35-44. 054175
Girardot SP; Ryan PB; Smith SM; Davis WT; Hamilton CB; Obenour RA; Renfro JR; Tromatore KA; Reed GD. (2006).
Ozone and PM25 exposure and acute pulmonary health effects: a study of hikers in the Great Smoky Mountains
National Park. Environ Health Perspect, 113: 612-617. 088271
December 2009 6-223
-------
Godleski JJ; Clarke RW; Coull BA; Saldiva PHN; Jiang NF; Lawrence J; Koutrakis P. (2002). Composition of inhaled
urban air particles determines acute pulmonary responses. Ann Occup Hyg, 46: 419-424. 156478
Godleski JJ; Verrier RL; Koutrakis P; Catalano P; Coull B; Reinisch U; Lovett EG; Lawrence J; Murthy GG; Wolfson JM;
Clarke RW; Nearing BD; Killingsworth C. (2000). Mechanisms of morbidity and mortality from exposure to
ambient air particles. Res Rep Health Eff Inst, 91: 5-88; discussion 89-103. 000738
Gold DR; Litonjua AA; Zanobetti A; Coull BA; Schwartz J; MacCallum G; Verrier RL; Nearing BD; Canner MJ; Suh H;
Stone PH. (2005). Air pollution and ST-segment depression in elderly subjects. Environ Health Perspect, 113: 883-
887. 087558
Goldberg MS; Giannetti N; Burnett RT; Mayo NE; Valois MF; Brophy JM. (2008). A panel study in congestive heart
failure to estimate the short-term effects from personal factors and environmental conditions on oxygen saturation
and pulse rate. Occup Environ Med, 65: 659-666. 180380
Gong H Jr; Sioutas C; Linn WS. (2003). Controlled exposures of healthy and asthmatic volunteers to concentrated ambient
particles in metropolitan Los Angeles. Health Effects Institute. Boston, MA. 087365
Gong H; Sioutas C; Linn WS; Clark KW; Terrell SL; Terrell LL; Anderson KR; Kim S; Chang MC. (2000). Controlled
human exposures to concentarted ambient fine particles in metropolitan Los Angeles: Methodology and preliminary
health-effect findings. Inhal Toxicol, 12: 107-119. 155799
Gong H Jr; Linn WS; Clark KW; Anderson KR; Geller MD; Sioutas C. (2005). Respiratory responses to exposures with
fine particulates and nitrogen dioxide in the elderly with and without COPD. Inhal Toxicol, 17: 123-132. 087921
Gong H Jr; Linn WS; Clark KW; Anderson KR; Sioutas C; Alexis NE; Cascio WE; Devlin RB. (2008). Exposures of
healthy and asthmatic volunteers to concentrated ambient ultrafine particles in Los Angeles. Inhal Toxicol, 20: 533-
545. 156483
Gong H Jr; Linn WS; Sioutas C; Terrell SL; Clark KW; Anderson KR; Terrell LL. (2003). Controlled exposures of healthy
and asthmatic volunteers to concentrated ambient fine particles in Los Angeles. Inhal Toxicol, 15: 305-325. 042106
Gong H Jr; Linn WS; Terrell SL; Anderson KR; Clark KW; Sioutas C; Cascio WE; Alexis N; Devlin RB. (2004).
Exposures of elderly volunteers with and without chronic obstructive pulmonary disease (COPD) to concentrated
ambient fine particulate pollution. Inhal Toxicol, 16: 731-744. 087964
Gong H Jr; Linn WS; Terrell SL; Clark KW; Geller MD; Anderson KR; Cascio WE; Sioutas C. (2004). Altered heart-rate
variability in asthmatic and healthy volunteers exposed to concentrated ambient coarse particles. Inhal Toxicol, 16:
335-343. 055628
Gordian ME; Choudhury AH. (2003). PM10 and asthma medication in schoolchildren. Arch Environ Occup Health, 58: 42-
47. 054842
Gowdy K; Krantz QT; Daniels M; Linak WP; Jaspers I; Gilmour MI. (2008). Modulation of pulmonary inflammatory
responses and antimicrobial defenses in mice exposed to diesel exhaust. Toxicol Appl Pharmacol, 229(3): 310-319.
097226
Graff D; Cascio W; Rappold A; Zhou H; Huang Y; Devlin R. (2009). Exposure to concentrated coarse air pollution
particles causes mild cardiopulmonary effects in healthy young adults . Environ Health Perspect, 117: 1089-1094.
191981
Gurgueira SA; Lawrence J; Coull B; Murthy GGK; Gonzalez-Flecha B. (2002). Rapid increases in the steady-state
concentration of reactive oxygen species in the lungs and heart after particulate air pollution inhalation. Environ
Health Perspect, 110: 749-755. 036535
Hajat S; Haines A; Atkinson RW; Bremner SA; Anderson HR; Emberlin J. (2001). Association between air pollution and
daily consultations with general practitioners for allergic rhinitis in London, United Kingdom. Am J Epidemiol,
153:704-714.016693
Halonen JI; Lanki T; Yli-Tuomi T; Kulmala M; Tiittanen P; Pekkanen J. (2008). Urban air pollution, and asthma and
COPD hospital emergency room visits. Thorax, 63: 635-641. 189507
Halonen JI; Lanki T; Yli-Tuomi T; Tiittanen P; Kulmala M; Pekkanen. (2009). Particulate air pollution and acute
cardiorespiratory hospital admissions and mortality among the elderly. Epidemiology, 20: 143-153. 180379
Hamada K; Suzaki Y; Leme A; Ito T; Miyamoto K; Kobzik L; Kimura H. (2007). Exposure of pregnant mice to an air
pollutant aerosol increases asthma susceptibility in offspring. J Toxicol Environ Health A, 70: 688-695. 091235
December 2009 6-224
-------
Hamade AK; Rabold R; Tankersley CG. (2008). Adverse cardiovascular effects with acute particulate matter and ozone
exposures: interstrain variation in mice. Environ Health Perspect, 116: 1033-1039. 156515
Hanigan 1C; Johnston FH; Morgan GG. (2008). Vegetation fire smoke, indigenous status and cardio respiratory hospital
admissions in Darwin, Australia, 1996-2005: a time-series study. Environ Health, 7: 42. 156518
Hao M; Cornier S; Wang M; Lee James J; Nel A. (2003). Diesel exhaust particles exert acute effects on airway
inflammation and function in murine allergen provocation models. J Allergy Clin Immunol, 112: 905-914. 096565
Hapcioglu B; Issever H; Kocyigit E; Disci R; Vatansever S; Ozdilli K. (2006). The effect of air pollution and
meteorological parameters on chronic obstructive pulmonary disease at an Istanbul hospital. Indoor Built Environ,
15: 147-153.093263
Happo MS; Salonen RO; Halinen AI; Jalava PI; Pennanen AS; Kosma VM; Sillanpaa M; Hillamo R; Brunekreef B;
Katsouyanni K; Sunyer J; Hirvonen MR. (2007). Dose and time dependency of inflammatory responses in the
mouse lung to urban air coarse, fine, and ultrafine particles from six European cities. Inhal Toxicol, 19: 227-246.
096630
Harder V; Gilmour P; Lentner B; Karg E; Takenaka S; Ziesenis A; Stampfl A; Kodavanti U; Heyder J; Schulz H. (2005).
Cardiovascular responses in unrestrained WKY rats to inhaled ultrafine carbon particles. Inhal Toxicol, 17: 29-42.
087371
Harkema JR; Keeler G; Wagner J; Morishita M; Timm E; Hotchkiss J; Marsik F; Dvonch T; Kaminski N; Barr E. (2004).
Effects of concentrated ambient particles on normal and hypersecretory airways in rats. Health Effects Institute.
Boston, MA. 056842
Harrod KS; Jaramillo RJ; Berger JA; Gigliotti AP; Seilkop SK; Reed MD. (2005). Inhaled diesel engine emissions reduce
bacterial clearance and exacerbate lung disease to Pseudomonas aeruginosa infection in vivo. Toxicol Sci, 83: 155-
165. 088144
Harrod KS; Jaramillo RJ; Rosenberger CL; Wang S-Z; Berger JA; McDonald JD; Reed MD. (2003). Increased
susceptibility to RSV infection by exposure to inhaled diesel engine emissions. Am J Respir Cell Mol Biol, 28: 451-
463. 097046
Hayek T; Oiknine J; Brook JG; Aviram M. (1994). Increased plasma and lipoprotein lipid peroxidation in apo E-deficient
mice. Biochem Biophys Res Commun, 201: 1567-1574. 156527
Hedley AJ; Wong C-M; Thach TQ; Ma S; Lam T-H; Anderson HR. (2002). Cardiorespiratory and all-cause mortality after
restrictions on sulphur content of fuel in Hong Kong: an intervention study. Lancet, 360: 1646-1652. 040284
Heidenfelder BL; Reif DM; Harkema JR; Cohen Hubal EA; Hudgens EE; Bramble LA; Wagner JG; Morishita M; Keeler
GJ; Edwards SW; Gallagher JE. (2009). Comparative microarray analysis and pulmonary changes in brown
Norway rats exposed to ovalbumin and concentrated air particulates. Toxicol Sci, 108: 207-221. 190026
Henneberger A; Zareba W; Ibald-Mulli A; Ruckerl R; Cyrys J; Couderc J-P; Mykins B; Woelke G; WichmannH-E; Peters
A. (2005). Repolarization changes induced by air pollution in ischemic heart disease patients. Environ Health
Perspect, 113: 440-446. 087960
Henrotin JB; Besancenot JP; Bejot Y; Giroud M. (2007). Short-term effects of ozone air pollution on ischaemic stroke
occurrence: A case-crossover analysis from a 10-year population-based study in Dijon, France. Occup Environ
Med, 64: 439-445. 093270
Hinwood AL; De Klerk N; Rodriguez C; Jacoby P; Runnion T; Rye P; Landau L; Murray F; Feldwick M; Spickett J.
(2006). The relationship between changes in daily air pollution and hospitalizations in Perth, Australia 1992-1998:
A case-crossover study. Int J Environ Health Res, 16: 27-46. 088976
Hirshon JM; Shardell M; Alles S; Powell JL; Squibb K; Ondov J; Blaisdell CJ. (2008). Elevated ambient air zinc increases
pediatric asthma morbidity. Environ Health Perspect, 116: 826-831. 180375
Hogervorst JG; de Kok TM; Briede JJ; Wesseling G; Kleinjans JC; van Schayck CP (2006). Relationship between radical
generation by urban ambient particulate matter and pulmonary function of school children. J Toxicol Environ
Health A, 69: 245-262. 156559
Holguin F; Flores S; Ross Z; Cortez M; Molina M; Molina L; Rincon C; Jerrett M; Berhane K; Granados A; Romieu I.
(2007). Traffic-related exposures, airway function, inflammation, and respiratory symptoms in children. Am J
Respir Grit Care Med, 176: 1236-1242. 099000
December 2009 6-225
-------
Holguin F; Tellez-Rojo MM; Hernandez M; Cortez M; Chow JC; Watson JG; Mannino D; Romieu I. (2003). Air pollution
and heart rate variability among the elderly in Mexico City. Epidemiology, 14: 521-527. 057326
Holloman CH; Bortnick SM; Morara M; Strauss WJ; Calder CA. (2004). A Bayesian hierarchical approach for relating
PM2.5 exposure to cardiovascular mortliaty in North Carolina. Environ Health Perspect, 112: 1282-1288. 087375
Hong Y-C; Hwang S-S; Kim JH; Lee K-H; Lee H-J; Lee K-H; Yu S-D; Kim D-S. (2007). Metals in particulate pollutants
affect peak expiratory flow of schoolchildren. Environ Health Perspect, 115: 430-434. 091347
Hopke PK; Ito K; Mar T; Christensen WF; Eatough DJ; Henry RC; Kim E; Laden F; Lall R; Larson TV; Liu H; Neas L;
Pinto J; Stolzel M; Suh H; Paatero P; Thurston GD. (2006). PM source apportionment and health effects: 1
Intercomparison of source apportionment results. J Expo Sci Environ Epidemiol, 16: 275-286. 088390
Host S; Larrieu S; Pascal L; Blanchard M; Declercq C; Fabre P; Jusot JF; Chardon B; Le Tertre A; Wagner V; Prouvost H;
Lefranc A. (2007). Short-term Associations between Fine and Coarse Particles and Cardiorespiratory
Hospitalizations in Six French Cities. Occup Environ Med, 18: S107-S108. 155851
Host S; Larrieu S; Pascal L; Blanchard M; Declercq C; Fabre P; Jusot JF; Chardon B; Le Tertre A; Wagner V; Prouvost H;
Lefranc A. (2008). Short-term associations between fine and coarse particles and hospital admissions for
Cardiorespiratory diseases in six French cities. Occup Environ Med, 65: 544-551. 155852
Huang Y-CT; Ohio AJ; Stonehuerner J; McGee J; Carter JD; Grambow SC; Devlin RB. (2003). The role of soluble
components in ambient fine particles-induced changes in human lungs and blood. Inhal Toxicol, 15: 327-342.
087377
Huber SA; Sakkinen P; Conze D; Hardin N; Tracy R. (1999). Interleukin-6 exacerbates early atherosclerosis in mice.
Arterioscler Thromb Vase Biol, 19: 2364-2367. 156575
Hwang J-S; Chan C-C. (2002). Effects of air pollution on daily clinic visits for lower respiratory tract illness. Am J
Epidemiol, 155: 1-10. 023222
Hwang J-S; Nadziejko C; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in
mice: III Acute and chronic effects of CAPs on heart rate, heart-rate fluctuation, and body temperature. Inhal
Toxicol, 17: 199-207. 087957
Ibald-Mulli A; Timonen KL; Peters A; Heinrich J; Wolke G; Lanki T; Buzorius G; Kreyling WG; De Hartog J; Hoek G;
Ten Brink HM; Pekkanen J. (2004). Effects of particulate air pollution on blood pressure and heart rate in subjects
with cardiovascular disease: a multicenter approach. Environ Health Perspect, 112: 369-377. 087415
Inoue K; Takano H; Yanagisawa R; Sakurai M; Ichinose T; Sadakane K; Yoshikawa T. (2005). Effects of nano particles on
antigen-related airway inflammation in mice. Respir Res, 6: 106. 088625
Ito K. (2003). Associations of particulate matter components with daily mortality and morbidity in Detroit, Michigan, In:
Revised analyses of time-series studies of air pollution and health. Special report. Health Effects Institute. Boston,
MA. R828112. http: //www.healtheffects.org/Pubs/TimeSeries.pdf. 042856
Ito K; Christensen WF; Eatough DJ; Henry RC; Kim E; Laden F; Lall R; Larson TV; Neas L; Hopke PK; Thurston GD.
(2006). PM source apportionment and health effects: 2 An investigation of intermethod variability in associations
between source-apportioned fine particle mass and daily mortality in Washington, DC. J Expo Sci Environ
Epidemiol, 16: 300-310. 088391
Ito K; Thurston GD; Silverman RA. (2007). Characterization of PM2.5, gaseous pollutants, and meteorological interactions
in the context of time-series health effects models. J Expo Sci Environ Epidemiol, 17 Suppl 2: S45-S60. 156594
Ito K; Thurston GD; Silverman RA;. (2007). Association between coarse particles and asthma emergency department (ED)
visits in New York City. Presented at American Thoracic Society international conference 2007, San Francisco, CA.
091262
Ito T; Okumura H; Tsukue N; Kobayashi T; Honda K; Sekizawa K. (2006). Effect of diesel exhaust particles on mRNA
expression of viral and bacterial receptors in rat lung epithelial L2 cells. Toxicol Lett, 165: 66-70. 096648
Ito T; Suzuki T; Tamura K; Nezu T; Honda K; Kobayashi T. (2008). Examination of mRNA expression in rat hearts and
lungs for analysis of effects of exposure to concentrated ambient particles on cardiovascular function. Toxicol Sci,
243:271-283.096823
Jaffe DH; Singer ME; Rimm AA. (2003). Air pollution and emergency department visits for asthma among Ohio Medicaid
recipients, 1991-1996. Environ Res, 91: 21-28. 041957
December 2009 6-226
-------
Jalaludin B; Morgan G; Lincoln D; Sheppeard V; Simpson R; Corbett S. (2006). Associations between ambient air
pollution and daily emergency department attendances for cardiovascular disease in the elderly (65+ years),
Sydney, Australia. J Expo Sci Environ Epidemiol, 16: 225-237. 189416
Jalaludin BB; O'Toole BI; Leeder SR. (2004). Acute effects of urban ambient air pollution on respiratory symptoms,
asthma medication use, and doctor visits for asthma in a cohort of Australian children. Environ Res, 95: 32-42.
056595
Jalava PI; Salonen RO; Halinen AI; Penttinen P; Pennanen AS; Sillanpaa M; Sandell E; Hillamo R; Hirvonen M-R. (2006).
In vitro inflammatory and cytotoxic effects of size-segregated particulate samples collected during long-range
transport of wildfire smoke to Helsinki. Toxicol Appl Pharmacol, 215: 341-353. 155872
Jalava PI; Salonen RO; Pennanen AS; Happo MS; Penttinen P; Halinen AI; Sillanpaa M; Hillamo R; Hirvonen MR. (2008).
Effects of solubility of urban air fine and coarse particles on cytotoxic and inflammatory responses in RAW 264.7
macrophage cell line. Toxicol Appl Pharmacol, 229: 146-60. 098968
Janes H; Sheppard L; Lumley T. (2005). Case-crossover analyses of air pollution exposure data: referent selection
strategies and their implications for bias. Epidemiology, 16: 717-726. 087535
Jansen KL; Larson TV; Koenig JQ; Mar TF; Fields C; Stewart J; Lippmann M. (2005). Associations between health effects
and particulate matter and black carbon in subjects with respiratory disease. Environ Health Perspect, 113: 1741-
1746. 082236
Janssen NAH; Schwartz J; Zanobetti A; Suh HH. (2002). Air conditioning and source-specific particles as modifiers of the
effect of PM10 on hospital admissions for heart and lung disease. Environ Health Perspect, 110: 43-49. 016743
Johnston FH; Bailie RS; Pilotto LS; Hanigan 1C. (2007). Ambient biomass smoke and cardio-respiratory hospital
admissions in Darwin, Australia. BMC Public Health, 7: 240. 155882
Johnston FH; Webby RJ; Pilotto LS; Bailie RS; Parry DL; Halpin SJ. (2006). Vegetation fires, particulate air pollution and
asthma: a panel study in the Australian monsoon tropics. Int J Environ Health Res, 16: 391-404. 091386
Just J; Segala C; Sahraoui F; Priol G; Grimfeld A; Neukirch F. (2002). Short-term health effects of particulate and
photochemical air pollution in asthmatic children. Eur Respir J, 20: 899-906. 035429
Kannel WB; Abbott RD; Savage DD; McNamara PM. (1983). Coronary heart disease and atrial fibrillation: the
Framingham Study. Am Heart J, 106: 389-396. 156623
Katsouyanni K; Touloumi G; Samoli E; Petasakis Y; Analitis A; Le Tertre A; Rossi G; Zmirou D; Ballester F; Boumghar A;
Anderson HR; Wojtyniak B; Paldy A; Braunstein R; Pekkanen J; Schindler C; Schwartz J. (2003). Sensitivity
analysis of various models of short-term effects of ambient particles on total mortality in 29 cities in APHEA2.
042807
Kelishadi R; Mirghaffari N; Poursafa P; Gidding S. (2009). Lifestyle and environmental factors associated with
inflammation, oxidative stress and insulin resistance in children. Atherosclerosis, 203: 311-319. 191960
Kettunen J; Lanki T; Tiittanen P; Aalto PP; Koskentalo T; Kulmala M; Salomaa V; Pekkanen J. (2007). Associations of fine
and ultrafine particulate air pollution with stroke mortality in an area of low air pollution levels. Stroke, 38: 918-
922. 091242
Kim SY; O'Neill MS; Lee JT; Cho Y; Kim J; Kim H. (2007). Air pollution, socioeconomic position, and emergency
hospital visits for asthma in Seoul, Korea. Int Arch Occup Environ Health, 80: 701-710. 092837
Klein-Patel ME; Diamond G; Boniotto M; Saad S; Ryan LK. (2006). Inhibition of beta-defensin gene expression in airway
epithelial cells by low doses of residual oil fly ash is mediated by vanadium. Toxicol Sci, 92: 115-125. 097092
Kleinman M; Sioutas C; Stram D; Froines J; Cho A; Chakrabarti B; Hamade A; Meacher D; Oldham M. (2005). Inhalation
of concentrated ambient particulate matter near a heavily trafficked road stimulates antigen-induced airway
responses in mice. J Air Waste Manag Assoc, 55: 1277-1288. 087880
Kleinman MT; Araujo JA; Nel A; Sioutas C; Campbell A; Cong PQ; Li H; Bondy SC. (2008). Inhaled ultrafine particulate
matter affects CNS inflammatory processes and may act via MAP kinase signaling pathways. Toxicol Lett, 178:
127-130. 190074
Kleinman MT; Sioutas C; Froines JR; Fanning E; Hamade A; Mendez L; Meacher D; Oldham M. (2007). Inhalation of
concentrated ambient particulate matter near a heavily trafficked road stimulates antigen-induced airway responses
in mice. Inhal Toxicol, 19 Suppl 1: 117-126. 097082
December 2009 6-227
-------
Klemm RJ; Lipfert FW; Wyzga RE; Gust C. (2004). Daily mortality and air pollution in Atlanta: two years of data from
ARIES. Inhal Toxicol, 16 Suppl 1: 131-141. 056585
Klemm RJ; Mason R. (2003). Replication of reanalysis of Harvard Six-City mortality study. In HEI Special Report:
Revised Analyses of Time-Series Studies of Air Pollution and Health, Part II (pp. 165-172). Boston, MA: Health
Effects Institute. 042801
Knuckles T; Lund A; Lucas S; Campen M. (2008). Diesel exhaust exposure enhances venoconstriction via uncoupling of
eNOS. Toxicol Appl Pharmacol, 230: 346-351. 191987
Ko FW; Tarn W; Wong TW; Lai CK; Wong GW; Leung TF; Ng SS; Hui DS. (2007). Effects of air pollution on asthma
hospitalization rates in different age groups in Hong Kong. Clin Exp Allergy, 37: 1312-9. 196606
Ko FWS; Tarn W; Wong TW; Chan DPS. (2007). Temporal relationship between air pollutants and hospital admissions for
chronic obstructive pulmonary disease in Hong Kong. Thorax, 62: 780-785. 091639
Ko FWS; Tarn W; Wong TW; Lai CKW;. (2007). Effects of air pollution on asthma hospitalization rates in different age
groups inHong Kong. Clin Exp Allergy, 37: 1312-1319. 092844
Kodavanti UP; Schladweiler MC; Ledbetter AD; Hauser R; Christiani DC; McGee J; Richards JR; Costa DL. (2002).
Temporal association between pulmonary and systemic effects of particulate matter in healthy and cardiovascular
compromised rats. J Toxicol Environ Health A, 65: 1545-1569. 025236
Kodavanti UP; Schladweiler MC; Ledbetter AD; McGee JK; Walsh L; Gilmour PS; Highfill JW; Davies D; Pinkerton KE;
Richards JH; Crissman K; Andrews D; Costa DL. (2005). Consistent pulmonary and systemic responses from
inhalation of fine concentrated ambient particles: roles of rat strains used and physicochemical properties. Environ
Health Perspect, 113: 1561-1568. 087946
Kodavanti UP; Schladweiler MCJ; Ledbetter AD; Hauser R; Christiani DC; Samet JM; McGee J; Richards JH; Costa DL.
(2002). Pulmonary and systemic effects of zinc-containing emission particles in three rat strains: multiple exposure
scenarios. Toxicol Sci, 70: 73-85. 035344
Koenig J; Allen R; Larson T; Liu S. (2005). Response to "Indoor- and Outdoor-Generated Particles and Children with
asthma" [letter]. Environ Health Perspect, 113: A581. 088999
Koenig JQ; Jansen K; Mar TF; Lumley T; Kaufman J; Trenga CA; Sullivan J; Liu LJ; Shapiro GG; Larson TV. (2003).
Measurement of offline exhaled nitric oxide in a study of community exposure to air pollution. Environ Health
Perspect, 111: 1625-1629. 156653
Koenig JQ; Mar TF; Allen RW; Jansen K; Lumley T; Sullivan JH; Trenga CA; Larson T; Liu LJ. (2005). Pulmonary effects
of indoor- and outdoor-generated particles in children with asthma. Environ Health Perspect, 113: 499-503. 087384
Koken PJM; Piver WT; Ye F; Elixhauser A; Olsen LM; Portier CJ. (2003). Temperature, air pollution, and hospitalization
for cardiovascular diseases among elderly people in Denver. Environ Health Perspect, 111: 1312-1317. 049466
Kongerud J; Madden MC; Hazucha M; Peden D. (2006). Nasal responses in asthmatic and nonasthmatic subjects following
exposure to diesel exhaust particles. Inhal Toxicol, 18: 589-594. 156656
Kooter IM; Boere AJ; Fokkens PH; Leseman DL; Dormans JA; Cassee FR. (2006). Response of spontaneously
hypertensive rats to inhalation of fine and ultrafine particles from traffic: experimental controlled study. Part Fibre
Toxicol, 15: 3-7. 097547
Kuo HW; Lai JS; Lee MC; Tai RC; Lee MC. (2002). Respiratory effects of air pollutants among asthmatics in central
Taiwan. Arch Environ Occup Health, 57: 194-200. 036310
Laden F; Neas LM; Dockery DW; Schwartz J. (2000). Association of fine particulate matter from different sources with
daily mortality in six US cities. Environ Health Perspect, 108: 941-947. 012102
Lagorio S; Forastiere F; Pistelli R; lavarone I; Michelozzi P; Fano V; Marconi A; Ziemacki G; Ostro BD. (2006). Air
pollution and lung function among susceptible adult subjects: a panel study. Environ Health, 5:11. 089800
Lai LW; Cheng WL. (2008). The impact of air quality on respiratory admissions during Asian dust storm periods. Int J
Environ Health Res, 18: 429-450. 180301
Lanki T; De Hartog JJ; Heinrich J; Hoek G; Janssen NAH; Peters A; Stolzel M; Timonen KL; Vallius M; Vanninen E;
Pekkanen J. (2006). Can we identify sources of fine particles responsible for exercise-induced ischemia on days
with elevated air pollution? The ULTRA study. Environ Health Perspect, 114: 655-660. 088412
December 2009 6-228
-------
Lanki T; Hoek G; Timonen K; Peters A; Tiittanen P; Vanninen E; Pekkanen J. (2008). Hourly variation in fine particle
exposure is associated with transiently increased risk of ST segment depression. Br Med J, 65: 782-786. 191984
Lanki T; Pekkanen J; Aalto P; Elosua R; Berglind N; D'Ippoliti D; Kulmala M; Nyberg F; Peters A; Picciotto S; Salomaa V;
Sunyer J; Tiittanen P; Von Klot S; Forastiere F; for the HEAPSS Study Group. (2006). Associations of traffic-
related air pollutants with hospitalisation for first acute myocardial infarction: the HEAPSS study. Occup Environ
Med, 63: 844-851.089788
Larrieu S; Jusot J-F; Blanchard M; Prouvost H; Declercq C; Fabre P; Pascal L; Le Tertre A; Wagner V; Riviere S; Chardon
B; Borelli D; Cassadou S; Eilstein D; Lefranc A. (2007). Short term effects of air pollution on hospitalizations for
cardiovascular diseases in eight French cities: The PSAS program. Sci Total Environ, 387: 105-112. 093031
Larrieu S; Lefranc A; Gault G; Chatignoux E; Couvy F; Jouves B; Filleul L. (2009). Are the Short-term Effects of Air
Pollution Restricted to Cardiorespiratory Diseases?. Am J Epidemiol, 169: 1-8. 180294
Larsson B-M; Sehistedt M; Grunewald J; Skold CM; Lundin A; Blomberg A; Sandstrom T; Eklund A; Svartengren M.
(2007). Road tunnel air pollution induces bronchoalveolar inflammation in healthy subjects. Eur Respir J, 29: 699-
705.091375
Last JA; Ward R; Temple L; Pinkerton KE; Kenyon NJ. (2004). Ovalbumin-induced airway inflammation and fibrosis in
mice also exposed to ultrafine particles. Inhal Toxicol, 16: 93-102. 097334
Laupacis A; Boysen G; Connolly S. (1994). Risk factors for stroke and efficacy of antithrombotic therapy in atrial
fibrillation: Analysis of pooled data from five randomized controlled trials. Arch Intern Med, 154: 1449-1457.
190901
Laurent O; Pedrono G; Filleul L; Segala C; Lefranc A; Schillinger C; Riviere E; Bard D. (2009). Influence of
socioeconomic deprivation on the relation between air pollution and beta-agonist sales for asthma. Chest, 135: 717-
723. 192129
Laurent O; Pedrono G; Segala C; Filleul L; Havard S; Deguen S; Schillinger C; Riviere E; Bard D. (2008). Air pollution,
asthma attacks, and socioeconomic deprivation: a small-area case-crossover study. Am J Epidemiol, 168: 58-65.
156672
Lay JC; Bennett WD; Kim CS; Devlin RB; Bromberg PA. (1998). Retention and intracellular distribution of instilled iron
oxide particles in human alveolar macrophages. Am J Respir Cell Mol Biol, 18: 687-695. 007683
Lay JC;ZemanKL; Ghio AJ; Bennett WD. (2001). Effects of inhaled iron oxide particles on alveolar epithelial
permeability in normal subjects. Inhal Toxicol, 13: 1065-1078. 020613
Le Tertre A; Medina S; Samoli E; Forsberg B; Michelozzi P; Boumghar A; Vonk JM; Bellini A; Atkinson R; Ayres JG;
Sunyer J; Schwartz J; Katsouyanni K. (2002). Short term effects of particulate air pollution on cardiovascular
diseases in eight European cities. J Epidemiol Community Health, 56: 773-779. 023746
Le Tertre A; Medina S; Samoli E; Forsberg B; Michelozzi P; Boumghar A; Vonk JM; Bellini A; Atkinson R; Ayres JG;
Sunyer J; Schwartz J; Katsouyanni K. (2003). Short-term effects of particulate air pollution on cardiovascular
diseases in eight European cities. In HEI Special Report: Revised Analyses of Time-Series Studies of Air Pollution
and Health, Part II (pp. 173-176). Boston: Health Effects Institute. 042820
Le Tertre A; Schwartz J; Touloumi G. (2005). Empirical Bayes and adjusted estimates approach to estimating the relation of
mortality to eposure of PM10. Risk Anal, 25: 711-718. 087560
Lee IM; Tsai SS; Ho CK; Chiu HF; Yang CY. (2007). Air pollution and hospital admissions for congestive heart failure in a
tropical city: Kaohsiung, Taiwan. Inhal Toxicol, 19: 899-904. 093271
Lee J-T; Kim H; Song H; Hong Y-C; Cho Y-S; Shin S-Y; Hyun Y-J; Kim Y-S. (2002). Air pollution and asthma among
children in Seoul, Korea. Epidemiology, 13: 481-484. 034826
Lee J-T; Son J-Y; Cho Y-S. (2007). A comparison of mortality related to urban air particles between periods with Asian dust
days and without Asian dust days in Seoul, Korea, 2000-2004. Environ Res, 105: 409-13. 093042
Lee JT; Kim H; Cho YS; Hong YC; Ha EH; Park H. (2003). Air pollution and hospital admissions for ischemic heart
diseases among individuals 64+years of age residing in Seoul, Korea. Arch Environ Health, 58: 617-623. 095552
Lee SL; Wong WHS; Lau YL. (2006). Association between air pollution and asthma admission among children in Hong
Kong. Clin Exp Allergy, 36: 1138-1146. 090176
December 2009 6-229
-------
Lei Y-C; Chan C-C; Wang P-Y; Lee C-T; Cheng T-J. (2004). Effects of Asian dust event particles on inflammation markers
in peripheral blood and bronchoalveolar lavage in pulmonary hypertensive rats. Environ Res, 95: 71-76. 087884
Lei YC; Chen MC; Chan CC; Wang PY; Lee CT; Cheng TJ. (2004). Effects of concentrated ambient particles on airway
responsiveness and pulmonary inflammation in pulmonary hypertensive rats. Inhal Toxicol, 16: 785-792. 087999
Levine AM; Gwozdz J; Stark J; Bruno M; Whitsett J; Korfhagen T. (1999). Surfactant protein-A enhances respiratory
syncytial virus clearance in vivo. JClin Invest, 103: 1015-1021. 156687
Le Vine AM; Whitsett JA. (2001). Pulmonary collectins and innate host defense of the lung. Microb Infect, 3: 161-166.
155928
Levy D; Sheppard L; Checkoway H; Kaufman J; Lumley T; Koenig J; Siscovick D. (2001). A case-crossover analysis of
particulate matter air pollution and out-of-hospital primary cardiac arrest. Epidemiology, 12: 193-199. 017171
Lewis CW; Klouda GA; Ellenson WD. (2004). Radiocarbon measurement of the biogenic contribution to summertime PM-
2.5 ambient aerosol in Nashville, TN. Atmos Environ, 38: 6053-6061. 097498
Lewis CW; Norris GA; Conner TL; Henry RC. (2003). Source apportionment of Phoenix PM2.5 aerosol with the Unmix
Receptor model. J Air Waste Manag Assoc, 53: 325-338. 088413
Lewis TC; Robins TG; Dvonch JT; Keeler GJ; Yip FY; Mentz GB; Lin X; Parker EA; Israel BA; Gonzalez L; Hill Y
(2005). Air pollution-associated changes in lung function among asthmatic children in Detroit. Environ Health
Perspect, 113: 1068-1075. 081079
Li N; Wang M; Bramble LA; Schmitz DA; Schauer JJ; Sioutas C; Harkema JR; Nel AE. (2009). The adjuvant effect of
ambient particulate matter is closely reflected by the particulate oxidant potential. Environ Health Perspect, 117:
1116-1123. 190457
Li Y-J; Kawada T; Matsumoto A; Azuma A; Kudoh S; Takizawa H; Sugawara I. (2007). Airway inflammatory responses to
oxidative stress induced by low-dose diesel exhaust particle exposure differ between mouse strains. Exp Lung Res,
33: 227-244. 155929
Liao D; Duan Y; Whitsel EA; Zheng Z-J; Heiss G; Chinchilli VM; Lin H-M. (2004). Association of higher levels of
ambient criteria pollutants with impaired cardiac autonomic control: a population-based study. Am J Epidemiol,
159:768-777.056590
Liao D; Heiss G; Chinchilli VM; Duan Y; Folsom AR; Lin HM; Salomaa V (2005). Association of criteria pollutants with
plasma hemostatic/inflammatory markers: a population-based study. J Expo Sci Environ Epidemiol, 15: 319-328.
088677
Liao D; Whitsel EA; Duan Y; Lin HM; Quibrera PM; Smith R; Peuquet DJ; Prineas RJ; Zhang ZM; Anderson G. (2009).
Ambient particulate air pollution and ectopy-the environmental epidemiology of arrhythmogenesis in Women's
Health Initiative Study, 1999-2004. J Toxicol Environ Health A, 72: 30-38. 199519
Lin M; Chen Y; Burnett RT; Villeneuve PJ; Krewski D. (2002). The influence of ambient coarse particulate matter on
asthma hospitalization in children: case-crossover and time-series analyses. Environ Health Perspect, 110: 575-581.
026067
Lin M; Stieb DM; Chen Y. (2005). Coarse particulate matter and hospitalization for respiratory infections in children
younger than 15 years in Toronto: a case-crossover analysis. Pediatrics, 116: 235-240. 087828
Linn WS; Szlachcic Y; Gong H Jr; Kinney PL; Berhane KT. (2000). Air pollution and daily hospital admissions in
metropolitan Los Angeles. Environ Health Perspect, 108: 427-434. 002839
Lipfert FW; Morris SC; Wyzga RE. (2000). Daily mortality in the Philadelphia metropolitan area and size-classified
particulate matter. J Air Waste Manag Assoc, 50: 1501-1513. 004088
Lippmann M. (2000). Environmental toxicants: human exposures and their health effects. New York: John Wiley and Sons.
024579
Lippmann M; Hwang J; Maclejczyk P; Chen L. (2005). PM source apportionment for short-term cardiac function changes
inApoE-/- mice. Environ Health Perspect, 113: 1575-1579. 087453
Lippmann M; Ito K; Hwang JS; Maciejczyk P; Chen LC. (2006). Cardiovascular effects of nickel in ambient air. Environ
Health Perspect, 114: 1662-1669. 091165
December 2009 6-230
-------
Lippmann M; Ito K; Nadas A; Burnett RT. (2000). Association of participate matter components with daily mortality and
morbidity in urban populations. Health Effects Institute. Cambridge, MA. 011938
Lipsett MJ; Tsai FC; Roger L; Woo M; Ostro BD. (2006). Coarse particles and heart rate variability among older adults
with coronary artery disease in the Coachella Valley, California. Environ Health Perspect, 114: 1215-1220. 088753
Lisabeth LD; Escobar JD; Dvonch JT; Sanchez BN; Majersik JJ; Brown DL; Smith MA; Morgenstern LB. (2008). Ambient
air pollution and risk for ischemic stroke and transient ischemic attack. Ann Neurol, 64: 53-59. 155939
Liu J; Ballaney M; Al-Alem U; Quan C; Jin X; Perera F; Chen LC; Miller RL. (2008). Combined Inhaled Diesel Exhaust
Particles and Allergen Exposure Alter Methylation of T Helper Genes and IgE Production In Vivo. Toxicol Sci, 102:
76-81. 156709
Liu L; Poon R; Chen L; Frescura AM; Montuschi P; Ciabattoni G; Wheeler A; Dales R. (2009). Acute effects of air
pollution on pulmonary function, airway inflammation, and oxidative stress in asthmatic children. Environ Health
Perspect, 117: 668-674. 192003
Liu L; Ruddy TD; Dalipaj M; Szyszkowicz M; You H; Poon R; Wheeler A; Dales R. (2007). Influence of personal
exposure to particulate air pollution on cardiovascular physiology and biomarkers of inflammation and oxidative
stress in subjects with diabetes. J Occup Environ Med, 49: 258-265. 156705
Liu SX; Hou FF; Guo ZJ; Nagai R; Zhang WR; Liu ZQ; Zhou ZM; Zhou M; Xie D; Wang GB; Zhang X. (2006).
Advanced oxidation protein products accelerate atherosclerosis through promoting oxidative stress and
inflammation. Arterioscler Thromb Vase Biol, 26: 1156-1162. 192002
Liu X; Meng Z. (2005). Effects of airborne fine particulate matter on antioxidant capacity and lipid peroxidation in
multiple organs of rats. Inhal Toxicol, 17: 467-473. 088650
Ljungman P; Bellander T; Schneider A; Breitner S; Forastiere F; Hampel R; Illig T; Jacquemin B; Katsouyanni K; von Klot
S. (2009). Modification of the interleukin-6 response to air pollution by interleukin-6 and fibrinogen
polymorphisms. Environ Health Perspect, 117: 1373-1379. 191983
Ljungman PLS; Berglind N; Holmgren C; Gadler F; Edvardsson N; Pershagen G; Rosenqvist M; Sjogren B; Bellander T.
(2008). Rapid effects of air pollution on ventricular arrhythmias. Eur Heart J, 29: 2894-2901. 180266
Llorca J; Salas A; Prieto-Salceda D; Chinchon-Bengoechea V; Delgado-Rodriguez M. (2005). Nitrogen dioxide increases
cardiorespiratory admissions in Torrelavega (Spain). J Environ Health, 68: 30-35. 087825
Lokken PR; Wellenius GA; Coull BA; Burger MR; Schlaug G; Suh HH; Mittleman MA. (2009). Air Pollution and Risk of
Stroke: Underestimation of Effect Due to Misclassification of Time of Event Onset. Epidemiology, 20: 137-142.
186774
Low RB; Bielory L; Qureshi AI; Dunn V; Stuhlmiller DF; Dickey DA. (2006). The relation of stroke admissions to recent
weather, airborne allergens, air pollution, seasons, upper respiratory infections, and asthma incidence, September
11, 2001, and day of the week. Stroke, 37: 951-957. 090441
Lucking A; Lundback M; Mills N; Faratian D; Barath S; Pourazar J; Cassee F; Donaldson K; Boon N; Badimon J;
Sandstorm T; Blomberg A; Newby D. (2008). Diesel exhaust inhalation increases thrombus formation in man. Eur
Heart J, 29: 3043-3051. 191993
Luginaah IN; Fung K Y; Gorey KM; Webster G; Wills C. (2005). Association of ambient air pollution with respiratory
hospitalization in a government designated "area of concern": the case of Windsor, Ontario. Environ Health
Perspect, 113: 290-296. 057327
Lund AK; Lucero J; Lucas S; Madden MC; McDonald JD; Seagrave JC; Knuckles TL; Campen MJ. (2009). Vehicular
emissions induce vascular MMP-9 expression and activity associated with endothelin-1 mediated pathways.
Arterioscler Thromb Vase Biol, 29: 511-517. 180257
Lundback M; Mills NL; Lucking A; Barath S; Donaldson K; Newby DE; Sandstrom T; Blomberg A. (2009). Experimental
exposure to diesel exhaust increases arterial stiffness in man. Part Fibre Toxicol, 6: 7. 191967
Luttmann-Gibson H; Suh HH; Coull BA; Dockery DW; Sarnet SE; Schwartz J; Stone PH; Gold DR. (2006). Short-term
effects of air pollution on heart rate variability in senior adults in Steubenville, Ohio. J Occup Environ Med, 48:
780-788. 089794
Maciejczyk P; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice: VIII
source-related daily variations in in vitro responses to CAPs. Inhal Toxicol, 17: 243-253. 087456
December 2009 6-231
-------
Mar TF; Ito K; Koenig JQ; Larson TV; Eatough DJ; Henry RC; Kim E; Laden F; Lall R; Neas L; Stolzel M; Paatero P;
Hopke PK; Thurston GD. (2006). PM source apportionment and health effects: 3 Investigation of inter-method
variations in associations between estimated source contributions of PM25 and daily mortality in Phoenix, AZ. J
Expo Sci Environ Epidemiol, 16: 311-320. 086143
Mar TF; Jansen K; Shepherd K; Lumley T; Larson TV; Koenig JQ. (2005). Exhaled nitric oxide in children with asthma
and short-term PM25 exposure in Seattle. Environ Health Perspect, 113: 1791-1794. 088759
Mar TF; Koenig JQ; Jansen K; Sullivan J; Kaufman J; Trenga CA; Siahpush SH; Liu L-JS; Neas L. (2005). Fine particulate
air pollution and cardiorespiratory effects in the elderly. Epidemiology, 16: 681-687. 087566
Mar TF; Larson TV; Stier RA; Claiborn C; Koenig JQ. (2004). An analysis of the association between respiratory
symptoms in subjects with asthma and daily air pollution in Spokane, Washington. Inhal Toxicol, 16: 809-815.
057309
Mar TF; Norris GA; Koenig JQ; Larson TV. (2000). Associations between air pollution and mortality in Phoenix, 1995-
1997. Environ Health Perspect, 108: 347-353. 001760
Mar TF; Norris GA; Larson TV; Wilson WE; Koenig JQ. (2003). Air pollution and cardiovascular mortality in Phoenix,
1995-1997. Health Effects Institute. Cambridge, MA. 042841
Martins LC; Latorre MRDO; Saldiva PHN; Braga ALF. (2002). Air pollution and emergency room visits due to chronic
lower respiratory diseases in the elderly: An ecological time-series study in Sao Paulo, Brazil. J Occup Environ
Med, 44: 622-627. 035059
Masjedi MR; Jamaati HR; Dokouhaki P; Ahmadzadeh Z; Taheri SA; Bigdeli M; Izadi S; Rostamian A; Aagin K; Ghavam
SM. (2003). The effects of air pollution on acute respiratory conditions. Respirology, 8: 213-230. 052100
Matsumoto A; Hiramatsu K; Li Y; Azuma A; Kudoh S; Takizawa H; Sugawara I. (2006). Repeated exposure to low-dose
diesel exhaust after allergen challenge exaggerates asthmatic responses in mice. Clin Immunol, 121: 227-235.
098017
McCreanor J; Cullinan P; Nieuwenhuijsen MJ; Stewart-Evans J; Malliarou E; Jarup L; Harrington R; Svartengren M; Han
I-K; Ohman-Strickland P; Chung KF; Zhang J. (2007). Respiratory effects of exposure to diesel traffic in persons
with asthma. N Engl J Med, 357: 2348-2358. 092841
McDonald J; Barr E; White R; Kracko D; Chow J; Zielinska B; Grosjean E. (2008). Generation and characterization of
gasoline engine exhaust inhalation exposure atmospheres. Inhal Toxicol, 20: 1157-1168. 191978
McDonald JD; Barr EB; White RK; Chow JC; Schauer JJ; Zielinska B; Grosjean E. (2004). Generation and
characterization of four dilutions of diesel engine exhaust for a subchronic inhalation study. Environ Sci Technol,
38: 2513-2522. 055644
McQueen DS; Donaldson K; Bond SM; McNeilly JD; Newman S; Barton NJ; Duffin R. (2007). Bilateral vagotomy or
atropine pre-treatment reduces experimental diesel-soot induced lung inflammation. Toxicol Appl Pharmacol, 219:
62-71.096266
Medina-Ramon M; Zanobetti A; Schwartz J. (2006). The effect of ozone and PM10 on hospital admissions for pneumonia
and chronic obstructive pulmonary disease: a national multicity study. Am J Epidemiol, 163: 579-588. 087721
Menache MG; Miller FJ; Raabe OG (1995). Particle inhalability curves for humans and small laboratory animals. Ann
Occup Hyg, 39: 317-328. 006533
Metzger KB; Klein M; Flanders WD; Peel JL; Mulholland JA; Langberg JJ; Tolbert PE. (2007). Ambient air pollution and
cardiac arrhythmias in patients with implantable defibrillators. Epidemiology, 18: 585-592. 092856
Metzger KB; Tolbert PE; Klein M; Peel JL; Flanders WD; Todd KH; Mulholland JA; Ryan PB; Frumkin H. (2004).
Ambient air pollution and cardiovascular emergency department visits. Epidemiology, 15: 46-56. 044222
Michaud JP; Grove JS; Krupitsky DCEH. (2004). Emergency department visits and "vog"-related air quality inHilo,
Hawai'i. Environ Res, 95: 11-9. 188530
Middleton N; Yiallouros P; Kleanthous S; Kolokotroni O; Schwartz J; Dockery DW; Demokritou P; Koutrakis P. (2008). A
10-year time-series analysis of respiratory and cardiovascular morbidity in Nicosia, Cyprus: the effect of short-term
changes in air pollution and dust storms. Environ Health, 7:39. 156760
December 2009 6-232
-------
Migliaretti G; Cadum E; Migliore E; Cavallo F. (2005). Traffic air pollution and hospital admission for asthma: a case-
control approach in a Turin (Italy) population. Int Arch Occup Environ Health, 78: 164-169. 088689
Migliaretti G; Cavallo F. (2004). Urban air pollution and asthma in children. Pediatr Pulmonol, 38: 198-203. 087425
Mills NL; Robinson SD; Fokkens PH; Leseman DL; Miller MR; Anderson D; Freney EJ; Heal MR; Donovan RJ;
Blomberg A; Sandstrom T; MacNee W; Boon NA; Donaldson K; Newby DE; Cassee FR. (2008). Exposure to
concentrated ambient particles does not affect vascular function in patients with coronary heart disease. Environ
Health Perspect, 116: 709-715. 156766
Mills NL; Tornqvist H; Gonzalez MC; Vink E; Robinson SD; Soderberg S; Boon NA; Donaldson K; Sandstrom T;
Blomberg A; Newby DE. (2007). Ischemic and thrombotic effects of dilute diesel-exhaust inhalation in men with
coronary heart disease. N Engl J Med, 357: 1075-1082. 091206
Mills NL; Tornqvist H; Robinson SD; Gonzalez M; Darnley K; MacNee W; Boon NA; Donaldson K; Blomberg A;
Sandstrom T; Newby DE. (2005). Diesel exhaust inhalation causes vascular dysfunction and impaired endogenous
fibrinolysis. Circulation, 112: 3930-3936. 095757
Min KB; Min JY; Cho SI; Paek D. (2008). The relationship between air pollutants and heart-rate variability among
community residents in Korea. Inhal Toxicol, 4: 435-444. 191901
Mohr LB; Luo S; Mathias E; Tobing R; Homan S; Sterling D. (2008). Influence of season and temperature on the
relationship of elementalcarbon air pollution to pediatric asthma emergency room visits. J Asthma, 45: 936-943.
180215
Moolgavkar SH. (2003). Air pollution and daily deaths and hospital admissions in Los Angeles and Cook counties. Health
Effects Institute. Boston, MA. 042864
Moolgavkar SH. (2003). Air pollution and daily mortality in two US counties: season-specific analyses and exposure-
response relationships. Inhal Toxicol, 15: 877-907. 051316
Moore KJ; Kunjathoor VV; Koehn SL; Manning JJ; Tseng AA; Silver JM; McKee M; Freeman MW. (2005). Loss of
receptor-mediated lipid uptake via scavenger receptor A or CD36 pathways does not ameliorate atherosclerosis in
hyperlipidemic mice. J Clin Invest, 115: 2192-2201. 156780
Morishita M; Keeler G; Wagner J; Marsik F; Timm E; Dvonch J; Harkema J. (2004). Pulmonary retention of particulate
matter is associated with airway inflammation in allergic rats exposed to air pollution in urban Detroit. Inhal
Toxicol, 16: 663-674. 087979
Morris RD; Naumova EN. (1998). Carbon monoxide and hospital admissions for congestive heart failure: evidence of an
increased effect at low temperatures. Environ Health Perspect, 106: 649-653. 086857
Morrison D; Skwarski D; Millar AM; Adams W; MacNee W. (1998). A comparison of three methods of measuring 99mTc-
DTPA lung clearance and their repeatability. Eur Respir J, 11: 1141-1146. 024924
Mortimer KM; Neas LM; Dockery DW; Redline S; Tager IB. (2002). The effect of air pollution on inner-city children with
asthma. Eur Respir J, 19: 699-705. 030281
Moshammer H; Hutter H-P; Hauck H; Neuberger M. (2006). Low levels of air pollution induce changes of lung function in
a panel of schoolchildren. Eur Respir J, 27: 1138-1143. 090771
Moshammer H; Neuberger M. (2003). The active surface of suspended particles as a predictor of lung function and
pulmonary symptoms in Austrian school children. Atmos Environ, 37: 1737-1744. 041956
Mudway IS; Stenfors N; Duggan ST; Roxborough H; Zielinski H; Marklund SL; Blomberg A; Frew AJ; Sandstrom T;
Kelly F J. (2004). An in vitro and in vivo investigation of the effects of diesel exhaust on human airway lining fluid
antioxidants. ArchBiochem Biophys, 423: 200-212. 180208
Muggenburg BA; Barr EB; Cheng YS; Seagrave JC; Tilley LP; Mauderley JL. (2000). Effect of inhaled residual oil fly ash
on the electrocardiogram of dogs. Inhal Toxicol, 12 Suppl 4: 189-208. 189163
Murata A; Kida K; Hasunuma H; Kanegae H; Ishimaru Y; Motegi T; Yamada K; Yoshioka H; Yamamoto K; Kudoh S.
(2007). Environmental influence on the measurement of exhaled nitric oxide concentration in school children:
special reference to methodology. J Nippon Med Sch, 74: 30-6. 189159
December 2009 6-233
-------
Mutlu GM; Green D; Bellmeyer A; Baker CM; Burgess Z; Rajamannan N; Christman JW; Foiles N; Kamp DW; Ghio AJ;
Chandel NS; Dean DA; Sznajder JI; Budinger GR. (2007). Ambient participate matter accelerates coagulation via
an IL-6-dependent pathway. J Clin Invest, 117: 2952-2961. 121441
Nadziejko C; Fang K; Chen LC; Cohen B; Karpatkin M; Nadas A. (2002). Effect of concentrated ambient particulate
matter on blood coagulation parameters in rats. 050587
Nadziejko C; Fang K; Nadziejko E; Narciso SP; Zhong M; Chen LC. (2002). Immediate effects of particulate air pollutants
on heart rate and respiratory rate in hypertensive rats. Cardiovasc Toxicol, 2: 245-252. 087460
Nadziejko C; Fang K; Narciso S; Zhong M; Su WC; Gordon T; Nadas A; Chen LC. (2004). Effect of particulate and
gaseous pollutants on spontaneous arrhythmias in aged rats. Inhal Toxicol, 16: 373-380. 055632
Nascimento LF; Pereira LA; Braga AL; Modolo MC; Carvalho JA Jr. (2006). Effects of air pollution on children's health in
a city in southeastern Brazil. Rev Saude Publica, 40: 77-82. 093247
NCHS. (2007). Classification of Diseases, Functioning, and Disability...Monitoring the Nation's Health. Retrieved 10-
DEC-09, from http://www.cdc.gov/nchs/icd9.htm. 157194
Neuberger M; Schimek MG; Horak F Jr; Moshammer H; Kundi M; Frischer T; Gomiscek B; Puxbaum H; Hauck H;
AUPHEP-Team. (2004). Acute effects of particulate matter on respiratory diseases, symptoms and functions:
epidemiological results of the Austrian Projects on Health Effects of Particulate Matter (AUPHEP). Atmos Environ,
38:3971-3981.093249
Nightingale JA; Maggs R; Cullinan P; Donnelly LE; Rogers DF; Kinnersley R; Chung KF; Barnes PJ; Ashmore M;
Newman-Taylor A. (2000). Airway inflammation after controlled exposure to diesel exhaust particulates. Am J
Respir Grit Care Med, 162: 161-166. 011659
Nikolov MC; Coull BA; Catalano PJ; Diaz E; Godleski JJ. (2008). Statistical methods to evaluate health effects associated
with major sources of air pollution: a case-study of breathing patterns during exposure to concentrated Boston air
particles. J Roy Stat Soc C Appl Stat, 57: 357-378. 156808
Nordenhall C; Pourazar J; Ledin M-C; Levin J-O; Sandstrom T; Adelroth E. (2001). Diesel exhaust enhances airway
responsiveness in asthmatic subjects. Eur Respir J, 17: 909-915. 025185
Nurkiewicz TR; Porter DW; Barger M; Castranova V; Boegehold MA. (2004). Particulate matter exposure impairs
systemic microvascular endothelium-dependent dilation. Environ Health Perspect, 112: 1299-1306. 087968
Nurkiewicz TR; Porter DW; Barger M; Millecchia L; Rao KM; Marvar PJ; Hubbs AF; Castranova V; Boegehold MA.
(2006). Systemic microvascular dysfunction and inflammation after pulmonary particulate matter exposure.
Environ Health Perspect, 114: 412-419. 088611
Nurkiewicz TR; Porter DW; Hubbs AF; Cumpston JL; Chen BT; Frazer DG; Castranova V. (2008). Nanoparticle inhalation
augments particle-dependent systemic microvascular dysfunction. Part Fibre Toxicol, 5:1. 156816
Nurkiewicz TR; Porter DW; Hubbs AF; Stone S; Chen BT; Frazer DG; Boegehold MA; Castranova V. (2009). Pulmonary
nanoparticle exposure disrupts systemic microvascular nitric oxide signaling. Toxicol Sci, 110: 191-203. 191961
Nygaard UC; Samuelsen M; Aase A; Lovik M. (2004). The capacity of particles to increase allergic sensitization is
predicted by particle number and surface area, not by particle mass. Toxicol Sci, 82: 515-524. 058558
O'Connor GT; Neas L; Vaughn B; Kattan M; Mitchell H; Grain EF; Evans R 3rd; Gruchalla R; Morgan W; Stout J; Adams
GK; Lippmann M. (2008). Acute respiratory health effects of air pollution on children with asthma in US inner
cities. JAllergy Clinlmmunol, 121: 1133-1139. 156818
O'Neill MS; Veves A; Sarnat JA; Zanobetti A; Gold DR; Economides PA; Horton ES; Schwartz J. (2007). Air pollution and
inflammation in type 2 diabetes: a mechanism for susceptibility. Occup Environ Med, 64: 373-379. 091362
O'Neill MS; Veves A; Zanobetti A; Sarnat JA; Gold DR; Economides PA; Horton ES; Schwartz J. (2005). Diabetes
enhances vulnerability to particulate air pollution-associated impairment in vascular reactivity and endothelial
function. Circulation, 111: 2913-2920. 088423
Odajima H; Yamazaki S; Nitta H. (2008). Decline in peak expiratory flow according to hourly short-term concentration of
particulate matter in asthmatic children. Inhal Toxicol, 20: 1263-1272. 192005
Oftedal B; Nafstad P; Magnus P; Bjorkly S; Skrondal A. (2003). Traffic related air pollution and acute hospital admission
for respiratory diseases in Drammen, Norway 1995-2000. Eur J Epidemiol, 18: 671-675. 055623
December 2009 6-234
-------
Okin PM; Devereux RB; Howard BV; Fabsitz RR; Lee ET; Welty TK. (2000). Assessment of QT interval and QT
dispersion for prediction of all-cause and cardiovascular mortality in American Indians: The Strong Heart Study.
Circulation, 101: 61-66. 156002
Ostro B. (1995). Fine particulate air pollution and mortality in two Southern California counties. Environ Res, 70: 98-104.
079197
Ostro B; Broadwin R; Green S; Feng W-Y; Lipsett M. (2006). Fine particulate air pollution and mortality in nine California
counties: results from CALFINE. Environ Health Perspect, 114: 29-33. 087991
Ostro B; Feng W-Y; Broadwin R; Green S; Lipsett M. (2007). The effects of components of fine particulate air pollution on
mortality in California: results from CALFINE. Environ Health Perspect, 115:13-19. 091354
Ostro B; Roth L; Malig B; Marty M. (2009). The effects of fine particle components on respiratory hospital admissions in
children. Environ Health Perspect, 117: 475-480. 191971
Ostro BD; Broadwin R; Lipsett MJ. (2003). Coarse particles and daily mortality in Coachella Valley, California. Health
Effects Institute. Boston, MA. 042824
Ostro BD; Feng WY; Broadwin R; Malig BJ; Green RS; Lipsett MJ. (2008). The impact of components of fine particulate
matter on cardiovascular mortality in susceptible subpopulations. Occup Environ Med, 65: 750-756. 097971
Park SK; O'Neill MS; Vokonas PS; Sparrow D; Schwartz J. (2005). Effects of air pollution on heart rate variability: The VA
normative aging study. Environ Health Perspect, 113: 304-309. 057331
Park SK; O'Neill MS; Vokonas PS; Sparrow D; Spiro A 3rd; Tucker KL; Suh H; Hu H; Schwartz J. (2008). Traffic-related
particles are associated with elevated homocysteine: the VA normative aging study. Am J Respir Crit Care Med,
178: 283-289. 156845
Park SK; O'Neill MS; Vokonas PS; Sparrow D; Wright RO; Coull B; Nie H; Hu H; Schwartz J. (2008). Air pollution and
heart rate variability: effect modification by chronic lead exposure. Epidemiology, 19: 111-120. 093027
Park SK; O'Neill MS; Wright RO; Hu H; Vokonas PS; Sparrow D; Suh H; Schwartz J. (2006). HFE genotype, particulate
air pollution, and heart rate variability; a gene-environmental interaction. Circulation, 114: 2798-2805. 091245
Peacock JL; Symonds P; Jackson P; Bremner SA; Scarlett JF; Strachan DP; Anderson HR. (2003). Acute effects of winter
air pollution on respiratory function in schoolchildren in southern England. Occup Environ Med, 60: 82-89. 042026
Peel JL; Metzger KB; Klein M; Flanders WD; Mulholland JA; Tolbert PE. (2007). Ambient air pollution and
cardiovascular emergency department visits in potentially sensitive groups. Am J Epidemiol, 165: 625-633. 090442
Peel JL; Tolbert PE; Klein M; Metzger KB; Flanders WD; Knox T; Mulholland JA; Ryan PB; Frumkin H. (2005). Ambient
air pollution and respiratory emergency department visits. Epidemiology, 16: 164-174. 056305
Pekkanen J; Peters A; Hoek G; Tiittanen P; Brunekreef B; de Hartog J; Heinrich J; Ibald-Mulli A; Kreyling WG; Lanki T;
Timonen KL; Vanninen E. (2002). Particulate air pollution and risk of ST-segment depression during repeated
submaximal exercise tests among subjects with coronary heart disease: the exposure and risk assessment for fine
and ultrafine particles in ambient air (ULTRA) study. Circulation, 106: 933-938. 035050
Peled R; Friger M; Bolotin A; Bibi H; Epstein L; Pilpel D; Scharf S. (2005). Fine particles and meteorological conditions
are associated with lung function in children with asthma living near two power plants. Public Health, 119: 418-
425. 156015
Peng R; Bell M; Geyh A; McDermott A; Zeger S; Samet J; Dominici F. (2009). Emergency admissions for cardiovascular
and respiratory diseases and the chemical composition of fine particle air pollution. Environ Health Perspect, 117:
957-963. 191998
Peng RD; Chang HH; Bell ML; McDermott A; Zeger SL; Samet JM; Dominici F. (2008). Coarse particulate matter air
pollution and hospital admissions for cardiovascular and respiratory diseases among Medicare patients. JAMA,
299: 2172-2179. 156850
Peng RD; Dominici F; Pastor-Barriuso R; Zeger SL; Samet JM. (2005). Seasonal analyses of air pollution and mortality in
100 US cities. Am J Epidemiol, 161: 585-594. 087463
December 2009 6-235
-------
Pennanen AS; Sillanpaa M; Hillamo R; Quass U; John AC; Branis M; Hunova I; Meliefste K; Janssen NA; Koskentalo T;
Castano-Vinyals G; Bouso L; Chalbot MC; Kavouras IG; Salonen RO. (2007). Performance of a high-volume
cascade impactor in six European urban environments: mass measurement and chemical characterization of size-
segregated particulate samples. Sci Total Environ, 374: 297-310. 155357
Penttinen P; Vallius M; Tiittanen P; Ruuskanen J; Pekkanen J. (2006). Source-specific fine particles in urban air and
respiratory function among adult asthmatics. Inhal Toxicol, 18: 191-198. 087988
Pereira CEL; Heck TG; Saldiva PHN; Rhoden CR. (2007). Ambient particulate air pollution from vehicles promotes lipid
peroxidation and inflammatory responses in rat lung. Braz JMed Biol Res, 40: 1353-1359. 156019
Peretz A; Kaufman JD; Trenga CA; Allen J; Carlsten C; Aulet MR; Adar SD; Sullivan JH. (2008). Effects of diesel exhaust
inhalation on heart rate variability in human volunteers. Environ Res, 107: 178-184. 156855
Peretz A; Peck EC; Bammler TK; Beyer RP; Sullivan JH; Trenga CA; Srinouanprachnah S; Farin FM; Kaufman JD.
(2007). Diesel exhaust inhalation and assessment of peripheral blood mononuclear cell gene transcription effects:
an exploratory study of healthy human volunteers. Inhal Toxicol, 19: 1107-1119. 156853
Peretz A; Sullivan JH; Leotta DF; Trenga CA; Sands FN; Allen J; Carlsten C; Wilkinson CW; Gill EA; Kaufman JD.
(2008). Diesel exhaust inhalation elicits acute vasoconstriction in vivo. Environ Health Perspect, 116: 937-942.
156854
Perez L; Tobias A; Querol X; Kunzli N; Pey J; Alastuey A; Viana M; Valero N; Gonzalez-Cabre M; Sunyer J. (2008).
Coarse particles from Saharan dust and daily mortality. Epidemiology, 19: 800-807. 156020
Peters A. (2005). Particulate matter and heart disease: evidence from epidemiological studies. Toxicol Appl Pharmacol,
207: S477-S482. 087759
Peters A; Dockery DW; Muller JE; Mittleman MA. (2001). Increased particulate air pollution and the triggering of
myocardial infarction. Circulation, 103: 2810-2815. 016546
Peters A; Greven S; Heid I; Baldari F; Breitner S; Bellander T; Chrysohoou C; Illig T; Jacquemin B; Koenig W. (2009).
Fibrinogen genes modify the fibrinogen response to ambient particulate matter. Am J Respir Crit Care Med, 179:
484-491. 191992
Peters A; Liu E; Verrier RL; Schwartz J; Gold DR; Mittleman M; Baliff J; Oh JA; Allen G; Monahan K; Dockery DW.
(2000). Air pollution and incidence of cardiac arrhythmia. Epidemiology, 11: 11-17. 011347
Peters A; Von Klot S; Heier M; Trentinaglia I; Hermann A; Wichmann HE; Lowel H. (2004). Exposure to traffic and the
onset of myocardial infarction. N Engl J Med, 351: 1721-1730. 087464
Petrovic S; Urch B; Brook J; Datema J; Purdham J; Liu L; Lukic Z; Zimmerman B; Toiler G; Downar E; Corey P; Tarlo S;
Broder I; Dales R; Silverman F. (2000). Cardiorespiratory effects of concentrated ambient PM25: A pilot study
using controlled human exposures. Inhal Toxicol, 1: 173-188. 004638
Piedrahita JA; Zhang SH; Hagaman JR; Oliver PM; Maeda N. (1992). Generation of mice carrying a mutant apolipoprotein
E gene inactivated by gene targeting in embryonic stem cells. Presented at Proceedings of the National Academy of
Science, 5/15/1992. 156868
Pietropaoli AP; Frampton MW; Hyde RW; Morrow PE; Oberdorster G; Cox C; Speers DM; Frasier LM; Chalupa DC;
Huang LS; Utell MJ. (2004). Pulmonary function, diffusing capacity, and inflammation in healthy and asthmatic
subjects exposed to ultrafine particles. Inhal Toxicol, 16: 59-72. 156025
Pinkerton KE; Zhou Y; Teague SV; Peake JL; Walther RC; Kennedy IM; Leppert VJ; Aust AE. (2004). Reduced lung cell
proliferation following short-term exposure to ultrafine soot and iron particles in neonatal rats: key to impaired lung
growth?. Inhal Toxicol, 1: 73-81. 087465
Pinkerton KE; Zhou Y; Zhong C; Smith KR; Teague SV; Kennedy IM; Menache MG (2008). Mechanisms of particulate
matter toxicity in neonatal and young adult rat lungs. Health Effects Institute. Boston, MA. 135. 190471
Pope C; Renlund D; Kfoury A; May H; Home B. (2008). Relation of heart failure hospitalization to exposure to fine
particulate air pollution. Am J Cardiol, 102: 1230-1234. 191969
Pope CA; Hansen ML; Long RW; Nielsen KR; Eatough NL; Wilson WE; Eatough DJ. (2004). Ambient particulate air
pollution, heart rate variability, and blood markers of inflammation in a panel of elderly subjects. Environ Health
Perspect, 112: 339-345. 055238
December 2009 6-236
-------
Pope CAIII; Muhlestein IB; May HT; Renlund DG; Anderson JL; Home BD. (2006). Ischemic heart disease events
triggered by short-term exposure to fine particulate air pollution. Circulation, 114: 2443-2448. 091246
Pourazar J; Blomberg A; Kelly FJ; Davies DE; Wilson SJ; Holgate ST; Sandstrom T. (2008). Diesel exhaust increases
EGFR and phosphorylated C-terminal Tyr 1173 in the bronchial epithelium. Part Fibre Toxicol, 5: 8. 156884
Pourazar J; Mudway IS; Samet JM; Helleday R; Blomberg A; Wilson SJ; Frew AJ; Kelly FJ; Sandstrom T. (2005). Diesel
exhaust activates redox-sensitive transcription factors and kinases in human airways. Am J Physiol, 289: L724-
L730. 088305
Power K; Balmes J; Solomon C. (2008). Controlled exposure to combined particles and ozone decreases heart rate
variability. J Occup Environ Med, 50: 1253-1260. 191982
Preutthipan A; Udomsubpayakul U; Chaisupamongkollarp T; Pentamwa P. (2004). Effect of PM10 pollution in Bangkok on
children with and without asthma. Pediatr Pulmonol, 37: 187-192. 055598
Pritchett LC; Cooper JA. (1985). Aerosol Characterization Study of Anchorage, Alaska: Chemical Analysis and Source
Apportionment. 156886
Prystowsky EN; Benson DW Jr; Fuster V; Hart RG; Kay GN; Myerburg RJ; Naccarelli GV; Wyse DG. (1996).
Management of patients with atrial fibrillation. A Statement for Healthcare Professionals. From the Subcommittee
on Electrocardiography and Electrophysiology, American Heart Association. Circulation, 93: 1262-1277. 156031
Raabe OG; Al-Bayati MA; Teague SV; Rasolt A. (1988). Regional deposition of inhaled monodisperse, coarse, and fine
aerosol particles in small laboratory animals. In Inhaled particles VI: Proceedings of an international symposium
and workshop on lung dosimetry (pp. 53-63). Cambridge, U.K.: PergamonPress. 001439
Rabinovitch N; Strand M; Gelfand EW (2006). Particulate levels are associated with early asthma worsening in children
with persistent disease. Am J Respir Grit Care Med, 173: 1098-1105. 088031
Rabinovitch N; Zhang LN; Murphy JR; Vedal S; Dutton SJ; Gelfand EW. (2004). Effects of wintertime ambient air
pollutants on asthma exacerbations in urban minority children with moderate to severe disease. J Allergy Clin
Immunol, 114: 1131-1137.096753
Ranzi A; Gambini M; Spattini A; Galassi C; Sesti D; Bedeschi M; Messori A; Baroni A; Cavagni G; Lauriola P. (2004). Air
pollution and respiratory status in asthmatic children: Hints for a locally based preventive strategy AIRE study. Eur
JEpidemiol, 19: 567-576. 089500
Reed MD; Barrett EG; Campen MJ; Divine KK; Gigliotti AP; McDonald JD; Seagrave JC; Mauderly JL; Seilkop SK;
Swenberg JA. (2008). Health effects of subchronic inhalation exposure to gasoline engine exhaust. Inhal Toxicol,
20: 1125-1143. 156903
Reed MD; Campen MJ; Gigliotti AP; Harrod KS; McDonald JD; Seagrave JC; Mauderly JL; Seilkop SK. (2006). Health
effects of subchronic exposure to environmental levels of hardwood smoke. Inhal Toxicol, 18: 523-539. 156043
Reed MD; Gigliotti AP; McDonald JD; Seagrave JC; Seilkop SK; Mauderly JL. (2004). Health effects of subchronic
exposure to environmental levels of diesel exhaust. Inhal Toxicol, 16: 177-193. 055625
Rhoden CR; Ghelfi E; Gonzalez-Flecha B. (2008). Pulmonary inflammation by ambient air particles is mediated by
superoxide anion. Inhal Toxicol, 20: 11-15. 190475
Rhoden CR; Lawrence J; Godleski JJ; Gonzalez-Flecha B. (2004). N-acetylcysteine prevents lung inflammation after
short-term inhalation exposure to concentrated ambient particles. Toxicol Sci, 79: 296-303. 087969
Rhoden CR; Wellenius GA; Ghelfi E; Lawrence J; Gonzalez-Flecha B. (2005). PM-induced cardiac oxidative stress and
dysfunction are mediated by autonomic stimulation. Biochim Biophys Acta, 1725: 305-313. 087878
Rich DQ; Freudenberger RS; Ohman-Strickland P; Cho Y; Kipen HM. (2008). Right heart pressure increases after acute
increases in ambient particulate concentration. Environ Health Perspect, 116: 1167-1171. 156910
Rich DQ; Kim MH; Turner JR; Mittleman MA; Schwartz J; Catalano PJ; Dockery DW. (2006). Association of ventricular
arrhythmias detected by implantable cardioverter defibrillator and ambient air pollutants in the St Louis, Missouri
metropolitan area. Occup Environ Med, 63: 591-596. 089814
Rich DQ; Mittleman MA; Link MS; Schwartz J; Luttmann-Gibson H; Catalano PJ; Speizer FE; Gold DR; Dockery DW.
(2006). Increased risk of paroxysmal atrial fibrillation episodes associated with acute increases in ambient air
pollution. Environ Health Perspect, 114: 120-123. 088427
December 2009 6-237
-------
Rich DQ; Schwartz J; Mittleman MA; Link M; Luttmann-Gibson H; Catalano PJ; Speizer FE; Dockery DW. (2005).
Association of short-term ambient air pollution concentrations and ventricular arrhythmias. Am J Epidemiol, 161:
1123-1132.079620
Rich KE; Petkau J; Vedal S; Brauer M. (2004). A case-crossover analysis of particulate air pollution and cardiac arrhythmia
in patients with implantable cardioverter defibrillators. Inhal Toxicol, 16: 363-372. 055631
Riediker M; Cascio WE; Griggs TR; Herbst MC; Bromberg PA; Neas L; Williams RW; Devlin RB. (2004). Particulate
matter exposure in cars is associated with cardiovascular effects in healthy young men. Am J Respir Crit Care Med,
169: 934-940. 056992
Riediker M; Devlin RB; Griggs TR; Herbst MC; Bromberg PA; Williams RW; Cascio WE. (2004). Cardiovascular effects
in patrol officers are associated with fine particulate matter from brake wear and engine emissions. Part Fibre
Toxicol, 1: 2. 091261
Riojas-Rodriguez H; Escamilla-Cejudo JA; Gonzalez-Hermosillo JA; Tellez-Rojo MM; Vallejo M; Santos-Burgoa C;
Rojas-Bracho L. (2006). Personal PM2.5 and CO exposures and heart rate variability in subjects with known
ischemic heart disease in Mexico City. JExpo Sci Environ Epidemiol, 16: 131-137. 156913
Roberts JR; Taylor MD; Castranova V; Clarke RW; Antonini JM. (2004). Soluble metals associated with residual oil fly ash
increase morbidity and lung injury after bacterial infection in rats. J Toxicol Environ Health A, 67: 251-263. 196994
Roberts JR; Young S-H; Castranova V; Antonini JM. (2007). Soluble metals in residual oil fly ash alter innate and adaptive
pulmonary immune responses to bacterial infection in rats. Toxicol Appl Pharmacol, 221: 306-319. 097623
Rodriguez C; Tonkin R; Heyworth J; Kusel M; De Klerk N; Sly PD; Franklin P; Runnion T; Blockley A; Landau L;
Hinwood AL. (2007). The relationship between outdoor air quality and respiratory symptoms in young children. Int
J Environ Health Res, 17: 351-360 . 092842
Romieu I; Garcia-Esteban R; Sunyer J; Rios C; Alcaraz-Zubeldia M; Velasco SR; Holguin F. (2008). The effect of
supplementation with omega-3 polyunsaturated fatty acids on markers of oxidative stress in elderly exposed to
PM(2.5). Environ Health Perspect, 116: 1237-1242. 156922
Romieu I; Tellez-Rojo MM; Lazo M; Manzano-Patino Cortez-Lugo M; Julien P; Belanger MC; Hernandez-Avila
MHolguin F. (2005). Omega-3 fatty acid prevents heart rate variability reductions associated with particulate
matter. Am J Respir Crit Care Med, 172: 1534-1540. 086297
Rosenthal FS; Carney JP; Olinger ML. (2008). Out-of-hospital cardiac arrest and airborne fine particulate matter: a case-
crossover analysis of emergency medical services data in Indianapolis, Indiana. Environ Health Perspect, 116: 631 -
636. 156925
Routledge HC; Manney S; Harrison RM; Ayres JG; Townend JN. (2006). Effect of inhaled sulphur dioxide and carbon
particles on heart rate variability and markers of inflammation and coagulation in human subjects. Heart, 92: 220-
227. 088674
Rowan 3rd WH; Campen MJ; Wichers LB; Watkinson WP. (2007). Heart rate variability in rodents: uses and caveats in
toxicological studies. Cardiovasc Toxicol, 7: 28-51. 191911
Roy D; Talajic M; Dubuc M; Thibault B; Guerra P; Made L; Khairy P. (2009). Atrial fibrillation and congestive heart
failure. Curr Opin Cardiol, 24: 29. 190902
Ruckerl R; Greven S; Ljungman P; Aalto P; Antoniades C; Bellander T; Berglind N; Chrysohoou C; Forastiere F;
JacqueminB; vonKlot S; Koenig W; Kuchenhoff H; Lanki T; Pekkanen J; Perucci CA; Schneider A; Sunyer J;
Peters A. (2007). Air pollution and inflammation (interleukin-6, C-reactive protein, fibrinogen) in myocardial
infarction survivors. Environ Health Perspect, 115: 1072-1080. 156931
Ruckerl R; Ibald-Mulli A; Koenig WSchneider A; Woelke G; Cyrys J; Heinrich J; Marder V; Frampton M; Wichmann HE;
Peters A. (2006). Air pollution and markers of inflammation and coagulation in patients with coronary heart
disease. Environ Health Perspect, 173: 432-441.088754
Rudell B; Blomberg A; Helleday R; Ledin M-C; Lundback B; Stjernberg N; Horstedt P; Sandstrom T. (1999).
Bronchoalveolar inflammation after exposure to diesel exhaust: Comparison between unfiltered and particle trap
filtered exhaust. Occup Environ Med, 56: 527-534. 001964
Rudell B; Ledin M-C; Hammarstrom U; Stjernberg N; Lundback B; Sandstrom T. (1996). Effects on symptoms and lung
function in humans experimentally exposed to diesel exhaust. Occup Environ Med, 53: 658-662. 056577
December 2009 6-238
-------
Rundell KW; Caviston R. (2008). Ultrafine and fine particulate matter inhalation decreases exercise performance in healthy
subjects. J Strength Cond Res, 22: 2-5. 191986
Rundell KW; Hoffman JR; Caviston R; Bulbulian R; Hollenbach AM. (2007). Inhalation of ultrafine and fine particulate
matter disrupts systemic vascular function. Inhal Toxicol, 19: 133-140. 156060
Safar M; Chamiot-Clerc P; Dagher G; Renaud IF. (2001). Pulse pressure, endothelium function, and arterial stiffness in
spontaneously hypertensive rats. Hypertension, 38: 1416-1421. 156068
Saldiva PHN; Clarke RW; Coull BA; Stearns RC; Lawrence J; Krishna-Murthy GG; Diaz E; Koutrakis P; Suh H; Tsuda A;
Godleski JJ. (2002). Lung inflammation induced by concentrated ambient air particles is related to particle
composition. Am J Respir Grit Care Med, 165: 1610-1617. 025988
Salvi S; Blomberg A; Rudell B; Kelly F; Sandstrom T; Holgate ST; Frew A. (1999). Acute inflammatory responses in the
airways and peripheral blood after short-term exposure to diesel exhaust in healthy human volunteers. Am J Respir
Grit Care Med, 159: 702-709. 058637
Samet JM; Dominici F; Zeger SL; Schwartz J; Dockery DW. (2000). National morbidity, mortality, and air pollution study
Part I: methods and methodologic issues. 005809
Samet JM; Rappold A; Graff D; Cascio WE; Berntsen JH; Huang YC; Herbst M; Bassett M; Montilla T; Hazucha MJ;
Bromberg PA; Devlin RB. (2009). Concentrated ambient ultrafine particle exposure induces cardiac changes in
young healthy volunteers. Am J Respir Grit Care Med, 179: 1034-1042. 191913
Samet JM; Zeger SL; Dominici F; Curriero F; Coursac I; Dockery DW; Schwartz J; Zanobetti A. (2000). The national
morbidity, mortality, and air pollution study Part II: morbidity, mortality, and air pollution in the United States.
010269
Samoli E; Analitis A; Touloumi G; Schwartz J; Anderson HR; Sunyer J; Bisanti L; Zmirou D; Vonk JM; Pekkanen J;
Goodman P; Paldy A; Schindler C; Katsouyanni K. (2005). Estimating the exposure-response relationships between
particulate matter and mortality within the APHEAmulticity project. Environ Health Perspect, 113: 88-95. 087436
Samoli E; Peng R; Ramsay T; Pipikou M; Touloumi G; Dominici F; Burnett R; Cohen A; Krewski D; Samet J. (2008).
Acute effects of ambient particulate matter on mortality in Europe and North America: results from the APHENA
study. Environ Health Perspect, 116: 1480-1486. 188455
Sandercock GRH; Brodie DA. (2006). The role of heart rate variability in prognosis for different modes of death in chronic
heart failure. Pacing Clin Electrophysiol, 29: 892-904. 197465
Santos U; Terra-Filho M; Lin C; Pereira L; Vieira T; Saldiva P; Braga A. (2008). Cardiac arrhythmia emergency room
visits and environmental air pollution in Sao Paulo, Brazil. J Epidemiol Community Health, 62: 267-272. 192004
Sarin S; UndemB; Sanico A; Togias A. (2006). The role of the nervous system in rhinitis. J Clin Immunol, 118: 999-1016.
191166
Sarnat JA; Marmur A; Klein M; Kim E; Russell AG; Sarnat SE; Mulholland JA; Hopke PK; Tolbert PE. (2008). Fine
particle sources and cardiorespiratory morbidity: An application of chemical mass balance and factor analytical
source-apportionment methods. Environ Health Perspect, 116: 459-466. 097972
Sarnat SE; Suh HH; Coull BA; Schwartz J; Stone PH; Gold DR. (2006). Ambient particulate air pollution and cardiac
arrhythmia in a panel of older adults in Steubenville, Ohio. Occup Environ Med, 63: 700-706. 090489
Sauerzapf V; Jones AP; Cross J. (2009). Environmental factors and hospitalisation for chronic obstructive pulmonary
disease in a rural county of England. J Epidemiol Community Health, 63: 324-328. 180082
Schaumann F; Borm PJA; Herbrich A; Knoch J; Pitz M; Schins RPF; Luettig B; Hohlfeld JM; Heinrich J; Krug N. (2004).
Metal-rich ambient particles (particulate matter 2.5) cause airway inflammation in healthy subjects. Am J Respir
Grit Care Med, 170: 898-903. 087966
Schildcrout JS; Sheppard L; Lumley T; Slaughter JC; Koenig JQ; Shapiro GG. (2006). Ambient air pollution and asthma
exacerbations in children: An eight-city analysis. Am J Epidemiol, 164: 505-517. 089812
Schins RPF; Lightbody JH; Borm PJA; Shi T; Donaldson K; Stone V. (2004). Inflammatory effects of coarse and fine
particulate matter in relation to chemical and biological constituents. Toxicol Appl Pharmacol, 195: 1-11. 054173
December 2009 6-239
-------
Schneider A; Neas L; Herbst M; Case M; Williams R; Cascio W; Hinderliter A; Holguin F; Buse J; Dungan K. (2008).
Endothelial dysfunction: associations with exposure to ambient fine particles in diabetic individuals. Environ
Health Perspect, 116: 1666-1674. 191985
Schreuder AB; Larson TV; Sheppard L; Claiborn CS. (2006). Ambient woodsmoke and associated respiratory emergency
department visits in Spokane, Washington. Int J Occup Environ Health, 12: 147-153. 097959
Schwartz J. (2003). Airborne particles and daily deaths in 10 US cities. 042800
Schwartz J. (2003). Daily deaths associated with air pollution in six US cities and short-term mortality displacement in
Boston. 042811
Schwartz J. (2004). Is the association of airborne particles with daily deaths confounded by gaseous air pollutants? An
approach to control by matching. Environ Health Perspect, 112: 557-561. 053506
Schwartz J. (2004). The effects of particulate air pollution on daily deaths: a multi-city case crossover analysis. Occup
Environ Med, 61: 956-961. 078998
Schwartz J; Coull B; Laden F; Ryan L. (2008). The effect of dose and timing of dose on the association between airborne
particles and survival. Environ Health Perspect, 116: 64-69. 156963
Schwartz J; Litonjua A; Suh H; Verrier M; Zanobetti A; Syring M; Nearing B; Verrier R; Stone P; MacCallum G; Speizer
FE; Gold DR. (2005). Traffic related pollution and heart rate variability in a panel of elderly subjects. Thorax, 60:
455-461.074317
Schwartz J; Morris R. (1995). Air pollution and hospital admissions for cardiovascular disease in Detroit, Michigan. Am J
Epidemiol, 142: 23-35. 046186
Schwartz J; Neas LM. (2000). Fine particles are more strongly associated than coarse particles with acute respiratory health
effects in schoolchildren. Epidemiology, 11: 6-10. 007625
Schwartz J; Park SK; O'Neill MS; Vokonas PS; SparrowD; Weiss S; Kelsey K. (2005). Glutathione-S-transferase Ml,
obesity, statins, and autonomic effects of particles: gene-by-drug-by-environment interaction. Am J Respir Crit Care
Med, 172: 1529-1533. 086296
Seagrave J; Campen M; McDonald J; Mauderly J; Rohr A. (2008). Oxidative stress, inflammation, and pulmonary function
assessment in rats exposed to laboratory-generated pollutant mixtures. J Toxicol Environ Health A, 71: 1352-1362.
191990
Seagrave JC; McDonald JD; Bedrick E; Edgerton ES; Gigliotti AP; Jansen JJ; Ke L; Naeher LP; Seilkop SK; Zheng M;
Mauderley JL. (2006). Lung toxicity of ambient particulate matter from southeastern US sites with different
contributing sources: relationships between composition and effects. Environ Health Perspect, 114: 1387-93.
091291
Segala C; Poizeau D; Neukirch F; Aubier M; Samson J; Gehanno P. (2004). Air pollution, passive smoking, and respiratory
symptoms in adults. Arch Environ Occup Health, 59: 669-676. 090449
Shah AP; Pietropaoli AP; Frasier LM; Speers DM; Chalupa DC; Delehanty JM; Huang LS; Utell MJ; Frampton MW.
(2008). Effect of inhaled carbon ultrafine particles on reactive hyperemia in healthy human subjects. Environ
Health Perspect, 116: 375-380. 156970
Sheppard L; Levy D; Norris G; Larson TV; Koenig JQ. (2003). Effects of ambient air pollution and nonelderly asthma
hospital admissions in Seattle, Washington, 1987-1994. Epidemiology, 10: 23-30. 042826
Shi JP; Harrison RM; Evans D. (2001). Comparison of ambient particle surface area measurement by epiphaniometer and
SMPS/APS. Atmos Environ, 35: 6193-6200. 078292
Sillanpaa M; Frey A; Hillamo R; Pennanen AS; Salonen RO. (2005). Organic, elemental and inorganic carbon in
particulate matter of six urban environments in Europe. Atmos Chem Phys, 5: 2869-2879. 156980
Simpson R; Williams G; Petroeschevsky A; Best T; Morgan G; Denison L; Hinwood A; Neville G. (2005). The short-term
effects of air pollution on hospital admissions in four Australian cities. AustNZ J Public Health, 29: 213-221.
087438
Sinclair AH; Tolsma D. (2004). Associations and lags between air pollution and acute respiratory visits in an ambulatory
care setting: 25-month results from the aerosol research and inhalation epidemiological study. J Air Waste Manag
Assoc, 54: 1212-1218. 088696
December 2009 6-240
-------
Sirivelu MP; MohanKumar SMJ; Wagner JG; Harkema JR; MohanKumar PS. (2006). Activation of the stress axis and
neurochemical alterations in specific brain areas by concentrated ambient particle exposure with concomitant
allergic airway disease. Environ Health Perspect, 114: 870-874. 111151
Slaughter JC; Kim E; Sheppard L; Sullivan JH; Larson TV; Claiborn C. (2005). Association between particulate matter and
emergency room visits, hospital admissions and mortality in Spokane, Washington. J Expo Sci Environ Epidemiol,
15: 153-159.073854
Slaughter JC; Lumley T; Sheppard L; Koenig JQ; Shapiro GG. (2003). Effects of ambient air pollution on symptom
severity and medication use in children with asthma. Ann Allergy Asthma Immunol, 91: 346-353. 086294
Smith AD; Taylor DR. (2005). Is exhaled nitric oxide measurement a useful clinical test in asthma?. Curr Opin Allergy Clin
Immunol, 5: 49-56. 192176
Smith KR; Kim S; Recendez JJ; Teague SV; Menache MG; Grubbs DE; Sioutas C; Pinkerton KE. (2003). Airborne
particles of the California Central Valley alter the lungs of healthy adult rats. Environ Health Perspect, 111: 902-
908. 042107
Smith KR; Veranth JM; Kodavanti UP; Aust AE; Pinkerton KE. (2006). Acute pulmonary and systemic effects of inhaled
coal fly ash in rats: comparison to ambient environmental particles. Toxicol Sci, 93: 390-399. 110864
Steerenberg P; Withagen C; Dalen W; Dormans J; Loveren H. (2004). Adjuvant Activity of Ambient Particulate Matter in
Macrophage Activity-Suppressed, N-Acetylcysteine-Treated, iNOS- and IL-4-Deficient Mice. Inhal Toxicol, 16:
835-843. 087981
Steerenberg PA; Van Amelsvoort L; Lovik M; Hetland RB; Alberg T; Halatek T; Bloemen HJT; Rydzynski K; Swaen G;
Schwarze P; Dybing E; Cassee FR. (2006). Relation between sources of particulate air pollution and biological
effect parameters in samples from four European cities: an exploratory study. Inhal Toxicol, 18: 333-346. 088249
Steerenberg PA; Withagen CE; van Dalen WJ; Dormans JA; Heisterkamp SH; van Loveren H; Cassee FR. (2005). Dose
dependency of adjuvant activity of particulate matter from five European sites in three seasons in an ovalbumin-
mouse model. Inhal Toxicol, 17: 133-145. 088649
Steerenberg PA; Withagen GET; van Dalen WJ; Dormans JAMA; Cassee FR; Heisterkamp SH; van Loveren H. (2004).
Adjuvant activity of ambient particulate matter of different sites, sizes, and seasons in a respiratory allergy mouse
model. Toxicol Appl Pharmacol, 200: 186-200. 096024
Steinvil A; Kordova-Biezuner L; Shapira I; Berliner S; Rogowski O. (2008). Short-term exposure to air pollution and
inflammation-sensitive biomarkers. Environ Res, 106: 51-61. 188893
Stenfors N; Nordenhall C; Salvi SS; Mudway I; Soderberg M; Blomberg A; Helleday R; Levin JO; Holgate ST; Kelly FJ;
Frew AJ; Sandstrom T. (2004). Different airway inflammatory responses in asthmatic and healthy humans exposed
to diesel. EurRespir J, 23: 82-86. 157009
Stevens T; Krantz QT; Linak WP; Hester S; Gilmour MI. (2008). Increased transcription of immune and metabolic
pathways in naive and allergic mice exposed to diesel exhaust. Toxicol Sci, 102: 359-370. 157010
Stolzel M; Breitner S; Cyrys J; Pitz M; Wolke G; Kreyling W; Heinrich J; Wichmann H-E; Peters A. (2007). Daily
mortality and particulate matter in different size classes in Erfurt, Germany. J Expo Sci Environ Epidemiol, 17:
458-467. 091374
Stolzel M; Peters A; Wichmann H-E. (2003). Daily mortality and fine and ultrafine particles in Erfurt, Germany. 042842
Strand M; Vedal S; Rodes C; Dutton SJ; Gelfand EW; Rabinovitch N. (2006). Estimating effects of ambient PM2.5
exposure on health using PM2.5 component measurements and regression calibration. J Expo Sci Environ
Epidemiol, 16: 30-38. 089203
Suglia SF; Gryparis A; Wright RO; Schwartz J; Wright RJ. (2008). Association of black carbon with cognition among
children in a prospective birth cohort study. Am J Epidemiol, 167: 280-286. 157027
Sullivan J; Ishikawa N; Sheppard L; Siscovick D; Checkoway H; Kaufman J. (2003). Exposure to ambient fine particulate
matter and primary cardiac arrest among persons with and without clinically recognized heart disease. Am J
Epidemiol, 157: 501-509. 043156
Sullivan J; Sheppard L; Schreuder A; Ishikawa N; Siscovick D; Kaufman J. (2005). Relation between short-term fine-
particulate matter exposure and onset of myocardial infarction. Epidemiology, 16: 41-48. 050854
December 2009 6-241
-------
Sullivan JH; Hubbard R; Liu SL; Shepherd K; Trenga CA; Koenig JQ; Chandler WL; Kaufman JD. (2007). A community
study of the effect of particulate matter on blood measures of inflammation and thrombosis in an elderly population.
Environ Health Perspect, 6:3. 100083
Sullivan JH; Schreuder AB; Trenga CA; Liu SL; Larson TV; Koenig JQ; Kaufman JD. (2005). Association between short
term exposure to fine particulate matter and heart rate variability in older subjects with and without heart disease.
Thorax, 60: 462-466. 109418
Sureshkumar V; Paul B; Uthirappan M; Pandey R; Sahu AP; Lai K; Prasad AK; Srivastava S; Saxena A; Mathur N; Gupta
YK. (2005). Proinflammatory and anti-inflammatory cytokine balance in gasoline exhaust induced pulmonary
injury in mice. Inhal Toxicol, 17: 161-168. 088306
Symons JM; Wang L; Guallar E; Howell E; Dominici F; Schwab M; Ange BA; Samet J; Ondov J; Harrison D; Geyh A.
(2006). A case-crossover study of fine particulate matter air pollution and onset of congestive heart failure symptom
exacerbation leading to hospitalization. Am J Epidemiol, 164: 421-433. 091258
Szyszkowicz M. (2008). Ambient air pollution and daily emergency department visits for ischemic stroke in Edmonton,
Canada. Int J Occup Med Environ Health, 21: 295-300. 192128
Szyszkowicz M. (2009). Air pollution and ED visits for chest pain. Am J Emerg Med, 27: 165-168. 191996
S0rensen M; Daneshvar B; Hansen M; Dragsted LO; Hertel O; Knudsen L; Loft S. (2003). Personal PM2.5 exposure and
markers of oxidative stress in blood. Environ Health Perspect, 111: 161-166. 157000
Tamagawa E; Bai N; Morimoto K; Gray C; Mui T; Yatera K; Zhang X; Xing L; Li Y; Laher I. (2008). Particulate matter
exposure induces persistent lung inflammation and endothelial dysfunction. Am J Physiol Lung Cell Mol Physiol,
295:L79-L85. 191988
Tan WC; Qiu D; Liam BL; Ng TP; Lee SH; Van Eeden SF; D'Yachkova Y; Hogg JC. (2000). The human bone marrow
response to acute air pollution caused by forest fires. Am J Respir Crit Care Med, 161: 1213-1217. 002304
Tankersley CG; Bierman A; Rabold R. (2007). Variation in heart rate regulation and the effects of particle exposure in
inbred mice. Inhal Toxicol, 19: 621-629. 097910
Tankersley CG; Campen M; Bierman A; Flanders SE; Broman KW; Rabold R. (2004). Particle effects on heart-rate
regulation in senescent mice. Inhal Toxicol, 16: 381-390. 094378
Tankersley CG; Champion HC; Takimoto E; Gabrielson K; Bedja D; Misra V; El-Haddad H; Rabold R; Mitzner W. (2008).
Exposure to inhaled particulate matter impairs cardiac function in senescent mice. Am J Physiol Regul Integr Comp
Physiol, 295: R252-R263. 157043
Tankersley CG; Irizarry R; Flanders SE; Rabold R; Frank R. (2003). Unstable heart rate and temperature regulation predict
mortality in AKR/J mice. Am J Physiol, 284: R742-R750. 053919
Tecer LH; Alagha O; Karaca F; Tuncel G; Eldes N. (2008). Particulate Matter (PM2.5, PM10-2.5, and PM10) and
Children's Hospital Admissions for Asthma and Respiratory Diseases: A Bidirectional Case-Crossover Study. J
Toxicol Environ Health A, 71: 512-520. 180030
TFESC. (1996). Heart rate variability: Standards of measurement, physiological interpretation, and clinical use. Task Force
of the European Society of Cardiology. North American Society of Pacing and Electrophysiology . Eur Heart J, 17:
354-381.003061
Thackara JW. (1981). An acute inhalation toxicity study of AC-7987-001 in rats with attachments and cover letter dated
1/24/92. 078988
Thong T; Raitt MH. (2007). Predicting imminent episodes of ventricular tachyarrhythmias using heart rate. Pacing Clin
Electrophysiol, 30: 874-884. 197462
Thurston G; Ito K; Mar T; Christensen WF; Eatough DJ; Henry RC; Kim E; Laden F; Lall R; Larson TV; Liu H; Neas L;
Pinto J; Stolzel M; Suh H; Hopke PK. (2005). Results and implications of the workshop on the source
apportionment of PM health effects. Epidemiology, 16: S134-S135. 097949
Thurston GD; Ito K; Hayes CG; Bates DV; Lippmann M. (1994). Respiratory hospital admissions and summertime haze air
pollution in Toronto, Ontario: consideration of the role of acid aerosols. Environ Res, 65: 271-290. 043921
December 2009 6-242
-------
Timonen KL; Hoek G; Heinrich J; Bernard A; Brunekreef B; De Hartog J; Hameri K; Ibald-Mulli A; Mirme A; Peters A;
Tiittanen P; Kreyling WG; Pekkanen J. (2004). Daily variation in fine and ultrafine particulate air pollution and
urinary concentrations of lung Clara cell protein CC16. Occup Environ Med, 61: 908-914. 087915
Timonen KL; Vanninen E; De Hartog J; Ibald-Mulli A; Brunekreef B; Gold DR; Henrich J; Hoek G; Lanki T; Peters A;
Tarkiainen T; Tiittanen P; Kreyling W; Pekkanen J. (2006). Effects of ultrafine and fine particulate and gaseous air
pollution on cardiac autonomic control in subjects with coronary artery disease: The ULTRA study. J Expo Sci
Environ Epidemiol, 16: 332-341.088747
Tolbert PE; Klein M; Metzger KB; Peel J; Flanders WD; Todd K; Mulholland JA; Ryan PB; Frumkin H. (2000). Interim
results of the study of particulates and health in Atlanta (SOPHIA). J Expo Sci Environ Epidemiol, 10: 446-460.
010320
Tolbert PE; Klein M; Peel JL; Sarnat SE; Sarnat JA. (2007). Multipollutant modeling issues in a study of ambient air
quality and emergency department visits in Atlanta. J Expo Sci Environ Epidemiol, 17: S29-S35. 090316
Tornqvist H; Mills NL; Gonzalez M; Miller MR; Robinson SD; Megson IL; MacNee W; Donaldson K; Soderberg S;
Newby DE; Sandstrom T; Blomberg A. (2007). Persistent endothelial dysfunction in humans after diesel exhaust
inhalation. Am J Respir Grit Care Med, 176: 395-400. 091279
Trenga CA; Sullivan JH; Schildcrout JS; Shepherd KP; Shapiro GG; Liu LJ; Kaufman JD; Koenig JQ. (2006). Effect of
particulate air pollution on lung function in adult and pediatric subjects in a Seattle panel study. Chest, 129: 1614-
1622. 155209
Tsai FC; Apte MG; Daisey JM. (2000). An exploratory analysis of the relationship between mortality and the chemical
composition of airborne particulate matter. Inhal Toxicol, 12: 121-135. 006251
Tsai S-S; Cheng M-H; Chiu H-F; Wu T-N; Yang C-Y. (2006). Air pollution and hospital admissions for asthma in a tropical
city: Kaohsiung, Taiwan. Inhal Toxicol, 18: 549-554. 089768
Tsai S-S; Goggins WB; Chiu H-F; Yang C-Y. (2003). Evidence for an association between air pollution and daily stroke
admissions in Kaohsiung, Taiwan. Stroke, 34: 2612-2616. 080133
Tunnicliffe WS; Harrison RM; Kelly FJ; Dunster C; Ayres JG (2003). The effect of sulphurous air pollutant exposures on
symptoms, lung function, exhaled nitric oxide, and nasal epithelial lining fluid antioxidant concentrations in normal
and asthmatic adults. Occup Environ Med, 60: 1-7. 088744
U.S. EPA. (1982). Air quality criteria for particulate matter and sulfur oxides, Vol 1, 2, 3. U.S. Environmental Protection
Agency, Environmental Criteria and Assessment Office. Washington, D.C.. EPA 600/8-82-029a; EPA 600/8-82-
029b; EPA 600/8-82-029c. http://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=3000188Z.txt;
http ://nepis. epa. gov/Exe/ZyPURL. cgi?Dockey=3 00018EV.txt;
http://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=300053KV.txt. 017610
U.S. EPA. (1996). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research Triangle
Park, NC. EPA/600/P-95/001aF-cF. 079380
U.S. EPA. (2002). Health assessment document for diesel engine exhaust. U.S. Environmental Protection Agency.
Washington, DC. 042866
U.S. EPA. (2004). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research Triangle
Park, NC. EPA/600/P-99/002aF-bF. 056905
Ulirsch GV; Ball LM; Kaye W; Shy CM; Lee CV; Crawford-Brown D; Symons M; Holloway T. (2007). Effect of
particulate matter air pollution on hospital admissions and medical visits for lung and heart disease in two southeast
Idaho cities. J Expo Sci Environ Epidemiol, 17: 478-487. 091332
Upadhyay D; Panduri V; Ghio A; Kamp DW. (2003). Particulate matter induces alveolar epithelial cell DNA damage and
apoptosis: role of free radicals and the mitochondria. Am J Respir Cell Mol Biol, 29: 180-187. 097370
Upadhyay S; Stoeger T; Harder V; Thomas RF; Schladweiler MC; Semmler-Behnke M; Takenaka S; Karg E; Reitmeir P;
Bader M; Stampfl A; Kodovanti U; Schulz H. (2008). Exposure to ultrafine carbon particles at levels below
detectablepulmonary inflammation affects cardiovascular performance inspontaneously hypertensive rats. Part
Fibre Toxicol, 5:19. 159345
Urch B; Brook JR; Wasserstein D; Brook RD; Rajagopalan S; Corey P; Silverman F. (2004). Relative contributions of
PM2.5 chemical constituents to acute arterial vasoconstriction in humans. Inhal Toxicol, 16: 345-352. 055629
December 2009 6-243
-------
Urch B; Silverman F; Corey P; Brook JR; Lukic KZ; Rajagopalan S; Brook RD. (2005). Acute blood pressure responses in
healthy adults during controlled air pollution exposures. Environ Health Perspect, 113: 1052-1055. 081080
Vedal S; Rich K; Brauer M; White R; Petkau J. (2004). Air pollution and cardiac arrhythmias in patients with implantable
cardiovascular defibrillators. Inhal Toxicol, 16: 353-362. 055630
Vegni FE; Ros O. (2004). Hospital accident and emergency burden is unaffected by today's air pollution levels. Eur J
Emerg Med, 11: 86-88. 087448
Veranth JM; Moss TA; Chow JC; Labban R; Nichols WK; Walton JC; Walton JG; Yost GS. (2006). Correlation of in vitro
cytokine responses with the chemical composition of soil-derived particulate matter. Environ Health Perspect, 114:
341-349. Q87479
Veronesi B; Makwana O; Pooler M; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles
VII Degeneration of dopaminergic neurons inApo E-/- mice. Inhal Toxicol, 17: 235-241. 087481
Vigotti MA; Chiaverini F; Biagiola P; Rossi G (2007). Urban air pollution and emergency visits for respiratory complaints
in Pisa, Italy. J Toxicol Environ Health A, 70: 266-269. 090711
Villeneuve PJ; Burnett RT; Shi Y; Krewski D; Goldberg MS; Hertzman C; Chen Y; Brook J. (2003). A time-series study of
air pollution, socioeconomic status, and mortality in Vancouver, Canada. J Expo Sci Environ Epidemiol, 13: 427-
435. 055051
Villeneuve PJ; Chen L; Stieb D; Rowe BH. (2006). Associations between outdoor air pollution and emergency department
visits for stroke in Edmonton, Canada. Eur J Epidemiol, 21: 689-700. 090191
Vincent R; Kumarathasan P; Goegan P; Bjarnason SG; Guenette J; Berube D; Adamson IY; Desjardins S; Burnett RT;
Miller FJ; Battistini B. (2001). Inhalation toxicology of urban ambient particulate matter: acute cardiovascular
effect in rats. 021184
Von Klot S; Peters A; Aalto P; Bellander T; Berglind N; DTppoliti D; Elosua R; Hermann A; Kulmala M; Lanki T; Lowel
H; Pekkanen J; Picciotto S; Sunyer J; Forastiere F; Health Effects of Particles on Susceptible Subpopulations
(HEAPSS) Study Group. (2005). Ambient air pollution is associated with increased risk of hospital cardiac
readmissions of myocardial infarction survivors in five European cities. Circulation, 112: 3073-3079. 088070
Von Klot S; Wolke G; Tuch T; Heinrich J; Dockery DW; Schwartz J; Kreyling WG; Wichmann HE; Peters A. (2002).
Increased asthma medication use in association with ambient fine and ultrafine particles. Eur Respir J, 20: 691-702.
034706
Ward DJ; Roberts KT; Jones N; Harrison RM; Ayres JG; Hussain S; Walters S. (2002). Effects of daily variation in outdoor
particulates and ambient acid species in normal and asthmatic children. Thorax, 57: 489-502. 025839
Ward PA. (2003). Acute lung injury: how the lung inflammatory response works. Eur Respir J, 44: 22s-23s. 157111
WegesserTC; Last JA. (2008). Lung response to coarse PM: bioassay in mice. Toxicol Appl Pharmacol, 230: 159-166.
190506
Welin L; Eriksson H; Larsson B; Svardsudd K; Wilhelmsen L; Tibblin G. (1993). Risk Factors for Coronary Heart Disease
during 25 Years of Follow-Up. Cardiology, 82: 223-228. 156151
Wellenius GA; Batalha JRF; Diaz EA; Lawrence J; Coull BA; Katz T; Verrier RL; Godleski JJ. (2004). Cardiac effects of
carbon monoxide and ambient particles in a rat model of myocardial infarction. Toxicol Sci, 80: 367-376. 087874
Wellenius GA; Bateson TF; Mittleman MA; Schwartz J. (2005). Particulate air pollution and the rate of hospitalization for
congestive heart failure among medicare beneficiaries in Pittsburgh, Pennsylvania. Am J Epidemiol, 161: 1030-
1036. 087483
Wellenius GA; Coull BA; Batalha JRF; Diaz EA; Lawrence J; Godleski JJ. (2006). Effects of ambient particles and carbon
monoxide on supraventricular arrhythmias in a rat model of myocardial infarction. Inhal Toxicol, 18: 1077-1082.
156152
Wellenius GA; Coull BA; Godleski JJ; Koutrakis P; Okabe K; Savage ST. (2003). Inhalation of concentrated ambient air
particles exacerbates myocardial ischemia in conscious dogs. Environ Health Perspect, 111: 402-408. 055691
Wellenius GA; Saldiva PHN; Batalha JRF; Murthy GGK; Coull BA; Verrier RL; Godleski JJ. (2002). Electrocardiographic
changes during exposure to residual oil fly ash (ROFA) particles in a rat model of myocardial infarction. Toxicol
Sci, 66: 327-335. 025405
December 2009 6-244
-------
Wellenius GA; Schwartz J; Mittleman MA. (2005). Air pollution and hospital admissions for ischemic and hemorrhagic
stroke among medicare beneficiaries. Stroke, 36: 2549-2553. 088685
Wellenius GA; Schwartz J; Mittleman MA. (2006). Particulate air pollution and hospital admissions for congestive heart
failure in seven United States cities. Am J Cardiol, 97: 404-408. 088748
Wellenius GA; Yeh GY; Coull BA; Suh HH; Phillips RS; Mittleman MA. (2007). Effects of ambient air pollution on
functional status in patients with chronic congestive heart failure: a repeated-measures study. Environ Health, 6: 1-
7. 092830
Welty LJ; Zeger SL. (2005). Are the acute effects of particulate matter on mortality in the National Morbidity, Mortality,
and Air Pollution study the result of inadequate control for weather and season? A sensitivity analysis using flexible
distributed lag models. Am J Epidemiol, 162: 80-88. 087484
Wheeler A; Zanobetti A; Gold DR; Schwartz J; Stone P; Suh HH. (2006). The relationship between ambient air pollution
and heart rate variability differs for individuals with heart and pulmonary disease. Environ Health Perspect, 114:
560-566. 088453
Whitekus MJ; Li N; Zhang M; Wang M; Horwitz MA; Nelson SK; Horwitz LD; Brechun N; Diaz-Sanchez D; Nel AE.
(2002). Thiol antioxidants inhibit the adjuvant effects of aerosolized diesel exhaust particles in a murine model for
ovalbumin sensitization. J Immigr Minor Health, 168: 2560-2567. 157142
Whitsel E; Quibrera P; Christ S; Liao D; Prineas R; Anderson G; Heiss G (2009). Heart rate variability, ambient particulate
matter air pollution, and glucose homeostasis: the environmental epidemiology of arrhythmogenesis in the Women's
Health Initiative. Am J Epidemiol, 169: 693-703. 191980
Wichmann H-E; Spix C; Tuch T; Wolke G; Peters A; Heinrich J; Kreyling WG; Heyder J. (2000). Daily mortality and fine
and ultrafine particles in Erfurt, Germany Part I: role of particle number and particle mass. Health Effects Institute.
Boston, MA. 013912
Widdicombe J. (2006). Reflexes from the lungs and airways: historical perspective. J Appl Physiol, 101: 628. 155519
Widdicombe JG (2003). Overview of neural pathways in allergy and asthma. Pulm Pharmacol Ther, 16: 23-30. 157145
Wilson WE; Mar TF; Koenig JQ. (2007). Influence of exposure error and effect modification by socioeconomic status on
the association of acute cardiovascular mortality with particulate matter in Phoenix. J Expo Sci Environ Epidemiol,
17: S11-S19. 157149
Win-Shwe TT; Yamamoto S; Fujitani Y; Hirano S; Fujimaki H. (2008). Spatial learning and memory function-related gene
expression in the hippocampus of mouse exposed to nanoparticle-rich diesel exhaust. Neurotoxicology, 29: 940-
947. 190146
Witten ML; Wong SS ; Sun NN; Keith I; Kweon C; Foster DE; Schauer JJ; Sherrill DL. (2005). Neurogenic responses in
rat lungs after nose-only exposure to diesel exhaust. 087485
Wong C-M; Atkinson RW; Anderson HR; Hedley AJ; Ma S; Chau PY-K; Lam T-H. (2002). A tale of two cities: effects of
air pollution on hospital admissions in Hong Kong and London compared. Environ Health Perspect, 110: 67-77.
023232
Wong SS; Sun NN; Keith I; Kweon C-B; Foster DE; Schauer James J; Witten ML. (2003). Tachykinin substance P
signaling involved in diesel exhaust-induced bronchopulmonary neurogenic inflammation in rats. Arch Toxicol, 77:
638-650. 097707
Wong TW; Lau TS; Yu TS; Neller A; Wong SL; Tarn W; Pang SW (1999). Air pollution and hospital admissions for
respiratory and cardiovascular diseases in Hong Kong. Occup Environ Med, 56: 679-683. 009172
Wordley J; Walters S; Ayres JG. (1997). Short term variations in hospital admissions and mortality and particulate air
pollution. Occup Environ Med, 54: 108-116. 082745
Wu J; M Winer A; J Delfino R. (2006). Exposure assessment of particulate matter air pollution before, during, and after the
2003 Southern California wildfires. Atmos Environ, 40: 3333-3348. 157156
Yan YH; Huang CH; Chen WJ; Wu MF; Cheng TJ. (2008). Effects of diesel exhaust particles on left ventricular function in
isoproterenol-induced myocardial injury and healthy rats. Inhal Toxicol, 20: 199-203. 098625
Yang C-Y; Chen Y-S; Yang C-H; Ho S-C. (2004). Relationship between ambient air pollution and hospital admissions for
cardiovascular diseases in Kaohsiung, Taiwan. J Toxicol Environ Health A, 67: 483-493. 094376
December 2009 6-245
-------
Yang CY. (2008). Air pollution and hospital admissions for congestive heart failure in a subtropical city: Taipei, Taiwan. J
Toxicol Environ Health A, 71: 1085-1090. 157160
Yang CY; Chen CC; Chen CY; Kuo HW. (2007). Air pollution and hospital admissions for asthma in a subtropical city:
Taipei, Taiwan. J Toxicol Environ Health A, 70: 111-117. 092848
Yang CY; Chen CJ. (2007). Air pollution and hospital admissions for chronic obstructive pulmonary disease in a
subtropical city: Taipei, Taiwan. J Toxicol Environ Health A, 70: 1214-1219. 092847
Yang CY; Cheng MH; Chen CC. (2009). Effects of Asian Dust Storm Events on Hospital Admissions for Congestive Heart
Failure in Taipei, Taiwan. J Toxicol Environ Health A, 72: 324-328. 190341
Yang Q; Chen Y; Krewski D; Shi Y; Burnett RT; McGrail KM. (2004). Association between particulate air pollution and
first hospital admission for childhood respiratory illness in Vancouver, Canada. Arch Environ Occup Health, 59: 14-
21.087488
Yeatts K; Svendsen E; Creason J; Alexis N; Herbst M; Scott J; Kupper L; Williams R; Neas L; Cascio W; Devlin RB;
Peden DB. (2007). Coarse particulate matter (PM25-10) affects heart rate variability, blood lipids, and circulating
eosinophils in adults with asthma. Environ Health Perspect, 115: 709-714. 091266
Yokota S; Seki T; Naito Y; Tachibana S; Hirabayashi N; Nakasaka T; Ohara N; Kobayashi H. (2008). Tracheal instillation
of diesel exhaust particles component causes blood and pulmonary neutrophilia and enhances myocardial oxidative
stress in mice. J Toxicol Sci, 33: 609-620. 190109
Yu O; Sheppard L; Lumley T; Koenig JQ; S. (2000). Effects of ambient air pollution on symptoms of asthma in Seattle-
area children enrolled in the CAMP study. Environ Health Perspect, 108: 1209-1214. 013254
Yue W; Schneider A; Stolzel M; Ruckerl R; Cyrys J; Pan X; Zareba W; Koenig W; Wichmann HE; Peters A. (2007).
Ambient source-specific particles are associated with prolonged repolarization and increased levels of inflammation
in male coronary artery disease patients. Mutat Res Fund Mol Mech Mutagen, 621: 50-60. 097968
Zabel M; Klingenheben T; Franz MR; Hohnloser SH. (1998). Assessment of QT Dispersion for Prediction of Mortality or
Arrhythmic Events After Myocardial Infarction : Results of a Prospective, Long-term Follow-up Study. Circulation,
97: 2543-2550. 156176
Zanchi AC; Venturini CD; Saiki M; Nascimento Saldiva PH; Tannhauser Barros HM; Rhoden CR. (2008). Chronic nasal
instillation of residual-oil fly ash (ROFA) induces brain lipid peroxidation and behavioral changes in rats. Inhal
Toxicol, 20: 795-800. 157173
Zanobetti A; Canner MJ; Stone PH; Schwartz J; Sher D; Eagan-Bengston E; Gates KA; Hartley LH; Suh H; Gold DR.
(2004). Ambient pollution and blood pressure in cardiac rehabilitation patients. Circulation, 110: 2184-2189.
087489
Zanobetti A; Schwartz J. (2003). Airborne particles and hospital admissions for heart and lung disease. 043119
Zanobetti A; Schwartz J. (2003). Airborne particles and hospital admissions for heart and lung disease. In: Revised
analyses of time-series studies of air pollution and health. Special Report. 157174
Zanobetti A; Schwartz J. (2005). The effect of particulate air pollution on emergency admissions for myocardial infarction:
A multicity case-crossover analysis. Environ Health Perspect, 113: 978-982. 088069
Zanobetti A; Schwartz J. (2006). Air pollution and emergency admissions in Boston, MA. J Epidemiol Community Health,
60: 890-895. 090195
Zanobetti A; Schwartz J. (2009). The effect of fine and coarse particulate air pollution on mortality: A national analysis.
Environ Health Perspect, 117: 1-40. 188462
Zareba W; Couderc JP; Oberdorster G; Chalupa D; Cox C; Huang LS; Peters A; Utell MJ; Frampton MW. (2009). EGG
parameters and exposure to carbon ultrafine particles in young healthy subjects. Inhal Toxicol, 21: 223-233. 190101
Zeka A; Sullivan JR; Vokonas PS; Sparrow D; Schwartz J. (2006). Inflammatory markers and particulate air pollution:
characterizing the pathway to disease. Int J Epidemiol, 35: 1347-1354. 157177
Zeka A; Zanobetti A; Schwartz J. (2005). Short term effects of particulate matter on cause specific mortality: effects of lags
and modification by city characteristics. Occup Environ Med, 62: 718-725. 088068
Zeka A; Zanobetti A; Schwartz J. (2006). Individual-level modifiers of the effects of particulate matter on daily mortality.
Am J Epidemiol, 163: 849-859. 088749
December 2009 6-246
-------
Zelikoff JT; Chen LC; Cohen MD; Fang K; Gordon T; Li Y; Nadziejko C; Schlesinger RB. (2003). Effects of inhaled
ambient participate matter on pulmonary antimicrobial immune defense. Inhal Toxicol, 15: 131-150. 039009
Zhang SH; Reddick RL; Piedrahita JA; Maeda N. (1992). Spontaneous hypercholesterolemia and arterial lesions in mice
lacking apolipoprotein E. Science, 258: 468-471. 157180
Zhang Z; Whitsel E; Quibrera P; Smith R; Liao D; Anderson G; Prineas R. (2009). Ambient fine particulate matter
exposure and myocardial ischemia in the Environmental Epidemiology of Arrhythmogenesis in the Women's Health
Initiative (EEAWHI) study. Environ Health Perspect, 117: 751-756. 191970
December 2009 6-247
-------
Chapter 7. Integrated Health Effects of
Long-Term PM Exposure
7.1. Introduction
This chapter reviews, summarizes, and integrates the evidence on relationships between health
effects and long-term exposures to various size fractions and sources of PM. Cardiopulmonary
health effects of long-term exposure to PM have been examined in an extensive body of
epidemiologic and toxicological studies. Both epidemiologic and toxicological studies provide a
basis for examining reproductive and developmental and cancer health outcomes with regard to
long-term exposure to PM. In addition, there is a large body of epidemiologic literature evaluating
the relationship between mortality and long-term exposure to PM.
Conclusions from the 2004 PM AQCD are summarized briefly at the beginning of each
section, and the evaluation of evidence from recent studies builds upon what was available during
the previous review. For each health outcome (e.g., respiratory infections, lung function), results are
summarized for studies from the specific scientific discipline, i.e., epidemiologic and toxicological
studies. The major sections (i.e., cardiovascular, respiratory, reproductive/developmental, cancer)
conclude with summaries of the evidence for the various health outcomes within that category and
integration of the findings that lead to conclusions regarding causality based upon the framework
described in Chapter 1. Determination of causality is made for the overall health effect category,
such as cardiovascular effects, with coherence and plausibility being based upon the evidence from
across disciplines and also across the suite of related health outcomes including cause-specific
mortality. Section 7.6 provides detailed discussions on the epidemiologic literature for long-term
exposure to PM and mortality. In each summary section (7.2.11, 7.3.9, 7.4.3, 7.5.4, and 7.6.5), the
evidence is briefly reviewed and independent conclusions drawn for relationships with PM2 5,
PMio.2.5, and UF particles (UFPs).
7.2. Cardiovascular and Systemic Effects
Studies examining associations between long-term exposure to ambient PM (over months to
years) and CVD morbidity had not been conducted and thus were not included in the 1996 or 2004
PM Air Quality Criteria Documents (U.S. EPA, 1996, 079380: U.S. EPA, 2004, 056905). A number
of studies were included in the 2004 PM AQCD that evaluated the effect of long-term PM2.5
exposure on cardiovascular mortality and found consistent associations. No toxicological studies
examined chronic atherosclerotic effects of PM exposure in animal models. However, a subchronic
study that evaluated atherosclerosis progression in hyperlipidemic rabbits was discussed and this
study provided the foundation for the subsequent work that has been conducted in this area (Suwa et
al., 2002, 028588). No previous toxicological studies evaluated effects of subchronic or chronic PM
exposure on diabetes measures, or HR or HRV changes, nor were there animal toxicological studies
included in the 2004 PM AQCD that evaluated systemic inflammatory or blood coagulation markers
following subchronic or chronic PM exposure.
Several new epidemiologic studies have examined the long-term PM-CVD association among
U.S. and European populations. The studies investigate the association of both PM2.5 and PMi0
exposures with a variety of clinical and subclinical CVD outcomes. Epidemiologic and toxicological
studies have provided evidence of the adverse effects of long-term exposure to PM2.5 on
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009 7-1
-------
cardiovascular outcomes, including atherosclerosis, clinical and subclinical markers of
cardiovascular morbidity, and cardiovascular mortality. The evidence of these effects from long-term
exposure to PMi0_2.5 is weaker.
7.2.1. Atherosclerosis
Atherosclerosis is a progressive disease that contributes to several adverse outcomes,
including acute coronary syndromes such as myocardial infarction, sudden cardiac death, stroke and
vascular aneurysms. It is multifaceted, beginning with an early injury or inflammation that promotes
the extravasation of inflammatory cells. Under conditions of oxidative or nitrosative stress and high
lipid or cholesterol concentrations, the vessel wall undergoes a chronic remodeling that is
characterized by the presence of foam cells, migrated and differentiated smooth muscle cells, and
ultimately a fibrous cap. The advanced lesion that develops from this process can occlude perfusion
to distal tissue, causing ischemia, and erode, degrade, or even rupture, revealing coagulant initiators
(tissue factor) that promote thrombosis, stenosis, and infarction or stroke. Several detailed reviews of
atherosclerosis pathology have been published elsewhere (Ross, 1999, 156926; Stacker and Keaney,
2004, 1570131
7.2.1.1. Epidemiologic Studies
Measures of Atherosclerosis
Although no study has examined the association between long-term PM exposure and
longitudinal change in subclinical markers of atherosclerosis, several cross sectional studies have
been conducted. Markers of atherosclerosis used in these studies include coronary artery calcium
(CAC), carotid intima-media thickness (CIMT), ankle-brachial index (ABI), and abdominal aortic
calcium (AAC). These measures are descried briefly below.
CAC represents the accumulation of calcium in coronary artery macrophages and represents
an advanced stage of atherosclerosis. As such CAC is a measure of atherosclerosis assessed by non-
contrast, cardiac-gated electron beam computed tomography (EBCT) or multidetector computed
tomography (MDCT) of the coronary arteries in the heart (Greenland and Kizilbash, 2005, 156496;
Hoffmann et al, 2005, 156556; Mollet et al, 2005, 155988). The prevalence of CAC is strongly
related to age. Few people have detectable CAC in their second decade of life but the prevalence of
CAC rises to approximately 100% by age 80 (Ardehali et al., 2007, 155662). Previous studies
suggest that while the absence of CAC does not rule out atherosclerosis, it does imply a very low
likelihood of significant arterial obstruction (Achenbach and Daniel, 2001, 156189; Arad et al.,
1996, 155661; Shaw et al., 2003, 156083; Shemesh et al., 1996, 156085). Conversely, the presence
of CAC confirms the existence of atherosclerotic plaque and the amount of calcification varies
directly with the likelihood of obstructive disease (Ardehali et al., 2007, 155662). CAC is a
quantified using the Agatston method (Agatston et al., 1990, 156197). Its repeatability depends on
the laboratory and the method of calculation (O'Rourke et al., 2000, 192159). Agatston scores are
frequently used to classify individuals into one of five groups (zero; mild; moderate; severe;
extensive) or according to age- and sex-specific percentiles of the CAC distribution (Erbel et al.,
2007, 155768).
CIMT is a measure of atherosclerosis assessed by high-resolution, B-mode ultrasonography of
the carotid arteries in the neck, the walls of which have inner (intimal), middle (medial) and outer
(adventitial) layers (Craven et al., 1990, 155740; O'Leary et al., 1999, 156826; Wendelhag et al.,
1993, 157136). CIMT estimates the distance in mm or (im between the innermost (blood-intima) and
outermost (media-adventitia) interfaces, often by averaging over three arterial segments in the
common carotid, carotid bulb, and internal carotid artery (Amato et al., 2007, 155656). CIMT has
been associated with atherosclerosis risk factors (Heiss et al., 1991, 156535; O'Leary et al., 1992,
156825; Salonen and Salonen, 1991, 156938). prevalent coronary heart disease (Chambless et al.,
1997, 156329; Geroulakos et al., 1994, 155788). and incident coronary and cerebral events (O'Leary
et al., 1999, 156826; van der Meer et al., 2004, 156129). Several studies have indicated that CIMT
measurements are accurate (Girerd et al., 1994, 156474; Pignoli et al., 1986, 156026; Wendelhag et
December 2009 7-2
-------
al., 1991, 157135) and reproducible (Montauban et al, 1999, 156777; Smilde et al, 1997, 156988;
Willekes et al., 1999, 157147), especially for the common carotid artery (Montauban et al., 1999,
156777).
ABI, which is also known as the ankle-arm or resting (blood) pressure index, is a measure of
lower extremity arterial occlusive disease commonly caused by advanced atherosclerosis (Weitz et
al., 1996, 156150). It is assessed by continuous wave Doppler and manual or automated
oscillometric sphygmomanometry, the latter having been shown to have higher repeatability and
validity (Weitz et al., 1996, 156150). ABI is defined as the unitless ratio of ankle to brachial systolic
blood pressures measured in mmHg. As ankle pressure is normally equal to or slightly higher than
arm pressure (resulting in an ABI > 1.0), epidemiologic studies typically define the normal ABI
range as 0.90 to 1.50 (Resnick et al., 2004, 156048). Low ABI has been associated with all-cause and
CVD mortality (Newman et al., 1993, 156805; Vogt et al., 1993, 157100). as well as myocardial
infarction and stroke (Karthikeyan and Lip, 2007, 156626).
AAC is a measure of atherosclerosis assessed by non-contrast, EBCT or MDCT of the
abdominal aorta. It is scored much like CAC (Agatston et al., 1990, 156197). but the age-specific
prevalence and extent of AAC is greater, particularly among women and at ages >50 yr. Although
AAC has not been studied as extensively as CAC, it is associated with carotid and coronary
atherosclerosis as well as cardiovascular morbidity and mortality (Allison et al., 2004, 156210;
Allison et al., 2006, 155653; Hollander et al., 2003, 156562; Khoury et al., 1997, 156636; Oei et al.,
2002, 156820; Walsh et al., 2002, 157103; Wilson et al., 2001, 156159; Witteman et al., 1986,
156161) and measurements are sufficiently reproducible to allow serial investigations over time
(Budoff et al., 2005, 192105).
Study Findings
Diez Roux et al. (2008, 156401) conducted cross-sectional analyses of the association of three
of these subclinical markers of atherosclerosis (CAC, CIMT and ABI), collected from 2000 to 2003
during baseline examinations of participants enrolled in the Multi-Ethnic Study of Atherosclerosis
(MESA), with long-term exposure to PM25 and PMi0. The study population included 5,172
ethnically diverse people (53% female) residing in Baltimore, MD; Chicago, IL; Forsyth County,
NC; Los Angeles, CA; New York, NY; and St. Paul, MN ranging in age from 45 to 84 yr old.
Authors used spatio-temporal modeling of pollutant concentrations, weather and demographic data
to impute 20-yr avg exposures to PM2 5 and PMi0. They reported small increases in CIMT of 1%
(95% CI: 0-1.4) and 0.5% (95% CI: 0-1), which correspond to absolute changes of 8 (95% CI: 0-12)
and 7 (95% CI: 0-14) urn, per 10 ug/m3 increase in 20-yr avg PMi0 and PM25 concentration,
respectively. Evidence of age-, gender-, lipid- and smoking-related susceptibility was lacking. They
also reported weak, non-significant increases in the relative prevalence of CAC of 1% (95% CI: -2 to
4) and 0.5% (95% CI: -2 to 3) per 10 ug/m3 increase in PMi0 and PM2.5, respectively. Among the
subset of 2,586 participants with EBCT-identified calcification, similarly weak associations were
observed. There was little evidence of modification of the CAC associations by demographic,
socioeconomic or clinical characteristics. Finally, the authors report no differences in mean ABI with
PMio or PM25 concentrations. The null findings for ABI exhibited little heterogeneity among
participant subgroups and were similarly null when ABI was modeled as a dichotomous outcome
using a cutpoint of 0.9 units.
MESA investigators also examined the chronic PM2 5-AAC association in a residentially stable
subset of 1,147 participants (mean age = 66 yr; 50% female) randomly selected from all MESA
centers, except Baltimore, MD for enrollment in its Aortic Calcium Ancillary Study (Allen et al.,
2009, 156209). The authors used kriging and inverse residence-to-monitor distance-weighted
averaging of EPAAQS data to estimate 2-yr mean exposures to PM25. In cross-sectional analyses,
the authors found a 6% (95% CI: -4 to 16) excess risk of a non-zero Agatston score and an 8% (95%
CI: -30 to 46) increase in AAC, i.e., approximately 50 (95% CI: -251 to 385) Agatston units, per
10 ug/m3 increase in PM25 concentration. These associations were stronger among users than non-
users of lipid lowering drugs.
Kunzli et al. (2005, 087387) used baseline data collected between 1998-2003 from two
randomized placebo-controlled clinical trials, the Vitamin E Atherosclerosis Progression Study
(VEAPS) and the B-Vitamin Atherosclerosis Intervention Trial (BVAIT), for their ancillary cross-
sectional analyses of the effect of long-term PM25 exposure on CIMT. The study population included
798 residents of the greater Los Angeles, CA area who were more than 40 yr old at baseline and 44%
were female. The authors used universal kriging of PM25 data from 23 state and local monitors
December 2009 7-3
-------
operating in 2000 to estimate 1-yr avg exposure to PM25 at each participant's geocoded U.S. Postal
Service ZIP code. They found a 4.2% (95% CI: -0.2 to 8.9) or approximately 32 (95% CI: -2 to
68) urn increase in CIMT per 10 ug/m increase in PM25 concentration. In contrast to findings from
the relatively large, ethnically diverse, yet geographically overlapping MESA ancillary study
described above, PM-related increases in CIMT were two- to three-fold larger among older and
female participants taking lipid lowering drugs in this study. PM-related increases in CIMT were
also higher in never smokers when compared with current or former smokers.
Hoffmann et al. (2007, 091163) conducted a cross-sectional analysis of data collected at
baseline (2000-2003) for 4,494 residents of Essen, Mulheim and Bochum, Germany enrolled in the
Heinz Nixdorf Recall Study from 2000 to 2003. The age of participants ranged from 45-74 yr and
51% were female. In this cross-sectional study the authors used dispersion and chemistry transport
modeling of emissions, climate and topography data to estimate 1-yr avg exposure to PM2 5 in 2002
(the midpoint of the baseline exam.) They reported an imprecise 43% (95% CI: -15 to 115) or 102
(95% CI: -77 to 273) Agatston unit increase in CAC per 10 ug/m3 increase in PM2.5. Differences in
strength of association between subgroups defined by demographic and clinical characteristics were
small. The authors reported a more consistent association of CAC with traffic exposure (distance
from a major roadway) than with PM2 5 in this study.
In a subsequent analysis of these data, Hoffmann et al. (2009, 190376) examined the PM-ABI
association in this population. In this cross-sectional study, no changes in ABI were observed in
association with PM2 5 concentration nor was evidence of effect modification by demographic and
clinical characteristics apparent. As in the previous study (Hoffmann et al., 2007, 091163). residing
near a major roadway was a stronger predictor of atherosclerotic changes. Absolute changes in ABI
of-0.024 (95% CI: -0.047 to -0.001) were associated with living within 50 m of a major roadway
compared to living more than 200 m away.
Each of the studies described above relied on cross-sectional analyses examining differences
in long-term average PM2 5 concentrations across space (as well as time to the extent baseline
examinations were conducted over time). Such associations may reflect the effect of compositional
differences in PM25 as well as the effect of higher PM25 concentrations. Most associations of PM25
with CAC (Diez et al., 2008, 156401: Hoffmann et al., 2007, 091163). CIMT (Diez et al., 2008,
156401: Kunzli et al., 2005, 087387). ABI (Diez et al., 2008, 156401: Hoffmann et al., 2009,
190376) and AAC (Allen et al., 2009, 156209) reviewed in this section were weak and/or imprecise.
However, several factors including exposure measurement error, variation in baseline measures
atherosclerosis, as well as limited power may contribute to the insensitivity of these cross-sectional
studies to detect small differences in CAC, CIMT, ABI and AAC. The study by Hoffmann et al.
(2007, 091163). which reported large, imprecise and non-significant increases in CAC in association
with PM2 5, is not distinguished from the other studies reviewed by a superior study design or larger
sample size. The several fold difference in the magnitude of CIMT associations reported by Kunzli
et al. (2005, 087387) and Diez Roux et al. (2008, 156401) may be related to differences between the
study populations. The ambient PM concentrations from these studies are characterized in Table 7-1.
7.2.1.2. lexicological Studies
In the only study of this kind described in the 2004 PM AQCD, Suwa et al. (2002, 028588)
demonstrated more advanced atherosclerotic lesions based on phenotype and volume fraction in the
left main and right coronary arteries of rabbits exposed to PMi0 (5 mg/kg, 2 times/wkx4 wk).
Although this study was conducted using IT exposure methodology at a relatively high dose, it
provided the first experimental evidence that PM exposure may result in progression of
atherosclerosis. Recent toxicological studies conducted using inhalation exposures have replicated
these findings at relevant concentrations and are discussed below.
CAPS
New studies have demonstrated increased atherosclerotic plaque area in aortas of ApoE"7" mice
exposed to PM25 CAPs for 4-6 mo (6 h/day><5 days/wk). Average CAPs concentrations ranged from
85 to 138 ug/m and all of the studies were conducted in Tuxedo or Manhattan, NY. Chen and
Nadziejko (2005, 087219) reported that the percentage of aortic intimal surface covered by
atherosclerotic lesions in ApoE"7" mice was increased. In male ApoE"77LDLR"7" mice, both lesion area
December 2009 7-4
-------
and cellularity in the aortic root were enhanced by Tuxedo, NY CAPs exposure, although there was
no change in lipid content. Genetic profiles within plaques recovered from ApoE"" mice included
many of the molecular pathways known to contribute to atherosclerosis, including inflammation
(Floyd et al., 2009, 190350). Sun (2005, 087952) similarly demonstrated an enhancement of
atherosclerosis in ApoE" mice exposed Tuxedo, NY CAPs. Plaque area in the aortic arch and
abdominal aorta was significantly increased in the PM-exposed, high fat-chow group compared to
air-exposed, high fat-chow group. Macrophage infiltration in the abdominal aorta was also observed
in the groups exposed to CAPs. A study conducted in Manhattan for 4 mo (May- September 2007)
showed that PM2.5 CAPs exposure increased atherosclerotic plaque area and led to higher levels of
macrophage infiltration, collagen deposition, and lipid composition in thoracic aortas of ApoE"7" mice
(Ying et al., 2009, 190111). which is consistent with the previous two studies described that were
conducted in Tuxedo, NY.
Alteration of vasomotor function has been observed in aortic rings of ApoE~'~ mice on a high
fat diet with long-term exposure to CAPs (Sun et al., 2005, 087952; Ying et al., 2009, 190111). Sun
(2005, 087952) reported that. PM2.5-exposed animals exhibited increased vasoconstrictor
responsiveness to serotonin and PE. Increased ROS and elevated iNOS protein expression in aortic
sections of CAPs-exposed mice may have resulted alterations in the NO pathway and generation of
peroxynitrite that could have affected vascular reactivity. In contrast, Ying, et al. (2009, 190111)
demonstrated decreased maximum constriction induced by PE following Manhattan CAPs exposure.
Pretreatment with the soluble guanylate cyclase (sGC) inhibitor ODQ attenuated the response,
indicating that CAPs exposure resulted in abnormal NO/sGC signaling. Expression of iNOS mRNA
and protein was increased in aortas of CAPs-exposed mice, further supporting a role for NO
production. In conjunction with increased NO, aortic superoxide production was demonstrated that
appeared to be partially driven by increased NADPH oxidase activity. The difference in
vasoconstrictor responses between these two studies may be attributable to varying durations
(6 versus 4 mo, respectively) or CAPs compositions.
Sun (2005, 087952) and Ying et al. (2009, 190111) reported similar relaxation responses to
ACh for air- and CAPs-exposed mice. However, Manhattan CAPs-exposed mice had a markedly
decreased response to A23187, indicating that NO release occurred via Ca2+-dependent mechanisms
(Ying et al., 2009, 190111). Abnormal eNOS function is likely responsible for the decreased
relaxation response, as activation of eNOS (but not iNOS) is Ca2+-dependent.
A recent study (Sun et al., 2008, 157033) that was part of the research described above (Sun et
al., 2005, 087952) investigated tissue factor (TF) expression in aortas, which is a major regulator of
hemostasis and thrombosis following vascular injury or plaque erosion. In PM2.5-exposed ApoE"7"
mice on a high-fat diet, TF was significantly elevated in the plaques of aortic sections compared to
air-exposed mice on the high-fat diet. TF expression was generally detected in (1) the extracellular
matrix surrounding macrophages and foam cell-rich areas; and (2) around smooth muscle cells.
One new study of CAPs PM25 or UFPs derived from traffic was conducted. Araujo et al.
(2008, 156222) compared the relative impact of UF (0.01-0.18 urn) and fine (0.01-2.5 urn) PM
inhalation on aortic lesion development in ApoE"" mice following a 40-day exposure (5
h/day><3 days/wk for 75 total h). Animals were on a normal chow diet and exposed to CAPs in a
mobile inhalation laboratory parked 300 m from a freeway in downtown Los Angeles. Exposure
concentrations were -440 ug/m3 for PM25 and -110 ug/m3 for UFPs, and the number concentrations
were roughly equivalent (4.56*105 and 5.59*105 particles/cm3 for PM25 and UFPs, respectively).
Significant increases in plaque size (estimated by lesions at the aortic root) were reported for mice
exposed to UFPs only. The lesions were largely comprised of macrophages with intracellular lipid
accumulation. Increased total cholesterol measured at the end of the exposure protocol was observed
only in the PM2 5 group. HDL isolated from the UF PM-exposed mice demonstrated decreased anti-
inflammatory protective capacity against LDL-induced monocyte chemotactic activity in an in vitro
assay. The livers from the UFP-exposed mice demonstrated significant increases in lipid
peroxidation and several stress-related gene products (catalase, glutathione S-transferase Ya,
NADPH-quinone oxidoreductasel, superoxide dismutase 2). Thus, UFPs in these exposures had a
substantially greater impact on the systemic response than did PM2 5.
December 2009 7-5
-------
Ambient Air
A study employing young BALB/c mice examined the effects of a 4-month exposure (24
h/dayx7 days/wk) to ambient air on arterial histopathology (Lemos et al., 2006, 088594). Outdoor
exposure chambers were located in downtown Sao Paulo, Brazil next to streets of high traffic
density. In the control chamber, PMi0 and NO2 were filtered with 50% and 75% efficiency,
respectively. The average pollutant concentrations were 2.06 ppm for CO (8-h mean), 104.75 ug/m3
for NO2 (24-h mean), 11.07 ug/m3 for SO2 (24-h mean), and 35.52 ug/m3 for PMi0 (24-h mean) at a
monitoring site within 100 m of the inhalation chambers. The pulmonary and coronary arteries
demonstrated significant decreases in L/W ratio for animals exposed to the entire ambient mixture
compared to controls, indicating thicker walls in these vessels. There was no difference reported for
the L/W ratio in renal arteries. Morphologic examination suggested that the increases in L/W ratio
were due to muscular hypertrophy rather than fibrosis. The results of this study indicate vascular
remodeling of the pulmonary and coronary arteries, as opposed to changes in tone.
To examine the role of systemic inflammation and recruitment of monocytes into plaque tissue
as a possible pathway for accelerated atherosclerosis, Yatera et al. (2008, 157162) exposed female
Watanabe heritable hyperlipidemic rabbits (42 week old) to Ottawa PM10 (EHC-93) via IT
instillation (5 mg/rabbit; approximately 1.56 mg/kg) twice a week for 4 wk. Transfusion of whole
blood harvested to from exposed and non-exposed animals to donor rabbits supplied labeled
monocytes for assessment of monocyte recruitment from the blood to the aortic wall. The fraction of
aortic surface and volume of aortic wall taken up by atherosclerotic plaque was increased and the
number of labeled monocytes in the atherosclerotic plaques was elevated in rabbits exposed to PMi0.
In addition, labeled monocytes were attached onto the endothelium overlying atherosclerotic plaques
and the number that migrated into the smooth muscle underneath plaques in aortic vessel walls was
greater with PMi0 exposure compared to control. These responses were not observed in normal
vessel walls. ICAM-1 and VCAM-1 expression was elevated in atherosclerotic lesions, likely
indicating enhanced monocyte adhesion to endothelium and migration into plaques. Monocytes in
plaque tissue stained with immunogold demonstrated foam cell characteristics, which were more
numerous in the rabbits exposed to PMi0.
Gasoline Exhaust
Lund and colleagues (2007, 125741) used whole emissions from gasoline exhaust to
investigate changes in the transcriptional regulation of several gene products with known roles in
both the chronic promotion and acute degradation/destabilization of atheromatous plaques. These
50-day exposures (6 h/day><7 days/wk) employed ApoE~'~ mice on high-fat chow and the
concentrations of the high exposure group were 61 ug/m3 for PM, 19 ppm for NOX, 80 ppm for CO,
and 12.0 ppm for total hydrocarbons. The average particle number median diameter was
approximately 15 nm (McDonald et al., 2007, 156746). Dilutions of gasoline engine emissions
induced a concentration-dependent increase in transcription of matrix metalloproteinase (MMP)
isoform 9, ET-1, and HO-1 in aortas; MMP-3 and -9 mRNA levels were only increased in animals in
the highest exposure group. Strong increases in oxidative stress markers (nitrotyrosine and TEARS)
in the aortas were also observed. However, using a high-efficiency particle trap, they established that
most of the effects were caused by the gaseous portion of the emissions and not the particles. This
study did not directly address lesion area.
7.2.2. Venous Thromboembolism
One epidemiologic study examined the relationship between long term PMi0 concentration,
venous thromboembolism, and laboratory measures of hemostasis (prothrombin and activated partial
thomboplastin times [PT; PTT]). PT and PTT measure the extrinsic and intrinsic blood coagulation
pathways, the former activated in response to blood vessel injury, the latter, key to subsequent
amplification of the coagulation cascade and propagation of thrombus (Mackman et al., 2007,
156723). Decreases in PT and PTT are consistent with a hypercoagulable, prothrombotic state.
December 2009 7-6
-------
7.2.2.1. Epidemiologic Studies
Baccarelli et al. (2008, 157984) studied 2,081 residents (56% female) of the Lombardy region
of Italy whose ages ranged from 18 to 84 yr old. In this case-control study of 871 patients with
ultrasonographically or venographically diagnosed lower extremity deep vein thrombosis (DVT) and
1,210 of their healthy friends or relatives (1995-2005), the authors used arithmetic averaging of PMi0
data available at 53 monitors in nine geographic areas to estimate 1-yr avg residence-specific
exposures. They found -0.06 (95% CI: -0.11 to 0) and -0.12 (95% CI: -0.23 to 0) decreases in
standardized correlation coefficients for PT as well as 0.01 (95% CI: -0.03 to 0.04) and -0.09 (95%
CI: -0.19 to 0.01) decreases in standardized correlation coefficients for PTT among cases and
controls, respectively, per 10 ug/m3 increase in PM10. Patients with DVT who were taking heparin or
coumarin anticoagulants were not asked to stop taking them before measurement of PT and aPTT. Of
additional note, PT was neither adjusted for differences in reagents used to determine it nor
conventionally reported as the International Normalized Ratio (INR). The ambient PM
concentrations from this study are characterized in Table 7-1.
7.2.3. Metabolic Syndromes
7.2.3.1. Epidemiologic Studies
Chen and Schwartz (2008, 190106) studied 2,978 residentially stable participants in 33 U.S.
communities (age range = 20-89 yr; 49% female) who were examined during phase 1 of the National
Health and Nutrition Examination Survey III (1989-1991). In this cross-sectional study, the authors
used inverse-distance weighted averaging of U.S. EPA AQS monitored data from participant and
adjacent counties of residence to estimate 1-yr avg exposures to PMi0. They found that after
adjustment, residents of communities with lower PMi0 concentrations had fewer white blood cells
than residents of communities with higher PMi0 concentrations. This difference increased with
increasing number of metabolic abnormalities (insulin resistance; hypertension;
hypertriglyceridemia; low high-density lipoprotein cholesterol; abdominal obesity) reported by the
participant. This observed difference across individuals with different degrees of metabolic
abnormalities supports the concept that the presence of a metabolic syndrome may impart greater
susceptibility to PM-associated long-term CVD effects.
7.2.3.2. lexicological Studies
Diabetics as a potentially susceptible subpopulation have only recently been evaluated. A
toxicological study of a diet-induced obesity mouse model (C57BL/6 fed high-fat chow for 10 wk)
examined the effects of a 128-day PM2.5 CAPs exposure (mean mass concentration 72.7 ug/m3;
Tuxedo, NY) on insulin resistance, adipose inflammation, and visceral adiposity (Sun et al., 2009,
190487). Elevated fasting glucose and insulin levels were observed in CAPs-exposed mice compared
to air-exposed during the glucose tolerance test. Aortic rings of mice exposed to CAPs demonstrated
decreased peak relaxation to ACh or insulin, which was associated with reduced NO bioavailability.
Additionally, insulin signaling was impaired in aortic tissue via lowered endothelial Akt
phosphorylation. Increases in adipokines and systemic inflammatory markers (i.e., TNF-a, IL-6,
E-selectin, ICAM-1, PAI-1, resistin, leptin) were reported for CAPs-exposed mice. CAPs resulted in
increased visceral and mesenteric fat mass, as well as greater adipose tissue macrophages in
epididymal fat pads and larger adipocyte size compared to mice in the filtered air group. The results
of this study demonstrate that PM2.5 exposure can exaggerate insulin resistance, visceral adiposity,
and inflammation in mice fed high-fat chow.
December 2009 7-7
-------
7.2.4. Systemic Inflammation, Immune Function, and Blood Coagulation
7.2.4.1. Epidemiologic Studies
As discussed in Section 7.2.3.1, Chen and Schwartz (2008, 190106) conducted a cross-
sectional study in 33 U.S. communities and used inverse-distance weighted averaging of U.S. EPA
AQS monitored data from participant and adjacent counties of residence to estimate 1-yr avg
exposures to PMi0 (median concentration within quartiles = 23.1, 31.2, 38.8 and 53.7 ug/m3). They
found that after adjustment, residents of communities in quartile 1 had 138 (95% CI: 2-273) fewer
white blood cells (xl06/L) than residents of communities in quartiles 2-4. This difference increased
with increasing number of metabolic abnormalities.
Forbes et al. (2009, 190351) studied approximately 25,000 adults (age > 16 yr; 53% female)
who were representatively sampled from 720 English postcode sectors and participated in the Health
Survey for England (1994, 1998 and 2003). In this fixed-effects meta-analysis of year-specific cross-
sectional findings, the authors used dispersion modeling of emissions and weather data to estimate
2-yr avg exposures to PMi0 at participant postcode sector centroids (median in 1994, 1998 and 2003
= 19.5, 17.9 and 16.2 ug/m3, respectively). They found little evidence of a PM10-inflammatory
marker association, i.e., only a -0.08% (95% CI: -0.25 to 0.10) decrease in fibrinogen concentration
and a 0.14% (95% CI: -1.00 to 1.30) increase in CRP concentration per 1 ug/m3 increase in PMi0.
Calderon-Garciduenas et al. (2007, 091252) compared residentially stable, non-smoking
healthy children (age range: 6-13 yr) living and attending school between 2003-2004 in Mexico City
(historically high PM; altitude 2,250 m) and Polotitlan (historically low PM; altitude 2,380 m). In
this ecologic study, residents of Mexico City (n = 59; 93% female) had fewer white blood cells and
neutrophils (xlOvL) than residents of Polotitlan (n = 22; 69% female): unadjusted mean 6.2 (95%
CI: 5.7-6.6) versus 6.9 (95% CI: 6.3-7.5) and 2.9 (95% CI: 2.3-3.5) versus 3.8 (95% CI: 3.2-4.4),
respectively.
Calderon-Garciduenas et al. (2009, 192107) subsequently compared 37 unadjusted mean
measures of immune function and inflammation among an expanded number of these participants.
They found that under a two-sided type I error rate (a) = 0.05, 16 (43%) of the measures were
significantly different in residents of southwest Mexico City (n = 66; 48% female) than those in
Polotitlan (n = 93; 57% female). However, only 8 measures were significantly different after Bon
Ferroni-correction (a = 0.05 / 37 = 0.001) and even fewer would be after adjustment for reported
correlation between the measures of immune function and inflammation, e.g., CRP and
lipopolysaccharide binding protein (Pearson's r = 0.71).
Two cross-sectional analyses of PMi0 concentration and markers of immune function or
inflammation have been conducted with significant changes observed in the NHANES population
(stronger effects among those with metabolic disorders) (Chen and Schwartz, 2008, 190106) but not
in a relative large survey of adults, which was conducted in England (Forbes et al., 2009, 190351).
Ecological analyses comparing children in high versus low pollution regions in Mexico show
differences in unadjusted blood markers that may be related to PM concentration or other
unmeasured risk factors that differs across the communities studied (Calderon-Garciduenas et al.,
2007, 091252; Calderon-Garciduenas et al., 2009, 192107).
7.2.4.2. lexicological Studies
In addition to the PM2.5 study mentioned previously that showed increased TF expression (an
important initiator of thrombosis) in aortas of ApoE"7" mice following subchronic CAPs exposure
(Sun et al., 2008, 157033). three recent studies examined hematology and clotting parameters in rats
and mice exposed to DE, gasoline exhaust, or hardwood smoke for 1 week or 6 mo (Reed et al.,
2004, 055625: Reed et al., 2006, 156043: Reed et al., 2008, 156903). In all studies, male and female
F344 rats were exposed to the mixtures by whole-body inhalation for 6 h/day, 7 day/wk. Respiratory
effects for these studies are presented in Section 7.3.3.
December 2009 7-8
-------
Diesel Exhaust
The target PM concentrations in the DE study was 30, 100, 300, and 1,000 ug/m3 and the
MMAD was 0.10-0.15 um (Reed et al., 2004, 055625). Male and female rats exposed to DE at the
highest concentration (NO concentration 45.3 ppm; NO2 concentration 4.0 ppm; CO concentration
29.8 ppm; SO2 concentration 365 ppb) for 6 mo demonstrated decreased serum Factor VII, but no
change in plasma fibrinogen or thrombin anti-thrombin complex (TAT) (Reed et al., 2004, 055625).
White blood cells were decreased only in female rats in the highest exposure group. Another DE
study of shorter duration (4 wk, 4 h/day, 5day/wk; PM mass concentration 507 or 2201 ug/m3, CO
1.3 and 4.8 ppm, NO <2.5 and 5.9 ppm, NO2 O.25 and 1.2 ppm, SO2 0.2 and 0.3 ppm for low and
high PM exposures, respectively) did not demonstrate changes in hematologic parameters or those
related to coagulation (i.e., PT, PPT, plasma fibrinogen, D-dimer) or inflammation (i.e., CRP) in SH
or WKY rats (Gottipolu et al., 2009, 190360). Together, these findings do not support a DE-related
stimulation of blood coagulation following 1 or 6 mo of exposure.
Hardwood Smoke
The target PM concentrations in the hardwood smoke study was 30, 100, 300, and 1,000 ug/m3
and the MMAD was 0.25-0.36 um (Reed et al., 2006, 156043). In male rats exposed to hardwood
smoke, the mid-low group (PM concentration 113 ug/m3; NO, NO2, SO2 concentrations 0 ppm; CO
concentration 1,832.3 ppm) had the greatest responses in hematology parameters, including
increased hematocrit, hemoglobin, lymphocytes, and decreased segmented neutrophils (Reed et al.,
2006, 156043). Platelets were elevated in male and female rats after 1 week of exposure, but this
response returned to control values following the 6-month exposure. No changes were observed for
any coagulation markers at 6 mo.
Gasoline Exhaust
PM mass in the gasoline exhaust study ranged from 6.6 to 59.1 ug/m3, with the corresponding
number concentration between 2.6><104 and 5.0><105 particles/cm3; the dilutions for the gasoline
exhaust were 1:10, 1:15 or 1:90 and filtered PM at the 1:10 dilution (Reed et al., 2008, 156903).
Similar to the responses observed with hardwood smoke, male and female rats in the mid- and high-
gasoline exhaust exposure groups (NO concentrations 11.9 and 18.4 ppm; NO2 concentrations 0.5
and 0.9 ppm; CO concentration 73.2 and 107.3 ppm; SO2 concentration 0.38 and 0.62 ppm,
respectively) demonstrated elevated hematocrit and hemoglobin; RBC count was also elevated in
these groups (Reed et al., 2008, 156903). The only response that appeared somewhat dependent on
the presence of particles was increased RBC in female rats at 6 mo, although the authors attributed
the observed increases to the high concentration of CO.
Collectively, these studies do not indicate robust systemic inflammation or coagulation
responses in F344 rats following 6-month exposures to diesel, hardwood smoke, or gasoline exhaust.
The limited effects that were observed could possibly be due to the varying gas concentrations in the
exposure mixtures.
7.2.5. Renal and Vascular Function
Two recent epidemiologic studies have tested associations between PM exposure and
indicators of renal and vascular function (urinary albumin to creatinine ratio [UACR] and blood
pressure). UACR is a measure of urinary albumin excretion (National Kidney Foundation, 2008,
156796). When calculated as the ratio of albumin to creatinine concentrations in untimed ("spot")
urine samples, UACR approximates 24-h urinary albumin excretion and can be used to identify
albuminuria, a marker of generalized vascular endothelial damage (Xu et al., 2008, 157157). Values
> 30 mg/g (3.5 mg/mmol) and > 300 mg/g (34 mg/mmol) usually define micro- and
macroalbuminuria, both of which are associated with increases in CVD incidence and mortality
(Bigazzi et al., 1998, 156272; Deckert et al., 1996, 156389; Dinneen and Gerstein, 1997, 156403;
Gerstein et al., 2001, 156466; Mogensen, 1984, 156769). Several researchers have called the
December 2009 7-9
-------
dichotomization of albuminuria into question, observing that there is no threshold below which risk
of cardiovascular and end-stage kidney disease disappears (Forman and Brenner, 2006, 156439;
Knight and Curhan, 2003, 179900; Ruggenenti and Remuzzi, 2006, 156933).
Systolic, diastolic, pulse, and mean arterial blood pressures (SBP; DBF; PP; MAP) in mmHg
have also been used as measures of cardiovascular disease. Franklin et al. (1997, 156446) suggested
that SBP and PP were the only two measures predictive of carotid stenosis in a multivariable analysis
considering all 4 measures, whereas Khattar et al. (2001, 155896) suggested that their prognostic
significance in hypertensive populations may differ by age, with SBP and PP being most predictive
among those > 60 yr and DBP among those <60 yr old (Khattar et al., 2001, 155896).
7.2.5.1. Epidemiologic Studies
O'Neill et al. (2007, 156006) examined the association of UACR with PM2.5 and PMi0 among
members of the MESA population described previously (Diez et al., 2008, 156401). For this study of
UACR, which included cross-sectional and longitudinal analyses, the study population was restricted
to a subset of 3,901 participants (mean age = 63 yr; 52% female) with complete covariate, outcome
and exposure data at their first through third exams (2000-2004). In cross-sectional analyses, the
authors found that after adjustment for demographic and clinical characteristics, 10 ug/m3 increases
in 20-yr imputed exposures to PM25 and PMi0 were associated with negligible 0.002 (95% CI:
-0.048 to 0.052) and -0.002 (95% CI: -0.038 to 0.035) mean differences in baseline log UACR,
respectively. Similarly, small statistically non-significant decreases in the prevalence of
microalbuminuria (defined in this setting as > 25 mg/g) provided little evidence of an effect on renal
function. These largely null cross-sectional findings mirrored those based on the study's shorter-term
(30- and 60-day) PM2.5 and PM10 exposures. Moreover, longitudinal analyses revealed only a weak
association between 3-yr change in log UACR and 20-yr PMi0 exposure. Evidence of effect
modification by demographic and geographic characteristics was not apparent in either the cross-
sectional or longitudinal analyses.
Auchincloss et al. (2008, 156234) focused on automated, oscillometric, sphygmomanometric
measures of blood pressures in mmHg (SBP; DBP; PP; MAP). Like O'Neill (2007, 156006). Diez et
al. (2008, 156401) and Allen et al. (2007, 156006). Auchincloss et al. (2008, 156234) based their
examination on the previously described MESA population. The authors included 5,112 study
participants (age range = 45-84 yr; 52% female) who were free of clinically manifested CVD at their
baseline exam in one of six primarily urban U.S. locations (2000-2002). In this cross-sectional study,
they used arithmetic averaging of EPA AQS PM2.5 data available at the monitor nearest to each
participant's geocoded U.S. Postal Service ZIP code centroid to estimate 30- and 60-day avg
exposures to PM25. They found small nonsignificant increases of 1.5 (95% CI: -0.2 to 3.2), 0.2
(95% CI: -0.7 to 1.0), 1.3 (95% CI: 0.1 to 2.6), and 0.6 (95% CI: -0.4 to 1.7) mmHg increases in
SBP, DBP, PP and MAP, respectively, per 10 (ig/m3 increase in 30-day avg PM25 exposure,
Associations were slightly weaker for 60-day avg PM2 5 exposure and among participants without
hypertension, during cooler weather, in the presence of low NO2, residing >300 m from a highway,
or surrounded by lower road density.
Finally, the Calderon-Garciduenas et al. (2007, 091252) ecologic study introduced in
Section 7.2.3.1 also found that children residing in Mexico City had higher mean pulmonary artery
pressure as assessed by Doppler echocardiography and fasting plasma endothelin-1 (ET-1) than
residents in Polotitlan: unadjusted mean 17.5 (95% CI: 15.7-19.4) versus 14.6 (95% CI: 13.8-15.4)
mmHg and 2.23 (95% CI: 1.93-2.53) versus 1.23 (95% CI: 1.11-1.35) pg/mL, respectively. Within
Mexico City, ET-1 was higher in residents of the Northeast (historically higher PM25) than those of
the Southwest (historically lower PM25).
The MESA analyses of UACR (O'Neill et al., 2007, 156006) and the ecologic study of
children living in a highly polluted area of Mexico (Calderon-Garciduenas et al., 2007, 091252)
provide little evidence that long-term exposure to PM2 5 had an effect on renal and vascular function,
respectively. Auchincloss et al. (2008, 156234) reports small nonsignificant associations of blood
pressure with 30- and 60-day avg PM2 5 concentrations. PM concentrations from the analyses are
characterized in Table 7-1.
December 2009 7-10
-------
Table 7-1. Characterization of ambient PM concentrations from studies of subclinical measures of
cardiovascular diseases and long-term exposure.
Study
Location
Mean Concentration (ug/m
MESA: Multi-Ethnic Study of Atherosclerosis
HNRS: Heinz Nixdorf Recall Study
VEAPS: Vitamin E Atherosclerosis Progression Study
BVAIT: B-Vitamin Atherosclerosis Intervention Trial
Upper Percentile
Concentrations (ug/m3
PM10
Diez Roux et al. (2008, 1564011
O'Neill et al. (2007, 1560061
Baccarelli et al. (2008, 157984)
Rosenlund et al. (2006, 0897961
Chen and Swartz (2008, 190106)
Forbes et al. (2009, 1903511
MESA: 6 Cities U.S.
MESA: 6 Cities U.S.
Lombardy Region Italy
Stockholm, Sweden
US Population (NHANES)
British Population
20 yr imputed mean: 34
Long-Term Exposure:
1982-2002: 34.7
1982-1987: 40.5
1988-1992: 38
1993-1997: 30.6
1998-2002: 29.7
Previous Month: 27.5
NR
30-yavgPMio (traffic)
Cases: 2.6
Controls: 2.4
Annual avg: 36.8
1994: 19.5 (median)
1998: 17.9 (median)
2003: 16.2 (median)
NR
NR
NR
5th-95th %: 0.5-6
0.6-5.9
NR
1994, Min-Max: 12.5-36.1
1998, Min-Max: 12.6-27.0
2003, Min-Max: 11. 0-22.7
PM2.5
Hoffmann etal.(2007, 091163)
Allen et al. (2009, 1562091
Kunzli et al. (2005, 0873871
Auchincloss et al. (2008, 1562341
O'Neill (2007, 1560061
Diez Roux et al. (2008, 156401)
Hoffmann et al. (2009, 190376)
Calderon-Garciduenas et al. (2009,
1921071
Calderon-Garciduenas et al. (2007,
0912521
HNRS, 3 Cities Germany
MESA: 5 Cities
VEAPS BVAIT
MESA: 6 Cities
MESA: 6 Cities U.S.
MESA: 6 Cities U.S.
HNRS: 3 Cities Germany
Southwest Mexico (high pollution)
Potitlan (low pollution)
Southwest Mexico (high pollution)
Potitlan (low pollution)
Annual avg: 22.8
Annual avg: 15.8
Annual avg: 20.3
Prior 30 days: 16.8
Prior 60 days: 16.7
Previous Month: 16.5
20-y imputed mean: 21.7
Annual avg: 22.8
Annual avg: 25
Annual avg: <15
NR
NR
NR
Min-Max: 10.6-24.7
Min-Max: 5.2-26.9
NR
NR
NR
Min-max: 19.8-26.8
NR
NR
NR
NR
7.2.5.2. Toxicological Studies
In a PM2.5 CAPs study of 10 wk (6 h/day><5 days/wk) in Tuxedo, NY (mean mass
concentration 79.1 ug/m3), there was no difference in mean arterial pressure (MAP) in SD rats
between groups (Sun et al., 2008, 157032). When angiotensin II (Ang II) was infused during the last
week of exposure to induce systemic hypertension, the MAP slope was consistently greater in the
CAPs-exposed rats compared to the filtered air group. Furthermore, thoracic aortic rings were more
responsive to phenylephrine-induced constriction and less responsive to ACh-induced relaxation in
the PM+Ang II vessels. In contrast to the latter findings, the relaxation response was exaggerated in
the PM+Ang II aortic segments with a Rho-kinase (ROCK) inhibitor. Superoxide production in
aortic rings increased in the PM+Ang II group compared to the filtered air group and the addition of
December 2009
7-11
-------
NAD(P)H oxidase inhibitor (apocynin) or aNOS inhibitor (L-NAME) attenuated the superoxide
generation. The levels of tetrahydrobiopterin (BH4) were decreased in mesenteric vasculature and the
heart by 46% and 41% in the PM+Ang II group compared to controls, respectively; furthermore,
levels of BH4 in the liver were similarly reduced, which is consistent with a systemic effect of CAPs.
Together, these findings indicate that CAPs potentiate Ang II-induced hypertension and alter
vascular reactivity, perhaps through activated NADPH oxidase and eNOS uncoupling that result in
oxidative stress generation and triggering of the Rho/ROCK signaling pathway.
7.2.6. Autonomic Function
7.2.6.1. lexicological Studies
Hwang et al. (2005, 087957) and Chen and Hwang (2005, 087218) used radiotelemetry to
examine the chronic changes in HR and HRV resulting from the same CAPs exposures described
previously (Chen and Nadziejko, 2005, 087219). The overall average CAPs exposure concentration
was 133 ug/m3 and results indicate differing responses to CAPs between ApoE"" mice and their
genetic background strain, C57BL/6J mice (Hwang et al., 2005, 087957). Using the time period of
1:30-4:30 a.m., C57BL/6J mice showed a HR increase only over the last month of exposure. In
contrast, ApoE"7" mice had chronic decreases of 33.8 beat/min for HR. Changes in HRV (SDNN and
rMSSD) were somewhat more complicated, with biphasic responses in ApoE"7" mice over the
5-month period (initial increase over first 6 wk, decrease over next 12 wk, and slight upward turn for
remainder of the study)(Chen and Hwang, 2005, 087218). Increasing linear trends were observed in
C57BL/6J mice for SDNN and rMSSD. The average CAPs concentration for the HRV study was
110 ug/m3. However, only three C57BL/6J mice in the exposure group were included in the analysis
compared to ten ApoE"" animals, thus making it difficult to interpret the C57BL/6J mice responses
(Chen and Hwang, 2005, 087218; Hwang et al., 2005, 087957).
7.2.7. Cardiac changes
7.2.7.1. Toxicological studies
Two recent toxicological studies have evaluated the effects of PM on cardiac effects including
pathology and gene expression. Cardiac mitochondrial function has also been evaluated following
PM exposure in rats.
Diesel Exhaust
A recent study of DE exposure (PM mass concentration 507 or 2,201 ug/m3, CO 1.3 or
4.8 ppm, NO <2.5 or 5.9 ppm, NO2 O.25 or 1.2 ppm, SO2 0.2 or 0.3 ppm for low and high PM
exposures, respectively; geometric median number diameter 85 nm) indicated a hypertensive-like
cardiac gene expression in WKY rats that mimicked baseline patterns in air-exposed SH rats
(Gottipolu et al., 2009, 190360). Exposure to the high concentration of DE for 4 wk (4 h/day,
5 day/wk) led to downregulation of genes involved in stress, antioxidant compensatory response,
growth and extracellular matrix regulation, membrane transport of molecules, mitochondrial
function, thrombosis regulation, and immune function. No genes were affected by DE in SH rats. A
dose-dependent inhibition of mitochondrial aconitase activity in both rat strains was observed,
indicating a DE effect on oxidative stress. It should be noted that while DE-related cardiovascular
effects were found in WKY rats only, pulmonary inflammation and injury were observed in both
strains (Sections 7.3.3.2 and 7.3.5.1).
December 2009 7-12
-------
Model Particles
Wallenborn et al. (2008, 191171) examined the subchronic (5 h/day, 3 day/wk, 16 wk)
pulmonary, cardiac, and systemic effects of nose-only exposure to particulate ZnSO4 (9, 35, or
120 ug/m ) in WKY rats. Particle size was reported to be 31-44 nm measured as number median
diameter. Although changes in pulmonary inflammation or injury and cardiac pathology were not
observed, effects on cardiac mitochondrial protein and enzyme levels were noted (i.e., increased
ferritin levels, decrease in succinate dehydrogenase activity), possibly indicating a small degree of
mitochondrial dysfunction. Glutathione peroxidase, an antioxidant enzyme, was also decreased in
the cardiac cytosol. Gene expression analysis identified alterations in cardiac genes involved in cell
signaling events, ion channels regulation, and coagulation in animals exposed to the highest ZnSO4
concentration only. This study demonstrates a possible direct effect of ZnSO4 on extrapulmonary
systems, as suggested by the lack of pulmonary effects (Section 7.3.3.2).
7.2.8. Left Ventricular Mass and Function
Van Hee et al. (2009, 192110) studied 3,827 participants (age range = 45-84 yr; 53% female)
who underwent magnetic resonance imaging (MRI) of the heart at the baseline examination of the
MESA cohort (2000-2002). This cross-sectional study focused on two MRI-based outcome
measures: left ventricular mass index (LVMI, g/m2) and ejection fraction (EF, %), the former
estimated using the DuBois formula for body surface area, the latter as the ratio of stroke volume to
end diastolic volume. The study also estimated annual mean exposures to PM2.5 at participants'
geocoded residential addresses in 2000 using ordinary kriging of U.S. EPA AQS concentration data.
In fully adjusted models, it found 3.8 (95% CI: -6.1 to 13.7) g/m2 and -3.0% (-8.0 to 2.0) differences
in LVMI and EF per 10 (ig/m3 increment in PM2.5. The findings were small and imprecise, albeit
suggestive of a slight, PM-associated increase in the mass and decrease in the function of the left
ventricle. The effect of living within 50 m of a major roadway on LVMI was greater than the effect
of PM2.5 (i.e., 1.4 g/m2 [95% CI: 0.3-2.5] per 10 ug/m3.)
7.2.9. Clinical Outcomes in Epidemiologic Studies
Several epidemiologic studies of U.S. and European populations have examined associations
between long-term PM exposures and clinical CVD events (Baccarelli et al., 2008, 157984;
Hoffmann et al., 2006, 091162: Hoffmann et al., 2009, 190376: Maheswaran et al., 2005, 088683:
Maheswaran et al., 2005, 090769: Miller et al., 2007, 090130: Rosenlund et al., 2006, 089796:
Solomon et al., 2003, 156994: Zanobetti and Schwartz, 2007, 091247). Results from these studies
are summarized in Figure 7-1. The ambient PM concentrations from these studies are characterized
in Table 7-2.
Coronary Heart Disease
Epidemiologic studies examining the association of coronary heart disease (CHD) with long-
term PM exposure are discussed below (Hoffmann et al., 2006, 091162: Maheswaran et al., 2005,
090769: Miller et al., 2007, 090130: Puett et al., 2008, 156891: Rosenlund et al., 2006, 089796:
Rosenlund et al., 2009, 190309: Zanobetti and Schwartz, 2007, 091247). Cases of CHD were
variably defined in these studies to include history of angina pectoris, MI, coronary artery
revascularization (bypass graft; angioplasty; stent; atherectomy), and congestive heart failure (CHF).
Results pertaining to death from CHD are described in Section 7.6.
Miller et al. (2007, 090130) studied incident, validated MI, revascularization, and CHD death,
both separately and collectively, among 58,610 post-menopausal female residents of 36 U.S.
metropolitan areas (age range = 50-79 yr) enrolled in the Women's Health Initiative Observational
Study (WHI OS, 1994-1998). In this prospective cohort study of participants free of CVD at baseline
(median duration of follow-up = 6 yr), the authors used arithmetic averaging of year 2000 EPA AQS
PM2.5 data available at the monitor nearest to each participant's geocoded U.S. Postal Service five-
digit ZIP code centroid to estimate 1-yr avg exposures. They found 6% (95% CI: -15 to 34), 20%
(95% CI: 0-43) and 21% (95% CI: 4-42) increases in the overall risk of MI, revascularization, and
December 2009 7-13
-------
their combination with CHD death per 10 ug/m3 increase in PM2.5, respectively. Hazards were higher
within than between cities and in the obese. For the combined CVD outcome (MI, revascularization,
stroke, CHD death, cerebrovascular disease), authors reported a 24% (95% CI: 9-41) increase in risk
that was higher among participants at higher than lower quintiles of body mass index, waist-to-hip
ratio, and waist circumference. The PM2.5-CVD association was stronger among non-diabetic than
diabetic participants.
Table 7-2. Characterization of ambient PM concentrations from studies of clinical cardiovascular
diseases and long-term exposure.
Study
Location
Mean Annual
Concentration (ug/m )
Upper Percentile
Concentrations (ug/m
PM10
Puett et al. (2008, 156891)
Zanobetti and Schwartz (2007, 091247)
Rosenlund et al. (2006, 0897961
Rosenlund et al. (2009, 1903091
Maheswaran et al. (2005, 0907691
Baccarelli et al. (2008, 1579841
13 U.S. States
21 U.S. Cities
Stockholm, Sweden
Stockholm, Sweden
Sheffield, U.K.
Lombardia Region, Italy
21.6
28.8
30 yavgPM10 (traffic)
Cases: 2.6
Controls: 2.4
5-yravg PM10 from traffic:
Cases: 2.4 (median)
Controls: 2.2 (median)
Range of means in each quintile: 16-23.3
NR
Overall range NR
5th-95th Percentile
0.5-6.0
0.6-5.9
NR
NR
PM2.5
Miller etal. (2007, 090130)
Hoffmann et al. (2006, 0911621
Hoffman et al. (2009, 1903761
WHI: 36 Metropolitan areas
HNRS: 2 Cities Germany
HRNS: 2 Cities German
Citywideavg(yr2000): 13.5
23.3
22.8
Min-max: 4-19.3
NR
NR
WHI: Womens Health Initiative
HNRS: Hans Nixdorf Recall Study
Puett et al. (2008, 156891) studied incident, validated CHD, CHD death, and non-fatal MI
among 66,250 female residents (mean age = 62 yr) of metropolitan statistical areas in thirteen
northeastern U.S. states who were enrolled in the Nurses' Health Study (NHS, 1992-2002). In this
prospective cohort study of women without a history of non-fatal MI at baseline (maximum duration
of follow-up = 4 yr), the authors used two-stage, spatially smoothed, land use regression to estimate
residence-specific, 1-yr ma PMi0 exposures from U.S. EPA AQS and emissions, IMPROVE, and
Harvard University monitor data. They found a 10% (95% CI: -6 to 29) increase in risk of first CHD
event per 10 u/m3 increase in 1-yr avg PMi0 exposure, while the association with MI was close to the
null value. The association with fatal CHD event of 30% (95% CI: 0-71) was stronger. Furthermore,
associations with CHD death were higher in the obese and in the never smokers.
Rosenlund et al. (2006, 089796) studied 2,938 residents of Stockholm County, Sweden (age
range = 45-70 yr; 34% female). In this case-control study of 1,085 patients with their first, validated
non-fatal MI and an age-, gender- and catchment-stratified random sample of 1,853 controls without
MI (1992-1994), the authors used street canyon-adjusted dispersion modeling of emissions data to
estimate 30-yr avg exposure to PMi0 (median = 2.4 ug/m3). They found that the OR for prevalent MI
per 10 ug/iri increase in PM10 was 0.85 (95% CI: 0.50-1.42). The OR for fatal MI was elevated, but
not statistically significant.
In a more recent study, Rosenlund et al. (2009, 190309) evaluated 554,340 residents (age
range = 15-79 yr; 49% female) of Stockholm County, Sweden (1984-1996). In this population-based,
case-control study of 43,275 cases of incident, validated MI, the authors used dispersion modeling of
traffic emissions and land use data to estimate 5-yr avg exposure to PMi0. They found that after
December 2009
7-14
-------
adjustment for demographic, temporal, and socioeconomic characteristics, the OR for MI per
5 ug/m3 increase in PMi0 was 1.04 (95% CI: 1.00-1.09). ORs were higher after restriction to fatal
cases, in- or out-of-hospital deaths, and participants who did not move between population censuses.
Authors state that control for confounding was superior in their previous study (Rosenlund et al.,
2006, 089796) although the size of the population was larger in this recent study (Rosenlund et al.,
2009, 190309).
Zanobetti and Schwartz (2007, 091247) studied ICD-coded recurrent MI (ICD 9 410) and
post-infarction CHF (ICD 9 428) among 196,131 Medicare recipients (age > 65 yr; 50% female)
discharged alive following MI hospitalization in 21 cities from 12 U.S. states (1985-1999). In this
ecologic, open cohort study of re-hospitalization among MI survivors (mean duration of follow-
up = 3.6 and 3.7 yr for MI and CHF, respectively), the authors used arithmetic averaging of EPA
AQS PMio data available in the county of hospitalization to estimate 1-yr avg exposures. They found
17% (95% CI: 5-31) and 11% (95% CI: 3-21) increases in the risk of recurrent MI and post-
infarction CHF, respectively, per 10 ug/m3 increase in PMi0 exposure. Hazards were somewhat
higher among persons aged >75 yr.
Hoffmann et al. (2006, 091162) studied self-reported CHD (MI or revascularization) among
3,399 residents of Essen and Mulheim, Germany (age range = 45-75 yr; 51% female) at the baseline
exam of the Heinz Nixdorf Recall Study (2000-2003) introduced previously. In this cross-sectional
ancillary study, the authors used dispersion modeling of emissions, climate and topography data to
estimate 1-yr avg exposure to PM2.5 (mean = 23.3 ug/m3). They found little evidence of an
association between PM2.5 and CHD in these data. After adjustment for geographic, demographic
and clinical characteristics, the OR for prevalent CHD per 10 ug/m3 increase in exposure was 0.55
(95% CI: 0.14-2.11).
Study
Avg Time Endpoint
Effect Estimate (95% CI)
Prospective Cohort Studies
Miller etal.(2007,090130)
58,610 post-menopausal women enrolled in WHI,
incident, validated cases, 36 US cities
Puettetal. (2008.156891)
75,809 women in the Nurses Health Study,
validated cases, 13 metro areas, NE states
Case Control Studies
Rosenlund et al. (2009,190X9)
24,347 cases, N=276,926 randomly selected
population based controls, Stockholm, Sweden
Rosenlund et al. (2006,089796)
2,246 cases, N=3,206 randomly selected
population controls, Stockholm, Sweden
Baccarelli et al. (2008,157984)
871 cases, N=1210 healthy friend controls,
Lombardy, Italy
Other Study Designs
Hoffman etal. (2006.091162)
Hoffman et al. (2009,190376)
Prevalent cases at baseline in 3,399 residents,
cross-sectional, 2 German cities
Zanobetti & Schwartz (2007,091247)
196,131 Ml survivors (ICD9410), readmitted for
Ml or CHF, ecologic, open cohort, 21 US cities
Maheswaran et al. (2005,090769:2005,088683)
Hospital admissions among 199,682 residents,
ecologic design, Sheffield, UK
1 yr Incident Ml
Revascularization
Stroke
Cerebrovascular Disease
All CVD
1yr 1st CHD Event
Syr Fatal/Non-Fatal Ml
30 yr Fatal/Non-Fatal Ml
1 yr DVT
1 yr Self-reported CHD
1 yr Self-reported PVD
1 yr Recurrent Ml Hospitalization
Post-Mi CHF Hospitalization
5 yr CHD Hospitalization
Stroke Hospitalization
PM,
PM,C
PM,,
PMZ
PM,,
I I i \ I I I I I I I I I I I I I \ \ I I I I I I
0,0 0,2 0.4 0,6 0.8 10 1.2 1.4 1,6 1.8 2,0 2.2 2.4
Relative Risk Estimate
Figure 7-1. Risk estimates for the associations of clinical outcomes with long-term exposure
to ambient PM2.s and PMio.
December 2009
7-15
-------
In the study of 1,030 census enumeration districts in Sheffield, U.K. described previously,
Maheswaran et al. (2005, 090769) studied 11,407 ICD-10-coded emergency hospitalizations for
CHD (ICD10 120-25) among 199,682 residents (age > 45 yr; 45% female). In this ecologic study, the
authors used dispersion modeling of emissions and climate data to estimate 5-yr avg exposure to
PMio. They found that after adjusting for smoking prevalence, controlling for socioeconomic factors,
and smoothing, the age- and gender-standardized rate ratios for CHD admission were 1.01 (95% CI:
0.92-1.11), 1.04 (95% CI: 0.93-1.15), 0.97 (95% CI: 0.87-1.08), and 1.07 (95% CI: 0.95-1.20) across
PMio quintiles. The linear trend was somewhat stronger for CHD mortality (Section 7.3).
The study of post-menopausal women enrolled in the WHI OS by Miller et al. (2007, 090130)
was the only U.S. study to examine the effect of PM2.5 rather than PMi0. This study, which provides
strong evidence of an association, was distinguished by its prospective cohort design, validation of
incident cases and large population. Puett et al. (2008, 156891). the other U.S. study with
comparable design features, provides evidence of an association of incident CHD with long- term
PMio exposure. Findings from Swedish case control studies of incident validated cases of MI were
not consistent. A cross-sectional study of self-reported CHD did not provide evidence of an
association with PM25> while findings from two ecologic studies of PMio indicated positive
associations of CHD hospitalizations with PMio (Maheswaran et al., 2005, 088683; Zanobetti and
Schwartz, 2007, 091247).
Stroke
Miller et al. (2007, 090130) found 28% (95% CI: 2-61) and 35% (95% CI: 8-68) increases in
the overall risk of validated stroke and cerebrovascular disease, respectively, per 10 ug/m3 increase
in 1-yr avg PM2.5 exposure. Risks were higher within than between cities. In the study of 1030
Census of enumeration districts in Sheffield, U.K. described previously, Maheswaran et al. (2005,
088683) studied 5,122 ICD-10-coded emergency hospital admissions for stroke (160-69) among
199,682 residents (age > 45 yr; 45% female) of 1,030 census enumeration districts in Sheffield, U.K.
(1994-1999). In this ecologic study, the authors used dispersion modeling of emissions and climate
data to estimate 5-yr avg exposure to PMi0. They found that the age- and gender-standardized rate
ratios for stroke admission were 1.05 (95% CI: 0.94-1.17), 1.07 (95% CI: 0.95-1.20), 1.06 (95% CI:
0.94-1.20), and 1.15 (95% CI: 1.01-1.31) across PMio quintiles. Linear trend was somewhat stronger
for stroke mortality (Section 7.6).
These studies examining the long-term PM-stroke relationship provide evidence of
association. Maheswaran et al. (2005, 088683) examined emergency room hospital admissions in
Sheffield, U.K. using an ecologic design while results reported by Miller et al. (2007, 090130) are
based on the prospective cohort study of the WHI OS population (both introduced previously).
Peripheral Arterial Disease
The German Heinz Nixdorf Recall cross-sectional study described in Section 7.2.1.1
(Hoffmann et al., 2009, 190376) also evaluated the association between 1-yr avg exposure to PM2.5
and peripheral arterial disease (self-reported history of a surgical or procedural intervention or an
ABI <0.9 in one or both legs). The authors found no evidence of an increase in risk. The OR for
peripheral arterial disease was 0.87 (95% CI: 0.57-1.34) per 3.9 ug/m3 increase in PM25. However,
evidence of an association with traffic exposure was present in these data. ORs of 1.77 (95% CI:
1.01-3.10), 1.02 (95% CI: 0.58-1.80), and 1.07 (95% CI: 0.68-1.68) for residing < 50, 50-100, and
100-200 m of a major road (reference category: >200 m), respectively were observed. ORs were
higher among participants with CAC scores < 75th percentile, women, and smokers.
Deep Vein Thrombosis
The Italian case-control study (introduced in Section 7.2.1.2) also examined the chronic PM10-
DVT association (Baccarelli et al., 2008 157984). The authors found a 70% (95% CI: 30-223)
increase in the odds of DVT per 10 ug/m increase in 1-yr avg PMio exposure. This finding was
consistent with the decreases in PT and PTT also observed among controls in this context as well as
December 2009 7-16
-------
the 47% (95% CI: 11-96) increase in the odds of DVT per inter-decile range (242 m) increase in the
residence-to-major-roadway distance observed among a subset of cases and controls (Baccarelli et
al., 2009, 188183). The PMi0-DVT and distance-DVT associations were both weaker among women
and among users of oral contraceptives or hormone therapy.
7.2.10.Cardiovascular Mortality
New epidemiologic evidence reports a consistent association between long-term exposure to
PM2.5 and increased risk of cardiovascular mortality. There is little evidence for the long-term effects
of PMio_2.5 on cardiovascular mortality. This section focuses on cardiovascular mortality outcomes in
response to long-term exposure to PM. The studies that investigate long-term exposure and mortality
due to any specific or all (nonaccidental) causes are evaluated in Section 7.6. A summary of the
mean PM concentrations reported for the studies characterized in this section is presented in Table
7-8, and the effect estimates are presented in Figure 7-7 and Figure 7-8.
A number of large, U.S. cohort studies have found consistent associations between long-term
exposure to PM2.5 and cardiovascular mortality. The American Cancer Society (ACS) (Pope et al.
(2004, 055880) reported positive associations with deaths from specific cardiovascular diseases,
particularly ischemic heart disease, and a group of cardiac conditions including dysrhythmia, heart
failure and cardiac arrest (RR for cardiovascular mortality = 1.12 [95% CI: 1.08-1.15] per 10 ug/m3
PM25). In an additional reanalysis that extended the follow-up period for the ACS cohort to 18 yr
(1982-2000) (Krewski et al., 2009, 191193), investigators found effect estimates that were similar,
though generally higher, than those reported in previous ACS analyses.
A follow-up to the Harvard Six Cities study (Laden et al., 2006, 087605) used updated air
pollution and mortality data and found positive associations between long-term exposure to PM2.5
and mortality. Of special note is a statistically significant reduction in mortality risk reported with
reduced long-term fine particle concentrations. This reduced mortality risk was observed for deaths
due to cardiovascular and respiratory causes, but not for lung cancer deaths.
The WHI cohort study (Miller et al., 2007, 090130) (described previously) found that each
10 ug/m3 increase of PM2.5 was associated with a 76% increase in the risk of death from
cardiovascular disease (hazard ratio, 1.76 [95% CI: 1.25-2.47]). The WHI study not only confirms
the ACS and Six City Study associations with cardiovascular mortality in yet another well
characterized cohort with detailed individual-level information, it also has been able to consider the
individual medical records of the thousands of WHI subjects over the period of the study. This has
allowed the researchers to examine not only mortality, but also related morbidity in the form of heart
problems (cardiovascular events) experienced by the subjects during the study. These morbidity co-
associations with PM2 5 in the same population lend even greater support to the biological
plausibility of the air pollution-mortality associations found in this study.
In an analysis for the Seventh-Day Adventist cohort in California (AHSMOG), a positive
association with coronary heart disease mortality was reported among females (92 deaths; RR = 1.42
[95% CI: 1.06-1.90] per 10 ug/m3 PM2.5), but not among males (53 deaths; RR = 0.90 [95% CI:
0.76-1.05] per 10 ug/m3 PM25) (Chen et al., 2005, 087942). Associations were strongest in the subset
of postmenopausal women (80 deaths; RR = 1.49 [95% CI: 1.17-1.89] per 10 ug/m3 PM2.5). The
authors speculated that females may be more sensitive to air pollution-related effects, based on
differences between males and females in dosimetry and exposure. As was found with PM2 5, a
positive association with coronary heart disease mortality was reported for PMi0_2 5 and PMi0 among
females (RR = 1.38 [95% CI: 0.97-1.95] per 10 ug/m3 PM10.25;RR = 1.22 [95% CI: 1.01-1.47] per
10 ug/m3 PM10), but not for males (RR = 0.92 [95% CI: 0.66-1.29] per 10 ug/m3 PM10.2.5; RR = 0.94
[95% CI: 0.82-1.08] per 10 ug/m3 PMio); associations were strongest in the subset of
postmenopausal women (80 deaths) (Chen et al., 2005, 087942).
Two additional studies explored the effects of PM10 on cardiovascular mortality. The Nurses'
Health Study (Puett et al., 2008, 156891) is an ongoing prospective cohort study examining the
relation of chronic PMi0 exposures with all-cause mortality and incident and fatal coronary heart
disease consisting of 66,250 female nurses in MS As in the northeastern region of the U.S. The
association with fatal CHD occurred with the greatest magnitude when compared with other
specified causes of death (hazard ratio 1.42 [95% CI: 1.11-1.81]). The North Rhine-Westphalia State
Environment Agency (LUANRW) initiated a cohort of approximately 4,800 women, and assessed
whether long-term exposure to air pollution originating from motorized traffic and industrial sources
was associated with total and cause-specific mortality (Gehring et al., 2006, 089797). They found
December 2009 7-17
-------
that cardiopulmonary mortality was associated with PM10 (RR = 1.52 [95% CI: 1.09-2.15] per
10 ug/m3PMi0).
In summary, the 2004 PM AQCD concluded that there was strong evidence that long-term
exposure to PM2.5 was associated with increased cardiopulmonary mortality. Recent studies
investigating cardiovascular mortality provide some of the strongest evidence for a cardiovascular
effect of PM. A number of large cohort studies have been conducted throughout the U.S. and
reported consistent increases in cardiovascular mortality related to PM2.5 concentrations. The results
of two of these studies have been replicated in independent reanalyses. These effects are coherent
with short-term epidemiologic studies of CVD morbidity and mortality and with long-term
epidemiologic studies of CVD morbidity. In addition, biological plausibility and coherence are
provided by toxicological studies demonstrating short-term cardiovascular effects as well as
PM2.5-related plaque progression in chronically exposed mice.
7.2.11.Summary and Causal Determinations
7.2.11.1. PM2.5
Epidemiologic studies examining associations between long-term exposure to ambient PM
(over months to years) and CVD morbidity had not been conducted and thus were not included in the
1996 or 2004 PM AQCDs (U.S. EPA, 1996, 079380; U.S. EPA, 2006, 157071). A number of studies
were included in the 2004 AQCD that evaluated the effect of long-term PM2.5 exposure on
cardiovascular mortality and found strong and consistent associations. No toxicological studies had
evaluated the effects of subchronic or chronic PM exposure on CVD effects in the 2004 PM AQCD.
Recently, epidemiologic and toxicological studies have provided evidence of the adverse effects of
long-term exposure to PM2 5 on cardiovascular outcomes and endpoints, including atherosclerosis
and clinical and subclinical markers of cardiovascular morbidity.
The strongest evidence for a CVD health effect related to long-term PM2 5 exposure comes
from epidemiologic studies of cardiovascular mortality. A number of large, multicity U.S. studies
(the ACS, Six Cities Study, WHI, and AHSMOG) provide consistent evidence of an effect between
long-term exposure to PM25 and cardiovascular mortality (Section 7.2.10). These studies were
conducted in urban areas across the U.S. where mean concentrations ranged from 10.2-29.0 (ig/m3
(Table 7-8). An epidemiologic study investigating the relationship between PM2 5 and clinical CVD
morbidity among post-menopausal women (Miller et al., 2007, 090130) provides evidence of an
effect that is coherent with the cardiovascular mortality studies. This large, prospective cohort study
of incident, validated cases found large increases in the adjusted risk of MI, revascularization, and
stroke using a 1-yr avg PM25 concentration (mean = 13.5 ug/m3). A cross-sectional analyses of self-
reported prevalence of CHD and peripheral arterial disease found no such increase in the odds of
CVD morbidity (Hoffmann et al., 2006, 091162); the inconsistency of these findings with Miller et
al. (2007, 090130) may be explained by differences in study design or location.
The effect of long-term PM25 exposure on pre-clinical measures of atherosclerosis (CIMT,
CAC, AAC or ABI) has been studied in several populations using a cross-sectional study design. The
magnitude of the PM25 effects and their consistency across different measures of atherosclerosis in
these studies varies widely, and they may be limited in their ability to discern small changes in these
measures. Kunzli et al. (2005, 087387) observed anon-significant 4.2% increase in CIMT associated
with long-term PM2 5 exposure among participants of a clinical trial in greater Los Angeles, which
was several fold higher than the 0.5% increase observed by Diez-Roux et al. (2008, 156401) in their
analyses of MESA baseline data. The associations in MESA of CAC and ABI with long-term PM25
exposure were largely null (Diez et al., 2008, 156401), while an increase in AAC with long-term
PM2 5 exposure was reported (Chang et al., 2008, 180393). By contrast, a 43% increase in CAC was
associated with long-term PM2 5 exposure in a German study, but no similar association with ABI
was observed (Hoffmann et al., 2009, 190376). Although the number of studies examining these
relationships is limited, effect modification by use of lipid lowering drugs and smoking status was
reported in more than one study of long-term PM25 and PMi0 exposure.
Evidence of enhanced atherosclerosis development was demonstrated in new toxicological
studies that report increased plaque and lesion areas, lipid deposition, and TF in aortas of ApoE"7"
mice exposed to CAPs (Section7.2.1.2). In addition, alterations in vasoreactivity were observed,
December 2009 7-18
-------
suggesting an impaired NO pathway. Additional toxicological studies of PM10 are consistent with
these results. Further support is provided by a study that reported decreased LAV ratio in the
pulmonary and coronary arteries of mice exposed to ambient air. However, PM2.5 CAPs derived from
traffic in Los Angeles did not affect plaque size (Araujo et al., 2008, 156222). Collectively, these
toxicological studies provide biological plausibility for the associations reported in epidemiologic
studies.
There is limited evidence for the effects of PM2.5 on renal or vascular function. Cross-sectional
and longitudinal epidemiologic analyses of PM2.5 and UACR revealed no evidence of an effect
(O'Neill et al., 2007, 156006). while small non-statistically significant increases in BP with 30- and
60-day avg PM25 concentrations were reported (Auchincloss et al., 2008, 156234). A toxicological
study did not show changes in MAP with CAPs, but indicated a CAPs-related potentiation of
experimentally-induced hypertension (Sun et al., 2008, 157032). In addition, CAPs has induced
changes in insulin resistance, visceral adiposity, and inflammation in a diet-induced obesity mouse
model (Sun et al., 2009, 190487), indicating that diabetics may be a potentially susceptible
population to PM exposure.
In summary, a number of large U.S. cohort studies report associations of long-term PM25
concentration with cardiovascular mortality. These studies provide the strongest evidence for an
effect of long-term PM25 exposure on CVD effects. Additional evidence comes from a
methodologically rigorous epidemiology study that demonstrates coherent associations between
long-term PM2 5 exposure and CVD morbidity among post-menopausal women. Toxicological
studies demonstrate that this effect is biologically plausible and the effect is coherent with studies of
short-term PM2 5 exposure and CVD morbidity and mortality, and with long-term exposure to PM2 5
and CVD mortality. Associations between PM2 5 and subclinical measures of atherosclerosis are
inconsistent, but cross-sectional studies may be limited in their ability to discern small changes in
these measures. In addition, potential modification of the PM25-CVD association by smoking status
and the use of lipid lowering drugs has been demonstrated in epidemiologic studies that used
individual-level data. Toxicological studies provide evidence for accelerated development of
atherosclerosis in ApoE"" mice exposed to CAPs and show effects on coagulation factors,
experimentally-induced hypertension, and vascular reactivity. Available studies of clinical
cardiovascular disease outcomes report inconsistent results. Based on the above findings, the
epidemiologic and toxicological evidence is sufficient to infer a causal relationship between
long-term PM2.5 exposures and cardiovascular effects.
7.2.11.2. PM10.2.5
One epidemiologic study evaluated the relationship between long-term exposure to PMi0_2.5
and cardiovascular mortality and found a positive association with coronary heart disease mortality
among females, but not for males; associations were strongest in the subset of post-menopausal
women (Chen et al., 2005, 087942). No toxicological studies of long-term exposure to ambient
PM10_2.5 and cardiovascular effects have been conducted to date. Evidence is inadequate to infer
the presence or absence of a causal relationship.
7.2.11.3. UFPs
A few toxicological studies of long-term exposure to UFPs have been conducted. Increased
plaque size was reported in mice exposed to UF CAPs derived from traffic (Araujo et al., 2008,
156222). Studies of diesel and gasoline exhaust reported relatively few changes in hematologic or
coagulation parameters (Section 7.2.4.2) and one DE study demonstrated altered cardiac gene
expression in normotensive rats that reflected the development of hypertension (Gottipolu et al.,
2009, 190360). Whole and filtered gasoline exhaust induced increases in gene products involved in
atheromatous plaque formation and/or degradation, but these effects were largely due to the gaseous
emissions (Lund et al., 2007, 125741). Evidence from these studies alone is inadequate to infer
the presence or absence of a causal relationship
December 2009 7-19
-------
7.3. Respiratory Effects
Several cohort studies reviewed in the 2004 PM AQCD provided evidence for relationships
between long-term PM exposure and effects on the respiratory system, though it did not rule out the
possibility that the observed respiratory effects may have been confounded by other pollutants. In 12
southern California communities in the Children's Health Study (CHS), Gauderman et al. (2000,
012531; 2002, 026013) found that decreases in lung function growth among schoolchildren were
associated with long-term exposure to PM. Declines in pulmonary function were reported with all
three major PM size classes - PMi0, PMi0_2.5 and PM2.5 - though the three PM measures were highly
correlated. In another analysis of data from the CHS cohort, McConnell et al. (1999, 007028).
reported an increased risk of bronchitis symptoms in children living in communities with higher
PMio and PM2.5 concentrations. These results were found to be consistent with results of cross-
sectional analyses of the 24-city study by Dockery et al. (1996, 046219) and Raizenne et al. (1996,
077268). that were assessed in the 1996 PM AQCD. These studies reported associations between
increased bronchitis rates and decreased peak flow with fine particle sulfate and fine particle acidity.
However, the high correlation of PMi0, acid vapor and NO2 precluded clear attribution of the
bronchitis effects reported by McConnell et al. (1999, 007028) to PM alone. In a prospective cohort
study among a subset of children in the CHS (n = 110) who moved to other locations during the
study period, Avol et al. (2001, 020552) reported that those subjects who moved to areas of lower
PMio showed increased growth in lung function compared with subjects who moved to communities
with higher PMio concentrations. Finally, the 2004 PM AQCD concluded that there was strong
epidemiologic evidence for associations between long-term exposures to PM2 5 and cardiopulmonary
mortality, though the respiratory effects were not separated from the cardiovascular effects in this
conclusion.
The 2004 PM AQCD (U.S. EPA, 2004, 056905) concluded that the evidence for an association
between long-term exposure to PM and respiratory effects may be confounded by other pollutants.
Gauderman et al. (2002, 026013) reported declines for FEVi and McConnell et al. (1999, 007028)
reported increased ORs for bronchi tic symptoms in asthmatics for PMio and PM2 5. Recent
epidemiologic literature includes results from several prospective cohort studies, which found
consistent, positive associations between long-term exposure to PM and respiratory morbidity.
Associations were reported with PM2 5 and PMio, and the studies showing associations only with
PMio were conducted in locations where the PM consisted predominantly of fine particles, providing
support for associations with long-term exposure to fine particles. These results are summarized
below; further details of these studies are summarized in Annex E.
Very few subchronic and chronic toxicological studies investigating respiratory effects were
available in the 2004 PM AQCD. However, the 2002 EPA Health Assessment Document for DE
reported that chronic exposure to DE was associated with histopathology including alveolar
histiocytosis, aggregation of alveolar macrophages, tissue inflammation, increased
polymorphonuclear leukocytes, hyperplasia of bronchiolar and Type 2 epithelial cells, thickened
alveolar septa, edema, fibrosis, emphysema and lesions of the trachea and bronchi. Since then a
number of animal toxicological studies have been conducted involving inhalation exposure to CAPs,
urban air, DE, gasoline exhaust, and wood smoke. These subchronic and chronic studies provide
evidence of altered pulmonary function, inflammation, histopathological changes and oxidative and
allergic responses following PM2 5 exposures. These results are summarized below; further details of
these studies are summarized in Annex D.
7.3.1. Respiratory Symptoms and Disease Incidence
7.3.1.1. Epidemiologic Studies
New longitudinal cohort studies provide the best evidence to evaluate the relationship between
long-term exposure to ambient PM and increased incidence of respiratory symptoms or disease. A
summary of the mean PM concentrations reported for the long-term exposure studies characterized
in this section is presented in Table 7-3.
December 2009 7-20
-------
Bayer-Oglesby et al. (2005, 086245) examined the decline of ambient pollution levels and
improved respiratory health demonstrated by a reduction in respiratory symptoms and diseases in
school children (n = 9,591) in Switzerland. Reduced air pollution exposure resulted in improved
respiratory health of children. Further, the average reduction of symptom prevalence was more
pronounced in areas with stronger reduction of air pollution levels. The average decline of PMi0
between 1993 and 2000 across the nine study regions was 9.8 (ig/m (29%). Declining levels of PMi0
were associated with declining prevalence of chronic cough, bronchitis, common cold, nocturnal dry
cough, and conjunctivitis symptoms, but no significant associations were reported for wheezing,
sneezing, asthma, and hay fever, as shown in Figure 7-2. In Figure 7-2, Panel (B) illustrates that on
an aggregate level across regions, the mean change in adjusted prevalence of chronic cough is
associated with the mean change in PMio levels (r = 0.78; p = 0.02). Similar associations were seen
for nocturnal dry cough and conjunctivitis symptoms and PMio levels. Roosli et al. (2000, 010296;
2001, 108738; 2005, 156923) have demonstrated that PM10 levels are homogeneously distributed
within regions of Basel, Switzerland and are not substantially affected by local traffic, justifying the
single-monitor approach for assignment of PMi0 exposures. Based on parallel measurements of
PM2.5 and PMio at seven sites in Switzerland, PM2 5 and PMio at all sites are generally highly
correlated (r2 ranging from 0.85 to 0.98) (Gehrig and Buchmann, 2003, 139678). indicating that
PMio consists predominantly of fine particles in these locations.
Schindler et al. (2009, 191950) reported that sustained reduction in ambient PM10
concentrations can lead to decreases in respiratory symptoms among Swiss adults in the SAPALDIA
study. They compared baseline data in 1991 to a follow-up interview in 2002 after a substantial
decline in PMio concentrations served as a natural experiment. Each subject was assigned model-
based estimates of PMio concentrations averaged over the 12 mo preceding each health assessment
with mean decline in PMi0 levels of 6.2 (ig/m (SD = 3.9 (ig/m3). When the authors tested the joint
hypothesis of no association between the PM10 difference and symptom incidence or persistence,
positive results were obtained for regular cough, chronic cough or phlegm and wheezing but not
regular phlegm or wheezing without a cold.
Pierse et al. (2006, 088757) studied the association between primary PMio (particles directly
emitted from local sources/traffic) and the prevalence and incidence of respiratory symptoms in a
randomly sampled cohort of 4,400 children (aged 1-5 yr) in Leicestershire, England surveyed in
1998 and again in 2001. Annual exposure to primary PMio was calculated for the home address
using the Airviro statistical dispersion model. After adjusting for confounders, mean annual exposure
to locally generated PMio was associated with an increased prevalence of cough without a cold in
both the 1998 (OR 1.21 [95% CI: 1.07-1.38], n = 2,164) and 2001 surveys (OR 1.56 [95% CI:
1.32-1.84], n= 1,756).
Nordling et al. (2008, 097998) examined the relationship between estimated PM exposure
levels and respiratory health effects in a Swedish birth cohort of preschool children (n = 4,089). The
spatial distributions of PM from traffic in the study area were estimated with emission databases and
statistical dispersion modeling. Children were examined at 2 mo and 1, 2, and 4 yr of age. Using GIS
methods, the average contribution of traffic-generated PMio above regional background to the
children's residential outdoor air pollution levels was determined. To evaluate the exposure
assessment, the authors compared the estimated levels of traffic-generated PMio with PM2.5
measurements from 42 locations (Hoek et al., 2002, 042364) and reported modeled traffic-generated
PMio correlated reasonably well with measured PM2.5 (r = 0.61). Persistent wheezing (cumulative
incidence up to age 4 yr) was associated with exposure to traffic-generated PMio (OR 2.28 [95% CI:
0.84-6.24] per 10 ug/m3 increase) while transient and late onset wheezing was not associated. This
study demonstrates that respiratory effects may be present in preschool children.
December 2009 7-21
-------
Table 7-3. Characterization of ambient PM concentrations from studies of respiratory
symptoms/disease and long-term exposures.
Study
Location
Mean Annual Concentration (ug/m3)
Upper Percentile
Concentrations
(ug/m3)
PM2.5
Annesi-Maesano et al. (2007, 0931801
Braueretal. (2007, 0906911
Cosset al. (2004, 055624)
Islam et al. (2007, 090697)
Janssen et al. (2003, 1335551
Kim et al. (2004, 0873831
McConnell et al. (2003, 049490)
Morgenstern et al. (2008, 1567821
6 French Cities
The Netherlands
U.S.
12 CHS/CA communities
The Netherlands
San Francisco, CA
12 CHS/CA communities
Munich, Germany
Range of means across sites: 8.7-23.0
Avg of means across sites: 15.5
16.9
13.7
20.5
Range of means across sites: 11-15
Avg of means across sites: 12
13.8
11.1
75th: 18.1
90th: 19.0
Max: 25.2
75th: 15.9
Max: 29.5
75th: 22.1
Max: 24.4
Max: 28.5
PM10
Bayer-Oglesby et al. (2005, 0862451
Kunzli et al. (2009, 1919491
Nordling et al. (2008, 097998)
Schindleretal. (2009, 191950)
McConnell et al. (2003, 049490)
Pierse et al. (2006, 0887571
Nine study regions in Switzerland
Switzerland
Sweden
Switzerland
12 CHS/CA communities
Leicestershire, U.K.
21.5
4*
**
30.8
1.33
Max: 46
Max: 63.5
75th: 1.84
*Source specific; PM10 from traffic
**0nly reported change in PM concentration
December 2009
7-22
-------
c
o
0)
E
^»
T3
V
3
^
Source: Bayer-Oglesby et al. (2005, 0862451
Figure 7-2. Adjusted ORs and 95% CIs of symptoms and respiratory diseases associated
with a decline of 10 ug/m3 PMio levels in Swiss Surveillance Program of
Childhood Allergy and Respiratory Symptoms1. Inset: Mean change in adjusted
prevalence (1998-2001 to 1992-1993) versus mean change in regional annual
averages of PM™ (1997-2000 to 1993) for chronic cough, across nine SCARPOL
regions (An: Anieres. Be: Bern. Bi: Biel. Ge: Geneva. La:, Langnau. Lu: Lugano.
Mo: Montana. Pa: Payerne. Zh: Zurich).
McConnell et al. (2003, 049490) conducted a prospective study examining the association
between air pollution and bronchitic symptoms in 475 school children with asthma in 12 Southern
California communities as part of the CHS from 1996 to 1999. They investigated both the
differences between- communities with 4-yr avg and within-communities yearly variation in PM
(i.e., PMio, PM25, PMio_25, EC, and OC). Based on a 10 ug/m3 change in PM25, within-communities
effects were larger (OR 1.90 [95% CI: 1.10-2.70]) than those for between-communities (OR 1.30
[95% CI: 1.10-1.50]). The OR for the 10 ug/m3 range in 4-yr avg PM25 concentrations across the 12
communities was 1.29 (95% CI: 1.06-1.58). Similar results were reported for PMi0 and PMi0_2.5 but
the effect estimates were smaller in magnitude and generally not statistically significant. Within-
community associations were not confounded by any time-fixed personal covariates. In two-
1 Adjusted for age, sex, nationality, parental education, number of siblings; farming status, low birth weight, breastfeeding, child who
smokes, family history of asthma, bronchitis, and/or atopy, mother who smokes, indoor humidity, mode of heating and cooking,
carpeting, pets allowed in bedroom, removal of carpet and/or pets for health reasons, person who completed questionnaire, month when
questionnaire was completed, number of days with the maximum temperature <0°C, and belief of mother that there is an association
between environmental exposures and children's respiratory health
December 2009
7-23
-------
pollutant models, the within-community effect estimates for PM2.5 and OC were significant in the
presence of several other pollutants. While the within-community single-pollutant effect of PM2.5
(B = 0.085/ug/m3) was only modestly attenuated after adjusting for some pollutants, it was markedly
reduced after adjusting for NO2 or OC. The between-community effect estimates generally were not
significant in the presence of other pollutants in copollutant models.
In the CHS, Islam et al. (2007, 090697) examined the hypothesis that ambient air pollution
attenuates the reduced risk for childhood asthma that is associated with higher lung function
(n = 2,057). At each age a distribution of pulmonary functions exists. Haland et al. (2006, 156511)
found evidence that children with high lung function have a reduced risk for asthma. Islam et al.
(2007, 090697) used the CHS data to study how the association of asthma incidence with lung
function is modified by long-term PM exposure. The incidence rate (IR) of newly diagnosed asthma
increased from 9.5/1,000 person-years for children with percent-predicted FEF25_75 values > 120% to
20.4/1,000 person-years for children with FEF25_75 value < 100%. Over the 10th-90th percentile
range for FEF25.75 (57.1%), the hazard ratio of new onset asthma was 0.50 (95% CI: 0.35-0.71). The
IR of asthma for FEF25_75> 120% in the "high" PM2.5 (13.7-29.5 (ig/m3) communities was 15.9/1,000
person-years compared to 6.4/1,000 person-years in "low" PM25 (5.7-8.5 (ig/m3) communities. Loss
of protection by high lung function against new onset asthma in the "high" PM25 communities was
observed for all the lung function measures. Figure 7-3 shows the effect of PM25 on the association
of lung function with asthma. Of all the pollutants examined (NO2, PM10, PM25, acid vapor, O3, EC,
and OC), PM2 5 appeared to have the strongest modifying effect on the association between lung
function with asthma as it had the highest R2 value (0.42). Over the 10th-90th percentile range of
FEF25_75, the hazard ratio of new onset asthma was 0.34 (95% CI: 0.21-0.56) in a community with
low PM2.5 (<13.7 ug/m3) and 0.76 (95% CI: 0.45-1.26) in a community with high PM2.5
(> 13.7 ug/m3). The data do not indicate that PM exposure increased rates of incident asthma among
children with poor lung function at study entry because rates among those with poor lung function
were similar in both low and high pollution communities.
1.00
1.50-
1.20-1
as
N
0.90-
0.60-
0.30-
0.00
High PM2 6 - Communities
Low PM2 5
Communities
ML
RV
LB
R2 = 0.42
P = 0.01
I
10
I
15
I
20
I
25
I
30
PM2.5(Mg/m3)
Source: Reprinted with Permission of BMJ Publishing Group Ltd & British Thoracic Society from Islam et al. (2007, 090697'
Figure 7-3. Effect of PM2.s on the association of lung function with asthma. Community-
specific hazard ratio of newly diagnosed asthma over 10-90th percentile range
(57.1%) of FEF25.75% by level of ambient PM2.5 (ug/m3). The 12 CHS communities
are shown.
In a prospective birth cohort study (n = 4,000) in The Netherlands, Brauer et al. (2007,
090691) assessed the development of asthma, allergic symptoms, and respiratory infection during
the first 4 yr of life in relation to long-term PM2 5 concentration at the home address with a validated
model using GIS. PM25 was associated with doctor-diagnosed asthma (OR = 1.32 [95% CI:
December 2009
7-24
-------
1.04-1.69]) for a cumulative lifetime indicator. These findings extend observations made at 2 yr of
age in the same cohort (Brauer et al., 2002, 035192) providing greater confidence in the association.
No associations were observed for bronchitis.
Kunzli et al. (2009, 191949) used the SAPALDIA cohort study discussed previously in this
section to evaluate the relationship between the 11-yr change (1991-2002) in traffic-related PMi0 and
asthma incidence-adult onset asthma. In a cohort of 2,725 never-smokers without asthma at baseline
(age: 18-60 yr in 1991), subjects reporting doctor-diagnosed asthma at follow-up were considered
incident cases. Modeled traffic-related PMi0 levels were used. Cox proportional hazard models for
time to asthma onset were used with adjustments for cofounders. The study findings suggest that PM
contributes to asthma development and that reductions in PM decrease asthma risk. A strong feature
of SAPALDIA is the ability to assign space, time, and source-specific pollution to each subject.
Further, Kunzli et al. (2008, 129258) discusses the impact of attributable health risk models for
exposures that are assumed to cause both chronic disease and its exacerbations. The added impact of
causing disease increases the risk compared to only exacerbations.
A matched case-control study of infant bronchiolitis (ICD 9 code 466.1) hospitalization and
two measures of long-term exposure - the month prior to hospitalization (subchronic) and the
lifetime average (chronic) - to PM2.5 and gaseous air pollutants in the South Coast Air Basin of
southern California was conducted by Karr et al. (2007, 090719) among 18,595 infants born between
1995-2000. For each case, 10 controls matched on date were randomly selected from birth records.
Exposure was based on PM2.5 measurements collected every third day. The mean distance between
the subjects' residential ZIP code and the assigned monitor was generally 4-6 mi with a maximum
distance of 30 mi. For 10 ug/m3 increases in both sub-chronic and chronic PM2.5 exposure, an
adjusted OR of 1.09 (95% CI: 1.04-1.14) was observed. In multipollutant model analyses, the
association with PM2.5 was robust to the inclusion of gaseous pollutants. Also, in a cohort of children
in Germany, Morgenstern et al. (2008, 156782) modeled PM2 5 data at birth addresses found
statistically significant effects for asthmatic bronchitis, hay fever, and allergic sensitization to pollen.
Goss et al. (2004, 055624) conducted a national study examining the relationship between air
pollutants and health effects in a cohort of cystic fibrosis (CF) patients (n = 11,484) over the age of
6 yr (mean age = 18.4, SD = 10) enrolled in the Cystic Fibrosis Foundation National Patient Registry
in 1999 and 2000. Exposure was assessed by linking air pollution values from the closest population
monitor from the Air Quality System (AQS) with the centroid of the patient's home ZIP code that
was within 30 mi. The mean distance from the patient's ZIP code to monitors for PM2 5 and PMi0
was 10.8 mi (SD 7.8) and 11.5 mi (SD 7.9), respectively. PM2.5 and PMi0 24-h avg were collected
every 1 to 12 days. CF diagnosis involves genetic screening panels and a common severe mutation
used is the loss of phenylalanine at the 508th position. Genotyping was available in 74% of the
population and of those genotyped, 66% carried one or more delta F508 deletions. After adjusting for
confounders, a 10 ug/m increase in PM25 or PM10 was associated with a 21% (95% CI: 7-33) or 8%
(95% CI: 2-15) increase in the odds of two or more exacerbations, respectively. The exacerbations
were defined as a CF-related pulmonary condition requiring admission to the hospital or use of home
intravenous antibiotics. The estimate for the associations between pulmonary exacerbations and
PM25 and PMi0 were attenuated when the models were adjusted for lung function. Brown et al.
(2001, 012307) found that particle deposition was increased in CF and that particle distribution in
the lungs was enhanced in poorly ventilated tracheobronchial regions in CF patients. Such focal
deposition may partially explain the association of PM and CF exacerbation.
Annesi-Maesano (2007, 093180) relate individual data on asthma and allergy from 5,338
school children (10.4 ± 0.7 yr) attending 108 randomly chosen schools in 6 French cities to the
concentration of PM25 monitored in school yards. Atopic asthma was related to PM25 (OR 1.43
[95% CI: 1.07-1.91]) when high PM2.5 concentrations (20.7 ug/m3) were compared to low PM2.5
concentrations (8.7 ug/m3). The report is consistent with the results in an earlier paper (Penard-
Morand et al., 2005, 087951) in the same sample of children that related the findings to PMi0.
Kim et al. (2004, 087383) conducted a school-based cross-sectional study in the San Francisco
metropolitan area in 2001 comprised of 10 neighborhoods to examine the relationship between
traffic-related pollutants and current bronchitic symptoms and asthma obtained by parental
questionnaire (n = 1,109). They related traffic-related pollutants (PM) and bronchitic and asthma
symptoms in the past 12 mo. No multipollutant models were evaluated because of the high
interpollutant correlations. PM25 levels ranged across the school sites from 11 to 15 ug/m3.
Schikowski et al. (2005, 088637) examined the relationship between both long-term air
pollution exposure and living close to busy roads and COPD in the Rhine-Ruhr Basin of Germany
December 2009 7-25
-------
from 1985 to 1994 using consecutive cross-sectional studies. Seven monitoring stations that were
<8 km to a woman's home address provided TSP data that PMi0 was estimated from using a
conversion factor (obtained from parallel measurement of TSP and PMi0 conducted at 7 sites in the
Ruhr area). Distance to a major road was determined using GIS. The results of the study suggest that
long-term exposure to air pollution from PMi0 and living near a major road might increase the risk of
developing COPD and can have a detrimental effect on lung function. All ORs for 5-yr exposures
were stronger than those for 1-yr exposures.
In summary, the 2004 PM AQCD evaluated the available studies which primarily related
effects to bronchitic symptoms in school-age children. New studies are using several different
methods to include individual estimates of exposure to ambient PM that may reduce the impact of
exposure error. The strength and consistency of the outcomes is enhanced by results being reported
by several different researchers in different countries using different designs. Most recent studies
have focused on children, but a few studies have also reported associations in adults.
The CHS (McConnell et al., 2003, 049490) provides evidence in a prospective longitudinal
cohort study that relates PM2.s and bronchitic symptoms and reports larger associations for within-
community effects that are less subject to confounding than between-community effects. Several
new studies report similar findings with long-term exposure to PMi0 in areas where fine particles are
the predominant fraction of PMi0. In England, in a cohort of 4,400 children (aged 1-5 yr), an
association is seen with an increased prevalence of cough without a cold. Further evidence includes a
reduction of respiratory symptoms corresponding to decreasing PM levels in "natural experiments"
in both a cohort of Swiss school children (Bayer-Oglesby et al., 2005, 086245) and adults (Schindler
etal, 2009, 191950).
In a separate analysis of the CHS, Islam et al. (2007, 090697) showed that PM2.5 had the
strongest modifying effect on the association between lung function with asthma such that loss of
protection by high lung function against new onset asthma in high PM2.s communities was observed
for all the lung function measures from 10 to 18 yr of age. This relates new onset asthma to long-
term PM exposure. In the Netherlands, Brauer et al. (2007, 090691) augments the literature with data
examining the first 4 yr of life in a birth cohort showing an association with doctor-diagnosed
asthma. Further, in an adult cohort in the SALPALDIA study, Kunzli et al. (2009, 191949) relate PM
to asthma incidence.
7.3.2. Pulmonary Function
Several cohort studies reviewed in the 2004 PM AQCD provided evidence for relationships
between long-term PM exposure and effects on the respiratory system. In 12 southern California
communities in the Children's Health Study (CHS), Gauderman et al. (2000, 012531: 2002, 026013)
found that decreases in lung function growth among school children were associated with long-term
exposure to PM. Declines in pulmonary function were reported with all three major PM size classes
- PMio, PMio_2.5 and PM2.5 - though the three PM measures were highly correlated. These results
were found to be consistent with results of cross-sectional analyses of Raizenne et al. (1996,
077268). that was assessed in the 1996 PM AQCD. That study reported associations between
decreased peak flow with fine particle sulfate and fine particle acidity. Finally, in a prospective
cohort study among a subset of children in the CHS (n = 110) who moved to other locations during
the study period, Avol et al. (2001, 020552) reported that those subjects who moved to areas of lower
PMio showed increased growth in lung function compared with subjects who moved to communities
with higher PMio concentrations who showed decrease growth in lung function.
7.3.2.1. Epidemiologic Studies
New longitudinal cohort studies have evaluated the relationship between long-term exposure
to PM and changes in measures of pulmonary function (FVC, FEVi, and measures of expiratory
flow). Cross-sectional studies also offer supportive information (Annex E) and may provide insights
derived from within community analysis. Lung function increases continue through early adulthood
with growth and development, then declines with aging (Stanojevic et al., 2008, 157007; Thurlbeck,
1982, 093260; Zeman and Bennett, 2006, 157178). A summary of the mean PM concentrations
reported for the long-term exposure studies characterized in this section is presented in Table 7-4.
December 2009 7-26
-------
Table 7-4. Characterization of ambient PM concentrations from studies of FEVi and long-term
exposures.
Study
Location
Mean Annual
Concentration (ug/m3)
Upper Percentile
Concentrations (ug/m3)
PM2.5
Gauderman et al. (2002, 0260131
12 CHS/CA communities
5-30
Gauderman et al. (2004, 0565691
12 CHS/CA communities
6-27
Cosset al. (2004, 0556241
U.S.
13.7
75th: 15.9
Range of means across sites: 3.7-44.7
Gotschi et al. (2008, 1803641
21 European cities
Avg of mean across sites: 16.8
PM10
Downs et al. (2007, 0928531
Gauderman et al. (2002, 0260131
Gauderman et al. (2004, 0565691
Nordling et al. (2008, 097998)
Avoletal. (2001,020552)
Rojas-Martinez et al. (2007, 0910641
8 cities in Switzerland
12 CHS/CA communities
12 CHS/CA communities
Sweden
Southern CA/CHS
Mexico City, Mexico
Range of means across sites: 9-46
Avg of mean across sites: 21.6
Range of means across sites: 13-78
Avg of mean across sites: NR
Range of means across sites: 18-68
Avg of mean across sites: NR
Modeled exposure
Range of means across sites: 15.0-66.2
75.6
75th: 92.2
90th: 112.7
The CHS prospectively examined the relationship between air pollutants and lung function
(FVC, FEVi, MMEF) in a cohort (n = 1,759) of children between the ages of 10 and 18 yr, a period
of rapid lung development (Gauderman et al., 2004, 056569). Air pollution monitoring stations
provided data in each of the 12 study communities from 1994-2000. The results for C^PM^NO^
PM2.5, acid vapor, and EC are depicted in Figure 7-4. In general, copollutant models for any pair of
pollutants did not provide a substantially better fit to the data than the corresponding single-pollutant
models due to the strong correlation between most pollutants. The pollution-related deficits in the
average growth in lung function over the 8-yr period resulted in clinically important deficits in
attained lung function at the age of 18.
Downs et al. (2007, 092853) prospectively examined 9,651 randomly selected adults (18-60 yr
of age) in eight cities in Switzerland (see alsoAckermann-Liebrich et al., 1997, 077537) to ascertain
the relationship between reduced exposure to PM10 and age-related decline in lung function (FVC,
FEVi, and FEF25_5o). An evaluated statistical dispersion model (Liu et al., 2007, 093093) provided
spatially resolved concentrations of PMi0 that enabled assignment to residential addresses for the
participant examinations in 1991 and 2002 that yielded a median decline of 5.3 ug/m3 (IQR 4.1-7.5).
Decreasing PMi0 concentrations attenuated the decline in lung function. Effects were greater in tests
reflecting small airway function. No other pollutant relationships were evaluated, though a related
study indicated that levels of NO2 also declined over the same period (Ackermann-Liebrich et al.,
2005, 087826). Generalized cross-validation essentially chose a linear fit for the concentration-
response curve for age-related decline in lung function.
These data show that improvement in air quality may slow the annual rate of decline in lung
function in adulthood indicating positive consequences for public health. Further evidence on
improvement in respiratory health with reduction in air pollution levels is provided from studies
conducted in East Germany related to dramatic emissions reductions after the reunification in 1990
(Fryer and Collins, 2003, 156454: Heinrich et al., 2002, 034825; Sugiri et al., 2006, 088760). This
type of "natural experiment" provides additional support for epidemiologic findings that relatively
low levels of airborne particles have respiratory effects.
December 2009
7-27
-------
1
1
R=0.04
P-0.89
LB
UP
SD
_*_
ID-,
* ML
• RV
» AT
» LM
» AL
• LE
I
LA
i-tN-
»UP
*ML
25
35 45 55 65
O3 from 10 a.m. to 6 p.m. (ppb)
10-,
R-0.75
P-0.005
75
UP
L;N
10 20 30 40 50 60
PM10 (^g/m'l
70
ML
1
»SD
ID-,
6-
4-
»LE
LN,
»LE
10
20
N02(ppb)
30
40
e. 10-,
I
"S
» LN
10
R=0.74
P=0.006
15
PM2
20
25
30
* UP
»SD
#LE
LN
4 6
Acid Vapor (ppb)
10 12
Figure 7-4.
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
Elemental Carbon (fg/m3)
Source: Adapted from Gauderman et al. (2004, 056569)
Copyright © 2004 Massachusetts Medical Society. All rights reserved.
Proportion of 18-yr olds with an FEV1 below 80% of the predicted value plotted
against the average levels of pollutants from 1994 through 2000 in the 12
southern California communities of the Children's Health Study. AL = Alpine;
AT = Atascadero; LA = Lake Arrowhead; LB = Long Beach; LE = Lake Elsinore;
LM = Lompoc; LN = Lancaster; ML = Mira Loma; RV = Riverside; SD = San Dimas;
SM = Santa Maria; UP = Upland.
In a prospective cohort study consisting of school-age children (n = 3,170) who were 8 yr of
age at the beginning of the study, had not been diagnosed with asthma, and were located in Mexico
City, Rojas-Martinez et al. (2007, 091064) evaluated the association between long-term exposure to
PMio, O3 and NO2 and lung function growth every 6 mo from April 1996 through May 1999.
Exposure data were provided by 10 air quality monitor stations located within 2 km of each child's
school. The multipollutant model effect of PMi0 over the age of 8-10 yr of life in this cohort on FVC,
FEVi, and FEF 25-75 showed an association. Single pollutant models showed an association between
ambient pollutants (O3, PMi0 and NO2) and deficits in lung function growth. The association
between PM10 and FEF25_75 was not statistically significant. While the estimates from copollutant
models were not substantially different than single pollutant models, independent effects for
pollutants could not be estimated accurately because the traffic-related pollutants were correlated.
December 2009
7-28
-------
Although no PM2.5 data were presented in this study, in a separate study Chow et al. (2002) report
that during the winter of 1997 approximately 50% of PMi0 was in the PM25 fraction in Mexico City.
Gotschi et al. (2008, 156485) examined the relationship between air pollution and lung
function in adults in the European Community Respiratory Health Survey (ECRHS). FEVi and FVC
were assessed at baseline and after 9 yr of follow-up from 21 European centers (followed-up sample
n = 5,610). No statistically significant associations were found between city-specific annual mean
PM2 5 and average lung function levels which is in contrast to the results seen by Ackermann-
Liebrich (1997, 077537) (SAPALDIA) and Schikowski et al. (2005, 088637) (SALIA) which
compared across far more homogenous populations than for the population assessed in the ECRHS.
Misclassification and confounding may partially explain the discrepancy in findings.
In a birth cohort (n = 2,170) in Oslo, Norway, Oftedal et al. (2008, 093202) examined effects
of exposure to PM2.5 and PMi0 on lung function (FVC, FEVi, FEF50°/0). Spirometry was performed in
2,307 children aged 9-10 yr in 2001-2002. Residential air pollution levels over the time period
1992-2002 were calculated using EPISODE dispersion models to provide three time scales of
exposure: (1) first year of life; (2) lifetime exposure; and (3) just before the lung function test. Only
single pollutant models were evaluated because air pollutants were highly correlated (r = 0.83-0.95).
PM exposure was associated with changes in adjusted peak respiratory flow, especially in girls. No
effect was found for forced volumes. Adjusting for contextual socioeconomic factors diminished
associations. Results for PMi0 were similar to those for PM2 5.
In an exploratory study, Mortimer et al. (2008, 187280) examined the association of prenatal
and lifetime exposure to air pollutants using geocoded monthly average PMi0 levels with pulmonary
function in a San Joaquin Valley, California cohort of 232 children (ages 6-11 yr) with asthma. First
and second trimester PMiq exposures (based on monthly average concentrations) had a negative
effect on pulmonary function and may relate to prenatal exposures affecting the lungs as they begin
to develop at 6-wk gestation.
Dales et al. (2008, 156378) in a cross-sectional prevalence study examined the relationship of
pulmonary function and PM measures, other pollutants, and indicators of motor vehicle emissions in
Windsor, Ontario, in a cohort of 2,402 school children. PM25 and PMi0 concentrations were
estimated for each child's residence at the postal code level. Each 10 ug/m3 increase in PM25 was
associated with a 7.0% decrease in FVC expressed in a percentage of predicted.
In Leicester, England, investigators examined the carbon content of airway macrophages in
induced sputum in 64 of 114 healthy 8-15 year-old children (Grigg et al., 2008, 156499; Kulkarni et
al., 2006, 089257). The carbon content of airway macrophages (Finch et al., 2002, 054603; Strom et
al., 1990, 157020) was used as a marker of individual exposure to PMip. Near each child's home,
exposure to PMi0 was estimated using a statistical dispersion model (Pierse et al., 2006, 088757).
The authors reported a dose-dependent inverse association between the carbon content of airway
macrophages and lung function in children and found no evidence that reduced lung function itself
causes an increase in carbon content. Consistent results were obtained for both FVC and FEF25_75.
Caution should be used when interpreting these results as the accuracy of the estimates on individual
PMio exposures were not validated; there is potential for confounding by ethnic origin; and there is
concern that the magnitude of the changes in pulmonary function associated with increased particle
area appear large (Boushey et al., 2008, 192162).
Nordling et al. (2008, 097998) discussed above in the respiratory symptoms section, also
reported that lower PEF at age 4 was associated with exposure to traffic-related PMi0 (-8.93 L/min
[95% CI: -17.78 to -0.088]). Goss et al. (2004, 055624). discussed in Section 7.3.1.1, found strong
inverse relationships between FEVi and PM25 concentrations in both cross-sectional and
longitudinal analyses.
In summary, recent studies have greatly expanded the evidence available for the 2004 PM
AQCD. The earlier CHS studies followed young children for 2-4 yr. New analyses have been
conducted that include longer follow-up periods of this cohort through 18 yr of age (considered early
adulthood for lung development (Stanojevic et al., 2008, 157007) and provide evidence that effects
from exposure to PM2 5 persist into early adulthood. Longitudinal studies follow effects over time
and are considered to provide the best evidence as opposed to studies across communities as in
cross-sectional studies. The longitudinal cohort studies in the 2004 PM AQCD provided data for
children in one location in one study and new longitudinal studies have been conducted in other
locations.
Gauderman et al. (2004, 056569) reported that PM2 5 exposure was associated with clinically
and statistically significant deficits in FEVi attained at the age of 18 yr. Clinical significance was
December 2009 7-29
-------
defined as a FEVi below 80% of the predicted value, a criterion commonly used in clinical settings
to identify persons at increased risk for adverse respiratory conditions. This clinical aspect is an
important enhancement over the earlier results reported in the 2004 PM AQCD. Further, the
association reported in this study that evaluated the 8-yr time period into early adulthood not only
provided evidence for the persistence of the effect, but in addition the strength and robustness of the
outcomes were more positive, larger, and more certain than previous CHS studies of shorter follow-
up.
Supporting this result are new longitudinal cohort studies conducted by other researchers in
other locations with different methods. Though these studies report results for PMi0, available data
discussed above indicate that the majority of PMi0 is composed of PM2.5 in these areas. New studies
provide positive results from Mexico City, Sweden, and a national cystic fibrosis cohort in the U.S.
One study reported null results in a European cohort described as having potential misclassification
and confounding concerns as well as lacking a homogenous population potentially rendering the
outcome as non-informative. A natural experiment in Switzerland, where PM levels had decreased,
reported that improvement in air quality may slow the annual rate of decline in lung function in
adulthood, indicating positive consequences for public health. These natural experiments are
considered especially supportive.
The relationship between long-term PM exposure and decreased lung function is thus seen
during lung growth and lung development in school-age children into adulthood. At adult ages
studies continue to show a relationship between decreased lung function and long-term PM
exposure. Some newer studies attempting to study the relationship of long-term PM exposure from
birth through preschool are reporting a relationship. Thus, the impact of long-term PM exposure is
seen over the time period of lung function growth and development and the decline of lung function
with aging.
Overall, effect estimates from these studies are negative (i.e., indicating decreasing lung
function) and the pattern of effects are similar between the studies for FVC and FEVi. Thus, the data
are consistent and coherent across several designs, locations, and researchers. With cautions noted,
the results relating carbon content of airway macrophages to decreased measures of pulmonary
function add plausibility to the epidemiologic findings. Some new studies are using individual
estimates of exposure to ambient PM to reduce the impact of exposure error (Downs et al., 2007,
092853: Jerrett et al., 2005, 087381).
As was found in the 2004 PM AQCD, the studies report associations with PM2.5 and PMi0,
while most did not evaluate PMi0_2.5. Associations have been reported with fine particle components,
particularly EC and OC. Source apportionment methods generally have not been used in these long-
term exposure studies. However, numerous studies have evaluated exposures to PM related to traffic
or motor vehicle sources. For example, Meng et al. (2007, 093275) investigated the associations
between traffic and outdoor pollution levels and poorly controlled asthma among adults who were
respondents to the California Health Interview Survey and found associations for traffic density and
o, but not PM2.5.
7.3.2.2. lexicological Studies
Urban Air
One new study evaluated the effects of chronic exposure to ambient levels of urban particles
on lung development in the mouse (Mauad et al., 2008, 156743). Both functional and anatomical
indices of lung development were measured. Male and female BALB/c mice were continuously
exposed to ambient or filtered Sao Paolo air for 8 mo. Concentrations in the "polluted chamber"
versus "clean chamber" were 16.8 versus 2.9 (ig/m3 PM25. Thus PM levels were reduced by
filtration but not entirely eliminated. Ambient concentrations of CO, NO2 and SO2 were 1.7 ppm,
89.4 (ig/m3 and 8.1 (ig/m3, respectively. Concentrations of gaseous pollutants were assumed to be
similar to ambient levels in both chambers. After 4 mo, the animals were mated and the offspring
were divided into 4 groups to provide for a prenatal exposure group, a postnatal exposure group, a
pre and postnatal exposure group and a control group. Animals were sacrificed at 15 and 90 days of
age for histological analysis of lungs. Pulmonary pressure-volume measurements were also
conducted in the 90-day-old offspring. Statistically significant reductions in inspiratory and
December 2009 7-30
-------
expiratory volumes were found in the group receiving both prenatal and postnatal exposure, but not
in the groups receiving only prenatal exposure or only postnatal exposure, compared with controls.
These changes in pulmonary function correlated with anatomical changes which are discussed in
Section 7.3.5.1.
Diesel Exhaust
Li et al. (2007, 155929) exposed BALB/c and C56BL/6 mice to clean air or to low-dose DE
(at a PM concentration of 100 (ig/m3) for 7 h/day and 5 days/week for 1, 4 and 8 wk. Average gas
concentrations were reported to be 3.5 ppm CO, 2.2 ppm NO2, and less than 0.01 ppm SO2. Airway
hyperresponsiveness (AHR) was evaluated by whole-body plethysmography at Day 0 and after 1, 4
and 8 wk of exposure. Short-term exposure responses are discussed in Section 6.3.2.3, 6.3.3.3 and
6.3.4.2. The increased sensitivity of airways to methacholine (measured as Penh) seen in C57BL/6
but not BALB/c mice at 1 week was also seen at 4 wk but not at 8 wk. This study suggests that
adaptation occurs during prolonged DE exposure. Influx of inflammatory cells, markers of oxidative
stress and effects of antioxidant intervention were also evaluated (Sections 7.3.3.2 and 7.3.4.1).
Although no attempt was made in this study to determine the effects of gaseous components of DE
on the measured responses, concentrations of gases were very low suggesting that PM may have
been responsible for the observed effects.
In many animal studies changes in ventilatory patterns are assessed using whole-body
plethysmography, for which measurements are reported as enhanced pause (Penh). Some
investigators report increased Penh as an indicator of AHR, but these are inconsistently correlated
and many investigators consider Penh solely an indicator of altered ventilatory timing in the absence
of other measurements to confirm AHR. Therefore use of the terms AHR or airway responsiveness
has been limited to instances in which the terminology has been similarly applied by the study
investigators.
Gottipolu et al. (2009, 190360) exposed WKY and SH rats to filtered air or DE (particulate
concentration 500 and 2,000 fig/m3) for 4 h/day and 5 days/wk over a 4-wk period. Concentrations
of gases were 1.3 and 4.8 ppm CO, NO <2.5 and 5.9 ppm NO, O.25 and 1.2 ppmNO2, 0.2 and
0.3 ppm SO2for low and high PM exposures, respectively. Particle size, measured as geometric
median number and volume diameters, was 85 and 220 nm, respectively. No DE-related effects were
found for breathing parameters measured by whole-body plethysmography. Other pulmonary effects
are described in Sections 7.3.3.2 and 7.3.5.1.
Woodsmoke
One study evaluated the effects of subchronic woodsmoke exposure on pulmonary function in
Brown Norway rats. Rats were exposed 3 h/day and 5 days/week for 4 and 12 wk to air or to
concentrated wood smoke from the piny on pine which is native to the U.S. Southwest (Tesfaigzi et
al., 2002, 025575). PM concentrations in the woodsmoke were 1,000 and 10,000 (ig/m . The
particles in this woodsmoke had a bimodal size distribution with the smaller size fraction (74%)
characterized by a MMAD of 0.405 (im and the larger size fraction (26%) characterized by a
MMAD of 6.7-11.7 (im. Many of these larger particles would not be inhalable by the rat since 8 (im
MMAD particles are about 50% inhalable (Menache et al., 1995, 006533). Concentrations of gases
were reported to be 15-106.4 ppm CO, 2.2-18.9 ppm NO, 2.4-19.7 ppm NOX and 3.5-13.8 ppm total
hydrocarbon in these exposures. Respiratory function measured by whole-body plethysmography
demonstrated a statistically significant increase in total pulmonary resistance in rats exposed to
1000 (ig/m3 woodsmoke. Additional effects were found at 10,000 (ig/m3. Inflammatory and
histopathological responses were also evaluated (Sections 7.3.3.2 and 7.3.5.1).
December 2009 7-31
-------
7.3.3. Pulmonary Inflammation
7.3.3.1. Epidemiologic Studies
One epidemiologic study examined the relationship of airway inflammation (eNO) and PM
measures, other pollutants, and indicators of motor vehicle emissions in Windsor, Ontario (Dales et
al., 2008, 156378). This cohort of 2,402 school children estimated PM2.5 and PMi0.2.5 for each child's
residence at the postal code level with an evaluated statistical model (Wheeler et al., 2006, 103905).
Each 10 ug/m3 increase in 1-yr PM25 was associated with a 39% increase in eNO (p = 0.058).
Associations between eNO and PMi0_2.5 were positive but not statistically significant.
7.3.3.2. lexicological Studies
CAPs Studies
A set of subchronic studies involved exposure of normal (C57BL1/6) mice, ApoE"7" and the
double-knockout ApoE"7"/LDLR"7" mice to Tuxedo, NY CAPs for 5-6 month (March, April or May
through September 2003 (Lippmann et al., 2005, 087452). The average PM25 exposure
concentration was 110 ug/m . Animals were fed a normal chow diet during the CAPs exposure
period. No pulmonary inflammation was observed in response to CAPs exposure as measured by
BALF cell counts and histology. The lack of a persistent pulmonary response may have been due to
adaptation of the lung following repeated exposures. In fact, a parallel study examined CAPs-related
gene expression in the double-knockout animals and found upregulation of numerous genes in lung
tissue (Gunnison and Chen, 2005, 087956). An in vitro study conducted simultaneously found daily
variations in CAPs-mediated NF-KB activation in cultured human bronchial epithelial cells,
suggesting that transcription factor-mediated gene upregulation could occur in response to CAPs
(Maciejczyk and Chen, 2005, 087456). It should be noted that significant cardiovascular effects
were observed in these subchronic studies which are discussed in Section 7.2.1.2.
Araujo et al. (2008, 156222) compared the relative impact of UF (0.01-0.18 um) versus fine
(0.01-2.5 um) PM inhalation in ApoE"7" mice following a 40 day exposure (5 h/day><3 days/wk for 75
total hours). Animals were fed a normal chow diet and exposed to PM from November 3 -December
12, 2005 in a mobile inhalation laboratory that was parked 300 m from the 110 Freeway in
downtown Los Angeles. Particles were concentrated to -440 ug/m3 for PM2 5 exposures and
-110 ug/m3 for the UF exposures, representing a roughly 15-fold increase in concentration from
ambient levels; the number concentration of PM in the fine and UF chambers were roughly
equivalent (4.56xl05 and 5.59xl05 particles/cm3, respectively). Over 50% of the UFPs were
comprised of OC compared to only 25% for PM25. No major increase in BALF inflammatory cells
was found in response to PM. However UFP exposure resulted in significant cardiovascular and
systemic effects (Section 7.2.1.2).
Diesel Exhaust
Gottipolu et al. (2009, 190360) exposed WKY and SH rats to filtered air or DE for 4 wk as
described in Section 7.3.2.2. Previous studies from this laboratory have shown enhanced effects of
PM in SH compared with WKY rats. Although the main focus of this recent study was on DE-
induced mitochondrial oxidative stress and hypertensive gene expression in the heart
(Section 7.2.7.1), some pulmonary effects were also found. Subchronic exposure to DE resulted in a
dose-dependent increase in BALF neutrophils in both rat strains although levels of measured
cytokines were not altered. Histological analysis of lung tissue from rats exposed to the higher
concentration of DE demonstrated accumulation of particle-laden macrophages as well as focal
alveolar hyperplasia and inflammation. Effect on indices of injury are discussed in Section 7.3.5.1.
Ishihara and Kagawa (2003, 096404) exposed Wistar rats to filtered air and DE containing
200, 1,000 and 3,000 ug/m3 PM for 16 h/day and 6 days/wk for 6, 12, 18 or 24 mo. The mass median
particle diameter was reported to be between 0.3 and 0.5 um. Concentrations of gases ranged from
December 2009 7-32
-------
2.93-35.67 ppmNOx, 0.23-4.57 ppm SO2, 1.8-21.9 ppm CO in the DE exposures. Statistically
significant increases in total numbers of inflammatory cells and neutrophils in BALF were observed
beginning at 6-12 mo of exposure to DE containing 1,000 and 3,000 (ig/m3 PM. When rats were
exposed to DE containing 1,000 (ig/m3 PM, which was filtered to remove PM, the inflammatory cell
response was significantly diminished. These results implicate the PM fraction of DE as a key
determinant of the inflammation. The PM fraction was also found to mediate the increase in protein
levels, the decrease in PGE2 levels and alterations in mucus and surfactant components observed in
BALF (Section 7.3.5.1).
Li et al. (2007, 155929) exposed BALB/c and C56BL/6 mice to low dose DE as described in
Section 7.3.2.2. for 1, 4 and 8 wk. Increases in numbers of BALF macrophages and total
inflammatory cells were observed in BALB/c mice at 8 wk but not 4 wk of DE exposure. Persistent
increases in numbers of BALF neutrophils and lymphocytes were observed in both strains at 4 and
8 wk of DE exposure. Corresponding increases in BALF cytokines differed between the two strains.
These results should be interpreted with caution since comparisons were made with Day 0 controls
rather than age-matched controls. No histopathological changes in the lungs were seen at any time
point after DE exposure. This study demonstrated differences in pulmonary responses to low dose
DE between two mouse strains. AHR, pulmonary inflammation, markers of oxidative stress and
effects of antioxidant intervention were also evaluated (Sections 7.3.2.2 and 7.3.4.1). Although no
attempt was made in this study to determine the effects of gaseous components of DE on the
measured responses, concentrations of gases were very low suggesting that PM may have been
responsible for the observed effects.
In a study by Hiramatsu et al. (2003, 155846). BALB/c and C57BL/6 mice were exposed to
DE (PM concentrations 100 and 3,000 (ig/m3) for 1 or 3 mo. Concentrations of gases were reported
to be 3.5-9.5 ppm CO, 2.2-14.8 ppm NOX, and less than 0.01 ppm SO2. Modest increases in BALF
neutrophils and lymphocytes were observed in response to DE in both mouse strains at 1 and 3 mo.
Histological analysis demonstrated diesel exposure particle-laden alveolar macrophages in alveoli
and peribronchial tissues at both time points. Bronchus-associated lymphoid tissue developed after
3-month exposure to the higher concentration of DE in both mouse strains. Mac-1 positive cells (a
marker of phagocytic activation of alveolar macrophages) were also increased in BALF of BALB/c
mice exposed to the higher concentration of DE for 1 and 3 mo. Increased expression of several
cytokines and decreased expression of iNOS mRNA was observed in DE-exposed mice at 1 and
3 mo. NF-KB activation was also noted following 1-month exposure to the lower concentration of
DE. No attempt was made in this study to determine the responses to gaseous components of the DE.
In a study by Reed et al. (2004, 055625). healthy Fisher 344 rats and A/J mice were exposed to
DE (PM concentration = 30, 100, 300 and 1,000 (ig/m3) by whole body inhalation for 6 h/day,
7 days/wk for either 1 week or 6 mo. Concentrations of gases were reported to be 2.0-45.3 ppm NO,
0.2-4.0 ppmNO2, 1.5-29.8 ppm CO and 8-365 ppb SO2. Short-term responses are discussed in
Section 6.3.3.3 and 6.3.7.2, and sub-chronic systemic effects are presented in Section 7.2.4.1. Six
months of exposure resulted in no measurable effects on pulmonary inflammation. However
numerous black particles were observed within alveolar macrophages after 6 mo of exposure.
Seagrave et al. (2005, 088000) evaluated pulmonary responses in male and female CDF
(F-344)/CrlBR rats exposed 6 h/day for 6 mo to filtered air or DE at concentrations ranging from
30-1000 (ig/m3 PM. Concentrations of gases were reported for the highest exposure as 45.3 ppm NO,
4.0 ppm NO2, 29.8 ppm CO and 2.2 ppm total vapor hydrocarbon. No changes in BALF cells were
noted. A small decrease in TNF-a was seen in BALF of female rats exposed to the highest
concentration of DE for 6 mo. Pulmonary injury also was evaluated (Section 7.3.5.1). Thus changes
in BALF markers were modest and gender-specific.
Woodsmoke
Seagrave et al. (2005, 088000) also evaluated pulmonary responses in male and female CDF
(F344)/CrlBR rats exposed 6 h/day for 6 mo to filtered air or hardwood smoke concentrations
ranging from 30-1,000 (ig/m3 PM. Concentrations of gases were reported for the highest exposure as
3.0 ppm CO and 3.1 ppm total vapor hydrocarbon. A small increase in BALF neutrophils was
observed in male rats exposed to the lowest concentration of hardwood smoke. Female rats exhibited
a decrease in BALF macrophage inflammatory protein-2 (MIP-2) at the highest concentration of
hardwood smoke. Pulmonary injury also was evaluated (Section 7.3.5.1). In general, responses to
December 2009 7-33
-------
hardwood smoke were more remarkable than responses to DE seen in a parallel study. However
these gender-specific responses were modest and difficult to interpret.
In a study by Reed et al. (2006, 156043). Fisher 344 rats, SHR rats, A/J mice and
C57BL/6 mice were exposed to clean air or hardwood smoke (PM concentrations 30, 100, 300 and
1,000 (ig/m3) by whole body inhalation for 6 h/day, 7 days/wk for either 1 week or 6 mo.
Concentrations of gases ranged from 229.0-14887.6 mg/m3 for CO, 54.9-139.3 (ig/m3 for ammonia,
and 177.6- 3455.0 (ig/m3 nonmethane VOC in these exposures. Short-term responses are discussed
in Section 6.3.7.2 and sub-chronic effects are presented in Section 7.2.4.1. Histological analysis of
lung tissue showed minimal increases in alveolar macrophages. The effects of hardwood smoke on
bacterial clearance are discussed below (Section 7.3.7.2).
Another study evaluated the effects of subchronic woodsmoke exposure in Brown Norway rats
and is described in detail in Section 7.3.2.2 (Tesfaigzi et al., 2002, 025575). Numbers of alveolar
macrophages in BALF were significantly increased in rats exposed to 1,000 (ig/m3 woodsmoke for
12 wk, but no changes were seen in numbers of other inflammatory cells. A large percent of BALF
macrophages contained carbonaceous material. Histological analysis of lung tissue showed minimal
to mild inflammation in the epiglottis of the larynx in rats exposed to both concentrations of
woodsmoke.
Ramos et al. (2009, 190116) examined the effects of subchronic woodsmoke exposure on the
development of emphysema in guinea pigs. Inflammation is thought to be involved in the
pathogenesis of this form of COPD. Statistically significant increases in total numbers of BALF cells
were observed in guinea pigs exposed to smoke for 1-7 mo, with numbers of macrophages increased
at 1-4 mo and numbers of neutrophils increased at 4-7 mo. At 4 mo, alveolar mononuclear
phagocytic and lymphocytic peribronchiolar inflammation were observed by histological analysis of
lung tissue. This study is discussed in depth in Section 7.2.5.1.
Model Particles
Wallenborn et al. (2008, 191171) examined the pulmonary, cardiac and systemic effects of
subchronic exposure to particulate ZnSO4. WKY rats were exposed nose-only to 10, 30, or
100 ug/m3UFP of ZnSO4 for 5 h/day and 3 day/wk over a 16-wk period. Particle size was reported
to be 31-44 nm measured as number median diameter. No changes in pulmonary inflammation or
injury were observed although cardiac effects were noted (Section 7.2.7.1). This study possibly
demonstrates a direct effect of ZnSO4 on extrapulmonary systems, as suggested by the lack of
pulmonary effects.
7.3.4. Pulmonary Oxidative Response
7.3.4.1. lexicological Studies
Urban Air
One new study evaluated the effects of subchronic exposure to ambient levels of urban
particles on the development of emphysema in papain-treated mice (Lopes et al., 2009, 190430).
Since oxidative stress is thought to contribute to the development of emphysema, 8-isoprostane
levels were measured in lung tissue from the four groups of mice used in this study. A statistically
significant increase in 8-isoprostane, a marker of oxidative stress, was observed in lungs from mice
treated with papain and exposed to ambient air compared with the other groups of mice. This study
is described in greater depth in Section 7.3.5.1.
December 2009 7-34
-------
Diesel Exhaust
Li et al. (2007, 155929) exposed mice to low dose DE for 1, 4 and 8 wk as described in
Section 7.3.2.2. Markers of oxidative stress and effects of antioxidant intervention were evaluated in
this model. While HO-1 mRNA and protein were increased in lung tissues of both mouse strains
after 1 week of DE exposure (Section 6.3.4.2), at 8 wk of DE exposure, HO-1 protein levels
remained high in C57BL/6 mice but returned to control values in BALB/c mice. This study
demonstrates differences in pulmonary responses to low dose DE between two mouse strains.
Furthermore, this study suggests that adaptation occurs in BALB/c mice during prolonged DE
exposure since the increase in HO-1 protein seen in both strains at 1 week of exposure was only seen
in C57BL/6 mice at 8 wk. AHR (Section 7.3.2.2) and pulmonary inflammation (Section 7.3.3.2)
were also evaluated. Although no attempt was made in this study to determine the effects of gaseous
components of DE on the measured responses, concentrations of gases were very low. This suggests
that PM may have been responsible for the observed effects.
7.3.5. Pulmonary Injury
7.3.5.1. lexicological Studies
Urban Air
One new study evaluated the effects of chronic exposure to ambient levels of urban particles
on lung development in the mouse (Mauad et al., 2008, 156743). Both functional and anatomical
indices of lung development were measured in mice exposed prenatally and/or postnatally as
described in Section 7.3.2.2. Animals were sacrificed at 15 and 90 days of age for histological
analysis of lungs. Histological analysis demonstrated the presence of mild foci of macrophages
containing black dots of carbon pigment in the prenatal and postnatal exposure group at 90 days. In
addition, the alveolar spaces of 15-day old mice in the prenatal and postnatal exposure group were
enlarged compared with controls. Morphometric analysis demonstrated statistically significant
decreases in surface to volume ratio at 15 and 90 days in the prenatal and postnatal exposure group
compared with controls. Since alveolarization is normally complete by 15 days of age, these results
suggest incomplete alveolarization in the 15-day-old group and an enlargement of air spaces in the
90-day-old group. These anatomical changes correlated with decrements in pulmonary function
which are discussed in Section 7.3.2.2.
Prolonged exposure to low levels of ambient air pollution beginning in early life has been
linked to secretory changes in the nasal cavity of mice, specifically increased production of acidic
mucosubstances (Pires-Neto et al., 2006, 096734). Six-day-old Swiss mice were continuously
chamber exposed to ambient or filtered Sao Paulo air for 5 mo. Concentrations in the "polluted
chamber" versus "clean chamber" were (in (ig/m3) 59.52 versus 37.08 for NO2, 12.52 versus 0 for
BC, and 46.49 versus 18.62 for PM2.5. Thus, pollutant levels were reduced by filtration but not
entirely eliminated. Compared to filtered air, exposure to ambient air resulted in increased total
mucus and acidic mucus in the epithelium lining the nasal septum, but no statistically significant
differences in other parameters (amount of neutral mucus, volume proportions of neutral mucus,
total mucus, or nonsecretory epithelium, epithelial thickness, or ratio between neutral and acidic
mucus). The physicochemical properties of mucus glycoproteins are critical to the protective
function of the airway mucus layer. Acidified mucus is more viscous, and is associated with a
decrease in mucociliary transport. Thus acidic mucosubstances may represent impaired defense
mechanisms in the respiratory tract.
One new study evaluated the effects of subchronic exposure to ambient levels of urban
particles on the development of emphysema in papain-treated mice (Lopes et al., 2009, 190430).
Emphysema is a form of COPD caused by the destruction of extracellular matrix in the alveolar
region of the lung which results in airspace enlargement, airflow limitation and a reduction of the
gas-exchange area of the lung. Inflammation, oxidative stress, protease imbalance and apoptosis are
thought to contribute to the development of emphysema. In this study, male BALB/c mice were
December 2009 7-35
-------
continuously exposed to ambient or filtered Sao Paulo air for 2 mo. Concentrations of PM2.5 in the
"polluted chamber" versus "clean chamber" were 33.86 ± 2.09 versus 2.68 ± 0.38 (ig/m3. Thus
filtration reduced PM levels considerably. Ambient concentrations of CO and SO2 were 1.7 ppm and
16.2 (ig/m3 respectively. No significant difference was observed in the concentrations of NO2 in the
"polluted chamber" versus "clean chamber" (60-80 (ig/m3). Half of the mice were pre-treated with
papain by intranasal instillation in order to induce emphysema. Morphometric analysis of lung tissue
demonstrated a statistically significant increase in mean linear intercept, a measure of airspace
enlargement, in papain-treated mice compared with saline-treated controls exposed to filtered air.
While exposure to ambient air failed to increase mean linear intercept values in saline-treated mice,
mean linear intercept values were significantly increased in papain-treated mice exposed to ambient
air compared with papain-treated mice exposed to filtered air. A similar pattern of responses was
observed for the volume proportion of collagen and elastin fibers in alveolar tissue, which are
markers of alveolar wall remodeling. Lung immunohistochemical analysis demonstrated an effect of
papain, but not ambient air, on macrophage cell density and matrix metalloproteinase 12-positive
cell density. No differences in caspase-3 positive cells, a marker of apoptosis, were observed
between the four groups of mice. Oxidative stress was evaluated in this model as described in
Section 7.3.4.1. Taken together, results of this study demonstrate that urban levels of PM, mainly
from traffic sources, worsen protease-induced emphysema in an animal model.
Pulmonary vascular remodeling, measured by a decrease in the lumen to wall ratio, was
observed in mice exposed to ambient Sao Paulo air for 4 mo (Lemos et al., 2006, 088594). This
study is described in greater detail in Section 7.2.1.2.
Kato and Kagawa (2003, 089563) exposed Wistar rats to roadside air contaminated mainly
with automobile emissions (55.7-65.2 ppb NO2 and 63-65 (ig/m3 suspended PM [SPM]) and
examined the effects on respiratory tissue after 24, 48, or 60 wk of exposure. The surface of the
lungs was light gray in color after all durations of exposure, and BC particle deposits accumulated
with prolonged exposure. These characteristics were not evident in filtered air-exposed control
animals, although filtered air contained low levels of air pollutants (< 6.2 ppb NO2 and 15 (ig/m3
SPM). The most common change observed using transmission electron microscopy was the presence
of particle laden (anthracotic) alveolar macrophages, or anthracosis, in a wide range of pulmonary
tissues, including the submucosa, tracheal- and bronchiole-associated lymph nodes, alveolar wall
and space, pleura, and perivascular connective tissue. These changes were evident after 24 wk and
increased with duration of exposure. Other changes included increases in the number of mucus
granules in goblet cells, mast cell infiltration (but no degranulation) after 24 wk, increased
lysosomes in ciliated cells, some altered morphology of Clara cells, and hypertrophy of the alveolar
walls after 48 wk. No goblet cell proliferation was observed, but slight, variable acidification of
mucus granules appeared after 24 and 48 wk and disappeared after 60 wk. Anthracotic macrophages
were seen in contact with plasma cells and lymphocytes in the lymphoid tissue, suggesting immune
cell interaction in the immediate vicinity of particles. Even after 60 wk, no lymph node anthracosis
was observed in the filtered air group.
In a post-mortem study of lung tissues from 20 female lifelong residents of Mexico City, a
high PM locale, histology demonstrated significantly greater amounts of fibrous tissue and muscle in
the airway walls compared to subjects from Vancouver (Churg et al., 2003, 087899). a city with
relatively low PM levels. Electron microscopy showed carbonaceous aggregates of UFPs, which the
authors conclude penetrate into and are retained in the walls of small airways. The study shows an
association between retained particles and airway remodeling in the form of excess muscle and
fibrotic walls. The subjects were deemed suitable for examination based on never-smoker status, no
use of biomass fuels for cooking, no known occupational particle/dust exposure, death by cause
other than respiratory disease, and extended residence in each locale (lifelong for Mexico City and
>20 yr for Vancouver). However, subjects from the two locales were not matched with respect to
ethnicity, sex (20 females from Mexico City versus 13 females and 7 males from Vancouver), or
mean age at death (66 ± 9 versus 76 ± 11), and other possibly influential factors such as exercise or
diet were not considered.
Diesel Exhaust
Gottipolu et al. (2009, 190360) exposed WKY and SH rats to filtered air or DE as described in
Section 7.3.2.2. Previous studies from this laboratory have shown enhanced effects of PM in SH
December 2009 7-36
-------
compared with WKY rats. Although the main focus of this recent study was on DE-induced
mitochondrial oxidative stress and hypertensive gene expression in the heart (Section 7.2.7.1), some
pulmonary effects were found. Inflammatory effects are described in Section 7.3.3.2. GGT activity
in BALF was increased in both strains in response to the higher concentration of DE. No DE-related
changes were observed in BALF protein or albumin. Histological analysis of lung tissue from rats
exposed to the higher concentration of DE demonstrated accumulation of particle-laden
macrophages as well as focal alveolar hyperplasia and inflammation. No effects on indices of
pulmonary function were observed (Section 7.3.2.2.)
Ishihara and Kagawa (2003, 096404) exposed rats to DE for up to 24 mo as described in
Section 7.3.3.2. A statistically significant increase in BALF protein was observed at 12 mo of
exposure to DE containing 1,000 ug/m3 PM. This response was attenuated when the DE was filtered
to remove PM. Pulmonary inflammation was noted and is described in Section 7.3.3.2.
Seagrave et al. (2005, 088000) evaluated pulmonary responses in rats exposed to DE for up to
6 mo as described in Section 7.3.3.2. A small increase in LDH was seen in BALF of female rats
exposed to the highest concentration of DE for 6 mo. Pulmonary inflammation was also evaluated
(Section 7.3.3.2). The changes in BALF markers in this study were modest and gender-specific.
Gasoline Exhaust
Reed et al. (2008, 156903) examined a variety of health effects following subchronic
inhalation exposure to gasoline engine exhaust. Male and female CDF (F344)/CrlBR rats, SHR rats
and male C57BL/6 mice were exposed for 6 h/day and 7 days/wk for a period of 3 days-6 mo. The
dilutions for the gasoline exhaust were 1:10, 1:15 and 1:90; filtered PM was at the 1:10 dilution. PM
mass ranged from 6.6 to 59.1 ug/m3, with the corresponding number concentration between 2.6><104
and 5.0><105 particles/cm3. Concentrations of gases ranged from 12.8-107.3 ppm CO, 2.0-17.9 ppm
NO, 0.1-0.9 ppmNO2, 0.09-0.62 ppm SO2 and 0.38-3.37 ppmNH3. Other effects are described in
Sections 7.2.4.1 and 7.3.6.1. No pulmonary inflammation or histopathological changes were noted in
the F344 rats and A/J mice, except for a time-dependent increase in the number of macrophages
containing PM. However statistically significant increases of 47% and 29% in BALF LDH were
observed in female and male F344 rats, respectively, after 6 mo of exposure to the highest
concentration of engine exhaust. This response was absent when gasoline exhaust was filtered,
implicating PM as a key determinant of this response. In addition, exposure to the highest
concentration of gasoline exhaust resulted in statistically significant decreases in hydrogen peroxide
and superoxide production in unstimulated and stimulated BALF macrophages. Hypermethylation of
lung DNA was observed in male F344 rats following 6 mo of exposure to gasoline exhaust
containing 30 ug/m3 PM. This response was PM-dependent since it was absent in mice exposed to
filtered gasoline exhaust. The significance of this epigenetic change in terms of respiratory health
effects is not known. However, altered patterns of DNA methylation can affect gene expression and
are sometimes associated with altered immune responses and/or the development of cancer.
Woodsmoke
Seagrave et al. (2005, 088000) also evaluated pulmonary responses in rats exposed to
hardwood smoke for 6 mo as described in Section 7.3.3.2. Increases in BALF LDH and protein were
seen in male but not female rats. Female rats exhibited a decrease in BALF glutathione at the highest
concentration of hardwood smoke. Decreases in BALF alkaline phosphatase were found in both
males and females exposed to 1,000 ug/m3 hardwood smoke. Male rats exposed to 100 and
300 ug/m3 hardwood smoke exhibited a decrease in BALF |3-glucuronidase activity. Pulmonary
inflammation was also evaluated (Section 7.3.3.2). These changes in BALF markers in this study
were modest and gender-specific.
Another study evaluated the effects of subchronic woodsmoke exposure in Brown Norway rats
as described in Section 7.3.2.2. (Tesfaigzi et al., 2002, 025575). Exposure to 1,000 ug/m3
woodsmoke for 12 wk resulted in a statistically significant increase in Alcian Blue- (AB) and
Periodic Acid Schiff- (PAS) positive airway epithelial cells compared to controls, indicating an
increase in mucous secretory cells containing neutral and acid mucus, respectively. More significant
histopathological responses were found following exposure to 10,000 ug/m3 of DE. Pulmonary
December 2009 7-37
-------
function and inflammation were evaluated also but are not discussed here due to the extremely high
exposure level (Sections 7.3.2.2. and 7.3.3.2).
Ramos et al. (2009, 190116) examined the effects of subchronic woodsmoke exposure on the
development of emphysema in guinea pigs. In particular, the involvement of macrophages and
macrophage-derived MMP in woodsmoke-related responses was investigated. Guinea pigs were
exposed to ambient air or to whole smoke from pine wood for 3 h/day and 5 days/wk over a 7-month
period. PMi0 and PM2.5 concentrations in the exposure chambers were reported to be 502 ± 34 and
363 ± 23 (ig/m3, respectively, while the concentration of CO was less than 80 ppm. COHb levels
were reported to be 6% in controls and 15-20% in smoke-exposed guinea pigs. Statistically
significant decreases in body weight were observed in guinea pigs exposed to smoke for 4 or more
months compared with controls. Statistically significant increases in total numbers of BALF cells
were observed in guinea pigs exposed to smoke for 1-7 mo, with numbers of macrophages increased
at 1-4 month and numbers of neutrophils increased at 4-7 mo. At 4 mo, alveolar mononuclear
phagocytic and lymphocytic peribronchiolar inflammation, as well as bronchiolar epithelial and
smooth muscle hyperplasia, were observed by histological analysis of lung tissue. Emphysematous
lesions, smooth muscle hyperplasia and pulmonary arterial hypertension were noted at 7 mo.
Morphometric analysis of lung tissue demonstrated statistically significant increases in mean linear
intercept values, a measure of airspace enlargement, in guinea pigs at 6 and 7 mo of exposure.
Statistically significant increases in elastolytic activity was observed in BALF macrophages and lung
tissue homogenates at 1-7 mo of exposure. Lung collagenolytic activity was also increased at 4-7 mo
of exposure and corresponded in time with the presence of active forms of MMP-2 and MMP-9 in
lung tissue homogenates and BALF. Furthermore, MMP-1 and MMP-9 immunoreactivity was
detected in macrophages, epithelial and interstitial cells in smoke-exposed animals at 7 mo.
Increased levels of MMP-2 and MMP-9 mRNA were also found in smoke-exposed guinea pigs after
3-7 mo. Apoptosis was found in BALF macrophages (TUNEL assay) from guinea pigs exposed to
smoke for 3-7 mo and in alveolar epithelial cells (caspase-3 immunoreactivity) after 7 mo. Taken
together, these results provide evidence that subchronic exposure to woodsmoke leads to the
development of emphysematous lesions accompanied by the accumulation of alveolar macrophages,
increased levels and activation of MMPs, connective tissue remodeling and apoptosis. However, the
high levels of CO and COHb reported in this study make it difficult to conclude that woodsmoke PM
alone is responsible for these dramatic effects.
7.3.6. Allergic Responses
7.3.6.1. Epidemiologic Studies
A number of epidemiologic studies have found associations between PM and allergic (or
atopic) indicators. Allergy is a major driver of asthma, which has been associated with PM in studies
discussed in previous sections. In a study by Annesi-Maesano (2007, 093180) (described in
Section 7.3.1.1) atopic asthma was related to PM2.5 (OR 1.43 [95% CI: 1.07-1.91]) and positive skin
prick test to common allergens was also increased with higher PM levels. This report is consistent
with the results from an earlier study (Penard-Morand et al., 2005, 087951) in the same sample of
children that associated allergic rhinitis and atopic dermatitis with PMi0. Also, Morgenstern et al.
(2008, 156782) found statistically significant effects for asthmatic bronchitis, hay fever, and allergic
sensitization to pollen in a cohort of children in Germany examining modeled PM2 5 data at birth
addresses. Distance to a main road had a dose-response relationship with sensitization to outdoor
allergens. Nordling et al. (2008, 097998) (discussed above in Section 7.3.2.1) reported a positive
association of PM10 exposure during the first year of life with allergenic sensitization (IgE
antibodies) to inhaled allergens, especially pollen. In a study by Brauer et al. (2007, 090691)
(discussed above in Section 7.3.1.1) an interquartile range increase in PM2.5 was associated with an
increased risk of sensitization to food allergens (OR 1.75 [95% CI 1.23-2.47]). A significant
association was found for sensitization to any allergen, but none was found for sensitization to
specific indoor or outdoor aeroallergens or atopic dermatitis (eczema). In a study by Janssen et al.
(2003, 133555). PM2 5 was associated with allergic indicators such as hay fever (ever), skin prick test
reactivity to outdoor allergens, current itchy rash, and conjunctivitis in Dutch children. These same
outcomes were also associated with proximity of the school to truck traffic but not car traffic,
December 2009 7-38
-------
suggesting a role for diesel-related pollution. Consistent with the aforementioned Dutch study by
Brauer et al. (2007, 090691). PM2 5 was not associated with eczema.
Mortimer et al. (2008, 187280) examined the association between prenatal and early-life
exposures to air pollutants with allergic sensitization in a cohort of 170 children with asthma, ages
6-11 yr, living in central California. Sensitization to at least one allergen was associated with higher
levels of PMio and CO during the entire pregnancy and 2nd trimester and higher PMi0 during the
first 2 yr of life. Sensitization to at least one indoor allergen was associated with higher exposures to
PMio and CO in during the entire pregnancy and during the 2nd trimester. However, no significant
associations remained for PMio after adjustment for copollutants, effect modifiers, or potential
cofounders in addition to year of birth. The authors advise that the large number of comparisons may
be of concern and this study should be viewed as an exploratory, hypothesis-generating undertaking.
In examining the National Health Interview Survey for the years 1997-2006, Bhattacharyya et al.
(2009, 180154) found relationships between air quality and the prevalence of hay fever and sinusitis.
However, the air quality data were not clearly defined and as such caution is required in
interpretation of these results. In contrast, Bayer-Oglesby et al. (2005, 086245) found no significant
association between declining levels of PMio and hay fever in Switzerland. In a study by Oftedal et
al. (2007, 191948) conducted in Oslo, Norway, early-life exposure to PMio or PM2.5 was generally
not associated with sensitization to allergens in 9- to 10-yr-old children; lifetime exposures to PMio
and PM2.5 were associated with dust mite allergy, but the association was diminished by adjustment
for socioeconomic factors . In Norway, wood burning in the wintertime is thought to account for
about half of the PM2.5 levels. Although associations between PM and reactivity to specific allergens
have been reported in long-term studies, there is a consistent lack of correlation between PM and
total IgE levels, indicating a selective enhancement of allergic responses.
7.3.6.2. lexicological Studies
Diesel Exhaust
Exposure to relatively low doses of DE has been shown to exacerbate asthmatic responses in
ovalbumin (OVA) sensitized and challenged BALB/c mice (Matsumoto et al., 2006, 098017). Mice
were intraperitoneally sensitized and intranasally challenged 1 day prior to inhalation exposure to
DE (PM concentration 100 ug/m3; CO, 3.5 ppm; NO2, 2.2 ppm; SO2 O.01 ppm) for 1 day or 1, 4, or
8 wk (7/h/day, 5 days/wk, endpoints 12 h post DE exposure). Results from the 1- and 4-wk
exposures are described in Section 6.3.6.3. It should be noted that control mice were left in a clean
room as opposed to undergoing chamber exposure to filtered air. The significant increases in AHR
and airway sensitivity observed following snorter exposure periods did not persist at 8 wk. BALF
cytokines were altered by DE exposure with only RANTES significantly elevated after 8 wk. DE
had no effect on OVA challenge-induced peribronchial inflammatory or mucin positive cells. These
results suggest that adaptive processes may have occurred during prolonged exposure to DE.
Gasoline Exhaust
In a study by Reed et al. (2008, 156903). BALB/c mice were exposed to whole gasoline
exhaust diluted 1:10 (H), 1:15 (M), or 1:90 (L), filtered exhaust at the 1:10 (HF), or clean air for
6 h/day (atmospheric characterization described in Section 6.3.6.3). GEE exposure from conception
through 4 wk of age induced slight but non-significant increases in OVA-specific IgGl in offspring
but had no significant effect on airway reactivity, BALF cytokine or cell concentrations, although
there were non-significant increases in lung neutrophils and eosinophils. Significant increases in
total serum IgE were observed, but this effect persisted after filtration of particles and was thus
attributed to gas phase components.
December 2009 7-39
-------
Woodsmoke
In a study by Tesfaigzi et al. (2005, 156116). Brown Norway rats were sensitized and
challenged with OVA. Rats were exposed for 70 days to filtered air or to 1,000 (ig/m3 hardwood
smoke. Particles were characterized by aMMAD of 0.36 (im. Concentrations of gases were reported
to be 13.0 ppm CO and 3.1 ppm total vapor hydrocarbon with negligible NOX. Respiratory function
was measured in anesthetized animals by whole-body plethysmography and demonstrated a
significant increase in functional residual capacity as well as a significant increase in dynamic lung
compliance in hardwood smoke-exposed animals compared to controls. No change in total
pulmonary resistance or airway responsiveness to methacholine was observed. BALF inflammatory
cells were not increased, although histological analysis demonstrated focal inflammation including
granulomatous lesion and eosinophilic infiltrations in hardwood smoke-exposed rats. Alterations of
several cytokines in BALF and plasma were noted. Changes in airway epithelial mucus cells and
intraepithelial stored mucosubstances were modest and did not achieve statistical significance.
Results of this study demonstrate that subchronic exposure to hardwood smoke had minimal effects
on pulmonary responses in a rat model of allergen sensitization and challenge.
7.3.7. Host Defense
7.3.7.1. Epidemiologic Studies
Epidemiologic studies of respiratory infections indicate an association with PM. This is more
evident when considering short-term exposures (Chapter 6), but studies of long-term exposures have
observed associations with general respiratory symptoms often caused by infection, such as
bronchitis. In a birth cohort study of approximately 4,000 Dutch children, Brauer et al. (2007,
09069Prescribed in Section 7.3.1.1) found significant positive associations for PM2.5 with
ear/nose/throat infections and doctor-diagnosed flu/serious cold in the first 4 yr of life. These results
are consistent with an earlier study by Brauer et al. (2006, 090757). which found that an increase of
10 (ig/m3 PM25 was associated with increased risk for ear infections in the Netherlands [OR 1.50
(95% CI, 1.00-2.22)]. A Swiss study by Bayer-Oglesby et al. (2005, 086245). discussed in
Section 7.3.1.1 above, demonstrated that declining levels of PMi0 were associated with declining
prevalence of common cold and conjunctivitis. Because traffic-related pollutants such as UFPs are
high near major roadways and then decay exponentially over a short distance, Williams, et al. (2009,
191945) assessed exposure according to residential proximity to major roads in a Seattle area study
of postmenopausal women. Proximity to major roads was associated with a 21% decrease in natural
killer cell function, which is an important defense against viral infection and tumors. This finding
was limited to women who reported exercising near traffic; other markers of inflammation and
lymphocyte proliferation did not consistently differ according to proximity to major roads. In the
Puget Sound region of Washington, Karr et al. (2009, 191946) reported that there may be a modest
increased risk of bronchiolitis related to PM2.5 exposure for infants born just before the peak
respiratory syncytial virus (RSV) season. Risk estimates were stronger when restricted to cases
specifically attributed to RSV and for infants residing closer to highways. Emerging evidence
suggests that respiratory infections, particularly infection by viruses such as RSV, can cause asthma
or trigger asthma attacks.
7.3.7.2. lexicological Studies
Diesel Exhaust
DE may affect systemic immunity. The proliferative response of A/J mouse spleen cells
following stimulation with T cell mitogens was suppressed by 6 mo of daily exposure to DE at
concentrations at or above 300 (ig/m3 PM (Burchiel et al., 2004, 055557). B cell proliferation was
increased at 300 (ig/m3 but unaffected at higher concentrations (up to 1,000 (ig/m ). Concentrations
of gases and were reported in the parallel study by Reed et al. (2004, 055625). described in
December 2009 7-40
-------
Section 7.3.3.2. The Reed study reported a decrease in spleen weight in male mice (27% reduction in
the 300 (ig/m3 exposure group). The immunosuppressive effects of DE were not due to PAHs or
benzo(a)pyrene (BaP)-quinones (BPQs) since there were little, if any, of these compounds present in
the chamber atmosphere. It should be noted that sentinel animals were negative for mouse
parvovirus at the start of the study, but seroconverted by the end of the study, indicating possible
infection. Parvovirus can interfere with the modulation of lymphocyte mitogenic responses (Baker,
1998, 156245). A 6-month exposure (6h/day, 7d/wk) to 30, 100, 300 or 1,000 (ig/m3 of PM in DE did
not significantly affect bacterial clearance in C57BL/6 mice infected withPseudomonas aeruginosa,
although all levels reduced bacterial clearance when the exposure only lasted a week (Harrod et al.,
2005, 088144). Characterization of the exposure atmosphere was given by Reed et al. (2004,
055625) (Section 7.3.3.2.).
Gasoline Exhaust
In a study by Reed et al. (2008, 156903) (described in Section 6.3.7.2) long-term exposure to
fresh gasoline exhaust (6h/day, 7d/wk for 6 mo) did not affect clearance of P. aeruginosa from the
lungs of C57BL/6 mice.
Hardwood Smoke
One study demonstrated immunosuppressive effects of hardwood smoke (Burchiel et al., 2005,
088090). Exposure to hardwood smoke increased proliferation of T cells from A/J mice exposed
daily to 100 (ig/m3 PM for 6 mo, but produced a concentration-dependent suppression of
proliferation at PM concentrations >300 (ig/m3. No effects on B cell proliferation were observed.
Concentrations of NO and NO2 were not detectable or <40 ppb for all exposure levels. CO was
reported to be 2, 4, and 13 ppm for the 100, 300 and 1,000 (ig/m3 PM concentrations, respectively.
Exposure atmospheres contained significant levels of naphthalene and methylated napthalenes,
fluorene, phenanthrene, and anthracene, as well as low concentrations of several metals (K, Ca, and
Fe) (Burchiel et al., 2005, 088090). It should be noted that serologic analysis of study sentinel
animals indicated infection with parvovirus , which can interfere with the modulation of lymphocyte
mitogenic responses (Baker, 1998, 156245). In another study by Reed et al. (2006, 156043)
C57BL/6 mice were exposed to 30-1,000 (ig/m3 hardwood smoke by whole-body inhalation for 6
mo prior to instillation of P. aeruginosa. Exposure characterizations are described in Section 7.3.3.2.
Although there was a trend toward increased clearance with increasing exposure concentrations,
there was no statistically significant effect of hardwood smoke exposure on bacterial clearance.
7.3.8. Respiratory Mortality
Two large U.S. cohort studies examined the effect of long-term exposure to PM2.5 on
respiratory mortality with mixed results. In the ACS study, Pope et al. (2004, 055880) reported
positive associations with deaths from specific cardiovascular diseases, but no PM2.5 associations
were found with respiratory mortality. A follow-up to the Harvard Six Cities study (Laden et al.,
2006, 087605) used updated air pollution and mortality data and found positive associations between
long-term exposure to PM2.5 and mortality. Of special note is a statistically significant reduction in
mortality risk reported with reduced long-term fine particle concentrations observed for deaths due
to cardiovascular and respiratory causes, but not for lung cancer deaths. There is some evidence for
an association between PM25 and respiratory mortality among post-neonatal infants (ages 1 month-1
year) (Section 7.4.1). In summary, when deaths due to respiratory causes are separated from all-
cause (nonaccidental) and cardiopulmonary deaths, there is limited and inconsistent evidence for an
effect of PM25 on respiratory mortality, with one large cohort study finding a reduction in deaths due
to respiratory causes associated with reduced PM2 5 concentrations, and another large cohort study
finding no PM2 5 associations with respiratory mortality.
December 2009 7-41
-------
7.3.9. Summary and Causal Determinations
7.3.9.1. PM2.5
The epidemiologic studies reviewed in the 2004 PM AQCD suggested relationships between
long-term PMi0 and PM2.5 (or PM2.i) exposures and increased incidence of respiratory symptoms and
disease. One of these studies indicated associations with bronchitis in the 24-city cohort (Dockery et
al., 1996, 046219). They also suggested relationships between long-term exposure to PM2.5 and
pulmonary function decrements in the CHS (Gauderman et al., 2000, 012531; Gauderman et al.,
2002, 026013). These findings added to the database of the earlier 22-city study of PM2.i (Raizenne
et al., 1996, 077268) that found an association between exposure to ambient particle strong acidity
and impairment of lung function in children. No long-term exposure toxicological studies were
reported in the 2004 PM AQCD.
Recent studies have greatly expanded the evidence available since the 2004 PM AQCD. New
analyses have been conducted that include longer follow-up periods of the CHS cohort through 18 yr
of age and provide evidence that effects from exposure to PM2 5 persist into early adulthood.
Gauderman et al. (2004, 056569) reported that PM2 5 exposure was associated with clinically and
statistically significant deficits in FEVi attained at the age of 18 yr. In addition, the strength and
robustness of the outcomes were larger in magnitude, and more precise than previous CHS studies
with shorter follow-up periods. Supporting this result are new longitudinal cohort studies conducted
by other researchers in other locations with different methods. These studies report results for PMi0
that is dominated by PM2 5. New studies provide positive associations from Mexico City, Sweden,
and a national cystic fibrosis cohort in the U.S. A natural experiment in Switzerland, where PM
levels had decreased, reported that improvement in air quality may slow the annual rate of decline in
lung function in adulthood, indicating positive consequences for public health. Thus, the data are
consistent and coherent across several study designs, locations and researchers. As was found in the
2004 PM AQCD, the studies report associations with PM2 5 and PMi0, while most did not evaluate
PMio_2.5. Associations have been reported with fine particle components, particularly EC and OC.
Source apportionment methods generally have not been used in these long-term exposure studies.
Coherence and biological plausibility for the observed associations with lung function
decrements is provided by toxicological studies (Section7.3.2.2). A recent study demonstrated that
pre- and postnatal exposure to ambient levels of urban particles affected mouse lung development, as
measured by anatomical and functional indices (Mauad et al., 2008, 156743). Another study
suggested that the developing lung may be susceptible to PM since acute exposure to UF iron-soot
decreased cell proliferation in the proximal alveolar region of neonatal rats (Pinkerton et al., 2004,
087465) (Section 6.3.5.3). Impaired lung development is a viable mechanism by which PM may
reduce lung function growth in children. Other animal toxicological studies have demonstrated
alterations in pulmonary function following exposure to DE and wood smoke (Section 7.3.2.2).
An expanded body of epidemiologic evidence for the effect of PM2 5 on respiratory symptoms
and asthma incidence now includes prospective cohort studies conducted by different researchers in
different locations, both within and outside the U.S. with different methods. The CHS provides
evidence in a prospective longitudinal cohort study that relates PM2 5 and bronchitic symptoms and
reports larger associations for within-community effects that are less subject to confounding than
between-community effects (McConnell et al., 2003, 049490). Several new studies report similar
findings with long-term exposure to PMi0 in areas where fine particles predominate. In England, an
association was seen with an increased prevalence of cough without a cold. Further evidence
includes a reduction of respiratory symptoms corresponding to decreasing PM levels in natural
experiments in cohorts of Swiss school children (Bayer-Oglesby et al., 2005, 086245) and adults
(Schindler et al., 2009, 191950).
New studies examined the relationship between long-term PM2 5 exposure and asthma
incidence. PM2 5 had the strongest modifying effect on the association between lung function with
asthma in an analysis of the CHS (Islam et al., 2007, 090697). The loss of protection by high lung
function against new onset asthma in high PM2 5 communities was observed for all the lung function
measures. In the Netherlands, an association with doctor-diagnosed asthma was found in a birth
cohort examining the first 4 yr of life (Brauer et al., 2007, 090691) Further, findings from an adult
cohort suggest that traffic-related PMi0 contributes to asthma development and that reductions in PM
decrease asthma risk (Kunzli et al., 2009, 191949).
December 2009 7-42
-------
A large proportion of asthma is driven by allergy, and the majority of recent epidemiologic
studies examining allergic (or atopic) indicators found positive associations with PM2 5 or PMi0
(Section 7.3.6.1). Limited evidence for PM-mediated allergic responses is provided by toxicological
studies of DE and woodsmoke, while effects of gasoline exhaust were attributed to gaseous
components (Section 7.3.6.2).
Long-term PM2.5 exposure is associated with pulmonary inflammation and oxidative
responses. An epidemiologic study found a relationship between PM2.5 and increased inflammatory
marker eNO among school children (Dales et al., 2008, 156378). Toxicological studies of pulmonary
inflammation have demonstrated mixed results, with subchronic DE exposures generating increases
and CAPs and wood smoke inducing little or no response (Section 7.3.3.2). The pulmonary
inflammation observed with DE was attributable to the particle fraction. Toxicological studies also
reported evidence of oxidative responses (Section 7.3.4.1). Adaptation to prolonged DE was
observed for some oxidative responses in addition to some allergic and pulmonary function
responses (Section 7.3.2.2 and 7.3.6.2).
Additional support for the relationship between long-term PM2.5 exposures and respiratory
outcomes is provided by pulmonary injury responses observed in toxicological studies (Section
7.3.5.1). Markers of pulmonary injury were increased in rats exposed to DE and gasoline exhaust;
and these changes were attributable to PM. Further, lung DNA methylation was observed in the
gasoline exhaust study. Histopathological changes have also been reported following exposure to
heavily-trafficked urban air and woodsmoke. Findings include nasal and airway mucous cell
hyperplasia accompanied by alterations in mucus production which can lead to a loss of mucus-
mediated protective functions; exacerbation of protease-induced emphysema; and mast cell
infiltration and hypertrophy of alveolar walls. These results provide biological plausibility for
adverse respiratory outcomes following long-term PM exposure.
Limited information is available on host defense responses (Section 7.3.7) and respiratory
mortality (Section 7.3.8) resulting from PM2.5 exposure. Several recent epidemiologic studies
suggest a relationship between long-term exposure to PM25 or PMi0 and infection in children and
infants (Section 7.3.7.1). A few toxicological studies suggest that DE exposure affects systemic
immunity, and although impaired bacterial clearance is associated with short-term exposures to DE,
neither DE or gasoline exhaust seems to have this effect after longer exposures (Section 7.3.7.2).
In summary, the strongest evidence for a relationship between long-term exposure to PM2 5
and respiratory morbidity is provided by epidemiologic studies demonstrating associations with
decrements in lung function growth in children and with respiratory symptoms and disease incidence
in adults. Mean PM25 concentrations in these study locations ranged from 13.8 to 30 ug/m3 during
the study periods. These studies provide evidence for associations in areas where PM is
predominantly fine particles. A major challenge to interpreting the results of these studies is that the
PM size fractions and concentrations of other air pollutants are often correlated; however, the
consistency of findings across different locations supports an independent effect of PM25. Recent
toxicological studies provide support for the associations with PM2 5 and decreases in lung function
growth in children. Pre- and postnatal exposure to ambient levels of urban particles was found to
affect mouse lung development, which provides biological plausibility for the epidemiologic
findings. Recent subchronic and chronic toxicological studies also demonstrate altered pulmonary
function, mild inflammation, oxidative responses, histopathological changes including mucus cell
hyperplasia and enhanced allergic responses in response to CAPs, DE, urban air and woodsmoke and
provide further coherence and biological plausibility. Exacerbation of emphysematous lesions was
noted in one study involving exposure to urban air in a heavily-trafficked area. Collectively, the
evidence is sufficient to conclude that the relationship between long-term PM exposure
and respiratory effects is likely to be causal.
7.3.9.2. PM10.2.5
The 2004 PM AQCD did not report long-term exposure studies for PMi0_2.s. The only recent
study to evaluate long-term exposure to PMi0_2.5 found positive, but not statistically significant
associations with eNO (Dales et al., 2008, 156378). The evidence is inadequate to determine if 3
causal relationship exists between long-term PMi025 exposures and respiratory effects
December 2009 7-43
-------
7.3.9.3. UFPs
The 2004 PM AQCD did not report long-term exposure studies for UFPs. The current
evidence for long-term UFP effects is limited to toxicological studies. Generally, subchronic
exposure to DE induced pulmonary inflammation, which was in contrast to UF CAPs exposure
(Section 7.3.3.2) It appeared that the PM fraction was responsible for the inflammatory response
with DE exposure. Long-term exposure to DE also resulted in oxidative and allergic responses,
although lung injury was not remarkable (Sections 7.3.4.1 and 7.3.6.2). The evidence is inadequate
to determine if a causal relationship exists between long-term UFP exposures and
respiratory effects
7.4. Reproductive, Developmental, Prenatal and Neonatal
Outcomes
7.4.1. Epidemiologic Studies
This section evaluates and summarizes the scientific evidence on PM and developmental and
pregnancy outcomes and infant mortality. Infants and fetal development processes may be
particularly vulnerable to PM exposure, and although the physical mechanisms are not fully
understood, several hypotheses have been proposed involving direct effects on fetal health, altered
placenta function, or indirect effects on the mother's health (Bracken et al., 2003, 156288; Clifton et
al., 2001, 156360: Maisonet et al., 2004, 156725: Schatz et al., 1990, 156073: Sram et al., 2005,
087442). Study of these outcomes can be difficult given the need for detailed data and potential
residential movement of mothers during pregnancy. Two recent articles have reviewed
methodological issues relating to the study of outdoor air pollution and adverse birth outcomes (Ritz
and Wilhelm, 2008, 156914: Slama et al., 2008, 156985). Some of the key challenges to
interpretation of these study results include the difficulty in assessing exposure as most studies use
existing monitoring networks to estimate individual exposure to ambient PM; the inability to control
for potential confounders such as other risk factors that affect birth outcomes (e.g., smoking);
evaluating the exposure window (e.g., trimester) of importance; and limited evidence on the
physiological mechanism of these effects (Ritz and Wilhelm, 2008, 156914; Slama et al., 2008,
156985). Another uncertainty is whether PM effects differ by the child's sex. A review of preterm
birth and low birth weight studies found limited indication that effects may differ by gender,
however sample size was limited (Ghosh et al., 2007, 091233).
Previous summaries of the association between PM concentrations and pregnancy outcomes
and infant mortality were presented in previous PM AQCDs. The 1996 PM AQCD concluded that
although few studies had been conducted on the link between PM and infant mortality, the research
"suggested an association," particularly for post-neonates (U.S. EPA, 1996, 079380). In the 2004 PM
AQCD, additional evidence was available on PM's effect on fetal and early postnatal development
and mortality (U.S. EPA, 2004, 056905) and although some studies indicated a relationship between
PM and pregnancy outcomes, others did not. Studies identifying associations found that exposure to
PMio early during pregnancy (first month of pregnancy) or late in the pregnancy (6 wk prior to birth)
were linked with higher risk of preterm birth, including models adjusted for other pollutants, and that
PM2.5 during the first month of pregnancy was associated with intrauterine growth restriction.
However, other work did not identify relationships between PMio exposure and low birth weight.
The state of the science at that time, as indicated in the 2004 PM AQCD, was that the research
provided mixed results based on studies from multiple countries, and that additional research was
required to better understand the impact of PM on pregnancy outcomes and infant mortality.
Considering evidence from recent studies discussed below, along with previous AQCD conclusions,
epidemiologic studies consistently report associations between PMio and PM2.5 exposure and low
birth weight and infant mortality, especially during the post-neonatal period. Animal toxicological
evidence supports these associations with PM2 5, but provides little mechanistic information or
biological plausibility. Information on the ambient concentrations of PMio and PM2.5 in these study
locations can be found in Table 7-5.
December 2009 7-44
-------
7.4.1.1. Low Birth Weight
A large number of studies have investigated exposure to ambient PM and low birth weight at
term, including a U.S. national study, as well as two studies in the northeast U.S., and four in
California. Parker and Woodruff (2008, 156846) linked U.S. birth records for singletons delivered at
40-wk gestation in 2001-2003 during the months of March, June, September and December to
quarterly estimates of PM exposure by county of residence and month of birth. They found an
association between PMi0_2.5 and birthweight (-13 g [95% CI: -18.3 to -7.6]) per 10 ug/m3 increase),
but no such association for PM2 5.
Maisonet et al. (2001, 016624) analyzed 89,557 births (1994-96) in six northeastern cities
(Boston and Springfield MA; Hartford CT; Philadelphia and Pittsburgh PA; and Washington DC).
Each city had three PMi0 monitors measuring every sixth day. Results from multiple monitors were
averaged in each city. Exposure was determined for each trimester of pregnancy and categorized by
quartiles (<25, 25-30, 31-35, 36-43 ug/m3) and 95th percentile (>43ug/m ). There was no increased
risk for low birth weight at term associated with PMi0 exposure during any trimester of pregnancy.
When birth weight was considered as a continuous outcome, exposure to PMi0 was not associated
with a reduction in mean birth weight.
In contrast, Bell et al. (2007, 093256) reported positive associations for both PM2.5 and PMi0
with birth weight in a study of births (n = 358,504) in Connecticut and Massachusetts (1999-2002).
Birth data indicated county, not street address or ZIP code, so women were assigned exposure based
on county residence at delivery. The difference in birth weight per 10 ug/m3 associated with PM25
was -66.8 (95% CI: -77.7 to -55.9) g. For PM10 it was -11.1 (95% CI: -15.0 to -7.2) g. The increased
risk for low birth weight was OR = 1.054 (95% CI: 1.022-1.087) for PM2.5 and OR = 1.027 (95% CI:
0.991-1.064) for PMi0, based on average exposure during pregnancy. Reductions in birth weight
were also associated with third trimester exposure to PMi0 and second and third trimester exposure
to PM25. Comparing this study to Maisonet et al. (2001, 016624). a larger sample size was able to
detect a small increase in risk. In addition, birth weight was reduced more by exposure to PM2 5 than
by exposure to PMi0. Measured PM2 5 concentrations were not available in the earlier study.
The Children's Health Study is a population based cohort of children living in 12 southern
California communities, selected on the basis of differing levels of air pollution (Salam et al., 2005,
087885). as previously discussed in Section 7.3. The children in grades 4, 7 and 10 were recruited
through schools. A subset of this cohort (n = 6,259) were born in California from 1975-1987. Of
these, birth certificates were located for 4,842, including 3,901 infants born at term and 72 cases of
low birth weight at term. Using the mother's ZIP code at the time of birth, exposure was determined
by inverse distance weighting of up to three PMi0 monitors within 50 km of the ZIP code centroid. If
there was a PMi0 monitor within 5 km of the ZIP code centroid (40% of data), exposure from that
monitor was used. Exposure was calculated for the entire pregnancy, and for each trimester of
pregnancy. A 10 ug/m increase in PMi0 during the third trimester reduced mean birth weight -10.9 g
(95% CI: -21.1 to -0.6) in single pollutant models, but became non-significant in copollutant models
controlling for the effects of O3. Increased risks of low birth weight (<2,500 g) were not statistically
significant (OR =1.3 [95% CI: 0.9-1.9]). A strength of this study was the cohort data available
included information on SES and smoking during pregnancy. A limitation is the assignment of
exposure based on monitoring stations up to 50 km distant; this may have introduced substantial
exposure misclassification obscuring some associations.
December 2009 7-45
-------
Table 7-5. Characterization of ambient PM concentrations from studies of reproductive,
developmental, prenatal and neonatal outcomes and long-term exposure.
Study
Location
Mean Annual
Concentration (ug/m3)
Upper Percentile
Concentrations (ug/m3)
PM2.5
Basu et al. (2004, 0878961
Bell et al. (2007, 0910591
Braueretal. (2008, 156292)
Huynh et al. (2006, 0912401
Jalaludin et al. (2007, 1566011
Liu (2007, 0904291
Loomisetal. (1999, 087288)
Mannes et al. (2005, 087895)
Parker et al. (2005, 0874621
Ritz et al. (2007, 0961461
Wilhelm and Ritz (2005, 0886681
Woodruff etal. (2006, 088758)
Woodruff etal. (2008, 0983861
CA
CIS MA
Vancouver, Canada
CA
Sydney, Australia
Multicity, Canada
Mexico City
Sydney, Australia
CA
Los Angeles, CA
Los Angeles, CA
CA
U.S.
Range of means across sites: 14.5-18.2
Avg of means across sites: 16.2
22.3
5.3
Range of means across trimesters: 17.5-18.8
Avg of means across trimesters: 18.2
9.0
12.2
27.4
9.4
15.4
20.0
21.0
19.2a
Range of means across effects: 1 4.5-1 4.9a
Avg of means across effects: 14.8a
Max: 26.3-34.1
Max: 37.0
75th: 15
Max: 85
75th: 11.2; Max:
Max: 38.9-48.5
75th: 22.7
75th: 18.5-18.7
82.1
PMiO-2.5
Parker et al. (2008, 156013)
U.S.
13.2
75th: 17.5
PM10
Bell et al. (2007, 093256)
Braueretal. (2008, 156292)
Chen et al. (2002, 024945)'
Gilboa et al. (2005, 0878921
Ha et al. (2003, 0425521
Hansen et al. (2006, 089818)
Hansen et al. (2007, 090703)
Jalaludin et al. (2007, 156601)
Kim et al. (2007, 1566421
Lee et al. (2003, 043202)
Leem et al. (2006, 089828)
Lipfert et al. (2000, 004103)
Maisonet etal. (2001,016624)
Mannes et al. (2005, 0878951
Pereiraetal. (1998, 007264)
Ritz et al. (2000, 012068)
Ritz et al. (2006, 0898191
CIS MA
Vancouver, Canada
Wfeshoe County, NV
TX
Seoul, South Korea
Brisbane, Australia
Brisbane, Australia
Sydney, Australia
Seoul, Korea
Seoul, Korea
Incheon, Korea
U.S.
NEU.S.
Sydney, Australia
Sao Paulo, Brazil
CA
CA
22.3
12.7
31.53
23.8a
69.2
19.6
19.6
16.3
Range of means across time: 88.7-89.7
Avg of means across time: 89.2
71.1
53.8a
33.1
31. Oa
16.8
65.04
49.3
46.3
Max: 35.4
75th: 39.35; Max: 157.32
75th: 29
75th: 87.7; Max:
Max: 171.7
75th: 22.7; Max:
75th: 89.3; Max:
75th: 64.6; Max:
Max: 59
75th: 36.1; Max:
75th: 19.9; Max:
Max: 192.8
Max: 178.8
Max: 83.5
245.4
171.7
236.9
106.39
46.5
104.0
December 2009
7-46
-------
Study
Rogers and Dunlop (2006, 0912321
Romieu et al. (2004, 093074)
Sagiv et al. (2005, 0874681
Salam et al. (2005, 0878851
Suh et al. (2008, 1920771
Tsai et al. (2006, 090709)
Wilhelm and Ritz (2005, 0886681
Woodruff etal. (2008, 0983861
Yang et al. (2006, 090760)
Location
GA
Ciudad Juarez, Mexico
PA
CA
Seoul, Korea
Kaohsiung, Taiwan
Los Angeles, CA
U.S.
Taipei, Taiwan
Mean Annual
Concentration (ug/m3)
3.75
33.0-45.9
Range of means across time: 25.3-27. 1
Avg of means across time: 26.2
Range of means across trimesters: 45.4-46.6
Avg of means across trimesters: 45.8
Range of means across trimesters: 54.6-61 . 1
Avg of means across trimesters: 58.27
81.5
38.1
Range of means across effects: 28.6-29.8a
Avg of means across effects: 29. f
53.2
Upper Percentile
Concentrations (ug/m3)
75th: 15.07
Max: 68.9-156.3
75th: 62.8-67.8
Max: 85.1-107.36
75th: 111. 5; Max: 232.0
Max: 74.6-103.7
75th: 33.8-36.5
75th: 64.9; Max: 234.9
3Median concentration
Parker et al. (2005, 087462) examined births in California within 5 miles of a monitoring
station (n = 18,247). Only infants born at 40 wk gestation were included. Thus all infants were the
same gestational age, and had been exposed in the same year. Exposure to PM2.5 in quartiles (<11.9,
11.9-13.9, 14.0-18.4, >18.4) was associated with decrements in birth weight. Infants exposed to
>13.9 ug/m3 experienced reductions in birth weight (third quartile -13.7 g (95% CI: -34.2 to 6.9),
fourth quartile -36.1 g (95% CI: -55.8 to -16.5). These are larger reductions than have been seen in
some other studies. However, this study reduced misclassification by including only women living
within 5 miles of a monitoring station, and only included births at 40 wk gestation. Reducing
misclassification should lead to a stronger association, if the association is causal.
The effects of spatial variation in exposure were also investigated by Wilhelm and Ritz (2005,
088668). Their study included all women living in ZIP codes where 60% of the ZIP code was within
two miles of a monitoring station in the Southern California Basin, and women with known
addresses in Los Angeles County within 4 miles of a monitoring station. Exposure to average PMi0
in the third trimester was analyzed for increased risk of low birth weight at term (> 37-wk gestation).
Analysis at the ZIP code level did not detect increased risk (per 10 ug/m3 PM10s OR = 1.03 [95% CI:
0.97-1.09]). However the analysis based on geocoded addresses indicated that increasing exposure to
PMio was associated with increased risk of low birth weight for women living within 1 mile of the
station where PMi0 was measured. For these women (n = 247 cases, 10,981 non-cases), each
10 ug/m3 increase in PMio was associated with a 22% increase in risk of term low birth weight
(OR = 1.22 [95% CI: 1.05-1.41]). In the categorical analysis, exposure to PM10 >44.4 ug/nrwas
associated with a 48% increase in risk (OR = 1.48 [95% CI: 1.00-2.19]). Increased risk of low birth
weight also was associated with exposure to CO in single pollutant models. However, when
multipollutant models were considered, the effects of CO were attenuated but the effects of PMio
increased. Controlling for CO, NO2, and O3, each 10 ug/m3 increase in exposure to PMio increased
risk of low birth weight 36% (OR = 1.36 [95% CI: 1.12-1.65]).
Spatial variation in PM2.5 exposure was investigated by Basu et al. (2004, 087896). They
included only mothers who lived within 5 miles of a PM2 5 monitor and within a California county
with at least 1 monitor. To minimize potential confounding, they included only white (n = 8,597) or
Hispanic (n = 8,114) women, who were married, between 20 and 30 yr of age, completed at least
high school and were having their first child. Consistently, PM2 5 exposure measured by the county
monitor was more strongly associated with reductions in birth weight than exposure measured by the
neighborhood monitor. The results were replicated in both the white and the Hispanic samples.
Reductions in birth weight ranged from 15.2 to 43.5 g per 10 ug/m3 increase in PM25.
In the remaining U.S. study, Chen et al. (2002, 024945) analyzed 33,859 birth certificates of
residents of Washoe County in northern Nevada (1991-1999). There were four sites monitoring
during the study period, it appears (not stated) that exposure was averaged over the county. A
December 2009
7-47
-------
10 ug/m3 increase in exposure to PM10 during the third trimester of pregnancy was associated with
an 11 g reduction in birth weight (95% CI: -2.3 to -19.8). Effects on risk of low birth weight were not
statistically significant. For exposure in the third trimester of 19.77 to 44.74 ug/m3 compared to
<19.74 ug/m3 the odds ratio for low birth weight was 1.05 (95% CI: 0.81-1.36). Comparing exposure
>44.74 to the same reference category, the odds ratio was 1.10 (95% CI: 0.71-1.71).
Misclassification of exposure may have occurred when exposure was averaged over a large
geographic area (16,968 km2).
Recent international studies investigating effects of particles on low birth weight include one
in Munich (Slama et al, 2007, 093216). two in Canada (Brauer et al, 2008, 156292; Dugandzic et
al., 2006, 088681). two in Australia (Hansen et al., 2007, 090703: Mannes et al., 2005, 087895). two
in Taiwan (Lin et al., 2004, 089827: Yang et al., 2003, 087886) one in Korea (Ha et al., 2003,
042552) and two in Sao Paulo, Brazil (Gouveia et al., 2004, 055613: Medeiros and Gouveia, 2005,
089824). The majority of these studies found that PM concentrations were associated with low birth
weight, though two studies (Hansen et al., 2007, 090703: Lin et al., 2004, 089827) found no
associations. The effect estimates were similar in magnitude to those reported in the U.S. studies.
Considerations in Interpreting Results of Low Birth Weight Studies
Studies included subjects at distances from monitoring stations varying from as close as 1 mile
or 2 km, to as far as 50 km or the size of the county. Studies that only included subjects living within
a short distance (1 mile, 2 km) of the monitoring station (thus likely reducing exposure measurement
error) were more likely to find that PM exposure was associated with increased risk of low birth
weight. However, Basu et al. (2004, 087896) reported a stronger association between PM2.5 exposure
and birth weight when exposure was estimated based on the county monitor, rather than the monitor
within 5 miles of the residence. They suggest that county level exposure may be more representative
of where women spend their time, including not only home, but also other time spent away from
home. Other pollutants also appeared to influence the risk associated with particle exposure. In one
study, exposure to PMi0 in a single pollutant model reduced birth weight by 11 g, but became non-
significant in copollutant models with O3 (Salam et al., 2005, 087885). In another study the risk
associated with PM10 exposure increased from 22% to 36% when other pollutants were included in
the model (Wilhelm and Ritz, 2005, 088668). All but one study in the U.S. found some association
between particle exposure and reduced birth weight (Maisonet et al., 2001, 016624). The results of
international studies were inconsistent. This might be related to the chemical composition of
particles in the U.S., or to differences in the pollutant mixture. Studies with null results must be
interpreted with caution when the comparison groups have significant exposure. This was certainly
the situation in studies in Taiwan and Korea (Lee et al., 2003, 043202: Lin et al., 2004, 089827: Yang
et al., 2003, 087886). Differences in geographical locations, study samples and linkage decisions
may contribute to the diverse findings in the literature on the association between PM and
birthweight, even within the U.S. (Parker and Woodruff, 2008, 156846).
7.4.1.2. Preterm Birth
A potential association of exposure to airborne particles and preterm birth has been
investigated in numerous epidemiologic studies, including some conducted in the U.S. and others in
foreign countries. Three U.S. studies have been carried out by the same group of investigators in
California.
A natural experiment occurred when an open-hearth steel mill in Utah Valley was closed from
August 1986 through September 1987. Parker et al. (2008, 156013) compared birth outcomes for
Utah mothers within and outside of the Utah Valley, before, during, and after the mill closure. They
report that mothers who were pregnant around the time of the closure of the mill were less likely to
deliver prematurely than mothers who were pregnant before or after. The strongest effect estimates
were observed for exposure during the second trimester (14% decrease in risk of preterm birth
during mill closure). Preterm birth outside of the Utah Valley did not change during the time of the
mill closure.
In 2000, Ritz et al. (2000, 012068) published the first study investigating the association of
preterm birth with PM in the U.S. The study population was women living in the southern California
Basin. There were eight monitoring stations measuring PMi0 every 6th day during the study period.
December 2009 7-48
-------
Birth certificates (1989-1993) were analyzed for women living in ZIP codes within 2 miles of a
monitoring station. Women with multiple gestations, chronic disease prior to pregnancy and women
who delivered by cesarean section were excluded resulting in a study population of 48,904 women.
The risk of preterm birth increased by 4% (RR = 1.04 [95% CI: 1.02-1.6]) per 10 ug/m3 increase in
PMio averaged in the 6 wk before birth. Exposure to PMi0 in the first month of pregnancy resulted in
a 3% increase in risk (RR = 1.03 [95% CI: 1.01-1.05]). These results were robust in multipollutant
models.
Wilhelm and Ritz (2005, 088668) reinvestigated this association among women in the same
area in 2005, when air pollution had declined from a mean level near 50 ug/m3 to a mean level near
40 ug/m3. Birth certificate data from 1994-2000 was analyzed for women living in ZIP codes within
2 miles of a monitoring station, or with addresses within 5 miles of a monitoring station. No
significant effects of exposure to PMi0 were reported. Exposure to PM2.5 6 wk before birth resulted
in an increase in preterm birth (RR = 1.19 [95% CI: 1.02-1.40]) for the highest quartile of exposure
(PM25 >24.3 ug/m3). Using a continuous measure of PM25, there was a 10% increase in risk for each
10 ug/m3 increase in PM2.5 (RR = 1.10 [95% CI: 1.00-1.21]).
There have been two major criticisms of air pollution studies using birth certificate data. First,
that birth certificates only indicate the address at birth and the exposure of women who moved
during pregnancy may be misclassified; second, that information about some important confounders
may not be available (e.g., smoking). To obtain more precise information about these variables, Ritz
et al. (2007, 096146) conducted a case-control study nested within a cohort of birth certificates (Jan
2003-Dec 2003) in Los Angeles County. Births to women residing in ZIP codes (n = 24) close to
monitoring stations or major population centers or roadways (n = 87) were eligible (n = 58,316
births). All cases of low birth weight or preterm birth and an equal number of randomly sampled
controls in the 24 ZIP codes close to monitors were selected. In the other 87 ZIP codes, 30% of cases
and an equal number of controls were randomly sampled. Of 6,374 women selected for the case
control study, 2,543 (40%) were interviewed. The association of preterm birth with exposure to
PM2 5 differed between women responding to the survey and women who did not respond. Among
responders, exposure to each 10 ug/m3 increase in PM25 concentration in the first trimester increased
risk to preterm birth by 23% (RR = 1.23 [95% CI: 1.02-1.48]). There was no increase in risk among
non-responders (RR = 0.95 [95% CI: 0.82-1.10]), or in the entire birth cohort (RR = 1.00 [95% CI:
0.94-1.07]).
An additional case control study of preterm birth and PM2s exposure (Huynh et al., 2006,
091240) used California birth certificate data. Singleton preterm infants (24-36-wk gestation) born in
California (1999-2000) whose mothers lived within 5 miles of a PM25 monitor were eligible. Each of
these 10,673 preterm infants were matched to three term (39- to 44-wk gestation) controls (having a
last menstrual period within 2 wk of the case infant), resulting in a study population of 42,692.
Controlling for maternal race/ethnicity, education, marital status, parity and CO exposure, exposure
to PM2.5 >17.7 ug/m3 increased the risk of preterm birth by 14% (OR =1.14 [95% CI: 1.07-1.23]).
Averaging PM25 exposure over the first month of pregnancy, the last 2 wk before birth, or the entire
pregnancy did not substantially change the risk estimate.
Two additional studies of preterm birth and exposure to particulate air pollution have been
conducted in the U.S. Each has used a unique methodology. Sagiv et al. (2005, 087468) used time
series to analyze births in four Pennsylvania counties between January 1997 and December 2001. In
this analysis, exposure to PMi0 is compared to the rate of preterm births each day. Both acute
exposure (on the day of birth) and longer term exposure (average exposure for the preceding 6 wk)
were considered in the analysis. An advantage of this analysis is that days, rather than individuals are
compared, so confounding by individual risk factors is minimized. For exposure averaged over the
6 wk prior to birth, there was an increase in risk (RR = 1.07 [95% CI: 0.98-1.18]), which persisted
for acute exposure with a 2-day lag (RR =1.10 [95% CI: 1.00-1.21]) and 5-day lag (RR = 1.07 95%
CI: 0.98-1.18]).
Rogers and Dunlop (2006, 091232) examined exposure to particles and risk of delivery of an
infant weighing less than 1,500 g (all of which were preterm) from 24 counties in Georgia. The study
included 69 preterm, small for gestational age (SGA) infants, 59 preterm appropriate for gestational
age (AGA) infants and 197 term AGA controls. Exposure was estimated using an environmental
transport model that considered PMi0 emissions from 32 geographically located industrial point
sources, meteorological factors, and geographic location of the birth home. Exposure was
categorized by quartiles. Comparing women who delivered a preterm AGA infant to those who
December 2009 7-49
-------
delivered a term AGA infant, exposure to PM10>15.07 ug/m3 tripled the risk (OR = 3.68 [95% CI:
1.44-9.44]).
Brauer et al. (2008, 156292) evaluated the impacts of PM25 on preterm birth using
spatiotemporal exposure metrics in Vancouver, Canada. The authors found similar results when they
used a land-use regression model or inverse distance weighting as the exposure metric. For preterm
births <37 wk, they reported an OR of 1.06 (95% CI: 1.01-1.11), and for preterm births <35 wk the
OR increased to 1.12 (95% CI: 1.02-1.24). There were no consistent trends for early or late
gestational period to be more strongly associated with preterm births.
Suh et al. (2008, 192077) conducted a study to determine if the effects of exposure to PMi0
during pregnancy on preterm delivery are modified by maternal polymorphisms in metabolic genes.
They analyzed the effects of the gene-environment interaction between the GSTM1, GSTT1,
CYPlal-T6235C and -1462V polymorphisms and exposure to PMi0 during pregnancy on preterm
birth in a case-control study in Seoul, Korea. PM10 concentration > 75th percentile alone was
significant in the third trimester of pregnancy (OR = 2.33 [95% CI: 1.33-4.80]), but not in the first or
second trimester. The risk of preterm delivery conferred by the GSTM1 null genotype was increased,
and the highest risk was found during the third trimester of pregnancy (OR = 2.58 [95% CI:
1.34-4.97]). There were no statistical associations with the GSTT1 or CYP1A1 genotypes. When the
gene-environment interaction was analyzed, the risk for preterm birth was substantially higher for
women who carried the GSTM1 null genotype and were exposed to high levels of PMi0 (> 75th
percentile) than for those who carried the GSTM1 positive genotype but were only exposed to low
levels of PMio (<75th percentile) during the third trimester of pregnancy (OR = 6.22, 95% CI:
2.14-18.08).
In Incheon, Korea, Leem et al. (2006, 089828) estimated PMi0 exposure spatially as well as
temporally. Exposure was based on 26 monitors and kriging was used to determine exposure for 120
dongs (administrative districts, mean area 7.82 km2, median area 1.42 km3). The sample included
52,113 births, from 2001-2002. PMio was very weakly correlated with other pollutants. Exposure
was compared in quartiles for the first and third trimester of pregnancy. In the first trimester, relative
risks for the second, third and fourth quartiles were RR = 1.14 (95% CI: 0.97-1.34), RR = 1.07 (95%
CI: 0.94-1.37), and RR = 1.24 (95% CI: 1.09-1.41), respectively. Exposure to PMio in quartile one
(reference group) was 26.9-45.9 ug/m3; fourth quartile exposure equaled 64.6-106.4 ug/m3. The
p-value for trend was 0.02. Exposure in the third trimester was not related to preterm birth, however
no information was provided to determine how exposure in the third trimester was adjusted for
women who delivered preterm.
Two studies investigating risks of preterm birth related to particle exposure have been reported
from Australia. In Brisbane, Hansen et al. (2006, 089818) studied 28,200 births (2000-2003) in an
area of low PMio concentrations. Exposure to an interquartile range increase in PMio exposure in the
first trimester resulted in a 15% increased risk of preterm birth (OR =1.15 [95% CI: 1.06-1.25]).
This result was strongly influenced by the effect of PMio exposure in the first month of pregnancy
(OR =1.19 [95% CI: 1.13-1.26]). PMi0 was correlated with O3 (r = 0.77) in this study and O3 also
increased risk in the first trimester. No effects were associated with exposure to PMio in the third
trimester.
In Sydney, associations between exposure to particles and preterm birth varied by season.
Jalaludin et al. (2007, 156601) obtained information on all births in metropolitan Sydney
(1998-2000). Exposure to PM2.5 in the 3 mo preceding birth was associated with an increased risk of
preterm birth (OR = 1.11 [95% CI: 1.04-1.19]). Additional effects were dependent on season of
conception. Both PM10 (OR = 1.3 [95% CI: 1.2-1.5]) and PM2.5 (OR = 1.4 [95% CI: 1.3-1.6]) were
associated with increased risk for conceptions in the winter. Conceptions in summer were associated
with reductions in risk (PM1(? OR = 0.91 [95% CI: 0.88-0.93]) (PM2.5 OR = 0.87 [95% CI:
0.84-0.92]). Due to both positive and negative findings, the authors recommend caution in
interpreting their results.
Considerations in Analyzing Environmental Exposures and Preterm Birth
A major issue in studying environmental exposures and preterm birth is selecting the relevant
exposure period, since the biological mechanisms leading to preterm birth and the critical periods of
vulnerability are poorly understood (Bobak, 2000, 011448). Exposures proximate to the birth may be
most relevant if exposure causes an acute effect. However, exposure occurring in early gestation
December 2009 7-50
-------
might affect placentation, with results observable later in pregnancy, or cumulative exposure during
pregnancy may be the most important determinant. The studies reviewed have dealt with this issue in
different ways. Many have considered several exposure metrics based on different periods of
exposure.
Often the time periods used are the first month (or first trimester) of pregnancy and the last
month (or 6 wk) prior to delivery. Using a time interval prior to delivery introduces an additional
problem since cases and controls are not in the same stage of development when they are compared.
For example, a preterm infant delivered at 36 wk is a 32-week fetus 4 wk prior to birth, while an
infant born at term (40 wk) is a 36-week fetus 4 wk prior to birth. Only one study (Huynh et al.,
2006, 091240) adjusted for this in the design.
Many of these studies compare exposure in quartiles, using the lowest quartile as the reference
(or control) group. No studies use a truly unexposed control group. If exposure in the lowest quartile
confers risk, than it may be difficult to demonstrate additional risk associated with a higher quartile.
Thus negative studies must be interpreted with caution.
Preterm birth occurs both naturally (idiopathic preterm), and as a result of medical
intervention (iatrogenicpreterm). Ritz et al. (2000, 012068; 2007, 096146) excluded all births by
Cesarean section, to limit their studies to idiopathic preterm. No other studies attempted to
distinguish the type of preterm birth, although PM exposure maybe associated with only one type.
This is a source of potential effect misclassification.
7.4.1.3. Growth Restriction
Low birth weight has often been used as an outcome measure because it is easily available and
accurately recorded on birth certificates. However, low birth weight may result from either short
gestation, or inadequate growth in utero. Most of the studies investigating air pollution exposure and
low birth weight, limited their analysis to term infants to focus on inadequate growth. A number of
studies were identified that specifically addressed growth restriction in utero by identifying infants
who failed to meet specific growth standards. Usually these infants had birth weights less than the
10th percentile for gestational age, using an external standard. Many of these studies have been
previously discussed, since they also examined other reproductive outcomes (low birth weight or
preterm delivery).
Three studies in the U.S. examined intrauterine growth. A recent study (Rich et al., 2009,
180122) investigated very small for gestational age (defined as a fetal growth ratio <0.75), small for
gestational age (defined as > 75 and <85) and "reference" births (> 85) to women residing in New
Jersey and mean air pollutant concentrations during the first, second and third trimesters. They
reported an increased risk of SGA associated with first and third trimester PM2 5 concentrations
(1.116 [95% CI: 1.012, 1.232], and 1.106 [1.008-1.212], per 10 ug/m3 PM2.5, respectively). Parker et
al. (2005, 087462) reported a positive association between exposure to PM2.5. Since this study only
included singleton live births at 40-wk gestation, birth weights less than 2,872 g for girls and 2,986 g
for boys were designated SGA, based on births in California. Infants exposed to the highest quartile
PM2.5 (>18.4 ug/m ) compared to the lowest quartile PM25 (<11.9 ug/m ) were 23% more likely to
be small for gestational age (OR = 1.23 [95% CI: 1.03-1.50]). Very similar results were found for
exposure in each of the three trimesters respectively (OR =1.26 [95% CI: 1.04-1.51], OR =1.24
[95% CI: 1.04-1.49], OR = 1.21 [95% CI: 1.02-1.43]). These results controlled for exposure to CO,
which did not increase risk for SGA.
In contrast, Salam et al. (2005, 087885) found no association between exposure to PMi0 and
intrauterine growth retardation (IUGR) in the California Children's Health Study. IUGR was defined
as less than the 15th percentile of predicted birth weight based on gestational age and sex in term
infants. Apparently no external standard was used since 15% of infants in the study were designated
as IUGR. An IQR increase in PMi0 exposure was not significantly associated with IUGR for the
whole pregnancy (OR =1.1 [95% CI: 0.9-1.3]) or for any specific trimester. Differences between
this study and the study by Parker et al. (2005, 087462) include measurement of PMi0 versus PM25,
a less stringent definition of IUGR, and exposures determined by monitors located much farther
away from the subjects' residences (up to 50 km versus within 5 mi). All of these factors could lead
to misclassification.
Two studies investigating particle exposure and SGA were conducted in Australia, with
differing results (Hansen et al., 2007, 090703: Mannes et al., 2005, 087895). Mannes et al. (2005,
087895) defined SGA as birth weight less than two standard deviations below the national mean
December 2009 7-51
-------
birth weight for gestational age. In this study there was a statistically significant effect of exposure to
both PM10 (OR = 1.10 [95% CI: 1.00-1.48], per 10 ug/m3 increase) and PM2.5 (OR = 1.34 [95% CI:
1.10-1.63], per 10 ug/m3 increase) for exposure during the second trimester. When analysis was
restricted to births within 5 km of the monitoring station, the association for PMi0 became slightly
stronger (OR = 1.22 [95% CI: 1.10-1.34]). Exposure during other trimesters of pregnancy was not
associated with IUGR.
In Brisbane, Hansen et al. (2007, 090703) examined head circumference (HC), crown heel
length (CHL) and risk of SGA, defined as less than the tenth percentile of weight for gestational age
and gender based on an Australian national standard. There was no consistent relationship between
PMio exposure and SGA, HC or CHL in any trimester of pregnancy. PMi0 exposure was determined
by averaging values from the five monitoring stations. Due to the sample size and limited number of
monitoring stations, it was not possible to analyze the data for women living within 5 km of a
monitoring station, as was done in Sydney.
In Canada, Liu et al. (2007, 090429) investigated the effect of PM2.5 exposure on fetal growth
in three cities, Calgary, Edmonton and Montreal. IUGR was defined as birth weight below the tenth
percentile, by sex and gestational week (37-42) for all singleton live births in Canada between 1986
and 2000. Models were adjusted for maternal age, parity, infant sex, season of birth, city of
residence, and year of birth. A 10 ug/m3 increase in PM2.5 was associated with an increased risk for
IUGR (OR = 1.07 [95% CI: 1.03-1.10]) in the first trimester, and similar risks were associated with
exposure in the second or third trimesters. The effect of PM2.5 was reduced in multipollutant models
including CO and NO2
Brauer et al. (2008, 156292) observed consistent increased risks of SGA for PM2.5 PMi0, NO2,
NO and CO in Vancouver, Canada (20% increase in risk in PM25 and PMip per 10 ug/m increase).
The effects were similar for exposure estimates based on nearest monitor, inverse distance
weighting, and land-use regression modeling. ORs for early or late pregnancy exposure windows
were remarkably similar to those for the full duration of pregnancy.
7.4.1.4. Birth Defects
Four recent studies examined PM and birth defects. The Seoul, Korea study discussed above
also considered congenital anomalies, defined as a defect in the child's body structure (Kim et al.,
2007, 156642). PMi0 levels were associated with higher risk of birth defects for the second trimester,
with a 16% (95% CI: 0-34) increase in risk per 10 ug/m3 in PMi0.
Two U.S. studies examined air pollution and risk of birth defects. Data were collected from the
California Birth Defects Monitoring Program for four counties in Southern California (Los Angeles,
Riverside, San Bernardino, and Orange) for the period 1987-1993, although each county included a
subset of this period (Ritz et al., 2002, 023227). Cases (i.e., infants with birth defects) were
identified as live birth infants and fetal deaths from 20-wk gestation to 1 yr post-birth, with isolated,
multiple, syndrome, or chromosomal cardiac or orofacial cleft defects. Cases were restricted to those
with registry data for gestational age and residence ZIP code, and those with residences <10 miles
from an air pollution monitor. Six types of categories were included: aortic defects; atrium and
atrium septum defects; endocrinal and mitral value defects; pulmonary artery and valve defects;
conotruncal defects; and ventricular septal defects not part of the conotruncal category. PMi0
measurements were available every 6 days. While results indicated increased risk of birth defects for
higher levels of CO or O3, the authors determined that results for PMio were inconclusive, finding no
consistent trend of effect after adjustment for CO and O3.
The other U.S. study examined birth defects through a case-control design in seven Texas
counties for the period 1997-2000 (Gilboa et al., 2005, 087892). Births were excluded for parents
<18 yr and several non-air pollution risk factors known to be associated with birth defects (e.g.,
maternal diabetes, holoprosencephaly in addition to oral clef). Comparison of the highest
(> 29.0 ug/m3) and lowest (<19.521 ug/m3) quartiles of PMio for exposure defined as the third to
eighth week of pregnancy generated an OR of 2.27 (95% CI: 1.43-3.60) for risk of isolated atrial
septal defects and 1.26 (95% CI: 1.03-1.55) for individual atrial septal defects. Including other
pollutants (CO, NO2, O3, SO2) in the model did not greatly alter results; numerical results for
copollutant analysis were not provided. Strong evidence was not observed for a relationship between
PMio and the other birth defect categories. Review articles have concluded that the scientific
literature is not sufficient to conclude a relationship between air pollution and birth defects (Sram et
al., 2005, 087442).
December 2009 7-52
-------
A recent study of oral clefts conducted in Taiwan found no association between this birth
defect and concentrations of PMi0 during the first or second gestational month (Hwang and
Jaakkola, 2008, 193794). This population-based case-control study included 653 cases and a random
sample of 6,530 controls born in Taiwan between 2001 and 2003.
7.4.1.5. Infant Mortality
Many studies have identified strong associations between exposure to particles and increased
risk of mortality in adults or the general population, including for short- and long-term exposure
(Sections 6.5 and 7.6). Less evidence is available for the potential impact on infant mortality,
although studies have been conducted in several countries. The results of these infant mortality
studies are presented here with the other reproductive and developmental outcomes because it is
likely that in vitro exposures contribute to this outcome. Both long-term and short-term exposure
studies of infant mortality are included in this section. Results on PM and infant mortality includes a
range of findings, with some studies finding associations and many statistically non-significant or
null effects. Yet, more consistency is observed when results are divided into the type of health
outcome based on the age of infant and cause of death.
An important question regarding the association between PM and infant mortality is the
critical window of exposure during development for which infants are susceptible. Several age
intervals have been explored: neonatal (<1 mo); infants (<1 yr); and postneonatal (1 mo-1 yr).
Within these various age categories, multiple causes of deaths have been investigated, particularly
total deaths and respiratory-related deaths. The studies reflect a variety of study designs, particle size
ranges, exposure periods, regions, and adjustment for confounders.
Stillbirth
Only one study of stillbirths and PM was identified. A prospective cohort of pregnant women
in Seoul, Korea from 2001 to 2004 was examined with respect to exposure to PMi0 (Kim et al.,
2007, 156642). Gestational age was estimated by the last menstrual period or by ultrasound.
Whereas many of the previously discussed studies of PM and pregnancy outcomes were based on
national registries, this study examined medical records and gathered individual information through
interviews on socioeconomic condition, medical history, pregnancy complications, smoking, second-
hand smoke exposure, and alcohol use. Mother's exposure to PMi0 was based on residence for each
month of pregnancy, each trimester defined as a three month period, and the 6 wk prior to death.
Exposure was assigned by the nearest monitor. A 10 (ig/m3 increase in PM10 in the third trimester
was associated with an 8% (95% CI: 2-14) increase in risk of stillbirth.
In Sao Paulo, Brazil, Poisson regression of stillbirth counts for the period 1991-1992 found
that a 10 (ig/m3 increase in PMi0 was associated with a 0.8% increase in stillbirth rates (Pereira et al.,
1998, 007264). When other pollutants (NO2, SO2, CO, O3) were included simultaneously in the
model, the association did not remain. Stillbirths were defined as fetal loss at >28 wk of pregnancy
age, weight >1,000 g, or length of fetus >35 cm.
Neonatal Mortality and Neonatal Respiratory Mortality, <1 Month
Studies on PM and neonatal mortality (<1 month) included a time-series analysis of PMi0 for
4 yr of data (1998-2000) for Sao Paulo, Brazil (Lin et al., 2004, 095787). The analysis used daily
counts of deaths from government registries and adjusted for temporal trend, day of the week,
weather, and holidays. Findings indicated that a 10 (ig/m3 increase in PMi0 was associated with a
1.71% (95% CI: 0.31-3.32) increase in risk of neonatal death.
A case-crossover study of 11 yr (1989-2000) in Southern California did not find an association
between PMi0 and neonatal deaths (Ritz et al., 2006, 089819). Quantitative results were not
provided. The authors considered adjustment for season, county, parity, gender, prenatal care, and
maternal age, education, and race/ethnicity.
These results add to previous work on PM and neonatal death, including studies identifying
higher risk of neonatal mortality with higher TSP in the Czech Republic in an ecological analysis
(Bobak and Leon, 1992, 044415) and case-crossover study (Bobak and Leon, 1999, 007678). and a
December 2009 7-53
-------
Poisson model study in Kagoshima City, Japan (Shinkura et al., 1999, 090050). An ecological study
evaluated U.S. PMi0 data for the year 1990 using long-term pollution levels in 180 U.S. counties
(Lipfert et al., 2000, 004103). Analysis considered birth weight, sex, month of birth, location by state
and county, prenatal care, and mother's race, age, educational level, marital status, and smoking
status. County-level variables were included for socioeconomic status, altitude, and climate. Results
indicate a 13.1% increase in neonatal mortality (95% CI: 4.4-22.6) per 10 (ig/m3 PMi0 for non-low
birth weight infants. Statistically significant associations were also observed considering all infants
or low birth weight infants. However, higher levels of SO2 were associated with lower risk of infant
mortality. When sulfate and an estimate of non-sulfate particles were included in the regression
simultaneously, associations were observed with non-sulfate particles and an inverse relationship
with sulfate particles. Respiratory neonatal mortality was not associated with higher TSP in the
Czech Republic case-control study (Bobak and Leon, 1999, 007678).
Infant Mortality and Infant Respiratory Mortality, <1Year
A literature search did not reveal new studies on PM and infant mortality (<1 year) since the
previous PM AQCD. Previously conducted studies include a case-control study that reported
associations between infant mortality and TSP levels over the period between birth and death for
infants in the Czech Republic (Bobak and Leon, 1999, 007678). An ecological study evaluated U.S.
PMio data for the year 1990 using long-term pollution levels in 180 U.S. counties (Lipfert et al.,
2000, 004103). The authors found a 9.64% (95% CI: 4.60-14.9) increase in risk of infant mortality
for non-low birth weight infants per 10 ug/m3 increase in PMi0, a 13.4% (95% CI: -10.3 to 43.5)
increase in non-low birth weight respiratory-disease related deaths (ICD 9 460-519) and a 19.5%
(95% CI: 0.07-42.8) increase in all non-low birth weight respiratory-related infant deaths (ICD 9
460-519,769,770).
Postneonatal Mortality and Postneonatal Respiratory Mortality, 1 Month-1 Year
Several studies have been conducted on PM and postneonatal mortality since the previous PM
AQCD, including three from the U.S., one from Mexico, and three from Asia. Two case-control
studies examined the risk of PM to postneonatal death in California. Research focused on Southern
California for the period 1989-2000 linked birth and death certificates and considered PMi0 2 mo
prior to death with adjustment for prenatal care, gender, parity, county, season, and mother's age,
race/ethnicity, and education (Ritz et al., 2006, 089819). As previously noted, this study did not find
an association between PM10 and neonatal mortality (<1 month), however an association was
observed for post-neonatal mortality, with a 10 (ig/m increase in PMi0 associated with a 4% (95%
CI: 1-6) increase in risk. The exposure period of 2 wk before death was also considered, producing
effect estimates of 5% (95% CI: 1-10) for the same PMi0 increment. Even larger effect estimates
were observed for those who died at ages 4-12 mo. When CO, NO2, and O3 were simultaneously
included with PMi0 in the model, the central estimate reduced to 2% for the 2-wk exposure period
and 4% for the 2-mo exposure period, and both estimates lost statistical significance. The other case-
control study of California considered PM2.5 from 1999 to 2000 for infants born to mothers within
five miles of a PM2 5 monitoring station (Woodruff et al., 2006, 088758). Infants who died during the
postneonatal period were matched to infants with date of birth within 2 wk and birth weight
category. Exposure was estimated from the time of birth to death. Models considered parity and
maternal race, education, age, and martial status. A 10 (ig/m3 increase in PM25 was associated with a
7% (95% CI: -7 to 24) increase in postneonatal death
County-level PMio and PM25 for the first 2 mo of life for births in urban U.S. counties
(> 250,000 residents) from 1999 to 2002 were evaluated in relation to postneonatal mortality with
GEE models (Woodruff et al., 2008, 098386). Births were restricted to singleton births with
gestational age < 44 wk, same county of residence at birth and death, and non-missing data on birth
order, birth weight, and maternal race, education, and martial status. Higher levels of either PM
metric were associated with higher risk of postneonatal mortality, with 4% (95% CI: -1 to 10)
increase in mortality risk per 10 (ig/m3 in PMio and 4% (95% CI: -2 to 11) increase in mortality risk
for the same increment of PM25. This work builds on a previous study of 86 U.S. urban areas from
December 2009 7-54
-------
1989 to 1991, finding a 4% (95% CI: 2-7) increase in postneonatal mortality per 10 (ig/m3 county-
level PMio over the first 2 mo of life (Woodruff et al., 1997, 084271).
In Ciudad Juarez, Mexico, a case-crossover approach was applied to data from 1997 to 2001
based on death certificates and the cumulative PMi0 for the day of death and previous two days
(Romieu et al., 2004, 093074). A case-crossover study of Kaohsiung, Taiwan from 1994 to 2000
compared the average of PMi0 on the day of death and two previous days to PMi0 in control periods
a week before and week after death (Tsai et al., 2006, 090709). A similar approach was also applied
to 1994-2000 data from Taipei, Taiwan, also using case-crossover methods for the lag 0-2 PMi0 with
referent periods the week before and after death (Yang et al., 2006, 090760). In these case-crossover
studies, season was addressed through matching in the study design. A 10 (ig/m3 increase in PMi0
was associated with a 2.0% (95% CI: -2.8 to 7.0) increase in the Mexico study, a 0.59 (95% CI: -15.0
to 18.8) increase in postneonatal death in the Kaohsiung study, and a 1.02% (95% CI: -13.2 to 17.6)
increase in the Taipei study. A study in Seoul, South Korea from 1995 to 1999 used time-series
approaches adjusted for temporal trend and weather, based on national death registries excluding
accidental deaths (Ha et al., 2003, 042552). A 10 (ig/m3 increase in PMi0 was associated with a
3.14% (95% CI: 2.16-4.14) increase in risk of death for postneonates.
A subset of the studies examining postneonatal mortality also considered the subset of
postneonatal deaths from respiratory causes. These include the time-series study in South Korea,
finding a 17.8% (95% CI: 14.4-21.2) increase in respiratory-mortality per 10 (ig/m3 increase in PM10
(Ha et al., 2003, 042552) and the case-crossover study in Mexico, for which the same increment in
PMi0 was associated with a 1.5% (95% CI: -14.1 to 13.0) decrease in risk (Romieu et al., 2004,
093074). Both California case-control studies identified associations, with a 5% (95% CI: 1-10)
increase in risk in Southern California (Ritz et al., 2006, 089819) and 57.4% (95% CI: 7.0-132)
increase in California per 10 (ig/m3 PM10 (Woodruff et al., 2006, 088758). The U.S. study found this
increment in PM10 to be linked with a 16% (95% CI: 6.0-28.0) increase in respiratory postneonatal
mortality, although effect estimates for PM2.5 were not statistically significant (Woodruff et al., 2008,
098386). Earlier studies on respiratory-related postneonatal mortality include the study of 86 U.S.
urban areas, finding statistically significant effects (Woodruff et al., 1997, 084271).
Sudden Infant Death Syndrome
Three studies examining the relationship between PM and sudden infant death syndrome
(SIDS) have been published from 2002 onward. These studies examined infant mortality and were
thereby discussed in this section previously. A case-control study over a 12-year period (1989 to
2000) matched 10 controls to deaths (cases) in Southern California (Ritz et al., 2006, 089819). A
10 (ig/m3 increase in PM10 the 2 mo prior to death was associated with a 3% (95% CI: -1 to 8)
increase in SIDS. Adjusted for other pollutants (CO, NO2, and O3), the effect estimate reduced to 1%
(95% CI: -5 to 7).
A case-control study, also based in California, found an OR of 1.008 (95% CI: 1.006-1.012)
per 10 (ig/m3 increase in PM2.5, considering a SIDS definition of ICD 10 R95 (Woodruff et al., 2006,
088758). Due to changes in SIDS diagnosis, another SIDS definition was explored for ICD 10 R99
in addition to ICD 10 R95. Under this SIDS definition, the effect estimate changed to 1.03 (95% CI:
0.79-1.35). The authors also examined whether the relationship between PM2.5 and SIDS differed by
season, finding no significant difference. PMi0 and PMi0_2.5 were not associated with risk of SIDS;
numerical results were not provided for these PM metrics. The third recent study of PM and SIDS
examined U.S. urban counties from 1999 to 2002 (Woodruff et al., 2008, 098386). Statistically non-
significant relationships were observed between SIDS and PMi0 or PM25 in the first 2 mo of life.
These studies add to earlier work, such as a U.S. study that found higher risk of SIDS with
higher annual PM25 levels, including in a separate analysis of normal birth weight infants (Lipfert et
al., 2000 004103). and a U.S. study identifying a 12% (95% CI: 7-17) increase in SIDS risk per
10 (ig/m3 in PMi0 for the first 2 mo of life for normal weight births (Woodruff et al., 1997, 084271).
A study based on Taiwan found higher SIDS risk with lower visibility (Knobel et al., 1995, 155905).
whereas a 12-city Canadian time-series study identified no significant associations (Dales et al.,
2004, 087342).
Deaths by SIDS were identified by different methods in the studies, partly due to transition
from ICD 9 to ICD 10 codes, but also due to different choices within the research design. Two
studies examined multiple approaches (ICD 10 R95, ICD 10 R95 and R99) (Woodruff et al., 2006,
December 2009 7-55
-------
088758: Woodruff et al, 2008, 098386). and other studies investigated ICD 9 798.0 and ICD 10 R95
(Ritz and Wilhelm, 2008, 156914). ICD 9 798.0 (Woodruff et al., 1997, 084271). ICD 9 798.0 and
799.0 (Knobel et al., 1995, 155905). as well as a sudden unexplained death of infant <1 year for
which an autopsy did not identify a specific cause of death (Dales et al., 2004, 087342). These
variations in the definition of health outcomes add to differences in populations and study designs.
Although some findings indicate a potential effect of PM on risk of SIDS, with the strongest
evidence perhaps from the case-control study in California (Woodruff et al., 2006, 088758). others
do not find an effect or observe an uncertain association. For the relationship between PM and SIDS,
a 2004 review article concluded consistent evidence exists compared to evidence for other infant
mortality effects (Glinianaia et al., 2004, 087898). whereas other reviews found weaker or
insufficient evidence (Heinrich and Slama, 2007, 156534). Another review concluded that the
scientific literature on air pollution and SIDS suggests an effect, but that further research is needed to
draw a conclusion (Tong and Colditz, 2004, 087883).
Considerations for Comparisons across Studies
Comparison of results across studies can be challenging due to several issues, including
differences in methodologies, populations and study areas, pollution levels, and the exposure
timeframes used. Given the large variation in study designs, the methods to address potential
confounders vary. For example, weather and season were addressed in the case-control studies by
matching, in the time-series study through non-linear functions of temperature and temporal trend,
and in the ecological study through county-level variables. All studies included consideration of
seasonality and weather. Researchers used different definitions of respiratory-related deaths,
including ICD 9 460-519 (Bobak and Leon, 1999, 007678: Lipfert et al., 2000, 004103): ICD 9
460-519, 769-770 (Lipfert et al., 2000, 004103): ICD 9 460-519, 769, 770.4, 770.7, 770.8, 770.9,
and ICD 10 JOO-J98, P22.0, P22.9, P27.1, P27.9, P28.0, P28.4, P28.5, and P28.9 (Ritz et al., 2006,
089819): and ICD 9 460-519 and ICD 10 JOO-J99 for any cause on death certificate (Romieu et al.,
2004, 093074): ICD 10 JOO-99 and P27.1 excluding J69.0 (Woodruff et al., 2006, 088758: Woodruff
et al., 2008, 098386): and ICD 9 460-519 (Woodruff et al., 1997, 084271).
Socioeconomic conditions were included at the individual level, typically maternal education,
in many studies (e.g., Bobak and Leon, 1999, 007678: Ritz and Wilhelm, 2008, 156914: Ritz et al.,
2006, 089819: Woodruff et al., 1997, 084271: Woodruff et al., 2006, 088758) and at the community-
level in others (e.g., Bobak and Leon, 1992, 044415: Penna and Duchiade, 1991, 073325) or for
both individual and community-level data (e.g., Lipfert et al., 2000, 004103). The time-series
approach is unlikely to be confounded by socioeconomic and other variables that do not exhibit day-
to-day variation. Similarly, case-crossover methods use each case as his/her own control, thereby
negating the need for individual-level confounders such as socioeconomic status (e.g., Romieu et al.,
2004, 093074: Tsai et al., 2006, 090709: Yang et al., 2006, 090760). All studies published after 2001
incorporated individual-level socioeconomic data or were of case-crossover or time-series design.
One study specifically examined whether socioeconomic status modified the PM and mortality
relationship, dividing subjects into three socioeconomic strata based on the ZIP code of residence at
death (Romieu et al., 2004, 093074). This work, based in Mexico, found that at lower socio-
economic levels the association between PMi0 and postneonatal mortality increased. Although the
overall association showed higher risk of death with higher PMi0 with statistical uncertainty, for the
lowest socio-economic group, a 10 (ig/m3 increment in cumulative PMi0 over the 2 days before death
was associated with a 60% (95% CI: 3-149) increase in postneonatal death. A trend of higher effect
for lower socio-economic condition is observed in all 3 lag structures.
Studies differ in terms of the time frame of pregnancy that was used to estimate exposure.
Exposure to PM for infant mortality (<1 yr) was estimated as the levels between birth and death
(Bobak and Leon, 1999, 007678). annual community levels (Lipfert et al., 2000, 004103: Penna
and Duchiade, 1991, 073325) and the 3-5 days prior to death (Loomis et al., 1999, 087288). For
neonatal deaths, exposure timeframes considered were the time between birth and death (Bobak and
Leon, 1992, 044415: Bobak and Leon, 1999, 007678). annual levels (Bobak and Leon, 1999,
007678: Lipfert et al., 2000, 004103). monthly levels (Shinkura et al., 1999, 090050). the same day
concentrations (Lin et al., 2004, 095787). and the 2 mo or 2 wk prior to death (Ritz et al., 2006,
089819). Postneonatal mortality was associated with PM concentrations based on annual levels
(Bobak and Leon, 1992, 044415: Lipfert et al., 2000, 004103). between birth and death (Bobak and
December 2009 7-56
-------
Leon, 1999, 007678; Woodruff et al, 2006, 088758). 2 mo before death (Ritz et al., 2006, 089819).
the first 2 mo of life (Woodruff et al., 1997, 084271; Woodruff et al., 2006, 088758). the day of death
(Ha et al., 2003, 042552). and the average of the same day as death and previous 2 days (Romieu et
al., 2004, 093074; Tsai et al., 2006, 090709; Yang et al., 2006, 090760). Thus, no consistent window
of exposure was identified across the studies.
PMio concentrations were highest in South Korea (69.2 (ig/m3) (Ha et al., 2003, 042552) and
Taiwan (81.45 (ig/m3) (Tsai et al., 2006, 090709). and lowest in the U.S. (29.1 (ig/m3) (Woodruff et
al., 2008, 098386) and Japan (21.6 ug/mj) (Shinkura et al., 1999, 090050). All studies used
community-level exposure information based on ambient monitors, as opposed to exposure
measured at the individual level (e.g., subject's home) or personal monitoring.
Given similar sources for multiple pollutants (e.g., traffic), disentangling the health responses
of copollutants is a challenge in the study of ambient air pollution. Several studies examined
multiple pollutants, most by estimating the effect of different pollutants through several univariate
models. Some studies noted the difficulty of separating PM effects from those of other pollutants,
but noted stronger evidence for particles than other pollutants (Bobak and Leon, 1999, 007678). A
few studies applied copollutant models by including multiple pollutants simultaneously in the same
model. Effect estimates for the relationship between PMi0 and neonatal deaths in Sao Paulo were
reduced to a null effect when SO2 was incorporated (Lin et al., 2004, 095787). Associations between
PM10 and postneonatal mortality or respiratory postneonatal mortality remained but lost statistical
significance in a multiple pollutant model with CO, NO2, and O3 (Ritz et al., 2006, 089819).
Several review articles in recent years have examined whether exposure to PM affects risk of
infant mortality, generally concluding that more consistent evidence has been observed for
postneonatal mortality, particularly from respiratory causes (Bobak and Leon, 1999, 007678;
Heinrich and Slama, 2007, 156534; Lacasana et al., 2005, 155914; Sram et al., 2005, 087442). In
one review authors identified 14 studies on infant mortality and air pollution and determined that
studies on PM and infant mortality do not provide consistent results, although more evidence was
present for an association for some subsets of infant mortality such as postneonatal respiratory-
related mortality (Bobak and Leon, 1999, 007678). The relationship between PM and postneonatal
respiratory mortality was concluded to be causal in one review (Sram et al., 2005, 087442). and
strong and consistent in another (Heinrich and Slama, 2007, 156534). Meta-analysis using inverse-
variance weighting of PM10 studies found that a 10 ug/m3 increase in acute PM10 exposure was
associated with 3.3% (95% CI: 2.4-4.3) increase in risk of postneonatal mortality, whereas the same
increment of chronic PMi0 exposure was linked with a 4.8% (95% CI: 2.2-7.2) increase in
postneonatal mortality and a 21.6% (95% CI: 10.2-34.2) increase for respiratory postneonatal
mortality (Lacasana et al., 2005, 155914).
Studies that examined multiple outcomes and ages of death allow a direct comparison based
on the same study population and methodologies, thereby negating the concern that inconsistent
results are due to underlying variation in population, approaches, etc. In this review, one study, based
in Southern California identified no association for neonatal effects (numerical results not provided)
but statistically significant results for postneonatal mortality (Ritz et al., 2006, 089819). Figure
7-5compares risk for the postneonatal period for respiratory and total mortality. In six of the seven
studies, higher effect estimates were observed for respiratory-related mortality. Results from the
neonatal period found higher effects for total mortality compared to respiratory mortality (Bobak
and Leon, 1999, 007678) and the reverse for a study examining infant mortality (Lipfert et al., 2000,
004103). Thus, there exists evidence for a stronger effect at the postneonatal period and for
respiratory-related mortality, although this trend is not consistent across all studies.
December 2009 7-57
-------
20
Neonatal
40
60
60
Figure 7-5. Percent increase in postneonatal mortality per 10 ug/m3 in PMio, comparing risk
for total and respiratory mortality. Panel a (left) provides central estimates; panel
b (right) also adds the 95% intervals. The points reflect central estimates and the
lines the 95% intervals. Solid lines represent statistically significant effect
estimates; dashed lines represent non-statistically significant estimates.1
7.4.1.6. Decrements in Sperm Quality
Limited research conducted in the Czech Republic on the effect of ambient air pollution on
sperm production has found associations between elevated air pollution and decrements in
proportionately fewer motile sperm, proportionately fewer sperm with normal morphology or normal
head shape, proportionately more sperm with abnormal chromatin (Selevan et al., 2000, 012578).
and an increase in the percentage of sperm with DNA fragmentation (Rubes et al., 2005, 078091).
These results were not specific to PM, but for exposure to a high-, medium- or low-polluted air
mixture. Similarly, in Salt Lake City, Utah, PM2.5 was associated with decreased sperm motility and
morphology (Hammoud et al., 2009, 192156). Research in Los Angeles, California examined 5,134
semen samples from 48 donors in relation to ambient air pollution measured 0-9, 10-14, 70-90 days
before semen collection over a 2-yr period (1996-1998). Ambient O3 during all exposure periods had
a significant negative correlation with average sperm concentration, and no other pollutant measures
were significantly associated with sperm quality parameters, or presented quantitatively (Sokol et al.,
2006, 098539).
7.4.2. lexicological Studies
This section summarizes recent evidence on reproductive health effects reported with exposure
to ambient PM; no evidence was presented in this area in the 2004 PM AQCD. Studies from
different toxicological rodent models allow for investigation of specific mechanisms and modes of
1 Studies included are Bobak and Leon (1999, 007678). Ha et al. (2003, 042552). Ritz et al. (2006, 089819). Romieu et al. (2004, 093074).
Romieu et al. (2008, 156922). Woodruff et al. (1997, 084271). Woodruff et al. (2006, 088758). Findings from Bobak and Leon (1999,
007678) were based on TSP and were converted to PMio estimates assuming PMio/TSP = 0.8 as per summary data in the original article
(Bobak and Leon, 1999, 007678). Findings from Woodruff et al. (1997, 084271) for respiratory-related mortality were based on non-low
birth weight infants. Results for Woodruff et al. (2006, 088758) were based on PM2.5 and were converted to PMio assuming
PM2.5/PM10 = 0.6.
December 2009
7-58
-------
action for reproductive changes. Emphasis is placed here on results from different windows of
development, i.e., exposure in utero, neonatally or as an adult can affect reproductive outcomes as an
adult. In addition, studies evaluating whether fertility is affected in female and/or male animals by a
similar exposure, and how exposures are transmitted to the fertility of the F i offspring, are
summarized. Hormonal changes which can lead to decreased sperm count or changes in the estrous
cycle are also of interest. Studies of pregnancy losses and placental sufficiency are also reported.
Most recently, the role of environmental chemicals in shifting sex ratios (also seen in epidemiologic
studies) and in affecting heritable DNA changes have become outcomes of interest.
7.4.2.1. Female Reproductive Effects
Urban Air
Windows of exposure are important in determining reproductive success as an adult. Exposure
as a neonate may have a drastically different impact than does a similar adult exposure. To test this,
female BALB/C mice were exposed to ambient air in Sao Paulo as neonates or as adults and then
were bred to non-exposed males (Mohallem et al., 2005, 088657). Ambient concentrations of the
pollutants CO, NO2 PNV and SO2 were 2.2 ± 1.0 ppm, 107.8 ± 42.3 ^g/rn3, 35.5 ± 12.8 j-ig/m3,
and 11.2 ± 5.3 |J,g/m , respectively. They reported decreased fertility in animals exposed as
newborns, but not in adult-exposed female BALB/c mice. There were a significantly higher number
of liveborn pups from dams housed in filtered chambers (PM and gaseous components removed)
versus animals exposed to ambient air as newborns. There was also a higher incidence of
implantation failures in dams reared as newborns in polluted chambers. Sex ratio, number of
pregnancies per group, resorptions, fetal deaths, and fetal placental weights did not differ
significantly by exposure group. Thus, in these studies, exposure to ambient air pollution affected
future reproductive success of females if they were exposed as neonates and not if exposed as adults.
Diesel Exhaust
Significant work has been done in male rodent models to determine the effect of PM exposure
on reproductive outcomes, with fewer studies conducted using female rodents. Tsukue et al. (2004,
096643) exposed pregnant C57-BL mice to DE (0.1 mg/m3) or to clean air (controls) for 8 h/day
from GD2-13. The concentration of the gaseous materials including NO, NOX, NO2, CO and SO2
are 2.2 ± 0.34 ppm, 2.5 ± 0.34 ppm, 0.0 ppm, 9.8 ± 0.69 ppm, and <0.1 ppm (not detectable),
respectively. At GD14 female fetuses were collected for analysis of mRNAfor two genes involved in
sexual differentiation (Ad4BP-l/SF-l and MIS), and found no significant changes. Work by Yoshida
et al. (2006, 097015) showed changes in these two transcripts in male ICR fetuses exposed to similar
concentrations of DE, albeit with different daily durations of exposure. Further work by Yoshida et
al. (2006, 097015) showed that of three mouse strains tested, ICR male fetuses were the most
sensitive to DE-dependent changes in these two genes. Nonetheless, strain sensitivity to DE particles
may also differ by sex. Thus, it appears that female mice exposed in utero to DE show a lack of
response at the mRNA level of MIS or Ad4bP-l/SF-l, important genes in male sexual differentiation
that showed DE-dependent changes in male pups from dams exposed in utero. Female fetuses have
shown a decrease in BMP-15, which is related to oocyte development (Tsukue et al., 2004, 096643).
A sensitive measure of androgenic activity in male rodents is anogenital distance (AGD), i.e.,
decreased AGD is seen with exposure to anti-androgenic environmental chemicals, the phthalates
(Foster et al., 1980, 094701: Foster et al., 2001, 156442). To assess the role of DE exposure on
reproductive success and anti-androgenic effects on offspring, Tsukue et al. (2002, 030593) exposed
6 week-old female C57-B1 mice to 4 mo of DE (0.3, 1.0, or 3.0 mg/m3; PM MMAD of 0.4 urn) or
filtered air. DE-exposed estrous females had significantly decreased uterine weight (1.0 mg/m ).
Some of the DE-exposed females were bred to unexposed males and DE-exposure led to increased,
albeit not significantly increased, rates of pregnancy loss in mated females (up to 25%). Offspring
were weighed after birth and decreases in body weight were observed at 6 and 8 wk (males and
females, 1.0 and 3.0 mg/m3) and 9 wk (females, 1.0 and 3.0 mg/m3). Anogenital distance was
decreased in 70-day old DE-exposed male offspring (0.3 mg/m3). In female offspring at 70 days of
December 2009 7-59
-------
age, lower organ weights (adrenals, liver and thymus) were observed (1.0 mg/m3) compared to
controls; thymus weight of the 0.3 mg/m females was also lower at 70 days. Crown to rump length
in females from dams exposed to DE (1.0 and 3.0 mg/m3) was less than the control group. In
conclusion, adult exposure to DE led to maternal-dependent reproductive changes that affected
outcomes in offspring that manifested as decreased pup body weight, anti-androgenic effects like
decreased AGD and decreased organ weight (which may have been confounded by changes in body
weight because weights were not reported as relative organ weights).
7.4.2.2. Male Reproductive Effects
Diesel Exhaust
Studies were performed to determine PM-dependent strain sensitivity of the male reproductive
tract using male steroidogenic enzymes as the model pathway. Three strains of pregnant mice (ICR,
C57B1/6J or ddY mice) were continuously exposed to DE at 0.1 mg/m3 via inhalation or clean air
over gestational days 2-13 (Yoshida et al., 2006, 156170). At GDI4, dams were euthanized and
fetuses were collected. Male fetuses were collected from each dam for mRNA analysis of genes
related to male gonad development including Mullerian inhibiting substance (MIS; crucial for sexual
differentiation including Mullerian duct regression in males), steroid transgenic factor (Ad4BP/SF-l,
an enzyme in the testosterone synthesis pathway), cytochrome P450 cholesterol side chain cleavage
enzyme (P450scc), and other steroidogenic enzymes [17(3-hydroxysteroid dehydrogenase (HSD),
cytochrome P450 17-a-hydroxylase (P450cl7), and 3-|3hydroxysteroid dehydrogenase (3(3HSD)].
There were significant decreases in MIS (ICR and C57BL/6 mice) and Ad4BP/SF-l (ICR mice)
compared to the control groups. The ddY strain demonstrated no changes in Ad4BP/SF-l or MIS,
which may be due to marked changes in 3(3-hD expression compared to non-DE exposed controls.
From these studies, it appears that mouse strains with in utero exposure to DE show differential
sensitivity in gonadal differentiation genes (mRNA) expression in male offspring; ICR are the most
sensitive, followed by C57BL/6, with ddY mice being the least sensitive.
Yoshida et al. (2006, 097015) also monitored changes in the male reproductive tract after in
utero exposure to DE. Timed-pregnant ICR dams were exposed during gestation (2 days post-coitus
[dpc]-16 dpc) to continuous DE (0.3, 1.0 or 3.0 mg/m3) or clean air. The reproductive tracts of male
offspring were monitored at 4 wk postnatally. These pups received possible continued exposure
through lactation as dams were exposed to DE during gestation and nursed pups. Exposure to 0.3
mg/m of DE had no effect on male reproductive organ weight or serum testosterone. The
intermediate concentration of 1.0 mg/m3 induced increases in serum testosterone. Exposure to the
higher concentration (1.0 and 3.0 mg/m3) of DE led to significant increases in reproductive gland
weight (testis, prostate, and coagulating gland). The organ weights are presented as absolute numbers
and not adjusted for body weight, which is sometimes problematic for complete representation of
hormonal changes, as body weight may confound absolute organ weight changes. Transcripts
relating to male sexual differentiation (MIS and AD4BP/SF-1, 1.0 and 3.0 mg/m3) were also
significantly decreased. Sexual differentiation is a tightly regulated process and these changes in
transcription may lead to changes that can affect genitalia development.
The effects of DE exposure on male spermatogenesis have also been demonstrated. Exposure
of pregnant ICR mice to DE (2-16 dpc continuous inhalation exposure to 1.0 mg/m3 or filtered clean
air) led to impaired spermatogenesis in offspring (Ono et al., 2007, 156007). Male offspring were
followed at PND 8, 16, 21 (3 wk), 35 (5 wk) and 84 (12 wk). After 16 dpc, but before termination of
the study, all of the animals were transferred to a regular animal care facility and received clean air
exposure until the termination of the study. No cross fostering was performed in this experiment, so
pups that were born to DE-exposed dams were also nursed on these dams and may have received
lactational exposure to DE. The gaseous components of the diluted DE included NO, NO2, SO2, and
CO2 at concentrations of 11.75 ± 1.18, 4.62 ± 0.36, 0.21 ± 0.01, and 4922 ± 244 ppm, respectively.
Body weight was significantly depressed at PNDs 8 and 35. Accessory gland relative weight was
significantly increased at PND8 and PND 16 only. Serum testosterone was significantly decreased at
3 wk and was significantly increased at 12 wk. At 5 and 12 wk, daily sperm production (DSP) was
significantly decreased. FSH receptor and StAR mRNA levels were significantly increased at 5 and
12 wk, respectively. Relative testis weight and relative epididymal weight were unchanged at all
December 2009 7-60
-------
time points. Histological changes showed sertoli cells with partial vacuolization and a significant
increase in testicular multinucleated giant cells in the seminiferous tubules of DE-exposed animals
compared to control. This study indicates that in utero exposure to DE had effects on
spermatogenesis in offspring at the histological, hormonal and functional levels.
In utero exposure to DE and its effect on adult body weight, sex ratio, and male reproductive
gland weight was measured by Yoshida et al. (2006, 097015). Pregnant ICR mice were exposed by
inhalation to DE (0.3, 1.0 or 3.0 mg/m3) or clean air from 2 dpc to 16 dpc. Pups were allowed to
nurse in clean air on exposed dams until weaning and at PND28, male pups were sacrificed. At this
time, serum testosterone and pup reproductive gland weight was determined. Significant increases in
relative reproductive organ weights were reported at 1.0 and 3.0 mg/m3 for the seminal vesicle,
testis, epididymis, coagulating gland, prostate and liver. Male pup serum testosterone was
significantly increased at 1.0 mg/m3. Mean testosterone positively correlated with testis weight, DSP,
aromatase and steroidogenic enzyme message levels (P450cc, c!7 lyase, and P450 aromatase). Sex
ratio did not differ in DE-exposed animals versus control. Male pup body weight of DE-exposed
animals was significantly increased at PND28 (1.0 and 3.0 mg/m3). These studies show that in utero
DE-exposure led to increased serum testosterone and increased reproductive gland weight in male
offspring early in life.
The effects of DE on murine adult male reproductive function were studied by exposing ICR
male mice (6 wk of age) to DE (clean air control, 0.3, 1.0 or 3.0 mg/m3) for 12 h/day for 6 mo with
another group receiving a 1-mo recovery of clean air post-exposure (Yoshida and Takedab, 2004,
097760). After 6 mo of DE exposure, there was a concentration-dependent increase in degeneration
of seminiferous tubules and a decrease in DSP/g of testis tissue. After 6 mo exposure to DE particles
plus 1 mo of recovery in clean air, significant decreases remained in DSP at the two highest
concentrations. The effect of ingestion of deposited PM on the fur with grooming cannot be ruled out
as a possible exposure pathway in this experiment.
To expand on PM-dependent changes in spermatogenesis, an eloquent DE-exposure model
was designed to determine if PM or the gaseous phase of DE was responsible for changes in sperm
production in rodents (Watanabe, 2005. 087985). Pregnant dams (F344/DuCrj rats) exposed to DE
(6 h/day exposure to 0.17 or 1.71 mg/m ; <90% of PM less than 0.5 um; NO2 concentrations 0.10
and 0.79 ppm, respectively) or filtered air (removing PM only, low concentration filtered air and
high concentration filtered air) from GD7 to parturition produced adult male offspring with a
decreased number of sertoli cells and decreased DSP (PND 96) when compared to control mice
exposed to clean air. The concentrations of NO2 for the low and high filtered exposure groups were
0.1 and 0.8 ppm, respectively. Because both PM-filtered and DE-exposure groups showed the same
outcomes, the effects are likely due to gaseous components of DE.
Motorcycle Exhaust
Adult male (8-wk old) Wistar rats were exposed to motorcycle exhaust (ME) for 1 h in the
morning and 1 h in the afternoon (5 day/wk) at 1:50 dilution for 4 wk (group A), 1:10 dilution for
2 wk (group B) or 4 wk (group C), or to clean air (Huang et al., 2008, 156574). After 4 wk of
exposure, both exposed groups had significantly decreased body weight compared to the control
group. All three ME exposure groups showed a decreased number of spermatids in the testis. Both
1:10 exposure groups also demonstrated decreased caudal epididymal sperm counts. Group C had
significant decreased testicular weight, decreased mRNA expression for the cytochrome P450
substrate 7-ehtoxycoumarin O-de-ethylase, and increased IL-6, IL-lp, and COX-2 mRNA levels.
Decreased protein levels of the antioxidant, superoxide dismutase, and increased IL-6 protein were
reported for group C when compared to control. In addition, serum testosterone was significantly
decreased in group C. Co-treatment with the antioxidant vitamin E resulted in partial attenuation of
serum testosterone levels and caudal epididymal sperm counts, and returned IL-6, IL-1(3, and COX-2
ME exposure-dependent message levels to baseline. The glutathione antioxidant system and lipid
peroxidation were unchanged. In conclusion, male animals exposed to ME showed significant
decrements in body weight, spermatid number, and serum testosterone with an increase in
inflammatory cytokines. Vitamin E co-treatment with ME-exposure led to an attenuation of
inflammation and a partial rescue of testosterone levels and sperm numbers.
December 2009 7-61
-------
Summary of Toxicological Study Findings for Male Reproductive Effects
In summary, laboratory animals exposed to DE in utero or as adults manifest with abnormal
effects on the male reproductive system. In utero exposure to DE induced increased reproductive
gland weight and increased serum testosterone in early life (PND28), which may lead to early
puberty (albeit not measured in this study). With similar in utero DE exposures, later life outcomes
include decreased DSP, aberrant sperm morphology, and hormonal changes (testosterone and FSHr
decrements). Chronic exposure of adult mice to DE also induced decreased DSP and seminiferous
tubule degeneration. DE-dependent effects on male reproductive function have been reported in
multiple animal models, with only one model separating exposure based on particulate versus
gaseous components. DE and filtered air (gaseous phase only) exposure in utero induced sertoli cell
and DSP decrements in both groups, indicating that the gaseous phase of DE was causative. Adult
male rats exposed to ME manifested with decreased spermatid number, serum testosterone, and an
increase in inflammatory cytokines. Significant effects on the male reproductive system have been
demonstrated after exposure to ambient PM sources (DE or ME). Nonetheless, these models often
include a complex mixture of gaseous component and PM exposure, which makes interpreting the
contribution from PM alone difficult.
7.4.2.3. Multiple Generation Effects
Urban Air
Veras et al. (2009, 190496) investigated pregnancy and female reproductive outcomes in
BALB/c female mice exposed to ambient air or PM-filtered ambient air at one of two different time
periods (before conception and during pregnancy) near an area of high traffic density in Sao Paulo,
Brazil. Exposures were 27.5 and 6.5 ug/m PM2.5 for ambient and PM-filtered air chambers,
respectively, with 101 ug/m3 NO2, 1.81 ug/m3 CO, and 7.66 ppm SO2 in both chambers. Two groups
of 2nd generation (G2) nulliparous female mice were continuously exposed from birth. Estrous
cyclicity and ovarian follicle classification were followed at PND60 (reproductive maturation) in one
group. A further group was subdivided into four groups by exposures during pregnancy following
reproductive capability and pregnancy outcomes of the G2 mice. Animals exposed to ambient air
versus PM-filtered air had an extended time in estrous and thus, a reduction in the number of cycles
during the study period. The number of antral follicles was significantly decreased in the ambient air
versus the PM-filtered air animals. Other follicular quantification (number of small, growing or
preovulatory follicles) showed no differences between the two chambers. There was an increase in
the time necessary for mating, a decrease in the fertility index, and an increase in the pregnancy
index in the ambient air group versus the PM-filtered group. Specifically, in the ambient air groups,
there was a significant increase in rate of the post-implantation loss in Gl and G2 groups. However,
there was no statistically significant change in number of pups in the litter. Fetal weight was
decreased in all treatment groups (ambient air groups Gl and G2, and PM-filtered G2) when
compared to the PM-filtered Gl group or animals raised entirely in filtered air, showing that fetal
weight was affected by both pre-gestational and gestational PM exposure.
PM exposure prior to conception is associated with increased time in estrous, which in other
animal models can be related to ovarian hormone dysfunction and ovulatory problems. These estrous
alterations can contribute to fecundity issues. There was no significant difference in number of
preovulatory follicles in the above model, but there was a statistically significant decrease in the
number of antral follicles (Veras et al., 2009, 190496). Antral follicles are the last stage in follicle
development prior to ovulation, and a decrease in antral follicle number can be related to premature
reproductive senescence, premature ovarian failure, or early menopause, which were not followed in
this study.
In this study (Veras et al., 2009, 190496). the males that were used to generate the Gl and G2
groups were also exposed to ambient air or PM-filtered ambient air, and thus the reproductive
contribution of these males to the overall fertility and mating changes in the females cannot be
totally eliminated as a possible confounder to the observed effects. Thus, these effects are hard to
differentiate as male- or female-dependent and likely indicate a general loss of reproductive fitness.
Interestingly, both pre- and gestational exposure to ambient air induced a significant loss in post-
December 2009 7-62
-------
implantation of fetuses and this may be related to placental insufficiency as has been described in
other work by this lab (Veras et al., 2008, 190493).
7.4.2.4. Receptor Mediated Effects
Arylhydrocarbon Receptor (AhR)
Diesel Exhaust Particles
The AhR is often activated by chemicals classified as endocrine disrupting compounds
(EDCs), exogenous chemicals that behave as hormonally active agents, disrupting the physiological
function of endogenous hormones. DE particles are known to activate the AhR. A recent study by
Izawa et al. (2007, 190387) showed that certain polyphenols (quercetin from the onion) and food
extracts (Ginkgo biloba extract) are able to attenuate DE particle-dependent AhR activation when
measured with the Ah-Immunoassay, thus possibly attenuating the EDC activity of DE particles.
7.4.2.5. Developmental Effects
Sex Ratio
Urban Air
A correlation between PM10 exposure and a decrease in standardized sex ratios (SSRs) has
been reported in humans exposed to air pollution (Lichtenfels et al., 2007, 097041; Wilson et al.,
2000, 010288). with fewer numbers of male births reported. To understand this shift, two groups
(control and exposed) of male Swiss mice were housed concurrently in Sao Paulo and received either
ambient air exposure or filtered air (chemical and particulate filtering) from PND10 for 4 mo
(Lichtenfels et al., 2007, 097041). Filtration efficiency for PM2.5, CB, and NO2 inside the chamber
was found to be 55%, 100%, and 35%, respectively. After this exposure, non-exposed females were
placed in either chamber to mate. After mating, the males were sacrificed and testes collected; males
exposed to ambient air showed decreased testicular and epididymal sperm counts, decreased total
number of germ cells, and decreased elongated spermatids, but no significant change in litter size.
Females were housed in the chambers and sacrificed on GD19 when the number of pups born alive
and the sex ratio were obtained. There was a significant decrease in the SSR for pups born after
living in the ambient air-exposed chamber compared to the filtered chamber. In this study, a shift in
SSR has been shown for both humans and rodents exposed to air pollution, but other studies with DE
exposure (Yoshida et al., 2006, 156170) or ambient air in Sao Paulo (Mohallem et al., 2005, 088657)
showed no changes in rodent sex ratio. Possible exposure to PM and other components of ambient
air via ingestion during grooming cannot be ruled out in this rodent model.
Immunological Effects: Placenta
Diesel Exhaust
Placental insufficiency can lead to the loss of a pregnancy or to adverse fetal outcomes. DE-
exposure has been shown to induce inflammation in various models. Fujimoto et al. (2005, 096556)
assessed cytokine/immunological changes of DE-dependent inhalation exposure on the placenta
during pregnancy. Pregnant Slc:CR mice were exposed to DE (0.3, 1.0, or 3.0 mg/m3; PM MMAD
of 0.4 (im) or clean air from 2 to 13 dpc and dams, placenta, and pups were collected at 14 dpc.
There was a significant increase in the number of absorbed placentas in DE-exposed animals (0.3
December 2009 7-63
-------
and 3.0 mg/m3) with a significant decrease in the number of absorbed placentas in DE-exposed
animals at the middle concentration (1.0 mg/m3). Absorbed placentas from DE exposed mice had
undetectable levels of CYP1A1 and twofold increases in TNF-a; CYP1A1 placental mRNAfrom
healthy placentas of DE-exposed mice was unchanged versus control. IL-2, IL-5, IL-12a, IL-12B and
GM-CSF mRNA significantly increased in placentas of DE-exposed animals (0.3 and 3.0 mg/m ).
Fujimoto et al. (2005, 096556) reported DE-induced significant increases in multiple inflammatory
markers in the placenta with significant increases in the number of absorbed placentas.
Immunological Effects: Asthma
Model Particles
In utero exposure may confer susceptibility to PM-induced asthmatic responses in offspring.
Exposure of pregnant BALB/c mice to aerosolized ROFA leachate by inhalation or to DE particles
intranasally increases asthma susceptibility to their offspring (Fedulov et al., 2008, 097482; Hamada
et al., 2007, 091235). The offspring from dams exposed for 30 min to 50 mg/mL ROFA 1, 3, or
5 days prior to delivery responded to OVA immunization and aerosol challenge with airway
hyperreactivity and increased antigen-specific IgE and IgGl antibodies (Hamada et al., 2007,
091235). Airway hyperreactivity was also observed in the offspring of dams intranasally instilled
with 50 (ig of DE particles or TiO2, or 250 (ig CB, indicating that the same effect could be
demonstrated using relatively "inert" particles (Fedulov et al., 2008, 097482). Pregnant mice were
particularly sensitive to exposure to DE or TiO2 particles, and genetic analysis indicated differential
expression of 80 genes in response to TiO2 in pregnant dams. Thus pregnancy and in utero exposure
may enhance responses to PM, and exposure to even relatively inert particles may result in offspring
predisposed to asthma.
Placental Morphology
Urban Air
Exposure to ambient air pollution during pregnancy is associated with reduced fetal weight in
both human and animal models. The effect of particulate urban air pollution on the functional
morphology of the mouse placenta was explored by exposing second generation mice in one of four
groups to urban Sao Paulo air (PM was 67% PM2.5, mainly of vehicular origin) or filtered air (Veras
et al., 2008, 190493). Experimental design was: group F-F comprised of mice that were raised in
filtered air chambers and completed pregnancy in filtered air chambers; group F-nF raised in filtered
air and pregnant in ambient air; group nF-nF raised and completed pregnancy in non-filtered air
chambers; and group nF-F mice raised in ambient air and received filtered air during pregnancy.
Mean PM25 concentrations in the F and nF chambers were 6.5 and 27.5 (ig/m3, respectively.
Exposure was from PND20-PND60. After this exposure, the animals were mated and then
maintained in their respective chambers during pregnancy. Pregnancy was terminated at GD8 (near
term) with placentas and fetuses collected for analysis.
Exposure to ambient PM pre-gestationally or gestationally led to significantly smaller fetal
weight (total litter weight). Pregestational exposure to ambient air induced significant increases in
fetal capillary surface area and total mass-specific conductance, but this may be explained by
reduced maternal/dam blood space and diameters. Gestational exposure to non-filtered air was
associated with reduced volume, diameter (caliber) and surface area of maternal blood space with
compensatory greater fetal capillary surface and oxygen diffusion conduction rates. Intravascular
barrier thickness, a quantitative relationship between trophoblast volume and the combined surfaces
of maternal blood spaces and fetal capillaries, was not reduced with ambient air exposure. This study
provides evidence that fetal/piacental circulatory adaptation to maternal blood deficits after ambient
PM exposure may not be sufficient to overcome PM-dependent birth weight deficits in mice exposed
to ambient air, with the magnitude of this effect greater in the gestationally-exposed groups.
December 2009 7-64
-------
Placenta! Weights and Birth Outcomes
Urban Air
Pregnant female Swiss mice were exposed to ambient air (Sao Paulo) or filtered air over
various portions of gestation to determine if there was an association between fetal or placental
weight or birth outcomes with exposure to air pollution (Rocha et al., 2008, 096685). The reported
ambient concentrations of PMi0 (42 ± 17 ug/mi ), NO2 (97 ± 39 ug/m3), and SO2 (9 ± 4 ug/mi ) were
measured 100 m away from the rodent exposure chambers. By using six time windows of exposure
that covered 1-3 wk of gestation (the entire gestation period in a mouse), a significant decrease in
near-term fetal weight (GDI9) was induced by ambient air-exposure during the first week of
gestation. Decreased placental weight could be induced by ambient air exposure during any of the
3 wk of gestation. This study points to possible windows of exposure that may be important in
evaluating epidemiologic study results.
Neurodevelopmental Effects
Diesel Exhaust
The diagnosis of autism is on the rise in the Western world with its etiology mostly unknown.
Autism-associated cell loss is brain region-specific and hypothesized to be developmental in origin.
Sugamata et al. (2006, 097166) exposed pregnant ICR mice to DE (0.3 mg/m3) continuously from
2 dpc to 16 dpc. Pups with in utero exposure to DE were nursed in clean air chambers, but may have
received gastro-intestinal exposure via lactational transfer of various components of DE. At 11 wk of
age, cerebellar brain tissue was collected. Earlier work has shown that DE particles (<0.1 urn) have
been detected in the brains (cerebral cortex and hippocampus) of newborn pups who were born to
dams exposed to DE during pregnancy (Sugamata et al., 2006, 097166). Histological analysis of DE-
exposed pup cerebella revealed significant increases in caspase-3 (c-3) positive cells compared to
control and significant decreases in cerebella Purkinje cell numbers in DE-exposed animals versus
control. The ratio of cells positive for apoptosis (c-3 positive) showed a nearly significant sex
difference with males displaying increased apoptosis versus females (p = 0.09). In humans with
autism, the cerebellum has a decreased number of Purkinje cells, which is thought to be fetal and
developmental in origin; further, these authors speculate that humans may be more sensitive to DE-
dependent neuronal brain changes, as the human placenta is two-layers thick compared to the mouse
placenta that is four layers thick.
Behavioral Effects
Diesel Exhaust Particles
Body weight decrements at birth have recently been associated through the Barker hypothesis
with adverse adult outcomes. Thus, many publications have begun to focus on decreased birth
weight for gestational age and associated adult changes. Hougaard et al. (2008, 156570) exposed 40
timed-pregnant C57BL/6 dams to DE particles reference materials (SRM 2975) via inhalation over
GD7-GD19 of pregnancy. They found significantly decreased pup weight at weaning, albeit not at
birth. PM-dependent liver changes were monitored by following various inflammatory and
genotoxicity-related mRNA transcripts and there were no significant differences in pups at PND2.
The comet assay from PND2 pup livers showed no significant differences in DNA damage between
DE particle-exposed and control animals. The prohormone, thyroxine, was unchanged in control and
DE particle-exposed dams and offspring at weaning. At 2 mo, female DE particle-exposed pups
required less time than controls to locate the platform in its new location during the first trial of the
spatial reversal learning task in the Morris water maze. Thus, DE particle exposure during in utero
development led to behavioral changes without body weight at weaning or changes in inflammatory
markers or thyroid hormone levels.
December 2009 7-65
-------
Diesel Exhaust
The effect of in utero DE exposure on CNS motor function was evaluated in male pups (ICR
mice) after dams received DE exposure (8h/d><5d/wk) from GD2-GD17 (Yokota et al., 2009,
190518). The exposure atmosphere contained concentrations of 1.0 mg/m3 for particle mass,
2.67 ppm CO, 0.23 ppmNO2, and <0.01 ppm SO2. Spontaneous motor activity was significantly
decreased in pups (PND35), as was the dopamine metabolite homovanillic acid measured in the
striatum and nucleus accumbens, indicating decreased dopamine (DA) turnover. However, DA levels
were unchanged in the same areas of the brain. The authors conclude that these data demonstrate that
maternal exposure to DE induced hypolocomotion, similar to earlier studies with adult and neonatal
DE particle exposure (Peters et al., 2000, 001756). with decreased extracellular DA release.
Lactation
Diesel Exhaust
Tozuka et al. (2004, 090864) monitored the transfer of PAHs to fetuses and breast milk of
F344 rats exposed to DE (6h/day) for 2 wk from GD7-GD 20 (minus 4 days for the weekend when
no exposure occurred) with PMi0 concentration of 1.73 mg/m3. At PND 14, breast milk was
collected. Fifteen PAHs were monitored in the DE exposure chamber and seven were quantified in
dam blood with levels of phenanthrene (Phe), anthracene (Ant) and benz[a]anthracene (BaA) in the
DE group being significantly higher than the control group. In breast milk, Ant, fluoranthene (Flu),
pyrene (Pyr), and chrysene (Chr) showed significant increases in the DE group compared to the
control group. BaA tended to be about fourfold higher than the control group in breast milk, but the
increase was not significant. PAHs in dam livers of DE versus control were not significantly
different. The results of this study demonstrate that PAHs derived from DE are transferred across the
placenta from the DE-exposed dam to the fetus. Lactational transfer through the breast milk is also
likely as PAHs were detected in dam breast milk, but this should be confirmed in future studies that
cross foster control and exposed dams and pups. The lipophilicity of the PAH based on its structure
likely affected its uptake in the dam, as PAHs with 3 or 4 rings were found in maternal blood and
PAHs with 5 or 6 rings were not detected.
Heritable DMA Changes and Epigenetic Changes
Ambient Air
To address the role of ambient air exposure on heritable changes, Somers et al. (2004, 078098)
exposed mice to ambient air in at a rural Canadian site or at an urban site near a steel mill. They
showed that offspring of mice exposed to ambient air in urban regions inherited paternal-origin
expanded simple tandem repeat (ESTR) mutations 1.9-2.1 times more frequently than offspring of
mice exposed to HEPA filtered air or those exposed to rural ambient air. Mouse expanded simple
tandem repeat (ESTR) DNA is composed of short base pair repeats which are unstable in the
germline and tend to mutate by insertion or deletion of repeat units. In vivo and in situ studies have
shown that murine ESTR loci are susceptible to ionizing radiation, and other environmental
mutatgen-dependent germline mutations, and are thus good markers of exposure to environmental
contaminants.
Expanding upon the above work and to determine if PM or the gaseous phase of the urban air
was responsible for heritable mutations, Yauk et al. (2008, 157164) exposed mature male
C57BlxCBA Fl hybrid mice to either HEPA-filtered air or to ambient air in Hamilton, Ontario,
Canada for 3 or 10 wk, or 10 wk plus 6 wk of clean air exposure (16 wk). Sperm DNA was
monitored for expanded simple tandem repeat (ESTR) mutations. In addition, male-germ line
(spermatogonial stem cell) DNA methylation was monitored post-exposure. This area in Hamilton is
near two steel mills and a major highway. Air quality data provided by the Ontario Ministry of the
Environment showed the highest concentrations of TSP and metals at week 4 (93.8 ±17 and
3.6 ± 0.7 (ig/m3, respectively) and PAH at week 3 (8.3 ± 1.7 ng/m3). Mutation frequency at ESTR
December 2009 7-66
-------
Ms6-hm locus in sperm DNA from mice exposed 3 or 10 wk did not show elevated ESTR mutation
frequencies, but there was a significant increase in ESTR mutation frequency at 16 wk in ambient
air-exposed males versus HEPA filter-exposed animals, pointing to a PM-dependent mechanism of
action. When compared to HEPA filter air-exposed males, ambient air-exposed males manifested
with hypermethylation of germ-line DNA at 10 and 16 wk. These PM-dependent epigenetic
modifications (hypermethylation) were not seen in the halploid stage (3 wk) of spermatogenesis, but
were nonetheless seen in early stages of spermatogenesis (10 wk) and remained significantly
elevated in mature sperm even after removal of the mouse from the environmental exposure (16 wk).
Thus, these studies indicate that the ambient PM phase and not the gaseous phase is responsible for
the increased frequency of heritable DNA mutations and epigenetic modifications.
7.4.3. Summary and Causal Determinations
7.4.3.1. PM2.5
The 1996 PM AQCD concluded that while few studies had been conducted on the link
between PM and infant mortality, the research "suggested an association," particularly for post-
neonates (U.S. EPA, 1996, 079380). In the 2004 PM AQCD, additional evidence was available on
PM's effect on fetal and early postnatal development and mortality and while some studies indicated
a relationship between PM and pregnancy outcomes, others did not (U.S. EPA, 2004, 056905).
Studies identifying associations found that exposure to PMi0 early during pregnancy (first month of
pregnancy) or late in the pregnancy (6 wk prior to birth) were linked with higher risk of preterm
birth, including models adjusted for other pollutants, and that PM2.5 during the first month of
pregnancy was associated with IUGR. However, other work did not identify relationships between
PMio exposure and low birth weight. The state of the science at that time, as indicated in the 2004
PM AQCD, was that the research provided mixed results based on studies from multiple countries.
Building on the evidence characterized in the previous AQCDs, recent epidemiologic studies
conducted in the U.S. and Europe were able to examine the effects of PM25, and all found an
increased risk of low birth weight (Section 7.4.1). Exposure to PM2.5 was usually associated with
greater reductions in birth weight than exposure to PM^. All of the studies that examined the
relationship between PM2.5 and preterm birth report positive associations, and most were statistically
significant. The studies evaluating the association between PM2 5 and growth restriction all found
positive associations, with the strongest evidence coming when exposure was assessed during the
first or second trimester (Section 7.4.1). For infant mortality (<1 yr), several studies examined PM25
and found positive associations (Section 7.4.1).
Animal toxicological studies reported effects including decreased uterine weight, limited
evidence of male reproductive effects, and conflicting reports of reproductive outcomes in male
offspring, particularly in studies of DE (Section 7.4.2). Toxicological studies also reported effects for
several development outcomes, including immunological effects (placental and related to asthma),
neurodevelopmental and behavioral effects (Section 7.4.2).
In summary evidence is accumulating from epidemiologic studies for effects on low birth
weight and infant mortality, especially due to respiratory causes during the post-neonatal period. The
mean PM25 concentrations during the study periods ranged from 5.3-27.4 (ig/m3. Exposure to PM25
was usually associated with greater reductions in birth weight than exposure to PMi0. Several U.S.
studies of PMio investigating fetal growth reported 11-g decrements in birth weight associated with
PMio exposure. Most of these studies were conducted in California, where PM25 and PMi0_2.5
contribute almost equally to the PMio mass concentration. So while these results can not be
attributed to one size fraction or the other, the consistency of the results strengthens the interpretation
that particle exposure may be causally related to reductions in birth weight. Similarly, animal
evidence supported an association between PM2 5 and PMio exposure and adverse reproductive and
developmental outcomes, but provided little mechanistic information or biological plausibility for an
association between long-term PM exposure and adverse birth outcomes, including low birth weight,
or infant mortality. Epidemiologic studies do not consistently report associations between PM
exposure and preterm birth, growth restriction, birth defects or decreased sperm quality. New
evidence from animal toxicological studies on heritable mutations is of great interest, and warrants
further investigation. Overall, the epidemiologic and toxicological evidence is SUQQBStiVB Of 3
December 2009 7-67
-------
causal relationship between long-term exposures to PM25 and reproductive and
developmental outcomes.
7.4.3.2. PM10.2.5
Evidence is inadequate to determine if a causal relationship exists between long-
term exposure to PMi0-2.5 and developmental and reproductive outcomes because studies
have not been conducted in sufficient quantity or quality to draw any conclusion. A single study
found an association between PMi0_2.5 and birthweight (-13 g [95% CI: -18.3 to -7.6] per 10 ug/m3
increase), but no such association for PM2.5 (Parker et al., 2008, 156013).
7.4.3.3. UFPs
The 2004 PM AQCD did not report long-term exposure studies for UFPs. No epidemiologic or
animal toxicology studies have been conducted to evaluate the effects of long-term UFP exposure
and reproductive and developmental effects. Ambient air exposures, which likely include UFPs, are
reported in this ISA but there is no delineation of the separate contribution from UFPs. The evidence
is inadequate to determine if a causal relationship exists between long-term UFP
exposures and reproductive and developmental effects
7.5. Cancer, Mutagenicity, and Genotoxicity
Evidence from epidemiologic and animal toxicological studies has been accumulating for
more than three decades regarding the mutagenicity and carcinogenicity of PM in the ambient air.
DE has been identified as one source of PM in ambient air, and has been extensively studied for its
carcinogenic potential. In 1989, the International Agency for Research on Cancer (IARC) found that
there was sufficient evidence that extracts of DE particles were carcinogenic in experimental animals
and that there was limited evidence for the carcinogenic effect of DE in humans (IARC, 1989,
002958). This conclusion was based on studies in which organic extracts of DE particles were used
to evaluate the effects of concentrates of the organic compounds associated with carbonaceous soot
particles. These extracts were applied to the skin or administered by IT instillation or intrapulmonary
implantation to mice, rats, or Syrian hamsters and an excess of tumors on the skin, lung or at the site
of injection were observed.
In 2002, the U.S. EPA reviewed over 30 epidemiologic studies that investigated the potential
carcinogenicity of DE. These studies, on average, found that long-term occupational exposures to
DE were associated with a 40% increase in the relative risk of lung cancer (U.S. EPA, 2002,
042866). In the same report the U.S. EPA concluded that extensive studies with salmonella had
unequivocally demonstrated mutagenic activity in both particulate and gaseous fractions of DE.
They further concluded that DE may present a lung cancer hazard to humans (U.S. EPA, 2002,
042866). The particulate phase appeared to have the greatest contribution to the carcinogenic effect.
Both the particle core and the associated organic compounds demonstrated carcinogenic properties,
although a role for the gas-phase components of DE could not be ruled out. Almost the entire diesel
particle mass is < 10 um in diameter (PMi0), with approximately 94% of the mass of these particles
<2.5 um in diameter (PM2.5), including a subgroup with a large number of UFPs (U.S. EPA, 2002,
042866). U.S. EPA considered the weight of evidence for potential human carcinogenicity for DE to
be strong, even though inferences were involved in the overall assessment, and concluded that DE is
"likely to be carcinogenic to humans by inhalation" and that this hazard applies to environmental
exposures (U.S. EPA, 2002, 042866).
Two recent reviews of the mutagenicity (Claxton et al., 2004, 089008) and carcinogenicity
(Claxton and Woodall, 2007, 180391) of ambient air have characterized the animal toxicological
literature on ambient air pollution and cancer. The majority of these toxicological studies have been
conducted using IT instillation or dermal routes of exposure. Generally, the toxicological evidence
reviewed in this ISA has been limited to inhalation studies conducted with lower concentrations of
December 2009 7-68
-------
PM (<2 mg/m3), relevant to current ambient concentrations and the current regulatory standard
(Section 1.3). Because this ISA focuses on toxicological studies which use the inhalation route of
exposure, it is possible that important evidence for the role of PM in mutagenicity, tumorigenicity,
and/or carcinogenicity may be missed. In order to accurately characterize the relationship between
PM and cancer and be consistent with the EPA Guidelines for Carcinogen Risk Assessment
(U.S. EPA, 2005, 086237). these reviews (that include studies that employ IT instillation and dermal
routes of exposure) are summarized briefly.
Claxton et al. (2004, 089008) reviewed the mutagenicity of air in the Salmonella (Ames)
assay, and showed that hundreds of compounds identified in ambient air from varying chemical
classes are mutagenic and that the commonly monitored PAHs could not account for the majority of
mutagenicity associated with most airborne particles. They concluded that the smallest particles have
the highest toxicity per particulate mass, with the PM2.5 size fraction having greater mutagenic and
cytotoxic potential than the PM10 size fraction, which had a higher mutagenic potential than the TSP
size fraction. One study reviewed by Claxton et al. (2004, 089008) found that the cytotoxic potential
of PM2.s was higher in wintertime samples than in summertime samples. A series of studies on
source apportionment for ambient particle mutagenic activity reviewed by Claxton et al. (2004,
089008) indicate that mobile sources (cars and diesel trucks) account for most of the mutagenic
activity.
Claxton and Woodall (2007, 180391) reviewed many studies that examined the rodent
carcinogenicity of extracts of ambient PM samples; the PM was of various size classes, often from
TSP samples. Among a variety of mouse and rat strains, application methods, and samples
employed, the authors found no pattern that would suggest the routine use of a particular strain or
protocol would be more informative than another. The primary conclusion that comes from the
analysis of rodent carcinogenicity studies is that the most polluted urban air samples tested to date
are carcinogenic; the contribution of PM and different size classes of PM to the carcinogenic effects
of ambient air has not been delineated. The differences in response by the various strains of inbred
mice indicate that the genetic background of an individual can influence tumorigenic response.
Studies examining different components of ambient PM (e.g., PAHs) confirm that ambient air
contains multiple carcinogens, and that the carcinogenic potential of particles from different airsheds
can be quite different. Therefore, one would expect the incidence of cancers related to ambient air
exposure in different metropolitan areas to differ.
Numerous epidemiologic and animal toxicological studies of ambient PM and their
contributing sources have been conducted to assess the relative mutagenic or genotoxic potential.
Studies previously reviewed in the 2004 PM AQCD (U.S. EPA, 2004, 056905) provide evidence that
ambient PM as well as PM from specific combustion sources (e.g., fossil fuels) is mutagenic in vivo
and in vitro. Building on these results, data from recent epidemiologic and animal toxicological
studies that evaluated the carcinogenic, mutagenic and/or genotoxic effects of PM, PM-constituents,
and combustion emission source particles are reviewed in this section.
7.5.1. Epidemiologic Studies
The 2004 PM AQCD reported on original and follow-up analyses for three prospective cohort
studies that examined the relationship between PM and lung cancer incidence and mortality. Based
on these findings, as well as on the results from case-control and ecologic studies, the 2004 PM
AQCD concluded that long-term PM exposure may increase the risk of lung cancer incidence and
mortality. The largest of the three prospective cohort studies included in the 2004 PM AQCD was the
ACS study (Pope et al., 2002, 024689). This study was the follow-up to the original ACS study
(Pope et al., 1995, 045159). and included a longer follow-up period and reported a statistically
significant association between PM2.5 exposure and lung cancer mortality.
A 14- to 16-yr prospective study conducted using the Six Cities Study cohort reported a
slightly elevated risk of lung cancer mortality for individuals living in the most polluted city (mean
PMi0: 46.5 (ig/m3; mean PM2.5 29.6 (ig/m3) as compared to the least polluted city (mean PMi0:
18.2 (ig/m3; mean PM25 11.0 (ig/m3) but the association was not statistically significant (Dockery et
al., 1993. 0444571
Re-analysis of the AHSMOG cohort, a study of non-smoking whites living in California,
concluded that elevated long-term exposure to PMi0 was associated with lung cancer incidence
among both men and women (Beeson et al., 1998, 048890). The original study had reported an
excess of incident lung cancers only among women (Abbey et al., 1991, 042668). Further reanalysis
December 2009 7-69
-------
of this cohort revealed an association between PM10 and lung cancer mortality among men but no
association among women (Abbey et al., 1999, 047559). In addition, McDonnell et al. (2000,
010319) reported increases in lung cancer mortality with long-term exposure to PM2.5 in the
AHSMOG cohort; no association was seen for PMi0_2.5.
7.5.1.1. Lung Cancer Mortality and Incidence
The following sections will examine extensions of the above mentioned cohort studies and
new studies published since the 2004 PM AQCD. The section includes discussion of both lung
cancer incidence and mortality, as well as markers of exposure/susceptibility. A summary of the
mean PM concentrations reported for the new studies is presented in Table 7-6. In addition, a
summary of the associations for lung cancer mortality and incidence are presented in Table 7-7 and
Figure 7-7 (Section 7.6) Further discussion of all-cause and cause-specific mortality is presented in
Section 7.6.
Table 7-6. Characterization of ambient PM concentrations from recent studies of cancer and long-
term exposures to PM.
Study
Brunekreef et al. (2009, 191947)
Bonneretal. (2005, 088993)
Jerret et al. (2005, 1894051
Laden et al. (2006, 0876051
Naess et al. (2007, 0907361
Palli et al. (2008, 1568371
Pedersen et al. (2006, 1568481
Sorensen et al. (2005, 0830531
Sram et al. (2007, 1884571
Sram et al. (2007, 1920841
Vineisetal. (2006, 1920891
Vinzents et al. (2005, 0874821
Location
The Netherlands
Wfestern NY State
Los Angeles, California
6 U.S. cities
Oslo, Norway
Florence, Italy
Czech Republic
Copenhagen, Denmark
Czech Republic
Czech Republic
Multi-city, Europe
Copenhagen, Denmark
Pollutant
PM25
TSP
PM25
PM25
PM25
PM10
PM10
PM25
PM10
PM25
PM10
PM25
PM10
PM25
PM10
PM10
Mean Annual
Concentration (ug/m '
28.3
44
Range of means across sites: 10.2-29.0
Avg of means across sites: 16.4
15
19
NR
Range of means across sites: 12.6-20.7
Avg of means across sites: 16.7
Range of means across sites: 36.4-55.6
Avg of means across sites: 46.0
Range of means across sites: 24.8-44.4
Avg of means across sites: 34.6
Range of means across sites: 19.9-73.4
Avg of means across sites: 35.4
Range of means across sites: 16.9-23.5
Avg of means across sites: 20.2
Upper Percentile
Concentrations (ug/m '
Max: 36.8
Max:27.1
Max: 22
Max: 30
Max: 46-120
Max: 120-238.6
75th: 24.3-27.7
Max: 55
Max: 38
A subset of the ACS cohort study from 1982 to 2000 that included only residents of Los
Angeles, California was used to examine the association between PM2.5 and lung cancer mortality
while adjusting for both individual and neighborhood covariates (Jerrett et al., 2005, 189405). There
was a positive association between PM2.5 and lung cancer mortality when adjusting for 44 individual
covariates (RR 1.44 [95% CI: 0.98-2.11] per 10 (ig/m3 increase in PM2.5). However, including all
potential individual and neighborhood covariates associated with mortality reduced the association
December 2009
7-70
-------
(RR 1.20 [95% CI: 0.79-1.82] per 10 ug/m3 increase in PM2.5). A recent re-analysis of the full ACS
cohort also demonstrated a positive association between PM25 and lung cancer mortality (RR 1.11
[95% CI: 1.04-1.18]) (Krewski et al., 2009, 191193). The authors observed modification of this risk
by educational attainment, with those completing a high school degree or less having greater risk. In
addition to utilizing the ACS cohort for a nationwide analysis, this same study conducted two
regional assessments, one in the New York City area and the other in the Los Angeles area. No
association was detected between PM25 and lung cancer mortality in the analysis of the region
included in the New York City analysis. A positive association was observed in the Los Angeles-area
analysis using an unadjusted model, but this association did not persist after control for individual,
ecologic, and copollutant covariates.
The Six Cities Study was extended to include data from 1990-1998, a period including 1,368
deaths and 54,735 person-years (Laden et al., 2006, 087605). An elevated risk ratio for lung cancer
mortality was reported when the entire follow-up period (1974-1998) was included in the analysis
(RR 1.27 [95% CI 0.96-1.69] per 10 ug/m3 increase in average annual PM2.5). However, estimated
decreases in PM2.5 were not associated with reduced lung cancer mortality (RR 1.06 (95% CI:
0.43-2.62] for every 10 ug/m3 reduction in PM25).
Naess et al. (2007, 090736) studied individuals aged 51-90 yr living in Oslo, Norway in 1992.
Death certificate data were obtained for 1992-1998 and information on PM was collected from
1992-1995. Women had a larger association of lung cancer mortality with PM25 compared to men.
Similar results were reported for PMi0.
Most recently, Brunekreef et al. (2009, 191947) used the Netherlands cohort study (NLCS) on
diet and cancer to conduct a re-analysis of the research performed by Beelen et al. (2008, 156263)
examining the association between PM and both lung cancer mortality and incidence. After 10 yr of
follow-up, there was no association between PM25 and lung cancer mortality for either the analysis
of the full cohort (n = 105,296) (RR 1.06 [95% CI: 0.82-1.38] per 10 ug/m3 increase in PM25) or the
case-cohort (n = 4,075) (RR 0.87 [95% CI: 0.52-1.47]). There was also no association with black
smoke or traffic density variables, although living near a major roadway was associated with an
elevated relative risk for lung cancer in the full cohort analysis (RR 1.20 [95% CI: 0.98-1.47]). The
association was not present in the case-cohort analysis (RR 1.07 [95% CI: 0.70-1.64]).
In addition to lung cancer mortality, Brunekreef et al. (2009, 191947) also examined the
association with lung cancer incidence using 11.3 yr of follow-up data. In both the full cohort and
the case-cohort analyses no association was reported between PM2 5 and lung cancer incidence (full
cohort: RR0.81 [95% CI: 0.63-1.04]; case-cohort: RR0.67 [95% CI: 0.41-1.10] per 10 ug/m3
increase in PM2s). The same was true for analyses of BS and traffic density variables.
The association between PM and incident lung cancers was examined in the European
Prospective Investigation into Cancer and Nutrition study (EPIC) (Vineis et al., 2006, 192089).
Within this cohort, a nested case-control study, the GenAir study, included cases of incident cancer
and controls matched on age, gender, smoking status, country of recruitment, and time between
recruitment and diagnosis. Only non-smokers and former smokers who had quit smoking at least
10 yr prior were included. The study included 113 cases and 312 controls. No association was seen
between PMi0 and lung cancer (OR 0.91 [95% CI: 0.70-1.18] per 10 ug/m3). The OR was elevated
when cotinine, a marker for cigarette exposure, was included in the model but the authors state that
this is probably due to small study size (OR 2.85 [95% CI: 0.97-8.33] comparing > 11 ug/m3 to
<11 ug/m3). Control for other potential confounders, such as BMI, education level, and intake of
fruit and vegetables, did not have a large impact on the estimate.
December 2009 7-71
-------
Table 7-7. Associations* between ambient PM concentrations from select studies of lung cancer
mortality and incidence.
Study
Cohort
Location
Years
Analysis subgroup
Effect Estimate (95% Cl)
MORTALITY -PM^
Dockeryetal. (1993, 044457)'
Krewski et al. (2000, 012281)'
Laden et al. (2006, 087605)
Beelen et al. (2008, 156263)
Beelen et al. (2008, 156263)
Brunekreef et al. (2009, 191947)
Brunekreef et al. (2009, 191947)
Pope etal. (1995,045159)'
Pope et al. (2002, 024689)'
Jerret et al. (2005, 189405)
Krewski et al. (2009, 191193)
Krewski et al. (2009, 191193)
Krewski et al. (2009, 191193)
McDonnell et al. (2000, 010319)'
Naess et al. (2007, 090736)
Naess et al. (2007, 090736)
Naess et al. (2007, 090736)
Naess et al. (2007, 090736)
Six-Cities
Six-Cities-Re-analysis
Six-Cities
NLCS
NLCS
NLCS-Re-analysis
NLCS-Re-analysis
ACS
ACS
ACS-LA
ACS-Re-analysis
ACS-Re-analysis
ACS-Re-analysis
AHSMOG
Six cities across the U.S.
Six cities across the U.S.
Six cities across the U.S.
Netherlands
Netherlands
Netherlands
Netherlands
U.S.
U.S.
Los Angeles
U.S.
New York City
Los Angeles
California
Oslo, Norway
Oslo, Norway
Oslo, Norway
Oslo, Norway
1974-1991
1974-1991
1974-1998
1987-1996
1987-1996
1987-1996
1987-1996
1982-1989
1982-2000
1982-2000
1982-2000
1982-2000
1982-2000
1973-1977
1992-1998
1992-1998
1992-1998
1992-1998
Full Cohort
Case Cohort
Full Cohort
Case Cohort
Intra-metro Los Angeles
Intra-metro New York City
Intra-metro Los Angeles
Men
Men, 51-70yrs
Men, 71-90yrs
V\fomen, 51-70yrs
Wfomen, 71-90yrs
1.18(0.89-1.57)
1.16(0.86-1.23)
1.27(0.96-1.69)
1.06(0.82-1.38)
0.87(0.52-1.47)
1.06(0.82-1.38)
0.87(0.52-1.47)
1.01 (0.91-1.12)
1.13(1.04-1.22)
1.44(0.98-2.11)
1.11 (1.04-1.18)
0.90 (0.29-2.78)
1.31 (0.90-1.92)
1.39(0.79-2.46)
1.18(0.93-1.52)
1.18(0.93-1.52)
1.83(1.36-2.47)
1.45(1.05-2.02)
MORTALITY- PMio
McDonnell et al. (2000, 010319)'
Naess et al. (2007, 090736) '
Naess et al. (2007, 090736)
Naess et al. (2007, 090736) '
Naess et al. (2007, 090736) '
AHSMOG
California
Oslo, Norway
Oslo, Norway
Oslo, Norway
Oslo, Norway
1973-1977
1992-1998
1992-1998
1992-1998
1992-1998
Men
Men, 51-70yrs
Men, 71-90yrs
V\fomen, 51-70yrs
Wfomen, 71-90yrs
1.23(0.84-1.80)
1.12(0.95-1.33)
1.14(0.97-1.36)
1.50(1.23-1.84)
1.29(1.03-1.60)
INCIDENCE - PM2.S
Beelen et al. (2008, 155681)
Beelen et al. (2008, 155681)
Brunekreef et al. (2009, 191947)
Brunekreef et al. (2009, 191947)
NLCS
NLCS
NLCS-Re-analysis
NLCS-Re-analysis
Netherlands
Netherlands
Netherlands
Netherlands
1987-1996
1987-1996
1987-1996
1987-1996
Full Cohort
Case Cohort
Full Cohort
Case Cohort
0.81 (0.63-1.04)
0.65(0.41-1.04)
0.81 (0.63-1.04)
0.67(0.41-1.10)
INCIDENCE -PMio
Beesonetal. (1998, 048890)
Vineis etal. (2006, 192089)
AHSMOG
GenAir
California
Europe
1977-1992
1993-1999
Men
Case-Control
1.99(1.32-3.00)
0.91 (0.70-1.18)
*per 10ug/m increase
fResults from the paper were standardized to10 ug/m3 [For McDonnell et al. (2000, 010319) the non-standardized results were reported based on IQR increments (24.3 ug/m3 for PM25 and 29.5 ug/m3 for
PM10). For Naess et al. (2007, 090736) the original hazard ratios were calculated based on quartiles of PM exposure. The results were converted to10 ug/m3 using the mean range of the four quartiles (3.95
ug/m3for PM25 and 5.88 ug/m3for PM10)].
y was included in the 2004 PM AQCD
December 2009
7-72
-------
7.5.1.2. Other Cancers
Bonner et al. (2005, 088993) conducted a population-based, case-control study of the
association between ambient exposure to PAHs in early life and breast cancer incidence among
women living in Erie and Niagara counties in the state of New York. Cases (n = 1,166 of which 841
were post-menopausal) were women with primary breast cancer, and controls (n = 2,105 of which
1,495 were post-menopausal) were frequency matched to the cases by age, race, and county of
residence. TSP was used as a proxy for PAH exposure. Annual average TSP concentrations
(1959-1997) were obtained from the New York State Department of Environmental Conservation for
Erie and Niagara Counties. Among postmenopausal women, exposure to high concentrations of TSP
(>140 (ig/m3) at birth was associated with an OR of 2.42 for breast cancer (95% CI: 0.97-6.09)
relative to low concentrations of TSP (<84 (ig/m3). ORs were elevated for pollution exposures at age
of menarche (OR: 1.45 [95% CI: 0.74-2.87]) and age at first birth (OR: 1.33 [95% CI: 0.87-2.06])
among postmenopausal women. Among premenopausal women, exposure to high concentrations of
TSP at birth was associated with an OR for breast cancer incidence of 1.79 (95% CI: 0.62-5.10)
relative to low exposure levels, exposure at age of menarche was associated with an OR of 0.66
(95% CI: 0.38-1.16), and exposure at age of first birth was associated with an OR of 0.52 (95% CI:
0.22-1.20).
7.5.1.3. Markers of Exposure or Susceptibility
Several studies looked at markers of exposure or susceptibility as the outcome associated with
short-term exposure. These studies are included here because they may be relevant to the mechanism
that leads to cancer associated with long-term exposures. For example, inflammation can contribute
to carcinogenesis by inducing genomic instability, which can then lead to altered gene expression,
enhanced proliferation, and resistance to apoptotic signals. Reactive oxygen and nitrogen species,
provided by PM components or inflammation pathways, can cause molecular damage leading to
cellular transformation. Elevated inflammatory cytokines, chemokines, and prostaglandins promote
tumor growth and angiogenesis, which in turn promotes metastasis and malignant invasion. In
particular, IL-6, IL-8, IL-lp, COX-2, and TNF-a have been implicated in these processes (Kundu
and Surh, 2008, 198840). Several lines of evidence support the involvement of COX-2 in the
pathogenesis of lung cancer (Lee et al., 2008, 198811). Both short- and long-term exposure studies
demonstrate relationships between various forms of PM and increased production of these
inflammatory mediators, both in the lungs and circulation. Additionally, limited evidence suggests
that exposure to PM (Chen and Schwartz, 2008, 190106). or traffic (Williams et al., 2009, 191945).
or residence in a polluted airshed (Calderon-Garciduenas et al., 2007, 091252; Calderon-
Garciduefias et al., 2009, 192107) are associated with decreases in the number or function of natural
killer cells or other white blood cells, indicating suppression of anti-tumor defenses.
A study performed in the Czech Republic compared 53 male policemen working at least
8 hours per day outdoors in urban air with age- and sex-matched controls who spent at least 90% of
their day indoors (n = 52) (Sram et al., 2007, 188457). During the sampling period, two monitors
from downtown and suburban areas detected levels of air pollutants in the following ranges: PMi0
32-55 (ig/m3, PM2.5 27-38 (ig/m3, c-PAHs 18-22 ng/m3, and B[a]P 2.5-3.1 ng/m3 using a VAPS
monitor (measurements taken with a HiVol monitor, which has a lower flow rate, had a mean for
PMio of 62.6 (ig/m3). c-PAHs detected on personal monitors during sampling days had a mean of
12.04 ng/m3 among the policemen and 6.17 ng/m3 among the controls. No difference in percent of
chromosomal aberrations was observed between the policemen and control group using conventional
cytogenetic analysis. However, using fluorescent in situ hybridization (FISH), a difference in
chromosomal aberrations between the policemen and control group was reported. For example, the
percentage of aberrant cells, as well as the genomic frequency of translocations per 100 cells, was
about 1.4-fold greater in the policemen. This was largely driven by a difference in chromosomal
aberrations between nonsmoking policemen and nonsmoking controls. A similar study that included
only the policemen (n = 60), reported that the mean exposure to c-PAHs and B[a]P for 40-50 days
before sampling was associated with chromosomal aberrations when analyzed with FISH (Sram et
al., 2007, 192084). However, when included in a model with other covariates, the association with
these variables was null. No association was present with use of conventional cytogenetic analysis.
Palli et al. (2008, 156837) investigated the correlation between ambient PMi0 concentrations
and individual levels of DNA bulky adducts. Study participants were 214 healthy adults aged
December 2009 7-73
-------
35-64 yr at enrollment who resided in the city of Florence, Italy. This study was conducted between
1993 and 1998. PMi0 exposure levels were based on daily environmental measures provided by two
types of urban monitoring stations (high-traffic and low-traffic). The researchers assessed correlation
between DNA bulky adducts measured in blood samples and PMi0 concentrations prior to blood
sample collection. Time windows of PMi0 exposure evaluated in this study were 0-5 days, 0-10 days,
0-15 days, 0-30 days, 0-60 days, and 0-90 days prior to blood sample collection. Overall, average
PMio concentrations decreased during the study period, with some fluctuations. Quantitative values
were not reported, but PMi0 appeared to range between approximately 30 and 100 (ig/m3 for high-
traffic stations, and between approximately 20 and 50 (ig/m3for low-traffic stations. This study
found that levels of DNA bulky adducts among non-smoking workers with occupational traffic
exposure were positively correlated with cumulative PMi0 levels from high-traffic stations during
approximately 2 wk preceding blood sample collection (0-5 days: r = 0.55, p = 0.03; 0-10 days:
r = 0.58, p = 0.02; 0-15 days: r = 0.56, p = 0.02). DNA bulky adducts were not associated with PMi0
levels among Florence residents with no occupational exposure to vehicle emissions or among
smokers. DNA bulky adducts were not associated with PMi0 levels assessed by low-traffic urban
monitoring stations.
The association between personal exposure to water-soluble transition metals in PM2.5 and
oxidative stress-induced DNA damage was investigated among 49 students from Central
Copenhagen, Denmark (Sorensen et al., 2005, 083053). Researchers assessed PM2.5 exposure by
personal sampling over two weekday periods twice in one year (November 1999 and August 2000),
and determined the concentration of water-soluble transition metals (V, Cr, Fe, Ni, Cu and Pt) in
these samples. In addition, lymphocyte and 24-h urine samples were analyzed for DNA damage by
measuring 7-hydro-8-oxo-2'-deoxyguanosine (8-oxodG). Mean concentrations and corresponding
IQR of these metals differed between months of sample collection. This study found that 8-oxodG
concentration in lymphocytes was significantly associated with V and Cr concentrations, with a
1.9% increase in 8-oxodG per 1 (ig/L increase in V concentration and a 2.2% increase in 8-oxodG
per 1 (ig/L increase in Cr concentration; these associations were independent of the PM2.5 mass
concentration. The other transition metals were not significantly associated with the 8-oxodG
concentration in lymphocytes, and none of the six measured transition metals was associated with
the 8-oxodG concentration in urine.
Vinzents et al. (2005, 087482) investigated the association between UFP and PM10
concentrations with levels of purine oxidation and strand breaks in DNA using a crossover design in
Copenhagen, Denmark. Study participants were 15 healthy nonsmoking individuals with a mean age
of 25 yr. UFP exposure was evaluated using number concentration obtained in the breathing zone by
portable instruments in six 18-h weekday periods from March to June 2003. Ambient concentrations
for PMio and UFP were also measured on all exposure days at curbside street stations and at one
urban background station. Oxidative DNA damage was assessed by evaluating strand breaks and
oxidized purines in mononuclear cells isolated from venous blood the morning after exposure
measurement. Mean number concentration of UFPs (street station) was 30.4xl03 UFPs/mL (standard
deviation [SD]: 1.38), mean mass concentration of PMi0at a background monitoring station was
16.9 (ig/m (SD: 1.53), and mean mass concentration of PMio at a street station was 23.5 (ig/m3 (SD:
1.48). Mean personal exposure to UFPs was 32.4xl03 UFPs/mL (SD: 1.49) while bicycling (5
occasions), 19.6xl03 UFPs/mL (SD: 1.78) during other outdoor activities (6 occasions), and
13.4><103 UFPs/mL (SD: 1.96) while indoors (6 occasions). The regression coefficients of the mixed-
effects models looking at level of purine oxidation were estimated as 1.50><10~3 (95% CI: 0.59xlO~3
to 2.42xlO~3;/> = 0.002) for cumulative outdoor exposure and 1.07xlO~3 (95% CI: 0.37xlO~3 to
1.77xlO~3;/? = 0.003) for cumulative indoor exposure. Neither cumulative outdoor nor cumulative
indoor exposures to UFPs were associated with strand breaks. Neither ambient air concentrations of
PMio nor number concentrations of UFPs at monitoring stations were significant predictors of DNA
damage.
Additionally, a number of studies employed ecologic study designs, comparing the prevalence
of biomarkers in populations from more polluted locations to those in less polluted locations. In a
pilot study conducted in the Czech Republic (Pedersen et al., 2006, 156848). children age 5-11 yr
provided 5 mL blood samples and the frequency of micronuclei (MN) in peripheral blood
lymphocytes was analyzed as a measure of cytogenetic effects. Significantly higher frequencies of
MN were found in younger children living in Teplice (PM2.5 concentration =120 (ig/m3) than in
Prachatice (PM? 5 concentration = 46 (ig/m3). The levels of c-PAHs were also much higher in Teplice
(nearly 30 ng/m in Teplice and about 15 ng/m3 in Prachatice). The difference in MN frequencies
December 2009 7-74
-------
observed in the children from the two locations may be attributable to differences in exposure to air
pollution, but could also be due to differences in diet or other environmental exposures. This finding
is noteworthy considering MN formation in peripheral blood lymphocytes is thought to be
biologically relevant for carcinogenesis.
Avogbe et al. (2005, 087811) showed a correlation between the level of oxidative DNA
damage in individuals and exposure to ambient UFPs. Formamidopyrimidine DNA glycosylase
sensitive sites and the presence of DNA strand breaks were assessed in blood and urine samples
obtained from healthy, non-smoking male volunteers that lived and worked in different areas of
Cotonou, Benin. Exposure to benzene was assessed by urinary excretion of S-phenylmercapturic
acid. There was a high degree of correlation between exposure to benzene and UFPs and the
presence of DNA strand breaks and formamidopyrimidine DNA glycosylase sensitive sites (rural
subjects < suburban subjects < residents living near high traffic roads < taxi drivers). Genotyping
studies showed that the magnitude of the effects of benzene and UFPs may be modified by
polymorphisms in GSTP1 andNQOl genes.
Tovalin et al. (2006, 091322) evaluated the association between exposure to air pollutants and
the level of DNA damage using the single cell gel electrophoresis (comet) assay. Mononuclear
lymphocytes from outdoor and indoor workers from two areas in Mexico, Mexico City (large city)
and Puebla (medium size city), were evaluated. The outcomes showed that the outdoor workers in
Mexico City exhibited greater DNA damage than indoor workers in the same region. Similar levels
of DNA damage were observed between indoor and outdoor workers in Puebla. The level of
observed DNA damage was correlated with exposure to O3 and PM2.5.
In summary, several recent studies have reported an association between lung cancer mortality
and long-term PM2.5 exposure. Although many of the estimates include the null in the confidence
interval, overall the results have shown a positive relationship. The two recent studies that looked at
lung cancer incidence did not report an association with PM2.5 (Brunekreef et al., 2009, 191947) or
PMio (Vineis et al., 2006, 192089). Studies of exposure/susceptibility markers have reported
inconsistent outcomes, with some markers being associated with PM and others not.
7.5.2. lexicological Studies
Over the past 30 yr numerous mutagenicity and genotoxicity studies of ambient PM and their
contributing sources have been conducted to assess the relative mutagenic or genotoxic potential.
Studies previously reviewed in the 2004 PM AQCD (U.S. EPA, 2004, 056905) provide compelling
evidence that ambient PM and PM from specific combustion sources (e.g., fossil fuels) are
mutagenic in vivo and in vitro. Research cited in the 2004 AQCD demonstrated mutagenic activity
of ambient PM from urban centers in California, Germany and the Netherlands. These studies
suggested that ubiquitous emission sources, particularly motor vehicle emissions, rather than isolated
point sources were largely responsible for the mutagenic effects. In addition, the mutagenicity was
dependent upon the chemical composition of the PM with unsubstituted poly aromatic compounds
and semi-polar compounds being highly mutagenic. Mutagenicity was also dependent on size, with
the fine fraction of urban PM having greater effects than the coarse fraction. Genotoxic activity was
demonstrated for ambient PM from two high traffic areas (one upwind and one downwind) and a
rural site. In addition, the 2004 AQCD reported that exhausts from gasoline and diesel engines were
mutagenic and that DE was more potent. More mutagenicity was observed for exhaust from cold
starts than starts at room temperature. Both gaseous and particulate fractions of DE were found to be
mutagenic. Sequential fractionation of extracts from gasoline and DE implicated the polar fractions,
especially nitrated polynuclear aromatic compounds, as contributing greatly to mutagenicity. Among
some of the other mutagenically active compounds found in the gas phase of DE are ethylene,
benzene, 1,3-butadiene, acrolein and several PAHs, all of which are also present in gasoline exhaust.
Also cited in the 2004 AQCD were studies demonstrating mutagenic effects of emissions from
wood/biomass burning, which were primarily attributable to the organic fraction and not the
condensate. It was noted that wood smoke induced both frameshift mutations and base pair
substitution but not DNA adducts. Further, emissions from coal combustion in China were found to
be mutagenic, with both polar and aromatic fractions contributing to effects. Little data were
available on the mutagenicity of coal fly ash emissions from U.S. conventional combustion plants. In
conclusion, these studies provide evidence that ambient PM and combustion-derived PM are
mutagenic/genotoxic. The 2004 AQCD noted that there is not a simple relationship between
December 2009 7-75
-------
mutagenic potential and carcinogenic potential in animals or humans. No studies evaluating
carcinogenic effects of PM were reported in the 2004 AQCD.
Building on results of earlier studies in the 2004 PM AQCD, data from newly published
studies that evaluated the mutagenic, genotoxic and carcinogenic effects of PM, PM-constituents,
and combustion emission source particles are reviewed. Pertinent studies are described briefly in the
following paragraphs. A summary table is provided in Annex D, Tables D7 and D8).
7.5.2.1. Mutagenesis and Genotoxicity
In Vitro Studies
In general, studies have focused on PM and PM extracts for mutagenicity testing using
bacteria and mammalian cell lines. PM and/or PM extracts from ambient air samples, wood smoke,
and coal, diesel, or gasoline combustion have all been reported to induce mutation in S. typhimurium
and in cultured human cells (Abou et al, 2007, 098819: Gabelova et al, 2007, 156457: Gabelova et
al., 2007, 156458: Hannigan et al., 1997, 083598: Hornberg et al., 1998, 155849). In addition, effects
associated with PM and PM-associated constituents include induction of MN formation, DNA
adduct formation, SCE, DNA strand breaks, frameshifts and inhibition of gap-junction intercellular
communication (Alink et al., 1998, 087159: Arlt et al., 2007, 097257: Avogbe et al., 2005, 087811:
Gabelova et al., 2007, 156457: Gabelova et al., 2007, 156458: Healey et al., 2006, 156532: Hornberg
et al., 1996, 087164: Hornberg et al., 1998, 155849: Sevastyanova et al., 2007, 1569691
Constituents adsorbed onto individual particles play a large role in the genotoxic potential of
PM. Poma et al. (2006, 096903) showed that fine CB particles were consistently less genotoxic than
similar concentrations of PM2.5 extracts, suggesting that the adsorbed components play a role in the
genotoxic potential of PM. Total PAH and carcinogenic PAH content were correlated with the
genotoxic effects of PM (De Kok et al., 2005, 088656: Sevastyanova et al., 2007, 156969).
Comparison of different extracts (water-soluble versus organic) by Gutierrez-Castillo et al. (2006,
089030) indicated that water-soluble extracts were more genotoxic than the corresponding organic
extracts. Sharma et al. (2007, 156975) reported that mutagenic activity of extracted PM samples
collected in and around a waste incineration plant was attributed to the moderately polar and polar
fractions. The polar and crude fractions were mutagenic without metabolic activation, suggesting a
direct mutagenic effect. No mutagenic activity was observed from any of the nonpolar samples
evaluated. Arlt and colleagues (2007, 097257) have shown that the PM constituents
2-nitrobenzanthrone (2-NB) and 3-nitrobenzanthrone were genotoxic in a variety of bacterial and
mammalian cell systems.
Conflicting data have been reported on the role of metabolic enzymes in the genotoxicity of
PM and their adsorbed constituents. Arlt et al. (2007, 097257) reported that the PM constituent 2-NB
was genotoxic in bacterial and mammalian cells. However, metabolic activation with the human N-
acetyltransferase 2 or sulfotransferase (SULT1A1) enzyme was needed for the effect to be observed
in human cells. Erdinger et al. (2005, 156423) demonstrated that mutagenic activity was not affected
when metabolism was induced, de Kok et al. (2005, 088656) evaluated the relationship between the
physical, chemical, and genotoxic effects of ambient PM. TSP, PMi0, and PM2.5 were sampled at
different locations and the extracts were assessed for mutagenicity and induction of DNA adducts in
cells. Overall, induction of rat liver S9 metabolism generally reduced the mutagenic potential via the
Ames assay of the particle fractions and DNA reactivity (induction of DNA adducts) was generally
higher after metabolic activation. Binkova et al. (2003, 156274) found that the addition of S9
increased PMio-dependent DNA adduct formation.
Ambient Air
A limited number of studies evaluated the impact of the season on the genotoxic effects of
ambient PM. A few studies have indicated that greater genotoxic effects were associated with
samples collected during the winter months compared to those collected in the summer (Abou et al.,
2007, 098819: Gabelova et al., 2007, 156457: Gabelova et al., 2007, 156458). In contrast, Hannigan
et al. (1997, 083598) indicated that no seasonal variation was observed. Studies have also shown that
greater genotoxic effects were associated with smaller particle size extracts (e.g., PM2.5>PMi0) and
December 2009 7-76
-------
from samples collected in urban areas or closer to higher trafficked areas (Abou et al., 2007, 098819;
Hornberg et al., 1998, 155849).
de Kok et al. (2005, 088656) found the direct mutagenicity (Ames assay) and the direct DNA
reactivity (DNA adduct formation) of the PM2.s size fraction was significantly higher than that of the
larger size fractions (TSP, PMi0) at most locations.
DNA damage was assessed by the Comet assay in A549 cells exposed to PM collected from a
high traffic area in Copenhagen, Denmark (TSP approximately 30 ug/m3) and compared to the
results from exposure of A549 cells to standard reference materials (SRM1650 or SRM2975) at the
same concentrations (2.5-250 ug/ml) (Danielsen et al., 2008, 192092). All three particles induced
strand breaks and oxidized purines in a dose-dependent manner and there were no obvious
differences in potency. In contrast, only the ambient PM formed 8-oxodG when incubated with calf
thymus DNA, which may be due to the concentration of transition metals.
Diesel and Gasoline Exhaust
Automobile DE particles (A-DE particles) was tested in S. typhimurium strains TA98, TA100,
and its derivatives (e.g., TA98NR and YG1021) and found to be more mutagenic than forklift DE
particles (f-DE particles, derivative SRM2975), based on PM mass. A-DE particles had 227 times
more PAH-type mutagenic activity and 8-45 times more nitroarene-type mutagenic activity
(DeMarini et al., 2004, 066329). Using a diesel engine without an oxidation catalytic converter
(OCC), the diesel engine exhaust particle extract produced the highest number of revertant colonies
in strains TA98 and TA100 with and without S9 at several tested loads when compared to extracts
from low-sulfur diesel fuel (LSDF), rapeseed oil methyl ester (RME), and soybean oil methyl ester
(SME). When an OCC was installed in the exhaust pipe of the engine, all extracts reduced the
number of revertant colonies in both strains with and without S9 at partial loads but increased the
number of revertant colonies without S9 at rated power. At idling, DE particles extracts increased the
number of revertant colonies with and without S9 (Bunger et al., 2006, 156303). In a separate study,
engine emissions (particle extracts and condensates) from rapeseed (canola) oil were found to
produce greater mutagenic effects in S typhimurium strains TA98 and TA100 than DE particles
(Bunger et al., 2007, 156304). Additionally, DE extract (DEE) from diesel fuel containing various
percentages of ethanol was also observed to induce mutational response in two Salmonella strains.
Base diesel fuel DEE and DEE from fuel with 20% ethanol caused more significant DNA damage in
rat fibrocytes L-929 cells than extracts containing 5, 10, or 15% ethanol (Song et al., 2007, 155306).
DE and gasoline engine exhaust particles, as well as their semi-volatile organic compound
(SVOC) extracts, induced mutations in the two S. typhimurium strains YG1024 and YG1029 in the
absence and presence of S9; the PM extracts were more mutagenic than the SVOC extracts.
Additionally, all extracts except the DE extract induced DNA damage and MN formation in Chinese
hamster lung V79 cells (Liu et al., 2005, 097019). Another study demonstrated that gasoline engine
exhaust significantly increased colony formation in TA98 with and without S9 (Zhang et al., 2007,
157186).
Jacobsen et al. (2008, 156597) used the FEl-Muta™ Mouse lung epithelial cell line to
investigate putative mechanisms of DE particle-induced mutagenicity. Mutation ion frequencies and
ROS were determined after cells were incubated with 37.5 or 75 ug/ml DE particles (SRM1650) for
72-h (n = 8). The mutation frequency at the 75 ug/ml dose was significantly increased (1.55-fold;
p<0.001) in contrast to cells treated with 37.5 ug/ml DE particles. DE particles-induced ROS
generation 1.6- to 1.9-fold in the epithelial cell cultures after 3 h of exposure compared with the 3- to
10-fold increase in ROS production previously reported for CB. The authors concluded that the
mutagenic activity of DE particles is likely attributable to activity from the organic fraction that both
contains reactive species and can generate ROS.
In human A549 and CHO-K1 cells, the organic fraction of DE particles significantly increased
the amount of Comet and MN formation, respectively, in the presence and absence of SKF-525A (a
CYP450 inhibitor) and S9, respectively (Oh and Chung, 2006, 088296). The organic base and
neutral fractions of DE particles also significantly induced DNA damage but only without SKF-
525A, and all fractions but the moderately polar fraction (phthalates and PAH oxyderivatives)
induced MN formation with and without S9 (Bao et al., 2007, 097258). Gasoline engine exhaust
significantly induced DNA damage as measured in the Comet assay and increased the frequency of
MN in human A549 cells (Zhang et al., 2007, 157186). In human-hamster hybrid (AL) cells, DE
particles (SRM 2975) dose-dependently increased the mutation yield at the CD59 locus; this was
December 2009 7-77
-------
significantly reduced by simultaneous treatment with phagocytosis inhibitors (Bao et al., 2007,
097258).
Wood Smoke
The mutagenicity of wood smoke and cigarette smoke (CS) extracts was assayed in
S. typhimurium strains TA98 and TA100 (Ames assay) using the pre-incubation assay with
exogenous metabolic activation (rat liver S-9). Extracts of both samples (62.5 or 125 (ig total PM
equivalent/ml) were equally mutagenic to strain TA98 but the wood smoke extract was less
mutagenic than the CS extracts in strain TA100 (Iba et al., 2006, 156582).
In Vivo studies
Ambient Air
The contribution of ambient urban roadside air exposure (4, 12, 24, 48 or 60 wk) to DNA
damage was examined in the lungs, nasal mucosa, and livers of adult male Wistar rats in Kawasaki,
Japan (Sato et al., 2003, 096615). Messenger RNA levels of CYP450 enzymes that catalyze the
transformation of PAHs to reactive metabolites were also evaluated. Concentrations of gases were
reported to be 12-182 ppb NO and 0-9 ppb NO2 in the filtered air chamber and 33-280 ppb NO and
42-81 ppb NO2 in the experimental group chamber. Suspended PM concentrations were 11-19 (ig/m3
in the filtered air chamber and 42-100 (ig/m3 (average 63 (ig/m3) in the experimental group chamber.
Body weight significantly decreased in exposed animals at 24, 48 and 60 wk. A 4-wk exposure to
urban roadside air resulted in significant increases in multiple DNA adducts (lung, nasal, and liver
DNA adducts). With longer exposures, there were significant increases in lung (48 wk), nasal
(60 wk), and liver DNA adducts (60 wk). Changes were seen in CYP1A2 mRNA at 4 wk with a
2.3-fold increase in exposed animals compared to the control group with no change observed at
60 wk; CYP1A1 mRNA was unchanged. These results indicate that exposure to ambient air in this
roadside area could induce DNA adduct formation, which may be important for carcinogenicity.
Earlier studies (Ichinose et al., 1997, 053264) have shown that 8-oxodG, a DNA adduct, is elevated
along with tumor formation in a dose-dependent manner in mice administered DE particles. The
finding of adducts in the liver indicated that deposition of PM and its associated PAHs in the lung
can have indirect effects on extrapulmonary organs. It should be noted that PM deposition on the fur
and ingestion during grooming cannot be ruled out as a possible exposure route.
Another animal toxicological study employed "non-carcinogenic" particles obtained from
pooled non-cancerous lung tissue collected during surgical lung resection from three non-smoking
male patients diagnosed with lung adenocarcinomas (Tokiwa et al., 2005, 191952). Particles were
partially purified to remove organic compounds. Morphologically the particles were similar to DE or
ambient air PM and the organic extracts from the particles were directly mutagenic in S.
typhimurium tester strains TA98, YG1021 and YG1024. BALB/c and ICR mice were intratracheally
instilled with particles at doses of 0.25, 0.5, 1.0, or 2.0 mg/mouse. After 24 h, 8-oxodG was
measured in lung DNA and found to be increased in ICR mice in a dose-dependent manner, reaching
a maximum of-2.75 8-oxodG/105 dG at the 2.0 mg dose. The response was statistically significant
at doses of 0.5, 1.0, and 2.0 mg. The increased 8-oxodG levels observed in vivo was reported to be
likely due to hydroxyl radicals presumed to be involved in phagocytosis of non-mutagenic particles
by inflammatory cells that could induce hydroxylation of guanine residue on DNA.
Diesel Exhaust
An in vivo study employed gtp delta transgenic mice carrying the lambda EG 10 on each
Chromosome 17 from a C57BL/6J background to investigate the effects of DE particles on mutation
frequency (Hashimoto et al., 2007, 097261). Mice were exposed via inhalation to DE particles or via
IT instillation to DE particles or DE particle extract and lambda EG10 phages were rescued; E. coli
YG6020 was infected with the phage and screened for 6-thioguanine resistance. The mutagenic
potency (mutation frequency per mg) caused by DE particle extract was twice that of DE particles,
suggesting that the mutagenicity of DE particles is attributed primarily to compounds in the extract,
December 2009 7-78
-------
since ~50% of the weight of DE particles was provided by the extract. There was no difference in
mutation frequency between the 1 and 3 mg/m DE particle groups after 12 wk of exposure.
Wood Smoke
One recent study measured the effect of freshly generated hardwood smoke on CYP1A1
activity based on ethoxyresorufin O-deethylase in pulmonary microsomes recovered from male
Sprague-Dawley rats exposed to hardwood smoke by nose-only inhalation exposure (Iba et al, 2006,
156582). CYP1A1 activity in rat lung explants treated with extracts of the total PM (TPM) from
hardwood smoke samples and from freshly generated cigarette smoke (CS) was also evaluated.
Unlike CS, hardwood smoke did not induce pulmonary CYP1A1 activity or mRNA (assessed by
northern blot analysis) nor did extracts of hardwood smoke TPM induce CYP1A1 protein (assessed
by western blot analysis) in cultured rat lung explants. The results suggest that unique constituents
that are activated by CYP1 Al may be present in CS but not hardwood smoke.
7.5.2.2. Carcinogenesis
Studies published prior to the 2004 AQCD that evaluated the carcinogenicity of ambient air
were reviewed by Claxton and Woodall (2007, 180391). Five studies involved chronic inhalation
exposures in rodents. No statistically significant increase in tumorigenesis was observed following
chronic exposure to urban air pollution in Los Angeles (Gardner, 1966, 015129; Gardner et al., 1969,
015130; Wayne and Chambers, 1968, 038537). However in a study conducted in Brazil, urban air
pollution was found to enhance the formation of urethane-induced lung tumors in mice (Cury et al.,
2000, 192100; Reymao et al., 1997, 084653).
Two recent studies evaluated the carcinogenic potential of chronic inhalation exposures to DE
(Reed et al., 2004, 055625) and hardwood smoke (Reed et al., 2006, 156043). Two indicators of
carcinogenic potential, formation of MN and tumorigenesis were measured in strain A/J mice, which
is a mouse model that spontaneously develops lung tumors. Exposure to DE or hardwood smoke at
concentrations of 1,000 ug/m3 and below did not cause increased formation of MN or an increased
rate of lung tumors in this cancer-prone rodent model. These studies are described below.
Diesel Exhaust
A/J mice were exposed to 30, 100, 300 and 1000 ug/m3 DE for 6 h/day and 7 days/wk for
6 mo (Reed et al., 2004, 055625). The concentration of gases in this including NOX, NO2, CO, SO2,
NH3, methane, non-methane VOC, and FID total hydrocarbon ranged from control to high dose
group values of 0 to 50.4 ± 0.6 ppm, 0.2 ± 0.2 to 6.9 ± 3.3 ppm, 0.3 ± 0.1 to 30.9 ± 4.5 ppm, not
detectable to 955.2 ± 58.4 ppb, 176.5 ± 8.8 to 9.1 ± 0.2 ug/m3, 1406.5 ± 253.2 to 2642.1 ±
455.9 ug/m3, 134.0 ± 52.1 to 1578.6 ± 256.2 ug/m3, 0.1 ± 0.1 to 2.2 ± 0.2 ppm, respectively. Particle
sizes in the four exposure groups ranged from 0.10-0.15 urn MMAD with geometric standard
deviations of 1.4-1.8. Following the 6-mo exposure and a 6-mo recovery period, mice were collected
and MN formation in blood and tumor multiplicity and tumor incidence were measured in lungs. No
increases in formation of MN or numbers of lung adenomas were observed in DE-exposed mice
compared with controls.
Wood Smoke
A/J mice were exposed to 30, 100, 300 and 1,000 ug/m3 hardwood smoke or to 30, 100, 300
and 1,000 ug/m3 DE for 6 h/day and 7 days/wk for 6 mo (Reed et al., 2006, 156043). Gaseous
components of the hardwood smoke included CO, NH3, and non-methane VOC with concentrations
from control levels to high dose hardwood smoke exposure ranging from 229 ± 31 to 14887.6 ±
832.3 ppm, 139.3 ± 2.3 to 54.9 ± 1.2 ug/m3 and 177.6 ± 10.4 to 3455.0 ± 557.2 ug/m3, respectively.
Concentrations of NOX, NO2and SO2were reported to be null. Particle sizes in the four exposure
groups ranged from 0.25-0.36 urn MMAD with geometric standard deviations of 2.0-3.3. Following
the 6-mo exposure and a 6-mo recovery period, mice were collected and MN formation in blood and
tumor multiplicity and tumor incidence were measured in lungs. No increases in formation of MN or
December 2009 7-79
-------
numbers of lung adenomas were observed in hardwood smoke-exposed mice compared with
controls. However, hardwood smoke from this study was mutagenic in the Ames reverse mutation
assay.
7.5.2.3. Summary of lexicological Studies
In summary, numerous new in vitro studies confirm and extend findings reported in the 2004
AQCD that ambient PM from urban sites and combustion-derived PM are mutagenic and genotoxic.
A small number of new studies were conducted in vivo. One of these studies demonstrated increased
mutagenic potency in mice exposed to DE particles and DE particle extract. Another study found
increased formation of 8-oxodG, a DNA adduct, following IT instillation of PM in mice. A chronic
inhalation study of rats exposed to urban roadside air reported increased formation of DNA adducts
in nose, lung, and liver and induction of CYP1A2. Inhalation exposure of rats to hardwood smoke
failed to induce CYP1A1 in another study. Finally, two chronic inhalation studies found no evidence
of carcinogenic potential for DE and hardwood smoke in a cancer-prone mouse model. Collectively,
these results provide some evidence, mainly from in vitro studies, to support the biological
plausibility of ambient PM-lung cancer relationships observed in epidemiology studies.
7.5.3. Epigenetic Studies and Other Heritable DNA mutations
Two epidemiologic epigenetic studies examined the effect of PM on DNA methylation. Both
studies examined methylation of Alu and long interspersed nuclear element-1 (LINE-1) sequences,
which are located in repetitive elements. In previous studies, methylation of these sequences has
been linked to global genomic DNA methylation content (Weisenberger et al., 2005, 192101; Yang et
al., 2004, 192102).
The first study included men age 55 and older who were part of the Normative Aging Study in
the Boston area (Baccarelli et al., 2009, 192155). A stationary monitoring site located 1 km from the
examination site was used to estimate ambient PM25 exposure for the duration of the study
(1999-2007). During the study period, the median level of PM25> averaged over 7-day periods, was
9.8 (ig/m3 (interquartile range 8.0-12.0 (ig/m3). There was no association between PM2 5 and Alu
methylation. LINE-1 methylation was associated with PM2.5 measured over the 7 days before the
examinations.
The second study included 63 healthy men aged 27-55 yr working at an electric furnace steel
plant (Tarantini et al., 2009, 192010). Blood samples were taken twice, once in the morning after
2 days of not working and once in the morning after 3 full days of work. PMi0 was measured in
11 work areas and individuals completed daily logs about the amount of time spent in each area. On
average, individuals had an estimated exposure of 233.4 (ig/m3 PM10 (range 73.4-1220.2 (ig/m3).
Short-term exposure did not alter the methylation of Alu and LINE-1. To examine effects of long-
term exposure, both blood samples were considered independent of time, and Alu and LINE-1 were
examined with respect to overall estimated PMi0 exposure using mixed effects models. There was a
negative association between increasing levels of PMi0 exposure and Alu and LINE-1 methylation,
indicating that PMi0 causes epigenetic changes to occur with long-term exposure. This study also
looked at levels of iNOS gene, which is a gene suppressed by DNA methylation. iNOS expression
was not associated with long term exposure to PMi0 but was affected by methylation in the short
term.
Animal toxicology studies evaluating the effect of PM exposure on changes in the epigenome
and other non-epigenetic heritable DNA changes have only recently been conducted. After earlier
work showed increased germline mutation rates in herring gulls nesting near steel mills on Lake
Ontario (Yauk and Quinn, 1996, 089093) further work was conducted to address air-dependent
contribution to germline mutations by housing male and female Swiss Webster mice in the same area
and comparing mutation rates in those animals with mutation rates of animals housed in a rural
setting with less air pollution (Somers et al., 2002, 078100). To determine if PM or the gaseous
phase of the urban air was responsible for heritable mutations, Yauk et al. (2008, 157164) exposed
mature male C57BlxCBA Fl hybrid mice to either HEPA-filtered air or to ambient air in Hamilton,
Ontario, Canada for 3 or 10 wk, or 10 wk plus 6 wk of clean air exposure (16 wk) (also discussed in
Section 7.4.2.5). Sperm DNA was monitored for ESTR mutations, testicular sample bulky DNA
adducts, and DNA single or double strand breaks. In addition, male-germ line (spermatogonial stem
December 2009 7-80
-------
cell) DNA methylation was monitored post-exposure. This area in Hamilton is near two steel mills
and a major highway. Air composition showed mean concentrations for TSP of 93.8 ± 17 (ig/m3,
PAH of 8.3 ±1.7 ng/m3, and metal of 3.6 ± 0.7 ug/m3. Mutation frequency at ESTR Ms6-hm locus in
sperm DNA from mice exposed 3 or 10 wk did not show elevated ESTR mutation frequencies, but
there was a significant increase in ESTR mutation frequency at 16 wk compared to HEPA-filter
control animals, pointing to a PM-dependent mechanism of action. No detectable DNA adducts were
observed in testes samples at any of the time points monitored. To verify inhalation exposure to
particles, DNA adducts were reported in the lungs of mice exposed for 3 wk to ambient air; no other
time points showed detectable DNA adduct formation. Hypermethylation of germ-line DNA was
also observed in mice exposed to ambient air for 10 and 16 wk. These PM-dependent epigenetic
modifications (hypermetnylation) were not seen in the halploid stage (3 wk) of spermatogenesis, but
were nonetheless seen in early stages of spermatogenesis (10 wk) and remained significantly
elevated in mature sperm even after removal of the mouse from the environmental exposure (16 wk).
Thus, these studies indicate that the ambient PM phase and not the gaseous phase is responsible for
the increased frequency of heritable DNA mutations and epigenetic modifications.
Based on the limited evidence from these epigenetics studies, long-term exposure to PMi0 may
result in epigenetic changes. PM2.5 also potentially affects some DNA methylation content. As
epigenetic research progresses, future studies examining the relationship between PM and DNA
methylation will be important in more thoroughly characterizing these associations.
The effect of ambient PM on heritable DNA mutations and the epigenome has been well
characterized in a Canadian steel mill area. Mice exposed to ambient PM plus gases developed
paternally-derived heritable DNA mutations and epigenetic changes in sperm DNA that were not
observed in mice exposed to ambient air that was HEPA-filtered. This is the first animal toxicology
study showing heritable effects of PM exposure on DNA mutation and the epigenome. Because the
epigenetics field is so new, further work in this emerging area will expand on these PM-dependent
methylation changes to determine if the results can be recapitulated at other urban sites.
7.5.4. Summary and Causal Determinations
7.5.4.1. PM2.5
The 2004 PM AQCD reported on original and follow-up analyses for three prospective cohort
studies that reported positive relationships between PM2.5 and lung cancer mortality. Several recent,
well-conducted epidemiologic studies have extended the evidence for a positive association between
PM2.5 and lung cancer mortality (Section 7.5.1.1). Generally, studies have not reported associations
between long-term exposure to PM2.5 or PM10 and lung cancer incidence (Section 7.5.1.1). Animal
toxicological studies did not focus on specific size fractions of PM, but rather examined ambient
PM, wood smoke, and DE particles (Section 7.5.2). A number of recent studies indicate that ambient
urban PM, emissions from wood/biomass burning, emissions from coal combustion, and gasoline
and DE are mutagenic and that PAHs are genotoxic (Section 7.5.2). These findings are consistent
with earlier studies that concluded that ambient PM and PM from specific combustion sources are
mutagenic and genotoxic and provide biological plausibility for the results observed in the
epidemiologic studies. A limited number of epidemiologic and toxicological studies on the
epigenome demonstrate that PM induces changes in methylation (Section 7.5.3), anew area of
research that will likely be expanded in the future. However, it has yet to be determined how these
alterations in the genome could influence the initiation and promotion of cancer. Overall, the
evidence is suggestive of a causal relationship between relevant PM exposures and
cancer, with the strongest evidence from the epidemiologic studies of lung cancer
mortality. This evidence is limited by the non-specific measure of PM size fraction in some of the
epidemiologic studies and most of the animal toxicological studies, and the inconsistency in
evidence with recent epidemiologic studies for an effect on cancer incidence. There is no
epidemiologic evidence for cancer related to long-term exposure to PM in organs or systems other
than the lung.
December 2009 7-81
-------
7.5.4.2. PM10.2.5
The 2004 PM AQCD did not report long-term exposure studies for PMio_2.s. No epidemiologic
studies have been conducted to evaluate the effects of long-term PMi0_2.5 exposure and cancer. The
evidence is inadequate to assess the association between PWWsand UFP exposures and
cancer.
7.5.4.3. UFPs
The 2004 PM AQCD did not report long-term exposure studies for UFPs. No epidemiologic
studies have been conducted to evaluate the effects of long-term UFP and cancer. The evidence is
inadequate to determine if a causal relationship exists between long-term UFP
exposures and cancer
7.6. Mortality
In the 1996 PM AQCD, results were presented for three prospective cohort studies of adult
populations: the Six Cities Study (Dockery et al., 1993, 044457); the ACS Study (Pope et al, 1995,
045159); and the AHSMOG Study (Abbey et al., 1995, 000669). The 1996 AQCD concluded that
the chronic exposure studies, taken together, suggested associations between increases in mortality
and long-term exposure to PM25, though there was no evidence to support an association with PMi0_
2.5 (U.S. EPA, 1996, 079380).
Discussions of mortality and long-term exposure to PM in the 2004 PM AQCD emphasized
the results of four U.S. prospective cohort studies, but the greatest weight was placed on the findings
of the ACS and the Harvard Six Cities studies, which had each undergone extensive independent
reanalysis, and which were based on cohorts that were broadly representative of the U.S. population.
The 2004 PM AQCD concluded that the results from the Seventh-Day Adventist (AHSMOG) cohort
provided some suggestive (but less conclusive) evidence for associations, while results from the
Veterans Cohort provided inconsistent evidence for associations between long-term exposures to
PM2.5 and mortality. Collectively, the 2004 PM AQCD found that these studies provided strong
evidence that long-term exposure to PM2.5 was associated with increased risk of human mortality.
Effect estimates for all-cause mortality ranged from 6 to 13% increased risk per 10 (ig/m3 PM2.5,
while effect estimates for cardiopulmonary mortality ranged from 6 to 19% per 10 (ig/m3 PM2.5. For
lung cancer mortality, the effect estimate was a 13% increase per 10 (ig/m3 PM25, based upon the
results of the extended analysis from the ACS cohort (Pope et al., 2002, 024689). With regard to
PM10_2.5, the 2004 PM AQCD reported that no association was observed between mortality and long-
term exposure to PMi0_2.s in the ACS study (Pope et al., 2002, 024689). while a positive but
statistically non-significant association was reported in males in the AHSMOG cohort (McDonnell et
al., 2000, 010319). Thus, the 2004 PM AQCD concluded that there was insufficient evidence for
associations between long-term exposure to PMi0_2.s and mortality. Overall, the 2004 PM AQCD
concluded that there was strong epidemiologic evidence for associations between long-term
exposures to PM2 5 and excess all-cause and cardiopulmonary mortality.
At the time of the 2004 PM AQCD, only a limited number of the chronic-exposure cohort
studies had considered direct measurements of constituents of PM, other than sulfates. With regard
to source-oriented evaluations of mortality associations with long-term exposure, the 2004 PM
AQCD noted only the study by Hoek et al. (2002, 042364). in which the authors concluded that
long-term exposure to traffic-related air pollution may shorten life expectancy. However, Hoek et al.
(2002, 042364) also noted that living near a major road might include other factors that contribute to
mortality associations. There was not sufficient evidence at the time of the 2004 PM AQCD to draw
conclusions on effects associated with specific components or sources of PM.
New epidemiologic evidence reports a consistent association between long-term exposure to
PM25 and increased risk of mortality. There is little evidence for the long-term effects of PMi0_2.s on
mortality. Although this section focuses on mortality outcomes in response to long-term exposure to
PM, it does not evaluate studies that examine the association between PM and infant mortality.
December 2009 7-82
-------
These studies are evaluated in Section 7.5 because it is possible that in utero exposures contribute to
infant mortality. A summary of the mean PM concentrations reported for the studies characterized in
this section is presented in Table 7-8.
Table 7-8. Characterization of ambient PM concentrations from studies of mortality and long-term
exposures to PM.
Study
Location
Mean Concentration
(ug/m3)
Upper Percentile Concentrations (ug/m3)
PM2.5
Brunekreef et al. (2009, 1919471
Chen et al. (2005, 0879421
Eftim et al. (2008, 099104)
Enstrom (2005, 087356)
Cosset al. (2004, 055624)
Janes etal. (2007, 0909271
Jerrett et al. (2005, 1894051
Krewski et al. (2009, 1911931
Laden et al. (2006, 0876051
Lipfert et al. (2006, 0882181
Miller etal. (2007, 0901301
Pope et al. (2004, 0558801
Schwartz et al. (2008, 1569631
Zeger etal. (2007,157176)
Zeger etal. (2008, 191951)
The Netherlands
Multicity, CA
U.S.
CA
U.S.
U.S.
Los Angeles, CA
U.S.
Multicity, U.S.
U.S.
U.S.
U.S.
Multicity, U.S.
U.S.
U.S.
28
29.0
13.6-14.1
23.4
13.7
14.0
14.02
10.2-29.0
14.3
13.5
17.1
17.5
13.2
95th: 32
99th: 33
Max: 37
Max: 19. 1-25.1
Max: 36.1
75th: 15.9
Max: 27.1
75th: 16.00
90th: 26.75
95th: 27.89
Max: 30.01
75th: 18.3
Max: 28.3
Max: 40
17.0
75th: 14.9
PMiO-2.5
Chen et al. (2005, 0879421
Lipfert et al. (2006, 088218)
Multicity, CA
U.S.
25.4
16.0
PM10
Chen et al. (2005, 087942)
Gehring et al. (2006, 089797)
Goss etal. (2004,055624)
Puett et al. (2008, 1568911
Zanobetti et al. (2008, 1561771
Multicity, CA
North Rhine, Germany
U.S.
NEU.S.
U.S.
52.6
43.7-48.0
24.8
21.6
29.4
Max: 52.5-56.1
75th: 28.9
December 2009
7-83
-------
7.6.1. Recent Studies of Long-Term Exposure to PM and Mortality
Studies since the last PM AQCD include results of new analyses and insights for the ACS and
Harvard Six Cities studies, further analyses from the AHSMOG and Veterans study cohorts, as well
as analyses of a Cystic Fibrosis cohort and subsets of the ACS from Los Angeles and New York City.
In the original analyses of the Six Cities and ACS cohort studies, no associations were found
between long-term exposure to PMi0_2.5 and mortality, and the extended and follow-up analyses did
not evaluate associations with PMi0_2.5. The historical and more recent results for PM2.5 of both the
ACS and the Harvard Six Cities studies are compiled in Figure 7-6. Moreover, since the last PM
AQCD, there is a major new cohort that investigates the effects of PM2.5 on cardiovascular mortality
in the literature: the WHI study (Miller et al, 2007, 090130). Most recently, an ecologic cohort study
of the nation's Medicare population has been completed (Eftim et al., 2008, 099104). These new
findings further strengthen the evidence linking long-term exposure to PM2 5 and mortality, while
providing indications that the magnitude of the PM2 5-mortality association is larger than previously
estimated (Figure 7-7). Two recent reports from the AHSMOG and Veterans study cohorts have
provided some limited evidence for associations between long-term exposure to PMi0_2.5 and
mortality. The original analyses of the AHSMOG cohort study found positive associations between
long-term concentrations of PMi0 and 15-yr mortality due to natural causes and lung cancer (Abbey
et al., 1999, 047559). McDonnell et al. (2000, 010319) reanalyzed these data and concluded that
previously observed association of long-term ambient PMi0 concentrations with mortality for males
were best explained by a relationship of mortality with the fine fraction of PMi0 rather than the
thoracic coarse fraction of PMi0. Recent reports from the AHSMOG study cohort, as well as the
Nurses' Health Study and a cohort of women in Germany have provided some evidence for
associations between long-term exposure to PMi0 and mortality among women.
Harvard Six Cities: Afollow-up study has used updated air pollution and mortality data; an
additional 1,368 deaths occurred during the follow-up period (1990-1998) versus 1,364 deaths in the
original study period (1974-1989) (Laden et al., 2006, 087605). Statistically significant associations
are reported between long-term exposure to PM2 5 and mortality for data for the two periods
(RR =1.16 [95% CI: 1.07-1.26] per 10 ug/m3 PM2.5). Of special note is a statistically significant
reduction in mortality risk reported with reduced long-term PM25 concentrations (RR = 0.73 [95%
CI: 0.57-0.95] per 10 ug/m3 PM25). This is equivalent to an RR of 1.27 for reduced mortality risks
with reduced long-term PM2 5 concentrations. This reduced mortality risk was observed for deaths
due to cardiovascular and respiratory causes, but not for lung cancer deaths. The PM2 5
concentrations for recent years were estimated from visibility data, which introduces some
uncertainty in the interpretation of the results from this study. Coupled with the results of the original
analysis (Dockery et al., 1993, 044457). this study strongly suggests that a reduction in PM25
pollution yields positive health benefits.
ACS Extended AnalyseS/ReanalysiS II: Two new analyses further evaluated the
associations of long-term PM25 exposures with risk of mortality in 50 U.S. cities reported by Pope
and colleagues (2002, 024689). adding new details about deaths from specific cardiovascular and
respiratory causes (Krewski, 2009, 190075; Pope et al., 2004, 055880). Pope et al. (2004, 055880)
reported positive associations with deaths from specific cardiovascular diseases, particularly
ischemic heart disease (IHD), and a group of cardiac conditions including dysrhythmia, heart failure
and cardiac arrest (RR for cardiovascular mortality = 1.12, 95% CI 1.08-1.15 per 10 ug/m3 PM25),
but no PM associations were found with respiratory mortality.
In an additional reanalysis that extended the follow-up period for the ACS cohort to 18 yr
(1982-2000) (Krewski et al., 2009, 191193). investigators found effect estimates that were similar,
though generally higher, than those reported in previous ACS analyses. This reanalysis also included
data for seven ecologic (neighborhood-level) contextual (i.e., not individual-lev el) covariates, each
of which represents local factors known or suspected to influence mortality, such as poverty level,
educational attainment, and unemployment. The effect estimate for all cause mortality, based on
PM2.5 concentrations measured in 1999-2000 was 1.03 (95% CI: 1.01-1.05). The corresponding
effect estimates for deaths due to IHD and lung cancer were 1.15 (95% CI: 1.04-1.18) and 1.11
(95% CI: 1.04-1.18), respectively. In earlier analyses of this cohort, investigators found that
increasing education levels appeared to reduce the effect of PM2 5 exposure on mortality. Results
from this reanalysis show a similar pattern, although with somewhat less certainty, for all causes of
death except IHD, for which the pattern was reversed. Overall, although the addition of random
effects modeling and contextual covariates to the ACS model made most effect estimates higher (but
December 2009 7-84
-------
some lower), they were not statistically different from the earlier ACS effect estimates. Thus, these
new analyses, with their more extensive consideration of potentially confounding factors, confirm
the published ACS PM2.5-mortality results to be robust.
California Cancer Prevention Study: In a cohort of elderly people in 11 California
counties (mean age 73 yr in 1983), an association was reported for long-term PM25 exposure with
all-cause deaths from 1973-1982 (RR = 1.04 [95% CI: 1.01-1.07] per 10 ug/m3 PM2.5) (Enstrom,
2005, 087356). However, no significant associations were reported with deaths in later time periods
when PM2.5 levels had decreased in the most polluted counties (1983-2002) (RR = 1.00 [95% CI:
0.98-1.02] per 10 (ig/m3 PM25). The PM25 data were obtained from the EPAs Inhalation Particle
Network (collected 1979-1983), and the locations represented a subset of data used in the 50-city
ACS study (Pope et al, 1995, 045159). However, the use of average values for California counties
as exposure surrogates likely leads to significant exposure error, as many California counties are
large and quite topographically variable.
Cohort
Study
Years Mean
Effect Estimate (95% CI)
SCS Original
Reanalysis
Temporal Changes
Extended
6-Cities Medicare
ACS Original
Reanalysis
Extended
Extended
Intra-metro LA
ACS Medicare
Reanalysis II
Reanalysis II - LA
Reanalysis II - NYC
SCS Original
Reanalysis
ACS Original
Reanalysis
Extended
Extended
Intra-metro LA
Reanalysis II
Reanalysis II - LA
Reanalysis II - NYC
SCS Extended
ACS Reanalysis
Extended
ACS Extended
Intra-metro LA
Reanalysis II
Reanalysis II - LA
Reanalysis II - NYC
SCS Original
Reanalysis
Extended
ACS Original
Extended
Extended
Intra-metro LA
Reanalysis II
Reanalysis II - LA
Reanalysis II - NYC
SCS Extended
ACS Extended
Extended
Intra-metro LA
Reanalysis II
CPD=Cardio-Pulmonary Disease
CVD=Cardiovascular Disease
IHD=lschemic Heart Disease
Dockery et al. (1993, 044457) 1974-1991
Krewski et al. (2000, 012281) 1974-1991
Villeneuve et al. (2002,042576) 1974-1991
Laden et al. (2006,087605) 1974-1998
Eftim et al. (2008,099104) 2000-2002
Pope et al. (1995,045159) 1982-1989
Krewski et al.(2000,012281) 1982-1989
Pope et al. (2002,024689) 1979-1983
Pope et al. (2002,024689) 1999-2000
Jerrett et al. (2005,189405) 1982-2000
Eftim et al. (2008,099104) 2000-2002
Krewski et al. (2009,190075) 1982-2000
Krewski et al. (2009,190075) 1982-2000
Krewski et al. (2009,190075) 1982-2000
Dockery etal. (1993,044457) 1974-1991
Krewski et al. (2000,012281) 1974-1991
Pope etal. (1995,045159) 1982-1989
Krewski et al. (2000,012281) 1982-1989
Pope et al. (2002,024689) 1979-1983
Pope et al. (2002,024689) 1999-2000
Jerrett et al. (2005,189405) 1982-2000
Krewski et al. (2009,190075) 1982-2000
Krewski et al. (2009,190075) 1982-2000
Krewski et al. (2009,190075) 1982-2000
Laden et al. (2006,087605) 1974-1998
Krewski et al. (2000, 012281) 1982-1989
Pope et al. (2004,055880) 1982-2000
Pope et al. (2004,055880) 1982-2000
Jerrett et al. (2005,189405) 1982-2000
Krewski et al. (2009,190075) 1982-2000
Krewski et al. (2009,190075) 1982-2000
Krewski et al. (2009,190075) 1982-2000
Dockery et al. (1993, 044457) 1974-1991
Krewski et al. (2000, 012281) 1974-1991
Laden et al. (2006,087605) 1974-1998
Pope et al. (1995,045159) 1982-1989
Pope et al. (2002,024689) 1979-1983
Pope et al. (2002,024689) 1999-2000
Jerrett et al. (2005,189405) 1982-2000
Krewski et al. (2009,190075) 1982-2000
Krewski et al. (2009,190075) 1982-2000
Krewski et al. (2009,190075) 1982-2000
Laden et al. (2006,087605) 1974-1998
Pope et al. (2002,024689) 1979-1983
Pope et al. (2002,024689) 1999-2000
Jerrett et al. (2005,189405) 1982-2000
Krewski et al. (2009,190075) 1982-2000
18.6
18.6
18.6
17.6
14.1
18.2
18.2
21.1
14.0
19.0
13.6
14.0
20.5
12.8
18.6
18.6
18.2
18.2
21.1
14.0
19.0
14.0
20.5
12.8 <
17.6
18.2
17.1
17.1
19.0
14.0
20.5
12.8
18.6
18.6
17.6
18.2
21.1
14.0
19.0
14.0
20.5
12.8 <
17.6
21.1
14.0
19.0
14.0
All Cause
CPD
CVD
IHD
Lung Cancer
Other
1
0.5
\ \
1.0 1.5
Relative Risk Estimate
I
2.0
Figure 7-6. Mortality risk estimates associated with long-term exposure to PM2.6 from the
Harvard Six Cities Study (SCS) and the American Cancer Society Study (ACS).
December 2009
7-85
-------
Study
Cohort
Subset
Mean Effect.Estimate (95% Cl)
McDonnell et al. (2000,010319) AHSMOG
Brunekreef et al. (2009,191947) NLCS-AIR
Enstrom (2005, 087356)
Jerrett etal. (2005.189405)
Krewski et al. (2009,191193)
Laden etal. (2006.087605)
Lipfertetal. (2006,088218)
Eftimetal. (2008.099104)
Krewski etal. (2009.191193)
Gossetal. (2004.055624)
Zegeretal. (2008,191951)
CA Cancer Prevention
ACS-LA
ACS Reanalysis II-LA
Harvard 6-Cities
Veterans Cohort
Medicare Cohort
ACS Reanalysis II
U.S. Cystic Fibrosis
MCAPS
Krewski etal. (2009. 191193)
Brunekreef et al. (2009, 191947
Pope etal. (2004. 055880)
Laden etal. (2006.087605)
Naess etal. (2007,090736)
Miller etal. (2007.090130)
Chen etal. (2005.087942)
Jerrett etal. (2005.189405)
Krewski etal. (2009.190075)
Pope etal. (2004.055880)
Krewski et al. (2009,191193)
McDonnell et al. (2000,010319)
Jerrett etal. (2005.189405)
Krewski et al. (2009,191193)
Brunekreef et al. (2009,191947)
Laden etal. (2006.087605)
McDonnell et al. (2000,010319)
Brunekreef et al. (2009,191947)
Jerrett etal. (2005.189405)
Krewski et al. (2009,191193)
Laden etal. (2006.087605)
Naess etal. (2007.090736)
Krewski et al. (2009,191193)
Brunekreef et al. (2009,191947
Jerrett etal. (2005.189405)
Laden etal. (2006.087605)
Krewski etal. (2009.191193)
ACS Reanalysis II-NYC
NLCS-AIR
ACS
Harvard 6-Cities
Oslo, Norway
WHI
AHSMOG
ACS-LA
ACS Reanalysis II-LA
ACS
ACS Reanalysis II
ACS Reanalysis II-NYC
AHSMOG
ACS-LA
ACS Reanalysis II-LA
ACS Reanalysis II
ACS Reanalysis II-NYC
NLCS-AIR
Harvard 6-Cities
AHSMOG
NLCS-AIR
ACS-LA
ACS Reanalysis II-LA
Harvard 6-Cities
Oslo, Norway
ACS Reanalysis II
ACS Reanalysis II-NYC
NLCS-AIR
ACS-LA
Harvard 6-Cities
ACS Reanalysis II
Males
Full Cohort
Case Cohort
1973-1982
1983-2002
1973-2002
ACS Sites
6-Cities sites
65+, Eastern
65+, Central
65+, Western
65-74, Eastern
65-74, Central
65-74, Western
65+, Eastern
75-84, Central
75-84, Western
85+, Eastern
85+, Central
85+, Western
Full Cohort
Case Cohort
Males, 51 -70 yrs
Males, 71 -90 yrs
Females, 51 -70 yrs
Females 71 -90 yrs
Females
Females
Males
Males
Full Cohort
Case Cohort
Males
Full Cohort
Case Cohort
32.0
28.3
28.3
23.4
23.4*
23.4*
19.0
20.5
16.4
14.3
14.3
13.6
14.1
14.0
13.7
14.0
10.7
13.1
14.0
10.7
13.1
14.0
10.7
13.1
14.0
10.7
13.1
12.8
28.3
28.3
17.1
16.4
14.3
14.3
14.3
14.3
13.5
29.0
29.0
19.0
20.5
17.1
14.0
12.8
19.0
20.5
14.0
12.8«r-
28.3
28.3 -
16.4
32.0
28.3
28.3-
19.0
20.5
All Cause
Males, 51-70 yrs 14.3
Males 71-90 yrs 14.3
Females, 51-70 yrs 14.3
Females, 71-90 yrs 14.3
14.0
12.8-s-
28.3
28.3
19.0
16.4
14.0
Full Cohort
Case Cohort
4
>.
i _
I.
-i.
4.
i
CV
CHD
IHD
CPD
Respiratory
-Lung Cancer
Other
CV=Cardiovascular; CHD= Coronary Heart Disease
IHD=lschemic Heart Disease; CPD=Cardio-Pulmonary Disease
* PM2 5 data from 1973-1982 applied to all subsequent time periods
0.5
1,0 1,5 2.0
Relative Risk Estimate
2.5
Figure 7-7. Mortality risk estimates, long-term exposure to PM2.s in recent cohort studies.
December 2009
7-86
-------
AHSMOG: In this analysis for the Seventh-Day Adventist cohort in California, a positive,
statistically significant, association with coronary heart disease mortality was reported among
females (92 deaths; RR = 1.42 [95% CI: 1.06-1.90] per 10 ug/m3 PM25), but not among males
(53 deaths; RR = 0.90 [95% CI: 0.76-1.05] per 10 ug/m3 PM2.5) (Chen et al., 2005, 087942).
Associations were strongest in the subset of postmenopausal women (80 deaths; RR = 1.49 [95% CI:
1.17-1.89] per 10 ug/m3 PM2.5). The authors speculated that females may be more sensitive to air
pollution-related effects, based on differences between males and females in dosimetry and
exposure. As was found with PM2.5, a positive association with coronary heart disease mortality was
10-2.5,
reported for PM10_25 and PM10 among females (RR = 1.38 [95% CI: 0.97-1.95] per 10 ug/m3 PM
RR = 1.22 [95% CI: 1.01-1.47] per 10 ug/m3 PM10), but not for males (RR = 0.92 [95% CI: 0.66-
1.29] per 10 ug/m3 PM10.2.5; RR = 0.94 [95% CI: 0.82-1.08] per 10 ug/m3 PM10); associations were
strongest in the subset of postmenopausal women (80 deaths) (Chen et al., 2005, 087942).
U.S. Cystic FibrOSJS cohort: A positive, but not statistically significant, association was
reported for PM2.5 in this study (RR = 1.32 [95% CI: 0.91-1.93] per 10 ug/m3 PM2.5) that primarily
focused on evidence of exacerbation of respiratory symptoms (Goss et al., 2004, 055624). No clear
association was reported for PMi0. However, only 200 deaths had occurred in the cohort of over
11,000 people (average age in cohort was 18.4 yr), so the power of this study to detect associations
was relatively low.
Women's Health Initiative (WHI) Study: This nationwide cohort study considered 65,893
post-menopausal women with no history of cardiovascular disease who lived in 36 U.S. metropolitan
areas from 1994 to 1998 (Miller et al., 2007, 090130). The study had a median subject follow-up
time of 6 years. Miller and colleagues assessed each woman's exposure to air pollutants using the
monitor located nearest to their residence. Hazard ratios were estimated for the first cardiovascular
event, adjusting for age, race or ethnic group, smoking status, educational level, household income,
body-mass index, and presence or absence of diabetes, hypertension, or hypercholesterolemia.
Overall, this study concludes that "long-term exposure to fine particulate air pollution is associated
with the incidence of cardiovascular disease and death among postmenopausal women." In terms of
effect size, the study found that each increase of 10 ug/m3 of PM25 was associated with a 24%
increase in the risk of a cardiovascular event (hazard ratio, 1.24 [95% CI: 1.09-1.41]) and a 76%
increase in the risk of death from cardiovascular disease (hazard ratio, 1.76 [95% CI: 1.25-2.47]).
While this study found results confirmatory to the ACS and Six Cities Study, it reports much larger
relative risk estimates per ug/m3 PM2 5. In addition, since the study included only women without
pre-existing cardiovascular disease, it could potentially be a healthier cohort population than that
considered by the ACS and Six Cities Study. Indeed, the WHI Study reported only 216
cardiovascular deaths in 349,643 women-yr of follow-up, or a rate of 0.075% deaths per year (Miller
et al., 2007, 090130). while the ACS Study reported that 10% of subjects died of cardiovascular
disease over a 16-yr follow-up period, yielding a rate of 0.625% per year, or approximately 8 times
the cardiovascular mortality rate of the WHI population (Pope et al., 2004, 055880). Thus, PM2.5
impacts may yield higher relative risk estimates in the WHI population because the PM2 5 risk is
being compared to a much lower prevailing risk of cardiovascular death in this select study
population.
The WHI study not only confirms the ACS and Six City Study associations with mortality in
yet another well characterized cohort with detailed individual-level information, it also has been able
to consider the individual medical records of the thousands of WHI subjects over the period of the
study. This has allowed the researchers to examine not only mortality, but also related morbidity in
the form of heart problems (cardiovascular events) experienced by the subjects during the study. As
reported in this paper, this examination confirmed that there is an increased risk of cardiovascular
morbidity, as well (Section 7.2.9). These morbidity co-associations with PM2 5 in the same
population lend even greater support to the biological plausibility of the air pollution-mortality
associations found in this study.
Medicare Cohort Studies: Using Medicare data, Eftim and co-authors (2008, 099104)
assessed the association of PM25 with mortality for the same locations included in the ACS and Six
City Study. For these locations, they estimated the chronic effects of PM25 on mortality for the
period 2000-2002 using mortality data for cohorts of Medicare participants and average PM25 levels
from monitors in the same counties included in the two studies. Using aggregate counts of mortality
by county for three age groups, they estimated mortality risk associated with air pollution adjusting
for age and sex and area-level covariates (education, income level, poverty, and employment), and
controlled for potential confounding by cigarette smoking by including standardized mortality ratios
December 2009 7-87
-------
for lung cancer and COPD. This study is, therefore, an ecological analysis, similar to past published
cross-sectional analyses, in that area-level covariates (education, income level, poverty, and
employment) are employed as controlling variables, since individual level information is not
available from the Medicare database (other than age and sex), which includes virtually all
Americans aged 65 or greater. Exposures are also ecological in nature, as central site data are used as
indices of exposure. These results indicated that a 10 (ig/m3 increase in the yearly average PM25
concentration is associated with 10.9% (95% CI: 9.0-12.8) and with 20.8% (95% CI: 14.8-27.1)
increases in all-cause mortality for the ACS and Six Cities Study counties, respectively. The
estimates are somewhat higher than those reported by the original investigators, and there may be
several possible explanations for this apparent increase, especially that this is an older population
than the ACS cohort. Perhaps the most likely explanation is that the lack of personal confounder
information (e.g., past personal smoking information) led to an insufficient control for the effects of
these other variables' effects on mortality, inflating the pollution effect estimates somewhat, similar
to what has been found in the ACS analyses when only ecological-level control variables were
included. The ability of the Eftim et al. (2008, 099104) study results to qualitatively replicate the
original individual-level cohort study (e.g., ACS and Six Cities Study) results suggests that past
ecological cross-sectional mortality study results may also provide useful insights into the nature of
the association, especially when used for consideration of time trends, or for comparisons of the
relative (rather than absolute) sizes of risks between different pollutants or PM components in health
effects associations.
Janes et al. (2007, 090927) used the same nationwide Medicare mortality data to examine the
association between monthly averages of PM2.5 over the preceding 12 mo and monthly mortality
rates in 113 U.S. counties from 2000 to 2002. They decomposed the association between PM2.5 and
mortality into two components: (1) the association between "national trends" in PM2.5 and mortality;
and (2) the association between "local trends," defined as county-specific deviations from national
trends. This second component is posited to provide evidence as to whether counties having steeper
declines in PM2 5 also have steeper declines in mortality relative to the national trend. They report
that the exposure effect estimates are different at these two spatiotemporal scales, raising concerns
about confounding bias in these analyses. The authors assert that the association between trends in
PM2 5 and mortality at the national scale is more likely to be confounded than is the association
between trends in PM25 and mortality at the local scale and, if the association at the national scale is
set aside, that there is little evidence of an association between 12-month exposure to PM25 and
mortality in this analysis. However, in response, Pope and Burnett (2007, 090928) point out that
such use of long-term time trends as the primary source of exposure variability has been avoided in
most other air pollution epidemiology studies because of such concerns about potential confounding
of such time-trend associations.
By linking monitoring data to the U.S. Medicare system by county of residence, Zeger et al.
(2007, 157176) analyzed Medicare mortality records, comprising over 20 million enrollees in the
250 largest counties during 2000-2002. The authors estimated log-linear regression models having
age-specific county level mortality rates as the outcome and, as the main predictor, the average PM2 5
pollution level in each county during 2000. Area-level covariates were used to adjust for socio-
economic status and smoking. The authors reported results under several degrees of adjustment for
spatial confounding and with stratification into eastern, central and western U.S. counties. A
10 ug/m3 increase in PM25 was associated with a 7.6% increase in mortality (95% CI: 4.4-10.8).
When adjusted for spatial confounding, the estimated log-relative risks dropped by 50%. Zeger et al.
(2007, 157176) found a stronger association in the eastern counties than nationally, with no evidence
of an association in western counties.
In a subsequent report, Zeger et al. (2008, 191951) created a new retrospective cohort, the
Medicare Cohort Air Pollution Study (MCAPS), consisting of 13.2 million persons residing in 4,568
ZIP codes in urban areas having geographic centroids within 6 miles of a PM2 5 monitor. Using this
cohort, they investigated the relationship between 6-yr avg exposure to PM2 5 and mortality risk over
the period 2000-2005. When divided by region, the associations between long-term exposure to
PM2 5 and mortality for the eastern and central ZIP codes were qualitatively similar to those reported
in the ACS and Six Cities Study, with 11.4% (95% CI: 8.8-14.1) and 20.4% (95% CI: 15.0-25.8)
increases per 10 (ig/m3 increase in PM25 in the eastern and central regions, respectively. The MCAPS
results included evidence of differing PM2 5 relative risks by age and geographic location, where risk
declines with increasing age category until there is no evidence of an association among persons
December 2009 7-88
-------
> 85 yr of age, and there is no evidence of a positive association for the 640 urban ZIP codes in the
western region of the U.S.
Using hospital discharge data, Zanobetti et al. (2008, 156177) constructed a cohort of persons
discharged with COPD using Medicare data between 1985 and 1999. Positive associations in the
survival analyses were reported for single year and multiple-year lag exposures, with a hazard ratio
for total mortality of 1.22 (95% CI: 1.17-1.27) per 10 ug/m3 increase in PMi0 over the previous
4 years.
Veterans Cohort: A recent reanalysis of the Veterans cohort data focused on exposure to
traffic-related air pollution (traffic density based on traffic flow rate data and road segment length)
reported a stronger relationship between mortality with long-term exposure to traffic than with PM2 5
mass (Lipfert et al., 2006, 088218). A significant association was reported between total mortality
and PM2.5 in single-pollutant models (RR =1.12 [95% CI: 1.04-1.20] per 10 ug/m3 PM2.5). This risk
estimate is larger than results reported in a previous study of this cohort. In multipollutant models
including traffic density, the association with PM2.5 was reduced and lost statistical significance.
Traffic emissions contribute to PM2 5 so it would be expected that the two would be highly
correlated, and, thus, these multipollutant model results should be interpreted with caution. In a
companion study, Lipfert et al. (2006, 088218) used data from EPA's fine particle speciation
network, and reported findings for PM2 5 which were similar to those reported by Lipfert et al. (2006,
088218). In this study (Lipfert et al., 2006, 088218). a significant association was reported between
long-term exposure to PMi0_2.5 and total mortality in a single-pollutant model (RR = 1.07, 95% CI:
1.01-1.12 per 10 ug/m3 PM10_2.5). However, the association became negative and not statistically
significant in a model that included traffic density. As it would be expected that traffic would
contribute to the PMi0_2.5 concentrations, it is difficult to interpret the results of these multipollutant
analyses.
Nurses' Health Study Cohort: The Nurses' Health Study (Puett et al., 2008, 156891) is an
ongoing prospective cohort study examining the relation of chronic PMi0 exposures with all-cause
mortality and incident and fatal CHD consisting of 66,250 female nurses in MSAs in the
northeastern region of the U.S. All cause mortality was statistically significantly associated with
average PMi0 exposures in the time period 3-48 mo preceding death. The association was strongest
with average PMi0 exposure in the 24 mo prior to death (hazard ratio 1.16 [95% CI: 1.05-1.28]) and
weakest with exposure in the month prior to death (hazard ratio 1.04 [95% CI: 0.98-1.11]). The
association with fatal CHD occurred with the greatest magnitude with mean exposure in the 24 mo
prior to death (hazard ratio 1.42 [95% CI: 1.11-1.81]).
Netherlands Cohort Study (NLCS): The Netherlands Cohort Study (Brunekreef et al.,
2009, 191947) estimates the effects of traffic-related air pollution on cause specific mortality in a
cohort of approximately 120,000 subjects aged 55-69 yr at enrollment. For a 10 ug/m3 increase in
PM25 concentration, the relative risk for natural-cause mortality in the full cohort was 1.06 (95% CI:
0.97-1.16), similar in magnitude to the results reported by the ACS. In a case-cohort analysis
adjusted for additional potential confounders, there were no associations between air pollution and
mortality.
German Cohort: The North Rhine-Westphalia State Environment Agency (LUANRW)
initiated a cohort of approximately 4,800 women, and assessed whether long-term exposure to air
pollution originating from motorized traffic and industrial sources was associated with total and
cause-specific mortality (Gehring et al., 2006, 089797). They found that cardiopulmonary mortality
was associated with PM10 (RR = 1.52 [95% CI: 1.09-2.15] per 10 ug/m3 PM10).
7.6.2. Composition and Source-Oriented Analyses of PM
As discussed in the 2004 PM AQCD, only a very limited number of the chronic exposure
cohort studies have included direct measurements of chemical-specific PM constituents other than
sulfates, or assessments of source-oriented effects, in their analyses. One exception is the Veterans
Cohort Study, which looked at associations with some constituents, and traffic.
Veterans Cohort: Using data from EPA's fine particle speciation network, Lipfert et al.
(2006, 088756) reported a positive association for mortality with sulfates. Using 2002 data from the
fine particle speciation network, positive associations were found between mortality and long-term
exposures to nitrates, EC, Ni and V, as well as traffic density and peak O3 concentrations. In
December 2009 7-89
-------
multipollutant models, associations with traffic density remained significant, as did nitrates, Ni and
V in some models.
Netherlands Cohort Study: Beelen et al. (2008, 156263) studied the association between
long-term exposure to traffic-related air pollution and mortality in a Dutch cohort. They used data
from an ongoing cohort study on diet and cancer with 120,852 subjects who were followed from
1987 to 1996. Exposure to BS, NO2, SO2, and PM2.5, as well as various exposure variables related to
traffic, were estimated at the home address. Traffic intensity on the nearest road was independently
associated with mortality. Relative risks (CI) for a 10 (ig/m increase in BS concentrations
(difference between 5th and 95th percentile) were 1.05 (95% CI: 1.00-1.11) for natural cause, 1.04
(95% CI: 0.95-1.13) for cardiovascular, 1.22 (95% CI: 0.99-1.50) for respiratory, 1.03 (95% CI:
0.88-1.20) for lung cancer, and 1.04 (95% CI: 0.97-1.12) for mortality other than cardiovascular,
respiratory, or lung cancer. Results were similar for NO2 and PM2 5, but no associations were found
for SO2. Traffic-related air pollution and several traffic exposure variables were associated with
mortality in the full cohort, although the relative risks were generally small. Associations between
natural-cause and respiratory mortality were statistically significant for NO2 and BS. These results
add to the evidence that long-term exposure to traffic-related particulate air pollution is associated
with increased mortality.
Given the general dearth of published source-oriented studies of the mortality impacts of long-
term PM exposure components, and given that the recent Medicare Cohort study now indicates that
such ecological cross-sectional studies can be useful for evaluating time trends and/or comparisons
across pollution components, it may well be that examining past cross-sectional studies comparing
source-oriented components of PM may be informative. In particular, Ozkaynak and Thurston (1987,
072960). utilized the chemical speciation conducted in the Inhalable Particle (IP) Network to
conduct a chemical constituent and source-oriented evaluation on long-term PM exposure and
mortality in the U.S. They analyzed the 1980 U.S. vital statistics and available ambient air pollution
data bases for sulfates and fine, inhalable, and TSP mass. Using multiple regression analyses, they
conducted a cross-sectional analysis of the association between various particle measures and total
mortality. Results from the various analyses indicated the importance of considering particle size,
composition, and source information in modeling of particle pollution health effects. Of the
independent mortality predictors considered, particle exposure measures most related to the
respirable fraction of the aerosols, such as fine particles and sulfates, were most consistently and
significantly associated with the reported SMSA-specific total annual mortality rates. On the other
hand, particle mass measures that included PMi0_2.5 (e.g., total suspended particles and inhalable
particles) were often found to be non-significant predictors of total mortality. Furthermore, based on
the application of PM25 source apportionment, particles from industrial sources and from coal
combustion were indicated to be more significant contributors to human mortality than fine soil-
derived particles.
7.6.3. Within-City Effects of PM Exposure
Much of the exposure gradient in the national-scale cohort studies was due to city-to-city
differences in regional air pollution, raising the possibility that some or all of the original PM-
survival associations may have been driven instead by cify-to-city differences in some unknown
(non-pollution) confounder variable. This has been evaluated by three recent studies.
ACS, LOS Angeles: To investigate this issue, two new analyses using ACS data focused on
neighborhood-to-neighborhood differences in urban air pollutants, using data from 23 PM2 5
monitoring stations in the Los Angeles area, and applying interpolation methods (Jerrett et al., 2005,
189405) or land use regression methods (Krewski et al., 2009, 191193) to assign exposure levels to
study individuals. This resulted in both improved exposure assessment and an increased focus on
local sources of PM25. Significant associations between PM25 and mortality from all causes and
cardiopulmonary diseases were reported with the magnitude of the relative risks being greater than
those reported in previous assessments. In general, the associations for PM2 5 and mortality using
these two methods for exposure assessment were similar, though the use of land use regression
resulted in somewhat smaller hazard ratios and tighter CIs (see Table 7-9). This indicates that city-to-
city confounding was not the cause of the associations found in the earlier ACS Cohort studies. This
provides evidence that reducing exposure error can result in stronger associations between PM2 5 and
mortality than generally observed in broader studies having less exposure detail.
December 2009 7-90
-------
Table 7-9. Comparison of results from ACS intra-urban analysis of Los Angeles and New York City
using kriging or land use regression to estimate exposure.
Cause of
Death
All Cause
IHD
CPD
Lung Cancer
Los Angeles:
Hazard Ratio1 and 95% Confidence
Interval Using Kriging
(Jerrett et al, 2005, 189405)
1.11 (0.99-1.25)
1.25(0.99-1.59)
1.07(0.91-1.26)
1.20(079-1.82)
Los Angeles:
Hazard Ratio1 and 95% Confidence
Interval Using Land Use Regression
(Krewskietal.. 2009. 191193)
1.13(1.01-1.25)
1.26(1.02-1.56)
1.09(0.94-1.26)
1.31 (0.90-1.92)
New York City:
Hazard Ratio1 and 95% Confidence
Interval Using Land Use Regression
(Krewskietal.. 2009. 191193)
0.86(0.63-1.18)
1.56(0.87-2.88)
0.66(0.41-1.08)
0.90 (0.29-2.78)
1Hazard ratios presented per 10 [jg/m3 increase in PM25
2Model included parsimonious contextual covariates
3Model included parsimonious individual level (23) and ecologic (4) covariates
^Model included all 44 individual level and 7 ecologic covariates.
ACS, New York: Krewski et al. (2009, 191193) applied the same techniques used in the land
use regression analysis of Los Angeles to an investigation conducted in New York City. Annual
average concentrations were calculated for each of 62 monitors from 3 yr of daily monitoring data
for 1999-2001. Those data were combined with land-use data collected from traffic counting
systems, roadway network maps, satellite photos of the study area, and local government planning
and tax-assessment maps to assign estimated exposures to the ACS participants. The investigators
did not observe elevated effect estimates for all cause, CPD or lung cancer deaths, but IHD did show
a positive association with PM2 5 concentration. The difference between the 90th and 10th percentiles
of the 3-yr avg PM2.5 concentration was 1.5 (ig/m3 and the difference between the minimum and
maximum values of the 3-yr avg PM2.5 concentration was 7.8 (ig/m3. This narrow range in PM2.5
exposure contrasts across the New York City metropolitan area and may well account for the
inconclusive results in this city-specific analysis. Relatively uniform exposures would reduce the
power of the statistical models to detect patterns of mortality relative to exposure and estimate the
association with precision.
WHI Study: This study also investigated the within- versus between-city effects in its cities.
As shown in Figure 7-8, similar effects for both the within and between-city analyses demonstrate
that this association is not due to some other (non-pollution) confounder differing between the
various cities, strengthening confidence in the overall pollution-effect estimates.
December 2009
7-91
-------
A Overall Effect
12-
11-
0 3 6 9 12 15 IS 21 24 11 30
B Between-City Effect
12-
C Within-City Effect
0 3 6 9 12 15 18 21 24 27 30
3 6 9 12 15 IS 21 24 27 30
PM |im)
Source: Miller et al. (2007, 090130)
Copyright © 2007 Massachusetts Medical Society. All rights reserved.
Figure 7-8. Plots of the relative risk of death from cardiovascular disease from the Women's
Health Initiative study displaying the between-city and within-city contributions
to the overall association between PM2.s and cardiovascular mortality windows of
exposure-effects.
7.6.4. Effects of Different Long-term Exposure Windows
The delay between changes in exposure and changes in health has important policy
implications. Schwartz et al. (2008, 156963) investigated this issue using an extended follow-up of
the Harvard Six Cities Study. Cox proportional hazards models were fit to control for smoking, body
mass index, and other covariates. Penalized splines were fit in a flexible functional form to the
concentration response to examine its shape, and the degrees of freedom for the curve were selected
based on Akaike's information criterion (AIC). The researchers also used model averaging as an
alternative approach, where multiple models are fit explicitly and averaged, weighted by their
probability of being correct given the data. The lag relationship by model was averaged across a
range of unconstrained distributed lag models (i.e., same year, 1 yr prior, 2 yr prior, etc.). Results of
the lag comparison are shown in Figure 7-9 indicating that the effects of changes in exposure on
mortality are seen within 2 yr. The authors also noted that the concentration-response curve was
linear, clearly continuing below the level of the current U.S. air quality standard of 15 (ig/m3.
December 2009
7-92
-------
1.20
1.15
1.10
JS
-------
LO
O -
o
O
Cfl
"l^
CD
_>
Is
CD
01
to
O5
CD
2000
Figure 7-10.
2002 2004 2006 2008
Time (years)
Source: Reprinted with Permission from Oxford University Press & the International Epidemiological Society from Roosli et al. (2005, 1 56923)
Time course of relative risk of death after a sudden decrease in air pollution
exposure during the year 2000, assuming a steady state model (solid line) and a
dynamic model (bold dashed line). The thin dashed line refers to the reference
scenario.
Table 7-10. Distribution of the effect of a hypothetical reduction of 10 ug/m PMio in 2000 on all-
cause mortality 2000-2009 in Switzerland.
Year
1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
Proportion of total effect (%)
39.3 23.9
14.5
5.3
3.2
2.0
1.2
0.7
0.4
Relative risk (per 10 pg/m reduction in PM10)
1.0
0.9775 0.9863 0.9917 0.9950 0.9969 0.9981 0.9989 0.9993 0.9996 0.9997
Relative risk and proportion of total effect in each year are shown, assuming a time constant k of 0.5
Source: Roosli et al. (2005,156923)
In the reanalysis of the ACS cohort, the investigators calculated time windows of exposure as
average concentrations during successive 5-yr periods preceding the date of death (Krewski et al.,
2009, 191193). The investigators considered the time window with the best-fitting model (judged by
the AIC statistic) to be the period during which pollution had the strongest influence on mortality.
Overall, the differences between the time periods were small and demonstrated no definitive
patterns. High correlations between exposure levels in the three periods may have reduced the ability
of this analysis to detect any differences in the relative importance of the time windows. The
investigators did not analyze any time periods smaller than 5 yr, so the results are not directly
comparable to those reported by Schwartz et al. (2008, 156963). Roosli et al. (2005, 156923). and
Puett et al. (2008, 156891).
December 2009
7-94
-------
Generally, these results indicate a developing coherence of the air pollution mortality
literature, suggesting that the health benefits from reducing air pollution do not require a long
latency period and would be expected within a few years of intervention.
7.6.5. Summary and Causal Determinations
7.6.5.1. PM2.5
In the 1996 PM AQCD (U.S. EPA, 1996, 079380). results were presented for three prospective
cohort studies of adult populations: the Six Cities Study (Dockery et al, 1993, 044457): the ACS
Study (Pope et al., 1995, 045159): and the AHSMOG Study (Abbey et al., 1995, 000669). The 1996
AQCD concluded that the chronic exposure studies, taken together, suggested associations between
increases in mortality and long-term exposure to PM2 5, though there was no evidence to support an
association with PMi0_2.5 (U.S. EPA, 1996, 079380). Discussions of mortality and long-term
exposure to PM in the 2004 PM AQCD emphasized the results of four U.S. prospective cohort
studies, but the greatest weight was placed on the findings of the ACS and the Harvard Six Cities
studies, which had undergone extensive independent reanalysis, and which were based on cohorts
that were broadly representative of the U.S. population. Collectively, the 2004 PM AQCD found that
these studies provided strong evidence that long-term exposure to PM2 5 was associated with
increased risk of human mortality.
The recent evidence is largely consistent with past studies, further supporting the evidence of
associations between long-term PM2.5 exposure and increased risk of human mortality (Section 7.6)
in areas with mean concentrations from 13.2 to 29 ug/m3 (Figure 7-7). New evidence from the Six
Cities cohort study shows a relatively large risk estimate for reduced mortality risk with decreases in
PM2.5 (Laden et al., 2006, 087605). The results of new analyses from the Six Cities cohort and the
ACS study in Los Angeles suggest that previous and current studies may have underestimated the
magnitude of the association (Jerrett et al., 2005, 189405). With regard to mortality by cause-of-
death, recent ACS analyses indicate that cardiovascular mortality primarily accounts for the total
mortality association with PM2.5 among adults, and not respiratory mortality. The recent WHI cohort
study shows even higher cardiovascular risks per ug/m3 than found in the ACS study, but this is
likely due to the fact that the study included only post-menopausal women without pre-existing
cardiovascular disease (Miller et al., 2007, 090130). There is additional evidence for an association
between PM25 exposure and lung cancer mortality (Section 7.5.1.1). The WHI study also considered
within versus between city mortality, as well as morbidity co-associations with PM2 5 in the same
population. The first showed that the results are not due to between city confounding, and the
morbidity analyses show the coherence of the mortality association across health endpoints,
supporting the biological plausibility of the air pollution-mortality associations found in these
studies.
Results from a new study examining the relationship between life expectancy and PM2 5 and
the findings from a multiyear expert judgment study that comprehensively characterizes the size and
uncertainty in estimates of mortality reductions associated with decreases in PM25 in the U.S draw
conclusions that are consistent with an association between long-term exposure to PM2 5 and
mortality (Pope et al., 2009, 190107: Roman et al., 2008, 156921). Pope et al. (2009, 190107) report
that a decrease of 10 ug/m3 in the concentration of PM25 is associated with an estimated increase in
mean (± SE) life expectancy of 0.61 ± 0.20 year. For the approximate period of 1980-2000, the
average increase in life expectancy was 2.72 yr among the 211 counties in the analysis. The authors
note that reduced air pollution was only one factor contributing to increased life expectancies, with
its effects overlapping with those of other factors.
Roman et al. (2008, 156921) applied state-of-the-art expert judgment elicitation techniques to
develop probabilistic uncertainty distributions that reflect the broader array of uncertainties in the
concentration-response relationship. This study followed best standard practices for expert
elicitations based on the body of literature accumulated over the past two decades. The resulting
PM25 effect estimate distributions, elicited from 12 of the world's leading experts on this issue, are
shown in Figure 7-11. They indicate both larger central estimates of mortality reductions for
decreases in long-term PM25 exposure in the U.S. (averaging almost 1% per ug/m3 PM25) than
reported (for example) by the ACS Study (i.e., 0.6% per ug/m3 PM2.5 in Pope et al. (2002, 024689).
December 2009 7-95
-------
and a wider distribution of uncertainty by each expert than provided by any one of the PM2.5
epidemiologic studies. However, a composite uncertainty range of the overall mean effect estimate
(i.e., based upon all 12 experts' estimates, but not provided in Figure 7-11) would be much narrower,
and closer to that derived from the ACS study than indicated for any one expert shown in Figure
7-11.
Group 1
Group 2
a.
•C
o
1
10-30
Causality Likelihood 99% 75% 99%
Expert E L
4-10 >10-30 4-30 4-30 4-30 4-16 >18-30
98% 98% 95% 95% 70% 35% 35%
B DIG K
4-7 >7-30 4-30 4-30 4-30 4-30 pope Dockery
100% 100% 99% 99% 95% 90% B| a| et al..
F C J A H 2002 1993
Key: Closed circle = median; Open circle = mean: Box = interquartile range: Solid line = 90% credible interval
Source: Reprinted with Permission of ACS from Roman et al. (2008,1569211
Figure 7-11. Experts' mean effect estimates and uncertainty distributions for the PM2.s
mortality concentration-response coefficient for a 1 ug/m3 change in annual
average PM2.s.
Overall, recent evidence supports the strong evidence reported in the 2004 PM AQCD
(U.S. EPA, 2004, 056905) that long-term exposure to PM2.5 is associated with an increased risk of
human mortality. When looking at the cause of death, the strongest evidence comes from mortality
due to cardiovascular disease, with additional evidence supporting an association between PM2 5 and
lung cancer mortality (Figure 7-7). Fewer studies evaluate the respiratory component of
cardiopulmonary mortality, and the evidence to support an association with long-term exposure to
PM2.5 and respiratory mortality is weak (Figure 7-7). Together these findings are consistent and
coherent with the evidence from epidemiologic, controlled human exposure, and animal
toxicological studies for the effects of short- and long-term exposure to PM on cardiovascular effects
presented in Sections 6.2 and 7.2, respectively. Evidence of short- and long-term exposure to PM2.5
and respiratory effects (Sections 6.3 and 7.3, respectively) and infant mortality (Section 7.4) are
coherent with the weak respiratory mortality effects. Additionally, the evidence for short- and long-
term cardiovascular and respiratory morbidity provides biological plausibility for mortality due to
cardiovascular or respiratory disease. The most recent evidence for the association between long-
term exposure to PM2 5 and mortality is particularly strong for women. Collectively, the evidence is
sufficient to conclude that the relationship between long-term PM25 exposures and
mortality is causal.
December 2009
7-96
-------
7.6.5.2. PM10.2.5
In the 2004 PM AQCD, results from the ACS and Six Cities study analyses indicated that
PMio_2.5 was not associated with mortality. Evidence is still limited to adequately characterize the
association between PM10_2.5 and PM sources and/or components. The new findings from AHSMOG
and Veterans cohort studies provide limited evidence of associations between long-term exposure to
PMio_2.5 and mortality in areas with mean concentrations from 16 to 25 (ig/m3. The evidence for
PMio25 is inadequate to determine if a causal relationship exists between long-term
exposures and mortality
7.6.5.3. UFPs
The 2004 PM AQCD did not report long-term exposure studies for UFPs. No epidemiologic
studies have been conducted to evaluate the effects of long-term UFP exposure and mortality. The
evidence is inadequate to determine if a causal relationship exists between long-term UFP
exposures and mortality
December 2009 7-97
-------
Chapter 7 References
Abbey DE; Lebowitz MD; Mills PK; Petersen FF; Beeson WL; Burchette RJ. (1995). Long-term ambient concentrations of
particulates and oxidants and development of chronic disease in a cohort of nonsmoking California residents.
Presented at In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on particulate air pollution and
human mortality and morbidity; January 1994; Irvine, CA. Inhalation Toxicol. 7: 19-34., Irvine, CA. 000669
Abbey DE; Mills PK; Petersen FF; Beeson WL. (1991). Long-term ambient concentrations of total suspended particulates
and oxidants as related to incidence of chronic disease in California Seventh-Day Adventists. Environ Health
Perspect, 94: 43-50. 042668
Abbey DE; Nishino N; McDonnell WF; Burchette RJ; Knutsen SF; Beeson WL; Yang JX. (1999). Long-term inhalable
particles and other air pollutants related to mortality in nonsmokers. Am J Respir Crit Care Med, 159: 373-382.
047559
Abou Chakra OR; Joyeux M; Nerriere E; Strub MP; Zmirou-Navier D. (2007). Genotoxicity of organic extracts of urban
airborne particulate matter: an assessment within a personal exposure study. Chemosphere, 66: 1375-81. 098819
Achenbach S; Daniel WG. (2001). Noninvasive coronary angiography—an acceptable alternative?. N Engl J Med, 345:
1909-1910. 156189
Ackermann-Liebrich U; Kuna-Dibbert B; Probst-Hensch NM; Schindler C; Dietrich DF; Stutz EZ; Bayer-Oglesby L;
Baum F; Brandli O; Brutsche M; Downs SH; Keidel D; Gerbase MW; Imboden M; Keller R; Knopfli B; Kunzli
Nellweger J-P; Leuenberger P; SALPADIA Team. (2005). Follow-up of the Swiss Cohort study on air pollution and
lung diseases in adults (SAPALDIA 2) 1991-2003: methods and characterization of participants. Int J Public Health,
50: 245-263. 087826
Ackermann-Liebrich U; Leuenberger P; Schwartz J; Schindler C; Monn C; Bolognini B; Bongard JP; Brandli O;
Domenighetti G; Elsasser S; Grize L; Karrer W; Keller R; Keller-Wossidlo H; Kunzli N; Martin BW; Medici TC;
Pliger B; Wuthrich B; Zellweger JP; Zemp E. (1997). Lung function and long term exposure to air pollutants in
Switzerland. Am J Respir Crit Care Med, 155: 122-129. 077537
Agatston AS; Janowitz WR; Hildner FJ; Zusmer NR; Viamonte M Jr; Detrano R. (1990). Quantification of coronary artery
calcium using ultrafast computed tomography. J Am Coll Cardiol, 15: 827-832. 156197
Alink GM; Sjogren M; Bos RP; Doekes G; Kromhout H; Scheepers PTJ. (1998). Effect of airborne particles from selected
indoor and outdoor environments on gap-junctional intercellular communication. Toxicol Lett, 96/97: 209-213.
087159
Allen RW; Criqui MH; Diez Roux AV; Allison M; Shea S; Detrano R; Sheppard L; Wong N; Hinckley Stukovsky K;
Kaufman JD. (2009). Fine particulate air pollution, proximity to traffic, and aortic atherosclerosis: The multi-ethnic
study of atherosclerosis. Epidemiology, 20: 254-264. 156209
Allison MA; Cheung P; Criqui MH; Langer RD; Wright CM. (2006). Mitral and aortic annular calcification are highly
associated with systemic calcified atherosclerosis. Circulation, 113: 861-866. 155653
Allison MA; Criqui MH; Wright CM. (2004). Patterns and risk factors for systemic calcified atherosclerosis. Arterioscler
Thromb Vase Biol, 24: 331-336. 156210
Amato M; Montorsi P; Ravani A; Oldani E; Galli S; Ravagnani PM; Tremoli E; Baldassarre D. (2007). Carotid intima-
media thickness by B-mode ultrasound as surrogate of coronary atherosclerosis: correlation with quantitative
coronary angiography and coronary intravascular ultrasound findings. Eur Heart J, 28: 2094-2101. 155656
Annesi-Maesano I; Moreau D; Caillaud D; Lavaud F; Le Moullec Y; Taytard A; Pauli G; Charpin D. (2007). Residential
proximity fine particles related to allergic sensitisation and asthma in primary school children. Respir Med, 101:
1721-1729.093180
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009 7-9
-------
Arad Y; Spadaro LA; Goodman K; Lledo-Perez A; Sherman S; Lerner G; Guerci AD. (1996). Predictive value of electron
beam computed tomography of the coronary arteries. 19-month follow-up of 1173 asymptomatic subjects.
Circulation, 93: 1951-1953. 155661
Araujo JA; Barajas B; Kleinman M; Wang X; Bennett BJ; Gong KW; Navab M; Harkema J; Sioutas C; Lusis AJ; Nel AE.
(2008). Ambient particulate pollutants in the ultrafine range promote early atherosclerosis and systemic oxidative
stress. Circ Res, 102: 589-596. 156222
Ardehali R; Nasir K; Kolandaivelu A; Budoff MJ; Blumenthal RS. (2007). Screening patients for subclinical
atherosclerosis with non-contrast cardiac CT. Atherosclerosis, 192: 235-242. 155662
Arlt VM; Glatt H; Gamboa da Costa G; Reynisson J; Takamura-Enya T; Phillips DH. (2007). Mutagenicity and DNA
adduct formation by the urban air pollutant 2-nitrobenzanthrone. Toxicol Sci, 98: 445-57. 097257
Auchincloss AH; Roux AV; Dvonch JT; Brown PL; Barr RG; Daviglus ML; Goff DC; Kaufman JD; O'Neill MS. (2008).
Associations between Recent Exposure to Ambient Fine Particulate Matter and Blood Pressure in the Multi-Ethnic
Study of Atherosclerosis (MESA). Environ Health Perspect, 116: 486-491. 156234
Avogbe PH; Ayi-Fanou L; Autrup H; Loft S; Fayomi B; Sanni A; Vinzents P; Moller P. (2005). Ultrafine particulate matter
and high-level benzene urban air pollution in relation to oxidative DNA damage. Carcinogenesis, 26: 613-620.
087811
Avol EL; Gauderman WJ; Tan SM; London SJ; Peters JM. (2001). Respiratory effects of relocating to areas of differing air
pollution levels. Am J Respir Grit Care Med, 164: 2067-2072. 020552
Baccarelli A; Martinelli I; Pegoraro V; Melly S; Grillo P; Zanobetti A; Hou L; Bertazzi PA; Mannucci PM; Schwartz J.
(2009). Living near Major Traffic Roads and Risk of Deep Vein Thrombosis. Circulation, 119: 3118-3124. 188183
Baccarelli A; Martinelli I; Zanobetti A; Grillo P; HouLF; Bertazzi PA; Mannucci PM; Schwartz J. (2008). Exposure to
Particulate Air Pollution and Risk of Deep Vein Thrombosis. Arch Intern Med, 168: 920-927. 157984
Baccarelli A; Wright RO; Bollati V; Tarantini L; Litonjua AA; Suh HH; Zanobetti A; Sparrow D; Vokonas PS; Schwartz J.
(2009). Rapid DNA Methylation Changes after Exposure to Traffic Particles . Am J Respir Grit Care Med, 179:
572-578. 192155
Baker DG (1998). Natural pathogens of laboratory mice, rats, and rabbits and their effects on research. Clin Microbiol Rev,
11:231-266. 156245
Bao L; Chen S; Wu L; Hei Tom K; Wu Y; Yu Z; Xu A. (2007). Mutagenicity of diesel exhaust particles mediated by cell-
particle interaction in mammalian cells. Toxicol Sci, 229: 91-100. 097258
Basu R; Woodruff TJ; Parker JD; Saulnier L; Schoendorf KG. (2004). Comparing exposure metrics in the relationship
between PM25 and birth weight in California. J Expo Sci Environ Epidemiol, 14: 391-396. 087896
Bayer-Oglesby L; Grize L; Gassner M; Takken-Sahli K; Sennhauser FH; Neu U; Schindler C; Braun-Fahrlander C. (2005).
Decline of ambient air pollution levels and improved respiratory health in Swiss children. Environ Health Perspect,
113: 1632-1637. 086245
Beelen R; Hoek G; van den Brandt PA; Goldbohm RA; Fischer P; Schouten LJ; Armstrong B; Brunekreef B. (2008). Long-
term exposure to traffic-related air pollution and lung cancer risk. Epidemiology, 19: 702-710. 155681
Beelen R; Hoek G; van den Brandt PA; Goldbohm RA; Fischer P; Schouten LJ; Jerrett M; Hughes E; Armstrong B;
Brunekreef B. (2008). Long-term effects of traffic-related air pollution on mortality in a Dutch cohort (NLCS-AIR
study). Environ Health Perspect, 116: 196-202. 156263
Beeson WL; Abbey DE; Knutsen SF. (1998). Long-term concentrations of ambient air pollutants and incident lung cancer
in California adults: results from the AHSMOG study. Environ Health Perspect, 106: 813-823. 048890
Bell ML; Ebisu K; Belanger K. (2007). Ambient air pollution and low birth weight in Connecticut and Massachusetts.
Environ Health Perspect, 115: 1118-24. 091059
Bell ML; Kim JY; Dominici F. (2007). Potential confounding of particulate matter on the short-term association between
ozone and mortality inmultisite time-series studies. Environ Health Perspect, 115: 1591-1595. 093256
Bhattacharyya N. (2009). Air quality influences the prevalence of hay fever and sinusitis. Laryngoscope, 119: 429-433.
180154
December 2009 7-99
-------
Bigazzi R; Bianchi S; Baldari D; Campese VM. (1998). Microalbuminuria predicts cardiovascular events and renal
insufficiency in patients with essential hypertension. J Hypertens, 16: 1325-1333. 156272
Binkova B; Cerna M; Pastorkova A; Jeli'nek R; Benes I; Novak J; Sram RJ. (2003). Biological activities of organic
compounds adsorbed onto ambient air particles: comparison between the cities of Teplice and Prague during the
summer and winter seasons 2000-2001. Mutat Res Fund Mol Mech Mutagen, 525: 43-59. 156274
Bobak M. (2000). Outdoor air pollution, low birth weight, and prematurity. Environ Health Perspect, 108: 173-176. 011448
Bobak M; Leon DA. (1992). Air pollution and infant mortality in the Czech Republic, 1986-1988. Lancet, 8826: 1010-
1014.044415
Bobak M; Leon DA. (1999). The effect of air pollution on infant mortality appears specific for respiratory causes in the
postneonatal period. Epidemiology, 10: 666-670. 007678
Bonner MR; Han D; Nie J; Rogerson P; Vena JE; Muti P; Trevisan M; Edge SB; Freudenheim JL. (2005). Breast cancer
risk and exposure in early life to poly cyclic aromatic hydrocarbons using total suspended particulates as a proxy
measure. Cancer Epidemiol Biomarkers Prev, 14: 53-60. 088993
Boushey HA; Armstrong B; Brauer M; Brunekreef B; Buckpitt A; Hoidal JR; London S; Reid N; Rom WN; Russell A.
(2008). Critique, Health Review Committee. In Grigg J; Kulkarni N; Pierse N; Rushton L; O?Callaghan C; Rutman
A (Ed.),Research report 134: Black-pigmented material in airway macrophages from healthy children: Association
with lung function and modeled PM10 (pp. 25-33). Boston, MA: Health Effects Institute. 192162
Bracken MB; Triche EW; Belanger K; Saftlas A; Beckett WS; Leaderer BP (2003). Asthma symptoms, severity, and drug
therapy: a prospective study of effects on 2205 pregnancies. Obstet Gynecol, 102: 739-753. 156288
Brauer M; Gehring U; Brunekreef B; De Jongste J; Gerritsen J; Rovers M; Wichmann H-E; Wijga A; Heinrich J. (2006).
Traffic-related air pollution and otitis media. Environ Health Perspect, 114: 1414-1418. 090757
Brauer M; Hoek G; Smit HA; De Jongste JC; Gerritsen J; Postma DS; Kerkhof M; Brunekreef B. (2007). Air pollution and
development of asthma, allergy and infections in a birth cohort. Eur Respir J, 29: 879-888. 090691
Brauer M; Hoek G; Van Vliet P; Meliefste K; Fischer PH; Wrjga A; Koopman LP; Neijens HJ; Gerritsen J; Kerkhof M;
Heinrich J; Bellander T; Brunekreef B. (2002). Air pollution from traffic and the development of respiratory
infections and asthmatic and allergic symptoms in children. Am J Respir Crit Care Med, 166: 1092-1098. 035192
Brauer M; Lencar C; Tamburic L; Koehoorn M; Demers P; Karr C. (2008). A cohort study of traffic-related air pollution
impacts on birth outcomes. Environ Health Perspect, 116: 680-686. 156292
Brown JS; Zeman KL; Bennett WD. (2001). Regional deposition of coarse particles and ventilation distribution in healthy
subjects and patients with cystic fibrosis. J Aerosol Med Pulm Drug Deliv, 14: 443-454. 012307
Brunekreef B; Beelen R; Hoek G; Schouten L; Bausch-Goldbohm S; Fischer P; Armstrong B; Hughes E; Jerrett M; van
den Brandt P. (2009). Effects of long-term exposure to traffic-related air pollution on respiratory and cardiovascular
mortality in the Netherlands: The NLCS-AIR Study. Health Effects Institute. Boston, MA. 139. 191947
Budoff MJ; Takasu J; Katz R; Mao S; Shavelle DM; O'Brien KD; Blumenthal RS; Carr JJ; Kronmal R. (2005).
Reproducibility of CT measurements of aortic valve calcification, mitral annulus calcification, and aortic wall
calcification in the multi-ethnic study of atherosclerosis. Acad Radiol, 13: 166-172. 192105
Bunger J; Krahl J; Weigel A; Schroder O; Bruning T; Muller M; Hallier E; Westphal G. (2006). Influence of fuel properties,
nitrogen oxides, and exhaust treatment by an oxidation catalytic converter on the mutagenicity of diesel engine
emissions. Arch Toxicol, 80: 540-546. 156303
Bunger J; Schappler-Scheele B; Hilgers R; Hallier E. (2007). A 5-year follow-up study on respiratory disorders and lung
function in workers exposed to organic dust from composting plants. Int Arch Occup Environ Health, 80: 306-312.
156304
Burchiel SW; Lauer FT; Dunaway SL; Zawadzki J; McDonald JD; Reed MD. (2005). Hardwood smoke alters murine
splenic T cell responses to mitogens following a 6-month whole body inhalation exposure. Toxicol Appl Pharmacol,
202: 229-236. 088090
Burchiel SW; Lauer FT; McDonald JD; Reed MD. (2004). Systemic immunotoxicity in AJ mice following 6-month whole
body inhalation exposure to diesel exhaust. Toxicol Appl Pharmacol, 196: 337-345. 055557
December 2009 7-100
-------
Calderon-Garciduenas L; Macias-Parra M; Hoffmann HJ; Valencia-Salazar G; Henriquez-Roldan C; Osnaya N; Monte OC;
Barragan-Mejia G; Villarreal-Calderon R; Romero L; Granada-Macias M; Torres-Jardon R; Medina-Cortina H;
Maronpot RR. (2009). Immunotoxicity and environment: immunodysregulation and systemic inflammation in
children. Toxicol Pathol, 37: 161-169. 192107
Calderon-Garciduenas L; Vincent R; Mora-Tiscareno A; Franco-Lira M; Henriquez-Roldan C; Barragan-Mejia G; Garrido-
Garcia L; Camacho-Reyes L; Valencia-Salazar G; Paredes R; Romero L; Osnaya H; Villarreal-Calderon R; Torres-
Jardon. (2007). Elevated plasma endothelin-1 and pulmonary arterial pressure in children exposed to air pollution.
Environ Health Perspect, 115: 1248-1253. 091252
Chambless LE; Heiss G; Folsom AR; Rosamond W; Szklo M; Sharrett AR; Clegg LX. (1997). Association of coronary
heart disease incidence with carotid arterial wall thickness and major risk factors: the Atherosclerosis Risk in
Communities (ARIC) Study, 1987-1993. Am J Epidemiol, 146: 483-494. 156329
Chang J; Delfino RJ; Gillen D; Tjoa T; Nickerson B; Cooper D. (2008). Repeated respiratory hospital encounters among
children with asthma and residential proximity to traffic. J Occup Environ Med, 66: 90-98. 180393
Chen JC; Schwartz J. (2008). Metabolic syndrome and inflammatory responses to long-term particulate air pollutants.
Environ Health Perspect, 116: 612-617. 190106
ChenL; Yang W; JennisonBL; Goodrich A; Omaye ST. (2002). Air pollution and birth weight in northern Nevada, 1991-
1999. Inhal Toxicol, 14: 141-157. 024945
Chen LC; Hwang JS. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice IV
Characterization of acute and chronic effects of ambient air fine particulate matter exposures on heart-rate
variability. Inhal Toxicol, 17: 209-216. 087218
Chen LC; Nadziejko C. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice V CAPs
exacerbate aortic plaque development in hyperlipidemic mice. Inhal Toxicol, 17: 217-224. 087219
Chen LH; Knutsen SF; Shavlik D; Beeson WL; Petersen F; Ghamsary M; Abbey D. (2005). The association between fatal
coronary heart disease and ambient particulate air pollution: Are females at greater risk?. Environ Health Perspect,
113: 1723-1729. 087942
Churg A; Brauer M; del Carmen Avila-Casado M; Fortoul TI; Wright JL. (2003). Chronic exposure to high levels of
particulate air pollution and small airway remodeling. Environ Health Perspect, 111: 714-718. 087899
Clancy L; Goodman P; Sinclair H; Dockery DW. (2002). Effect of air pollution control on death rates in Dublin, Ireland: an
intervention study. Lancet, 360: 1210-1214. 035270
Claxton LD; Matthews PP; Warren SH. (2004). The genotoxicity of ambient outdoor air, a review: Salmonella
mutagenicity. DNARepair (Amst), 567: 347-399. 089008
Claxton LD; Woodall GM. (2007). A review of the mutagenicity and rodent carcinogenicity of ambient air. Mutat Res Rev
Mutat Res, 636: 36-94. 180391
Clifton VL; Giles WB; Smith R; Bisits AT; Hempenstall PA; Kessell CG; Gibson PG (2001). Alterations of placental
vascular function in asthmatic pregnancies. Am J Respir Crit Care Med, 164: 546-554. 156360
Craven TE; Ryu JE; Espeland MA; Kahl FR; McKinney WM; Toole JF; McMahan MR; Thompson CJ; Heiss G; Grouse JR
3rd. (1990). Evaluation of the associations between carotid artery atherosclerosis and coronary artery stenosis. A
case-control study. Circulation, 82: 1230-1242. 155740
Cury PM; Lichtenfels AJ; Reymao MS; Conceicao GM; Capelozzi VL; Saldiva PH. (2000). Urban levels of air pollution
modifies the progression of urethane-induced lung tumours in mice. Pathol Res Pract, 196: 627-633. 192100
Dales R; Burnett RT; Smith-Doiron M; Stieb DM; Brook JR. (2004). Air pollution and sudden infant death syndrome.
Pediatrics, 113: 628-631. 087342
Dales R; Wheeler A; Mahmud M; Frescura AM; Smith-Doiron M; Nethery E; Liu L. (2008). The Influence of Living Near
Roadways on Spirometry and Exhaled Nitric Oxide in Elementary Schoolchildren. Environ Health Perspect, 116:
1423-1427. 156378
Danielsen PH; Loft S; M011er P. (2008). DNA damage and cytotoxicity in type II lung epithelial (A549) cell cultures after
exposure to diesel exhaust and urban street particles. Part Fibre Toxicol, 5: 6. 192092
December 2009 7-101
-------
Deckert T; Yokoyama H; Mathiesen E; Ronn B; Jensen T; Feldt-Rasmussen B; Borch-Johnsen K; Jensen JS. (1996). Cohort
study of predictive value of urinary albumin excretion for atherosclerotic vascular disease in patients with insulin
dependent diabetes. Br Med J, 312: 871-874. 156389
DeMarini DM; Brooks LR; Warren SH; Kobayashi T; Gilmour MI; Singh P. (2004). Bioassay-directed fractionation and
Salmonella mutagenicity of automobile and forklift diesel exhaust particles. Environ Health Perspect, 112: 814-819.
066329
De Kok TM; Hogervorst JG; Briede JJ; Van Herwijnen MH; Maas LM; Moonen EJ; Driece HA; Kleinjans JC. (2005).
Genotoxicity and physicochemical characteristics of traffic-related ambient particulate matter. Environ Mol
Mutagen, 46: 71-80. 088656
Diez Roux AV; Auchincloss AH; Franklin TG; Raghunathan T; Barr RG; Kaufman J; Astor B; Keeler J. (2008). Long-term
exposure to ambient particulate matter and prevalence of subclinical atherosclerosis in the Multi-Ethnic Study of
Atherosclerosis. Am J Epidemiol, 167: 667-675. 156401
Dinneen SF; Gerstein HC. (1997). The association of microalbuminuria and mortality in non-insulin-dependent diabetes
mellitus. A systematic overview of the literature. Arch Intern Med, 157: 1413-1418. 156403
Dockery DW; Damokosh AI; Neas LM; Raizenne M; Spengler JD; Koutrakis P; Ware JH; Speizer FE. (1996). Health
effects of acid aerosols on North American children: respiratory symptoms and illness. Environ Health Perspect,
104: 500-505. 046219
Dockery DW; Pope CAIII; Xu X; Spengler JD; Ware JH; Fay ME; Ferris BG Jr; Speizer FE. (1993). An association
between air pollution and mortality in six US cities. N Engl J Med, 329: 1753-1759. 044457
Downs SH; Schindler C; Liu L-JS; Keidel D; Bayer-Oglesby L; Brutsche MH; Gerbase MW; Keller R; Kunzli N;
Leuenberger P; Probst-Hensch NM; Tschopp J-M; Zellweger J-P; Rochat T; Schwartz J; Ackermann-Liebrich U;.
(2007). Reduced exposure to PM10 and attenuated age-related decline in lung function. J Expo Sci Environ
Epidemiol, 15: 185-204. 092853
Dugandzic R; Dodds L; Stieb D; Smith-Doiron M. (2006). The association between low level exposures to ambient air
pollution and term low birth weight: A retrospective cohort study. Environ Health, 5:3. 088681
Eftim SE; Samet JM; Janes H; McDermott A; Dominici F. (2008). Fine particulate matter and mortality: a comparison of
the six cities and American Cancer Society cohorts with a medicare cohort. Epidemiology, 19: 209-216. 099104
Enstrom JE. (2005). Fine particulate air pollution and total mortality among elderly Californians, 1973-2002. Inhal Toxicol,
17: 803-816. 087356
Erbel R; Mohlenkamp S; Kerkhoff G; Budde T; Schmermund A. (2007). Non-invasive screening for coronary artery
disease: calcium scoring. Heart, 93: 1620-1629. 155768
Erdinger L; Durr M; Hopker KA. (2005). Correlations between mutagenic activity of organic extracts of airborne
particulate matter, NOx and sulphur dioxide in southern Germany: results of a two-year study. Environ Sci Pollut
Reslnt, 12: 10-20. 156423
Fedulov AV; Leme A; Yang Z; Dahl M; Lim R; Mariani TJ; Kobzik L. (2008). Pulmonary exposure to particles during
pregnancy causes increased neonatal asthma susceptibility. Am J Respir Cell Mol Biol, 38: 57-67. 097482
Finch GL; Hobbs CH; Blair LF; Barr EB; Hahn FF; Jaramillo RJ; Kubatko JE; March TH; White RK; Krone JR; Menache
MG; Nikula KJ; Mauderly JL; Van Gerpen J; Merceica MD; Zielinska B; Stankowski L; Burling K; Howell S;
Mauderly JL. (2002). Effects of subchronic inhalation exposure of rats to emissions from a diesel engine burning
soybean oil-derived biodiesel fuel. Inhal Toxicol, 14: 1017-1048. 054603
Floyd HS; Chen LC; Vallanat B; Dreher K. (2009). Fine ambient air particulate matter exposure induces molecular
alterations associated with vascular disease progression within plaques of atherosclerotic susceptible mice. Inhal
Toxicol, 21: 394-403. 190350
Forbes LJ; Patel MD; Rudnicka AR; Cook DG; Bush T; Stedman JR; Whincup PH; Strachan DP; Anderson RH. (2009).
Chronic exposure to outdoor air pollution and markers of systemic inflammation. Epidemiology, 20: 245-253.
190351
Forman JP; Brenner BM. (2006). 'Hypertension' and 'microalbuminuria': the bell tolls for thee. Kidney Int, 69: 22-28.
156439
December 2009 7-102
-------
Foster PMD; Mylchreest E; Gaido KW; Sar M. (2001). Effects of phthalate esters on the developing reproductive tract of
male rats. Hum Reprod Update, 7: 231-235. 156442
Foster PMD; Thomas LV; Cook MW; Gangolli SD. (1980). Study of the testicular effects and changes in zinc excretion
produced by some n-alkyl phthalates in the rat. Toxicol Appl Pharmacol, 54: 392-398. 094701
Franklin SS; Sutton-Tyrrell K; Belle SH; Weber MA; Kuller LH. (1997). The importance of pulsatile components of
hypertension in predicting carotid stenosis in older adults. J Hypertens, 15: 1143-1150. 156446
Fryer ME; Collins CD. (2003). Model intercomparison for the uptake of organic chemicals by plants. Environ Sci Technol,
37: 1617-1624. 156454
Fujimoto A; Tsukue N; Watanabe M; Sugawara I; Yanagisawa R; Takano H; Yoshida S; Takeda K. (2005). Diesel exhaust
affects immunological action in the placentas of mice. Environ Toxicol, 20: 431-440. 096556
Gabelova A; Valovicova Z; Bacova G; Labaj J; Binkova B; Topinka J; Sevastyanova O; Sram RJ; Kalina I; Habalova V;
Popov TA; Panev T; Farmer PB. (2007). Sensitivity of different endpoints for in vitro measurement of genotoxicity
of extractable organic matter associated with ambient airborne particles (PM10). Mutat Res Fund Mol Mech
Mutagen, 620: 103-113. 156458
Gabelova A; Valovicova Z; Labaj J; Bacova G; Binkova B; Farmer Peter B. (2007). Assessment of oxidative DNA damage
formation by organic complex mixtures from airborne particles PM(10). Mutat Res, 620: 135-144. 156457
Gardner MB. (1966). Biological effects of urban air pollution III Lung tumors in mice. Arch Environ Occup Health, 12:
305-313.015129
Gardner MB; Loosli CG; Hanes B; Blackmore W; Teebken D. (1969). Histopathologic findings in rats exposed to ambient
and filtered air. Arch Environ Occup Health, 19: 637-647. 015130
Gauderman WJ; Avol E; Gilliland F; Vora H; Thomas D; Berhane K; McConnell R; Kuenzli N; Lurmann F; Rappaport E;
Margolis H; Bates D; Peters J. (2004). The effect of air pollution on lung development from 10 to 18 years of age.
NEnglJMed, 351: 1057-1067. 056569
Gauderman WJ; Gilliland GF; Vora H; Avol E; Stram D; McConnell R; Thomas D; Lurmann F; Margolis HG; Rappaport
EB; Berhane K; Peters JM. (2002). Association between air pollution and lung function growth in southern
California children: results from a second cohort. Am J Respir Crit Care Med, 166: 76-84. 026013
Gauderman WJ; McConnell R; Gilliland F; London S; Thomas D; Avol E; Vora H; Berhane K; Rappaport EB; Lurmann F;
Margolis HG; Peters J. (2000). Association between air pollution and lung function growth in southern California
children. Am J Respir Crit Care Med, 162: 1383-1390. 012531
Gehrig R; Buchmann B. (2003). Characterising seasonal variations and spatial distribution of ambient PM10 and PM25
concentrations based on long-term Swiss monitoring data. Atmos Environ, 37: 2571-2580. 139678
Gehring U; Heinrich J; Kramer U; Grote V; Hochadel M; Sugiri D; Kraft M; Rauchfuss K; Eberwein HG; Wichmann H-E.
(2006). Long-term exposure to ambient air pollution and cardiopulmonary mortality in women. Epidemiology, 17:
545-551.089797
Geroulakos G; O'Gorman DJ; Kalodiki E; Sheridan DJ; Nicolaides AN. (1994). The carotid intima-media thickness as a
marker of the presence of severe symptomatic coronary artery disease. Eur Heart J, 15: 781-785. 155788
Gerstein HC; Mann JF; Yi Q; Zinman B; Dinneen SF; Hoogwerf B; Halle JP; Young J; Rashkow A; Joyce C; Nawaz S;
Yusuf S. (2001). Albuminuria and risk of cardiovascular events, death, and heart failure in diabetic and nondiabetic
individuals. JAMA, 286: 421-426. 156466
Ghosh R; Rankin J; Pless-Mulloli T; Glinianaia S. (2007). Does the effect of air pollution on pregnancy outcomes differ by
gender? A systematic review. Environ Res, 105: 400-408. 091233
Gilboa SM; Mendola P; Olshan AF; Langlois PH; Savitz DA; Loomis D; Herring AH; Fixler DE. (2005). Relation between
ambient air quality and selected birth defects, seven county study, Texas, 1997-2000. Am J Epidemiol, 162: 238-
252. 087892
Girerd X; Mourad JJ; Acar C; Heudes D; Chiche S; Bruneval P; Mignot JP; Billaud E; Safar M; Laurent S. (1994).
Noninvasive measurement of medium-sized artery intima-media thickness in humans: in vitro validation. J Vase
Res, 31: 114-120. 156474
December 2009 7-103
-------
Glinianaia FV; Rankin J; Bell R; Pless-Mulloli T; Howel D. (2004). Does particulate air pollution contribute to infant
death? A systematic review. Environ Health Perspect, 112: 1365-1371. 087898
Goss CH; Newsom SA; Schildcrout JS; Sheppard L; Kaufman JD. (2004). Effect of ambient air pollution on pulmonary
exacerbations and lung function in cystic fibrosis. Am J Respir Crit Care Med, 169: 816-821. 055624
Gotschi T; Heinrich J; Sunyer J; Kunzli N. (2008). Long-term effects of ambient air pollution on lung function: A review.
Epidemiology, 19: 690-701. 156485
Gotschi T; Sunyer J; Chinn S; de Marco R; Forsberg B; Gauderman JW; Garcia-Esteban R; Heinrich J; Jacquemin B; Jarvis
D; Ponzio M; Villani S; Kunzli N. (2008). Air pollution and lung function in the European Community Respiratory
Health Survey. Int J Epidemiol, 37: 1349-1358. 180364
Gottipolu RR; Wallenborn JG; Karoly ED; Schladweiler MC; Ledbetter AD; Krantz T; Linak WP; Nyska A; Johnson JA;
Thomas R; Richards JE; Jaskot RH; Kodavanti UP. (2009). One-month diesel exhaust inhalation produces
hypertensive gene expression pattern in healthy rats. Environ Health Perspect, 117: 38-46. 190360
Gouveia N; Bremner SA; Novaes HMD. (2004). Association between ambient air pollution and birth weight in Sao Paulo,
Brazil. J Epidemiol Community Health, 58: 11-17. 055613
Greenland P; Kizilbash MA. (2005). Coronary computed tomography in coronary risk assessment. J Cardpulm Rehabil, 25:
3-10. 156496
Grigg J; Kulkarni N; Pierse N; Rushton L; O'Callaghan C; Rutman A. (2008). Black-pigmented material in airway
macrophages from healthy children: association with lung function and modeled PM10. Heath Effects Institute.
Boston, MA. Research Report 134. 156499
Gunnison A; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice VI Gene
expression in heart and lung tissue. Inhal Toxicol, 17: 225-233. 087956
Gutierrez-Castillo ME; Roubicek DA; Cebrian-Garcia ME; De Vizcaya-Ruiz A; Sordo-Cedeno M; Ostrosky-Wegman P.
(2006). Effect of chemical composition on the induction of DNA damage by urban airborne particulate matter.
Environ Mol Mutagen, 47: 199-211. 089030
Ha E-H; Lee J-T; Kim H; Hong Y-C; Lee. (2003). Infant susceptibility of mortality to air pollution in Seoul, South Korea.
Pediatrics, 111: 284-290. 042552
Haland G; Carlsen KG; Sandvik L; Devulapalli CS; Munthe-Kaas MC; Pettersen M; Carlsen KH. (2006). Reduced lung
function at birth and the risk of asthma at 10 years of age. N Engl J Med, 355: 1682-1689. 156511
Hamada K; Suzaki Y; Leme A; Ito T; Miyamoto K; Kobzik L; Kimura H. (2007). Exposure of pregnant mice to an air
pollutant aerosol increases asthma susceptibility in offspring. J Toxicol Environ Health A, 70: 688-695. 091235
Hammoud A; Carrell DT; Gibson M; Sanderson M; Parker-Jones K; Matthew Peterson C. (2009). Decreased sperm
motility is associated with air pollution in Salt Lake City. Fertil Steril, TBD: TBD. 192156
Hannigan MP; Cass GR; Penman BW; Crespi CL; Lafleur AL; Busby WF Jr; Thilly WG (1997). Human cell mutagens in
Los Angeles air. Environ Sci Technol, 31: 438-447. 083598
Hansen C; Neller A; Williams G; Simpson R. (2006). Maternal exposure to low levels of ambient air pollution and preterm
birth in Brisbane, Australia. BJOG, 113: 935-941. 089818
Hansen C; Neller A; Williams G; Simpson R. (2007). Low levels of ambient air pollution during pregnancy and fetal
growth among term neonates in Brisbane, Australia. Environ Res, 103: 383-389. 090703
Harrod KS; Jaramillo RJ; Berger JA; Gigliotti AP; Seilkop SK; Reed MD. (2005). Inhaled diesel engine emissions reduce
bacterial clearance and exacerbate lung disease to Pseudomonas aeruginosa infection in vivo. Toxicol Sci, 83: 155-
165. 088144
Hashimoto AH; Amanuma K; Hiyoshi K; Sugawara Y; Goto S; Yanagisawa R; Takano H; Masumura K-I; Nohmi T; Aoki
Y. (2007). Mutations in the lungs of gpt delta transgenic mice following inhalation of diesel exhaust. Environ Mol
Mutagen, 48: 682-693. 097261
Healey K; Smith EC; Wild CP; Routledge MN. (2006). The mutagenicity of urban particulate matter in an enzyme free
system is associated with the generation of reactive oxygen species. Mutat Res Fund Mol Mech Mutagen, 602: 1-6.
156532
December 2009 7-104
-------
Heinrich J; Hoelscher B; Frye C; Meyer I; Pitz M; Cyrys J; Wjst M; Neas L; Wichmann H-E. (2002). Improved air quality
in reunified Germany and decreases in respiratory symptoms. Epidemiology, 13: 394-401. 034825
Heinrich J; Slama R. (2007). Fine particles, a major threat to children. Int J Hyg Environ Health, 210: 617-622. 156534
Heiss G; Sharrett AR; Barnes R; Chambless LE; Szklo M; Alzola C. (1991). Carotid atherosclerosis measured by B-mode
ultrasound in populations: associations with cardiovascular risk factors in the ARIC study. Am J Epidemiol, 134:
250-256. 156535
Hiramatsu K; Azuma A; Kudoh S; Desaki M; Takizawa H; Sugawara I. (2003). Inhalation of diesel exhaust for three
months affects major cytokine expression and induces bronchus-associated lymphoid tissue formation in murine
lungs. Exp Lung Res, 29: 607-622. 155846
Hoek G; Brunekreef B; Goldbohm S; Fischer P; Van den Brandt PA. (2002). Association between mortality and indicators
of traffic-related air pollution in the Netherlands: A cohort study. Lancet, 360: 1203-1209. 042364
Hoffmann B; Moebus S; Kroger K; Stang A; Mohlenkamp S; Dragano N; Schmermund A; Memmesheimer M; Erbel R;
Jockel K-H. (2009). Residential exposure to urban pollution, ankle-brachial index, and peripheral arterial disease.
Epidemiology, 20: 280-288. 190376
Hoffmann B; Moebus S; Mohlenkamp S; Stang A; Lehmann N; Dragano N; Schmermund A; Memmesheimer M; Mann K;
Erbel R; Jockel K-H; Heinz Nixdorf Recall Study Investigative Group. (2007). Residential exposure to traffic is
associated with coronary atherosclerosis. Circulation, 116: 489-496. 091163
Hoffmann B; Moebus S; Stang A; Beck E-M; Dragano N; Mohlenkamp S; Schmermund A; Memmesheimer M; Mann K;
Erbel R; Jockel K-H; Heinz Nixdorf RECALL Study Investigative Group. (2006). Residence close to high traffic
and prevalence of coronary heart disease. Eur Heart J, 27: 2696-2702. 091162
Hoffmann MH; Shi H; Schmitz BL; Schmid FT; Lieberknecht M; Schulze R; Ludwig B; Kroschel U; Jahnke N; Haerer W;
Brambs HJ; Aschoff AJ. (2005). Noninvasive coronary angiography with multislice computed tomography. JAMA,
293:2471-2478. 156556
Hollander M; Hak AE; Koudstaal PJ; Bots ML; Grobbee DE; Hofman A; Witteman JC; Breteler MM. (2003). Comparison
between measures of atherosclerosis and risk of stroke: the Rotterdam Study. Stroke, 34: 2367-2372. 156562
Hornberg C; Maciuleviciute L; Seemayer NH. (1996). Sister chromatid exchanges in rodent tracheal epithelium exposed in
vitro to environmental pollutants. Toxicol Lett, 88: 45-53. 087164
Hornberg C; Maciuleviciute L; Seemayer NH; Kainka E. (1998). Induction of sister chromatid exchanges (SCE) in human
tracheal epithelial cells by the fractions PM-10 and PM-2.5 of airborne particulates. Toxicol Lett, 96: 215-220.
155849
Hougaard KS; Jensen KA; Nordly P; Taxvig C; Vogel U; Saber AT; Wallin H. (2008). Effects of prenatal exposure to diesel
exhaust particles on postnatal development, behavior, genotoxicity and inflammation in mice. Part Fibre Toxicol, 5:
3. 156570
Huang JY; Liao JW; Liu YC; Lu SY; Chou CP; Chan WH; Chen SU; Ueng TH. (2008). Motorcycle exhaust induces
reproductive toxicity and testicular interleukin-6 in male rats. Toxicol Sci, 103: 137-148. 156574
Huynh M; Woodruff TJ; Parker JD; Schoendorf KC. (2006). Relationships between air pollution and preterm birth in
California. Paediatr Perinat Epidemiol, 20: 454-461. 091240
Hwang BF; Jaakkola JJ. (2008). Ozone and other air pollutants and the risk of oral clefts. Environ Health Perspect, 116:
1411-1415. 193794
Hwang J-S; Nadziejko C; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in
mice: III Acute and chronic effects of CAPs on heart rate, heart-rate fluctuation, and body temperature. Inhal
Toxicol, 17: 199-207. 087957
IARC. (1989). Diesel and gasoline engine exhausts. Lyon, France: International Agency for Research on Cancer. 002958
Iba MM; Fung J; Chung L; Zhao J; Winnik B; Buckley BT; Chen LC; Zelikoff JT; Kou YR. (2006). Differential
inducibility of rat pulmonary CYP1 Al by cigarette smoke and wood smoke. Mutat Res Fund Mol Mech Mutagen,
606: 1-11. 156582
Ichinose T; Yajima Y; Nagashima M; Takenoshita S; Nagamachi Y; Sagai M. (1997). Lung carcinogenesis and formation of
8-hydroxy-deoxyguanosine in mice by diesel exhaust particles. Carcinogenesis, 18: 185-192. 053264
December 2009 7-105
-------
Ishihara Y; Kagawa J. (2003). Chronic diesel exhaust exposures of rats demonstrate concentration and time-dependent
effects on pulmonary inflammation. Inhal Toxicol, 15: 473-492. 096404
Islam T; Gauderman WJ; Berhane K; McConnell R; Avol E; Peters JM; Gilliland FD. (2007). The relationship between air
pollution, lung function and asthma in adolescents. Thorax, 62: 957-963. 090697
Izawa H; Watanabe G; Taya K; Sagai M. (2007). Inhibitory effects of foods and polyphenols on activation of aryl
hydrocarbon receptor induced by diesel exhaust particles. Environ Sci J Integr Environ Res, 14: 149-156. 190387
Jacobsen NR; Mrller P; Cohn CA; Loft S; Vogel U; Wallin H. (2008). Diesel exhaust particles are mutagenic in FE1 -
Muta™ Mouse lung epithelial cells. Mutat Res Fund Mol Mech Mutagen, 641: 54-57. 156597
Jalaludin B; Marines T; Morgan G; Lincoln D; Sheppeard V; Corbett S. (2007). Impact of ambient air pollution on
gestational age is modified by season in Sydney, Australia. Environ Health, 6:16. 156601
Janes H; Dominici F; Zeger SL. (2007). Trends in air pollution and mortality: an approach to the assessment of unmeasured
confounding. Epidemiology, 18: 416-423. 090927
Janssen NAH; Brunekreef B; van Vliet P; Aarts F; Maliefste K; Harssema H; Fischer P. (2003). The relationship between
air pollution from heavy traffic and allergic sensitization, bronchial byperresponsiveness, and respiratory symptoms
in Dutch schoolchildren. Epidemiology, 111: 1512-1518. 133555
Jerrett M; Burnett RT; Ma R; Pope 3rd CA; Krewski D; Newbold KB; Thurston G; Shi Y; Finkelstein N; Calle EE; Thun
MJ. (2005). Spatial analysis of air pollution and mortality in Los Angeles. Epidemiology, 16: 727-36. 189405
Jerrett M; Buzzelli M; Burnett RT; DeLuca PF. (2005). Particulate air pollution, social confounders, and mortality in small
areas of an industrial city. Soc Sci Med, 60: 2845-2863. 087381
Karr C; Lumley T; Schreuder A; Davis R; Larson T; Ritz B; Kaufman J. (2007). Effects of subchronic and chronic
exposure to ambient air pollutants on infant bronchiolitis. Am J Epidemiol, 165: 553-560. 090719
Karr C; Rudra C; Miller K; Gould T; Larson T; Sathyanarayana S; Koenig J. (2009). Infant exposure to fine particulate
matter and traffic and risk of hospitalization for RSV bronchiolitis in a region with lower ambient air pollution.
Environ Res, 109: 321-327. 191946
Karthikeyan VJ; Lip GYH. (2007). Peripheral artery disease and hypertension: the relation between ankle-brachial index
and mortality. J Hum Hypertens, 21: 762-765. 156626
Kato A; Kagawa J. (2003). Morphological effects in rat lungs exposed to urban roadside air. Inhal Toxicol, 15: 799-818.
089563
Khattar RS; Swales JD; Dore C; Senior R; Lahiri A. (2001). Effect of aging on the prognostic significance of ambulatory
systolic, diastolic, and pulse pressure in essential hypertension. Circulation, 104: 783-789. 155896
Khoury Z; Schwartz R; Gottlieb S; Chenzbraun A; Stern S; Keren A. (1997). Relation of coronary artery disease to
atherosclerotic disease in the aorta, carotid, and femoral arteries evaluated by ultrasound. Am J Cardiol, 80: 1429-
1433. 156636
Kim JJ; Smorodinsky S; Lipsett M; Singer BC; Hodgson AT; Ostro B. (2004). Traffic-related air pollution near busy roads:
the East Bay children's Respiratory Health Study. Am J Respir Grit Care Med, 170: 520-526. 087383
Kim OJ; Ha EH; Kim BM; Park HS; Jung WJ; Lee BE; Suh YJ; Kim YJ; Lee JT; Kim H; Hong YC. (2007). PM10 and
pregnancy outcomes: a hospital-based cohort study of pregnant women in Seoul. J Occup Environ Med, 49:13 94-
1402. 156642
Knight EL; Curhan GC. (2003). Albuminuria: moving beyond traditional microalbuminuria cut-points. Current Opinion
Nephrol Hypertens, 12: 283-284. 179900
Knobel HH; Chen CJ; Liang KY (1995). Sudden infant death syndrome in relation to weather and optimetrically measured
air pollution in Taiwan. Pediatrics, 96: 1106-1110. 155905
Krewski D. (2009). Evaluating the effects of ambient air pollution on life expectancy. N Engl J Med, 360: 413-415. 190075
Krewski D; Burnett RT; Goldberg MS; Hoover K; Siemiatycki J; Jerrett M; Abrahamowicz M; White WH. (2000).
Reanalysis of the Harvard Six Cities study and the American Cancer Society study of particulate air pollution and
mortality. Health Effects Institute. Cambridge, MA.http://pubs.healtheffects.org/view.php?id=6. 012281
December 2009 7-106
-------
Krewski D; Jerrett M; Burnett RT; Ma R; Hughes E; Shi Y; Turner MC; Pope AC III; Thurston G; Calle EE; Thun MJ.
(2009). Extended follow-up and spatial analysis of the American Cancer Society study linking particulate air
pollution and mortality. Health Effects Institute. Cambridge, MA. Report Nr. 140. 191193
Kulkarni N; Pierse N; Rushton L; Grigg J. (2006). Carbon in airway macrophages and lung function in children. N Engl J
Med, 355: 21-30. 089257
Kundu JK; Surh YJ. (2008). Inflammation: Gearing the journey to cancer. Mutat Res, 659: 15-30. 198840
Kunzli N; Bridevaux P-O; Liu S; Garcia-Esteban R; Schindler C; Gerbase M; Sunyer J; keidel D; Rochat T. (2009).
Traffic-related air pollution correlates with adult-onset asthma among never-smokers. Thorax, 64: 664-670. 191949
Kunzli N; Jerrett M; Mack WJ; Beckerman B; LaBree L; Gilliland F; Thomas D; Peters J; Hodis HN. (2005). Ambient air
pollution and atherosclerosis in Los Angeles. Environ Health Perspect, 113: 201-206. 087387
Kunzli N; Perez L; Lurmann F; Hricko A; Penfold B; McConnell R. (2008). An Attributable Risk Model for Exposures
Assumed to Cause Both Chronic Disease and its Exacerbations. Epidemiology, 19: 179-185. 129258
Lacasana M; Esplugues A; Ballester F. (2005). Exposure to ambient air pollution and prenatal and early childhood health
effects. Eur J Epidemiol, 20: 183-199. 155914
Laden F; Schwartz J; Speizer FE; Dockery DW. (2006). Reduction in fine particulate air pollution and mortality: extended
follow-up of the Harvard Six Cities study. Am J Respir Grit Care Med, 173: 667-672. 087605
Lee BE; Ha EH; Park HS; Kim YJ; Hong YC; Kim H; Lee JT. (2003). Exposure to air pollution during different gestational
phases contributes to risks of low birth weight. Hum Reprod, 18: 638-643. 043202
Lee JS; Kim KI; Baek SH. (2008). Nuclear receptors and coregulators in inflammation and cancer . Cancer Lett, 267: 189-
196. 198811
Leem J-H; Kaplan BM; Shim YK; Pohl HR; Gotway CA; Bullard SM; Rogers JF; Smith MM; Tylenda CA. (2006).
Exposures to air pollutants during pregnancy andpreterm delivery. Environ Health Perspect, 114: 905-910. 089828
Lemos M; Mohallen S; Macchione M; Dolhnikoff M; Assunaao J; Godleski J; Saldiva P. (2006). Chronic exposure to
urban air Pollution induces structural alterations in murine pulmonary and coronary arteries. Inhal Toxicol, 18: 247-
253. 088594
Li Y-J; Kawada T; Matsumoto A; Azuma A; Kudoh S; Takizawa H; Sugawara I. (2007). Airway inflammatory responses to
oxidative stress induced by low-dose diesel exhaust particle exposure differ between mouse strains. Exp Lung Res,
33: 227-244. 155929
Lichtenfels AJFC; Gomes JB; Fieri PC; Miraglia SGEK; Hallak J; Saldiva PHN. (2007). Increased levels of air pollution
and a decrease in the human and mouse male-to-female ration in Sao Paulo, Brazil. Fertil Steril, 87: 230-232.
097041
Lin C-M; Li C-Y; Mao I-F. (2004). Increased risks of term low-birth-weight infants in a petrochemical industrial city with
high air pollution levels. Arch Environ Occup Health, 59: 663-668. 089827
Lin CA; Pereira LAA; Nishioka DC; Conceicao GMS; Graga ALF; Saldiva PHN. (2004). Air pollution and neonatal deaths
in Sao Paulo, Brazil. Braz J Med Biol Res, 37: 765-770. 095787
Lipfert FW; Baty JD; Miller JP; Wyzga RE. (2006). PM2.5 constituents and related air quality variables as predictors of
survival in a cohort of U.S. military veterans. Inhal Toxicol, 18: 645-657. 088756
Lipfert FW; Wyzga RE; Baty JD; Miller JP. (2006). Traffic density as a surrogate measure of environmental exposures in
studies of air pollution health effects: long-term mortality in a cohort of US veterans. Atmos Environ, 40: 154-169.
088218
Lipfert FW; Zhang J; Wyzga RE. (2000). Infant mortality and air pollution: a comprehensive analysis of US data for 1990.
J Air Waste Manag Assoc, 50: 1350-1366. 004103
Lippmann M; Gordon T; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles in mice. IX.
Integral assessment and human health implications of subchronic exposures of mice to CAPs. Inhal Toxicol, 17:
255-261.087452
Liu L-JS; Curjuric I; Keidel D; Heldstab J; Kunzli N; Bayer-Oglesby L; Ackermann-Liebrich U; Schindler C; SAPALDIA
team. (2007). Characterization of source-specific air pollution exposure for a large population-based Swiss cohort
(SAPALDIA). Environ Health Perspect, 115: 1638-1645. 093093
December 2009 7-107
-------
Liu S; Krewski D; Shi Y; Chen Y; Burnett R. (2007). Association between maternal exposure to ambient air pollutants
during pregnancy and fetal growth restriction. J Expo Sci Environ Epidemiol, 17: 426-432. 090429
Liu Y-Q; Keane M; Ensell M; Miller W; Kashon M; Ong T-m; Mauderly J; Lawson D; Gautam M; Zielinska B; Whitney
K; Eberhardt J; Wallace W. (2005). In vitro genotoxicity of exhaust emissions of diesel and gasoline engine vehicles
operated on a unified driving cycle. J Environ Monit, 7: 60-66. 097019
Loomis D; Castillejos M; Gold DR; McDonnell W; Borja-Aburto VH. (1999). Air pollution and infant mortality in Mexico
City. Epidemiology, 10: 118-123. 087288
Lopes FD; Pinto TS; Arantes-Costa FM; Moriya HT; Biselli PJ; Ferraz LF; Lichtenfels AJ; Saldiva PH; Mauad T; Martins
MA. (2009). Exposure to ambient levels of particles emitted by traffic worsens emphysema in mice. Environ Res,
109:544-551. 190430
Lund AK; Knuckles TL; Obot Akata C; Shohet R; McDonald JD; Gigliotti A; Seagrave JC; Campen MJ. (2007). Gasoline
exhaust emissions induce vascular remodeling pathways involved in atherosclerosis. Toxicol Sci, 95: 485-94.
125741
Maciejczyk P; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice: VIII
source-related daily variations in in vitro responses to CAPs. Inhal Toxicol, 17: 243-253. 087456
Mackman N; Tilley RE; Key NS. (2007). Role of the extrinsic pathway of blood coagulation in hemostasis and thrombosis.
Arterioscler Thromb Vase Biol, 27: 1687-1693. 156723
Maheswaran R; Haining RP; Brindley P; Law J; Pearson T; Fryers PR; Wise S; Campbell MJ. (2005). Outdoor air pollution
and stroke in Sheffield, United Kingdom: A small-area level geographical study. Stroke, 36: 239-243. 088683
Maheswaran R; Haining RP; Brindley P; Law J; Pearson T; Fryers PR; Wise S; Campbell MJ. (2005). Outdoor air
pollution, mortality, and hospital admissions from coronary heart disease in Sheffield, UK: A small-area level
ecological study. Eur Heart J, 26: 2543-2549. 090769
Maisonet M; Bush TJ; Correa A; Jaakkola JJK. (2001). Relation between ambient air pollution and low birth weight in the
northeastern United States. Environ Health Perspect, 1093: 351-356. 016624
Maisonet M; Correa A; Misra D; Jaakkola JJ. (2004). A review of the literature on the effects of ambient air pollution on
fetal growth. Environ Res, 95: 106-115. 156725
Mannes T; Jalaludin B; Morgan G; Lincoln D; Sheppeard V; Corbett S. (2005). Impact of ambient air pollution on birth
weight in Sydney, Australia. Occup Environ Med, 62: 524-530. 087895
Matsumoto A; Hiramatsu K; Li Y; Azuma A; Kudoh S; Takizawa H; Sugawara I. (2006). Repeated exposure to low-dose
diesel exhaust after allergen challenge exaggerates asthmatic responses in mice. Clin Immunol, 121: 227-235.
098017
Mauad T; Rivero DH; de Oliveira RC; Lichtenfels AJ; Guimaraes ET; de Andre PA; Kasahara DI; Bueno HM; Saldiva PH.
(2008). Chronic exposure to ambient levels of urban particles affects mouse lung development. Am J Respir Crit
Care Med, 178: 721-728. 156743
McConnell R; Berhane K; Gilliland F; London SJ; Vora H; Avol E; Gauderman WJ; Margolis HG; Lurmann F; Thomas
DC; Peters JM. (1999). Air pollution and bronchitic symptoms in southern California children with asthma. Environ
Health Perspect, 107: 757-760. 007028
McConnell R; Berhane K; Gilliland F; Molitor J; Thomas D; Lurmann F; Avol E; Gauderman WJ; Peters JM. (2003).
Prospective study of air pollution and bronchitic symptoms in children with asthma. Am J Respir Crit Care Med,
168: 790-797. 049490
McDonald JD; Reed MD; Campen MJ; Barrett EG; Seagrave J; Mauderly JL. (2007). Health effects of inhaled gasoline
engine emissions. Inhal Toxicol, 19 Suppl 1: 107-116. 156746
McDonnell WF; Nishino-Ishikawa N; Petersen FF; Chen LH; Abbey DE. (2000). Relationships of mortality with the fine
and coarse fractions of long-term ambient PM10 concentrations in nonsmokers. J Expo Sci Environ Epidemiol, 10:
427-436. 010319
Medeiros A; Gouveia N. (2005). Relacao entre baixo peso ao nascer e a poluicao do ar no Municipio de Sao Paulo
[Relationship between low birthweight and air pollution in the city of Sao Paulo, Brazil]. Rev Saude Publica, 39:
965-972. 089824
December 2009 7-108
-------
Menache MG; Miller FJ; Raabe OG. (1995). Particle inhalability curves for humans and small laboratory animals. Ann
Occup Hyg, 39: 317-328. 006533
Meng YY; Wilhelm M; Rull RP; English P; Ritz B. (2007). Traffic and outdoor air pollution levels near residences and
poorly controlled asthma in adults. Ann Allergy Asthma Immunol, 98: 455-463. 093275
Miller KA; Siscovick DS; Sheppard L; Shepherd K; Sullivan JH; Anderson GL; Kaufman JD. (2007). Long-term exposure
to air pollution and incidence of cardiovascular events in women. N Engl J Med, 356: 447-458. 090130
Mogensen CE. (1984). Microalbuminuria predicts clinical proteinuria and early mortality in maturity-onset diabetes. N
Engl J Med, 310: 356-360. 156769
Mohallem SV; de Araujo Lobo DJ; Pesquero CR; Assuncao JV; de Andre PA; Saldiva PH; Dolhnikoff M. (2005).
Decreased fertility in mice exposed to environmental air pollution in the city of Sao Paulo. Environ Res, 98: 196-
202. 088657
Mollet NR; Cademartiri F; de Feyter PJ. (2005). Non-invasive multislice CT coronary imaging. Heart, 91: 401-407.
155988
Montauban van Swijndregt AD; De Lange EE; De Groot E; Ackerstaff RG (1999). An in vivo evaluation of the
reproducibility of intima-media thickness measurements of the carotid artery segments using B-mode ultrasound.
Ultrasound Med Biol, 25: 323-330. 156777
Morgenstern V; Zutavern A; Cyrys J; Brockow I; Koletzko S; Kramer U; Behrendt H; Herbarth O; von Berg A; Bauer CP;
Wichmann HE; Heinrich J. (2008). Atopic diseases, allergic sensitization, and exposure to traffic-related air
pollution in children. Am J Respir Grit Care Med, 177: 1331-1337. 156782
Mortimer K; Neugebauer R; Lurmann F; Alcorn S; Balmes J; Tager I. (2008). Early-Lifetime exposure to air pollution and
allergic sensitization in children with asthma. J Asthma, 45: 874-881. 187280
Naess O; Nafstad P; Aamodt G; Claussen B; Rosland P. (2007). Relation between concentration of air pollution and cause-
specific mortality: four-year exposures to nitrogen dioxide and particulate matter pollutants in 470 neighborhoods
in Oslo, Norway. Am J Epidemiol, 165: 435-443. 090736
National Kidney Foundation. (2008). Chronic Kidney Disease: Evaluation, Classification, and Stratification. The National
Kidney Foundation . Kidney Disease Outcomes Quality Initiative. 156796
Newman AB; Sutton-Tyrrell K; VogtMT; Kuller LH. (1993). Morbidity and mortality in hypertensive adults with a low
ankle/arm blood pressure index. JAMA, 270: 487-489. 156805
Nordling E; Berglind N; Melen E; Emenius G; Hallberg J; Nyberg F; Pershagen G; Svartengren M; Wickman M; Bellander
T. (2008). Traffic-related air pollution and childhood respiratory symptoms, function and allergies. Epidemiology,
19:401-8.097998
O'Leary DH; Polak JF; Kronmal RA; Kittner SJ; Bond MG; Wolfson SK Jr; Bommer W; Price TR; Gardin JM; Savage PJ.
(1992). Distribution and correlates of sonographically detected carotid artery disease in the Cardiovascular Health
Study. The CHS Collaborative Research Group. Stroke, 23: 1752-1760. 156825
O'Leary DH; Polak JF; Kronmal RA; Manolio TA; Burke GL; Wolfson SK Jr. (1999). Carotid-artery intima and media
thickness as a risk factor for myocardial infarction and stroke in older adults. Cardiovascular Health Study
Collaborative Research Group. N Engl J Med, 340: 14-22. 156826
O'Neill MS; Diez-Roux AV; Auchincloss AH; Franklin TG; Jacobs Jnr DR; Astor BC; Dvonch JT; Kaufman J. (2007).
Airborne particulate matter exposure and urinary albumin excretion: The Multi-Ethnic Study of Atherosclerosis.
Occup Environ Med, 65: 534-540. 156006
O'Rourke RA; Brundage BH; Froelicher VF; Greenland P; Grundy SM; Hachamovitch R; Pohost GM; Shaw LJ; Weintraub
WS; Winters WL; Forrester JS; Douglas PS; Faxon DP; Fisher JD; Gregoratos G; Hochman JS; Hutter AM; Kaul S;
Wo Ik MJ. (2000). American College of Cardiology/American Heart Association Expert Consensus document on
electron-beam computed tomography for the diagnosis and prognosis of coronary artery disease. Circulation, 102:
126-140. 192159
Oei HH; Vliegenthart R; Hak AE; Iglesias del Sol A; Hofman A; Oudkerk M; Witteman JC. (2002). The association
between coronary calcification assessed by electron beam computed tomography and measures of extracoronary
atherosclerosis: the Rotterdam Coronary Calcification Study. J Am Coll Cardiol, 39: 1745-1751. 156820
December 2009 7-109
-------
Oftedal B; Brunekreef B; Nystad W; Madsen C; Walker S-E; Nafstad P. (2008). Residential outdoor air pollution and lung
function in schoolchildren. Epidemiology, 19: 129-137. 093202
Oftedal B; Brunekreef B; Nystad W; Nafstad P. (2007). Residential outdoor air pollution and allergen sensitization in
schoolchildren in Oslo, Norway. Clin Exp Allergy, 37: 1632. 191948
Oh S-M; Chung K-H. (2006). Identification of mammalian cell genotoxins in respirable diesel exhaust particles by
bioassay-directed chemical analysis. Toxicol Lett, 161: 226-235. 088296
Ono N; Oshio S; Niwata Y; Yoshida S; Tsukue N; Sugawara I; Takano H; Takeda K. (2007). Prenatal exposure to diesel
exhaust impairs mouse spermatogenesis. Inhal Toxicol, 19: 275-281. 156007
Ozkaynak H; Thurston GD. (1987). Associations between 1980 US mortality rates and alternative measures of airborne
particle concentration. Risk Anal, 7: 449-461. 072960
Palli D; Saieva C; Munnia A; Peluso M; Grechi D; Zanna I; Caini S; Decarli A; Sera F; Masala G (2008). DNA adducts
and PM10 exposure in traffic-exposed workers and urban residents from the EPIC-Florence City study. Sci Total
Environ, 403: 105-112. 156837
Parker JD; Mendola P; Woodruff TJ. (2008). Preterm birth after the Utah Valley Steel Mill closure: a natural experiment.
Epidemiology, 19: 820-823. 156013
Parker JD; Woodruff TJ. (2008). Influences of study design and location on the relationship between particulate matter air
pollution and birthweight. Paediatr Perinat Epidemiol, 22: 214-227. 156846
Parker JD; Woodruff TJ; Basu R; Schoendorf KG. (2005). Air pollution and birth weight among term infants in California.
Pediatrics, 115: 121-128. 087462
Pedersen M; Vinzents P; Petersen JH; Kleinjans JC; Plas G; Kirsch-Volders M; Dostal M; Rossner P; Beskid O; Sram RJ;
Merlo DF; Knudsen LE. (2006). Cytogenetic effects in children and mothers exposed to air pollution assessed by
the frequency of micronuclei and fluorescence in situ hybridization (FISH): a family pilot study in the Czech
Republic. Mutat Res Fund Mol Mech Mutagen, 608: 112-120. 156848
Penard-Morand C; Charpin D; Raherison C; Kopferschmitt C; Caillaud D; Lavaud F; Annesi-Maesano I. (2005). Long-
term exposure to background air pollution related to respiratory and allergic health in schoolchildren. Clin Exp
Allergy, 35: 1279-1287. 087951
Penna MLF; Duchiade MP. (1991). Air pollution and infant mortality from pneumonia in the Rio de Janeiro metropolitan
area. Bull Pan Am Health Organ, 25: 47-54. 073325
Pereira LAA; Loomis D; Conceicao GMS; Braga ALF; Areas RM; Kishi HS; Singer JM; Bohm GM; Saldiva PHN. (1998).
Association between air pollution and intrauterine mortality in Sao Paulo, Brazil. Environ Health Perspect, 106:
325-329. 007264
Peters A; Skorkovsky J; Kotesovec F; Brynda J; Spix C; Wichmann HE; Heinrich J. (2000). Associations between
mortality and air pollution in central Europe. Environ Health Perspect, 108: 283-287. 001756
Pierse N; Rushton L; Harris RS; Kuehni CE; Silverman M; Grigg J. (2006). Locally-generated particulate pollution and
respiratory symptoms in young children. Thorax, 61: 216-220. 088757
Pignoli P; Tremoli E; Poli A; Oreste P; Paoletti R. (1986). Intimal plus medial thickness of the arterial wall: a direct
measurement with ultrasound imaging. Circulation, 74: 1399-1406. 156026
Pinkerton KE; Zhou Y; Teague SV; Peake JL; Walther RC; Kennedy IM; Leppert VJ; Aust AE. (2004). Reduced lung cell
proliferation following short-term exposure to ultrafine soot and iron particles in neonatal rats: key to impaired lung
growth?. Inhal Toxicol, 1: 73-81. 087465
Pires-Neto RC; Lichtenfels AJ; Scares SR; Macchione M; Saldiva PHN; Dolhnikoff M. (2006). Effects of Sao Paulo air
pollution on the upper airways of mice. Environ Res, 101: 356-361. 096734
Poma A; Limongi T; Pisani C; Granato V; Picozzi P. (2006). Genotoxicity induced by fine urban air particulate matter in
the macrophages cell line RAW 264.7. Toxicol In Vitro, 20: 1023-1029. 096903
Pope CA III. (1989). Respiratory disease associated with community air pollution and a steel mill, Utah Valley. Am J
Public Health, 79: 623-628. 044461
Pope CAIII; Burnett RT. (2007). Confounding in air pollution epidemiology: the broader context. Epidemiology, 18: 424-
426. 090928
December 2009 7-110
-------
Pope CAIII; Burnett RT; Thun MJ; Calle EE; Krewski D; Ito K; Thurston GD. (2002). Lung cancer, cardiopulmonary
mortality, and long-term exposure to fine particulate air pollution. JAMA, 287: 1132-1141. 024689
Pope CAIII; Thun MJ; Namboodiri MM; Dockery DW; Evans JS; Speizer FE; Heath CW Jr. (1995). Particulate air
pollution as a predictor of mortality in a prospective study of US adults. Am J Respir Crit Care Med, 151: 669-674.
045159
Pope III CA; Burnett RT; Thurston GD; Thun MJ; Calle EE; Krewski D; Godleski JJ. (2004). Cardiovascular mortality and
long-term exposure to particulate air pollution: epidemiological evidence of general pathophysiological pathways of
disease. Circulation, 109: 71-77. 055880
Pope III CA; Ezzati M; Dockery DW. (2009). Fine-particulate air pollution and life expectancy in the United States. N Engl
JMed, 360: 376-386. 190107
Puett RC; Schwartz J; Hart JE; Yanosky JD; Speizer FE; Suh H; Paciorek CJ; Neas LM; Laden F. (2008). Chronic
particulate exposure, mortality, and coronary heart disease in the nurses' health study. Am J Epidemiol, 168: 1 Al-
lies. L56891
Raizenne M; Neas LM; Damokosh AI; Dockery DW; Spengler JD; Koutrakis P; Ware JH; Speizer FE. (1996). Health
effects of acid aerosols on North American children: pulmonary function. Environ Health Perspect, 104: 506-514.
077268
Ramos C; Cisneros J; Gonzalez-Avila G; Becerril C; Ruiz V; Montano M. (2009). Increase of matrix metalloproteinases in
woodsmoke-induced lung emphysema in guinea pigs. Inhal Toxicol, 21: 119-132. 190116
Reed MD; Barrett EG; Campen MJ; Divine KK; Gigliotti AP; McDonald JD; Seagrave JC; Mauderly JL; Seilkop SK;
Swenberg JA. (2008). Health effects of subchronic inhalation exposure to gasoline engine exhaust. Inhal Toxicol,
20: 1125-1143. 156903
Reed MD; Campen MJ; Gigliotti AP; Harrod KS; McDonald JD; Seagrave JC; Mauderly JL; Seilkop SK. (2006). Health
effects of subchronic exposure to environmental levels of hardwood smoke. Inhal Toxicol, 18: 523-539. 156043
Reed MD; Gigliotti AP; McDonald JD; Seagrave JC; Seilkop SK; Mauderly JL. (2004). Health effects of subchronic
exposure to environmental levels of diesel exhaust. Inhal Toxicol, 16: 177-193. 055625
Resnick HE; Lindsay RS; McDermott MM; Devereux RB; Jones KL; Fabsitz RR; Howard BV. (2004). Relationship of
high and low ankle brachial index to all-cause and cardiovascular disease mortality: the Strong Heart Study.
Circulation, 109: 733-739. 156048
Reymao MSF; Cury PM; Lichtenfels AJFC; Lemos M; Battlehner CN; Conceicao GMS; Capelozzi VL; Monies GS; Junior
MF; Martins MA; Bohm GM; Saldiva PHN. (1997). Urban air pollution enhances the formation of urethane-
induced lung tumors in mice. Environ Res, 74: 150-158. 084653
Rich DQ; Demissie K; Lu SE; Kamat L; Wartenberg D; Rhoads GG (2009). Ambient air pollutant concentrations during
pregnancy and the risk of fetal growth restriction. J Epidemiol Community Health, 63: 488-496. 180122
Ritz B; Wilhelm M. (2008). Ambient air pollution and adverse birth outcomes: methodologic issues in an emerging field.
Basic ApplEcol, 102: 182-190. 156914
Ritz B; Wilhelm M; Hoggatt KJ; Ghosh JK. (2007). Ambient air pollution and preterm birth in the environment and
pregnancy outcomes study at the University of California, Los Angeles. Am JEpidemiol, 166: 1045-1052. 096146
Ritz B; Wilhelm M; Zhao Y. (2006). Air pollution and infant death in southern California, 1989-2000. Pediatrics, 118: 493-
502. 089819
Ritz B; Yu F; Chapa G; Fruin S. (2000). Effect of air pollution on preterm birth among children born in Southern California
between 1989 and 1993. Epidemiology, 11: 502-511. 012068
Ritz B; Yu F; Fruin S; Chapa G; Shaw GM; Harris JA. (2002). Ambient air pollution and risk of birth defects in Southern
California. Am J Epidemiol, 155: 17-25. 023227
Rocha ESIR; Lichtenfels AJ; Amador Pereira LA; Saldiva PH. (2008). Effects of ambient levels of air pollution generated
by traffic on birth and placental weights in mice. Fertil Steril, 90: 1921-1924. 096685
Rogers JF; Dunlop AL. (2006). Air pollution and very low birth weight infants: a target population?. Pediatrics, 118: 156-
164. 091232
December 2009 7-111
-------
Rojas-Martinez R; Perez-Padilla R; Olaiz-Fernandez G; Mendoza-Alvarado L; Moreno-Macias H; Fortoul T; McDonnell
W; Loomis D; Romieu I. (2007). Lung function growth in children with long-term exposure to air pollutants in
Mexico City. Am J Respir Grit Care Med, 176: 377-384. 091064
Roman HA; Walker KD; Walsh TL; Conner L; Richmond HM; Hubbell BJ; Kinney PL. (2008). Expert judgment
assessment of the mortality impact of changes in ambient fine particulate matter in the U. S. Environ Sci Technol,
42: 2268-2274. 156921
Romieu I; Garcia-Esteban R; Sunyer J; Rios C; Alcaraz-Zubeldia M; Velasco SR; Holguin F. (2008). The effect of
supplementation with omega-3 polyunsaturated fatty acids on markers of oxidative stress in elderly exposed to
PM(2.5). Environ Health Perspect, 116: 1237-1242. 156922
Romieu I; Ramirez-Aguilar M; Moreno-Macias H; Barraza-Villarreal A; Miller P; Hernandez-Cadena L; Carbajal-Arroyo
LA; Hernandez-Avila M. (2004). Infant mortality and air pollution: modifying effect by social class. J Occup
Environ Hyg, 46: 1210-1216. 093074
Roosli M; Braun-Fahrlander C; Kunzli N; Oglesby L; Theis G; Camenzind M; Mathys P; Staehelin J. (2000). Spatial
variability of different fractions of particulate matter within an urban environment and between urban and rural
sites. J Air Waste Manag Assoc, 50: 1115-1124. 010296
Roosli M; Kunzli N; Braun-Fahrlander C; Egger M. (2005). Years of life lost attributable to air pollution in Switzerland:
dynamic exposure-response model. Int J Epidemiol, 34: 1029-1035. 156923
Roosli M; Theis G; Kunzli N; Staehelin J; Mathys P; Oglesby L; Camenzind M; Braun-Fahrlander C. (2001). Temporal
and spatial variation of the chemical composition of PM10 at urban and rural sites in the Basel area, Switzerland.
Atmos Environ, 35: 3701-3713. 108738
Rosenlund M; Bellander T; Nordquist T; Alfredsson L. (2009). Traffic-generated air pollution and myocardial infarction.
Epidemiology, 20: 265-71. 190309
Rosenlund M; Berglind N; Pershagen G; Hallqvist J; Jonson T; Bellander T. (2006). Long-term exposure to urban air
pollution and myocardial infarction. Epidemiology, 17: 383-390. 089796
Ross R. (1999). Atherosclerosis-an inflammatory disease. N Engl J Med, 340: 115-126. 156926
Rubes J; Selevan SG; Evenson DP; Zudova D; Vozdova M; Zudova Z; Robbins WA; Perreault SD. (2005). Episodic air
pollution is associated with increased DNA fragmentation in human sperm without other changes in semen quality.
Hum Reprod, 20: 2776-2783. 078091
Ruggenenti P; Remuzzi G. (2006). Time to abandon microalbuminuria?. Kidney Int, 70: 1214-1222. 156933
Sagiv SK; Mendola P; Loomis D; Herring AH; Neas LM; Savitz DA; Poole C. (2005). A time-series analysis of air
pollution and preterm birth in Pennsylvania, 1997-2001. Environ Health Perspect, 113: 602-606. 087468
Salam MT; Millstein J; Li Y-F; Lurmann FW; Margolis HG; Gilliland FD. (2005). Birth outcomes and prenatal exposure to
ozone, carbon monoxide, and particulate matter: results from the Children's Health Study. Environ Health Perspect,
113: 1638-1644. 087885
Salonen JT; SalonenR. (1991). Ultrasonographically assessed carotid morphology and the risk of coronary heart disease.
Arterioscler Thromb Vase Biol, 11: 1245-1249. 156938
Sato H; Suzuki Kazuo T; Sone H; Yamano Y; Kagawa J; Aoki Y (2003). DNA-adduct formation in lungs, nasal mucosa,
and livers of rats exposed to urban roadside air in Kawasaki City, Japan. Environ Res, 93: 36-44. 096615
Schatz M; Zeiger RS; Hoffman CP (1990). Intrauterine growth is related to gestational pulmonary function in pregnant
asthmatic women. Kaiser-Permanente Asthma and Pregnancy Study Group. Chest, 98: 389-392. 156073
Schikowski T; Sugiri D; Ranft U; Gehring U; Heinrich J; Wichmann HE; Kramer U. (2005). Long-term air pollution
exposure and living close to busy roads are associated with COPD in women. Respir Res, 22: 152-161. 088637
Schindler C; Keidel D; Gerbase MW; Zemp E; Bettschart R; Brandli O; Brutsche MH; Burdet L; Karrer W; Knopfli B;
Pons M; Rapp R; Bayer-Oglesby L; Kunzli N; Schwartz J; Liu L-JS; Ackermann-Liebrich U; Rochat T; the
SAPALDIA Team. (2009). Improvements in PM10 exposure and reduced rates of respiratory symptoms in a cohort
of swiss adults (SAPALDIA). Am J Respir Grit Care Med, 179: 579-587. 191950
Schwartz J; Coull B; Laden F; Ryan L. (2008). The effect of dose and timing of dose on the association between airborne
particles and survival. Environ Health Perspect, 116: 64-69. 156963
December 2009 7-112
-------
Seagrave J; McDonald JD; Reed MD; Seilkop SK; Mauderly JL. (2005). Responses to subchronic inhalation of low
concentrations of diesel exhaust and hardwood smoke measured in rat bronchoalveolar lavage fluid. Inhal Toxicol,
17: 657-670. 088000
Selevan SG; Borkovec L; Slott VL; Zudova Z; Rubes J; Evenson DP; Perreault SD. (2000). Semen quality and
reproductive health of young Czech men exposed to seasonal air pollution. Environ Health Perspect, 108: 887-894.
012578
Sevastyanova O; Binkova B; Topinka J; Sram RJ; Kalina I; Popov T; Novakova Z; Farmer PB. (2007). In vitro
genotoxicity of PAH mixtures and organic extract from urban air particles part II: human cell lines. Mutat Res Fund
Mol Mech Mutagen, 620: 123-134. 156969
Sharma AK; Jensen KA; Rank J; White PA; Lundstedt S; Gagne R; Jacobsen NR; Kristiansen J; Vogel U; Wallin H. (2007).
Genotoxicity, inflammation and physico-chemical properties of fine particle samples from an incineration energy
plant and urban air. Mutat Res Fund Mol Mech Mutagen, 633: 95-111. 156975
Shaw LJ; Raggi P; Schisterman E; Berman DS; Callister TQ. (2003). Prognostic value of cardiac risk factors and coronary
artery calcium screening for all-cause mortality. Radiology, 228: 826-833. 156083
Shemesh J; Tenenbaum A; Fisman EZ; Apter S; Rath S; Rozenman J; Itzchak Y; Motro M. (1996). Absence of coronary
calcification on double-helical CT scans: predictor of angiographically normal coronary arteries in elderly women?.
Radiology, 199: 665-668. 156085
Shinkura R; Fujiyama C; Akiba S. (1999). Relationship between ambient sulfur dioxide levels and neonatal mortality near
the Mt Sakurajima volcano in Japan. J Epidemiol, 9: 344-349. 090050
Slama R; Darrow L; Parker J; Woodruff TJ; Strickland M; Nieuwenhuijsen M; Glinianaia S; Hoggatt KJ; Kannan S; Hurley
F; Kalinka J; Sram R; Brauer M; Wilhelm M; Heinrich J; Ritz B. (2008). Meeting report: atmospheric pollution and
human reproduction. Environ Health Perspect, 116: 791-798. 156985
Slama R; Morgenstern V; Cyrys J; Zutavern A; Herbarth O; Wichmann HE; Heinrich J; LISA Study Group. (2007). Traffic-
related atmospheric pollutants levels during pregnancy and offspring's term birth weight: a study relying on a land-
use regression exposure model. Environ Health Perspect, 115: 1283-1292. 093216
Smilde TJ; Wollersheim H; Van Langen H; Stalenhoef AF. (1997). Reproducibility of ultrasonographic measurements of
different carotid and femoral artery segments in healthy subjects and in patients with increased intima-media
thickness. Clin Sci (Lond), 93: 317-324. 156988
Sokol RZ; Kraft P; Fowler IM; Mamet R; Kim E; Berhane KT. (2006). Exposure to environmental ozone alters semen
quality. Environ Health Perspect, 114: 360-5. 098539
Solomon P; Baumann K; Edgerton E; Tanner R; Eatough D; Modey W; Marin H; Savoie D; Natarajan S; Meyer MB.
(2003). Comparison of integrated samplers for mass and composition during the 1999 Atlanta supersites project. J
Geophys Res, 108: 8423. 156994
Somers CM; McCarry BE; Malek F; Quinn JS. (2004). Reduction of particulate air pollution lowers the risk of heritable
mutations in mice. Science, 304: 1008-1010. 078098
Somers CM; Yauk CL; White PA; Parfett CLJ; Quinn JS. (2002). Air pollution induces heritable DNA mutations. PNAS,
99: 15904-15907.078100
Song CL; Zhou YC; Huang RJ; Wang YQ; Huang QF; Lu G; Liu KM. (2007). Influence of ethanol-diesel blended fuels on
diesel exhaust emissions and mutagenic and genotoxic activities of particulate extracts. J Hazard Mater, 149: 355-
363. 155306
Sorensen M; Schins RPF; Hertel O; Loft S. (2005). Transition metals in personal samples of PM25 and oxidative stress in
human volunteers. Cancer Epidemiol Biomarkers Prev, 14: 1340-1343. 083053
Sram R; Beskid O; Binkova B; Chvatalova I; Lnenickova Z; Milcova A; Solansky I; Tulupova E; Bavorova H; Ocadlikova
D. (2007). Chromosomal aberrations in environmentally exposed population in relation to metabolic and DNA
repair genes polymorphisms. Mutat Res Fund Mol Mech Mutagen, 620: 22-33. 188457
Sram RJ; Beskid O; Rossnerova A; Rossner P; Lnenickova Z; Milcova A; Solansky I; Binkova B. (2007). Environmental
exposure to carcinogenic polycyclic aromatic hydrocarbons—the interpretation of cytogenetic analysis by FISH.
Toxicol Lett, 172: 12-20. 192084
December 2009 7-113
-------
Sram RJ; Binkova B; Dejmek J; Bobak M. (2005). Ambient air pollution and pregnancy outcomes: a review of the
literature. Environ Health Perspect, 113: 375-382. 087442
Stanojevic S; Wade A; Stocks J; Hankinson J; Coates AL; Pan H; Rosenthal M; Corey M; Lebecque P; Cole TJ. (2008).
Reference ranges for spirometry across all ages: a new approach. Am J Respir Crit Care Med, 177: 253-260.
157007
Stacker R; Keaney Jr IF. (2004). Role of oxidative modifications in atherosclerosis. Physiol Rev, 84: 1381-1478. 157013
Strom KA; Garg BD; Johnson JT; D'Arcy JB; Smiler KL. (1990). Inhaled particle retention in rats receiving low exposures
of diesel exhaust. J Toxicol Environ Health, 29: 377-398. 157020
Sugamata M; Ihara T; Takano H; Oshio S; Takeda K. (2006). Maternal diesel exhaust exposure damages newborn murine
brains. Eisei Kagaku, 52: 82-84. 097166
Sugiri D; Ranft U; Schikowski T; Kramer U. (2006). The influence of large-scale airborne particle decline and traffic-
related exposure on children's lung function. Environ Health Perspect, 114: 282-288. 088760
Suh YJ; Ha EH; Park H; Kim YJ; Kim H; Hong YC. (2008). GSTM1 polymorphism along with PM10 exposure contributes
to the risk of preterm delivery. Mutat Res, 656: 62-67. 192077
Sun Q; Wang A; Jin X; Natanzon A; Duquaine D; Brook RD; Aguinaldo JG; Fayad ZA; Fuster V; Lippmann M; Chen LC;
Rajagopalan S. (2005). Long-term air pollution exposure and acceleration of atherosclerosis and vascular
inflammation in an animal model. JAMA, 294: 3003-3010. 087952
Sun Q; Yue P; Deiuliis JA; Lumeng CN; Kampfrath T; Mikolaj MB; Cai Y; Ostrowski MC; Lu B; Parthasarathy S; Brook
RD; Moffatt-Bruce SD; Chen LC; Rajagopalan S. (2009). Ambient air pollution exaggerates adipose inflammation
and insulin resistance in a mouse model of diet-induced obesity. Circulation, 119: 538-546. 190487
Sun Q; Yue P; Kirk RI; Wang A; Moatti D; Jin X; Lu B; Schecter AD; Lippmann M; Gordon T; Chen LC; Rajagopalan S.
(2008). Ambient air particulate matter exposure and tissue factor expression in atherosclerosis. Inhal Toxicol, 20:
127-137. 157033
Sun Q; Yue P; Ying Z; Cardounel AJ; Brook RD; Devlin R; Hwang JS; Zweier JL; Chen LC; Rajagopalan S. (2008). Air
Pollution Exposure Potentiates Hypertension Through Reactive Oxygen Species-Mediated Activation of
Rho/ROCK. Arterioscler Thromb Vase Biol, 28: 1760-1766. 157032
Suwa T; Hogg JC; Quinlan KB; Ohgami A; Vincent R; Van Eeden SF. (2002). Particulate air pollution induces progression
of atherosclerosis. J Am Coll Cardiol, 39: 935-942. 028588
Tarantini L; Bonzini M; Apostoli P; Pegoraro V; Bollati V; Marinelli B; Cantone L; Rizzo G; Hou L; Schwartz J; Bertazzi
PA; Baccarelli A. (2009). Effects of particulate matter on genomic DNA methylation content and iNOS promoter
methylation. Environ Health Perspect, 117: 217-222. 192010
Tesfaigzi Y; McDonald JD; Reed MD; Singh SP; De Sanctis GT; Eynott PR; Hahn FF; Campen MJ; Mauderly JL. (2005).
Low-level subchronic exposure to wood smoke exacerbates inflammatory responses in allergic rats. Toxicol Sci,
88: 505-513. 156116
Tesfaigzi Y; Singh SP; Foster JE; Kubatko J; Barr EB; Fine PM; McDonald JD; Hahn FF; Mauderly JL. (2002). Health
effects of subchronic exposure to low levels of wood smoke in rats. Toxicol Sci, 65: 115-125. 025575
Thurlbeck WM. (1982). Postnatal human lung growth. Thorax, 37: 564-571.093260
Tokiwa H; Sera N; Nakanishi Y. (2005). Involvement of alveolar macrophages in the formation of 8-oxodeoxyguanosine
associated with exogenous particles in human lungs. Inhal Toxicol, 17: 577-585. 191952
Tong S; Colditz P. (2004). Air pollution and sudden infant death syndrome: a literature review. Paediatr Perinat Epidemiol,
18: 327-335. 087883
Tovalin H; Valverde M; Morandi MT; Blanco S; Whitehead L; Rojas E. (2006). DNA damage in outdoor workers
occupationally exposed to environmental air pollutants. Occup Environ Med, 63: 230-236. 091322
Tozuka Y; Watanabe N; Ohsawa M; Toriba A; Kizu R; Hayakawa K. (2004). Transfer of poly cyclic aromatic hydrocarbons
to fetuses and breast milk of rats exposed to diesel exhaust. Eisei Kagaku, 50: 497-502. 090864
Tsai S-S; Chen C-C; Hsieh H-J; Chang C-C; Yang C-Y (2006). Air pollution and postneonatal mortality in a tropical city:
Kaohsiung, Taiwan. Inhal Toxicol, 18: 185-189. 090709
December 2009 7-114
-------
Tsukue N; Tsubone H; Suzuki AK. (2002). Diesel exhaust affects the abnormal delivery in pregnant mice and the growth of
their young. Inhal Toxicol, 14: 635-651.030593
Tsukue N; Yoshida S; Sugawara I; Taked K. (2004). Effect of diesel exhaust on development of fetal reproductive function
in ICR female mice. Eisei Kagaku, 50: 174-80. 096643
U.S. EPA. (1996). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research Triangle
Park, NC. EPA/600/P-95/001aF-cF. 079380
U.S. EPA. (2002). Health assessment document for diesel engine exhaust. U.S. Environmental Protection Agency.
Washington, DC. 042866
U.S. EPA. (2004). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research Triangle
Park, NC. EPA/600/P-99/002aF-bF. 056905
U. S. EPA. (2005). Guidelines for carcinogen risk assessment, Risk Assessment Forum Report. U. S. Environmental
Protection Agency. Washington, DC. EPA/630/P-03/001F. http://cfpub.epa.gov/ncea/index.cfm. 086237
U. S. EPA. (2006). Provisional Assessment of Recent Studies on Health Effects of Particulate Matter Exposure. U. S.
Environmental Protection Agency. Research Triangle Park, NC. 157071
van der Meer IM; Bots ML; Hofman A; del Sol AI; van der Kuip DA; Witteman JC. (2004). Predictive value of
noninvasive measures of atherosclerosis for incident myocardial infarction: the Rotterdam Study. Circulation, 109:
1089-1094. 156129
Van Hee VC; Adar SD; Szpiro AA; Barr RG; Bluemke DA; Diez Roux AV; Gill EA; Sheppard L; Kaufman JD. (2009).
Exposure to traffic and left ventricular mass and function: the Multi-Ethnic Study of Atherosclerosis. Am J Respir
Grit Care Med, 179: 827-834. 192110
Veras MM; Damaceno-Rodrigues NR; Caldini EG; Maciel Ribeiro AA; Mayhew TM; Saldiva PH; Dolhnikoff M. (2008).
Particulate urban air pollution affects the functional morphology of mouse placenta. Biol Reprod, 79: 578-584.
190493
Veras MM; Damaceno-Rodrigues NR; Guimaraes Silva RM; Scoriza JN; Saldiva PH; Caldini EG; Dolhnikoff M. (2009).
Chronic exposure to fine particulate matter emitted by traffic affects reproductive and fetal outcomes in mice.
Environ Res, 109: 536-543. 190496
Villeneuve PJ; Goldberg MS; Krewski D; Burnett RT; Chen Y. (2002). Fine particulate air pollution and all-cause mortality
within the Harvard six-cities study: variations in risk by period of exposure. Ann Epidemiol, 12: 568-576. 042576
Vineis P; Hoek G; Krzyzanowski M; Vigna-Taglianti F; Veglia F; Airoldi L; Autrup H; Dunning A; Garte S; Hainaut P;
Malaveille C; Matullo G; Overvad K; Raaschou-Nielsen O; Clavel-Chapelon F; Linseisen J; Boeing H;
Trichopoulou A; Palli D; Peluso M; Krogh V; Tumino R; Panico S; Bueno-De-Mesquita HB; Peeters PH; Lund EE;
Gonzalez CA; Martinez C; Dorronsoro M; Barricarte A; Cirera L; Quiros JR; Berglund G; Forsberg B; Day NE;
Key TJ; Saracci R; Kaaks R; Riboli E. (2006). Air pollution and risk of lung cancer in a prospective study in
Europe. Int J Cancer, 119: 169-174. 192089
Vinzents PS; Moller P; Sorensen M; Knudsen LE; Herte LQ; Jensen FP; Schibye B; Loft S. (2005). Personal exposure to
ultrafine particles and oxidative DNA damage. Environ Health Perspect, 113: 1485-1490. 087482
Vogt MT; Cauley JA; Newman AB; Kuller LH; Hulley SB. (1993). Decreased ankle/arm blood pressure index and
mortality in elderly women. JAMA, 270: 465-469. 157100
Wallenborn JG; Evansky P; Shannahan JH; Vallanat B; Ledbetter AD; Schladweiler MC; Richards JH; Gottipolu RR;
Nyska A; Kodavanti UP. (2008). Subchronic inhalation of zinc sulfate induces cardiac changes in healthy rats.
Toxicol Appl Pharmacol, 232: 69-77. 191171
Walsh CR; Cupples LA; Levy D; Kiel DP; Hannan M; Wilson PW; O'Donnell CJ. (2002). Abdominal aortic calcific
deposits are associated with increased risk for congestive heart failure: the Framingham Heart Study. Am Heart J,
144: 733-739. 157103
Watanabe N. (2005). Decreased number of sperms and Sertoli cells in mature rats exposed to diesel exhaust as fetuses.
Toxicol Lett, 155: 51-58. 087985
Wayne LG; Chambers LA. (1968). Biological effects of urban air pollution: V a study of effects of Los Angeles atmosphere
on laboratory rodents. Arch Environ Occup Health, 16: 871-885. 038537
December 2009 7-115
-------
Weisenberger DJ; Campan M; Long TI; Kim M; Woods C; Fiala E; Ehrlich M; Laird PW. (2005). Analysis of repetitive
element DNA methylation by MethyLight. Nucleic Acids Res, 33: 6823-36. 192101
Weitz JI; Byrne J; Clagett GP; Farkouh ME; Porter JM; Sackett DL; Strandness DE Jr; Taylor LM. (1996). Diagnosis and
treatment of chronic arterial insufficiency of the lower extremities: a critical review. Circulation, 94: 3026-3049.
156150
Wendelhag I; Gustavsson T; Suurkula M; Berglund G; Wikstrand J. (1991). Ultrasound measurement of wall thickness in
the carotid artery: fundamental principles and description of a computerized analysing system. Clinical Physiol, 11:
565-577. 157135
Wendelhag I; Wiklund O; Wikstrand J. (1993). Atherosclerotic changes in the femoral and carotid arteries in familial
hypercholesterolemia. Ultrasonographic assessment of intima-media thickness and plaque occurrence. Arterioscler
Thromb Vase Biol, 13: 1404-1411. 157136
Wheeler AJ; Villeneuve P; Smith-Doiron M; Mahmud M; Dales R; Brook JR. (2006). Children's exposure to ambient air
pollution: the application of different exposure assessment methodologies. Epidemiology, 17: S33-S34. 103905
Wilhelm M; Ritz B. (2005). Local variations in CO and particulate air pollution and adverse birth outcomes in Los Angeles
County, California, USA. Environ Health Perspect, 113: 1212-1221. 088668
Willekes C; Brands PJ; Willigers JM; Hoeks AP; Reneman RS. (1999). Assessment of local differences in intima-media
thickness in the human common carotid artery. J Vase Res, 36: 222-228. 157147
Williams L; Ulrich C; Larson T; Wener M; Wood B; Campbell P; Potter J; McTiernan A; De Roos A. (2009). Proximity to
Traffic, Inflammation, and Immune Function among Women in the Seattle, Washington, Area. Environ Health
Perspect 117: 373. 191945
Wilson PW; Kauppila LI; O'Donnell CJ; Kiel DP; Hannan M; Polak JM; Cupples LA. (2001). Abdominal aortic calcific
deposits are an important predictor of vascular morbidity and mortality. Circulation, 103: 1529-1534. 156159
Wilson WE; Mage DT; Grant LD. (2000). Estimating separately personal exposure to ambient and nonambient particulate
matter for epidemiology and risk assessment: why and how. J Air Waste Manag Assoc, 50: 1167-1183. 010288
Witteman JC; Kok FJ; van Saase JL; Valkenburg HA. (1986). Aortic calcification as a predictor of cardiovascular mortality.
Lancet, 2: 1120-1122. 156161
Woodruff T J; Darrow LA; Parker JD. (2008). Air pollution and postneonatal infant mortality in the United States, 1999-
2002. Environ Health Perspect, 116: 110-115. 098386
Woodruff TJ; Grillo J; Schoendorf KG. (1997). The relationship between selected causes of postneonatal infant mortality
and particulate air pollution in the United States. Environ Health Perspect, 105: 608-612. 084271
Woodruff TJ; Parker JD; Schoendorf KC. (2006). Fine particulate matter (PM25) air pollution and selected causes of
postneonatal infant mortality in California. Environ Health Perspect, 114: 785-790. 088758
Xu J; Lee ET; Devereux RB; Umans JG; Bella JN; Shara NM; Yeh J; Fabsitz RR; Howard BV (2008). A longitudinal study
of risk factors for incident albuminuria in diabetic American Indians: the Strong Heart Study. Am J Kidney Dis, 51:
415-424. 157157
Yang AS; Estecio MR; Doshi K; Kondo Y; Tajara EH; Issa JP. (2004). A simple method for estimating global DNA
methylation using bisulfite PCR of repetitive DNA elements. Nucleic Acids Res, 32: e38. 192102
Yang C-Y; Hsieh H-J; Tsai S-S; Wu T-N; Chiu H-F. (2006). Correlation between air pollution and postneonatal mortality in
a subtropical city: Taipei, Taiwan. J Toxicol Environ Health A, 69: 2033-2040. 090760
Yang C-Y; Tseng Y-T; Chang C-C. (2003). Effects of air pollution on birthweight among children born between 1995 and
1997 inKaohsiung, Taiwan. J Toxicol Environ Health A, 66: 807-816. 087886
Yatera K; Hsieh J; Hogg James C; Tranfield E; Suzuki H; Shih C-H; Behzad Ali R; Vincent R; van Eeden Stephan F.
(2008). Particulate matter air pollution exposure promotes recruitment of monocytes into atherosclerotic plaques.
Am J Physiol Heart Circ Physiol, 294: H944-H953. 157162
Yauk C; Polyzos A; Rowan-Carroll A; Somers CM; Godschalk RW; Van Schooten FJ; Berndt ML; Pogribny IP; Koturbash
I; Williams A; Douglas GR; Kovalchuk O. (2008). Germ-line mutations, DNA damage, and global
hypermethylation in mice exposed to particulate air pollution in an urban/industrial location. PNAS, 105: 605-610.
157164
December 2009 7-116
-------
Yauk CL; Quinn JS. (1996). Multilocus DNA fingerprinting reveals high rate of heritable genetic mutation in herring gulls
nesting in an industrialized urban site. PNAS, 93: 12137-12141. 089093
Ying Z; Kampfrath T; Thurston G; Farrar B; Lippmann M; Wang A; Sun Q; Chen LC; Rajagopalan S. (2009). Ambient
particulates alter vascular function through induction of reactive oxygen and nitrogen species. Toxicol Sci, 1: 1-36.
190111
Yokota S; Mizuo K; Moriya N; Oshio S; Sugawara I; Takeda K. (2009). Effect of prenatal exposure to diesel exhaust on
dopaminergic system in mice. Neurosci Lett, 449: 38-41. 190518
Yoshida S; Ono N; Tsukue N; Oshio S; Umeda T; Takano H; Takeda K. (2006). In utero exposure to diesel exhaust
increased accessory reproductive gland weight and serum testosterone concentration in male mice. Environ Sci J
Integr Environ Res, 13: 139-147. 097015
Yoshida S; Takedab K. (2004). The effects of diesel exhaust onmurine male reproductive function. Eisei Kagaku, 50: 210-
214. 097760
Yoshida S; Yoshida M; Sugawara I; Takeda K. (2006). Mice strain differences in effects of fetal exposure to diesel exhaust
gas on male gonadal differentiation. Environ Sci J Integr Environ Res, 13: 117-123. 156170
Zanobetti A; Bind MAC; Schwartz J. (2008). Particulate air pollution and survival in a COPD cohort. Environ Health
Perspect, 7: 48. 156177
Zanobetti A; Schwartz J. (2007). Particulate air pollution, progression, and survival after myocardial infarction. Environ
Health Perspect, 115: 769-775. 091247
Zeger S; Dominici F; McDermott A; Samet J. (2008). Mortality in the Medicare population and chronic exposure to fine
particulate air pollution in urban centers (2000-2005). Environ Health Perspect, 116: 1614-1619. 191951
Zeger S; McDermott A; Dominici F; Samet J. (2007). Mortality in the medicare population and chronic exposure to fine
particulate air pollution. Johns Hopkins University. Baltimore.http://www.bepress.com/jhubiostat/paperl33. 157176
Zeman KL; Bennett WD. (2006). Growth of the small airways and alveoli from childhood to the adult lung measured by
aerosol-derived airway morphometry. J Appl Physiol, 100: 965-971. 157178
Zhang Z; Che W; Liang Y; Wu M; Li N; Shu Y; Liu F; Wu D. (2007). Comparison of cytotoxicity and genotoxicity induced
by the extracts of methanol and gasoline engine exhausts. Toxicol In Vitro, 21: 1058-1065. 157186
December 2009 7-117
-------
Chapter 8. Populations Susceptible to
PM-related Health Effects
Interindividual variation in human responses to air pollutants indicates that some populations
are at increased risk for the detrimental effects of ambient exposure to an air pollutant (e.g., PM)
(Kleeberger and Ohtsuka, 2005, 130489). The NAAQS are intended to provide an adequate margin
of safety for both general populations and sensitive subgroups, or those individuals potentially at
increased risk for health effects in response to ambient air pollution (see Section 1.1). To facilitate
the identification of populations at the greatest risk for PM-related health effects, studies have
evaluated factors that contribute to the susceptibility and/or vulnerability of an individual to PM. The
definition for both of these terms has been found to vary across studies, but in most instances
susceptibility refers to biological or intrinsic factors (e.g., lifestage, gender) while vulnerability
refers to non-biological or extrinsic factors (e.g., socioeconomic status [SES]) (see Table 8-1).
Additionally, in some cases, the terms "at-risk" and sensitive populations have been used to
encompass these concepts more generally. However, in many cases a factor identified that increases
an individual's risk for morbidity or mortality effects from exposure to an air pollutant (e.g., PM) can
not be easily categorized as a susceptibility or vulnerability factor. For example, a population that is
characterized as having low SES, traditionally defined as a vulnerability factor, may have less access
to healthcare resulting in the manifestation of disease (i.e., a susceptibility factor), but they may also
reside in a location that results in exposure to higher concentrations of an air pollutant, increasing
their vulnerability. Therefore, the terms susceptibility and vulnerability are intertwined and at times
can not be distinguished from one another.
As a result of the inconsistencies in the definition of susceptibility and vulnerability presented
in the literature as well as the inability to clearly delineate whether an identified factor increases an
individual's susceptibility or vulnerability to an air pollutant, in this ISA, the term 'susceptible
population' will be used as a blanket term and defined as the following:
Populations that have a greater likelihood of experiencing health effects related to exposure to
an air pollutant (e.g., PM) due to a variety of factors including, but not limited to: genetic or
developmental factors, race, gender, lifestage, lifestyle (e.g., smoking status and nutrition) or
preexisting disease; as well as, population-level factors that can increase an individual's
exposure to an air pollutant (e.g., PM) such as socioeconomic status [SES], which
encompasses reduced access to health care, low educational attainment, residential location,
and other factors.
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009 8-1
-------
Table 8-1. Definitions of susceptible and vulnerable in the PM literature.
Definition
Reference
Susceptible: predisposed to develop a noninfectious disease
Vulnerable: capable of being hurt: susceptible to injury or disease
Merriam-Webster (2009, 1921461
Susceptible: greater likelihood of an adverse outcome given a specific exposure, in comparison with the general
population. Includes both host and environmental factors (e.g., genetics, diet, physiological state, age, gender,
social, economic, and geographic attributes).
Vulnerable: periods during an individual's life when they are more susceptible to environmental exposures.
American Lung Association (2001,0166261
Susceptible: innate (e.g., genetic or developmental) or acquired (e.g., age, disease or smoking or smoking) factors
that make individuals more likely to experience effects with exposure to PM.
Vulnerable: PM-related effects due to factors including socioeconomic status (e.g., reduced access to health care) or
particularly elevated exposure levels.
Susceptible: greater or lesser biological response to exposure.
Vulnerable: more or less exposed.
Vulnerable: to be susceptible to harm or neglect, that is, acts of commission or omission on the part of others that
can wound.
Susceptible:may be those who are significantly more liable than the general population to be affected by a stressor
due to life stage (e.g., children, the elderly, or pregnant women), genetic polymorphisms (e.g., the small but
significant percentage of the population who have genetic susceptibilities), prior immune reactions (e.g., individuals
who have been "sensitized" to a particular chemical), disease state (e.g., asthmatics), or prior damage to cells or
systems (e.g., individuals with damaged ear structures due to prior exposure to toluene, making them more sensitive
to damage by high noise levels).
Vulnerable: differential exposure and differential preparedness (e.g., immunization).
Susceptible: intrinsic (e.g., age, gender, pre-existing disease (e.g., asthma) and genetics) and extrinsic (previous
exposure and nutritional status) factors.
Susceptible:characteristics that contribute to increased risk of PM-related health effects (e.g., genetics, pre-existing
disease, age, gender, race, socioeconomic status, healthcare availability, educational attainment, and housing
characteristics).
U.S. EPA. (2008, 1570721
U.S. EPA (2009, 1921491
Adav, LA. (2001,1921501
U.S. EPA (2003, 1921451
Kleeberger and Ohtsuka (2005, 1304891
Pope and Dockery (2006,1568811
To examine whether air pollutants (e.g., PM) differentially affect certain populations,
epidemiologic studies conduct stratified analyses to identify the presence or absence of effect
modification. A thorough evaluation of potential effect modifiers may help identify populations that
are more susceptible to an air pollutant (e.g., PM). Although the design of toxicological and
controlled human exposure studies do not allow for an extensive examination of effect modifiers, the
use of animal models of disease and the study of individuals with underlying disease or genetic
polymorphisms do allow for comparisons between subgroups. Therefore, the results from these
studies, combined with those results obtained through stratified analyses in epidemiologic studies,
contribute to the overall weight of evidence for the increased susceptibility of specific populations to
an air pollutant (e.g., PM).
This chapter discusses the epidemiologic, controlled human exposure, and toxicological
studies evaluated in Chapters 6 and 7 that provide information on potentially susceptible
populations. The studies highlighted include only those studies that present stratified results (e.g.,
males vs. females or <65 vs. > 65). This approach allowed for a comparison between populations
exposed to similar PM concentrations and within the same study design. In addition, numerous
studies that focus on only one potentially susceptible population can provide supporting evidence on
whether a population is susceptible to PM exposure and are described in Chapters 6 and 7, but these
studies are not discussed in detail in this chapter. Table 8-2 provides an overview of the factors
examined in the current toxicological, controlled human exposure, and epidemiologic literature and
the direction of the underlying evidence in determining whether a factor increases the susceptibility
of a population to PM-related health effects.
December 2009
8-2
-------
Table 8-2. Susceptibility factors evaluated.
Factor Collective Evidence (+/-)2
Older Adults (> 65) +
Children (<18)1 +
Pregnancy and Developmental Effects +*
Gender
Race/Ethnicity
Genetic factors
- Genetic polymorphisms +
- Epigenetics
Cardiovascular Diseases +
Respiratory Illnesses +
Respiratory Contributions to Cardiovascular Effects
Diabetes +*
Obesity +*
Socioeconomic Status (SES) +
Health Status (e.g., Nutrition)3 +*
1 The age range that defines a child varies from study to study. In some cases it is <21 years old while in others it is <18
years old (Firestone et al., 2007,192071). For the purposes of this exercise children are defined as those individuals <18
years old because the majority of epidemiologic studies consider individuals under the age of 18 children.
2 This column identifies whether the "collective" evidence from studies evaluated found that a specific factor increased (+)
or did not increase (-) a population's susceptiblity to PM exposure (i.e., PM exposure to all size fractions combined). In
instances where only a few studies were evaluated for a speicifc factor it was not possible to clearly assign a (+) or (-) as a
result the direction of the preliminary evidence is identified along with (*) to represent that more information is warranted.
3 These factors are surrogates of socioeconomic status and are discussed within this subsection of the chapter.
8.1. Potentially Susceptible Populations
8.1.1. Lifestage
8.1.1.1. OlderAdults
Evidence for PM-related health effects in older adults spans epidemiologic, controlled human
exposure, and toxicological studies. The 2004 PM AQCD found evidence for increased risk of
cardiovascular effects in older adults exposed to PM (U.S. EPA, 2004, 056905). Older adults
represent a potentially susceptible population due to the higher prevalence of pre-existing
cardiovascular and respiratory diseases found in this age range compared to younger age groups. The
increased susceptibility in this population can primarily be attributed to the gradual decline in
physiological processes as part of the aging process (U.S. EPA, 2006, 192082). Therefore, some
overlap exists between potentially susceptible older adults and the population that encompases
individuals with pre-existing diseases (Kan et al., 2008, 156621). Epidemiologic studies that conduct
age stratified analyses primarily focus on the association between short-term exposure to PM and
cardiovascular morbidity, but additional studies have examined the association between PM and
respiratory morbidity and mortality.
December 2009 8-3
-------
In recent publications, the epidemiologic evidence for cardiovascular effects in older adults in
response to short-term exposure to PMi0_25 and PM25 is limited, but taken together with evidence
from studies of PMi0 (e.g., Larrieu et al., 2007, 093031; Le Tertre et al., 2002, 023746). supports the
increased risk of cardiovascular morbidity in older adults. Host et al. (2007, 155851) found an
increase in cardiovascular disease (CVD) hospital admissions in individuals >65 yr compared to all
ages for short-term exposure to both PMi0_2.5 and PM2 5. Barnett et al. (2006, 089770) analyzed data
from several cities across Australia and New Zealand and found that the excess risk of
hospitalizations for cardiac diseases, congestive heart failure (CHF), ischemic heart disease (IHD),
myocardial infarction (MI), and all CVD was greater among patients aged > 65 yr as compared to
those individuals <65 years in response to short-term exposure to PM2.5. U.S.- and Canadian-based
studies that examined the association between short-term exposure to PM and cardiovascular
morbidity primarily found no evidence for increased risk among older adults. Metzger et al. (2004,
044222) found no evidence of effect modification by age for cardiovascular outcomes and short-term
exposure to PM2 5 in Atlanta, Georgia, which is supported by the results from other studies that
focused on short-term exposure to PMi0 (Fung et al., 2005, 074322; Zanobetti and Schwartz, 2005,
088069). However, Pope et al. (2008, 191969) observed an increased risk of HF hospital admissions
in older adults (i.e., > 65 yr) in Utah, but the study used a 14-day lagged cumulative moving average
of PM2.5, which is much longer than the lags examined by the other U.S.- and Canadian-based
studies. Although studies have not consistently found an association between short-term exposure to
PM and respiratory-related health effects in older adults, some studies have reported an increase in
respiratory hospital admissions in individuals 65 years of age and older (e.g., Fung et al., 2005,
093262).
Additional evidence for an increase in cardiovascular and respiratory effects among older
adults has been observed in controlled human exposure and dosimetry studies. Devlin et al. (2003,
087348) found that older subjects exposed to PM2 5 concentrated ambient particles (CAPs)
experienced significant decreases in heart rate variability (HRV) (both in time and frequency)
immediately following exposure, when compared to healthy young subjects. In addition, Gong et al
(2004, 055628) reported that older subjects demonstrated significant decreases in HRV when
exposed to PM25 CAPs, but this study did not compare the response in older subjects to those
elicited by young, healthy individuals. However, the study did find that healthy older adults were
more susceptible to decreases in HRV compared to those with an underlying health condition (i.e.,
chronic obstructive pulmonary disease [COPD]) in response to PM exposure (Gong et al., 2004,
055628). Dosimetry studies have shown a depression of PM25 and PMi0_2.5 clearance in all regions of
the respiratory tract with increasing age beyond young adulthood in humans and laboratory animals.
These results suggest that older adults are also susceptible to PM-related respiratory health effects
(Section 4.3.4.1).
Animal toxicological studies have attempted to characterize the relationship between age and
PM-related health effects through the development of models that mimic the physiological
conditions associated with older individuals. For example, Nadziejko et al. (2004, 055632) observed
arrhythmias in older, but not younger, rats exposed to PM2 5 CAPs. In addition, another study
(Tankersley et al., 2004, 094378) that used a mouse model of terminal senescence demonstrated
altered baseline autonomic tone in response to carbon black exposure, which may subsequently
affect the quality and severity of cardiovascular responses (Tankersley et al., 2007, 097910).
Reductions in cardiac fractional shortening and significant pulmonary vascular congestion upon
exposure to carbon black were also reported in older mice (Tankersley et al., 2008, 157043). Overall,
these studies provide biological plausibility for the increase in cardiovascular effects in older adults
observed in the controlled human exposure and epidemiologic studies.
Recent epidemiologic studies have also found that individuals >65 years of age are more
susceptible to all-cause (nonaccidental) mortality upon short-term exposure to both PM2 5 (Ostro et
al., 2006, 087991) and PM10 (Samoli et al., 2008, 188455; Zeka et al., 2006, 088749). which is
consistent with the findings of the 2004 PM AQCD. Of note are the results from Ostro et al. (2006,
087991) that reported a slight increase in mortality for older adults compared to all ages in single-
pollutant models, but a robust effect estimate in co-pollutant models with gaseous pollutants
(i.e., PM25+CO and PM25+NO2). These results differ from those in the all ages model (i.e.,
attenuation of the effect estimate in co-pollutant models with CO and NO2), which suggests that
older adults are more susceptible to PM exposures, even though the age-stratified effect estimates in
single-pollutant models did not significantly differ. Epidemiologic studies that examined the
association between mortality and long-term exposure to PM (i.e., PM25) have found results
December 2009 8-4
-------
contradictory to those obtained in the short-term exposure studies. Villeneuve et al. (2002, 042576).
Naess et al. (2007, 090736) and Zeger et al. (2008, 191951) report evidence of differing PM2.5
relative risks by age, where risk declines with increasing age starting at age 60 until there is no
evidence of an association among persons > 85 yr.
The evidence from epidemiologic, controlled human exposure, and toxicological studies that
focused on exposures to PM2.5,PMi0_2.5, and PMi0, provide coherence and biological plausibility for
the association between PM and cardiovascular morbidity in older adults. The clear pattern of
positive associations only being observed in epidemiologic studies conducted in non-U.S. locations
brings into question the influence of PM composition on health effects. However, the difference in
effects observed between U.S. and Canadian, and international studies could also be due to possible
differences in the identification of CVD-related morbidity and mortality between the studies
evaluated. Although most studies examined the effect of PM on CVD outcomes in older adults, the
additional evidence from epidemiologic studies that focus on respiratory morbidity and mortality in
response to short-term exposure to PM also indicate that older adults represent a susceptible
population. As the demographics of the U.S. population shift over the next 20 years with a larger
percentage of the population (i.e., 13% of the population in 2011 and a projected 20% in 2030)
encompassing individuals > 65 yr (U.S. Census, 2000, 157064). an increase in the number of
PM-related health effects (e.g., cardiovascular and respiratory morbidity, and mortality) in
individuals > 65 years of age could occur.
8.1.1.2. Children
Children have generally been considered more susceptible to PM exposure due to multiple
factors including more time spent outdoors, greater activity levels, exposures resulting in higher
doses per body weight and lung surface area, and the potential for irreversible effects on the
developing lung (U.S. EPA, 2004, 056905). The 2004 PM AQCD found that studies which stratify
results by age typically report associations between PM and respiratory-related health effects in
children, specifically asthma (U.S. EPA, 2004, 056905). Of the recent epidemiologic studies
evaluated, only a few have examined the association between PMi0_2.5 and PM2.5 and respiratory
effects in children. Mar et al. (2004, 057309) found increased respiratory effects (e.g., wheeze,
cough, lower respiratory symptoms) in children 7-12 years of age compared to individuals 20-51
years of age in response to exposure to both PMi0_2.5 and PM2.5 in Spokane, Washingon. In addition,
Host et al. (2007, 155851) found an increase in respiratory-related hospital admissions with short-
term exposure to PMi0_2.s among children ages 0-14 yr in 6 French cities. An examination of studies
that also focused on PMi0 provide additional support for PM-induced respiratory effects in children
(Mar et al., 2004, 057309; Peel et al., 2005, 056305). A recent toxicological study provides
biological plausibility for the increase in PM-related respiratory effects in children observed in the
epidemiologic studies. Mauad et al. (2008, 156743) using both prenatal and postnatal mice exposed
to ambient PM2.5 in a "polluted chamber" found evidence for changes in lung function and
pulmonary injury (e.g., incomplete alveolarization). Additionally, Pinkerton et al. (2004, 087465;
2008, 190471) found evidence suggesting that the developing lung is more susceptible to PM by
demonstrating that neonatal rats exposed to iron-soot PM had a reduction in cell proliferation in the
lung. Overall, the evidence from epidemiologic studies that have examined the health effects
associated with all size fractions of PM and toxicological studies that have examined individual PM
components provide additional support to the hypothesis that children are more susceptible to
respiratory effects from exposure to PM.
8.1.2. Pregnancy and Developmental Effects
While the majority of the literature focuses on epidemiologic studies that examine the
potential health effects (e.g., low birth weight, growth restriction) attributed to in utero exposure to
PM (see Section 7.4), it is unclear if the health effects observed are due to soluble fractions of PM
that cross the placenta or physiological alterations in the pregnant woman. In the case of exposure to
PM, adverse health effects in the offspring could be mediated by potentially greater susceptibility in
the pregnant woman. For example, an inflammatory response leads to differential activation of
multiple genes involved in immune response and regulation, cell metabolism, and proliferation all of
which can lead to health effects in the developing fetus (Fedulov et al., 2008, 097482). Toxicological
December 2009 8-5
-------
studies have recently examined whether exposure to air pollutants during pregnancy leads to
increased allergic susceptibility in the offspring. Fedulov et al. (2008, 097482) used an animal model
to examine the effect of diesel exhaust particles (DEPs) along with an immunologically "inert"
particle (TiO2) on pregnant mice. The authors found that pregnant mice exhibited a local and
systemic inflammatory response when exposed to either DEP or TiO2, which was not observed in
control, non-pregnant mice. In addition, the offspring of exposed pregnant mice developed AHR and
allergic inflammation. This study suggests that exposure to PM2 5, and even relatively inert particles,
during pregnancy can potentially lead to increased allergic susceptibility in offspring and
subsequently the development of asthma.
8.1.3. Gender
The 2004 PM AQCD did not find consistent evidence for a difference in health effects by
gender. However, there appeared to be gender differences in the localization of particles when
deposited in the respiratory tract and the deposition rate due to differences in body size, conductive
airway size, and ventilatory parameters (U.S. EPA, 2004, 056905). For example, females have
proportionally smaller airways and slightly greater airway reactivity than males (Yunginger et al.,
1992. 192074).
Few recent epidemiologic studies have conducted gender-stratified analyses when examining
the association between either short- or long-term exposure to PMi0_2.5 or PM2.5. Similar to the
studies evaluated in the 2004 PM AQCD, the current literature has not found a consistent pattern of
associations by gender for any health outcome. Pope et al. (2006, 091246) observed a slightly larger,
non-significant, association between short-term exposure to PM2.5 and daily hospital admission for
acute IHD events in males. An examination of gender-specific effects by both Ostro et al. (2006,
087991) and Franklin et al. (2007, 091257) found conflicting associations by gender for multiple
cause-specific mortality outcomes. The inconsistency in associations between males and females is
further highlighted in studies that examined the health effects associated with long-term exposure to
PMio_2.5 and PM2 5. Chen et al. (2005, 087942) found larger effects in females for congestive heart
disease (CHD) mortality upon long-term exposure to PMi0_2.5 in three California cities. Naess et al.
(2007, 090736). also observed slightly larger effect estimates in females for CVD and lung cancer
mortality upon long-term exposure to PM2 5, but for COPD mortality the greatest association was
found in males.
The majority of the epidemiologic studies that examined the association between exposure to
PM and gender focused on exposure to PMi0. Although most of these studies do not attribute the
association to specific size fractions (i.e., PMi0_2.5 or PM2 5) or provide insight as to whether one size
fraction may be driving the observed effect, the studies of PMi0 provide further support that gender
does not appear to differentially affect PM-related health outcomes. Neither Zanobetti and Schwartz
(2005, 088069) nor Wellenius et al. (2006, 088748) found gender to be a significant effect modifier
of the risk estimates associated with short-term exposure to PMi0 and cardiovascular hospital
admissions. These results are consistent with those found in other studies that examined the
association between short-term exposure to PMi0 and both cardiovascular and respiratory hospital
admissions (Luginaah et al., 2005, 057327; Middleton et al., 2008, 156760). Additional studies that
examined the effects of short-term and long-term exposure to PMi0 on respiratory morbidity and
mortality (Boezen et al., 2005, 087396; Chen et al., 2005, 087942; Zanobetti and Schwartz, 2005,
088069; Zeka et al., 2006, 088749) found results that are consistent with those reported in studies of
PMio_2.5 and PM25 (i.e., gender is not likely to be an effect modifier).
Although human clinical studies are not typically powered to detect differences in response
between males and females, one study did report significantly greater decreases in blood monocytes,
basophils, and eosinophils in females compared to males following controlled exposures to UF EC
(Frampton et al., 2006, 088665). Overall, the evidence from primarily epidemiologic studies that
examined the association between short- and long-term exposure to PMi0_2.5 and PM2 5, along with
the supporting evidence from PMi0 studies, further confirms that although differences in dosimetry
exist between males and females, neither gender consistently exhibits a higher disposition for PM-
related health effects.
December 2009 8-6
-------
8.1.4. Race/Ethnicity
The 2004 PM AQCD (U.S. EPA, 2004, 056905) did not evaluate the potential susceptibility of
individuals of different races and ethnicities to PM exposure. The results from epidemiologic studies
evaluated in this review that examined the potential effect modification of the PM-morbidity and -
mortality relationships by race and ethnicity varied depending on the study location. In an analysis of
the PM2.5-mortality relationship, Ostro et al. (2006, 087991) stratified the association by race and
ethnicity, and observed a positive and marginally significant effect for whites and Hispanics, but not
for blacks, in response to short-term exposure to PM2.5 in 9 California counties. An additional
analysis performed by Ostro et al. (2008, 097971) in 6 California counties using PM2.5 and various
PM2.5 components, also found a significant association between mortality, specifically cardiovascular
mortality, and Hispanic ethnicity (Ostro et al., 2008, 097971). It should be noted that neither study,
Ostro et al. (2006, 087991) nor Ostro et al. (2008, 097971). controlled for potential confounders
(e.g., SES factors and location of residence) of the association observed between PM25 exposure and
Hispanic ethnicity. As a result, Ostro et al (2008, 097971) speculated that the increased PM2 5-
mortality risks observed for Hispanics could be due to a variety of factors including, higher rates of:
non-high school graduates, obesity, no leisure-time activity, and alcohol consumption within the
Hispanic population in California. Additional evidence for the potential susceptibility of individuals
by race and ethnicity were derived from studies on the health effects associated with short-term
exposure to PMi0. Wellenius et al. (2006, 088748) observed that race (i.e., white vs. other) did not
significantly modify the association between short-term exposure to PMi0and CHF hospital
admissions. Additionally, Zeka et al. (2006, 088749) did not observe any difference in mortality
effect estimates when stratifying by race (i.e., black and white) upon short-term exposure to PMi0.
To date, dosimetry studies have not extensively examined differences in particle deposition between
races or ethnicities to confirm the epidemiologic findings. Although not extensively analyzed,
toxicological studies have examined PM responses in different mouse and rat strains, and reported
greater CV effects (Kodavanti et al., 2003, 051325; Tankersley et al., 2007, 097910) and
compromised host defense (Ohtsuka et al., 2000, 004409) for some strains. These studies provide
some support, in terms of biological plausibility, for differences in PM-induced health effects by race
or ethnicity. However, it is unclear how the difference in the response to PM in different mouse or rat
strains extrapolates to PM-induced differences between races or ethnicities. Overall, the results from
the studies that examined the potential effect modification of PM associations by race and ethnicity
provide some evidence for increased risk of mortality in Hispanics upon short-term exposure to
PM2 5. However, the evidence for this association is derived from two studies conducted in
California, and it is unclear if the studies adequately controlled for potential confounders. Additional
studies conducted in other locations that stratify results by ethnicity have not yet been conducted to
substantiate these results.
8.1.5. Gene-Environment Interaction
A consensus now exists that gene-environment interactions merit serious consideration when
examining the relationship between ambient exposures to air pollutants and the development of
health effects (Gilliland et al., 1999, 155792: Kauffmann et al., 2004, 090968). These potential
interactions were not evaluated in the 2004 PM AQCD. Inter-individual variation in human
responses to air pollutants suggests that some populations are at increased risk of detrimental effects
due to pollutant exposure, and it has become clear that the genetic makeup of an individual can
increase their susceptibility (Kleeberger and Ohtsuka, 2005, 130489). Gene-environment
interactions can result in health effects due to: genetic polymorphisms, which result in the lack of a
protein or a change that makes a functionally important protein dysfunctional; or genetic damage in
response to an exposure which potentially leads to a health response (e.g., formation of benzo [a]
pyrene DNA adducts in response to PM exposure). In this review, the majority of studies examine
gene-environment interactions due to genetic polymorphisms. In order to establish useful links
between polymorphisms in candidate genes and adverse health effects, several criteria must be
satisfied: the product of the candidate gene must be significantly involved in the pathogenesis of the
adverse effect of interest; and polymorphisms in the gene must produce a functional change in either
the protein product or in the level of expression of the protein (U.S. EPA, 2008, 157075). Further, the
issue of confounding by other environmental exposures must be carefully considered.
December 2009 8-7
-------
It has been hypothesized that the cardiovascular and respiratory health effects that occur in
response to short-term PM exposure are mediated by oxidative stress (see Section 5.1.1). Research
has examined this hypothesis by primarily focusing on the glutathione-S transferase (GST) genes
because they have common, functionally important polymorphic alleles that significantly affect
antioxidant defense function in the lung, and approximately half of the white population has a
polymorphic null allele, resulting a large potential study population (Schwartz et al., 2005, 086296).
Exposure to free radicals and oxidants in air pollution leads to a cascade of events, which can result
in a reduction in glutathione (GSH), and an increase in the transcription of GSTs. Individuals with
genotypes that result in reduced or absent enzymatic activity are likely to have reduced antioxidant
defenses and potentially increased susceptibility to inhaled oxidants and free radicals.
Numerous studies have examined the role of genetic polymorphisms on PM-related
cardiovascular health effects using the Normative Aging Study cohort. Schwartz et al. (2005,
086296) and Chahine et al. (2007, 156327) found that individuals with null GSTM1 alleles had a
larger decrease in HRV upon short-term exposure to PM2.5 compared to individuals with at least one
allele. Polymorphisms in the HO-1 promoter resulted in lowered HRV upon short-term exposure to
PM2 5 in individuals with the long repeat polymorphism compared to those individuals with the short
repeat polymorphism (Chahine et al., 2007, 156327). In addition, Schneider et al. (2008, 191985)
found that diabetic individuals with null GSTM1 alleles had larger decrements in FMD (i.e., flow-
mediated dialation of the brachial artery), suggesting alterations in endothelial function. A controlled
human exposure study (Gilliland et al., 2004, 156471) also examined whether genetic
polymorphisms increase the susceptibility of individuals to respiratory morbidity in response to PM
exposure. Gilliland et al. (2004, 156471) examined the effect of allergens and DEPs on individuals
with either null genotypes for GSTM1 and GSTT1 or GSTP1 codon 105 variants. The authors found
that individuals with the GSTM1 null or the GSTP1 1105 wildtype genotypes were more susceptible
to allergic inflammation upon exposure to allergen and DEPs. Additional genes within the GST
pathway have also been examined (e.g., NQO1), but the sample sizes are relatively small, which
prohibits the analysis of the potential effect modification of PM-related health effects by these genes
(e.g., Schneider et al., 2008, 191985).
The interaction between GST genes and PM exposure has recently been extended to studies
that examined the effect of PM exposure on birth outcomes. A recent study (Suh et al., 2008,
192077) that examined the effect of high PM10 exposures during the third trimester of pregnancy on
the risk of preterm delivery, found that women with the GSTM1 null genotype were at an increased
risk of preterm birth. When examining the interaction between high PMi0 concentrations during the
third trimester of pregnancy and the presence of the GSTM1 null genotype on the risk of preterm
delivery, there was evidence for a synergistic gene-environment interaction in pregnant women. This
effect could occur due to oxidative stress induced by metals contained in PMi0, and subsequently
could be modified by polymorphisms of the GSTM1 gene. This oxidative stress causes oxidative
DNA damage in fetal tissues, which may lead to preterm delivery via a reduction in placental blood
flow.
An examination of other genes outside the GST pathway have also been conducted to
determine if specific polymorphisms increase the susceptibility of individuals to PM. Baccarelli et al.
(2008, 157984) in the Normative Aging Study observed that individuals with polymorphisms in
MTHFR (C677T methylenetetrahydrofolate reductase), an alteration associated with reduced
enzyme activity, and cSHMT (cytoplasmic serine hydroxymethyltransferase) (i.e., [CT/TT] MTHFR
and [CC] cSHMT genotypes), alterations associated with higher homocysteine levels, have a
reduction in SDNN, upon exposure to PM2.5. Peters et al. (2009, 191992) examined single nucleotide
polymorphisms (SNPs) in the fibrinogen gene in myocardial infarction survivors to assess whether
exposure to PMi0 altered steady state levels of fibrinogen, which has been implicated in promoting
atherothrombosis. The authors found that individuals with single nucleotide polymorphisms (SNPs)
in the fibrinogen gene have higher baseline fibrinogen levels which when combined with the
inflammatory effects (i.e., increased fibrinogen levels) associated with exposure to PM could
increase their risk of PM-related cardiovascular health effects. These results taken together suggest
that individuals with null alleles or specific polymorphisms in genes that mediate the antioxidant
response to oxidative stress, regulate enzyme activity, or regulate levels of procoagulants are more
susceptible to PM. However, in some cases genetic polymorphisms may actually reduce an
individual's susceptibility to PM-related health effects. For example, Park et al. (2006, 091245)
found that individuals with two hemochromatosis (HFE) polymorphisms (C282Y and H63D), which
result in an increase in iron uptake, had smaller reductions in HRV upon exposure to PM2.5. This
December 2009 8-8
-------
effect could possibly be due to the reduction in free iron that enters oxidation-reduction (redox)
reactions and the subsequent reduction in reactive oxygen species (ROS).
More recently, studies have begun to focus on epigenetic effects associated with PM exposure
(i.e., the effect of PM on DNA methylation) due to the fact that DNA methylation can result in gene
alterations. The limited number of epidemiologic studies that examined epigenetic effects have
found some evidence that long-term exposure to PM2 5 and PMi0 can influence DNA methylation
(Baccarelli et al, 2009, 188183; Tarantini et al., 2009, 192010). Additionally, a toxicological study
found some evidence of hypermethylation of spermatogonial stem cells in response to the PM
component of ambient urban air (Yauk et al., 2008, 157164). Although epigenetic effects have been
observed in response to PM exposure in some studies additional research is needed to more
accurately characterize these associations.
Overall, the evidence suggests that specific genetic polymorphisms can potentially increase
the susceptibility of an individual to PM exposure, but protective polymorphisms also exist, which
may diminish the health effects attributed to PM exposure in some individuals. In addition, the
studies that examine genetic polymorphisms or epigenetics can potentially provide additional
information that can aid in identifying the specific pathways and mechanisms by which PM initiates
health effects.
8.1.6. Pre-Existing Disease
In 2004, the National Research Council (NRC) published a report that emphasized the need to
evaluate the effect of air pollution on susceptible populations, including those with respiratory
illnesses and cardiovascular diseases (NRC, 2004, 156814). The 2004 PM AQCD included
epidemiologic evidence suggesting that individuals with pre-existing heart and lung diseases, as well
as diabetes may be more susceptible to PM exposure. In addition, toxicological studies that used
animal models of cardiopulmonary diseases and heightened allergic sensitivity found evidence of
enhanced susceptibility. More recent epidemiologic and human clinical studies have directly
examined the effect of PM on individuals with pre-existing diseases and toxicological studies have
employed disease models to identify whether exposure to PM disproportionately effects certain
populations.
8.1.6.1. Cardiovascular Diseases
The potential effect of underlying cardiovascular diseases on PM-related health responses has
been examined using epidemiologic studies that stratify effect estimates by underlying conditions or
secondary diagnoses, and toxicological studies that use animal models to mimic the physiological
conditions associated with various cardiovascular diseases (e.g., MI, ischemia, and atherosclerosis).
A limited number of controlled human exposure studies have also examined the potential
relationship between CVD and exposure to PM in individuals with underlying cardiovascular
conditions, but these studies have provided somewhat inconsistent evidence for these associations.
The majority of the epidemiologic literature that examined the association between short-term
exposure to PM and cardiovascular outcomes focuses on cardiovascular-related hospital admissions
and emergency department (ED) visits. Hypertension is the pre-existing condition that has been
considered to the greatest extent when examining the association between short-term exposure to PM
and cardiovascular-related HAs and ED visits. Pope et al. (2006, 091246) found no evidence of
effect modification of the IHD ED visit association with PM2.5in individuals with secondary
hypertension in Utah. This is consistent with the results of both Wellenius et al. (2006, 088748) in 7
U.S. cities and Lee et al. (2008, 192076) in Taipei, which found that hypertension did not modify the
association between PM10 and cardiovascular-related health outcomes. These results differ from
those presented by Peel et al. (2007, 090442). in Atlanta, which observed that exposure to PMi0
resulted in an increase in ED visits for arrhythmias and CHF in individuals with underlying
hypertension. An additional study conducted by Park et al. (2005, 057331) in Boston found that
underlying hypertension increased associations between HRV, specifically a reduction in the HF
parameter, and short-term exposure to PM2.5.
Park et al. (2005, 057331). in the analysis mentioned above, examined other underlying
cardiovascular conditions and found associations between PM2.5 and HRV in individuals with
pre-existing IHD. In a toxicological study, Wellenius et al. (2003, 055691) examined the effects of
December 2009 8-9
-------
PM2.5 CAPs exposure on induced myocardial ischemia in dogs, which mimics the effects associated
with IHD. The authors found that exposure to PM2.5 prior to the induced ischemia increased
ST-segment elevation, indicating greater ischemia than air-exposed animals (Wellenius et al., 2003,
055691). A follow-up study implicated impaired myocardial blood flow in the response (Bartoli et
al., 2009, 179904).
Additional studies examined the effects of PM on cardiac function in individuals with
dysrhythmia. Peel et al. (2007, 090442) observed some evidence for an increase in ED visits for IHD
for individuals with secondary dysrhythmia and PMi0 exposure. However, when examining CHF
hospital admissions in 7 U.S. cities, Wellenius et al. (2006, 088748) found no evidence for effect
modification of PMi0 exposure in individuals with secondary dysrhythmia.
Limited evidence is available from epidemiologic studies that examined other pre-existing
cardiovascular conditions, such as CHF and MI. Pope et al. (2006, 091246) observed an increase in
hospital admissions for acute IHD in individuals with underlying CHF upon short-term exposure to
PM2.5. However, Peel et al. (2007, 090442) did not find that underlying CHF contributed to an
increase in the association between IHD ED visits and short-term exposure to PMi0. Zanobetti and
Schwartz (2005, 088069) also examined the potential effect modification of the association between
PMio and cardiovascular-related health effects in individuals with CHF, but used MI hospital
admissions as the outcome of interest. Underlying CHF was not found to increase MI hospital
admissions for exposure to PMi0 in the cohort of more than 300,000 hospital admissions.
Wellenius et al. (2006, 088748) examined the effect of previous diagnoses of acute MI on the
association between CHF hospital admissions and short-term exposure to PMi0 in 7 U.S. cities. In
this study, Wellenius et al. (2006, 088748) found no evidence of effect modification of the
relationship between PMi0 and CHF hospital admissions by previous acute MI. Toxicological studies
have provided additional evidence for the cardiovascular health effects associated with exposure to
PM in individuals with underlying MI. Anselme et al. (2007, 097084) and Wellenius et al. (2006,
156152) examined the arrhythmic effects of PM on rats that experienced an MI using two different
models. Wellenius et al. (2006, 156152) used a post-myocardium sensitivity model (acute MI) and
observed that exposure to PM2 5 CAPs decreased ventricular premature beats and spontaneous
supraventricular ectopic beats. In contrast, the MI model of chronic heart failure (i.e., rats that
experienced an MI 3 mo prior to exposure), demonstrated a prominent increase in the incidence of
premature ventricular contraction when exposed to DE (Anselme et al., 2007, 097084). The
discrepancy in effects observed between studies could be due to differences in the MI model or the
PM exposure (i.e., CAPs vs. DE).
Additional toxicological studies examined the association between PM and pre-existing
cardiovascular diseases using amurine model of atherosclerosis (ApoE~'~ mouse). For example,
Campen et al. (2005, 083977: 2006, 096879) examined the heart rate and ECG effects of acute
exposure to PM on ApoE"'" mice. With DE, dramatic bradycardia and T-wave depression were
observed that were attributable to the gases (Campen et al., 2005, 083977), while whole gasoline
emissions induced T-wave alterations that required particles (Campen et al., 2006, 096879).
However, these studies along with others that used this mouse model (see Section 6.2 and 7.2) did
not compare the effects observed with the ApoE"" mouse to other non-diseased mouse models, so it is
unclear if the responses would differ if other strains were used in the same experimental protocol.
Controlled human exposure studies that examined the effect of pre-existing diseases on
cardiovascular outcomes with exposure to PM are less consistent and difficult to interpret in the
context of the results from the epidemiologic and toxicological studies. Mills et al. (2007, 091206;
2008, 156766) investigated the effects of dilute DE, or fine and ultrafine CAPs, respectively, on
subjects with coronary artery disease and prior MI. Exposure to dilute DE was found to promote
exercise-induced ST-segment changes indicating myocardial ischemia, as well as inhibit endogenous
fibrinolytic capacity (Mills et al., 2007, 091206). The physiological responses observed in Mills et
al. (2007, 091206) provides a measure of coherency with the cardiovascular effects observed in
epidemiologic studies, including increases in hospital admissions and ED visits for IHD and stroke
associated with exposure to PM. An examination of fine and ultrafine CAPs that were low in
combustion derived particles, were not found to exhibit any significant effects on vascular function
(Mills et al., 2008, 156766). Routledge et al. (2006, 088674) reported no change in HRV in a group
of adults with coronary artery disease following exposure to ultrafine carbon particles, which may be
explained in part by the use of medication (beta blockers) among the majority of the subjects.
Although the epidemiologic studies did not examine potential effect modification of pre-existing
cardiovascular conditions on effects associated with long-term exposure to PM, a few
December 2009 8-10
-------
toxicological studies exposed animals with underlying cardiovascular conditions to PM for
months. In studies that focused on the cardiovascular effects following subchronic exposure to
PM in ApoE"A mice, relatively consistent physiological effects were observed across studies.
Araujo et al. (2008, 156222) exposed mice to ultrafme CAPs and observed enhanced size of
early atherosclerotic lesions. Similarly, Chen and Nadziejko (2005, 087219) and Sun et al. (2005,
087952: 2008, 157033) exposed mice to PM2.5 CAPs with the same results. An additional
long-term exposure study observed a decreasing trend in heart rate, physical activity, and
temperature along with biphasic responses in HRV (SDNN and rMSSD) upon exposure to CAPs
(Chen and Hwang, 2005, 087218).
While the majority of the literature examines the potential modification of the association
between PM and non-fatal cardiovascular health effects, a few new studies have also examined
effect modification in mortality associations. Zeka et al. (2006, 088749) found an increase in risk
estimates for associations between PMi0 and mortality in individuals with underlying stroke, while
Bateson et al. (2004, 086244) found evidence for effect modification of the PM-mortality association
in individuals with CHF.
Collectively, the evidence from epidemiologic and toxicological, and to a lesser extent,
controlled human exposure studies indicates that individuals with underlying cardiovascular diseases
are susceptible to PM exposure. Although the evidence for some outcomes was inconsistent across
epidemiologic and toxicological studies, this could be due to a variety of issues including the PM
size fraction used in the study along with the study location. Even with these caveats, a large
proportion of the U.S. population has been diagnosed with cardiovascular diseases (i.e.,
approximately 51.6 million people with hypertension, 24.1 million with heart disease, and 14.1
million with coronary heart disease [see Table 8-3]), and therefore represents a large population that
is potentially more susceptible to PM exposure than the general population.
Table 8-3. Percent of the U.S. population with respiratory diseases, cardiovascular diseases, and
diabetes.
Age
Regional
Adults (18+)*
18-44 45-64
65-74
75+
NE
MW
W
Chronic Condition/
Disease
Number (x 1(T
RESPIRATORY DISEASES
Asthma*
24.2
11.0
11.5
10.5
11.7
9.3
11.7
11.5
10.5 10.8
Asthma (<18yrs)
6.8*
9.3*
COPD
Chronic bronchitis
9.5
4.3
2.9
5.5
5.6
6.7
3.8
4.4
4.9
3.5
Emphysema
4.1
0.3
2.4
5.0
6.4
1.4
2.3
1.9
1.6
CARDIOVASCULAR DISEASES
All heart disease
24.1
10.9
3.6
12.3
26.1
36.3
10.8
12.7
10.9
9.2
Coronary heart disease
14.1
6.4
0.9
7.2
18.4
25.5
6.4
7.6
6.6
4.7
Hypertension
51.6
23.4
77
32.4
52.7
53.5
22.2
23.7
25.3
20.6
Stroke
5.6
2.6
0.5
2.4
7.6
11.2
2.1
2.8
2.9
2.2
Diabetes
17.1
7.8
2.6
10.4
18.2
17.9
7.2
8.0
7.4
* All data for adults except asthma prevalence data for children under 18 years of age, from CDC (2008,156324: 2008,156325). For adults prevalence data based off adults
responding to "ever told had asthma."
Source: Data from Pleis and Lethbridge-Qejku (2007,1568751: CDC (2008,156324: 2008,1563251.
December 2009
8-11
-------
8.1.6.2. Respiratory Illnesses
Investigators have examined the effect of pre-existing respiratory illnesses on multiple health
outcomes (e.g., mortality, asthma symptoms, CHF) in response to exposure to ambient levels of PM.
Animal models have been developed and/or human clinical studies conducted to examine the
possible PM effects on pre-existing respiratory conditions in a controlled setting.
Epidemiologic studies have examined the effect of short-term exposure to PM on the
respiratory health of asthmatic individuals measuring a variety of respiratory outcomes. Asthmatic
individuals were found to have an increase in medication use (Rabinovitch et al., 2006, 088031).
respiratory symptoms (i.e., asthma symptoms, cough, shortness of breath, and chest tightness) (Gent
et al., 2003, 052885). and asthma symptoms (Delfino et al., 2002, 093740; 2003, 050460) with short-
term exposure to PM2.5; and morning symptoms (Mortimer et al., 2002, 030281) and asthma attacks
(Desqueyroux et al., 2002, 026052) with short-term exposure to PMi0.
Toxicological studies that have used ovalbumin-induced allergic airway disease models
provide evidence which supports the findings of the epidemiologic literature. Morishita et al. (2004,
087979) used this model to assess the health effects of PM2.5 components. In response to short-term
exposure to CAPs from Detroit, an area with pediatric asthma rates three times the national average,
rats with allergic airway disease were found to preferentially retain PM derived from identified local
combustion sources in association with eosinophil influx and BALF protein content after an acute
exposure (Morishita et al., 2004, 087979). These findings suggest that individuals with allergic
airways conditions are more susceptible to allergic airways responses upon exposure to PM2.5, which
may be partially attributed to increased pulmonary deposition and localization of particles in the
respiratory tract (Morishita et al., 2004, 087979). An additional study (Heidenfelder et al., 2009,
190026) examined whether genes are differentially expressed upon exposure to PM. They found that
exposure to CAPs increased the expression of genes associated with inflammation and airway
remodeling in rats with allergic airway disease. Although the evidence is much more limited, not all
of the toxicological studies evaluated that examined the effect of underlying respiratory conditions
on PM-related respiratory morbidity focused on allergic airways disease. Using an animal model of
emphysema (i.e., papain-treated mice), Lopes et al. (2009, 190430) found that papain-treated mice
exposed to urban ambient PM demonstrated a statistically significant increase in mean linear
intercept, a measure of airspace enlargement, compared to saline-treated controls exposed to filtered
air. These results provide preliminary evidence, which suggests that non-allergic respiratory
morbidities may also increase the susceptibility of an individual to PM-related respiratory effects.
The results from the epidemiologic and toxicological studies that focused on underlying
allergic airways disease is supported by a series of controlled human exposure studies which have
shown that exposure to DEPs increases the allergic inflammatory response in atopic individuals
(Bastain et al., 2003, 098690; Diaz-Sanchez et al., 1997, 051247; Nordenhall et al., 2001, 025185).
However, not all controlled human exposure studies have found evidence for differences between the
respiratory effects exhibited by healthy and asthmatic individuals. Studies by Gong et al. (2003,
042106; 2004, 055628; Gong et al., 2008, 156483) reported that healthy and asthmatic subjects
exposed to coarse, fine and ultrafine CAPs, exhibited similar respiratory responses. However, it
should be noted that these studies excluded moderate and severe asthmatics that would be expected
to show increased susceptibility to PM exposure.
In addition to examining the association between exposure to PM and respiratory effects in
asthmatics, some studies examined whether individuals with COPD represent a potentially
susceptible population. Desqueyroux et al. (2002, 026052) did not observe an increase in the
exacerbation1 of COPD in response to short-term exposure to PM2 5. However, studies that examined
the effect of PM on lung function in individuals with COPD (Lagorio et al., 2006, 089800; Trenga et
al., 2006, 155209) observed declines in FEVi, and FEVi and FVC, respectively in response to PMi0
and/or PM25. Silkoff et al. (2005, 087471) observed associations between PMi0 and a reduction in
FEVi and PM25 and a reduction in PEF, in those with COPD, but only during one winter of the
analysis. Only one controlled human exposure study examined the effects of PM on COPD subjects
and found no significant difference in respiratory effects between healthy and individuals with
COPD upon exposure to PM2.5 CAPs (Gong et al., 2004, 055628). On the other hand the results from
dosimetry studies have shown that COPD patients have increased dose rates and impaired
1 Desqueyroux et al. (2002, 026052) defined a COPD exacerbation as (a) decrease in "vesicular" breath sound, (b) bronchial obstruction,
(c) tachycardia or arrhythmia, or (d) cyanosis.
December 2009 8-12
-------
mucociliary clearance relative to age matched healthy subjects, suggesting that individuals with
COPD are potentially at a greater risk of PM-related health effects (Sections 4.2.4.5 and 4.3.4.3).
A few of the epidemiologic studies examined the effect of underlying respiratory illnesses on
the association between short- and long-term exposure to PM and mortality. Using different
pre-existing respiratory illnesses, Zeka et al. (2006, 088749) and De Leon et al. (2003, 055688)
found that short-term exposure to PMi0 increased the risk of nonaccidental mortality for pneumonia
and circulatory mortality for all respiratory illnesses, respectively. Additionally, Zanobetti et al.
(2008, 156177) observed an association between long-term exposure to PMi0 and mortality in
individuals that had previously been hospitalized for COPD. Although these studies do not examine
additional size fractions of PM, together they highlight the potential effect of underlying respiratory
illnesses on the PM-mortality relationship.
Overall, the epidemiologic, controlled human exposure, and toxicological studies evaluated
provide biological plausibility for the increased health effects observed in epidemiologic studies
among asthmatic individuals in response to PM exposure. Although, the evidence from studies that
examined associations between PM and health effects in individuals with COPD is inconsistent,
taken together individuals with COPD and asthma represent a large percent of the U.S. population
(~45 million people), which may be more susceptible to PM-related health effects (Table 8-3).
8.1.6.3. Respiratory Contributions to Cardiovascular Effects
Although the majority of health effects observed in individuals with pre-existing respiratory
illnesses were associated with respiratory illness exacerbations, studies also examined whether
underlying respiratory illnesses can lead to cardiovascular effects in response to PM exposure.
Controlled human exposure and toxicological studies have also observed some cardiovascular effects
in individuals with pre-existing respiratory illnesses. Gong et al. (2003, 042106) observed acute
responses in the cardiovascular system and systemic circulation among asthmatic individuals after
exposure to PM2.5 CAPs. However, respiratory disease has not consistently been observed to affect
cardiovascular response in controlled human exposure studies. In a toxicological study, Batalha et al.
(2002, 088109). using a chronic bronchitis animal model, found that the pulmonary artery
lumen-to-wall ratio was decreased in rats exposed to PM2.5 CAPs, although the induced bronchitis
didn't seem to affect the response. The majority of epidemiologic studies that examined whether
underlying respiratory illnesses contributed to the manifestation of PM-related cardiovascular
hospital admission or ED visits, did not report increases in effects for a variety of cardiovascular
outcomes (e.g., IHD, arrhythmias, CHF, MI) for individuals with underlying respiratory infection
(Wellenius et al., 2006, 088748). pneumonia (Zanobetti and Schwartz, 2005, 088069). or COPD
(Peel et al., 2007, 090442; Wellenius et al., 2005, 087483). However, Yeatts et al. (2007, 091266). in
a panel study, found evidence for cardiovascular effects, specifically reductions in HRV parameters,
in asthmatic adults upon short-term exposure to PMi0_2.5. It must be noted that most of the
aforementioned epidemiologic studies focused on exposure to PMi0, and, therefore, it is unclear how
these results compare to those found in the controlled human exposure and toxicological studies that
focused on exposure to PM2.5 (e.g., CAPs). Thus, it is unclear if individuals with underlying
respiratory illnesses represent a population that is potentially susceptible to PM-related
cardiovascular effects.
8.1.6.4. Diabetes and Obesity
It has been hypothesized that the systemic inflammatory cascade leads to an increase in
cardiovascular risk (Dubowsky et al., 2006, 088750). As a result, individuals with conditions linked
to chronic inflammation (i.e., diabetes and obesity), have been examined to determine whether
diabetes or obesity facilitate the manifestation of PM-mediated health effects, and, therefore,
represent a potentially susceptible population.
Epidemiologic studies have examined whether diabetes modifies the association between
cardiovascular health effects and PM exposure, but these studies have primarily focused on short-
term exposure to PMi0. Time-series studies have provided evidence through an examination of
hospital admission and ED visits and mortality, which suggests an increase in health effects in
diabetic individuals in response to PM exposure. Multicity studies have found upwards of 75%
greater risk of hospitalization for cardiac diseases in individuals with diabetes upon exposure to
(Zanobetti and Schwartz, 2002, 034821). Studies conducted in Atlanta, Georgia have also
December 2009 8-13
-------
found increased risk for cardiovascular-related ED visits in diabetics, specifically for IHD,
arrhythmias, and CHF (Peel et al., 2007, 090442). Additional studies found some evidence that
individuals with diabetes are at increased risk of mortality upon exposure to PMi0 (Zeka et al., 2006,
088749) and PM2.5 (Goldberg et al., 2006, 088641). However, some studies (both multicity and
single-city) have not observed a modification of the risk of cardiovascular ED visits and hospital
admissions in response to exposure to PMi0 in diabetics (Pope et al., 2006, 091246; Wellenius et al.,
2006, 088748; Zanobetti and Schwartz, 2005, 088069).
Panel and cohort studies have been conducted to determine the physiological changes that
occur in individuals with diabetes in response to PM exposure. These studies examined both changes
in inflammatory markers along with specific physiological alterations in the cardiovascular system.
Schneider et al. (2008, 191985) in a panel study of 22 individuals with type 2 diabetes mellitus in
Chapel Hill, NC found evidence that ambient exposure to PM2.5 enhanced the reduction in various
markers of endothelial function. Liu et al. (2007, 156705) observed an increase in end-diastolic FMD
and end-systolic FMD, and decreases in end-diastolic basal diameter and end-systolic basal diameter
in diabetics upon exposure to PMi0. The authors also observed positive associations with FMD and
blood pressure in diabetic individuals. A controlled human exposure study conducted by Carlsten et
al. (2008, 156323) found that DE did not elicit any prothrombotic effects in subjects with metabolic
syndrome, which consists of physiological alterations similar to those observed in both diabetic and
obese individuals. An examination of biomarkers found mixed results, with Liao et al. (2005,
088677) observing an increase in vWF; Liu et al. (2007, 156705) observing an increase in TEARS,
but not CRP or TNF-a; and Dubowsky et al. (2006, 088750) observing an increase in CRP and
WBCs. Overall, it is unclear how differences in each of the aforementioned biomarkers contribute to
the potential overall cardiovascular effect observed in diabetic individuals; however, an increase in
inflammation, oxidative stress, and acute phase response may contribute to cardiovascular effects. A
recent toxicological study (Sun et al., 2009, 190487), also demonstrated the potential for PM-related
health effects in diabetics. Sun et al. (2009, 190487) found that PM2.5 CAPs exposure for 4 mo can
exaggerate insulin resistance, visceral adiposity, and inflammation in a diet-induced obesity mouse
model.
Overall, epidemiologic studies have reported evidence for increased effects in diabetics in
response to PM exposure, with preliminary evidence for pathophysiologic alterations from
toxicological studies. This potentially susceptible population is large, with an estimated 17.1 million
diabetic individuals in the U.S. (Table 8-3). However, the limited evidence from toxicological and
controlled human exposure studies along with the lack of studies that examined additional PM size
fractions warrants additional research to confirm the associations observed and to identify the
biological pathway(s) that may result in a greater response to PM in diabetics.
In addition to diabetes, obesity has been examined as a health condition with the potential to
lead to an increase in PM-related health effects. Only a few recent studies have examined the
potential effect modification of PM risk estimates by obesity. Schwartz et al. (2005, 086296)
reported a change in HRV in obese (i.e., BMI > 30 kg/m2) compared to non-obese subjects, while
Dubowsky et al. (2006, 088750) observed an increase in inflammatory markers (i.e., CRP, IL-6, and
WBC) in response to short-term exposure to PM2 5 among obese individuals. Additionally, Schneider
et al. (2008, 191985) found some evidence for a larger reduction in FMD in individuals with a BMI
>30 kg/m3 in response to PM2 5 exposure. These effects could be due, in part, to a higher PM dose
rate in obese individuals, which has been demonstrated in children by Bennett and Zeman (2004,
155686). These investigators also reported that tidal volume and resting minute ventilation increased
with body mass index. Although a limited amount of research has been conducted to examine PM-
related health effects in obese individuals there is an increasing trend of individuals within the U.S.
that have been defined as overweight (BMI > 25.0) or obese (BMI > 25.0), with the prevalence of
overweight individuals increasing from 20-74% from 1960 to 2004, and the prevalence of obese
individuals increasing from 13.3-32.1% (NCHS, 2006, 198921).
8.1.7. Socioeconomic Status
SES is a composite measure that usually consists of economic status, measured by income;
social status measured by education; and work status measured by occupation (Dutton and Levine,
1989, 192052). Based on data from the U.S. Census Bureau in 2006, from among commonly-used
indicators of SES, about 12% of individuals and 11% of families are below the poverty line
(U.S. Census, 2009, 192147). Although the measure of SES is composed of a multitude of
December 2009 8-14
-------
surrogates, each of these linked factors can influence an individual's susceptibility to PM-related
health effects. Additionally, low SES individuals have been found to have a higher prevalence of
pre-existing diseases; inadequate medical treatment; and limited access to fresh foods leading to a
reduced intake of antioxidant polyunsaturated fatty acids and vitamins, which can increase this
population's susceptibility to PM (Kan et al., 2008, 156621).
Surrogates of SES, such as educational attainment, have been shown in some studies to
modify health outcomes of PM exposure for a population. Within the U.S. approximately 16% of the
population does not have a high school degree and only 27% have a bachelor's degree or higher
level of education (U.S. Census, 2009, 192148). Educational attainment generally coincides with an
individual's income level, which is correlated to other surrogates of SES, such as residential
environment (Jerrett et al., 2004, 087379). Franklin et al. (2008, 155779) noted an increased risk in
mortality associated with short-term exposure to PM2.5 and its components for individuals with low
SES while additional analyses stratified by education level have also observed consistent trends of
increased mortality for PM25 and PM2 5 species for individuals with low educational attainment
(Ostro et al., 2006, 087991; Ostro et al., 2008, 097971; Zeka et al., 2006, 088749). This is further
supported by a reanalysis of the ACS cohort (Krewski et al., 2009, 191193). which found moderate
evidence for increased lung cancer mortality in individuals with a high school education or less
compared to individuals with more than a high school education in response to long-term exposure
to PM2 5. However, when examining education level and IHD mortality due to long-term exposure to
PM2 5 Krewski et al. (2009, 191193) observed an inverse relationship.
Epidemiologic studies have also examined additional surrogates of SES, such as residential
location and nutritional status to identify their influence on the susceptibility of a population. Jerrett
et al. (2004, 087379) examined the modification of acute mortality effects due to particulate air
pollution exposure by residential location in Hamilton, Canada using educational attainment as a
surrogate for SES. The authors found that the area of the city with the highest SES characteristics
displayed no evidence of effect modification while the area with the lowest SES characteristics had
the largest health effects. Likewise, Wilson et al. (2007, 157149) examined the effect of SES on the
association between mortality and short-term exposure to PM in Phoenix, but used educational
attainment and income to represent SES. When stratifying Phoenix into central, middle, and outer
rings of varying urban density central Phoenix, the area with the lowest SES, was found to exhibit
the greatest association with PM2 5. However, the association with urban density differed when
examining PMi0_2.5, with the greatest effect being observed for the middle ring. Yanosky et al. (2008,
192081) examined whether long-term exposure to traffic-related pollutants, using NO2 as a
surrogate, varied by SES at the block group level. The authors found higher levels of NO2 associated
with lower SES areas, which suggests that lower SES individuals are disproportionately exposed to
traffic-related pollutants, which includes PM.
Nutritional deficiencies have been associated with increased susceptibility to a variety of
infectious diseases and chronic health effects. Low SES may decrease access to fresh foods, and thus
be related to nutritional deficiencies that could increase susceptibility to PM-related health effects.
Baccarelli et al. (2008, 157984) examined the association between exposure to PM2 5 and HRV in
individuals with polymorphisms in MTHFR and cSHMT genes, which are associated with reduced
enzyme activity and increased risk of CVD. The authors found that when individuals with these
genetic polymorphisms increased their intake (above median levels) of B6, B12, or methionine no
PM2 5 effect on HRV was observed.
8.1.8. Summary
Upon evaluating the association between short- and long-term exposure to PM and various
health outcomes, studies also attempted to identify populations that are more susceptible to PM.
These studies did so by: conducting stratified analyses; examining individuals with an underlying
health condition; or developing animal models that mimic the physiological conditions associated
with an adverse health effect. These studies identified a multitude of factors that could potentially
contribute to whether an individual is susceptible to PM (Table 8-2). Although studies have primarily
used exposures to PMi0 or PM2 5, the available evidence suggests that the identified factors may also
enhance susceptibility to PMi0_2.5.
The majority of observations made during the evaluation of the literature reviewed in this ISA
are consistent with those reported in the 2004 PM AQCD. An evaluation of age-related health effects
suggests that older adults have heightened responses for cardiovascular morbidity with PM exposure.
December 2009 8-15
-------
In addition, epidemiologic and toxicological studies provide evidence, which indicates that children
are at an increased risk of PM-related respiratory effects. It should be noted that the health effects
observed in children could be initiated by exposures to PM that occurred during key windows of
development, such as in utero. Studies that focus on exposures during development have reported
inconsistent findings (see Section 7.4.), but a recent toxicological study suggests that inflammatory
responses in pregnant women due to exposure to PM could result in health effects in the developing
fetus.
Epidemiologic studies have also examined whether additional factors, such as gender, race, or
ethnicity modify the association between PM and morbidity and mortality outcomes. Consistent with
the findings of the 2004 PM AQCD, gender and race do not seem to modify the association between
PM and morbidity and mortality outcomes. However, some evidence, albeit from two studies
conducted in California, suggest that Hispanic ethnicity may modify the association between PM and
mortality.
Recent epidemiologic and toxicological studies provided evidence that individuals with null
alleles or polymorphisms in genes that mediate the antioxidant response to oxidative stress (i.e.,
GSTM1), regulate enzyme activity (i.e., MTHFR and cSHMT), or regulate levels of procoagulants
(i.e., fibrinogen) are more susceptible to PM exposure. However, some studies have shown that
polymorphisms in genes (e.g., HFE) can have a protective effect upon PM exposure. Additionally,
preliminary evidence suggests that PM exposure can impart epigenetic effects (i.e., DNA
methylation), however, this requires further investigation.
Collectively, the evidence from epidemiologic and toxicological, and to a lesser extent,
controlled human exposure studies indicate increased susceptibility of individuals with underlying
cardiovascular diseases and respiratory illnesses, specifically asthma, to PM exposure. Additional
controlled human exposure and toxicological studies provide some evidence for increased PM-
related cardiovascular effects in individuals with underlying respiratory health conditions. However,
the results are not consistent with epidemiologic studies, resulting in the need for further
investigation.
Recently studies have begun to examine the influence of preexisting chronic inflammatory
conditions, such as diabetes and obesity, on PM-related health effects. These studies have found
some evidence for increased associations for cardiovascular outcomes along with physiological
alterations in markers of inflammation, oxidative stress, and acute phase response. However more
research is needed to thoroughly examine the effect of PM exposure on obese individuals and to
identify the biological pathway(s) that could lead to increased susceptibility of diabetic and obese
individuals to PM.
There is also evidence that SES, measured using surrogates such as educational attainment or
residential location, modifies the association between PM and morbidity and mortality outcomes. In
addition, nutritional status, another surrogate of SES, has been shown to have protective effects
against PM exposure in individuals that have a higher intake in some vitamins and nutrients.
Overall, the epidemiologic, controlled human exposure, and toxicological studies evaluated in
this review provide evidence for increased susceptibility for various populations. Although the level
of evidence varies depending on the factor being evaluated collectively, it can be concluded that
some populations are more susceptible to PM than the general population.
December 2009 8-16
-------
Chapter 8 References
Aday LA. (2001). At risk in America: The health and health care needs of vulnerable populations in the United States. San
Francisco, CA: Jossey-Bass, Inc. 192150
American Lung Association. (2001). Urban air pollution and health inequities: A workshop report. Environ Health
Perspect, 3: 357-374. 016626
Anselme F; Loriot S; Henry J-P; Dionnet F; Napoleoni J-G; Thnillez C; Morin J-P (2007). Inhalation of diluted diesel
engine emission impacts heart rate variability and arrhythmia occurrence in a rat model of chronic ischemic heart
failure. Arch Toxicol, 81: 299-307. 097084
Araujo JA; Barajas B; Kleinman M; Wang X; Bennett BJ; Gong KW; Navab M; Harkema J; Sioutas C; Lusis AJ; Nel AE.
(2008). Ambient particulate pollutants in the ultrafine range promote early atherosclerosis and systemic oxidative
stress. Circ Res, 102: 589-596. 156222
Baccarelli A; Martinelli I; Pegoraro V; Melly S; Grillo P; Zanobetti A; Hou L; Bertazzi PA; Mannucci PM; Schwartz J.
(2009). Living near Major Traffic Roads and Risk of Deep Vein Thrombosis. Circulation, 119: 3118-3124. 188183
Baccarelli A; Martinelli I; Zanobetti A; Grillo P; HouLF; Bertazzi PA; Mannucci PM; Schwartz J. (2008). Exposure to
Particulate Air Pollution and Risk of Deep Vein Thrombosis. Arch Intern Med, 168: 920-927. 157984
Barnett AG; Williams GM; Schwartz J; Best TL; Neller AH; Petroeschevsky AL; Simpson RW (2006). The effects of air
pollution on hospitalizations for cardiovascular disease in elderly people in Australian and New Zealand cities.
Environ Health Perspect, 114: 1018-1023. 089770
Bartoli CR; Wellenius GA; Coull BA; Akiyama I; Diaz EA; Lawrence J; Okabe K; Verrier RL; Godleski JJ. (2009).
Concentrated ambient particles alter myocardial blood flow during acute ischemia in conscious canines. Environ
Health Perspect, 117: 333-337. 179904
BastainTM; Gilliland FD; Li Y-F; Saxon A; Diaz-Sanchez D. (2003). Intraindividual reproducibility of nasal allergic
responses to diesel exhaust particles indicates a susceptible phenotype. Clinlmmunol, 109: 130-136. 098690
Batalha JR; Saldiva P H; Clarke RW; Coull BA; Stearns RC; Lawrence J; Murthy GG; Koutrakis P; Godleski JJ. (2002).
Concentrated ambient air particles induce vasoconstriction of small pulmonary arteries in rats. Environ Health
Perspect, 110: 1191-1197. 088109
Bateson TF; Schwartz J. (2004). Who is sensitive to the effects of particulate air pollution on mortality? A case-crossover
analysis of effect modifiers. Epidemiology, 15: 143-149. 086244
Bennett WD; Zeman KL. (2004). Effect of body size on breathing pattern and fine-particle deposition in children. J Appl
Physiol, 97: 821-826. 155686
Boezen HM; Vonk JM; Van Der Zee SC; Gerritsen J; Hoek G; Brunekreef B; Schouten JP; Postma DS. (2005).
Susceptibility to air pollution in elderly males and females. Eur Respir J, 25: 1018-1024. 087396
Campen MJ; Babu NS; Helms GA; Pett S; Wernly J; Mehran R; McDonald JD. (2005). Nonparticulate components of
diesel exhaust promote constriction in coronary arteries from ApoE -/- mice. Toxicol Sci, 88: 95-102. 083977
Campen MJ; McDonald JD; Reed MD; Seagrave J. (2006). Fresh gasoline emissions, not paved road dust, alter cardiac
repolarization in ApoE-/- mice. Cardiovasc Toxicol, 6: 199-210. 096879
Carlsten C; Kaufman JD; Trenga CA; Allen J; Peretz A; Sullivan JH. (2008). Thrombotic markers in metabolic syndrome
subjects exposed to diesel exhaust. Inhal Toxicol, 20: 917-921. 156323
CDC. (2008). 2006 NHIS data: table 3-1: current asthma population estimates, in thousands by age, United States: national
health interview survey, 2006. Retrieved Ol-DEC-09, from http://www.cdc.gov/ASTHMA/nhis/06/table3-l .htm.
156324
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009 8-17
-------
CDC. (2008). 2006 NHIS data: table 4-1: current asthma prevalence percents by age, United States: national health
interview survey, 2006. Retrieved Ol-DEC-09, from http://www.cdc.gov/ASTHMA/nhis/06/table4-l.htm. 156325
Chahine T; Baccarelli A; Litonjua A; Wright RO; Suh H; Gold DR; Sparrow D; Vokonas P; Schwartz J. (2007). Particulate
air pollution, oxidative stress genes, and heart rate variability in an elderly cohort. Environ Health Perspect, 115:
1617-1622. 156327
Chen LC; Hwang JS. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice IV
Characterization of acute and chronic effects of ambient air fine particulate matter exposures on heart-rate
variability. Inhal Toxicol, 17: 209-216. 087218
Chen LC; Nadziejko C. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice V CAPs
exacerbate aortic plaque development in hyperlipidemic mice. Inhal Toxicol, 17: 217-224. 087219
Chen LH; Knutsen SF; Shavlik D; Beeson WL; Petersen F; Ghamsary M; Abbey D. (2005). The association between fatal
coronary heart disease and ambient particulate air pollution: Are females at greater risk?. Environ Health Perspect,
113: 1723-1729. 087942
Delfino RJ; Gone H; Linn WS; Pellizzari ED; Hu Y. (2003). Asthma symptoms in Hispanic children and daily ambient
exposures to toxic and criteria air pollutants. Environ Health Perspect, 111: 647-656. 050460
Delfino RJ; Zeiger RS; Seltzer JM; Street DH; McLaren CE. (2002). Association of asthma symptoms with peak
particulate air pollution and effect modification by anti-inflammatory medication use. Environ Health Perspect,
110: A607-A617. 093740
Desqueyroux H; Pujet J-C; Prosper M; Squinazi F; Momas I. (2002). Short-term effects of low-level air pollution on
respiratory health of adults suffering from moderate to severe asthma. Environ Res, 89: 29-37. 026052
Devlin RB; Ghio AJ; Kehrl H; Sanders G; Cascio W. (2003). Elderly humans exposed to concentrated air pollution
particles have decreased heart rate variability. Eur Respir J, 40: 76S-80S. 087348
De Leon SF; Thurston GD; Ito K. (2003). Contribution of respiratory disease to nonrespiratory mortality associations with
air pollution. Am J Respir Grit Care Med, 167: 1117-1123. 055688
Diaz-Sanchez D; Tsien A; Fleming J; Saxon A. (1997). Combined diesel exhaust particulate and ragweed allergen
challenge markedly enhances human in vivo nasal ragweed-specific IgE and skews cytokine production to a T
helper cell 2-type pattern. J Immunol, 158: 2406-2413. 051247
Dubowsky SD; Suh H; Schwartz J; Coull BA; Gold DR. (2006). Diabetes, obesity, and hypertension may enhance
associations between air pollution and markers of systemic inflammation. Environ Health Perspect, 114: 992-998.
088750
Dutton DB; Levine S. (1989). Overview, methodological critique, and reformulation. In JP Bunker; DS Gomby; BH Kehrer
(Ed.),Pathways to health (pp. 29-69). Menlo Park, CA: The Henry J. Kaiser Family Foundation. 192052
Fedulov AV; Leme A; Yang Z; Dahl M; Lim R; Mariani TJ; Kobzik L. (2008). Pulmonary exposure to particles during
pregnancy causes increased neonatal asthma susceptibility. Am J Respir Cell Mol Biol, 38: 57-67. 097482
Firestone M; Moya J; Cohen-Hubal E; Zartarian V; Xue J. (2007). Identifying childhood age groups for exposure
assessments and monitoring. Risk Anal, 27: 701-714. 192071
Frampton MW; Stewart JC; Oberdorster G; Morrow PE; Chalupa D; Pietropaoli AP; Frasier LM; Speers DM; Cox C;
Huang LS; Utell MJ. (2006). Inhalation of ultrafine particles alters blood leukocyte expression of adhesion
molecules in humans. Environ Health Perspect, 114: 51-58. 088665
Franklin M; Koutrakis P; Schwartz J. (2008). The role of particle composition on the association between PM2.5 and
mortality. Epidemiology, 19: 680-689. 155779
Franklin M; Zeka A; Schwartz J. (2007). Association between PM2.5 and all-cause and specific-cause mortality in 27 US
communities. JExpo Sci Environ Epidemiol, 17: 279-287. 091257
Fung KY; Luginaah I; Gorey KM; Webster G. (2005). Air pollution and daily hospital admissions for cardiovascular
diseases in Windsor, Ontario. Can J Public Health, 96: 29-33. 074322
Fung KY; Luginaah IKMG; Webster G. (2005). Air pollution and daily hospitalization rates for cardiovascular and
respiratory diseases in London, Ontario. Int J Environ Stud, 62: 677-685. 093262
December 2009 8-18
-------
Gent JF; Triche EW; Holford TR; Belanger K; Bracken MB; Beckett WS; Leaderer BP. (2003). Association of low-level
ozone and fine particles with respiratory symptoms in children with asthma. JAMA, 290: 1859-1867. 052885
Gilliland FD; Li YF; Saxon A; Diaz-Sanchez D. (2004). Effect of glutathione-S-transferase Ml and PI genotypes on
xenobiotic enhancement of allergic responses: randomised, placebo-controlled crossover study. Lancet, 363: 119-
125. 156471
Gilliland FD; McConnell R; Peters J; Gong Jr H. (1999). A Theoretical Basis for Investigating Ambient Air Pollution and
Children's Respiratory Health. Environ Health Perspect, 107: 403-407. 155792
Goldberg MS; Burnett RT; Yale JF; Valois MF; Brook JR. (2006). Associations between ambient air pollution and daily
mortality among persons with diabetes and cardiovascular disease. Environ Res, 100: 255-267. 088641
Gong H Jr; Linn WS; Clark KW; Anderson KR; Sioutas C; Alexis NE; Cascio WE; Devlin RB. (2008). Exposures of
healthy and asthmatic volunteers to concentrated ambient ultrafine particles in Los Angeles. Inhal Toxicol, 20: 533-
545. 156483
Gong H Jr; Linn WS; Sioutas C; Terrell SL; Clark KW; Anderson KR; Terrell LL. (2003). Controlled exposures of healthy
and asthmatic volunteers to concentrated ambient fine particles in Los Angeles. Inhal Toxicol, 15: 305-325. 042106
Gong H Jr; Linn WS; Terrell SL; Clark KW; Geller MD; Anderson KR; Cascio WE; Sioutas C. (2004). Altered heart-rate
variability in asthmatic and healthy volunteers exposed to concentrated ambient coarse particles. Inhal Toxicol, 16:
335-343. 055628
Heidenfelder BL; Reif DM; Harkema JR; Cohen Hubal EA; Hudgens EE; Bramble LA; Wagner JG; Morishita M; Keeler
GJ; Edwards SW; Gallagher JE. (2009). Comparative microarray analysis and pulmonary changes in brown
Norway rats exposed to ovalbumin and concentrated air particulates. Toxicol Sci, 108: 207-221. 190026
Host S; Larrieu S; Pascal L; Blanchard M; Declercq C; Fabre P; Jusot JF; Chardon B; Le Tertre A; Wagner V; Prouvost H;
Lefranc A. (2007). Short-term Associations between Fine and Coarse Particles and Cardiorespiratory
Hospitalizations in Six French Cities. Occup Environ Med, 18: S107-S108. 155851
Jerrett M; Burnett RT; Brook J; Kanaroglou P; Giovis C; Finkelstein N; Hutchison B. (2004). Do socioeconomic
characteristics modify the short term association between air pollution and mortality? Evidence from a zonal time
series in Hamilton, Canada. J Epidemiol Community Health, 58: 31-40. 087379
Kan H; London SJ; Chen G; Zhang Y; Song G; Zhao N; Jiang L; Chen B. (2008). Season, sex, age, and education as
modifiers of the effects of outdoor air pollution on daily mortality in Shanghai, China: The Public Health and Air
Pollution in Asia (PAPA) Study. Environ Health Perspect, 116: 1183-1188. 156621
Kauffmann F; Anto J; Baur MP; Bickeboller H; Clayton D; Cookson WOC; Demenais F; Helms PJ; Humphery-Smith I;
Imbeaud S; Knoppers BM; Lathrop M; Little J; Pearce N; Schaid D; Silverman E; Weiss S; Wjst M. (2004). Post-
genome respiratory epidemiology: a multidisciplinary challenge. Eur Respir J, 24: 471-480. 090968
Kleeberger SR; Ohtsuka Y. (2005). Gene-particulate matter-health interactions. Toxicol Appl Pharmacol, 207: S276-S281.
130489
Kodavanti UP; Moyer CF; Ledbetter AD; Schladweiler MC; Costa DL; Hauser R; Christiani DC; Nyska A. (2003). Inhaled
environmental combustion particles cause myocardial injury in the Wistar Kyoto rat. Toxicol Sci, 71: 237-245.
051325
Krewski D; Jerrett M; Burnett RT; Ma R; Hughes E; Shi Y; Turner MC; Pope AC III; Thurston G; Calle EE; Thun MJ.
(2009). Extended follow-up and spatial analysis of the American Cancer Society study linking particulate air
pollution and mortality. Health Effects Institute. Cambridge, MA. Report Nr. 140. 191193
Lagorio S; Forastiere F; Pistelli R; lavarone I; Michelozzi P; Fano V; Marconi A; Ziemacki G; Ostro BD. (2006). Air
pollution and lung function among susceptible adult subjects: a panel study. Environ Health, 5:11. 089800
Larrieu S; Jusot J-F; Blanchard M; Prouvost H; Declercq C; Fabre P; Pascal L; Le Tertre A; Wagner V; Riviere S; Chardon
B; Borelli D; Cassadou S; Eilstein D; Lefranc A. (2007). Short term effects of air pollution on hospitalizations for
cardiovascular diseases in eight French cities: The PSAS program. Sci Total Environ, 387: 105-112. 093031
Lee IM; Tsai SS; Ho CK; Chiu HF; Wu TN; Yang CY (2008). Air pollution and hospital admissions for congestive heart
failure: are there potentially sensitive groups?. Environ Res, 108: 348-353. 192076
December 2009 8-19
-------
Le Tertre A; Medina S; Samoli E; Forsberg B; Michelozzi P; Boumghar A; Vonk JM; Bellini A; Atkinson R; Ayres JG;
Sunyer J; Schwartz J; Katsouyanni K. (2002). Short term effects of participate air pollution on cardiovascular
diseases in eight European cities. J Epidemiol Community Health, 56: 773-779. 023746
Liao D; Heiss G; Chinchilli VM; Duan Y; Folsom AR; Lin HM; Salomaa V. (2005). Association of criteria pollutants with
plasma hemostatic/inflammatory markers: a population-based study. J Expo Sci Environ Epidemiol, 15: 319-328.
088677
Liu L; Ruddy TD; Dalipaj M; Szyszkowicz M; You H; Poon R; Wheeler A; Dales R. (2007). Influence of personal
exposure to particulate air pollution on cardiovascular physiology and biomarkers of inflammation and oxidative
stress in subjects with diabetes. J Occup Environ Med, 49: 258-265. 156705
Lopes FD; Pinto TS; Arantes-Costa FM; Moriya HT; Biselli PJ; Ferraz LF; Lichtenfels AJ; Saldiva PH; Mauad T; Martins
MA. (2009). Exposure to ambient levels of particles emitted by traffic worsens emphysema in mice. Environ Res,
109:544-551. 190430
Luginaah IN; Fung KY; Gorey KM; Webster G; Wills C. (2005). Association of ambient air pollution with respiratory
hospitalization in a government designated "area of concern": the case of Windsor, Ontario. Environ Health
Perspect, 113: 290-296. 057327
Mar TF; Larson TV; Stier RA; Claiborn C; Koenig JQ. (2004). An analysis of the association between respiratory
symptoms in subjects with asthma and daily air pollution in Spokane, Washington. Inhal Toxicol, 16: 809-815.
057309
Mauad T; Rivero DH; de Oliveira RC; Lichtenfels AJ; Guimaraes ET; de Andre PA; Kasahara DI; Bueno HM; Saldiva PH.
(2008). Chronic exposure to ambient levels of urban particles affects mouse lung development. Am J Respir Crit
Care Med, 178: 721-728. 156743
Merriam-Webster. (2009). Merriam-Webster on-line. Retrieved 15-JUN-09, from http://www.merriam-
webster.com/medical/susceptible; http://www.merriam-webster.com/medical/vulnerable. 192146
Metzger KB; Tolbert PE; Klein M; Peel JL; Flanders WD; Todd KH; Mulholland JA; Ryan PB; Frumkin H. (2004).
Ambient air pollution and cardiovascular emergency department visits. Epidemiology, 15: 46-56. 044222
Middleton N; Yiallouros P; Kleanthous S; Kolokotroni O; Schwartz J; Dockery DW; Demokritou P; Koutrakis P. (2008). A
10-year time-series analysis of respiratory and cardiovascular morbidity in Nicosia, Cyprus: the effect of short-term
changes in air pollution and dust storms. Environ Health, 7:39. 156760
Mills NL; Robinson SD; Fokkens PH; Leseman DL; Miller MR; Anderson D; Freney EJ; Heal MR; Donovan RJ;
Blomberg A; Sandstrom T; MacNee W; Boon NA; Donaldson K; Newby DE; Cassee FR. (2008). Exposure to
concentrated ambient particles does not affect vascular function in patients with coronary heart disease. Environ
Health Perspect, 116: 709-715. 156766
Mills NL; Tornqvist H; Gonzalez MC; Vink E; Robinson SD; Soderberg S; Boon NA; Donaldson K; Sandstrom T;
Blomberg A; Newby DE. (2007). Ischemic and thrombotic effects of dilute diesel-exhaust inhalation in men with
coronary heart disease. N Engl J Med, 357: 1075-1082. 091206
Morishita M; Keeler G; Wagner J; Marsik F; Timm E; Dvonch J; Harkema J. (2004). Pulmonary retention of particulate
matter is associated with airway inflammation in allergic rats exposed to air pollution in urban Detroit. Inhal
Toxicol, 16: 663-674. 087979
Mortimer KM; Neas LM; Dockery DW; Redline S; Tager IB. (2002). The effect of air pollution on inner-city children with
asthma. Eur Respir J, 19: 699-705. 030281
Nadziejko C; Fang K; Narciso S; Zhong M; Su WC; Gordon T; Nadas A; Chen LC. (2004). Effect of particulate and
gaseous pollutants on spontaneous arrhythmias in aged rats. Inhal Toxicol, 16: 373-380. 055632
Naess O; Nafstad P; Aamodt G; Claussen B; Rosland P. (2007). Relation between concentration of air pollution and cause-
specific mortality: four-year exposures to nitrogen dioxide and particulate matter pollutants in 470 neighborhoods
in Oslo, Norway. Am J Epidemiol, 165: 435-443. 090736
NCHS. (2006). Chartbook on Trends in the Health of Americans. National Center for Health Statistics, Centers for Disease
Control and Prevention. Hyattsville, MD.http://www.cdc.gov/nchs/data/hus/hus06.pdf. 198921
Nordenhall C; Pourazar J; Ledin M-C; Levin J-O; Sandstrom T; Adelroth E. (2001). Diesel exhaust enhances airway
responsiveness in asthmatic subjects. Eur Respir J, 17: 909-915. 025185
December 2009 8-20
-------
NRC. (2004). Research priorities for airborne participate matter: IV Continuing research progress. National Academies
Press. Washington, DC. 156814
Ohtsuka Y; Brunson KJ; Jedlicka AE; Mitzner W; Clarke RW; Zhang L-Y; Eleff SM; Kleeberger SR. (2000). Genetic
linkage analysis of susceptibility to particle exposure in mice. Am J Respir Cell Mol Biol, 22: 574-581. 004409
Ostro B; Broadwin R; Green S; Feng W-Y; Lipsett M. (2006). Fine particulate air pollution and mortality in nine California
counties: results from CALFINE. Environ Health Perspect, 114: 29-33. 087991
Ostro BD; Feng WY; Broadwin R; Malig BJ; Green RS; Lipsett MJ. (2008). The impact of components of fine particulate
matter on cardiovascular mortality in susceptible subpopulations. Occup Environ Med, 65: 750-756. 097971
Park SK; O'Neill MS; Vokonas PS; Sparrow D; Schwartz J. (2005). Effects of air pollution on heart rate variability: The VA
normative aging study. Environ Health Perspect, 113: 304-309. 057331
Park SK; O'Neill MS; Wright RO; Hu H; Vokonas PS; Sparrow D; Suh H; Schwartz J. (2006). HFE genotype, particulate
air pollution, and heart rate variability; a gene-environmental interaction. Circulation, 114: 2798-2805. 091245
Peel JL; Metzger KB; Klein M; Flanders WD; Mulholland JA; Tolbert PE. (2007). Ambient air pollution and
cardiovascular emergency department visits in potentially sensitive groups. Am J Epidemiol, 165: 625-633. 090442
Peel JL; Tolbert PE; Klein M; Metzger KB; Flanders WD; Knox T; Mulholland JA; Ryan PB; Frumkin H. (2005). Ambient
air pollution and respiratory emergency department visits. Epidemiology, 16: 164-174. 056305
Peters A; Greven S; Heid I; Baldari F; Breitner S; Bellander T; Chrysohoou C; Illig T; Jacquemin B; Koenig W. (2009).
Fibrinogen genes modify the fibrinogen response to ambient particulate matter. Am J Respir Grit Care Med, 179:
484-491. 191992
Pinkerton KE; Zhou Y; Teague SV; Peake JL; Walther RC; Kennedy IM; Leppert VJ; Aust AE. (2004). Reduced lung cell
proliferation following short-term exposure to ultrafine soot and iron particles in neonatal rats: key to impaired lung
growth?. Inhal Toxicol, 1: 73-81. 087465
Pinkerton KE; Zhou Y; Zhong C; Smith KR; Teague SV; Kennedy IM; Menache MG (2008). Mechanisms of particulate
matter toxicity in neonatal and young adult rat lungs. Health Effects Institute. Boston, MA. 135. 190471
Pleis JR; Lethbridge-Cejku M. (2007). Summary health statistics for US adults: National Health Interview Survey, 2006.
National Centr for Health Statistics, Centers for Disease Control and Prevention, U.S. Department of Health and
Human Services. Hyattsville, Maryland. Vital Health Stat Series 10, Number 235; DHHS 2008-1563.
http://www.cdc.gov/nchs/data/series/sr_l 07srlO_235.pdf. 156875
Pope C; Renlund D; Kfoury A; May H; Home B. (2008). Relation of heart failure hospitalization to exposure to fine
particulate air pollution. Am J Cardiol, 102: 1230-1234. 191969
Pope CA; Dockery DW. (2006). Health effects of fine particulate air pollution: Lines that connect. J Air Waste Manag
Assoc, 56: 709-742. 156881
Pope CAIII; Muhlestein JB; May HT; Renlund DG; Anderson JL; Home BD. (2006). Ischemic heart disease events
triggered by short-term exposure to fine particulate air pollution. Circulation, 114: 2443-2448. 091246
Rabinovitch N; Strand M; Gelfand EW (2006). Particulate levels are associated with early asthma worsening in children
with persistent disease. Am J Respir Grit Care Med, 173: 1098-1105. 088031
Routledge HC; Manney S; Harrison RM; Ayres JG; Townend JN. (2006). Effect of inhaled sulphur dioxide and carbon
particles on heart rate variability and markers of inflammation and coagulation in human subjects. Heart, 92: 220-
227. 088674
Samoli E; Peng R; Ramsay T; Pipikou M; Touloumi G; Dominici F; Burnett R; Cohen A; Krewski D; Samet J. (2008).
Acute effects of ambient particulate matter on mortality in Europe and North America: results from the APHENA
study. Environ Health Perspect, 116: 1480-1486. 188455
Schneider A; Neas L; Herbst M; Case M; Williams R; Cascio W; Hinderliter A; Holguin F; Buse J; Dungan K. (2008).
Endothelial dysfunction: associations with exposure to ambient fine particles in diabetic individuals. Environ
Health Perspect, 116: 1666-1674. 191985
Schwartz J; Park SK; O'Neill MS; Vokonas PS; Sparrow D; Weiss S; Kelsey K. (2005). Glutathione-S-transferase Ml,
obesity, statins, and autonomic effects of particles: gene-by-drug-by-environment interaction. Am J Respir Grit Care
Med, 172: 1529-1533. 086296
December 2009 8-21
-------
Silkoff PE; Zhang L; Dutton S; Langmack EL; Vedal S; Murphy J; Make B. (2005). Winter air pollution and disease
parameters in advanced chronic obstructive pulmonary disease panels residing in Denver, Colorado. J Allergy Clin
Immunol, 115: 337-344. 087471
Suh YJ; Ha EH; Park H; Kim YJ; Kim H; Hong YC. (2008). GSTM1 polymorphism along with PM10 exposure contributes
to the risk of preterm delivery. Mutat Res, 656: 62-67. 192077
Sun Q; Wang A; Jin X; Natanzon A; Duquaine D; Brook RD; Aguinaldo JG; Fayad ZA; Fuster V; Lippmann M; Chen LC;
Rajagopalan S. (2005). Long-term air pollution exposure and acceleration of atherosclerosis and vascular
inflammation in an animal model. JAMA, 294: 3003-3010. 087952
Sun Q; Yue P; Deiuliis JA; Lumeng CN; Kampfrath T; Mikolaj MB; Cai Y; Ostrowski MC; Lu B; Parthasarathy S; Brook
RD; Moffatt-Bruce SD; Chen LC; Rajagopalan S. (2009). Ambient air pollution exaggerates adipose inflammation
and insulin resistance in a mouse model of diet-induced obesity. Circulation, 119: 538-546. 190487
Sun Q; Yue P; Kirk RI; Wang A; Moatti D; Jin X; Lu B; Schecter AD; Lippmann M; Gordon T; Chen LC; Rajagopalan S.
(2008). Ambient air particulate matter exposure and tissue factor expression in atherosclerosis. Inhal Toxicol, 20:
127-137. 157033
Tankersley CG; Bierman A; Rabold R. (2007). Variation in heart rate regulation and the effects of particle exposure in
inbred mice. Inhal Toxicol, 19: 621-629. 097910
Tankersley CG; Campen M; Bierman A; Flanders SE; Broman KW; Rabold R. (2004). Particle effects on heart-rate
regulation in senescent mice. Inhal Toxicol, 16: 381-390. 094378
Tankersley CG; Champion HC; Takimoto E; Gabrielson K; Bedja D; Misra V; El-Haddad H; Rabold R; Mitzner W. (2008).
Exposure to inhaled particulate matter impairs cardiac function in senescent mice. Am J Physiol Regul Integr Comp
Physiol, 295: R252-R263. 157043
Tarantini L; Bonzini M; Apostoli P; Pegoraro V; Bollati V; Marinelli B; Cantone L; Rizzo G; Hou L; Schwartz J; Bertazzi
PA; Baccarelli A. (2009). Effects of particulate matter on genomic DNA methylation content and iNOS promoter
methylation. Environ Health Perspect, 117: 217-222. 192010
Trenga CA; Sullivan JH; Schildcrout JS; Shepherd KP; Shapiro GG; Liu LJ; Kaufman JD; Koenig JQ. (2006). Effect of
particulate air pollution on lung function in adult and pediatric subjects in a Seattle panel study. Chest, 129: 1614-
1622. 155209
U.S. Census Bureau. (2000). Census Bureau projects doubling of nation's population by 2100. Retrieved 13-MAY-01, from
http://www.census.gov/Press-Release/www/2000/cbOO-05.html. 157064
U.S. Census. (2009). Educational Attainment in the United States: 2007 (http://www.census.gov/prod/2009pubs/p20-
560.pdf). U.S. Genus Bureau, Economics and Statistics Administration, U.S. Department of Commerce.
Washington, DC. P20-560. http://www.census.gov/prod/2009pubs/p20-560.pdf. 192148
U.S. Census. (2009). U.S. census (2000): poverty. Retrieved Ol-JUL-09, from
http://www.census.gov/hhes/www.porvety/hispov/hstpov2.html. 192147
U.S. EPA. (2003). Framework for cumulative risk assessment. National Center for Environmental Assessment, Office of
Research and Development, U.S. Environmental Protection Agency. Washington, DC. EPA/630/P-02/001F. 192145
U.S. EPA. (2004). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research Triangle
Park, NC. EPA/600/P-99/002aF-bF. 056905
U.S. EPA. (2006). Aging and Toxic Response: Issues Relevant to Risk Assessment (Final). U.S. Environmental Protection
Agency. Washington, DC.http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=l 56648. 192082
U.S. EPA. (2008). Integrated review plan for the national ambient air quality standards for particulate matter. U.S.
Environmental Protection Agency, Office of Research and Development, National Center for Environmental
Assessment. Research Triangle Park, NC. 157072
U.S. EPA. (2008). Integrated Science Assessment for Sulfur Oxides - Health Criteria. U.S. Environmental Protection
Agency. Research Triangle Park, NC. EPA/600/R-08/047F.
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=l 98843. 157075
U.S. EPA. (2009). Susceptible subpopulations. Retrieved 15-MAY-09, from
http://www.epa.gov/nerl/goals/health/populations.html. 192149
December 2009 8-22
-------
Villeneuve PJ; Goldberg MS; Krewski D; Burnett RT; Chen Y. (2002). Fine particulate air pollution and all-cause mortality
within the Harvard six-cities study: variations in risk by period of exposure. Ann Epidemiol, 12: 568-576. 042576
Wellenius GA; Bateson TF; Mittleman MA; Schwartz J. (2005). Particulate air pollution and the rate of hospitalization for
congestive heart failure among medicare beneficiaries in Pittsburgh, Pennsylvania. Am J Epidemiol, 161: 1030-
1036. 087483
Wellenius GA; Coull BA; Batalha JRF; Diaz EA; Lawrence J; Godleski JJ. (2006). Effects of ambient particles and carbon
monoxide on supraventricular arrhythmias in a rat model of myocardial infarction. Inhal Toxicol, 18: 1077-1082.
156152
Wellenius GA; Coull BA; Godleski JJ; Koutrakis P; Okabe K; Savage ST. (2003). Inhalation of concentrated ambient air
particles exacerbates myocardial ischemia in conscious dogs. Environ Health Perspect, 111: 402-408. 055691
Wellenius GA; Schwartz J; Mittleman MA. (2006). Particulate air pollution and hospital admissions for congestive heart
failure in seven United States cities. Am J Cardiol, 97: 404-408. 088748
Wilson WE; Mar TF; Koenig JQ. (2007). Influence of exposure error and effect modification by socioeconomic status on
the association of acute cardiovascular mortality with particulate matter in Phoenix. J Expo Sci Environ Epidemiol,
17: S11-S19. 157149
Yanosky JD; Schwartz J; Suh HH. (2008). Associations between measures of socioeconomic position and chronic nitrogen
dioxide exposure in Worcester, Massachusetts. J Toxicol Environ Health A, 71: 1593-1602. 192081
Yauk C; Polyzos A; Rowan-Carroll A; Somers CM; Godschalk RW; Van Schooten FJ; Berndt ML; Pogribny IP; Koturbash
I; Williams A; Douglas GR; Kovalchuk O. (2008). Germ-line mutations, DNA damage, and global
hypermethylation in mice exposed to particulate air pollution in an urban/industrial location. PNAS, 105: 605-610.
157164
Yeatts K; Svendsen E; Creason J; Alexis N; Herbst M; Scott J; Kupper L; Williams R; Neas L; Cascio W; Devlin RB;
Peden DB. (2007). Coarse particulate matter (PM25-10) affects heart rate variability, blood lipids, and circulating
eosinophils in adults with asthma. Environ Health Perspect, 115: 709-714. 091266
Yunginger JW; Reed CE; O'Connell EJ; Melton LJ 3rd; O'Fallon WM; Silverstein MD. (1992). A community-based study
of the epidemiology of asthma. Incidence rates, 1964-1983. Am Rev Respir Dis, 146: 888-894. 192074
Zanobetti A; Bind MAC; Schwartz J. (2008). Particulate air pollution and survival in a COPD cohort. Environ Health
Perspect, 7: 48. 156177
Zanobetti A; Schwartz J. (2002). Cardiovascular damage by airborne particles: are diabetics more susceptible?.
Epidemiology, 13: 588-592. 034821
Zanobetti A; Schwartz J. (2005). The effect of particulate air pollution on emergency admissions for myocardial infarction:
A multicity case-crossover analysis. Environ Health Perspect, 113: 978-982. 088069
Zeger S; Dominici F; McDermott A; Samet J. (2008). Mortality in the Medicare population and chronic exposure to fine
particulate air pollution in urban centers (2000-2005). Environ Health Perspect, 116: 1614-1619. 191951
Zeka A; Zanobetti A; Schwartz J. (2006). Individual-level modifiers of the effects of particulate matter on daily mortality.
Am J Epidemiol, 163: 849-859. 088749
December 2009 8-23
-------
Chapter 9. Welfare Effects
9.1. Introduction
This chapter is a synthesis and evaluation of the most policy-relevant science used to help
form the scientific foundation for review of the secondary (welfare-based) NAAQS aimed at
protecting against welfare effects of ambient airborne PM. Specifically, Chapter 9 assesses the
effects of atmospheric PM on the environment, including: (1) effects on visibility; (2) effects on
climate; (3) ecological effects; and (4) effects on materials. These sections initially highlight the
conclusions from the 2004 PM AQCD (U.S. EPA, 2004, 056905). followed by an evaluation of
recent publications and assessment of the expanded body of evidence. In some sections, few new
publications are available, and the discussion is primarily a brief overview of the key conclusions
from the previous review.
As discussed in Chapter 1, the effects of particulate NOX and SOX have recently been
evaluated in the ISA for Oxides of Nitrogen and Sulfur - Ecological Criteria (U.S. EPA, 2008,
157074). That ISA focused on the effects from deposition of gas- and particle-phase pollutants
related to ambient NOX and SOX concentrations that can lead to acidification and nutrient
enrichment, as well as on the potential for increased mercury methylation from SO42~ deposition.
Thus, emphasis in this document is placed on the effects of airborne PM on visibility and climate,
and on the deposition effects of PM constituents other than NOX and SOX, primarily metals and
carbonaceous compounds.
Chapter 2 of this assessment provides an integrative overview of the major welfare effects
evaluated. EPA's framework for causality, described in Chapter 1, is applied throughout the
evaluation and the causal determinations are highlighted.
9.2. Effects on Visibility
9.2.1. Introduction
In recent years, most visibility research involved characterizing visibility conditions and trends
over broad regional scales, improving the understanding of the atmospheric processes and pollutants
responsible for the regional impacts, and attribution of visibility-impairing pollutants to emission
sources, source types, and regions. The motivation for much of this work has come from the
visibility protection provisions of the 1977 Clean Air Act Amendments (CAAA) that called for the
development of regulations to address reduction of regional haze in 156 NPs and wilderness areas to
natural conditions, and from the subsequent Regional Haze Rule (RHR) promulgated in 1999 by
EPA in response to the CAA mandate. Implementation of the RHR entails planned emissions
reductions to reach natural haze conditions in these protected areas by 2064 in six 10-year planning
steps.
Haze conditions caused solely by PM from natural sources are generally much lower than
contemporary conditions. The largest difference is between natural and current conditions for the
inorganic salts ammonium sulfate and ammonium nitrate, with natural concentrations taken to be
just a few tenths of a (ig/m3 each (Trijonis et al., 1990, 157058). while current conditions of both
over large regions of the country are an order of magnitude or more larger (DeBell, 2006, 156388).
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009 9-1
-------
However, natural source PM can be substantial on an episodic basis for crustal mineral PM
components during high windblown dust conditions and for carbonaceous PM from biomass
combustion during wildfire and prescribed burning episodes. The need for information to generate
RHR implementation plans has resulted in extensive use of continental-scale air quality simulation
modeling and assessment of expanded ambient monitoring data sets.
Unlike the substantial remote-area visibility investigations that have been conducted in
response to the RHR, relatively little work on urban visibility effects has been done in recent years.
For example, there has been relatively little new research on the optical and human perceptual
aspects of atmospheric visibility over the last decade or more. These topics have been the subjects of
numerous earlier investigations that have been summarized in detail elsewhere (Latimer and Ireson,
1980, 035723: Middleton, 1952, 016324: Tombach and McDonald, 2004, 157054: Trijonis et al,
1990, 157058: U.S. EPA, 1979, 157065: Watson et al., 2002, 035623). including past criteria
documents on PM, SO2 and NOX (U.S. EPA, 1982, 017610: U.S. EPA, 1993, 017649: U.S. EPA,
2004, 056905).
In spite of this fact, the understanding of urban visibility conditions has continued to improve.
By applying a well established algorithm that relates PM and haze conditions to data currently
collected from routine filter-based PM chemical speciation monitors located in numerous urban areas
(Jayanty, 2003, 156605). and to data collected from the more recently deployed high time- and
size-resolved PM speciation monitors located in several cities such as those in the PM Supersites
program (Solomon and Hopke, 2008, 156997) urban visibility conditions can be better
characterized. Comparisons between urban and remote area data in the same region afford the
opportunity to differentiate between regional and local visibility impacts. The availability of better
size and time resolution PM composition data, compared to that available from the routine
monitoring programs, reduces the number of simplifying assumptions required to estimate visibility
conditions in these areas, thereby reducing the uncertainty of the estimates. Thus, the state of the
science supporting urban visibility assessments continues to improve.
The background section below contains an overview of long-available information to help
provide context to the more recently published literature summarized in subsequent sections.
9.2.2. Background
Air pollution-induced visibility impairment is caused by the loss of image-forming light (i.e.,
signal) and the addition of non-image forming light (i.e., noise) between an observer and the object
being viewed. These changes to the light reaching the observer are a result of light being scattered
and absorbed by particles and gases in the sight path (see the schematic in Figure 9-1).
Electromagnetic theory developed to characterize the interaction of light with matter (Mie, 1908,
155983) permits the calculation of light scattering and absorption by particles and gas molecules
where the index of refraction and shape of particles by size are known (Van de Hulst, 1981, 191972).
The ability of human observers to visually detect distant objects or identify changes in their
appearance depends on the apparent contrast of the object against its background. The apparent
contrast is affected by changes in the physicochemical characteristics of the atmosphere caused by
air pollution as well as factors not related to air quality such as length of the sight path, scenic
lighting and the physical characteristics of the viewed object and other elements of the scene. To
rigorously determine the perceived visual effects of changes in the optical properties of the
atmosphere requires the use of radiative transfer modeling to determine changes in light from the
field of view experienced by the observer, followed by the use of psychophysical modeling to
determine the response to the light by the eye-brain system. The complexity of such an approach
discourages its common use.
Atmospheric light extinction is a fundamental atmospheric optics metric used to characterize
air pollution impacts on visibility. It is the fractional loss of intensity in a light beam per unit distance
due to scattering and absorption by the gases and particles in the air. Light extinction (6ext) can be
expressed as the sum of light scattering by particles (&s,p), scattering by gases (bStg), absorption by
particles (bAf) and absorption by gases (b^. Light extinction and its components are expressed in
units of inverse length, typically either inverse kilometers (km"1) or, as will be the convention in this
document, inverse megameters (Mm"1). Traditionally, for visibility-protection applications, the most
sensitive portion of the spectrum for human vision (550 nm) has been used to characterize light
extinction and its components.
December 2009 9-2
-------
Light from clouds
scattered into
sight path
Image-forming
light scattered f
out of sight path
Sunlight ^- ^
scattered Ljght ref,eet^d
from ground
scattered into
sight path
Image-forming
light absorbed
Source: Malm (1999, 0250371.
Figure 9-1. Important factors involved in seeing a scenic vista are outlined. Image-forming
information from an object is reduced (scattered and absorbed) as it passes
through the atmosphere to the human observer. Air light is also added to the
sight path by scattering processes. Sunlight, light from clouds, and
ground-reflected light all impinge on and scatter from particulates located in the
sight path. Some of this scattered light remains in the sight path, and at times it
can become so bright that the image essentially disappears. A final important
factor in seeing and appreciating a scenic vista are the characteristics of the
human observer.
A parametric analysis has shown that a constant fractional change in light extinction results in
a similar perceptual change regardless of certain baseline conditions (Pitchford et al., 1990, 156871).
From this assessment, the deciview haze index, which is a log transformation of light extinction,
similar in many ways to the decibel index for acoustic measurements, was developed (Pitchford and
Malm, 1994, 044922). A one deciview (Idv) change is about a 10% change in light extinction, which
is a small change that is detectable for sensitive viewing situations. The haze index in deciview units
is an appropriate metric for expressing the extent of haze changes where the perceptibility of the
change is an issue. The RHR has adopted the deciview haze index as the metric for tracking
long-term haze trends of visibility-protected federal lands (U.S. EPA, 2001, 157068). Light
extinction and its components are more useful metrics for characterizing the apportionment of haze
to its pollutant components due to the approximately linear relationship between pollutant species
concentrations and their contributions to light extinction.
December 2009
9-3
-------
starlight
absorbed
moonlight
scattering
","*! by particles
starlight
scattered
urban light pollution
reflocttv* cloud
observer in » non-urban setting
r»ll«cliv« cloud
affected by
unnaturally
bright
nighttime
sky
moonlight f
scattering
by particles
S puticl* \,'
' Sid. X
starlight
absorbed
starlight
scattered
bright urban nighttime sky
> A particle
\ back
Figure 9-2. Schematic of remote-area (top) and urban (bottom) nighttime sky visibility
showing the effects of PM and light pollution.
Daytime visibility has dominated the attention of those who have studied the visibility effects
of air pollution, though nighttime visibility is also known to be affected by air pollution. Stargazing
is a popular human activity in urban and remote settings. The reduction in visibility of the night sky
is primarily dependent on the addition of light into the sight path, the brightness of the night sky, and
the reduction in contrast of stars against the background (see the schematic in Figure 9-2). These are
December 2009
9-4
-------
controlled by the addition of PM, which enhances scattering, and the addition of anthropogenic
sources of light. Scattering of anthropogenic light contributes to the "skyglow" within and over
populated areas, adding to the total sky brightness. The visual result is a reduction of the number of
visible stars and the disappearance of diffuse or subtle phenomena such as the Milky Way. The
extinction of starlight is a secondary and minor effect also caused by increased scattering and
absorption. Anthropogenic light sources include artificial outdoor lighting, which varies dramatically
across space. Natural sources include the Moon, planets, and stars that have a predictable rhythm
across time.
The nighttime visual environment has some important differences to note. Light sources and
ambient conditions are typically five to seven orders of magnitude dimmer at night than in sunlight.
Moonlight, like sunlight, introduces light throughout an observer's sight path at a constant angle. On
the other hand, dim starlight emanates from all over the celestial hemisphere while artificial lights
are concentrated in cities and illuminate the atmosphere from below. Sight paths are often inclined
upward at night as targets may be nearby terrain features or celestial phenomena. Extinction behaves
the same at night as during the day, lowering the contrast of scenes through scattering and
absorption; nevertheless the different light sources will yield variable changes in visibility as
compared to what has been established for the daytime scenario. Little research has been conducted
on nighttime visibility. Even if the air quality-visibility interactions are shown to be similar between
day and night settings, the human psychophysical response at night is expected to differ. Though
recent advances in the ability to instrument and quantify nighttime scenes (Duriscoe et al., 2007,
156411) have been made and can be utilized to evaluate nocturnal visibility, the state of the science
is not yet comparable to that associated with daytime visibility impairment. The remainder of this
document focuses exclusively on daytime visibility.
9.2.2.1. Non-PM Visibility Effects
Light extinction due to the gaseous components of the atmosphere is relatively well
understood and well estimated for any atmospheric conditions. Absorption of visible light by gases
in the atmosphere is primarily by NO2, and can be directly and accurately estimated from NO2
concentrations by multiplying by the absorption efficiency. Scattering by gases is described by the
Rayleigh scattering theory.
NO2 absorbs more light in the short wavelength blue portion of the spectrum than at longer
wavelengths. For this reason a plume or layer of NO2 removes more of the blue light from the scene
viewed through the layer giving a yellow or brown appearance to the layer or plume. This filtering of
blue light by NO2 can deepen the brown appearance of hazes over urban areas, although it is not the
sole cause of such discoloration (U.S. EPA, 1993, 017649). The photopic-weighted absorption
efficiency at the 550 nm wavelength is incorporated into the revised version of the algorithm for
estimating light extinction from aerosol data that is used for implementing the RHR (Pitchford et al.,
2007, 098066). However, NO2 is not routinely measured at any of the monitoring sites representing
visibility protected areas where its impacts are assumed to be inconsequential compared to those of
PM. At background concentrations NO2 absorption is generally less than five percent of the light
scattering by clean air (Rayleigh scattering), making it imperceptible. Plume visibility models are
available to assess both achromatic contrast and discoloration associated with NO2 light absorption,
for point source emissions (Latimer and Ireson, 1980, 035723; Seigneur et al., 1984, 156965).
9.2.2.2. PM Visibility Effects
Particle light extinction is more complex than that caused by gaseous components. PM is
responsible for most visibility impairment except under near-pristine conditions, where Rayleigh
scattering is the largest contributor to light extinction or in plumes of combustion sources that are
well-controlled for particulate emissions (e.g., coal-fired power plants with bag houses), where light
absorption by NO2 may dominate the light extinction.
Light-absorbing carbon (e.g., DE soot and smoke) and some crustal minerals are the only
commonly occurring airborne particle components that absorb light. All particles scatter light, and
generally particle light scattering is the largest of the four light extinction components. While a
larger particle scatters more light than a similar shaped smaller particle of the same composition, the
light scattered per unit of mass concentration (i.e., mass scattering efficiency in units of
Mm'Vfug/m3] which reduces to m2/g) is greatest for particles with diameters from -0.3-1.0 um. If
December 2009 9-5
-------
the index of refraction, particle shape and concentration as a function of particle size are well
characterized, Mie theory can be used to accurately calculate the light scattering and absorption by
those particles. However, it is rare that these particle properties are known, so assumptions are used
in place of missing information to develop a simplified calculation scheme that provides an estimate
of the particle light extinction from available data sets.
Particles composed of water soluble inorganic salts (i.e., ammoniated sulfate, ammonium
nitrate, sodium chloride, etc.) are hygroscopic in that they absorb water as a function of relative
humidity to form a liquid solution droplet. Aside from the chemical consequences of this water
growth, the droplets become larger when relative humidity increases, resulting in increased light
scattering. Hence, the same PM dry concentration produces more haze. Figure 9-3 shows the effect
of water growth as a function of relative humidity on light scattering for two size distributions of
ammonium nitrate and ammonium sulfate particles as well as for internal and external mixtures (i.e.,
mixed within the same particle and in separate particles, respectively) of the two components. This
figure illustrates a number of important points. The water growth effect is substantial with an
increase in light scattering by about a factor of 10 between 40% and 97% relative humidity for the
same dry particle concentrations. The amount of scattering is significantly dependent on the dry
particle size distribution. However the growth curves for ammonium sulfate, ammonium nitrate and
mixtures of the two particle components are similar at any of the dry particle size distributions.
Water growth curves are also available for sodium chloride, the major component in sea salt, which
is an important PM component at coastal locations.
0.1
JQ
'o
o
O
O)
c
I
"ro
o
0.01
0.001
NH,NO,
(NH4)2S04
- External Mixture
(55.5% nitrate + 44.5% sulfate)
Internal mixture
40 50 60 70 80 90 100
%RH
Source: Reprinted with Permission of the American Geophysical Union from Tang (1996,1570421.
Figure 9-3. Effect of relative humidity on light scattering by mixtures of ammonium nitrate
and ammonium sulfate.
Using Mie theory, the scattering and absorption of any wavelength of light by a particle of
known size and index of refraction (a function of the wavelength of light) can be calculated
(Van de Hulst, 1981, 191972). Particle density is used to convert the particle light extinction to its
mass extinction efficiency (i.e., the ratio of particle light extinction to its mass). To expand the
calculations from one particle at a time to the multitude of particles in ambient aerosol, information
December 2009
9-6
-------
about the aerosol size and composition distributions are needed. Aerosol mixture refers to how the
major components that make up the particles are mixed. Methods have been developed to treat
simple mixture models ranging from external mixtures where the various components are assumed to
be in separate particles, to multi-component mixtures where individual particles contain several
components (Ouimette and Flagan, 1982, 025047). The latter includes internally mixed particles
where two or more components are mixed within the particles, and layered aerosol with a core of
one component covered by a shell of another component.
The Mie theory solution for an external mixture can be simplified to a linear relationship
where the light extinction is the sum of the mass concentration of each species multiplied by its
specific mass extinction efficiency (Ouimette and Flagan, 1982, 025047). This formulation
promotes the concept of apportioning the light extinction among the various PM species. For
internally mixed aerosol, the light extinction response to adding or removing mass of any component
to the aerosol is dependent on how such changes would affect the particle size, density and index of
refraction distribution of the aerosol. However, a number of investigators have shown that the
differences among the calculated light extinction values using external and various internal mixture
assumptions are generally less than about 10% (Lowenthal et al., 1995, 045134; Ramsey, 1966,
013946: Sloane, 1983, 025039: Sloane, 1984, 025040: Sloane, 1986, 045954: Sloane and Wolff,
1985, 045953: Wolff, 1985, 044680). This provides a basis to accept the apportionment of light
extinction to the PM components calculated using an external mixture assumption as a meaningful
surrogate for their contributions.
Ambient aerosols are usually a complex and unknown combination of both internal and
external mixtures of the particle components. Despite these complexities, PM light scattering can be
accurately calculated for any relative humidity if the chemical composition as a function of dry
particle size is known (Hand et al., 2002, 190367: Malm and Pitchford, 1997, 002519). However,
most routinely available ambient monitoring programs do not include data with sufficient detail to
make such calculations. The IMPROVE network with its greater than 150 remote area monitoring
sites (DeBell, 2006, 156388) and the CSN (Jayanty, 2003, 156605) with its greater than 150 urban
area monitoring sites collect 24-h duration fine particle samples (PM2.5) that are analyzed for the
major PM components including SO42~, nitrate, and carbonaceous particulate. CSN also analyzes for
ammonium ion, but does not monitor coarse mass (PMi0_2.s), while IMPROVE measures coarse mass
but does not analyze for ammonium ion. Neither data set has sufficient size resolution to make Mie
theory calculations of light extinction, nor does either program routinely monitor NO2
concentrations, which would be required to calculate its contribution to light extinction by
absorption.
A simple algorithm similar in form to the linear equation that results from Mie theory applied
with an external mixture assumption is frequently used to estimate light extinction from the
concentrations of the major components. The concentration of each of the major aerosol components
is multiplied by a dry extinction efficiency value and for the hygroscopic components (e.g.,
ammoniated sulfate and ammonium nitrate) an additional multiplicative term to account for the
water growth to estimate that components contribution to light extinction. Both the dry extinction
efficiency and water growth terms are developed by some combination of empirical assessment and
theoretical calculation using typical particle size distributions associated with each of the major
aerosol components, and they are evaluated by comparing the algorithm estimates of light extinction
with coincident optical measurements. Summing the contribution of each component gives the
estimate of total light extinction. The most commonly used of these is referred to as the IMPROVE
algorithm because it was developed specifically to use the IMPROVE aerosol monitoring data and
was evaluated using IMPROVE optical measurements at the subset of sites that make those
measurements (Malm et al., 1994, 044920). The formula for the traditional IMPROVE algorithm is
shown below.
December 2009 9-7
-------
bext * 3 x /(/iff) x [Sulfate]
+ 3xf(RH)x[Nitrate]
+ 4 x [Organic Mass]
+ 10 x [Elemental Carbon]
+ 1 x [Fine Soil]
+ 0.6 x [Coarse Mass]
+ 10
Equation 9-1
Source: DeBell (2006.1563881
Light extinction (6ext) is in units of Mm"1, the mass concentrations of the components indicated
in brackets are in ug/m , andf(RH) is the unitless water growth term that depends on relative
humidity. The dry extinction efficiency for particulate organic mass is larger than those for
particulate SO42~ and nitrate principally because the density of the dry inorganic compounds is
higher than that assumed for the PM organic mass components. Since IMPROVE does not include
ammonium ion monitoring, the assumption is made that all SO42~ is fully neutralized ammonium
sulfate and all nitrate is assumed to be ammonium nitrate. Though often reasonable, neither
assumption is always true (see Section 9.2.3.1). In the eastern U.S. during the summer there is
insufficient ammonia in the atmosphere to neutralize the SO42~ fully. Fine particle nitrates can
include sodium or calcium nitrate, which are the fine particle fraction of generally much coarser
particles due to nitric acid interactions with sea salt at near-coastal areas (sodium nitrate) or nitric
acid interactions with calcium carbonate in crustal aerosol (calcium nitrate). Despite the simplicity of
the algorithm, it performs reasonably well and permits the contributions to light extinction from each
of the major components (including the water associated with the SO42~ and nitrate compounds) to
be separately approximated.
Thef(RH) terms inflate the particulate SO42~ and nitrate light scattering for high relative
humidity conditions. For relative humidity below 40% thQ/(RH) value is 1, but it increases to 2 at
-66%, 3 at -83%, 4 at -90%, 5 at -93% and 6 at -95% relative humidity. The result is that both
particulate SO42~ and nitrate are more efficient per unit mass than any other aerosol component for
relative humidity above -85% where its total light extinction efficiency exceeds the 10m2/g
associated with EC. Based on this algorithm, particulate SO42~ and nitrate are estimated to have
comparable light extinction efficiencies (i.e., the same dry extinction efficiency andf(RH) water
growth terms), so on a per unit mass concentration basis at any specific relative humidity they are
treated as equally effective contributors to visibility effects. The strong relationship demonstrated
between dry light scattering and fine PM mass concentration or ambient light extinction and fine PM
mass concentration under low relative humidity conditions noted by a number of investigators
(Charlson et al, 1968, 095355; Chow et al, 2002, 037784; Chow et al., 2002, 036166; McMurry,
2000, 081517; Samuels et al., 1973, 070601; Waggoner and Weiss, 1980, 070152; Waggoner et al.,
1981, 095453) is reasonable based on this algorithm when the PM fractional composition is either
relatively constant or varies most among PM components with similar dry extinction efficiency
values (e.g., SO42~, nitrate and organic mass efficiencies).
9.2.2.3. Direct Optical Measurements
Light extinction and its components (i.e., scattering and absorption by particles and gases) can
be determined directly by optical measurements using commercially available instruments (Trijonis
et al., 1990, 157058). Though these measurements are all wavelength dependent, the convention for
visibility monitoring purposes is to make measurement at or near 550 nm, which is the wavelength
of maximum eye response. Direct PM light extinction, scattering and absorption measurements offer
a number of advantages compared to estimates using an algorithm applied to PM speciation data.
The direct optical measurements are considered more accurate because they do not depend on the
assumed particle characteristics (e.g., size, shape, density, component mixture, etc.) thought to be
associated with the major PM species. Also the optical measurements are made with high time
resolution (e.g., minutes to hourly) compared with the filter composition based estimates that are
December 2009 9-8
-------
typically 24-h duration, allowing the former to better characterize sub-daily temporal patterns which
can help in identifying influential source categories and characterize atmospheric phenomenon. The
higher time resolution attainable with direct light extinction measurements are also more
commensurate than the 24-h light extinction estimates from PM samples with the short exposure
time associated with perceived visibility effects.
Path-averaged light extinction can be determined by long-path transmissometers that monitors
the intensity of light that has traversed a known distance from a known initial intensity light source.
Transmission (i.e., the ratio of the final to the initial light intensity) is the natural logarithm of the
product of the path-averaged light extinction and the distance the light has traversed.
Transmissometer path-length establishes the useful range of light extinction over which the
measurements can be accurately measured, with path-lengths of 10 km or more required for pristine
conditions and <1 km more appropriate for hazier situations or to measure the visibility impacts
associated with fogs or precipitation events. The National Park Service (NPS) operated long-path
transmissometers at up to 25 locations from 1986 through 2004 (DeBell, 2006, 156388). but have
more recently discontinued their use at all but one remote area location due to the cost of
maintenance and the difficulties of performing calibration. Transmissometers are currently in routine
service at five urban areas.
A number of instruments measure the light scattered by particles and gases from a source of
known intensity. These include forward scattering, back scattering, polar, and integrating
nephelometers. Of these the integrating nephelometers with its high sensitivity and sample control
options has been more widely used for air quality-related visibility and PM monitoring purposes,
while the robust design of the open air forward scattering instruments have seen extensive use by the
National Weather Service (NWS) Automated Surface Observing System (ASOS) for characterizing
visibility principally for transportation safety purposes (NOAA, 1998). The potential utility of the
ASOS visibility network at about 900 locations for air quality monitoring has been established, but
the lack of resolution in the reported data is a serious impediment to this use of the data (Richards et
al., 1996. 190476).
Integrating nephelometers draw air into a sample chamber, making it possible to modify the
sample either by changing its humidity or controlling the particle size range that is measured. This
feature makes it possible to use sample-controlled nephelometers to investigate the effects of
ambient PM size and water growth characteristic on light scattering (Covert et al., 1972, 072055;
Malm and Day, 2001, 190431; Rood et al., 1987, 046397). For instance the coarse particle
contribution to light scattering can be estimated using a nephelometer that alternately samples
through a 2.5 um size selective inlet and a 10 urn size selective inlet. This separation by size may be
useful in that it would allow correction of the underestimated light scattering of larger particles due
to nephelometer angular truncation errors (Anderson and Ogren, 1998, 156213). For routine
monitoring, integrating nephelometers are typically either used to measure the PM component of
light scattering when operated at ambient relative humidity or to measure dry PM light scattering as
a high-time resolution surrogate for PM mass concentration when operated with a heater or other
sample air drier. Integrating nephelometers operated at ambient conditions by the IMPROVE
program have replaced the long-path transmissometer as the principal optical measurement at about
30 locations (DeBell, 2006, 156388).
PM light absorption can also be inferred from measured changes in the light transmitted
through a filter used to sample the PM compared to an identical clean filter (Bond et al., 1999,
156281). Such measurements can be made subsequent to sampling (Campbell et al., 1995, 190171)
or continuously during sampling by using specifically designed sampler (Hansen et al., 1982,
190368; Hansen et al., 1984, 002396). All of the filter-based methods require adjustments to the
optical measurements to account for filter and sampled particle light scattering effects associated
with particles concentrated on and within the matrix of the filters (Bond et al., 1999, 156281). Often
PM light absorption measurements are used to infer BC concentration by assuming it is the dominant
PM contributor to light absorption with a near constant absorption efficiency (Allen et al., 1999,
048923; Babich et al., 2000, 156239). In fact commercially available aethalometers incorporate an
absorption efficiency value so they can directly report BC concentrations. Like nephelometers,
commercially available aethalometers can be obtained with either single or multiple-wavelength
measurement capabilities, where the multi-wavelength data can be used to better characterize the
PM. More recently these have been used to distinguish BC that absorbs light strongly over the full
visible light spectrum (e.g., DE) from brown carbon that absorbs appreciably more at shorter
wavelengths than at long wavelengths (e.g., WS) (Andreae and Gelencser, 2006, 156215).
December 2009 9-9
-------
Other approaches to measure light absorption include a photoacoustic instrument that
measures the heating associated with absorbed light by suspended PM (Arnott et al., 1999, 020650;
Moosmuller et al., 1998, 020657), as well as by the difference between light extinction and light
scattering measurements (Bond et al., 1999, 156281).
9.2.2.4. Value of Good Visual Air Quality
The term visual air quality (VAQ) is used here to refer to the visibility effects caused solely by
air quality conditions. For example, it excludes the reduced visibility caused by fog. Two broadly
different approaches have traditionally been used to define and quantify the value of good VAQ. One
approach assesses the monetary value associated with visibility changes; the other assesses the
psychological value of visual air quality. With respect to the latter, reduced VAQ is considered an
environmental stressor (Campbell, 1983, 190172) that is associated with heightened amounts of
anxiety, tension, anger, fatigue, depression, and feelings of helplessness (Evans et al., 1987, 190347;
Zeidner and Shechter, 1988, 189973). Though the relationship between impaired VAQ and mental
health is poorly understood, there are greater emergency calls associated with psychiatric
disturbances during periods with reduced VAQ (Rotton and Frey, 1982, 190477). Studies have
shown that reduced VAQ affects people's behavior, including reductions in outdoor activities, and
increased hostility and aggressive behavior (Cunningham, 1979, 191974; Evans et al., 1982, 190521;
Jones and Bogat, 1978, 190396; Rotton et al., 1979, 190478).
The value of VAQ (both monetary and non-monetary) has been investigated in two broadly
different settings, non-recreational or urban settings and recreational settings, such as the NPs and
wilderness areas where visibility is protected by the RHR (Trijonis et al., 1990, 157058). In urban
settings, public surveys have shown that greater than 80% of the participants are aware of poor VAQ
conditions (Cohen et al., 1986, 190182). though attitudes towards poor VAQ have been shown to
vary by socio-economic status, health, and length of residence in the urban setting (Barker, 1976,
072137). The economic importance of urban visibility has been examined by a number of studies
designed to quantify the benefits (or willingness to pay) associated with potential improvements in
urban visibility. Urban visibility valuation research prior to 1997 was summarized in Chestnut and
Dennis (1997, 014525). and was also described in the 2004 PM AQCD (U.S. EPA, 2004, 056905)
and the 2005 PM Staff Paper (U.S. EPA, 2005, 090209). These reviews summarize 34 estimates
(based on different cities or model specifications) from six different studies. Since the mid 1990s,
however, only one new valuation study of urban visibility has been published (Beron et al., 2001,
156270) which is summarized below (Section 9.2.4.6).
In recreational settings, experience based demand models have been developed using on-site
and mail-in surveys to judge the relative importance to NP visitors of various park attributes
including good VAQ, to assess visitor awareness of VAQ conditions, and to explore possible
relationships between VAQ and visitor satisfaction (Ross et al., 1985, 044287; Ross et al., 1987,
037420). At the three western and two eastern NPs where this survey was conducted, visitors rated
the attribute identified as "clean, clear air" among the most important features of the parks. A
random sample of 1,800 visitors at one of the parks (Grand Canyon) showed that visitor awareness
of VAQ impacts increased as measured visibility conditions decreased, and that overall park
enjoyment and satisfaction decreased with reduced VAQ. Grand Canyon visitors when asked to
indicate how they would budget their time (e.g., between visiting an archaeological site or a view
lookout point) indicated that they would to be willing to significantly alter their behavior to
experience views under improved VAQ (Malm et al., 1984, 044292).
9.2.3. Monitoring and Assessment
Monitoring and the assessment of monitoring data serve a number of goals with regard to the
visibility effects of PM, including improving the understanding of the physio/chemical/optical
properties of the aerosol, characterizing spatial and temporal air quality patterns, and assessing the
causes (i.e., pollution sources and atmospheric processes) that are responsible for visibility
impairment. Information generated by special studies employing sophisticated instrumentation are
typically needed to advance the understanding of aerosol properties, while characterizing trends is
the product of analyzing routine monitoring data. Whereas, assessing the causes of haze usually
involves a weight-of-evidence approach applied to special study and/or routine monitoring data sets
December 2009 9-10
-------
plus the use of air quality simulation modeling. This section summarizes recently available
information that is based on monitoring data.
9.2.3.1. Aerosol Properties
Particle size is the most influential physical property of aerosols with respect to their dry light
extinction efficiency. Chemical composition by size is used to ascertain density (needed to convert
aerodynamic to physical size and to determine particle mass as a function of size) and to identify the
water growth characteristics of the aerosol (needed to calculate the particle size, density and index of
refraction at ambient RH). To characterize aerosol properties of interest for visibility effects, field
monitoring programs typically include particle size distribution monitoring, high size resolution
particle sampling with subsequent compositional analysis, and optical monitoring. These generate
data that permit optical closure assessments where the light scattering and/or light extinction
estimates from the aerosol data are compared to corresponding optical data. Since component
contributions to visibility are generally assessed by applying the IMPROVE or some similar
algorithm to measured or modeled aerosol concentration data, this section will include recent
investigations that evaluate or address various assumptions inherent in the use of these simple
algorithms.
One component of the Big Bend Regional Aerosol and Visibility Observational (BRAVO)
Study, conducted at Big Bend NP, TX in the summer and fall of 1999, entailed use of detailed
measurements of aerosol chemical composition, size distribution, water growth, and optical
properties to characterize the aerosol and assess the relationship between aerosol physical, chemical
and optical properties (Malm et al, 2003, 190434; Schichtel et al, 2004, 179902). Fine ammoniated
sulfate during the BRAVO Study was about half the fine particle mass concentration and was shown
to be responsible for about 35% of the light extinction. Rayleigh scattering was the second largest
contributor at about 25%, followed by coarse particle (about 18%), and organic compounds (about
13%). There was little fine particle nitrate (less than 5% of the mass concentration) and most of it is
apparently in the form of sodium nitrate and two thirds of it was found in the coarse mode where it
comprises about 8% of the coarse particle mass concentration. Both the composition of the nitrate
and the fact of much of it being in the coarse size mode (2.5 um >D>10 um) are inconsistent with
the implied assumptions of the IMPROVE algorithm.
A year-long special study of coarse particle speciation was conducted at nine IMPROVE
remote area monitoring sites during 2003-2004 to provide additional information about the
geographic and seasonal variations in coarse particle composition (Malm et al., 2007, 156730). The
same sampling and analytical methodologies procedures were used for the PMi0 samples as are
routinely used on the IMPROVE PM2.5 samples. The IMPROVE coarse particle speciation study did
not include ammonium analysis, so SO42~ and nitrate ions were assumed to be ammonium sulfate
and ammonium nitrate. As expected crustal minerals were the largest contributors to coarse mass
overall (about 60%), though at Mt. Rainier the fraction of coarse PM that was organic exceeded the
crustal mineral by nearly two to one (i.e., 59.2% compared to 33.5%) On average across sites the
organic particulate contributed significantly at about one quarter of the coarse mass, while
ammonium nitrate was the third largest contributor to coarse mass (about 8%). Sea salt was
negligible overall but high at the one coastal site (i.e., 12% at Brigantine, NJ). The two California
sites, San Gorgonio and Sequoia, had the highest coarse nitrate concentrations, 0.74 ug/m3 and
0.69 ug/m3, and high fine nitrates concentrations on average, 2.66 ug/m3 and 2.14 ug/m3,
respectively. Brigantine, a coastal site in New Jersey, had the highest fraction of total nitrate in the
coarse size range (36%). The authors speculate that Brigantine's particulate nitrate is likely sodium
nitrate, the result of nitric acid reactions with sodium chloride. The nine-site average fraction of total
nitrate in the coarse size range is 26%. By contrast, coarse SO42~ concentrations are small with only
about -1% of the total SO42~ in the coarse fraction.
Routine IMPROVE monitoring data include the mass concentration, but not the composition
of the coarse PM fraction, so the algorithm used to estimate light extinction does not include any
provision for varied coarse PM composition as shown in this study. This study shows that about 10%
of the coarse mass across the nine monitoring sites is composed of hygroscopic materials (i.e.,
ammonium sulfate, ammonium nitrate and sea salt), which during high humidity conditions will
scatter more light than estimated by the current algorithm (e.g., -20% bias at -90% relative
humidity). However, at coastal sites such as the Brigantine, NJ, IMPROVE site where the combined
concentration of the inorganic salts (i.e., sea salt, nitrate and SO42~) constitute a significant fraction
December 2009 9-11
-------
(-24% on average) of the coarse mass concentration, the IMPROVE algorithm underestimation of
light extinction by coarse PM can be significant for high relative humidity conditions (-60% at
-90% relative humidity). The resulting underestimation of total light extinction can be much smaller
since fine particle light extinction generally exceeds that contributed by coarse particles. Another
issue with regard to estimating light extinction from coarse PM concentration when the composition
is not crustal minerals, as has been assumed, has to do with the lower average density of the coarse
mode particles that results in greater particle numbers and/or larger particles and therefore a greater
light extinction efficiency (Malm and Hand, 2007, 155962).
Special studies with more complete, higher time resolution and size resolved particulate
inorganic ion species chemistry and precursor gases were conducted at seven of the nine sites with
IMPROVE coarse particle speciation monitoring (Lee et al., 2008, 156686). This work confirmed
the presence of sodium and calcium nitrate (referred to as mineral nitrate) primarily in the coarse
particle size range in addition to fine particle ammonium nitrate where low temperatures, high
humidity and excess ammonium (beyond that required to neutralize the particulate SO42~) favored
particle phase equilibrium. Figure 9-4 is a map snowing the locations and sample times and
estimated composition of the total particulate nitrate for the seven locations for this special study.
Sites with a high fraction of ammonium nitrate (e.g., San Gorgonio, Bondville, and Brigantine) have
the highest nitrate contributions to total mass concentration and haze, whereas sites with high
mineral nitrates tend to have low total nitrate contributions. This work shows that the common
assumption that particulate nitrate is in the fine particle size range and consists principally of
ammonium nitrate is not necessarily true.
Yossnte(July/Au
Brigantine (November, 2003)
0 NaNCh
0 Ca(NOs)2
San Gorgonio (July, 2003)
Source: Reprinted with Permission of Atmospheric Environment from Lee et al. (2008,156686)
Figure 9-4. Estimated fractions of total particulate nitrate during each field campaign
comprised of ammonium nitrate, reacted sea salt nitrate (shown as NaNOs), and
reacted soil dust nitrate (shown as Ca(NOs)2).
Extinction efficiencies for individual particle species can be theoretically calculated from
sized-resolved aerosol measurements and can be inferred using multiple linear regression applied to
aerosol composition and light extinction measurement data. In a recent publication, Hand and Malm
(2007, 155825) reviewed the literature since 1990 in which aerosol mass scattering efficiency values
were calculated or inferred. From these they have compiled normalized dry scattering efficiency
December 2009
9-12
-------
values for the individual species. Based on 93 separate determinations including marine, remote
continental and urban areas data sets, the average dry mass scattering efficiency for ammonium
sulfate is 2.5 ± 0.6 m2/g. Average values tended to be somewhat lower for the marine aerosol
(~2 m2/g) than for remote continental (-2.1 m2/g) and urban (2.6 m2/g) areas, and values also tended
to be lower for fairly clean arid locations compared with more humid polluted areas.
Based on 48 separate determinations including remote area and urban area data sets, the
average dry mass scattering efficiency for ammonium nitrate is 2.7 ± 0.5 m2/g (Hand and Malm,
2007, 155825). Average values were higher in remote locations (2.8 ± 0.5 m /g) compared to urban
locations (2.2 ± 0.5 m /g) though this might be accounted for by the predominate use of multiple
linear regression for the remote areas, which can be biased high, compared to the use of theoretical
calculations for the urban data sets.
Organic fine PM extinction efficiency of 3.9 ±1.5 m2/g is based on 58 separate
determinations, though much higher values (~6 m2/g) resulted for locations influenced by industrial
and biomass combustion sources (Hand and Malm, 2007, 155825). These organic fine PM
extinction efficiency values were adjusted to use a consistent ratio of organic mass to OC (OC) of
1.8 for each determination of the mass concentration. This value is generally associated with aged
organic PM, while for more freshly emitted PM, such as in an urban environment, a smaller ratio
(e.g., 1.4) would be more appropriate. This could explain the discrepancy between two approaches
used to estimate the organic PM light extinction efficiency for Phoenix (Hand and Malm, 2006,
156517). which resulted in a significantly lower value where a site specific regression method was
used compared to the value obtained from a method optimized for remote-area monitoring (2.47
m2/g compared to 3.71 m2/g). However in Fresno both the mass balance and light scattering balance
was improved by using a ratio of 1.8 instead of 1.4 to estimate the organic compound mass (Watson
and Chow, 2007, 157127). Another possible or partial factor with respect to urban light extinction
efficiency for organic PM may be that the size distribution of freshly emitted organic PM in urban
areas extends significantly into the ultra-fine particle size range (Demerjian and Mohnen, 2008,
156392) that is less efficient per mass concentration at light scattering than the generally larger-sized
aged organic PM such as from a distant forest fire as was measured at the Baltimore Supersite.
Hand and Malm (2007, 155825) also reviewed and made recommendations for extinction
efficiencies for the other PM components including mixed coarse mode (1.0 ± 0.9 m2/g based on 51
determinations) and fine mode dust or soil (3.3 ± 0.6 m2/g based on 23 determinations, but
recommending 1.0 m2/g for use with data from realistic collection efficiency samplers) and fine sea
salt (4.5 ±0.9 m2/g based on 25 determinations, but recommending 1.0 m2/g to 1.3 m2/g for use with
data from realistic collection efficiency samplers). This work did not address light absorption
efficiency of EC, CB, or crustal PM.
The Hand and Malm (2007, 155825) average dry mass light scattering efficiency values are
generally consistent with the values for the IMPROVE algorithm (as shown in Equation 9-1).
However the adoption of the IMPROVE algorithm by EPA for calculating the haze metric used to
track trends and assess the nominal pace of progress for the RHR (U.S. EPA, 2001, 157068) resulted
in much greater scrutiny of its performance in estimating extinction (Lowenthal and Kumar, 2003,
156712: Malm, 1999, 025037: Malm and Hand, 2007, 155962: Ryan et al, 2005, 156934). Among
the issues raised is that the algorithm tended to underestimate the light extinction for the haziest
conditions and overestimate light extinction for the clearest conditions in regions such as the
southeastern U.S., though it generally worked well in the arid western U.S. Furthermore, they
showed the lack of mass or light scattering closure at coastal sites due to sea salt that was not
accounted for by the IMPROVE algorithm. These assessments used mass concentration and light
extinction closure and regression analysis methods to infer that the dry extinction efficiency for the
major fine particle components would need to vary in order to avoid the biased estimates of light
extinction at the extremes. Theoretical calculations of SO42~ dry extinction efficiencies for 41 days of
size-resolved chemical composition data for Big Bend, TX as part of the BRAVO Study produced a
range of results from ~2.4 m2/g to ~4.1 m2/g, with the larger dry extinction efficiency values tending
to be associated with higher ammonium sulfate mass concentration and narrower size distributions
(Schichtel et al., 2004, 179902).
In response to the technical concerns raised about the performance of the IMPROVE
algorithm, a revised algorithm was developed (Pitchford et al., 2007, 098066). The revised version
of the algorithm differs from the original algorithm by: including a fine sea salt term related to the
measured chloride ion concentration; increasing by about 30% the mass concentration of the organic
aerosol component by changing the ratio of organic compound mass to OC mass from 1.4 to 1.8;
December 2009 9-13
-------
using site elevation dependent Rayleigh scattering in place of 10 Mm"1 that had been used at every
site; adding a NO2 light absorption term; and employing a split component model for the secondary
particulate components (i.e., SO42~, nitrate and organic species) with new water growth terms to
better estimate their extinction at the high and low extremes of the range. The revised algorithm is
displayed below in Equation 9-2.
bext * 2.2 x fa(Rti) x [Small Sulfati + 4.8 x fL(RH) x [LargeSuJfatd
+ 2.4 x fs(Rti) x [SmallNitrat^ + 5.1 x fL(Rti) x [Large Nitrati
+ 2.8 x [Small Organic Mas& + 6.1 x [Large Organic Mas^
+ 10 x [ElementalCarbori
+ 1 x [FineSoiA
+ 1.7 x4(/Z#) x [SeaSaM
+ 0.6 x [Coarse Massi
+ RayleighScattering (Site Specific)
+ 0.33 x [NO2 (ppb)}
Equation 9-2
Source: Reprinted with Permission of the Air & Wast Management Association from Pitchford et al. (2007.0980661
Small and large SO42~, nitrate and organic mass are used to refer to the splitting of the
concentrations of each of those three species into two size distributions. This approach accounts for
increased light extinction efficiency with mass by using a simple mixing model that assume that each
of these three components are comprised of an external mixture of small and large particle size
modes. Conceptually, the large mode particles represent aged or cloud-processed aerosol, while the
small mode particles represent relatively newly generated particles from the gas phase precursors.
The former are more likely to be associated with high concentrations while the latter are likely to be
at relatively low concentration.
The geometric mean diameter and standard deviations assumed for these two size modes are
0.5 um and 1.5 for the large mode particles and 0.2 urn and 2.2 for the small mode particles. Mie
theory applied to these size distributions for the three species results in dry extinction efficiencies for
the small and large mode ammonium sulfate (2.2 m2/g and 4.8 m2/g), ammonium nitrate (2.4 m2/g
and 5.1 m2/g) and organic mass (2.8 m2/g and 6.1 m2/g). Water growth terms specifically derived for
the small and large size distribution using the upper branch of the hygroscopic growth curves for
ammonium sulfate are applied to both the SO42~ and nitrate PM. No water growth is assumed for
organic PM.
A simple empirically developed apportionment approach that was evaluated by testing the new
algorithms estimated light scattering at the 21 IMPROVE sites that have nephelometer-measured
light scattering data. For each sample, the fraction of the fine particle component (SO42~, nitrate, or
organic mass) that is assigned to the large mode is calculated by dividing the total concentration of
the component by 20 ug/m3 (e.g., if the total fine particle nitrate concentration is 4 ug/m3, the large
mode concentration is 1/5 of 4 ug/m3 or 0.8 ug/m , leaving 3.2 ug/m3 in the small mode). If the total
concentration of a component exceeds 20 ug/m3, all of it is assumed to be in the large mode.
Figure 9-5 and Figure 9-6 are scatterplots of the estimated versus measured light scattering for
the two algorithms. The revised algorithm has noticeably reduced bias at the upper and lower
extremes. However, the new algorithm estimates have somewhat reduced precision (i.e., the points
are more broadly scattered). States have adopted the new algorithm for the technical assessments that
support their RHR State Implementation Plans, but the revised algorithm was too recently developed
to be incorporated into any of the peer-reviewed technical literature reported on below. In general the
differences resulting from use of the original versus the revised IMPROVE algorithm in identifying
best and worst haze conditions and the apportionment of the various PM components are small with
exception of coastal locations where sea salt may be a significant contributor.
December 2009 9-14
-------
350 -i
300
50
100
150 200
Measured Bsp
250
300
350
Source: Reprinted with Permission of the Air & Waste Management Associaiton from Pitchford, et al. (2007,
Figure 9-5. A scatter plot of the original IMPROVE algorithm estimated particle light
scattering versus measured particle light scattering.
100
250
300
350
150 200
Measured Bsp
Source: Reprinted with Permission of the Air & Waste Management Associaiton from Pitchford, et al. (2007,
Figure 9-6. Scatter plot of the revised algorithm estimates of light scattering versus
measured light scattering.
December 2009
9-15
-------
9.2.3.2. Spatial Patterns
The IMPROVE network is the basis for much of what is known about particulate species
spatial and temporal patterns for remote areas of the U.S. Though IMPROVE includes some urban
monitoring sites, most of what is known about urban particle speciation trends is based on the EPA
Speciation Trend Network (STN) and other similarly operated state particle speciation sites jointly
referred to as the Chemical Speciation Network (CSN) (Jayanty, 2003, 156605). The number of
IMPROVE network sites has increased considerably beginning in 2000, first to increase its ability to
generate data representative of the 156 visibility-protected NPs and wilderness areas, then later as
the states in the central U.S. requested additional remote-area monitoring to better understand their
contributions to regional haze. The expansion of the network into the central U.S. significantly
improved the understanding of spatial trends in a region of the country that had little speciation
monitoring. Except as otherwise noted most of the information in this section was from the
IMPROVE Report IV (DeBell, 2006, 156388) and displays of data that are readily generated using
the Visibility Information Exchange Web Site (VIEWS). VIEWS, the ambient monitoring data
system, is one of several websites (as described in Table 9-1) sponsored by the Regional Planning
Organizations (RPO) that documents substantial, though often otherwise unpublished, technical
information generated to support implementation of the RHR.
Figure 9-7 shows maps of remote area light extinction estimates from PM speciation data for
two years selected to demonstrate the additional information available due to the expansion of the
IMPROVE network into the central U.S. The locations of monitoring sites supplying the data shown
as color contours are shown as dot on the maps. Users of such contour maps are usually cautioned
that the contours are only there to guide the eye to sites with similar measurements and that nothing
should be implied about spatial patterns where there are no monitoring sites. Certainly these plots
give proof to the wisdom of such warnings. Prior to 2001 there were no IMPROVE or any other
remote-area aerosol speciation monitoring sites in the central states between northern Minnesota and
Michigan to the north and Arkansas and Kentucky to the south. The lack of monitoring over such a
large region in the center of the country hid the presence of high average regional haze over the
midwestern U.S. Smaller scale differences are seen in the rest of the country and some of those are
due to interannual variations as well as to better spatial resolution made possible by a more dense
monitoring network.
December 2009 9-16
-------
Table 9-1. Regional Planning Organization websites with visibility characterization and source
attribution assessment information.
Type of
Information
RPO Home Pages
Visibility -Air Quality
Monitoring Data
Emission Inventory
Data
Monitoring Data
Assessment
Visibility Modeling
Name and Web Address
Western Regional Air Partnership
http://www.wrapair.org/
Central Regional Air Planning Association
http://www.cenrap.org/
Midwest Regional Planning Organization
http://64.27.125.175/mrpo.html
Visibility Improvement State and Tribal
Association of the Southeast
http://www.vistas-sesarm.org/
Mid-Atlantic/Northeast Visibility Union
http://www.manevu.org/
http://www.nescaum.org/topics/regional-haze
http://www.marama.org/visibilitv/
Visibility Information Exchange Web Site
http://vista.cira.colostate.edu/views/
Emissions Data Management System
http://www.wrapedms.org/default login. asp
Causes of Haze Assessment
http://www.coha.dri.edu/
U. of California-Riverside Modeling Center
http://pah.cert.ucr.edu/agm/308/
http://pah.cert.ucr.edu/agm/cenrap/index.shtml
http://pah.cert.ucr.edu/vistas/
RPO
WRAP
CENRAP
MRPO
VISTAS
MANE-VU
NESCAUM
MARAMA
All RPOs
WRAP
WRAP
CENRAP
WRAP
CENRAP
VISTAS
Information Content and Comments
Organizational structure, plans, projects, reports and links to other sites with
additional information.
Air Use Management (NESCAU M) and Mid-Atlantic Regional Air
Management Association (MARAMA) to develop the technical information for
RHR in the Northeast. All three web sites contain unique technical support
information.
All IMPROVE and most other PM speciation data, RHR compatible derived
parameters, and user-friendly tools to summarize and display data.
WRAP emission inventory data warehouse and tools that provides a
consistent approach to regional emissions tracking
Monitoring site-specific descriptive characterizations and maps, seasonal and
trends analysis, air flow analysis, & receptor modeling.
Descriptions of input data, performance, and results of regional scale
modeling (CMAQ & CAMx) & source attribution for base and future year
regional haze.
Integrated Information Technical Support System
to Support RHR SIP
Preparations http://vista.cira.colostate.edu/tss/
WRAP
Provides access and common formats to display and summarize emissions
inventory information, monitoring data/ assessment and regional haze
modeling result to aid state and tribal analyst prepare RHR implementation
plans.
December 2009
9-17
-------
Figure 9-7.
1/Mtr
Source: VIEWS (http://vista.cira.colostate.edu/view5/)
IMPROVE network PM species estimated light extinction for 2000 (left) and for
2004 (right).
Source: VIEWS (http://vista.cira.colostate.edu/views/)
Figure 9-8.
Mean estimated light extinction from PM speciation measurements for the first
(top left), second (top right), third (bottom left), and fourth (bottom right) calendar
quarters of 2004.
Figure 9-8 shows the seasonal pattern of PM species estimated light extinction using maps of
mean values for each of the calendar quarter for 2004. The first quarter has the highest region of
haze centered in the midwestern U.S.; the warmer second and third quarters have the region of
highest haze over the Ohio River Valley; and the fourth quarter is a composite with high haze in both
December 2009
9-18
-------
the Midwest and Ohio River Valley. Smaller regions of haze show up in the Columbia River Valley
(border between Washington and Oregon) in the colder first and fourth quarters and in Southern
California in the warmer second and third quarters.
The IMPROVE algorithm permits each PM component contribution to light extinction to be
separately estimated. Figure 9-9, Figure 9-10, and Figure 9-11 display the seasonal variation of the
percent contribution to aerosol light extinction by the various component estimates. Figure 9-9
shows the contributions by SO42~ and nitrate particulate including the haze enhancement caused by
the absorbed water in humid conditions. As shown in Figure 9-9, a large regional pattern of high
contribution to haze by nitrate PM is centered in the Midwest, and during the cooler months the
nitrate PM is the dominant cause of haze in the region responsible for a third to a half of the
particulate light extinction. Midwestern particulate nitrate is responsible for the regional pattern of
the highest haze conditions shifting from the Ohio River Valley during summer to the Midwest in the
winter as shown in Figure 9-8. Particulate nitrate is also a significant contributor to particulate light
extinction year-around in parts of California, where it generally contributes 20%-40%. The Pacific
Northwest, parts of Idaho and Utah experience large contributions to particulate light extinction by
nitrates during the colder seasons, with contributions of 20%-30%. Figure 9-9 also shows that
particulate SO42~ is the predominate contributor in the eastern U.S., where it contributes 40% or
more on average and during the summer months up to three quarters of the particulate light
extinction over much of the East. In the western U.S. particulate SO42~ generally contribute 20-50%
of the particle light extinction. Regions of the lowest fractional contributions by particulate SO42~
and nitrate for any calendar quarter are generally in the western U.S.
Figure 9-10 shows the contributions to haze by the carbonaceous PM components (i.e., organic
mass and EC). They show broadly similar patterns with the greatest contributions in the western U.S.
especially during the warmer months of the year. For the most part this spatial pattern results from
the dominant contributions to haze by SO42~ and nitrate PM in the eastern half of the U.S., leaving
relatively little for other component contributions. The fractional contribution to haze by organic PM
is generally two to five times that of EC. In absolute terms, both carbonaceous components tend to
have two to three times higher concentrations in the eastern U.S. than in the non-coastal western
states.
December 2009 9-19
-------
Source: VIEWS (http://vista.cira.colostate.edu/view5/)
Figure 9-9.
Percent contributions of ammonium nitrate (left column) and ammonium sulfate
(right column) to particulate light extinction for each calendar quarter of 2004
(first through fourth quarter arranged from top to bottom). Note that the contour
intervals are not the same for the two species contributions.
December 2009
9-20
-------
Source: VIEWS (http://vista.cira.colostate.edu/view5/)
Figure 9-10.
Percent contributions of organic mass (left column) and EC (right column) to
particulate light extinction for each calendar quarter of 2004 (first through fourth
quarter arranged from top to bottom). Note that the contour intervals are not the
same for the two species contributions.
December 2009
9-21
-------
Source: VIEWS (http://vista.cira.colostate.edu/view5/)
Figure 9-11.
Percent contributions of coarse mass (left column) and fine soil (right column) to
particulate light extinction for each calendar quarter of 2004 (first through fourth
quarter arranged from top to bottom). Note that the contour intervals are not the
same for the two species contributions.
December 2009
9-22
-------
Figure 9-11 shows the contributions to haze by coarse mass and fine soil components. As with
the carbonaceous components, these crustal dominated components have a similar spatial pattern
with regions of highest contribution to haze in the western U.S., and just as for the carbonaceous
PM, the explanation for low contributions in the eastern U.S. is the dominant contributions to haze
by SO42~ and nitrate PM leaving relatively little for other components. The crustal components
contribute more to haze in the arid regions of the west including the southwestern deserts. In
absolute terms, coarse mass concentrations are as high in the rural areas of the center of the country
(including Oklahoma, Arkansas, Kansas, Missouri, and Iowa) as they are in the Desert Southwest.
Typically coarse mass contributions to haze exceed those of fine mass by a factor of 2-4.
9.2.3.3. Urban and Regional Patterns
Using a combination of IMPROVE and CSN data, it is possible to compare urban PM2.5
concentrations and composition to corresponding remote-area regional values. These are shown as
paired color contours maps for IMPROVE and IMPROVE plus CSN (see Figure 9-12 through
Figure 9-23). The degree of comparability of the data from these 2 networks was assessed by an
analysis of two years of co-located IMPROVE and CSN data from 6 urban areas. The CSN organic
mass data were adjusted for a positive sampling artifact prior to inclusion in this assessment, in a
fashion similar to that used for the IMPROVE data set (pages 29-30, DeBell, 2006, 156388). Note
that the contour scales are different between the two maps for each component pair of maps so that
each contains as much information as possible using ten concentration contours. To assess the degree
to which urban areas have higher PM component concentrations compared to regional background
note how many contour intervals surround the urban monitoring sites. The U.S. EPA (2004, 190219)
used the pairing of IMPROVE and CSN monitoring sites at 13 selected urban areas to separate local
and regional contributions of three major PM2.5 components as shown in Figure 9-24.
In Figure 9-12 and Figure 9-13, urban PM2.5 concentrations are systematically higher than
those in the surrounding non-urban regions. The urban excess is generally much higher in the
western U.S. than in the East (e.g., there are five contour intervals separating Salt Lake City from its
remote regional area compared to only two for Indianapolis). This implies that eastern and western
urban PM2.5 concentrations and resulting visibility are less different than the eastern and western
regional concentrations and visibility.
December 2009 9-23
-------
IMPROVE Site
IMPROVE Urban Site
-l->
*^
146
10.9
9.68
8.47
7.26
6.05
4.84
3.63
2.42
1.21
0.00
Mg/m3
,
® Alaska
Puerto Rico /
Virgin Islands
Source: Debell (2006,156388).
Figure 9-12. IMPROVE Mean PM2.s mass concentration determined by summing the major
components for the 2000-2004.
Alaska
Hawaii
(2vW1'>'~ *''*'*'- '\
\ ;v-^«x^. *r^\
\J
ASTNSite
^—M • IMPROVE Site
• IMPROVE Urban Site
,
•^
— 307
B 14.7
130
11 4
9.78
8.15
6.52
4.89
13.26
1.63
0.00
(jg/m3
e
Puerto Rico /
Virgin Islands
Source: Debell (2006,156388).
Figure 9-13. IMPROVE and CSN (STN) mean PM2.s mass concentration determined by
summing the major components for 2000-2004.
December 2009
9-24
-------
Alaska
• IMPROVE Site
• IMPROVE Urban Site
Hawaii
e
Puerto Rico /
Virgin Islands
Source: Debell (2006,156388).
Figure 9-14. IMPROVE mean ammonium nitrate concentrations for 2000-2004.
Alaska
Hawaii
A STN Site
• IMPROVE Site
• IMPROVE Urban Site
e e
Puerto Rico /
Virgin Islands
Source: Debell (2006,1563881.
Figure 9-15. IMPROVE and CSN (STN) mean ammonium nitrate concentrations for 2000-2004.
December 2009
9-25
-------
Hawaii
IMPROVE Site
IMPROVE Urban Site
Puerto Rico /
Virgin Islands
Source: Debell (2006,156388).
Figure 9-16. IMPROVE mean ammonium sulfate concentrations for 2000-2004.
* Alaska
Hawaii
A STN Site
• IMPROVE Site
• IMPROVE Urban Site
e
Puerto Rico /
Virgin Islands
Source: Debell (2006,1563881.
Figure 9-17. IMPROVE and CSN (STN) mean ammonium sulfate concentrations for 2000-2004.
Figure 9-14, Figure 9-15, and Figure 9-24 show the PM2.5 nitrate in remote and urban areas.
Here the western states have urban particulate nitrate concentrations that far exceed twice the remote
December 2009
9-26
-------
area regional concentrations. For the Central Valley of California and Los Angeles areas, the urban
excess of ammonium nitrate exceeds regional concentrations by 2 ug/m3 to 12 ug/m3. In the region
of the Midwest nitrate bulge, the urban concentrations were less than twice the regional
concentrations for an annual urban excess of about 1 ug/m3. Northeast and southeast of the Midwest
nitrate bulge, annual urban particulate nitrate concentrations are several tenths to about 1 ug/m3
above the remote area regional concentrations, with warmer southern locations tending to have the
smaller concentrations of both regional and urban excess particulate nitrate.
As shown in Figure 9-16, Figure 9-17, and Figure 9-24, annual-averaged urban particulate
SO42~ concentrations are generally not much higher than the regional values, with urban excess
generally of less than about 0.5 ug/m3. The exceptions apparent by comparing Figure 9-16 and
Figure 9-17 are in Texas and Louisiana where urban excess particulate SO42~ are >1 ug/m3, perhaps
caused by local emissions (e.g., from oil refineries). Urban contributions are a larger fraction of the
total particulate SO42~ concentrations in the western U.S. because the regional concentrations are
much lower than in the East. The modest additional particulate SO42~ concentrations associated with
urban areas suggests that most particulate SO42~ is regionally distributed, and that IMPROVE and
CSN monitoring sites can be used together to enhance the ability to delineate particulate SO42~
spatial distributions. For example, note that the additional data from urban sites shown in Figure
9-17 extends north and south of the distribution of the high particulate SO42~ loading shown in
Figure 9-16 over Tennessee and Kentucky, as well as the high loadings over southern Pennsylvania,
eastern West Virginia and northern Virginia. (The color-contour suggested dip in concentrations
between the two eastern particulate SO42~ high concentrations regions may not exist in the
atmosphere, but this cannot be verified without speciation monitoring sites in southern Ohio, the
border of Kentucky and West Virginia and western Virginia.)
Urban and remote area carbonaceous PM2.5 are displayed in Figure 9-18 and Figure 9-19
(organic mass), Figure 9-20 and Figure 9-21 (EC), and Figure 9-24 (total carbon = organic + EC
concentration). Just as with particulate nitrate, both organic mass and EC concentrations are more
than twice the remote-area background concentrations for western urban monitoring locations. One
of the more interesting pairing of sites is for the Virgin Islands compared to the urban site at San
Juan, Puerto Rico (see the map cutout Figure 9-18 through Figure 9-21). The San Juan urban excess
OC is moderate, while the EC value is among the most extreme inferred in this manner. For eastern
urban areas, approximately half the total carbon is local while the other half is regional. In eastern
urban areas, carbonaceous and SO42~ particulate are the two major components of PM2.5, with
roughly equal contributions, and account for over 80% of the mass concentration. Edgerton et al.
(2004, 156413) showed that carbonaceous PM2.5 is responsible for most of the urban excess above
regional concentrations at four urban/rural paired Southeastern Aerosol Research and
Characterization (SEARCH) monitoring sites in the southeastern U.S. However, the higher overall
light extinction efficiency for SO42~ resulting from its hydrophilicity gives it ~ 2:1 dominance in
responsibility for eastern urban light extinction.
Urban and remote area soil PM2.5 concentrations are displayed in Figure 9-22 and Figure 9-23.
Urban fine soil concentrations are at most a few tenths of a ug/m higher than the regional
background concentrations and in some regions they are much less. Just as with carbonaceous PM2 5,
the Virgin Island, San Juan, Puerto Rico pair are interesting for fine soil. In this case, both of these
island monitoring sites have high concentrations of fine soil, which is caused by the influence of the
trans-Atlantic transport path of dust from Africa (Prospero, 1996, 156889).
No urban-remote pair of coarse mass concentration maps is available because CSN does not
monitor coarse mass. In Malm et al. (2004, 156728) a map of annual mean coarse mass
concentration is shown for 2003 which includes the values for IMPROVE urban sites, including two
in the western U.S. with much more coarse mass than the nearby remote areas monitoring sites
(i.e., ~24 ug/m3 compared to ~9 ug/m3 for Phoenix, AZ, and ~6 ug/m3 compared to ~2 ug/m3 for
Puget Sound, WA) and one eastern IMPROVE site at Washington, DC with less coarse mass than the
surrounding remote area values (~2 ug/m3 compared to ~4 ug/m3).
December 2009 9-27
-------
I
9
* Alaska
Hawaii
IMPROVE Site
IMPROVE Urban Site
O
Puerto Rico /
Virgin Islands
Source: Debell (2006,1563881.
Figure 9-18. IMPROVE monitored mean organic mass concentrations for 2000-2004.
P12.4
6.47
5.75
5.03
4.31
3.60
2.88
2.16
11.44
0.72
0.00
Mg/m3
Puerto Rico/
Virgin Islands
Alaska
Hawaii
A STN Site
• IMPROVE Site
• IMPROVE Urban Site
Source: Debell (2006,1563881.
Figure 9-19. IMPROVE and CSN (STN) mean organic mass concentrations for 2000-2004.
December 2009
9-28
-------
• Alaska
Hawaii
• IMPROVE Site
• IMPROVE Urban Site
0
Puerto Rico/
Virgin Islands
Source: Debell (2006,1563881.
Figure 9-20. IMPROVE mean EC concentrations for 2000-2004.
V1.74
0.95
0.84
0.74
0.63
• 0.53
0.42
0.32
10.21
0.11
000
(jg/m3
_
° Alaska
Hawaii
A STN Site
• IMPROVE Site
• Urban IMPROVE Site
a e
Puerto Rico /
Virgin Islands
Source: Debell (2006,156388).
Figure 9-21. IMPROVE and CSN (STN) mean EC concentrations for 2000-2004.
December 2009
9-29
-------
4
• •
..
3.08
144
1.28
1.12
0.96
0.80
0.64
0.48
0.32
0.16
0.00
|jg/m3
,
• IMPROVE Site
• IMPROVE Urban Site
O
Puerto Rico /
Virgin Islands
Source: Debell (2006,156388).
Figure 9-22. IMPROVE mean fine soil concentrations for 2000-2004.
Alaska
Hawaii
A STN Site
• IMPROVE Site
• IMPROVE Urban Site
Puerto Rico /
Virgin Islands
Source: Debell (2006,156388).
Figure 9-23. IMPROVE and CSN (STN) fine soil concentrations, 2000-2004.
Figure 9-25 shows the remote area coarse mass concentrations as measured by the IMPROVE
network. The pattern of high coarse mass concentrations from Oklahoma to Iowa is comparable to
the high concentrations in the desert southwest, though as shown in Figure 9-11 it contributes a
December 2009
9-30
-------
smaller relative share of the light extinction because of the higher contributions to haze by
particulate nitrate and sulfate in this agricultural region of the country. Comparing Figure 9-22 and
Figure 9-25 shows that the coarse mass and fine soil concentration patterns are similar for the desert
southwest but there is a much lower fine soil to coarse mass concentration ratio for the agricultural
center of the country, suggesting a regional difference in the size distribution of the coarse mass,
perhaps due to differences in suspendable soil materials.
9.2.3.4. Temporal Trends
Visibility trend analysis requires relatively long data records to avoid having meteorologically
driven interannual variability obscure more meaningful emissions-driven air quality trends. A
requirement for long-term data limits the number of monitoring sites useful for trend analysis. Maps
that show haze trends for IMPROVE sites for the 10-year period 1995-2004 for the mean of the 20%
best and the 20% worst haze days where sites are required to have a minimum of 6 complete years of
data during the 10-year period are shown in Figure 9-26 and Figure 9-27, respectively. The best haze
days have improving haze at most sites (32 of 47), no trend at several sites (10 of 47) and degrading
visibility at just one site (Great Sand Dunes, CO). The worst haze days have improving haze
conditions at several sites (13 of 47), no trend at most sites (30 of 47), and degrading visibility at a
few western sites (4 of 47).
December 2009 9-31
-------
Nitrates
Sulfates
Fresno
Missoula
Salt Lake City
Tulsa
St. Louis
Birmingham
Indianapolis
Atlanta
Cleveland
Charlotte
Richmond
Baltimore
New York City
I
13
\ \
j — i
[ WEST
U E^7
1
~^H
~|
| |
3D
— I — i n Regional
Contribution
— 1 1 • Local
Contribution
| |
i i i i i i i i i i i i
1 2 4 6 8 10 12
Fresno
Missoula
Salt Lake City
Tulsa
St Louis
Birmingham
Indianapolis
Atlanta
Cleveland
Charlotte
Richmond
Baltimore
New York City
(
I I
|
— i — i
| | WEST
I I EAST
3
II
I
II
ID Regional
Contribution
D Local
|
I 2 4 6 8 10 12
Annual Average Concentration
of Nitrates, ug/m3
Annual Average Concentration
of Sulfates, ug/m3
Carbon
Fresno
Missoula
Salt Lake City
Tulsa
St. Louis
Birmingham
Indianapolis
Atlanta
Cleveland
Charlotte
Richmond
Baltimore
New York City
WEST
EAST
i n Regional
1 Contribution
FJ Local
Contribution
2 4 6 8 10 12
Annual Average Concentration
of Carbon,
Source: U.S. EPA (2004, 1902191
Figure 9-24. Regional and local contributions to annual average PM2.s by particulate S042~,
nitrate and total carbon (i.e., organic plus EC) for select urban areas based on
paired IMPROVE and CSN monitoring sites.
December 2009
9-32
-------
IMPROVE Site
IMPROVE Urban Site
e
Puerto Rico /
Virgin Islands
Source: Debell (2006,1563881.
Figure 9-25. IMPROVE mean coarse mass concentrations for 2000-2004.
Eight-, ten-, and sixteen-year trends analysis conducted for the Western Regional Air
Partnership (WRAP) as part of the Causes of Haze Assessment (http ://www. wrapair. or) show that
improving trends for the 20% best haze days for the sites in the western U.S. generally correspond to
improving trends for all of the major components with the exception of parti culate nitrate. Trends
assessment for the worst haze days at western sites show consistent reductions in parti culate SO42~,
but otherwise have mixed increasing and decreasing haze component trends, many of which are not
statistically significant. Substantial interannual and shorter term spatial and temporal wildfire
activity variations have been shown to have a significant impact on the variability of haze in the
western U.S. (Park et al., 2006, 190469: Spracklen et al., 2007, 190485). Edgerton et al. (2004,
156413) showed a decreasing trend in PM2.5 of about 18% (corresponding to 1 ug/m3 to 2 ug/m3) for
4 urban-rural paired SEARCH sites in the Southeastern U.S. corresponding to similar reductions in
SO42~ and carbonaceous particulate.
December 2009
9-33
-------
f-.
* Improving Trend, p<-0.05
Improving Trend, G.05
-------
r/:
* Improving irend,D<=005
Improving Trend, 0 05
-------
atmospheric ammonia or elevated temperatures, trends in nitrogen may be principally in nitric acid
with no net change in nitrate light extinction. Alternately with abundant ammonia and low
temperatures the trend in nitrogen may be principally in particulate nitrate and the nitrate component
of haze.
Western US
North Eastern US
1.38
- 1.15
in
-0.92 in
- 0.69
LLJ
CN
O
V)
0.46
16.2
- 13.5
10.8 in
o
'i/i
ui
E
LU
c*j
O
U)
1988 1990 1992 1994 1996 1998
1988 1990 1992 1994 1996 1998
South Middle US
South Eastern US
1.83
0.61
1988 1990 1992 1994 1996 1998
-*— Sulfate Ion
«~ 7.5
"Si
w 4.5
6.75
5.625
4.5
3.375 E
O
CO
2.25
1988 1990 1992 1994 1996 1998
SO2 Emissions
Source: Reprinted with Permission of the American Geophysical Union from Malm et al. (2002,1567271.
Figure 9-28. Ten-year trends in the 80th percentile particulate S042" concentration based on
IMPROVE and CASTNet monitoring and net S02 emissions from the National
Emissions Trends (NET) data base by region of the U.S.
Ten-year trends (1994-2003) of particulate nitrate contribution to light extinction during the
20% worst haze conditions conducted as part of the Causes of Haze Assessment (see the link in
Table 9-1) are shown in Figure 9-29. This indicates that haze from particulate nitrates is increasing
across the western U.S. at a rate of several Mm"1 per year in parts of California and at a rate of
several tenths of an Mm"1 across the Four-Corners states. A similar particulate nitrate trends map (not
shown here) for the 20% best haze conditions that is available at the same web site shows decreasing
particulate nitrate contribution to light extinction at nearly all of the western monitoring sites. While
statistically significant, these trends for both the 20% worst and 20% best haze periods are
influenced by an unexplained nationwide period of depressed nitrate concentrations measured by the
IMPROVE network during a 4-yr period from the winter of 1996-1997 through the winter of
2000-2001. Extensive examinations of plausible monitoring methodological explanations have failed
December 2009
9-36
-------
to offer any evidence that the data are invalid (McDade, 2004, 192075). but no satisfactory
atmospheric or emissions-related explanation has been offered to account for this 4-yr depression of
nitrate. Similar analyses of particulate nitrate haze trends are not available for the rest of the country.
Maps of remote-area 8-, 10-, and 16-yr trends for carbonaceous and crustal PM species based
on IMPROVE monitoring are available for the western U.S. conducted as part of the Causes of Haze
Assessment. Generally these show a broad range of results (i.e., a mixture of statistically significant
upward or downward trends and insignificant trends often with neighboring sites having opposing
trends) that vary considerably depending on the number of years selected (i.e., 8, 10, or 16) and
whether trends are for the best, worst, or middle of the haze distribution data. The scatter in these
results is undoubtedly due to the high interannual variability and varying locations of wildfire and
wind-suspended dust emissions that dominate the remote-area concentrations of the carbonaceous
and crustal PM species in the western U.S.
ft t
ft
t * * *
A
A f Legend
t
n
A 4 , WNSIope
ft " " '
A$ ^ i ; ft 0.064-0,093
I)1 0.094 -
'ft 0.171 -
'(V 0.521 -
. . f]1 0.094-0170
If
0.171 - 0.520
1509
^h 1.510-5089
WNPValue
I 0.21-1.00
0.11 -0.20
0.06-0.10
• 0.00 - 0.05
Source: http://wvwv.coha.dri.edu/
Figure 9-29. Map of 10-yr trends (1994-2003) in haze by participate nitrate contribution to haze
for the worst 20% annual haze periods. The orientation, size and color of the
arrows indicate the direction, magnitude and statistical level of significance of
the trends. These consistent upward trends may be a misleading result due to an
unexplained sampling issue (see text for additional information).
9.2.3.5. Causes of Haze
In order to attribute haze to emissions from individual sources, source types, or source regions,
generally, any of a number of receptor and air quality simulation modeling approaches are applied.
When using multiple approaches, the results of each are reconciled using a weight-of-evidence
methodology. Commonly this methodology has been applied to the extensive datasets generated by
special studies designed to estimate source-receptor relationships for a few receptor locations or for
individual emission sources (Pitchford et al., 1999, 156873: Pitchford et al., 2005, 156874: Schichtel
December 2009 9-37
-------
et al., 2005, 156957). More recently the Regional Planning Organizations (RPOs) have sponsored
extensive regional haze source attribution assessments using weight-of-evidence methodologies to
reconcile attribution results for virtually all of the remote-area IMPROVE sites to support the
development of State Implementation Plans for the RHR. Additionally, a number of recent urban
special studies, including those sponsored by EPA PM Supersites program (Solomon and Hopke,
2008, 156997). have addressed the causes of and sources contributing to urban excess PM
concentrations above region concentrations. Attribution results uncertainties are generally larger than
those of the measurements upon which they are based because they also include model uncertainties
and assumptions, as well as issues of representativeness (e.g., How well does the data used in the
analysis represent the typical emissions, air quality and meteorological conditions of interest?). As
such it is advisable to treat the attribution results reported below as semi-quantitative.
The relative importance of the PM species that contribute to haze varies by region of the U.S.
and time of year as shown in Figure 9-9, Figure 9-10, and Figure 9-11, above. Generally haze in the
western half of the U.S. is not dominated by any one or two PM species. In the eastern half of the
U.S., SO42~, especially during summer, and nitrate during the winter in the Midwest are the dominant
haze species. As described above, urban haze can be viewed as a composite of the regional and local
contributions where local contributions seem to be dominated by carbonaceous and, to a lesser
extent, nitrate and crustal PM components. There have been far fewer urban investigations that
explicitly consider visibility impacts, though there are numerous studies of urban PM source
attribution. The order of discussion below on the cause of haze is by region beginning in the western
U.S. and proceeding to the east, analogous to dominant air flow patterns across the lower 48 states
and will include information from urban studies along side those of remote-area haze investigations.
Based on modeling of an episode (September 23-25, 1996) in the California South Coast Air
Basin (SCAB) and another episode (January 4-6, 1996) in the San Joaquin Valley (SJV) by Ying and
Kleeman (2006, 098359), about 80% of the particulate SO42~ for both regions is from upwind
sources (i.e., likely from offshore sources including marine shipping, long-range transport and
natural marine sources), with most of the remaining is associated with diesel and high-sulfur fuel
combustion. Kleeman et al. (1999, 011286). using a combination of measurements and modeling,
showed that the upwind particulate SO4Z~ source region for the SCAB was over the Pacific Ocean
(confirmed by measurements on Santa Catalina Island) and that these particles subsequently grew
with accumulation of additional secondary aerosol material, principally ammonium nitrate as they
traversed the SCAB. The majority of the nitric acid that forms particulate nitrate in the SCAB is
from diesel and gasoline combustion (-63%), while much of the ammonia is from agricultural
sources (-40%), catalyst equipped gasoline combustion (-16%), and upwind sources (-18%). The
majority of the OC found in SCAB was attributed in this study to primary emissions by
transportation-related sources, including diesel (-13%) and gasoline (-44%) engines and paved road
dust (-12%). At the Fullerton site in the middle of the SCAB the concentration of locally generated
organics is roughly double that of the locally generated nitrates (-5.6 ug/m3 compared to
-2.4 ug/m3), while at Riverside on the east edge of the SCAB and near the large agricultural sources
of ammonia emissions, the particulate nitrate concentrations are nearly double that of organic PM
(-17 ug/m3 compared to -10 ug/m3).
Ying and Kleeman (2006, 098359) showed that during the winter 1996 episode in the SJV
most of the nitric acid that forms particulate nitrate is from upwind sources (-57%) with diesel and
gasoline combustion contributing most of the rest (30%), while much of the ammonia is from
upwind sources (-39%) and a combination of area, soil and fertilizer sources (-52%). In an
assessment of PM particle size and composition in the SJV during the winter of 2000-2001, Herner
et al. (2006, 135981) showed that fresh emissions of carbonaceous PM from combustion sources in
urban locations (Sacramento, Modesto, and Bakersfield, CA) move quickly from ultrafme particle
size (i.e., diameter -0.1 um) to accumulation mode by condensation with accumulation mode
(i.e., diameter -0.5 um) particles, and that secondary nitrate particle formation occurs preferentially
on the surface of hydrated ammonium sulfate particles during the afternoon when gas-phase nitric
acid is at peak photo-chemical production from NOX. Given the abundance of ammonia emissions
and low ambient temperatures, particulate nitrate production in this way is only limited by the
availability of nitric acid. Due to the cool winter conditions there was little SOA production during
this study. Sea salt was shown to dominate the larger coarse particle mode during on-shore wind at
the background coastal monitoring site at Bodega Bay, north of San Francisco, CA.
Using a regression analysis to find the dependence of particulate SO42~ concentration
measured over a 3-year period (2000-2002) at 84 western IMPROVE monitoring sites on the
December 2009 9-38
-------
modeled transport trajectories to the sites for each sample period, Xu et al. (2006, 102706) were able
to infer the source regions that supplied particulate SO42~ in the western U.S. Among the source
regions included in this analysis was the near coastal Pacific Ocean (i.e., a 300-km zone off the coast
of California, Oregon, and Washington). Up to half of the parti culate SO42~ measured at Southern
California monitoring sites was associated with this source region. As shown in Figure 9-30 the zone
of impact from this source region included large regions of California, Arizona, and Nevada. The
authors made the case that high sulfur content fuel used in marine shipping and port emissions may
be largely responsible. As a result, the WRAP RPO emissions inventory was modified to include
marine snipping and a Pacific Offshore source region was added to source attribution by air quality
simulation modeling.
The SO42~ attribution results of the WRAP air quality modeling (results available on the
Technical Support System (TSS) website, see Table 9-1 for the web-link) credit the Pacific offshore
source region with somewhat smaller contributions than those from the trajectory regression work by
Xu et al. (2006, 102706) with concentrations at the peak impact site in California that are about 45%
compared to 50% by regressions and even greater differences for more distant monitoring sites.
Based on the modeling attribution the Pacific offshore source region was responsible for 10-20% of
the nitrate measured in Southern California.
Contribution of the Pacific Coast to Siilntc Concentration
0.1-0.2
0.2-0.4
0.4-0.6
0.6-0.8
0.8 - 1 .4
Source: Reprinted with Permission of Atmospheric Environment from Xu et al. (2006,1027061.
Figure 9-30.
Contributions of the Pacific Coast area to the ammonium sulfate (ug/m3) at 84
remote-area monitoring sites in western U.S. based on trajectory regression for
all sample periods from 2000-2002 (dots denote locations of the IMPROVE
aerosol monitoring sites).
A coordinated effort by federal, state, and county air quality organizations to determine the
causes of haze in the Columbia River Gorge (a deep and narrow gap in the Cascade Mountains on
the Washington/Oregon border) through extensive multiyear measurements and high spatial
resolution air quality modeling of typical episodes demonstrated the multitude of emission sources
that contribute to its impairment (Pitchford et al., 2007, 098066). During the summer, gorge winds
are generally from the west and relatively dry. More than half of the haze during a typical summer
episode is from a combination of international and other distant sources (-22% at the western end of
the gorge) plus regional natural sources including wildfire and secondary organic PM from biogenic
emissions (-39% at the eastern end of the gorge). The Portland/Vancouver metropolitan area was
responsible for a significant amount of the haze during the summer (-20% on in the western end of
December 2009
9-39
-------
the gorge), while sources within the gorge were responsible for a moderate amount of haze (-6% and
-9% at the western and eastern ends of the gorge). The wind is much more often from the east
during the winter. The highest haze conditions in the gorge are during the winter and are associated
with fog conditions that rapidly convert precursor gaseous emissions of NOX and SO2 from local and
regional combustion sources and NH4 from local and regional agricultural activities to secondary
nitrate and SO42~ PM that persist as a post-fog intense haze. Contributions by these sources east of
the gorge contribute -57% of the haze on the eastern end of the gorge, with half of the nitrate and
SO4 ~ particulate from electric utility emissions and most of the rest from transportation sources.
Other sources contributing during the winter haze at the eastern end of the gorge are from sources
outside the modeling domain (i.e., most of Washington and Oregon) and within the gorge (-23% and
-10%, respectively).
An assessment of concurrent measurements at the nearby Mt. Hood IMPROVE monitoring
site (45 km south of the Columbia River at 1,531 m ASL) shows that Columbia River Gorge haze
conditions and especially the wintertime high nitrate/SO42~ contributions to haze are not typical of
the generally higher elevation remote areas of the region (Pitchford et al., 2007, 098066). However
Gorge-like high wintertime nitrate and SO42~ are found at the Hells Canyon IMPROVE site, which is
similarly situated in a narrow canyon of the Snake River almost 400 km east of the Gorge (from the
VIEWS web site, see Table 9-1), implying that there may be a substantial vertical concentration
gradient during winter in this complex terrain.
Several example monitoring locations distributed across the northern and southern portions of
the western U.S. have been selected to illustrate the attribution results from the WRAP-sponsored
attribution analysis tools that estimate the relative responsibility for haze of the various PM species
by source region and source type. The selected sites include Olympic NP, WA; Yellowstone NP, WY;
and Badlands NP, SD across the north, and San Gorgonio Wilderness (W), CA; Grand Canyon NP,
AZ and Salt Creek W, NM across the south as shown in Figure 9-31.
December 2009 9-40
-------
2000-2004 Baseline Average
20% Worst Days
IMPROVE Aerosol Extinction (Mrn-1)
-------
area and wildfire emissions contributing from a few percent at the furthest eastern sites to about half
at San Gorgonio.
By comparison, the particulate nitrate is much more from domestic regional emission sources,
with -60-80% being from emissions within the WRAP region. For the west coast sites about 25% of
the nitrate is from a combination of Pacific offshore emissions (i.e., marine shipping) and outside
domain regions. Canadian emissions are responsible for about 10-30% of the particulate nitrate for
the three northern sites, but Mexican emissions do not contribute appreciably to particulate nitrate
for the three southern sites. Motor vehicles are the largest contributing NOX source category
responsible for particulate nitrate for these six WRAP sites, with a combination of point, area and
wildfire source categories also contributing from about 10-50% of the WRAP regional emissions.
WRAP only used the virtual tracer approach to investigate source locations and categories for
SO2 and NOX emissions. A different type of virtual tracer modeling tool was used to track the
various OC compounds and sort them into three groups for 2002. The first group labeled primary
organics includes all of the organics that are emitted directly as PM from any source type and
location. The second group labeled anthropogenic secondary organics is PM produced in the
atmosphere by aromatic VOCs. The third category labeled biogenic secondary organics is PM
produced in the atmosphere by biogenic VOCs. Organic PM in the biogenic secondary category
includes those that would functionally be considered man-made emissions (e.g., those from
agricultural crops and urban landscapes), though in most remote areas of the west these man-made
VOC emissions are small compared to those of the natural biogenic sources. Figure 9-33, Figure
9-34, and Figure 9-35 show the monthly averaged apportionment of organic PM for the 6 selected
monitoring locations.
December 2009 9-42
-------
Olympic NP
2002
Sulfate 1.1 ua/m3
Yellowstone NP
'2002
Sulfate0.6ua/m3
Bad
NP
San Gorgon 10 W Grand Canyon NP
D02
Sulfate 1.2 uo/m3
Salt Creek W
2002
Sulfate 0.8 ua/m3
\^l
2002
Sulfate 0.5 uci/m3
2002
Sulfate 1.4 ua/m3
(a)
WRAP
D Pacific Off shore
CENRAP
Eastern U.S.
JCanada
Mexico
Outside Domain
Olympic NP
Yellowstone NP
Badlands NP
V
2002 2002 2002
Nitrate 1.6 ua/m3 Nitrate 0.6 ua/m3 Nitrate 1.2 ua/m3
\
San Gorgonio W Grand Canyon NP Salt Creek W
(b)
WRAP
Pacific Offshore
CENRAP
Eastern U.S.
D Canada
Mexico
Outside Domain
2002
Nitrate 1.5 uo/m3
2002
Nitrate 0.4 ua/m3
2002
Nitrate 0.5 ua/m3
Figure 9-32. Particulate $(>42~ (a) and nitrate (b) source attribution by region using CAMx
modeling for six western remote area monitoring sites : top left to right Olympic
NP, WA; Yellowstone NP, WY; Badlands NP, SD; bottom left to right San Gorgonio
W, CA; Grand Canyon NP, AZ; and Salt Creek W, NM. WRAP includes ND, SD, WY,
CO, NM and all states further west. CENRAP includes all states east of WRAP and
west of the Mississippi River including MN. Eastern U.S. includes all states east
of CENRAP. The Pacific Offshore extends 300km to the west of CA, OR, and WA.
Outside Domain refers to the modeling domain, which extend hundreds of
kilometers into the Pacific and Atlantic Oceans and from Hudson Bay Canada to
just north of Mexico City. This figure was assembled from site-specific diagrams
produced on the TSS web site (see Table 9-1) for 2002.
December 2009
9-43
-------
Organic Aerosol for All Days
Class I Area - Olympic NP, WA
3.20
Organic Aerosol for All Days
Class I Areas - San Gorgonio W, CA: San Jacinto W, CA
2.80
Source: From the TSS website, see Table 9-1.
Figure 9-33.
Monthly averaged model predicted organic mass concentration apportioned into
primary PM and anthropogenic and biogenic secondary PM categories for the
Olympic NP (top) and San Gorgonio W (bottom) monitoring sites.
Based on the modeling results for these six sites and confirmed by measurements (see, e.g.,
Figure 9-10), a west-to-east decreasing gradient of organic mass exists with annual concentrations
from -2 ug/m3 for the coastal state sites to ~1 ug/m3 for the intermountain west sites, to <1 ug/m3 for
the sites just east of the Rocky Mountains, discounting the large fire impacts for July at Yellowstone
NP which raised its annual mean to -2 ug/m3. At all of these remote-area sites anthropogenic
secondary PM is estimated to be a small fraction of the organic mass, with the largest fractional
contribution at the San Gorgonio monitoring site immediately downwind of the major Southern
California urban areas, yet having <10% of the monthly mean organic mass from anthropogenic
secondary PM. Of the six selected monitoring sites, San Gorgonio has the highest fraction of the
organic PM from primary emissions (-57%), followed by Yellowstone (-55%), then the two
eastern-most sites (Badlands -42% and Salt Creek 41%), and with Grand Canyon and Olympic NPs
the lowest fraction by primary emissions (-37%). Yellowstone NP would have had the lowest
fraction of organic PM by primary emissions had it not been for the month of July (the 11-mo mean
was 29%) when wild fire smoke contributed. Results from recent chamber and field studies, and
modeling would seem to call into question apportionment of primary and secondary carbon done by
traditional air quality model simulations of OC, as described above, due to the combined effects of
December 2009
9-44
-------
extensive evaporation of semivolatile primary emissions when diluted and to photochemical
reactions of low volatility gas phase species that substantially increases the amount of secondary
organic PM (Robinson et al., 2007, 191975).
14.00
12.00
J= 10.00
3
^ 8.00
o
I 6'°°
-------
0.80
Organic Aerosol for All Days
Class I Area - Badlands NP, 3D
1.20
1.00
Organic Aerosol for All Days
Class I Area - Salt Creek NVVRW, NM
Source: From the TSS website, see Table 9-1.
Figure 9-35.
Monthly averaged model predicted organic mass concentration apportioned into
primary PM and anthropogenic and biogenic secondary PM categories for the
Badland NP (top) and Salt Creek W (bottom) monitoring sites.
Radiocarbon (14C) dating techniques were used to group ambient PM carbon into fossil and
contemporary source categories at 12 IMPROVE monitoring sites across the U.S., 8 of which are in
the WRAP region (Schichtel et al., 2008, 156958). Results of this work showed that contemporary
carbon accounts for about half the carbon in urban areas, 70-97% in near-urban areas (i.e., San
Gorgonio) and 82-100% in remote areas. Comparing these radiocarbon dating results with the
WRAP virtual tracer modeling results for organic aerosol (above), and presuming that the modeled
anthropogenic secondary organic is fossil carbon and the biogenic secondary is contemporary
carbon, suggests that a large fraction of the model-determined regional primary organic PM is from
contemporary carbon sources (e.g., smoke from wildfires).
Schichtel et al. (2008, 156958) compared radiocarbon measurements at two sets of urban/rural
paired sites in the west (Mount Rainer/Seattle, and Tonto/Phoenix). Figure 9-36 shows that most of
total carbon urban excess (i.e., urban site concentration minus the regional site concentration) in the
summer is from fossil carbon sources (87% and 79%, respectively), while in the winter there is a
surprisingly high fraction of the urban excess at both sites that is from contemporary carbon sources
(41% and 47%, respectively). This implies that urban, and therefore anthropogenic, activities
December 2009
9-46
-------
generate almost as much PM2.5 carbon from contemporary sources (e.g., residential wood
combustion) as from the fossil sources during the winter for these two western urban areas.
5 -
-™1
"E 4 -
O)
5 3 -
c
O •)
JQ ^
ro
0 1 -
n
Summer
c? c?
o^ N~ t—
15 b o
"E "Ł
-g 0) 0)
| TO 3. 3.
u n
c
15 o
*~ _Q
O i-
|_ (0
O
oo ^^H in ^^^ CNI
f" ^^^^1 T— O
II ^^^1 II II
CO HH C/) 05
X |^^ X X
1» E
S
-------
Puget Sound
0.51
O
action Contemporary
> 0.90 t 0.24
0.75 - 0,90 + 0.22
0.60 - 0.75 + 0.21
0.45 - 0.60 ± 0.20
< 0.45 t 0.19
'
Phoenix Tonlo
0-56 0.84
O O
Pint Sound
0.53
Mount P
Rainier
0.95
*•
Fraction Contemporary
> 0.90 t 0.15
0.75 - 0.90 + 0.14
0.60 - 0.75 t 0.13
0.45 - 0.60 -t 0.12
< 0.4S t 0.11
Phoenix Tonto
0.49 0.71
O O
Figure 9-37. Average contemporary fraction of PM2.s carbon for the summer (top) and winter
(bottom), estimated from IMPROVE monitoring data (June 2004-February 2006)
based on EC/TC ratios. The contemporary values from radiocarbon dating for the
12 monitoring sites are indicated in by colored circles with the site names. Color
contours are shown to aid in showing sites with similar values. Site locations are
indicated by circles for remote area sites and triangles for urban sites.
December 2009
9-48
-------
Contemporary carbon estimates for all of the IMPROVE network monitoring sites for data
from two summer seasons (June, July and August, 2004/2005) and two winter seasons (December,
2004/2005, January and February, 2005/2006) were calculated from the measured EC to total carbon
(EC/TC) ratios using the 12-site empirical relationship between radiocarbon determined
contemporary carbon fraction and IMPROVE measured EC/TC ratio (Schichtel et al., 2008,
156958). The results are displayed in color contour maps in Figure 9-37, which also shows the
radiocarbon determined contemporary carbon for the 12 sites. The lowest contemporary carbon
estimates (<60%) in both seasons are for urban areas. In the rural West, most of the sites have over
90% of their PM carbon from contemporary carbon sources during the summer and from 60% to
over 90% during the winter. In the rural East, most of the sites have 45-90% of their PM carbon from
contemporary carbon sources during the summer and from 60% to over 90% in the winter.
Schichtel et al. (2008, 156958) showed a strong relationship between the site-averaged EC/TC
ratios and the site-averaged fraction of fossil carbon separately for the summer and winter data sets
(i.e., R2 of 0.71 and 0.87, respectively). Using regression analysis they estimated that the summer
and winter EC/TC ratios associated with purely fossil carbon were 0.35 ± 0.039 and 0.46 ± 0.028,
respectively, and for purely contemporary carbon the EC/TC ratios were 0.12 ± 0.011 and
0.19 ± 0.0095. These ratios are shown to be consistent with corresponding ratios from the literature
for source testing primary fossil and contemporary combustion sources respectively. They are also
shown to be consistent with the 90 percentile value of the EC/TC ratio from the urban IMPROVE
monitoring sites (0.41 and 0.44 for summer and winter) and the 10th percentile values of the EC/TC
ratio for remote areas IMPROVE monitoring sites (0.07 and 0.16 for summer and winter), which
they argue are dominated by fossil and contemporary carbon, respectively.
The largest sources of contemporary carbon are primary emissions from biomass burning and
SOAfrom biogenic precursor gases (e.g., terpenes from conifer forests). Schichtel et al. (2008,
156958) estimated the 12-site overall contribution by secondary organic PM to the summer
contemporary carbon fraction as 36 ± 6.4% by assuming the EC/TC ratio for contemporary carbon
during the winter represented the ratio of primary emissions only (i.e., no secondary organic PM
formation in the winter) and that the EC/TC ratio for primary emissions is independent of seasons.
This approach should provide a lower bound estimate of the secondary OC species. The same
method applied to the fossil carbon fraction yielded an estimate of 23 ± 10% of the fossil carbon PM
from secondary organic formation in the atmosphere during the summer. These estimates correspond
to over 40% of the contemporary and over 35% of the fossil OC being from secondary PM
formation.
December 2009 9-49
-------
100.00
90.00
30.00
70.00
60.00
50.00
40.00
Sources and Areas of Potential Organic Carbon Emissions on Worst 20% Visibility Days
Class I Area - Olympic NP, WA
30.00
20.00
10.00
0.00
WRAPTSS- 3O1MIE
70.00
60.00
Sources and Areas of Potential Elemental Carbon Emissions on Worst 20% Visibility Days
Class I Area-Olympic NP.WA
• WBDust
Fugitive Dust
| Road Dust
Off-Road Mobile
| On-Road Mobile
Off-Shore
] WRAP Area Q&G Biogenic | Anthro Fire
JArea Natural Fire | Point
Source: From the TSS website (see Table 9-1)
Figure 9-38. Results of the weighted emissions potential tool applied to primary OC emissions
(top) and EC emissions (bottom) for the baseline and projected 2018 emissions
inventories for Olympic NP. Only source regions (WRAP states and other
regions) with the largest estimated contributions are shown (i.e., Canada,
Oregon, Pacific Off-Shore, and Washington from left to right). The scale is
normalized (i.e., unitless) one over distance weighted emissions multiplied by
trajectory residence time.
December 2009
9-50
-------
Sources and Areas of Potential Organic Carbon Emissions on Worst 20% Visibility Days
Class I Areas - San Gorgonio W, CA: San Jacinto W, CA
CA - 2000-04
2018PRP
PO - 2000-04
2018 PRP
Sources and Areas of Potential Elemental Carbon Emissions on Worst 20% Visibility Days
Class I Areas - San Gorgonio W, CA: San Jacinto W, CA
CA - 2000-04
2018 PRP
PO - 2000-04
2018 PRP
•WBDust
D Fugitive Dust
Figure 9-39.
Road Dust | On-Road Mobile Q WRAP Area O&G
Off-Road Mobile Off-Shore []Area
Biogenic • Antrim Fire
Natural Fire | Point
Source: From the TSS website (see Table 9-1).
Results of the weighted emissions potential tool applied to primary OC emissions
(top) and EC emissions (bottom) for the baseline and projected 2018 emissions
inventories for San Gorgonio W. Only source regions (WRAP states and other
regions) with the largest estimated contributions are shown (i.e., California and
Pacific Off-Shore from left to right). The scale is normalized (i.e., unitless) one
over distance weighted emissions multiplied by trajectory residence time.
December 2009
9-51
-------
WRAP applied a weighted emissions potential analysis tool that combined gridded emissions data with back-trajectory analysis that simulated the transport pathway to the
various monitoring sites to infer likely source region and emission categories for the 20% best and 20% worst haze conditions for each of the IMPROVEPMspeciation monitoring
locations in the West. Unlike the virtual tracer approach that uses a full regional air quality simulation model, this method does not explicitly account for chemistry or removal
processes and it does not incorporate the sophisticated dispersion estimates (i.e., it uses one over distance weighting for dispersion), so it should be considered a screening tool
that has been found to be helpful in identifying the likely sources contributing to haze. More information on this approach is available elsewhere (see the link to the TSS in Table
9-1). Primary OC and EC PM species results from the weighted emissions potential tool for the worst 20% haze days using the 2000-2004 base years' emissions and trajectories,
and the same trajectories with 2018-projected emissions for each of the 6 selected western monitoring locations are shown in Source: From the TSS website (see Table 9-1)
Figure 9-38 through Figure 9-43.
For Olympic NP (Source: From the TSS website (see Table 9-1)
Figure 9-38), most of the primary OC as well as most of the EC PM is likely to be from the
state of Washington during the worst haze days. This is because the multi-day trajectories that
transport emissions on its worst days tend to be short (within 200 km based on maps available on
TSS, see Table 9-1). Area sources, which include emissions from residential wood heating,
watercraft, non-mobile urban and other sources too small to be labeled as point sources, are the big
contributors to primary organic, while on- and off-road mobile emissions plus area sources are large
contributors to the EC at Olympic NP. The 2018 projected growth in area sources and decrease in
emissions of mobile source emissions is anticipated to increase the haze by primary OC while
reducing the haze by EC at Olympic NP The same analysis applied to San Gorgonio (Figure 9-39), is
similar in that the majority of the emissions with the potential to contribute to primary OC and EC
PM is from the home state, California in this case. However, the likely importance of natural fire
emissions for carbonaceous PM species sites is substantially greater at San Gorgonio W. than it was
for Olympic NP.
The weighted emissions potential results applied to Yellowstone NP and Grand Canyon NP
(Figure 9-40 and Figure 9-41) show the likely dominance of natural fire emissions in the
intermountain western U.S. to primary OC and EC PM during worst haze conditions for these two
locations. Numerous states have emissions that have the potential to contribute noticeably to these
carbonaceous species, due to relatively long multi-day trajectories (500-1,000 km) on worst haze
days, though for both sites the home state has the greatest potential based on this inverse distance
weighting approach. On- and off-road mobile sources in Arizona and California have significant
potential to contribute to Grand Canyon carbonaceous particles, especially EC concentrations,
probably due to some of the trajectories being over the populated areas of these two states to the
south and southwest of Grand Canyon.
December 2009 9-52
-------
Sources and Areas of Potential Organic Carbon Emissions on Worst 20% Visibilty Days
Class 1 Areas - Grand Teton NP, WY: Red Rock Lakes NVVRVV, MT: Teton W, VW: Yellowstone NP, WY
O
1
0.00
WRAP TSS-3
70.00
33
O
-
^ 0 =• | |
Sources and Areas of Potential Elemental Carbon Emissions on Worst 20% Visibility Days
Class I Areas - Grand Teton NP, VW: Red Rock Lakes NWRVV, MT: Teton W, WY: Yellowstone NP, WY
•*r o_
^_
-^™
ff
^— ^—
Q_
^^ I— ^
•^-Q_T3-Q_-^-Q_^-Q_
^m
•ft
^=
^^
CL
™*
•
PTSS-M10&
WBDust
Fugitive Dust
Q
• Road Dust
Off-Road Mobile
i—
5
| On-Road Mobile
Off-Shore
i—
•WRAP Area OSG
DArea
1
Biogenic
Natural Fire
|
| Anthro Fire
• Point
Source: From the TSS website (see Table 9-1).
Figure 9-40. Results of the weighted emissions potential tool applied to primary OC emissions
(top) and EC emissions (bottom) for the baseline and projected 2018 emissions
inventories for Yellowstone NP. Only source regions (WRAP states and other
regions) with the largest estimated contributions are shown (i.e., California,
Idaho, Montana, Oregon, Utah, Washington, and Wyoming from left to right). The
scale is normalized (i.e., unitless) one over distance weighted emissions
multiplied by trajectory residence time.
December 2009
9-53
-------
BO.00
Sources and Areas of Potential Organic Carbon Emissions on Worst 20% Visibility Days
Class I Area - Grand Canyon NP, AZ
50.00
40.00
30.00
20.00
10.00
0.00
Sources and Areas of Potential Elemental Carbon Emissions on Worst 20% Visibility Days
Class I Area - Grand Canyon NP, AZ
50.00
g
•WBDust • Road Dust | On-Road Mobile QwRAP Area OSG Biogenic
DFugitive Dust Off-Road Mobile ^Off-Shore DArea Natural Fire
Source: From the TSS website (see Table 9-1).
| Anthro Fire
I Point
Figure 9-41. Results of the weighted emissions potential tool applied to primary OC emissions
(top) and EC emissions (bottom) for the baseline and projected 2018 emissions
inventories for Grand Canyon NP. Only source regions (WRAP states and other
regions) with the largest estimated contributions are shown (i.e., Arizona,
California, Mexico, New Mexico, Nevada, Oregon, Pacific Off-shore and Utah from
left to right). The scale is normalized (i.e., unitless) one over distance weighted
emissions multiplied by trajectory residence time.
December 2009
9-54
-------
30.00
Sources and Areas of Potential Organic Carbon Emissions on Worst 20% Visibility Days
Class I Area - Badlands NP, SD
WRAP Tss-aaimE o ^ uj
30.00
1
3.00
0.00
Sources and Areas of Potential Elemental Carbon Emissions on Worst 20% Visibility Days
Class I Area - Badlands NP, SD
AZ - 2000-04 -
201 8 PRP -
CA - 2000-04 -
201 8 PRP -~||
WRAP TSS-3O1OI13
CAN - 2000-04 . ' I
201 8 PRP '[ |
_
-
Hn
nn
CEN -2000-04 |
201 3 PRP f
CO -2000-04
201 8 PRP- yj
BUS - 2000-04 Jj]
201 8 PRP -F
nn
F
O CO S
CN O CN
i CN i
Q •"
p
rn
=
•
iVOlVOIVOfVOly
Q-^O-^GL^CL^CL
coocoocoocooco
OCNOCNOCNDCNO
CN , CN , CN , CN i CN
Q > ce Q
-z. -z. o no
UT - 2000-04 - I]
201 8 PRP- ||
RR
S Ł 3 Ł
o °- o °-
0 CO 0 CO
O T- O T-
CN O CN O
i CN i CN
• WBDust • Road Dust | On-Road Mobile [] WRAP Area 0&G Biogenic HAnthroFire
D Fugitive Dust Off-Road Mobile ^Off-Shore Area Natural Fire | Point
Source: From the TSS website (see Table 9-1).
Figure 9-42. Results of the weighted emissions potential tool applied to primary OC emissions
(top) and EC emissions (bottom) for the baseline and projected 2018 emissions
inventories for Badlands NP. Only source regions (WRAP states and other
regions) with the largest estimated contributions are shown (i.e., Arizona,
California, Canada, CenRAP, Colorado, eastern U.S., Idaho, Montana, North
Dakota, Nevada, Oregon, South Dakota, Utah, Washington, and Wyoming from
left to right). The scale is normalized (i.e., unitless) one over distance weighted
emissions multiplied by trajectory residence time.
December 2009
9-55
-------
Sources and Areas of Potential Organic Carbon Emissions on Worst 20% Visibility Days
Class I Area - Salt Creek NVVRVV, NM
40.00
36.00
36.00
Sources and Areas of Potential Elemental Carbon Emissions on Worst 20% Visibility Days
Class I Area - Salt Creek NWRW, NM
•WBDust • Road Dust | On-Road Mobile O WRAP Area Q&3 Biogenic • Arrthro Fire
D Fugitive Dust Off-Road Mobile ~ Off-Shore []Area Natural Fire | Point
Source: From the TSS website (see Table 9-1).
Figure 9-43. Results of the weighted emissions potential tool applied to primary OC emissions
(top) and EC emissions (bottom) for the baseline and projected 2018 emissions
inventories for Salt Creek W. Only source regions (WRAP states and other
regions) with the largest estimated contributions are shown (i.e., Arizona,
California, CenRAP, Colorado, eastern U.S., Idaho, Mexico, Montana, New Mexico,
Nevada, Oregon, Utah, and Wyoming from left to right). The scale is normalized
(i.e., unitless) one over distance weighted emissions multiplied by trajectory
residence time.
December 2009
9-56
-------
For the most easterly of the selected WRAP sites, Badlands NP and Salt Creek, the weighted
emissions potential results for primary OC and EC (Figure 9-42 and Figure 9-43) show potential
contributions from a greater number of states and multi-state regions than for selected sites further to
the west. This may be due in part to trajectories associated with worst haze conditions for these two
sites being moderately long (-500 km) and in multiple directions. Natural fire emissions have the
greatest potential to contribute to organic species PM at Badlands NP, but are less likely to be
dominant at Salt Creek or at either site in its contribution to EC PM concentrations. The
contributions by emissions from area and mobile sources from the home states and states to the east
(Central States Regional Air Partnership states are labeled "CEN" in the figures) are potentially
greater than by natural fire; this is especially true for contributions to EC PM.
WRAP applied the weighted emissions potential tool to assess likely source types and regions
contributing to coarse mass concentrations. The results for the six selected monitoring sites (not
shown) are as follows. Most dust at Olympic NP is likely to be from fugitive dust sources in
Washington state, while at San Gorgonio it is likely more from road dust with smaller amounts from
fugitive dust sources. The amount from wind-blown dust is small for both of these far westerly sites.
Wind-blown dust is likely the largest source contributing to coarse mass at Grand Canyon NP,
Badlands NP and Salt Creek W with most of it originating in the home-state for those sites. The
weighted emissions potential results for coarse mass at Yellowstone are different from those of the
other five selected sites in that Idaho and Montana each have a higher potential to contribute to
coarse mass on the worst haze days than the home state (Wyoming), and that wind-blown and road
dust both contribute substantially as does fugitive dust and natural fire.
In another WRAP-sponsored effort to better understand the causes of remote area haze in the
western U.S., each of the worst haze days for all western IMPROVE monitoring sites where dust
(defined as the sum of coarse mass and fine soil PM) was the largest contributor to light extinction
was separately assessed to categorize the most likely dust source (Kavouras et al., 2007, 156630;
Kavouras et al., 2009, 191976) and the Causes of Haze Website - see Table 9-1). Elemental
composition was used to assess the likelihood that the dust was associated with long-range transport
from Asia. A regression analysis at each site between dust concentrations and coincident local wind
speed was used to generate site-specific estimates of local windblown dust for each sample period.
Finally, back trajectory analysis combined with maps constructed of wind erosion potential
(i.e., developed by combining soil types and land cover classifications) are used in a manner similar
to the weighted emissions potential analysis to identify the likelihood of regionally transported
wind-blown dust as the source. These assessments were conducted on each of the 610 so-called
"worst dust haze days" at 70 monitoring sites for data from 2001-2003 to classify each day by its
likely contributions from Asian dust, local windblown dust, upwind transport and undetermined. The
undetermined category includes those sample periods that failed to be classified into one of the other
three source categories suggesting that mechanically suspended dust activities (e.g., unpaved road
dust, agricultural, construction and mining activities) may be responsible.
Of the 610 "worst dust haze days" at the 70 WRAP monitoring sites, 55 sample periods are
classified as Asian dust influenced, almost exclusively in the spring; 201 sample period are classified
as local windblown dust, mostly in the spring but some in all seasons; 240 sample periods are
classified as upwind transported dust, with a broader seasonal distribution centered on summer and
few instances during winter; and 114 are in the undetermined category with most in summer and
least in winter. Most dust days occurred in the deserts of Arizona, New Mexico, Colorado, western
Texas and southern California, and these were dominated by local and regionally transported
wind-blown dust (e.g., 84% for Salt Creek W). Asian dust caused only a few of the worst dust days
during the 3-year assessment period, though it is an important source (i.e., 10-40% of the worst dust
days) for sites in the more northern regions of the West with greater vegetative land-cover where
local and regionally transported wind-blown dust was infrequent. The frequency of worst dust events
classified as undetermined was greatest for sites in the vicinity of large urban and agricultural areas
such as those in California and southern Arizona.
December 2009 9-57
-------
O
*=
u
X
LU
O)
100 -
.2 80 -
Ł 60 H
40 -
Sulfate Haze Source Attribution
• Carbon • Other Mexico
• Texas • Eastern US
D Western US D Other
Organics + LAC + Nitrates +
Fine Soil + Coarse
July 9
August 9
September 9
October 9
Source: Reprinted with Permission of the Air & Waste Management Associaiton from Pitchford et al. (2005,156874)
Figure 9-44. BRAVO study haze contributions for Big Bend NP, TX during a 4-mo period in
1999. Shown are impacts by various particulate SO/' sources, as well as the
total light extinction (black line) and Rayleigh or clear air light scattering.
December 2009
9-58
-------
2004 Annual
ammNOSf
Figure 9-45. Maps of spatial patterns for average annual particulate nitrate measurements
(top), and for ammonia emissions for April 2002 from the WRAP emissions
inventory (bottom).
December 2009
9-59
-------
NO
WRAP 36k base02a All Sources Emissions
2002 Yearly Total
1.250
0.000
LOG(tons/yea»)
December 31,2002 0:00:00
Min= 0.000 at (4,1), Max= -1.995 at (130,68)
NO2
WRAP 36k base02aAII Sources Emissions
2002 Yearly Total
0.000 1
LOG(tons/yeal)
PAUE
bv
WCNC
December31,2002 0:00:00
Min= 0.000 at (4,1 ),Max= 4.041 at (130,68)
Figure 9-46. Maps of spatial patterns of annual NO (left) and N02 (right) emissions for 2002
from the WRAP emissions inventory.
Source attribution of the participate SO42~ contribution to haze at Big Bend NP, TX was a
primary motivation for the BRAVO study. Schichtel et al. (2005, 156957) showed that during the
four-month field monitoring study (July-October 1999), SO2 emissions sources in the U.S. and
Mexico were responsible for -55% and -38% of the particulate SO42~, respectively. Among U.S.
source regions, Texas was responsible for -16%, eastern U.S. -30%, and the western U.S. -9%. A
large coal fired power plant, the Carbon facility in Mexico, just south of Eagle Pass, TX, was
responsible for -19%, making it the largest single contributor. Pitchford et al. (2005, 156874) put
these results into the context of other component contributions to regional haze, plus seasonal and
longer-term variations in haze by particulate components. Figure 9-44 shows the temporal variation
of the contributions by the various SO2 emissions source regions plus the Carbon facility during the
BRAVO study period. The largest particulate SO42~ peak haze periods are dominated by infrequent
large contribution by emission sources in TX and the eastern U.S., while Mexican sources including
the Carbon facility are more frequent contributors to haze, but at generally lower light extinction
values. Particulate nitrate contributions to haze at Big Bend NP are among the lowest measured in
the U.S. (-3% of light extinction on average and for worst haze episodes).
Nitrate concentrations are a significant contributor to light extinction further to the north of
Texas in the center of the country. While SO42~ can be in particulate form though not fully
neutralized by ammonia, nitric acid from NOX emissions requires neutralization by ammonium to
become particulate ammonium nitrate. One way to explore the causes of the Midwest nitrate bulge is
to compare its spatial distribution with the spatial distributions of NOX and ammonia emissions.
Figure 9-45 shows a map of the annual average particulate nitrate concentrations (top) with a map of
ammonia emissions directly below it. Animal agriculture is responsible for most of the ammonia
emissions in the Midwest. The striking similarity between the ambient particulate nitrate
concentration and the ammonia emissions spatial patterns with regional maximum centered on Iowa
is in contrast to the NOX (i.e., NO + NO2) emissions spatial patterns, shown in Figure 9-46. NOX
emissions are high over a broad region of the country associated with the larger population densities
and greater numbers of fossil fuel electric generation plant generally to the east of the Midwest
nitrate bulge. While both ammonia and nitric acid are needed to form particulate ammonium nitrate,
the maps suggest the Midwest nitrate bulge is due primarily to the abundance of free ammonia (i.e.,
the amount beyond what is required to neutralize the acidic particulate SO42~). By contrast, the
region to the east of the Midwest nitrate bulge should have plenty of nitric acid given the higher
emissions of NOX, but apparently has a deficiency of free ammonia. The few eastern monitoring
sites with locally high particulate nitrate (near southeastern PA) are located within a small region of
high density animal agriculture that shows up as a high ammonia emissions region in Figure 9-45.
Note that California's South Coast and Central Valley have both high ammonia and high NOX
emissions, explaining the high particulate nitrate contribution to haze there.
December 2009
9-60
-------
To better understand the role of ammonia in the formation of the Midwest nitrate bulge, the
Midwest RPO and Central States Regional Air Partnership deployed a measurement program from
late 2003 through early 2005 at 10 locations (9 rural and 1 urban) in the region (see Figure 9-47) to
monitor particulate SO42~, nitrate, and ammonium ions, plus the precursor gases sulfur dioxide, nitric
acid, and ammonia (Kenski et al, 2004, 192078; Sweet et al., 2005, 180038). These data have been
used as input for thermodynamic equilibrium modeling to assess the changes in PM concentrations
that would result from changes to precursor concentrations (Blanchard and Tanenbaum, 2006,
190005; Blanchard et al., 2007, 098659). Blanchard and Tanenbaum (2006, 190005) and Blanchard
et al. (2007, 098659) conclude that the current conditions at nine of the ten sites are near the point of
transition between the precursor species (nitric acid and ammonia) that limits the formation of
particulate nitrate. If excess ammonia increases, either by greater ammonia emissions or by
anticipated decreases in SO2 emissions, then nitric acid concentration would need to be reduced (via
lower NOX emissions) in order to reduce the particulate nitrate concentration.
Given the comparability of particulate SO42~ and nitrate with regard to their light extinction
efficiencies, their visibility impacts are proportional to the sum of their mass concentrations. A
reduction in SO42~ caused by SO2 emission reductions would reduce the particulate SO42~
concentration, though according to the thermodynamic equilibrium modeling for these sites the
particulate nitrate concentration will be increased somewhat. However, the total particulate SO42~
plus nitrate concentration would be reduced so visibility impacts would be decreased. At current
ammonium concentrations the predicted response of changes to SO42~ and nitric acid concentrations
(i.e., SO2 and NOX emissions changes) are similar in respect to the resulting magnitude of changes to
the total particulate SO42~ plus nitrate concentrations. At all but two sites the total particulate SO42~
plus nitrate concentrations would decrease if either ammonia or nitric acid where reduced.
Source: Kenski et al. (2004,1920781
Figure 9-47. Midwest ammonia monitoring network.
A further degree of complications in understanding the response of particulate nitrate to
changes in precursor concentrations results from the temperature and humidity dependence of the
partition between particulate ammonium nitrate and the disassociated gaseous nitric acid and
ammonia. This dependence causes seasonal and even diurnal differences in the expected responses
of particulate nitrate concentrations to changes in precursor concentrations. As expected, during the
colder times of the year the total particulate concentrations are more sensitive to changes in ammonia
and nitric acid concentrations than during the warmer seasons when SO42~ concentrations are greater.
December 2009
9-61
-------
As shown in Figure 9-48, results of an air transport assessment to identify emission source
areas associated with high particulate nitrate at five monitoring locations in the East (four
remote-area sites and Toronto, Canada) implicate the high ammonia emissions region of the Midwest
as a common source region (Canada-US Air Quality Committee, 2004, 190519). This assessment
does not preclude local sources of the precursor gases responsible for parti culate ammonium nitrate,
but does suggest that long-range transport of parti culate nitrate or ammonia from the high emissions
region of the Midwest is also contributing to eastern nitrate episodes.
Upwind Probability Fields for Ammonium Nitrate
Lye Brook VT,
real Smoky Mountains NP
Source: Canada-U.S. Air Committee (2004,190519).
Figure 9-48. Upwind transport probability fields associated with high participate nitrate
concentrations measured at Toronto, Canada; Boundary Water Canoe Area, MN;
Shenandoah NP, VA; Lye Brook, VT; and Great Smoky Mountains NP, TN.
In a similar air transport assessment for measurements at Underbill, VT and at Brigantine, NJ,
Hopke et al. (2005, 156567) identified separate regions associated with particulate SO42~
accompanied by trace particulate components associated with coal burning (e.g., Se) and
accompanied by trace particulate components associated with oil burning (e.g., V). As shown in
Figure 9-49, the coal-burning related particulate SO42~ for these two monitoring sites is associated
with long-range transport from the Ohio River Valley, while oil-burning related particulate SO42~ is
from more nearby emissions in the high population region of coastal New York, New Jersey,
Massachusetts, and Connecticut.
December 2009
9-62
-------
Source: Reprinted with Permission of Environmental Science and Technology from Hopke, et al. (2005,156567).
Figure 9-49. Trajectory probability fields for periods with high participate SQ*~ measured at
Underbill, VT and Brigantine, NJ (shown as white stars) associated with
oil-burning trace components (left) and with coal-burning trace components
(right). Shown for comparison are the interpolated 862 emissions areal density
contours for oil combustion sources (emissions times 10) and coal combustion
sources, displayed as yellow and red contour lines, respectively.
The Regional Aerosol Intensive Network (RAIN) was established by MANE-VU to generate
enhanced continuous visibility, plus fine particle mass and composition monitoring data at a string of
three monitoring locations along the transport path from the Ohio River Valley to coastal Maine
(NESCAUM, 2006, 156802). The dominant role of particulate SO42~ in the northeast is well
demonstrated by a scatter plot of RAIN data that shows the relationship between particulate SO42~
extinction, calculated using the IMPROVE algorithm plotted against directly measured particle light
scattering for hourly data over a 8-mo period, beginning in July 2004 at the Acadia NP, ME
monitoring site (see Figure 9-50). Particulate SO42~ explains 90% of the variance in particulate light
scattering even though it is responsible for only about 64% of the total light extinction (annual
averaged value from the VIEWS web site). Adding the contribution by the second-largest regional
contributor to light extinction, particulate OC with about 14%, does not improve the variance
explained, but does increase the slope to 0.78. The noticeable difference between these two plots is
that the particulate SO42~ alone underestimates light scattering during low haze periods (points on the
plot are below the regression line for light scattering <70 Mm"1), while the agreement is improved
with the addition of particulate OC contributions to haze (regression slope is nearer to one and
reduced bias for low haze periods). This implies that particulate SO42~ and any other co-varying PM
species are largely responsible for the highly impacted periods, while OC and other co-varying PM
species contribute more during the less extreme haze periods. Particulate nitrate contribution to light
extinction at Acadia is about 10% on average.
December 2009
9-63
-------
Nephelometer Bsp versus SULFATE Bsp
Nephelometer Bsp vs {SULFATE + OMC)
50 100 150 200 250 300 3:
50 100 150 280 250 300 3!
2-Hour Nephelometer Bsp (1/Mmeters)
Figure 9-50.
2-
2-Hr Nephelometer (1,'Mmeters)
Source: RAIN Preliminary Data Analysis Report (NESCAUM, 2006,156802'
2-
Scatter plots of participate SO/' (left) and participate S04Z" and organic mass
(right) versus nephelometer measured particle light scattering for Acadia NP, ME.
Particulate nitrate concentrations are considerably lower in the SO42 -dominated warmer
southeastern U.S. than in the Northeast and upper Midwest. Blanchard et al. (2007, 098659)
conducted thermodynamic equilibrium modeling on data from the eight SEARCH monitoring sites
and found that total particulate nitrate plus SO42~ is much more responsive to changes in SO2
concentrations than to changes in nitric acid concentrations, which in turn is more responsive than
changes in ammonia concentrations.
The VISTAS RPO commissioned an emissions sensitivity study using CMAQ modeling on
winter and summer 2009 emissions projected from the 2002 emissions inventory (NCDENR, 2007,
156798). Figure 9-51 contains bar plots for two North Carolina Class I areas that indicate projected
changes in light extinction for the worst haze day due to 30% emissions reductions by particulate
species, source types and location across the Southeastern states modeling domain (i.e., as far west
as Texas, as far north as Pennsylvania, as far south as the Florida Keys, and as far east as -300 km
from the North Carolina coast). Great Smoky Mountains in the southern Appalachian Mountains has
the greatest sensitivity to changes by SO2 emissions from electrical generation units (EGU) and to a
lesser extent other SO2 emission sources in the region. Reductions of NOX emissions from ground or
point sources are not nearly as effective as SO2 reductions in reducing the light extinction at Great
Smoky Mountains. This is due principally to the worst days at Great Smoky Mountains occurring
during the summer, when temperatures are too high to support high particulate nitrate
concentrations. For the same reason, ammonia emission reductions are also ineffective. Swanquarter
W, NC is a coastal location where some of the worst haze days are during the winter and include
contributions from particulate ammonium nitrate. Both SO2 and ammonia emissions reductions
would be effective at reducing worst haze days at the Swanquarter W, though NOX emissions are not
as effective presumably because the atmosphere is ammonia-limited for particulate nitrate
production.
December 2009
9-64
-------
Great Smoky Mtns, TN (20% Worst Days)
• Bio VOC.
D Anthro VOC
• BCs
D MRPO
• M-VU
DCEN
• VISTAS
• wv
n VA
DTN
DSC
• NC
• MS
DKY
DGA
• FL
DAL
Swanquarter, NC (20% Worst Days)
• Bio VOC
D Anthro VOC
• BCs
• MRPO
• M-VU
DCEN
• VISTAS
• WV
DVA
DTN
DSC
• NC
• MS
DKY
DGA
• FL
DAL
Source: NCDENR (2007, 156798).
Figure 9-51. CMAQ air quality modeling projections of visibility responses on the 20% worst
haze days at Great Smoky Mountains NP, NC (top) and Swanquarter W, NC
(bottom) to 30% reductions. This is from a projected 2009 emission inventory of
visibility-reducing pollutants by source category and geographic areas.
9.2.4. Urban Visibility Valuation and Preference
The Clean Air Act §302(h) defines public welfare to include the effects of air pollution on
"... visibility, ... and personal comfort and wellbeing." Though good visibility conditions in Class I
(e.g., NPs) and wilderness areas have long been recognized as important to the public welfare (see
discussions in EPA (2004, 056905; 2005, 090209) and Chestnut and Dennis (1997, 014525).
visibility conditions in urban areas also contribute to the public welfare. Although visibility
impairment may be caused by either natural or manmade conditions (or both), it is only impairment
that occurs as a result of air pollution (either alone or in combination with water vapor or other
atmospheric conditions) that can be mitigated by regulations such as the RHR (40 CFR 51.300
through 309) or the Secondary NAAQS. The term visual air quality (VAQ) is used here to refer to
December 2009
9-65
-------
the visibility effects caused solely by air quality conditions, so for example it excludes the reduced
visibility caused by fog. Visibly poor air quality causes people to be concerned about substantive
health risks, but degraded VAQ adversely affects people in additional ways. These include the
aesthetic and wellbeing benefits of better visibility, improved road and air safety, and enhanced
recreation in activities like hiking and bicycling. Because the human health impacts of air pollution
are assessed under the Primary NAAQS, it is necessary to separate out these non-health components
associated with the visibility condition produced by a given amount of air pollution when assessing
the need for additional regulation to protect the public welfare effect of visibility under the
Secondary NAAQS. The degree to which previous human preference and valuation studies for VAQ
have adequately made this distinction and separation is an important issue in applying results from
available studies in a Secondary NAAQS (or benefits estimation for any policy affecting VAQ)
context. The remainder of this discussion is focused on those aesthetic and wellbeing qualities
associated with a given VAQ in urban areas.
The term "urban visibility" is used to refer to VAQ throughout a city or metropolitan area.
Urban visibility includes the VAQ conditions in all locations that people experience in their daily
lives, including scenes such as residential streets and neighborhood parks, commercial and industrial
areas, highway and commuting corridors, central downtown areas, and views from elevated locations
providing a broad overlook of the metropolitan area. Thus urban visibility includes VAQ conditions
in major cities and smaller towns and encompasses all the VAQ an individual resident sees on a
regular basis. Visibility conditions in urban and suburban locations are therefore distinct from
visibility in rural or wilderness settings such as the Class 1 areas defined by the Clean Air Act, which
include NPs and similar natural settings.
Visibility has direct significance to people's enjoyment of daily activities and their overall
wellbeing. Visibility conditions can be described both as an aesthetic quality as well as a
scientifically measurable set of atmospheric conditions. Due to the subjective nature of aesthetics,
people's preferences with respect to visibility are difficult to express or quantify, but people have
expressed in many different ways that they enjoy and value a clear view. A number of social science
studies have been undertaken to link perceived urban visibility to an array of effects reflecting the
overall desire for good VAQ, and the benefits of improving currently degraded VAQ. This wide
range of diverse studies have identified types of benefits of good VAQ.
For example, psychological research has demonstrated that people are emotionally affected by
low VAQ such that their overall sense of wellbeing is diminished (e.g., Bickerstaff and Walker,
2001, 156271). Researchers have also shown that perception of pollution is correlated with stress,
annoyance, and symptoms of depression (Evans and Jacobs, 1982, 179899; Jacobs et al, 1984,
156596; Mace et al., 2004, 180255). Sociological research has demonstrated that VAQ is deeply
intertwined with a "sense of place," effecting people's sense of the desirability of a neighborhood
quite apart from the actual physical conditions of the area (e.g., ABT, 2002, 156186; Day, 2007,
156386; Elliot et al., 1999, 010716; Howel et al., 2002, 156571). Public policy research finds that
people think it is important to protect visibility, and accept the concept of setting standards to protect
visibility (e.g., ABT, 2001, 156185; BBC Research & Consulting, 2002, 156258; Ely et al., 1991,
156417; Pry or, 1996, 056598). Finally, economic valuation research has measured the amount of
money that people are willing to pay to protect or improve both urban visibility (e.g., summary
review in Beron et al., 2001, 156270; Chestnut and Dennis, 1997, 014525) and natural locations
such as NPs and other locations defined by the Clean Air Act as Class I visibility area (e.g., summary
review in Chestnut and Dennis, 1997, 014525).
Urban visibility has been examined in two types of studies directly relevant to the NAAQS
review process: urban visibility preference studies and urban visibility valuation studies. The
purpose of the remainder of this section is to review preference studies in four urban areas, as well as
one new urban visibility valuation study not previously discussed in previous EPA Criteria
Documents or OAQPS Staff Papers.
Both types of studies are designed to evaluate individuals' desire (or demand) for good VAQ
where they live, using different metrics to evaluate demand. Urban visibility preference studies
examine individuals' demand by investigating what amount of visibility degradation is unacceptable
while economic studies examine demand by investigating how much one would be willing to pay to
improve visibility.
December 2009 9-66
-------
9.2.4.1. Urban Visibility Preference Studies
One group of urban visibility research projects focused on identifying preferences for urban
VAQ without necessarily estimating the economic value of improving visibility. This group of
preference studies used a common focus group method to estimate the visibility impairment
conditions that respondents described as "acceptable." The specific definition of acceptable was
largely left to each individual respondent, allowing each to identify their own preferences.
There are three completed studies that used this method, and two pilot studies that provided
additional information (Table 9-2). The completed studies were conducted in Denver, Colorado (Ely
et al, 1991, 156417). two cities in British Columbia, Canada (Pryor, 1996, 056598). and Phoenix,
Arizona (BBC Research & Consulting, 2002, 156258). The additional studies were conducted in
Washington, DC (ABT, 2001, 156185; Smith and Howell, 2009, 198803).
Each study collected information in a focus group setting, presenting slides depicting various
visibility conditions. All four studies used photographs of a single scene from the study's city; each
photo included images of the broad downtown area and spreading out to the hills or mountains
composing the scene's backdrop. The maximum sight distance under good conditions varied by city,
ranging from 8 km in Washington, DC to mountains hundreds of kilometers away in Denver.
Multiple photos of the same scene were used to present approximately 20 different visibility
impairment conditions. The Denver and British Columbia studies used actual photographs taken in
the same location to depict various visibility conditions. The Phoenix and Washington, DC studies
used photographs prepared using the WinHaze software from Air Resource Specialists (ARS).
WinHaze is a computer-imaging software program that simulates visual air quality differences of
various scenes, allowing the user to "degrade" an original near-pristine visibility condition
photograph to create a photograph of each desired VAQ condition.
A common characteristic of the three visibility preference studies was that each was conducted
in the West where distant mountains were shown in the photograph used to elicit local participant
responses about visibility. Among other issues, the Washington D.C. pilot study was the first step in
a process to expand the results to other regions where typical scenes may have different sensitivity to
perceived visibility changes in PM air quality and where participants may have different acceptable
visibility preference values.
The range of median preference values for an acceptable amount of visibility degradation from
the 4 urban areas was approximately 19-33 dv. Measured in terms of visual range (VR), these
median acceptable values were between approximately 59 and 20 km.
Table 9-2. Summary of urban visibility preference studies.
Report Date
Duration of
session
Compensation
# focus group
sessions
# participants
Age range
Annual or
seasonal
# total scenes
presented
# of total visibility
conditions
presented
Denver, CO
1991
None (civic groups)
17
214
Adults
Wintertime
Single scene of
downtown with
mountains in
background
20 conditions (+ 5
duplicates)
Phoenix, AZ
2003
45min
$50
27 total at 6 locations,
Including 3 in Spanish
385
18-65+
Annual
Single scene of
downtown and
mountains, 42 km
maximum distance
21 conditions (+ 4
duplicates)
2 British Columbia cities
1996
50 min
None (class room exercise)
4
180
University students
Summertime
Single scene from each city
20 conditions (10 each from each city)
Washington, DC
(2001)
2001
2h
$50
1
9
27-58
Annual
Single scene of DC Mall
and downtown, 8 km
maximum sight
20 conditions (+ 5
duplicates)
Washington, DC
(2009)
2009
None
3 tests
64
Adults
Annual
Single scene of DC
Mall and downtown, 8
km maximum sight
22 conditions
December 2009
9-67
-------
Source of slides
Medium of
presentation
Ranking scale
used
Visibility range
presented
Health issue
directions
Key questions
asked
Mean dv found
"acceptable"
Denver, CO
Actual photos taken
between 9am and
3pm
Slide projection
7 point scale
11 to 40 dv
Ignore potential
health impacts;
visibility only
a)RankVAQ(1-7
scale)
b) Is each slide
"acceptable"
c) "How much haze
is too much?"
20.3dv
Phoenix, AZ
WinHaze
Slide projection
7 point scale
15 to 35 dv
Judge solely on
visibility, do not
consider health
a)RankVAQ(1-7
scale)
b) Is each slide
"acceptable"
c) How many days a
year would this picture
be "acceptable"
23-25 dv
2 British Columbia cities
Actual photos taken at 1 p.m. or 4 p.m.
Slide projection
7 point scale
13-25 dv (Chilliwack) 13.5-31. 5 dv
(Abbotsford)
Judge solely on visibility, do not
consider health
a) Rank VAQ (1-7 scale)
b) Is each slide "acceptable"
-23 dv(Chilliwack),
~19dv(Abbotsford)
Washington, DC
(2001)
WinHaze
Slide projection
7 point scale
9-38 dv
Health never
mentioned, "Focus only
on visibility"
a) Rank VAQ (1-7
scale)
b) Is each slide
"acceptable"
c) if this hazy, how
many hs would it be
acceptable (3 slides
only)
d) valuation question
-20 dv (range 20-25)
Washington, DC
(2009)
WinHaze
Slide projection
7 point scale
9-45 dv
Health never
mentioned, "Focus only
on visibility"
a) Rank VAQ (1-7
scale)
b) Is each slide
"acceptable"
-30 dv
9.2.4.2. Denver, Colorado Urban Visibility Preference Study
The Denver urban visibility preference study (Ely et al, 1991, 156417) was conducted on
behalf of the Colorado Department of Public Health and Environment (CDPHE). The study
conducted a series of focus group sessions with 17 civic and community groups in which a total of
214 individuals were asked to rate slides. The slides depicted varying values of VAQ for a
well-known Denver vista, including a broad view of downtown Denver with the mountains to the
west composing the scene's background. The participants were instructed to base their judgments on
three factors:
1. The standard was for an urban area, not a pristine NP area where the standards might be
more strict;
2. The value of an urban visibility standard violation should be set at a VAQ value considered
to be unreasonable, objectionable, and unacceptable visually; and
3. Judgments of standards violations should be based on visibility only, not on health effects.
Participants were shown 25 randomly ordered slides of actual photographs. The visibility
conditions presented in the slides ranged from 11-40 dv, approximating the 10th-90th percentile of
wintertime visibility conditions in Denver. The participants rated the 25 slides based on a scale of
1 (poor) to 7 (excellent), with 5 duplicates included. They were then asked to judge whether the slide
would violate what they would consider to be an appropriate urban visibility standard (i.e., whether
the amount of impairment was "acceptable" or "unacceptable"). The individual's judgment of a
slide's VAQ and whether the slide violated a visibility standard were highly correlated (Pearson
correlation coefficient >80%), as were the VAQ ratings and the yes/no "acceptable" response. The
participant's median response was that a visibility condition of 20.3 dv (extinction coefficient
bext= 76Mm"1, or VR ~51 km) was judged as "acceptable." The CDPHE subsequently established a
Denver visibility standard at this value (defined as bext = 76Mm"1), based on the median 50%
acceptability findings from the study.
December 2009
9-68
-------
9.2.4.3. Phoenix, Arizona Urban Visibility Preference Study
The Phoenix urban visibility preference study (BBC Research & Consulting, 2002, 156258)
was conducted on behalf of the Arizona Department of Environmental Quality. The Phoenix study
patterned its focus group survey process after the Denver study. The study included 385 participants
in 27 separate focus group sessions. Participants were recruited using random digit dialing to obtain
a sample group designed to be demographically representative of the larger Phoenix population.
Focus group sessions were held at six neighborhood locations throughout the metropolitan area to
improve the participation rate. Three sessions were held in Spanish in one region of the city with a
large Hispanic population (25%), although the final overall participation of native Spanish speakers
(18%) in the study was modestly below the targeted value. Participants received $50 as an
inducement to participate.
Participants were shown a series of 25 images of the same vista of downtown Phoenix, with
South Mountain in the background at a distance of about 40 km. Photographic slides of the images
were developed using WinHaze. The visibility impairment conditions ranged from 15-35 dv (the
extinction coefficient, bext, range was approximately 45 Mm"1 to 330 Mm" , or a visual range of
87-12 km). Participants first individually rated the randomly shown slides on a VAQ scale of
1 (unacceptable) to 7 (excellent). Participants were instructed to rate the photographs solely on
visibility, and to not base their decisions on either health concerns or what it would cost to have
better visibility. Next, the participants individually rated the randomly ordered slides as "acceptable"
or "unacceptable," defined as whether the visibility in the slide is unreasonable or objectionable.
Better visibility conditions (15 dv and 20 dv) were judged "acceptable" by 90% of all participants.
At 24 dv nearly half of all participants thought the VAQ was "unacceptable," with almost
three-quarters judging 26 dv as unacceptable.
The Phoenix urban visibility study formed the basis of the decision of the Phoenix Visibility
Index Oversight Committee for a visibility index for the Phoenix Metropolitan Area (Arizona DEQ,
2000, 019164). The Phoenix Visibility Index establishes an indexed system with 5 categories of
visibility conditions, ranging from "Excellent" (14 dv or less) to "Very Poor" (29 dv or greater). The
"Good" range is 15-20 dv. The environmental goal of the Phoenix urban visibility program is to
achieve continued progress through 2018 by moving the number of days in lower quality categories
into better quality categories.
9.2.4.4. British Columbia, Canada Urban Visibility Preference Study
The British Columbia urban visibility preference study (Pryor, 1996, 056598) was conducted
on behalf of the Ministry of Environment. The study conducted focus group sessions that were also
developed following the methods used in the Denver study. Participants were students at the
University of British Columbia, who participated in one of four focus group sessions with between 7
and 95 participants. A total of 180 respondents completed surveys (29 did not complete the survey).
Participants in the study were shown slides of two suburban locations in British Columbia:
Chilliwack and Abbotsford. Using the same general protocol as the Denver study, Pry or found that
responses from this study found the acceptable level of visibility was 23 dv in Chilliwack and 19 dv
in Abbotsford. Pry or (1996, 056598) discusses some possible reasons for the variation in standard
visibility judgments between the two locations. Factors discussed include the relative complexity of
the scenes, potential bias of the sample population (only University students participated), and the
different amounts of development at each location. Abbotsford (population 130,000) is an ethnically
diverse suburb adjacent to the Vancouver Metro area, while Chilliwack (population 70,000) is an
agricultural community 100 km east Vancouver in the Frazier Valley.
The British Columbia urban visibility preference study is being considered by the B.C.
Ministry of the Environment as a part of establishing urban and wilderness visibility goals in British
Columbia.
9.2.4.5. Washington, DC Urban Visibility Preference Studies
The Washington, DC urban visibility pilot study (ABT, 2001, 156185) was conducted on
behalf of the EPA, and was designed to be a pilot focus group study, an initial developmental trial
run of a larger study. The intent of the pilot study was to study both focus group method design and
potential survey questions. Due to funding limitations, only a single focus group session was held,
December 2009 9-69
-------
consisting of one extended session with nine participants. No further urban visibility focus group
sessions were held in Washington, DC.
Due to the small number of participants, it is not possible to make statistical inferences about
the opinions of the general population. The study does, however, provide additional useful
information about urban visibility studies, potentially helping to both better understand previous
studies as well as design future studies.
The study also adopted the general Denver study method, modifying it as appropriate to be
applicable in an eastern urban setting which has substantially different visibility conditions than any
of the three western locations of the other preference studies. Washington's (and the entire East)
visibility is typically substantially worse than western cities, and has different characteristics.
Washington's visibility impairment is primarily a uniform whitish haze dominated by sulfates,
relative humidity values are higher, the low lying terrain provides substantially shorter maximum
sight distances, and many residents are not well informed that anthropogenic emissions impair
visibility on hazy days.
The Washington focus group session included questions on valuation, as well as on
preferences. The focus group was asked to state its preferences measured in an increase in the
general cost of living for certain increments of improvement in visibility on atypical summer day. A
general cost of living approach is one payment vehicle approach that can be used in willingness to
pay studies, especially for environmental issues arising from multiple diverse emission sources (e.g.,
transportation, electricity generation, industry, etc.) making a specific price increase potentially
misleading.
The first part of the focus group session was designed to be an hour long, and was comparable
to the focus group sessions in the Denver and Phoenix studies. A single scene was used; a panoramic
shot of the Potomac River, Washington mall and downtown Washington, DC. In the first part of the
session people were asked to rate the VAQ of 25 photographs (prepared using WinHaze, and
projected on a large screen), judge the acceptability of visibility condition in each slide, and answer
the valuation questions. The second half of the session, however, was a moderated discussion session
about the format and content of the first phase of the session. In this moderated discussion,
participants were asked about their understanding of each question asked in the first half of the
session. Particular issues in designing a focus group session were also explored. Important
participant comments included:
1. Participants had been asked how they reacted to the initial direction to base their answers
only on visibility, but health was never explicitly mentioned by the focus group moderator.
Participants strongly agreed with the decision to not mention that health effects are
associated with visibility impairment. They understood the directions as meaning they should
ignore health issues, and said their answers would have been different if they included health
as well as visibility in their judgments.
2. Differentiating between haze and weather conditions was difficult. Weather was not
discussed in the focus group session, and the photographs were WinHaze altered photos with
identical weather conditions. Participants mentioned they were still confused about the role
of weather and humidity in the different visibility conditions presented in the photos.
3. Questions about how many hours an impairment level would be acceptable were confusing.
Most participants were normally indoors during most of the day, so questions about duration
of outdoor conditions were difficult to answer.
4. Participants strongly agreed that not mentioning the purpose of the study, or the sponsor,
until the very end (after all the questions were answered) was viewed as very important.
Most felt this information would have influenced their answers.
The Smith and Howell (2009, 198803) study recreated the same WinHaze images used in the
2001 Washington, DC urban visibility preference study, and followed a shortened version of the
same question protocol as the 2001 study. The WinHaze images were presented to a total of
64 participants who were all employees of CRA International, Inc. (Smith and Howell also are CRA
International employees).
The stated purpose of the Smith and Howell (2009, 198803) study was to explore the
robustness of the 2001 pilot study results. To investigate this issue, Smith and Howell (2009,
198803) conducted three different tests concerning urban visibility preferences. Each participant was
involved with only one test. Test 1 was designed to replicate the 2001 study. Test 2 reduced the
December 2009 9-70
-------
upper end of the range of VAQ by eliminating the 11 images used in Test 1 with a VAQ above
27.1 dv. Test 3 increased the upper end of the range of VAQ by including two new images of worse
VAQ; the two new images had a VAQ of 42 dv and 45 dv. Smith and Howell (2009, 198803)
concluded that changing the range of VAQ presented to the participants affects the responses about
whether a particular VAQ is acceptable.
9.2.4.6. Urban Visibility Valuation Studies
The one recent urban visibility benefit assessment not included in earlier reviews is "The
Benefits of Visibility Improvement: New Evidence from the Los Angeles Metropolitan Area" (Beron
et al., 2001, 156270). Rather than a contingent valuation method (CVM) technique used in the
majority of other urban visibility valuation studies, Beron et al. (2001, 156270) used a housing
market hedonic technique. The housing hedonic methods were used in previous urban visibility
studies by Murdoch and Thayer (1988, 156788) and Trijonis et al. (1985, 078468). A housing market
hedonic study views a housing unit as composed of a bundle of attributes, and uses housing sale
price data from a large number of units in a metropolitan area to estimate the value of each
component. Hedonic pricing has been used to estimate economic values for environmental effects
that have a direct effect on housing market values. It relies on the measurement of differentials in
property values under various environmental quality conditions including air pollution, visibility and
other environmental amenities such as access to nearby beaches and parks, as well as by physical
attributes of the house and attributes of the neighborhood.
Beron et al. (2001, 156270) obtained data on approximately 840,000 owner-occupied, single
family housing sales between 1980 and 1995 from the California South Coast Air Basin (composed
of Los Angeles and Orange Counties, and the portions of Riverside and San Bernardino Counties in
the greater metropolitan area). The real estate data included information on the sale price of the
house, 13 housing attributes (square footage, number of bathrooms, etc.), 9 neighborhood attributes
(percent poverty, school quality, FBI crime index, etc.), and three air pollution variables: ozone,
particulates (measured by total suspended particulates, or TSP), and visibility. Visibility was
measured as the annual average of visual range, measured in miles, and was obtained from seven
airports within the study region. The visibility range was from 12.4 miles (Los Angeles International
Airport, 1991) to 31.9 miles (Palm Springs Airport, 1995). Ozone data (39 monitors) and TSP data
(40 monitors) were obtained from the South Coast Air Quality Management District. Annual mean
values for each year were calculated for ozone and TSP.
Beron et al. (2001, 156270) presented results for a hypothetical basin-wide 20% visibility
improvement, or an increase from 15.3 to 18.4 miles, which is equivalent to approximately 27.6 dv
to 25.8 dv. The initial results reflect the change in the purchase price of a house associated with this
difference in VAQ, which can be interpreted as a present value of a stream of annual values over the
lifetime of the house. The authors therefore selected a time horizon (30 yrs) and an interest rate (8%)
to calculate an annual per household benefit per dv ranging from $484 to $1,756. The Beron results
are higher than the CVM-based values summarized in Chestnut and Dennis (1997, 014525). which
ranged from $12 to $132 per dv. It should be noted that the $132 CVM values cited by Chestnut and
Dennis (1997, 014525) is from a study in the Los Angeles area (Brookshire, 1979, 156298). The
Beron et al. (2001, 156270) results are also higher than the Trijonis et al. (1990, 157058) hedonic
study in the Los Angeles area, which had a range of $134 to $360 per dv per year. All values
reported here are in terms of 1994 prices.
A critical question for all urban visibility valuation studies is the extent to which the estimated
values strictly reflect preferences for visibility, and do not include a component of preferences for
reducing health risk from air pollution. The ability to isolate the value of visibility from within the
collection of intertwined benefits from visual air quality, which is inherently multi-attributed, is a
challenge for all visibility valuation studies. Each study attempts to isolate visibility from other
effect categories, but different studies take different approaches.
Beron et al. (2001, 156270) include two measures of air pollution directly related to health
effects in their housing market hedonic study, ozone and particulates (using TSP as the metric for
particulates), as well as visibility. They argue that the presence of the two health-related pollution
conditions results in an estimated hedonic demand function for visibility that successfully separates
the health component of demand for overall air quality from the visibility component. An alternative
interpretation is that the estimated visibility function still includes a component of health risk
because the housing market data does not support completely isolating the demand for visibility (due
December 2009 9-71
-------
to correlated variables, omitted variables, measurement error, model specification error, etc.) from
demand for health risk reductions measured by the two health related air quality metrics.
A key issue in interpreting the Beron et al. (2001, 156270) results is whether the objective
measures of air quality characteristics (e.g., visibility, PM concentrations, etc.) capture people's
perceptions of the different aspects of air quality in a given location. To the extent the people
simultaneously use what they see regarding VAQ as an indicator of the overall air quality including
potential health risks, then including all the measures in the equation is not necessarily sufficient to
isolate one effect from the other.
9.2.5. Summary of Effects on Visibility
Visibility impairment is caused by light scattering and absorption by suspended particles and
gases. NO2 is the only commonly occurring atmospheric pollutant gas that absorbs visible spectrum
radiation, though in most situations the amount of light absorption by NO2 is overwhelmed by the
higher amounts of particulate light extinction (i.e., the combination of scattering and absorption)
usually accompanying high NO2 concentrations. Light scattering by gases in a pollutant-free
atmosphere provides a limit to visibility in pristine conditions and is the largest contributor to the
total light extinction during the least visibility-impaired periods in remote regions of the western
U.S. There is strong and consistent evidence that PM is the overwhelming source of visibility
impairment in both urban and remote areas. EC and some crustal minerals are the only commonly
occurring airborne particle components that absorb light. All particles scatter light, and generally
light scattering by particles is the largest of the four light extinction components. Although a larger
particle scatters more light than a similarly shaped smaller particle of the same composition, the light
scattered per unit of mass is greatest for particles with diameters from approximately 0.3-1.0 um.
For studies where detailed data on particle composition by size data are available, accurate
calculations of light extinction can be made. However, routinely available PM speciation data can be
used to make reasonable estimates of light extinction using relatively simple algorithms that multiply
the concentrations of each of the major PM species by its dry extinction efficiency and by a water
growth term that accounts for particle size change as a function of relative humidity for hygroscopic
species (e.g., SO42~, nitrate, and sea salt). This permits the visibility impairment associated with each
of the major PM components to be separately approximated from PM speciation monitoring data.
There are six major PM components: PM2 5 SO4 ~ usually assumed to be ammonium sulfate, PM2 5
nitrate usually assumed to be ammonium nitrate, PM2 5 OC, PM2 5 EC, PM2 5 crustal material
(referred to as fine soil), and PMi0_2.s or coarse mass.
Direct optical measurement of light extinction measured by transmissometer, or by combining
the PM light scattering measured by integrating nephelometers with the PM light absorption
measured by an aethalometer offer a number of advantages compared to algorithm estimates of light
extinction based on PM composition and relative humidity data. The direct measurements are not
subject to the uncertainties associated with assumed scattering and absorption efficiencies used in the
PM algorithm approach. The direct measurements have higher time resolution (i.e., minutes to
hours), which is more commensurate with the visibility effects compared with calculated light
extinction using routinely available PM speciation data (i.e., 24-h duration).
Particulate SO42~ and nitrate are produced in the atmosphere from gaseous precursors, making
them secondary PM species. They both have comparable light extinction efficiencies (haze impacts
per unit mass concentration) at any relative humidity value, their light scattering per unit mass
concentration increases with increasing relative humidity, and at sufficiently high humidity values
(RH>85%) they are the most efficient particulate species contributing to haze. Particulate SO42~ is
the dominant source of regional haze in the eastern U.S. (>50% of the particulate light extinction)
and an important contributor to haze elsewhere in the country (>20% of particulate light extinction).
Particulate nitrate is a minor component of remote-area regional haze in the non-California
western and eastern U.S., but an important contributor in much of California and in the upper
Midwestern U.S. especially during winter when it is the dominant contributor to particulate light
extinction. While both nitric acid (a reaction product of NOX emissions) and ammonia are needed to
form ammonium nitrate, the apparent reason for the Midwest nitrate bulge (i.e., region of high winter
PM nitrate) is an abundance of atmospheric ammonia in this region principally from agricultural
emissions. There is evidence that transport from the Midwest nitrate bulge region is responsible for
some of the ammonium nitrate episodes experienced in downwind regions far to the east. Urban
particulate nitrate concentrations are significantly elevated above surrounding remote-area
December 2009 9-72
-------
background concentrations with the largest urban contributions in the western U.S. Particulate
ammonium nitrate concentrations in California and the Midwestern nitrate bulge region are an order
of magnitude greater than estimated natural ammonium nitrate concentrations. Thermodynamic and
air quality simulation modeling show that particulate nitrate concentrations are sensitive to changes
in either NOX emissions (from a combination of mobile and point sources) or ammonia emissions
(principally from agricultural sources), with the responsiveness of particulate nitrate to emissions
changes depending on the relative abundance of ammonia and nitric acid in the atmosphere.
EC and OC have the highest dry extinction efficiencies of the major PM species and are
responsible for a large fraction of the haze especially in the Northwestern U.S., though absolute
concentrations are as high in the eastern U.S. Both are a product of incomplete combustion of fuels,
including those used in internal combustion processes (gasoline and diesel emissions) and open
biomass burning (smoke from wild and prescribed fire). OC PM species are also produced by
atmospheric transformation of precursor gaseous emissions. Smoke plume impacts from large
wildfires dominate many of the worst haze periods in the western U.S. Carbonaceous PM is
generally the largest component of urban excess PM2.5 (i.e., the difference between urban and
regional background concentration). Western urban areas have more than twice the average
concentrations of carbonaceous PM than remote areas sites in the same region. In eastern urban
areas, PM25 is dominated by about equal concentrations of carbonaceous and SO42~ components,
though the usually high relative humidity in the East causes the hydrated SO42~ particles to be
responsible for about twice as much of the urban haze as that caused by the carbonaceous PM.
Radiocarbon dating of carbonaceous PM from twelve sites (eight in the West, two of which are
urban) showed that about half of the urban area carbonaceous PM is from contemporary as opposed
to fossil sources, while in remote areas the fraction that is contemporary ranges from 82-100%.
Summer urban excess carbonaceous PM is dominated by fossil carbon for the two western urban
areas (Phoenix, AZ and Seattle, WA), but nearly half of the winter urban excess for these two urban
areas are from contemporary carbon sources (e.g., residential wood combustion). An empirical
relationship between the radiocarbon analysis results and the more widely measured EC and OC data
set was used to estimate the fraction of contemporary carbon at about 150 monitoring locations
nationwide. The highest fraction of contemporary carbon is for the western remote areas sites during
the summer (>90% contemporary) and the least was for eastern urban areas during the summer
(<45% contemporary). Winter tended to have less extreme fractions of contemporary carbon for both
remote and urban areas. Alower bound estimate of 40% of the contemporary and 35% of the fossil
carbon is from secondary conversion of gaseous precursor during the summer at the twelve
radiocarbon monitoring sites, suggesting that primary carbonaceous PM whether from fossil or
contemporary sources represent less than two thirds of the total carbonaceous PM.
PM2.5 crustal material (referred to as fine soil) and coarse mass (i.e., PMi0 minus PM2.5) are
significant contributors to haze for remote areas sites in the arid Southwestern U.S. where they
contribute a quarter to a third of the haze, with coarse mass usually contributing twice that of fine
soil. Coarse mass concentrations are as high in the Central Great Plains as in the Southwestern
deserts though there are no corresponding high concentrations of fine soil as in the Southwest. Also,
the relative contribution to haze by the high coarse mass in the Great Plains is much smaller because
of the generally higher haze values caused by the high concentrations of SO42~ and nitrate PM in that
region.
A comprehensive assessment of the 610 worst haze sample periods over a 3-yr period in the
western U.S. where dust is the major contributor categorized each site/sampler period into four
causal groups: Asian dust, local windblown dust, transported regional windblown dust, and
undetermined dust (i.e., not in one of the three other groups). Most dust days occurred at sites in
Arizona, New Mexico, Colorado, western Texas, and southern California, and these were dominated
by local and regionally transported wind-blown dust. Asian dust caused only a few of the worst dust
days during the 3-year assessment period, though it is an important source of dust for the more
northerly regions of the West (responsible for 10-40% of their worst dust periods) where there is
rarely any windblown dust probably due to the greater ground cover. The frequency of worst dust
events classified as undetermined was greatest for sites in the vicinity of large urban and agricultural
areas such as those in California and Southern Arizona.
Visibility has direct significance to people's enjoyment of daily activities and their overall
sense of wellbeing. For example, psychological research has demonstrated that people are
emotionally affected by poor VAQ such that their overall sense of wellbeing is diminished. Urban
visibility has been examined in two types of studies directly relevant to the NAAQS review process:
December 2009 9-73
-------
urban visibility preference studies and urban visibility valuation studies. Both types of studies are
designed to evaluate individuals' desire for good VAQ where they live, using different metrics.
Urban visibility preference studies examine individuals' preferences by investigating the amount of
visibility degradation considered unacceptable, while economic studies examine the value an
individual places on improving VAQ by eliciting how much the individual would be willing to pay
for different amounts of VAQ improvement.
There are three urban visibility preference studies and two additional pilot studies that have
been conducted to date that provide useful information on individuals' preferences for good VAQ in
the urban setting. The completed studies were conducted in Denver, Colorado (Ely et al, 1991,
156417). two cities in British Columbia, Canada (Pryor, 1996, 056598) and Phoenix, AZ
(BBC Research & Consulting, 2002, 156258). The additional studies were conducted in Washington,
DC (ABT, 2001, 156185: Smith and Howell, 2009, 198803). The range of median preference values
for an acceptable amount of visibility degradation from the 4 urban areas was approximately
19-33 dv. Measured in terms of visual range (VR), these median acceptable values were between
approximately 59 and 20 km.
The economic importance of urban visibility has been examined by a number of studies
designed to quantify the benefits (or willingness to pay) associated with potential improvements in
urban visibility. Urban visibility valuation research prior to 1997 was summarized in Chestnut and
Dennis (1997, 014525). and was also described in the 2004 PM AQCD (U.S. EPA, 2004, 056905)
and the 2005 PM Staff Paper (U.S. EPA, 2005, 090209). Since the mid-1990s, little new information
has become available regarding urban visibility valuation.
Collectively, the evidence is sufficient to conclude that 3 C3USal relationship exists
between PM and visibility impairment.
9.3. Effects on Climate
While most of this ISA is restricted to consideration of the emissions, transport and
transformation, resulting concentrations, and effects from PM in the U.S., because the effects
endpoint here is climate, a larger spatial domain is needed. However, this assessment is not intended
to be comprehensive even as a survey of the enormous range and volume of science related to
climate effects from PM; rather, particular attention has been paid to data relevant to the U.S.
The two principal sources for material in this section are Chapter 2, "Changes in Atmospheric
Constituents and in Radiative Forcing," (Forster et al., 2007, 092936) in the comprehensive Working
Group I report in the Fourth Assessment Report (AR4) from the Intergovernmental Panel on Climate
Change (IPCC), Climate Change 2007: The Physical Science Basis (IPCC, 2007, 092765). hereafter
IPCC AR4; and the U.S. Climate Change Science Program Synthesis and Assessment Product 2.3,
"Atmospheric Aerosol Properties and Climate Impacts," by Chin et al. (2009, 192130). hereafter
CCSP SAP2.3. The EPA is a constituent agency member of the U.S. federated CCSP along with
NOAA and NASA, which led production of CCSP SAP2.3 incorporating significant sections from
EPA data and reports related particularly to U.S. emissions and measurements. Sections from each of
these recent comprehensive reports are included here in their entirety or as emended as noted where
they represent the most thorough summary of the climate effects of aerosols. (In the sections
included from IPCC AR4 and CCSP SAP2.3, 'aerosols' is more frequently used than "PM" and that
word is retained.)
9.3.1. The Climate Effects of Aerosols
Section 9.3.1 comes directly from CCSP SAP2.3 Chapter 1 Section 1.2, with section, table,
and figure numbers changed to be internally consistent with this ISA.
Aerosols exert a variety of impacts on the environment. Aerosols (sometimes referred to
particulate matter or "PM," especially in air quality applications), when concentrated near the
surface, have long been recognized as affecting pulmonary function and other aspects of
human health. Sulfate and nitrate aerosols play a role in acidifying the surface downwind of
gaseous sulfur and odd nitrogen sources. Particles deposited far downwind might fertilize iron-
poor waters in remote oceans, and Saharan dust reaching the Amazon Basin is thought to
contribute nutrients to the rainforest soil.
December 2009 9-74
-------
Aerosols also interact strongly with solar and terrestrial radiation in several ways. Figure
9-52 offers a schematic overview. First, they scatter and absorb sunlight (Charlson and Pilat,
1969, 190025; McCormick and Ludwig, 1967, 190528; Mitchell, 1971, 190546); these are
described as "direct effects" on shortwave (solar) radiation. Second, aerosols act as sites at
which water vapor can accumulate during cloud droplet formation, serving as cloud
condensation nuclei or CCN. Any change in number concentration or hygroscopic properties
of such particles has the potential to modify the physical and radiative properties of clouds,
altering cloud brightness (Twomey, 1977, 190533) and the likelihood and intensity with which
a cloud will precipitate (e.g.,Albrecht, 1989, 045783.; Gunn and Phillips, 1957, 190595; Liou
and Ou, 1989,190407).
Collectively changes in cloud processes due to anthropogenic aerosols are referred to as
aerosol indirect effects. Finally, absorption of solar radiation by particles is thought to
contribute to a reduction in cloudiness, a phenomenon referred to as the semi-direct effect.
This occurs because absorbing aerosol warms the atmosphere, which changes the atmospheric
stability, and reduces surface flux.
The primary direct effect of aerosols is a brightening of the planet when viewed from
space, as much of Earth's surface is dark ocean, and most aerosols scatter more than 90% of
the visible light reaching them. The primary indirect effects of aerosols on clouds include an
increase in cloud brightness, change in precipitation and possibly an increase in lifetime; thus
the overall net impact of aerosols is an enhancement of Earth's reflectance (shortwave albedo).
This reduces the sunlight reaching Earth's surface, producing a net climatic cooling, as well as
a redistribution of the radiant and latent heat energy deposited in the atmosphere. These effects
can alter atmospheric circulation and the water cycle, including precipitation patterns, on a
variety of length and time scales (e.g., Ramanathan et al, 2001, 042681; Zhang et al, 2006,
190933).
o
• •; :•!•: : ^-^r:>
Scattering & Unperturbed
absorption of cloud
radiation
1 Direct effects j
^--^.-^
Increased CDNC
(constant LWC)
(Twomey, 1974)
I Cloud albedo effect/ 1
I 1st indirect effect/ I
\ Twomey effect j
~^. A^_^» ' ^ \ /
Surface
^•-- •*?!>' Indirect effect_ ""•' •".;..•
on ice clouds ""."'.
and contrails . . "
Drizzle Increased cloud height Increased cloud Heating causes
suppression. (Pincus & Baker, 1994) lifetime cloud burn-off
Increased LWC (Albrecht, 1989) (Ackerman et al., 2000)
I \Cloud lifetime effect/ 2nd indirect effect/ Albrecht effect J I Semi-direct effect I
Source: IPCC (2007, 092765) modified from Haywood and Boucher (2000,156531).
Figure 9-52. Aerosol radiative forcing. Airborne particles can affect the heat balance of the
atmosphere, directly, by scattering and absorbing sunlight, and indirectly, by
altering cloud brightness and possibly lifetime. Here small black dots represent
aerosols, circles represent cloud droplets, and straight lines represent short-
wave radiation, and wavy lines, long-wave radiation. LWC is liquid water content,
and CDNC is cloud droplet number concentration. Confidence in the magnitudes
of these effects varies considerably (see Chapter 3). Although the overall effect of
aerosols is a net cooling at the surface, the heterogeneity of particle spatial
distribution, emission history, and properties, as well as differences in surface
reflectance, mean that the magnitude and even the sign of aerosol effects vary
immensely with location, season and sometimes inter-annually. The human-
induced component of these effects is sometimes called "climate forcing."
Several variables are used to quantify the impact aerosols have on Earth's energy balance;
these are helpful in describing current understanding, and in assessing possible future steps.
For the purposes of this report, aerosol radiative forcing (KF) is defined as the net energy
flux (downwelling minus upwelling) difference between an initial and a perturbed aerosol
loading state, at a specified level in the atmosphere. (Other quantities, such as solar radiation,
December 2009
9-75
-------
are assumed to be the same for both states.) This difference is defined such that a negative
aerosol forcing implies that the change in aerosols relative to the initial state exerts a cooling
influence, whereas a positive forcing would mean the change in aerosols exerts a warming
influence.
There are a number of subtleties associated with this definition:
(1) The initial state against which aerosol forcing is assessed must be specified. For direct
aerosol radiative forcing, it is sometimes taken as the complete absence of aerosols. IPCC AR4
(2001, 156587) uses as the initial state their estimate of aerosol loading in 1750. That year is
taken as the approximate beginning of the era when humans exerted accelerated influence on
the environment.
(2) A distinction must be made between aerosol RF and the anthropogenic contribution to
aerosol RF. Much effort has been made to distinguishing these contributions by modeling and
with the help of space-based, airborne, and surface-based remote sensing, as well as in situ
measurements. These efforts are described in subsequent chapters (of the CCSP SAP2.3).
(3) In general, aerosol RF and anthropogenic aerosol RF include energy associated with
both the shortwave (solar) and the long-wave (primarily planetary thermal infrared)
components of Earth's radiation budget. However, the solar component typically dominates, so
in this document, these terms are used to refer to the solar component only, unless specified
otherwise. The wavelength separation between the short- and long-wave components is usually
set at around three or four micrometers.
(4) The IPCC AR4 (2007, 092765) defines radiative forcing as the net downward minus
upward irradiance at the tropopause due to an external driver of climate change. This
definition excludes stratospheric contributions to the overall forcing. Under typical conditions,
most aerosols are located within the troposphere, so aerosol forcing at TOA and at the
tropopause are expected to be very similar. Major volcanic eruptions or conflagrations can
alter this picture regionally, and even globally.
(5) Aerosol radiative forcing can be evaluated at the surface, within the atmosphere, or at
top-of-atmosphere (TOA). In this document, unless specified otherwise, aerosol radiative
forcing is assessed at TOA.
(6) As discussed subsequently, aerosol radiative forcing can be greater at the surface than
at TOA if the aerosols absorb solar radiation. TOA forcing affects the radiation budget of the
planet. Differences between TOA forcing and surface forcing represent heating within the
atmosphere that can affect vertical stability, circulation on many scales, cloud formation, and
precipitation, all of which are climate effects of aerosols. In this document, unless specified
otherwise, these additional climate effects are not included in aerosol radiative forcing.)
(7) Aerosol direct radiative forcing can be evaluated under cloud-free conditions or under
natural conditions, sometimes termed "all-sky" conditions, which include clouds. Cloud-free
direct aerosol forcing is more easily and more accurately calculated; it is generally greater than
all-sky forcing because clouds can mask the aerosol contribution to the scattered light. Indirect
forcing, of course, must be evaluated for cloudy or all-sky conditions. In this document, unless
specified otherwise, aerosol radiative forcing is assessed for all-sky conditions.
(8) Aerosol radiative forcing can be evaluated instantaneously, daily (24 h) averaged, or
assessed over some other time period. Many measurements, such as those from polar-orbiting
satellites, provide instantaneous values, whereas models usually consider aerosol RF as a daily
average quantity. In this document, unless specified otherwise, daily averaged aerosol radiative
forcing is reported.
(9) Another subtlety is the distinction between a "forcing" and a "feedback." As different
parts of the climate system interact, it is often unclear which elements are "causes" of climate
change (forcings among them), which are responses to these causes, and which might be some
of each. So, for example, the concept of aerosol effects on clouds is complicated by the impact
clouds have on aerosols; the aggregate is often called aerosol-cloud interactions. This
distinction sometimes matters, as it is more natural to attribute responsibility for causes than
for responses. However, practical environmental considerations usually depend on the net
result of all influences. In this report, "feedbacks" are taken as the consequences of changes in
surface or atmospheric temperature, with the understanding that for some applications, the
accounting may be done differently.
December 2009 9-76
-------
RF Terms
RF values (W m"2)
Spatial scale
LOSU
Long-lived
greenhouse gases"
Ozone
Stratospheric water
vapour from CH4
Surface albedo
{Direct effect
Cloud albedo
effect
Linear contrails
1.66 [1.49 to 1.83]
0.48 [0.43 to 0.53]
0.16 [0.14 to 0.18]
0.34 [0.31 to 0.37]
-0.05 [-0.15 to 0.05]
0.35 [0.25 to 0.65]
0.07 [0.02 to 0.12]
-0.2 [-0.4 to 0.0]
0.1 [0.0 to 0.2]
-0.5 [-0.9 to-0.1]
-0.7 [-1.8 to -0.3]
0.01 [0.003 to 0.03]
Global
Global
Continental
to global
Global
Local to
continental
Continental
to global
Continental
to global
Continental
High
High
Med
Low
Med
- Low
Med
- Low
Low
Low
Solar irradiance
0.12 [0.06 to 0.30]
Global
Low
Total net
anthropogenic
1.6 [0.6 to 2.4]
-2-101
Radiative Forcing (W nrr2)
Source: IPCC (2007, 092765).
Figure 9-53.
Global average radiative forcing (RF) estimates and uncertainty ranges in 2005,
relative to the pre-industrial climate. Anthropogenic C02, methane (ChU), nitrous
oxide (N20), ozone, and aerosols as well as the natural solar irradiance variations
are included. Typical geographical extent of the forcing (spatial scale) and the
assessed level of scientific understanding (LOSU) are also given. Forcing is
expressed in units of watts per square meter (W/m2). The total anthropogenic
radiative forcing and its associated uncertainty are also given.
December 2009
9-77
-------
.a
CO
.a
o
.i 1
CD
cc
Total aerosol
•adiative forcing -
cooling
warming
Long-lived greenhouse ,
gases and ozone i
radiative forcings i
1 ' •
Total anthropogenic radiative forcing i
i
i
i
ii
n
n
i
i
Ok
-3
-1 0 1
Radiative Forcing (W nr2)
Source: Adapted from IPCC (2007, 0927651.
Figure 9-54. Probability distribution functions (PDFs) for anthropogenic aerosol and GHG
RFs. Dashed red curve: RF of long-lived greenhouse gases plus ozone; dashed
blue curve: RF of aerosols (direct and cloud albedo RF); red filled curve:
combined anthropogenic RF. The RF range is at the 90% confidence interval.
In summary, aerosol radiative forcing, the fundamental quantity about which this report is
written, must be qualified by specifying the initial and perturbed aerosol states for which the
radiative flux difference is calculated, the altitude at which the quantity is assessed, the
wavelength regime considered, the temporal averaging, the cloud conditions, and whether total
or only human-induced contributions are considered. The definition given here, qualified as
needed, is used throughout the report.
Although the possibility that aerosols affect climate was recognized more than 40 years
ago, the measurements needed to establish the magnitude of such effects, or even whether
specific aerosol types warm or cool the surface, were lacking. Satellite instruments capable of
at least crudely monitoring aerosol amount globally were first deployed in the late 1970s. But
scientific focus on this subject grew substantially in the 1990s (e.g., Charlson and Wigley,
1994, 189989: Charlson et al, 1991, 045793: 1992, 045794: Penner et al, 1992, 045825), in
part because it was recognized that reproducing the observed temperature trends over the
industrial period with climate models requires including net global cooling by aerosols in the
calculation (IPCC, 1995, 190991: 1996,190990), along with the warming influence of
enhanced atmospheric greenhouse gas (GHG) concentrations - mainly carbon dioxide,
methane, nitrous oxide, chlorofluorocarbons, and ozone.
Improved satellite instruments, ground- and ship-based surface monitoring, more
sophisticated chemical transport and climate models, and field campaigns that brought all
these elements together with aircraft remote sensing and in situ sampling for focused,
coordinated study, began to fill in some of the knowledge gaps. By the Fourth IPCC
Assessment Report, the scientific community consensus held that in global average, the sum of
direct and indirect top-of-atmosphere (TOA) forcing by anthropogenic aerosols is negative
(cooling) of about -1.3 W/m2 (-2.2 to -0.5 W/m2). This is significant compared to the positive
forcing by anthropogenic GHGs (including ozone), about 2.9 ± 0.3 W/rri (IPCC, 2007,
092765). However, the spatial distribution of the gases and aerosols are very different, and
they do not simply exert compensating influences on climate.
The IPCC aerosol forcing assessments are based largely on model calculations,
constrained as much as possible by observations. At present, aerosol influences are not yet
quantified adequately, according to Figure 9-53, as scientific understanding is designated as
"Medium-Low" and "Low" for the direct and indirect climate forcing, respectively. The IPCC
AR4 (2007, 092765) concluded that uncertainties associated with changes in Earth's radiation
budget due to anthropogenic aerosols make the largest contribution to the overall uncertainty
in radiative forcing of climate change among the factors assessed over the industrial period
(Figure 9-54).
December 2009
9-78
-------
-20
South Smiin A-CO.I A-n.z A -0.3- Harm Łur0p« e«al
Artivrica Africa 0.35 Amorica
BkHnii&s burning Mineral dust Pollution
Source: Adapted with Permission of theAmerican Geophysical Union from from Zhou et al. (2005, 1561831.
Figure 9-55. The clear-sky forcing efficiency ET, defined as the diurnally averaged aerosol
direct radiative effect (W/m2) per unit AOD at 550 nm, calculated at both TOA and
the surface, for typical aerosol types over different geographical regions. The
vertical black lines represent ± one standard deviation of ET for individual aerosol
regimes and A is surface broadband albedo.
Although AOD, aerosol properties, aerosol vertical distribution, and surface reflectivity
all contribute to aerosol radiative forcing, AOD usually varies on regional scales more than the
other aerosol quantities involved. Forcing efficiency (Ex), defined as a ratio of direct aerosol
radiative forcing to AOD at 550 nm, reports the sensitivity of aerosol radiative forcing to
AOD, and is useful for isolating the influences of particle properties and other factors from
that of AOD. ET is expected to exhibit a range of values globally, because it is governed
mainly by aerosol size distribution and chemical composition (which determine aerosol single-
scattering albedo and phase function), surface reflectivity, and solar irradiance, each of which
exhibits pronounced spatial and temporal variations. To assess aerosol RF, ET is multiplied by
the ambient AOD.
Figure 9-55 shows a range of ET, derived from AERONET surface sun photometer
network measurements of aerosol loading and particle properties, representing different
aerosol and surface types, and geographic locations. It demonstrates how aerosol direct solar
radiative forcing (with initial state taken as the absence of aerosol) is determined by a
combination of aerosol and surface properties. For example, ET due to southern African
biomass burning smoke is greater at the surface and smaller at TOA than South American
smoke because the southern African smoke absorbs sunlight more strongly, and the magnitude
of ET for mineral dust for several locations varies depending on the underlying surface
reflectance. Figure 9-55 illustrates one further point, that the radiative forcing by aerosols on
surface energy balance can be much greater than that at TOA. This is especially true when the
particles have SSA substantially less than 1, which can create differences between surface and
TOA forcing as large as a factor of five (e.g., Zhou et al., 2005, 156183).
December 2009
9-79
-------
Table 9-3. Top-of-atmosphere, cloud-free, instantaneous direct aerosol radiative forcing
dependence on aerosol and surface properties. Here TWP, SGP, and NSA are the
Tropical West Pacific island, Southern Great Plains, and North Slope Alaska
observation stations maintained by the DOE ARM program, respectively.
Instantaneous values are given at specific solar zenith angle. Upper and middle
parts are from McComiskey et al. (2008,190523). Representative, parameter-
specific measurement uncertainty upper bounds for producing 1 W/m2
instantaneous TOA forcing accuracy are given in the lower part, based on
sensitivities at three sites from the middle part of the table.
Parameters TWP SGP NSA
AEROSOL PROPERTIES (AOD, SSA, G), SOLAR ZENITH ANGLE (SZA), SURFACE ALBEDO (A), AND AEROSOL DIRECT RF AT
TOA (F)
AOD 0.05 0.1 0.05
SSA 0.97 0.95 0.95
g 0.8 0.6 0.7
A 0.05 0.1 0.9
SZA 30 45 70
F(W/m2) -2.2 -6.3 2.6
SENSITIVITY OF CLOUD-FREE, INSTANTANEOUS, TOA DIRECT AEROSOL RADIATIVE FORCING TO AEROSOL AND SURFACE
PROPERTIES, W/fif PER UNIT CHANGE IN PROPERTY
dFld(AOD) -45 -64 51
dFld(SSA) -11 -50 -60
dFldg 13 23 2
dFldA 8 24 6
REPRESENTATIVE MEASUREMENT UNCERTAINTY UPPER BOUNDS FOR PRODUCING 1 W/M2 ACCURACY OF AEROSOL RF
AOD 0.022 0.016 0.020
SSA 0.091 0.020 0.017
g 0.077 0.043
A 0.125 0.042 0.167
Table 9-3 presents estimates of cloud-free, instantaneous, aerosol direct RF dependence
on AOD, and on aerosol and surface properties, calculated for three sites maintained by the
U.S. Department of Energy's Atmospheric Radiation Measurement (ARM) program, where
surface and atmospheric conditions span a significant range of natural environments
(McComiskey et al., 2008, 190523). Here aerosol RF is evaluated relative to an initial state
that is the complete absence of aerosols. Note that aerosol direct RF dependence on individual
parameters varies considerably, depending on the values of the other parameters, and in
particular, that aerosol RF dependence on AOD actually changes sign, from net cooling to net
warming, when aerosols reside over an exceedingly bright surface. Sensitivity values are given
for snapshots at fixed solar zenith angles, relevant to measurements made, for example, by
polar-orbiting satellites.
The lower portion of Table 9-3 presents upper bounds on instantaneous measurement
uncertainty, assessed individually for each of AOD, SSA, g, and A, to produce a 1 W/m2 top-of
atmosphere, cloud-free aerosol RF accuracy. The values are derived from the upper portion of
the table, and reflect the diversity of conditions captured by the three ARM sties. Aerosol RF
sensitivity of 1 W/m2 is used as an example; uncertainty upper bounds are obtained from the
partial derivative for each parameter by neglecting the uncertainties for all other parameters.
These estimates produce an instantaneous AOD measurement uncertainty upper bound
between about 0.01 and 0.02, and SSA constrained to about 0.02 over surfaces as bright as or
brighter than the ARM Southern Great Plains site, typical of mid-latitude, vegetated land.
Other researchers, using independent data sets, have derived ranges of ET and aerosol RF
December 2009 9-80
-------
sensitivity similar to those presented here, for a variety of conditions (Christopher and Jones,
2008, 189985: Yu et al, 2006, 156173: Zhou et al, 2005, 156183).
These uncertainty bounds provide a baseline against which current and expected near-
future instantaneous measurement capabilities are assessed in Chapter 2 (of the CCSP
SAP2.3). Model sensitivity is usually evaluated for larger-scale (even global) and longer-term
averages. When instantaneous measured values from a randomly sampled population are
averaged, the uncertainty component associated with random error diminishes as something
like the inverse square root of the number of samples. As a result, the accuracy limits used for
assessing more broadly averaged model results corresponding to those used for assessing
instantaneous measurements would have to be tighter, as discussed in Chapter 4 (of the CCSP
SAP2.3).
In summary, much of the challenge in quantifying aerosol influences arises from large
spatial and temporal heterogeneity, caused by the wide variety of aerosol sources, sizes and
compositions, the spatial non-uniformity and intermittency of these sources, the short
atmospheric lifetime of most aerosols, and the spatially and temporally non-uniform chemical
and microphysical processing that occurs in the atmosphere. In regions having high
concentrations of anthropogenic aerosol, for example, aerosol forcing is much stronger than
the global average, and can exceed the magnitude of GHG warming, locally reversing the sign
of the net forcing. It is also important to recognize that the global-scale aerosol TOA forcing
alone is not an adequate metric for climate change (NRC, 2005, 057409). Due to aerosol
absorption, mainly by soot, smoke, and some desert dust particles, the aerosol direct radiative
forcing at the surface can be much greater than the TOA forcing, and in addition, the radiative
heating of the atmosphere by absorbing particles can change the atmospheric temperature
structure, evolution, and possibly large-scale dynamical systems such as the monsoons (Kim et
al., 2006, 190917: Lau et al., 2009, 190229). By realizing aerosol's climate significance and
the challenge of charactering highly variable aerosol amount and properties, the U.S. Climate
Change Research Initiative (CCPJ) identified research on atmospheric concentrations and
effects of aerosols specifically as a top priority (NRC, 2001, 053303.).
9.3.2. Overview of Aerosol Measurement Capabilities
9.3.2.1. Satellite Remote Sensing
Section 9.3.2 with the exception of the final paragraph, comes directly from CCSP SAP2.3
Chapter 2, Section 2.2 with section, table, and figure numbers changed to be internally consistent
with this ISA.
A measurement-based characterization of aerosols on a global scale can be realized only
through satellite remote sensing, which is the only means of characterizing the large spatial
and temporal heterogeneities of aerosol distributions. Monitoring aerosols from space has been
performed for over two decades and is planned for the coming decade with enhanced
capabilities (Forster et al., 2007, 092936: King et al., 1999, 190635: Lee et al., 2006, 190358:
Mischenko et al., 2007, 190543). Table 9-4 summarizes major satellite measurements currently
available for the tropospheric aerosol characterization and radiative forcing research.
Early aerosol monitoring from space relied on sensors that were designed for other
purposes. The Advanced Very High Resolution Radiometer (AVHRR), intended as a cloud and
surface monitoring instrument, provides radiance observations in the visible and near infrared
wavelengths that are sensitive to aerosol properties over the ocean (Husar et al., 1997, 045900:
Mishchenko et al., 1999, 190541). Originally intended for ozone monitoring, the ultraviolet
(UV) channels used for the Total Ozone Mapping Spectrometer (TOMS) are sensitive to
aerosol UV absorption with little surface interferences, even over land (Torres et al., 1998,
190503). This UV-technique makes TOMS suitable for monitoring biomass burning smoke
and dust, though with limited sensitivity near the surface (Herman et al., 1997, 048393) and
for retrieving aerosol single-scattering albedo from space (Torres et al., 2005, 190507). (Anew
sensor, the Ozone Monitoring Instrument (OMI) aboard Aura, has improved on such UV-
technique advantages, providing higher spatial resolution and more spectral channels; see
(Veihelmann et al., 2007, 190627). Such historical sensors have provided multi-decadal
climatology of aerosol optical depth that has significantly advanced the understanding of
aerosol distributions and long-term variability (e.g.,Geogdzhayev et al., 2002, 190574: Massie
et al., 2004, 190492: Mishchenko and Geogdzhayev, 2007, 190545: Mishchenko et al., 2007,
190542: Torres et al., 2002, 190505: Zhao et al., 2008, 190935).
Over the past decade, satellite aerosol retrievals have become increasingly sophisticated.
Now, satellites measure the angular dependence of radiance and polarization at multiple
wavelengths from UV through the infrared (IR) at fine spatial resolution. From these
observations, retrieved aerosol products include not only optical depth at one wavelength, but
also spectral optical depth and some information about particle size over both ocean and land,
as well as more direct measurements of polarization and phase function. In addition, cloud
screening is much more robust than before and onboard calibration is now widely available.
December 2009 9-81
-------
Examples of such new and enhanced sensors include the MODerate resolution Imaging
Spectroradiometer (MODIS, see Box 2.1 of CCSP SAP2.3), the Multi-angle Imaging
SpectroRadiometer (MISR, see Box 2.2 of CCSP SAP2.3), Polarization and Directionality of
the Earth's Reflectance (POLDER, see Box 2.3 of CCSP SAP2.3), and OMI, among others.
The accuracy for AOD measurement from these sensors is about 0.05 or 20% of AOD (Kahn
et al, 2005, 190966: Remer et al, 2005, 190221) and somewhat better over dark water, but
that for aerosol microphysical properties, which is useful for distinguishing aerosol air mass
types, is generally low. The Clouds and the Earth's Radiant Energy System (CERES, see Box
2.4 of CCSP SAP2.3) measures broadband solar and terrestrial radiances. The CERES
radiation measurements in combination with satellite retrievals of aerosol optical depth can be
used to determine aerosol direct radiative forcing.
Table 9-4. Summary of major satellite measurements currently available for the tropospheric
aerosol characterization and radiative forcing research.
Category Properties Sensor/platform
AVHRR/NOAA-series
TOMS/Nimbus,ADEOSI, EP
POLDER-1, 2, PARASOL
Loading MODISATerra.Aqua
MISR/Terra
OMI/Aura
AVHRR/NOAA-series
Column-integrated POLDER-1 ,2, PARASOL
Size, shape MODIS/Terra.Aqua
Parameters
Optical depth
Angstrom exponent
Fine-mode fraction,
Angstrom exponent,
non-spherical fraction
Fine-mode fraction
Angstrom exponent
Effective radium
Asymmetry factor
Spatial coverage
~daily coverage of
global ocean
-daily coverage of
global land and ocean
-weekly coverage of
global land and ocean,
including bright desert
and nadir sun-glint
-daily coverage of
global land and ocean
Global ocean
Global land and ocean
Global land and ocean
ocean)
Temporal coverage
1981 -present
1979-2001
1997-present
2000-present (Terra)
2002-present (Aqua)
2000-present
2005-present
1981 -present
1997-present
2000-present (Terra)
2002-present (Aqua)
MISR/Terra
Angstrom exponent,
small, medium large
fractions, non-spherical
fractions
Global land and ocean 2000-present
TOMS/Nimbus,ADEOSI, EP
Absorption OMI/Aura
MISR/Terra
GLAS/ICESat
Vertical-resolved shape"9' °IZ°' ^
CALIOP/CALIPSO
Absorbing aerosol index,
absorbing optical depth
Single-scattering albedo
(2-4 bins)
Extinction/backscatter
Extinction/backscatter,
color ratio,
depolarization ratio
Global land and ocean
Global land and ocean,
• 16-day repeating cycle,
single-nadir
measurement
1979-2001
2005-present
2000-present
2003-present
(-3 mo/yr)
2006-present
December 2009
9-82
-------
Complementary to these passive sensors, active remote sensing from space is also now
possible and ongoing (see Box 2.5 of CCSP SAP2.3). Both the Geoscience Laser Altimeter
System (GLAS) and the Cloud and Aerosol Lidar with Orthogonal Polarization (CALIOP) are
collecting essential information about aerosol vertical distributions. Furthermore, the
constellation of six afternoon-overpass spacecrafts (as illustrated in Figure 9-60), the so-called
A-Train (Stephens et al., 2002, 190412) makes it possible for the first time to conduct near
simultaneous (within 15 minutes) measurements of aerosols, clouds, and radiative fluxes in
multiple dimensions with sensors in complementary capabilities.
The improved accuracy of aerosol products (mainly AOD) from these new-generation
sensors, together with improvements in characterizing the earth's surface and clouds, can help
reduce the uncertainties associated with estimating the aerosol direct radiative forcing (Yu et
al., 2006, 156173): and references therein). The retrieved aerosol microphysical properties,
such as size, absorption, and non-spherical fraction can help distinguish anthropogenic
aerosols from natural aerosols and hence help assess the anthropogenic component of aerosol
direct radiative forcing (Bellouin et al., 2005, 155684; Christopher et al., 2006, 155729;
Kaufman et al., 2005, 155891: Yu et al., 2006, 156173). However, to infer aerosol number
concentrations and examine indirect aerosol radiative effects from space, significant efforts are
needed to measure aerosol size distribution with much improved accuracy, characterize aerosol
type, account for impacts of water uptake on aerosol optical depth, and determine the fraction
of aerosols that is at the level of the clouds (Kapustin et al., 2006, 190961: Rosenfeld, 2006,
190233). In addition, satellite remote sensing is not sensitive to particles much smaller than
0.1 micrometer in diameter, which comprise of a significant fraction of those that serve as
cloud condensation nuclei.
Finally, algorithms are being developed to retrieve aerosol absorption or SSA from
satellite observations (e.g., Kaufman et al., 2002, 190955: Torres et al., 2005, 190507). The
NASA Glory mission, scheduled to launch in 2009 and to be added to the A-Train, will deploy
a multi-angle, multispectral polarimeter to determine the global distribution of aerosol and
clouds. It will also be able to infer microphysical property information, from which aerosol
type (e.g., marine, dust, pollution, etc.) can be inferred for improving quantification of the
aerosol direct and indirect forcing on climate (Mischenko et al., 2007, 190543).
In summary, major advances have been made in both passive and active aerosol remote
sensing from space in the past decade, providing better coverage, spatial resolution, retrieved
AOD accuracy, and particle property information. However, AOD accuracy is still much
poorer than that from surface-based sun photometers (0.01-0.02), even over vegetated land and
dark water where retrievals are most reliable. Although there is some hope of approaching this
level of uncertainty with a new generation of satellite instruments, the satellite retrievals entail
additional sensitivities to aerosol and surface scattering properties. It seems unlikely that
satellite remote sensing could exceed the sun photometer accuracy without introducing some
as-yet-unspecified new technology. Spacebased lidars are for the first time providing global
constraints on aerosol vertical distribution, and multi-angle imaging is supplementing this with
maps of plume injection height in aerosol source regions. Major advances have also been made
during the past decade in distinguishing aerosol types from space, and the data are now useful
for validating aerosol transport model simulations of aerosol air mass type distributions and
transports, particularly over dark water. But particle size, shape, and especially SSA
information has large uncertainty; improvements will be needed to better distinguish
anthropogenic from natural aerosols using space-based retrievals. The particle microphysical
property detail required to assess aerosol radiative forcing will come largely from targeted in
situ and surface remote sensing measurements, at least for the near-future, although estimates
of measurement-based aerosol RF can be made from judicious use of the satellite data with
relaxed requirements for characterizing aerosol microphysical properties.
MODerate resolution Imaging Spectroradiometer
MODIS performs near global daily observations of atmospheric aerosols. Seven of 36
channels (between 0.47 and 2.13 um) are used to retrieve aerosol properties over cloud and
surface-screened areas (Li et al., 2004, 190386: Martins et al., 2002, 190470). Over vegetated
land, MODIS retrieves aerosol optical depth at three visible channels with high accuracy of ±
0.05 ± 0.2x (Chu et al., 2002, 190001: Kaufman and Fraser, 1997, 190958: Levy et al., 2007,
190379: Remer et al., 2005, 190221). Most recently a deep blue algorithm (Hsu et al., 2004,
190622) has been implemented to retrieve aerosols over bright deserts on an operational basis,
with an estimated accuracy of 20-30%. Because of the greater simplicity of the ocean surface,
MODIS has the unique capability of retrieving not only aerosol optical depth with greater
accuracy, i.e., ± 0.03 ± O.OSx (Remer et al., 2002, 190218: 2005, 190221: Remer et al., 2008,
190224: Tanre et al., 1997, 190452), but also quantitative aerosol size parameters (e.g.,
effective radius, fine-mode fraction of AOD) (Kaufman et al., 2002, 190956: Kleidman et al.,
2005, 190175: Remer et al., 2005, 190221). The fine-mode fraction has been used as a tool for
separating anthropogenic aerosol from natural ones and estimating the anthropogenic aerosol
direct climate forcing (Kaufman et al., 2005, 155891). Figure 9-56 shows composites of
MODIS AOD and fine-mode fraction that illustrate seasonal and geographical variations of
aerosol types. Clearly seen from the figure is heavy pollution over East Asia in both months,
biomass burning smoke over South Africa, South America, and Southeast Asia in August,
December 2009 9-83
-------
heavy dust storms over North Atlantic in both months and over Arabian Sea in August, and a
mixture of dust and pollution plume swept across North Pacific in April.
Multi-Angle Imaging SpectroRadiometer
MISR, aboard the sun-synchronous, polar orbiting satellite Terra, measures upwelling
solar radiance in four visible-near-IR spectral bands and at nine view angles spread out in the
forward and aft directions along the flight path (Diner et al, 2002, 189967). It acquires global
coverage about once per week. A wide range of along-track view angles makes it feasible to
more accurately evaluate the surface contribution to the TOA radiances and hence retrieve
aerosols over both ocean and land surfaces, including bright desert and sunglint regions (Diner
et al., 1998, 189962; Kahn et al., 2005, 190966; Martonchik et al., 1998, 190472; 2002,
190490). MISR AODs are within 20% or ± 0.05 of coincident AERONET measurements
(Abdou et al., 2005, 190028; Kahn et al., 2005, 189961). The MISR multi-angle data also
sample scattering angles ranging from about 60° to 160° in midlatitudes, yielding information
about particle size (Chen et al., 2008, 189984; Kahn et al., 1998, 190970; 2001, 190969; 2005,
190966) and shape (Kalashnikova and Kahn, 2006, 190962). The aggregate of aerosol
microphysical properties can be used to assess aerosol airmass type, a more robust
characterization of MISR-retrieved particle property information than individual attributes.
MISR also retrieves plume height in the vicinity of wildfire, volcano, and mineral dust aerosol
sources, where the plumes have discernable spatial contrast in the multi-angle imagery (Kahn
et al., 2007, 190964). Figure 9-57 is an example that illustrates MISR's ability to characterize
the load, optical properties, and stereo height of near-source fire plumes.
POLarization and Directionality of the Earth's Reflectance
POLDER is a unique aerosol sensor that consists of wide field-of-view imaging spectro-
radiometer capable of measuring multi-spectral, multi-directional, and polarized radiances
(Deuze et al., 2001, 192013). The observed radiances can be exploited to better separate the
atmospheric contribution from the surface contribution over both land and ocean. POLDER -1
and -2 flew onboard the ADEOS (Advanced Earth Observing Satellite) from November 1996
to June 1997 and April to October of 2003, respectively. A similar POLDER instrument flies
on the PARASOL satellite that was launched in December 2004.
Figure 9-58 shows global horizontal patterns of AOD and Angstrom exponent over the
oceans derived from the POLDER instrument for June 1997. The oceanic AOD map (Figure
9-58a) reveals near-coastal plumes of high AOD, which decrease with distance from the coast.
This pattern arises from aerosol emissions from the continents, followed by atmospheric
dispersion, transformation, and removal in the downwind direction. In large-scale flow fields,
such as the trade winds, these continental plumes persist over several thousand kilometers. The
Angstrom exponent shown in Figure 9-58 exhibits a very different pattern from that of the
aerosol optical depth; specifically, it exhibits high values downwind of industrialized regions
and regions of biomass burning, indicative of small particles arising from direct emissions
from combustion sources and/or gas-to-particle conversion, and low values associated with
large particles in plumes of soil dust from deserts and in sea salt aerosols.
December 2009 9-84
-------
Source: Adapted from (Chin et al., 2007,1900621: original figure from Yoram Kaufman and Reto Stockli.
Figure 9-56. A composite of MODIS/Terra observed aerosol optical depth (at 550 nm, green
light near the peak of human vision) and fine-mode fraction that shows spatial
and seasonal variations of aerosol types. Industrial pollution and biomass
burning aerosols are predominately small particles (shown as red), whereas
mineral dust and sea salt consist primarily of large particles (shown as green).
Bright red and bright green indicate heavy pollution and dust plumes,
respectively.
December 2009
9-85
-------
Source: Reprinted with Permission of theAmerican Geophysical Union from Kahn et al. (2007,190964).
Figure 9-57. Oregon fire on September 4,2003, as observed by MISR: (a) MISR nadir view of
the fire plume, with five patch locations numbered and wind-vectors superposed
in yellow; (b) MISR aerosol optical depth at 558 nm; and (c) MISR stereo height
without wind correction for the same region.
9.3.2.2. Focused Field Campaigns
Over the past two decades, numerous focused field campaigns have examined the
physical, chemical, and optical properties and radiative forcing of aerosols in a variety of
aerosol regimes around the world, as listed in Table 9-5. These campaigns, which have been
designed with aerosol characterization as the main goal or as one of the major themes in more
interdisciplinary studies, were conducted mainly over or downwind of known continental
aerosol source regions, but in some instances in low-aerosol regimes, for contrast. During each
of these comprehensive campaigns, aerosols were studied in great detail, using combinations
of in situ and remote sensing observations of physical and chemical properties from various
platforms (e.g., aircraft, ships, satellites, and ground-based stations) and numerical modeling.
In spite of their relatively short duration, these field studies have acquired comprehensive data
sets of regional aerosol properties that have been used to understand the properties and
evolution of aerosols within the atmosphere and to improve the climatology of aerosol
microphysical properties used in satellite retrieval algorithms and CTMs.
December 2009
9-86
-------
Optical Depth
\ - 865 nm
Angstrom Exponent
•x = -d In rid In K
-0.2
Source: Reproduced with permission of Laboratoire d'Optique Atmospherique (LOA), Lille, FR; Laboratoire des Sciences du Climat et de I'Environnement
(LSCE), Gif sur Yvette, FR; Centre National d'etudes Spatiales (ONES), Toulouse, FR; and NAtional Space Development Agency (NASDA), Japan.
Figure 9-58. Global maps at 18 km resolution showing monthly average (a) AOD at 865 nm
and (b) Angstrom exponent of AOD over water surfaces only for June, 1997,
derived from radiance measurements by the POLDER.
9.3.2.3. Ground-Based In Situ Measurement Networks
Major U.S.-operated surface in situ and remote sensing networks for tropospheric aerosol
characterization and climate forcing research are listed in Table 9-6. These surface in situ
stations provide information about long-term changes and trends in aerosol concentrations and
properties, the influence of regional sources on aerosol properties, climatologies of aerosol
radiative properties, and data for testing models and satellite aerosol retrievals. The NOAA
Earth System Research Laboratory (ESRL) aerosol monitoring network consists of baseline,
regional, and mobile stations. These near-surface measurements include submicrometer and
sub-10 micrometer scattering and absorption coefficients from which the extinction coefficient
and single-scattering albedo can be derived. Additional measurements include particle
concentration and, at selected sites, CCN concentration, the hygroscopic growth factor, and
chemical composition.
Several of the stations, which are located across North America and world-wide, are in
regions where recent focused field campaigns have been conducted. The measurement
protocols at the stations are similar to those used during the field campaigns. Hence, the station
data are directly comparable to the field campaign data so that they provide a longer-term
measure of mean aerosol properties and their variability, as well as a context for the shorter-
duration measurements of the field campaigns.
The Interagency Monitoring of Protected Visual Environment (IMPROVE), which is
operated by the NP Service Air Resources Division, has stations across the U.S. located within
NPs (Malm et al., 1994, 044920). Although the primary focus of the network is air pollution,
the measurements are also relevant to climate forcing research. Measurements include fine and
coarse mode (PM2.5 and PMi0) aerosol mass concentration; concentrations of elements, sulfate,
nitrate, organic carbon, and elemental carbon; and scattering coefficients.
December 2009
9-87
-------
In addition, to these U.S.-operated networks, there are other national and international
surface networks that provide measurements of aerosol properties including, but not limited to,
the World Meteorological Organization (WMO) Global Atmospheric Watch (GAW) network
('http://www.wmo.int/pages/prog/arep/gaw/monitoring.html'), the European Monitoring and
Evaluation Programme (EMEP) (http ://wwwemep.int/), the Canadian Air and Precipitation
Monitoring Network (CAPMoN) (http://www.msc-smc.ec.gc.ca/capmon/index_e. cfin). and the
Acid Deposition Monitoring Network in East Asia (EANET) (http://www.eanet.cc/eanet.html).
Clouds and the Earth's Radiant Energy System
CERES measures broadband solar and terrestrial radiances at three channels with a large
footprint (e.g., 20 km for CERES/Terra) (Wielicki et al, 1996, 190637). It is co-located with
MODIS and MISR aboard Terra and with MODIS on Aqua. The observed radiances are
converted to TOA irradiances or fluxes using the Angular Distribution Models (ADMs) that
are functions of viewing angle, sun angle, and scene type (Loeb and Kato, 2002, 190432:
Loeb et al., 2005, 190436; Zhang et al., 2005, 190929). Such estimates of TOA solar flux in
clear-sky conditions can be compared to the expected flux for an aerosol-free atmosphere, in
conjunction with measurements of aerosol optical depth from other sensors (e.g., MODIS and
MISR) to derive the aerosol direct radiative forcing (Christopher et al., 2006, 155729: Loeb
and Manalo-Smith, 2005, 190433: Patadia et al., 2008, 190558: Zhang and Christopher, 2003,
190928: Zhang et al., 2005, 190930). The derived instantaneous value is then scaled to obtain
a daily average. A direct use of the coarse spatial resolution CERES measurements would
exclude aerosol distributions in partly cloudy CERES scenes. Several approaches that
incorporate coincident, high spatial and spectral resolution measurements (e.g., MODIS) have
been employed to overcome this limitation (Loeb and Manalo-Smith, 2005, 190433: Zhang et
al., 2005, 190930).
Active Remote Sensing of Aerosols
Following the success of a demonstration of lidar system aboard the U.S. Space Shuttle
mission in 1994, i.e., Lidar In-space Technology Experiment (LITE) (Winker et al., 1996,
190914), the Geoscience Laser Altimeter System (GLAS) was launched in early 2003 to
become the first polar orbiting satellite lidar. It provides global aerosol and cloud profiling for
a one-month period out of every three-to-six months. It has been demonstrated that GLAS is
capable of detecting and discriminating multiple layer clouds, atmospheric boundary layer
aerosols, and elevated aerosol layers (e.g., Spinhirne et al., 2005, 190410). The Cloud-Aerosol
Lidar and Infrared Pathfinder Satellite Observations (CALIPSO), launched on April 28, 2006,
is carrying a lidar instrument (Cloud and Aerosol Lidar with Orthogonal Polarization -
CALIOP) that has been collecting profiles of the attenuated backscatter at visible and near-
infrared wavelengths along with polarized backscatter in the visible channel (Winker et al.,
2003, 192017). CALIOP measurements have been used to derive the above-cloud fraction of
aerosol extinction optical depth (Chand et al., 2008, 189974), one of the important factors
determining aerosol direct radiative forcing in cloudy conditions. Figure 9-59 shows an event
of trans-Atlantic transport of Saharan dust captured by CALIPSO. Flying in formation with the
Aqua, AURA, POLDER, and CloudSat satellites, the vertically resolved information is
expected to greatly improve passive aerosol and cloud retrievals as well as allow the retrieval
of vertical distributions of aerosol extinction, fine- and coarse-mode separately (Huneeus and
Boucher, 2007, 190624: Kaufman et al., 2003, 190954: Leon et al., 2003, 190366).
December 2009 9-88
-------
Source: Reprinted with Permission of the American Geophysical Union from Liu et al. (2008,156709).
Figure 9-59. A dust event that originated in the Sahara desert on 17 August 2007 and was
transported to the Gulf of Mexico. Red lines represent back trajectories
indicating the transport track of the dust event. Vertical images are 532 nm
attenuated backscatter coefficients measured by CALIOP when passing over the
dust transport track. The letter "D" designates the dust layer, and "S" represents
smoke layers from biomass burning in Africa (17-19 August) and South America
(22 August). The track of the high-spectral-resolution-lidar(HSRL) measurement
is indicated by the white line superimposed on the 28 August CALIPSO image.
The HSRL track is coincident with the track of the 28 August CALIPSO
measurement off the coast of Texas between 28.75°N and 29.08°N.
9.3.2.4. In Situ Aerosol Profiling Programs
In addition to long-term ground based measurements, regular long-term aircraft in situ
measurements recently have been implemented at several locations. These programs provide a
statistically significant data set of the vertical distribution of aerosol properties to determine
spatial and temporal variability through the vertical column and the influence of regional
sources on that variability. In addition, the measurements provide data for satellite and model
validation. As part of its long-term ground measurements, NOAA has conducted regular flights
over Bondville, Illinois since 2006. Measurements include light scattering and absorption
coefficients, the relative humidity dependence of light scattering, aerosol number
concentration and size distribution, and chemical composition. The same measurements with
the exception of number concentration, size distribution, and chemical composition were made
by NOAA during regular overflights of DOE ARM's Southern Great Plains (SGP) site from
2000-2007 (Andrews et al., 2004, 190058)
(http://www.esrl.noaa.gov/gmd/aero/net/index.html).
In summary of Sections 9.3.2.2, 9.3.2.3, and 9.3.2.4, in situ measurements of aerosol
properties have greatly expanded over the past two decades as evidenced by the number of
focused field campaigns in or downwind of aerosol source regions all over the globe, the
continuation of existing and implementation of new sampling networks worldwide, and the
implementation of regular aerosol profiling measurements from fixed locations. In addition, in
situ measurement capabilities have undergone major advancements during this same time
December 2009
9-89
-------
period. These advancements include the ability to measure aerosol chemical composition as a
function of size at a time resolution of seconds to minutes (e.g., Jayne et al., 2000, 190978).
the development of instruments able to measure aerosol absorption and extinction coefficients
at high sensitivity and time resolution and as a function of relative humidity (e.g., Baynard et
al., 2007, 151669; Lack et al., 2006, 096032), and the deployment of these instruments across
the globe on ships, at ground-based sites, and on aircraft. However, further advances are
needed to make this newly developed instrumentation more affordable and turn-key so that it
can be deployed more widely to characterize aerosol properties at a variety of sites worldwide.
Figure 9-60. A constellation of five spacecraft that overfly the Equatorat about 1:30 p.m., the
so-called A-Train, carries sensors having complementary capabilities, offering
unprecedented opportunities to study aerosols from space in multiple
dimensions.
December 2009
9-90
-------
Table 9-5. List of major intensive field experiments that are relevant to aerosol research in a variety
of aerosol regimes around the globe conducted in the past two decades.
Aerosol Regimes
Anthropogenic aerosol
and boreal forest from
North America and West
Europe
Brown haze in South Asia
Anthropogenic aerosol
and desert dust mixture
from East Asia
Biomass burning smoke
in the tropics
Mineral dusts from North
Africa and Arabian
Peninsula
Remote oceanic aerosol
Intensive Field Experiments
Name
TARFOX
NEAQS
SCAR-A
CLAMS
INTEX-NA, ICARTT
DOEAIOP
MILAGRO
TexAQS/GoMACCS
ARCTS
ARCTAS
MINOS
LACE98
Aerosols99
INDOEX
ABC
EAST-AIRE
INTEX-B
ACE-Asia
TRACE-P
PEM-WestA&B
BASE-A
SCAR-B
LBA-SMOCC
SAFARI2000
SAFARI92
TRACE-ADABEX
DABEX
SAMUM
SHADE
PRIDE
UAE2
ACE-1
Location
North Atlantic
North Atlantic
North America
East Coast of U.S.
North America
Northern Oklahoma
Mexico City, Mexico
Texas and Gulf of Mexico
North-central Alaska to
Greenland (Arctic haze)
Northern Canada (smoke)
Mediterranean region
Lindberg, Germany
Atlantic
Indian subcontinent and
Indian Ocean
South and East Asia
China
Northeastern Pacific
East Asian and Northwest
Pacific
Western Pacific off East
Asia
Brazil
Brazil
Amazon basin
South Africa and South
South Atlantic
West Africa
Southern Morocco
West coast of North Africa
Puerto Rico
Arabian Peninsula
Southern Oceans
Time Period
July 1996
July-August 2002
1993
July-August 2001
Summer 2004
May 2003
March 2006
August-September 2006
March-April 2008
June-July 2008
July-August 2001
July-August 1998
January-February 1999
January-April 1 998 & 1999
Ongoing
March-April 2005
April 2006
April 2001
March-April 2001
September-October 1991
February-March 1994
1989
August-September 1995
September-November
2002
August-September 2000
September-October 1992
September-October 1992
January-February 2006
May-June 2006
September 2000
June-July 2000
August-September 2004
December 1995
Russell etal. (1999. 190363)
Quinn and Bates (2003, 049189)
Remer etal. (1997,190216)
Smith et al. (2005, 190401)
Fehsenfeld et al. (2006, 190531)
Ferrare etal. (2006. 190561)
Molina et al. (2008, 192019)
Jiang etal. (2008. 156609):
Lu et al. (2008, 190455)
http://www.espo.nasa.gov/arctas/
Lelieveld et al. (2002. 190361)
Ansmann et al. (2002)
Bates et al. (2001 , 043385)
Ramanathan et al. (2001 , 190196)
Ramanathan and Crutzen (2003, 190198)
Li etal. (2007. 190392)
Singh et al. (2008, 190394)
Huebert etal. (2003,190623);
Seinfeld et al. (2004, 190388)
Jacob et al. (2003, 190987)
Hoell et al. (1 996, 190607: 1 997, 057373)
Kaufman etal. (1992. 044557)
Kaufman etal. (1998. 089989)
Andreae etal. (2004, 155658)
King et al. (2003, 094395)
Lindesay etal. (1996,190403)
Fishman etal. (1996,190566)
Haywood etal. (2008,190602)
Heintzenberg et al. (2008, 190605)
Tanre et al. (2003, 190454)
Reid etal. (2003, 190213)
Reid etal. (2008. 190214)
Bates etal. ((1998. 190063) ;
Quinn and Coffman (1998.190918)
Source: Yu (2006,156173)
December 2009
9-91
-------
Table 9-6. Summary of major U.S. surface in situ and remote sensing networks for the
tropospheric aerosol characterization and radiative forcing research. All the reported
quantities are column-integrated or column-effective, except as indicated.
In situ
Remote
Sensing
Surface Network
NOAA ESRL aerosol monitoring
(http://www.esrl.noaa.gov/amd/aero/)
NPS/EPA IMPROVE
(http://vista.cira.colostate.edu/improve/)
NASAAERONET
(http ://aeronet.gsfc .nasa .gov)
DOE ARM (http://www.arm.gov)
NOAASURFRAD
(http://www.srrb.noaa.gov/surfrad/)
AERONET-MAN
(http ://aeronet.gsfc .nasa .gov/maritime aero
sol network.html)
NASAMPLNET
(http ://mplnet.gsfc .nasa .gov/)
Measured/Derived Parameters
Loading
Near-surface
extinction
coefficient,
optical depth,
CN/CNN
number
concentrations
Near-surface
mass
concentrations
and derived
extinction
coefficients by
species
Optical depth
Vertical profiles
of backscatter/
extinction
coefficient
Size, Shape
Angstrom
exponent,
hemispheric
backscatter
fraction,
asymmetry
factor,
hygroscopic
growth
Fine and coarse
separately
Fine-mode
fraction,
Angstrom
exponents,
asymmetry
factor, phase
function, non-
spherical
fraction
N/A
N/A
Absorption
Single-
scattering
albedo,
absorption
coefficient
Single-scattering
albedo,
absorption
coefficient
Single-
scattering
albedo,
absorption
optical depth,
refractive
indices
N/A
N/A
Chemistry
Chemical
composition in
selected sites
and periods
Ions ammonium
S04 ,
ammonium
nitrate organics,
EC, fine soil
N/A
N/A
N/A
Spatial
Coverage
5 baseline
stations,
several
regional
stations,
aircraft and
mobile
platforms
156NPsand
wilderness
areas in the
U.S.
-200 sites
over global
land and
islands
6 sites and 1
mobile facility
in N.America,
Europe, and
Asia
7 sites in the
U.S.
Global Ocean
-30 sites in
major
continents,
usually
co-located
with
AERONET
and ARM sites
Temporal
Coverage
1976 onward
1988 onward
1993 onward
1989 onward
1995 onward
2004-present
periodically
2000 onward
December 2009
9-92
-------
Figure 9-61. Geographical coverage of active AERONET sites in 2006.
9.3.2.5. Ground-Based Remote Sensing Measurement Networks
The Aerosol Robotic Network (AERONET) program is a federated ground-based remote
sensing network of well-calibrated sun photometers and radiometers
dittp://aeronet. gsfc.nasa. gov).
AERONET includes about 200 sites around the world, covering all major tropospheric
aerosol regimes (Holben et al., 1998, 155848: 2001, 190618), as illustrated in Figure 9-61.
Spectral measurements of sun and sky radiance are calibrated and screened for cloud-free
conditions (Smirnov et al., 2000, 190397). AERONET stations provide direct, calibrated
measurements of spectral AOD (normally at wavelengths of 440, 670, 870, and 1020 nm) with
an accuracy of ± 0.015 (Eck et al., 1999, 190390). In addition, inversion-based retrievals of a
variety of effective, column-mean properties have been developed, including aerosol single-
scattering albedo, size distributions, fine-mode fraction, degree of non-sphericity, phase
function, and asymmetry factor (Dubovik and King, 2000, 190197; Dubovik et al., 2000,
190177: Dubovik et al., 2002, 190202: O'Neill et al., 2003, 180187). The SSA can be retrieved
with an accuracy of ± 0.03, but only for AOD >0.4 (Dubovik et al., 2002,190202), which
precludes much of the planet. These retrieved parameters have been validated or are
undergoing validation by comparison to in situ measurements (e.g., Haywood et al., 2003,
190599: Leahy et al., 2007, 190232: Magi et al., 2005, 190468).
Recent developments associated with AERONET algorithms and data products include:
• simultaneous retrieval of aerosol and surface properties using combined AERONET
and satellite measurements (Sinyuk et al., 2007, 190395) with surface reflectance
taken into account (which significantly improves AERONET SSA retrieval
accuracy) (Eck et al., 2008, 190409):
• the addition of ocean color and high frequency solar flux measurements; and
• the establishment of the Maritime Aerosol Network (MAN) component to monitor
aerosols over the World oceans from ships of- opportunity (Smirnov et al., 2006,
190400).
Because of consistent calibration, cloud-screening, and retrieval methods, uniformly
acquired and processed data are available from all stations, some of which have operated for
over 10 years. These data constitute a high-quality, ground-based aerosol climatology and, as
such, have been widely used for aerosol process studies as well as for evaluation and
validation of model simulation and satellite remote sensing applications (e.g., Chin et al.,
2002, 189996: Kahn et al., 2005, 190966: Remer et al., 2005, 190221: Yu et al., 2003, 156171:
2006, 156173). In addition, AERONET retrievals of aerosol size distribution and refractive
indices have been used in algorithm development for satellite sensors (Levy et al., 2007,
190377: Remer et al., 2005, 190221). A set of aerosol optical properties provided by
AERONET has been used to calculate the aerosol direct radiative forcing (Procopio et al.,
2004, 190571: Zhou et al., 2005, 156183), which can be used to evaluate both satellite remote
sensing measurements and model simulations.
AERONET measurements are complemented by other ground-based aerosol networks
having less geographical or temporal coverage, such as the Atmospheric Radiation
Measurement (ARM) network (Ackerman and Stokes, 2003, 192080), NOAA's national
surface radiation budget network (SURFRAD) (Augustine et al., 2008, 189913) and other
networks with multifilter rotating shadowband radiometer (MFRSR) (Harrison et al., 1994,
045805: Michalsky et al., 2001,190537), and several lidar networks including:
December 2009 9-93
-------
• NASA Micro Pulse Lidar Network (MPLNET) (Welton et al., 2001, 157133; Welton
et al., 2002, 190631);
• Regional East Atmospheric Lidar Mesonet (REALM) in North America (Hoff and
McCann, 2002, 190612; Hoff et al., 2004, 190617);
• European Aerosol Research Lidar Network (EARLINET) (Matthias et al., 2004,
155971); and
Asian Dust Network (AD-Net) (e.g., Murayama et al., 2001, 155992).
Obtaining accurate aerosol extinction profile observations is pivotal to improving aerosol
radiative forcing and atmospheric response calculations. The values derived from these lidar
networks with state-of-the-art techniques (Schmid et al., 2006, 190372) are helping to fill this
need.
9.3.2.6. Synergy of Measurements and Model Simulations
Individual approaches discussed above have their own strengths and limitations, and are
usually complementary. None of these approaches alone is adequate to characterize large
spatial and temporal variations of aerosol physical and chemical properties and to address
complex aerosol-climate interactions. The best strategy for characterizing aerosols and
estimating their radiative forcing is to integrate measurements from different satellite sensors
with complementary capabilities from in situ and surface based measurements. Similarly,
while models are essential tools for estimating regional and global distributions and radiative
forcing of aerosols at present as well as in the past and the future, observations are required to
provide following, several synergistic approaches to studying aerosols and their radiative
forcing are discussed.
Closure Experiments
During intensive field studies, multiple platforms and instruments are deployed to sample
regional aerosol properties through a well-coordinated experimental design. Often, several
independent methods are used to measure or derive a single aerosol property or radiative
forcing. This combination of methods can be used to identify inconsistencies in the methods
and to quantify uncertainties in measured, derived, and calculated aerosol properties and
radiative forcings. This approach, often referred to as a closure experiment, has been widely
employed on both individual measurement platforms (local closure) and in studies involving
vertical measurements through the atmospheric column by one or more platforms (column
closure) (Quinn et al., 1996, 192021; Russell et al., 1997, 190359).
Past closure studies have revealed that the best agreement between methods occurs for
submicrometer, spherical particles such that different measures of aerosol optical properties
and optical depth agree within 10-15% and often better (e.g., Clarke et al., 1996, 190003;
Collins et al., 2000, 190059; Quinn et al., 2004, 190937; Schmid et al., 2000, 190369). Larger
particle sizes (e.g., sea salt and dust) present inlet collection efficiency issues and non-
spherical particles (e.g., dust) lead to differences in instrumental responses. In these cases,
differences between methods for determining aerosol optical depth can be as great as 35%
(Doherty et al., 2005, 190027; Wang et al., 2003, 157106). Closure studies on aerosol clear-sky
DRF reveal uncertainties of about 25% for sulfate/carbonaceous aerosol and 60% for dust-
containing aerosol (Bates et al., 2006, 189912). Future closure studies could integrate surface-
and satellite-based radiometric measurements of AOD with in situ optical, microphysical, and
aircraft radiometric measurements for a wide range of situations. There is also a need to
maintain consistency in comparing results and expressing uncertainties (Bates et al., 2006,
189912).
Constraining Models with In Situ Measurements
In situ measurements of aerosol chemical, microphysical, and optical properties with
known accuracy, based in part on closure studies, can be used to constrain regional CTM
simulations of aerosol direct forcing, as described by Bates et al. (2006, 189912). A key step in
the approach is assigning empirically derived optical properties to the individual chemical
components generated by the CTM for use in a Radiative Transfer Model (RTM). Specifically,
regional data from focused, short-duration field programs can be segregated according to
aerosol type (sea salt, dust, or sulfate/carbonaceous) based on measured chemical composition
and particle size. Corresponding measured optical properties can be carried along in the sorting
process so that they, too, are segregated by aerosol type. The empirically derived aerosol
properties for individual aerosol types, including mass scattering efficiency, single-scattering
albedo, and asymmetry factor, and their dependences on relative humidity, can be used in place
of assumed values in CTMs. Short-term, focused measurements of aerosol properties (e.g.,
aerosol concentration and AOD) also can be used to evaluate CTM parameterizations on a
regional basis, to suggest improvements to such uncertain model parameters, such as emission
factors and scavenging coefficients (e.g., Koch et al., 2007, 190185). Improvements in these
parameterizations using observations yield increasing confidence in simulations covering
regions and periods where and when measurements are not available. To evaluate the aerosol
December 2009 9-94
-------
properties generated by CTMs on broader scales in space and time, satellite observations and
long-term in situ measurements are required.
Improved Model Simulations with Satellite Measurements
Global measurements of aerosols from satellites (mainly AOD) with well-defined
accuracies offer an opportunity to evaluate model simulations at large spatial and temporal
scales. The satellite measurements can also be used to constrain aerosol model simulations and
hence the assessment of aerosol DRF through data assimilation or objective analysis process
(e.g., Collins et al, 2001, 189987; Liu et al, 2005, 190414; Yu et al, 2003, 156171; 2004,
190926: 2006, 156173: Zhang et al., 2008, 190932). Both satellite retrievals and model
simulations have uncertainties. The goal of data integration is to minimize the discrepancies
between them, and to form an optimal estimate of aerosol distributions by combining them,
typically with weights inversely proportional to the square of the errors of individual
descriptions. Such integration can fill gaps in satellite retrievals and generate global
distributions of aerosols that are consistent with ground-based measurements (Collins et al.,
2001,189987; Liu et al., 2005, 190414: Yu et al., 2003, 156171: 2006, 156173). Recent efforts
have also focused on retrieving global sources of aerosol from satellite observations using
inverse modeling, which may be valuable for reducing large aerosol simulation uncertainties
(Dubovik et al., 2007, 190211). Model refinements guided by model evaluation and integration
practices with satellite retrievals can then be used to improve aerosol simulations of the pre-
and post-satellite eras. Current measurement-based understanding of aerosol characterization
and radiative forcing is assessed in Section 9.3.3 through intercomparisons of a variety of
measurement-based estimates and model simulations published in literature. This is followed
by a detailed discussion of major outstanding issues in Section 9.3.4.
• STEM-INDOEX
CD Observation - INDOEX
• STEM - ACE-Asia
cn Observation - ACE-AsiE
• STEM - ICARTT
CD Observation - ICARTT
Sulfate Ammonium
POM
Sea Salt
E 100
Ł 10
0
§
o
• STEM-INDOEX
CD Observation - INDOEX
• STEM - ACE-Asia
CD Observation - ACE-Asij
• STEM - ICARTT
CD Observation - ICARTT
Source: Bates etal. (2006,1899121.
Figure 9-62. Comparison of the mean concentration (ug/m3) and standard deviation of the
modeled (STEM) aerosol chemical components with shipboard measurements
during INDOEX, ACE-Asia, and ICARTT.
Further complexity is added when attempting to relate surface PM2.5 to aerosol optical depths.
The main approach to derive surface PM2.5 from satellite optical depths from MISR is based on the
December 2009
9-95
-------
use of model derived profiles to determine ratios of aerosol optical depth to surface PM25 (Liu and
Koutrakis, 2007, 187007; Liu et al, 2007, 098197; Van Donkelaar et al, 2006, 192108).
Van Donkelaar et al. (2006, 192108) derived r = 0.69 (MODIS) and 0.58 (MISR) with annual
average ground based PM2.5 across the U.S. and Canada. For comparison, r between AERONET total
AOD and surface measurements was 0.71. On average, MODIS tended to overestimate surface PM2.5
by ~5 (ig/m3, while MISR estimates were biased high by about 3 (ig/m3. Liu et al. (2007, 098197)
and Liu and Koutrakis (2007, 187007) used MISR derived fractional AODs in their analysis and
found improvement in the retrievals when fractional AODs were used instead of total AOD, allowing
for better fits to the radiance data. They found that fractional AODs can explain 13-62% of the
variability in PM2.5 and its components in the eastern U.S. and 28-56% of the variability in the
western U.S. The models tended to underpredict PM25 by -7-8% in both the East and West. The
relative errors in surface PM25 were estimated to be 30% in the East and 34% in the West. For AODs
>0.15 (nominal continental background values), dust particles could be distinguished from other
particles with an estimated error of 4%. Performance improves substantially over polluted urban
areas because they have much larger AOD. For example, Gupta et al. (2006, 137694) derived a
Pearson r between MODIS AOD and surface PM25 of 0.96 for several urban areas around the world.
The MODIS Aerosol Optical Depth (AOD)/ In-situ PM2 5 correlation summary plot shown in Figure
9-63 below illustrates the correlation between AOD and surface PM25 across the U.S. and parts of
Canada. The correlation is based on coincident MODIS AOD pixels and 1-h PM25 concentrations
from the in-situ continuous surface monitors. The parameter plotted is the monitoring site-specific
running correlation coefficient during the preceding 60 days (in color scale). The correlation
coefficient has values between 1 (perfectly correlated) and -1 (perfectly anti-correlated). A value of
zero indicates that the two measurements vary independently of each other.
The running time period of the correlation determination is given in the plot title, 20090704-
20090901. The size of the point at each site indicates the number of coincidences between MODIS
AOD pixels and the measured surface 1-h PM25 concentrations for that period. Correlation
significance generally increases with increasing number of coincidences. Higher correlations suggest
the MODIS AOD pixel is reflective of in-situ surface PM25 mass concentrations at the monitor
location.
December 2009 9-96
-------
20190704-20090901 Correlation between AIRNOW 1 -hour PM2.5 and MODIS AOD
ff Coincide ness
• 4-
• 30-40
* 20-30
• 10-20
-1.0
-0.5
0.0
Correlation
0.5
1.0
Figure 9-63. Correlations between one-hour PMzs surface measurements in the U.S. and
southern Canada reported to AIRNOW and MODIS satellite AOD values for the
period between 4 July and 1 September 2009. Symbol size indicates number of
coincident points at that location in that period; symbol color is indexed to
degree of correlation from -1 (cooler) to +1 (warmer).
The data and image of the aerosol comparison shown in Figure 9-63 were taken from the
multi-agency project, Infusing satellite Data into Environmental Air Quality Applications (IDEA), a
partnership of NASA, NOAA, and EPA designed to improve air quality assessment, management,
and prediction by infusing NASA satellite measurements into NOAA and EPA analyses for public
benefit. IDEA is funded by these three agencies and managed by the University of Maryland
Baltimore County and NOAA.
9.3.3. Assessments of Aerosol Characterization and Climate Forcing
Sections 9.3.3 through 9.3.6 come directly from CCSP SAP2.3 Chapters 2, Section 2.3
through Chapter 3, Section 3.8, with section, table, and figure numbers changed to be internally
consistent with this ISA.
December 2009
9-97
-------
This section focuses on the assessment of measurement-based aerosol characterization
and its use in improving estimates of the direct radiative forcing on regional and global scales.
In situ measurements provide highly accurate aerosol chemical, microphysical, and optical
properties on a regional basis and for the particular time period of a given field campaign.
Remote sensing from satellites and ground-based networks provide spatial and temporal
coverage that intensive field campaigns lack. Both in situ measurements and remote sensing
have been used to determine key parameters for estimating aerosol direct radiative forcing
including aerosol single scattering albedo, asymmetry factor, optical depth remote sensing has
also been providing simultaneous measurements of aerosol optical depth and radiative fluxes
that can be combined to derive aerosol direct radiative forcing at the TOA with relaxed
requirement for characterizing aerosol properties. Progress in using both satellite and surface-
based measurements to study aerosol-cloud interactions and aerosol indirect forcing is also
discussed.
9.3.3.1. The Use of Measured Aerosol Properties to Improve Models
The wide variety of aerosol data sets from intensive field campaigns provides a rigorous
"testbed" for model simulations of aerosol properties and distributions and estimates of DRF.
As described in Section 9.3.2.6, in situ measurements can be used to constrain regional CTM
simulations of aerosol properties, DRF, anthropogenic component of DRF, and to evaluate
CTM parameterizations. In addition, in situ measurements can be used to develop simplifying
parameterizations for use by CTMs.
Several factors contribute to the uncertainty of CTM calculations of size-distributed
aerosol composition including emissions, aerosol removal by wet deposition, processes
involved in the formation of secondary aerosols and the chemical and microphysical evolution
of aerosols, vertical transport, and meteorological fields including the timing and amount of
precipitation, formation of clouds, and relative humidity. In situ measurements made during
focused field campaigns provide a point of comparison for the CTM-generated aerosol
distributions at the surface and at discrete points above the surface. Such comparisons are
essential for identifying areas where the models need improvement.
December 2009 9-98
-------
Pittsburgh, PA PSP, NY Chebogue Edinburgh UK Manchester, UK TORCH 1, UK TORCH 2, UK Hyytiala QUEST Taunus, Germany
CAN " ' (Winter) (Summer) ' Finland Finland 3
12 ug/m 2.9 ug/m 3.0 |jg/m 5.2 ug/m 14.0 ug/m 5.3 ug/m 7.6 ug/m 2.0 ug/m 2.6 ug/m 16 ug/m
»«• • • • •
7 0 ug/m >v \ \ ^v \ /
I N150 \ \00 ^ \ /
12 ug/m3 13 ug/m3 31 ug/m3 2.8 ug/m3 8.5 ug/m3 1.5 ug/m3 2.3 ug/m3 11 ug/m3 13 ug/m3 13 ug/m
*
Houston Mexico City Duke Forest Off Coast NE Mace Head Jungfraujoch Cheju Island Okinawa Island Fukue Island
TX NC US Ireland Switzerland Korea Japan Japan
Source: Data from Zhang et al. (2007,189998).
Figure 9-64. Location of aerosol chemical composition measurements with aerosol mass
spectrometers. Colors for the labels indicate the type of sampling location:
urban areas (blue), <100 mi downwind of major cites (black), and rural/remote
areas >100 miles downwind (pink). Pie charts show the average mass
concentration and chemical composition: organics (green), S042~(red), nitrate
(blue), ammonium (orange), and chloride (purple), of non-refractory PMi.
Figure 9-62 shows a comparison of submicrometer and supermicrometer aerosol chemical
components measured during INDOEX, ACEAsia, and ICARTT onboard a ship and the same
values calculated with the STEM Model (e.g., Bates et al., 2004, 189958: Carmichael et al.,
2002, 148319; Carmichael et al., 2003, 190042; Streets et al., 2006, 157019; Tang et al., 2003,
190441: Tang et al., 2004, 190445). To permit direct comparison of the measured and modeled
values, the model was driven by analyzed meteorological data and sampled at the times and
locations of the shipboard measurements every 30 min along the cruise track. The best
agreement was found for submicrometer sulfate and BC. The agreement was best for sulfate;
this is attributed to greater accuracy in emissions, chemical conversion, and removal for this
component. Underestimation of dust and sea salt is most likely due to errors in model-
calculated emissions. Large discrepancies between the modeled and measured values occurred
for submicrometer particulate organic matter (POM) (INDOEX), and for particles in the
supermicrometer size range such as dust (ACE-Asia), and sea salt (all regions). The model
underestimated the total mass of the supermicrometer aerosol by about a factor of 3. POM
makes up a large and variable fraction of aerosol mass throughout the anthropogenically
influenced northern hemisphere, and yet models have severe problems in properly representing
this type of aerosol. Much of this discrepancy follows from the models inability to represent
the formation of secondary organic aerosols (SOA) from the precursor volatile organic
compounds (VOC). Figure 9-64 shows a summary of the results from aerosol mass
spectrometer measurements at 30 sites over North America, Europe, and Asia. Based on
aircraft measurements of urban-influenced air over New England, de Gouw et al. (2005,
190020) found that POM was highly correlated with secondary anthropogenic gas phase
species suggesting that the POM was derived from secondary anthropogenic sources and that
the formation took one day or more.
December 2009
9-99
-------
25-
120-
O la '
CL
I 10-
J3
13
W 5-
0-
0 10 20 30
Photochemical age (h)
I I
0.0 0.5 1.0 1.5
Acetylene (ppbv)
J I
O
CL
25-
20-
10-
5-
0-
I i l
10 20 30
Iso-propyl nitrate (pptv)
Source: Data from de Guowet al. (2005,190020).
Figure 9-65. Scatterplots of the submicrometer POM measured during NEAQS versus
A) acetylene and B) iso-propyl nitrate. The colors of the data points in A) denote
the photochemical age as determined by the ratios of compounds of known OH
reactivity. The gray area in A) shows the range of ratios between submicrometer
POM and acetylene observed by Kirchstetter et al. (1999, 010642) in tunnel
studies.
Figure 9-65 shows scatterplots of submicrometer POM versus acetylene (a gas phase
primary emitted VOC species) and isopropyl nitrate (a secondary gas phase organic species
formed by atmospheric reactions). The increase in submicrometer POM with increasing
photochemical age could not be explained by the removal of VOC alone, which are its
traditionally recognized precursors. This result suggests that other species must have
contributed and/or that the mechanism for POM formation is more efficient than assumed by
models. Similar results were obtained from the 2006 MILAGRO field campaign conducted in
Mexico City (Kleinman et al., 2008, 190074X and comparisons of GCM results with several
long-term monitoring stations also showed that the model underestimated organic aerosol
concentrations (Koch et al., 2007, 190185). Recent laboratory work suggests that isoprene may
be a major SOA source missing from previous atmospheric models (Henze and Seinfeld,
2006, 190606: Kroll et al., 2006, 190195X but underestimating sources from certain economic
sectors may also play a role (Koch et al., 2007, 190185). Models also have difficulty in
representing the vertical distribution of organic aerosols, underpredicting their occurrence in
the free troposphere (FT) (Heald et al., 2005, 190603). While organic aerosol presents models
with some of their greatest challenges, even the distribution of well-characterized sulfate
aerosol is not always estimated correctly in models (Shindell et al., 2008, 190391).
Comparisons of DRF and its anthropogenic component calculated with assumed optical
properties and values constrained by in situ measurements can help identify areas of
uncertainty in model parameterizations. In a study described by Bates et al. (2006, 189912X
two different CTMs (MOZART and STEM) were used to calculate dry mass concentrations of
the dominant aerosol species (sulfate, organic carbon, BC, sea salt, and dust).
In situ measurements were used to calculate the corresponding optical properties for each
aerosol type for use in a radiative transfer model. Aerosol DRF and its anthropogenic
component estimated using the empirically derived and a priori optical properties were then
compared. The DRF and its anthropogenic component were calculated as the net downward
solar flux difference between the model state with aerosol and of the model state with no
aerosol. It was found that the constrained optical properties derived from measurements
increased the calculated AOD (34 ± 8%), TOADRF (32 ± 12%), and anthropogenic
component of TOADRF (37 ± 7%) relative to runs using the a priori values. These increases
were due to larger values of the constrained mass extinction efficiencies relative to the a priori
values. In addition, differences in AOD due to using the aerosol loadings from MOZART
versus those from STEM were much greater than differences resulting from the a priori vs.
constrained RTM runs. In situ observations also can be used to generate simplified
parameterizations for CTMs and RTMs thereby lending an empirical foundation to uncertain
parameters currently in use by models. CTMs generate concentration fields of individual
aerosol chemical components that are then used as input to radiative transfer models (RTMs)
for the calculation of DRF. Currently, these calculations are performed with a variety of
simplifying assumptions concerning the RH dependence of light scattering by the aerosol.
December 2009
9-100
-------
Chemical components often are treated as externally mixed each with a unique RH
dependence of light scattering. However, both model and measurement studies reveal that
POM, internally mixed with water-soluble salts, can reduce the hygroscopic response of the
aerosol, which decreases its water content and ability to scatter light at elevated relative
humidity (e.g., Carrico et al, 2005, 190052; Saxena et al, 1995, 077273).
The complexity of the POM composition and its impact on aerosol optical properties
requires the development of simplifying parameterizations that allow for the incorporation of
information derived from field measurements into calculations of DRF (Quinn et al., 2005,
156033). Measurements made during INDOEX, ACE-Asia, and ICARTT revealed a
substantial decrease in fasp(RH) with increasing mass fraction of POM in the accumulation
mode. Based on these data, a parameterization was developed that quantitatively describes the
relationship between POM mass fraction and fasp(RH) for accumulation mode sulfate-POM
mixtures (Quinn et al., 2005, 156033). This simplified parameterization may be used as input
to RTMs to derive values of fasp(RH) based on CTM estimates of the POM mass fraction.
Alternatively, the relationship may be used to assess values of fasp(RH) currently being used
in RTMs.
9.3.3.2. Intercomparisons of Satellite Measurements and Model Simulation of Aerosol
Optical Depth
As aerosol DRF is highly dependent on the amount of aerosol present, it is of first-order
importance to improve the spatial characterization of AOD on a global scale. This requires an
evaluation of the various remote sensing AOD data sets and comparison with model-based
AOD estimates. The latter comparison is particularly important if models are to be used in
projections of future climate states that would result from assumed future emissions. Both
remote sensing and model simulation have uncertainties and satellite-model integration is
needed to obtain an optimum description of aerosol distribution.
Figure 9-65 shows an intercomparison of annual average AOD at 550 nm from two recent
satellite aerosol sensors (MODIS and MISR), five model simulations (GOCART, GISS,
SPRINTARS, LMDZ-LOA, LMDZ-INCA) and three satellite-model integrations (MO_GO,
MI_GO, MO_MI_GO). These model-satellite integrations are conducted by using an optimum
interpolation approach (Yu et al., 2003, 156171) to constrain GOCART simulated AOD with
that from MODIS, MISR, or MODIS over ocean and MISR over land, denoted as MO_GO,
MI_GO, and MO_MI_GO, respectively. MODIS values of AOD are from Terra Collection 4
retrievals and MISR AOD is based on early post launch retrievals. MODIS and MISR
retrievals give a comparable average AOD on the global scale, with MISR greater than
MODIS by 0.01-0.02 depending on the season. However, differences between MODIS and
MISR are much larger when land and ocean are examined separately: AOD from MODIS is
0.02-0.07 higher over land but 0.03-0.04 lower over ocean than the AOD from MISR. Several
major causes for the systematic MODIS-MISR differences have been identified, including
instrument calibration and sampling differences, different assumptions about ocean surface
boundary conditions made in the individual retrieval algorithms, missing particle property or
mixture options in the look-up tables, and cloud screening (Kahn et al., 2007, 190963). The
MODIS-MISR AOD differences are being reduced by continuous efforts on improving
satellite retrieval algorithms and radiance calibration. The new MODIS aerosol retrieval
algorithms in Collection 5 have resulted in a reduction of 0.07 for global land mean AOD
(Levy et al., 2007, 190379), and improved radiance calibration for MISR removed —40% of
AOD bias over dark water scenes (Kahn et al., 2005, 190965).
The annual and global average AOD from the five models is 0.19 ± 0.02 (mean ± standard
deviation) over land and 0.13 ± 0.05 over ocean, respectively. Clearly, the model-based mean
AOD is smaller than both MODIS- and MISR-derived values (except the GISS model). A
similar conclusion has been drawn from more extensive comparisons involving more models
and satellites (Kinne et al., 2006, 155903). On regional scales, satellite-model differences are
much larger. These differences could be attributed in part to cloud contamination (Kaufman et
al., 2005, 155891; Zhang et al., 2005, 190931) and 3D cloud effects in satellite retrievals
(Kaufman et al., 2005, 155891: Wen et al., 2006, 179964) or to models missing important
aerosol sources/sinks or physical processes (Koren et al., 2007, 190192). Integrated satellite-
model products are generally in-between the satellite retrievals and the model simulations, and
agree better with AERONET measurements (e.g., Yu et al., 2003, 156171). As in comparisons
between models and in situ measurements (Bates et al., 2006, 189912), there appears to be a
relationship between uncertainties in the representation of dust in models and the uncertainty
in AOD, and its global distribution.
For example, the GISS model generates more dust than the other models (Figure 9-67), resulting in a closer agreement with MODIS and MISR in the global mean (Source: Data
taken from Kinne et al. (2006,155903)
Figure 9-66). However, the distribution of AOD between land and ocean is quite different
from MODIS- and MISR-derived values.
Figure 9-67 shows larger model differences in the simulated percentage contributions of
individual components to the total aerosol optical depth on a global scale, and hence in the
simulated aerosol single-scattering properties (e.g., single-scattering albedo, and phase
function), as documented in Kinne et al. (2006, 155903). This, combined with the differences
December 2009 9-101
-------
in aerosol loading (as characterized by AOD) determines the model diversity in simulated
aerosol direct radiative forcing, as discussed later. However, current satellite remote sensing
capability is not sufficient to constrain model simulations of aerosol components.
• Und BOcean • Global
^ rf 0°° JP &* ^
Source: Data taken from Kinne et al. (2006,1559031
Figure 9-66. Comparison of annual mean aerosol optical depth (AOD).
100%
80%
60%
am
• • • •
• ss
DU
POM
• BC
SU
.^ X
Source: Data taken from Kinne et al. (2006,1559031.
Figure 9-67. Percentage contributions of individual aerosol components. SU - sulfate, BC -
BC, POM - particulate organic matter, DU - dust, SS - sea salt; to the total
aerosol optical depth (at 550 nm) on a global scale simulated by the five models.
9.3.3.3. Satellite-Based Estimates of Aerosol Direct Radiative Forcing
Table 9-7 summarizes approaches to estimating the aerosol direct radiative forcing,
including a brief description of methods, identifies major sources of uncertainty, and provides
references. These estimates fall into three broad categories, namely (A) satellite-based, (B)
satellite-model integrated, and (C) model-based. As satellite aerosol measurements are
generally limited to cloud-free conditions, the discussion here focuses on assessments of clear-
sky aerosol direct radiative forcing, a net (downwelling minus upwelling) solar flux difference
between with aerosol (natural + anthropogenic) and in the absence of aerosol.
December 2009
9-102
-------
Global Distributions
Figure 9-68 shows global distributions of aerosol optical depth at 550 nm (left panel) and
diurnally averaged clear-sky TOA DRF (right panel) for March-April-May (MAM) based on
the different approaches. The DRF at the surface follows the same pattern as that at the TOA
but is significantly larger in magnitude because of aerosol absorption. It appears that different
approaches agree on large-scale patterns of aerosol optical depth and the direct radiative
forcing. In this season, the aerosol impacts in the Northern Hemisphere are much larger than
those in the Southern Hemisphere. Dust outbreaks and biomass burning elevate the optical
depth to more than 0.3 over large parts of North Africa and the tropical Atlantic. In the tropical
Atlantic, TOA cooling as large as -10 W/m2 extends westward to Central America. In eastern
China, the optical depth is as high as 0.6-0.8, resulting from the combined effects of industrial
activities and biomass burning in the south, and dust outbreaks in the north. The Asian impacts
also extend to the North Pacific, producing a TOA cooling of more than -10 W/m . Other areas
having large aerosol impacts include Western Europe, midlatitude North Atlantic, and much of
South Asia and the Indian Ocean. Over the "roaring forties" in the Southern Hemisphere, high
winds generate a large amount of sea salt. Elevated optical depth, along with high solar zenith
angle and hence large backscattering to space, results in a band of TOA cooling of more than -
4 W/m2. However, there is also some question as to whether thin cirrus (e.g., Zhang et al,
2005, 190931) and unaccounted-for whitecaps contribute to the apparent enhancement in AOD
retrieved by satellite. Some differences exist between different approaches. For example, the
early post-launch MISR retrieved optical depths over the southern hemisphere oceans are
higher than MODIS retrievals and GOCART simulations. Over the "roaring forties," the
MODIS derived TOA solar flux perturbations are larger than the estimates from other
approaches.
December 2009 9-103
-------
Table 9-7. Summary of approaches to estimating the aerosol direct radiative forcing in three
categories: (1) satellite retrievals; (2) satellite-model integrations; and (3) model
simulations.
Category Product
Brief Description
Identified Sources of Uncertainty
Major References
MODIS
Using MODIS retrievals of a linked
set of AOD, u0, and phase function
consistently in conjunction with a
radiative transfer model (RTM) to
calculate TOA fluxes that best match
the observed radiances.
Radiance calibration, cloud-aerosol discrimination,
instantaneous-to-diurnal scaling, RTM
parameterizations
Remer and Kaufman (2006,
190222)
MODIS A
Splitting MODIS AOD over ocean into
mineral dust, sea salt, and biomass-
burning and pollution; using AERONET
measurements to derive the size
distribution and single-scattering albedo ^ngieAERONET site'tocharacte'rFe a large r^Jion
for individual components.
Satellite AOD and FMF retrievals, overestimate due to
summing up the compositional direct forcing, use of a
Bellouin et al. (2005,1556841
A Satellite
retrievals
CERES A
Using CERES fluxes in combination
with standard MODIS aerosol.
Using CERES fluxes in combination
CERES_B with NOAA NESDIS aerosol from
MODIS radiances.
Using CERES fluxes in combination
„___- „ with MODIS (ocean) and MISR (non-
- desert land) aerosol with new angular
models for aerosols.
Calibration of CERES radiances, large CERES
footprint, satellite AOD retrieval, radiance-to-flux
conversion (ADM), instantaneous-to-diurnal scaling,
narrow-to-broadband conversion
Loeb and Manalo-Smith
(2005,190433); Loeb and
Kato (2002,190432)
Zhang etal. (2005, 086743;
2005,157185); Zhang and
Christopher (2003,190928);
Christopher etal. (2006,
155729); Patadiaetal. (2008,
190558)
Using POLDER AOD in combination
POLDER with prescribed aerosol models
(similar to MODIS).
Similar to MODIS
Boucher and Tanre (2000,
190041): Bellouin etal.
(2003.189911)
MODIS_G
MISR G
Using GOCART simulations to fill
AOD gaps in satellite retrievals.
B. Satellite-
model
integrations
MO GO
Integration of MODIS and GOCART
AOD.
Integration of GOCART AOD with
MO_MI_GO retrievals from MODIS (Ocean) and
MISR (Land).
- Propagation of uncertainties associated with both
satellite retrievals and model simulations (but the
model-satellite integration approach does result in
- improved AOD quality for MO_GO, and 0_MI_GO)
'Aerosol single-scattering
albedo and asymmetry factor
are taken from GOCART
simulations
*Yuetal. (2003.156171:
2004.190926:2006.156173)
SeaWiFS
Using SeaWiFS AOD and assumed
aerosol models.
Similar to MODIS_G and MISR_G, too weak aerosol
absorption
Chou et al. (2002,190008)
GOCART
Offline RT calculations using
monthly avg aerosols with a time step
of 30 min (without the presence of
clouds).
CDDIMTADC Online RT calculations every 3 hrs
SPRINTARS
C. Model
simulations
GISS
Online model simulations and
weighted by clear-sky fraction.
Emissions, parameterizations of a variety of sub-grid
" aerosol processes (e.g., wet and dry deposition, cloud
convection, aqueous-phase oxidation), assumptions on
aerosol size, absorption, mixture, and humidification of
. particles, Meteorology fields, not fully evaluated
surface albedo schemes, RT parameterizations
LMDZ-INCA
Online RT calculations every 2 hrs
(cloud fraction = 0).
LMDZ-LOA
Online RT calculations every 2 hrs
(cloud fraction = 0).
Chin etal. (2002,189996) ;Yu
etal. (2004,190926)
Takemura et al. (2002,
190438:2005.190439)
Koch and Hansen (2005,
190183): Koch etal. (2006.
190184)
Balkanskietal. (2007,
189979);Schulzetal. (2006,
190381): Kinne etal. (2006.
155903)
Reddy et al. (2005,190207:
2005.190208)
Source: Adapted from Yu et al. (2006,1561731.
December 2009
9-104
-------
MODIS
MODIS
MO Ml GO
GOCART
CERES A
MO MI_GO
•.
.0 ,05 ,1 .15 .2 ,3 .4 .5 .6 .8 I. -30 -20 -15 -10 -8 -6 -4 -2 0 5Wms
Source: Yu et al. (2006,156173)
Figure 9-68. Geographical patterns of seasonally (MAM) averaged aerosol optical depth at 550
nm (left panel) and the diurnally averaged clear-sky aerosol direct radiative (solar
spectrum) forcing (W/m2) at the TOA (right panel) derived from satellite (Terra)
retrievals. MODIS (Remer and Kaufman, 2006,190222; Remeret al., 2005,
190221): MISR (Kahn et al., 2005,190966): and CERES_A (Loeb and Manalo-
Smith, 2005,190433): GOCART simulations (Chin et al., 2002,189996: Yu et al.,
2004,190926): and GOCART-MODIS-MISR integrations (MO_MI_GO) (Yu et al.,
2006.156173).
December 2009
9-105
-------
Global Mean
Figure 9-69 summarizes the measurement- and model-based estimates of clear-sky
annual-averaged DRF at both the TOA and surface from 60°S to 60°N. Seasonal DRF values
for individual estimates are summarized in Table 9-8 and Table 9-9 for ocean and land,
respectively. Mean, median and standard error e (s=a/(n-1)1/2), where a is standard deviation
and n is the number of methods) are calculated for measurement- and model-based estimates
separately. Note that although the standard deviation or standard error reported here is not a
fully rigorous measure of a true experimental uncertainty, it is indicative of the uncertainty
because independent approaches with independent sources of errors are used (see Table 9-7; in
the modeling community, this is called the "diversity;" see Section 9.3.6).
Ocean
For the TOA DRF, a majority of measurement-based and satellite-model integration-based
estimates agree with each other within about 10%. On annual average, the measurement-based
estimates give the DRF of -5.5 ± 0.2 W/m2 (mean ± e) at the TOA and -8.7 ± 0.7 W/m2 at the
surface. This suggests that the ocean surface cooling is about 60% larger than the cooling at
the TOA. Model simulations give wide ranges of DRF estimates at both the TOA and surface.
The ensemble of five models gives the annual average DRF (mean± e) of-3.2 ± 0.6 W/m2 and
-4.9 ±0.8 W/m2 at the TOA and surface, respectively. On average, the surface cooling is about
37% larger than the TOA cooling, smaller than the measurement-based estimate of surface and
TOA difference of 60%. However, the 'measurement-based' estimate of surface DRF is
actually a calculated value, using poorly constrained particle properties.
Land
It remains challenging to use satellite measurements alone for characterizing complex
aerosol properties over land surfaces with high accuracy. As such, DRF estimates over land
have to rely largely on model simulations and satellite-model integrations. On a global and
annual average, the satellite-model integrated approaches derive a mean DRF of-4.9 W/m2 at
the TOA and -11.9 W/m at the surface respectively. The surface cooling is more than a factor
of 2 larger than the TOA cooling because of aerosol absorption. Note that the TOA DRF of -
4.9 W/m2 agrees quite well with the most recent satellite-based estimate of -5.1 ± 1.1 W/m2
over non-desert land based on coincident measurements of MISR AOD and CERES solar flux
(Patadia et al., 2008, 190558). For comparisons, an ensemble of five model simulations
derives a DRF (mean ± e) over land of-3.0 ± 0.6 W/m2 at the TOA and -7.6 ± 0.9 W/m2 at the
surface, respectively. Seasonal variations of DRF over land, as derived from both
measurements and models, are larger than those over ocean.
aaa
0
DOBS
MOD
-15
-12
DOBS
MOD
-3
0
OCEAN
LAND
OCEAN
LAND
Source: Yu et al. (2006,156173)
Figure 9-69. Summary of observation- and model-based (denoted as OBS and MOD,
respectively) estimates of clear-sky, annual average DRF at the TOA and at the
surface. The box and vertical bar represent median and standard error,
respectively.
December 2009
9-106
-------
Table 9-8.
Summary of seasonal and annual average clear-sky DRF (W/m2) at the TOA and the
surface (SFC) over global OCEAN derived with different methods and data.
DJF
MAM
JJA
SON
ANN
MODIS
MODIX_A*
CERES_A
CERES_B
CERES_C
MODIS_G
MISR_G**
MO_GO
MO_MI_GO
POLDER
SeaWiFS
Obs. Mean
Obs. Median
Obs.o
Obs.t
GOCART
SPRINTARS
GISS
LMDZ-INCA
LMDZ-LOA
Mod. Mean
Mod Median
Mod.o
Mod.t
Mod ./Obs.
TOA
-5.9
-6.0
-5.3
-3.8
-5.3
-5.5
-6.4
-4.9
-4.9
-5.7
-6.0
-5.4
-5.5
0.72
0.23
-3.6
-1.5
-3.3
-4.6
-2.2
-3.0
-3.3
1.21
0.61
.60
SFC
-8.2
-9.1
-10.3
-7.8
-7.9
-6.6
-8.3
-8.1
1.26
0.56
-5.7
-2.5
-4.1
-5.6
-4.1
-4.4
-41.
1.32
0.66
.51
TOA
-5.8
-6.4
-6.1
-4.3
-5.4
-5.7
-6.5
-5.1
-5.1
-5.7
-5.2
-5.6
-5.7
0.64
0.20
-4.0
-1.5
-3.5
-4.7
-2.2
-3.2
-3.5
1.31
0.66
.61
SFC
-8.9
-10.4
-11.4
-9.3
-9.2
-5.8
-9.2
-9.3
1.89
0.85
-7.2
-2.5
-4.6
-5.9
-3.7
-4.8
-4.6
1.84
0.92
.50
TOA
-6.0
-6.5
-5.4
-3.5
-5.2
-6.0
-7.0
-5.4
-5.5
-5.8
-4.9
-5.6
-5.5
0.91
0.29
-4.7
-1.9
-3.5
-5.0
-2.5
-3.5
-3.5
1.35
0.67
.64
SFC
-9.3
-10.6
-11.9
-9.4
-9.5
-5.6
-9.4
-9.5
2.10
0.94
-8.0
-3.3
-4.9
-6.3
-4.4
-5.4
-4.9
1.82
0.91
.52
TOA
-5.8
-6.4
-5.1
-3.6
-5.5
-6.3
-5.0
-5.0
-5.6
-5.3
-5.4
-5.4
0.79
0.26
-4.0
-1.5
-3.8
-4.8
-2.2
-3.3
-3.8
1.36
0.68
.70
SFC
-8.9
-9.8
-10.9
-8.7
-8.6
-5.7
-8.8
-8.8
1.74
0.78
-6.8
-2.5
-5.4
-5.5
-4.1
-4.9
-5.4
1.63
0.81
.61
TOA
-5.9
-6.4
-5.5
-3.8
-5.3
-5.7
-6.5
-5.1
-5.1
-5.7
-5.2***
-5.4
-5.5
-5.5
0.70
0.21
-4.1
-1.6
-3.5
-4.7
-2.3
-3.2
-3.5
1.28
0.64
.64
SFC
-8.9
-10.0
-11.1
-8.8
-8.7
-7.7***
-5.9
-8.7
-8.8
1.65
0.67
-6.9
-2.7
-4.8
-5.8
-4.1
-4.9
-4.8
1.60
0.80
.55
* High bias may result from adding the DRF of individual components to derive the total DRF (Bellouin et al., 2005,1556841.
**High bias most likely results from an overall overestimate of 20% in early post-launch MISR optical depth retrievals (Kahn et al., 2005,190966).
*** Bellouin et al. (2003,1899111 use AERONET retrieval of aerosol absorption as a constraint to the method in Boucher and Tanre (2000,1900411, deriving aerosol direct radiative forcing both at the TOA and
the surface.
Sources of data: MODIS (Remer & Kaufman, 2006), MODIS_A (Bellouin et al., 2005,1556841. POLDER (Bellouin et al., 2003,189911: Boucher and Tanre, 2000,1900411. CERES_A and CERES_B (Loeb
and Manalo-Smith, 2005, 1904331. CERES_C (Zhang et al., 2005, 1909301. MODIS_G, MISR_G, MO_GO, MO_MI_GO (Yu et al, 2004,190926: 2006, 1561731. SeaWFS (Chou et al, 2002, 1900081.
GOCART (Chin et al, 2002,189996: Yu et al, 2004,1909261. SPRINTARS (Takemura et al, 2002,1904381. GISS (Koch and Hansen, 2005,190183: Koch et al, 2006,1901841. LMDZ-INCA (Kinne et al,
2006,155903: Schulz et al, 2006,1903811. LMDZ-LOA (Reddy et al, 2005,190207: Reddy et al, 2005,1902081. Mean, median, standard deviation (o), and standard error (E) are calculated for observations
(Obs) and model simulations (Mod) separately. The last row is the ratio of model median to observational median (taken from Yu et al, 2006,156173).
December 2009
9-107
-------
Table 9-9. Summary of seasonal and annual average clear-sky DRF (W/m2) at the TOA and the
surface (SFC) over global LAND derived with different methods and data.
DJF MAM JJA SON ANN
Products ^^^^^^^^^^^^^^^^^^^^^^^^—^^^^^^^^^^^^^^^^^^^^^^^^—
TOA SFC TOA SFC TOA SFC TOA SFC TOA SFC
MODIS_G
MISR_G
MO_GO
MO_MI_GO
Obs. Mean
Obs. Median
Obs.o
Obs.t
GOCART
SPRINTARS
GISS
LMDZ-INCA
LMDZ-LOA
Mod. Mean
Mod Median
Mod.o
Mod. Ł
Mod ./Obs.
-4.1
-3.9
-3.5
-3.4
-3.7
-3.7
0.33
0.17
02.9
-1.4
-1.6
-3.0
-1.3
-2.0
-1.6
0.84
0.42
0.43
-9.1
-8.7
-7.5
-7.4
-8.2
-8.1
0.85
0.49
-6.1
-4.0
-3.9
-5.8
-5.4
-5.0
-5.4
1.03
0.51
0.67
-5.8
-5.1
-5.1
-4.7
-5.2
-5.1
0.46
0.26
-4.4
-1.5
-3.2
-4.0
-1.8
-3.0
-3.2
1.29
0.65
0.63
-14.9
-13.0
-12.9
-11.8
-13.2
-13.0
1.29
0.74
-10.9
-4.6
-7.9
-9.2
-6.4
-7.8
-7.9
2.44
1.22
0.61
-6.6
-5.8
-5.8
-5.3
-5.9
-5.8
0.54
0.31
-4.8
-2.0
-3.6
-6.0
-2.7
-3.8
-3.6
1.61
0.80
0.62
-17.4
-14.6
-14.9
-13.5
-15.1
-14.8
1.65
0.85
-12.3
-6.7
-9.3
-13.5
-8.9
-10.1
-9.3
2.74
1.37
0.62
-5.4
-4.6
-4.8
-4.3
-4.8
-4.7
0.46
0.27
-4.3
-1.7
-2.5
-4.3
-2.1
-3.0
-2.5
1.24
0.62
0.53
-12.8
-101.7
-10.9
-9.7
-11.0
-10.8
1.29
0.75
-9.3
-5.2
-6.6
-8.2
-6.7
-7.2
-6.7
1.58
0.79
0.62
-5.5
-4.9
-4.8
-4.4
-4.9
-4.9
0.45
0.26
-4.1
-1.7
-2.8
-4.3
-2.0
-3.0
-2.8
1.19
0.59
0.58
-13.5
-11.8
-11.6
-10.6
-11.9
-11.7
1.20
0.70
-9.7
-5.1
-7.2
-9.2
-6.9
-7.6
-7.2
1.86
0.93
0.62
Sources of data: MODIS_G, MISR_G, MO_GO, MO_MI_GO (Yu et al, 2004,190926: 2006, 156173), GOCART (Chin et al, 2002, 189996: Yu et al, 2004, 190926),
SPRINTARS (Takemura et al, 2002,190438), GISS (Koch and Hansen, 2005,190183: Koch et al, 2006,190184), LMDZ-INCA (Balkanski et al, 2007,189979: Kinne et al,
2006,155903: Schulz et al, 2006,190381), LMDZ-LOA (Reddy et al, 2005,190207: Reddy et al, 2005,190208). Mean, median, standard deviation (a), and standard error (E)
are calculated for observations (Obs) and model simulations (Mod) separately. The last row is the ratio of model median to observational median. (Taken from Yu et al, 2006,
156173).
The above analyses show that, on a global average, the measurement-based estimates of
DRF are 55-80% greater than the model-based estimates. The differences are even larger on
regional scales. Such measurement-model differences are a combination of differences in
aerosol amount (optical depth), single-scattering properties, surface albedo, and radiative
transfer schemes (Yu et al., 2006, 156173). As discussed earlier, MODIS retrieved optical
depths tend to be overestimated by about 10-15% due to the contamination of thin cirrus and
clouds in general (Kaufman et al., 2005, 155891). Such overestimation of optical depth would
result in a comparable overestimate of the aerosol direct radiative forcing. Other satellite AOD
data may have similar contamination, which however has not yet been quantified. On the other
hand, the observations may be measuring enhanced AOD and DRF due to processes not well
represented in the models including humidification and enhancement of aerosols in the vicinity
of clouds (Koren et al., 2007, 190192).
From the perspective of model simulations, uncertainties associated with
parameterizations of various aerosol processes and meteorological fields, as documented under
the AEROCOM and Global Modeling Initiative (GMI) frameworks (Kinne et al., 2006,
155903: Liu et al., 2007, 190427: Textor et al., 2006,190456), contribute to the large
measurement-model and model-model discrepancies. Factors determining the AOD should be
major reasons for the DRF discrepancy and the constraint of model AOD with well evaluated
and bias reduced satellite AOD through a data assimilation approach can reduce the DRF
discrepancy significantly. Other factors (such as model parameterization of surface reflectance,
and model-satellite differences in single-scattering albedo and asymmetry factor due to
satellite sampling bias toward cloud-free conditions) should also contribute, as evidenced by
the existence of a large discrepancy in the radiative efficiency (Yu et al., 2006, 156173).
Significant effort will be needed in the future to conduct comprehensive assessments.
December 2009 9-108
-------
9.3.3.4. Satellite-Based Estimates of Anthropogenic Component of Aerosol Direct
Radiative Forcing
Satellite instruments do not measure the aerosol chemical composition needed to
discriminate anthropogenic from natural aerosol components. Because anthropogenic aerosols
are predominantly sub-micron, the fine-mode fraction derived from POLDER, MODIS, or
MISR might be used as a tool for deriving anthropogenic aerosol optical depth. This could
provide a feasible way to conduct measurement-based estimates of anthropogenic component
of aerosol direct radiative forcing (Kaufman et al., 2002, 190956). Such method derives
anthropogenic AOD from satellite measurements by empirically correcting contributions of
natural sources (dust and maritime aerosol) to the sub-micron AOD (Kaufman et al., 2005,
155891). The MODIS-based estimate of anthropogenic AOD is about 0.033 over oceans,
consistent with model assessments of 0.030—0.036 even though the total AOD from MODIS is
25-40% higher than the models (Kaufman et al., 2005, 155891). This accounts for 21 ± 7% of
the MODIS-observed total aerosol optical depth, compared with about 33% of anthropogenic
contributions estimated by the models. The anthropogenic fraction of AOD should be much
larger over land (i.e., 47 ± 9% from a composite of several models) (Bellouin et al., 2005,
155684), comparable to the 40% estimated by Yu et al. (2006, 156173). Similarly, the non-
spherical fraction from MISR or POLDER can be used to separate dust from spherical aerosol
(Kahn et al., 2001, 190969; Kalashnikova and Kahn, 2006, 190962), providing another
constraint for distinguishing anthropogenic from natural aerosols.
There have been several estimates of anthropogenic component of DRF in recent years.
Table 9-10 lists such estimates of anthropogenic component of TOADRF that are from model
simulations (Schulz et al., 2006, 190381) and constrained to some degree by satellite
observations (Bellouin et al., 2005, 155684: Bellouin et al., 2008, 189999: Christopher et al.,
2006, 155729: Chung et al., 2005, 155733: Kaufman et al., 2005, 155891: Matsui et al., 2006,
190495: Quaas et al., 2008, 190916: Yu et al., 2006, 156173: Zhao et al., 2008, 190936). The
satellite-based clear-sky DRF by anthropogenic aerosols is estimated to be -1.1 ±0.37 W/m2
over ocean, about a factor of 2 stronger than model simulated -0.6 W/m2. Similar DRF
estimates are rare over land, but a few studies do suggest that the anthropogenic DRF over
land is much more negative than that over ocean (Bellouin et al., 2005, 155684: Bellouin et al.,
2008, 189999: Yu et al., 2006, 156173). On global average, the measurement-based estimate of
anthropogenic DRF ranges from -0.9 to -1.9 W/m2, again stronger than the model-based
estimate of-0.8 W/m2. Similar to DRF estimates for total aerosols, satellite-based estimates of
anthropogenic component of DRF are rare over land.
On global average, anthropogenic aerosols are generally more absorptive than natural
aerosols. As such the anthropogenic component of DRF is much more negative at the surface
than at TOA. Several observation-constrained studies estimate that the global average, clear-
sky, anthropogenic component of DRF at the surface ranges from -4.2 to -5.1 W/m (Bellouin
et al., 2005, 155684: Chung et al., 2005, 155733: Matsui et al., 2006, 190495: Yu et al., 2004,
190926), which is about a factor of 2 larger in magnitude than the model estimates (e.g.,
Reddy et al., 2005, 190208).
Uncertainties in estimates of the anthropogenic component of aerosol DRF are greater
than for the total aerosol, particularly over land. An uncertainty analysis (Yu et al., 2006,
156173) partitions the uncertainty for the global average anthropogenic DRF between land and
ocean more or less evenly. Five parameters, namely fine-mode fraction (ff) and anthropogenic
fraction of fine-mode fraction (faf) over both land and ocean, and T over ocean, contribute
nearly 80% of the overall uncertainty in the anthropogenic DRF estimate, with individual
shares ranging from 13-20% (Yu et al., 2006, 156173). These uncertainties presumably
represent a lower bound because the sources of error are assumed to be independent.
Uncertainties associated with several parameters are also not well defined. Nevertheless, such
uncertainty analysis is useful for guiding future research and documenting advances in
understanding.
December 2009 9-109
-------
Table 9-10. Estimates of anthropogenic components of aerosol optical depth (Tant) and clear-sky
DRF at the TOA from model simulations.
Ocean
Land
Global
Estimated uncertainty or
Tan.
Kaufman et al. (2005. 155891) 0.033
Bellouin et al. (2005, 155684) 0.028
Chung etal. (2005. 155733)
Yu et al. (2006, 156173) 0.031
Christopher et al. (2006,
155729)
Matsui and Pielke (2006,
190495)
Quaas etal. (2008. 190916)
Bellouin et al. (2008, 1899991 0.021
Zhao etal. (2008, 190936)
Schulz et al. (2006, 190381J 0-022
DRF T
(W/m2) lant
-1.4
-0.8 0.13
-1.1 -.088
-1.4
-1.6
-0.7
-0.6 0.107
-1.25
-0.59 0.065
model diversity tor UKh
DRF -.- DRF
(W/m2) lant (W/m2)
30%
0.062 -1.9 15%
-1.1
,, a nn/|o 1, 47% (ocean), 845 (land), and
"1'a U'U4B "1-J 62% (global)
65%
30°S-30°N oceans
-1.8 -0.9 45%
Update to Bellouin etal. (2005,
-3.3 0.043 -1.3 1556841 with MODIS Collection 5
data
35%
-114 0036 -077 30-40%; same emissions
prescribed for all models
Sources: Schulz et al., (2006,1903811 approaches constrained by satellite observations, Kaufman et al. (2005,155891
156173): Christopher et al. (2006,1557291: Matsui and Pielke (2006,1904951: Quaas et al. (2008,1909161: Zhao et al.
I; Bellouin et al. (2005,1556841 2008; Chung et al. (2005,1557331: Yu et al. (2006,
(2008, 1909361.
9.3.3.5. Aerosol-Cloud Interactions and Indirect Forcing
Satellite views of the Earth show a planet whose albedo is dominated by dark oceans and
vegetated surfaces, white clouds, and bright deserts. The bright white clouds overlying darker
oceans or vegetated surface demonstrate the significant effect that clouds have on the Earth's
radiative balance. Low clouds reflect incoming sunlight back to space, acting to cool the
planet, whereas high clouds can trap outgoing terrestrial radiation and act to warm the planet.
In the Arctic, low clouds have also been shown to warm the surface (Garrett and Zhao, 2006,
190570). Changes in cloud cover, in cloud vertical development, and cloud optical properties
will have strong radiative and therefore, climatic impacts. Furthermore, factors that change
cloud development will also change precipitation processes. These changes may alter amounts,
locations and intensities of local and regional rain and snowfall, creating droughts, floods and
severe weather.
Cloud droplets form on a subset of aerosol particles called cloud condensation nuclei
(CCN). In general, an increase in aerosol leads to an increase in CCN and an increase in drop
concentration. Thus, for the same amount of liquid water in a cloud, more available CCN will
result in a greater number but smaller size of droplets (Twomey, 1977, 190533). A cloud with
smaller but more numerous droplets will be brighter and reflect more sunlight to space, thus
exerting a cooling effect. This is the first aerosol indirect radiative effect, or "albedo effect."
The effectiveness of a particle as a CCN depends on its size and composition so that the degree
to which clouds become brighter for a given aerosol perturbation, and therefore the extent of
cooling, depends on the aerosol size distribution and its size-dependent composition. In
addition, aerosol perturbations to cloud microphysics may involve feedbacks; for example,
smaller drops are less likely to collide and coalesce; this will inhibit growth, suppressing
precipitation, and possibly increasing cloud lifetime (Albrecht, 1989, 045783). In this case
clouds may exert an even stronger cooling effect.
A distinctly different aerosol effect on clouds exists in thin Arctic clouds (LWP <25 gm"2)
having low emissivity. Aerosol has been shown to increase the longwave emissivity in these
clouds, thereby warming the surface (Garrett and Zhao, 2006, 190570; Lubin and Vogelmann,
2006, 190466).
Some aerosol particles, particularly black carbon and dust, also act as ice nuclei (IN) and
in so doing, modify the microphysical properties of mixed-phase and ice-clouds. An increase
in IN will generate more ice crystals, which grow at the expense of water droplets due to the
December 2009
9-110
-------
difference in vapor pressure over ice and water surfaces. The efficient growth of ice particles
may increase the precipitation efficiency. In deep convective, polluted clouds there is a delay
in the onset of freezing because droplets are smaller. These clouds may eventually precipitate,
but only after higher altitudes are reached that result in taller cloud tops, more lightning and
greater chance of severe weather (Andreae et al., 2004, 155658; Rosenfeld and Lansky, 1998,
190230). The present state of knowledge of the nature and abundance of IN, and ice formation
in clouds is extremely poor. There is some observational evidence of aerosol influences on ice
processes, but a clear link between aerosol, IN concentrations, ice crystal concentrations and
growth to precipitation has not been established. This report therefore only peripherally
addresses ice processes. More information can be found in a review by the WMO/IUGG
International Aerosol-Precipitation Scientific Assessment (Levin and Cotton, 2008, 190375).
In addition to their roles as CCN and IN, aerosols also absorb and scatter light, and
therefore they can change atmospheric conditions (temperature, stability, and surface fluxes)
that influence cloud development and properties (Ackerman et al., 2000, 002987: Hansen et
al., 1997, 043104). Thus, aerosols affect clouds through changing cloud droplet size
distributions, cloud particle phase, and by changing the atmospheric environment of the cloud.
9.3.3.6. Remote Sensing of Aerosol-Cloud Interactions and Indirect Forcing
The AVHRR satellite instruments have observed relationships between columnar aerosol
loading, retrieved cloud microphysics, and cloud brightness over the Amazon Basin that are
consistent with the theories explained above (Feingold et al., 2001, 190544: Kaufman and
Fraser, 1997, 190958: Kaufman and Nakajima, 1993, 190959), but do not necessarily prove a
causal relationship. Other studies have linked cloud and aerosol microphysical parameters or
cloud albedo and droplet size using satellite data applied over the entire global oceans (Han et
al., 1998, 190594: Nakajima et al., 2001,190552: Wetzel and Stowe, 1999, 190636). Using
these correlations with estimates of aerosol increase from the pre-industrial era, estimates of
anthropogenic aerosol indirect radiative forcing fall into the range of-0.7 to -1.7 W/m2
(Nakajima et al., 2001,190552).
Introduction of the more modern instruments (POLDER and MODIS) has allowed more
detailed observations of relationships between aerosol and cloud parameters. Cloud cover can
both decrease and increase with increasing aerosol loading (Kaufman et al., 2005, 155891:
Koren et al., 2004, 190187: Koren et al., 2005, 190188: Matheson et al., 2005, 190494:
Sekiguchi et al., 2003, 190385: Yu et al., 2007, 093173). The same is true of LWP (Han et al.,
2002, 049181: Matsui et al., 2006, 190498). Aerosol absorption appears to be an important
factor in determining how cloud cover will respond to increased aerosol loading (Jiang and
Feingold, 2006, 190976: Kaufman and Koren, 2006, 190951: Koren et al., 2008, 190193).
Different responses of cloud cover to increased aerosol could also be correlated with
atmospheric thermodynamic and moisture structure (Yu et al., 2007, 093173). Observations in
the MODIS data show that aerosol loading correlates with enhanced convection and greater
production of ice anvils in the summer Atlantic Ocean (Koren et al., 2005, 190188), which
conflicts with previous results that used AVHRR and could not isolate convective systems
from shallow clouds (Sekiguchi et al., 2003, 190385).
In recent years, surface-based remote sensing has also been applied to address aerosol
effects on cloud microphysics. This method offers some interesting insights, and is
complementary to the global satellite view. Surface remote sensing can only be applied at a
limited number of locations, and therefore lacks the global satellite view. However, these
surface stations yield high temporal resolution data and because they sample aerosol below,
rather than adjacent to clouds they do not suffer from "cloud contamination." With the
appropriate instrumentation (lidar) they can measure the local aerosol entering the clouds,
rather than a column-integrated aerosol optical depth. Under well-mixed conditions, surface in
situ aerosol measurements can be used. Surface remote-sensing studies are discussed in more
detail below, although the main science issues are common to satellite remote sensing.
Feingold et al. (2001, 190544) used data collected at the ARM Southern Great Plains
(SGP) site to allow simultaneous retrieval of aerosol and cloud properties. A combination of a
Doppler cloud radar and a microwave radiometer was used to retrieve cloud drop effective
radius re profiles in non-precipitating (radar reflectivity Z <-17 dBZ), ice-free clouds.
Simultaneously, sub-cloud aerosol extinction profiles were measured with a lidar to quantify
the response of drop sizes to changes in aerosol properties. Cloud data were binned according
to liquid water path (LWP) as measured with a microwave radiometer, consistent with
Twomey's (1977, 190533) conceptual view of the aerosol impact on cloud microphysics. With
high temporal/spatial resolution data (on the order of 20s or 1 OOs of meters), realizations of
aerosol-cloud interactions at the large eddy scale were obtained, and quantified in terms of the
relative decrease in re in response to a relative increase in aerosol extinction (din re/din
extinction), as shown in Figure 9-70. Examining the dependence in this way reduces reliance
on absolute measures of cloud and aerosol parameters and minimizes sensitivity to
measurement error, provided errors are unbiased. This formulation permitted these responses
to be related to cloud microphysical theory. Restricting the examination to updrafts only (as
determined from the radar Doppler signal) permitted examination of the role of updraft in
determining the response of re to changes in aerosol (via changes in drop number
December 2009 9-111
-------
concentration Nd). Analysis of data from 7 days showed that turbulence intensifies the aerosol
impact on cloud microphysics.
In addition to radar/microwave radiometer retrievals of aerosol and cloud properties,
measurements of cloud optical depth by surface based radiometers such as the MFRSR
(Michalsky et al., 2001, 190537) have been used in combination with measurements of cloud
LWP by microwave radiometer to measure an average value of re during daylight when the
solar elevation angle is sufficiently high (Min and Harrison, 1996, 190538). Using this
retrieval, Kim et al. (2003, 155899) performed analyses of the re response to changes in
aerosol at the same continental site, using a surface measurement of the aerosol light scattering
coefficient instead of using extinction near cloud base as a proxy for CCN. Variance in LWP
was shown to explain most of the variance in cloud optical depth, exacerbating detection of an
aerosol effect. Although a decrease in re was observed with increasing scattering coefficient,
the relation was not strong, indicative of other influences on re and/or decoupling between the
surface and cloud layer. A similar study was conducted by Garrett et al. (2004, 190568) at a
location in the Arctic.
E
=1
April 3 1998
IE
0.07 • LWP: 100-1 10 g m"
- 0.09 •LWP: 110-121 g m"
0.09 eLWP: 121-133 g m'
i i i i i i i
.00
Source: Adapted with Permission of the American Geophysical Union from Feingold et al. (2003, 190551).
Figure 9-70.
Scatter plots showing mean cloud drop effective radius (re) versus aerosol
extinction coefficient (unit: km-1) for various liquid water path (LWP) bands on
April 3, 1998 at ARM SGP site.
They suggested that summertime Arctic clouds are more sensitive to aerosol perturbations
than clouds at lower latitudes. The advantage of the MFRSR/microwave radiometer
combination is that it derives re from cloud optical depth and LWP and it is not as sensitive to
large drops as the radar is. A limitation is that it can be applied only to clouds with extensive
horizontal cover during daylight hours.
More recent data analyses by Feingold et al. (2003, 190551X Kim et al. (2008, 130785)
and McComiskey et al. (2008, 190525) at a variety of locations, and modeling work (Feingold,
2003, 190547) have investigated (i) the use of different proxies for cloud condensation nuclei,
such as the light scattering coefficient and aerosol index; (ii) sensitivity of cloud
microphysical/optical properties to controlling factors such as aerosol size distribution,
entrainment, LWP, and updraft velocity; (iii) the effect of optical- as opposed to radar-
retrievals of drop size; and (iv) spatial heterogeneity. These studies have reinforced the
importance of LWP and vertical velocity as controlling parameters. They have also begun to
reconcile the reasons for the large discrepancies between various approaches, and platforms
(satellite, aircraft in situ, and surface-based remote sensing). These investigations are
important because satellite measurements that use a similar approach are being employed in
GCMs to represent the albedo indirect effect (Quaas and Boucher, 2005, 190573). In fact, the
weakest albedo indirect effect in IPCC (2007, 092765) derives from satellite measurements
that have very weak responses of re to changes in aerosol. The relationship between these
aerosol-cloud microphysical responses and cloud radiative forcing has been examined by
December 2009
9-112
-------
McComiskey and Feingold (2008, 190517). They showed that for plane-parallel clouds, a
typical uncertainty in the logarithmic gradient of a re-aerosol relationship of 0.05 results in a
local forcing error of-3 to -10 W/m2, depending on the aerosol perturbation. This sensitivity
reinforces the importance of adequate quantification of aerosol effects on cloud microphysics
to assessment of the radiative forcing, i.e., the indirect effect. Quantification of these effects
from remote sensors is exacerbated by measurement errors. For example, LWP is measured to
an accuracy of 25 g m-2 at best, and since it is the thinnest clouds (i.e., low LWP) that are most
susceptible (from a radiative forcing perspective) to changes in aerosol, this measurement
uncertainty represents a significant uncertainty in whether the observed response is related to
aerosol, or to differences in LWP. The accuracy and spatial resolution of satellite-based LWP
measurements is much poorer and this represents a significant challenge. In some cases
important measurements are simply absent, e.g., updraft is not measured from satellite-based
remote sensors.
Finally, cloud radar data from CloudSat, along with the A-train aerosol data, is providing
great opportunity for inferring aerosol effects on precipitation (e.g., Stephens and Haynes,
2007, 190413). The aerosol effect on precipitation is far more complex than the albedo effect
because the instantaneous view provided by satellites makes it difficult to establish causal
relationships.
9.3.3.7. In Situ Studies of Aerosol-Cloud Interactions
In situ observations of aerosol effects on cloud microphysics date back to the 1950s and
1960s (Brenguier et al, 2000, 189966; Gunn and Phillips, 1957, 190595; Leaitch et al, 1992,
045270: Radke et al., 1989, 156034: Squires, 1958, 045608: Warner, 1968, 157114: Warner
and Twomey, 1967, 045616: to name a few). These studies showed that high concentrations of
CCN from anthropogenic sources, such as industrial pollution or the burning of sugarcane can
increase cloud droplet number concentration Nd, thus increasing cloud microphysical stability
and potentially reducing precipitation efficiency. As in the case of remote sensing studies, the
causal link between aerosol perturbations and cloud microphysical responses (e.g., re or Nd) is
much better established than the relationship between aerosol and changes in cloud fraction,
LWC, and precipitation (see also Levin and Cotton, 2008, 190375).
In situ cloud measurements are usually regarded as "ground truth" for satellite retrievals
but in fact there is considerable uncertainty in measured parameters such liquid water content
(LWC), and size distribution, which forms the basis of other calculations such as drop
concentration, re and extinction. It is not uncommon to see discrepancies in LWC on the order
of 50% between different instruments, and cloud drop size distributions are difficult to
measure, particularly for droplets <10 um where Mie scattering oscillations generate
ambiguities in drop size. Measurement uncertainty in re from in situ probes is assessed, for
horizontally homogeneous clouds, to be on the order of 15- 20%, compared to 10% for
MODIS and 15-20% for other spectral measurements (Feingold et al., 2003, 190551). As with
remote measurements it is prudent to consider relative (as opposed to absolute) changes in
cloud microphysics related to relative changes in aerosol. An added consideration is that in situ
measurements typically represent a very small sample of the atmosphere akin to a thin pencil
line through a large volume. For an aircraft flying at 100 m/s and sampling at 1 Hz, the sample
volume is on the order of 10 cm . The larger spatial sampling of remote sensing has the
advantage of being more representative but it removes small-scale (i.e., sub sampling volume)
variability, and therefore, may obscure important cloud processes.
Measurements at a wide variety of locations around the world have shown that increases
in aerosol concentration lead to increases in Nd. However the rate of this increase is highly
variable and always sub-linear, as exemplified by the compilation of data in Ramanathan et al.
(2001, 042681). This is because, as discussed previously, Nd is a function of numerous
parameters in addition to aerosol number concentration, including size distribution, updraft
velocity (Leaitch et al., 1996, 190354), and composition. In stratocumulus clouds,
characterized by relatively low vertical velocity (and low supersaturation) only a small fraction
of particles can be activated whereas in vigorous cumulus clouds that have high updraft
velocities, a much larger fraction of aerosol particles is activated. Thus the ratio of Nd to
aerosol particle number concentration is highly variable.
In recent years there has been a concerted effort to reconcile measured Nd concentrations
with those calculated based on observed aerosol size and composition, as well as updraft
velocity. These so-called "closure experiments" have demonstrated that on average, agreement
in Nd between these approaches is on the order of 20% (Conant et al., 2004, 190010). This
provides confidence in theoretical understanding of droplet activation, however, measurement
accuracy is not high enough to constrain the aerosol composition effects that have magnitudes
<20%.
One exception to the rule that more aerosol particles result in larger Nd is the case of giant
CCN (sizes on the order of a few microns), which, in concentrations on the order of 1 cm3 (i.e.,
~ 1% of the total concentration) can lead to significant suppression in cloud supersaturation
and reductions in Nd (O'Dowd et al., 1999, 090414). The measurement of these large particles
is difficult and hence the importance of this effect is hard to assess. These same giant CCN, at
concentrations as low as 1/liter, can significantly affect the initiation of precipitation in
December 2009 9-113
-------
moderately polluted clouds (Johnson, 1982, 190973) and in so doing alter cloud albedo
(Feingold et al, 1999, 190540).
The most direct link between the remote sensing of aerosol-cloud interactions discussed
in Section 9.3.3.6 and in situ observations is via observations of relationships between drop
concentration Nd and CCN concentration. Theory shows that if re-CCN relationships are
calculated at constant LWP or LWC, their logarithmic slope is -1/3 that of the Nd-CCN
logarithmic slope (i.e., dlnre/dlnCCN = -1/3 dlnNd/ dlnCCN). In general, Nd-CCN slopes
measured in situ tend to be stronger than equivalent slopes obtained from remote sensing -
particularly in the case of satellite remote sensing (McComiskey and Feingold, 2008,
190517). There are a number of reasons for this: (i) in situ measurements focus on smaller
spatial scales and are more likely to observe the droplet activation process as opposed to
remote sensing that incorporates larger spatial scales and includes other processes such as drop
coalescence that reduce Nd, and therefore the slope of the Nd- CCN relationship
(McComiskey et al., 2008, 190525). (ii) Satellite remote sensing studies typically do not sort
their data by LWP, and this has been shown to reduce the magnitude of the re-CCN response
(Feingold, 2003, 190547).
In conclusion, observational estimates of aerosol indirect radiative forcings are still in
their infancy. Effects on cloud microphysics that result in cloud brightening have to be
considered along with effects on cloud lifetime, cover, vertical development and ice
production. For in situ measurements, aerosol effects on cloud microphysics are reasonably
consistent (within ~ 20%) with theory but measurement uncertainties in remote sensing of
aerosol effects on clouds, as well as complexity associated with three-dimensional radiative
transfer, result in considerable uncertainty in radiative forcing. The higher order indirect
effects are poorly understood and even the sign of the microphysical response and forcing may
not always be the same. Aerosol type and specifically the absorption properties of the aerosol
may cause different cloud responses. Early estimates of observationally based aerosol indirect
forcing range from -0.7 to -1.7 W/m2 (Nakajima et al., 2001,190552) and -0.6 to -1.2 W/m2
(Sekiguchi et al., 2003, 190385), depending on the estimate for aerosol increase from pre-
industrial times and whether aerosol effects on cloud fraction are also included in the estimate.
9.3.4. Outstanding Issues
Despite substantial progress, as summarized in Sections 9.3.2 and 9.3.3, most
measurement-based studies so far have concentrated on influences produced by the sum of
natural and anthropogenic aerosols on solar radiation under clear sky conditions. Important
issues remain:
• Because accurate measurements of aerosol absorption are lacking and land surface
reflection values are uncertain, DRF estimates over land and at the ocean surface are
less well constrained than the estimate of TOA DRF over ocean.
• Current estimates of the anthropogenic component of aerosol direct radiative forcing
have large uncertainties, especially over land.
• Because there are very few measurements of aerosol absorption vertical distribution,
mainly from aircraft during field campaigns, estimates of direct radiative forcing of
above-cloud aerosols and profiles of atmospheric radiative heating induced by
aerosol absorption are poorly constrained.
• There is a need to quantify aerosol impacts on thermal infrared radiation, especially
for dust.
• The diurnal cycle of aerosol direct radiative forcing cannot be adequately
characterized with currently available, sun-synchronous, polar orbiting satellite
measurements.
• Measuring aerosol, cloud, and ambient meteorology contributions to indirect
radiative forcing remains a major challenge.
• Long-term aerosol trends and their relationship to observed surface solar radiation
changes are not well understood.
The current status and prospects for these areas are briefly discussed below.
Measuring Aerosol Absorption and Single-Scattering Albedo
Currently, the accuracy of both in situ and remote sensing aerosol SSA measurements is
generally ± 0.03 at best, which implies that the inferred accuracy of clear sky aerosol DRF
would be larger than 1 W/m2 (see Chapter 1 of the CSSP SAP2.3). Recently developed
photoacoustic (Arnott et al., 1999, 020650) and cavity ring down extinction cell (Strawa et al.,
2002, 190421) techniques for measuring aerosol absorption produce SSA with improved
accuracy over previous methods. However, these methods are still experimental, and must be
deployed on aircraft. Aerosol absorption retrievals from satellites using the UV-technique have
large uncertainties associated with its sensitivity to the height of the aerosol layer(s) (Torres et
al., 2005, 190507), and it is unclear how the UV results can be extended to visible
wavelengths. Views in and out of sunglint can be used to retrieve total aerosol extinction and
scattering, respectively, thus constraining aerosol absorption over oceans (Kaufman et al.,
2002, 190955). However, this technique requires retrievals of aerosol scattering properties,
December 2009 9-114
-------
including the real part of the refractive index, well beyond what has so far been demonstrated
from space. In summary, there is a need to pursue a better understanding of the uncertainty in
SSA from both in situ measurements and remote sensing retrievals and, with this knowledge,
to synthesize different data sets to yield a characterization of aerosol absorption with well-
defined uncertainty (Leahy et al., 2007, 190232). Laboratory studies of aerosol absorption of
specific known composition are also needed to interpret in situ measurements and remote
sensing retrievals and to provide updated database of particle absorbing properties for models.
Estimating the Aerosol Direct Radiative Forcing over Land
Land surface reflection is large, heterogeneous, and anisotropic, which complicates
aerosol retrievals and DRF determination from satellites. Currently, the aerosol retrievals over
land have relatively lower accuracy than those over ocean (Section 9.3.2.5) and satellite data
are rarely used alone for estimating DRF over land (Section 9.3.3). Several issues need to be
addressed, such as developing appropriate angular models for aerosols over land (Patadia et
al., 2008, 190558) and improving land surface reflectance characterization. MODIS and MISR
measure land surface reflection wavelength dependence and angular distribution at high
resolution (Martonchik et al., 1998, 190484; Martonchik et al., 2002, 190490; Moody et al.,
2005, 190548). This offers a promising opportunity for inferring the aerosol direct radiative
forcing over land from satellite measurements of radiative fluxes (e.g., CERES) and from
critical reflectance techniques (Fraser and Kaufman, 1985, 190567: Kaufman, 1987, 190960).
The aerosol direct radiative forcing over land depends strongly on aerosol absorption and
improved measurements of aerosol absorption are required.
Distinguishing Anthropogenic from Natural Aerosols
Current estimates of anthropogenic components of AOD and direct radiative forcing have
larger uncertainties than total aerosol optical depth and direct radiative forcing, particularly
over land (see Section 9.3.3.4), because of relatively large uncertainties in the retrieved aerosol
microphysical properties (see Section 9.3.2). Future measurements should focus on improved
retrievals of such aerosol properties as size distribution, particle shape, and absorption, along
with algorithm refinement for better aerosol optical depth retrievals. Coordinated in situ
measurements offer a promising avenue for validating and refining satellite identification of
anthropogenic aerosols (Anderson et al., 2005, 189993: 2005, 189991). For satellite-based
aerosol type characterization, it is sometimes assumed that all biomass-burning aerosol is
anthropogenic and all dust aerosol is natural (Kaufman et al., 2005, 155891). The better
determination of anthropogenic aerosols requires a quantification of biomass burning ignited
by lightning (natural origin) and mineral dust due to human induced changes of land
cover/land use and climate (anthropogenic origin). Improved emissions inventories and better
integration of satellite observations with models seem likely to reduce the uncertainties in
aerosol source attribution.
Profiling the Vertical Distributions of Aerosols
Current aerosol profile data are far from adequate for quantifying the aerosol radiative
forcing and atmospheric response to the forcing. The data have limited spatial and temporal
coverage, even for current spaceborne lidar measurements. Retrieving aerosol extinction
profile from lidar measured attenuated backscatter is subject to large uncertainties resulting
from aerosol type characterization. Current space-borne Lidar measurements are also not
sensitive to aerosol absorption. Because of lack of aerosol vertical distribution observations,
the estimates of DRF in cloudy conditions and dust DRF in the thermal infrared remain highly
uncertain (Lubin et al., 2002, 190463: Schulz et al., 2006, 190381: Sokolik et al., 2001,
190404). It also remains challenging to constrain the aerosol-induced atmospheric heating rate
increment that is essential for assessing atmospheric responses to the aerosol radiative forcing
(e.g., Feingold et al., 2005, 190550: Lau et al., 2006, 190223: Yu et al., 2002, 190923).
Progress in the foreseeable future is likely to come from (1) better use of existing, global,
space-based backscatter lidar data to constrain model simulations, and (2) deployment of new
instruments, such as high-spectral-resolution lidar (HSRL), capable of retrieving both
extinction and backscatter from space. The HSRL lidar system will be deployed on the
EarthCARE satellite mission tentatively scheduled for 2013
Chttp://asimov/esrin.esi.it/esaLP/ASESMYNW9SC_Lpearthcare_l .html).
Characterizing the Diurnal Cycle of Aerosol Direct Radiative Forcing
The diurnal variability of aerosol can be large, depending on location and aerosol type
(Smirnov et al., 2002, 190398), especially in wildfire situations, and in places where boundary
layer aerosols hydrate or otherwise change significantly during the day. This cannot be
captured by currently available, sun-synchronous, polar orbiting satellites. Geostationary
satellites provide adequate time resolution (Christopher and Zhang, 2002, 190031: Wang et
al., 2003, 157106), but lack the information required to characterize aerosol types. Aerosol
December 2009 9-115
-------
type information from low earth orbit satellites can help improve accuracy of geostationary
satellite aerosol retrievals (Costa et al, 2004, 190006: 2004, 192022). For estimating the
diurnal cycle of aerosol DRF, additional efforts are needed to adequately characterize the
anisotropy of surface reflection (Yu et al., 2004, 190926) and daytime variation of clouds.
Studying Aerosol-Cloud Interactions and Indirect Radiative Forcing
Remote sensing estimates of aerosol indirect forcing are still rare and uncertain.
Improvements are needed for both aerosol characterization and measurements of cloud
properties, precipitation, water vapor, and temperature profiles. Basic processes still need to be
understood on regional and global scales. Remote sensing observations of aerosol-cloud
interactions and aerosol indirect forcing are for the most part based on simple correlations
among variables, from which cause-and-effects cannot be deduced. One difficulty in inferring
aerosol effects on clouds from the observed relationships is separating aerosol from
meteorological effects, as aerosol loading itself is often correlated with the meteorology. In
addition, there are systematic errors and biases in satellite aerosol retrievals for partly cloud-
filled scenes. Stratifying aerosol and cloud data by liquid water content, a key step in
quantifying the albedo (or first) indirect effect, is usually missing. Future work will need to
combine satellite observations with in situ validation and modeling interpretation. A
methodology for integrating observations (in situ and remote) and models at the range of
relevant temporal/spatial scales is crucial to improve understanding of aerosol indirect effects
and aerosol-cloud interactions.
Quantifying Long-Term Trends of Aerosols at Regional Scales
Because secular changes are subtle and are superposed on seasonal and other natural
variability, this requires the construction of consistent, multi-decadal records of climate-quality
data. To be meaningful, aerosol trend analysis must be performed on a regional basis. Long-
term trends of aerosol optical depth have been studied using measurements from surface
remote sensing stations (e.g., Augustine et al., 2008, 189913: Hoyt and Frohlich, 1983,
190621: Luo et al., 2001, 190467) and historic satellite sensors (Massie et al., 2004, 190492:
Mishchenko and Geogdzhayev, 2007, 190545: Mishchenko et al., 2007, 190542: Zhao et al.,
2008, 190935). An emerging multiyear climatology of high quality AOD data from modern
satellite sensors (e.g., Kahn et al., 2005, 190966: Remer et al., 2008, 190224) has been used to
examine the interannual variations of aerosol (e.g., Koren et al., 2007, 190189: Mishchenko
and Geogdzhayev, 2007, 190545)
and contribute significantly to the study of aerosol trends. Current observational
capability needs to be continued to avoid any data gaps. A synergy of aerosol products from
historical, modern and future sensors is needed to construct as long a record as possible. Such
a data synergy can build upon understanding and reconciliation of AOD differences among
different sensors or platforms (Jeong et al., 2005, 190977). This requires overlapping data
records for multiple sensors. A close examination of relevant issues associated with individual
sensors is urgently needed, including sensor calibration, algorithm assumptions, cloud
screening, data sampling and aggregation, among others.
Linking Aerosol Long-Term Trends with Changes of Surface Solar Radiation
Analysis of the long-term surface solar radiation record suggests significant trends during
past decades (e.g., Alpert et al., 2005, 190047: Pinker et al., 2005, 190569: Stanhill and
Cohen, 2001, 042121: Wild et al., 2005, 156156). Although a significant and widespread
decline in surface total solar radiation (the sum of direct and diffuse irradiance) occurred up to
1990 (so-called solar dimming), a sustained increase has been observed during the subsequent
decade. Speculation suggests that such trends result from decadal changes of aerosols and the
interplay of aerosol direct and indirect radiative forcing (Norris and Wild, 2007, 190555:
Ruckstuhl et al., 2008, 190356: Stanhill and Cohen, 2001, 042121: Streets et al., 2006,
190425: Wild et al., 2005, 156156). However, reliable observations of aerosol trends are
required to test these ideas. In addition to aerosol optical depth, changes in aerosol
composition must also be quantified, to account for changing industrial practices,
environmental regulations, and biomass burning emissions (Novakov et al., 2003, 048398:
Streets and Aunan, 2005, 156106: Streets et al., 2004, 190423). Such compositional changes
will affect the aerosol SSA and size distribution, which in turn will affect the surface solar
radiation (e.g., Qian et al., 2007, 190572). However, such data are currently rare and subject to
large uncertainties. Finally, a better understanding of aerosol-radiation-cloud interactions and
trends in cloudiness, cloud albedo, and surface albedo is badly needed to attribute the observed
radiation changes to aerosol changes with less ambiguity.
December 2009 9-116
-------
9.3.5. Concluding Remarks
Since the concept of aerosol-radiation-climate interactions was first proposed around
1970, substantial progress has been made in determining the mechanisms and magnitudes of
these interactions, particularly in the last 10 years. Such progress has greatly benefited from
significant improvements in aerosol measurements and increasing sophistication of model
simulations. As a result, knowledge of aerosol properties and their interaction with solar
radiation on regional and global scales is much improved. Such progress plays a unique role in
the definitive assessment of the global anthropogenic radiative forcing, as "virtually certainly
positive" in IPCC AR4 (Haywood and Schulz, 2007, 190600).
In Situ Measurements of Aerosols
New in situ instruments such as aerosol mass spectrometers, photoacoustic techniques,
and cavity ring down cells provide high accuracy and fast time resolution measurements of
aerosol chemical and optical properties. Numerous focused field campaigns and the emerging
ground-based aerosol networks are improving regional aerosol chemical, microphysical, and
radiative property characterization. Aerosol closure studies of different measurements indicate
that measurements of submicrometer, spherical sulfate and carbonaceous particles have a
much better accuracy than that for dust-dominated aerosol. The accumulated comprehensive
data sets of regional aerosol properties provide a rigorous "test bed" and strong constraint for
satellite retrievals and model simulations of aerosols and their direct radiative forcing.
Remote Sensing Measurements of Aerosols
Surface networks, covering various aerosol regimes around the globe, have been
measuring aerosol optical depth with an accuracy of 0.01-0.02, which is adequate for
achieving the accuracy of 1 W/m2 for cloud-free TOA DRF. On the other hand, aerosol
microphysical properties retrieved from these networks, especially SSA, have relatively large
uncertainties and are only available in very limited conditions. Current satellite sensors can
measure AOD with an accuracy of about 0.05 or 15-20% in most cases. The implementation of
multi-wavelength, multi-angle, and polarization measuring capabilities has also made it
possible to measure particle properties (size, shape, and absorption) that are essential for
characterizing aerosol type and estimating anthropogenic component of aerosols. However,
these microphysical measurements are more uncertain than AOD measurements.
Observational Estimates of Clear-Sky Aerosol Direct Radiative Forcing
Closure studies based on focused field experiments reveal DRF uncertainties of about
25% for sulfate/carbonaceous aerosol and 60% for dust at regional scales. The high-accuracy
of MODIS, MISR and POLDER aerosol products and broadband flux measurements from
CERES make it feasible to obtain observational constraints for aerosol TOA DRF at a global
scale, with relaxed requirements for measuring particle microphysical properties. Major
conclusions from the assessment are:
• A number of satellite-based approaches consistently estimate the clear-sky diurnally
averaged TOA DRF (on solar radiation) to be about -5.5 ± 0.2 W/m2 (mean ±
standard error from various methods) over global ocean. At the ocean surface, the
diurnally averaged DRF is estimated to be -8.7 ± 0.7 W/m2. These values are
calculated for the difference between today's measured total aerosol (natural plus
anthropogenic) and the absence of all aerosol.
• Overall, in comparison to that over ocean, the DRF estimates over land are more
poorly constrained by observations and have larger uncertainties. A few satellite
retrieval and satellite-model integration yield the overland clear-sky diurnally
averaged DRF of -4.9 ± 0.7 W/m2 and -11.8 ± 1.9 W/m2 at the TOA and surface,
respectively. These values over land are calculated for the difference between total
aerosol and the complete absence of all aerosol.
• Use of satellite measurements of aerosol microphysical properties yields that on a
global ocean average, about 20% of AOD is contributed by human activities and the
clear-sky TOA DRF by anthropogenic aerosols is -1.1 ± 0.4 W/m . Similar DRF
estimates are rare over land, but a few measurement-model integrated studies do
suggest much more negative DRF over land than over ocean.
• These satellite-based DRF estimates are much greater than the model-based
estimates, with differences much larger at regional scales than at a global scale.
Measurements of Aerosol-Cloud Interactions and Indirect Radiative Forcing
In situ measurement of cloud properties and aerosol effects on cloud microphysics suggest
that theoretical understanding of the activation process for water cloud is reasonably well-
understood. Remote sensing of aerosol effects on droplet size associated with the albedo effect
tends to underestimate the magnitude of the response compared to in situ measurements.
December 2009 9-117
-------
Recent efforts trace this to a combination of lack of stratification of data by cloud water, the
relatively large spatial scale over which measurements are averaged (which includes variability
in cloud fields, and processes that obscure the aerosol-cloud processes), as well as
measurement uncertainties (particularly in broken cloud fields). It remains a major challenge
to infer aerosol number concentrations from satellite measurements. The present state of
knowledge of the nature and abundance of IN, and ice formation in clouds is extremely poor.
Despite the substantial progress in recent decades, several important issues remain, such
as measurements of aerosol size distribution, particle shape, absorption, and vertical profiles,
and the detection of aerosol long-term trend and establishment of its connection with the
observed trends of solar radiation reaching the surface, as discussed in Section 9.3.4.
Furthering the understanding of aerosol impacts on climate requires a coordinated research
strategy to improve the measurement accuracy and use the measurements to validate and
effectively constrain model simulations. Concepts of future research in measurements are
discussed in Chapter 4 "Way Forward" (of the CCSP SAP2.3).
9.3.6. Modeling the Effect of Aerosols on Climate
9.3.6.1. Introduction
The IPCC Fourth Assessment Report (AR4) (IPCC, 2007, 092765) concludes that man's
influence on the warming climate is in the category of "very likely". This conclusion is based
on, among other things, the ability of models to simulate the global and, to some extent,
regional variations of temperature over the past 50-100 years. When anthropogenic effects are
included, the simulations can reproduce the observed warming (primarily for the past 50
years); when they are not, the models do not get very much warming at all. In fact, all of the
models runs for the IPCC AR4 assessment (more than 20) produce this distinctive result,
driven by the greenhouse gas increases that have been observed to occur.
These results were produced in models whose average global warming associated with a
doubled CO2 forcing of 4 W/m was about 3°C. This translates into a climate sensitivity
(surface temperature change per forcing) of about 0.75°C/(W/m2). The determination of
climate sensitivity is crucial to projecting the future impact of increased greenhouse gases, and
the credibility of this projected value relies on the ability of these models to simulate the
observed temperature changes over the past century. However, in producing the observed
temperature trend in the past, the models made use of very uncertain aerosol forcing. The
greenhouse gas change by itself produces warming in models that exceeds that observed by
some 40% on average (IPCC, 2007, 092765). Cooling associated with aerosols reduces this
warming to the observed level. Different climate models use differing aerosol forcings, both
direct (aerosol scattering and absorption of short and longwave radiation) and indirect (aerosol
effect on cloud cover reflectivity and lifetime), whose magnitudes vary markedly from one
model to the next. Kiehl (2007, 190949) using nine of the IPCC (2007, 092765) AR4 climate
models found that they had a factor of three forcing differences in the aerosol contribution for
the 20th century. The differing aerosol forcing is the prime reason why models whose climate
sensitivity varies by almost a factor of three can produce the observed trend. It was thus
concluded that the uncertainty in IPCC (2007, 092765) anthropogenic climate simulations for
the past century should really be much greater than stated (Kerr, 2007, 190950: Schwartz et al.,
2007, 190384), since, in general, models with low/high sensitivity to greenhouse warming
used weaker/stronger aerosol cooling to obtain the same temperature response (Kiehl, 2007,
190949). Had the situation been reversed and the low/high sensitivity models used strong/
weak aerosol forcing, there would have been a greater divergence in model simulations of the
past century.
December 2009 9-118
-------
Figure 9-71. Sampling the Arctic Haze. Pollution and smoke aerosols can travel long
distances, from mid-latitudes to the Arctic, causing "Arctic Haze." Photo taken
from the NASA DC-8 aircraft during the ARCTAS field experiment over Alaska in
April 2008. Credit: Mian Chin, NASA.
Therefore, the fact that a model has accurately reproduced the global temperature change
in the past does not imply that its future forecast is accurate. This state of affairs will remain
until a firmer estimate of radiative forcing (RF) by aerosols, in addition to that by greenhouse
gases, is available.
Two different approaches are used to assess the aerosol effect on climate. "Forward
modeling" studies incorporate different aerosol types and attempt to explicitly calculate the
aerosol RF. From this approach, IPCC (2007, 092765) concluded that the best estimate of the
global aerosol direct RF (compared with preindustrial times) is -0.5 (-0.9 to -0.1) W/m2. The
RF due to the cloud albedo or brightness effect (also referred to as first indirect or Twomey
effect) is estimated to be -0.7 (-1.8 to -0.3) W/m . No estimate was specified for the effect
associated with cloud lifetime. The total negative RF due to aerosols according to IPCC (2007,
092765) estimates is then -1.3 (-2.2 to -0.5) W/m2. In comparison, the positive radiative
forcing (RF) from greenhouse gases (including tropospheric ozone) is estimated to be +2.9 ±
0.3 W/m2; hence tropospheric aerosols reduce the influence from greenhouse gases by about
45% (15-85%). This approach however inherits large uncertainties in aerosol amount,
composition, and physical and optical properties in modeling of atmospheric aerosols. The
consequences of these uncertainties are discussed in the next section.
The other method of calculating aerosol forcing is called the "inverse approach" - it is
assumed that the observed climate change is primarily the result of the known climate forcing
contributions. If one further assumes a particular climate sensitivity (or a range of
sensitivities), one can determine what the total forcing had to be to produce the observed
temperature change. The aerosol forcing is then deduced as a residual after subtraction of the
greenhouse gas forcing along with other known forcings from the total value. Studies of this
nature come up with aerosol forcing ranges of-0.6 to -1.7 W/m2 (Knutti et al., 2002, 190178;
December 2009
9-119
-------
Knutti et al., 2003, 190180); IPCC AR4 Chap.9); -0.4 to -1.6 W/m2 (Gregory et al., 2002,
190593): and -0.4 to -1.4 W/m2 (Stott, 2006, 190419). This approach however provides a
bracket of the possible range of aerosol forcing without the assessment of current knowledge
of the complexity of atmospheric aerosols.
This chapter of the CC SP SAP2.3 reviews the current state of aerosol RF in the global
models and assesses the uncertainties in these calculations. First representation of aerosols in
the forward global chemistry and transport models and the diversity of the model simulated
aerosol fields are discussed; then calculation of the aerosol direct and indirect effects in the
climate models is reviewed; finally the impacts of aerosols on climate model simulations and
their implications are assessed.
9.3.6.2. Modeling of Atmospheric Aerosols
The global aerosol modeling capability has developed rapidly in the past decade. In the
late 1990s, there were only a few global models that were able to simulate one or two aerosol
components, but now there are a few dozen global models that simulate a comprehensive suite
of aerosols in the atmosphere. As introduced in Chapter 1 (of the CCSP SAP2.3), aerosols
consist of a variety of species including dust, sea salt, sulfate, nitrate, and carbonaceous
aerosols (black and organic carbon) produced from natural and man-made sources with a wide
range of physical and optical properties. Because of the complexity of the processes and
composition, and highly inhomogeneous distribution of aerosols, accurately modeling
atmospheric aerosols and their effects remains a challenge. Models have to take into account
not only the aerosol and precursor emissions, but also the chemical transformation, transport,
and removal processes (e.g., dry and wet depositions) to simulate the aerosol mass
concentrations. Furthermore, aerosol particle size can grow in the atmosphere because the
ambient water vapor can condense on the aerosol particles. This "swelling" process, called
hygroscopic growth, is most commonly parameterized in the models as a function of relative
humidity.
Estimates of Emissions
Aerosols have various sources from both natural and anthropogenic processes. Natural
emissions include wind-blown mineral dust, aerosol and precursor gases from volcanic
eruptions, natural wild fires, vegetation, and oceans. Anthropogenic sources include emissions
from fossil fuel and biofuel combustion, industrial processes, agriculture practices, and human-
induced biomass burning.
Following earlier attempts to quantify manmade primary emissions of aerosols (Penner et
al., 1993, 045457; Turco et al., 1983, 190529) systematic work was undertaken in the late
1990s to calculate emissions of black carbon (BC) and organic carbon (OC), using fuel-use
data and measured emission factors (Cooke and Wilson, 1996, 190046; Cooke et al., 1999,
156365; Liousse et al., 1996, 078158). The work was extended in greater detail and with
improved attention to source-specific emission factors in Bond et al. (2004, 056389), which
provides global inventories of BC and OC for the year 1996, with regional and source-
category discrimination that includes contributions from industrial, transportation, residential
solid-fuel combustion, vegetation and open biomass burning (forest fires, agricultural waste
burning, etc.), and diesel vehicles.
Emissions from natural sources—which include wind-blown mineral dust, wildfires, sea
salt, and volcanic eruptions—are less well quantified, mainly because of the difficulties of
measuring emission rates in the field and the unpredictable nature of the events. Often,
emissions must be inferred from ambient observations at some distance from the actual source.
As an example, it was concluded (Lewis and Schwartz, 2004, 192023) that available
information on size-dependent sea salt production rates could only provide order-of-magnitude
estimates. The natural emissions in general can vary dramatically over space and time.
Aerosols can be produced from trace gases in the atmospheric via chemical reactions, and
those aerosols are called secondary aerosols, as distinct from primary aerosols that are directly
emitted to the atmosphere as aerosol particles. For example, most sulfate and nitrate aerosols
are secondary aerosols that are formed from their precursor gases, sulfur dioxide (SO2) and
nitrogen oxides (NO and NO2, collectively called NOX), respectively. Those sources have been
studied for many years and are relatively well known. By contrast, the sources of secondary
organic aerosols (SOA) are poorly understood, including emissions of their precursor gases
(called volatile organic compounds, VOC) from both natural and anthropogenic sources and
the atmospheric production processes.
Globally, sea salt and mineral dust dominate the total aerosol mass emissions because of
the large source areas and/or large particle sizes. However, sea salt and dust also have shorter
atmospheric lifetimes because of their large particle size, and are radiatively less active than
aerosols with small particle size, such as sulfate, nitrate, BC, and particulate organic matter
(POM, which includes both carbon and non-carbon mass in the organic aerosol), most of
which are anthropogenic in origin.
Because the anthropogenic aerosol RF is usually evaluated (e.g., by the IPCC) as the
anthropogenic perturbation since the pre-industrial period, it is necessary to estimate the
December 2009 9-120
-------
historical emission trends, especially the emissions in the pre-industrial era. Compared to
estimates of present-day emissions, estimates of historical emission have much larger
uncertainties. Information for past years on the source types and strengths and even locations
are difficult to obtain, so historical inventories from preindustrial times to the present have to
be based on limited knowledge and data. Several studies on historical emission inventories of
BC and OC (e.g., Bond et al, 2007, 190050: Fernandes et al, 2007, 190554: Ito and Penne,
2005, 190626: Junker and Liousse, 2008, 190971: Novakov et al., 2003, 048398), SO2 (Stern,
2005,190416), and various species (Dentener et al., 2006, 088434: Van Aardenne et al., 2001,
055564) are available in the literature; there are some similarities and some differences among
them, but the emission estimates for early times do not have the rigor of the studies for
present-day emissions. One major conclusion from all these studies is that the growth of
primary aerosol emissions in the 20th century was not nearly as rapid as the growth in CO2
emissions. This is because in the late 19th and early 20th centuries, particle emissions such as
BC and POM were relatively high due to the heavy use of biofuels and the lack of particulate
controls on coal-burning facilities; however, as economic development continued, traditional
biofuel use remained fairly constant and particulate emissions from coal burning were reduced
by the application of technological controls (Bond et al., 2007, 190050). Thus, particle
emissions in the 20th century did not grow as fast as CO2 emissions, as the latter are roughly
proportional to total fuel use—oil and gas included. Another challenge is estimating historical
biomass burning emissions. A recent study suggested about a 40% increase in carbon
emissions from biomass burning from the beginning to the end of last century (Mouillot et al.,
2006,190549), but it is difficult to verify.
Table 9-11. Anthropogenic emissions of aerosols and precursors for 2000 and 1750.
Source
Biomass burning
Biofuel
Fossil fuel
Species*
BC
POM
S
BC
POM
S
BC
POM
S
Emission*2000
(Tg/yr)
3.1
34.7
4.1
9.1
9.6
3.0
3.2
98.9
1.03
12.8
1.46
0.39
1.56
0.12
Emission 1750
(Tg/yr)
# Data source for 2000 emission: biomass burning - Global Fire Emission Dataset fGFED); biofuel BC and POM - Speciated Pollutant Emission Wzard (SPEW); biofuel sulfur - International Institute for
Applied System Analysis (NASA); fossil fuel BC and POM - SPEW; fossil fuel sulfur - Emission Database for Global Atmospheric Research (EDGAR) and NASA. Fossil fuel emission of sulfur (S) is the sum of
emission from industry, power plants, and transportation listed in Dentener et al. (2006, 088434V
* S=sulfur, including S02 and particulate S0j2~. Most emitted as S02, and 2.5% emitted as S0j2~.
Source: Adapted from Dentener et al. (2006, 088434)
As an example, Table 9-11 shows estimated anthropogenic emissions of sulfur, BC and
POM in the present day (year 2000) and pre-industrial time (1750) compiled by Dentener et al.
(2006, 088434) These estimates have been used in the Aerosol Comparisons between
Observations and Models (AeroCom) project (Experiment B, which uses the year 2000
emission; and Experiment PP^E, which uses pre-industrial emissions), for simulating
atmospheric aerosols and anthropogenic aerosol RF. The AeroCom results are discussed below
and in Section 9.3.6.3.
Aerosol Mass Loading and Optical Depth
In the global models, aerosols are usually simulated in the successive steps of sources
(emission and chemical formation), transport (from source location to other area), and removal
processes (dry deposition, in which particles fall onto the surface, and wet deposition by rain)
that control the aerosol lifetime. Collectively, emission, transport, and removal determine the
amount (mass) of aerosols in the atmosphere. Aerosol optical depth (AOD), which is a
measure of solar or thermal radiation being attenuated by aerosol particles via scattering or
absorption, can be related to the atmospheric aerosol mass loading as follows:
December 2009 9-121
-------
AOD = MEE • M
where M is the aerosol mass loading per unit area (g m-2), MEE is the mass extinction
efficiency or specific extinction in unit of m2/g, which is
MEE =
Equation 9-3
where Qext is the extinction coefficient (a function of particle size distribution and
refractive index), reff is the aerosol particle effective radius, p is the aerosol particle density,
and f is the ratio of ambient aerosol mass (wet) to dry aerosol mass M. Here, M is the result
from model-simulated atmospheric processes and MEE embodies the aerosol physical
(including microphysical) and optical properties. Since Qext varies with radiation wavelength,
so do MEE and AOD. AOD is the quantity that is most commonly obtained from remote
sensing measurements and is frequently used for model evaluation (see Chapter 2 of the CCSP
SAP2.3). AOD is also a key parameter determining aerosol radiative effects.
Here the results from the recent multiple global-model studies by the AeroCom project are
summarized, as they represent the current assessment of model-simulated atmospheric aerosol
loading, optical properties, and RF for the present-day. AeroCom aims to document differences
in global aerosol models and compare the model output to observations. Sixteen global models
participated in the AeroCom Experiment A (AeroCom-A), for which every model used their
own configuration, including their own choice of estimating emissions (Kinne et al., 2006,
155903: Textor et al., 2006, 190456). Five major aerosol types: sulfate, BC, POM, dust, and
sea salt, were included in the experiments, although some models had additional aerosol
species. Of those major aerosol types, dust and sea-salt are predominantly natural in origin,
whereas sulfate, BC, and POM have major anthropogenic sources.
Table 9-12 summarizes the model results from the AeroCom-A for several key
parameters: Sources (emission and chemical transformation), mass loading, lifetime, removal
rates, and MEE and AOD at a commonly used, mid-visible, wavelength of 550 nanometer
(nm). These are the globally averaged values for the year 2000. Major features and conclusions
are:
• Globally, aerosol source (in mass) is dominated by sea salt, followed by dust, sulfate,
POM, and BC. Over the non-desert land area, human activity is the major source of
sulfate, black carbon, and organic aerosols.
• Aerosols are removed from the atmosphere by wet and dry deposition. Although sea
salt dominates the emissions, it is quickly removed from the atmosphere because of
its large particle size and near-surface distributions, thus having the shortest lifetime.
The median lifetime of sea salt from the AeroCom-A models is less than half a day,
whereas dust and sulfate have similar lifetimes of 4 days and BC and POM 6-7 days.
• Globally, small-particle-sized sulfate, BC, and POM make up a little over 10% of
total aerosol mass in the atmosphere. However, they are mainly from anthropogenic
activity, so the highest concentrations are in the most populated regions, where their
effects on climate and air quality are major concerns.
• Sulfate and BC have their highest MEE at mid-visible wavelengths, whereas dust is
lowest among the aerosol types modeled. That means for the same amount of aerosol
mass, sulfate and BC are more effective at attenuating (scattering or absorbing) solar
radiation than dust. This is why the sulfate AOD is about the same as dust AOD even
though the atmospheric amount of sulfate mass is 10 times less than that of the dust.
• There are large differences, or diversities, among the models for all the parameters
listed in Table 9-12. The largest model diversity, shown as the % standard deviation
from the all-model-mean and the range (minimum and maximum values) in Table
9-12, is in sea salt emission and removal; this is mainly associated with the
differences in particle size range and source parameterizations in each model. The
diversity of sea salt atmospheric loading however is much smaller than that of
sources or sinks, because the largest particles have the shortest lifetimes even though
they comprise the largest fraction of emitted and deposited mass.
Equation 9-4
December 2009
9-122
-------
Table 9-12. Summary of statistics of AeroCom Experiment A results from 16 global models.
Quantity
Mean
Median
Range
Stddev/mean*
SOURCES (TG/YR)
scV~
BC
Organic matter
Dust
Sea salt
179
11.9
96.6
1840
16600
186
11.3
96.0
1640
6280
98-232
7.8-19.4
53-138
672-4040
2180-121000
22%
23%
26%
49%
199%
REMOVAL RATE (DAY-)
scV~
BC
Organic matter
Dust
Sea salt
0.25
0.15
0.16
0.31
5.07
0.24
0.15
0.16
0.25
2.50
0.19-0.39
0.066-0.19
0.09-0.23
0.14-0.79
0.95-35.0
18%
21%
24%
62%
188%
LIFETIME (DAY)
S04"
BC
Organic matter
Dust
Sea salt
4.12
7.12
6.54
4.14
0.48
4.13
6.54
6.16
4.04
0.41
2.6-5.4
5.3-15
4.3-11
1.3-7.0
0.03-1.1
18%
33%
27%
43%
58%
MASS LOADING (TG)
S04"
BC
Organic matter
Dust
Sea salt
1.99
0.24
1.70
19.2
7.52
1.98
0.21
1.76
20.5
6.37
0.92-2.70
0.046-0.51
0.46-2.56
4.5-29.5
2.5-13.2
25%
42%
27%
40%
54%
MEEAT550NM(M2G-1)
SO/~
BC
Organic matter
Dust
Sea salt
11.3
9.4
5.7
0.99
3.0
9.5
9.2
5.7
0.95
3.1
4.2-28.3
5.3-18.9
3.7-9.1
0.46-2.05
0.97-7.5
56%
36%
26%
45%
55%
AODAT550NM
S04"
BC
Organic matter
Dust
Sea salt
TOTAL AOT AT 550 NM
0.035
0.004
0.018
0.032
0.033
0.724
0.034
0.004
0.019
0.033
0.030
0.727
0.015-0.051
0.002-0.009
0.006-0.030
0.012-0.054
0.02-0.067
0.056-0.151
33%
46%
36%
44%
42%
18%
Stddev/mean was used as the term "diversity" in Textor et al. (2006,190456).
Source: Textor et al. (2006,1904561 and Kinne et al. (2006,1559031, and AeroCom website http://nansen.ipsl.iussieu.fr/AEROCOM/data.html
December 2009
9-123
-------
• Among the key parameters compared in Table 9-12, the models agree best for
simulated total AOD - the % of standard deviation from the model mean is 18%,
with the extreme values just a factor of 2 apart. The median value of the multi-model
simulated global annual mean total AOD, 0.127, is also in agreement with the global
mean values from recent satellite measurements. However, despite the general
agreement in total AOD, there are significant diversities at the individual component
level for aerosol optical thickness, mass loading, and mass extinction efficiency. This
indicates that uncertainties in assessing aerosol climate forcing are still large, and
they depend not only on total AOD but also on aerosol absorption and scattering
direction (called asymmetry factor), both of which are determined by aerosol
physical and optical properties. In addition, even with large differences in mass
loading and MEE among different models, these terms could compensate for each
other (Equation 9-3) to produce similar AOD. This is illustrated in Figure 9-72. For
example, model LO and LS have quite different mass loading (44 and 74 mg m" ,
respectively), especially for dust and sea salt amount, but they produce nearly
identical total AOD (0.127 and 0.128, respectively).
• Because of the large spatial and temporal variations of aerosol distributions, regional
and seasonal diversities are even larger than the diversity for global annual means.
To further isolate the impact of the differences in emissions on the diversity of simulated
aerosol mass loading, identical emissions for aerosols and their precursor were used in the
AeroCom Experiment B exercise in which 12 of the 16 AeroCom-A models participated
(Textor et al., 2007, 190458). The comparison of the results and diversity between AeroCom-A
and -B for the same models showed that using harmonized emissions does not significantly
reduce model diversity for the simulated global mass and AOD fields, indicating that the
differences in atmospheric processes, such as transport, removal, chemistry, and aerosol
microphysics, play more important roles than emission in creating diversity among the models.
This outcome is somewhat different from another recent study, in which the differences in
calculated clear-sky aerosol RF between two models (a regional model STEM and a global
model MOZART) were attributed mostly to the differences in emissions (Bates et al., 2006,
189912), although the conclusion was based on only two model simulations for a few focused
regions. It is highly recommended from the outcome of AeroCom-A and -B that, although
more detailed evaluation for each individual process is needed, multi-model ensemble results,
e.g., median values of multi-model output variables, should be used to estimate aerosol RF,
due to their greater robustness, relative to individual models, when compared to observations
(Schulz et al., 2006, 190381: Textor et al., 2006, 190456: Textor et al., 2007, 190458).
9.3.6.3. Calculating Aerosol Direct Radiative Forcing
The three parameters that define the aerosol direct RF are the AOD, the single scattering
albedo (SSA), and the asymmetry factor (g), all of which are wavelength dependent. AOD is
indicative of how much aerosol exists in the column, SSA is the fraction of radiation being
scattered versus the total attenuation (scattered and absorbed), and the g relates to the direction
of scattering that is related to the size of the particles (see Chapter 1 of the CCSP SAP2.3). An
indication of the particle size is provided by another parameter, the Angstrom exponent (A),
which is a measure of differences of AOD at different wavelengths. For typical tropospheric
aerosols, A tends to be inversely dependent on particle size; larger values of A are generally
associated with smaller aerosols particles. These parameters are further related; for example,
for a given composition, the ability of a particle to scatter radiation decreases more rapidly
with decreasing size than does its ability to absorb, so at a given wavelength varying A can
change SSA. Note that AOD, SSA, g, A, and all the other parameters in Equation 9-3 and
Equation 9-4 vary with space and time due to variations of both aerosol composition and
relative humidity, which influence these characteristics.
December 2009 9-124
-------
n is
0.1ft
• 0,12
I fl-i
" 0.08
c Q.Qi
1 U.Q4
D.02
0
ll'l •
Mill
I Hi
,1 i
•AH
ss
DU
•POM
•BC
LO LS UL SP
-------
Table 9-13. S042" mass loading, MEE and AOD at 550 nm, shortwave radiative forcing at the top of
the atmosphere, and normalized forcing with respect to AOD and mass. All values refer
to anthropogenic perturbation.
Model
Mass load
(mg m" )
MEE
(mV1)
AOD att
550 nm
TOA Forcing
(W/m2)
Forcing/AOD
(W/m2)
Forcing/Mass
(Wg-1)
PUBLISHED SINCE IPCC 2001
ACCM3
B GEOSCHEM
CGISS
DGISS
EGISS*
FSPRINTARS
GLMD
HLOA
I GATORG
JPNNL
KUIO-CTM
LUIO-GCM
2.23
1.53
3.30
3.27
2.12
1.55
2.76
3.03
3.06
5.50
1.79
2.28
11.8
6.7
9.7
9.9
7.6
10.6
0.018
0.022
0.015
0.03
0.042
0.019
-0.56
-0.33
-0.65
-0.96
-0.57
-0.21
-0.42
-0.41
-0.32
-0.44
-0.37
-0.29
-18
-30
-14
-10
-19
-251
-216
-197
-294
-269
-135
-152
-135
-105
-80
-207
-127
AEROCOM: IDENTICAL EMISSIONS USED FOR YEAR 2000 AND 1750
MUMI
N UIO-CTM
OLOA
PLSCE
Q ECHAMS-HAM
R GISS**
S UIO-GCM
T SPRINTARS
UULAQ
Average A-L
Average M-U
Minimum A-U
Maximum A-U
StddevA-L
Std dev M-U
%Stddev/avgA-L
%Stddev/avg M-U
2.64
1.70
3.64
3.01
2.47
1.34
1.72
1.19
1.62
2.70
2.15
1.19
5.50
1.09
0.83
40%
39%
7.6
11.2
9.6
7.6
6.5
4.5
7.0
10.9
12.3
9.4
8.6
4.5
12.3
1.9
2.6
20%
30%
0.02
0.019
0.035
0.023
0.016
0.006
0.012
0.013
0.02
0.024
0.018
0.006
0.042
0.010
0.008
41%
45%
-0.58
-0.36
-0.49
-0.42
-0.46
-0.19
-0.25
-0.16
-0.22
-00.46
-0.35
-0.96
-0.16
0.202
0.149
44%
43%
-29
-19
-14
-18
-29
-32
-21
-12
-11
-18
-21
-32
-10
7
8
38%
37%
-220
-212
-135
-140
-186
-142
-145
-134
-136
-181
-161
-294
-80
68
35
385
22%
Model abbreviations: CCM3=Community Climate Model; GEOSCHEM=Goddard Earth Observing System-Chemistry; GISS=Goddard nstitute for Space Studies; SPRINTARS=Spectral Radiation-Transport
Model for Aerosol Species; LMD=Laboratoire de Meteorologie Dynamique; LOA=Laboratoire d'OptiqueAtmospherique; GATORG=Gas, Aerosol Transport and General circulation model; PNNL=Pacific
Northwest National Laboratory; UIO-CTM=Univeristy of Oslo CTM; UIO-GCM=University of Oslo GCM; UMI=University of Michigan; LSCE=Laboratoire des Sciences du Climat et de {'Environment;
ECHAMS5-HAM=European Centre Hamburg with Hamburg Aerosol Module; ULAQ=University of IL'Aquila.
Source: Adapted from IPCC AR4 (2007, 0927651 and Schulz et al. (2006,1903811
December 2009 9-126
-------
Table 9-14. Particulate organic matter (POM) and BC mass loading, AOD at 550 nm, shortwave
radiative forcing at the top of the atmosphere, and normalized forcing with respect to
AOD and mass.
Model
Mass
load
(mg rrf2)
MEE
(mV1)
AOD at J
CCQ nm Forcma
ssunm (W/ 2j
Forcing/
AOD
(W/m2)
Forcing/
Mass
(Wg-1)
Mass
load
(mg m"2)
MEE
(mV1)
AOD at
550 nm
TOA
Forcina
(W/m2)
Forcing/
AOD
(W/m2)
Forcing/
Mass
(Wg-1)
PUBLISHED SINCE IPCC 2001
A SPRINTARS
BLOA
CGISS
DGISS
EGISS*
FGISS
G SPRINTARS
H GATORG
I MOZGN
JCCM
KUIO-CTM
2.33
1.86
1.86
2.39
2.49
2.67
2.56
3.03
6.9
9.1
8.1
10.9
5.9
-0.24
0.016 -0.25
0.017 -0.26
0.015 -0.30
-0.18
-0.23
0.029 -0.27
-0.06
0.018 -0.34
-16
-15
-20
-9
-19
-107
-140
-161
-75
-92
-101
-23
-112
0.37
0.29
0.29
0.39
0.43
0.53
0.39
0.33
0.30
0.36
0.55
0.61
0.35
0.50
0.53
0.42
0.55
0.34
0.19
AEROCOM: IDENTICAL EMISSIONS FOR YEAR 2000 & 1750
LUMI
M UIO-CTM
NLOA
OLSCE
P ECHAMS-HAM
QGISS**
R UIO-GCM
S SPRINTARS
TULAQ
Average A-K
Average L-T
Minimum A-T
Maximum A-T
StddevA-K
Std dev L-T
%Stddev/avgA-K
%Stddev/avg L-T
1.16
1.12
1.41
1.50
1.00
1.22
0.88
1.84
1.71
2.40
1.32
0.88
3.03
0.39
0.32
16%
25%
5.2
5.2
6.0
5.3
7.7
4.9
5.2
10.9
4.4
8.2
6.1
4.4
10.9
1.7
2.0
21%
33%
0.0060 -0.23
0.0058 -0.16
0.0085 -0.16
0.0079 -0.17
0.0077 -0.10
0.0060 -0.14
0.0046 -0.06
0.0200 -0.10
0.0075 -0.09
0.019 -0.24
0.008 -0.13
0.005 -0.34
0.029 -0.06
0.006 0.09
0.005 0.05
30% 36%
56% 39%
-38
-28
-19
-22
-13
-23
-13
-5
-12
-16
-19
-38
-5
4
10
26%
52%
-198
-143
-113
-113
-100
-115
-68
-54
-53
-102
-106
-198
-23
41
46
41%
43%
0.19
0.19
0.25
0.25
0.16
0.24
0.19
0.37
0.38
0.37
0.25
0.16
0.53
0.08
0.08
22%
32%
6.8
7.1
7.9
4.4
7.7
7.6
10.3
9.5
7.6
7.7
4.4
10.3
1.6
21%
1.29
1.34
1.98
1.11
1.23
1.83
1.95
3.50
2.90
1.90
1.11
3.50
0.82
43%
0.25
0.22
0.32
0.30
0.20
0.22
0.36
0.32
0.08
0.44
0.25
0.08
0.61
0.06
0.09
23%
34%
194
164
162
270
163
120
185
91
28
153
28
270
68
45%
1316
1158
1280
1200
1250
917
1895
865
211
1242
1121
211
2103
384
450
31%
40%
Source: Based on IPCC AR4 (2007, 092765) and Schulz et al. (2006,190381).
December 2009
9-127
-------
The IPCC AR4 (2007, 092765) assessed anthropogenic aerosol RF based on the model
results published after the IPCC TAR in 2001, including those from the AeroCom study
discussed above. These results (adopted from IPCC AR4) are shown in Table 9-13 for sulfate
and Table 9-14 for carbonaceous aerosols (BC and POM), respectively. All values listed in
Table 9-13 and Table 9-14 refer to anthropogenic perturbation, i.e., excluding the natural
fraction of these aerosols. In addition to the mass burden, MEE, and AOD, Table 9-13 and
Table 9-14 also list the "normalized forcing", also known as "forcing efficiency", one for the
forcing per unit AOD, and the other the forcing per gram of aerosol mass (dry). For some
models, aerosols are externally mixed, that is, each aerosol particle contains only one aerosol
type such as sulfate, whereas other models allow aerosols to mix internally to different
degrees, that is, each aerosol particle can have more than one component, such as black carbon
coated with sulfate. For models with internal mixing of aerosols, the component values for
AOD, MEE, and forcing were extracted (Schulz et al, 2006, 190381).
Considerable variation exists among these models for all quantities in Table 9-13 and
Table 9-14. The RF for all the components varies by a factor of 6 or more: Sulfate from 0.16 to
0.96 W/m2, POM from -0.06 to -0.34 W/m2, and BC from +0.08 to +0.61 W/m2, with the
standard deviation in the range of 30 to 40% of the ensemble mean. It should be noted that
although BC has the lowest mass loading and AOD, it is the only aerosol species that absorbs
strongly, causing positive forcing to warm the atmosphere, in contrast to other aerosols that
impose negative forcing to cool the atmosphere. As a result, the net anthropogenic aerosol
forcing as a whole becomes less negative when BC is included. The global average
anthropogenic aerosol direct RF at the top of the atmosphere (TOA) from the models, together
with observation-based estimates (see Chapter 2 of the CCSP SAP2.3), is presented in Figure
9-73. Note the wide range for forcing in Figure 9-73. The comparison with observation-based
estimates shows that the model estimated forcing is in general lower, partially because the
forcing value from the model is the difference between present-day and pre-industrial time,
whereas the observation-derived quantity is the difference between an atmosphere with and
without anthropogenic aerosols, so the "background" value that is subtracted from the total
forcing is higher in the models. The discussion so far has dealt with global average values. The
geographic distributions of multi-model aerosol direct RF has been evaluated among the
AeroCom models, which are shown in Figure 9-74 for total and anthropogenic AOD at 550 nm
and anthropogenic aerosol RF at TOA, within the atmospheric column, and at the surface.
Globally, anthropogenic AOD is about 25% of total AOD (Figure 9-74A and B) but is more
concentrated over polluted regions in Asia, Europe, and North America and biomass burning
regions in tropical southern Africa and South America. At TOA, anthropogenic aerosol causes
negative forcing over mid-latitude continents and oceans with the most negative values (-1 to
-2 W/m2) over polluted regions (Figure 9-74C). Although anthropogenic aerosol has a cooling
effect at the surface with surface forcing values down to -10 W/m over China, India, and
tropical Africa (Figure 9-74E), it warms the atmospheric column with the largest effects again
over the polluted and biomass burning regions. This heating effect will change the atmospheric
circulation and can affect the weather and precipitation (e.g., Kim et al., 2006, 190917).
December 2009 9-128
-------
Aerosol Direct Radiative Forcing
§
11
%S
s
Bellouinetal. (2005.155684)
Chung etal. (2005,155733)
Yuetal. (2006.156173)
GISSJ (Liao and Seinfeld, 2005,199892)
GISS_2 (Liao and Seinfeld, 2005,199892)
LOA (Reddy et al., 2005,190208)
SPRINTARS (Takemura et al., 2005,190439)
UIO-GCM (Kirkevag and Iversen, 2002,199893;
GATORG (Jacobson, 2001,043110)
GISS (Hansen et al., 2005,190596)
GISS (Koch. 2001.192054)
UMI
UIO_CTM
LOA
LSCE
ECHAM50HAM
GISS
UIO-GCM
SPRINTARS
ULAQ
-0.8
-0.6 -0.4 -0.2
Radiative Forcing (W/m2)
Source: IPCC (2007,092763.
Figure 9-73.
Aerosol direct radiative forcing in various climate and aerosol models. Observed
values are shown in the top section.
December 2009
9-129
-------
A
1
I
•Hfml
5-0
4.0
.1.0
!.0
1 II
05
Pd
o.o
-0.2
-o.s
-1,0
-3.0
-J.O
-*n
-i.O
-10,0
•45 8 45 M
loi^tulfi
Source: Schulz et al. (2006,1903811 and AeroCom image catalog (http://nansen.ipsl.iussieu.fr/AEROCOM/aerocomhome.htmll
Figure 9-74. Aerosol optical thickness and anthropogenic shortwave all-sky radiative forcing
from the AeroCom study. Shown in the figure: total AOD (A) and anthropogenic
AOD (B) at 550 nm, and radiative forcing at TOA (C), atmospheric column (D), and
surface (E).
December 2009
9-130
-------
Aerosol Direct Radiative Forcing
I
I
Hadley (Jones etal., 2001,1Ł
Hadley (Williams et al., 2001,199898)
CSIRO (Rotstayn and Penner, 2001,193754)
GISS (Menon et al., 2002,155978)
CSIRO (Rotstayn and Liu, 2003,199905)
LMDZ (Quaas et al., 2004,199907)
LMDZ/POLDER (Dufresne et al, 2005,199908)
GFDL (Ming et al, 2005,199919)
ECHAM (Lohmann et al, 2000,199910)
PNNL (Ghan et al, 2001,199911)
NCAR-CCM (Chuang et al, 2002,199912)
NCAR-CCM (Kristjansson et al, 2002,199916)
SPRINTARS (Suzuki et al, 2004,199917)
SPIRNTARS (Takemura et al, 2005,190439)
GISS (Hansen et al, 2005,190596)
LMDZ/CTRL (Quaas and Boucher, 2005,190573)
LMDZ/POLDER (Quaas and Boucher, 2005,19057
LMDZ/MODIS (Quaas and Boucher, 2005,190573)
UM_ctrl(Chen and Penner. 2005.199918)
UM-1 (Chen and Penner, 2005,199918)
UM_2 (Chen and Penner, 2005,199918)
UM_3 (Chen and Penner, 2005,199918)
UM_4 (Chen and Penner, 2005,199918)
UM_5 (Chen and Penner, 2005,199918)
UM_6 (Chen and Penner, 2005,199918)
Oslo (Penner et al, 2006,190564)
LMDZ (Penner et al, 2006,190564)
CCSR (Penner et al, 2006,190564)
-2.0 -1.5 -1.0 -0.5
Radiative Forcing (W/m2)
Source: IPCC (2007, 092765).
Figure 9-75. Radiative forcing from the cloud albedo effect (1st aerosol indirect effect) in the
global climate models used from IPCC (2007, 092765). Chapter 2, Figure 2.14, of
the IPCC AR4. Species included in the lower panel are S042", sea salt, organic
and BC, dust and nitrates; in the top panel, only S042", sea salt and OC are
included.
December 2009
9-131
-------
Basic conclusions from forward modeling of aerosol direct RF are:
• The most recent estimate of all-sky shortwave aerosol direct RF at TOA from
anthropogenic sulfate, BC, and POM (mostly from fossil fuel/biofuel combustion and
biomass burning) is -0.22 ±0.18 W/m averaged globally, exerting a net cooling
effect. This value would represent the low-end of the forcing magnitude, since some
potentially significant anthropogenic aerosols, such as nitrate and dust from human
activities are not included because of their highly uncertain sources and processes.
IPCC AR4 had adjusted the total anthropogenic aerosol direct RF to -0.5 ± 0.4 W/m2
by adding estimated anthropogenic nitrate and dust forcing values based on limited
modeling studies and by considering the observation-based estimates (see Chapter 2
of the CCSP SAP2.3).
• Both sulfate and POM cause negative forcing whereas BC causes positive forcing
because of its highly absorbing nature. Although BC comprises only a small fraction
of anthropogenic aerosol mass load and AOD, its forcing efficiency (with respect to
either AOD or mass) is an order of magnitude stronger than sulfate and POM, so its
positive shortwave forcing largely offsets the negative forcing from sulfate and
POM. This points out the importance of improving the model ability to simulate each
individual aerosol components more accurately, especially black carbon. Separately,
it is estimated from recent model studies that anthropogenic sulfate, POM, and BC
forcings at TOA are -0.4, -0.18, +0.35 W/m2, respectively. The anthropogenic nitrate
and dust forcings are estimated at -0.1 W/m2 for each, with uncertainties exceeds
100% (IPCC, 2007, 092765).
• In contrast to long-lived greenhouse gases, anthropogenic aerosol RF exhibits
significant regional and seasonal variations. The forcing magnitude is the largest over
the industrial and biomass burning source regions, where the magnitude of the
negative aerosol forcing can be of the same magnitude or even stronger than that of
positive greenhouse gas forcing.
• There is a large spread of model-calculated aerosol RF even in the global annual
averaged values. The AeroCom study shows that the model diversity at some
locations (mostly East Asia and African biomass burning regions) can reach ± 3
W/m2, which is an order of magnitude above the global averaged forcing value of -
0.22 W/m2. The large diversity reflects the low level of current understanding of
aerosol radiative forcing, which is compounded by uncertainties in emissions,
transport, transformation, removal, particle size, and optical and microphysical
(including hygroscopic) properties.
• In spite of the relatively small value of forcing at TOA, the magnitudes of
anthropogenic forcing at the surface and within the atmospheric column are
considerably larger: -1 to -2 W/m2 at the surface and +0.8 to +2 W/m2 in the
atmosphere. Anthropogenic aerosols thus cool the surface but heat the atmosphere,
on average. Regionally, the atmospheric heating can reach annually averaged values
exceeding 5 W/m2. Source: Schulz et al. (2006, 190381) and AeroCom Image
Catalog (littp://nansen.ipsl.jussieu.fr/AEROCOM7aerocomhome.html)
• Figure 9-74D). These regional effects and the negative surface forcing are expected
to exert an important effect on climate through alteration of the hydrological cycle.
9.3.6.4. Calculating Aerosol Indirect Forcing
Aerosol Effects on Clouds
A subset of the aerosol particles can act as cloud condensation nuclei (CCN) and/or ice
nuclei (IN). Increases in aerosol particle concentrations, therefore, may increase the ambient
concentrations of CCN and IN, affecting cloud properties. For a fixed cloud liquid water
content, a CCN increase will lead to more cloud droplets so that the cloud droplet size will
decrease. That effect leads to brighter clouds, the enhanced albedo then being referred to as the
"cloud albedo effect" (Twomey, 1977, 190533), also known as the first indirect effect. If the
droplet size is smaller, it may take longer to rainout, leading to an increase in cloud lifetime,
hence the "cloud lifetime" effect (Albrecht, 1989, 045783), also called the second indirect
effect. Approximately one-third of the models used for the IPCC 20th century climate change
simulations incorporated an aerosol indirect effect, generally (though not exclusively)
considered only with sulfates.
Shown in Figure 9-75 are results from published model studies indicating the different RF
values from the cloud albedo effect. The cloud albedo effect ranges from -0.22 to -1.85 W/m2;
the lowest estimates are from simulations that constrained representation of aerosol effects on
clouds with satellite measurements of drop size vs. aerosol index. In view of the difficulty of
quantifying this effect remotely (discussed later), it is not clear whether this constraint
provides an improved estimate. The estimate in the IPCC AR4 ranges from +0.4 to -1.1 W/m2,
with a "best-guess" estimate of 0.7 W/m2.
The representation of cloud effects in GCMs is considered below. However, it is
becoming increasingly clear from studies based on high resolution simulations of aerosol-
December 2009 9-132
-------
cloud interactions that there is a great deal of complexity that is unresolved in climate models.
This point is examined below (High Resolution Modeling).
Most models did not incorporate the "cloud lifetime effect." Hansen et al. (2005, 059087)
compared this latter influence (in the form of time-averaged cloud area or cloud cover
increase) with the cloud albedo effect. In contrast to the discussion in IPCC (2007, 092765),
they argue that the cloud cover effect is more likely to be the dominant one, as suggested both
by cloud-resolving model studies (Ackerman et al., 2004, 190056) and satellite observations
(Kaufman et al., 2005, 155891). The cloud albedo effect may be partly offset by reduced cloud
thickness accompanying aerosol pollutants, producing a meteorological (cloud) rather than
aerosol effect (see the discussion in Lohman and Feichter, 2005, 155942). The distinction
between meteorological feedback and aerosol forcing can become quite opaque; as noted
earlier, the term feedback is restricted here to those processes that are responding to a change
in temperature. Nevertheless, both aerosol indirect effects were utilized in Hansen et al. (2005,
059087), with the second indirect effect calculated by relating cloud cover to the aerosol
number concentration, which in turn is a function of sulfate, nitrate, black carbon and organic
carbon concentration. Only the low altitude cloud influence was modeled, principally because
there are greater aerosol concentrations at low levels, and because low clouds currently exert
greater cloud RE The aerosol influence on high altitude clouds, associated with IN changes, is
a relatively unexplored area for models and as well for process-level understanding.
Hansen et al. (2005, 059087) used coefficients to normalize the cooling from aerosol
indirect effects to between -0.75 and -1 W/m2, based on comparisons of modeled and observed
changes in the diurnal temperature range as well as some satellite observations. The response
of the GISS model to the direct and two indirect effects is shown in Figure 9-76. As
parameterized, the cloud lifetime effect produced somewhat greater negative RF (cooling), but
this was the result of the coefficients chosen. Geographically, it appears that the "cloud cover"
effect produced slightly more cooling in the Southern Hemisphere than did the "cloud albedo"
response, with the reverse being true in the Northern Hemisphere (differences on the order of a
few tenths °C).
Model Experiments
There are many different factors that can explain the large divergence of aerosol indirect
effects in models (Figure 9-75). To explore this in more depth, Penner et al. (2006, 190564)
used three general circulation models to analyze the differences between models for the first
indirect effect, as well as a combined first plus second indirect effect. The models all had
different cloud and/or convection parameterizations. In the first experiment, the monthly
average aerosol mass and size distribution of, effectively, sulfate aerosol were prescribed, and
all models followed the same prescription for parameterizing the cloud droplet number
concentration (CDNC) as a function of aerosol concentration. In that sense, the only difference
among the models was their separate cloud formation and radiation schemes. The different
models all produced similar droplet effective radii, and therefore shortwave cloud forcing, and
change in net outgoing whole sky radiation between pre-industrial times and the present.
Hence the first indirect effect was not a strong function of the cloud or radiation scheme. The
results for this and the following experiments are presented in Figure 9-77, where the
experimental results are shown sequentially from left to right for the whole sky effect and in
Table 9-15 for the clear-sky and cloud forcing response as well.
December 2009 9-133
-------
Aerosol Direct and Indirect Effects
Cloud Cover (%) Planetary Al bedo (%) Fs (W/mJ) ATs (°C)
Direct .16 .14 (E3 Run) -.64 Direct (I E3 Run) -.32
-.77 AlECIdAlb (I Run) -.36
•5 -4 ,2 -I -.5-.2 .2 .51247 -2-! -.5-.2-.I .1 .2 .5 I 2 -23-5 -2 -I -.S-.2 .2 .5 I 1 5 19 -5-2 -I ..S-.2-.I.I .2 .5 I 2 S
Source: Reprinted with Permission of Bioresource Technology from Hansen et al. (2005, 0590871.
Figure 9-76. Anthropogenic impact on cloud cover, planetary albedo, radiative flux at the
surface (while holding sea surface temperatures and sea ice fixed) and surface
air temperature change from the direct aerosol forcing (top row), the first indirect
effect (second row) and the second indirect effect (third row). The temperature
change is calculated from year 81-120 of a coupled atmosphere simulation with
the GISS model.
The change in cloud forcing is the difference between whole sky and clear sky outgoing
radiation in the present day minus pre-industrial simulation. The large differences seen
between experiments 5 and 6 are due to the inclusion of the clear sky component of aerosol
scattering and absorption (the direct effect) in experiment 6.
In the second experiment, the aerosol mass and size distribution were again prescribed,
but now each model used its own formulation for relating aerosols to droplets. In this case one
of the models produced larger effective radii and therefore a much smaller first indirect aerosol
effect (Figure 9-77, Table 9-15. However, even in the two models where the effective radius
change and net global forcing were similar, the spatial patterns of cloud forcing differ,
especially over the biomass burning regions of Africa and South America.
The third experiment allowed the models to relate the change in droplet size to change in
precipitation efficiency (i.e., they were now also allowing the second indirect effect - smaller
droplets being less efficient rain producers - as well as the first). The models utilized the same
relationship for autoconversion of cloud droplets to precipitation. Changing the precipitation
efficiency results in all models producing an increase in cloud liquid water path, although the
effect on cloud fraction was smaller than in the previous experiments. The net result was to
increase the negative radiative forcing in all three models, albeit with different magnitudes: for
two of the models the net impact on outgoing shortwave radiative increased by about 20%,
whereas in the third model (which had the much smaller first indirect effect), it was magnified
by a factor of three.
December 2009
9-134
-------
Exp I Exp 2 6xp» 3 E*p -4 Exp 5 Exp 6
Source: Adapted from Penner et al. (2006,1905641.
Figure 9-77. Global average present-day short wave cloud forcing at TOA (top) and change in
whole sky net outgoing shortwave radiation (bottom) between the present-day
and pre-industrial simulations for each model in each experiment.
December 2009
9-135
-------
Table 9-15. Differences in present day and pre-industrial outgoing solar radiation (W/m2) in the
different experiments.
Model
EXP1
EXP2
EXP3
EXP4
EXP5
EXP6
WHOLE-SKY
CAM-Oslo
-0.648
-0.726
-0.833
-0.580
-0.365
-0.518
LMD-Z
-0.682
-0.597
-0.722
-1.194
-1.479
-1.553
CCSR
-0.739
-0.218
-0.733
-0.350
-1.386
-1.386
CLEAR-SKY
CAM-Oslo
-0.063
-0.066
-0.026
0.014
-0.054
-0.575
LMD-Z
-0.054
0.019
0.030
-0.066
-0.126
-1.034
CCSR
0.018
-0.007
-0.045
-0.008
0.018
-1.160
CLOUD-FORCING
CAM-Oslo
-0.548
-0.660
-0.807
-0.595
-0.311
0.056
LMD-Z
-0.628
-0.616
-0.752
-1.128
-1.353
-0.518
CCSR
-0.757
-0.212
-0.728
-0.345
-1.404
-0.200
EXP1: tests cloud formation and radiation schemes
EXP2: tests formulation for relating aerosols to droplets
EXP3: tests inclusion of droplet size influence on precipitation efficiency
EXP4: tests formulation of droplet size influence on precipitation efficiency
EXP5: tests model aerosol formulation from common sources
EXP6: added the direct aerosol effect
Source: Adapted from Penner et al. (2006,1905641.
In the fourth experiment, the models were now each allowed to use their own formulation
to relate aerosols to precipitation efficiency. This introduced some additional changes in the
whole sky shortwave forcing (Figure 9-77).
In the fifth experiment, models were allowed to produce their own aerosol concentrations,
but were given common sources. This produced the largest changes in the RF in several of the
models. Within any one model, therefore, the change in aerosol concentration has the largest
effect on droplet concentrations and effective radii. This experiment too resulted in large
changes in RF.
In the last experiment, the aerosol direct effect was included, based on the full range of
aerosols used in each model. While the impact on the whole-sky forcing was not large, the
addition of aerosol scattering and absorption primarily affected the change in clear sky
radiation (Table 9-15).
The results of this study emphasize that in addition to questions concerning cloud physics,
the differences in aerosol concentrations among the models play a strong role in inducing
differences in the indirect effect(s), as well as the direct one.
Observational constraints on climate model simulations of the indirect effect with satellite
data (e.g., MODIS) have been performed previously in a number of studies (e.g., Lohmann et
al., 2006, 190451; Menon et al., 2008, 190534; Quaas et al., 2006, 190915; Storlevmo et al.,
2006, 190418).
These have been somewhat limited since the satellite retrieved data used do not have the
vertical profiles needed to resolve aerosol and cloud fields (e.g., cloud droplet number and
liquid water content); the temporal resolution of simultaneous aerosol and cloud product
retrievals are usually not available at a frequency of more than one a day; and higher level
clouds often obscure low clouds and aerosols. Thus, the indirect effect, especially the second
indirect effect, remains, to a large extent, unconstrained by satellite observations. However,
improved measurements of aerosol vertical distribution from the newer generation of sensors
on the A-train platform may provide a better understanding of changes to cloud properties
from aerosols. Simulating the top-of-atmosphere reflectance for comparison to satellite
measured values could be another way to compare model with observations, which would
eliminate the inconsistent assumptions of aerosol optical properties and surface reflectance
encountered when compared the model calculated and satellite retrieved AOD values.
December 2009
9-136
-------
Additional Aerosol Influences
Various observations have empirically related aerosols injected from biomass burning or
industrial processes to reductions in rainfall (e.g., Andreae et al, 2004, 155658: Eagan et al.,
1974, 190231; Rosenfeld, 2000, 002234; Warner, 1968, 157114). There are several potential
mechanisms associated with this response.
In addition to the two indirect aerosol effects noted above, a process denoted as the
"semidirect" effect involves the absorption of solar radiation by aerosols such as black carbon
and dust. The absorption increases the temperature, thus lowering the relative humidity and
producing evaporation, hence a reduction in cloud liquid water. The impact of this process
depends strongly on what the effective aerosol absorption actually is; the more absorbing the
aerosol, the larger the potential positive forcing on climate (by reducing low level clouds and
allowing more solar radiation to reach the surface). This effect is responsible for shifting the
critical value of SSA (separating aerosol cooling from aerosol warming) from 0.86 with fixed
clouds to 0.91 with varying clouds (Hansen et al., 1997, 043104). Reduction in cloud cover
and liquid water is one way aerosols could reduce rainfall.
More generally, aerosols can alter the location of solar radiation absorption within the
system, and this aspect alone can alter climate and precipitation even without producing any
change in net radiation at the top of the atmosphere (the usual metric for climate impact). By
decreasing solar absorption at the surface, aerosols (from both the direct and indirect effects)
reduce the energy available for evapotranspiration, potentially resulting in a decrease in
precipitation. This effect has been suggested as the reason for the decrease in pan evaporation
over the last 50 years (Roderick and Farquhar, 2002, 042788). The decline in solar radiation at
the surface appears to have ended in the 1990s (Wild et al., 2005, 156156), perhaps because of
reduced aerosol emissions in industrial areas (Kruger and Grasl, 2002, 190200), although this
issue is still not settled.
Energy absorption by aerosols above the boundary layer can also inhibit precipitation by
warming the air at altitude relative to the surface, i.e., increasing atmospheric stability. The
increased stability can then inhibit convection, affecting both rainfall and atmospheric
circulation (Chung and Zhang, 2004, 190054; Ramanathan et al., 2001, 042681). To the extent
that aerosols decrease droplet size and reduce precipitation efficiency, this effect by itself could
result in lowered rainfall values locally.
In their latest simulations, Hansen et al. (2007, 190597) did find that the indirect aerosol
effect reduced tropical precipitation; however, the effect is similar regardless of which of the
two indirect effects is used, and also similar to the direct effect. So it is likely that the
reduction of tropical precipitation is because of aerosol induced cooling at the surface and the
consequent reduced evapotranspiration. Similar conclusions were reached by Yu et al. (2002,
190923) and Feingold et al. (2005, 190550). In this case, the effect is a feedback and not a
forcing.
The local precipitation change, through its impacts on dynamics and soil moisture, can
have large positive feedbacks. Harvey (2004, 190598) concluded from assessing the response
to aerosols in eight coupled models that the aerosol impact on precipitation was larger than on
temperature. He also found that the precipitation impact differed substantially among the
models, with little correlation among them.
Recent GCM simulations have further examined the aerosol effects on hydrological cycle.
Ramanathan et al. (2005, 190199) showed from fully coupled ocean-atmosphere GCM
experiments that the "solar dimming" effect at the surface, i.e., the reduction of solar radiation
reaching the surface, due to the inclusion of absorbing aerosol forcing causes a reduction in
surface evaporation, a decrease in meridional sea surface temperature (SST) gradient and an
increase in atmospheric stability, and a reduction in rainfall over South Asia. Lau and Kim
(2006, 190226) examined the direct effects of aerosol on the monsoon water cycle variability
from GCM simulations with prescribed realistic global aerosol forcing and proposed the
"elevated heat pump" effect, suggesting that atmospheric heating by absorbing aerosols (dust
and black carbon), through water cycle feedback, may lead to a strengthening of the South
Asia monsoon. These model results are not necessarily at odds with each other, but rather
illustrate the complexity of the aerosol-monsoon interactions that are associated with different
mechanisms, whose relative importance in affecting the monsoon may be strongly dependent
on spatial and temporal scales and the timing of the monsoon. These results may be model
dependent and should be further examined.
High Resolution Modeling
Largely by its nature, the representation of the interaction between aerosol and clouds in
GCMs is poorly resolved. This stems in large part from the fact that GCMs do not resolve
convection on their large grids (order of several hundred km), that their treatment of cloud
microphysics is rather crude, and that as discussed previously, their representation of aerosol
needs improvement. Superparametrization efforts (where standard cloud parameterizations in
the GCM are replaced by resolving clouds in each grid column of the GCM via a cloud
resolving model) (e.g.,Grabowski, 2004, 190590) could lead the way for the development of
more realistic cloud fields and thus improved treatments of aerosol cloud interactions in large-
scale models. However, these are just being incorporated in models that resolve both cloud and
December 2009 9-137
-------
aerosols. Detailed cloud parcel models have been developed to focus on the droplet activation
problem (that asks under what conditions droplets actually start forming) and questions
associated with the first indirect effect. The coupling of aerosol and cloud modules to
dynamical models that resolve the large turbulent eddies associated with vertical motion and
clouds [large eddy simulations (LES) models, with grid sizes of—100 m and domains —10 km]
has proven to be a powerful tool for representing the details of aerosol-cloud interactions
together with feedbacks (e.g., Ackerman et al, 2004, 190056; Feingold et al, 1994, 190535;
1999, 190540: Kogan et al., 1994, 190186: Stevens et al., 1996, 190417).
This section explores some of the complexity in the aerosol indirect effects revealed by
such studies to illustrate how difficult parameterizing these effects properly in GCMs could
really be.
The First Indirect Effect
The relationship between aerosol and drop concentrations (or drop sizes) is a key piece of
the first indirect effect puzzle. (It should not, however, be equated to the first indirect effect
which concerns itself with the resultant RF). A huge body of measurement and modeling work
points to the fact that drop concentrations increase with increasing aerosol. The main
unresolved questions relate to the degree of this effect, and the relative importance of aerosol
size distribution, composition and updraft velocity in determining drop concentrations (for a
review, see Mcfiggans et al., 2006, 190532). Studies indicate that the aerosol number
concentration and size distribution are the most important aerosol factors. Updraft velocity
(unresolved by GCMs) is particularly important under conditions of high aerosol particle
number concentration.
Although it is likely that composition has some effect on drop number concentrations,
composition is generally regarded as relatively unimportant compared to the other parameters
(Dusek et al., 2006, 155756: Ervens et al., 2005, 190527: Feingold et al., 2003, 190551:
Fitzgerald, 1975, 095417). Therefore, it has been stated that the significant complexity in
aerosol composition can be modeled, for the most part, using fairly simple parameterizations
that reflect the soluble and insoluble fractions (e.g., Rissler et al., 2004, 190225). However,
composition cannot be simply dismissed. Furthermore, chemical interactions also cannot be
overlooked. A large uncertainty remains concerning the impact of organic species on cloud
droplet growth kinetics, thus cloud droplet formation. Cloud drop size is affected by wet
scavenging, which depends on aerosol composition especially for freshly emitted aerosol. And
future changes in composition will presumably arise due to biofuels/biomass burning and a
reduction in sulfate emissions, which emphasizes the need to include composition changes in
models when assessing the first indirect effect. The simple soluble/insoluble fraction model
may become less applicable than is currently the case.
The updraft velocity, and its change as climate warms, may be the most difficult aspect to
simulate in GCMs because of the small scales involved. In GCMs it is calculated in the
dynamics as a grid box average, and parameterized on the small scale indirectly because it is a
key part of convection and the spatial distribution of condensate, as well as droplet activation.
Numerous solutions to this problem have been sought, including estimation of vertical velocity
based on predicted turbulent kinetic energy from boundary layer models (Larson et al., 2001,
190212: Lohmann et al., 1999, 190443) and PDF representations of subgrid quantities, such as
vertical velocity and the vertically-integrated cloud liquid water ('liquid water path,' or LWP)
(Golaz et al., 2002, 190587: 2002, 190589: Larson et al., 2005, 190220: Pincus and Klein,
2000, 190565). Embedding cloud-resolving models within GCMs is also being actively
pursued (Grabowski et al., 1999, 190592: Randall et al., 2003, 190201). Numerous other
details come into play; for example, the treatment of cloud droplet activation in GCM
frameworks is often based on the assumption of adiabatic conditions, which may overestimate
the sensitivity of cloud to changes in CCN (Sotiropoulou et al., 2006, 190406: Sotiropoulou et
al., 2007, 190405). This points to the need for improved theoretical understanding followed by
new parameterizations.
Other Indirect Effects
The second indirect effect is often referred to as the "cloud lifetime effect", based on the
premise that non-precipitating clouds will live longer. In GCMs the "lifetime effect" is
equivalent to changing the representation of precipitation production and can be parameterized
as an increase in cloud area or cloud cover (e.g., Hansen et al., 2005, 059087). The second
indirect effect hypothesis states that the more numerous and smaller drops associated with
aerosol perturbations, suppress collision-induced rain, and result in a longer cloud lifetime.
Observational evidence for the suppression of rain in warm clouds exists in the form of
isolated studies (e.g., Warner, 1968, 157114) but to date there is no statistically robust proof of
surface rain suppression (Levin and Cotton, 2008, 190375). Results from ship-track studies
show that cloud water may increase or decrease in the tracks (Coakley and Walsh, 2002,
192025) and satellite studies suggest similar results for warm boundary layer clouds (Han et
al., 2002, 049181). Ackerman et al. (2004, 190056) used LES to show that in stratocumulus,
cloud water may increase or decrease in response to increasing aerosol depending on the
relative humidify of the air overlaying the cloud. Wang et al. (2003, 157106) showed that all
December 2009 9-138
-------
else being equal, polluted stratocumulus clouds tend to have lower water contents than clean
clouds because the small droplets associated with polluted clouds evaporate more readily and
induce an evaporation-entrainment feedback that dilutes the cloud. This result was confirmed
by Xue and Feingold (2006, 190920) and Jiang and Feingold (2006, 190976) for shallow
cumulus, where pollution particles were shown to decrease cloud fraction. Furthermore, Xue et
al. (2008, 190921) suggested that there may exist two regimes: the first, a precipitating regime
at low aerosol concentrations where an increase in aerosol will suppress precipitation and
increase cloud cover (Albrecht, 1989, 045783): and a second, non-precipitating regime where
the enhanced evaporation associated with smaller drops will decrease cloud water and cloud
fraction.
The possibility of bistable aerosol states was proposed earlier by Baker and Charlson
(1990, 190016) based on consideration of aerosol sources and sinks. They used a simple
numerical model to suggest that the marine boundary layer prefers two aerosol states: a clean,
oceanic regime characterized by a weak aerosol source and less reflective clouds; and a
polluted, continental regime characterized by more reflective clouds. On the other hand, study
by Ackerman et al. (1994, 189975) did not support such a bistable system using a somewhat
more sophisticated model. Further observations are needed to clarify the nature of
cloud/aerosol interactions under a variety of conditions.
Finally, the question of possible effects of aerosol on cloud lifetime was examined by
Jiang et al. (2006, 133165), who tracked hundreds of cumulus clouds generated by LES from
their formative stages until they dissipated. They showed that in the model there was no effect
of aerosol on cloud lifetime, and that cloud lifetime was dominated by dynamical variability.
It could be argued that the representation of these complex feedbacks in GCMs is not
warranted until a better understanding of the processes is at hand. Moreover, until GCMs are
able to represent cloud scales, it is questionable what can be obtained by adding microphysical
complexity to poorly resolved clouds. A better representation of aerosol-cloud interactions in
GCMs therefore depends on the ability to improve representation of aerosols and clouds, as
well as their interaction, in the hydrologic cycle. This issue is discussed further in the next
chapter.
9.3.6.5. Aerosol in the Climate Models
Aerosol in the IPCC AR4 Climate Model Simulations
To assess the atmospheric and climate response to aerosol forcing, e.g., changes in surface
temperate, precipitation, or atmospheric circulation, aerosols, together with greenhouse gases
should be an integrated part of climate model simulation under the past, present, and future
conditions. Table 9-16 lists the forcing species that were included in 25 climate modeling
groups used in the IPCC AR4 (2007, 092765) assessment. All the models included long-lived
greenhouse gases, most models included sulfate direct forcing, but only a fraction of those
climate models considered other aerosol types. In other words, aerosol RF was not adequately
accounted for in the climate simulations for the IPCC AR4. Put still differently, the current
aerosol modeling capability has not been fully incorporated into the climate model
simulations. As pointed out in Section 9.3.6.4, fewer than one-third of the models incorporated
an aerosol indirect effect, and most considered only sulfates.
The following discussion compares two of the IPCC AR4 climate models that include all
major forcing agencies in their climate simulation: the model from the NASA Goddard
Institute for Space Studies (GISS) and from the NOAA Geophysical Fluid Dynamics
Laboratory (GFDL). The purpose in presenting these comparisons is to help elucidate how
modelers go about assessing their aerosol components, and the difficulties that entail. A
particular concern is how aerosol forcings were obtained in the climate model experiments for
IPCC AR4. Comparisons with observations have already led to some improvements that can
be implemented in climate models for subsequent climate change experiments (e.g., Koch et
al., 2006, 190184, for GISS model). This aspect is discussed further in Chapter 4 of the CCSP
SAP2.3.
December 2009 9-139
-------
Table 9-16. Forcings used in IPCC
adapted from SAP 1.1
participating modelini
documentation/ipcc
AR4 simulations of 20th century climate change. This table is
Table 5.2 (compiled using information provided by the
3 centers, see http://www.pcmdi.llnl.gov/ipcc/model-
model documentation. php) plus additional information from
that website. Eleven different forcings are listed: well-mixed greenhouse gases (G),
tropospheric and stratospheric ozone (0), S042" aerosol direct (SD) and indirect
effects (S), black carbon (BC) and organic carbon aerosols (OC), mineral dust (MD),
sea salt (SS), land use/land cover (LU), solar irradiance (SO), and volcanic aerosols
(V). Check mark denotes inclusion of a specific forcing. As used here, "inclusion"
means specification of a time-varying forcing, with changes on interannual and
longer timescales.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Model
BCC-CMI
BCCR-BCM2.0
CCSM3
CGCM3.1(T47)
CGCM3.1(T63)
CNRM-CM3
CSIRO-MkS.O
CSIRO-Mk3.5
ECHAMS/MPI-OM
ECHO-G
FGOALS-g1 .0
GFDL-CM2.0
GFDL-CM2.1
GISS-AOM
GISS-EH
GISS-ER
INGV-SXG
INM-CM3.0
IPSL-CM4
MICROC3.2(hires)
MICROC3.2(medres)
MRI-CGCM2.3.2
PCM
UKMO-HasCM3
UKMO-HadGEM1
Country
China
Norway
USA
Canada
Canada
France
Australia
Australia
Germany
Germany/
Korea
China
USA
USA
USA
USA
USA
Italy
Russia
France
Japan
Japan
Japan
USA
U.K.
U.K.
G
A/
A/
A/
V
V
V
V
V
V
A/
V
V
A/
V
V
V
A/
V
A/
V
V
V
V
AI
V
0 SD SI BC OC MD SS LU SO V
V V
A/ A/ A/
V V V V V V
V
V
V V V
V
V
V V V
V V V V V
V
V V V V V V
A/ A/ A/ A/ A/ A/
A/ A/
//////////
AA'A'A'A'A'A'A'A'A'
/ / / / / / / / / /
AA'A'A'A'A'A'A'A'A'
V V
A/ A/
A/ A/
V V V V V V V V V
V V V V V V V V V
V V V
V V V V
A/ A/ A/
A/ V V V V V V
77?e G/SS Mode/
There have been many different configurations of aerosol simulations in the GISS model
over the years, with different emissions, physics packages, etc., as is apparent from the
multiple GISS entries in the preceding figures and tables. There were also three different GISS
GCM submissions to IPCC AR4, which varied in their model physics and ocean formulation.
(Note that the aerosols in these three GISS versions are different from those in the
AeroCom simulations described in Sections 9.3.6.2 and 9.3.6.3.) The GCM results discussed
December 2009
9-140
-------
below all relate to the simulations known as GISS model ER (Schmidt et al, 2006, 190373)
(see Table 9-16). Although the detailed description and model evaluation have been presented
in Liu et al. (2006, 190422), below are the general characteristics of aerosols in the GISS ER:
Aerosol fields: The aerosol fields used in the GISS ER is a prescribed "climatology"
which is obtained from chemistry transport model simulations with monthly averaged mass
concentrations representing conditions up to 1990. Aerosol species included are sulfate, nitrate,
BC, POM, dust, and sea salt. Dry size effective radii are specified for each of the aerosol types,
and laboratory-measured phase functions are employed for all solar and thermal wavelengths.
For hygroscopic aerosols (sulfate, nitrate, POM, and sea salt), formulas are used for the
particle growth of each aerosol as a function of relative humidity, including the change in
density and optical parameters. With these specifications, the AOD, single scattering albedo,
and phase function of the various aerosols are calculated. While the aerosol distribution is
prescribed as monthly mean values, the relative humidity component of the extinction is
updated each hour. The global averaged AOD at 550 nm is about 0.15.
Global distribution: When comparing with AOD from observations by multiple satellite
sensors of MODIS, MISR, POLDER, and AVHRR and surface based sunphotometer network
AERONET (see Chapter 2 of the CCSP SAP2.3 for detailed information about data),
qualitative agreement is apparent, with generally higher burdens in Northern Hemisphere
summer, and seasonal variations of smoke over southern Africa and South America, as well as
wind blown dust over northern African and the Persian Gulf. Aerosol optical depth in both
model and observations is smaller away from land. There are, however, considerable
discrepancies between the model and observations. Overall, the GISS GCM has reduced
aerosol optical depths compared with the satellite data (a global, clear-sky average of about
80% compared with MODIS and MISR data), although it is in better agreement with
AERONET ground-based measurements in some locations (note that the input aerosol values
were calibrated with AERONET data). The model values over the Sahel in Northern
Hemisphere winter and the Amazon in Southern Hemisphere winter are excessive, indicative
of errors in the biomass burning distributions, at least partially associated with an older
biomass burning source used (the source used here was from (Liousse et al., 1996, 078158)).
Seasonal variation: A comparison of the seasonal distribution of the global AOD between
the GISS model and satellite data indicates that the model seasonal variation is in qualitative
agreement with observations for many of the locations that represent major aerosol regimes,
although there are noticeable differences. For example, in some locations the seasonal
variations are different from or even opposite to the observations.
Particle size parameter: The Angstrom exponent (A), which is determined by the contrast
between the AOD at two or more different wavelengths and is related to aerosol particle size
(discussed in Section 9.3.6.3). This parameter is important because the particle size
distribution affects the efficiency of scattering of both short and long wave radiation, as
discussed earlier. A from the GISS model is biased low compared with AERONET, MODIS,
and POLDER data, although there are technical differences in determining the A. This low
bias suggests that the aerosol particle size in the GISS model is probably too large. The
average effective radius in the GISS model appears to be 0.3-0.4 um, whereas the
observational data indicates a value more in the range of 0.2-0.3 um (Liu et al., 2006, 190422).
Single scattering albedo: The model-calculated SSA (at 550 nm) appears to be generally
higher than the AERONET data at worldwide locations (not enough absorption), but lower
than AERONET data in Northern Africa, the Persian Gulf, and the Amazon (too much
absorption). This discrepancy reflects the difficulties in modeling BC, which is the dominant
absorbing aerosol, and aerosol sizes. Global averaged SSA at 550 nm from the GISS model is
at about 0.95.
Aerosol direct RF: The GISS model calculated anthropogenic aerosol direct shortwave RF
is -0.56 W/m2 at TOA and -2.87 W/m2 at the surface. The TOA forcing (upper left, Figure
9-78) indicates that, as expected, the model has larger negative values in polluted regions and
positive forcing at the highest latitudes. At the surface (lower left, Figure 9-78) GISS model
values exceed -4 W/m over large regions. Note there is also a longwave RF of aerosols (right
column), although they are much weaker than the shortwave RF.
There are several concerns for climate change simulations related to the aerosol trend in
the GISS model. One is that the aerosol fields in the GISS AR4 climate simulation (version
ER) are kept fixed after 1990. In fact, the observed trend shows a reduction in tropospheric
aerosol optical thickness from 1990 through the present, at least over the oceans (Mishchenko
and Geogdzhayev, 2007, 190545). Hansen et al. (2007, 190597) suggested that the deficient
warming in the GISS model over Eurasia post-1990 was due to the lack of this trend. Indeed, a
possible conclusion from the Penner et al. (2002, 190562) study was that the GISS model
overestimated the AOD (presumably associated with anthropogenic aerosols) poleward of
30°N. However, when an alternate experiment reduced the aerosol optical depths, the polar
warming became excessive (Hansen et al., 2007, 190597). The other concern is that the GISS
model may underestimate the organic and sea salt AOD, and overestimate the influence of
black carbon aerosols in the biomass burning regions (Liu et al., 2006, 190422; deduced from
Penner et al., 2002, 190562). To the extent that is true, it would indicate the GISS model
underestimates the aerosol direct cooling effect in a substantial portion of the tropics, outside
of biomass burning areas. Clarifying those issues requires numerous modeling experiments
and various types of observations.
December 2009 9-141
-------
The GFDL Model
A comprehensive description and evaluation of the GFDL aerosol simulation are given in
Ginoux et al. (2006, 190582). Below are the general characteristics:
Aerosol fields: The aerosols used in the GFDL climate experiments are obtained from
simulations performed with the MOZART 2 model (Model for Ozone and Related chemical
Tracers) (Horowitz, 2006, 190620; Horowitz et al., 2003, 057770). The exceptions were dust,
which was generated with a separate simulation of MOZART 2, using sources from Ginoux et
al. (2001, 190579) and wind fields from NCEP/NCAR reanalysis data; and sea salt, whose
monthly mean concentrations were obtained from a previous study by Haywood et al. (1999,
040453). It includes most of the same aerosol species as in the GISS model (although it does
not include nitrates), and, as in the GISS model, relates the dry aerosol to wet aerosol optical
depth via the model's relative humidity for sulfate (but not for organic carbon); for sea salt, a
constant relative humidity of 80% was used. Although the parameterizations come from
different sources, both models maintain a very large growth in sulfate particle size when the
relative humidity exceeds 90%.
Global distributions: Overall, the GFDL global mean aerosol mass loading is within 30%
of that of other studies (Chin et al., 2002, 189996; Reddy et al., 2005, 190207; Tie et al., 2005,
190459), except for sea salt, which is 2 to 5 times smaller. However, the sulfate AOD (0.1) is
2.5 times that of other studies, whereas the organic carbon value is considerably smaller (on
the order of 1/2). Both of these differences are influenced by the relationship with relative
humidity. In the GFDL model, sulfate is allowed to grow up to 100% relative humidity, but
organic carbon does not increase in size as relative humidify increases. Comparison of AOD
with AVHRR and MODIS data for the time period 1996-2000 shows that the global mean
value over the ocean (0.15) agrees with AVHRR data (0.14) but there are significant
differences regionally, with the model overestimating the value in the northern mid latitude
oceans and underestimating it in the southern ocean. Comparison with MODIS also shows
good agreement globally (0.15), but in this case indicates large disagreements over land, with
the model producing excessive AOD over industrialized countries and underestimating the
effect over biomass burning regions. Overall, the global averaged AOD at 550 nm is 0.17,
which is higher than the maximum values in the AeroCom-A experiments (Table 9-12) and
exceeds the observed value too (Ae and S* in Figure 9-72).
December 2009 9-142
-------
net totor TOA
-56
JS7
netsdar BOA
-1.87
net thermal BOA
.13
-J -.2 -J 0 .1 3 ,5 .7 I.I
Source: Figure provided by A. Lads, GISS.
Figure 9-78. Direct radiative forcing by anthropogenic aerosols in the GISS model (including
sulfates, BC, OC and nitrates). Short wave forcing at TOA and surface are shown
in the top left and bottom left panels. The corresponding thermal forcing is
indicated in the right hand panels.
Composition: Comparison of GFDL modeled species with in situ data over North
America, Europe, and over oceans has revealed that the sulfate is overestimated in spring and
summer and underestimated in winter in many regions, including Europe and North America.
Organic and black carbon aerosols are also overestimated in polluted regions by a factor of
two, whereas organic carbon aerosols are elsewhere underestimated by factors of 2 to 3. Dust
concentrations at the surface agree with observations to within a factor of 1 in most places
where significant dust exists, although over the southwest U.S. it is a factor of 10 too large.
Surface concentrations of sea salt are underestimated by more than a factor of 2. Over the
oceans, the excessive sulfate AOD compensates for the low sea salt values except in the
southern oceans.
Size and single-scattering albedo: No specific comparison was given for particle size or
single-scattering albedo, but the excessive sulfate would likely produce too high a value of
reflectivity relative to absorption except in some polluted regions where black carbon (an
absorbing aerosol) is also overestimated.
As in the case of the GISS model, there are several concerns with the GFDL model. The
good global-average agreement masks an excessive aerosol loading over the Northern
Hemisphere (in particular, over the northeast U.S. and Europe) and an underestimate over
biomass burning regions and the southern oceans. Several model improvements are needed,
including better parameterization of hygroscopic growth at high relative humidity for sulfate
and organic carbon; better sea salt simulations; correcting an error in extinction coefficients;
and improved biomass burning emissions inventory (Ginoux et al., 2006, 190582).
Comparisons between GISS and GFDL Model
Both GISS and GFDL models were used in the IPCC AR4 climate simulations for climate
sensitivity that included aerosol forcing. It would be constructive, therefore, to compare the
similarities and differences of aerosols in these two models and to understand what their
December 2009
9-143
-------
impacts are in climate change simulations. Figure 9-79 shows the percentage AOD from
different aerosol components in the two models.
Sulfate: The sulfate AOD from the GISS model is within the range of that from all other
models (Table 9-13), but that from the GFDL model exceeds the maximum value by a factor of
2.5. An assessment in SAP 3.2 (CCSP, 2008, 192028; Shindell et al, 2008, 190393) also
concludes that GFDL had excessive sulfate AOD compared with other models. The sulfate
AOD from GFDL is nearly a factor of 4 large than that from GISS, although the sulfate burden
differs only by about 50% between the two models. Clearly, this implies a large difference in
sulfate MEE between the two models.
BC and POM: Compared to observations, the GISS model appears to overestimate the
influence of BC and POM in the biomass burning regions and underestimate it elsewhere,
whereas the GFDL model is somewhat the reverse: it overestimates it in polluted regions, and
underestimates it in biomass burning areas. The global comparison shown in Table 9-14
indicates the GISS model has values similar to those from other models, which might be the
result of such compensating errors. The GISS and GFDL models have relatively similar
global-average black carbon contributions, and the same appears true for POM.
Sea salt: The GISS model has a much larger sea salt contribution than does GFDL (or
indeed other models).
Global and regional distributions: Overall, the global averaged AOD is 0.15 from the
GISS model and 0.17 from GFDL. However, as shown in Figure 9-79, the contribution to this
AOD from different aerosol components shows greater disparity. For example, over the
Southern Ocean where the primary influence is due to sea salt in the GISS model, but in the
GFDL it is sulfate. The lack of satellite observations of the component contributions and the
limited available in situ measurements make the model improvements at aerosol composition
level difficult.
Climate simulations: With such large differences in aerosol composition and distribution
between the GISS and GFDL models, one might expect that the model simulated surface
temperature might be quite different. Indeed, the GFDL model was able to reproduce the
observed temperature change during the 20th century without the use of an indirect aerosol
effect, whereas the GISS model required a substantial indirect aerosol contribution (more than
half of the total aerosol forcing) (Hansen et al., 2007, 190597). It is likely that the reason for
this difference was the excessive direct effect in the GFDL model caused by its overestimation
of the sulfate optical depth. The GISS model direct aerosol effect (see Section 9.3.6.6) is close
to that derived from observations (Chapter 2 of the CCSP SAP2.3); this suggests that for
models with climate sensitivity close to 0.75°C/(W/m2) (as in the GISS and GFDL models), an
indirect effect is needed.
December 2009 9-144
-------
nun=U96 BC ADD Fraction imix=l96
ire Siff ADD Fr?r:t.~>n
18 61 mui-OBt OunAUD hnsmon m»x*90.7
180 IJdW ftM 0 60E ilDE 180
Figure 9-79. Percentage of aerosol optical depth in the GISS, left, based on Liu et
al. (2006,190422). provided by A. Lacis, GISS, and GFDL, right, from Ginoux et al.
(2006,190582). Models associated with the different components: S(>42~ (1st
row), BC (2nd row), OC (3rd row), sea-salt (4th row), dust (5th row), and nitrate
(last row). Nitrate not available in GFDL model). Numbers on the GISS panels are
global average, but on the GFDL panels are maximum and minimum.
December 2009
9-145
-------
Additional Considerations
Long wave aerosol forcing: So far only the aerosol RF in the shortwave (solar) spectrum
has been discussed. Figure 9-78 (right column) shows that compared to the shortwave forcing,
the values of anthropogenic aerosol long wave (thermal) forcing in the GISS model are on the
order of 10%. Like the shortwave forcing, these values will also be affected by the particular
aerosol characteristics used in the simulation.
Aerosol vertical distribution: Vertical distribution is particularly important for absorbing
aerosols, such as BC and dust in calculating the RF, particularly when longwave forcing is
considered (e.g., Figure 9-78) because the energy they reradiate depends on the temperature
(and hence altitude), which affects the calculated forcing values. Several model inter-
comparison studies have shown that the largest difference among model simulated aerosol
distributions is the vertical profile (e.g., Lohmann et al., 2001, 190448; Penner et al., 2002,
190562: Textor et al., 2006, 190456X due to the significant diversities in atmospheric
processes in the models (e.g., Table 9-12). In addition, the vertical distribution also varies with
space and time, as illustrated in Figure 9-80 from the GISS ER simulations for January and
July showing the most probable altitude of aerosol vertical locations. In general, aerosols in the
northern hemisphere are located at lower altitudes in January than in July, and vice versa for
the southern hemisphere.
Mixing state: Most climate model simulations incorporating different aerosol types have
been made using external mixtures, i.e., the evaluation of the aerosols and their radiative
properties are calculated separately for each aerosol type (assuming no mixing between
different components within individual particles). Observations indicate that aerosols
commonly consist of internally mixed particles, and these "internal mixtures" can have very
different radiative impacts. For example, the GISS-1 (internal mixture) and GISS-2 (external
mixture) model results shows very different magnitude and sign of aerosol forcing from
slightly positive (implying slight warming) to strong negative (implying significant cooling)
TOA forcing (Figure 9-73), due to changes in both radiative properties of the mixtures, and in
aerosol amount. The more sophisticated aerosol mixtures from detailed microphysics
calculations now being used/developed by different modeling groups may well end up
producing very different direct (and indirect) forcing values.
Cloudy sky vs. clear sky: The satellite or AERONET observations are all for clear sky
only because aerosol cannot be measured in the remote sensing technique when clouds are
present. However, almost all the model results are for all-sky because of difficulty in extracting
cloud-free scenes from the GCMs. So the AOD comparisons discussed earlier are not
completely consistent. Because AOD can be significantly amplified when relative humidity is
high, such as near or inside clouds, all-sky AOD values are expected to be higher than clear
sky AOD values. On the other hand, the aerosol RF at TOA is significantly lower for all-sky
than for clear sky conditions; the IPCC AR4 and AeroCom RF study (Schulz et al., 2006,
190381) have shown that on average the aerosol RF value for all-sky is about 1/3 of that for
clear sky although with large diversity (63%). These aspects illustrate the complexity of the
system and the difficulty of representing aerosol radiative influences in climate models whose
cloud and aerosol distributions are somewhat problematic. And of course aerosols in cloudy
regions can affect the clouds themselves, as discussed in Section 9.3.6.5.
December 2009 9-146
-------
•
1
100 JAC 410 480 540 400 660 72C 780 640 900
Source: A. Lacis, GISS.
Figure 9-80. Most probable aerosol altitude (in pressure, hPa) from the GISS model in January
(top) and July (bottom).
9.3.6.6. Impacts of Aerosols on Climate Model Simulations
Surface Temperature Change
It was noted in the introduction that aerosol cooling is essential in order for models to
produce the observed global temperature rise over the last century, at least models with climate
sensitivities in the range of 3°C for doubled CO2 (or ~0.75°C/(W/m2)). The implications of
this are discussed here in somewhat more detail. Hansen et al. (2007, 190597) show that in the
GISS model, well-mixed greenhouse gases produce a warming of close to 1°C between 1880
and the present (Table 9-17). The direct effect of tropospheric aerosols as calculated in that
model produces cooling of close to -0.3°C between those same years, while the indirect effect
(represented in that study as cloud cover change) produces an additional cooling of similar
magnitude (note that the general model result quoted in IPCC AR4 is that the indirect RF is
twice that of the direct effect).
December 2009
9-147
-------
Table 9-17. Climate forcings (1880-2003) used to drive GISS climate simulations, along with the
surface air temperature changes obtained for several periods.
Forcing Agent Forcing Wm"2 (1880-2003) AT Surface °C (year to 2003)
Fa Fs Fe 1880 1900 1950 1979
Well-mixed GHGs 2.62 2.50 2.65 2.72 0.96 0.93 0.74 0.43
Stratospheric H20 0.06 0.05 0.03 0.01 0.05 0.00
03 0.44 0.28 0.26 0.23 O.pOS 0.05 0.00 -0.01
Land use -0.09 -0.09 -0.05 -O.p07 -0.04 -0.02
Snow albedo 0.05 0.05 0.14 0.14 0.03 0.00 0.02 -0.01
Solar irradiance 0.23 0.24 0.23 0.22 0.07 0.07 0.01 0.02
Stratospheric aerosols 0.00 0.00 0.00 0.00 -0.08 -0.03 -0.06 0.04
Trop. aerosol direct forcing -0.41 -0.38 -0.52 -0.60 -0.28 -0.23 -0.18 -0.10
Trop. aerosol indirect forcing -0.87 -0.77 -0.27 -0.29 -0.14 -0.05
Sum of above 1.86 1.90 0.49 0.44 0.40 0.30
All forcings at once 1.77 1.75 0.53 0.61 0.44 0.29
Source: Reprinted with Permission of Springer Publishing from Hansen et al. (2007,1905971.
Instantaneous (Fi), adjusted (Fa), fixed SST (Fs) and effective (Fe) forcings are defined in Hansen et al. (2005, 0590871
The time dependence of the total aerosol forcing used as well as the individual species
components is shown in Figure 9-81. The resultant warming, 0.53 (± 0.04) °C including these
and other forcings (Table 9-17), is less than the observed value of 0.6-0.7°C from 1880-2003.
Hansen et al. (2007, 190597) further show that a reduction in sulfate optical thickness and the
direct aerosol effect by 50%, which also reduced the aerosol indirect effect by 18%, produces a
negative aerosol forcing from 1880-2003 of-0.91 W/m2 (down from-1.37 W/m2 with this
revised forcing). The model now warms 0.75°C over that time. Hansen et al. (2007, 190597)
defend this change by noting that sulfate aerosol removal over North America and western
Europe during the 1990s led to a cleaner atmosphere. Note that the comparisons shown in the
previous section suggest that the GISS model already underestimates aerosol optical depths; it
is thus trends that are the issue here.
The magnitude of the indirect effect used by Hansen et al. (2005, 190596) is roughly
calibrated to reproduce the observed change in diurnal temperature cycle and is consistent with
some satellite observations. However, as Anderson et al. (2003, 054820) note, the forward
calculation of aerosol negative forcing covers a much larger range than is normally used in
GCMs; the values chosen, as in this case, are consistent with the inverse reasoning estimates of
what is needed to produce the observed warming, and hence generally consistent with current
model climate sensitivities. The authors justify this approach by claiming that paleoclimate
data indicate a climate sensitivity of close to 0.75 (± 0.25)°C/(W/m2), and therefore something
close to this magnitude of negative forcing is reasonable. Even this stated range leaves
significant uncertainty in climate sensitivity and the magnitude of the aerosol negative forcing.
Furthermore, IPCC (2007, 092765) concluded that paleoclimate data are not capable of
narrowing the range of climate sensitivity, nominally 0.375-1.13 °C/(W/m2), because of
uncertainties in paleoclimate forcing and response; so from this perspective the total aerosol
forcing is even less constrained than the GISS estimate. Hansen et al. (2007, 190597)
acknowledge that "an equally good match to observations probably could be obtained from a
model with larger sensitivity and smaller net forcing, or a model with smaller sensitivity and
larger forcing".
December 2009 9-148
-------
BC Industrial
BC Biomass
OC Industrial
OC Biomass
Nitrates
Sulfates
— BC Industrial
BC Biomass
—•— OC Industrial
OC Biomass
Nitrates
Sulfates
1875
1900
1925
1950
1975
2000
- Black Carbon Forcings
- Reflective Aerosol Forcings
- All Aerosols
1875
1900
1925
1950
1975
2000
Source: Reprinted with Permission of Springer from Hansen et al. (2007,1905971.
Figure 9-81. Time dependence of aerosol optical thickness (left) and climate forcing (right).
Note that as specified, the aerosol trends are all "flat" from 1990-2000.
The GFDL model results for global mean ocean temperature change (down to 3 km depth)
for the time period 1860-2000 is shown in Figure 9-82, along with the different contributing
factors (Delworth et al., 2005, 190055). This is the same GFDL model whose aerosol
distribution was discussed previously. The aerosol forcing produces a cooling on the order of
50% that of greenhouse warming (generally similar to that calculated by the GISS model,
Table 9-17). Note that this was achieved without any aerosol indirect effect.
The general model response noted by IPCC, as discussed in the introduction was that the
total aerosol forcing of -1.3 W/m2 reduced the greenhouse forcing of near 3 W/m by about
45%, in the neighborhood of the GFDL and GISS forcings. Since the average model sensitivity
was close to 0.75 °C/(W/m ), similar to the sensitivities of these models, the necessary
negative forcing is therefore similar. The agreement cannot therefore be used to validate the
actual aerosol effect until climate sensitivity itself is better known.
Is there some way to distinguish between greenhouse gas and aerosol forcing that would
allow the observational record to indicate how much of each was really occurring? This
question of attribution has been the subject of numerous papers, and the full scope of the
discussion is beyond the range of this (CCSP SAP2.3) report. It might be briefly noted that
Zhang et al. (2006, 157722) using results from several climate models and including both
spatial and temporal patterns, found that the climate responses to greenhouse gases and sulfate
aerosols are correlated, and separation is possible only occasionally, especially at global scales.
This conclusion appears to be both model and method-dependent: using time-space
distinctions as opposed to trend detection may work differently in different models (Gillett et
al., 2002, 190576). Using multiple models helps primarily by providing larger-ensemble sizes
for statistics (Gillett et al., 2002, 190578). However, even separating between the effects of
different aerosol types is difficult. Jones et al. (2005, 155885) concluded that currently the
pattern of temperature change due to black carbon is indistinguishable from the sulfate aerosol
pattern. In contrast, Hansen et al. (2005, 059087) found that absorbing aerosols produce a
different global response than other forcings, and so may be distinguishable. Overall, the
similarity in response to all these very different forcings is undoubtedly due to the importance
of climate feedbacks in amplifying the forcing, whatever its nature.
December 2009
9-149
-------
-0.06 -
1860 1880 1900 1920 1940 1960 1980 2000
Source: Courtesy of the American Geogphysical Union from Delworth et al. (2005,1900551.
Figure 9-82. Change in global mean ocean temperature (left axis) and ocean heat content
(right axis) for the top 3000 m due to different forcings in the GFDL model.
WMGG includes all greenhouse gases and ozone; NATURAL includes solar and
volcanic aerosols (events shown as green triangles on the bottom axis).
Observed ocean heat content changes are shown as well.
Distinctions in the climate response do appear to arise in the vertical, where absorbing
aerosols produce warming that is exhibited throughout the troposphere and into the
stratosphere, whereas reflective aerosols cool the troposphere but warm the stratosphere
(Hansen et al., 2005, 059087; IPCC, 2007, 092765). Delworth et al. (2005, 190055) noted that
in the ocean, the cooling effect of aerosols extended to greater depths, due to the thermal
instability associated with cooling the ocean surface. Hence the temperature response at levels
both above and below the surface may provide an additional constraint on the magnitudes of
each of these forcings, as may the difference between Northern and Southern Hemisphere
changes (IPCC, 2007, 092765, Chapter 9). The profile of atmospheric temperature response
will be useful to the extent that the vertical profile of aerosol absorption, an important
parameter to measure, is known.
Implications for Climate Model Simulations
The comparisons in Sections 9.3.6.2 and 9.3.6.3 suggest that there are large differences in
model calculated aerosol distributions, mainly because of the large uncertainties in modeling
the aerosol atmospheric processes in addition to the uncertainties in emissions. The fact that
the total optical depth is in better agreement between models than the individual components
means that even with similar optical depths, the aerosol direct forcing effect can be quite
different, as shown in the AeroCom studies. Because the diversity among models and
discrepancy between models and observations are much larger at the regional level than in
global average, the assessment of climate response (e.g., surface temperature change) to
aerosol forcing would be more accurate for global average than for regional or hemispheric
differentiation. However, since aerosol forcing is much more pronounced on regional than on
global scales because of the highly variable aerosol distributions, it is insufficient or even
misleading to just get the global average right.
The indirect effect is strongly influenced by the aerosol concentrations, size, type, mixing
state, microphysical processes, and vertical profile. As shown in previous sections, very large
differences exist in those quantities even among the models having similar AOD. Moreover,
modeling aerosol indirect forcing presents more challenges than direct forcing because there is
so far no rigorous observational data, especially on a global scale, that one can use to test the
model simulations. As seen in the comparisons of the GISS and GFDL model climate
simulations for IPCC AR4, aerosol indirect forcing was so poorly constrained that it was
completely ignored by one model (GFDL) but used by another (GISS) at a magnitude that is
more than half of the direct forcing, in order to reproduce the observed surface temperature
trends. A majority of the climate models used in IPCC AR4 do not consider indirect effects;
the ones that did were mostly limited to highly simplified sulfate indirect effects (Table 9-16).
Improvements must be made to at least the degree that the aerosol indirect forcing can no
longer be used to mask the deficiencies in estimating the climate response to greenhouse gas
and aerosol direct RF.
December 2009
9-150
-------
9.3.6.7. Outstanding Issues
Clearly there are still large gaps in assessing the aerosol impacts on climate through
modeling. Major outstanding issues and prospects of improving model simulations are
discussed below.
Aerosol composition
Many global models are now able to simulate major aerosol types such as sulfate, black
carbon, and POM, dust, and sea salt, but only a small fraction of these models simulate nitrate
aerosols or consider anthropogenic secondary organic aerosols. And it is difficult to quantify
the dust emission from human activities. As a result, the IPCC AR4 estimation of the nitrate
and anthropogenic dust TOA forcing was left with very large uncertainty. The next generation
of global models should therefore have a more comprehensive suite of aerosol compositions
with better-constrained anthropogenic sources.
Aerosol absorption
One of the most critical parameters in aerosol direct RF and aerosol impact on
hydrological cycles is the aerosol absorption. Most of the absorption is from BC despite its
small contribution to total aerosol short and long-wave spectral ranges, whereas POM absorbs
in the near UV. The aerosol absorption or SSA, will have to be much better represented in the
models through improving the estimates of carbonaceous and dust aerosol sources, their
atmospheric distributions, and optical properties.
Aerosol indirect effects
The activation of aerosol particles into CCN depends not only on particle size but
chemical composition, with the relative importance of size and composition unclear. In current
aerosol-climate modeling, aerosol size distribution is generally prescribed and simulations of
aerosol composition have large uncertainties. Therefore the model estimated "albedo effect"
has large uncertainties. How aerosol would influence cloud lifetime/ cover is still in debate.
The influence of aerosols on other aspects of the climate system, such as precipitation, is even
more uncertain, as are the physical processes involved. Processes that determine aerosol size
distributions, hygroscopic growth, mixing state, as well as CCN concentrations, however, are
inadequately represented in most of the global models. It will also be difficult to improve the
estimate of indirect effects until the models can produce more realistic cloud characteristics.
Aerosol impacts on surface radiation and atmospheric heating
Although these effects are well acknowledged to play roles in modulating atmospheric
circulation and water cycle, few coherent or comprehensive modeling studies have focused on
them, as compared to the efforts that have gone to assessing aerosol RF at TOA. They have not
yet been addressed in the previous IPCC reports. Here, of particular importance is to improve
the accuracy of aerosol absorption.
Long-term trends of aerosol
To assess the aerosol effects on climate change the long-term variations of aerosol amount
and composition and how they are related to the emission trends in different regions have to be
specified. Simulations of historical aerosol trends can be problematic since historical emissions
of aerosols have shown large uncertainties - as information is difficult to obtain on past source
types, strengths, and even locations. The IPCC AR4 simulations used several alternative
aerosol emission histories, especially for BC and POM aerosols.
Climate modeling
Current aerosol simulation capabilities from CTMs have not been fully implemented in
most models used in IPCC AR4 climate simulations. Instead, a majority employed simplified
approaches to account for aerosol effects, to the extent that aerosol representations in the
GCMs, and the resulting forcing estimates, are inadequate. The oversimplification occurs in
part because the modeling complexity and computing resource would be significantly
increased if the full suite of aerosols were fully coupled in the climate models.
Observational constraints
Model improvement has been hindered by a lack of comprehensive datasets that could
provide multiple constraints for the key parameters simulated in the model. The extensive
AOD coverage from satellite observations and AERONET measurements has helped a great
December 2009 9-151
-------
deal in validating model-simulated AOD over the past decade, but further progress has been
slow. Large model diversities in aerosol composition, size, vertical distribution, and mixing
state are difficult to constrain, because of lack of reliable measurements with adequate spatial
and temporal coverage (see Chapter 2 of the CCSP SAP2.3).
Aerosol radiative forcing
Because of the large spatial and temporal differences in aerosol sources, types, emission
trends, compositions, and atmospheric concentrations, anthropogenic aerosol RF has profound
regional and seasonal variations. So it is an insufficient measure of aerosol RF scientific
understanding, however useful, for models (or observation-derived products) to converge only
on globally and annually averaged TOA RF values and accuracy. More emphasis should be
placed on regional and seasonal comparisons, and on climate effects in addition to direct RF at
TOA.
9.3.6.8. Conclusions
From forward modeling studies, as discussed in the IPCC (2007, 092765) , the direct
effect of aerosols since pre-industrial times has resulted in a negative RF of about -0.5 ± 0.4
W/m2. The RF due to cloud albedo or brightness effect is estimated to be -0.7 (-1.8 to -0.3)
W/m2. Forcing of similar magnitude has been used in some modeling studies for the effect
associated with cloud lifetime, in lieu of the cloud brightness influence. The total negative RF
due to aerosols according to IPCC (2007, 092765) estimates is therefore -1.3 (-2.2 to -0.5)
W/m2. With the inverse approach, in which aerosols provide forcing necessary to produce the
observed temperature change, values range from -1.7 to -0.4 W/m2 (IPCC, 2007, 092765) .
These results represent a substantial advance over previous assessments (e.g., IPCC TAR), as
the forward model estimated and inverse approach required aerosol TOA forcing values are
converging. However, large uncertainty ranges preclude using the forcing and temperature
records to more accurately determine climate sensitivity.
There are now a few dozen models that simulate a comprehensive suite of aerosols. This
is done primarily in the CTMs. Model inter-comparison studies have shown that models have
merged at matching the global annual averaged AOD observed by satellite instruments, but
they differ greatly in the relative amount of individual components, in vertical distributions,
and in optical properties. Because of the great spatial and temporal variations of aerosol
distributions, regional and seasonal diversities are much larger than that of the global annual
mean. Different emissions and differences in atmospheric processes, such as transport,
removal, chemistry, and aerosol microphysics, are chiefly responsible for the spread among the
models. The varying component contributions then lead to differences in aerosol direct RF, as
aerosol scattering and absorption properties depend on aerosol size and type. They also impact
the calculated indirect RF, whose variations are further amplified by the wide range of cloud
and convective parameterizations in models. Currently, the largest aerosol RF uncertainties are
associated with the aerosol indirect effect. Most climate models used for the IPCC AR4
simulations employed simplified approaches, with aerosols specified from stand-alone CTM
simulations. Despite the uncertainties in aerosol RF and widely varying model climate
sensitivity, the IPCC AR4 models were generally able to reproduce the observed temperature
record for the past century. This is because models with lower/higher climate sensitivity
generally used less/more negative aerosol forcing to offset the greenhouse gas warming. An
equally good match to observed surface temperature change in the past could be obtained from
a model with larger climate sensitivity and smaller net forcing, or a model with smaller
sensitivity and larger forcing (Hansen et al., 2007, 190597). Obviously, both greenhouse gases
and aerosol effects have to be much better quantified in future assessments.
Progress in better quantifying aerosol impacts on climate will only be made when the
capabilities of both aerosol observations and representation of aerosol processes in models are
improved. The primary concerns and issues discussed in this chapter of the CCSP SAP2.3
include:
Better representation of aerosol composition and absorption in the global models
Improved theoretical understanding of subgrid-scale processes crucial to aerosol-
cloud interactions and lifetime
Improved aerosol microphysics and cloud parameterizations
Better understanding of aerosol effects on surface radiation and hydrological cycles
More focused analysis on regional and seasonal variations of aerosols
More reliable simulations of aerosol historic long-term trends
More sophisticated climate model simulations with coupled aerosol and cloud
processes
Enhanced satellite observations of aerosol type, SSA, vertical distributions, and
aerosol radiative effect at TOA; more coordinated field experiments to provide
constraints on aerosol chemical, physical, and optical properties.
December 2009 9-152
-------
9.3.7. Fire as a Special Source of PM Welfare Effects
Much interest has developed in defining more precisely the role of pyrogenic C in the boreal
C cycle. This is due to: (1) the resistance of pyrogenic C to decomposition; (2) its influence on soil
processes; and (3) the absorption of solar radiation by soot aerosols (Preston and Schmidt, 2006,
156030).
Preston and Schmidt (2006, 156030) reviewed the current state of knowledge regarding
atmospheric emissions of pyrogenic C in the boreal zone. They considered chemical structures,
analytical methods, formation, characteristics in soil, loss mechanisms, and longevity. Biomass is
largely converted to gaseous forms during burning, but up to several percent is converted to
pyrogenic C, and this includes charcoal and BC. Charcoal is defined visually; BC is defined
chemically by its resistance to oxidation in the laboratory. Andreae and Gelencser (2006, 156215)
reviewed a different category of light-absorbing carbon, referred to as brown carbon.
Within the boreal zone, fire is a critical driver of ecosystem process and nutrient cycling
(Hicke et al, 2003, 156545). For example, Bachelet et al. (2005, 156241) estimated that 61% of the
C gained in Alaska by primary production of boreal forests between 1922 and 1996 was lost to fire.
An updated modeling effort to evaluate the radiative effects of aerosols was presented by Stier
et al. (2007, 157012). Inclusion of refractive indices recommended byBond and Bergstrom (2005,
155696) significantly increased aerosol RF and resulted in better agreement with sun-photometer
estimates. Although this stage of climate modeling improved the representation of aerosols, large
uncertainties remain regarding the effects of aerosol mixing and aerosol-cloud interactions.
Furthermore, Stier et al. (2007, 157012) emphasized that these types of modeling efforts are
dependent upon emission estimates that are likely to vary by a factor of 2 or more.
One important reason for the acknowledged uncertainty in estimating global emissions of
carbonaceous aerosols is the influence of intermittent fires that can occur at scales large enough to
affect hemispheric aerosol concentrations. To better quantify the effects of large-scale fire, Generoso
et al. (2007, 155786) used satellite observations of boreal fires in Russia in 2003 to evaluate the
performance of a global chemistry and transport model in simulating aerosol optical thickness,
transport, and deposition. Emissions estimates of BC and OC were adjusted in the model to better
match satellite observations of pollutant transport over the North Pacific. This resulted in an increase
in optical thickness and BC deposition by a factor of 2. The adjusted model estimated that the fires
contributed 16-33% of the optical thickness and 40-56% of BC deposition north of 75° N in the
spring and summer of 2003.
Large fires also occurred over the Iberian Peninsula and Mediterranean coast during 2003. A
meso-scale atmospheric transport model was used with ground-based measurements and satellite
optical measurements to characterize the dispersion of emitted smoke particles and quantify radiative
effects across Europe (Hodzic et al., 2007, 156553). The modeled wildfire emissions resulted in
increases in PM2.5 concentrations from 20 to 200%. The increased aerosol concentration was
estimated to increase radiative forcing by 10-35 W/m2 during the period of strong fire influence.
Absorption of radiation by BC was also estimated to decrease rates of photolysis by 30%. In this
simulation, all particles were assumed to be internally mixed, and secondary aerosol formation was
not considered. Meteorological conditions in Europe during the exceptionally hot summer of 2003
were linked to enhanced photochemically derived pollutants, increased wild fires, and elevated
aerosol concentrations in an analysis by Vautard et al. (2007, 106012).
In addition to incidental fires, routine biomass burning, usually associated with agriculture in
eastern Europe, also has been shown to contribute to hemispheric concentrations of carbonaceous
aerosols. In the spring of 2006, the most severe air pollution levels in the Arctic to date were
recorded (Stohl et al., 2006, 156100). Atmospheric transport modeling coupled with satellite fire
detection data identified biomass burning for agriculture as the primary cause of the high pollution
levels. Concentrations of PM2.5 peaked during the pollution episode at values of an order of
magnitude greater than those recorded prior to the episode. The increased transport of pollution into
the Arctic during 2006 was attributed to weather conditions that delayed preparations for crop
planting into May. Weather patterns favorable for pollutant transport into the Arctic were related to
unusually warm weather in late April and May, when the majority of agricultural biomass burning
took place that year.
In the summer of 2004, 2.7 million ha were burned by wildfire in Alaska and 3.1 million ha
were burned in Canada. Effects on atmospheric air quality were measured throughout the Arctic,
although the concentrations of particulates varied considerably. Aerosol optical depths were also
December 2009 9-153
-------
increased at all measurement stations, which indicated that the fires were likely to have had a
significant effect on the atmospheric radiation budget for the Arctic (Stohl et al., 2006, 156100). At
one site, a pronounced drop in albedo was observed due presumably to high deposition of light
absorbing particulates on the snow surface by the North American fires in 2004.
Investigations of the effects of large fires on climate forcing have typically focused on the
absorptive effects of BC. However, these fires also release large amounts of CO2 and CH4, as well as
light scattering compounds such as OC, and can enhance cloud formation. These fires also increase
radiative surface absorption through BC deposition on snow and ice, and alter surface albedo and
ecosystem energy budgets within the burn perimeter. Randerson et al. (2006, 156038) estimated the
net climate forcing of greenhouse gases, aerosols, BC deposition on snow and ice and changes in
albedo for the year subsequent to a fire and for 80 years in the future in interior Alaska. The net
effect of the fire in the first year was an increase in radiative forcing, but over the 80-year recovery
period, average net annual radiative forcing was decreased by the fire.
9.3.8. Radiative Effects of Volcanic Aerosols
Section 9.3.8.1. comes directly from IPCC AR4 Chapter 2, Section 2.7.2, with section, table,
and figure numbers changed to be internally consistent with this ISA.
9.3.8.1. Explosive Volcanic Activity
Radiative Effects of Volcanic Aerosols
Volcanic sulfate aerosols are formed as a result of oxidation of the sulfur gases emitted by
explosive volcanic eruptions into the stratosphere. The process of gas-to-particle conversion
has an e-folding time of roughly 35 days (Bluth et al., 1992, 192029: Read et al., 1993,
192031). The e-folding time (by mass) for sedimentation of sulfate aerosols is typically about
12 to 14 months (Baran and Foot, 1994, 192032: Barnes and Hofmann, 1997, 192044: Bluth
et al., 1997, 192045: Lambert et al., 1993, 192231). Also emitted directly during an eruption
are volcanic ash particulates (siliceous material). These are particles usually larger than 2 um
that sediment out of the stratosphere fairly rapidly due to gravity (within three months or so),
but could also play a role in the radiative perturbations in the immediate aftermath of an
eruption. Stratospheric aerosol data incorporated for climate change simulations tends to be
mostly that of the sulfates (Ammann et al., 2003, 192057: Hansen et al., 2002, 049177:
Ramachandran et al., 2000, 192050: Sato et al., 1993, 192046: Stenchikov et al., 1998,
192049: Tett et al., 2002, 192053). As noted in the Second Assessment Report (SAR) and the
TAR, explosive volcanic events are episodic, but the stratospheric aerosols resulting from them
yield substantial transitory perturbations to the radiative energy balance of the planet, with
both shortwave and longwave effects sensitive to the microphysical characteristics of the
aerosols (e.g., size distribution).
Long-term ground-based and balloon-borne instrumental observations have resulted in an
understanding of the optical effects and microphysical evolution of volcanic aerosols (Deshler
et al., 2003, 192058: Hofmann et al., 2003, 192062). Important ground-based observations of
aerosol characteristics from pre-satellite era spectral extinction measurements have been
analysed by Stothers (2001, 192233: 2001, 192232), but they do not provide global coverage.
Global observations of stratospheric aerosol over the last 25 years have been possible owing to
a number of satellite platforms, for example, TOMS and TOYS have been used to estimate
SO2 loadings from volcanic eruptions (Krueger et al., 2000, 192234: Prata et al., 2003,
192235). The Stratospheric Aerosol and Gas Experiment (SAGE) and Stratospheric Aerosol
Measurement (SAM) projects (e.g., McCormick and Trepte, 1987, 192328) have provided
vertically resolved stratospheric aerosol spectral extinction data for over 20 years, the longest
such record. This data set has significant gaps in coverage at the time of the El Chichon
eruption in 1982 (the second most important in the 20th century after Mt. Pinatubo in 1991)
and when the aerosol cloud is dense; these gaps have been partially filled by lidar
measurements and field campaigns (e.g., Antuna et al., 2003, 192251: Thomason and Peter,
2006, 192248).
Volcanic aerosols transported in the atmosphere to polar regions are preserved in the ice
sheets, thus recording the history of the Earth's volcanism for thousands of years (Kruysse,
1971,192236: Mosley-Thompson et al., 2003, 192255: Palmer et al., 2002, 192319). However,
the atmospheric loadings obtained from ice records suffer from uncertainties due to imprecise
knowledge of the latitudinal distribution of the aerosols, depositional noise that can affect the
signal for an individual eruption in a single ice core, and poor constraints on aerosol
microphysical properties. The best-documented explosive volcanic event to date, by way of
reliable and accurate observations, is the 1991 eruption of Mt. Pinatubo. The growth and decay
December 2009 9-154
-------
of aerosols resulting from this eruption have provided a basis for modeling the RF due to
explosive volcanoes. There have been no explosive and climatically significant volcanic events
since Mt. Pinatubo. As pointed out in Ramaswamy et al. (2001, 156899), stratospheric aerosol
concentrations are now at the lowest concentrations since the satellite era and global coverage
began in about 1980. Altitude dependent stratospheric optical observations at a few
wavelengths, together with columnar optical and physical measurements, have been used to
construct the time-dependent global field of stratospheric aerosol size distribution formed in
the aftermath of volcanic events. The wavelength-dependent stratospheric aerosol single-
scattering characteristics calculated for the solar and longwave spectrum are deployed in
climate models to account for the resulting radiative (shortwave plus longwave) perturbations.
Using available satellite- and ground-based observations, Hansen et al. (2002, 049177)
constructed a volcanic aerosols data set for the 1850-1999 period (Sato et al., 1993, 192046).
This has yielded zonal mean vertically resolved aerosol optical depths for visible wavelengths
and column average effective radii. Stenchikov et al. (2006, 192260) introduced a slight
variation to this data set, employing UARS observations to modify the effective radii relative
to Hansen et al. (2002, 049177), thus accounting for variations with altitude. Ammann et al.
(2003, 192057) developed a data set of total aerosol optical depth for the period since 1890
that does not include the Krakatau eruption. The data set is based on empirical estimates of
atmospheric loadings, which are then globally distributed using a simplified parameterization
of atmospheric transport, and employs a fixed aerosol effective radius (0.42 um) for
calculating optical properties. The above data sets have essentially provided the bases for the
volcanic aerosols implemented in virtually all of the models that have performed the 20th-
century climate integrations (Stenchikov et al., 2006, 192260). Relative to Sato et al. (1993,
192046), the Ammann et al. (2003, 192057) estimate yields a larger value of the optical depth,
by 20 to 30% in the second part of the 20th century, and by 50% for eruptions at the end of
19th and beginning of 20th century, for example, the 1902 Santa Maria eruption (Figure 9-83).
The global mean RF calculated using the Sato et al. (1993, 192046) data yields a peak in
radiative perturbation of about -3 W/m2 for the strong (rated in terms of emitted SO2) 1860 and
1991 eruptions of Krakatau and Mt. Pinatubo, respectively. The value is reduced to about -2
W/m2 for the relatively less intense El Chichon and Agung eruptions (Hansen et al., 2002,
049177). As expected from the arguments above, Ammann's RF is roughly 20 to 30% larger
than Sato's RF.
Not all features of the aerosols are well quantified, and extending and improving the data
sets remains an important area of research. This includes improved estimates of the aerosol
size parameters (Bingen et al., 2004, 192262), a new approach for calculating aerosol optical
characteristics using SAGE and UARS data (Bauman et al., 2003, 192265). and
intercomparison of data from different satellites and combining them to fill gaps (Randall et
al., 2001, 192268). While the aerosol characteristics are better constrained for the Mt. Pinatubo
eruption, and to some extent for the El Chichon and Agung eruptions, the reliability degrades
for aerosols from explosive volcanic events further back in time as there are few, if any,
observational constraints on their optical depth and size evolution.
The radiative effects due to volcanic aerosols from major eruptions are manifest in the
global mean anomaly of reflected solar radiation; this variable affords a good estimate of
radiative effects that can actually be tested against observations. However, unlike RF, this
variable contains effects due to feedbacks (e.g., changes in cloud distributions) so that it is
actually more a signature of the climate response. In the case of the Mt. Pinatubo eruption,
with a peak global visible optical depth of about 0.15, simulations yield a large negative
perturbation as noted above of about -3 W/m2 (Hansen et al., 2002, 049177: Ramachandran et
al., 2000, 192050) (see also Section 9.2 of the IPCC AR4). This modeled estimate of reflected
solar radiation compares reasonably with ERBS observations (Minnis et al., 1993, 190539).
However, the ERBS observations were for a relatively short duration, and the model-
observation comparisons are likely affected by differing cloud effects in simulations and
measurements. It is interesting to note (Stenchikov et al., 2006, 192260) that, in the Mt.
Pinatubo case, the Goddard Institute for Space Studies (GISS) models that use the Sato et al.
(1993, 192046) data yield an even greater solar reflection than the National Center for
Atmospheric Research (NCAR) model that uses the larger (Ammann et al., 2003, 192057)
optical depth estimate.
December 2009 9-155
-------
0.2
0.18
0.16
0.14
Volcanic Aerosol Total Visible Optical Depth
Sato et -g|. (1993)
1870 1880 1800 1900 1810 1920 1830 1040 1950 1960 1970 1960 1990
lima (years)
Figure 9-83. Visible (wavelength 0.55 urn) optical depth estimates of stratospheric SC>42~
aerosols formed in the aftermath of explosive volcanic eruptions that occurred
between 1860 and 2000. Results are shown from two different data sets that have
been used in recent climate model integrations. Note that the Ammann et al.
(2003,192057) data begins in 1890.
Thermal, Dynamical and Chemistry Perturbations Forced by Volcanic Aerosols
Four distinct mechanisms have been invoked with regards to the climate response to
volcanic aerosol RF. First, these forcings can directly affect the Earth's radiative balance and
thus alter surface temperature. Second, they introduce horizontal and vertical heating
gradients; these can alter the stratospheric circulation, in turn affecting the troposphere. Third,
the forcings can interact with internal climate system variability (e.g., El Nino-Southern
Oscillation, North Atlantic Oscillation, Quasi- Biennial Oscillation) and dynamical noise,
thereby triggering, amplifying or shifting these modes (see Section 9.2 of the IPCC AR4)
(Stenchikov et al., 2004, 192274: Yang and Schlesinger, 2001,192270). Fourth, volcanic
aerosols provide surfaces for heterogeneous chemistry affecting global stratospheric ozone
distributions (Chipperfield et al., 2003, 192275) and perturbing other trace gases for a
considerable period following an eruption. Each of the above mechanisms has its own spatial
and temporal response pattern. In addition, the mechanisms could depend on the background
state of the climate system, and thus on other forcings (e.g., due to well-mixed gases) (Meehl
et al., 2004, 192279X or interact with each other.
The complexity of radiative-dynamical response forced by volcanic impacts suggests that
it is important to calculate aerosol radiative effects interactively within the model rather than
prescribe them (Andronova et al., 1999, 192286; Broccoli et al., 2003, 192283). Despite
differences in volcanic aerosol parameters employed, models computing the aerosol radiative
effects interactively yield tropical and global mean lower-strata spheric warmings that are fairly
consistent with each other and with observations (Hansen et al., 2002, 049177: Ramachandran
et al., 2000, 192050: Ramaswamy et al., 2006, 192284: Stenchikov et al., 2004, 192274: Yang
and Schlesinger, 2001, 192270): however, there is a considerable range in the responses in the
December 2009
9-156
-------
polar stratosphere and troposphere. The global mean warming of the lower stratosphere is due
mainly to aerosol effects in the longwave spectrum, in contrast to the flux changes at the TOA
that are essentially due to aerosol effects in the solar spectrum. The net radiative effects of
volcanic aerosols on the thermal and hydrologic balance (e.g., surface temperature and
moisture) have been highlighted by recent studies (Free and Angell, 2002, 192281; Jones et
al, 2003, 192278) (see Chapter 6 of the IPCC AR4). See Chapter 9 (of the IPCC AR4) for
significance of the simulated responses and model-observation comparisons for 20th-century
eruptions. A mechanism closely linked to the optical depth perturbation and ensuing warming
of the tropical lower stratosphere is the potential change in the cross-tropopause water vapour
flux (Joshi and Shine, 2003, 192327) (see Section 2.3.7 of the IPCC AR4).
Anomalies in the volcanic-aerosol induced global radiative heating distribution can force
significant changes in atmospheric circulation, for example, perturbing the equator-to-pole
heating gradient (Ramaswamy et al., 2006, 192273; Stenchikov et al., 2002, 192277) (see
Section 9.2 of the IPCC AR4) and forcing a positive phase of the Arctic Oscillation that in turn
causes a counterintuitive boreal winter warming at middle and high latitudes over Eurasia and
North America (Miller et al., 2005, 192258: Perlwitz and Graf, 2001, 192271: Perlwitz and
Harnik, 2003, 192264: Rind et al., 2005, 192261: Shindell et al., 2003, 192069: Shindell et al.,
2004, 192267: Stenchikov et al., 2002, 192277: Stenchikov et al., 2004, 192274: Stenchikov et
al., 2006, 192260).
Stratospheric aerosols affect the chemistry and transport processes in the stratosphere,
resulting in the depletion of ozone (Brasseur and Granier, 1992, 192256: Chipperfield et al.,
2003, 192275: Solomon et al., 1996, 192252: Tie et al., 1994, 192253). Stenchikov et al.
(2002, 192277) demonstrated a link between ozone depletion and Arctic Oscillation response;
this is essentially a secondary radiative mechanism induced by volcanic aerosols through
stratospheric chemistry. Stratospheric cooling in the polar region associated with a stronger
polar vortex initiated by volcanic effects can increase the probability of formation of polar
stratospheric clouds and therefore enhance the rate of heterogeneous chemical destruction of
stratospheric ozone, especially in the NH (Tabazadeh et al., 2002, 192250). The above studies
indicate effects on the stratospheric ozone layer in the wake of a volcanic eruption and under
conditions of enhanced anthropogenic halogen loading. Interactive microphysics-chemistry-
climate models (Dameris et al., 2005, 192055: Rozanov et al., 2002, 192247: Rozanov et al.,
2004, 192245: Shindell et al., 2003, 192069: Timmreck et al., 2003, 192254)
indicate that aerosol-induced stratospheric heating affects the dispersion of the volcanic
aerosol cloud, thus affecting the spatial RF. However the models' simplified treatment of
aerosol microphysics introduces biases; further, they usually overestimate the mixing at the
tropopause level and intensity of meridional transport in the stratosphere (Douglass et al.,
2003, 057260; Schoeberl et al., 2003, 057262). For present climate studies, it is practical to
utilize simpler approaches that are reliably constrained by aerosol observations.
Because of its episodic and transitory nature, it is difficult to give a best estimate for the
volcanic RF, unlike the other agents. Neither a best estimate nor a level of scientific
understanding was given in the TAR. For the well-documented case of the explosive 1991 Mt.
Pinatubo eruption, there is a good scientific understanding. However, the limited knowledge of
the RF associated with prior episodic, explosive events indicates a low level of scientific
understanding.
9.3.9. Other Special Sources and Effects
International shipping has been identified as an additional source of carbonaceous aerosols.
Simulations with a climate model that included aerosol effects and 3 different emissions inventories
showed that shipping contributed 2.3-3.6% of the total SO42~ atmospheric aerosol content and
0.4-1.4% of the total BC atmospheric aerosol content, based on global means in 2000. This modeling
also showed that aerosol optical thickness over the Indian Ocean, the Gulf of Mexico, and the
northeastern Pacific Ocean varied by 8 to 10%. The corresponding all-sky (that includes both cloudy
and clear skies) direct radiative forcings ranged from -0.011 to -0.013 W/m . The greatest effect of
aerosols emitted from global shipping is likely to be an increase in cloud formation and the resulting
change in reflectivity of shortwave radiation. Aerosols from shipping were estimated to contribute
17-39% of the total anthropogenic aerosol radiation forcing effect.
When BC is deposited to the surface of ice or snow, solar absorption and heating occur at the
surface. This can melt additional snow or ice at the surface and the reflectivity of the surface can
change. Both factors affect aspects of climate. Jacobson (2003, 155866; 2004, 155870) and Jacobson
et al. (2004, 180362) estimated the warming due to fossil fuel BC and organic matter using the Gas,
Aerosol, Transport, Radiation, General Circulation, Mesoscale and Ocean Model
(GATOR-GCMOM). The modeling effort included consideration of the BC cycle, accounting for
emissions, transport, aerosol coagulation, aerosol growth, cloud activation, aerosol-cloud
coagulation, cloud-cloud coagulation, rainout, washout, dry deposition, and processes of precipitated
and dry-deposited BC in snow and sea ice. Results suggested that BC absorption in snow and sea ice
December 2009 9-157
-------
increased near-surface temperatures over a 10-year simulation by about 0.06°K (Jacobson, 2003,
189421).
BC soot is a potentially important agent of climate warming in the Arctic, and northern boreal
wildfires may contribute substantially to this effect. Soot is approximately twice as effective as CO2
in altering surface air temperature, and can reduce sea ice formation and snow albedo (Hansen and
Nazarenko, 2004, 156521).
Kim et al. (2005, 155900) investigated the relationships between northern boreal wildfires and
reductions in Arctic sea ice and glacial coverage. They modeled the FROSTFIRE boreal forest
control burn (Hinzman et al., 2003, 155845) with respect to BC aerosol transport, dispersion, and
deposition. Model results suggested that boreal wildfires could be a major source of BC soot to sea
ice and glaciers in Alaska. This may exacerbate summer melting of sea ice and reduce recruitment of
first-year ice into multiyear ice, thereby leading to an overall reduction in sea ice. Similarly,
increased BC soot on glaciers would be expected to increase summer melting and lead to an overall
reduction in glacial coverage (Kim et al., 2005, 155900).
Jacobson (2002, 155865) proposed, based on model simulations with 12 identifiable effects of
aerosol particles on climate, emission reductions of fossil fuel particulate BC and associated organic
matter could potentially slow warming for a specific period more than reduction of CO2 or CH4 for a
specific period. Jacobson's (2006, 156599) calculations suggested that fossil fuel BC plus organic
matter emissions reductions could eliminate 8-18% of total anthropogenic warming, and 20-45% of
net warming after accounting for aerosol cooling, within a period of 3-5 years (Chock et al., 2003,
155727). See also conflicting discussions (Feichter et al., 2003, 155772; Penner, 2003, 156851); and
further responses (Jacobson, 2003, 155867; Jacobson, 2003, 155868; Jacobson, 2003, 155869;
Penner, 2003, 156851).
Bond and Sun (2005, 156282) reviewed published data regarding the warming potential of
BC, compared with CO2 and other GHG. Climatic effects of GHG are generally compared on the
basis of top-of-the-atmosphere, globally averaged changes in radiative balance. On that basis, BC is
one of the largest individual warming agents, after CO2 and perhaps CH4 (Bond and Sun, 2005,
156282; Jacobson, 2000, 056378; Sato et al., 2003, 156947).
Reddy and Boucher (2007, 156042) conducted an analysis that provided regional estimates of
BC emissions from fossil fuels and biofuels. These estimates indicated that East and Southeast Asia
contributed over 50% of the global BC burden and its associated direct radiative forcing. Europe was
found to be the largest BC contributor in the northern latitudes. The indirect effect of BC deposition
on snow was also estimated to be highest for Europe.
To improve understanding of the role of aerosols in climate forcing, Chung and Seinfeld
(2002, 155732) estimated the global distribution of BC, primary organic particles (those directly
emitted from combustion), secondary organic particles (primary organic compounds partially
oxidized in the atmosphere), and SO42~ aerosols to model the overall radiative forcing of these
groups of compounds. The model was run with the assumption that the BC particles do not combine
with OC or SO42~ particles (termed an external mixture), and with the assumption that the particles
are represented by a core of BC surrounded by a shell of light scattering aerosols. Modeling results
suggested an overall radiative cooling effect from aerosols ranging from -0.39 to -0.78 W/m2.
Roberts and Jones (2004, 156052) used a climate modeling approach to compare possible
effects of BC on climate warming to those attributable to emissions from greenhouse gases. Results
suggested that the warming effect from atmospheric BC aerosols may not be large relative to that
from greenhouse gases. A different modeling approach by Roeckner et al. (2006, 156920) evaluated
the effects of BC and primary OC on climate under two scenarios of carbonaceous aerosol
emissions. In the first scenario, BC and primary OC emissions decreased over Europe and China, but
increased at lower latitudes. In the second scenario, emissions were frozen at 2000 levels. The effects
of both scenarios on mean global temperature were found to be small, but higher aerosol emissions
at low latitudes did result in atmospheric heating and corresponding land surface cooling that led to
increased precipitation and runoff in this simulation.
Study of BC effects in tropical climates was undertaken by Wang (2007, 156147). Substantial
effects of direct radiative forcing by BC on the tropical Pacific were shown in model results that
were similar to the El Nino Southern Oscillation activities both in the nature and scale of effects with
enhancement of the Indian monsoon circulation. The model suggested that atmospheric heating by
radiation absorption by BC can form temperature and pressure anomalies that favor propagation of
convection from western to central and eastern Pacific. More work will be needed to distinguish
between the aerosol signal and natural factors in controlling tropical precipitation in this region.
December 2009 9-158
-------
Table 9-18. Overview of the different aerosol indirect effects and their sign of the net radiative flux
change at the top of the atmosphere (TOA).
Effect
Cloud albedo
effect
Cloud lifetime
effect
Semi-direct effect
Glaciation indirect
effect
Thermodynamic
effect
Cloud Types
Affected
All clouds
All clouds
All clouds
Mixed-phase
clouds
Mixed-phase
clouds
Process
For the same cloud water or ice content more but smaller
cloud particles reflect more solar radiation
Smaller cloud particles decrease the precipitation efficiency
thereby presumably prolonging cloud lifetime
Absorption of solar radiation by absorbing aerosols affects
static stability and the surface energy budget, and may lead to
an evaporation of cloud particles
An increase in IN increases the precipitation efficiency
Smaller cloud droplets delay freezing causing super-cooled
clouds to extend to colder temperatures
Sign of Change
in TOA Radiation
Negative
Negative
Positive or Negative
Positive
Positive or Negative
Potential
Magnitude
Medium
Medium
Small
Medium
Medium
Scientific
Understanding
Low
Very low
Very low
Very low
Very low
Source: Reprinted with Permission of Cambridge University Press from Denman (2007,1563941
Table 9-19. Overview of the different aerosol indirect effects and their implications for the global
mean net shortwave radiation of the surface Fsfc (columns 2-4) and for precipitation
(columns 5-7).
Effect
Cloud albedo effect
Cloud lifetime effect
Semi-direct effect
Glaciation indirect
effect
Sign of Change
in Fsfc
Negative
Negative
Negative
Positive
Potential
Magnitude
Medium
Medium
Large
Medium
Scientific
Understanding
Low
Very low
Very low
Very low
Sign of Change in
Precipitation
n.a.
Negative
Negative
Positive
Potential
Magnitude
n.a.
Small
Large
Medium
Scientific
Understanding
n.a.
Very low
Very low
Very low
Thermodynamic Positive or
effect Negative
Medium
Very low
Positive or Negative
Medium
Very low
Source: Reprinted with Permission of Cambridge University Press from Denman (2007,156394)
There are several other kinds of climate effects from aerosol PM. None is well understood or
well quantified. The semi-direct effect, which involves absorption of solar radiation by soot particles
followed by re-emission as thermal radiation, is expected to heat the air mass and increase its static
stability relative to the surface. The semi-direct effect can also cause evaporation of cloud droplets,
thereby partially offsetting the cloud albedo indirect effect. The glaciation effect involves an increase
in IN, which is expected to cause rapid glaciation of a super-cooled liquid water cloud due to the
differences in vapor pressure over ice and water. Unlike cloud droplets, these ice crystals can quickly
reach precipitation size, with the potential to turn a non-precipitating cloud into a precipitating cloud.
The thermodynamic effect involves a delay in freezing by the smaller cloud droplets, which can
cause super cooled clouds to occur under colder temperatures. The possible consequences to
radiative flux of all of the processes are outlined in Table 9-18 (top of the atmosphere effects) and
Table 9-19 (surface radiative and precipitation effects) (Denman et al., 2007, 156394). though
significant uncertainties remain. Nevertheless, the individual processes cannot be considered in
isolation because of the numerous feedbacks, and because atmospheric aerosol concentrations and
climate are intimately coupled (Denman et al., 2007, 156394; Dentener et al., 2006, 088434).
December 2009
9-159
-------
9.3.9.1. Glaciers and Snowpack
Organic compounds are incorporated into snow by wet and dry deposition processes (Lei and
Wania, 2004, 127880; Roth et al, 2004, 056431). Atmospherically deposited organics appear to be
ubiquitous in snowpacks at appreciable concentrations (Grannas et al., 2007, 156492). Examples
include PAHs, phthalates, alkanes, phenols, low molecular weight carbonyls, POPs, and low
molecular weight organic acids (Halsall, 2004, 155822; Nakamura et al., 2000, 156792; Villa et al.,
2003, 156139). Humic-like substances found in the snowpack may release VOCs into the
atmosphere via photo-oxidation (Grannas et al., 2004, 155803; Grannas et al., 2007, 156492).
Several thousand organic species were identified by Grannas et al. (2006, 155804). based on
molecular weight, from a single ice core collected in Russia. Little information is available, however,
regarding the chemical properties of these chemical constituents. In addition to the diversity of
chemicals that are deposited into the snowpack, there are also biological organisms, including
bacteria and algae. Their role in influencing snow chemistry and volatilization processes is not
understood (Grannas et al., 2007, 156492).
Recent research has explored connections between the atmosphere and the cryosphere (land or
sea covered by snow or ice). A seasonal maximum of 40% of the Earth's land surface is covered by
snow or ice, as well as several percent of the oceans. Particulate deposition to snow and ice surfaces
can affect melting rates. Deposition of PM to glacial ice surfaces can affect the subsequent rate of
melting. A thin cover of debris contributes to accelerated melting. A thicker cover of debris, such as
that which may result from a volcanic eruption, retards melting. The difference is due to the
changing balance between enhanced absorption of shortwave radiation by PM and conductive heat
flow (insulation) through a buildup of material having low heat conduction (Kirkbride and
Dugmore, 2003, 156645).This issue is particularly important for deposition of large quantities of
volcanic material. To a lesser extent, however, the same principles apply to PM deposition derived
from air pollution. Under a thin layer of debris, ablation rates are higher than for clean ice. However,
as the thickness of the debris layer increases, ablation rates systematically decline (Nicholson and
Benn, 2006, 156806). The threshold debris thickness that separates ablation increase from decrease
is site specific and depends on local climate and the nature of the debris particles. Nicholson and
Benn (2006, 156806) presented a surface energy balance model to calculate ice melt beneath a
surface debris layer, based on meteorological data and basic debris characteristics. Modeled melting
rates matched observed rates, suggesting that the model produced useful results.
Long-range atmospheric transport of PM delivers a large fraction of the total input of POPs to
the Arctic region (Halsall, 2004, 155822). These contaminants can accumulate in Arctic food webs
and have become the focus of international research and concern. Nevertheless, fate and transport of
POPs within terrestrial and marine Arctic ecosystems are not well understood and are strongly
affected by the presence of snow and ice. Sea ice provides a barrier to air-water exchange, and this
hinders volatilization and re-emission of previously deposited contaminants (Halsall, 2004, 155822).
Thus, the effects of greenhouse gasses and PM on climate in the Arctic region have feedbacks to
POP fate, transport, and toxicity. The transfer of POPs among the major abiotic environmental
compartments in the Arctic are summarized in Figure 9-84 from Halsall (2004, 155822). Recent
studies detailing rate and transport of POPs are summarized in Table 9-20.
December 2009 9-160
-------
Table 9-20. Recent studies highlighting POP occurrence and fate in the major arctic compartments.
ATMOSPHERE
1 Annual time-series of OC and PCB concentrations in the Norwegian Arctic
2 Long-term analysis of the chlordane-group and their input to the Arctic with changing sources
3 PAH occurrence at monitoring sites across the Arctic, seasonality and gas/particle partitioning
4 PCB occurrence at monitoring sites across the Arctic, spatial differences and seasonality
5 Long-term analysis of PCB and OC trends in the Canadian Arctic and seasonal patterns
6 Trans-Pacific LRAT and impact of Asian sources on the western Canadian Arctic
Oehmeetal. (1996, 1560011
Bidleman et al. (2002, 155691)
Halsalletal. (1997, 155821)
Stern etal. (1997, 1560961
Hung et al. (2001, 155856: 2002, 1558571
Bailey et al. (2000, 1556701
FRESHWATER
7 Annual avg water concentrations in major Russian rivers for selected OC pesticides
8 Long-term (decades) PCB deposition profile in Arctic lake sediments
9 Mass balance of selected DCs in Canadian Arctic lake conducted with data collected over 3 yrs
10 Examining the biodegradation of HCHs in Canadian Arctic watersheds
Alexeeva et al. (2001, 155651)
Muir etal. (1996,155991)
Helm et al. (2002, 155835)
Helm et al. (2000, 155834)
MARINE
1 1 Transport and entry of (3-HCH into western Arctic Ocean via Pacific surface waters
1 2 Occurrence of current use pesticides in air, fog and surface seawater in the western Arctic Ocean
13 Resolving petrogenic and anthropogenic PAH input to marine sediments in coastal Arctic seas
14 Quantifying abiotic and biotic degradation of HCHs in the Arctic Ocean water column
15 PCBs and DCs in surface ocean water— Bering and Chukchi seas
16 Spatial patterns of HCHs and toxaphene in Arctic Ocean surface water
Li et al. (2002, 156691)
Chermyak etal. (1996, 155726)
Yunker etal. (1996, 1561751
Harner et al. (2000, 1558291
Strachan etal. (2001, 1561031
Jantunen and Bidleman (1998, 155877)
SNOW/AIR-FRESHWATER
17 PAHs (and inorganics) in surface snow layers (snowpit) at Summit, Greenland
18 PAHs measured in snow and ice layers on Agassiz ice-cap, Ellesmere Island, Canada
19 Modeling OC behaviour and fate in the surface seasonal snow pack at Amituk Lake, Canada
20 DCs, PCBs and PAHs in snow and ice of the Ob-Yenisey watershed of the Russian Arctic
Masclet et al. (2000, 1559661
Peters etal. (1995, 156856)
Wania etal. (1998, 156148)
Melnikov et al. (2003, 1567531
OCEAN/AIR
21 Transfer of a-HCH across the air/water interface in the western Arctic ocean
22 Calculated seasonality of OC air/water fluxes in the Canadian high Arctic
Jantunen and Bidleman (1996, 155876)
Hargrave etal. (1997, 155827)
OCEAN/ICE
23 Transport potential of contaminants across the Arctic ocean via sea-ice drift
24 The importance of eastern Arctic sea-ice drift as a source of contaminants to the Norwegian sea
Pfirman etal. (1997, 156864)
Korsnes et al. (2002, 156657)
Source: Reprinted with Permission of Elsevier Ltd. from Halsall (2004,15582:
December 2009
9-161
-------
LRAT
Atmosphere
(21-22)
Snow/ice caps
TTTrnn
Terrestrial po>
Catchment r.
rivers
(Ii-24.)
Snow/ sea ice
n
L. (1M6) l>?
p> Suriace E
Ocean m
vl
9
Deep Ocean
\
j
**g
-------
for the BC on snow RF of +0.10 ± 0.10 W/m2, with a low level of scientific understanding
(Section 2.9, Table 2.11, of the IPCC AR4).
9.3.9.3. Effects on Local and Regional Climate
Most effects of PM on climate, as assessed by IPCC (Stohl et al., 2007, 157015) and
summarized in this assessment, focus on global-scale processes and responses. In addition, it is also
possible that PM emissions contribute to local and regional climate changes. These might include
short-term cycles in rainfall or temperature and rainfall suppression, especially near cities and for
orographic precipitation. Rainfall suppression, in particular, is believed to exacerbate water supply
problems which are substantial in many regions, especially in the western U.S.
Aerosol particles, directly and through cloud enhancement, may reduce near-surface wind
speeds locally. Slower winds, in turn, reduce evaporation. The overall impact can be a reduction in
local precipitation. Jacobson and Kaufman (2006, 090942) investigated the effects of PM on
spatially-distributed wind speeds and resulting feedbacks to precipitation using the
GATOR-GCMOM (Jacobson, 2001, 155864) and supporting evidence from satellite data. The study
focused on the South Coast Air Basin (SCAB) in California during February and August, 2002-2004.
The modeled precipitation decrease over land in California was 2% of the baseline 1.5 mm/day due
to emissions of anthropogenic aerosol particle and precursor gasses in the SCAB domain. However,
the reduction over much of the Sierra Nevada, where most precipitation falls, was up to 0.5 mm/day,
or 4-5% of the baseline 10-13 mm/day in that mountainous region (Jacobson, 2006, 156599). The
probable mechanism was described as follows. Aerosol particles and aerosol-enhanced clouds reduce
wind speeds below them by stabilizing the air, reducing the vertical transport of horizontal
momentum. In turn, the reduced wind speeds, and associated reduced evaporation and increased
cloud lifetime, contributes to reduced local and regional precipitation (Jacobson, 2006, 156599).
Effects of air pollution on regional precipitation were quantified by Givati and Rosenfeld
(2004, 156475). They found a 15-25% reduction in the orographic component of precipitation
downwind of major coastal urban areas during the 20th century. Their study focused on
orographically-forced clouds because these short-lived, shallow clouds are expected to exhibit the
largest effect of air pollution on precipitation. Substantially larger precipitation suppression due to
aerosol participate pollution was found between Fresno and Sacramento in California by Givati and
Rosenfeld (2004, 156475). Precipitation losses over topographical barriers in the Sierra Nevada
amounted to 15-25% of the annual precipitation at elevations less than 2,000 m. This precipitation
suppression occurred mainly in the relatively shallow orographic clouds within the cold air mass of
cyclones. The suppression that occurred on the upslope side of the mountains was coupled with
similar percentage (but lower absolute volume) enhancement on the drier downslope eastern side
(Givati and Rosenfeld, 2004, 156475). Similar results were found in studies by Griffith et al. (2005,
156497). Jirak and Cotton (2006, 156612). Rosenfeld and Givati (2006, 156924). and Rosenfeld et
al. (2007, 156057). At all of these study locations (California, Israel, Utah, Colorado, China),
orographic precipitation decreased by 15-30% downwind of pollution sources, likely due to creation
of more and smaller cloud droplets and resulting suppression of precipitation.
The study of Givati and Rosenfeld (2004, 156475) was the first to quantify the microphysical
effect of mesoscale precipitation. Following the findings of Givati and Rosenfeld (2004, 156475).
the effects of aerosol air pollution on precipitation at high elevation sites in the Front Range of
Colorado adjacent to urban areas were investigated by Jirak and Cotton (2006, 156612).
Examination of precipitation trends showed that the ratio of upslope precipitation during easterly
flows at high elevation west of Denver and Colorado Springs to the upwind urban sites decreased by
about 30% over the past half century. These results provide further support for the hypothesis that
aerosol pollution suppresses orographic precipitation downwind of pollution source areas.
Griffith et al. (2005, 156497) found similar reductions in mountainous precipitation in Utah,
downwind of Salt Lake City and Provo. The ratio of precipitation at mountain stations located in
rural settings in Utah and Nevada remained stable, supporting the hypothesis that air pollution
decreases Ro (the ratio of precipitation at the downwind site to precipitation at the upwind pollution
source) over the mountains to the east of Salt Lake City.
Rosenfeld and Givati (2006, 156924) extended the investigation of the suppression of
precipitation by aerosol pollutants to a larger scale by examining the ratio between precipitation
amounts over the hills to precipitation over upwind lowland areas throughout the western U.S. from
the Pacific Coast to the Rocky Mountains. They found in these paired analyses a pattern of
December 2009 9-163
-------
decreasing precipitation by as much as 24% from the Mexican border to central California, with no
decrease in northern California and Oregon and smaller decrease of 14% in Washington east of
Seattle and Puget Sound. Similar decreases were found over Arizona and New Mexico (Rosenfeld,
2006, 190233). Utah (Griffith et al., 2005, 156497). and the east slope of the Colorado Rockies
(Jirak and Cotton, 2006, 156612).
Suppression of winter orographic precipitation appears to occur up to hundreds of kilometers
inland of coastal urban areas (Rosenfeld, 2006, 190233). Decreases in this precipitation ratio
occurred during winter orographic precipitation, but not during convective summer precipitation
over the same mountain ranges. This finding agrees with the expectation that aerosol-induced
changes in the rate of precipitation formation would cause a decrease in precipitation from shallow
and short-lived orographic clouds, but not necessarily from deeper and longer-lived thermally-driven
convective clouds.
Results of these studies of aerosol effects on orographic precipitation suggest that
human-caused air pollution, and fine particles in particular, have had a large effect on precipitation
well beyond the local scales of the pollution sources (Rosenfeld, 2006, 190233).
9.3.10.Summary of Effects on Climate
Aerosols affect climate through direct and indirect effects. The direct effect is primarily
realized as planet brightening when seen from space because most aerosols scatter most of the
visible spectrum light that reaches them. The IPCC AR4 reported that the radiative forcing from this
direct effect was -0.5 (±0.4) W/m2 and identified the level of scientific understanding of this effect as
'Medium-low'. The global mean direct radiative forcing effect from individual components of
aerosols was estimated for the first time in the IPCC AR4 where they were reported to be (all in
W/m2 units): -0.4 (±0.2) for sulfate, -0.05 (±0.05) for fossil fuel-derived organic carbon, +0.2 (±0.15)
for fossil fuel-derived black carbon, +0.03 (±0.12) for biomass burning, -0.1 (±0.1) for nitrates, and
-0.1 (±0.2) for mineral dust. Global loadings of anthropogenic dust and nitrates remain very
troublesome to estimate, making the radiative forcing estimates for these constituents particularly
uncertain.
Numerical modeling of aerosol effects on climate has sustained remarkable progress since the
time of the last PM AQCD, though model solutions still display large heterogeneity in their estimates
of the direct radiative forcing effect from anthropogenic aerosols. The clear-sky direct radiative
forcing over ocean due to anthropogenic aerosols is estimated from satellite instruments to be on the
order of-1.1 (±0.37) W/m2 while model estimates are -0.6 W/m2. The models' low bias over ocean is
carried through for the global average: global average direct radiative forcing from anthropogenic
aerosols is estimated from measurements to range from -0.9 to -1.9 W/m2, larger than the estimate of
-0.8 W/m2 from the models.
Aerosol indirect effects on climate are primarily realized as an increase in cloud brightness
(termed the 'first indirect' or Twomey effect), changes in precipitation, and possible changes in cloud
lifetime. The IPCC AR4 reported that the radiative forcing from the Twomey effect was -0.7 (range:
-1.1 to +4) and identified the level of scientific understanding of this effect as 'Low' in part owing to
the very large unknowns concerning aerosol size distributions and important interactions with
clouds. Other indirect effects from aerosols are not considered to be radiative forcing.
December 2009 9-164
-------
Taken together, direct and indirect effects from aerosols increase Earth's shortwave albedo or
reflectance thereby reducing the radiative flux reaching the surface from the Sun. This produces net
climate cooling from aerosols. The current scientific consensus reported by IPCC AR4 is that the
direct and indirect radiative forcing from anthropogenic aerosols computed at the top of the
atmosphere, on a global average, is about -1.3 (range: -2.2 to -0.5) W/m2. While the overall global
average effect of aerosols at the top of the atmosphere and at the surface is negative, absorption and
scattering by aerosols within the atmospheric column warms the atmosphere between the Earth's
surface and top of the atmosphere. In part, this is owing to differences in the distribution of aerosol
type and size within the vertical atmospheric column since aerosol type and size distributions
strongly affect the aerosol scattering and reradiation efficiencies at different altitudes and
atmospheric temperatures. And, although the magnitude of the overall negative radiative forcing at
the top of the atmosphere appears large in comparison to the analogous IPCC AR4 estimate of
positive radiative forcing from anthropogenic GHG of about +2.9 (± 0.3) W/m2, the horizontal,
vertical, and temporal distributions and the physical lifetimes of these two very different radiative
forcing agents are not similar; therefore, the effects do not simply off-set one another.
Overall, the evidence is sufficient to conclude that a causal relationship exists between
PM and effects on climate, including both direct effects on radiative forcing and indirect
effects that involve cloud feedbacks that influence precipitation formation and cloud
lifetimes.
9.4. Ecological Effects of PM
9.4.1. Introduction
PM is heterogeneous with respect to chemical composition and size; therefore, it can cause a
variety of ecological effects, which have been previously described by the U.S. EPA (2004, 056905)
and by Grantz et al. (2003, 155805). Atmospheric PM has been defined, for regulatory purposes,
mainly by size fractions and less clearly so in terms of chemical nature, structure, or source. Both
fine and coarse-mode particles may affect plants and other organisms; however, PM size classes do
not necessarily relate to ecological effects (U.S. EPA, 1996, 079380). More often the chemical
constituents drive the ecosystem response to PM (Grantz et al., 2003, 155805).
The previous PM assessment (U.S. EPA, 2004, 056905) included the acidifying effects of
particulate N and S. The 2008 NOXSOX ISA (U.S. EPA, 2008, 157074) assessed the effects of
particle- and gas-phase N and S pollution on acidification, N enrichment, and Hg methylation.
Acidification of ecosystems is driven primarily by deposition resulting from SOX, NOX, and NHX
pollution. Acidification from the deposition resulting from current emission levels causes a cascade
of effects that harm susceptible aquatic and terrestrial ecosystems, including slower growth and
injury to forests and localized extinction of fishes and other aquatic species. In addition to
acidification, atmospheric deposition of reactive N resulting from current NOX and NHX emissions
along with other non-atmospheric sources (e.g., fertilizers and wastewater), causes a suite of
ecological changes within sensitive ecosystems. These include increased primary productivity in
most N-limited ecosystems, biodiversity losses, changes in C cycling, and eutrophication and
harmful algal blooms in freshwater, estuarine, and ocean ecosystems. In some watersheds, additional
SO42~ from atmospheric deposition increases Hg methylation rates by increasing both the number
and activity of S-reducing bacteria. Methylmercury is a powerful toxin that can bioaccumulate to
toxic amounts in food webs at higher trophic levels.
This assessment of PM effects on ecosystems considers both direct and indirect exposure
pathways. Atmospheric PM may affect ecological receptors directly following deposition on surfaces
or indirectly by changing the soil chemistry or by changing the amount of radiation reaching the
Earth's surface. Indirect effects acting through the soil are often thought to be most significant
because they can alter nutrient cycling and inhibit nutrient uptake (U.S. EPA, 2004, 056905;
U.S. EPA, 2008, 157074). The U.S. EPA (2004, 056905) reported that the effects of PM can be both
chemical and physical. Physical effects of particle deposition on vegetation may include abrasion
and radiative heating. However, chemical effects may be more significant (U.S. EPA, 2008, 157074).
December 2009 9-165
-------
In general, anthropogenic stressors can result in damaged ecosystems that do not recover
readily (Odum, 1993, 076742; Rapport and Whitford, 1999, 004595). Ecosystems sometimes lack
the capacity to adapt to an anthropogenic stress and are unable to maintain their normal structure and
functions unless the stressor is removed (Rapport and Whitford, 1999, 004595). These stresses
result in a process of ecosystem degradation marked by a decrease in biodiversity, reduced primary
and secondary production, and a lower capacity to recover and return to the original ecosystem state.
In addition, there can be an increased prevalence of disease, reduced nutrient cycling, increased
dominance of exotic species, and increased dominance by smaller, short-lived opportunistic species
(Odum, 1985, 039482; Rapport and Whitford, 1999, 004595).
Ecosystems are often subjected to multiple stressors, of which atmospheric PM deposition is
only one. Additional stressors are also important, including O3 exposure, climatic variation, natural
and human disturbance, the occurrence of invasive non-native plants, native and non-native insect
pests, disease, acidification, and eutrophication among others. PM deposition interacts with these
other stressors to affect ecosystem patterns and processes.
The possible effects of particulate (and other) air pollutants on ecosystems have been
categorized by Guderian (1977, 004150) as follows:
• accumulation of pollutants in plants and other ecosystem components (such as soil and
surface- and groundwater),
• damage to consumers as a result of pollutant accumulation,
• changes in species diversity because of shifts in competition,
• disruption of biogeochemical cycles,
• disruption of stability and reduction in the ability to self-regulate,
• breakdown of stands and associations, and
• expansion of denuded zones.
The general conclusion of the last PM assessment (U.S. EPA, 2004, 056905) was that
ecosystem response to PM can be difficult to determine because the changes are often subtle. For
example, changes in the soil may not be observed until pollutant deposition has occurred for many
decades, except in the most severely polluted areas around heavily industrialized point sources. The
presence of co-occurring pollutants generally makes it difficult to attribute ecological effects to PM
alone or to one constituent in the deposited PM. In other words, the potential for alteration of
ecosystem function and structure exists but can be difficult to quantify except in cases of extreme
amounts of deposition, especially when there are other pollutants present in the ambient air that may
produce additive or synergistic responses.
New information on the ecological effects of coarse and fine particle PM is presented in the
following discussion in the context of effects that were known from the last PM AQCD (U.S. EPA,
2004, 056905). The general effects of the chemical constituents of PM are discussed; however, a
rigorous assessment of each chemical constituent (e.g., Hg, Cd, Pb, etc.) is not given. Both direct and
indirect effects will be discussed and the strength of the scientific evidence will be evaluated using
the causality framework.
9.4.1.1. Ecosystem Scale, Function, and Structure
Information presented in this section was collected at multiple scales, ranging from the
physiology of a given species to population, community, and ecosystem-level investigations. For this
assessment, "ecosystem" is defined as a functional entity consisting of interacting groups of living
organisms and their abiotic (chemical and physical) environment. Ecosystems cover a hierarchy of
spatial scales and can comprise the entire globe, biomes at the continental scale, or small, well-
circumscribed systems such as a small pond.
Ecosystems have both structure and function. Structure may refer to a variety of measurements
including the species richness, abundance, community composition and biodiversity as well as
December 2009 9-166
-------
landscape attributes. Competition among and within species and their tolerance to environmental
stresses are key elements of survivorship. When environmental conditions are shifted, for example,
by the presence of anthropogenic air pollution, these competitive relationships may change and
tolerance to stress may be exceeded. "Function" refers to the suite of processes and interactions
among the ecosystem components and their environment that involve nutrient and energy flow as
well as other attributes including water dynamics and the flux of trace gases. Plant processes
including photosynthesis, nutrient uptake, respiration, and C allocation, are directly related to
functions of energy flow and nutrient cycling. The energy accumulated and stored by vegetation (via
photosynthetic C capture) is available to other organisms. Energy moves from one organism to
another through food webs, until it is ultimately released as heat. Nutrients and water can be
recycled. Air pollution alters the function of ecosystems when elemental cycles or the energy flow
are altered. This alteration can also be manifested in changes in the biotic composition of
ecosystems.
There are at least three levels of ecosystem response to pollutant deposition: (1) the individual
organism and its environment; (2) the population and its environment; and (3) the biological
community composed of many species and their environment (Billings, 1978, 034165). Individual
organisms within a population vary in their ability to withstand the stress of environmental change.
The response of individual organisms within a population is based on their genetic constitution, stage
of growth at time of exposure to stress, and the microhabitat in which they are growing (Levine and
Pinto, 1998, 029599). The range within which organisms can exist and function determines the
ability of the population to survive. Those able to cope with the stresses survive and reproduce.
Competition among different species results in succession (community change over time) and,
ultimately, produces ecosystems composed of populations of species that have the capability to
tolerate the stresses (Guderian, 1985, 019325: Rapport and Whitford, 1999, 004595). Available
information on individual, population and community response to PM will be discussed.
9.4.1.2. Ecosystem Services
Ecosystem structure and function may be translated into ecosystem services. Ecosystem
services identify the varied and numerous ways that ecosystems are important to human welfare.
Ecosystems provide many goods and services that are of vital importance for the functioning of the
biosphere and provide the basis for the delivery of tangible benefits to human society. Hassan et al.
(2005, 092759) define these to include supporting, provisioning, regulating, and cultural services:
• Supporting services are necessary for the production of all other ecosystem services.
Some examples include biomass production, production of atmospheric O2, soil
formation and retention, nutrient cycling, water cycling, and provisioning of habitat.
Biodiversity is a supporting service that is increasingly recognized to sustain many of the
goods and services that humans enjoy from ecosystems. These provide a basis for three
higher-level categories of services.
• Provisioning services, such as products (Gitay et al., 2001, 092761), i.e., food (including
game, roots, seeds, nuts and other fruit, spices, fodder), fiber (including wood, textiles),
and medicinal and cosmetic products (including aromatic plants, pigments).
• Regulating services that are of paramount importance for human society such as (a) C
sequestration, (b) climate and water regulation, (c) protection from natural hazards such
as floods, avalanches, or rock-fall, (d) water and air purification, and (e) disease and pest
regulation.
• Cultural services that satisfy human spiritual and aesthetic appreciation of ecosystems
and their components.
9.4.2. Deposition of PM
Deposition of PM is discussed in Chapter 3.3.4. Additional material specifically related to
ecosystems is discussed in this section.
December 2009 9-167
-------
9.4.2.1. Forms of Deposition
Research summarized by the previous NAAQS PM assessment illustrated the complexity of
deposition processes. Airborne particles, their gas-phase precursors, and their transformation
products are removed from the atmosphere by wet and dry deposition processes. These deposition
processes transfer PM pollutants to other environmental media where they can alter the structure,
function, diversity, and sustainability of complex ecosystems. Dry deposition of PM is most effective
for coarse particles. These include primary geologic materials and elements such as iron and
manganese. By contrast, wet deposition is more effective for fine particles of secondary atmospheric
origin and elements such as cadmium, chromium, lead, nickel, and vanadium (Reisinger, 1990,
046737: Smith, 1990, 084015; U.S. EPA, 2004, 056905). The relative magnitudes of the different
deposition modes vary with ecosystem type, location, elevation, and chemical burden of the
atmosphere (U.S. EPA, 2004, 056905). There are differences in the deposition behavior of fine and
coarse particles. Coarse particles generally settle nearer their site of formation than do fine particles.
In addition, the chemical constitution of individual particles is correlated with size. For example,
much of the base cation and heavy metal burden is present on coarse particles.
Fine PM is often a secondary pollutant that forms within the atmosphere, rather than being
directly emitted from a pollution source. It derives from atmospheric gas-to-particle conversion
reactions involving nucleation, condensation, and coagulation, and from evaporation of water from
contaminated fog and cloud droplets. Fine PM may also contain condensates of VOCs, volatilized
metals, and products of incomplete combustion, including poly cyclic aromatic hydrocarbons (PAH)
and BC (soot) (U.S. EPA, 2004, 056905).
Fine PM may act as a carrier for materials such as herbicides that are phytotoxic. Fine PM
provides much of the surface area of particles suspended in the atmosphere, whereas coarse PM
provides much of the mass of airborne particles. Surface area can influence ecological effects
associated with the oxidizing capacity of fine particles, their interactions with other pollutants, and
their adsorption of organic compounds. Fine and coarse particles also respond to changes in
atmospheric humidity, precipitation, and wind, and these can alter their deposition characteristics.
Coarse PM is mainly a primary pollutant, having been emitted from pollution sources as fully
formed particles derived from abrasion and crushing processes, soil disturbances, desiccation of
marine aerosol emitted from bursting bubbles, hygroscopic fine PM expanding with humidity to
coarse mode, and/or gas condensation directly onto preexisting coarse particles. Suspended primary
coarse PM may contain iron, silica, aluminum, and base cations from soil, plant and insect
fragments, pollen, fungal spores, bacteria, and viruses, as well as fly ash, brake lining particles, and
automobile tire fragments. Coarse mode particles can be altered by chemical reactions and/or
physical interactions with gaseous or liquid contaminants.
Exposure to a given mass concentration of PM may lead to widely differing phytotoxic and
other environmental outcomes depending upon the particular mix of PM constituents involved.
Especially important in this regard are S and N components of PM, which are addressed in the 2008
NOXSOX ISA, and effects of particulate heavy metals and organic contaminants. This variability has
not been characterized adequately. Though effects of specific chemical fractions of PM have been
described to some extent, there has been relatively little research aimed at defining the effects of
unspeciated PM on plants or ecosystems.
9.4.2.2. Components of PM Deposition
Trace Metals
Atmospheric deposition can be the primary source of some metals to some watersheds. Metal
inputs can include the primary crustal elements (Al, Ca, K, Fe, Mg, Si, Ti) and the primary
anthropogenic elements (Cu, Zn, Cd, Cr, Mn, Pb, V, Hg). The crustal elements are derived largely
from weathering and erosion, whereas the anthropogenic elements are derived from combustion,
industrial sources, and other man-made sources (Goforth and Christoforou, 2006, 088353).
Heavy metal deposition to ecosystems depends on their location as well as upwind emissions
source strength. The deposition velocity tends to be dependent on particle size and chemical species.
Larger particles deposit more efficiently than smaller particles. Heavy metals preferentially associate
December 2009 9-168
-------
with fine particles. Fine particles also have the longest atmospheric residence times. Depending on
climate and topography, fine particles may remain airborne for days to months and may be
transported thousands of kilometers from their source.
Ecosystems immediately downwind of major heavy metal emissions sources may receive
locally heavy dry deposition. Trace element investigations conducted in roadside, industrial, and
urban environments have also shown that substantial amounts of particulate heavy metals can
accumulate on surfaces.
A significant trace metal component of PM is mercury (Hg). Hg is toxic and can move readily
through environmental compartments. Atmospheric and depositional inputs of Hg include both
natural and anthropogenic sources. Natural geologic contributions to Hg in the environment include
geothermal and volcanic activity, geologic metal deposits, and organic-rich sedimentary rocks. These
natural emissions combine with anthropogenic emissions from such sources as power plants,
landfills, sewage sludge, mine waste, and incineration (Gustin, 2003, 155816; Schroeder and
Munthe, 1998, 014559). Emissions from natural sources are controlled by geologic features,
including substrate Hg content, rock type, the degree of hydrothermal activity, and the presence of
heat sources (Gustin, 2003, 155816). The significance of natural Hg sources relative to
anthropogenic sources varies geographically. For example, Nevada occurs within a global
mercuriferous belt, with area emissions about three times higher than the value assumed for global
modeling (Gustin, 2003, 155816). In Nevada, natural and anthropogenic Hg emissions are
approximately equal (Gustin, 2003, 155816).
The U.S. EPA (1997, 157066) compiled an assessment of the sources and environmental
effects of Hg in the U.S. A variety of factors were found to influence Hg deposition, fate and
transport (Table 9-21). Such factors relate in particular to speciation of the Hg that is emitted, the
forms in which it is deposited from the atmosphere, and transformations that occur within the
atmosphere and within the aquatic, transitional, and terrestrial compartments of the receiving
watershed. There have been studies that have reconstructed, from lake sediment records, the
atmospheric depositional history of trace metals and PAHs in lakes adjacent to coal-fired power
plants. For example, Donahue et al. (2006, 155751) analyzed sediment from Wababun Lake, which
is located in Alberta, Canada in proximity (within 35 km) to 4 power plants built since 1950. Trace
metal concentrations of Hg, Cu, Pb, As, and Se in lake sediment increased by 1.2- to 4-fold. The
total PAH flux to surface sediments was 730-1,100 ug/m2/yr, which was two to five times higher
than in 2 lakes situated 20 km to the north and 70 km to the south. Further discussion of Hg effects
on ecosystems can be found in Section 9.4.5.
December 2009 9-169
-------
Table 9-21. Factors potentially important in estimating Hg exposure.
Factor
Importance and Possible Effect on Mercury Exposure
Type of anthropogenic source of mercury
Mercury emission rates from stack
Different combustion and industrial process sources are anticipated to have different local scale impacts due to
physical source characteristics (e.g., stack height), the method of waste generation (e.g., incineration or mass
burn) or mercury control devices and their effectiveness.
Increased emissions will result in a greater chance of adverse impacts on environment.
Mercury species emitted from stack
More soluble species will tend to deposit closer to the source.
Form of mercury emitted from stack
Transport properties can be highly dependent on form.
Deposition differences between vapor and
particulate-bound mercury
Vapor-phase forms may deposit significantly faster than particulate-bound forms.
Transformations of mercury after emission from
source
Relatively nontoxic forms emitted from source may be transformed into more toxic compounds.
Transformation of mercury in watershed soil
Reduction and revolatilization of mercury in soil limits the buildup of concentration.
Transport of mercury from watershed soils to
water body
Mercury in watershed soils can be a significant source to water bodies and subsequently to fish.
Transformation of mercury in water body
Reduction, methylation, and demethylation of mercury in water bodies affect the overall concentration and the
MHg fraction, which is bioaccumulated in fish.
Facility locations
Effects of meteorology and terrain may be significant.
Location relative to local mercury source
Receptors located downwind are more likely to have higher exposures. Influence of distance depends on source
type.
Contribution from non-local sources of mercury Important to keep predicted impacts of local sources in perspective.
Uncertainty
Reduces confidence in ability to estimate exposure accurately.
Source: Modified from U.S. EPA (1997,157066)
Organics
Organic compounds that may be associated with deposited PM include persistent organic
pollutants (POPs), pesticides, SOCs, polyaromatic hydrocarbons (PAHs) and flame retardants among
others. Organic compounds partition between gas and particle phases, and organic particulate
deposition depends largely on the particle sizes available for adsorption (U.S. EPA, 2004, 056905).
Dry deposition of organic materials is often dominated by the coarse fraction. Gas-particle phase
interconversions are important in determining the amount of dry deposition.
Most persistent organic pollutants (POPs) enter the biosphere via human activities, including
synthetic pesticide application, output of poly chlorinated dibenzo dioxins (PCDD) from incinerators,
and accidental release of PCBs from transformers (Lee, 2006, 088968). Once they are introduced
into the environment, their accumulation and magnification in biological systems are determined by
physiochemical properties and environmental conditions (Section 9.4.6). Uptake by plants can occur
at the soil/plant interface and at the air/plant interface (Krupa et al., 2008, 198696). For lipophilic
POPs, such as PCDDs and PCBs, the air/plant response route generally dominates (Lee, 2006,
088968; Thomas et al., 1998, 156118). but uptake through above-ground plant tissue also occurs. In
a study of zucchini (Cucurbita pepo), Lee et al. (2006, 088968) found chlordane pesticide
components in all vegetation tissues examined: root, stem, leaves, fruits.
Many pesticides and SOCs are carcinogenic or estrogenic and pose potential threats to aquatic
and terrestrial biota. Although deposition of SOCs was previously reported for the Sierra Nevada
Mountains in California and the Rocky Mountains in Colorado, little was previously known about
the occurrence, distribution, or sources of SOCs in alpine, sub-Arctic, and Arctic ecosystems in the
western U.S. The snowpack is efficient at scavenging of both particulate and gas phase pesticides
from the atmosphere (Halsall, 2004, 155822: Lei and Wania, 2004, 127880). Analysis of pesticides
in snowpack samples from seven NPs in the western U.S. by Hageman et al. (2006, 156509)
illustrated the deposition and fate of 47 pesticides and their degradation products. Correlation
analysis with latitude, temperature, elevation, PM, and two indicators of regional pesticide use
suggested that regional patterns in historic and current agricultural practices are largely responsible
December 2009
9-170
-------
for the distribution of pesticides in the NPs. Pesticide deposition to parks in Alaska was attributed to
long-range atmospheric transport.
PAHs include hundreds of different compounds that are characterized by possessing two or
more fused benzene rings. They are widespread contaminants in the environment, and are formed by
incomplete combustion of fossil fuels and other organic materials. Eight PAHs are considered
carcinogenic and 16 are classified by EPA as priority pollutants. They are common air pollutants in
metropolitan areas, derived from vehicular traffic and other urban sources. Especially high
concentrations have been found near Soderberg aluminum production industries and areas where
heating during winter via wood burning is common. Other sources, in addition to gasoline and diesel
engines, include forest fires and various forms of fossil fuel combustion (Sanderson and Farant,
2004, 1569421
The behavior of PAHs is strongly determined by their chemical characteristics, especially their
nonpolarity and hydrophobicity. They readily adsorb to particulates in the air and to sediments in
water. Srogi (2007, 180049) provided a thorough review of PAH concentrations in various
environmental compartments and their use for assessing environmental risks and possible effects on
ecosystems and human health.
Deposition and fate of PAH has been an important area of research. Because they are
carcinogenic, PAHs are important environmental contaminants. Root-soil behavior of PAHs is an
area of active study. Soil-bound PAHs are associated with soil organic matter and are therefore
generally not easily available for root uptake. PAHs are readily adsorbed to root surfaces but there
seems to be little movement to the interior of the root or movement up to the shoots (Gao and Zhu,
2004, 155782). Paddy rice is the main food crop planted in China. As an aquatic plant having aerial
roots, the movement of PAHs into rice roots may be different than their movement into more widely
studied land-grown food crops. PAH concentrations in the rice roots were more correlated with the
water and air compartments than with the soil (Jiao et al, 2007, 155879).
The total PAH concentration in grasses adjacent to a highway have been measured to be about
eight times higher than in grasses from reference sites not close to a highway (Crepineau et al., 2003,
155741). Howe et al. (2004, 155854) found that concentrations of PAHs and hexachlorobenzene
(HCB) in spruce (Picea spp.) needles at 36 sites in eastern Alaska varied by an order of magnitude.
Samples collected near the city of Fairbanks generally had higher concentrations than samples
collected from rural areas. The relative importance of combustion sources versus petrogenic sources
was highest in the near-coastal areas, as reflected in variation in the concentration of ratios of
isomeric PAHs.
Use of flame retardants has increased in recent years in response to fire product safety
regulations. However, some flame retardant chemicals are toxic and are readily transported
atmospherically. Use of some has been banned in Europe and some of the United States because of
their persistence and tendency to bioaccumulate (Hoh et al., 2006, 190378).
Base Cations
With respect to ecosystem effects from PM deposition, the inclusion of base cations
(especially Ca, Mg, and K) in atmospheric deposition is generally considered to be a positive effect.
Base cations are important plant nutrients that are in some locations present in short supply and that
are further depleted by the acidic components of deposition. Increased base cation deposition can
help to ameliorate adverse effects of acidification of soils and surface waters and reduce the toxicity
of inorganic Al to plant roots and aquatic biota. These topics are covered in detail in the recent 2008
NOxSOx ISA (U.S. EPA, 2008, 157074).
Although the effects of base cation deposition inputs to terrestrial ecosystems are most
commonly considered to be positive, under very high base cation deposition, plant health can be
adversely affected. Dust that is high in base cations can settle on leaves and other plant structures
and remain for extended periods of time. This is especially likely in arid environments because
rainfall can serve to wash dry deposited materials off the foliage. Extended dust coverage can result
in a variety of adverse impacts on plant physiology (Grantz et al., 2003, 155805). For example, van
Heerden et al. (2007, 156131) documented decreased chlorophyll content, inhibition of CO2
assimilation, and uncoupling of the oxygen-evolving complex in desert shrubs exposed to high
limestone dust deposition near a limestone quarry in Namibia.
December 2009 9-171
-------
Based on the Integrated Forest Study (IPS) data, the U.S. EPA (2004, 056905) concluded that
particulate deposition has a greater effect on base cation inputs to soils than on base cation losses
associated with the inputs of sulfur, nitrogen, and H+. These atmospheric inputs of base cations have
considerable significance, not only to the base cation status of these ecosystems, but also to the
potential of incoming precipitation to acidify or alkalize the soils in these ecosystems. This topic is
discussed in detail in the recent NOXSOX ISA (U.S. EPA, 2008, 157074).
9.4.2.3. Magnitude of Dry Deposition
Using Vegetation for Estimating Atmospheric Deposition
Whereas direct real-time measurement of deposition or air concentrations of atmospheric
contaminants is desirable, it is not always practical (Howe et al., 2004, 155854). Instead, passive
time-integrative methods are frequently used. These can involve analysis of vegetative tissues as a
record of pollutant exposure, or analysis of lake sediment cores or ice cores to determine changes in
pollutant input over time. There is a general assumption that the concentration of an analyte in
vegetation reflects the time-integrated concentration of that analyte in the air. The development of
deposition layers in sediment or ice cores allows the possibility of determining the effects of changes
in the atmospheric concentration over periods of years, decades, or longer.
Biomonitoring methods are important in air pollution assessment and provide a complement
for more typical instrumental analyses. It is well known that mosses can accumulate large amounts
of heavy metals in response to atmospheric deposition. Mosses accumulate dissolved materials and
PM deposited from the atmosphere and have been used extensively in Europe as surrogate collectors
for estimating bulk (wet plus dry) deposition of metals. The ease and low cost of this method has
enabled regional assessments to be conducted throughout Europe.
Despite its wide use, however, several papers have pointed out complications in the use of
mosses to quantify metal deposition rates. Zechmeister (1998, 156178) found that the uptake
efficiency for 12 heavy metals in three species of moss was similar, but that uptake efficiency in a
fourth species was uncorrelated with the other species for about half the metals considered.
Zechmeister (1998, 156178) also showed that productivity of an individual species can vary greatly
among sites. To calculate atmospheric deposition of metals from accumulation in mosses, both the
metal concentration and the rate of biomass production is needed. Further complication was shown
in the study of Shakya et al. (2008, 156081). which revealed that accumulation of Cu, Zn and Pb
decreased chlorophyll content. Sites with greater deposition amounts may therefore have lower rates
of productivity than cleaner sites.
Differences in uptake efficiencies among species and productivity among sites has led to the
use of a single moss species placed in mesh bags that can be distributed to areas where that species
of moss does not grow naturally. Studies to standardize this passive deposition monitoring approach
have been limited. Adamo et al. (2007, 155644) evaluated the effects of washing with water, oven
drying, and acid washing as preteatments and found little difference in uptake efficiencies, although
the ratio of the collecting surface area to mass was found to be a key factor in uptake efficiency.
Couto et al. (2004, 155739) investigated dry versus bulk deposition of metals using
transplanted moss bags. This study showed that at some sites dry deposition exceeded bulk
deposition, a likely outcome of wash-off of dry deposited particles. This study also documented
intercationic displacement and leaching as a result of acidic precipitation. The authors concluded that
the accumulated metal concentration represented an unstable equilibrium between inputs and outputs
of elements that were a function of the local environment and weather during the exposure period.
They also concluded that it was not possible to extrapolate calibrations between metal accumulation
in moss and atmospheric deposition of metals to areas with different weather conditions,
precipitation pH, and air contaminant concentrations. Zechmeister et al. (2003, 157175) also
presented results demonstrating the problems with dry deposited particles that can be washed off by
rain. These studies indicate that moss is not a completely effective collector of total particle
deposition. Deposition estimates from moss accumulation probably represent values that fall
between wet deposition and total deposition.
A European moss biomonitoring network has been in place since 1990 (Harmens et al., 2007,
155828). Sampling surveys are repeated every five years. The survey conducted in 2005/2006
December 2009 9-172
-------
occurred in 32 countries at over 7,000 sites. The network reports metal concentrations associated
with live moss tissue. Trends analysis of these data showed statistically significant decreases over
time in moss concentrations for As, Cu, V, and Zn. Trends were not observed for Cr, Fe, or Ni.
Results for individual countries participating in the survey have also been published. In Hungary,
major pollution sources were readily detected by moss sampling (Otvos et al., 2003, 156831).
Somewhat higher metal concentrations in mosses in 1997 than in other European countries were
attributed to the use of a different moss species in the Hungarian survey (Otvos et al., 2003, 156831).
Similar sampling in Romania showed regions with contamination that were among the highest in
Europe. These results were consistent with known air quality problems in Romania (Lucaciu et al.,
2004, 155947). Because particulate deposition is not well characterized using this method, spatial
patterns and temporal trends for particulate metal deposition in Europe only provide crude estimates
of relative deposition patterns.
The use of moss to assess heavy metal deposition has received much less attention in the U.S.
than in Europe. A study conducted in the Blue Ridge Mountains, VA, found that metal concentrations
in moss were related to elevation and canopy species at some sites (Schilling and Lehman, 2002,
113075). However, metal concentrations in moss were not related to concentrations in the O horizon
of the soil. Other measurement methods for trace metal deposition were not available to compare
with moss concentrations.
Epiphytic lichens have also been used to evaluate heavy metal accumulation. Helena et al.
(2004, 155833) found substantially increased concentrations of metals in lichens transplanted from a
relatively clean region to an area in proximity to a metal smelter. The presence of specific species of
bryophyte or lichen can serve as an effective bioindicator of metal contamination (Cuny et al., 2004,
155742). In some studies, tree bark has been used as a biomonitor for atmospheric deposition of
heavy metals (Baptista et al., 2008, 155673: Pacheco and Freitas, 2004, 156011: Rusu et al., 2006,
156062).
Biomonitoring using mosses, lichens, or other types of vegetation has been established as a
means of identifying spatial patterns in atmospheric deposition of heavy metals in relation to power
plants, industry, and other point and regional emissions sources. More recently, a number of studies
(Lopez et al., 2002, 155943: 2003, 155944: 2003, 155945) have used cattle that have been reared
predominantly on local forage as a means of monitoring atmospheric inputs of Cu, Ar, Zn, and Hg.
For example, Hg emissions from coal fired power plants in Spain had a substantial effect on Hg
accumulation by calves (Lopez et al., 2003, 155944). Accumulation of Hg by cattle extended to
-140-200 km downwind from the source.
Yang and Zhu (2007, 156168) investigated the effectiveness of pine needles as passive air
samplers for SOCs, such as PAHs, that are partially or completely particle-associated in the
atmosphere. PAH distribution patterns are complicated by their properties, which span a broad range
of octanol-air partition coefficients. This allows them to be present in both vapor and particle phases.
In addition, the air-plant partitioning of PAHs is affected by air temperature and atmospheric stability
(Krupa et al., 2008, 198696: Yang and Chen, 2007, 092847). DeNicola et al. (2005, 155747)
documented the suitability of a Mediterranean evergreen oak (Quercus ilex) to serve as a passive
biomonitor for atmospheric contamination with PAH in Italy.
Deposition to Canopies
Tree canopies have been shown to increase dry deposition from the atmosphere, including
deposition of PM. Dry deposition rates in the canopy are commonly estimated by the difference
between throughfall deposition and deposition measured by an open collector, although the use of
this approach to specifically quantify particulate deposition is complicated by gaseous deposition to
leaf surfaces and, for some elements, leaching and uptake. Avila and Rodrigo (2004, 155664) found
that trace metal deposition in throughfall in a Spanish oak forest were higher than bulk deposition for
Cu, Pb, Mn, V, and Ni, but not for Cd and Zn. This study also found that dry deposition of Cu, Pb,
Zn, Cd and V occurred, but that canopy uptake of Zn and Cd also occurred. Leaching of Mn and Ni
from the foliage was observed as well. Leaching of Ni, Cu, Mn, Rb, and Sr from a red spruce-balsam
fir canopy by acidic cloud water was also measured in a study by Lawson et al. (2003, 089371).
These studies suggest that leaching of trace metals from forest canopies varies with tree species and
the acidity of precipitation. Throughfall therefore cannot be assumed to represent total deposition of
heavy metals without evaluating uptake and leaching at the specific study site. Physical models have
December 2009 9-173
-------
provided an alternative to estimating dry deposition to canopies with throughfall measurements.
Recently, Pry or and Binkowski (2004, 116805) identified an additional complication in that models
typically hold particle size constant. Nevertheless, there may be significant modification of particle
size distributions during the deposition process.
The use of pine and oak canopies as bioindicators of atmospheric trace metal pollution was
investigated by Aboal et al. (2004, 155642). As an ecosystem pool, metals in leaves were likely to be
much more important than those in mosses. The authors concluded, however, that these tree species
were not effective bioindicators of atmospheric deposition of heavy metals. Metal concentrations in
leaves were found to be one to three orders of magnitude lower than in mosses collected in this
study.
The effectiveness of tree canopies in capturing particulates was investigated as a method for
improving air quality by Freer-Smith et al. (2004, 156451). This study showed that with
consideration of planting design, location of pollution source, and tree species, planting of trees can
be effective at reducing particulate air pollution. However, this approach does not address the
possible effects of the captured pollution on trees, soils and surface waters.
High-elevation forests generally receive larger particulate deposition loadings than equivalent
low elevation sites. Higher wind speeds at high elevation enhance the rate of aerosol impaction.
Orographic effects enhance rainfall intensity and composition and increase the duration of occult
deposition. High-elevation forests are often dominated by coniferous species with needle-shaped
leaves that enhance impaction and retention of PM delivered by all three deposition modes.
Deposition to Soil
As with mosses, accumulation of heavy metals in surface soils provides a general reflection of
the spatial distribution of industrial pollution. The distribution of toxic elements in urban soils has
been an important area of study (Madrid et al., 2002, 155956; Markiewicz et al., 2005, 155963).
Generally, Cu, Pb, Zn, and Ni have accumulated in urban soils compared with their rural
counterparts (Yuangen et al., 2006, 156174). In the study of Romic and Romic (2003, 156055).
relationships were found between urban activities and concentrations of metals in soils in developed
areas surrounding Zagreb, Croatia. Goodarzi et al. (2002, 155801) compared deposition estimated by
moss bags to concentrations of metals in A-horizon soils in the vicinity of a large smelter.
Statistically significant correlations were observed between the moss bag deposition estimates and
the soil metal concentrations for Cd, Pb, Zn, and in some cases also Cu. These correlations suggested
that atmospheric deposition of metals caused elevated metal concentrations in the upper mineral
horizon of these soils. No correlations were found for Hg or As in this study.
Studies have also been conducted to assess metal accumulation in peat because of the tendency
of most metals to be immobilized through binding with organic matter. Steinnes et al. (2005,
156095) presented geographical patterns of metal concentrations in surface peat throughout Norway
that corresponded to pollution sources, although the peat samples were collected in 1979. Zaccone et
al. (2007, 179930) found that variations of metal concentrations with depth in a single Swiss peat
core corresponded with the depositional history that would be expected from the industrial
revolution, although Cs137 activity exhibited a distribution in the profile that was not fully consistent
with the Chernobyl nuclear reactor accident. A detailed study of Finnish peat showed that
relationships between depth profiles of metal concentrations and deposition history can match well
for some metals at some sites, but not well for the same metals at other sites (Roberts et al., 2003,
156051). They also found that Zn and Cd accumulation rates were independent of deposition history
at each of three study sites.
Metal deposition to soil is also a significant concern adjacent to roadways. Urban stormwater
can be rich in heavy metals and other contaminants derived from atmospheric deposition, and can be
a major source of pollutant inputs to water bodies in urban settings. Urban stormwater runoff can
also be toxic to aquatic biota, partly due to trace metal concentrations (Greenstein et al., 2004,
155808: Sabin et al., 2005, 088300: Schiff et al., 2002, 156959). These processes are largely a
function of the impervious nature of much of the ground surface in urban areas (i.e., buildings, roads,
sidewalks, parking lots, construction sites). Dry-deposited pollutants can build up, especially in arid
and semi-arid environments, and then be washed into surface waters with the first precipitation
event. The concentrations of Cd, Ca, Cu, Pb, and Zn in road runoff were found to be significantly
higher during winter in Sweden. This seasonal pattern was attributed to the intense wearing of the
December 2009 9-174
-------
pavement that occurred during winter due to the use of studded tires in combination with chemical
effects of deicing salts (Backstrom et al., 2003, 156242).
9.4.3. Direct Effects of PM on Vegetation
Exposure to airborne PM can lead to a range of phytotoxic responses, depending on the
particular mix of deposited particles. This was well known at the time of the previous PM criteria
assessment, as summarized below. Most direct phytotoxic effects occur in severely polluted areas
surrounding industrial point sources, such as limestone quarries and other mining activities, cement
kilns, and metal smelting facilities (U.S. EPA, 2004, 056905). Experimental application of PM
constituents to foliage typically elicits little response at the more common ambient concentrations.
The diverse chemistry and size characteristics of ambient PM and the lack of clear distinction
between effects attributed to phytotoxic particles and to other air pollutants further confound the
understanding of the direct effects on foliar surfaces.
Deposition of PM can cause the accumulation of heavy metals on vegetative surfaces. Low
solubility limits foliar uptake and direct heavy metal toxicity because trace metals must be brought
into solution before they can enter into the leaves or bark of vascular plants. In those instances when
trace metals are absorbed, they are frequently bound in leaf tissue and are lost when the leaf drops
off (Hughes, 1981. 053595).
Depending on the size of the particles, the PM deposited on the leaf surface can affect the
plant's metabolism and photosynthesis by blocking light, obstructing stomatal apertures, increasing
leaf temperature and altering pigment and mineral content (Naidoo and Chirkoot, 2004, 190449)
(Section 9.4.3.1.). Fine PM has been shown to enter the leaf through the stomata and penetrate into
the mesophyll layers where it alters leaf chemistry (Da et al., 2006, 190190). Kuki et al. (2008,
155346) also showed increased leaf permeability and increased activity of enzymes in response to
fine PM lead (Section 9.4.5.).
Studies of the direct toxic effects of particles on vegetation have not yet advanced to the stage
of reproducible exposure experiments. In general, phytotoxic gases are deposited more readily,
assimilated more rapidly, and lead to greater direct injury of vegetation as compared with most
common particulate materials. The dose-response functions obtained in early experiments following
the exposure of plants to phytotoxic gases generally have not been observed following the
application of particles (U.S. EPA, 2004, 056905).
9.4.3.1. Effects of Coarse-mode Particles
The current state of scientific knowledge regarding the direct effects of coarse PM on plants
has not changed since publication of the previous PM criteria assessment (U.S. EPA, 2004, 056905).
The summary provided here is taken from that report. In many rural areas and some urban areas, the
majority of the mass in the coarse particle mode derives from the elements Si, Al, Ca, and Fe,
suggesting a crustal origin as fugitive dust from disturbed land, roadways, agriculture tillage, or
construction activities. Large particles tend to deposit near their source (Grantz et al., 2003, 155805)
and rapid sedimentation of coarse particles tends to restrict their direct effects on vegetation largely
to roadsides and forest edges, which often receive the greatest deposition (U.S. EPA, 2004, 056905)
Dust
Dust can cause both physical and chemical effects. Consequences are often mediated via
impacts on leaf cuticles and waxes. Deposition of inert PM on above-ground plant organs sufficient
to coat them with a layer of dust may result in changes in radiation received, a rise in leaf
temperature, and the blockage of stomata. Crust formation can reduce photosynthesis and the
formation of carbohydrates needed for normal growth, induce premature leaf-fall, damage leaf
tissues, inhibit growth of new tissue, and reduce starch storage. Dust may decrease photosynthesis,
respiration, and transpiration; and it may result in the condensation and reactivity of gaseous
pollutants with PM, thereby causing visible injury symptoms and decreased productivity (U.S. EPA,
2004, 056905). Leaves with trichomes may be more prone to the accumulation of dust on leaf
surfaces (Kuki et al., 2008, 155346).
December 2009 9-175
-------
The chemical composition of PM is usually the key phytotoxic factor leading to plant injury.
For example, cement-kiln dust liberates calcium hydroxide on hydration. It can then penetrate the
epidermis and enter the mesophyll, causing an increase in leaf surface pH. In turn, surface pH can be
important for surface microbial colonization and wax formation and degradation.
Salt
Sea-salt particles can serve as nuclei for the absorption and subsequent reaction of other
gaseous and particulate air pollutants. Direct effects on vegetation reflect these inputs and salt injury
caused by the sodium and chloride that constitute the bulk of these particles. The source of most salt
spray near the coast is aerosolized ocean water. Sea salt can cause damage to plants; however, it is
not covered in this assessment because it is not of anthropogenic origin. However particulate salt
may be input to an ecosystem from deicing salt.
Injury to vegetation from the application of deicing salt is caused by salt spray blown or
drifting from the highways (Viskari and Karenlampi, 2000, 019101). The most severe injury is often
observed nearest the highway. Conifers planted near roadway margins in the eastern U.S. often
exhibit foliar injury due to toxic amounts of saline aerosols deposited from deicing solutions
(U.S. EPA, 2004, 056905).
9.4.4. PM and Altered Radiative Flux
The effects of PM on radiative flux and the subsequent effects on vegetation have been
described in Section 4.2.3.2 of the previous PM assessment (U.S. EPA, 2004, 056905): a brief
overview is presented below. Atmospheric PM can affect ambient radiation, which can be considered
in both its direct and diffuse components. Foliar interception by canopy elements occurs for both up-
and down-welling radiation. Therefore, the effect of atmospheric PM on atmospheric turbidity
influences canopy processes both by radiation attenuation and by changing the efficiency of
radiation interception in the canopy through conversion of direct to diffuse radiation (Hoyt, 1978,
046638). Diffuse radiation is more uniformly distributed throughout the canopy and increases
canopy photosynthetic productivity by distributing radiation to lower leaves. The enrichment in
photosynthetically active radiation (PAR) present in diffuse radiation may offset a portion of the
effect of an increased atmospheric albedo due to atmospheric particles. Mercado et al. (2009,
190444) estimated the effects of variations in diffuse light on the terrestrial carbon sink during the
last century using a global model. The results indicated that the terrestrial carbon sink increased by
approximately 25% during the "global dimming'" period (1950-1980), likely driven by increased
diffuse light despite decreased PAR. However, under a future scenario in which SO42~ and BC
aerosols decline, the diffuse-radiation and the associated terrestrial C sink also decline (Mercado et
al., 2009, 190444).
The effects of regional haze on the yield of crops because of reduction in solar radiation were
examined by Chameides et al. (1999, 011184) in China, where regional haze is especially severe.
Based on model results, it was estimated that approximately 70% of crops were being depressed by
at least 5-30% by regional scale air pollution and its associated haze (Chameides et al., 1999,
011184: U.S. EPA, 2004, 056905).
9.4.5. Effects of Trace Metals on Ecosystems
Trace metals may enter the ecosystems as both fine and coarse particles. All but 10 of the 90
elements that comprise the inorganic fraction of the soil occur at concentrations of <0.1%
(1,000 ug/g) and are termed "trace" elements or trace metals. Trace metals with a density greater
than 6 g/cm3, referred to as "heavy metals" (e.g., Cd, Cu, Pb, Cr, Hg, Ni, Zn), are of particular
interest because of their potential toxicity to plants and animals. Although some trace metals are
essential for vegetative growth or animal health, they are all toxic in large quantities. Most trace
elements exist in the atmosphere in particulate form as metal oxides (Ormrod, 1984, 046892).
Aerosols containing trace elements derive predominantly from industrial activities. Generally, only
the heavy metals Cd, Cr, Ni, and Hg are released from stacks in the vapor phase (McGowan et al.,
1993, 046731). Atmospherically deposited PM can interact with a variety of biogeochemical
December 2009 9-176
-------
processes. The potential pathways of accumulation of trace metals in terrestrial ecosystems, as well
as the possible consequences of trace metal deposition on ecosystem functions, are summarized in
Figure 9-85 (U.S. EPA, 2004, 056905). A number of mass balance approaches (Macleod et al., 2005,
155954; Toose and Mackay, 2004, 156123). and metal speciation and transport models (Bhavsar et
al., 2004, 155689: Bhavsar et al., 2004, 155690: Gandhi et al., 2007, 155781) have been developed
in recent years.
Atmospheric Pb is a component of PM in some regions. The effects of Pb on ecosystems were
discussed in the 2006 Pb AQCD (U.S. EPA, 2006, 090110). which concluded that, due to the
deposition of Pb from past human practices (e.g., leaded gasoline, ore smelting) and the long
residence time of Pb in many aquatic and terrestrial ecosystems, a legacy of environmental Pb
burden exists, over which is superimposed much lower contemporary atmospheric Pb loadings. The
potential for ecological effects of the combined legacy and contemporary Pb burden to occur is a
function of the bioavailability or bioaccessibility of the Pb. This, in turn, is highly dependent upon
numerous site factors (e.g., soil OC content, pH, water hardness). Although the more localized
ecosystem impacts observed around smelters are often striking, effects generally cannot be attributed
solely to Pb, because of the presence of many other stressors (e.g., other heavy metals, oxides of
sulfur and nitrogen) that can also act singly or in concert with Pb to cause readily observable
environmental effects (U.S. EPA, 2004, 056905: U.S. EPA, 2008, 157074).
Effects of fine particle trace elements were described by the U.S. EPA (2004, 056905). and
some additional more recent research has also been conducted, especially on the topic of vegetative
uptake of trace elements from the soil. The state of scientific understanding as presented by the U.S.
EPA (2004, 056905) as well as a discussion of more recent research findings are presented below.
December 2009 9-177
-------
1. Wet/Dry Deposition
Atmosphere
9. Retranslocation
3. Litterfall, Resuspension,
Deposition, Leaching,
Stem Flow
Plant Surface
Phyllosphere
2. Foliar Uptake
Above-
Ground
Storage,
Metabolism
4. Translocation
Biologically
Unavailable
IX.
Soil Organic
X.
Primary
Minerals
Biologically / /
Available / /
IV.
Upper Soil
5. Mass Flow,
Diffusion
.10. Root
Turnover
7. Leaching
VII.
Lower Soil
V.
Rhizosphere/
Rhizoplane
^-
6. Root"
Uptake
VI.
Root Storage,
Metabolism
5. Mass Flow,
Diffusion
X 7. Leaching
VIII.
Groundwater
Source: U.S. EPA (2004, 0569051
Figure 9-85. Relationship of plant nutrients and trace metals with vegetation. Compartments
(roman numerals) represent potential storage sites; whereas arrows (Arabic
numerals) represent potential transfer routes.
9.4.5.1. Effects on Soil Chemistry
Trace metals are naturally found in small amounts in soils, ground water, and vegetation.
Many are essential micronutrients required for growth by plants and animals. Naturally occurring
mineralization can produce metal concentrations in soils and vegetation that are high compared to
atmospheric sources. Many metals are bound by chemical processes in the soil, reducing their
availability to biota. However, epiphytic or parasitic root colonizing microorganisms can solubilize
and transport metals for root uptake (Lingua et al., 2008, 155935). It can be difficult to assess the
extent to which observed heavy metal concentrations in soil are of anthropogenic origin. This is
because soil parent material, pedogenesis, and anthropogenic inputs all influence the amounts and
distribution of trace elements in soil. Trace element concentrations in some natural soils that are
remote from air pollution can be higher than soils derived from other parent materials that receive
anthropogenic inputs (Burt et al., 2003, 155709). The general effects of metals from atmospheric
deposition are presented in the following discussion.
There is not a standard method available for quantifying the bioavailability of heavy metals in
soil. A variety of models, isotopic studies, and sequential extraction methods have been used (Collins
et al., 2003, 155737: Feng et al., 2005, 155774; Shan et al., 2003, 156972). Total metal concentration
in soil does not give a good indication of potential biological effects because soils vary in their
December 2009
9-178
-------
ability to bind metals in forms that are not bioavailable. There are various methods available for
assessing bioavailability of metals, but soils are heterogeneous and there is no ideal method for
evaluating what conditions the soil biota experience. Almas et al. (2004, 155654) argued that the
actual measurement of biological effects is the best criterion for determining bioavailability. In
particular, the replacement of metal-sensitive microorganisms by metal-tolerant organisms within
each functional group may be one of the most sensitive indicators of metal exposure. An increase in
microbial trace metal tolerance per se would not be problematic if it was not for the fact that this
increase in tolerance is generally accompanied by a decrease in microbial diversity (Almas et al.,
2004, 155654; Lakzian et al., 2002, 156671).
Heavy metals deposited from the atmosphere to forests accumulate either in the organic forest
floor or in the upper mineral soil layers and metal concentration tends to decrease with soil depth.
The accumulation of heavy metals in soil is influenced by a variety of soil characteristics, including
pH, Fe and Al oxide content, amount of clay and organic material, and cation exchange capacity
(CEC) (Hernandez et al., 2003, 155841). Thus, the pattern of distribution of heavy metals in soils
depends on both the soil characteristics and the metal characteristics.
Burt et al. (2003, 155709) investigated the concentrations and chemical forms of trace metals
in smelter-contaminated soils collected in the Anaconda and Deer Lodge Valley area of Montana,
one of the major mining districts of the world for over a century (1864-1983). The relative
distributions of trace metals within the more soluble soil extraction forms were similar to their
respective total concentrations. This suggested a relationship between the concentrations of total
trace elements and concentrations of soluble mobile fractions. Sequential extractions do not provide
direct characterization of trace metal speciation, but rather an indication of chemical reactivity (Burt
et al., 2003, 155709; Ramos et al., 1994, 046736). Soluble and exchangeable forms are considered
readily mobile and bioavailable. Those bound to clay minerals or organic matter are considered
generally unavailable.
There is concern that Pb contamination of forest soil could move into groundwater. This would
be important in view of the large quantity of Pb deposited from the atmosphere in the 1960s and
1970s in response to combustion of leaded gasoline. This issue was investigated by Watmough et al.
(2004, 077809) who applied a stable isotope (207Pb) to the forest floors of white pine (Pinus strobus)
and sugar maple (Acer saccharum) stands. Added Pb was rapidly lost from the forest floor, likely
due to high litter turnover in these forest types. However, Pb concentrations in the upper 30 cm of
mineral soil were strongly correlated with soil OM, suggesting that Pb does not readily move down
the soil profile to the ground water, but rather is associated with the organic content of the upper soil
layers (Watmough et al., 2004, 077809).
The upper soil layers are typically an active site of litter decomposition and plant root uptake,
both processes may be affected by metal components of PM. Surface litter decomposition is reduced
in soils having high metal concentrations. This is likely due to the sensitivity to metals of microbial
decomposers and reduced palatability of plant litter having high metal concentration (Johnson and
Hale, 2008, 155881). Root decomposition is a key component of nutrient cycling. Johnson and Hale
(2008, 155881) measured in situ fine root decomposition at Sudbury, Ontario and Rouyn-Noranda,
Quebec. Elevated soil metal concentrations (Cu, Ni, Pb, Zn) did not necessarily reduce fine root
decomposition. Only at sites having high concentrations of metals did decomposing roots show
increased metal concentrations over time.
9.4.5.2. Effects on Soil Microbes and Plant Uptake via Soil
Upon entering the soil environment, PM pollutants can alter ecological processes of energy
flow and nutrient cycling, inhibit nutrient uptake, change ecosystem structure, and affect ecosystem
biodiversity. Many of the most important effects occur in the soil. The soil environment is one of the
most dynamic sites of biological interaction in nature. It is inhabited by microbial communities of
bacteria, fungi, and actinomycetes. These organisms are essential participants in the nutrient cycles
that make elements available for plant uptake. Changes in the soil environment that influence the
role of the bacteria and fungi in nutrient cycling determine plant and ultimately ecosystem response.
Many of the major indirect plant responses to PM deposition are chiefly soil-mediated and
depend on the chemical composition of the individual components of deposited PM. Effects may
result in changes in biota and in soil conditions that affect ecological processes, such as nutrient
cycling and uptake by plants.
December 2009 9-179
-------
The soil environment is rich in biota. Bacteria and fungi are usually most abundant in the
rhizosphere, the soil around plant roots that all mineral nutrients must pass through. Bacteria and
fungi benefit from the nutrients that are present in root exudates and make mineral nutrients
available for plant uptake. The soil-mediated ecosystem impacts of PM are largely determined by
effects on the growth of bacteria and mycorrhizal fungi that are involved in nutrient cycling and
plant nutrient uptake.
Soil Nutrient Cycling
Accumulation of heavy metals in litter can interfere with nutrient cycling. Microorganisms are
responsible for decomposition of organic matter, which contributes to soil fertility. Toxic effects on
the microflora can be caused by Zn, Cd, and Cu. The U.S. EPA (2004, 056905) judged that addition
of only a few mg of Zn per kg of soil can inhibit sensitive microbial processes. Enzymes involved in
the cycling of N, P, and S (especially arylsulfatase and phosphatase) seem to be most affected
(Kandeler et al., 1996, 094392).
Soil organic matter cycling is known to be sensitive to disturbance due to heavy metal
pollution. This can cause increased litter accumulation at sites close to metal emissions point
sources. The relative importance of the various processes that might be responsible for this
observation is poorly known. Boucher et al. (2005, 155699) conducted CO2 evolution studies in
microcosms having metal-rich and metal-poor plant materials. Their results suggested that there was
a pool of less readily decomposable C that appeared to be preferentially preserved in the presence of
high metal (Zn, Pb, Cd) concentrations in the leaves of the metallophyte Arabidopsis halleri. An
additional possibility is that increased lignification of the cell walls increased the amount of
insoluble C (Mayo et al., 1992, 155974).
Yuangen et al. (2006, 156174) found that urban soil basal respiration rates were positively
correlated with soil acetic acid-extractable Cd, Cu, Ni, and Zn. The soil microbial biomass was
negatively correlated with the concentrations of Pb fractions, but not with other metals. Overall
microbial biomass was lower for urban soils as compared with rural soils (Yuangen et al., 2006,
156174).
Metal Toxicity to Microbial Communities
It is believed that increased accumulation of litter in metal-contaminated areas is due to the
effects of metal toxicity on microorganisms. Smith (1991, 042566) reported the effects of Cd, Cu,
Ni, and Zn on the symbiotic activity of fungi, bacteria, and actinomycetes. In particular, the
formation of mycorrhizae has been shown to be reduced when Zn, Cu, Ni, and Cd were added to the
soil.
Most studies of the effects of heavy metals on soils have been conducted under laboratory
conditions. However, Oliveira and Pampulha (2006, 156827) performed a field study to evaluate
long-term changes in soil microbiological characteristics in response to heavy metal contamination.
Dehydrogenase activity, soil ATP content, and enumeration of major soil microbial groups illustrated
the effects of contamination. There was a marked decrease in total numbers of the different microbial
groups. In particular, asymbiotic nitrogen-fixers and heterotrophic bacteria were found to be
sensitive. Dehydrogenase activity was confirmed to be a good assay for determining the effect of
heavy metals on physiologically active soil microbial biomass.
The toxic effects of heavy metals on soil microorganisms are well known. However, less is
known about the relative sensitivity of different types of soil microorganisms (Rajapaksha et al.,
2004, 156035). Vaisvalavicius et al. (2006, 157080) assessed the toxicity of high concentrations of
Pb (839 mg/kg), Zn (844 mg/kg), and Cu (773 mg/kg) in the upper 0-0.1 m soil layer. Microbial
abundance of all groups was reduced and enzymatic activity was lower than for uncontaminated soil.
In particular, actinomycetes, oligonitrophobic and mineral N assimilating bacteria were most
affected.
Effects of heavy metals in soil on microbes depends on soil pH, organic content, and the type
of heavy metal exposure (Kucharski and Wyszkowska, 2004, 156662). Some studies have shown
that heavy metals inhibit microbial activity in soil (Smejkalova et al., 2003, 156987; Vasundhara et
al., 2004, 156133). However, Wyszkowska et al. (2007, 179948) showed that heavy metals can either
December 2009 9-180
-------
inhibit or stimulate the growth of soil microbes. Populations of Azotobacter spp. decreased, but
populations of oligotrophic and copiotrophic bacteria, actinomyces, and fungi increased in response
to heavy metal exposure. Acute metal stress causes a decrease in microbial biomass as
metal-sensitive microbes are inhibited (Joynt et al., 2006, 155887).
Studies of the impacts of metal stress on the microbial community composition in soil have
generally been based on microbial culturing techniques that can select only a subset of the natural
soil population of microbes. More recent culture-independent studies have been conducted using
phospholipids or nucleic acid biomarkers to reveal information regarding changes in microbial
community structure (Joynt et al., 2006, 155887). Using this approach, Joynt et al.(2006, 155887)
demonstrated that soils contaminated with both metals (Pb, Cr) and organic solvent compounds over
a period of several decades had undergone changes in community composition, but still contained a
phytogenetically diverse group of bacteria. This may reflect adaptation to the potentially toxic
conditions through such processes as natural selection, gene exchange, and immigration.
Comparison between a severely contaminated soil with a similar soil that had much lower amounts
of contamination showed considerably lower microbial diversity in the contaminated soil,
particularly for asymbiotic nitrogen fixers and heterotrophic bacteria (Oliveira and Pampulha, 2006,
156827).
As pollution increases, it is expected that the more sensitive species will be lost and the more
tolerant species remain. This gives rise to the concept of pollution-induced community tolerance
(PICT), which has been demonstrated for populations of bacteria and fungi (Davis et al., 2004,
155744). These researchers assessed the effects of long-term Zn exposure on the metabolic diversity
and tolerance to Zn of a soil microbial community across a gradient of Zn pollution. PICT was found
to correlate better with total soil Zn than with the concentration of Zn in soil pore water.
Soil Microbe Interactions with Plant Uptake of Metals
Atmospherically-deposited metals accumulate in upper soil horizons where fine roots are most
developed. The availability for plant uptake of metals in soil depends on metal speciation and soil
pH. In addition, metal binding to dissolved organic matter (DOM) reduces bioavailability (Sauve,
2001, 156948). Because organic matter typically decreases with soil depth, the affinity of metals for
organic matter can influence metal bioavailability at different soil depths. Fine roots (<2 mm
diameter) provide the major site of uptake and transport to the above-ground plant and generally
contain a large proportion of the total metals found in plants (Gordon and Jackson, 2000, 155802).
Fine roots are often colonized by mycorrhiza and interact with other soil microbes. Recent
published evidence supports that mycorrizha and bacteria influence plant uptake and tolerance of
metals. Mycorrhiza are fungi that colonize plant roots to form a symbiosis. Mycorrhiza take up
nutrients from the soil and transfer them to the plant in exchange for carbon from the plant. Like
plants, some species and strains of mycorrhiza are more tolerant of metals in the soil (Ray et al.,
2005, 190473). so that unpolluted and polluted sites may host different species and strains of
mycorrhiza (Vogel-Mikus et al., 2005, 190501).
Mycorrhiza have been observed to cause a range of effects on plants. In some cases, plants
colonized with mycorrhiza showed improved nutrient uptake and decreased metal uptake (Berthelsen
et al., 1995, 078058; Nogueira et al., 2004, 190460; Vogel-Mikus et al., 2006, 190502). Mycorrhiza
have been shown to accumulate metals and act as a sink (Berthelsen et al., 1995, 078058; Carvalho
et al., 2006, 155715) often preventing the metals in the roots from allocation to shoots (Kaldorf et
al., 1999, 190399; Scares and Siqueira, 2008, 190482; Zhang et al., 2005, 192083). For example,
estuarine salt marshes are often located close to urban and industrial areas and receive elevated
amounts of trace metal contaminants from point and non-point (including atmospheric deposition)
sources. Vegetation is important in the retention and accumulation of heavy metals in salt marshes.
Carvalho et al. (2006, 155715) conducted experiments on the effects of arbuscular mycorrhizal fungi
(AMF) on the uptake of Cd and Cu by Aster tripolium, a common plant species in polluted salt
marshes and a host of AMF. Carvalho et al. (2006, 155715) found that AMF colonization increased
metal accumulation in the root system of Aster tripolium without enhancing translocation to the
shoot. By trapping toxic metals in the roots, this plant species may reduce the extent of vegetative
stress caused by metal exposure and act as an effective sink for these metals. In a review paper
Christie et al. (2004, 190174) concluded that mycorrhiza may directly improve plant tolerance to
December 2009 9-181
-------
metals by binding and immobilizing metals and indirectly improve plant tolerance by improving
uptake of nutrients that increase plant growth.
There is recent evidence that bacteria and mycorrhiza act together to improve plant tolerance
to metals. Like mycorrhiza some bacteria are more tolerant to metals than others (Vivas et al., 2003,
190499). Combined inoculation of Trifolium sp. by the arbuscular mycorrhiza, Glomus mosseae, and
the bacterium, Brevivacillus sp., conferred tolerance to Cd by increasing nutrient status and rooting
development and by decreasing Cd uptake by the plant (Vivas et al., 2003, 190499). A similar result
was observed for Zn uptake (Vivas et al., 2006, 190500).
In some cases, mycorrhizae will not prevent metal uptake (Weissenhorn et al., 1995, 073826).
In fact, mycorrhiza may facilitate the accumulation of metals in plants and enhance the translocation
of metals from the root to the shoot (Citterio et al., 2005, 190176; Vogel-Mikus et al., 2005, 190501;
Zimmer et al., 2009, 192085). There is evidence of variable responses depending on the combination
of mycorrhiza and bacteria species. Zimmer et al. (2009, 192085) recently showed that the willow
tree (Salix sp.) colonized with the ectomycorrhizal fungus, Hebeloma crustuliniforme, and the
bacteria, Microcuccus Inters, increased total Cd and Zn accumulation due to enhanced mycorrhizal
formation. In these cases where soil microbes cause increased metal accumulation, there is a
potential to use the system for phytoremediation.
Plants also vary in the extent to which they take up heavy metals from the soil. Variability has
been shown to occur in response to different plant species and different metals. For example, Szabo
and Fodor (2006, 156109) exposed winter wheat (Triticum aestivum), maize (Zea mays) and
sunflower (Helianthus annuus) to a variety of micro-pollutants. Cadmium accumulation was
significant in both vegetative and reproductive plant parts. Vegetative winter wheat accumulated
substantial amounts of Hg, but the other species did not. Lead, Cu, and Zn showed only moderate
enrichment in crops (Szabo and Fodor, 2006, 156109).
There is some evidence to support that shallow-rooted plant species are most likely to take up
metals from the soil (Martin and Coughtrey, 1981, 047727). However, there is little evidence
confirming this observation. It may be more likely that shallow roots of species are likely to take up
metals because the metal often accumulates in shallow soil layers. Even though atmospheric PM will
usually deposit on soils before being taken up by plants, it could also be deposited to aquatic systems
with subsequent transfer to terrestrial plants. Contamination of stream sediments by heavy metals
can impact adjacent terrestrial ecosystems when high flows cause resuspension and subsequent
streamside deposition of sediment particles. For example, Ozdilek et al. (2007, 156010) showed that
metal concentrations in vegetation along the Blackstone River in Massachusetts and Rhode Island
were generally inversely related to the distance from the riverbank, with higher metal concentrations
in plant tissues located near the river. The ability of plants to take up metals from soil is an important
part of metal cycling in the environment. This uptake process allows the metals to enter the food
web, where they might exert mutagenic, carcinogenic, and teratogenic effects (Hunaiti et al., 2007,
156579).
9.4.5.3. Plant Response to Metals
Some metals, including Cu, Co, Ni, and Zn, are essential micronutrients needed for plant
growth. Others, including Hg, Cd, and Pb are not essential for plants. Though all heavy metals can
be directly toxic at sufficiently high concentrations, only Cu, Ni, and Zn have been documented as
being frequently toxic to plants (U.S. EPA, 2004, 056905). while toxicity due to Cd, Co, and Pb has
been observed less frequently (Smith, 1990, 046896). Toxic doses depend on the type of ion, ion
concentration, plant species and the stage of plant growth (Memon and Schroder, 2009, 190442).
Toxicity response is also dependent on the nutritional status of the plant and the development of
mycorrhizae (Strandberg et al., 2006, 156105). Plants respond to high concentrations of metals in
soil through a variety of mechanisms and there are substantial differences among plant species in
their response to heavy metal exposure. Mechanisms of metal tolerance included exclusion or
excretion rates, genetics (Patra et al., 2004, 081976; Yang et al., 2005, 192104). mycorrhizal
interactions (Gohre and Paszkowski, 2006, 190355). storage capability and accumulation (Clemens,
2006, 190179). and various cellular detoxification mechanisms (Gratao et al., 2005, 190364; Hall,
2002, 190365).
One of the most important mechanisms that increases plant tolerance to metals is chelation
with phytochelatins, such as metallothioneins and peptide ligands that are synthesized within the
plant from glutathione (Memon and Schroder, 2009, 190442). Phytochelatins are intracellular
December 2009 9-182
-------
metal-binding peptides that act as specific indicators of metal stress. Because they are produced by
plants as a response to sublethal concentrations of heavy metals, they can indicate that heavy metals
play a role in forest decline (Gawel et al., 1996, 012278). Phytochelatin concentrations have
previously been measured in coniferous trees in the northeastern U.S. The U.S. EPA (2004, 056905)
and Grantz et al. (2003, 155805) summarized studies indicating that both the number of dead red
spruce trees and phytochelatin concentrations increased sharply with elevation in the northeastern
U.S. Red spruce stands showing varying degrees of decline indicated a systematic and significant
increase in phytochelatin concentrations associated with the extent of tree injury. These data suggest
that metal stress might contribute to tree injury and forest decline in the northeastern U.S. The extent
to which low to moderate amounts of heavy metal deposition, which might occur at locations that are
not in close proximity to a large point source, contribute to adverse impacts on forest vegetation is
not known. Although the phytochelatin data suggest a linkage, more direct experimental data would
be needed to confirm such a finding.
In general, plant growth is negatively correlated with trace metal and heavy metal
concentration in soils and plant tissue (Audet and Charest, 2007, 190169). Trace metals, particularly
heavy metals, can influence forest growth. Growth suppression of foliar microflora has been shown
to result from Fe, Al, and Zn. These three metals can also inhibit fungal spore formation, as can Cd,
Cr, Mg, and Ni (see Smith, 1990, 046896). Metals cause stress and decreased photosynthesis
(Kucera et al., 2008, 190408) and disrupt numerous enzymes and metabolic pathways (Strydom et
al., 2006, 190486). Excessive concentrations of metals result in phytotoxicity through: (i) changes in
the permeability of the cell membrane; (ii) reactions of sulfydryl (-SH) groups with cations; (iii)
affinity for reacting with phosphate groups and active groups of ADP or ATP; and (iv) replacement
of essential ions (Patra et al., 2004, 081976).
In addition to disrupting photosynthesis and other metabolic pathways, metals have been
shown to alter frost hardiness and impair nutrition. A recent review by Taulavuori (2005, 190489)
suggests that metal- induced stress reduces frost hardiness of plants, a particular concern at high
elevation sites. Kim et al. (2003, 155899) found decreased concentration of K in needles and Ca in
stems ofPinus sylvestris seedlings exposed to Cd addition. This response suggests a disturbance of
nutrition in response to Cd. Pollutant-caused needle loss can reduce the interception of pollutants
from the atmosphere, and therefore reduce their concentrations in stemflow. This may be responsible
for the observation that species diversity of lichens is sometimes higher on trees affected by die-back
(Hauck, 2003, 155830).
Da Silva et al. (2006, 190190) have shown that PM had anatomical and physiological effects
on plants growing near an iron pelletization factory in Brazil. The effects of PM occurred due to
foliar uptake. Structural characteristics such as peltate trichomes may have formed a barrier
lessening the penetration of metallic iron into the mesophyll in some species. Iron was shown to
penetrate the trichomes, epidermic cells (adaxial and abaxial surfaces), stomata, xylem cells,
collenchyma, endodermis and mesophyll tissues. Once entering the stomata, the PM penetrates
within the mesophyll, it may modify the chemical balance of the mesophyll (Da et al., 2006,
190190).
A greenhouse study evaluated the combined effects of iron dust on restinga vegetation (coastal
vegetation of Brazil) that commonly grows near iron ore industries. Kuki et al. (2008, 155346) found
that iron dust had differing effects on gas exchange, chlorophyll content, iron content and antioxidant
enzyme activity on two plant species common to the restinga. Schinus terebinthifolius (an invasive
exotic in the U.S.) was not affected by the iron dust. However, Sophora tomentosa showed increased
iron content and membrane permeability to the leaves, increasing activity of antioxidant enzymes.
These results showed that the plants used different strategies to cope with PM pollution. S.
terebinthifolius avoided stress, while S. tomentosa used antioxidant enzyme systems to partially
neutralize oxidative stress.
Plant foliage can accumulate elemental Hg over time in response to air exposure and
concentrations in soil (Ericksen et al., 2003, 155769; Frescholtz et al., 2003, 190352). Amesocosm
experiment was conducted by Ericksen et al. (2003, 155769) where aspen trees were grown in
gas-exchange chambers in Hg-enriched soil (12.3 ±1.3 ug/g) and the Hg content in the foliage was
analyzed. Foliar Hg increased with leafage for 2-3 mo and then stabilized at leaf concentrations near
150 ng/g. About 80% of the Hg found in above-ground biomass was present in the leaves. The
concentration of Hg in trees grown in the same mesocosms in containers of low Hg soil
(0.03 ±0.01 ug/g) exhibited foliar Hg concentrations that were similar to those of trees grown in
Hg-enriched soil. Almost all of the foliar Hg originated from the atmosphere. Clearly, plant foliage
December 2009 9-183
-------
can be a major sink for airborne Hg, which can subsequently enter the soil after litterfall (Ericksen et
al., 2003, 155769). However, this study did not determine the extent to which atmospheric Hg was
dry-deposited on the foliage, as opposed to gaseous uptake through the stomata. Foliar/air Hg
exchange has been shown to be dynamic and bi-directional (Millhollen et al., 2006, 190447). These
investigators compared foliar Hg accumulation over time in three tree species with fluxes measured
using a plant gas-exchange system subsequent to soil amendment with HgCl2. Root tissue Hg
concentrations were strongly correlated with soil Hg concentrations, suggesting that below-ground
accumulation of Hg by roots may be an important process in the biogeochemical cycling of Hg in
soil systems. Nevertheless, measured foliar Hg fluxes indicated that deposition of atmospheric Hg
constituted the dominant flux of Hg to the leaf surface (Millhollen et al., 2006, 190447). Grigal
(2003, 155811) also found that Hg in vegetation is derived almost exclusively from the atmosphere.
Mercury uptake from soil is limited, partly because roots adsorb Hg but transport it to foliage very
poorly (Grigal, 2002, 156498). Grigal (2003, 155811) provided a thorough review of the
sequestration of Hg in forest and peatland ecosystems. A fundamental aspect of Hg cycling is its
strong relationship to organic matter. For that reason, peatlands sequester much larger quantities of
Hg than would be expected on the basis of their land area. Thus, if global climate change affects C
storage, it may indirectly affect Hg storage because of the strong relationship between Hg and
organic matter (Grigal, 2003, 155811).
Arbuscular mycorrhizal (AM) fungi can play important roles in mitigating toxicity of heavy
metals in plants. For example, AM symbiosis is known to be involved in plant adaptation to
As-contaminated soils. Higher plants that are adapted to As contaminated soils are generally
associated with mycorrhizal fungi (Gonzalez-Chavez et al., 2002, 155800). It has also been shown
that AM symbioses can influence plant coexistence and community diversity (O'Connor, 2002).
Some plants associated with AM fungi can successfully colonize sites that are heavily contaminated
by heavy metals (Pennisi, 2004, 156018).
Dong et al. (2008, 192106) cultivated white clover (Trifolium repens) and ryegrass (Lolium
perenne) in As-contaminated soil (water extractable As 82.7 mg/kg). The growth and P nutrition of
both species largely depended on AM symbiosis. The AM-inoculated plants showed selective uptake
and transfer of P over As.
PM pollution has the potential to alter species composition over long time scales. Kuki et al.
(2009, 190411) showed that early establishment stages ofSophora tomentosa species were
negatively affected by the combination of iron ore and acidifying particles. The deleterious effects of
the PM included deficient germination and toxic concentrations in roots. In contrast, S.
terebinthifolius was not affected by the PM revealing species resistance to the pollution. The
difference among species response suggests that over a long time period the imbalance will likely
change the species composition (Kuki et al., 2009, 190411).
The process of removing toxins from soil or water using photoautotrophs is referred to as
phytoremediation. Some plant species have good ability to extract heavy metals from soil, thereby
offering potential for phytoremediation (Clemens, 2006, 190179; Hooda, 2007, 190382;
Padmavathiamma and Li, 2007, 190465). For example, several species of willow (Salix spp.)
accumulate high concentrations of Zn and Cd in aboveground biomass (Lunackova et al., 2003,
155948; Meers et al., 2007, 155977; Rosselli et al., 2003, 156058). A first estimation of the order of
magnitude of potential metal removal by willow was 2 to 27 kg/ha/yr of Zn and 0.25 to 0.65 kg/ha/yr
for Cd (Meers et al., 2007, 155977). Build-up of high concentrations of trace metals in soil is
difficult to remediate because of the long residence times of metals in the environment. Plants that
survive on heavy metal-contaminated soils have been studied to elucidate the mechanisms that allow
them to tolerate such conditions and interactions between soil contamination and vegetation
composition (Becker and Brandel, 2007, 156260; Hall, 2002, 190365). There are numerous other
plants that have been investigated for application to phytoremediation. Plants that hyperaccumulate
metals have special potential for remediation of metal-contaminated sites. About 400 species have
been reported. Brassicaceae has the largest numbers of taxa, with 11 genera and 87 species known to
hyperaccumulate one or more metal contaminants (Prasad and DeOliveira, 2003, 156885).
Plant uptake is often the first step for a metal to enter higher levels of the food web.
Consumers of vegetation may often receive heavy loading of metals from their diets. Metals may
also bioaccumulate in some species and tissue concentrations are magnified at the higher trophic
levels, so-called biomagnifications (see Section 9.4.5.7. on Biomagnification).
December 2009 9-184
-------
9.4.5.4. Effects on Aquatic Ecosystems
The atmospheric deposition of PM into the ocean has important implications for primary
productivity and carbon sequestration. In part, metals in PM deposition may limit phytoplankton
growth in parts of the ocean (Crawford et al, 2003, 156370). In particular, Fe and Zn can influence
the productivity of algae that are involved in CaCO3 production. The production of both particulate
organic C and CaCO3 drive the ocean's biological carbon pump (Shulz et al., 2004, 156087). Thus,
in oceanic areas of trace metal limitation, changes in trace metal atmospheric deposition can affect
biogenic calcification, with potential consequences for CO2 partitioning between the ocean and
atmosphere.
A study by Sheesley et al. (2004, 156084) illustrated the value of bioassay procedures to
provide an initial screening of ambient PM toxicity. They used two species of green algae and two
extraction methods to compare the toxicities of atmospheric PM collected at two urban/industrial
sites and one rural site near the southern shore of Lake Michigan. Toxi cities varied by site, by
extraction solvent, and by bioassay. Results suggested that toxicity was not related to the total mass
of PM in the extract, but to the chemical components of the PM. It is noteworthy that the
concentrations of contaminants in PM in this type of short-term and acute toxicity testing are much
higher than would be found in the natural environment. Thus, the purpose of this type of testing is to
provide an initial screening-level comparison of relative toxi cities of atmospheric PM from different
source areas. It does not provide the data that would be needed to assess risk (Sheesley et al., 2004,
156084).
9.4.5.5. Effects on Animals
There has been little work focusing on animal indicators of PM effects in the field. However,
there have been several recent studies on snails, amphibians, earthworms, and bivalves that are
discussed below.
Bioindicator organisms can be especially useful for monitoring PM effects over geographical
and temporal scales. Terrestrial invertebrates have been used to monitor contaminants in both air and
soil. Snails (Helix sp.) accumulate trace metals and agrochemicals, and can be used as effective
biomonitors for urban air pollution (Beeby and Richmond, 2002, 155680; Regoli et al., 2006,
156046; Viard et al., 2004, 055675). Demonstrated biological effects include growth inhibition,
impairment of reproduction, and induction of metallothioneins that are involved in metal
detoxification (Gomot-de and Kerhoas, 2000, 155798; Regoli et al., 2006, 156046). The use of
sentinel species to detect the effects of complex mixtures of air pollutants is of particular value
because the chemical constituents are difficult to characterize, exhibit varying bioavailability, and are
subject to various synergistic effects.
Regoli et al. (2006, 156046) caged land snails (Helix aspersa) at five locations in the urban
areas of Ancona, Italy. After four weeks of exposure to ambient air pollution, the snails were
analyzed for trace metals and PAHs. Biomarkers were measured that correlated with contaminant
accumulation, including concentrations of metallothioneins, activity of biotransformation enzymes,
and peroxisomal proliferation. In addition, indicators of oxidative stress were measured, such as
oxyradical scavenging capacity, onset of cellular damage, and loss of DNA integrity. Results
documented substantial accumulation of metals and PAHs in snail digestive tissues in urban areas
having high traffic congestion. Cellular reactivity was also found, suggesting that this species is an
effective bioindicator for multipollutant air quality and PM monitoring.
Some amphibian ecotoxicological research has focused on heavy metal exposure. Contaminant
uptake can occur by oral, pulmonary, and dermal exposure (James et al., 2004, 155874; Johnson et
al., 1999, 155880; Lambert, 1997, 155916). This is potentially important because of documented
declines in amphibian populations in the U.S. and elsewhere in recent decades (Houlahan et al.,
2000, 155853). Toads were shown to be fairly tolerant of Cd exposure (James et al., 2004, 155874).
It is not clear whether current amounts of terrestrial metal contamination pose an increased risk to
amphibians in general.
Estuarine and marine bivalves provide potential bioindicators for Hg bioaccumulation. For
example, Coelho et al. (2006, 190181) investigated Hg concentrations in Scrobicularia plana, a
long-lived, deposit-feeding bivalve in southern Europe. Annual bioaccumulation rates were shown to
be strongly correlated with Hg concentrations in suspended particulate matter (SPM), a response to
December 2009 9-185
-------
their deposit-feeding tactics (Verdelhos et al., 2005, 190497). The ability to predict annual
accumulation rates for indicator species, such as this bivalve, may facilitate management actions to
avoid deleterious effects on humans through consumption of bivalves above a certain age/size class.
Earthworms often constitute a large percentage of soil animal biomass and they are considered
to be relatively sensitive indicators of soil metal contamination. They are continuously exposed to
the soil via dermal contact in the soil solution or ingestion of large quantities of soil pore water,
polluted food and/or soil particles (Lanno et al., 2004, 190415). Hobbelen et al. (2006, 190371)
determined the important metal pools for bioaccumulation by earthworms Lumbricus rubellus,
which live in the upper 5cm of soil and Aporrectodea caliginosa, which live in the upper 25 cm of
soil. Soil concentration explained much of earthworm concentrations, however Cd concentration in
A. calinginosa was best explained by pore water concentrations and no variable tested explained Zn
tissue concentrations. Massicotte et al. (2003, 155968) compared the cell viability and phagocytic
potential of three earthworm species (Lumbricus terrestris, Eisenia andrei, andAporrectodea
tuberculatd) in response to atmospheric emissions of metals from a cement factory in Quebec,
Canada. Cell viability actually increased in proximity (0.5 km) to the cement factory for A.
tuberculata, and this might have been due to beneficial effects of increased Ca deposition. There
were no significant differences observed for the other two species (Massicotte et al., 2003, 155968).
Biogeochemical cycling of Hg in the Arctic has been investigated, in part because observed
Hg concentrations in marine animals may pose health risks for local human populations. The lifetime
of gaseous elemental Hg (GEM) in the atmosphere, which constitutes about 95% of atmospheric Hg,
is generally about one year (Lin and Pehkonen, 1999, 190426). However, during spring (typically
March through June), the lifetime of GEM in the Arctic is much shorter, and atmospheric GEM can
be depleted in less than one day during atmospheric Hg depletion episodes (AMDE) (Lindberg et al.,
2002, 190429: Skov et al., 2004, 190481). During the AMDE, GEM is rapidly oxidized to reactive
gaseous Hg that can be deposited to the ground surface (Skov et al., 2004, 190481). Because of the
increased solar flux to the Arctic during spring and seasonal melting of sea ice, there may be an
increased efficiency of Hg bioaccumulation in Arctic food webs than would be expected based on
data collected at mid-latitudes. Skov et al. (2004, 190481) developed a simple parameterization for
AMDE and included it in the Danish Eulerian Hemispheric Model (DEHM). The model was shown
to reproduce the general structure of AMDE, suggesting that the limiting factor for AMDE may be
the surface temperature of sea ice.
9.4.5.6. Biomagnification across Trophic Levels
Biomagnification is the progressive accumulation of chemicals with increasing trophic level
(LeBlanc, 1995, 155921). Organic Hg is the most likely metal to biomagnify, in part because
organisms can efficiently assimilate methylmercury and it is slowly eliminated (Croteau et al., 2005,
156373; Reinfelder et al., 1998, 156047). Of the trace metals, there is also evidence that Cd, Pb, Zn,
Cu and Se biomagnify.
The study of trophic transfer and biomagnification is limited by the difficulty in discriminating
food webs and the uncertainty associated with assignment of trophic position to individual species
(Croteau et al., 2005, 156373). Use of stable isotopes can help to establish linkages. However, it is
difficult to determine the extent to which biomagnification occurs in a given ecosystem without
thoroughly investigating physiological biodynamics, habitat, food web structure, and trophic position
of relevant species. Thus, development of an understanding of ecosystem complexity is necessary to
determine what species might be at greatest risk from toxic metal exposure (Croteau et al., 2005,
156373).
Terrestrial
Bioaccumulation of heavy metals can occur through the plant-herbivore and litter-detrivore
food webs. The U.S. EPA (2004, 056905) concluded that Cd and Zn can bioaccumulate in
earthworms. Other invertebrates inhabiting soil litter may also accumulate metals. Although food
web accumulation of a metal may not result in mortality, it might reduce breeding potential or result
in other non-lethal effects that adversely affect organism responses to environmental cues.
Metal accumulation in litter can be found mainly around brass works, cement factories, and Pb
and Zn smelters. Organisms that feed on earthworms living in soils with elevated metal
December 2009 9-186
-------
concentrations may also accumulate Pb and Zn. Increased concentrations of heavy metals have been
found in a variety of mammals living in areas with elevated heavy metal concentrations in the soils.
The transfer of metals from plants to terrestrial snails is an interesting system for
biomagnifications because snails accumulate metals in their soft tissue and can contribute
significantly to the transfer of pollutants to primary consumers and terrestrial predators (Dallinger et
al., 2001, 192109). Notten et al. (2005, 190461) studied the transfer of Cu, Zn, Cd and Pb in
terrestrial soil-plant-snail food chains in metal-polluted soils of the Netherlands. The food chain
included perennial plant species Urtica dioica and the herbivorous snail Cepaea nemoralis. The
transfer of metal from the soil to the plant compartment was low (coefficient of determination R2 =
0.20). Total concentration of metals in soils was a poor predictor of leaf concentration. Low metal
concentration in the leaves was thought to be due to low pore water metal concentrations and was
also thought to be partly caused by low translocation from roots within the plant. The Cu, Zn and Cd
concentrations in the snails were always higher than concentrations in the leaves indicating
bioaccumulation. The metal transfer from the leaf to snail was highest among all routes tested,
suggesting that transfer from diet is important. Similar results were found by Beeby and Richmond
(2002, 155680) with the snail, Helix aspersa, and the plant, Taraxacum sp., for Zn, Pb, Cd, but not
for Cu.
Many types of predators including shrews, thrushes and beetle larvae include snails as part of
their diet (Gomot-De and Pihan, 2002, 190357: Seifert et al., 1999, 190480). Seifert et al. (1999,
190480) found the shrews eating snails with elevated Cd had critical levels of Cd in their kidneys.
Scheifler et al. (2007, 190379) found that Cd in snails lead to toxic levels in beetle larvae that caused
increased amounts of mortality.
Aquatic
In general, it has been assumed that metal biomagnification in aquatic ecosystems is an
exception rather than the rule (Gray, 2002, 155806). More recent research has demonstrated aquatic
biomagnification of certain metals. For example, Stewart et al. (2004, 156097) used stable isotopes
of C and N to show biomagnification of Se in San Francisco Bay food webs. Croteau et al. (2005,
156373) identified trophic position of estuarine organisms and food web structure in the delta of San
Francisco Bay to document Cd biomagnification in invertebrates that live on macrophytes and also
in fish. Concentrations of Cd were biomagnified 15 times within two trophic links in each food web.
In contrast, no tendency towards biomagnification was observed for Cu.
In aquatic ecosystems, biomagnification of trace metals does not necessarily occur. Nguyen et
al. (2005, 155997) found biodiminution for most metals in Lake Balaton, Hungary, with the
exception of slight enrichment of Zn from PM to zooplankton and of Cd from sediment to mussels.
Once transported to aquatic ecosystems, trace metals often preferentially bind to sediment
particles. Some of these sediment-bound metals may be unavailable to biota; in contrast, metals
bound to sediment organic matter may exhibit varying degrees of bioavailability (Di Toro et al.,
2005, 155750). Piol et al. (2006, 156028) studied the bioavailability of sediment-bound Cd to the
freshwater oligochaete Lumbriculus variegatus. They found that Cd uptake depended on the amount
of free dissolved Cd(II), and the Cd contribution from sedimentary particles to biological uptake was
negligible.
Marine bivalve mollusks bioaccumulate trace metals and other contaminants (LaBrecque et
al., 2004, 155913) and therefore may be used as bioindicators of contamination. In addition, they
constitute an important link to human health by virtue of their importance as a food source
(Cheggour et al., 2005, 155723: Li et al., 2002, 156691).
9.4.5.7. Effects near Smelters and Roadsides
The high PM concentrations in proximity to mining, smelting, roadsides and other industrial
sources result in heavy metal loadings that may be particularly damaging to nearby ecosystems.
December 2009 9-187
-------
Smelters
The Harjavalta region is one of the most intensively studied heavy metal polluted areas in the
world. Kiikkila et al. (2003, 156637) reviewed available data on heavy metal deposition and
environmental effects in this area. Emissions from the smelter were as high as 1,100 t/yr of dust, 140
t/yr Cu, 96 t/yr Ni, 162 t/yr Zn, and 94 t/yr Pb in 1987. Deposition amounts decreased substantially
after 1990, to only a few percent of the amounts that occurred during the 1980s.
Kiikkila (2003, 156637) investigated the effects of heavy metal pollution in proximity to a
Cu-Ni smelter at Harjavalta, Finland. The deposition of heavy metals increased within 30 km of the
smelter. Only slight changes in the understory vegetation were observed at distances greater than 8
km from the smelter. At 4 km distance, species composition of vegetation, insects, birds, and soil
microbiota changed and tree growth was reduced. Within about 1 km, only the most resistant
organisms were surviving.
The number of soil organisms clearly decreased and their community structure was altered
close to the Harjavalta smelter (Kiikkila, 2003, 156637). However, this effect was only pronounced
within about 2 km of the smelter. This suggests that the soil microfauna are relatively resistant to
metal pollution effects.
Soil microbial activity decreased close to the Harjavalta smelter (Kiikkila, 2003, 156637). as
reflected by microbial respiration, distribution of species within physiological groups, and microbial
and fungal biomass. The fungi appeared to be more sensitive to metal contamination than the
bacteria (Pennanen et al., 1996, 156016). The rate of litter decomposition decreased, causing an
accumulation of needle litter on top of the forest floor near the smelter (Fritze et al., 1989, 079635).
Inhibition of nutrient cycling and displacement by Cu and Ni of base cations from cation
exchange sites on the soil resulted in a decrease in base cation concentrations in the organic soil
layer (Derome and Lindroos, 1998, 155749; Kiikkila, 2003, 156637) close to the Harjavalta smelter.
In addition, Mg, Ca, and Mn concentrations in Scots pine (Pinus sylvestris) needles were low, and
this was attributed by Kiikkila (2003, 156637) to the toxic effects of Cu and Ni to plant fine roots
and also to ectomycorrhizal root tips (Helmisaari et al., 1999, 155836). Nutrient translocation during
fall was also affected close to the smelter; as a consequence needle concentrations of K were
relatively high (Nieminen et al., 1999, 155998).
Tree growth (Scots pine) has been poor (Malkonen et al., 1999, 155961) and most vegetation
was absent within 0.5 km of the smelter. Effects on plant species occurrence close to the smelter
were almost entirely negative. In contrast, some animal species responded positively, including a leaf
miner, three species of aphid, and some ants, beetles, and spiders.
Salemaa et al. (2004, 156069) investigated heavy metal concentrations in understory plant
species growing at varying distances from the Harjavalta Cu-Ni smelter. Heavy metal concentrations
(except Mn) were highest in bryophytes, followed by lichens, and were lowest in vascular plants.
Vascular plants are generally able to restrict the uptake of toxic elements, and therefore were able to
grow closer to the smelter than lichens. A pioneer moss (Pohlia nutans) was unusual in that it
survived close to the smelter despite its accumulation of high amounts of Cu and Ni.
Changes in breeding success of cavity-nesting passerine birds close to the Harjavalta smelter
were attributed to habitat changes in response to metal toxicity (Eeva et al., 2000, 155761; Kiikkila,
2003, 156637). Calcium supply is also well known to be important for breeding success in passerine
bird species. Eggshell thickness, egg size, clutch size, and hatchability of pied flycatcher (Ficedula
hypoleucd) were found to be depressed near the Cu smelter at Harjavalta, SW Finland (Eeva and
Lehikoinen, 2004, 155762). Availability of Ca-rich food to the birds was estimated by counting snail
shells in the nests postfledging. The number of snail shells correlated positively with the Ca
concentration of nestling feces and adult breeding success. In addition, the negative impact of Cu on
the number of fledglings was stronger at locations where Ca concentration was low (Eeva and
Lehikoinen, 2004, 155762).
Documentation of effects on individual species, such as was reported above, does not reveal
what the impacts might be on ecosystem function. Nevertheless, the mere fact that multiple species,
operating at different trophic levels, have been shown to be affected by the ambient deposition in
proximity to the smelter suggests that effects on ecosystem function may indeed have occurred.
More research is needed, however, to fully evaluate effects on function as opposed to abundance of
individual species.
December 2009 9-188
-------
Roadsides
Heavy metal particles are important constituents of road dust. These particles accumulate on
the road surface from brake linings, road paint, tire debris, diesel exhaust, road construction
materials, and catalyst materials. Road dust can be suspended in the atmosphere and contribute
metals to soil, air, and urban runoff (Adachi and Tainosho, 2004, 081380; Davis et al., 2001,
024933; Smolders and Degryse, 2002, 156091). In particular, Zn oxide comprises 0.4-4.3% of tire
tread (Smolders and Degryse, 2002, 156091) and tire wear is a substantial source of environmental
Zn pollution. Adachi and Tainosho (2004, 081380) used a field emission screening electron
microscope equipped with an energy dispersive x-ray spectrometer to characterize heavy metal
particles embedded in tire dust. Samples were classified into four likely source categories, based on
cluster analysis. Based on morphology and chemical composition, the samples were identified as
having derived from yellow paint (CrPbO4 particles), brake dust (particulate Ti, Fe, Cu, Sb, Zr, Ba
and heavy minerals [Y, Zr, La, Ce]), and tire tread (Zn oxide).
Since publication of EPA's 2004 PM criteria assessment, some additional research has been
conducted on the effects of windblown PM. Effects on physical, chemical, and biological attributes
of both plants and animals have been documented (Englert, 2004, 087939; Gleason et al., 2007,
155794; Kappos et al., 2004, 087922). Experiments by Gleason et al. (2007, 155794) suggest that
most direct effects on plants of windblown PM originating from on-road surfaces occur within 40 m
of the source. Windblown PM from roads or agriculture can cover plant photosynthetic structures
(Sharifi et al., 1999, 156082). cause impact damage (Armbrust and Retta, 2002, 156225). or
interfere with physiological mechanisms (Burkhardt et al., 2002, 155708). As previously discussed
in Section 9.4.5.5, land snails in urban areas have been shown to be a good indicator of traffic
pollution.
9.4.5.8. Toxicity to Mosses and Lichens
At the time of the most recent air quality criteria report for PM (U.S. EPA, 2004, 056905).
trace metal toxicity to lichens had been demonstrated in relatively few cases. Nash (1975, 016763)
documented Zn toxicity in the vicinity of a Zn smelter near Palmerton, PA. Experimental data had
suggested that lichen tolerance to Zn and Cd generally ranges between 200 and 600 ppm (Nash,
1975. 016763).
The effects of deposited metals on the mosses have not been well studied. Tremper et al.
(2004, 156126) exposed mosses of two species to roadside conditions and sampled them over a
period of 3 mo. Under field conditions, chlorophyll concentrations in moss tissue were not affected
by metal contamination and accumulation
Mosses and lichens readily take up metals from atmospheric deposition. Otnyukova (2007,
156009) demonstrated vertical gradients within a coniferous forest canopy in the fruticose lichen
genus Usnea with respect to lichen thallus morphology and heavy metal concentration. Abnormal
thalli at the tree-top level contained higher concentrations of Al, Fe, Zn, F, Sr, and Pb. This vertical
pattern within the tree canopy is in general accordance with known deposition of PM to plants
(Otnyukova, 2007, 156009).
There is an extensive literature on the use of mosses and lichens for estimating deposition
(biomonitors) and indicating metal exposure in ecosystems (bioindicators) (see Section 9.4.2.3.).
9.4.6. Organic Compounds
VOCs in the atmosphere are partitioned between the gas and particle phases. As described by
the U.S. EPA (2004, 056905). the partitioning depends on vapor pressure, temperature, surface area
of the particles, and the nature of the particles and of the chemical being adsorbed. A wide variety of
organic contaminants are deposited from the atmosphere. These include chemicals such as DDT,
PCBs, and PAHs.
Important organic atmospheric contaminants are generally those that are transported long
distances in the atmosphere, subsequently deposited into remote locations, and bioaccumulated to
sufficient concentrations that they can affect humans, wildlife, or other biota (Swackhamer et al.,
2004, 190488). Certain physical and chemical properties facilitate the movement of these
contaminants from land and water surfaces into the atmosphere, provide stability, and enhance
December 2009 9-189
-------
accumulation in lipids. Some, including the relatively small (up to 4 rings) PAHs degrade relatively
rapidly in the atmosphere or at the surface subsequent to atmospheric deposition. Below is a
summary of the findings of the U.S. EPA (2004, 056905). followed by discussion of more recent
research findings.
Plants may be used as passive monitors to compare the deposition of organic compounds
between sites. Vegetation can be used semi-quantitatively to indicate organic pollutant amounts if the
mechanism of accumulation is considered. Organic compounds can enter the plant via the roots or be
deposited as particles on the leaves and be taken up through the cuticle or stomata. The pathways
depend on the chemical and its physical properties. These include, for example, lipophilicity, water
solubility, vapor pressure, and Henry's law constant. Environmental conditions can also be
important, including temperature and organic content of soil, plant species, and the foliar surface
area and lipid content.
Organic particulates in the atmosphere are diverse in their makeup and sources. Vegetation
itself is an important source of hydrocarbon aerosols. Terpenes, particularly a-pinene, (3-pinene, and
limonene, released from tree foliage may react in the atmosphere to form submicron particles. These
naturally generated organic particles contribute significantly to the blue-haze aerosols formed
naturally over forested areas (Geron et al, 2000, 019095: U.S. EPA, 2004, 056905).
The low water solubility and high lipo-affinity of many organic xenobiotics control their
interaction with the vegetative components of natural ecosystems. Foliar surfaces are covered with a
waxy cuticle layer that helps reduce moisture loss and short-wave radiation stress. This epicuticular
wax consists largely of long-chain esters, polyesters, and paraffins, which accumulate lipophilic
compounds. Organic air contaminants in the particulate or vapor phase can be adsorbed to, and
accumulate in, the epicuticular wax of leaf surfaces. Direct uptake of organic contaminants through
the cuticle and vapor-phase uptake through the stomata are not well characterized for most trace
organics.
Soil acts as an important storage compartment for POPs, including PCBs and PAHs. There is a
continuous process of partitioning between the soil pool and the atmosphere, and this controls the
regional and global transport of these compounds (Backe et al., 2004, 155668; Wania and Mackay,
1993, 157110). Over time, POPs move towards equilibrium between the environmental
compartments, and this process can be described using the fugacity concept (Backe et al., 2004,
155668; Mackay, 1991, 042941). Fugacity reflects the tendency of a chemical constituent to escape
one environmental compartment and move to another. When an equilibrium distribution is achieved,
the fugacity quotient values in each compartment will be equal. Soil/air partitioning is controlled by
a variety of factors. These include soil properties, such as organic matter content, moisture, porosity,
texture, and structure, as well as the physiochemical properties of the pollutant, including vapor
pressure and water solubility.
The accumulation of PAHs in vegetation, due to their lipophilic nature, could contribute to
human and other animal exposure via food consumption. As a result, plant uptake of PAHs has been
an important area of research (Gao and Zhu, 2004, 155782). Most bioaccumulation of PAHs by
plants occurs by leaf uptake (Tao et al., 2006, 156112). Root uptake also occurs. It appears that roots
preferentially accumulate the lower molecular weight PAHs due to their greater water solubility
(Wild and Jones, 1992, 156155).
Various models have been developed to simulate plant uptake of organic contaminants. The
simple partition-limited model of Chiou et al. (2001, 156342) has been further expanded to increase
complexity and to include root uptake pathways (e.g., Fryer and Collins, 2003, 156454; Yang et al.,
2005, 192104; Zhu et al., 2004, 156184).
In evaluating receptor choice for studies of contaminant exposure to plants, and also
remediation potential, it is important to consider differences among species. For example, Parrish et
al. (2006, 156014) assessed the bioavailability of PAHs in soil. During the first growing season,
zucchini (Cucurbita pepo ssp. pepo) accumulated significantly more PAHs than did other related
plant species, including up to three orders of magnitude greater concentrations of the six-ring PAHs.
Parrish et al. (2006, 156014) also noted differences in PAH uptake by two different species of
earthworm.
The leaves of Quercus ilex have been shown to readily accumulate PAHs in situ. Young leaves
accumulated PAHs within three weeks of bud break. Mature leaves showed seasonality, with higher
PAH concentrations during winter (Alfani et al., 2005, 154319).
It is difficult to discriminate between PAHs that are adsorbed to plant root surfaces as opposed
to those that are actually taken up by the roots. In general, soil bound PAHs are associated with soil
December 2009 9-190
-------
organic matter and are therefore not readily available for root uptake (Fismes et al., 2002, 141156;
Jiao et al., 2007, 155879). Wild et al. (2005, 156156) used two-photon excitation microscopy to
visualize the uptake and transport of two PAHs (anthracene and phenanthrene) from a contaminated
soil into living wheat and maize roots. Jiao et al. (2007, 155879) developed a sequential extraction
method to discriminate between PAH adsorption in rice roots.
Maize roots and tops of plants have been shown to directly accumulate PAHs from aqueous
solution and from air in proportion to exposure amounts. Root concentration factors are log-linear
functions of log-based octanol-water partition coefficients (log Kow); similarly, leaf concentration
factors are log-linear functions of log-based octanol-air partition coefficients (log Koa) (Lin et al.,
2007, 155933). Although the bulk concentrations of PAHs in various plant tissues can differ greatly,
the observed differences disappear after they are normalized to lipid content (Lin et al., 2007,
155933). This suggests that the lipid content of different plant tissues may influence PAH
distribution within the plant.
Previously, there was relatively little information available regarding incorporation of
atmospherically deposited PAHs into aquatic food webs. It is known that PAHs can be transferred to
higher trophic levels, including fish, and that this transfer can be mediated by aquatic invertebrates,
which generally comprise an important part offish diets. High mountain lakes offer an effective
receptor for quantification of biomagnification in aquatic ecosystems from atmospheric PM
deposition. There are typically no sources of organic contaminants in their watersheds, and
atmospheric inputs dominate as sources of contamination. In addition, such lakes tend to have
relatively simple food webs. Vives et al. (2005, 157099) investigated PAH content of brown trout
(Salmo truttd) and their food items. Total PAH concentrations tended to be highest in organisms that
occupy littoral habitats, and lowest in pelagic organisms. This may reflect more efficient transfer of
PAHs to underlying sediments in shallower water and associated degradation within the water
column.
Some atmospheric organic contaminants have been shown to accumulate in biota at remote
locations. For example, polybrominated diphenyl ethers (PBDEs), which are man-made chemicals
used as flame retardants in materials manufacturing, have been found to accumulate in lichens and
mosses collected at King George Island, maritime Antarctica (Yogui and Sericano, 2008, 189971).
Because contaminant concentrations were not statistically different at sites close to and distant from
human facilities in Antarctica, the authors concluded that long-range atmospheric transport was the
likely primary source of PBDEs to King George Island. Law et al. (2003, 190420) reviewed
available data for accumulation of PBDEs and other brominated flame retardants in wildlife. These
compounds have become widely distributed in the environment, including in the deep-water, oceanic
food webs.
Ohyama et al. (2004, 190462) chose salmonid fish, mainly rainbow trout (Oncorhyncus
mykiss), as an indicator species to evaluate the transport and bioaccumulation of organochloride
compounds in the northern and central Sierra Nevada. They found that elevation was an important
factor affecting residual concentrations of poly chlorinated biphenyls (PCBs) in fish muscle tissue.
On this basis, Ohyama et al. (2004, 190462) concluded that PCB residue in rainbow trout, a widely
distributed salmonid species, provided a good monitoring tool for studying the effects of
mountainous topography on the long-range transport and distribution of persistent organic pollutants.
Semivolatile compounds can undergo repeated volatilization on surfaces, such as plant foliage,
in response to diel changes in temperature. As a consequence, such compounds can be deposited, re-
emitted, and re-deposited multiple times. This behavior can cause these compounds to move large
distances in a leap-frog fashion (Krupa et al., 2008, 198696). It is believed that POPs can be
atmospherically transported throughout the world because of their volatility and response to changes
in temperature. This "global distillation theory" (Holmqvist et al., 2006, 190380; Wania and
Mackay, 1993, 157110) predicts that POPs in the northern hemisphere are generally transported
towards the Arctic, and in the southern hemisphere they are transported toward the Antarctic. In
general, POP concentrations measured in the Arctic are higher than in the Antarctic. They have been
detected in all levels of the Arctic food web (Oehme et al., 1995, 011267). Bioconcentration of
organochlorines has been shown in the Arctic food web, including fish, seals, and polar bears
(Oehme et al., 1995, 011267). Concentrations measured in Arctic polar bears are especially high
(AMAP, 2004, 190168).
Holmqvist et al. (2006, 190380) measured levels of PCBs in longfin eels (Anguilla
dieffenbachii) in 17 streams on the west coast of South Island, New Zealand. The PCBs were at low
levels, and were believed to originate from atmospheric transport from industrial areas in Asia.
December 2009 9-191
-------
Characteristics of the longfin eel that make it susceptible to bioaccumulation of lipophilic persistent
pollutants include high lipid content (up to 40%), long lifespan (up to 90 yrs), and position near the
top of the food chain (Holmqvist et al, 2006, 190380).
Long-range transport of atmospherically deposited contaminants can be augmented by
biotransport. A good example of this phenomenon was documented by Ewald et al. (1998, 190348).
who showed that biotransport by migrating sockeye salmon (Oncorhynchus nerkd) in the Copper
River watershed, Alaska, had a greater influence than atmospheric transport on bioaccumulation of
PCBs and DDT in lake food webs. Organic pollutants accumulated by salmon during their ocean
residence were effectively transferred 410 km inland to their spawning lake. Arctic grayling
(Thymallus arcticus) in the salmon spawning lake were found to contain organic pollutants more
than twice as high as arctic grayling in a near-by salmon-free lake. The pollutant composition of the
grayling in the salmon spawning lake was similar to that of the migrating salmon (Ewald et al., 1998,
190348), suggesting that salmon migration contributed to bioaccumulation of organic contaminants
in the lake used for spawning by the salmon.
An assessment of the ecological effects of airborne metals and SOCs was conducted for eight
NPs by the Western Airborne Contaminants Assessment Project (WACAP) (Landers et al., 2008,
191181). From 2002-2007, WACAP researchers conducted analysis of the biological effects of
airborne contaminants in seven ecosystem compartments: air, snow, water, sediments, lichens,
conifer needles and fish. The goals were to identify where the pollutants were accumulating, identify
ecological indicators for those pollutants causing ecological harm, and to determine the source of the
air masses most likely to have transported the contaminants to the parks.
The results from WACAP were summarized by Landers et al. (2008, 191181). which
concluded that bioaccumulation of SOCs were observed throughout park ecosystems. Vegetation
tended to accumulate PAHs, CUPs, and HCHs. Conifer needles were a good indicator of pesticides,
however the ecological consequences of this accumulation are unexamined. SOCs in vegetation and
air showed different patterns, possibly because each medium absorbs different types of SOCs with
varying efficiencies. Mean ammonium nitrate concentration in ambient fine particulates <2.5 (im
diameter was a good predictor of dacthal, endosulfan, chloradane, trifluralin, DDT and PAH
concentrations in vegetation.
Concentrations of SOCs were five to seven orders of magnitude higher in fish tissue than in
sediments. Fish accumulated more PCBs, chlordanes, DDT and dieldrin than vegetation. Fish lipid
and age were the most reliable predictors of SOC concentrations. Most fish appeared normal during
field necropsies; however, individuals with both male and female reproductive organs were collected
at two sites. The incidence of this condition has increased since the pre-organic pollutant era.
Additionally, elevated concentrations of vitellogenin, a female protein involved in egg production,
were found in male fish from three sites, and directly related to the concentration of several
organochlorines at one site.
The lake sediment records showed steadily increasing mercury deposition over time at lakes in
two parks, Mt. Ranier NP and Rocky Mountain NP Apportionment of the mercury to its atmospheric
sources is not quantified at this time; however, the pattern in the sediment suggests a local source
rather than a global source. Mercury concentrations in fish exceeded contaminant health thresholds
for some piscivorous fish, mammals and birds in most parks. The average mercury concentration in
fish from one site and individual fish from three additional sites exceeded the U.S. EPA contaminant
health thresholds for humans.
Although this assessment focuses on chemical species that are components of PM, it does not
specifically assess the effects of particulate versus gas-phase forms; therefore, in most cases it is
difficult to apply the results to this assessment based on particulate concentration and size fraction.
9.4.7. Summary of Ecological Effects of PM
Ecological effects of PM include direct effects to metabolic processes of plant foliage;
contribution to total metal loading resulting in alteration of soil biogeochemistry and microbiology,
plant and animal growth and reproduction; and contribution to total organics loading resulting in
bioaccumulation and biomagnification across trophic levels. These effects were well-characterized in
the 2004 PM AQCD (U.S. EPA, 2004, 056905). Thus, the summary below builds upon the
conclusions provided in that review.
PM deposition comprises a heterogeneous mixture of particles differing in origin, size, and
chemical composition. Exposure to a given concentration of PM may, depending on the mix of
December 2009 9-192
-------
deposited particles, lead to a variety of phytotoxic responses and ecosystem effects. Moreover, many
of the ecological effects of PM are due to the chemical constituents (e.g., metals, organics, and ions)
and their contribution to total loading within an ecosystem.
Investigations of the direct effects of PM deposition on foliage have suggested little or no
effects on foliar processes, unless deposition levels were higher than is typically found in the
ambient environment. However, consistent and coherent evidence of direct effects of PM has been
found in heavily polluted areas adjacent to industrial point sources such as limestone quarries,
cement kilns, and metal smelters (Sections 9.4.3 and 9.4.5.7). Where toxic responses have been
documented, they generally have been associated with the acidity, trace metal content, surfactant
properties, or salinity of the deposited materials.
An important characteristic of fine particles is their ability to affect the flux of solar radiation
passing through the atmosphere, which can be considered in both its direct and diffuse components.
Foliar interception by canopy elements occurs for both up- and down-welling radiation. Therefore,
the effect of atmospheric PM on atmospheric turbidity influences canopy processes both by radiation
attenuation and by changing the efficiency of radiation interception in the canopy through
conversion of direct to diffuse radiation. Crop yields can be sensitive to the amount of radiation
received, and crop losses have been attributed to increased regional haze in some areas of the world
such as China. On the other hand, diffuse radiation is more uniformly distributed throughout the
canopy and may increase canopy photosynthetic productivity by distributing radiation to lower
leaves. The enrichment in photosynthetically active radiation (PAR) present in diffuse radiation may
offset a portion of the effect of an increased atmospheric albedo due to atmospheric particles. Further
research is needed to determine the effects of PM alteration of radiative flux on the growth of
vegetation in the U.S.
The deposition of PM onto vegetation and soil, depending on its chemical composition, can
produce responses within an ecosystem. The ecosystem response to pollutant deposition is a direct
function of the level of sensitivity of the ecosystem and its ability to ameliorate resulting change.
Many of the most important ecosystem effects of PM deposition occur in the soil. Upon entering the
soil environment, PM pollutants can alter ecological processes of energy flow and nutrient cycling,
inhibit nutrient uptake, change ecosystem structure, and affect ecosystem biodiversity. The soil
environment is one of the most dynamic sites of biological interaction in nature. It is inhabited by
microbial communities of bacteria, fungi, and actinomycetes, in addition to plant roots and soil
macro-fauna. These organisms are essential participants in the nutrient cycles that make elements
available for plant uptake. Changes in the soil environment can be important in determining plant
and ultimately ecosystem response to PM inputs.
There is strong and consistent evidence from field and laboratory experiments that metal
components of PM alter numerous aspects of ecosystem structure and function. Changes in the soil
chemistry, microbial communities and nutrient cycling, can result from the deposition of trace
metals. Exposures to trace metals are highly variable, depending on whether deposition is by wet or
dry processes. Although metals can cause phytotoxicity at high concentrations, few heavy metals
(e.g., Cu, Ni, Zn) have been documented to cause direct phytotoxicity under field conditions.
Exposure to coarse particles and elements such as Fe and Mg are more likely to occur via dry
deposition, while fine particles, which are more often deposited by wet deposition, are more likely to
contain elements such as Ca, Cr, Pb, Ni, and V. Ecosystems immediately downwind of major
emissions sources can receive locally heavy deposition inputs. Phytochelatins produced by plants as
a response to sublethal concentrations of heavy metals are indicators of metal stress to plants.
Increased concentrations of phytochelatins across regions and at greater elevation have been
associated with increased amounts of forest injury in the northeastern U.S.
Overall, the ecological evidence is sufficient to conclude that 3 Causal relationship JS
likely to exist between deposition of PM and a variety of effects on individual organisms
and ecosystems, based on information from the previous review and limited new
findings in this review. However, in many cases, it is difficult to characterize the nature and
magnitude of effects and to quantify relationships between ambient concentrations of PM and
ecosystem response due to significant data gaps and uncertainties as well as considerable variability
that exists in the components of PM and their various ecological effects.
December 2009 9-193
-------
9.5. Effects on Materials
Effects of air pollution on materials are related to both aesthetic appeal and physical damage.
Deposited particles, primarily carbonaceous compounds, cause soiling of building materials and
culturally important items, such as statues and works of art. Physical damage from dry deposition of
PM also can accelerate natural weathering processes. The major deterioration phenomenon affecting
building materials in response to atmospheric deposition is most likely sulfation, leading to
secondary salt crystallization which forms gypsum (Marinoni et al., 2003, 092520).
This section (a) summarizes information on exposure-related effects on materials associated
with particulate pollutants as addressed in the 2004 PM AQCD (U.S. EPA, 2004, 056905): and (b)
presents relevant information derived from very limited research conducted and published since
completion of that document. Most recent work on this topic has been conducted outside the U.S.
There is a variety of factors that contribute to the deterioration of monuments and buildings of
cultural significance. They include: (1) biodeterioration processes; (2) weathering of materials
exposed to the air; and (3) air pollution from both anthropogenic and natural sources (Herrera and
Videla, 2004, 155843). Because of the diversity in climate, proximity to marine aerosol sources, and
pollution of various types, the magnitude and relative importance of these causal agents vary by
location.
Much existing literature on damage to structural materials of cultural heritage has not seriously
considered the importance of biodeterioration processes and the relationship that often exists
between environmental characteristics and the microbial communities that colonize monuments and
buildings. In general, high humidity, high temperature, and air pollution often enhance the
biodeterioration hazard. Herrera and Videla (2004, 155843) concluded that heterotrophic bacteria,
fungi, and cyanobacteria were the main microbial colonizers of buildings that they investigated in
Latin America. Their analyses suggested that the major deterioration mechanism of limestone at the
Mayan site of Uxmal in a non-polluted rural environment was biosolubilization induced by
metabolic acids produced by bacteria and fungi. The rock decay at Tulum, near the seashore, was
mainly attributed to the marine influence. At Medellin, it appeared that biodeterioration effects from
microbes synergistically enhanced the effects of atmospheric factors on material decay. Deterioration
of structural material in the Cathedral of La Plata, located in a mixed urban/industrial environment,
was attributed mainly to atmospheric pollutants (Herrera and Videla, 2004, 155843).
Ambient particles can cause soiling of man-made surfaces. Soiling generally is considered an
optical effect. Soiling changes the reflectance from opaque materials and reduces the transmission of
light through transparent materials. Soiling can represent a significant detrimental effect, requiring
increased frequency of cleaning of glass windows and concrete structures, washing and repainting of
structures, and, in some cases, reduces the useful life of the object. Particles, especially carbon, may
also help catalyze chemical reactions that result in the deterioration of materials (U.S. EPA, 2004,
056905).
Soiling is dependent on atmospheric particle concentration, particle size distribution,
deposition rate, and the horizontal or vertical orientation and texture of the exposed surface (Haynie,
1986, 157198). The chemical composition and morphology of the particles and the optical properties
of the surface being soiled will determine the time at which soiling is perceived by human observers
(Nazaroff and Cass, 1991, 044577).
Perm et al. (2006, 155135) reported development of a simple passive particle collector for
estimating dry deposition to objects of cultural heritage. The observed mass of deposited particles
mainly belonged to the coarse particulate mode. The sampler collects particles from all directions. It
replicates at least some of the complexity of particle deposition to actual objects, and is easier to
analyze than a precious object (Perm et al., 2006, 155135).
Soiling of urban buildings constitutes a visual nuisance that leads to the loss of architectural
value. Soiling can include reversible darkening of the building surfaces and also irreversible damage.
Water runoff patterns on the building surfaces are influenced by the type of surface material,
architectural elements, and climate. Therefore, soiling does not occur uniformly across the building.
Public perception of soiling entails complex interactions between the extent of soiling, architecture,
and aesthetics (Grossi and Brimblecombe, 2004, 155813).
One of the most significant air pollution damage features affecting urban buildings and
monuments is the formation of black crusts. Quantification of different forms of carbon in black
crusts is difficult. There is often a carbonate component which is derived from the building material,
plus OC and EC, derived from air pollution. EC is considered to be a tracer for combustion sources,
December 2009 9-194
-------
whereas OC may derive from multiple sources, including atmospheric deposition of primary and
secondary pollutants, and the decay of protective organic treatments (Bonazza et al., 2005, 155695).
Bonazza et al. (2005, 155695) quantified OC and EC in damage layers on European cultural heritage
structures. OC predominated over EC at almost all locations investigated. Traffic appeared to be the
major source of fine carbonaceous particles, with organic matter as the main component (Putaud et
al., 2004, 055545). Viles and Gorbushina (2003, 156138) found that soiling in Oxford, U.K. showed
a relationship with traffic and NO2 concentrations.
In addition to the soiling effects of EC, much soiling appears to be largely of microbiological
origin (Viles and Gorbushina, 2003, 156138). Microbial biofilms, composed mainly of fungi, can
stain exposed rock surfaces with yellow, orange, brown, gray, or black colors. Microorganisms may
be able to trap PM more efficiently than the stone surface itself. In addition, microbial growth may
be stimulated by organic or nutrient constituents in PM deposition.
Viles et al. (2002, 156137) investigated the nature of soiling on limestone tablets in relation to
ambient air pollution and climate at three contrasting sites in Great Britain over periods of one to
eight years. Spectrophotometer and microscope observations suggested that there were not consistent
trends in soiling over time at the study sites. Each site behaved differently in terms of the temporal
development of soiling and the differences between sheltered and exposed limestone tablets. In
addition, organisms played important roles in the soiling response, even at the highly polluted site.
Some work has been conducted on public perception regarding the lightness of historic
buildings and the aesthetic need for cleaning subsequent to soiling by air pollution. Brimblecombe
and Grossi (2005, 155703) found a strong relationship between the perceived lightness of a building
and the opinion that it was dirty. This relationship was used to establish levels of blackening that
might be publicly acceptable.
Recently, the importance of organic contaminant deposition to the overall air pollution damage
to building materials has been recognized. Low molecular weight organic anions such as formate,
acetate, and oxalate are ubiquitous in black crusts in damage layers on stones and mortars sampled
from monuments and buildings throughout Europe (Sabbioni et al., 2003, 049282). This has been
observed at urban, suburban, and rural sites.
9.5.1. Effects on Paint
Studies have evaluated the soiling effects of particles on painted surfaces (U.S. EPA, 2004,
056905). Particles composed of EC, acids, and various other constituents are responsible for the
soiling of structural painted surfaces. Coarse-mode particles (>2.5 urn) initially contribute more
soiling of horizontal and vertical painted surfaces than do fine-mode particles (<2.5 um), but are
more easily removed by rain (Haynie and Lemmons, 1990, 044579). Rain interacts with coarse
particles, dissolving the particle and leaving stains on the painted surface (Creighton et al., 1990,
044578; Haynie and Lemmons, 1990, 044579). Particle deposition contributes to increased
frequency of cleaning of painted surfaces and physical damage to the painted surface. Air pollution
affects the durability of paint finishes by promoting discoloration, chalking, loss of gloss, erosion,
blistering, and peeling (U.S. EPA, 2004, 056905). There have been no new developments in this field
subsequent to the review of the U.S. EPA (2004, 056905).
9.5.2. Effects on Metal Surfaces
Metals undergo natural weathering processes. The effects of air pollutants on natural
weathering processes depend on the nature of the pollutant(s), the deposition rate, and the presence
of moisture (U.S. EPA, 2004, 056905). Pollutant effects on metal surfaces are governed by such
factors as the presence of protective corrosion films and surface electrolytes, the orientation of the
metal surface, and surface moisture. Surface moisture facilitates particulate deposition and promotes
corrosive reactions. Formation of hygroscopic salts increases the duration of surface wetness and
enhances corrosion.
A corrosion film, such as for example the rust layer on the surface of some metals, may
provide some protection against further corrosion. Its effectiveness in retarding the corrosion process
is affected by the solubility of the corrosion layer and the pollutant exposure. Other than the effects
of acidifying compounds, there has not been additional research conducted in recent years on the
effects of PM deposition on metal corrosion.
December 2009 9-195
-------
9.5.3. Effects on Stone
Air pollutants can enhance the natural weathering processes on building stone. The
development of crusts on stone monuments has been attributed to the interaction of the stone's
surface with pollutants, wet or dry deposition of atmospheric particles, and dry deposition of gypsum
particles. Because of a greater porosity and specific surface, mortars have a high potential for
reacting with environmental pollutants (Zappia et al., 1998, 012037).
Most research evaluating the effects of air pollutants on stone structures has concentrated on
gaseous pollutants (U.S. EPA, 2004, 056905). The dark color of gypsum is attributed to soiling by
carbonaceous particles. A lighter gray colored crust is attributed to soil dust and metal deposits
(Ausset et al., 1998, 040480: Camuffo, 1995, 076278: Lorusso et al., 1997, 084534: Moropoulou et
al., 1998, 040485). Lorusso et al. (1997, 084534) attributed the need for frequent cleaning and
restoration of historic monuments in Rome to exposure to total suspended particulates.
Grossi et al. (2003, 155812) investigated the black soiling rates of building granite, marble,
and limestone in two urban environments with different climates. Horizontal specimens were
exposed, both sheltered and unsheltered from rainfall. Limestone showed soiling proportional to the
square root of the time of exposure, but granite and marble did not.
Black soiling is caused mainly by particulate EC (PEC). For that reason, it is most prevalent in
urban environments due to the formation of carbonaceous fine particles from the incomplete
combustion of fossil fuels. Traffic emissions, especially from diesel engines, and wood burning are
important sources of PEC (Grossi et al., 2003, 155812).
Kamh (2005, 155888) studied the effects of weathering on Conway Castle, an historical
structure in Great Britain built about 1289 AC. The weathering was identified as honeycomb,
blackcrust, exfoliation, and discoloration, with white salt efflorescence at some parts. These features
are diagnostic for salt weathering (Goudie et al., 2002, 156486). and this was confirmed by
laboratory analyses, including scanning electron microscopy and x-ray diffraction. The authors
concluded that the salt was derived from three sources: sea spray, chemical alteration of the
carbonate in mortar into SO42~ salts by acidic deposition, and wet deposition of air pollutants on the
stone surface. The salt content on the rock surface fills the rock pores and then exerts high pressure
on the rock texture due to hydration of the salt in the cold humid environment. In particular, CaSO4
and Na2SO4 exert enough pressure on hydration as to deteriorate construction rock at both the micro-
and macroscale (Moses and Smith, 1994, 156785).
9.5.4. Summary of Effects on Materials
Building materials (metals, stones, cements, and paints) undergo natural weathering processes
from exposure to environmental elements (wind, moisture, temperature fluctuations, sunlight, etc.).
Metals form a protective film of oxidized metal (e.g., rust) that slows environmentally induced
corrosion. However, the natural process of metal corrosion is enhanced by exposure to
anthropogenic pollutants. For example, formation of hygroscopic salts increases the duration of
surface wetness and enhances corrosion.
A significant detrimental effect of particle pollution is the soiling of painted surfaces and other
building materials. Soiling changes the reflectance of opaque materials and reduces the transmission
of light through transparent materials. Soiling is a degradation process that requires remediation by
cleaning or washing, and, depending on the soiled surface, repainting. Particulate deposition can
result in increased cleaning frequency of the exposed surface and may reduce the usefulness of the
soiled material. Attempts have been made to quantify the pollutant exposure at which materials
damage and soiling have been perceived. However, to date, insufficient data are available to advance
the knowledge regarding perception thresholds with respect to pollutant concentration, particle size,
and chemical composition. Nevertheless, the evidence is sufficient to conclude that 3 C3US3I
relationship exists between PM and effects on materials.
December 2009 9-196
-------
Chapter 9 References
Abdou W; Diner D; Martonchik J; Bruegge C; Kahn R; Gaitley B; Crean K; Remer L; Holben B. (2005). Comparison of
coincident Multiangle Imaging Spectroradiometer and Moderate Resolution Imaging Spectroradiometer aerosol
optical depths over land and ocean scenes containing Aerosol Robotic Network sites. J Geophys Res, 110: D10S07.
190028
Aboal JR; Fernandez JA; Carballeira A. (2004). Oak leaves and pine needles as biomonitors of airborne trace elements
pollution. Environ Exp Bot, 51: 215-225. 155642
ABT. (2001). Assessing Public Opinions on Visibility Impairment Due to Air Pollution: Summary Report. 156185
ABT. (2002). Sense of Place and Stewardship: Final Focus Group Report. 156186
Ackerman A; Kirkpatrick M; Stevens D; Toon O. (2004). The impact of humidity above stratiform clouds on indirect
aerosol climate forcing. Nature, 432: 1014-1017. 190056
Ackerman AS; Toon OB; Hobbs PV. (1994). Reassessing the dependence of cloud condensation nucleus concentration on
formation rate. Nature, 367: 445-447. 189975
Ackerman AS; Toon OB; Stevens DE; Heymsfield AJ; Ramanathan V; WeltonEJ. (2000). Reduction of tropical cloudiness
by soot. Sci New York, 288: 1042-1047. 002987
Ackerman T; Stokes G (2003). The atmospheric radiation measurement program. Phys Today, 56: 38-44. 192080
Adachi K; Tainosho Y. (2004). Characterization of heavy metal particles embedded in tire dust. Environ Int, 30: 1009-1017.
081380
Adamo P; Crisafulli P; Giordano S; Minganti V; Modenesi P; Monaci F; Pittao E; Tretiach M; Bargagli R. (2007). Lichen
and moss bags as monitoring devices in urban areas. Part II: Trace element content in living and dead biomonitors
and comparison with synthetic materials. Environ Pollut, 146: 392-399. 155644
Albrecht B. (1989). Aerosols, cloud microphysics, and fractional cloudiness. Science, 245: 1227-1230. 045783
AlexeevaLB; Strachan WMJ; Shlychkova VV; Nazarova AA; Nikanorov AM; Korotova LG (2001). Organochlorine
pesticides and trace metal monitoring of Russian rivers flowing to the Arctic Ocean. Mar Pollut Bull, 43: 71 -85.
155651
Alfani A; De Nicola F; Maisto G; Prati MV. (2005). Long-term PAH accumulation after bud break in Quercus ilex L leaves
in a polluted environment. Atmos Environ, 39: 307-314. 154319
Allen GA; Lawrence J; Koutrakis P. (1999). Field validation of a semi-continuous method for aerosol black carbon
(aethalometer) and temporal patterns of summertime hourly black carbon measurements in southwestern PA. Atmos
Environ, 33: 817-823. 048923
Almas AR; Bakken LR; Mulder J. (2004). Changes in tolerance of soil microbial communities in Zn and Cd contaminated
soils. Soil Biol Biochem, 36: 805-813. 155654
Alpert P; Kishcha P; Kaufman Y; Schwarzbard R. (2005). Global dimming or local dimming?: Effect of urbanization on
sunlight availability. Geophys Res Lett, 32: 17. 190047
AMAP. (2004). AMAP Assessment 2002: Persistent Organic Pollutants in the Arctic. Arctic Monitoring and Assessment
Programme. Oslo, Norway. 190168
Ammann CM; Meehl GA; Washington WM; Zender CS. (2003). A monthly and latitudinally varying volcanic forcing
dataset in simulations of 20th century climate. Geophys Res Lett, 30: 1657. 192057
Anderson HR; Atkinson RW; Bremner SA; Marston L. (2003). Particulate air pollution and hospital admissions for
cardiorespiratory diseases: are the elderly at greater risk?. Eur Respir J, 40: 39s-46s. 054820
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009 9-197
-------
Anderson T; Charlson R; BellouinN; Boucher O; Chin M; Christopher S; Hay wood J; Kaufman Y; Kinne S; Ogren J.
(2005). An "A-Train" strategy for quantifying direct climate forcing by anthropogenic aerosols. Bull Am Meteorol
Soc, 86: 1795-1809. 189993
Anderson T; Wu Y; Chu D; Schmid B; Redemann J; Dubovik O. (2005). Testing the MODIS satellite retrieval of aerosol
fine-mode fraction. J Geophys Res, 110: D18204. 189991
Anderson TL; Ogren JA. (1998). Determining aerosol radiative properties using the TSI 3563 integrating nephelometer.
Aerosol Sci Technol, 29: 57-69. 156213
Andreae MO; Gelencser A. (2006). Black carbon or brown carbon? The nature of light-absorbing carbonaceous aerosols.
Atmos Chem Phys, 6: 3131-3148. 156215
Andreae MO; Rosenfeld D; Artaxo P; Costa AA; Frank GP; Longo KM; Silva-Dias MAP. (2004). Atmospheric science:
smoking rain clouds over the Amazon. Sci New York, 303: 1337-1341. 155658
Andrews E; Sheridan PJ; Ogren JA; Ferrare R. (2004). In situ aerosol profiles over the Southern Great Plains cloud and
radiation test bed site: 1. Aerosol optical properties. J Geophys Res, 109: D06208. 190058
Andronova NG; Rozanov EV; Yang F; Schlesinger ME; Stenchikov GL. (1999). Radiative forcing by volcanic aerosols
from 1850 to 1994. J Geophys Res, 104: 16807-16826. 192286
Antufia JC; Robock A; Stenchikov G; Zhou J; David C; Barnes J; Thomason L. (2003). Spatial and temporal variability of
the stratospheric aerosol cloud produced by the 1991 Mount Pinatubo eruption. J Geophys Res, 108: 4624. 192251
Arizona DEQ. (2000). Phoenix metropolitan area visibility briefing paper. Arizone Department of Environmental Quality.
Phoenix, AZ. 019164
Armbrust DV; Retta A. (2002). Wind and sandblast damage to growing vegetation. Ann Arid Zone, 39: 273-284. 156225
Arnott WP; Moosmuller H; Rogers CF; Jin T; Bruch R. (1999). Photoacoustic spectrometer for measuring light absorption
by aerosol; instrument description. Atmos Environ, 33: 2845-2852. 020650
Audet P; Charest C. (2007). Heavy metal phytoremediation from a meta-analytical perspective. Environ Pollut, 147: 231-
237. 190169
Augustine J; Hodges G; Dutton E; Michalsky J; Cornwall C. (2008). An aerosol optical depth climatology for NOAA's
national surface radiation budget network (SURFRAD). J Geophys Res, 113: D11204. 189913
Ausset P; Bannery F; Del Monte M; Lefevre RA. (1998). Recording of pre-industrial atmospheric environment by ancient
crusts on stone monuments. Atmos Environ, 32: 2859-2863. 040480
Avila A; Rodrigo A. (2004). Trace metal fluxes in bulk deposition, throughfall and stemflow at two evergreen oakstands in
NE Spain subject to different exposure to the industrial environment. Atmos Environ, 38: 171-180. 155664
Babich P; Wang PY; Allen G; Sioutas C; Koutrakis P. (2000). Development and Evaluation of a Continuous Ambient
PM2.5 Mass Monitor. Aerosol Sci Technol, 32: 309-324. 156239
Bachelet D; Lenihan J; Neilson R; Drapek R; Kittel T. (2005). Simulating the response of natural ecosystems and their fire
regimes to climatic variability in Alaska. Can J For Res, 35: 2244-2257. 156241
Backe C; Cousins IT; Larsson P. (2004). PCB in soils and estimated soil-air exchange fluxes of selected PCB congeners in
the south of Sweden. Environ Pollut, 128: 59-72. 155668
Backstrom M; Nilsson U; Hakansson K; Allard B; Karlsson S. (2003). Speciation of heavy metals in road runoff and
roadside total deposition. Water Air Soil Pollut, 147: 343-366. 156242
Bailey R; Barrie LA; Halsall C J; Fellin P; Muir DCG (2000). Atmospheric organochlorine pesticides in the western
Canadian Arctic: evidence of trans-Pacific transport. J Geophys Res, 105: 11805-11811. 155670
Baker M; Charlson R. (1990). Bistability of CCN concentrations and thermodynamics in the cloud-topped boundary layer.
Nature, 345: 142-145. 190016
Balkanski Y; Schulz M; Claquin T; Guibert S. (2007). Reevaluation of mineral aerosol radiative forcings suggests a better
agreement with satellite and AERONET data. Atmos Chem Phys, 7: 81-95. 189979
Baptista MS; Vasconcelos MTSD; Cabral JP; Freitas MC; Pacheco AMG (2008). Copper, nickel and lead in lichen and tree
bark transplants over different periods of time. Environ Pollut, 151: 408-413. 155673
December 2009 9-198
-------
Baran AJ; Foot JS. (1994). New application of the operational sounder HIRS in determining a climatology of sulphuric
acid aerosol from the Pinatubo eruption. J Geophys Res, 99: 25673-25679. 192032
Barker ML. (1976). Planning for environmental indices: observer appraisals of air quality. In K Craig; E Zube
(Ed.),Perceiving Environmental Quality (pp. 175-203). New York: Plenum Press. 072137
Barnes JE; Hofmann DJ. (1997). Lidar measurements of stratospheric aerosol over Mauna Loa Observatory. Geophys Res
Lett, 24: 192371926. 192044
Bates T; Anderson T; Baynard T; Bond T; Boucher O; Carmichael G; Clarke A; Erlick C; Guo H; Horowitz L. (2006).
Aerosol direct radiative effects over the northwest Atlantic, northwest Pacific, and North Indian Oceans: estimates
based on in-situ chemical and optical measurements and chemical transport modeling. Atmos Chem Phys Discuss,
6: 175-362. 189912
Bates T; Huebert B; Gras J; Griffiths B; Durkee P. (1998). The International Global Atmospheric Chemistry (IGAC)
Project?s First Aerosol Characterization Experiment (ACE l)-Overview. J Geophys Res, 103: 16297-16318.
190063
Bates T; Quinn P; Coffman D; Covert D; Miller T; Johnson J; Carmichael G; Uno I; Guazzotti S; Sodeman D. (2004).
Marine boundary layer dust and pollutant transport associated with the passage of a frontal system over eastern
Asia. J Geophys Res, 109: D19S19. 189958
Bates TS; Quinn PK; Coffman DJ; Johnson JE; Miller TL; Covert DS; Wiedensohler A; Leinert S; Nowak A; Neususs C.
(2001). Regional physical and chemical properties of the marine boundary layer aerosol across the Atlantic during
Aerosols99: an overview. J Geophys Res, 106: 20,767-20,782. 043385
Bauman JJ; Russell PB; Geller MA; Hamill P. (2003). A stratospheric aerosol climatology from SAGE II and CLAES
measurements: 2. Results and comparisons. J Geophys Res, 108: 4383. 192265
Baumgardner RE Jr; Isil SS; Bowser JJ; Fitzgerald KM. (1999). Measurements of rural sulfur dioxide and particle sulfate:
analysis of CASTNet data, 1987 through 1996. J Air Waste Manag Assoc, 49: 1266-1279. 011308
Baynard T; Lovejoy ER; Pettersson A; Brown SS; Lack D; Osthoff H; Massoli P; Ciciora S; Dube WP; Ravishankara AR.
(2007). Design and application of a pulsed cavity ring-down aerosol extinction spectrometer for field
measurements. Aerosol Sci Technol, 41: 447-462. 151669
BBC Research & Consulting. (2002). Phoenix Area Visibility Survey. Draft Report. Available at
http://www.azdeq.gov/environ/air/download/vis_021903f.pdf. 156258
Becker T; Brandel M. (2007). Vegetation-environment relationships in a heavy metal-dry grassland complex. Folia Geobot,
42: 11-28. 156260
Beeby A; Richmond L. (2002). Evaluating Helix aspersa as a sentinel for mapping metal pollution. Ecol Indicat, 1: 261-
270. 155680
BellouinN; Boucher O; Haywood J; Reddy MS. (2005). Global emissions of aerosol direct radiative forcing from satellite
measurements. Nature, 438: 1138-1141. 155684
BellouinN; Boucher O; Tanre D; Dubovik O. (2003). Aerosol absorption over the clear-sky oceans deduced from
POLDER-1 and AERONET observations. Geophys Res Lett, 30: 1748. 189911
Bellouin N; Jones A; Haywood J; Christopher S. (2008). Updated estimate of aerosol direct radiative forcing from satellite
observations and comparison against the Hadley Centre climate model. J Geophys Res, 113: D10205. 189999
Beron K; Murdoch J; Thayer M. (2001). The Benefits of Visibility Improvement: New Evidence from the Los Angeles
Metropolitan Area. J R Estate Finance Econ, 22: 319-337. 156270
Berthelsen BO; Olsen RA; Steinnes E. (1995). Ectomycorrhizal heavy metal accumulation as a contributing factor to heavy
metal levels in organic surface soils. Sci Total Environ, 170: 141-149. 078058
Bhavsar SP; Diamond ML; Evans LJ; Gandhi N; Nilsen J; Antunes P. (2004). Development of a coupled metal speciation-
fate model for surface aquatic systems. Environ Toxicol Chem, 23: 1376- 1385. 155689
Bhavsar SP; Diamond ML; Gandhi N; Nilsen J. (2004). Dynamic coupled metal transport-speciation model: application to
assess a zinc-contaminated lake. Environ Toxicol Chem, 23: 2410-2420. 155690
Bickerstaff K; Walker G. (2001). Public understandings of air pollution: the 'localisation'of environmental risk. Global
Environ Change, 11: 133-145. 156271
December 2009 9-199
-------
Bidleman TF; Jantunen LMM; Helm PA; Bronstrom-Lunden E; Juntto S. (2002). Chlordane isomers and enantiomers
suggest changing sources in Arctic air. Environ Sci Technol, 36: 539-544. 155691
Billings WD. (1978). Plants and the ecosystem. Belmont, CA: Wadsworth Publishing Company, Inc. 034165
Bingen C; Fussen D; Vanhellemont F. (2004). A global climatology of stratospheric aerosol size distribution parameters
derived from SAGE II data over the period 198472000: 2. Reference data . J Geophys Res, 102: D06202. 192262
Blanchard C; Tanenbaum S. (2006). Weekday/weekend differences in ambient air pollutant concentrations in Atlanta and
the southeastern United States. J Air Waste Manag Assoc, 56: 271-284. 190005
Blanchard CL; Tanenbaum S; Hidy GM. (2007). Effects of sulfur dioxide and oxides of nitrogen emission reductions on
fine particulate matter mass concentrations: regional comparisons. J Air Waste Manag Assoc, 57: 1337-50. 098659
Bluth GJS; Doiron SD; Schnetzler CC; Krueger AJ; Walter LS. (1992). Global tracking of the SO2 clouds from the June,
1991 Mount Pinatubo eruptions . Geophys Res Lett, 19: 151-154. 192029
Bluth GJS; Rose WI; SprodlE. (1997). Stratospheric loading of sulfur from explosive volcanic eruptions. J Geol, 105: 671-
683. 192045
Bonazza A; Sabbioni C; Ghedini N. (2005). Quantitative data on carbon fractions in interpretation of black crusts and
soiling on European built heritage. Atmos Environ, 39: 2607-2618. 155695
Bond T; Bhardwaj E; Dong R; Jogani R; Jung S; Roden C; Streets D; TrautmannN. (2007). Historical emissions of black
and organic carbon aerosol from energy-related combustion, 1850-2000. Global Biogeochem Cycles, 21: GB2018.
190050
Bond TC; Anderson TL; CampbellD. (1999). Calibration and Intercomparison of Filter-Based Measurements of Visible
Light Absorption by Aerosols. Aerosol Sci Technol, 30: 582-600. 156281
Bond TC; Bergstrom RW (2005). Light absorption by carbonaceous particles: an investigative review. Aerosol Sci
Technol, 39: 1-41. 155696
Bond TC; Streets DG; Yarber KF; Nelson SM; Woo J-H; Klimont Z. (2004). A technology-based global inventory of black
and organic carbon emissions from combustion. J Geophys Res, 109: D14203. 056389
Bond TC; Sun H. (2005). Can reducing black carbon emissions counteract global warming?. Environ Sci Technol, 39:
5921-5926. 156282
Boucher O; Tanre D. (2000). Estimation of the aerosol perturbation to the Earth's radiative budget over oceans using
POLDER satellite aerosol retrievals. Geophys Res Lett, 27: 1103-1106. 190041
Boucher U; Balabane M; Lamy I; Cambier P. (2005). Decomposition in soil microcosms of leaves of the metallophyte
Arabidopsis halleri: Effect of leaf-associated heavy metals on biodegradation. Environ Pollut, 135: 187-194.
155699
Brasseur G; Granier C. (1992). Mount Pinatubo aerosols, chlorofluorocarbons, and ozone depletion. Science, 257: 1239-
1242. 192256
Brenguier J; Chuang P; Fouquart Y; Johnson D; Parol F; Pawlowska H; Pelon J; Schuller L; Schroder F; Snider J. (2000).
An overview of the ACE-2 CLOUD YCOLUMN closure experiment. Tellus B Chem Phys Meteorol, 52: 815-827.
189966
Brimblecombe P; Grossi CM. (2005). Aesthetic thresholds and blackening of stone buildings. Sci Total Environ, 349: 175-
189. 155703
Broccoli AJ; Dixon KW; Delworth TL; Knutson TR; Stouffer RJ; Zeng F. (2003). Twentieth-century temperature and
precipitation trends in ensemble climate simulations including natural and anthropogenic forcing. J Geophys Res,
108:4798. 192283
Brookshire D. (1979). Methods Development for Assessing Air Pollution Control Benefits. Volume 2: Experiments in
Valuing Nonmarket Goods. A Case Study of Alternative Benefit Measures of Air Pollution in the South Coast Air
Basin of Southern California. EPA. Washington, DC. 156298
Burkhardt J; Kaiser H; Kappen L; Goldbach HE. (2002). The possible role of aerosols on stomatal conductivity for water
vapour. Basic Appl Ecol, 2: 351-364. 155708
December 2009 9-200
-------
Burt R; Wilson MA; Mays MD; Lee CW. (2003). Major and trace elements of selected pedons in the USA. J Environ Qual,
32:2109-2121. 155709
Campbell D; Copeland S; Cahill T. (1995). Measurement of Aerosol Absorption Coefficient from Teflon Filters using
Integrating Plate and Integrating Sphere Techniques. Aerosol Sci Technol, 22: 287-292. 190171
Campbell FW. (1983). Ambient Stressors. Environ Behav, 15: 355-380. 190172
Camuffo D. (1995). Physical weathering of stones. Presented at In: Saiz-Jimenez, C., ed. The deterioration of monuments:
proceedings of the 2nd international symposium on biodeterioration and biodegradation; January 1994; Sevilla,
Spain. Sci. Total Environ. 167: 1-14. 076278
Canada-US Air Quality Committee. (2004). Canada-United States transboundary particulate matter science assessment.
Meteorological Service of Canada. Toronto, Ontario, Canada. En56-203/2004E. http://www.msc-
smc.ec.gc.ca/saib/smog/transboundary/toc_e.html. 190519
Carmichael G; Tang Y; Kurata G; Uno I; Streets D; Thongboonchoo N; Woo J; Guttikunda S; White A; Wang T. (2003).
Evaluating regional emission estimates using the TRACE-P observations. J Geophys Res, 108: 8810. 190042
Carmichael GR; Calori G; Hayami H; Uno I; Cho SY; Engardt M; Kim SB; Ichikawa Y; Ikeda Y; Woo JH; Ueda H; Amann
M. (2002). The MICS-Asia study: model intercomparison of long-range transport and sulfur deposition in East
Asia. Atmos Environ, 36: 175-199. 148319
Carrico C; Kreidenweis S; Malm W; Day D; Lee T; Carrillo J; McMeeking G; Collett J. (2005). Hygroscopic growth
behavior of a carbon-dominated aerosol in Yosemite National Park. Atmos Environ, 39: 1393-1404. 190052
Carvalho LM; Cacador I; Martins-Loucao MA. (2006). Arbuscular mycorrhizal fungi enhance root cadmium and copper
accumulation in the roots of the salt marsh plant Aster tripolium L. Plant Soil, 285: 161-169. 155715
CCSP (2008). Climate Projections Based on Emissions Scenariosfor Long-lived and Short-lived Radiatively Active Gases
andAerosols. A Report by the U.S. Climate Change Science, Programand the Subcommittee on Global Change
Research. Department of Commerce. Washington, DC. 192028
Chameides WL; Yu H; Liu SC; Bergin M; Zhou X; Mearns L; Wang G; Kiang CS; Saylor RD; Luo C; Huang Y; Steiner A;
Giorgi F. (1999). Case study of the effects of atmospheric aerosols and regional haze on agriculture: an opportunity
to enhance crop yields in China through emission controls?. PNAS, 96: 13626-13633. 011184
Chand D; Anderson T; Wood R; Charlson R; Hu Y; Liu Z; Vaughan M. (2008). Quantifying above-cloud aerosol using
spaceborne lidar for improved understanding of cloudy-sky direct climate forcing. J Geophys Res, 113: D13206.
189974
Charlson R; Pilat M. (1969). Climate: the influence of aerosols. J Appl Meteorol, 8: 1001-1002. 190025
Charlson R; Wigley T. (1994). Sulfate aerosol and climate change. Sci Am, 270: 28-35. 189989
Charlson RJ; Ahlquist NC; Horvath H. (1968). On the generality of correlation of atmospheric aerosol mass concentration
and light scatter. Atmos Environ, 2: 455-464. 095355
Charlson RJ; Langner J; Rodhe H; Leovy CB; Warren SG (1991). Perturbation of the northern hemisphere radiative
balance by backscattering from anthropogenic sulfate aerosols. Tellus Dyn Meteorol Oceanogr, 43AB: 152-163.
045793
Charlson RJ; Schwartz SE; Hales JM; Cess RD; Coakley JA Jr; Hansen JE; Hofmann DJ. (1992). Climate forcing by
anthropogenic aerosols. Science, 255: 423-430. 045794
Cheggour M; Chafik A; Fisher NS; Benbrahim S. (2005). Metal concentrations in sediments and clams in four Moroccan
estuaries. Mar Environ Res, 59: 119-137. 155723
Chen W; Kahn R; Nelson D; Yau K; Seinfeld J. (2008). Sensitivity of multiangle imaging to the optical and microphysical
properties of biomass burning aerosols. J Geophys Res, 113: D10203. 189984
Chen Y; Penner JE. (2005). Uncertainty analysis of the first indirect aerosol effect. Atmos Chem Phys Discuss, 5: 4507-
4543. 199918
Chernyak SM; Rice CP; McConnell LL. (1996). Evidence of currently used pesticides in air, ice, fog, seawater and surface
microlayer in the Bering and Chukchi Seas. Mar Pollut Bull, 32: 410-419. 155726
December 2009 9-201
-------
Chestnut LG; Dennis RL. (1997). Economic benefits of improvements in visibility: acid rain Provisions of the 1990 Clean
Air Act Amendments. J Air Waste Manag Assoc, 47: 395-402. 014525
Chin M; Diehl T; Ginoux P; Malm W. (2007). Intercontinental transport of pollution and dust aerosols: implications for
regional air quality. Atmos Chem Phys, 7: 5501-5517. 190062
Chin M; Ginoux P; Kinne S; Torres O; Holben B; Duncan B; Martin R; Logan J; Higurashi A; Nakajima T. (2002).
Tropospheric aerosol optical thickness from the GOCART model and comparisons with satellite and Sun
photometer measurements. J Atmos Sci, 59: 461-483. 189996
Chin M; Kahn RA; Remer LA; Yu H; Rind D; Feingold G; Quinn PK; Schwartz SE; Streets DG; DeCola P; Halthorne R.
(2009). Atmospheric aerosol properties and climate impacts. Synthesis and assessment product 2.3. A report by the
U.S. Climate Change ScienceProgram and the Subcommittee on Global Change Research. National Aeronautics
and Space Administration. Washington, DC. 192130
Chiou CT; Sheng G; Manes M. (2001). A partition-limited model for the plant uptake of organic contaminants from soil
and water. Environ Sci Technol, 35: 1437-1444. 156342
Chipperfield MP; Randel WJ; Bodeker GE; Johnston P. (2003). Global ozone: past and future. In Scientific Assessment of
Ozone Depletion: 2002 (pp. 4.1-4.90). Geneva: World Meteorological Organization. 192275
Chock DP; Song Q; Hass H; Schell B; Ackermann I. (2003). Comment on "Control of fossil-fuel particulate black carbon
and organic matter, possibly the most effective method of slowing global warming" by Jacobson, M. Z. J Geophys
Res, 108: ACH12.1-ACH13.4. L55727
Chou M; Chan P; Wang M. (2002). Aerosol radiative forcing derived from SeaWiFS-retrieved aerosol optical properties. J
Atmos Sci, 59: 748-757. 190008
Chow JC; Watson JG; Edgerton SA; Vega E. (2002). Chemical composition of PM10 and PM25 in Mexico City during
winter 1997. Sci Total Environ, 287: 177-201. 036166
Chow JC; Watson JG; Lowenthal DH; Richards LW. (2002). Comparability between PM25 and particle light scattering
measurements. Environ Monit Assess, 79: 29-45. 037784
Christie P; Li XL; Chen BD. (2004). Arbuscular mycorrhiza can depress translocation of zinc to shoots of host plants in
soils moderately polluted with zinc. Plant Soil, 261: 209-217. 190174
Christopher S; Jones T. (2008). Short-wave aerosol radiative efficiency over the global oceans derived from satellite data.
Tellus B Chem Phys Meteorol, 60: 636-640. 189985
Christopher SA; Zhang J. (2002). Daytime variation of shortwave direct radiative forcing of biomass burning aerosols from
GOES-8 imager. J Atmos Sci, 59: 681-691. 190031
Christopher SA; Zhang J; Kaufman YJ; Remer L. (2006). Satellite-based assessment of the top of the atmosphere
anthropogenic aerosol radiative forcing over cloud-free oceans. Geophys Res Lett, 111: L15816. 155729
Chu D; Kaufman Y; Ichoku C; Remer L; Tanre D; Holben B. (2002). Validation of MODIS aerosol optical depth retrieval
over land. Geophys Res Lett, 29: 8007. 190001
Chuang CC; Penner JE; Prospero JM; Grant KE; Rau GH; Kawamoto K. (2002). Cloud susceptibility and the first aerosol
indirect forcing: Sensitivity to black carbon and aerosol concentrations. J Geophys Res, 107: 4564. 199912
Chung C; Zhang G. (2004). Impact of absorbing aerosol on precipitation: Dynamic aspects in association with convective
available potential energy and convective parameterization closure and dependence on aerosol heating profile. J
Geophys Res, 109: D22103. 190054
Chung CE; Ramanathan V; Kim D; Podgorny IA. (2005). Global anthropogenic aerosol direct forcing derived from
satellite and ground-based observations. J Geophys Res, 110: D24207. 155733
Chung SH; Seinfeld JH. (2002). Global distribution and climate forcing of carbonaceous aerosols. J Geophys Res, 107:
4407. 155732
Citterio S; Prato N; Fumagalli P; Aina R; Massa N; Santagostino A; Sgorbati S; Berta G. (2005). The arbuscular
mycorrhizal fungus Glomus mosseae induces growth and metal accumulation changes in Cannabis sativa L.
Chemosphere, 59: 21-29. 190176
December 2009 9-202
-------
Clarke A; Porter J; Valero F; Pilewskie P. (1996). Vertical profiles, aerosol microphysics, and optical closure during the
Atlantic Stratocumulus Transition Experiment: Measured and modeled column optical properties. J Geophys Res,
101:4443-4453. 190003
Clemens S. (2006). Toxic metal accumulation, responses to exposure and mechanisms of tolerance in plants. Biochimie,
88: 1707-1719. 190179
Coakley JA; Walsh CD. (2002). Limits to the aerosol indirect radiative effect derived from observations of ship tracks. J
Atmos Sci, 59: 668-680. 192025
Coelho JP; Rosa M; Pereira E; Duarte A; Pardal MA. (2006). Pattern and annual rates of Scrobicularia plana mercury
bioaccumulation in a human induced mercury gradient (Ria de Aveiro, Portugal). Estuar Coast Mar Sci, 69: 629-
635. 190181
Cohen S; Evans GW; Stokols D; Krantz DS. (1986). Behavior, Health, and Environmental Stress. New York: Plenum Press.
190182
Collins D; Johnsson H; Seinfeld J; Flagan R; Gasso S; Hegg D; Russell P; Schmid B; Livingston J; Ostrom E. (2000). In
situ aerosol-size distributions and clear-column radiative closure during ACE-2. Tellus B Chem Phys Meteorol, 52:
498-525. 190059
Collins RN; Merrington G; McLaughlin MJ; Morel JL. (2003). Organic ligand and pH effects on isotopically exchangeable
cadmium in polluted soils. Soil Sci Soc Am J, 67: 112-121. 155737
Collins W; Rasch P; Eaton B; Khattatov B; Lamarque J; Zender C. (2001). Simulating aerosols using a chemical transport
model with assimilation of satellite aerosol retrievals: Methodology forlNDOEX. J Geophys Res, 106: 7313-7336.
189987
Conant W; VanReken T; Rissman T; Varutbangkul V; Jonsson H; Nenes A; Jimenez J; Delia A; Bahreini R; Roberts G.
(2004). Aerosol-cloud drop concentration closure in warm cumulus. J Geophys Res, 109: D13204. 190010
Cooke W; Wilson J. (1996). A global black carbon aerosol model. J Geophys Res, 101: 19,395-19,409. 190046
Cooke WF; Liousse C; Cachier H; Feichter J. (1999). Construction of a 1 ° x 1 ° fossil fuel emission data set for
carbonaceous aerosol and implementation and radiative impact in the ECHAM 4 model. J Geophys Res, 104:
22137-22162. 156365
Costa M; Silva A; Levizzani V (2004). Aerosol characterization and direct radiative forcing assessment over the ocean.
Part I: Methodology and sensitivity analysis. J Appl Meteorol, 43: 1799-1817. 190006
Costa MJ; Levizzani V; Silva AM. (2004). Aerosol characterization and direct radiative forcing assessment over the ocean.
Part IIApplication to test cases and validation. J Appl Meteorol, 43: 1818-1833. 192022
Couto JA; Fernandez JA; Aboal JR; Carballeira A. (2004). Active biomonitoring of element uptake with terrestrial mosses:
a comparison of bulk and dry deposition. Sci Total Environ, 324: 211-222. 155739
Covert DS; Charlson RJ; Ahlquist NC. (1972). A study of the relationship of chemical composition and humidity to light
scattering by aerosols. J Appl Meteorol, 11: 968-976. 072055
Crawford DW; Lipsen MS; Purdie DA; Lohan MC; Statham PJ; Whitney FA; Putland JN; Johnson WK; Sutherland N;
Peterson TD; Harrison PJ; Wong CS. (2003). Influence of zinc and iron enrichments on phytoplankton growth in
the northeastern subarctic Pacific. Limnol Oceanogr, 48: 1583-1600. 156370
Creighton PJ; Lioy PJ; Haynie FH; Lemmons TJ; Miller JL; Gerhart J. (1990). Soiling by atmospheric aerosols in an urban
industrial area. J Air Waste Manag Assoc, 40: 1285-1289. 044578
Crepineau C; Rychen G; Feidt C; Le Roux Y; Lichtfouse E; Laurent F. (2003). Contamination of pastures by poly cyclic
aromatic hydrocarbons (PAHs) in the vicinity of a highway. J Agric Food Chem, 51: 4841-4845. 155741
Croteau MN; Luoma SN; Stewart AR. (2005). Trophic transfer of metals along freshwater food webs: Evidence of
cadmium biomagnification in nature. Limnol Oceanogr, 50: 1511-1519. 156373
Cunningham M. (1979). Weather, mood, and helping behavior: Quasi experiments with the sunshine Samaritan. J Pers Soc
Psychol, 37: 1947-1956. 191974
Cuny D; Denayer FO; De Foucault B; Schumacker R; Colein P; Van Haluwyn C. (2004). Patterns of metal soil
contamination and changes in terrestrial cryptogamic communities. Environ Pollut, 129: 289-297. 155742
December 2009 9-203
-------
Da Silva LC; Oliva MA; Azevedo AA; De Araujo JM. (2006). Responses of restinga plant species to pollution from an iron
pelletization factory. Water Air Soil Pollut, 175: 241-256. 190190
Dallinger R; Wang Y; Berger B; Mackay EA; Kagi JH. (2001). Spectroscopic characterization of metallothionein from the
terrestrial snail, Helix pomatia. Eur J Med Chem, 268: 4126-4133. 192109
Dameris M; Grewe V; Ponater M; Deckert R; Eyring V; Mager F; Matthes S; Schnadt C; Stenke A; Steil B; Bruhl C;
Giorgetta MA. (2005). Long-term changes and variability in a transient simulation with a chemistry-climate model
employing realistic forcing. Atmos Chem Phys Discuss, 5: 2121-2145. 192055
Davis AP; Shokouhian M; Ni S. (2001). Loading estimates of lead, copper, cadmium, and zinc in urban runoff from
specific sources. Chemosphere, 44: 997-1009. 024933
Davis MRH; Zhao FJ; McGrath SP (2004). Pollution-induced community tolerance of soil microbes in response to a zinc
gradient. Environ Toxicol Chem, 23: 2665-2672. 155744
Day R. (2007). Place and the experience of air quality. Health Place, 13: 249-260. 156386
DeBell L. (2006). Spatial and seasonal patterns and temporal variaibility of haze and its constituents in the United States:
Report IV. 1 Cooperative Institute for Research in the Atmosphere. Colorado State University Fort Collins, CO.
156388
Delworth T; Ramaswamy V; Stenchikov G (2005). The impact of aerosols on simulated ocean temperature and heat
content in the 20th century. Geophys Res Lett, 32: N/A. 190055
Demerjian KL; Mohnen VA. (2008). Synopsis of the temporal variation of particulate matter composition and size. J Air
Waste Manag Assoc, 58: 216-233. 156392
Denman KL; Brasseur G; Chidthaisong A; Ciais P; Cox PM; Dickinson RE; Hauglustaine D; Heinze C; Holland E; Jacob
D; Lohmann U; Ramachandran S; da Silva Dias PL; Wofsy SC; Zhang X. (2007). Couplings Between Changes in
the Climate System and Biogeochemistry. In Climate Change 2007: The Physical Science Basis. Contribution of
Working Group I to the Fourth AssessmentReport of the Intergovernmental Panel on Climate Change (pp. 499-
587). Cambridge and New York: Cambridge University Press. 156394
Dentener F; Stevenson D; Ellingsen K; Van Noije T; Schultz M; Amann M; Atherton C; Bell N; Bergmann D; Bey I;
Bouwman L; Butler T; Cofala J; Collins B; Drevet J; Doherty R; Eickhout B; Eskes H; Fiore A; Gauss M;
Hauglustaine D; Horowitz L; Isaksen ISA; Josse B; Lawrence M; Krol M; Lamarque JF; Montanaro V; Muller JF;
Peuch VH; Pitari G; Pyle J; Rast S; Rodriguez J; Sanderson M; Savage NH; Shindell D; Strahan S; Szopa S; Sudo
K; Van Dingenen R; Wild O; Zeng G. (2006). The global atmospheric environment for the next generation. Environ
Sci Technol, 40: 3586-3594. 088434
Derome J; Lindroos A-J. (1998). Effects of heavy metal contamination on macronutrient availability and acidification
parameters in forest soil in the vicinity of the Harjavalta Cu-Ni smelter, SW Finland. Environ Pollut, 99: 225-232.
155749
Deshler T; Hervig ME; Hofmann DJ; Rosen JM; Liley JB. (2003). Thirty years of in situ stratospheric aerosol size
distribution measurements from Laramie, Wyoming (41°N), using balloon-borne instruments . J Geophys Res, 108:
4167. 192058
Deuze J; Breon F; Devaux C; Goloub P; Herman M; Lafrance B; Maignan F; Marchand A; Nadal F; Perry G. (2001).
Remote sensing of aerosols over land surfaces from POLDER-ADEOS-1 polarized measurements. J Geophys Res,
106:4913^1926. 192013
De Gouw J; Middlebrook A; Warneke C; Goldan P; Kuster W; Roberts J; Fehsenfeld F; Worsnop D; Canagaratna M;
Pszenny A. (2005). Budget of organic carbon in a polluted atmosphere: Results from the New England Air Quality
Study in 2002. J Geophys Res, 110: D16305. 190020
De Nicola F; Maisto G; Prati MV; Alfani A. (2005). Temporal variations in PAH concentrations in Quercus ilex L. (holm
oak) leaves in an urban area. Chemosphere, 61: 432-440. 155747
Diner D; Beckert J; Bothwell G; Rodriguez J. (2002). Performance of the MISR instrument during its first 20 months in
Earth orbit. IEEE Trans Geosci Remote Sens, 40: 1449-1466. 189967
Diner D; Beckert J; Reilly T; Bruegge C; Conel J; Kahn R; Martonchik J; Ackerman T; Davies R; Gerstl S. (1998). Multi-
angle Imaging SpectroRadiometer(MISR) instrument description and experiment overview. IEEE Trans Geosci
Remote Sens, 36: 1072-1087. 189962
December 2009 9-204
-------
Di Toro DM; McGrath JH; Berry WJ; Paquin PR; Mathew R; Wu KB; Santore RC. (2005). Predicting sediment metal
toxicity using a sediment biotic ligand model: Methodology and initial application. Environ Toxicol Chem, 24:
2410-2427. 155750
Doherty S; Quinn P; Jefferson A; Carrico C; Anderson T; Hegg D. (2005). A comparison and summary of aerosol optical
properties as observed in situ from aircraft, ship, and land during ACE-Asia. J Geophys Res, 110: D04201. 190027
Donahue WF; Allen EW; Schindler DW. (2006). Impacts of coal-fired power plants on trace metals and poly cyclic
aromatic hydrocarbons (PAHs) in lake sediments in central Alberta, Canada. J Paleolimnol, 35: 111-128. 155751
Dong Y; Zhu YG; Smith FA; Wang Y; Chen B. (2008). Arbuscular mycorrhiza enhanced arsenic resistance of both white
clover (Trifolium repens Linn.) and ryegrass (Lolium perenne L.) plants in an arsenic-contaminated soil. Environ
Pollut, 155: 174-181. 192106
Douglass AR; Schoeberl MR; Rood RB; Pawson S. (2003). Evaluation of transport in the lower tropical stratosphere in a
global chemistry and transport model. J Geophys Res, 108: 4259. 057260
Dubovik O; Holben B; Eck TF; Smirnov A; Kaufman YJ; King MD; Tanre D;Slutsker I. (2002). Variability of absorption
and optical properties of key aerosol types observed in worldwide locations . J Atmos Sci, 59: 590-608. 190202
Dubovik O; King MD. (2000). A flexible inversion algorithm for retrieval of aerosol optical properties from sun and sky
radiance measurements. J Geophys Res, 105: 20673-20696. 190197
Dubovik O; Lapyonok T; Kaufman Y; Chin M; Ginoux P; Sinyuk A. (2007). Retrieving global sources of aerosols from
MODIS observations by inverting GOCART model. Atmos Chem Phys Discuss, 7: 3629-3718. 190211
Dubovik O; Smirnov A; Holben BN; King MD;Kaufman YJ; Eck TF; Slutsker I. (2000). Accuracy assessments of aerosol
optical properties retrieved from AERONET sun and sky radiance measurements. J Geophys Res, 105: 9791-9806.
190177
Dufresne L-L; Quaas J; Boucher O; Denvil S; Fairhead L. (2005). Contrasts in the effects on climate of anthropogenic
sulfate aerosols between the 20th and the 21st century. Geophys Res Lett, 32: L21703. 199908
Duriscoe DM; Luginbuhl CB; Moore CA. (2007). Measuring Night-Sky Brightness with a Wide-Field CCD Camera. Publ.
Astron. Soc. Pac., 119: 192-213. 156411
Dusek R; Frank GP; Hildebrandt L; Curtius J; Schneider J; Walter S; Chand D; Drewnick F; Kings S; Jung D; Borrmann S;
Andreae MO. (2006). Size matters more than chemistry for cloud-nucleating ability of aerosol particles. Science,
312: 1375-1378. 155756
Eagan RC; Hobbs PV; Radke LF (1974). Measurements of cloud condensation nuclei and cloud droplet size distributions
in the vicinity of forest fires. J Appl Meteorol, 13: 553-557. 190231
Eck TF; Holben BN; Reid JS; Dubovik O; Smirnov A; O'Neill NT; Slutsker I; Kinne S. (1999). Wavelength dependence of
the optical depth of biomass burning, urban, and desert dust aerosols . J Geophys Res, 104: 31333-31350. 190390
Eck TF; Holben BN; Reid JS; Sinyuk A; Dubovik O; Smirnov A; Giles D; O'Neill NT; Tsay SC; Ji Q; Mandoos AAI; Khan
MR; Reid EA; Schafer JS; Sorokine M; Newcomb W; Slutsker I. (2008). Spatial and temporal variability of
column- integrated aerosol optical properties in the southern Arabian Gulf and United Arab Emirates in summer. J
Geophys Res, 113: in press. 190409
Edgerton ES; Hartsell BE; Jansen JJ; Hansen DA; Waid CJ; Kandasamy K. (2004). 5-Year Trend Analysis of PM2.5 Data
from the SEARCH Network. Presented at Regional and Global Perspectives on Haze: Causes, Consequences and
Controversies, Visibility, Asheville, NC. 156413
Eeva T; Lehikoinen E. (2004). Rich calcium availability diminishes heavy metal toxicity in Pied Flycatcher. Funct Ecol,
18: 548-553. 155762
Eeva T; Tanhuanpaa S; Rabergh C; Airaksinen S; Nikinmaa M; Lehikoinen E. (2000). Biomarkers and fluctuating
asymmetry as indicators of pollution- induced stress in two hole-nesting passerines. Funct Ecol, 14: 235-243.
155761
Elliot SJ; Cole DC; Krueger P; Voorberg N; Wakefield S. (1999). The power of perception: health risk attributed to air
pollution in an urban industrial neighborhood. Risk Anal, 19: 621-634. 010716
Ely DW; Leary JT; Stewart TR; Ross DM. (1991). The Establishment of the Denver Visibility Standard. Presented at.
156417
December 2009 9-205
-------
Englert N. (2004). Fine particles and human health - a review of epidemiological studies. Toxicol Lett, 149: 235-242.
087939
Ericksen JA; Gustin MS; Schorran DE; Johnson DW; Lindberg SE; Coleman JS. (2003). Accumulation of atmospheric
mercury in forest foliage. Atmos Environ, 37: 1613-1622. 155769
Ervens B; Feingold G; and SMKreidenweis SM. (2005). The influence of water-soluble organic carbon on cloud drop
number concentration. J Geophys Res, 110: D18211.1-D18211.14. 190527
Evans GW; Jacobs SV. (1982). Air Pollution and Human Behavior. In Evans, GW (Ed.),Environmental Stress (pp. 105-
132). New York, NY: Cambridge University Press. 179899
Evans GW; Jacobs SV; Dooley D; Catalano R. (1987). The Interaction of Stressful Life Events and Chronic Strains on
Community Mental Health. Am J Community Psychol, 15: 23-24. 190347
Evans GW; Jacobs SV; Frager NB. (1982). Behavioral responses to air pollution. In A Baum; J Singer (Ed.),Advances in
Environmental Psychology (pp. 237-269). Hillsdale, NJ: Erlbaum. 190521
Ewald G; Larsson P; Linge H; Okla L; Szarzi M. (1998). Biotransport of organic pollutants to an inland Alaska lake by
migrating sockeye salmon (Oncorhynchus nerka). Arctic, 51: 40-47. 190348
Fehsenfeld FC; Ancellet G; Bates TS; Goldstein AH; Hardesty RM; Honrath R; Law KS; Lewis AC; Leaitch R; McKeen S;
Meagher J; Parrish DD; Pszenny AAP; Russell PB; Schlager H; Seinfeld J; Talbot R; Zbinden R. (2006).
International consortium for atmospheric research on transport and transformation (ICARTT): North America to
Europe-overview of the 2004 summer field study . J Geophys Res, 111: D23S01.1-D23S01.36. 190531
Feichter J; Sausen R; GraBl H; Fiebig M. (2003). Comment on "Control of fossil-fuel particulate black carbon and organic
matter, possibly the most effective method of slowing global warming" by M. Z. Jacobson. J Geophys Res, 108:
4767. 155772
Feingold G. (2003). Modeling of the first indirect effect: Analysis of measurement requirements. Geophys Res Lett, 30:
ASC7.1-ASC7.4. 190547
Feingold G; Furrer R; Pilewskie P; Remer LA; Min Q; Jonsson; Haflidi. (2003). Aerosol indirect effect studies at southern
great plains during the May 2003 intensive operations period . J Geophys Res, 111: D5. 190551
Feingold G; Jiang H; Harrington J. (2005). On smoke suppression of clouds in Amazonia. Geophys Res Lett, 32:
L02804.1-L02804.4. 190550
Feingold G; Remer LA; Ramaprasad J; Kaufman YJ. (2001). Analysis of smoke impact on clouds in Brazilian biomass
burning regions: An extension of Twomey's approach. J Geophys Res, 106: 22907-22922. 190544
Feingold G; Stevens B; Cotton WR; Walko RL. (1994). An explicit microphysics/LES model designed to simulate the
Twomey Effect. Atmos Res, 33: 207-233. 190535
Feingold GW; Cotton WR; Kreidenweis SM; Davis JT. (1999). The impact of giant cloud condensation nuclei on drizzle
formation in stratocumulus: Implications for cloud radiative properties. J Atmos Sci, 56: 4100-4117. 190540
Feng MH; Shan XQ; Zhang SZ; Wen B. (2005). Comparison of a rhizosphere-based method with other one-step extraction
methods for assessing the bioavailability of soil metals to wheat. Chemosphere, 59: 939-949. 155774
Perm M; Watt J; O'Hanlon S; De Santis F: Varotsos C. (2006). Deposition measurement of particulate matter in connection
with corrosion studies. Anal Bioanal Chem, 384: 1320-30. 155135
Fernandes SD; TrautmannNM; Streets DG; Roden CA; Bond TC. (2007). Global biofuel use, 1850-2000. Global
Biogeochem Cycles, 21: GB2019.1-GB2019.15. 190554
Ferrare R; Feingold G; Ghan S; Ogren J Schmid B; Schwartz SE; Sheridan P. (2006). Preface to special section:
Atmospheric radiation measurement program May 2003 intensive operations period examining aerosol properties
and radiative influences . J Geophys Res, 111: D05S01. 190561
Fiore AM; Jacob DJ; Mathur R; Martin RV (2003). Application of empirical orthogonal functions to evaluate ozone
simulations with regional and global models. J Geophys Res, 108: 4431. 047805
Fishman J; Hoell JM; Bendura RD; McNeal RJ; Kirchhoff V (1996). NASA GTE TRACE A experiment (September-
October 2002): Overview. J Geophys Res, 101: 23865-23880. 190566
December 2009 9-206
-------
Fismes J; Perrin-Ganier C; Empereur-Bissonnet P; Morel JL. (2002). Soil-to-root transfer and translocation of polycyclic
aromatic hydrocarbons by vegetables grown on industrial contaminated soils. J Environ Qual, 31: 1649-1656.
141156
Fitzgerald JW. (1975). Approximation formulas for the equilibrium size of an aerosol particle as a function of its dry size
and composition and the ambient relative humidity. J Appl Meteor Climatol, 14: 1044-1049. 095417
Forster P; Ramaswamy V; Artaxo P; Berntsen T; Betts R; Fahey DW; Haywood J; Lean J; Lowe DC; Myhre G; Nganga J;
Prinn R; Raga G; Schultz M; Van Dorland R. (2007). Changes in atmospheric constituents and in radiative forcing,
Chapter 2. In Solomon S, Qin D; Manning M; Chen Z; Marquis M; Averyt KB; Tignor M; Miller HL (Ed.)JPCC
Fourth Assessment Report (AR4): Climate Change 2007: Working Group I Report: The Physical Science Basis.
(pp. 129-234). Cambridge, U.K. and New York, NY: Intergovernmental Panel on Climate Change; Cambridge
University Press. 092936
Fraser R; Kaufman Y (1985). The relative importance of aerosol scattering and absorption in remote sensing. IEEE Trans
Geosci Remote Sens, GE-23: 625-633. 190567
Free M; Angell JK. (2002). Effect of volcanoes on the vertical temperature profile in radiosonde data. J Geophys Res, 107:
4101.192281
Freer-Smith PH; El-khatib A; Taylor G. (2004). Capture of particulate pollution by trees: a comparison of species typical of
semi-arid areas (Ficus nitida and Eucalyptus globulus) with European and North American species. Water Air Soil
Pollut, 155: 173-187. 156451
Frescholtz TF; Gustin MS; Schorran DE; Fernandez GCJ. (2003). Assessing the source of mercury in the foliar tissue of
quaking aspen. Environ Toxicol Chem, 22: 2114-2119. 190352
Fritze H; Niini S; Mikkola K; Makinen A. (1989). Soil microbial effects of a Cu-Ni smelter in southwestern Finland. Biol
Fertil Soils, 8: 87-94. 079635
Fryer ME; Collins CD. (2003). Model intercomparison for the uptake of organic chemicals by plants. Environ Sci Technol,
37: 1617-1624. 156454
Gandhi N; Bhavsar SP; Diamond ML; Kuwabara JS; Marvin-DiPasquale M; Krabbenhoft DP. (2007). Development of a
mercury speciation, fate, and biotic uptake (BIOTRANSPEC) model: Application to Lahontan Reservoir (Nevada,
USA). Environ Toxicol Chem, 26: 2260-2273. 155781
Gao YZ; Zhu LZ. (2004). Plant uptake, accumulation and translocation of phenanthrene and pyrene in soils. Chemosphere,
55: 1169-1178. 155782
Garrett T; Zhao C; Dong X; Mace G; Hobbs P. (2004). Effects of varying aerosol regimes on low-level Arctic stratus.
Geophys Res Lett, 31: L17105.1-L17105.4. 190568
Garrett TJ; Zhao C. (2006). Increased Arctic cloud longwave emissivity associated with pollution from mid-latitudes.
Nature, 440: 787-789. 190570
Gawel JE; Ahner BA; Friedland AJ; Morel FMM. (1996). Role for heavy metals in forest decline indicated by
phytochelatin measurements. Nature, 381: 64-65. 012278
Generoso S; Bey I; Attie J-L; Breon F-M. (2007). A satellite- and model-based assessment of the 2003 Russian fires:
impact on the arctic region. J Geophys Res, 112: 5302. 155786
Geogdzhayev I; Mishchenko M; Rossow W; Cairns B; Lacis AA. (2002). Global two-channel AVHRR retrievals of aerosol
properties over the ocean for the period of NCAA-9 observations and preliminary retrievals using NOAA-7 and
NOAA-11 data. J Atmos Sci, 59: 262-278. 190574
Geron C; Rasmussen R; Arnts RR; Guenther A. (2000). A review and synthesis of monoterpene speciation from forests in
the United States. Atmos Environ, 34: 1761-1781.019095
Ghan S; Laulainen N; Easter R; Wagener R; Nemesure S; Chapman E; Zhang Y; Leung R. (2001). Evaluation of aerosol
direct radiative forcing in MIRAGE. J Geophys Res, 106: 5295-5316. 199911
Gillett NP; Hegerl GC; Allen MR; Stott PA; Schnur R. (2002). Reconciling two approaches to the detection of
anthropogenic influence on climate. J Clim, 15: 326-329. 190576
Gillett NP; Zwiers FW; Weaver AJ; Hegerl GC; Allen MR; Stott PA. (2002). Detecting anthropogenic influence with a
multi-model ensemble. J Geophys Res, 29: 31.1-31.4. 190578
December 2009 9-207
-------
Ginoux P; Chin M; Holben B; Lin SJ; Tegen I; Prospero JM; Dubovik O. (2001). Sources and distributions of dust aerosols
simulated with the GOCART model. J Geophys Res, 20: 20255-20273. 190579
Ginoux P; Horowitz LW; Ramaswamy V; Geogdzhayev IV; Holben BN; Stenchikov G; Tie X. (2006). Evaluation of
aerosol distribution and optical depth in the geophysical fluid dynamics laboratory coupled model CM2.1 for
present climate. J Geophys Res, 111: D22210. 190582
Gitay H; Brown S; Easterling W; Jallow B. (2001). Ecosystems and their goods and services. In Climate change 2001:
impacts, adaptation and vulnerability. Contribution of working group II to the third assessment report of the
Intergovernmental Panel on Climate Change (pp. 237-342). Cambridge, United Kingdom: Cambridge University
Press. 092761
Givati A; Rosenfeld D. (2004). Quantifying precipitation suppression due to air pollution. J Appl Meteorol, 43: 1038-1056.
156475
Gleason SM; Faucette DT; Toyofuku MM; Torres CA; Bagley CF. (2007). Assessing and mitigating the effects of
windblown soil on rare and common vegetation. Environ Manage, 40: 1016-1024. 155794
Goforth MR; Christoforou CS. (2006). Particle size distribution and atmospheric metals measurements in a rural area in the
South Eastern USA. Sci Total Environ, 356: 217-227. 088353
Gohre V; Paszkowski U. (2006). Contribution of the arbuscular mycorrhizal symbiosis to heavy metal phytoremediation.
Planta, 223: 1115-1122. 190355
Golaz JC; Larson VE; Cotton WR. (2002). A PDF based model for boundary layer clouds. Part I: Method and model
description. JAtmos Sci, 59: 3540-3551. 190587
Golaz JC; Larson VE; Cotton WR. (2002). A PDF-Based Model for Boundary Layer Clouds. Part II: Model Results. J
Atmos Sci, 59: 3552-3571. 190589
Gomot-de Vaufleury A; Kerhoas I. (2000). Effects of cadmium on the reproductive system of the land snail Helix aspersa.
Bull Environ Contam Toxicol, 64: 434-442. 155798
Gomot-De Vaufleury A; Pihan F. (2002). Methods for toxicity assessment of contaminated soil by oral or dermal uptake in
land snails: Metal bioavailability and bioaccumulation. Environ Toxicol Chem, 21: 820-827. 190357
Gonzalez-Chavez C; Harris PJ; Dodd J; Meharg AA. (2002). Arbuscular mycorrhizal fungi confer enhanced arsenate
resistance onHolcus lanatus. NewPhytol, 155: 163-171. 155800
Goodarzi F; Sanei H; Garrett RG; Duncan WF. (2002). Accumulation of trace elements on the surface soil around the Trail
smelter, British Columbia, Canada. Environ Geol, 43: 29-38. 155801
Gordon W; Jackson R. (2000). Nutrient concentrations in fine roots. Ecology, 81: 275-280. 155802
Goudie AS; Elaine W; Viles HA. (2002). The roles of salt and fog in weathering: a laboratory simulation of conditions the
northern Atacama Desert, Chile. Catena, 48: 255-266. 156486
Grabowski WW. (2004). An improved framework for superparameterization. [embedding of cloud-resolving model in each
column of large-scale model for small-scale and mesoscale processes representation] . JAtmos Sci, 61: 1940-1952.
190590
Grabowski WW; Wu X; Moncrieff MW (1999). Cloud resolving modeling of tropical cloud systems during phase III of
GATE. Part III: Effects of cloud microphysics . JAtmos Sci, 56: 2384-2402. 190592
Grannas AM; Hockaday WC; Hatcher PG; Thompson LG; Mosley-Thompson E. (2006). New revelations on the nature of
organic matter in ice cores. J Geophys Res, 111: D04304. 155804
Grannas AM; Jones AE; Dibb J; Ammann M; Anastasio C; Beine HJ; Bergin M; Bottenheim J; Boxe CS; Carver G; Chen
G; Crawford JH; Domine F; Frey MM; Guzman MI; Heard DE; Helmig D; Hoffmann MR; Honrath RE; Huey LG;
Hutterli M; Jacobi HW; Klan P; Lefer B; McCo. (2007). An overview of snow photochemistry: evidence,
mechanisms and impacts. Atmos Chem Phys, 7: 4329-4373. 156492
Grannas AM; ShepsonPB; Filley TR. (2004). Photochemistry and nature of organic matter in Arctic and Antarctic snow.
Global Biogeochem Cycles, 18: GB1006. 155803
Grantz DA; Garner JHB; Johnson DW. (2003). Ecological effects of particulate matter. Environ Int, 29: 213-239. 155805
December 2009 9-208
-------
Gratao PL; Polle A; Lea PJ; Azevedo RA. (2005). Making the life of heavy metal-stressed plants a little easier. Funct Plant
Biol 32: 481-494. 190364
Gray JS. (2002). Biomagnification in marine systems: The perspective of an ecologist. Mar Pollut Bull, 45: 46-52. 155806
Greenstein D; Tiefenthaler L; Bay S. (2004). Toxicity of parking lot runoff after application of simulated rainfall. Arch
Environ Contam Toxicol, 47: 199-206. 155808
Gregory JM; Stouffer RJ; Raper SCB; Stott PA; Rayner NA. (2002). An observationally based estimate of the climate
sensitivity. J dim, 15: 3117-3121. 190593
Griffith DA; Solak ME; Yorty DP. (2005). Is air pollution impacting winter orographic precipitation in Utah? Weather
modification association. J Weather Modif, 37: 14-20. 156497
Grigal DF. (2002). Inputs and outputs of mercury from terrestrial watersheds: a review. Environ Rev, 10: 1-39. 156498
Grigal DF. (2003). Mercury sequestration in forests and peatlands: A review. J Environ Qual, 32: 393-405. 155811
Grossi CM; Brimblecombe P. (2004). Aesthetics of simulated soiling patterns on architecture. Environ Sci Technol, 38:
3971-3976. 155813
Grossi CM; Esbert RM; Diaz-Pache F; Alonso FJ. (2003). Soiling of building stones in urban environments. Build Environ,
38: 147-159. 155812
GuderianR. (1977). Accumulation of pollutants in plant organs. In Air pollution: phytotoxicity of acidic gases and its
significance in air pollution control (pp. 66-74). Berlin, Germany: Springer-Verlag. 004150
Guderian R. (1985). Air pollution by photochemical oxidants: formation, transport, control, and effects on plants. New
York: Springer-Verlag. 019325
GunnR; Phillips BB. (1957). An experimental investigation of the effect of air pollution on the initiation of rain. JAtmos
Sci, 14: 272-280. 190595
Gupta P; Christopher SA; Wang J; Gehrig R; Lee Y; Kumar N. (2006). Satellite remote sensing of particulate matter and air
quality assessment over global cities. Atmos Environ, 40: 5880-5892. 137694
Gustin MS. (2003). Are mercury emissions from geologic sources significant? A status report. Sci Total Environ, 304: 153-
167. 155816
Hageman KJ; Simonich SL; Campbell DH; Wilson GR; Landers DH. (2006). Atmospheric deposition of current-use and
historic-use pesticides in snow at national parks in the western United States. Environ Sci Technol, 40: 3174-3180.
156509
Hall JL. (2002). Cellular mechanisms for heavy metal detoxification and tolerance. JExp Bot, 53: 1-11. 190365
Halsall CJ. (2004). Investigating the occurrence of persistent organic pollutants (POPs) in the Arctic: their atmospheric
behaviour and interaction with the seasonal snow pack. Environ Pollut, 128: 163-175. 155822
Halsall CJ; Barrie LA; Fellin P; Muir DCG; Rovinski FY; Kononov EY (1997). Spatial and temporal variation of
polycyclic aromatic hydrocarbons in the arctic atmosphere. Environ Sci Technol, 31: 3593-3599. 155821
HanQ; Rossow WB; Chou J; Welch RM. (1998). Global survey of the relationship of cloud albedo and liquid water path
with droplet size using ISCCP. J Clim, 11: 1516-1528. 190594
Han Q; Rossow WB; Zeng J. (2002). Three different behaviors on liquid water path of clouds in aerosol-cloud interactions.
JAtmos Sci, 59: 726-735. 049181
Hand JL; Kreidenweis SM; Sherman DE; Collett JL Jr; Hering SV; Day DE; Malm WC. (2002). Aerosol size distributions
and visibility estimates during the big bend regional aerosol visibility and observational study (BRAVO). Atmos
Environ, 36: 5043-5055. 190367
Hand JL; Malm WC. (2006). Review of the IMPROVE Equation for Estimating Ambient Light Extinction Coefficients-
Final Report. 156517
Hand JL; Malm WC. (2007). Review of aerosol mass scattering efficiencies from ground-based measurements since 1990.
JGeophysRes, 112:D16203. 155825
Hansen AC; Zhang Q; Lyne PWL. (2005). Ethanol-diesel fuel blends-a review. Bioresour Technol, 96: 277-285. 059087
December 2009 9-209
-------
Hansen AD; Rosen AH; Novakov T. (1982). Real-Time Measurement of the Absorption Coefficient of Aerosol Particles.
Appl Opt, 21: 3060-3062. 190368
Hansen ADA; Rosen H; Novakov T. (1984). The aethalometer - an instrument for the real-time measurement of optical
absorption by aerosol particles. Sci Total Environ, 36: 191-196. 002396
Hansen J; Nazarenko L. (2004). Soot climate forcing via snow and ice albedos. PNAS, 101: 423-428. 156521
Hansen J; Sato M; Mazarenko L; Ruedy R; Lacis A; Koch D; Tegen I; Hall T; Shindell D; Santer B; Stone P; Novakov T;
Thomason L; Wang R; Wang Y; Jacob D; Hollands-worth S; Bishop L; Logan J; Thompson A; Stolarski R; Lean J;
Willson R; Levitus S; Antonov J; Rayner N; Parker D; Christy J. (2002). Climate forcings in Goddard Institute for
Space Studies S12000 simulations. J Geophys Res, 107: 4347. 049177
Hansen J; Sato M; Ruedy R; Kharecha P; Lacis A; Miller R; Nazarenko L; Lo K; Schmidt GA; Russell G; Aleinov I; Bauer
S; Baum E; Cairns B; Canuto V; Chandler M; Cheng Y; Cohen A; Del Genio A; Faluvegi G; Fleming E; Friend A;
Hall T; Jackman C; Jonas J; Kelley M; Kiang NY; Koch D; Labow G; Lerner J; Menon S; Novakov T; Oinas V;
Perlwitz JA; Perlwitz Ju; Rind D; Romanou A; Schmunk R; Shindell D; Stone P; Sun S; Streets D; Tausnev N;
Thresher D; Unger N; Yao M; Zhang S. (2007). Climate simulations for 1880-2003 with GISS modelE. Clim
Dynam, 29: 661-696. 190597
Hansen J; Sato M; Ruedy R; Lacis A; Oinas V. (2000). Global warming in the twenty-first century: an alternative scenario.
PNAS, 97: 9875-9880. 042683
Hansen J; Sato M; Ruedy R; Nazarenko L; Lacis A; Schmidt GA; Russell G; Aleinov I; Bauer M; Bauer S; Bell N; Cairns
B; Canuto V; Chandler M; Cheng Y; Del Genio A; Faluvegi G; Fleming E; Friend A; Hall T; Jackman C; Kelley M;
Kiang N; Koch D; Lean J; Lerner J; Lo K; Menon S; Miller R; Minnis P; Novakov T; Oinas V; Perlwitz JA;
Perlwitz Ju; Rind D; Romanou A; Shindell D; Stone P; Sun S; Tausnev N; Thresher D; Wielicki B; Wong T; Yao M;
Zhang S. (2005). Efficacy of climate forcings. J Geophys Res, 110: D18104. 190596
Hansen JE; Sato M; Ruedy R. (1997). Radiative forcing and climate response. J Geophys Res, 102: 6831-6864. 043104
Hargrave BT; Barrie LA; Bidleman TF; Welch HE. (1997). Seasonality in exchange of organochlorines between arctic air
and seawater. Environ Sci Technol, 31: 3258-3266. 155827
Harmens H; Norris DA; Koerber GR; Buse A; Steinnes E; Ruhling A. (2007). Temporal trends in the concentration of
arsenic, chromium, copper, iron, nickel, vanadium and zinc in mosses across Europe between 1990 and 2000.
Atmos Environ, 41: 6673-6687. 155828
Harner T; Jantunen LMM; Bidleman TF; Barrie LA; Kylin H; Strachan WMJ; Macdonald RW. (2000). Microbial
degradation is a key elimination pathway of hexachlorocyclohexanes from the Arctic Ocean. Geophys Res Lett, 27:
1155-1158. 155829
Harrison L; Michalsky J; Berndt J. (1994). Automated multifilter rotating shadow-band radiometer: an instrument for
optical depth and radiation measurements. Appl Opt, 33: 5118-5125. 045805
Harvey LDD. (2004). Characterizing the annual-mean climatic effect of anthropogenic CO2 and aerosol emissions in eight
coupled atmosphere-ocean GCMs. Clim Dynam, 23: 569-599. 190598
Hassan R; Scholes R; Ash N. (2005). Ecosystems and human well-being: current state and trends, volume 1. United
Kingdom: Shearwater Books. 092759
Hauck M. (2003). Epiphytic lichen diversity and forest dieback: The role of chemical site factors. Bryologist, 106: 257-
269. 155830
Haynie FH. (1986). Environmental factors affecting corrosion of weathering steel. Presented at Materials degradation
caused by acid rain: developed from the 20th state-of-the-art symposium of the American Chemical Society; June
1985; Arlington, VA., Washington, DC. 157198
Haynie FH; Lemmons TJ. (1990). Particulate matter soiling of exterior paints at a rural site. Aerosol Sci Technol, 13: 356-
367. 044579
Haywood J; Boucher O. (2000). Estimates of the direct and indirect radiative forcing due to tropospheric aerosols: A
review. Rev Geophys, 38: 513-543. 156531
Haywood J; Francis P; Osborne S; Glew M; Loeb N; Highwood E; Tanre D; Myhre G; Formenti P; Hirst E. (2003).
Radiative properties and direct radiative effect of Saharan dust measured by the C-l 30 aircraft during SHADE: 1.
Solar spectrum : The Saharan Dust Experiment (SHADE). J Geophys Res, 108: SAH 4-1. 190599
December 2009 9-210
-------
Haywood J; Schulz M. (2007). Causes of the reduction in uncertainty in the anthropogenic radiative forcing of climate
between IPCC (2001) and IPCC (2007). Geophys Res Lett, 34: L20701. 190600
Haywood JM; Pelon J; Formenti P; Bharmal N; Brooks M; Capes G; Chazette P; Chou C; Christopher S; Coe H; Cuesta J;
Derimian Y; Desboeufs K; Greed G; Harrison M; Heese B; Highwood EJ; Johnson B; Mallet M; Marticorena B;
Marsham J; Milton S; Myhre G; Osborne SR; Parker DJ; Rajot JL; Schulz M; Slingo A; Tanre D; Tulet P. (2008).
Overview of the dust and biomass burning experiment and African monsoon multidisciplinary analysis special
observing Period-0. J Geophys Res, 113: DOOC17. 190602
Haywood JM; Ramaswamy V; Soden BJ. (1999). Tropospheric aerosol climate forcing in clear-sky satellite observations
over the oceans. Science, 283: 1299-1303. 040453
Heald CL; Jacob DJ; Park RJ; Russell LM; Huebert BJ; Seinfeld JH; Liao H; Weber RJ. (2005). A large organic aerosol
source in the free troposphere missing from current models. Geophys Res Lett, 32: L18809.1-L18809.4. 190603
Heintzenberg J. (2008). The SAMUM-1 experiment over Southern Morocco: overview and introduction. Tellus B Chem
Phys Meteorol, 61:2-11. 190605
Helena P; Franc B; Cvetka RL. (2004). Monitoring of short-term heavy metal deposition by accumulation in epiphytic
lichens (Hypogymnia Physodes (L.) Nyl). J Atmos Chem, 49: 223-230. 155833
Helm PA; Diamond ML; Semkin R; Bidleman TF. (2000). Degradation as a loss mechanism in the fate of a-
hexachlorocyclohexane in Arctic watersheds. Environ Sci Technol, 34: 812-818. 155834
Helm PA; Diamond ML; Semkin R; Strachan WMJ; Teixeira C; Gregor D. (2002). Amass balance model describing multi-
year fate of organochlorine compounds in a high arctic lake. Environ Sci Technol, 36: 996-1003. 155835
Helmisaari H-S; Makkonen K; Olsson M; Viksna A; Malkonen E. (1999). Fine-root growth, mortality and heavy metal
concentrations in limed and fertilized Pinus silvestris (L.) stands in the vicinity of a Cu-Ni smelter in SW Finland.
Plant Soil, 209: 193-200. 155836
Henze DK; Seinfeld JH. (2006). Global secondary organic aerosol from isoprene oxidation. Geophys Res Lett, 33:
L09812.1-L09812.4. 190606
Herman JR; Bhartia PK; Torres O; Hsu C; Seftor C; Celarier E. (1997). Global distribution of UV-absorbing aerosols from
Nimbus 7/TOMS data. J Geophys Res, 102: 16911-16922. 048393
Hernandez L; Probst A; Probst JL; Ulrich E. (2003). Heavy metal distribution in some French forest soils: Evidence for
atmospheric contamination. Sci Total Environ, 312: 195-219. 155841
Herner JD; Ying Q; Aw J; Gao O; Chang DPY; Kleeman MJ. (2006). Dominant Mechanisms that Shape the Airborne
Particle Size and Composition Distribution in Central California. Aerosol Sci Technol, 40: 827-844. 135981
Herrera LK; Videla HA. (2004). The importance of atmospheric effects on biodeterioration of cultural heritage
constructional materials. Int Biodeterior Biodegradation, 54: 125-134. 155843
Hicke JA; Asner GP; Kasischike ES; French NHF; Randerson JT; Collatz GJ; Stocks BJ; Tucker CJ; Los SO; Field CB.
(2003). Postfire response of North American boreal forest net primary productivity analysed with satellite
observations. Global Change Biol, 9: 1145-1157. 156545
Hinzman LD; Fukuda M; Sandberg DV; Chapin III FS; Dash D. (2003). FROSTFIRE: an experimental approach to
predicting the climate feedbacks from the changing boreal fire regime. J Geophys Res, 108: 8153. 155845
Hobbelen PHF; Koolhaas JE; van Gestel CAM. (2006). Bioaccumulation of heavy metals in the earthworms Lumbricus
rubellus and Aporrectodea caliginosa in relation to total and available metal concentrations in field soils. Environ
Pollut, 144: 639-646. 190371
Hodzic A; Madronich S; Bohn B; Massie S; Menut L; Wiedinmyer C. (2007). Wildfire particulate matter in Europe during
summer 2003: meso-scale modeling of smoke emissions, transport and radiative effects. Atmos Chem Phys, 7:
4043-4064. 156553
Hoell JM; Davis DD; Liu SC; Newell R; Shipham M; Akimoto H; McNeal RJ; Bemdura RJ; Drewry JW. (1996). Pacific
exploratory mission-west A (PEM-WEST A): September-October, 1991. J Geophys Res, 101: 1641-1653. 190607
Hoell JM; Davis DD; Liu SC; Newell RE; Akimoto H; McNeal RJ; Bendura RJ. (1997). The Pacific exploratory mission-
west phase B: February-March, 1994. J Geophys Res, 102: 28223-28239. 057373
December 2009 9-211
-------
Hoff R; Engel-Cox J; Krotkov N; Palm S; Rogers R; McCann K; Sparling L; Jordan N; Torres O; Spinhirne J. (2004).
Long-range transport observations of two large forest fire plumes to the northeastern U.S. In Proceedings of the
22nd International Laser Radar Conference (pp. 683-686.). Paris: European Space Agency. 190617
Hoff RM; McCann KJ. (2002). Regional East Atmospheric Lidar Mesonet: REALM. In L. Bissonette; G Roy; G Vallee
(Ed.),Lidar Remote Sensing in Atmospheric and Earth Sciences (pp. 281-284). Val-Belair, Quebec: Def. R&D Can.
Valcartier. 190612
Hofmann D; Barnes J; Dutton E; Deshler T; Jager H; Keen R; Osborn M. (2003). Surface-based observations of volcanic
emissions to stratosphere. In Volcanism and the Earth's Atmosphere (pp. 57-73). Washington, DC: American
Geophysical Union. 192062
Hoh E; Zhu L; Kites RA. (2006). Dechlorane plus, a chlorinated flame retardant, in the Great Lakes. Environ Sci Technol,
40: 1184-1189. 190378
Holben BN; Eck TF; Slutsker I; Tanre D; Buis JP; Setzer A; Vermote E; Reagan JA; Kaufman YJ; Nakajima T; Lavenu F;
Jankowiak I; Smirnov A. (1998). AERONET: A federated instrument network and data archive for aerosol
characterization. Rem Sens Environ, 66: 1-16. 155848
Holben BN; Smirnov A; Eck TF; Slutsker I; Abuhassan N; Newcomb WW; Schafer JS; Tanre D; Chatenet B; Lavenu F.
(2001). An emerging groundbased aerosol climatology: aerosol optical depth from AERONET. J Geophys Res, 106:
12067-12098. 190618
Holland EA; Dentener FJ; Braswell BH; Sulzman JM. (1999). Contemporary and pre-industrial global reactive nitrogen
budgets. Biogeochemistry, 46: 7-43. 092051
Holmqvist M; Stenroth P; Berglund O; Nystrom P; Olsson K; Jellyman D; Mclntosh AR; Larsson P. (2006). Low levels of
persistent organic pollutants (POPs) in New Zealand eels reflect isolation from atmospheric sources. Environ
Pollut, 141: 532-538. 190380
Hooda V. (2007). Phytoremediation of toxic metals from soil and waste water. J Environ Biol, 28: 367-376. 190382
Hopke PK; Zhou L; Poirot RL. (2005). Reconciling Trajectory Ensemble Receptor Model Results with Emissions. Environ
Sci Technol, 39: 7980-7983. 156567
Horowitz L. (2006). Past, present, and future concentrations of tropospheric ozone and aerosols: Methodology, ozone
evaluation, and sensitivity to aerosol wet removal. J Geophys Res, 111: D22211.1 -D22211.16. 190620
Horowitz LW; Walters S; Mauzerall DL; Emmons LK; Rasch PJ; Granier C; Tie X; Lamarque J-F; Schultz MG; Tyndall
GS; Orlando JJ; Brasseur GP (2003). A global simulation of tropospheric ozone and related tracers: Description
and evaluation of MOZART, version 2. J Geophys Res, 108: D24. 057770
Houlahan JE; Findlay CS; Schmidt BR; Meyer AH; Kuzmin SL. (2000). Quantitative evidence for global amphibian
population declines. Nature, 404: 752-755. 155853
Howe TS; Billings S; Stolzberg RJ. (2004). Sources of poly cyclic aromatic hydrocarbons and hexachlorobenzene in spruce
needles of eastern Alaska. Environ Sci Technol, 38: 3294-3298. 155854
Howel D; Moffatt S; Prince H; Bush J; Dunn CE. (2002). Urban Air Quality in North-East England: Exploring the
Influences on Local Views and Perceptions. Risk Anal, 22: 121-130. 156571
HoytD;FrohlichC. (1983). Atmospheric transmission at Davos, Switzerland 1909-1979. Clim Change, 5: 61-71. 190621
Hoyt DV. (1978). Amodel for the calculation of solar global insolation. Solar Energy, 21: 27-35. 046638
HsuNC; Tsay SC; King MD; Herman JR. (2004). Aerosol properties over bright-reflecting source regions. IEEE Trans
Geosci Remote Sens, 42: 557-569. 190622
Huebert BJ; Bates T; Russell PB; Shi G; Kim YJ; Kawamura K; Carmichael G; Nakajima T. (2003). An overview of ACE-
Asia: Strategies for quantifying the relationships between Asian aerosols and their climatic impacts. J Geophys Res,
108: 8633. 190623
Hughes MK. (1981). Cycling of trace metals in ecosystems. InNWLepp (Ed.),Effect of heavy metal pollution on plants.
Volume 2: metals in the environment (pp. 95-118). London, United Kingdom: Applied Science Publishers. 053595
Hunaiti AA; Al-Oqlah A; Shannag NM; Abukhalaf IK; Silvestrov NA; Von Deutsch DA; Bayorh MA. (2007). Toward
understanding the influence of soil metals and sulfate content on plant thiols. J Toxicol Environ Health A, 70: 559-
567. 156579
December 2009 9-212
-------
Huneeus N; Boucher O. (2007). One-dimensional variational retrieval of aerosol extinction coefficient from synthetic
LIDAR and radiometric measurements. J Geophys Res, 112: D14303.1-D14303.14. 190624
Hung HH; Halsall CJ; Blanchard P. (2001). Are PCBs in the Canadian Arctic atmosphere declining? Evidence from 5 years
of monitoring. Environ Sci Technol, 35: 1303-1311. 155856
Hung HH; Halsall CJ; Blanchard P; Li HH; Fellin P; Stern G; Rosenberg B. (2002). Temporal trends of organochlorine
pesticides in the Canadian Arctic atmosphere. Environ Sci Technol, 36: 862-868. 155857
Husar RB; Prospero JM; Stowe LL. (1997). Characterization of tropospheric aerosols over the oceans with the NOAA
advanced very high resolution radiometer optical thickness operational product. J Geophys Res, 102: 16,889-
16,909. 045900
IPCC. (1995). Radiative forcing of climate change and an evaluation of the IPCC IS92 emission scenarios. In Climate
Change 1994 (pp. 193-204). New York: Cambridge University Press. 190991
IPCC. (1996). Radiative forcing of climate change. In Climate Change 1995 (pp. 65-131). New York: Intergovernmental
Panel on Climate Change. 190990
IPCC. (2001). Climate Change 2001: The Scientific Basis. Cambridge, UK: Intergovernmental Panel on Climate Change,
Cambridge University Press. 156587
IPCC. (2007). IPCC Fourth Assessment Report (AR4): Climate Change 2007: Working Group I Report: The Physical
Science Basis. Cambridge, UK and New York, NY: Intergovernmental Panel on Climate Change, Cambridge
University Press. 092765
Ito A; Penne JE. (2005). Historical emissions of carbonaceous aerosols from biomass and fossil fuel burning for the period
1870-2000. Global Biogeochem Cycles, 19: GB2028.1-GB2028.14. 190626
Jacob DJ; Crawford JH; Kleb MM; Connors VS; Bendura RJ; Raper JL; Sachse GW; Gille JC; Emmons L; Heald CL.
(2003). The Transport and Chemical Evolution over the Pacific (TRACE-P) aircraft mission: design, execution, and
first results. J Geophys Res, 108: 9000. 190987
Jacobs SV; Evans GW; Catalano R; Dooley D. (1984). Air pollution and depressive symptomatology: Exploratory analyses
of intervening psychosocial factors. Popul Environ, 7: 260-272. 156596
Jacobson MZ. (2000). A physically-based treatment of elemental carbon optics: implications for global direct forcing of
aerosols. Geophys Res Lett, 27: 217-220. 056378
Jacobson MZ. (2001). GATOR-GCMM: a global through urban scale air pollution and weather forecast model: 1. Model
design and treatment of subgrid soil, vegetation, roads, rooftops, water, sea ice, and snow. J Geophys Res, 106:
5385-5402. 155864
Jacobson MZ. (2001). Global direct radiative forcing due to multicomponent anthropogenic and natural aerosols. J
Geophys Res, 106: 1551-1568. 043110
Jacobson MZ. (2002). Control of fossil-fuel particulate black carbon and organic matter, possibly the most effective
method of slowing global warming. J Geophys Res, 107: 4410. 155865
Jacobson MZ. (2003). Development of mixed-phase clouds from multiple aerosol size distributions and the effect of the
clouds on aerosol removal. J Geophys Res, 108: 4245. 155866
Jacobson MZ. (2003). Reply to comment by D. P. Chock et al. on "Control of fossil-fuel particulate black carbon and
organic matter, possibly the most effective method of slowing global warming". J Geophys Res, 108: 4770. 155867
Jacobson MZ. (2003). Reply to comment by J. E. Penner on "Control of fossil-fuel particulate black carbon and organic
matter, possibly the most effective method of slowing global warming". J Geophys Res, 108: 4772. 155868
Jacobson MZ. (2003). Reply to comment by J. E. Penner on "Control of fossil-fuel particulate black carbon and organic
matter, possibly the most effective method of slowing global warming. J Geophys Res, 108: 4772. 189421
Jacobson MZ. (2003). Reply to comment by J. Feichter et al. on "Control of fossil-fuel particulate black carbon and organic
matter, possibly the most effective method of slowing global warming". J Geophys Res, 108: 4768. 155869
Jacobson MZ. (2004). Climate response of fossil fuel and biofuel soot, accounting for soot's feedback to snow and sea ice
albedo and emissivity. J Geophys Res, 109, : D21201. 155870
December 2009 9-213
-------
Jacobson MZ. (2006). Effects of externally-through-internally-mixed soot inclusions within clouds and precipitation on
global climate. J Phys Chem B, 110: 6860-6873. 156599
Jacobson MZ; Kaufman YJ. (2006). Wind reduction by aerosol particles. Geophys Res Lett, 33: L24814. 090942
Jacobson MZ; Seinfeld JH; Carmichael GR; Streets DG (2004). The effect on photochemical smog of converting the U.S.
fleet ofgasoline vehicles to modern diesel vehicles. Geophys Res Lett, 31: L02116. 180362
James SM; Little EE; Semlitsch RD. (2004). Effects of multiple routes of cadmium exposure on the hibernation success of
the American toad (Bufo americanus). Arch Environ Contam Toxicol, 46: 518-527. 155874
Jantunen LMM; Bidleman TF. (1996). Air-water gas exchange of hexachlorocyclohexanes (HCHs) and the enantiomers of
a-HCH in Arctic regions. J Geophys Res, 101: 28837-28846. 155876
Jantunen LMM; Bidleman TF. (1998). Organochlorine pesticides and enantiomers of chiral pesticides in Arctic Ocean
water. Arch Environ Contam Toxicol, 35: 218-228. 155877
Jayanty R. (2003). Overview of PM2.5 Chemical Speciation Nationwide Network Program. Presented at Specialty
Conference of the American Association for Aerosol Research, Pittsburgh, PA, April, 2003, Research Triangle Park,
NC. 156605
Jayne JT; Leard DC; Zhang X; Davidovits P; Smith KA; Kolb CE; Worsnop DR. (2000). Development of an aerosol mass
spectrometer for size and composition analysis of submicron particles. Aerosol Sci Technol, 33: 49-70. 190978
Jeong MJ; Li Z; Chu DA; Tsay SC. (2005). Quality and Compatibility Analyses of Global Aerosol Products Derived from
the Advanced Very High Resolution Radiometer and Moderate Resolution Imaging Spectroradiometer. J Geophys
Res, 110: D10S09. 190977
Jiang H; Feingold G (2006). Effect of aerosol on warm convective clouds: Aerosol-cloud-surface flux feedbacks in a new
coupled large eddy model. J Geophys Res, 111: D01202. 190976
Jiang J; Oberdrster G; Elder A; Gelein R; Mercer P; Biswas P. (2008). Does nanoparticle activity depend upon size and
crystal phase?. Nanotoxicology, 2: 33-42. 156609
Jiang W; Smyth S; Giroux E; Roth H; Yin D. (2006). Differences between CMAQ fine mode particle and PM 25
concentrations and their impact on model performance evaluation in the lower Fraser valley. Atmos Environ, 40:
4973-4985. 133165
Jiao XC; Xu FL; Dawson R; Chen SH; Tao S. (2007). Adsorption and absorption of polycyclic aromatic hydrocarbons to
rice roots. Environ Pollut, 148: 230-235. 155879
Jirak IL; Cotton WR. (2006). Effect of air pollution on precipitation along the Front Range of the Rocky Mountains. J Appl
Meteorol, 45: 236-245. 156612
Johnson D; Hale B. (2008). Fine root decomposition and cycling of Cu, Ni, Pb, and Zn at forest sites near smelters in
Sudbury, ON, and Rouyn-Noranda, QU, Canada. Hum Ecol Risk Assess, 14: 41-53. 155881
Johnson DB. (1982). The role of giant and ultragiant aerosol particles in warm rain initiation. J Atmos Sci, 39: 448-460.
190973
Johnson MS; Franke LS; Lee RB; Holladay SD. (1999). Bioaccumulation of 2,4,6-trinitrotoluene and poly chlorinated
biphenyls through two routes of exposure in a terrestrial amphibian: Is the dermal route significant?. Environ
Toxicol Chem, 18: 873-878. 155880
Jones A; Roberts DL; Woodage MJ. (2001). Indirect sulphate aerosol forcing in a climate model with an interactive sulphur
cycle. J Geophys Res, 106: 20293-20301. 199895
Jones GS; Jones A; Roberts DL; Stott PA; Williams KD. (2005). Sensitivity of global-scale climate change attribution
results to inclusion of fossil fuel black carbon aerosol. Geophys Res Lett, 32: LI 4701. 155885
Jones JW; Bogat GA. (1978). Air pollution and human aggression. Psychol Rep, 43: 721-722. 190396
Jones PD; Moberg A; Osborn TJ; Briffa KR. (2003). Surface climate responses to explosive volcanic eruptions seen in long
European temperature records and mid-to-high latitude tree-ring density around the Northern Hemisphere. In
Robock A; Oppenheimer C (Ed.),Volcanism and the earth's atmosphere (pp. 239-254). Washington, DC: American
Geophysical Union. 192278
December 2009 9-214
-------
JoshiM; Shine K. (2003). A GCM study of volcanic eruptions as a cause of increased stratospheric water vapor. J Clim, 16:
3525-3534. 192327
Joynt J; Bischoff M; Turco R; Konopka A; Nakatsu CH. (2006). Microbial community analysis of soils contaminated with
lead, chromium and petroleum hydrocarbons. Microb Ecol, 51: 209-219. 155887
Junker C; Liousse C. (2008). A global emission inventory of carbonaceous aerosol from historic records of fossil fuel and
biofuel consumption for the period 1860-1997. Atmos Chem Phys, 8: 1195-1207. 190971
Kahn R; Banerjee P; McDonald D. (2001). The sensitivity of multiangle imaging to natural mixtures of aerosols over
ocean. J Geophys Res, 106: 18219-18238. 190969
Kahn R; Banerjee P; McDonald D; Diner D. (1998). Sensitivity of multiangle imaging to aerosol optical depth, and to
pure-particle size distribution and composition over ocean. J Geophys Res, 103: 32195-32213. 190970
Kahn R; Gaitley B; Martonchik J; Diner D; Crean K; Holben B. (2005). Multiangle Imaging Spectroradiometer (MISR)
global aerosol optical depth validation based on 2 years of coincident Aerosol Robotic Network (AERONET)
observations. J Geophys Res, 110: D10S04. 189961
Kahn R; Li W; Martonchik J; Bruegge C; Diner D; Gaitley B; Abdou W; Dubovik O; Holben B; Smirnov A; Jin Z; Clark
D. (2005). MISR low-light-level calibration, and implications for aerosol retrieval over dark water. Science, 62:
1032-1052. 190965
Kahn R; Li W; Moroney C; Diner D; Martonchik J; Fishbein E. (2007). Aerosol source plume physical characteristics from
space-based multiangle imaging. J Geophys Res, 112: D11205. 190964
Kahn RA; Gaitley BJ; Martonchik JV; Diner DJ; Crean KA; Holben B. (2005). MISR global aerosol optical depth
validation based on two years of coincident AERONET observations. J Geophys Res, 110: D10S04. 190966
Kahn RA; Garay MJ; Nelson DL; Yau KK; Bull MA; Gaitley BJ; Martonchik JV; Levy R. (2007). Satellite-derived aerosol
optical depth over dark water from MISR and MODIS: comparisons with AERONET and implications for
climatological studies. J Geophys Res, 112: D18205. 190963
Kalashnikova O; Kahn R. (2006). Ability of multiangle remote sensing observations to identify and distinguish mineral
dust types: Part 2 Sensitivity over dark water. J Geophys Res, 111: Dl 1207. 190962
Kaldorf M; Kuhn AJ; Schroder WH; Hildebrandt U; Bothe H. (1999). Selective element deposits in maize colonized by a
heavy metal tolerance conferring arbuscular mycorrhizal fungus. J Plant Physiol, 154: 718-728. 190399
Kamh GME. (2005). The impact of landslides and salt weathering on Roman structures at high latitudes - Conway Castle,
Great Britain: A case study. Environ Geol, 48: 238-254. 155888
Kandeler E; Kampichler C; Horak O. (1996). Influence of heavy metals on the functional diversity of soil microbial
communities. Biol Fertil Soils, 23: 299-306. 094392
Kappos AD; Bruckmann P; Eikmann T; Englert N; Heinrich U; Hoppe P; Koch E; Krause GH; Kreyling WG; Rauchfuss
K; Rombout P; Schulz-Klemp V; Thiel WR; Wichmann HE. (2004). Health effects of particles in ambient air. Int J
Hyg Environ Health, 207: 399-407. 087922
Kapustin VN; Clarke AD; Shinozuka Y; Howell S; V Brekhovskikh V; Nakajima T; Higurashi A. (2006). On the
determination of a cloud condensation nuclei from satellite: Challenges and possibilities. J Geophys Res, 111:
D04202. 190961
Kaufman Y. (1987). Satellite sensing of aerosol absorption. J Geophys Res, 92: 4307-4317. 190960
Kaufman Y; Boucher O; Tanre D; Chin M; Remer L; Takemura T. (2005). Aerosol anthropogenic component estimated
from satellite data. Geophys Res Lett, 32: L17804. 155891
Kaufman Y; Fraser R. (1997). The effect of smoke particles on clouds and climate forcing. Science, 277: 1636-1639.
190958
Kaufman Y; Haywood J; Hobbs P; Hart W; Kleidman R; Schmid B. (2003). Remote sensing of vertical distributions of
smoke aerosol off the coast of Africa. Geophys Res Lett, 30: 1831. 190954
Kaufman YD; Tanre D; Boucher O. (2002). A satellite view of aerosols in the climate system. Nature, 419: 215-223.
190956
December 2009 9-215
-------
Kaufman YJ; Hobbs PV; Kirchoff VWJH; Artaxo P; Remer LA; Holben BN; King MD; Ward DE; Prins EM; Longo KM;
Mattos LF; Nobre CA; Spinhirne JD; Ji Q; Thompson AM; Gleason IF; Christopher SA; Tsay S-C. (1998). Smoke,
clouds, and radiation-Brazil (SCAR-B) experiment. J Geophys Res, 103: 31783-31808. 089989
Kaufman YJ; Koren I. (2006). Smoke and pollution aerosol effect on cloud cover. Science, 313: 655-658. 190951
Kaufman YJ; Martins JV; Remer LA; Schoeberl MR; Yamasoe MA. (2002). Satellite retrieval of aerosol absorption over
the oceans using sunglint. Geophys Res Lett, 29: 34.1-34.4. 190955
Kaufman YJ; Nakajima T. (1993). Effect of Amazon smoke on cloud microphysics and albedo—Analysis from satellite
imagery. J Appl Meteorol, 32: 729-744. 190959
Kaufman YJ; Setzer A; Ward D; Tanre D; Holben BN; Menzel P; Pereira MC; Rasmussen R. (1992). Biomass Burning
Airborne and Spaceborne Experiment in the Amazonas (BASE-A). J Geophys Res, 97: 14581-14599. 044557
Kavouras IG; Etyemezian V; DuBois DW; Xu J; Pitchford M. (2009). Source reconciliation of atmospheric dust causing
visibility impairment in Class I areas of the western United States. J Geophys Res, 114: D02308. 191976
Kavouras IG; Etyemezian V; Xu J; DuBois DW; Green M; Pitchford M. (2007). Assessment of the local windblown
component of dust in the western United States. J Geophys Res, 112: D08211. 156630
Kenski DM; Gay D; Fitzsimmons S. (2004). Ammonia and its role in midwestern haze. Presented at Regional and Global
Perspectives on Haze: Causes, Consequences and Controversies Visibility Specialty Conference, Asheville, NC.
192078
Kerr R. (2007). Another global warming icon comes under attack. Science, 317: 28-29. 190950
Kiehl JT. (2007). Twentieth century climate model response and climate sensitivity. Geophys Res Lett, 34: 1-4. 190949
Kiikkila O. (2003). Heavy-metal pollution and remediation of forest soil around the Harjavalta Cu-Ni smelter, in SW
Finland. Silva Fennica, 37: 399-415. 156637
Kim CG; Bell JNB; Power SA. (2003). Effects of soil cadmium on Pinus sylvestris L. seedlings. Plant Soil, 257: 443-449.
155899
Kim MK; Lau KM; Chin M; Kim KM; Sud Y; Walker GK. (2006). Atmospheric teleconnection over Eurasia induced by
aerosol radiative forcing during boreal spring. PNAS, 19: 4700-4718. 190917
Kim SW; Yoon SC; Kim J. (2008). Columnar Asian dust particle properties observed by sun/sky radiometers from 2000 to
2006 in Korea. Atmos Environ, 42: 492-504. 130785
Kim Y; Hatsushika H; Muskett RR; Yamazaki K. (2005). Possible effect of boreal wildfire soot on Arctic sea ice and
Alaska glaciers. Atmos Environ, 39: 3513-3520. 155900
King M; Kaufman Y; Tanre D;Nakajima T. (1999). Remote sensing of tropospheric aerosols from space: Past, present, and
future. Bull Am Meteorol Soc, 80: 2229-2259. 190635
King MD; Platnick S; Moeller CC; Revercomb HE; Chu DA. (2003). Remote sensing of smoke, land, and clouds from the
NASAER-2 during SAFARI 2000. J Geophys Res, 108: SAF38.1-SAF38.12. 094395
Kinne S; Schulz M; Textor C; Guibert S; Balkanski Y; Bauer SE; Berntsen T; Berglen TF; Boucher O; Chin M; Collins W;
Dentener F; Diehl T; Easter R; Feichter J; Fillmore D; Ghan S; Ginoux P; Gong S; Grini A; Hendricks J; Herzog M;
Horowitz L; Isaksen I; Iversen T; Kirkevag A; Kloster S; Koch D; Kristjansson JE; Krol M; Lauer A; Lamarque JF;
Lesins G; Liu X; Lohmann U; Montanaro V; Myhre G; Penner J; Pitari G; Reddy S; Seland O; Stier P; Takemura T;
Tie X. (2006). An AeroCom initial assessment—optical properties in aerosol component modules of global models.
Atmos Chem Phys, 6: 1815-1834. 155903
Kirchstetter TW; Harley RA; Kreisberg NM; Stolzenberg MR; Hering SV. (1999). On-road measurement of fine particle
and nitrogen oxide emissions from light- and heavy-duty motor vehicles. Atmos Environ, 33: 2955-2968. 010642
Kirkbride MP; Dugmore JA. (2003). Glaciological response to distal tephra fallout from the 1947 eruption of Hekla, south
Iceland. J Glaciol, 49: 420-428. 156645
Kirkevag A; Iversen T. (2002). Global direct radiative forcing byprocess-parametrized aerosol optical properties. J Geophys
Res, 107: 4433. 199893
Kleeman MJ; Hughes LS; Allen JO; Cass GR. (1999). Source contributions to the size and composition distribution of
atmospheric particles: southern California in September 1996. Environ Sci Technol, 33: 4331-4341. 011286
December 2009 9-216
-------
Kleidman R; O'Neill N; Remer L; Kaufman Y; Eck T; Tame D; Dubovik O; Holben B. (2005). Comparison of Moderate
Resolution Imaging Spectroradiometer (MODIS) and Aerosol Robotic Network (AERONET) remote-sensing
retrievals of aerosol fine mode fraction over ocean. J Geophys Res, 110: 1-16. 190175
Kleinman MT; Araujo JA; Nel A; Sioutas C; Campbell A; Cong PQ; Li H; Bondy SC. (2008). Inhaled ultrafine particulate
matter affects CNS inflammatory processes and may act via MAP kinase signaling pathways. Toxicol Lett, 178:
127-130. 190074
Knutti R; Stacker TF; Joos F; Plattner GK. (2002). Constraints on radiative forcing and future climate change from
observations and climate model ensembles. Nature, 416: 719-723. 190178
Knutti R; Stacker TF; Joos F; Plattner GK. (2003). Probabilistic climate change projections using neural networks. Clim
Dynam, 21: 257-272. 190180
Koch D. (2001). Transport and direct radiative forcing of carbonaceous and sulfate aerosols in the GISS GCM. J Geophys
Res, 106: 20311-20332. 192054
Koch D; Bond TC; Streets D; Unger N; van der Werf GR. (2007). Global impact of aerosols from particular source regions
and sectors. J Geophys Res, 112: D02205. 190185
Koch D; Hansen J. (2005). Distant origins of Arctic black carbon: A Goddard Institute for Space Studies ModelE
experiment. J Geophys Res, 110: D04204. 190183
Koch D; Schmidt G; Field C. (2006). Sulfur, sea salt, and radionuclide aerosols in GISS ModelE. J Geophys Res, 111:
D06206. 190184
Kogan YL; Lilly DK; Kogan ZN; Filyushkin V. (1994). The effect of CCN regeneration on the evolution of stratocumulus
cloud layers. Atmos Res, 33: 137-150. 190186
Koren I; Kaufman YJ; Remer LA; Martins JV. (2004). Measurement of the effect of Amazon smoke on inhibition of cloud
formation. Science, 303: 1342-1345. 190187
Koren I; Kaufman YJ; Rosenfeld D; Remer LA; Rudich Y. (2005). Aerosol invigoration and restructuring of Atlantic
convective clouds. Geophys Res Lett, 32: L14828. 190188
Koren I; Martins JV; Remer LA; Afargan H. (2008). Smoke invigoration versus inhibition of clouds over the Amazon.
Science, 321: 946-949. 190193
Koren I; Remer KA; Kaufman YJ; Rudich Y; Martins JV. (2007). On the twilight zone between clouds and aerosols.
Geophys Res Lett, 34: L08805. 190192
Koren I; Remer LA; Longo K. (2007). Reversal of trend of biomass burning in the Amazon. Geophys Res Lett, 34:
L20404. 190189
Korsnes R; Pavlova O; Godtliebsen F. (2002). Assessment of potential transport of pollutants into the Barents Sea via sea
ice—an observational approach. Mar Pollut Bull, 44: 861-869. 156657
Kristjansson JE; Staple A; Kristiansen J; Kaas E. (2002). A new look at possible connections between solar activity, clouds
and climate. Geophys Res Lett, 29: 2107. 199916
Kroll JH; Ng NL; Murphy SM; Flagan RC; Seinfeld JH. (2006). Secondary organic aerosol formation from isoprene
photooxidation. Environ Sci Technol, 40: 1869-1877. 190195
Krueger AJ; Schaefer SJ; Krotkov N; Bluth G; Barker S. (2000). Ultraviolet remote sensing of volcanic emissions. In
Remote Sensing of Active Volcanism (pp. 25-43). Washington, DC: American Geophysical Union. 192234
Kruger O; Grasl H. (2002). The indirect aerosol effect over Europe. Geophys Res Lett, 29: 1925. 190200
Krupa S; Booker F; Bowersox V; Lehmann CT; Grantz D. (2008). Uncertainties in the current knowledge of some
atmospheric trace gases associated with U.S. agriculture: A review. J Air Waste Manag Assoc, 58: 986-993. 198696
Kruysse A. (1971). Acute inhalation toxicity of acrolein in hamsters. Central Institute for Nutrition and Food Research. The
Netherlands. R 3516. 192236
Kucera T; Horakova H; Sonska A. (2008). Toxic metal ions in photoautotrophic organisms. Photosynthetica, 46: 481-489.
190408
Kucharski J; Wyszkowska J. (2004). Inter-relationship between number of microorganisms and spring barley yield and
degree of soil contamination with copper. Plant Soil Environ, 50: 243-249. 156662
December 2009 9-217
-------
Kuki KN; Oliva MA; Costa AC. (2009). The simulated effects of iron dust and acidity during the early stages of
establishment of two coastal plant species. Water Air Soil Pollut, 196: 287-295. 190411
Kuki KN; Oliva MA; Pereira EG; Costa AC; Cambraia J. (2008). Effects of simulated deposition of acid mist and iron ore
particulate matter on photosynthesis and the generation of oxidative stress in Schinus terebinthifolius Radii and
Sophora tomentosa L. Sci Total Environ, 403: 207-214. 155346
LaBrecque JJ; Benzo Z; Alfonso JA; Cordoves Manuelita Quintal PR; Gomez CV; Marcano E. (2004). The concentrations
of selected trace elements in clams, Trivela mactroidea along the Venezuelan coast in the state of Miranda. Mar
Pollut Bull, 49: 659-667. 155913
Lack DA; Lovejoy ER; Baynard T; Pettersson A; Ravishankara AR. (2006). Aerosol Absorption Measurement using
Photoacoustic Spectroscopy: Sensitivity, Calibration, and Uncertainty Developments. Aerosol Sci Technol, 40: 697
- 708. 096032
Lakzian A; Murphy P; Turner A; Beynon JL; Giller KE. (2002). Rhizobium leguminosarum bv. viciae populations in soils
with increasing heavy metal contamination: abundance, plasmid profiles, diversity and metal tolerance. Soil Biol
Biochem, 34: 519-529. 156671
Lambert A; Grainger RG; Remedios JJ; Rodgers CD; Corney M; Taylor FW. (1993). Measurements of the evolution of the
Mt. Pinatubo aerosol cloud by ISAMS. Geophys Res Lett, 20: 1287-1290. 192231
Lambert MRK. (1997). Environmental effects of heavy spillage from a destroyed pesticide store near Hargeisa
(Somaliland) assessed during the dry season, using reptiles and amphibians as bioindicators. Arch Environ Contam
Toxicol, 32: 80-93. 155916
Landers DH; Simonich SL; Jaffe DA; Geiser LH; Campbell DH; Schwindt AR; Schreck CB; Kent ML; Hafner WD; Taylor
HE; Hageman KJ; Usenko S; Ackerman LK; Schrlau JE; Rose NL; Blett TF; Erway MM. (2008). The Fate,
Transport and Ecological Impacts of Airborne Contaminants in Western National Parks (USA). U.S.
Environmental Protection Agency, Office of Research and Development, NHEERL, Western Ecology Division.
Corvallis, Oregon. 191181
Lanno R; Wells J; Conder J; Bradham K; Basta N. (2004). The bioavailability of chemicals in soil for earthworms.
Ecotoxicol Environ Saf, 57: 39-47. 190415
Larson VE; Golaz J-C; Jiang H; Cotton WR. (2005). Supplying local microphysics parameterizations with information
about subgrid variability: Latin hypercube sampling. JAtmos Sci, 62: 4010-4026. 190220
Larson VE; Wood R; Field PR; Golaz J-C. (2001). Small-scale and mesoscale variability of scalars in cloudy boundary
layers: One-dimensional probability density functions. JAtmos Sci, 58: 1978-1996. 190212
Latimer DA; Ireson RG. (1980). Workbook for estimating visibility impairment. U. S. Enviornmental Protection Agency.
Research Triangle Park, NC. EPA-450/4-80-031. 035723
Lau K-M; Kim K-M. (2006). Observational relationships between aerosol and Asian monsoon rainfall, and circulation.
Geophys Res Lett, 33: L21810. 190226
Lau K; Kim M; Kim K. (2006). Asian summer monsoon anomalies induced by aerosol direct forcing—the role of the
Tibetan Plateau. Clim Dynam, 36: 855-864. 190223
Lau KM; Kim KM; Sud YC; Walker GK. (2009). A GCM study of the response of the atmospheric water cycle ofWest
Africa and the Atlantic to Saharan dust radiative forcing. Ann Geophys, 27: 4023-4037. 190229
Law RJ; Alaee M; Allchin CR; Boon JP; Lebeuf M; Lepom P; Ster GA. (2003). Levels and trends of polybrominated
diphenylethers and other brominated flame retardants in wildlife. Environ Int, 29: 757-770. 190420
Lawson ST; Scherbatskoy TD; Malcolm EG; Keeler GJ. (2003). Cloud water and throughfall deposition of mercury and
trace elements in a high elevation spruce-fir forest at Mt Mansfield, Vermont. J Environ Monit, 5: 578-583. 089371
Leahy L; Anderson T; Eck T; Bergstrom R. (2007). A synthesis of single scattering albedo of biomass burning aerosol over
southern Africa during SAFARI 2000. Geophys Res Lett, 34: L12814. 190232
Leaitch WR; Banic CM; Isaac GA; Couture MD; Liu PSK; Gultepe I; Li S-M; Kleinman L; Daum PH; MacPherson JI.
(1996). Physical and chemical observations in marine stratus during the 1993 North Atlantic Regional Experiment:
Factors controlling cloud droplet number concentrations. J Geophys Res, 101: 29123-29135. 190354
December 2009 9-218
-------
Leaitch WR; Isaac GA; Strapp JW; Banic CM; Wiebe HA. (1992). The relationship between cloud droplet number
concentrations and anthropogenic pollution: observations and climatic implications. J Geophys Res, 97: 2463-2474.
045270
LeBlanc GA. (1995). Trophic-level differences in the bioconcentration of chemicals: Implications in assessing
environmental biomagnification. Environ Sci Technol, 29: 154-160. 155921
Lee JK. (2006). Toxicity and tissue distribution of magnetic nanoparticles in mice [erratum]. Toxicol Sci, 90: 267. 088968
Lee T; Yu XY; Ayres B; Kreidenweis SM; Malm WC; Collett JL. (2008). Observations of fine and coarse particle nitrate at
several rural locations in the United States. Atmos Environ, 42: 2720-2732. 156686
Lee TE; Miller SD; Turk FJ; Schueler C; Julian R; Deyo S; Dills P; Wang S. (2006). The NPOESS VIIRS day/night visible
sensor. Bull Am Meteorol Soc, 87: 191-199. 190358
Lei YD; Wania F. (2004). Is rain or snow a more efficient scavenger of organic chemicals?. Atmos Environ, 38: 3557-3571.
127880
Lelieveld J; Berresheim H; Borrmann S; Crutzen PJ; Dentener FJ; Fischer H; Feichter J; Flatau PJ; Heland J; Holzinger R;
Korrmann R; Lawrence MG; Levin Z; Markowicz KM; Mihalopoulos N; Minikin A; Ramanathan V; De Reus M;
Roelofs GJ; Scheeren HA; Sciare J; Schlager H; Schultz M; Siegmund P; Steil B; Stephanou EG; Stier P; Traub M;
Warneke C; Williams J; Ziereis H. (2002). Global air pollution crossroads over the Mediterranean . Science, 298:
794-799. 190361
Leon J; Tanre D; Pelon J; Kaufman Y; Haywood J; Chatenet B. (2003). Profiling of a Saharan dust outbreak based on a
synergy between active and passive remote sensing. J Geophys Res, 108: 8575. 190366
Levin Z; Cotton WR. (2008). Report from the WMO/ IUGG International Aerosol Precipitation Science, Assessment
Group (IAPSAG). In Aerosol pollution impact on precipitation: A scientific review (pp. 482). Geneva, Switzerland:
World Meteorological Organization. 190375
Levine JS; Pinto JP. (1998). The production of CO by biomass burning. Presented at Proceedings of the International
Conference on Atmospheric Carbon Monoxide and Its Environmental Effects, Washington, DC. 029599
Levy R; Remer L; Dubovik O. (2007). Global aerosol optical properties and application to MODIS aerosol retrieval over
land. J Geophys Res, 112: D13210. 190377
Levy R; Remer L; Mattoo S; Vermote E; Kaufman Y. (2007). Second-generation algorithm for retrieving aerosol properties
over land from MODIS spectral reflectance. J Geophys Res, 112: D13211. 190379
Lewis ER; Schwartz SE. (2004). Sea salt aerosol production: Mechanisms, methods, measurements and models: A critical
review. Washington, DC: American Geophysical Union. 192023
Li R-R; Kaufman YJ; Hao WM; Salmon JM; Gao B-C . (2004). A technique for detecting burn scars using MODIS data.
IEEE Trans Geosci Remote Sens, 42: 1300-1308. 190386
Li Y-F; MacDonald RW; Jantunen LMM; Harner T; Bidleman TF; Stachen WMJ. (2002). The transport of P-
hexachlorocyclohexane to the western Arctic Ocean: a contrast to aHCH. Sci Total Environ, 291: 229-246. 156691
Li Z; Chen H; Cribb M; Dickerson R; Holben B; Li C; Lu D; Luo Y; Maring H; Shi G; Tsay S-C; Wang P; Wang Y; Xia X;
Zheng Y; Yuan T; Zhao F. (2007). Preface to special section on East Asian studies of tropospheric aerosols: An
international regional experiment (EAST-AIRE). J Geophys Res, 112: D22SOO. 190392
Liao H; Seinfeld JH. (2005). Global impacts of gas-phase chemistryaerosolinteractions on direct radiative forcing by
anthropogenicaerosols and ozone. J Geophys Res, 110: D18208. 199892
Lin CJ; Pehkonen SO. (1999). The chemistry of atmospheric mercury: a review. Atmos Environ, 33: 2067-2070. 190426
Lin H; Tao S; Zuo Q; Coveney RM. (2007). Uptake of polycyclic aromatic hydrocarbons by maize plants. Environ Pollut,
148:614-619. 155933
Lindberg SE; Brooks S; Linn CJ; Scott KJ; Landis MS; Stevens RK; Goodsite M; Richter A. (2002). Dynamic oxidation of
gaseous mercury in the Arctic troposphere at polar sunrise. Environ Sci Technol, 36: 1245-1256. 190429
Lindesay JA; Andreae MO; Goldammer JG; Harris G; Annegarn HJ; Garstang M; Scholes RJ; van Wilgen BW (1996).
International Geosphere Biosphere Programme/International Global Atmospheric Chemistry SAFARI-92 field
experiment: Background and overview. J Geophys Res, 23521-23530: 101. 190403
December 2009 9-219
-------
Lingua G; Franchin C; Todeschini V; Castiglione S; SBiondi; Burlando B; Parravicini V; PTorrigiani; GBerta. (2008).
Arbuscular mycorrhizal fungi differentially affect the response to high zinc concentrations of two registered poplar
clones. EnvironPollut, 153: 137-147. 155935
Liou KN; Ou S-C. (1989). The role of cloud microphysical processes in climate: an assessment from a one-dimensional
perspective. J Geophys Res, 94: 8599-8607. 190407
Liousse C; Penner JE; Chuang C; Walton JJ; Eddleman H; Cachier H. (1996). A global three-dimensional model study of
carbonaceous aerosols. J Geophys Res, 101: 19,411-19,432. 078158
Liu H; Pinker R; Holben B. (2005). A global view of aerosols from merged transport models, satellite, and ground
observations. J Geophys Res, 110: D10S15. 190414
Liu J; Ballaney M; Al-Alem U; Quan C; Jin X; Perera F; Chen LC; Miller RL. (2008). Combined Inhaled Diesel Exhaust
Particles and Allergen Exposure Alter Methylation of T Helper Genes and IgE Production In Vivo. Toxicol Sci, 102:
76-81. 156709
Liu L; Lacis AA; Carlson BE; Mishchenko MI; Cairns B. (2006). Assessing Goddard Institute for Space Studies ModelE
aerosol climatology using satellite and ground-based measurements. J Geophys Res, 111: D20212. 190422
Liu X; Penner J; Das B; Bergmann D; Rodriguez J; Strahan S; Wang M; Feng Y. (2007). Uncertainties in global aerosol
simulations: Assessment using three meteorological data sets. J Geophys Res, 112: D11212. 190427
Liu Y; Koutrakis P. (2007). Estimating Fine Particulate Matter Component Concentrations and Size Distributions Using
Satellite- Retrieved Fractional Aerosol Optical Depth: Part 1—Method Development. J Air Waste Manag Assoc, 57:
1351-1359. 187007
Liu Y; Koutrakis P; Kahn R; Turquety S; Yantosca RM. (2007). Estimating fine particulate matter component
concentrations and size distributions using satellite-retrieved fractional aerosol optical depth: part 2—a case study. J
Air Waste Manag Assoc, 57: 1360-9. 098197
Loeb N; Kato S. (2002). Top-of-atmosphere direct radiative effect of aerosols over the tropical oceans from the Clouds and
the Earth's Radiant Energy System (CERES) satellite instrument. PNAS, 15: 1474-1484. 190432
Loeb N; Manalo-Smith N. (2005). Top-of-Atmosphere direct radiative effect of aerosols over global oceans from merged
CERES and MODIS observations. J Clim, 18: 3506-3526. 190433
Loeb NG; Kato S; Loukachine K; Smith NM. (2005). Angular distribution models for top-of-atmosphere radiative flux
estimation from the Clouds and the Earth's Radiant Energy System instrument on the Terra Satellite, part I:
Methodology. J Atmos Ocean Tech, 22: 338-351. 190436
Lohman U; Feichter J. (2005). Global indirect aerosol effects: a review. Atmos Chem Phys, 5: 715-737. 155942
Lohmann U; Feichter J; Chuang CC; Penner JE. (1999). Prediction of the number of cloud droplets in the ECHAM GCM. J
Geophys Res, 104: 9169-9198. 190443
Lohmann U; Feichter J; Penner JE; Leaitch WR. (2000). Indirect effect of sulfate and carbonaceous aerosols: A mechanistic
treatment. J Geophys Res, 105: 12193-12206. 199910
Lohmann U; Koren I; Kaufman YJ. (2006). Disentangling the role of microphysical and dynamical effects in determining
cloud properties over the Atlantic. Geophys Res Lett, 33: L09802. 190451
Lohmann U; Leaitch WR; Barrie L; Law K; Yi Y; Bergmann D; Bridgeman C; Chin M; Christensen J; Easter R; Feichter J;
Jeuken A; Kjellstrom E; Koch D; Land C; Rasch P; Roelofs GJ. (2001). Vertical distributions of sulfur species
simulated by large scale atmospheric models in COSAM: comparison with observations. Tellus B Chem Phys
Meteorol, 53: 646 - 672. 190448
Lopez Alonso M; Benedito JL; Miranda M; Castillo C; Hernandez J; Shore RF. (2002). Cattle as biomonitors of
environmental semi-metal and trace metal concentrations in Galicia (NW Spain). Arch Environ Contam Toxicol,
43: 103-108. 155943
Lopez Alonso M; Benedito JL; Miranda M; Castillo C; Hernandez J; Shore RF. (2003). Mercury concentrations in cattle
from NW Spain. Sci Total Environ, 302: 93-100 . 155945
Lopez Alonso M; Benedito JL; Miranda M; Fernandez JA; Castillo C; Hernandez J; Shore RF. (2003). Large-scale spatial
variation inmercury concentrations in cattle inNW Spain. Environ Pollut, 125: 173-181. 155944
December 2009 9-220
-------
Lorusso S; Marabelli M; Troili M. (1997). Air pollution and the deterioration of historic monuments. J Environ Pathol
Toxicol Oncol, 16: 171-173. 084534
Lowenthal DH; Kumar N. (2003). PM2.5 mass and light extinction reconstruction in IMPROVE. J Air Waste Manag
Assoc, 53: 1109-1120. 156712
Lowenthal DH; Rogers CF; Saxena P; Watson JG; Chow JC. (1995). Sensitivity of estimated light extinction coefficients to
model assumptions and measurement errors. Atmos Environ, 29: 751-766. 045134
Lu M-L; Feingold G; Jonsson H; Chuang P; Gates H; Flagan RC; Seinfeld JH. (2008). Aerosol-cloud relationships in
continental shallow cumulus. J Geophys Res, 113: D15201. 190455
Lubin D; Vogelmann AM. (2006). A climatologically significant aerosol longwave indirect effect in the Arctic. Nature, 439:
453-456. 190466
Lubin DS; Satheesh S; McFarquar G; Heymsfield A. (2002). Longwave radiative forcing of Indian Ocean tropospheric
aerosol. J Geophys Res, 107: 8004. 190463
Lucaciu A; Timofte L; Culicov O; Frontasyeva MV; Oprea C; Cucu-Man S; Mocanu R; Steinnes E. (2004). Atmospheric
deposition of trace elements in Romania studied by the moss biomonitoring technique. J Atmos Chem, 49: 533-548.
155947
Lunackova L; Masarovicova E; Kraova K; Stresko V. (2003). Response of fast growing woody plants from family
Salicaceae to cadmium treatment. Bull Environ Contain Toxicol, 70: 576-585. 155948
Luo Y; LuD; ZhouX; Li W; He Q. (2001). Characteristics of the spatial distribution and yearly variation of aerosol optical
depth over China in last 30 years. J Geophys Res, 106: 14501. 190467
Mace BL; Bell PA; Loomis RJ. (2004). Visibility and Natural Quiet in National Parks and Wilderness Areas: Psychological
Considerations. Environ Behav, 36: 5-31. 180255
Mackay D. (1991). Multimedia environmental models: the fugacity approach. Chelsea, MI: Lewis Publishers. 042941
Macleod M; McKone TE; Mackay D. (2005). Mass balance for mercury in the San Francisco Bay area. Environ Sci
Technol, 39: 6721-6729. 155954
Madrid L; Diaz-Barrientos E; Madrid F. (2002). Distribution of heavy metal contents of urban soils in parks of Seville.
Chemosphere, 49: 1301-1309. 155956
Magi B; Hobbs P; Kirchstetter T; Novakov T; Hegg D; Gao S; Redemann J; Schmid B. (2005). Aerosol properties and
chemical apportionment of aerosol optical depth at locations off the United States East Coast in July and August
2001. JAtmos Sci, 62: 919-933. 190468
Malkonen E; Derome J; Fritze H; Helmisaari H-S; Kukkola M; Kyto M; Saarsalmi A; Salemaa M. (1999). Compensatory
fertilization of Scots pine stands polluted by heavy metals. Nutr Cycling Agroecosyst, 55: 239-268. 155961
Malm W; Bell P; McGlothin GE. (1984). Field testing a methodology for assessing the importance of good visual air
quality. Presented at Air Pollution Control Association, June, 1984, San Francisco, CA. 044292
Malm WC. (1999). Introduction to visibility. Cooperative Institute for Research in the Atmosphere. Fort Collins, CO.
025037
Malm WC; Day DE. (2001). Estimates of aerosol species scattering characteristics as a function of relative humidity.
Atmos Environ, 35: 2845-2860. 190431
Malm WC; Day DE; Kreidenweis SM; Collett JL; Lee T. (2003). Humidity dependent optical properties of fine particle
during the Big Bend Regional Aerosol and Visibility Observational Study (BRAVO). J Geophys Res, 108: 4279.
190434
Malm WC; Hand JL. (2007). An examination of the physical and optical properties of aerosols collected in the IMPROVE
program. Atmos Environ, 41: 3407-3427. 155962
Malm WC; Pitchford ML. (1997). Comparison of calculated sulfate scattering efficiencies as estimated from size-resolved
particle measurements at three national locations. Atmos Environ, 31: 1315-1325. 002519
Malm WC; Pitchford ML; McDade C; Ashbaugh LL. (2007). Coarse particle speciation at selected locations in the rural
continental United States. Atmos Environ, 41: 2225-2239. 156730
December 2009 9-221
-------
Malm WC; Schichtel BA; Ames RB; Gebhart KA. (2002). A ten-year spatial and temporal trend of sulfate across the
United States. J Geophys Res, 107: 4627. 156727
Malm WC; Schichtel BA; Pitchford ML; Ashbaugh LL; Eldred RA. (2004). Spatial and monthly trends in speciated fine
particle concentration in the United States. J Geophys Res, 109: 3306. 156728
Malm WC; Trijonis J; Sisler J; Pitchford M; Dennis RL. (1994). Assessing the effect of SO2 emission changes on visibility.
Atmos Environ, 28: 1023-1034. 044920
Marinoni N; Birelli MP; Rostagno C; Pavese A. (2003). The effects of atmospheric multipollutants on modern concrete.
Atmos Environ, 37: 4701-4712. 092520
Markiewicz Patkowska J; Hursthouse A; Przybyla-Kij H. (2005). The interaction of heavy metals with urban soils: sorption
behaviour of Cd, Cu, Cr, Pb and Zn with a typical mixed brownfield deposit. Environ Int, 31: 513-521. 155963
Martin MH; Coughtrey PJ. (1981). Impact of metals on ecosystem function and productivity. InLepp, N. W. (Ed.),Effect of
heavy metal pollution on plants: volume 2, metals in the environment (pp. 119-158). Barking, United Kingdom:
Applied Science Publishers. 047727
Martins J; Tanre D; Remer L; Kaufman Y; Mattoo S; Levy R. (2002). MODIS cloud screening for remote sensing of
aerosol over oceans using spatial variability. Geophys Res Lett, 29: MOD4. 190470
Martonchik J; Diner D; Crean K; Bull M. (2002). Regional aerosol retrieval results from MISR. IEEE Trans Geosci
Remote Sens, 40: 1520-1531. 190490
Martonchik J; Diner D; Pinty B; Verstraete M; Myneni R; Knjazikhin Y; Gordon H. (1998). Determination of land and
ocean reflective, radiative, and biophysical properties using multiangle imaging. IEEE Trans Geosci Remote Sens,
36: 1266-1281. 190484
Martonchik JD; Diner D; Kahn R; Verstraete M; Pinty B; Gordon H; Ackerman T. (1998). Techniques for the Retrieval of
aerosol properties over land and ocean using multiangle data. IEEE Trans Geosci Remote Sens, 36: 1212-1227.
190472
Masclet PP; Hoyau V; Jaffrezo JL; Cachier H. (2000). Poly cyclic aromatic hydrocarbons deposition on the ice sheet of
Greenland. Part I. Superficial snow. Atmos Environ, 34: 3195-3207. 155966
Massicotte R; Robidoux PY; Sauve S; Flipo D; Fournier M; Trottier B. (2003). Immune response of earthworms
(Lumbricus terrestris, Eisenia andrei and Aporrectodea tuberculata) following in situ soil exposure to atmospheric
deposition from a cement factory. J Environ Monit, 5: 774-779. 155968
Massie S; Torres O; Smith S. (2004). Total ozone mapping spectrometer (TOMS) observations of increases in Asian
aerosol in winter from 1979 to 2000. J Geophys Res, 109:D18211. 190492
Matheson MA; Coakley JA; Tahnk WR. (2005). Aerosol and cloud property relationships for summertime stratiform clouds
in the northeastern Atlantic from Advanced Very High Resolution Radiometer observations. J Geophys Res, 110:
D24204. 190494
Matsui T; Masunaga H; Kreidenweis SM; Pielke RA; Tao W-K; Chin M; Kaufman YJ. (2006). Satellite-based assessment
of marine low cloud variability associated with aerosol, atmospheric stability, and the diurnal cycle. J Geophys Res,
111 :D17204. 190498
Matsui T; Pielke R;. (2006). Measurement-based estimation of the spatial gradient of aerosol radiative forcing. Geophys
Res Lett, 33: L11813. 190495
Matthias I; Ansmann A; Milller D; Wandinger U; Althausen D. (2004). Multiyear aerosol observations with dual
wavelength Raman lidar in the framework of EARLINET. J Geophys Res, 109: 1-15. 155971
Mayo JM; Legge AH; Yeung EC; Krupa SV; Bogner JC. (1992). The effects of sulphur gas and elemental sulphur dust
deposition on Pinus contorta x Pinus banksiana: Cell walls and water relations. Environ Pollut, 76: 43-50. 155974
McComiskey A; Feingold G (2008). Quantifying error in the radiative forcing of the first aerosol indirect effect. Geophys
Res Lett, 35: L20810. 190517
McComiskey A; Feingold G; Frisch AS; Turner D; Miller M; Chiu JC; Min Q; Ogren JA. (2008). An assessment of
aerosol- cloud interactions in marine stratus clouds based on surface remote sensing. J Geophys Res, 114: D09203.
190525
December 2009 9-222
-------
McComiskey A; Schwartz SE; Schmid B; Guan H; Lewis ER; Ricchiazzi P; Ogren JA. (2008). Direct aerosol forcing:
Calculation from observables and sensitivity to inputs. J Geophys Res, 113: D09202. 190523
McCormick M; Trepte C. (1987). Polar stratospheric optical depth observed between 1978 and 1985. J Geophys Res, 92:
4297-4306. 192328
McCormick R; Ludwig J. (1967). Climate modification by atmospheric aerosols. Science, 9: 1358 - 1359. 190528
McDade C. (2004). Summary of IMPROVE nitrate measurements. Retrieved 29-JUL-09, from
http://vista.cira.colostate.edu/improve/publications/GrayLit/gray _literature.htm. 192075
Mcfiggans G; Artaxo P; Baltensperger U; Coe H; Facchini MC; Feingold G; Fuzzi S; Gysel M; Laaksonen A; Lohmann U;
Mentel TF; Murphy DM; O'Dowd CD; Snider JR; Weingartner E. (2006). The effect of physical and chemical
aerosol properties on warm cloud droplet activation. Atmos Chem Phys, 6: 2593-2649. 190532
McGowan TF; Lipinski GE; Santoleri JJ. (1993). New rules affect the handling of waste fuels. Chem Eng, 100: 122-128.
046731
McMurry PH. (2000). A review of atmospheric aerosol measurements. Atmos Environ, 34: 1959-1999. 081517
Meehl GA; Washington WM; Ammann CM; Arblaster JM; Wigley TML; Tebaldi C. (2004). Combinations of natural and
anthropogenic forcings in twentieth-century climate. J Clim, 17: 3721-3727. 192279
Meers E; Vangronsveld J; Tack FMG; Vandecasteele B; Ruttens A. (2007). Potential of five willow species (Salix spp.) for
phytoextraction of heavy metals. Environ Exp Bot, 60: 57-68. 155977
Melnikov S; Carroll J; Gorshkov A; Vlasov S; Dahle S. (2003). Snow and ice concentrations of selected persistent organic
pollutants in the Ob-Yenisey River watershed. Sci Total Environ, 306: 27-37. 156753
Memon AR; Schroder P. (2009). Implications of metal accumulation mechanisms to phytoremediation. Environ Sci Pollut
Reslnt, 16: 162-175. 190442
Menon S; Del Genio AD; Kaufman Y; Bennartz R; Koch D; Loeb N; Orlikowski D. (2008). Analyzing signatures of
aerosol cloud interactions from satellite retrievals and the GISS GCM to constrain the aerosol indirect effect. J
Geophys Res, 113: D14S22. 190534
Menon S; Hansen J; Nazarenko L; Luo Y. (2002). Climate effects of black carbon aerosols in China and India. Science,
297: 2250-2253. 155978
Mercado LM; BellouinN; Sitch S; Boucher O; Huntingford C; Wild M; Cox PM. (2009). Impact of changes in diffuse
radiation on the global land carbon sink. Nature, 458: 1014-U1087. 190444
Michalsky J; Schlemmer J; Berkheiser W;. (2001). Multiyear measurements of aerosol optical depth in the Atmospheric
Radiation Measurement and Quantitative Links program. J Geophys Res, 106: 12099-12108. 190537
Middleton WEK. (1952). Vision through the atmosphere. Toronto, ON, Canada: University of Toronto Press. 016324
Mie G. (1908). Beitrage zur Optik trilber Medien, speziell kolloidaler Metallosungen. Ann Phys, 330: 377-445. 155983
Miller RL; Schmidt GA; Shindell DT. (2005). Forced annular variations in the 20th century Intergovernmental Panel on
Climate Change Fourth Assessment Report models. J Geophys Res, 111: D18101. 192258
Millhollen AG; Gustin MS; Obrist D. (2006). Foliar mercury accumulation and exchange for three tree species. Environ
Sci Technol, 40: 6001-6006. 190447
Min Q; Harrison LC. (1996). Cloud properties derived from surface MFRSR measurements and comparison with GEOS
results at the ARM SGP site. Geophys Res Lett, 23: 1641- 1644. 190538
Ming Y; Ramaswamy V; Ginoux PA; Larry W. Horowitz LW; Russell LM. (2005). Geophysical Fluid Dynamics
Laboratory generalcirculation model investigation of the indirect radiative effectsof anthropogenic sulfate aerosol. J
Geophys Res, 110: D22206. 199919
Minnis P; Harrison EF; Stowe LL; Gibson GG; Derm FM; Doelling DR; Smith WL Jr. (1993). Radiative Climate Forcing
by the Mount Pinatubo Eruption. Science, 259: 1411-1415. 190539
Mischenko M; Cairns B; Kopp G; Schueler CF; Fafaul BA; Hansen JE; Hooker RJ; Itchkawich T; Maring HB; Travis LD.
(2007). Accurate monitoring of terrestrial aerosols and total solar irradiance. Bull Am Meteorol Soc, 88: 677-691.
190543
December 2009 9-223
-------
Mishchenko MI; Geogdzhayev IV. (2007). Satellite remote sensing reveals regional tropospheric aerosol trends. Optics
Express, 15: 7423-7438. 190545
Mishchenko MI; Geogdzhayev IV; Cairns B; Rossow WB; Lacis AA. (1999). Aerosol retrievals over the ocean by use of
channels 1 and 2 AVHRR data: sensitivity analysis and preliminary results. Appl Opt, 38: 7325-7341. 190541
Mishchenko MI; Geogdzhayev IV; Rossow WB; Cairns B; Carlson BE; Lacis AA; Liu L; Travis LD. (2007). Long-term
satellite record reveals likely recent aerosol trend. Science, 315: 1543. 190542
Mitchell JM Jr. (1971). The effect of atmospheric aerosols on climate with special reference to temperature near the Earth's
surface. J Appl Meteorol, 10: 703-714. 190546
Molina LT; Madronich S; Gaffney JS; Singh HB. (2008). Overview of MILAGRO/INTEX-B Campaign. International
Global Atmospheric Chemistry (Newsletter). Seattle,
WA.http://www.igac.noaa.gov/newsletter/igac38/Apr_2008_IGAC_38.pdf. 192019
Moody E; King M; Platnick S; Schaaf C; Gao F. (2005). Spatially complete global spectral surface albedos: value-added
datasets derived from Terra MODIS land products. IEEE Trans Geosci Remote Sens, 43: 144-158. 190548
Moosmuller H; Arnott WP; Rogers CF. (1998). Photoacoustic and filter measurements related to aerosol light absorption
during the Northern Front Range Air Quality Study (Colorado 1996/1997). J Geophys Res, 103: 28,149-28,157.
020657
Moropoulou A; Bisbikou K; Torfs K; Van Grieken R; Zezza F; Macri F. (1998). Origin and growth of weathering crusts on
ancient marbles in industrial atmosphere. Atmos Environ, 32: 967-982. 040485
Moses CA; Smith BJ. (1994). Limestone weathering in the supratidal zone: an example from Mallorca. In DA Robinson;
RBG Williams (Ed.),Rock weathering and landform evolution (pp. 432-451). Chichester, West Sussex, United
Kingdom: John Wiley & Sons. 156785
Mosley-Thompson E; Mashiotta TA; Thompson LG (2003). High resolution ice core records of Late Holocene volcanism:
Current and future contributions from the Greenland PARCA cores. In Volcanism and the Earth's Atmosphere (pp.
153-164). Washington, DC: American Geophysical Union. 192255
Mouillot F; Narasimha A; Balkanski Y; Lamarque J-F. (2006). Global carbon emissions from biomass burning in the 20th
century. Geophys Res Lett, 33: L01801. 190549
Muir DCG; Omelchenko A; Grift NP; Savoie DA; Lockhart WL; Wilkinson P; Brunskill GJ. (1996). Spatial trends and
historical deposition of poly chlorinated biphenyls in Canadian midlatitude and arctic lake sediments. Environ Sci
Technol, 30: 3609-3617. 155991
Murayama T; Sugimoto N; Uno I; Kinoshita K; Aoki K; Hagiwara N; Liu Z; Matsui I; Sakai T; Shibata T; Arao K; Sohn B-
J; Won J-G; Yoon S-C; Li T; Zhou J; Hu H; Abo M; lokibe K; Koga R; Iwasaka Y. (2001). Ground-based network
observation of Asian dust events of April 1998 in East Asia. J Geophys Res, 106: 18345-18360. 155992
Murdoch JC; Thayer MA. (1988). Hedonic price estimation of variable urban air quality. J Environ Econ Manage, 15: 143-
146. 156788
Naidoo G; Chirkoot D. (2004). The effects of coal dust on photo synthetic performance of the mangrove, Avicennia marina
in Richards Bay, South Africa. Environ Pollut, 127: 359-366. 190449
Nakajima T; Higurashi A; Kawamoto K; Penner JE. (2001). A possible correlation between satellite-derived cloud and
aerosol microphysical parameters. Geophys Res Lett, 28: 1171-1174. 190552
Nakamura K; Nakawo M; Ageta Y; Goto-Azuma K; Kamiyam K. (2000). Post-depositional loss of nitrate in surface snow
layers of the Antarctic ice sheet. Bull. Glaciol. Res., 17: 11-16. 156792
Nash TH. (1975). Influence of effluents from a zinc factory on lichens. Ecol Monogr, 45: 183-198. 016763
Nazaroff WW; Cass GR. (1991). Protecting museum collections from soiling due to the deposition of airborne particles.
Atmos Environ, 25: 841-852. 044577
NCDENR. (2007). Technical Analyses Supporting Regional Haze State Implementation Plan. , North Carolina Department
of Environment and Natural Resources. NC. 156798
NESCAUM. (2006). Regional Aerosol Intensive Network (RAIN) Preliminary Data Analysis. Northeast States for
Coordinated Air Use Management. Boston, MA.http://www.nescaum.org/documents/2006-05-memo8-rain.pdf/.
156802
December 2009 9-224
-------
Nguyen HL; Leermakers M; Elskens M; De Ridder F; Doan TH; Baeyens W. (2005). Correlations, partitioning and
bioaccumulation of heavy metals between different compartments of Lake Balaton. Sci Total Environ, 341: 211-
226. 155997
Nicholson L; Benn DI. (2006). Calculating ice melt beneath a debris layer using meteorological data. J Glaciol, 52: 463-
470. 156806
Nieminen T; Derome J; Helmisaari H-S. (1999). Interactions between precipitation and Scots pine canopies along a heavy-
metal pollution gradient. Environ Pollut, 106: 129-137. 155998
Nogueira MA; Magalhaes GC; Cardoso EJBN. (2004). Manganese toxicity in mycorrhizal and phosphorus-fertilized
soybean plants. J Plant Nutr, 27: 141-156. 190460
Norris J; Wild M. (2007). Trends in aerosol radiative effects over Europe inferred from observed cloud cover, solar
"dimming", and solar "brightening". J Geophys Res, 112: D08214. 190555
Notten MJM; Oosthoek AJP; Rozema J; Aerts R. (2005). Heavy metal concentrations in a soil-plant-snail food chain along
a terrestrial soil pollution gradient. Environ Pollut, 138: 178-190. 190461
Novakov T; Ramanathan V; Hansen JE; Kirchstetter TW; Sato M; Sinton JE; Sathaye JA. (2003). Large historical changes
of fossil-fuel black carbon aerosols. Geophys Res Lett, 30: 1-4. 048398
NRC. (2001). Climate change science: an analysis of some key questions. National Research Council. National Academy
Press. Washington, DC. 053303
NRC. (2005). Radiative forcing of climate change: expanding the concept and addressing uncertainties. Washington DC:
National Research Council; The National Academies Press. 057409
O'Dowd C; McFiggans G; Creasey DJ; Pirjola L; Hoell C; Smith MH; Allan BJ; Plane JMC; Heard DE; Lee JD; Pilling
MJ; Kulmala M. (1999). On the photochemical production of new particles in the coastal boundary layer. Geophys
Res Lett, 26: 1707-1710. 090414
O'Neill NT; Eck TF; Smirnov A; Holben BN; Thulasiraman S. (2003). Spectral discrimination of coarse and fine mode
optical depth. J Geophys Res, 108: 4559. 180187
Odum EP (1985). Trends expected in stressed ecosystems. Bioscience, 35: 419-422. 039482
Odum EP. (1993). Major ecosystem types of the world. In Odum, E. P. (Ed.),Ecology and our endangered life-support
systemsSunderland, MA: Sinauer Associates, Inc. 076742
Oehme M; Biseth A; SchlabachM; Wiig O. (1995). Concentrations of poly chlorinated dibenzo-p-dioxins, dibenzofurans
and non-ortho substituted biphenyls in polar bear milk from Svalbard (Norway). Environ Pollut, 90: 401-407.
011267
Oehme M; Haugen J-E; Schlabach M. (1996). Seasonal changes and relation between levels of organochlorines in Arctic
ambient air: First results from an all-round-year monitoring program at Ny-Alesund, Norway. Environ Sci Technol,
30: 2294-2304. 156001
Ohyama K; Angermann J; Dunlap DY; Matsumura F. (2004). Distribution of poly chlorinated biphenyls and chlorinated
pesticide residues in trout in the Sierra Nevada. J Environ Qual, 33: 1752-1764. 190462
Oliveira A; Pampulha ME. (2006). Effects of long-term heavy metal contamination on soil microbial characteristics. J
Biosci Bioeng, 102: 157-161. 156827
Ormrod DP. (1984). Impact of trace element pollution on plants. In Treshow, M. (Ed.),Air pollution and plant life (pp. 291-
314). Chichester, United Kingdom: John Wiley & Sons Ltd. 046892
Otnyukova T. (2007). Epiphytic lichen growth abnormalities and element concentrations as early indicators of forest
decline. Environ Pollut, 146: 359-365. 156009
Otvos E; Pazmandi T; Tuba Z. (2003). First national survey of atmospheric heavy metal deposition in Hungary by the
analysis of mosses. Sci Total Environ, 309: 151-160. 156831
Ouimette JR; Flagan RC. (1982). The extinction coefficient of multicomponent aerosols. Atmos Environ, 16: 2405-2420.
025047
December 2009 9-225
-------
Ozdilek HG; Mathisen PP; Pellegrino D. (2007). Distribution of heavy metals in vegetation surrounding the Blackstone
River, USA: Considerations regarding sediment contamination and long term metals transport in freshwater riverine
ecosystems. J Environ Biol, 28: 493-502. 156010
Pacheco AMG; Freitas MC. (2004). Are lower epiphytes really that better than higher plants for indicating airborne
contaminants? An insight into the elemental contents of lichen thalli and tree bark by INAA. Journal of Radioanal
Chem, 259: 27-33. 156011
Padmavathiamma PK; Li LY. (2007). Phytoremediation technology: Hyper-accumulation metals in plants. Water Air Soil
Pollut, 184: 105-126. 190465
Palmer AS; Morgan VI; Curran MAJ; van Ommen TD; Mayewski PA. (2002). Antarctic volcanic flux ratios from Law
Dome ice cores. Ann Glaciol, 35: 329-332. 192319
Park RJ; Jacob DJ; Kumar N; Yantosca RM. (2006). Regional visibility statistics in the United States: Natural and
transboundary pollution influences, and implications for the regional haze rule. Atmos Environ, 40: 5405-5423.
190469
Parrish ZD; White JC; Isleyen M; Gent MPN; lannucci-Berger W; Eitzer BD; Kelsey JW; Martina MI. (2006).
Accumulation of weathered poly cyclic aromatic hydrocarbons (PAHs) by plant and earthworm species.
Chemosphere, 64: 609-618. 156014
Patadia F; Gupta P; Christopher SA. (2008). First observational estimates of global clear-sky shortwave aerosol direct
radiative effect over land. J Geophys Res, 35: L04810. 190558
Patra M; Bhowmik N; Bandopadhyay B; Sharma A. (2004). Comparison of mercury, lead and arsenic with respect to
genotoxic effects on plant systems and the development of genetic tolerance. Environ Exp Bot, 52: 199-223.
081976
Pennanen T; Frostegard A; Fritze H; Baath E. (1996). Phospholipid fatty acid composition and heavy metal tolerance of
soil microbial communities along two heavy metal-polluted gradients in coniferous forests. Appl Environ
Microbiol, 62: 420-428. 156016
Penner J; Quaas J; Storelvmo T; Takemura T; Boucher O; Guo H; Kirkevag A; Kristjansson JE; Seland O. (2006). Model
intercomparison of indirect aerosol effects. Atmos Chem Phys, 6: 3391-3405. 190564
Penner JE. (2003). Comment on "Control of fossil-fuel particulate black carbon and organic matter, possibly the most
effective method of slowing global warming. J Geophys Res, 108: 4771. 156851
Penner JE; Dickinson RE; O'Neill CA. (1992). Effects of aerosol from biomass burning on the global radiation budget.
Science, 256: 1432-1433. 045825
Penner JE; Eddleman H; Novakov T. (1993). Towards the development of a global inventory for black carbon emissions.
Atmos Environ, 27: 1277-1295. 045457
Penner JE; Zhang SY; Chin M; Chuang CC; Feichter J; Feng Y; Geogdzhayev IV; Ginoux P; Herzog M; Higurashi A; Koch
D; Land C; Lohmann U; Mishchenko M; Nakajima T; Pitari G; Soden B; Tegen I; Stowe L. (2002). A comparison
of model- and satellite derived aerosol optical depth and reflectivity. J Atmos Sci, 59: 441-460. 190562
Pennisi E. (2004). The secret life of fungi. Science, 304: 1620-1622. 156018
Perlwitz J; Graf HF. (2001). Trosposphere-stratosphere dynamic coupling under strong and weak polar vortex conditions .
Geophys Res Lett, 28: 271-274. 192271
Perlwitz J; Harnik N. (2003). Observational evidence of a stratospheric influence on the troposphere by planetary wave
reflection. J Clim, 16: 3011-3026. 192264
Peters AJ; Gregor DJ; Teixeira CF; Jones NP; Spencer C. (1995). The recent depositional trend of polycyclic aromatic
hydrocarbons and elemental carbon to the Agassiz Ice cap, Ellesmere Island, Canada. Sci Total Environ, 160/161:
167-179. 156856
Pfirman SL; Kogeler JW; Rigor I. (1997). Potential for rapid transport of contaminants from the Kara Sea. Sci Total
Environ, 202: 111-122. 156864
Pincus R; Klein SA. (2000). Unresolved spatial variability and microphysical process rates in large-scale models. J
Geophys Res, 105: 27059-27065. 190565
Pinker R; Zhang B; Dutton E. (2005). Do satellites detect trends in surface solar radiation?. Science, 308: 850-854. 190569
December 2009 9-226
-------
Piol MN; Lopez AG; Mino LA; Dos Santos Afonso M; Verrengia Guerrero NR. (2006). The impact of particle-bound
cadmium on bioavailability and bioaccumulation: A pragmatic approach. Environ Sci Technol, 40: 6341-6347.
156028
Pitchford M; Green M; Kuhns H; Tombach I; Malm W; Scruggs M; Farber R; Mirabella V; White WH; McDade C. (1999).
Project MOHAVE Final Report. 156873
Pitchford M; Maim W; Schichtel B; Kumar N; Lowenthal D; Hand J. (2007). Revised algorithm for estimating light
extinction from IMPROVE particle speciation data. J Air Waste Manag Assoc, 57: 1326-36. 098066
Pitchford ML; Malm WC. (1994). Development and applications of a standard visual index. Atmos Environ, 28: 1049-
1054. Q44922
Pitchford ML; Polkowsky BV; McGown MR; Malm WC; Molenar JV; Mauch L. (1990). Percent change in extinction
coefficient: a proposed approach for federal visibility protection strategy. In CV Mathai (Ed.),Transactions:
Visibility and Fine Particles (pp. 37-49). Pittsburgh, PA: Air and Waste Management Association. 156871
Pitchford ML; Schichtel BA; Gebhart KA; Barna MG; Malm WC; Tombach IH; Knipping EM. (2005). Reconciliation and
Interpretation of the Big Bend National Park Light Extinction Source Apportionment: Results from the Big Bend
Regional Aerosol and Visibility Observational Study- Part II. J Air Waste Manag Assoc, 55: 1726-1732. 156874
Prasad MNV; DeOliveira FHM. (2003). Metal hyperaccumulation in plants: Biodiversity prospecting for phytoremediation
technology. Electron J Biotechnol, 6: 110-146. 156885
Prata A; Rose W; Self S; O'Brien D. (2003). Global, long-term sulphur dioxide measurements from TOVS data: a new tool
for studying explosive volcanism and climate. In Volcanism and the Earth's Atmosphere (pp. 75-92). Washington,
DC: American Geophysical Union. 192235
Preston CM; Schmidt MWI. (2006). Black (pyrogenic) carbon: a synthesis of current knowledge and uncertainties with
special consideration of boreal regions. Biogeosciences, 3: 397-420. 156030
Procopio AS; Artaxo P; Kaufman YJ; Remer LA; Schafer JA; Holben BN. (2004). Multiyear analysis of Amazonian
biomass burning smoke radiative forcing of climate. J Geophys Res, 31: L03108. 190571
Prospero JM. (1996). Saharan dust transport over the North Atlantic Ocean and Mediterranean: an overview. In Guerzoni
S; Chester R (Ed.),The Impact of Desert Dust Across the Mediterranean (pp. 133-152). The Netherlands: Kluwer
Academic Publishers. 156889
Pry or SC. (1996). Assessing public perception of visibility for standard setting exercises. Atmos Environ, 30: 2705-2716.
056598
Pryor SC; Binkowski FS. (2004). An analysis of the time scales associated with aerosol processes during dry deposition.
Aerosol Sci Technol, 38: 1091-1098. 116805
Putaud J-P; Raes F; Van Dengenen R; Bruggemann E; Facchini M-C; Decesari S; Fuzzi S; Gehrig R; Huglin C; Laj P;
Lorbeer G; Maenhaut W; Mihalopoulos N; Muller K; Querol X; Rodriguez S; Schneider J; Spindler G; ten Brink H;
Torseth K; Wiedensohler A. (2004). A European aerosol phenomenology—2: chemical characteristics of particulate
matter at kerbside, urban, rural and background sites in Europe. Atmos Environ, 38: 2579-2595. 055545
Qian Y; Wang W; Leung L; Kaiser D. (2007). Variability of solar radiation under cloud-free skies in China: The role of
aerosols. Geophys Res Lett, 34: L12804. 190572
Quaas J; Boucher O. (2005). Constraining the first aerosol indirect radiative forcing in the LMDZ GCM using POLDER
and MODIS satellite data. Geophys Res Lett, 32: L17814. 190573
Quaas J; Boucher O; BellouinN; Kinne S. (2008). Satellite- based estimate of the direct and indirect aerosol climate
forcing. J Geophys Res, 113: D05204. 190916
Quaas J; Boucher O; Breon F-M. (2004). Aerosol indirect effects in POLDER satellite data and the Laboratoire de
Meteorologie Dynamique-Zoom (LMDZ) general circulation model. J Geophys Res, 109: D08205. 199907
Quaas J; Boucher O; Lohmann U. (2006). Constraining the total aerosol indirect effect in the LMDZ GCM and ECHAM4
GCMs using MODIS satellite data. Atmos Chem Phys Discuss, 5: 9669-9690. 190915
Quinn PK; Anderson T; Bates T; Dlugi R; Heintzenberg J; Von Hoyningen-Huene W; Kumula M; Russel P; Swietlicki E.
(1996). Closure in tropospheric aerosol-climate research : A review and future needs for addressing aerosol direct
shortwave radiative forcing. Contrib Atmos Phys, 69: 547-577. 192021
December 2009 9-227
-------
Quinn PK; Bates TS. (2003). North American, Asian, and Indian haze: Similar regional impacts on climate?. Geophys Res
Lett, 30: 1555.049189
Quinn PK; Bates TS; Baynard T; Clarke AD; Onasch TB; Wang W; Rood MJ; Andrews E; Allan J; Carrico CM; Coffman
D; Worsnop D. (2005). Impact of particulate organic matter on the relative humidity dependence of light scattering:
A simplified parameterization. Geophys Res Lett, 32: 1-4. 156033
Quinn PK; Coffman D; Kapustin V; Bates TS; Covert DS. (1998). Aerosol optical properties in the marine boundary layer
during ACE 1 and the underlying chemical and physical aerosol properties. J Geophys Res, 103: 16547-16563.
190918
Quinn PK; Coffman DJ; Bates TS; Welton EJ; Covert DS; Miller TL; Johnson JE; Maria S; Russell L; Arimoto R; Carrico
CM; Rood MJ; Anderson J. (2004). Aerosol optical properties measured aboard the Ronald H. Brown during ACE-
Asia as a function of aerosol chemical composition and source region. J Geophys Res, 109: 109. 190937
Radke LF; Coakley JA; King MD. (1989). Direct and remote sensing observations of the effects of ships on clouds.
Science, 246: 1146-1149. 156034
Rajapaksha R; MT-K; Baath E. (2004). Metal toxicity affects fungal and bacterial activities in soil differently. Appl Environ
Microbiol, 5: 2966 - 2973. 156035
Ramachandran S; Ramaswamy V; Stenchikov GL; Robock A. (2000). Radiative impact of the Mount Pinatubo volcanic
eruption: Lower stratospheric response. J Geophys Res, 105: 24409-24429. 192050
Ramanathan V; Chung C; Kim D; Bettge T; Buja L; Kiehl JT; Washington WM; Fu Q; Sikka DR; Wild M. (2005).
Atmospheric brown clouds: Impact on South Asian climate and hydrologic cycle. PNAS, 102: 5326-5333. 190199
Ramanathan V; Crutzen P. (2003). Atmospheric Brown "Clouds". Atmos Environ, 37: 4033-4035. 190198
Ramanathan V; Crutzen PJ; Kiehl JT; Rosenfeld D. (2001). Aerosols, climate and the hydrological cycle. Science, 294:
2119-2124.042681
Ramanathan V; Crutzen PJ; Lelieveld J; Mitra AP; Althausen D; Anderson J; Andreae MO; Cantrell W; Cass GR; Chung
CE; Clarke AD; Coakley JA; Collins WD; Conant WC; Dulac F; Heintzenberg J; Heymsfield AJ; Holben B;
Howell S; Hudson J; Jayaraman A; Kiehl JT; Krishnamurti TN; Lubin D; McFarquhar G; Novakov T; Ogren JA;
Podgorny IA; Prather K; Priestley K; Prospero JM; Quinn PK: Rajeev K; Rasch P; Rupert S; Sadourny R; Satheesh
SK; Shaw GE; Sheridan P; Valero FPJ. (2001). Indian Ocean Experiment: An integrated analysis of the climate
forcing and effects of the great Indo-Asian haze. J Geophys Res, 106: 28371-38398. 190196
Ramaswamy V; Boucher O; Haigh J; Hauglustaine D; Haywood J; Myhre G; Nakajima T; Shi G; Solomon S. (2001).
Radiative forcing of climate change, Chapter 6. In Houghton JT; Ding Y; Griggs DJ; Noguer M; van der Linden PJ;
Da X; Maskell K; Johnson CA (Ed.),IPCC Third Assessment Report (TAR): Climate Change 2001: Working Group
I: The Scientific Basis (pp. 349-416). Cambridge, U.K. and New York, NY: Intergovernmental Panel on Climate
Change; Cambridge University Press. 156899
Ramaswamy V; Ramachandran S; Stenchikov G; Robock A. (2006). A model study of the effect of Pinatubo volcanic
aerosols on the stratospheric temperatures+. In Kiehl JT; Ramanathan V (Ed.),Frontiers of climate modeling (pp.
152-178). Cambridge, UK: Cambridge University Press. 192273
Ramaswamy V; Schwarzkopf MD; Randel WJ; Santer BD; Soden BJ; Stenchikov GL. (2006). Anthropogenic and natural
influences in the evolution of lower stratospheric cooling. Science, 311: 1138-1141. 192284
Ramos L; Hernandez LM; Gonzalez MJ. (1994). Sequential fractionation of copper, lead, cadmium and zinc in soils from
or near Donana National Park. J Environ Qual, 23: 50-57. 046736
Ramsey JM. (1966). Concentrations of carbon monoxide at traffic intersections in Dayton, Ohio. Arch Environ Occup
Health, 13: 44-46. 013946
Randall CE; Bevilacqua RM; Lumpe JD; Hoppel KW. (2001). Validation of POAM III aerosols: Comparison to SAGE II
and HALOE. J Geophys Res, 106: 27525-27536. 192268
Randall D; Khairoutdinov M; Arakawa A; Grabowski W. (2003). Breaking the cloud parameterization deadlock. Bull Am
Meteorol Soc, 84: 1547-1564. 190201
Randerson JT; Liu H; Planner MG; Chambers SD; Jin Y; Hess PG; Pfister G; Mack MC; Treseder KK; Welp LR; Chapin
FS; Harden JW; Goulden ML; Lyons E; Neff JC; Schuur BAG; Zender CS. (2006). The impact of boreal forest fire
on climate warming. Science, 314: 1130-1132. 156038
December 2009 9-228
-------
Rapport DJ; Whitford WG. (1999). How ecosystems respond to stress: Common properties of arid and aquatic systems.
Bioscience, 49: 193-203. 004595
Ray P; Reddy UG; Lapeyrie F; Adholeya A. (2005). Effect of coal ash on growth and metal uptake by some selected
ectomycorrhizal fungi in vitro. Int J Phytoremediation, 7: 199-216. 190473
Read WG; Froidevaux L; Waters JW (1993). Microwave limb sounder measurement of stratospheric SO2 from the Mt.
Pinatubo Volcano. Geophys Res Lett, 20: 1299-1302. 192031
Reddy M; Boucher O; Balkanski Y; Schulz M. (2005). Aerosol optical depths and direct radiative perturbations by species
and source type. Geophys Res Lett, 32: LI2803. 190208
Reddy M; Boucher O; BellouinN; Schulz M; Balkanski Y; Dufresne J; Pham M. (2005). Estimates of multi-component
aerosol optical depth and direct radiative perturbation in the LMDZT general circulation model. J Geophys Res,
110: D10S16.190207
Reddy MS; Boucher O. (2007). Climate impact of black carbon emitted from energy consumption in the world's regions.
Geophys Res Lett, 34: 1802. 156042
Regoli F; Gorbi S; Fattorini D; Tedesco S; Notti A; Machella N; Bocchetti R; Benedetti M; Piva F. (2006). Use of the land
snail Helix aspersa sentinel organism for monitoring ecotoxicologic effects of urban pollution: An integrated
approach. Environ Health Perspect, 114: 63-69. 156046
Reid J. (2008). An overview of UAE2 flight operations: Observations of summertime atmospheric thermodynamic and
aerosol profiles of the southern Arabian Gulf. J Geophys Res, 113: D14213. 190214
Reid JS; Kinney JE; Westphal DL; Holben BN; Welton EJ; Tsay S-C; Eleuterio DP; Campbell JR; Christopher SA; Colarco
PR; Jonsson HH; Livingston JM; Maring HB; Meier ML; Pilewskie P; Prospero JM; Reid EA; Remer LA; Russell
PB; Savoie DL; Smirnov A; Tanre D. (2003). Analysis of measurements of Saharan dust by airborne and ground-
based remote sensing methods during the Puerto Rico Dust Experiment (PRIDE). J Geophys Res, 108: 8586.
190213
Reinfelder JR; Fisher NS; Luoma SN; Nichols JW; Wang W-X. (1998). Trace element trophic transfer in aquatic
organisms: Acritique of the kinetic model approach. Sci Total Environ, 219: 117-135. 156047
Reisinger LM. (1990). Analysis of airborne particles sampled in the southern Appalachian Mountains. Water Air Soil
Pollut, 50: 149-162. 046737
Remer L; Gasso; Hegg D; Kaufman Y; Holben B. (1997). Urban/industrial aerosol: ground based sun/sky radiometer and
airborne in situ measurements. J Geophys Res, 102: 16849-16859. 190216
Remer L; Kaufman Y. (2006). Aerosol direct radiative effect at the top of the atmosphere over cloud free ocean derived
from four years of MODIS data. Atmos Chem Phys, 6: 237-253. 190222
Remer L; Kaufman Y; Tanre D; Mattoo S; Chu D; Martins J; Li R; Ichoku C; Levy R; Kleidman R; Eck T; Vermote E;
Holben B. (2005). The MODIS aerosol algorithm, products and validation. J Atmos Sci, 62: 947-973. 190221
Remer L; Kleidman RG; Levy RC; Kaufman YJ; Tanre D; Mattoo S; Vanderlei Martins J; Ichoku C; Koren I; Yu H;
Holben BN. (2008). An emerging aerosol climatology from the MODIS satellite sensors. J Geophys Res, 113:
D14S01. 190224
Remer L; Tanre D; Kaufman Y; Ichoku C; Mattoo S; Levy R; Chu D; Holben B; Dubovik O; Smirnov A; Martins J; Li R;
Ahman Z. (2002). Validation of MODIS aerosol retrieval over ocean. Geophys Res Lett, 29: 8008. 190218
Richards LW; Dye TS; Arthur M; Byars MS. (1996). Analysis of ASOS Data for Visibility Purposes, Final Report.
Systems Applications International, Inc.. San Rafael, CA. STI-996231-1610-FR. 190476
Rind D; Perlwitz J; Lonergan P. (2005). AO/NAO response to climate change: 1. Respective influences of stratospheric and
tropospheric climate changes. J Geophys Res, 110: D12107. 192261
Rissler J; Swietlicki E; Zhou J; Roberts G; Andreae MO; Gatti LV; Artaxo P. (2004). Physical properties of the sub-
micrometer aerosol over the Amazon rain forest during the wet-todry season transition—comparison of modeled
and measured CCN concentrations. Atmos Chem Phys, 4: 2119-2143. 190225
Roberts DL; Jones A. (2004). Climate sensitivity to black carbon aerosol from fossil fuel combustion. J Geophys Res, 109:
6202. 156052
December 2009 9-229
-------
Roberts ES; Richards JH; Jaskot R; Dreher KL. (2003). Oxidative stress mediates air pollution particle-induced acute lung
injury and molecular pathology. Inhal Toxicol, 15: 1327-1346. 156051
Robinson A; Donahue N; Shrivastava M; Weitkamp E; Sage A; Grieshop A; Lane T; Pierce J; Pandis S. (2007). Rethinking
organic aerosols: Semivolatile emissions and photochemical aging. Science, 315: 1259. 191975
Roderick ML; Farquhar GD. (2002). The cause of decreased pan evaporation over the past 50 years. Science, 298: 1410-
1411.042788
Roeckner E; Stier P; Feichter J; Kloster S; Esch M; Fischer-Bruns L. (2006). Impact of carbonaceous aerosol emissions on
regional climate change. Clim Dynam, 27: 553-571. 156920
Romic M; Romic D. (2003). Heavy metals distribution in agricultural topsoils in urban area. Environ Geol, 43: 795-805.
156055
Rood MJ; Covert DS; Larson TV. (1987). Hygroscopic properties of atmospheric aerosol in Riverside, California. Tellus B
Chem Phys Meteorol, 39B: 383-397. 046397
Rosenfeld D. (2000). Suppression of rain and snow by urban and industrial air pollution. Science, 287: 1793-1796. 002234
Rosenfeld D. (2006). Aerosols, clouds, and climate. Science,312: 1. 190233
Rosenfeld D; Dai J; Yu X; Yao Z; Xu X; Yang X; Du C. (2007). Inverse relations between amounts of air pollution and
orographic precipitation. Science, 315: 1396-1398. 156057
Rosenfeld D; Givati A. (2006). Evidence of orographic precipitation suppression by air pollution-induced aerosols in the
western United States. J Appl Meteor Climatol, 45: 893-911. 156924
Rosenfeld D; Lansky I. (1998). Satellite-based insights into precipitation formation processes in continental and maritime
convective clouds. Bull Am Meteorol Soc, 79: 2457-2476. 190230
Ross DM; Malm WC; Loomis RJ. (1985). The psychological valuation of good visual air quality by national park visitors.
Presented at 78th annual meeting of the Air Pollution Control Association, June, 1985, Detroit, MI. 044287
Ross DM; Malm WC; Loomis RJ. (1987). An examination of the relative importance of park attributes at several national
parks. In Bhardwaja, P. S. (Ed.),Visibility protection: research and policy aspects, an APCA international specialty
conferenceGrand Teton National Park, WY; Pittsburgh, PA: Air Pollution Control Association. 037420
Rosselli W; Keller C; Boschi K. (2003). Phytoextraction capacity of trees growing on a metal contaminated soil. Plant Soil,
256: 265-272. 156058
Roth CM; Goss K-U; Schwarzenbach RP. (2004). Sorption of diverse organic vapors to snow. Environ Sci Technol, 38:
4078-4084. 056431
Rotstayn LD; Liu Y (2003). Sensitivity of the first indirect aerosol effect to an increase of the cloud droplet spectral
dispersion with droplet number concentration. J Clim, 16: 3476-3481. 199905
Rotstayn LD; Penner JE. (2001). Indirect aerosol forcing, quasi-forcing, and climate response. J Clim, 14: 2960-2975.
193754
Rotton J; Barry T; Milligan M; Fitzpatrick M. (1979). The air pollution experience and interpersonal aggression. Appl
Psychol, 9: 397-412. 190478
Rotton J; Frey J. (1982). Atmospheric conditions, seasonal trends, and psychiatric emergencies. In Replications and
Extensions (pp. Unknown). Washington, DC: American Psychological Association. 190477
Rozanov EV; Schlesinger ME; Andronova NG; Yang F; Malyshev SL; Zubov VA; Egorova TA; Li B. (2002).
Climate/chemistry effects of the Pinatubo volcanic eruption simulated by the UIUC stratosphere/troposphere GCM
with interactive photochemistry. J Geophys Res, 107: 4594. 192247
Rozanov EV; Schlesinger ME; Egorova TA; Li B; Andronova N; Zubov VA. (2004). Atmospheric response to the observed
increase of solar UV radiation from solar minimum to solar maximum simulated by the University of Illinois at
Urbana-Champaign climate-chemistry model. J Geophys Res, 109: D01110. 192245
Ruckstuhl C; Philipona R; Behrens K; Coen MC; Dilrr B; Heimo A; Matzler C; Nyeki S; Ohmura A; Vuilleumier L; Weller
M; Wehrli C; Zelenka A. (2008). Aerosol and cloud effects on solar brightening and recent rapid warming. Geophys
Res Lett, 35: LI2708. 190356
December 2009 9-230
-------
Russell P; Kinne S; Bergstrom R. (1997). Aerosol climate effects: local radiative forcing and column closure experiments. J
Geophys Res, 102: 9397-9407. 190359
Russell P; Livingston J; Hignett P; Kinne S; Wong J; Chien A; Bergstrom R; Durkee P; Hobbs P. (1999). Aerosol-induced
radiative flux changes off the United States mid-Atlantic coast: comparison of values calculated from sun
photometer and in situ data with those measured by airborne pyranometer. J Geophys Res, 104: 2289-2307. 190363
Rusu A-M; Jones GC; Chimonides PDJ; Purvis OW. (2006). Biomonitoring using the lichen Hypogymnia physodes and
bark samples near Zlatna, Romania immediately following closure of a copper ore-processing plant. Environ Pollut,
143: 81-88. 156062
Ryan PA; Lowenthal D; Kumar N. (2005). Improved light extinction reconstruction in interagency monitoring of protected
visual environments. JAir Waste Manag Assoc, 55: 1751-1759. 156934
Sabbioni C; Ghedini N; Bonazza A. (2003). Organic anions in damage layers on monuments and buildings. Atmos Environ,
37: 1261-1269. 049282
Sabin LD; Lim JH; Stolzenbach KD; Schiff KC. (2005). Contribution of trace metals from atmospheric deposition to
stormwater runoff in a small impervious urban catchment. Water Res, 39: 3929-3937. 088300
Salemaa M; Derome J; Helmisaari HS; Nieminen T; Vanha-Majamaa I. (2004). Element accumulation in boreal
bryophytes, lichens and vascular plants exposed to heavy metal and sulfur deposition in Finland. Sci Total Environ,
324: 141-160. 156069
Samuels HJ; Twiss S; Wong EW. (1973). Visibility, light scattering and mass concentration of particulate matter: report of
the California Tri-City Aerosol Sampling Project. Air Resources Board. CA. 070601
Sanderson EG; Farant JP. (2004). Indoor and outdoor poly cyclic aromatic hydrocarbons in residences surrounding a
Soderberg aluminum smelter in Canada. Environ Sci Technol, 38: 5350-5356. 156942
Sato M; Hansen J; Koch D; Lucis A; Ruedy R; Dubovik O; Holben B; Chin M; Novakov T. (2003). Global atmospheric
black carbon inferred from AAEONET. Presented at Proceedings of the National Academy of Science. 156947
Sato M; Hansen JE; McCormick MP; Pollack JB. (1993). Stratospheric aerosol optical depths, 1850-1990. J Geophys Res,
98: 22987-22994. 192046
Sauve S. (2001). Speciation of metals in soils. Presented at, Pensacola, FL. 156948
Saxena P; Hildemann LM; McMurry PH; Seinfeld JH. (1995). Organics alter hygroscopic behavior of atmospheric
particles. J Geophys Res, 100: 18,755-18,770. 077273
Schichtel BA; Gebhart K; Barna MG; Malm WC; Green MC. (2004). Big Bend Regional Aerosol and Visibility
Observational (BRAVO) Study Results: Air Quality Data and Source Attribution Analyses Results from the
National Park Service. Colorado State University CIRA. Ft. Collins, CO
.http://vista.cira.colostate.edu/improve/Studies/BRAVO/Studybravo.htm. 179902
Schichtel BA; Gebhart KA; Malm WC; Barna MG; Pitchford ML; Knipping EM; Tombach IH. (2005). Reconciliation and
interpretation of big bend national park particulate sulfur source apportionment: Results from the big bend regional
aerosol and visibility observational study-parti. JAir Waste Manag Assoc, 55: 1709-1725. 156957
Schichtel BA; Malm WC; Bench G; Fallen S; McDade CE; Chow JC; Watson JG (2008). Fossil and contemporary fine
particulate carbon fractions at 12 rural and urban sites in the United States. J Geophys Res, 113: D02311. 156958
Schiff K; Bay S; Stransky C. (2002). Characterization of stormwater toxicants from an urban watershed to freshwater and
marine organisms. Urban Water, 4: 215-227. 156959
Schilling JS; Lehman ME. (2002). Bioindication of atmospheric heavy metal deposition in the Southeastern US using the
moss Thuidium delicatulum. Atmos Environ, 36: 1611-1618. 113075
Schmid B; Ferrare R; Flynn C; Elleman R; Covert D; Strawa A; Welton E; Turner D; Jonsson H; Redemann J; Eilers J;
Ricci K; Hallar AG; Clayton M; Michalsky J; Smirnov A; Holben B; Barnard J. (2006). How well do state-of-the-
art techniques measuring the vertical profile of tropospheric aerosol extinction compare?. J Geophys Res, 111:1 -25.
190372
December 2009 9-231
-------
Schmid B; Livingston JM; Russell PB; Durkee PA; Jonsson HH; Collins DR; Flagan RC; Seinfeld JH; Gasso S; Hegg DA;
Ostrom E; Noone KJ; Welton EJ; Voss KJ; Gordon HR Formenti, P(9) Andreae, M.O. (2000). Clear sky closure
studies of lower tropospheric aerosol and water vapor during ACE-2 using airborne sunphotometer, airborne in situ,
space-borne, and ground-based measurements. Tellus B Chem Phys Meteorol, 52: 568-593. 190369
Schmidt GA; Ruedy R; Hansen JE; Aleinov I; Bell N; Bauer M; Bauer S; Cairns B; Canuto V; Cheng Y; del Genio A;
Faluvegi G; Friend AD; Hall TM; Hu Y; Kelley M; Kiang NY; Koch D; Lacis AA; Lerner J; Lo KK; Miller RL;
Nazarenko L; Oinas V; Perlwitz J; Perlwitz J; Rind D; Romanou A; Russell GL; Sato MKI; Shindell DT; Stone PH;
Sun S; Tausnev N; Thresher D; Yao MS. (2006). Present-day atmospheric simulations using GISS Model E:
Comparison to in situ, satellite and reanalysis data. J Clim, 19: 153-192. 190373
Schoeberl MR; Douglass AR; Zhu Z; Pawson S. (2003). A comparison of the lower stratospheric age spectra derived from
a general circulation model and two data assimilation systems. J Geophys Res, 108: 4113. 057262
Schroeder WH; Munthe J. (1998). Atmospheric mercury - an overview. Atmos Environ, 32: 809-822. 014559
Schulz M; Textor C; Kinne S; Balkanski Y; Bauer S; Berntsen T; Berglen T; Boucher O; Dentener F; Guibert S; Isaksen
ISA; Iversen T; Koch D; Kirkevag A; Liu X; Montanaro V; Myhre G; Penner JE; Pitari G; Reddy S; Seland O; Stier
P; Takemura T. (2006). Radiative forcing by aerosols as derived from the AeroCom present-day and preindustrial
simulations. Atmos Chem Phys, 6: 5225-5246. 190381
Schwartz SE; Charlson RJ; Rodhe H. (2007). Quantifying climate change- Too rosy a picture?. Nature, 2: 23-24. 190384
Seifert M; Anke S; Holzinger S; Jaritz M; Arnhold W; Anke M. (1999). Cadmium and strontium content of mice, shrews.
and some invertebrates. J. Trace Microprobe Tech., 17: 357-365. 190480
Seigneur C; Johnson CD; Latimer DA; Bergstrom RW; Hogo H. (1984). Users manual for the Plume Visibility Model
(PLUVUE II). Final report 23 Feb-29 Aug 83. Systems Applications, Inc. San Rafael, CA . PB-84-158302. 156965
Seinfeld JH; Carmichael GR; Arimoto R; Conant WC; Brechtel FJ; Bates TS; Cahill TA; Clarke AD; Doherty SJ; Flatau
PJ; Huebert BJ; Kim J; Markowicz KM; Quinn PK; Russell LM; Russell PB; Shimizu A; Shinozuka Y; Song CH;
Tang Y; Uno I; Vogelmann AM; Weber RJ; Woo J-H; Zhang XY (2004). ACEAsia: Regional climatic and
atmospheric chemical effects of Asian dust and pollution. Bull Am Meteorol Soc, 85: 367-380. 190388
Sekiguchi M; Nakajima T; Suzuki K; Kawamoto K; Higurashi A; Rosenfeld D; Sano I; Mukai S. (2003). A study of the
direct and indirect effects of aerosols using global satellite data sets of aerosol and cloud parameters. J Geophys
Res, 108 D22: 4699. 190385
Shakya K; Chettri MK; Sawidis T. (2008). Impact of heavy metals (copper, zinc, and lead) on the chlorophyll content of
some mosses. Arch Environ Contam Toxicol, 54: 412-421. 156081
Shan XQ; Wang ZW; Wang WS; Zhang SZ; Wen B. (2003). Labile rhizosphere soil solution fraction for prediction of
bioavailability of heavy metals and rare earth elements to plants. Anal Bioanal Chem, 375: 400-407. 156972
Sharifi MR; Gibson AC; Rundel PW. (1999). Phenological and physiological responses of heavily dusted creosote bush
(Larrea tridentata) to summer irrigation in the Mojave Desert. Flora, 14: 369-378. 156082
Sheesley RJ; Schauer JJ; Hemming JD; Barman MA; Geis SW; Tortorelli JJ. (2004). Toxicity of ambient atmospheric
particulate matter from the Lake Michigan (USA) airshed to aquatic organisms. Environ Toxicol Chem, 23: 133-
140. 156084
Shindell DT; Levy H; Schwarzkopf II MD; Horowitz LW; Lamarque JF; Faluvegi G. (2008). Multimodel projections of
climate change from short-lived emissions due to human activities. J Geophys Res, 113: Dill09. 190393
Shindell DT; Schmidt GA; Mann ME; Faluvegi G. (2004). Dynamic winter climate response to large tropical volcanic
eruptions since 1600. J Geophys Res, 109: D05104. 192267
Shindell DT; Schmidt GA; Miller RL; Mann M. (2003). Volcanic and solar forcing of climate change during the
preindustrial era. J Clim, 16: 4094-4107. 192069
Shindell DT; Teich H; Chin M; Dentener F; Doherty RM; Faluvegi G; Fiore AM; Hess P; MacKenzie IA; Sanderson MG;
Schultz MG; Schulz M; Stevenson DS; Textor C; Wild O; Bergmann DJ; Bian H; Cuvelier C; Duncan BN; Folberth
G; Horowitz LW; Jonson J; Kaminski JW; Manner E; Park R; Pringle KJ; Schroeder S; Szopa S; Takemura T; Zeng
G; Keating TJ; Zuber A. (2008). A multi-model assessment of pollution transport to the Arctic. Atmos Chem Phys,
8: 5353-5372. 190391
December 2009 9-232
-------
Shulz KG; Zondervan I; Gerringa LJA; Timmermans KR; Veldhuls MJW; Riebesell U. (2004). Effect of trace metal
availability on coccolithophorid calcification. Nature, 430: 673-676. 156087
Singh HB; Brune WH; Crawford JH; Flocke F; Jacob DJ. (2008). Chemistry and transport of pollution over the Gulf of
Mexico and the Pacific: spring 2006 INTEX-B campaign overview and first results. Atmos Chem Phys Discuss, 9:
2301-2318. 190394
Sinyuk A; Dubovik O; Holben B; Eck TF; Breon FM; Martonchik J; Kahn R; Diner DJ; Vermote EF; Roger JC; Lapyonok
T; Slutsker I. (2007). Simultaneous retrieval of aerosol and surface properties from a combination of AERONET
and satellite data. Rem Sens Environ, 107: 90-108. 190395
Skov H; Christensen JH; Goodsite ME; Heidam NZ; Jensen B; Wahlin P; Geernaert G (2004). Fate of elemental mercury
in the Arctic during atmospheric mercury depletion episodes and the load of atmospheric mercury to the Arctic.
Environ Sci Technol, 38: 2373-2382. 190481
Sloane CS. (1983). Optical properties of aerosols—comparison of measurements with model calculations. Atmos Environ,
17: 409-416. 025039
Sloane CS. (1984). Optical properties of aerosols of mixed composition. Atmos Environ, 18: 871-878. 025040
Sloane CS. (1986). Effect of composition on aerosol light scattering efficiencies. Atmos Environ, 20: 1025-1037. 045954
Sloane CS; Wolff GT. (1985). Prediction of ambient light scattering using a physical model responsive to relative humidity:
validation with measurements from Detroit. Atmos Environ, 19: 669-680. 045953
Smejkalova M; Mikanova O; Boruvka L. (2003). Effects of heavy metal concentrations on biological activity of soil
microorganisms. Plant Soil Environ, 49: 3217326. 156987
Smirnov A; Holben B; Eck T; Dubovik O; Slutsker I. (2000). Cloud screening and quality control algorithms for the
AERONET database. Rem Sens Environ, 73: 337-349. 190397
Smirnov A; Holben B; Eck T; Slutsker I; Chatenet B; Pinker R. (2002). Diurnal variability of aerosol optical depth
observed at AERONET (Aerosol Robotic Network) sites. Geophys Res Lett, 29: 2115. 190398
Smirnov A; Holben B; Sakerin S; Kabanov DM; Slutsker I; Chin M; Diehl TL; Remer LA; Kahn R; Ignatov A; Liu L;
Mishchenko M; Eck TF; Kucsera TL; Giles D; Kopelevich OV. (2006). Shipbased aerosol optical depth
measurements in the Atlantic Ocean, comparison with satellite retrievals and GOCART model. Geophys Res Lett,
33: L14817.190400
Smith AE; Howell S. (2009). An assessment of the robustness of visual air quality preference study results. CRA
International. Washington,
DC .http://yosemite.epa.gov/sab/sabproduct.nsf/B5591 lDF9796E5E385257592006FB737/$File/CRA+VAQ+Pref+
Robustness+Studv+3+30+09+final.pdf. 198803
Smith WH. (1990). Forest nutrient cycling: toxic ions. In Air pollution and forests: interactions between air contaminants
and forest ecosystemsNew York, NY: Springer-Verlag. 046896
Smith WH. (1990). Forests as sinks for air contaminants: vegetative compartment. In Air pollution and forests: interactions
between air contaminants and forest ecosystems (pp. 147-180). New York, NY: Springer-Verlag. 084015
Smith WH. (1991). Air pollution and forest damage. Chem Eng News, 6945: 30-43. 042566
Smith WL Jr; Charlock TP; Kahn R; Martins JV; Remer LA; Hobbs PV; Redemann J; Rutledge CK. (2005). EOS Terra
aerosol and radiative flux validation: An overview of the Chesapeake Lighthouse and aircraft measurements from
satellites (CLAMS) experiment. J Atmos Sci, 62: 903-918. 190401
Smolders E; Degryse F. (2002). Fate and effect of zinc from tire debris in soil. Environ Sci Technol, 36: 3706-3710.
156091
Scares CRFS; Siqueira JO. (2008). Mycorrhiza and phosphate protection of tropical grass species against heavy metal
toxicity in multi-contaminated soil. Biol Fertil Soils, 44: 833-841. 190482
Sokolik I; Winker D; Bergametti G; Gillette DA; Carmichael G; Kaufman YJ; Gomes L; Schuetz L; Penner JE. (2001).
Introduction to special section: outstanding problems in quantifying the radiative impacts of mineral dust. J
Geophys Res, 106: 18015-18027. 190404
December 2009 9-233
-------
Solomon PA; Hopke PK. (2008). A Special Issue of JA&WMA Supporting Key Scientific and Policy-and Health-Relevant
Findings from EPA's Particulate Matter Supersites Program and Related Studies: An Integration and Synthesis of
Results. J Air Waste ManagAssoc, 58: 137. 156997
Solomon S; Portmann RW; Garcia RR; Thomason LW; Poole LR; McCormick MP. (1996). The role of aerosol variations
in anthropogenic ozone depletion at northern midlatitudes. J Geophys Res, 101: 6713-6727. 192252
Sotiropoulou REP; Medina J; Nenes A. (2006). CCN predictions: is theory sufficient for assessments of the indirect effect?.
Geophys Res Lett, 33: L05816. 190406
Sotiropoulou REP; Nenes A; Adams PJ; Seinfeld JH. (2007). Cloud condensation nuclei prediction error from application
of Kohler theory: Importance for the aerosol indirect effect. J Geophys Res, 112: D12202. 190405
Spinhirne J; Palm S; Hart W; Hlavka D; Welton E. (2005). Cloud and Aerosol Measurements from the GLAS Space Borne
Lidar: initial results. Geophys Res Lett, 32: L22S03. 190410
Spracklen DV; Logan JA; Mickley LJ; Park RJ; Yevich R; Westerling AL; Jaffe DA. (2007). Wildfires drive interannual
variability of organic carbon aerosol in the western U.S. in summer. Geophys Res Lett, 34: LI6816. 190485
Squires P. (1958). The microstructure and colloidal stability of warm clouds: Part I - the relation between structure and
stability. Tellus DynMeteorol Oceanogr, 10: 256-261.045608
Srogi. (2007). Monitoring of environmental exposure to polycyclic aromatic hydrocarbons: a review. Environ Chem Lett,
5: 169-195. 180049
Stanhill G; Cohen S. (2001). Global dimming: a review of the evidence for a widespread and significant reduction in global
radiation with discussion of its probable causes and possible agricultural consequences. Agr Forest Meteorol, 107:
255-278. 042121
Steinnes E; Hvatum O0; B01viken B; Varskog P. (2005). Atmospheric supply of trace elements studied by peat samples
from ombrotrophic bogs. J Environ Qual, 34: 192-197. 156095
Stenchikov G; Hamilton K; Robock A; Ramaswamy V; Schwarzkopf MD. (2004). Arctic Oscillation response to the 1991
Pinatubo eruption in the SKYHI GCM with a realistic quasi-biennial oscillation. J Geophys Res, 109: D03112.
192274
Stenchikov G; Hamilton K; Stouffer RJ; Robock A; Ramaswamy V; Santer B; Graf H-F. (2006). Arctic Oscillation
response to volcanic eruptions in the IPCC AR4 climate models. J Geophys Res, 111: D07107. 192260
Stenchikov G; Robock A; Ramaswamy V; Schwarzkopf MD; Hamilton K; Ramachandran S. (2002). Arctic Oscillation
response to the 1991 Mount Pinatubo eruption: Effects of volcanic aerosols and ozone depletion. J Geophys Res,
107:024. 192277
Stenchikov GL; Kirchner I; Robock A; Graf H-F; Antuna JC; Grainger RG; Lambert A; Thomason L. (1998). Radiative
forcing from the 1991 Mount Pinatubo volcanic eruption. J Geophys Res, 103: 13837-13857. 192049
Stephens G; Vane DG; Boain RJ; Mace GG; Sassen K; Wang Z; Illingworth AJ; O'Connor EJ; Rossow WB; Durden SL;
Miller SD; Austin RT; Benedetti A; Mitrescu C; The CloudSat Science Team. (2002). The CloudSat mission and the
A-Train. Bull Am Meteorol Soc, 83: 1771-1790. 190412
Stephens GL; Haynes JM. (2007). Near global observations of the warm rain coalescence process. Geophys Res Lett, 34:
L20805. 190413
Stern DI. (2005). Global sulfur emissions from 1850 to 2000. Chemosphere, 58: 163-175. 190416
Stern GA; Halsall CJ; Barrie LA; Muir DCG; Fellin P; Rosenberg B. (1997). Polychlorinated biphenyls in arctic air. 1.
Temporal and spatial trends: 1992-1994. Environ Sci Technol, 31: 3619-3628. 156096
Stevens B; Feingold G; Walko RL; Cotton WR. (1996). On elements of the microphysical structure of numerically
simulated non-precipitating stratocumulus. JAtmos Sci, 53: 980-1006. 190417
Stewart AR; Luoma SN; Schlekat CE; Doblin MA; Hieb KA. (2004). Food web pathway determines how selenium affects
aquatic ecosystems. Environ Sci Technol, 38: 4519-4526. 156097
Stier P; Seinfeld JH; Kinne S; Boucher O. (2007). Aerosol absorption and radiative forcing. Atmos Chem Phys, 7: 5237-
5261. 157012
December 2009 9-234
-------
Stohl A; Andrews E; Burkhart IF; Forster C; Herber A; Hoch SW; Kowal D; Lunder C; Mefford T; Ogren JA; Sharma S;
SpichtingerN; Stebel K; Stone R; Strom J; T0rsethK; Wehrli C; Yttri KE. (2006). Pan-Arctic enhancements of
light absorbing aerosol concentrations due to North American boreal forest fires during summer 2004. J Geophys
Res, 111:022214.156100
Stohl A; Berg T; Burkhart JF; Fjajraa AM; Forster C; Herber A; Hov 0; Lunder C; McMillan WW; Oltmans S; Shiobara M;
Dimpson D; Solberg S; Stebel K; Strom J; T0rseth K; Treffeisen R; Virkkunen K; Yttri KE. (2007). Arctic smoke -
record high air pollution levels in the European Arctic due to agricultural fires in Eastern Europe in spring 2006.
Atmos Chem Phys, 7: 511-534. 157015
Storlevmo T; Kristjansson JE; Myhre G; Johnsud M; Stordal F. (2006). Combined observational and modeling based study
of the aerosol indirect effect. Atmos Chem Phys, 6: 3583-3601. 190418
Stothers RB. (2001). A chronology of annual mean effective radii of stratospheric aerosols from volcanic eruptions during
the twentieth century as derived from ground-based spectral extinction measurements. J Geophys Res, 106: 32043-
32049. 192233
Stothers RB. (2001). Major optical depth perturbations to the stratosphere from volcanic eruptions: Stellar extinction
period, 1961-1978. J Geophys Res, 106: 2993-3003. 192232
Stott PA. (2006). Observational constraints on past attributable warming and predictions of future global warming. J Clim,
19: 3055-3069. 190419
Strachan WMJ; Burniston DA; Williamson M; Bohdanowicz H. (2001). Spatial difference in persistent organochlorine
pollutant concentrations between the Bering and Chukchi Seas. Mar Pollut Bull, 43: 132-142. 156103
Strandberg B; Axelsen JA; Pedersen MB; Jensen J; Attrill MJ. (2006). Effect of a copper gradient on plant community
structure. Environ Toxicol Chem, 25: 743-753. 156105
Strawa A; Castaneda R; Owano T; Baer P; Paldus B. (2002). The measurement of aerosol optical properties using
continuous wave cavity ring-down techniques. J Atmos Ocean Tech, 20: 454-465. 190421
Streets D; Bond T; Lee T; Jang C. (2004). On the future of carbonaceous aerosol emissions. J Geophys Res, 109: D24212.
190423
Streets D; Wu Y; Chin M. (2006). Two-decadal aerosol trends as a likely explanation of the global dimming/brightening
transition. Geophys Res Lett, 33: LI5806. 190425
Streets DG; Aunan K. (2005). The importance of China's household sector for black carbon emissions. Geophys Res Lett,
32: L12708. L56106
Streets DG; Zhang Q; Wang L; He K; Hao J; Wu Y; Tang Y; Carmichael GR. (2006). Revisiting China?s CO emissions
after the Transport and Chemical Evolution over the Pacific (TRACE-P) mission: Synthesis of inventories,
atmospheric modeling, and observations. J Geophys Res, 111: Dl4306. 157019
Strydom C; Robinson C; Pretorius E; Whitcutt JM; Marx J; Bornman MS. (2006). The effect of selected metals on the
central metabolic pathways in biology: A review. WaterSA, 32: 543-554. 190486
Suzuki K; Nakajima T; Numaguti A; Takemura T; Kawamoto K; Higurashi A. (2004). A study of the aerosol effect on a
cloud field with simultaneous use of GCM modeling and satellite observation. J Atmos Sci, 61: 179-194. 199917
Swackhamer DL; Paerl HW; Eisenreich SJ; Hurley J; Hornbuckle KG; McLachlan M; Mount D; Muir D; Schindler D.
(2004). Impacts of atmospheric pollutants on aquatic ecosystems. Issues in Ecology, 12:1 -24. 190488
Sweet C; Caughey M; Gay D. (2005). Midwest Ammonia Monitoring Project: Summary for October 2003 through
November 2004. Illinois State Water Survey.
IL.http://www.ladco.org/reports/rpo/monitoring/ammonia_study _monitoring_report.pdf. 180038
Szabo L; Fodor L. (2006). Uptake of microelements by crops grown on heavy metal-amended soil. Commun Soil Sci Plant
Anal, 37: 2679-2689. 156109
Tabazadeh A; Drdla K; Schoeberl MR; Hamill P; Toon OB. (2002). Arctic "ozone hole" in a cold volcanic stratosphere.
PNAS, 99: 2609-2612. 192250
Takemura T; Nakajima T; Dubovik O; Holben B; Kinne S. (2002). Single-scattering albedo and radiative forcing of various
aerosol species with a global three-dimensional model. PNAS, 15: 333-352. 190438
December 2009 9-235
-------
Takemura T; Nozawa T; Emori S; Nakajima T; Nakajima T. (2005). Simulation of climate response to aerosol direct and
indirect effects with aerosol transport-radiation model. J Geophys Res, 110: D02202. 190439
Tang IN. (1996). Chemical and size effects of hygroscopic aerosols on light scattering coefficients. J Geophys Res, 101:
2457219. 157042
Tang Y; Carmichael GR; Seinfeld JH; Dabdub D; Weber RJ; Huebert B; Clarke AD; Guazzotti SA; Sodeman DA; Prather
KA; Uno I; Woo JH; Yienger JJ; Streets DG; Quinn PK; Johnson JE; Song CH; Grassian VH; Sandu A; Talbot RW;
Dibb JE. (2004). Three-dimensional simulations of inorganic aerosol distributions in East Asia during spring 2001.
J Geophys Res, 109: D19S23. 190445
Tang Y; Carmichael GR; Woo JH; Thongboonchoo N; Kurata G; Uno I; Streets DG; Blake DR; Weber RJ; Talbot RW;
Kondo Y; Singh HB; Wang T. (2003). Influences of biomass burning during the Transport and Chemical Evolution
Over the Pacific (TRACE-P) experiment identified by the regional chemical transport model. J Geophys Res, 108:
8824. 190441
Tanre D; Haywood J; Pelon J; Leon JF; Chatenet B; Formenti P; Francis P; Goloub P; Highwood EJ; Myhre G. (2003).
Measurement and modeling of the Saharan dust radiative impact: Overview of the Saharan Dust Experiment
(SHADE). J Geophys Res, 108: 8574. 190454
Tanre D; Kaufman Y; Herman M; Mattoo S. (1997). Remote sensing of aerosol properties over oceans using the MODIS/
EOS spectral radiances. J Geophys Res, 102: 16971-16988. 190452
Tao S; Jiao X; Chen S; Xu F; Li Y; Liu F. (2006). Uptake of vapor and particulate poly cyclic aromatic hydrocarbons by
cabbage. Environ Pollut, 140: 13-15. 156112
Taulavuori K; Prasad MNV; Taulavuori E; Laine K. (2005). Metal stress consequences on frost hardiness of plants at
northern high latitudes: a review and hypothesis. Environ Pollut, 135: 209-220. 190489
Tett SFB; Jones GS; Stott PA; Hill DC; Mitchell JFB; Allen MR; Ingram WJ; Johns TC; Johnson CE; Jones A; Roberts DL;
Sexton DMH; Woodage MJ. (2002). Estimation of natural and anthropogenic contributions to twentieth century
temperature change. J Geophys Res, 107: 4306. 192053
Textor C; Schulz M; Guibert S; Kinne S; Balkanski Y; Bauer S; Berntsen T; Berglen T; Boucher O; Chin M; Dentener F;
Diehl T; Easter R; Feichter H; Fillmore D; Ghan S; Ginoux P; Gong S; Grini A; Hendricks J; Horowitz L; Huang P;
Isaksen I; Iversen T; Kloster S; Koch D; Kirkevag A; Kristjansson JE; Krol M; Lauer A; Lamarque JF; Liu X
Montanaro V; Myhre G; Penner J; Pitari G; Reddy S; Sel 0; Stier P; Takemura T; Tie X. (2006). Analysis and
quantification of the diversities of aerosol life cycles within AEROCOM. Atmos Chem Phys, 6: 1777-1813. 190456
Textor C; Schulz M; Guibert S; Kinne S; Balkanski Y; Bauer S; Berntsen T; Berglen T; Boucher O; Chin M; Dentener F;
Diehl T; Feichter J; Fillmore D; Ginoux P; Gong S; Grini A; Hendricks J; Horowitz L; Huang P; Isaksen ISA;
Iversen T; Kloster S; Koch D; Kirkevag A; Kristjansson JE; Krol M; Lauer A; Lamarque JF; Liu X; Montanaro V;
Myhre G; Penner JE; Pitaril4 G; Reddy MS; Seland O; Stier P; Takemura T; Tie X. (2007). The effect of
harmonized emissions on aerosol properties in global models?an AeroCom experiment. Atmos Chem Phys, 7:
4489-4501. 190458
Thomas GO; Smith KEC; SweetmanAJ; Jones KC. (1998). Further studies of the air-pasture transfer of poly chlorinated
biphenyls. Environ Pollut, 102: 119-128. 156118
Thomason LW; Peter T. (2006). Assessment of Stratospheric Aerosol Properties (ASAP): Report on the Assessment Kick-
Off Workshop, Paris, France, 4-6 November 2001. Presented at Assessment of Stratospheric Aerosol Properties ,
Paris, France. 192248
Tie X; Madronich S; Walters S; Edwards DP; Ginoux P; Mahowald N; Zhang R; Lou C; Brasseur C. (2005). Assessment of
the global impact of aerosols on tropospheric oxidants. J Geophys Res, 110: 1-32. 190459
Tie XX; Brasseur GP; Briegleb B; Granier C. (1994). Two-dimensional simulation of Pinatubo aerosol and its effect on
stratospheric ozone. J Geophys Res, 99: 20545-20562. 192253
Timmreck C; Graf HF; Steil B. (2003). Aerosol chemistry interactions after the Mt. Pinatubo eruption. In Robock A;
Oppenheimer C (Ed.),Volcanism and the earth's atmosphere (pp. 213-227). Washington, DC: American Geophysical
Union. 192254
Tombach I; McDonald K. (2004). Visibility and Radiative Balance Effects. In PH. McMurry MFS, and J.S. Vickery
(Ed.),Particulate Matter Science for Policy Makers: ANARSTO Assessment (pp. 325-354). Cambridge, MA:
Cambridge University Press. 157054
December 2009 9-236
-------
loose LK; Mackay D. (2004). Adaptation of fugacity models to treat speciating chemicals with constant species
concentration ratios. Environ Sci Technol, 38: 4619-4626. 156123
Torres O; Bhartia P; Herman J; Ahmad Z; Gleason J. (1998). Derivation of aerosol properties from satellite measurements
of backscattered ultraviolet radiation: Theoretical bases. J Geophys Res, 103: 17009-17110. 190503
Torres O; Bhartia P; Herman J; Sinyuk A; Ginoux P; Holben B. (2002). A long-term record of aerosol optical depth from
TOMS observations and comparison to AERONET measurements. JAtmos Sci, 59: 398-413. 190505
Torres O; Bhartia P; Sinyuk A; Welton E; Holben B. (2005). Total Ozone Mapping Spectrometer measurements of aerosol
absorption from space: Comparison to SAFARI 2000 groundbased observations. J Geophys Res, 110: D10S18.
190507
Tremper AH; Agneta M; Burton S; Higgs DEB. (2004). Field and laboratory exposures of two moss species to low level
metal pollution. JAtmos Chem, 49: 111-120. 156126
Trijonis J; Thayer M; Murdoch J; Hagemen R. (1985). Air quality benefit analysis for Los Angeles and San Francisco
based on housing values and visibility. California Air Resources Board. Sacramento, CA. 078468
Trijonis JC; Malm WC; Pitchford M; White WH; Charlson R. (1990). Acidic deposition: State of science and technology.
Report 24. Visibility: Existing and historical conditions-causes and effects. Final report. 157058
Turco RP; Toon OB; WhittenRC; Pollack JB; Hamill P. (1983). The global cycle of particulate elemental carbon: a
theoretical assessment. In Pruppacher HR (Ed.),Precipitation Scavenging, Dry Deposition, and Resuspension (pp.
1337- 1351). New York: Elsevier Science. 190529
Twomey S. (1977). The influence of pollution on the shortwave albedo of clouds. JAtmos Sci, 34: 1149-1152. 190533
U.S. EPA. (1979). Protecting Visibility, an EPA Report to Congress. U.S. Environmental Protection Agency. Washington,
DC. 157065
U.S. EPA. (1982). Air quality criteria for particulate matter and sulfur oxides, Vol 1, 2, 3. U.S. Environmental Protection
Agency, Environmental Criteria and Assessment Office. Washington, D.C.. EPA 600/8-82-029a; EPA 600/8-82-
029b; EPA 600/8-82-029c. http://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=3000188Z.txt;
http ://nepis. epa. gov/Exe/ZyPURL. cgi?Dockey=3 00018EV.txt;
http://nepis.epa.gov/Exe/ZyPURL.cgi?Dockev=300053KV.txt. 017610
U.S. EPA. (1993). Air quality criteria for oxides of nitrogen, 3 Volumes. Environmental Criteria and Assessment Office,
Office of Health and Environmental Assessment, U.S. Environmental Protection Agency. Research Triangle Park,
NC. EPA/600/8-91/049. http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=40179. 017649
U.S. EPA. (1996). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research Triangle
Park, NC. EPA/600/P-95/001aF-cF. 079380
U.S. EPA. (1997). Mercury Study Report to Congress. U.S. Environmental Protection Agency. Washington, DC. 157066
U.S. EPA. (2000). National air pollutant emission trends, 1900-1998. U.S. Environmental Protection Agency. Washington,
DC. 012211
U.S. EPA. (2001). Draft Guidance for tracking progress under the regional haze rule. U.S. Environmental Protection
Agency. Washington, DC. 157068
U.S. EPA. (2004). Air quality criteria for particulate matter. U.S. Environmental Protection Agency. Research Triangle
Park, NC. EPA/600/P-99/002aF-bF. 056905
U.S. EPA. (2004). The Particle Pollution Report: Current Understanding of Air Quality and Emissions through 2003. U.S.
Environmental Protection Agency. Research Triangle Park, NC. EPA454-R-04-002. 190219
U.S. EPA. (2005). Review of the national ambient air quality standards for particulate matter: Policy assessment of
scientific and technical information OAQPS staff paper. U.S. Environmental Protection Agency. Washington, DC.
EPA/452/R-05-005a. http://www.epa.gov/ttn/naaqs/standards/pm/data/pmstaffpaper_20051221 .pdf. 090209
U.S. EPA. (2006). Air quality criteria for lead, in 2 Volumes. Office of Health and Environmental Assessment,
Environmental Criteria and Assessment Office, Office of Research and Development, U.S. Environmental
Protection Agency. Research Triangle Park, NC. EPA-600/R-5/144aF-bF. 090110
U.S. EPA. (2008). Integrated science assessment for oxides of nitrogen and sulfur: Ecological criteria. U.S. Environmental
Protection Agency. Research Triangle Park, NC. EPA/600/R-08/082F. 157074
December 2009 9-237
-------
Vaisvalavicius R; Motuzas A; Prosycevas I; Levinskaite L; Zakarauskaite D; Grigaliuniene K; Butkus V. (2006). Effect of
heavy metals on microbial communities and enzymatic activity in soil column experiment. Arch Agron Soil Sci, 52:
161-169. 157080
Van Aardenne JA; Dentener FJ; Olivier JGJ; Klein Goldewijk CGM; Lelieveld J. (2001). A 1 "degree" x 1 "degree"
resolution data set of historical anthropogenic trace gas emissions for the period 1980-1990. Global Biogeochem
Cycles, 15: 909-928. 055564
Van de Hulst H. (1981). Light scattering by small particles. New York: Dover. 191972
Van Donkelaar A; Martin RV; Park RJ. (2006). Estimating ground-level PM2.5 using aerosol optical depth determined from
satellite remote sensing . J Geophys Res, 111: D21. 192108
van Heerden PDR; Krilger GHJ; Kilbourn Louw M. (2007). Dynamic responses of photosystem II in the Namib Desert
shrub, Zygophyllum prismatocarpum, during and after foliar deposition of limestone dust. Environ Pollut, 146: 34-
45. 156131
Vasundhara G; Jayashree G; Muraleedhara-Kurup G (2004). Sequestration of nickel and copper by Azotobacter
chroococcum SB1. Bull Environ Contam Toxicol, 72: 1122-1127. 156133
Vautard R; Maidi M; Menut L; Beekmann M; Colette A. (2007). Boundary layer photochemistry simulated with a two-
stream convection scheme. Atmos Environ, 41: 8275-8287. 106012
Veihelmann B; Levelt PF; Stammes P; Veefkind JP. (2007). Simulation study of the aerosol information content in OMI
spectral reflectance measurements. Atmos Chem Phys, 7: 3115-3127. 190627
Verdelhos T; Neto JM; Marques JC; Pardal MA. (2005). The effect of eutrophication abatement on the bivalve
Scrobicularia plana. Estuar Coast Shelf Sci, 63: 261-268. 190497
Viard B; Pihan F; Promeyrat S; Pihan J-C. (2004). Integrated assessment of heavy metal (Pb, Zn, Cd) highway pollution:
bioaccumulation in soil, graminaceae and land snails. Chemosphere, 55: 1349-1314. 055675
Viles HA; Gorbushina AA. (2003). Soiling and microbial colonisation on urban roadside limestone: A three year study in
Oxford, England. Build Environ, 38: 1217-1224. 156138
Viles HA; Taylor MP; Yates TJS; Massey SW. (2002). Soiling and decay of N.M.E.P limestone tablets. Sci Total Environ,
292:215-229. 156137
Villa S; Vighi M; Maggi V; Finizio A; Bolzacchini E. (2003). Historical trends of organochlorine pesticides in an Alpine
glacier. J Atmos Chem, 46: 295-311. 156139
Viskari E-L; Karenlampi L. (2000). Roadside Scots pine as an indicator of deicing salt use - a comparative study from two
consecutive winters. Water Air Soil Pollut, 122: 405-419. 019101
Vivas A; Barea JM; Biro B; Azcon R. (2006). Effectiveness of autochthonous bacterium and mycorrhizal fungus on
Trifolium growth, symbiotic development and soil enzymatic activities in Zn contaminated soil. J Appl Microbiol,
100: 587-598. 190500
Vivas A; Voros A; Biro B; Barea JM; Ruiz-Lozano JM; Azcon R. (2003). Beneficial effects of indigenous Cd-tolerant and
Cd-sensitive Glomus mosseae associated with a Cd-adapted strain of Brevibacillus sp in improving plant tolerance
to Cd contamination. Appl Soil Ecol, 24: 177-186. 190499
Vives I; Grimalt JO; Ventura M; Catalan J. (2005). Distribution of poly cyclic aromatic hydrocarbons in the food web of a
high mountain lake, Pyrenees, Catalonia, Spain. Environ Toxicol Chem, 24: 1344-1352. 157099
Vogel-Mikus K; Drobne D; Regvar M. (2005). Zn, Cd and Pb accumulation and arbuscular mycorrhizal colonisation of
penny cress Thlaspi praecox Wulf. (Brassicaceae) from the vicinity of a lead mine and smelter in Slovenia. Environ
Pollut, 133: 233-242. 190501
Vogel-Mikus K; Pongrac P; Kump P; Necemer M; Regvar M. (2006). Colonisation of a Zn, Cd and Pb hyperaccumulator
Thlaspi praecox Wulfen with indigenous arbuscular mycorrhizal fungal mixture induces changes in heavy metal
and nutrient uptake. Environ Pollut, 139: 362-371. 190502
Waggoner AP; Weiss RE. (1980). Comparison of fine particle mass concentration and light scattering extinction in ambient
aerosol. Atmos Environ, 14: 623-626. 070152
Waggoner AP; Weiss RE; AhlquistNC; Covert DS; Will S; CharlsonRJ. (1981). Optical characteristics of atmospheric
aerosols. Atmos Environ, 15: 1891-1909. 095453
December 2009 9-238
-------
Wang C. (2007). Impact of direct radiative forcing of black carbon aerosols on tropical convective precipitation. Geophys
Res Lett, 34: 5709. 156147
Wang Y-Z; Ingram JL; Walters DM; Rice AB; Santos JH; Van Houten B; Bonner JC. (2003). Vanadium-induced STAT-1
activation in lung myofibroblasts requires H2O2 and P38 MAP kinase. Free Radic Biol Med, 35: 845-855. 157106
Wania F; Hoff JT; Jia CQ; Mackay D. (1998). The effects of snow and ice on the environmental behaviour of hydrophobic
organic chemicals. Environ Pollut, 102: 43-51. 156148
Wania F; Mackay D. (1993). Global fractionation and cold condensation of low volatility organochlorine compounds in
polar regions. Ambio, 22: 10-18. 157110
Warner J. (1968). A reduction in rainfall associated with smoke from sugar-cane fires—an inadvertent weather
modification?. JAppl Meteorol, 7: 247-251. 157114
Warner J; Twomey S. (1967). The production of cloud nuclei by cane fires and the effect on cloud droplet concentration. J
Atmos Sci, 24: 704-706. 045616
Watmough SA; Hutchinson TC; Dillon PJ. (2004). Lead dynamics in the forest floor and mineral soil in south-central
Ontario. Biogeochemistry, 71: 43-68. 077809
Watson JG; Chow JC. (2007). Receptor models for source apportionment of suspended particles. In Introduction to
Environmental Forensics (pp. 273-298). San Diego, CA: Academic Press. 157127
Watson JG; Zhu T; Chow JC; Engelbrecht J; Fujita EM; Wilson WE. (2002). Receptor modeling application framework for
particle source apportionment. Chemosphere, 49: 1093-1136. 035623
Weissenhorn I; Leyval C; Berthelin J. (1995). Bioavailability of heavy metals and abundance of arbuscular mycorrhiza in a
soil polluted by atmospheric deposition from a smelter. Biol Fertil Soils, 19: 22-28. 073826
Welton E; Voss K; Quinn P; Flatau P; Markowicz K; Campbell J; Spinhirne J; Gordon H; Johnson J. (2002). Measurements
of aerosol vertical profiles and optical properties during INDOEX 1999 using micro-pulse lidars. J Geophys Res,
107: 8019. 190631
Welton EJ; Campbell JR; Spinhirne JD; Scott VS. (2001). Global monitoring of clouds and aerosols using a network of
micro-pulse lidar systems. In Singh, U.N. T. Itabe N. Sugimoto (Ed.),Lidar remote sensing for industry and
environment monitoring (pp. 151-158). Bellingham, WA: Society of Photo-Optical Instrumentation Engineers.
157133
Wen T; Wang Y; Chang SY; Liu G (2006). On-line measurement of water-soluble ions in ambient particles. Adv Atmos Sci,
23: 586-592. 179964
Wetzel MA; Stowe LL. (1999). Satellite-observed patterns in stratus microphysics, aerosol optical thickness, and shortwave
radiative forcing. J Geophys Res, 104: 31287-31299. 190636
Wielicki B; Barkstrom B; Harrison E; Lee R; Smith G; Cooper J. (1996). Clouds and the Earth's radiant energy system
(CERES): An Earth observing system experiment. Bull Am Meteorol Soc, 77: 853-868. 190637
Wild E; Dent J; Thomas GO; Jones KC. (2005). Direct observation of organic contaminant uptake, storage, and metabolism
within plant roots. Environ Sci Technol, 39: 3695-3702. 156156
Wild SR; Jones KC. (1992). Polynuclear aromatic hydrocarbons uptake by carrots grown in sludge amended soil. J Environ
Qual, 21:217-225. 156155
Williams KD; Jones A; Roberts DL; Senior CA; Woodage MJ. (2001). The response of the climate system to the indirect
effects of anthropogenic sulfate aerosols. Clim Dynam, 17: 846-856. 199898
Winker D; Couch R; McCormick M. (1996). An overview of LITE: NASA's Lidar In-Space Technology Experiment. IEEE
Micro, 84: 164-180. 190914
Winker DM; Pelon J; McCormick MP. (2003). The CALIPSO mission: spaceborne lidar for observation of aerosols and
clouds. PNAS, 4893: 1-11. 192017
Wolff GT. (1985). Characteristics and consequences of soot in the atmosphere. Environ Int, 11: 259-269. 044680
Wyszkowska J; Boros E; Kucharski J. (2007). Effect of interactions between nickel and other heavy metals on the soil
microbiological properties. Plant Soil Environ, 53: 544-552. 179948
December 2009 9-239
-------
Xu J; DuBois D; Pitchford M; Green M; Etyemezian V. (2006). Attribution of sulfate aerosols in Federal Class I areas of
the western United States based on trajectory regression analysis. Atmos Environ, 40: 3433-3447. 102706
Xue H; Feingold G. (2006). Large eddy simulations of tradewind cumuli: Investigation of aerosol indirect effects. J Atmos
Sci, 63: 1605-1622. 190920
Xue H; Feingold G; Stevens B. (2008). Aerosol effects on clouds, precipitation, and the organization of shallow cumulus
convection. J Atmos Sci, 65: 392- 406. 190921
Yang CY; Chen CJ. (2007). Air pollution and hospital admissions for chronic obstructive pulmonary disease in a
subtropical city: Taipei, Taiwan. J Toxicol Environ Health A, 70: 1214-1219. 092847
Yang F; Schlesinger M. (2001). Identification and separation of Mount Pinatubo and El Nino-Southern Oscillation land
surface temperature anomalies. J Geophys Res, 106: 14757-14770. 192270
Yang XE; Jin XF; Feng Y; Islam E. (2005). Molecular mechanisms and genetic basis of heavy metal
tolerance/hyperaccumulation in plants. J Integr Plant Biol, 47: 1025-1035. 192104
Yang Z; Zhu L. (2007). Performance of the partition-limited model on predicting ryegrass uptake of polycyclic aromatic
hydrocarbons. Chemosphere, 67: 402-409. 156168
Ying Q; Kleeman MJ. (2006). Source contributions to the regional distribution of secondary particulate matter in
California. Atmos Environ, 40: 736-752. 098359
Yogui G; Sericano J. (2008). Polybrominated diphenyl ether flame retardants in lichens and mosses from King George
Island, maritime Antarctica. Chemosphere, 73: 1589-1593. 189971
Yu H; Dickinson R; Chin M; Kaufman Y; Zhou M; Zhou L; Tian Y; Dubovik O; Holben B. (2004). The direct radiative
effect of aerosols as determined from a combination of MODIS retrievals and GOCART simulations. J Geophys
Res, 109: D03206. 190926
Yu H; Kaufman YJ; Chin M; Feingold G; Remer LA; Anderson TL; Balkanski Y; Bellouin N; Boucher O; Christopher S;
Decola P; Kahn R; Koch D; Loeb N; Reddy MS; Schulz M; Takemura T; Zhou M. (2006). A review of
measurement-based assessments of the aerosol direct radiative effect and forcing. Atmos Chem Phys, 6: 613-666.
156173
Yu H; Liu S; Dickinson R. (2002). Radiative effects of aerosols on the evolution of the atmospheric boundary layer. J
Geophys Res, 107: 4142. 190923
Yu IJ; Park JD; Park ES; Song KS; Han KT; Han JH; Chung YH; Choi BS; Chung KH; Cho MH. (2003). Manganese
distribution in brains of Sprague-Dawley rats after 60 days of stainless steel welding-fume exposure.
Neurotoxicology, 24: 777-785. 156171
Yu LE; Yung LL; Ong C; Tan Y; Balasubramaniam KS; Hartono D; Shui G; Wenk MR; Ong W. (2007). Translocation and
effects of gold nanoparticles after inhalation exposure in rats. Nanotoxicology, 1: 235-242. 093173
Yuangen Y; Campbell CD; Clark L; Camerson CM; Paterson E. (2006). Microbial indicators of heavy metal contamination
in urban and rural soils. Chemosphere, 63: 1942-1952. 156174
Yunker MB; Snowdon LR; MacDonald RW; Smith JN; Fowler MG; Skibo DN. (1996). Polycyclic aromatic hydrocarbon
composition and potential sources for sediment samples from the Beufort and Barents Seas. Environ Sci Technol,
30: 1310-1320. 156175
Zaccone C; Cocozza C; Cheburkin AK; Shotyk W; Miano TM. (2007). Highly Organic Soils as "Witnesses" of
Anthropogenic Pb, Cu, Zn, and 137Cs Inputs During Centuries. Water Air Soil Pollut, 186: 263-271. 179930
Zappia G; Sabbioni C; Riontino C; Gobbi G; Favoni O. (1998). Exposure tests of building materials in urban atmosphere.
Sci Total Environ, 224: 235-244. 012037
Zechmeister HG (1998). Annual growth of four pleurocarpous moss species and their applicability for biomonitoring
heavy metals. Environ Monit Assess, 52: 441-451. 156178
Zechmeister HG; Hohenwallner D; Riss A; Hanus-Illnar A. (2003). Variation in heavy metal concentrations in the moss
species Abietinella abietina (Hedw.) Fleisch. according to sampling time, within site variability and increase in
biomass. Sci Total Environ, 301: 55-65. 157175
Zeidner M; Shechter M. (1988). Psychological responses to air pollution: some personality and demographic correlates. J
Environ Psychol, 8: 191-208. 189973
December 2009 9-240
-------
Zhang J; Christopher S. (2003). Longwave radiative forcing of Saharan dust aerosols estimated from MODIS, MISR, and
CERES observations on Terra. Geophys Res Lett, 30: 2188. 190928
Zhang J; Christopher S; Remer L; Kaufman Y. (2005). Shortwave aerosol radiative forcing over cloud-free oceans from
Terra. I: Angular models for aerosols. J Geophys Res, 110: D10S23. 190929
Zhang J; Christopher S; Remer L; Kaufman Y. (2005). Shortwave aerosol radiative forcing over cloud-free oceans from
Terra. II: Seasonal and global distributions. J Geophys Res, 110: D10S24. 190930
Zhang J; Reid JS; Holben BN. (2005). An analysis of potential cloud artifacts in MODIS over ocean aerosol optical
thickness products. Geophys Res Lett, 32: L15803. 190931
Zhang J; Reid JS; Westphal DL; Baker NL; Hyer EJ. (2008). A system for operational aerosol optical depth data
assimilation over global oceans. J Geophys Res, 113: Dl0208. 190932
Zhang KM; Wexler AS; Niemeier DA; Zhu YF; Sioutas W; Sioutas C. (2005). Evolution of particle number distribution
near roadways Part III: traffic on-road size resolved particulate emission factors. Atmos Environ, 39: 4155-4166.
086743
Zhang Q; Jimenez JL; Canagaratna MR; Allan JD; Coe HL Ulbrich I; Alfarra MR; Takami A; Middlebrook AM; Sun YL;
Dzepina K; Dunlea E; Docherty K; DeCarlo PF; Salcedo D; Onasch T; Jayne JR; Miyoshi T; Shimono A;
Hatakeyama S; Takegawa N; Kondo Y; Schneider J; Drewnick F; Borrmann S; Weimer S; Demerjian K; Williams
P; Bower K; Bahreini R; Cottrell L; Griffin RJ; Rautiainen J; Sun JR; Zhang YM; Worsnop DR. (2007). Ubiquity
and dominance of oxygenated species in organic aerosols in anthropogenically-influenced Northern Hemisphere
midlatitudes. Geophys Res Lett, 34: LI3801.189998
Zhang Q; Jimenez JL; Canagaratna MR; Jayne JT; Worsnop DR. (2005). Time- and size-resolved chemical composition of
submicron particles in Pittsburgh: Implications for aerosol sources and processes. J Geophys Res, 110: 1-19.
157185
Zhang X; Zwiers FW; Stott PA. (2006). Multi-model multisignal climate change detection at regional scale. J Clim, 19:
4294-4307. 190933
Zhang XH; Zhu YG; Chen BD; Lin AJ; Smith SE; Smith FA. (2005). Arbuscular mycorrhizal fungi contribute to resistance
of upland rice to combined metal contamination of soil. J Plant Nutr, 28: 2065-2077. 192083
Zhang XZ; Sun HW; Zhang ZY (2006). [Bioaccumulation of titanium dioxide nanoparticles in carp]. Huanjing Kexue, 27:
1631-5. 157722
Zhao TXP; Laszlo I; Guo W; Heidinger A; Cao C; Jelenak A; Tarpley D; Sullivan J. (2008). Study of long-term trend in
aerosol optical thickness observed from operational AVHRR satellite instrument. J Geophys Res, 113: D07201.
190935
Zhao TXP; Yu H; Laszlo I; Chin M; Conant WC. (2008). Derivation of component aerosol direct radiative forcing at the
top of atmosphere for clear-sky oceans. J Quant Spectrosc Radiat Transf, 109: 1162-1186. 190936
Zhou M; Yu H; Dickinson RE; Dubovik O; Holben BN. (2005). A normalized description of the direct effect of key aerosol
types on solar radiation as estimated from aerosol robotic network aerosols and moderate resolution imaging
spectroradiometer albedos. J Geophys Res, 110: D19202. 156183
Zhu Y; Hinds WC; Shen S; Sioutas C. (2004). Seasonal trends of concentration and size distribution of ultrafine particles
near major highways in Los Angeles. Aerosol Sci Technol, 38: 5-13. 156184
Zimmer D; Baum C; Leinweber P; Hrynkiewicz K; Meissner R. (2009). Associated bacteria increase the phytoextraction of
cadmium and zinc from a metal-contaminated soil by mycorrhizal willows. Int J Phytoremediation, 11: 200-213.
192085
December 2009 9-241
-------
Annex A. Atmospheric Science
A.1. Ambient Air Particle Monitoring
A.1.1. Measurements and Analytical Specifications
Table A-1. Summary of integrated and continuous samplers included in the field comparison.
Abbreviation Instrument
Manufacturer/ Research Institute
INTEGRATED PARTICLE OR GAS/PARTICLE INSTRUMENTS
Dichot Dichotomous Sampler with Virtual Impactor
AND-241 Dichot Thermo Andersen Series 241 Dichotomous Sampler
AND-246 Dichot Thermo Andersen SA-246B Dichotomous Sampler
FRM~hlVOL1° Thermo Andersen GMW-1200 HiVol PM10FRM Sampler
ARA-PCM ARA Particle Composition Monitor
CMU CMU Speciation Sampler
DRI-SFS DRI Sequential Filter Sampler
HEADS (or HI) Harvard EPAAnnular Denuder System (or Harvard Impactor)
IMPROVE_SSb IMPROVE Speciation Sampler
URG-3000N" Modified IMPROVE Module C Sampler for Carbon
MASS-400b URG Mass Aerosol Speciation Sampler Model 400
MASS-450" URG Mass Aerosol Speciation Sampler Model 450
MiniVol Battery-Powered Portable Low-Volume Sampler
or cncc Particle Concentrator-Brigham Young University Organic Sampling
PC-BOSS Sys(em
Andersen Instruments (Smyrna, GA)
Andersen Instruments
Andersen Instruments
Andersen Instruments
Atmospheric Research and Analysis Inc. (Piano, TX)
Carnegie Mellon University (CMU), (Pittsburgh, PA)
Desert Research Institute (Reno, NV)
Harvard School of Public Health (Boston, MA)
URG Corp. (Chapel Hill, NC)
URG Corp.
URG Corp.
URG Corp.
Air Metrics Inc. (Eugene, OR)
Brigham Young University (Provo, UT)
SAMPLING SYSTEM
PQ-200 FRM BGI PQ-200 FRM Sampler
PQ-200 FRMA BGI PQ-200A FRM Audit Sampler
R&P-ACCU R&P-Automated Cartridge Collector Unit Sampler
R&P-2000 FRM R&P Partisol-2000 FRM Sampler
R&P-2000 FRMA R&P Partisol-2000 FRM Audit Sampler
R&P-2025 Dichotb R&P Partisol 2025 Dichotomous Sequential Air Sampler
R&P-2025 FRM R&P Partisol-Plus Model 2025 PM25 Sequential Samplers
R&P-2300b R&P Partisol 2300 Chemical Speciation Sampler
BGI Inc. (Waltham, MA)
BGI Inc.
Rupprecht & Patashnick, Co. (Albany, NY)
Rupprecht & Patashnick, Co.
Rupprecht & Patashnick, Co.
Rupprecht & Patashnick, Co.
Rupprecht & Patashnick, Co.
Rupprecht & Patashnick, Co.
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009
A-1
-------
Abbreviation Instrument
RAAS-1 00 FRM Thermo Andersen Reference Ambient Air Sampler Model 1 00
Manufacturer/ Research Institute
Andersen Instruments
FRM SAMPLER
RAAS-200 FRM Thermo Andersen RAAS Model 200 FRM Audit Sampler
RAAS-300 FRM Thermo Andersen RAAS Model 300 FRM Sampler
RAAS-400b Thermo Andersen RAAS Model 400 Speciation Sampler
SASSb MetOne Spiral Ambient Speciation Sampler
SCS PM2.5 Sequential Cyclone Sampler
URG-PCMb URG Particle Composition Monitor
VAPS URG Versatile Air Pollution Sampler
Andersen Instruments
Andersen Instruments
Andersen Instruments
Met One Instruments (Grants Pass, OR)
New York University (New York, NY)
URG Corp. (Chapel Hill, NC)
URG Corp.
CONTINUOUS MASS INSTRUMENTS
BAM B-Attenuation Monitor Model 1 020
nano-BAM Met One BAM Model 1 020 with 1 50 nm impactor
CAMM Continuous Ambient Mass Monitor
p, yg Real-Time Ambient Mass Sampler (modified Tapered Element
Oscillation Microbalance with diffusion denuder and Nation dryer)
TEOM Tapered Element Oscillating Microbalance
30 "C-TEOM TEOM operated at 30 °C
50 °C-TEOM TEOM operated at 50 °C
SES-TEOM TEOM 1 400a Series with Sample Equilibration System
D-TEOM Differential TEOM
FDMS-TEOM Filter Dynamics Measurement System TEOM
ACCU-TEOM TEOM 1 400 Series with an automated cartridge collection unit
Met One Instruments
Met One Instruments
Developed by Harvard School of Public Health, commercialized
by Thermo Andersen Instruments; now withdrawn from market
Brigham Young University
Rupprecht & Patashnick, Co.
Rupprecht & Patashnick, Co.
Rupprecht & Patashnick, Co.
Rupprecht & Patashnick, Co.
Rupprecht & Patashnick, Co.
Rupprecht & Patashnick, Co.
Rupprecht & Patashnick, Co.
CONTINUOUS PARTICLE LIGHT SCATTERING INSTRUMENTS
Dust Trak Dust Trak nephelometer
EcoTech EcoTech Model M9003 nephelometer
NGN NGN-2 nephelometer
RR-M903 Radiance Research Nephelometer Model M903
TSI Inc. (Shoreview, MN)
EcoTech Ry Ltd., Australia (American EcoTech, Warren, Rl)
Optec Inc. (Lowell, Ml)
Radiance Research Inc. (Seattle, WA)
CONTINUOUS ELEMENT INSTRUMENTS
Graphite Furnace Atomic Absorption Spectrometry— aerosol
collection as preconcentrate slurry
University of Maryland (College Park, MD)
SEAS
Semicontinuous Elements in Aerosol Sampler
University of Maryland
CONTINUOUS NITRATE INSTRUMENTS
ADI-N
Aerosol Dynamics Inc. Flash Volatilization Analyzer
Aerosol Dynamics Inc. (Berkeley, CA)
ARA-N
Atmospheric Research and Analysis N03-Analyzer
Atmospheric Research and Analysis Inc.
R&P-8400N
R&P-8400N Flash Volatilization Continuous N03- Analyzer
Rupprecht & Patashnick, Co.
CONTINUOUS SULFATE INSTRUMENTS
ADI-S
Aerosol Dynamics Inc. Flash Volatilization Analyzer
Aerosol Dynamics Inc.
Continuous Ambient Sulfate Monitor (prototype of the TE-5020 by
Thermo Electron [Franklin, MA])
Harvard School of Public Health
R&P-8400S
R&P-8400S Flash Volatilization Continuous S042~ Analyzer
Rupprecht & Patashnick, Co.
TE-5020
Thermo Electron Model 5020 S042" Particulate Analyzer
Thermo Electron Corp. (Franklin, MA)
December 2009
A-2
-------
Abbreviation
Instrument
Manufacturer/ Research Institute
CONTINUOUS MULTI-ION INSTRUMENTS
AIM
Ambient Ion Monitor Model 9000 (Cf,N02" ,N03",P043",
S042", NH4+,Na+,Mg2+'K+,Ca24)
URG Corp.
Dionex-IC
~,Bf'N03~, P043~, S042~,
Dionex CorP'
ECN
Energy Research Center of the Netherlands IC-based sampler (Cl, Energy Research Center of the Netherlands (Petten, the
N03", S042", NH4+ ,Na+, Mg2+,K+, Ca24) Netherlands
PILS-IC
'° (C'"' N°2" ' N°3"' P°4 "'
Geor9ia lnstitute °f Technology (Atlanta, GA)
Texas Tech IC-based sampler (N03, S04~)
Texas Tech University (Lubbock, TX)
CONTINUOUS CARBON INSTRUMENTS
OC and EC
ADI-C
ADI Flash Volatilization Carbon Analyzer
Aerosol Dynamics Inc.
RU-OGI
Rutgers University/Oregon Graduate Institute in-situ carbon analyzer Rutgers University (Camden, NJ)/Oregon Graduate Institute
(OC, EC) (Beaverton, OR)
R&P-5400
R&P-5400 continuous ambient carbon analyzer
Rupprecht & Patashnick, Co.
Sunset OCEC
Sunset Semi-Continuous Real-Time Carbon Aerosol Analysis
Instrument
Sunset Laboratory, Inc. (Tigard, OR)
EC
Aethalometer
Magee Scientific Co. (Berkeley, CA)
AE-16
Magee AE-16 aethalometer (BC)
Magee Scientific Co.
AE-20
Magee AE-20 dual wavelength aethalometer (BC)
Magee Scientific Co.
AE-21
Magee AE-21 dual-wavelength aethalometer (BC)
Magee Scientific Co.
AE-31
Magee AE-31 seven color aethalometer (BC)
Magee Scientific Co.
DRI-PA
DRI Photoacoustic Analyzer (BC)
Droplet Measurement Technologies, Inc. (Boulder, CO)
MAAP
Multi-Angle Absorption Photometer, Model 5012 (BC)
Thermo Scientific Corp. (Franklin, MA)
PSAP
Particle Soot Absorption Photometer (BC)
Radiance Research Inc. (Seattle, WA)
Other Carbon
PAS-PAH
Photo-lonization Monitor for PAHs (Model PAS 2000)
EcoChem Analytics (League City, TX)
PILS-WSOC
PILS-WSOC Analyzer, combination of PILS and total organic
analyzer (TOA)
Georgia Institute of Technology
PARTICLE SIZING INSTRUMENTS FOR MASS AND CHEMICAL SPECIATION
DRUM-3
Davis Rotating-Drum Uniform Size-Cut Monitor (0.1-2.5 urn in 3
stages)
University of California-Davis (Davis, CA)
DRUM-8
Davis Rotating-Drum Uniform Size-Cut Monitor (0.09- > 5.0 urn in 8 Universi(y Qf Ca|ifornia_Davis
ELPI
Electrical Low Pressure Impactor (0.007-10 urn in 12 stages)
Dekati (Tampere, Finland)
LPI
Low Pressure Impactor (0.03-10 urn in 13 stages)
Aerosol Dynamics, Inc.
MOUDI
Micro Orifice Uniform Deposit Impactor
MSP Corp. (Minneapolis, MN)
MOUDI-100
MOUDI Model 100 (0.18-18 urn in 8 stages)
MSP Corp.
MOUDI-110
MOUDI Model 110 (0.056-18 urn in 10 stages)
MSP Corp.
Nano-MOUDI Nano MOUDI (°-01 °-°-056 M™ in 3 stages coupled to MOUDI Model Msp c
PARTICLE NUMBER / VOLUME INSTRUMENTS
APS
Aerodynamic Particle Sizer
TSI Inc.
APS-3320
TSI Model 3320 (0.5-20 urn)
TSI Inc.
December 2009
A-3
-------
Abbreviation Instrument Manufacturer/Research Institute
APS-3321 TSI Model 3321 (0.5-20 um; replaced TSI Model 3320) TSI Inc.
DMA Differential Mobility Analyzer TSI Inc.
DMA-3081 TSI Model 3081 (0.01-1 .Oum) TSI Inc.
DMA-3085 TSI Model 3085 (0.002-0.15 um) TSI Inc.
EEPS Engine Exhaust Particle Sizer (EEPS 0.056-0.56 um) TSI Inc.
FMPS Fast Mobility Particle Sizer (FMPS 0.056-0.56 um) TSI Inc.
GRIMM-1108 Optical Particle Counter (OPC; 0.3-20 um) GRIMM Technologies, Inc. (Douglasville, GA)
SMPS Scanning Mobility Particle Sizer TSI Inc.
SMPS-3936 TSI Model 3936L (0.01-1 .Oum) TSI Inc.
Nano-SMPS-3936 TSI Model 3936N (0.002-0.15 um) TSI Inc.
SMPS and Condensation Nucleus Counter (0.005-0.35 or ^DIMMT hi- i
0.01 -0.875 um) GRIMM Technologies, Inc.
SMPS-custom DMA Model 3071 and CPC Model 3010 TSI Inc.
WPS Wide-Range Particle Spectrometer (0.01-10.0 um) MSP Corp.
SINGLE PARTICLE INSTRUMENTS
AMS Aerosol Mass Spectrometer (0.04-2 um) Aerodyne Research Inc. (Billerica, MA)
ATOFMS Aerosol Time of Flight Mass Spectrometer (0.3-2.5 um) TSI Inc.
CNC, CPC Condensation Nucleus Counters, Condensation Particle Counter Various vendors
One Aree(?OoTl0°S6 SpeCtr°meter C°nSiSting °f tW° SMPS 3nd Carnegie Mellon University
LIBS Laser-Induced Breakdown Spectroscopy (toKte^QuS^airadt)81'181 "*"'** ^^
PALMS Particle Analysis by Laser Mass Spectrometer (0.22-2.5 um) NOAA (Boulder, CO)
RSMS-II Rapid Single Particle Mass Spectrometer -II (0.035-1.1 um) University of Delaware (Newark, DE)
RSMS-III Rapid Single Particle Mass Spectrometer III (0.01-2.0 um) University of Delaware
LABORATORY INSTRUMENTS
DRI Model 2001 Thermal/Optical Carbon Analyzer (OC, EC, Eight
DRI Model 2001 Carbon Fractions with reflectance and transmittance laser Atmoslytic, Inc. (Calabasas, CA)
correction)
SEM Scanning Electron Microscopy Various vendors
3Now with Thermo Scientific, Franklin, MA.
bEPA-approved speciation sampler used in the Speciation Trends Network (STN).
cNow commercialized by Applikon Analytical, the Netherlands, and marketed under the name "MARGA" (Monitor for Aerosols and Gases in Ambient Air).
dNot available.
Source: Chowetal. (2008,156355)
December 2009 A-4
-------
Table A-2. Summary of PM2.6 and PMio FRM and FEM samplers.
Manufacturer' *™ff ?ize
Description
Desi9nation#
FRN
BGIInc.
BGI Inc.
PQ-100 PM10 Louvered PM10 inlet; operates at flow rate of 16.7 L/min; 24-h FRM
- integrated sampler; uses a mass flow meter to adjust to equiva-
RFPS19P8194 Vol. 63, p. 69625,
RFPS-1298-124 12/17/98^
PQ 200 PM lent volumetric flow at ambient temperature and pressure. FRM
1?8i:>n Vol. 63, p. 31991,
RFPS-0598-120 06/11/98
December 2009
A-5
-------
Manufacturer'
Description
™J,?r Designation #
FRN
Thermo Scientific,
Inc.
Thermo Scientific,
Inc.
Thermo Scientific,
Inc.
Thermo Scientific,
Inc.
URG Corp.
CAPS
RAAS 100-
VSCC
RAAS 200-
VSCC
RAAS 300-
VSCC
MASS-100
PM2.5
PM2.5
PM2.5
PM2.5
PM2.5
Model 605 Computer Assisted Particle Sampler (CAPS), 24-h
integrated. Not available commercially.
Same as RAAS-100 PM15 sampler but with BGI VSCC, instead
ofWINS impactor.
Same as RAAS-200 PM25 sampler but with BGI VSCC instead
ofWINS impactor.
Same as RAAS-300 PM15 sampler but with BGI VSCC instead
ofWINS impactor.
Model MASS100 PM25 sampler with louvered PM10 inlet
followed by WINS impactor, operates at 16.7 L/min;24-h
integrated, volumetric flow measured by dry test meter at pump
outlet modulates pump speed to maintain flow rate; single
channel.
FRM
FEM (II)
FEM (II)
FEM (II)
FRM
RFPS-1098-123
EQPM-0804-153
EQPM-0804-154
EQPM-0804-155
RFPS-0400-135
Vol. 63, p.
10/29/98
Vol. 69, p.
08/06/04
Vol. 69, p.
08/06/04
Vol. 69, p.
08/06/04
Vol. 65, p.
05/08/00
8036,
47924,
47924,
47925,
26603,
Model MASS300 PM25 sampler with louvered PM10 inlet fol-
URGCorp. MASS-300 PM15 lowed by WINS impactor, operates at 16.7 L/min; 24-h inte- FRM
grated, sequential sampler with circular tray holding six filters.
Tisch Environ-
mental, Inc.
TE-6070
HiVol
PMn
Model TE-6070 PM10 High-Volume Sampler, with TE-6001 PM1(
size selective inlet; 8" x 10" filter holder.
FRM
RFPS 0400 136 Vol. 65, p. 26603,
Ki-rsuiuu i jo 05/08/00
RFPS 0202 141 Vol. 67, p. 15566,
KrHb-U4^-141 04/02/02
Models BAM 1020, GBAM 1020, BAM 1020-1, and GBAM
Met One BAM PM10 1020-1, with BX-802 inlet; glass-fiber filter tape with 1-h filter
change frequency.
FEM
FOPM n?QR 199 Vol. 63, p. 41253,
EQPM-0798-122 08/03/98M
3 BGI Inc.: BGI Incorporated, Waltham, MA. R&P: Rupprecht & Patashnick Company, Inc., Albany, NY, now Thermo Scientific, Inc., Franklin, MA. Andersen: Graseby Andersen, later Andersen Instruments,
Inc., Smyrna, GA, now Thermo Scientific, Inc., Franklin, MA. Thermo Environmental Instruments, Inc., now Thermo Scientific, Inc., Franklin, MA. URG Corp.: URG Corporation, Chapel Hill, NC. Tisch
Environmental, Inc., Cleves, OH. Met One Instruments, Inc., Grants Pass, OR
b The efficiency of an inlet (Watson et al., 1983, 045084) is determined by its 50% cut-point (d50, the diameter at which half of the particles penetrate through the inlet, while the other half is retained by the
inlet, while the other half is retained by the inlet) and the geometric standard deviation (GSD, which is an indicator of the sharpness of the separation, and is derived by the square root of the ratio of particle
diameters at penetrations of 16% andc 84%, [d16/dB/5).
c FRM: Federal Reference Method; FEM: Federal Equivalent Method. Roman numeral within parenthesis indicates FEM class.
d Particle separation in WINS is achieved by means of a single-jet round nozzle with flow directed into an impaction reservoir. The impaction surface consists of a GelmanTypeA/E glass-fiber filter immersed
in 1 mL of Dow Corning (Midland, Ml) 704 diffusion pump oil housed in a reservoir.
Note: The geometric standard deviation (GSD, which is an indicator of the sharpness of the separation, and is derived by the square root of the ratio of particle diameters at penetrations of 16% and 84%,
[d16/d84]0.5).
Source: Chowetal. (2008,156355)
December 2009
A-6
-------
Table A-3.
Observable
PM2.5 mass
Elements
Nitrate
Sulfate
Ammonium
OC, EC.TC
Total mass of
WSOC
Elements in
water soluble
matter: C, H, N,
andS
Dissolved
organic
nitrogen
Measurement and analytical specifications for filter analysis of mass, elements, ions,
and carbon.
Analytical
Accuracy3
± 5% 4
± 2-5% 4
1 6% with spiked
concentrations on
Teflon4
and 1 1-1 4% on
nylon filters! 3
± 5% 4
± 5% 4
1 5% for TC and
OC. No standard
exists to determine
EC accuracy
DPI Model 2001
Carbon
Analyzer: + 5%
TOA: ± 3-7% 24'25
C:1.5%;
H:3%;
N:3%;
S:5%30
N/A
Preclslonb
±10%4
±10%4
±5 to 10% on repli-
cate analysis 4l13'
co-located
precision 1 5-7%14'16
±6to10%4'14'15
±10%4
OC: ± 20%
EC: ± 20%
TC: ± 10% 4
DPI Model 2001
Carbon
Analyzer: + 10%
Sunset Carbon
Analyzer: ± 3%26
TOA:±5-10%27
±2%30
± 5-30%31
Minimum
Detectable Limit
(MDL)
0.04 ug/m3 c to
~1ug/m3d5<6
XRF: 0.4-30 ng/m3 g
8PIXE:6-360ng/m3
d9ICP/MS: 0.004-25
ng/m3 1° 0.05-11. 7
ng/m39'11 AAS: 0.02-
7.15 ng/m312
0.06 ug/m3 e to
0.2 ug/m3 d1'6'17
0.06ug/m3;to
0.2ug/m3d1'6'13
0.06 ug/m3e to
0. 07 ug/m3 d1'6
OC:0.1 ug/m3 f to
0.8 ug/m3 d
EC : 0.03 ug/m3 d to
0.1 ug/m3'
TC:0.8ug/m3d1'6
DRI Model 2001
Carbon Analyzer:
0.1-0.23 ugC/m323
Sunset Carbon
Analyzer: 0.05-
0.22ugC/m326'28
Elemental High TOC
N:0.05ugC/m329
TOA:0.12 |jg
C/m326
C: 0.3 ug/m3
H:0.09 ug/m3
N: 0.03 ug/m3
S: 0.10 ug/m330
0.001 ug N/m3 while
inorganic nitrogen is
low;>0.071ugN/m3
while inorganic
nitrogen is high31
Interferences
Electrostatic charges need to
be neutralized before
measurement; positive (e.g.,
OC adsorption) and negative
artifacts (e.g., nitrate
volatilization)
Volatile compounds may evap-
orate from filters due to vac-
uum in XRF and PIXE.
Potential contamination during
extraction and incomplete
extraction efficiency for ICP-
MS and AAS. Matrix
interference and peak overlap
may occur on heavily loaded
samples.
Subject to volatilization from
Teflon or quartz-fiber filters
N/A
Subject to volatilization from
Teflon or quartz-fiber filters
Subject to adsorption (positive
artifact) and volatilization
(negative artifact) of organic
gases to and from quartz-fiber
• filters
Extraction efficiency and
volume reduction steps
Contamination during sample
drying step
Concentration of inorganic
nitrogen
Comparability
Within 20% 4
10 to 30% depending
on species 4
Within 35% and
probably greater4
Typically within 10%;
MOUDIs13to20%
lower than speciation
samplers4'1*9
Within 30% 4
OC: Within 20 to
50%
EC: Within 20 to
200%
TC: Within 20% 4'17'22
Within! 7% 26
N/A
Good correlation
between UV and
persulfate oxidation
methods (R2 = 0.87)
Data
Completeness
90 to 1 00% h6'7
90to100%h6J
85 to 1 00% "
l^00%
86 to 1 00% 6'7
86 to 1 00% 6'7
N/A
N/A
N/A
December 2009
A-7
-------
and polyether
Mono- and
Di-carboxylic N/A
acids
Amino acids N/A
Mass of humic-
like substances N/A
(HULIS)
^!VS Precisionb
GC/MS: 1 23% 33'34
Typically + 20%,
3. + 48o/32 ranged from + 10
='±4-°/0 to + 30%'32'3"6'37'38
HPLC/MS:
± 5-26%39
GC/MS: ±5-11% on
3 rerjlicates, + 8 % in
avg
IC:±10-15%45
± 9% 48
N/A
Minimum
Detectable Limit
(MDL)
GC/MS:
Levoo,lucosan:10
2. 08 ng/m3 j31
0.01 -0.03 ng/m333'41
HPLC/MS: 9-648
pg/m339
GC/MS: 0.04-1 .12
ng/m346
1C: 0.01 -0.1 2 ng/m3
1.65-23.6
pg/m3k48
0.083 ng/m3149
Interferences
GCMS: Extraction recovery
interfered by sample matrix
Derivatization efficiency
IC/PAD: Overlapping peaks in
chromatogram
GC/MS: Extraction recovery
interfered by sample matrix
Derivatization efficiency
1C: Overlapping peaks in
chromatogram
Derivatization efficiency
Stability of derivatives
Overlapping peaks in
chromatogram
Separation efficiency
Comparability
IC/PAD: Good
correlation
(R2 = 0.97) with
HPLC/MS; and
(R2 = 0.89) with
GC/MS Method 42
GC/MS: Within 50%
for less volatile
compounds 46
N/A
N/A
r it
omp
N/A
N/A
N/A
N/A
3 Accuracy is the ability of analytical methods to quantify the observable of a standard reference material correctly; it does not refer to measurement accuracy if no standards available.50
b Refers to precision of co-located measurements, unless specified otherwise.
c Based on 1 ug/filter limit of detection for 24-h samples, assuming a flow rate of 16.7 L/min
d Based on field blanks collected with FRM samplers; ug/filter converted to ug/m3 basis assuming a flow rate of 16.7 L/min for 24-h
s Based on '/? of a 47-mm filter extracted in 15 mLdeionized-distilled water (DDW) for 24-h samples, assuming a flow rate of 16.7 L/min
f Based on 0.2 ug/cm2 detection limit and 13.8 cm2 deposit area for a 47-mm filter, assuming a flow rate of 16.7 L/min for 24-h
9 Based on 24-h samples at a flow rate of 16.7 L/min and analyzed by XRF
h Except for samples from one FRM sampler at Atlanta Supersite, for which data recovery was 50%7; reason not reported.
Reported as uncertainty in literature
J Based on 24-h samples at a flow rate of 16.7 L/min
k Based on 13.8 cm2 deposit area for a 47-mm filter and extracted into a final volume of 200 uL, assuming a flow rate of 16.7 L/min for 24-h and molecular weight of amino acid = 150
Based on 13.8 cm2 deposit area for a 47-mm filter and extracted into a final volume of 200 uL, assuming a flow rate of 16.7 L/min for 24-h
N/A: Not available
'Chow (1995, 0770121:2Watson and Chow (2001,1571231:3 Watson et al. (1983, 0450841:4Fehsenfeld et al. (2004,1573601:5Solomon et al. (2001,1571931: BWatson et al. (2005,1571241:7Mikel (2001,
1567621: "Watson et al. (1999, 0209491:8Solomon and Sioutas (2006,1569951:10Graney et al. (2004, 0537561: "lanaka et al. (1998,1570411:12Pancras et al. (2005, 0981201: "John et al. (1988, 0459031:
"Hering and Cass (1999, 084958): 15Fitz et al. (1989, 077387): "Hering et al. (1988, 036012): "Solomon et al. (2003,156994): "tabada et al. (2004,148859): "Fine et al. (2003,155775): 20Hogrefe et al.
(2004, 0990031:21Drewnick et al. (2003, 0991601:22Watson et al. (2005,1571251:23Ho et al. (2006,1565521:24Decesari et al. (2005,1445361:25 Mayol-Bracero et al. (2002, 0450101:2BYang et al. (2003,
1561671:27 Tursic et al. (2006,1570631:2BMader et al.(2004,1567241:28Xiao and Liu (2004, 0568011:30Kiss et al. (2002,1566461:31Cornell and Jickells (1999,1563671:32 Zheng et al. (2002, 0261001:
33Fraser et al. (2002,140741): x Fraser et al. (2003, 042231) 35Schauer er al. (2000, 012225): 3BFine et al. (2004,141283): 37Yue et al. (2004,157169): 3BRinehart et al. (2006,115184): 38Wan and Yu (2006,
1571041:40Poore (2000, 0128391:41Fraser et al. (2003, 0402661:42Engling et al. (2006,1564221:43Yu et al. (2005,1571671:44Tran et al. (2000, 0130251:45Yao et al. (2004,1022131:4BLi and Yu (2005,
1566921: "Henning et al. (2003,1565391:
-------
Table A-4. Measurement and analytical specifications for filter analysis of organic species.
Organic Analytical Accuracy
TD Solvent
Extraction
PAHs ±2.8-24.1%51 Z-score
„ values 0 to -
±4.4-29.4% ig56
13.8-26.5%53 +4_8%32
±0.5-12.9%54 ±6.5.22%57
0.05-4.83%55
n-Alkanes N/A ± 4-8%32
Hopanes N/A N/A
Steranes N/A N/A
Organic acids N/A +4-8%32
(including n-
alkanoic
acids, n-
alkenoic
acids, alkane
dicarboxylic
acids,
aromatic
carboxylic
acids, resin
acids)
Precision
TD
Avg ± 3.2%,
ranged
from ±0.05 to
±11.5%55
Avg ± 3.2%,
ranged
from ± 0.05
to±11.5%55
Avg ±3.2%,
ranged
from + 0.05
to + 11.5%55
Avg + 3.2%,
ranged
from + 0.05
to + 11.5%55
±10 55
to + 29%55
Solvent
Extraction
Avg ±8%,
ranged from
±3.8
to±15%56
±23%56Avg
±2.6%,
ranged
from ±0.6 to
±9.5%57
typically
± 20%,
ranged
from± 10 to
± 30%c 32'35"37
±23%56
Typically ± 20
%, from ±10
to ± 30%c 32'35"
37
+ 23%56
Typically + 20
%, from ±10
t3o + 30%c32'35"
N/A
+ 24%41
+ 23 %56
Typically + 20
%, from + 10
* _L onn/ c 32,35-
to + 30%
MDL
TD
0.016-0.48
ng/m3358
0.030-0.45
ng/m3355
0.081-0.86
ng/m3358
ng^'M7
0.030-0.14
ng/m3355
0.018-0.063
ng/m3355
Mono-
carboxylic
acids (C8,
C12, and
C16):
0.79, 2.0, and
3. 2 ng/m3354
Solvent
Extraction
0.83-1.66
ng/m3b38
0.033-3.85
ng/m3b56
0.01-0.03
ng/m333'34'37
0.76-276
Pg/m3b57
0.01-0.03
ng/m333'34'37
0.83-1 .66
ng/m3b38
°-01-3033°431
ng/m333'41
0.01 ng/m337
0.83-1 .66
ng/m3b60
0.01-0.03
ng/m333'41
Interferences
TD
Fragmentatio
n of labile
com-
pounds
Same as
PAHs
Same as
PAHs
Same as
PAHs
Fragmentatio
n of labile
compounds.
Loss of polar
species due
to absorption
onto the
surface of the
injector.
Improper sta-
tionary phase
column used
during TD
analysis.
Incomplete
thermal
desorption of
analytes
because of
strong affinity
with filter
matrix.
Solvent
Extraction
Possible
contaminants
from solvents
and compli-
cated extrac-
tion
procedures.
Loss of
volatile
compounds
during the ex-
traction and
pretreatment
steps.
Possible
carryover
from injection
port.
Same as
PAHs
Same as
PAHs
Same as
PAHs
Possible
contaminants
from solvents
and com-
plicated
extraction
procedures.
Loss of
volatile
compounds
during the ex-
traction and
pretreatment
steps.
Possible
carryover
from injection
port .
Low
derivatization
efficiency .
Comparabil
ity
R2s for solvent
extraction
were 0.95 58,
0.9755. and
0.98 59
R2s for solvent
extraction are
0.94 58 and
0.98 55'59
R2s for solvent
extraction are
0.99 55 and
0.998 59
R2s for
solvent
extraction are
0.97 55 and
0.998 59
Correlation
with solvent
extraction
method
R2 = 0.731 59
December 2009
A-9
-------
Analytical Accuracy
Polyols and N/A ± 4-8%32
sugars,
including gua-
iacol and sub-
stituted
guaiacols, sy-
ringol and
substituted
syringols,
an hydro-
sugars
Precision
N/A ± 23%56 N/A
Typically ± 20
%, from ±10
t3o±30%c32'35"
MDL Interferences
Levoglucosa: Same as Same as N/A
10ng/m361 organic acids organic acids
2.08ng/m3b38
0.01-0.03
nn/m333'41
3 Assumes 2.9 cm filter used in analysis from a deposit area of 13.8 cm , and sample collection at a flow rate of 16.7 L/min for 24-h
bAssumes sample collection at a flow rate of 16.7 L/min for 24-h.
c Reported as uncertainty in literature.
dAssumes a final extract volume of 1 mL and sample collection at a flow rate of 16.7 L/min for 24-h.
N/A: Not available
'Chow (1995, 0770121:2Watson and Chow (2001,1571231:3 Watson et al. (1983, 0450841:4Fehsenfeld et al. (2004,1573601:5Solomon et al. (2001,1571931: BWatson et al. (2005,1571241:7Mikel (2001,
156762): "Watson et al. (1999, 020949): 9Solomon and Sioutas (2006,156995): '"Graney et al. (2004, 053756): "Tanaka et al. (1998,157041): 12Pancras et al. (2005, 098120): "John et al. (1988, 045903):
"Hering and Cass (1999, 0849581:15Fitz et al. (1989, 0773871:1BHering et al. (1988, 0360121: "Solomon et al. (2003,1569941:1BCabada et al. (2004,1488591: "Fine et al. (2003,1557751:20Hogrefe et al.
(2004, 0990031:21Drewnick et al. (2003, 0991601:22Watson et al. (2005,1571251:23Ho et al. (2006,1565521:24Decesari et al. (2005,1445361:25 Mayol-Bracero et al. (2002, 0450101:2BYang et al. (2003,
156167): 27Tursic et al. (2006,157063): 2BMader et al.(2004,156724): 29Xiao and Liu (2004, 056801): 30Kiss et al. (2002,156646): 31Cornell and Jickells (1999,156367):32 Zheng et al. (2002, 026100):
33Fraser et al. (2002,140741): x Fraser et al. (2003, 042231135Schauer er al. (2000, 0122251:3BFine et al. (2004,1412831:37Yue et al. (2004,1571691:3BRinehart et al. (2006,1151841:39Wan and Yu (2006,
1571041:40Poore (2000, 0128391:41Fraser et al. (2003, 0402661:42Engling et al. (2006,1564221:43Yu et al. (2005,1571671: ""Iran et al. (2000, 0130251:45Yao et al. (2004,1022131:4BLi and Yu (2005,
156692): "Henning et al. (2003,156539):
-------
Table A-5. Measurement and analytical specifications for continuous mass and mass surrogate
instruments.
Instrument and Averaging
Measurement Principle Time
Analytical Precisionb
Accuracy3
MDL Interferences Comparability Data
Completeness
INERTIA INSTRUMENTS
TEOM Air is drawn through a 10min-24h ±0.75%c
size-selective inlet onto the
filter mounted on an
oscillating hollow tube. The
oscillation frequency
changes with mass loading
on the filter, which is used to
calculate mass concentration
by calibrating measured
frequency with standards.
FDMS TEOM. A self-refer- 1 h-24 h ± 0.75%c
encing TEOM with a filter at
4 °C that accounts for
volatile species. It is
equipped with a diffusion
Nafion dryer to remove
particle-bound water. The
Teflon (PTFE)-coated boro-
silicate glass-fiber filter that
is maintained at 4 °C re-
moves particles during the
reference flow cycle. The
flow alternates between a
base and reference flow
every 6 min. If a negative
mass is measured during the
reference flow, due to loss of
volatiles from the filter, it is
added to the mass made
during the prior particle-
laden samples to obtain total
PM2.5 concentration.
Differential Tapered Element 1 h-24 h ±0.75%c
Oscillating Microbalance (D-
TEOM)
Similar to FDMS, but an
electrostatic precipitator is
used in place of the glass-
fiber filter to remove particles
during the 6 min reference
flow cycle.
RAMS.ATEOMwitha 10min-24h N/A
cyclone inlet, diffusion
denuders, and Nafion dryer.
Particles are collected on a
"sandwich" filter (Teflon fol-
lowed by carbon-
impregnated glass-fiber
filter) on the tapered
oscillating element. The
various denuders remove
gas phase organic com-
pounds, nitric acid, sulfur
dioxide, nitrogen dioxide,
ammonia, and ozone, which
could otherwise be adsorbed
by the TEOM filter.
± 5 ug/m for 0.01 ug, which is
10-minavgc,d 0.06 ug/m for 1-h
avgc
± 1 .5 ug/m for
1-havg^d
<10%65 0.01 ug, which is
0.06 ug/m3
for1-havgc
<1()%e 65,69,70 Q.01 Ug, Or
0.06 ug/m3 for 1-h
avgc
<10%f71 ± 1 to 2 ug/m3 for
30-min avg 72
Loses semi-
volatile species
at both 30°C
and 50°C.
SESTEOM,
while less
sensitive to
relative
humidity, does
not completely
eliminate loss
of semi-volatile
species
N/A
N/A
N/A
Underestimated 99%6587-92%6
FRM mass by
20 to 35% 62-<4
9 to 30% higher 95-99%65'68
than FRM mass 67
Within 10% of 57-65%
mass by D-
TEOM, PC-
BOSS, RAMS
and BAM 66'67
Within 10% of 86%65
FDMS-TEOM
10 to 20% N/A
higher than avg
72 FRM mass
December 2009
A-11
-------
Instrument and Averaging Analytical Precisionb
Measurement Principle Time Accuracy3
MDL Interferences Comparability Data
Completeness
PRESSURE DROP INSTRUMENT
Continuous Ambient Mass 1 h-24 h N/A
Monitor (CAMM)
Air is drawn through a Tef-
lon-membrane filter tape and
the pressure drop across the
filter is monitored
continuously. The proportion
of pressure drop to aerosol
loading is related to the PM
concentration. The filter tape
advances every 30-60 min to
minimize volatilization and
adsorption artifacts during
sampling.
28.1% for 1-h <5ug/m3for1h
avg avg
15.9% for 24-h
avg
(-3.5 ug/m )
Needs effective
sealing for
good
performance;
even slight
leaks may
result in highly
variable
baseline.
Probably less
sensitive than
DTEOM or
RAMS. 75'77
Varied perfor- N/A
ma nee: within
2%ofSES-
TEOM and FRM
at Houston, TX,
while not
correlated with
D-TEOM or
FRM at
Rubidoux,
CA.76'77
B-ATTENUATION INSTRUMENT
B Attenuation Monitor (BAM) 1 h-24 h
B rays electrons are passed
through a quartz-fiber filter
tape on which particles are
collected. The loss of
electrons (B attenuation)
caused by the particle
loading on the filter is
converted to mass
concentration, after subtrac-
tion of blank filter
attenuation.
± 3 ug for ±2 ug/m3 c'h
24-h avg
concentrations
< 100 ug/m3
and 2% for
100 to
1 ,000 ug/m3
±8 ug for 1-h
concentrations
< 1 00 ug/m3
and 8% for
100 to
1000 ug/m3
5 ug/m3 for 1-h avg1 Water Up to 30% 93-99%6'65'67
absorption by higher than FRM
particles may mass and within
result in higher 2%ofFDMS
mass measure- TEOM 63'67
ments; maybe
important at RH
>85%
LIGHT-SCATTERING INSTRUMENT
Nephelometers (including 5 min-24 h
DustTrak)
A light source illuminates the
sample air and the scattered
light is detected at an angle
(usually 90°) relative to the
source. The signal is related
to the concentration of the
particles giving an estimate
of the particle light-scattering
coefficient. Zero air calibra-
tions can be performed using
particle-free air.
N/A
Nephelometers:
<5%forTSI
andNGNi
nephelo meters
DustTrak:
Greater of 0.1%
oM ug/m3c'h
Nephelo meter:
< 1.5 Mm'1
Conversion fac-
tor to calculate
r, „ , . ,3 massconcen-
DustTrak: ± 1 ug/m (ration from
for 24-h avg1 bscat may vary
depending on
particle size,
shape and
composition.
Light scattering
by DustTrak
proportional to
dp 6 for dp
< 0.25 urn 79
Typically good
correlation with
SES-TEOM and
D-TEOM (R2
>0.80).
Comparability
depends on
conversion
factor used.
>80-98% for
NGN2, RR-M903
and GreenTek
Nephelometers 6
>80% for
DustTrak6
98% for GRIMM
optical particle
counter65
December 2009
A-12
-------
Instrument and Averaging Analytical Precisionb
Measurement Principle Time Accuracy3
MDL Interferences Comparability Data
Completeness
3 Accuracy is the ability of analytical methods to quantify the observable of a standard reference material correctly; does not refer to measurement accuracy, since no standards available.
b Refers to precision of co-located measurements, unless specified otherwise.
c Manufacturer-specified measurement parameter.
d Details not available on how the precision was obtained and whether it refers to co-located precision.
e Includes a combination of estimates: based on co-located precision and based on regression slopes.
f Co-located precision with respect to PC-BOSS reconstructed PM2 5 mass.
9 Using glass-fiber "sandwich" filter.
h Specified as "resolution" by the manufacturer.
'Co-located precision estimate based on regression slope for NGN nephelometer (slope = 1.01, intercept = -1.64 ug/m3, R2 = 0.99).
J Specified as "Zero stability" by the manufacturer.
N/A: Not available.
1Chow (1995, 0770121: 2Watson and Chow (2001 , 157123): 3 Watson et al. (1983, 0450841: Vehsenfeld et al. (2004, 1573601: 5Solomon et al. (2001 , 1571931: "Watson et al. (2005, 1571241: 7Mikel (2001 ,
1567621: BWatson et al. (1999, 0209491: 'Solomon and Sioutas (2006, 1569951: 10Graney et al. (2004, 0537561: "lanaka et al. (1998, 1570411: 12Pancras et al. (2005, 0981201: "John et al. (1988, 0459031:
"Hering and Cass (1999, 0849581: 15Fitz et al. (1989, 0773871: "Hering et al. (1988, 0360121: "Solomon et al. (2003, 156994): 1BCabada et al. (2004, 148859): "Fine et al. (2003, 1557751: 20Hogrefe et al.
(2004, 0990031: 21Drewnick et al. (2003, 0991601: 22Watson et al. (2005, 1571251: 23Ho et al. (2006, 1565521: 2
-------
Table A-6. Measurement and analytical specifications for continuous elements.
Precision
MDL
Data
Interferences Comparability Comp|eteness
Semi-continuous 15-30 min ±10%bforMn, 20-43%c80
Elements in Aerosol Fe, Ni, Cu, Zn, Se,
System (SEAS) Cd, and Sb
Particles are collected at + 20%b for Cr, As,
30-min interval for and Pb 80
subsequent laboratory
atomic absorption
analysis for elements.
Aerosol collection is
through condensational
growth by direct steam
injection. The grown
particles are separated
from the airstream using
virtual impactor. The
droplets accumulate in a
slurry that is pumped to
a separate sample vial
for each time period.
Laser-Induced A few N/A N/A
Breakdown seconds
Spectroscopy (LIBS)
Used for in-situ single
particle analysis. A high-
power pulsed laser is
projected into particles
producing high-
temperature plasma.
Photons emission from
relaxing atoms in the
excited states provides
characteristics of
individual elements.
Al:440pg Spectral N/A N/A
Cr: 6.7 pg interferences limit
Mn:9.9pg the number of
Fe: 85 pg elements detected
Ni: 42 pg simultaneously
Cu: 26 pg
Zn: 43 pg
As: 27 pg
Se: 33 pg
Cd: 3.2 pg
SbieOpg,,
Pb: 31 pg50
Na:143fg N/A N/A N/A
Mg:53fg
Al: 184 fg
Ca: 50 fg
Cr: 166fg
Mn: 176 fg
CuMSfg91
3 Accuracy is the ability of analytical methods to quantify the observable of a standard reference material correctly; does not refer to measurement accuracy, since no standards are available.
b Based on analysis of standard reference material (SRM) 1643d from National Institute of Standards and Technology (NIST).
c Based on error propagation.
N/A: Not available
(Kidwell and Ondov, 2004,155898)' (Lithgowet al., 2004,
Source: Chowetal. (2008,1563551
December 2009
A-14
-------
Table A-7. Measurement and analytical specifications for continuous N03-.
Instrument and Measurement Averaging Analytical D • •
Principle Time Accuracy3 Precision
MDL
Interferences Comparability
Data
Completion
FLASH VOLATIZATION INSTRUMENTS
Aerosol Dynamics Inc. continuous nitrate
analyzer (ADIN)
Particle collection by humidification and lnmin M/A
impaction followed by flash volatilization ' u mln N/rt
and detection of the evolved gases in a
chemiluminescent NOX analyzer.
N/A 0.1 ug/m3 for
N/A 10-minavg82
Within ?n% nf
filter and
N/A continuous N03". 93%7
See Weber et al.
82 for details
Rupprecht and Patashnick continuous
nitrate analyzer (R&P-8400N)
Particle collection by impaction followed
by flash volatilization and detection of
the evolved gases in a
chemiluminescent NOX analyzer. A
carbon honeycomb denuder, installed at
the inlet to the Nafion humidifier
removes nitric acid and ammonia vapor.
10 min
N/A
Conversion and
volatilization
efficiency appears
to depend on
ambient
0.24 ug/m to composition; extent
n/c'"•'•"'" of underestimation
increases with
higher
concentrations. '
0.17 to
0.3 ug/m3 for
24-h avg 83'84
0.45 ug/m for
10-minavg
20 to 45% lower
>80->94%6
85
DENUDER-DIFFERENCE INSTRUMENT
Atmospheric Research and Analysis
nitrate analyzer (ARAN)
Sampled air passes through a 350°C
molybdenum (Mo) mesh that converts
particulate nitrate into NO. A pre-split
stream with a Teflon filter installed
upstream of an identical converter
(i.e., particle-free air) is used as a
reference. NO in both streams is
quantified by chemiluminescence and
their difference determines the
particulate nitrate concentration. The
instrument inlet contains a potassium
iodide- coated denuder to remove HN03
and N02.
30s
N/A
N/A
0.5 ug/m3 for
30-s avg 82
N/A
Within 30% of
filter and
continuous N03~.
See Weber et al.
82 for details.
76%7
SAMPLE DISSOLUTION FOLLOWED BY 1C ANALYSIS INSTRUMENTS
Energy Research Center of the
Netherlands (ECN) IC-based ion
analyzer
Collects particles into water drops using
a steam jet aerosol collector, via
cyclone. The combined flow from
collected droplets containing dissolved
aerosol components and wall steam
condensate is directed to an anion 1C for
analysis of nitrate. Interfering gases are
pre-removed by a rotating wet annular
denuder system.
1h
N/A
N/A 0.1 ug/m3
N/A
Within 30% of
filter and
continuous N03~.
See Weber et al.
82 for details.
100%7
Texas Tech University (TT) ion analyzer
Particles in the sample stream are
processed through a cyclone and a
parallel plate wet denuder, then collected
alternatively on one of two 2.5 cm pre-
washed glass fiber filters for a period of
15 min. The particles on the freshly
sampled filter are automatically
extracted for 6.5 min with water and
analyzed for nitrate by 1C.
15-30 min
N/A
N/A
0.010 ug/m
N/A
Within 30% of
filter and
continuous N03~.
See Weber et al.
82 for details.
97%7
Particle into Liquid Sampler-Ion
Chromatography (PILS-IC)
Ambient particles are mixed with
saturated water vapor to produce
droplets collected by impaction. The
resulting liquid stream is analyzed with
an 1C to quantify aerosol ionic
components.
1 h
N/A
l/nr
Within110% of
Consistent water nylon-filter N03
quality is essential and 37% higher
for good precision, than R&P-8400N
65-70%"
December 2009
A-15
-------
Instrument and Measurement Averaging Analytical D • •
Principle Time Accuracy3 Precision
MDL
Interferences Comparability
Data
Completion
Dionex-IC
The gas-denuded air stream enters the
annular channel of a concentric nozzle,
where deionized water generates a
spray that entrains the particles. The
flow is then drawn through a 0.5 urn pore
size PTFE filter. The remaining solution
is aspirated by a peristaltic pump and
sent to 1C for ion analysis.
1h
N/A
14%"
N/A
Consistent water
quality is essential
for good precision.
Bias of < 10%
relative to filter
NOs'65
N/A
Ambient Ion Monitor (AIM; Model 9000)
Air is drawn through a size-selective inlet
into a liquid diffusion denuder where
interfering gases are removed. The
stream enters a supersaturation
chamber where the resulting droplets
are collected through impaction. The col-
lected particles and a fraction of the
condensed water are accumulated until
the particles can be injected into 1C for
hourly analysis.
1h
N/A
N/A
0.1 ug/m3 for
1-havge
N/A
N/A
N/A
PARTICLE MASS SPECTROMETER INSTRUMENT
Aerosol Mass Spectrometer (AMS)
Air stream is drawn through an
aerodynamic lens and focused into a
beam in a vacuum chamber. This
aerosol beam is chopped by a
mechanical chopper and the flight time
of the particles through a particle-sizing
chamber is determined by the time-
resolved mass spectrometer
measurement. The particle impacts onto
a 600 °C heated plate where it
decomposes and is analyzed by a
quadruple mass spectrometer. The
nitrate ion, along with other ions, is
detected by the mass spectrometer.
A few
seconds
N/A
N/A 0.03 ug/m3
Subject to
interferences from
fragments of other
species with mass
to charge ratio in
the same range as
fragments of nitrate.
Highly refractory
materials are not
detected.
Within 10% of
nylon-filter N03",
and within 15% of
PILS-IC and 30%
of R&P8400N 20
94-98%20
3 Accuracy is the ability of analytical methods to quantify the observable of a standard reference material correctly; does not refer to measurement accuracy, since no standards are available.
b Overall uncertainty estimated by error propagation.
c Uncertainty estimated from uncertainties in flow rates and calibrations; does not refer to co-located precision.
a Co-located precision with respect to PC-BOSS PM25 total particulate N03 (the sum of the denuded front filter [non-volatilized N03-] and HN03-absorbing backup filter [volatilized N03]).
s Manufacturer specified measurement parameter
N/A: Not available.
1Chow (1995, 077012): 2Watson and Chow (2001,1571231:3 Watson et al. (1983, 045084): 4Fehsenfeld et al. (2004,1573601:5Solomon et al. (2001,1571931: BWatson et al. (2005,1571241:7Mikel (2001,
156762): "Watson et al. (1999, 020949): 'Solomon and Sioutas (2006,156995): 10Graney et al. (2004, 053756): "Tanaka et al. (1998,157041): 12Pancras et al. (2005, 098120): "John et al. (1988, 045903):
"Hering and Cass (1999, 0849581:15Fitz et al. (1989, 0773871:1BHering et al. (1988, 0360121: "Solomon et al. (2003,1569941:1BCabada et al. (2004,1488591: "Fine et al. (2003,1557751:2DHogrefe et al.
(2004, 0990031:21Drewnick et al. (2003, 0991601:22Watson et al. (2005,1571251:23Ho et al. (2006,1565521:24Decesari et al. (2005,1445361:25 Mayol-Bracero et al. (2002, 0450101:26Yang et al. (2003,
156167): 27Tursic et al. (2006,157063): 2BMader et al.(2004,156724): 28Xiao and Liu (2004, 056801): 30Kiss et al. (2002,156646): 31Cornell and Jickells (1999,156367):32 Zheng et al. (2002, 026100):
33Fraser et al. (2002,1407411: x Fraser et al. (2003, 042231135Schauer er al. (2000, 0122251:3BFine et al. (2004,1412831:37Yue et al. (2004,1571691:3BRinehart et al. (2006,1151841:39Wan and Yu (2006,
1571041: "Poore (2000, 0128391:41Fraser et al. (2003, 0402661:42Engling et al. (2006,1564221:43Yu et al. (2005,1571671:44Tran et al. (2000, 0130251:45Yao et al. (2004,1022131:4BLi and Yu (2005,
156692): "Henning et al. (2003,156539): 4BZhang and Anastasio (2003,157182): "Emmenegger et al. (2007,156418):50 Watson et al, (1989,157119): 51Greaves et al. (1985,156494): 52Waterman et al.
(2000,1571161:53Waterman et al. (2001,1571171: aFalkovich and Rudich (2001,1564271:55Chow et al. (2007,1572091:5BMiguel et al. (2004,1232601:57Crimmins and Baker (2006, 0970081:5BHo and Yu
(2004,1565511:59Jeon et al. (2001, 0166361: BOMazzoleni et al. (2007, 0980381: B1Poore (2002, 0514441: B2Butler et al. (2003,1563131: B3Chow et al. (2006,1466221: MRussell et al. (2004, 0824531: B5Grover
et al. (2006,138080): BBGrover et al. (2005, 090044): B7Schwab et al. (2006, 098449): BBHauck et al. (2004,156525): B8Jaques et al. (2004,155878): 70Rupprecht and Patashnick (2003,157207): 71Pang et al
(2002, 0303531:72Eatough et al. (2001, 0103031:73Lee et al. (2005,128139); MLee et al. (2005,1566801:75Babich et al. (2000,1562391:7BLee et al. (2005,1559251:77Lee et al. (2005,1281391:7BAnderson
and Ogren (1998,1562131:79Chung et al. (2001,1563571: BOKidwell and Ondov (2004,1558981: B1Lithgow et al. (2004,1266161: B2Weber et al. (2003,1571291: B3Harrison et al. (2004,1367871: "Rattigan et
al. (2006,115897): B5Wttig et al. (2004,103413): BBVaughn et al. (2005,157089): B7Chow et al. (2005, 099030): BBWeber et al (2001, 024640); B8Schwab et al. (2006, 098785): ™Lim et al. (2003, 037037):
"Watson and Chow (2002, 0378731:82Venkatachari et al. (2006,1059181:83Bae et al. (2004,1562431: MArhami et al. (2006,1562241:85Park et al. (2005,1568431:8BBae et al. (2004, 0986801:87Chow et al.
(2006,1563501:8BArnott et al. (2005,1562271:88Bond et al. (1999,1562811:1DDVirkkula et al. (2005,1570971:1D1Petzold et al. (2002,1568631:102Park et al. (2006, 0981041:103Arnott et al. (1999, 0206501:
"Veters et al. (2001, 016925): 105Pitchford et al. (1997,156872): 10BRees et al. (2004, 097164): 107Watson et al. (2000, 010354): 10BLee et al. (2005,156680): 108Hering et al. (2004,155837): ™Watson et al.
(1998,1988051:111Chakrabarti et al. (2004,1574261:112Mathai et al. (1990,1567411: '"Kidwell and Ondov (2001.0170921: '"stanier et al. (2004, 0959551: '"Khlystovet al. (2005,1566351:11BTakahama et
al. (2004,1570381:117Chow et al. (2005,1563481: '"Zhang et al. (2002,1571811: '"Subramanian et al. (2004, 0812031:120Chow et al. (2006,1552071: ™Birch and Cary (1996, 0260041:122Birch (1998,
024953): 123Birch and Cary (1996, 002352): ™NIOSH (1996, 156810): 125NIOSH (1999, 156811): 12BChow et al. (1993, 077459): 127Chow et al. (2007, 156354): ™Ellis and Novakov (1982, 156416):
128Peterson and Richards (2002,1568611:130Schauer et al. (2003, 0370141: "iMiddlebrook et al. (2003, 0429321: "2Wenzel et al. (2003,1571391: "3Jimenez et al. (2003,1566111: "Vhares et al. (2003,
1568661:135Qin and Prather (2006,1568951: "BZhang et al. (2005,1571851:137Bein et al. (2005,1562651: "BDrewnick et al. (2004,1557541: "8Drewnick et al. (2004,1557551: "°Lake et al. (2003,1566691:
"'Lake etal. (2004, 088411)
Source: Chow etal. (2008,156355)
December 2009
A-16
-------
Table A-8. Measurement and analytical specifications for continuous S04
2-
Instrument and Measurement Principle
Averaging Analytical
Time Accuracy3
Precision MDL
Data
Interferences Comparability Comp|eteness
FLASH VOLATILIZATION INSTRUMENTS
Aerosol Dynamics, Inc. continuous sulfate
analyzer (ADIS) , . ,3
Particle collection by impaction followed by flash 10 min N/A N/A ^VS"" N/A
volatilization and detection of the evolved gases
by a UV-fluorescence S02 analyzer.
Within 15% of
filter and
continuous
S042" 100%7
See Weber etal.
for details.
Rupprecht and Patashnick continuous sulfate
analyzer (R&P-8400S)
Particle collection by impaction followed by flash
volatilization and detection of the evolved gases 10 min
by a UV-fluorescence S02 analyzer. An
activated carbon denuder at the inlet to the
Nafion humidifier removes S02.
N/A
9 ug/m3 0.48 ug/m3 efficiency
and >30% appears to
at cone. depend on
<2ug/m3b ambient
composition
10-30% lower
than filter S042
20,21,84
84_ gg0/0 6,20,21,84,85
THERMAL REDUCTION INSTRUMENTS
Continuous Ambient Sulfate Monitor (CASM)
Sampled air passes through a Na2C03 coated
annular denuder to remove ambient S02 and is
subsequently split into independent sample and
filter flows. The sample flow passes through a
quartz tube containing a stainless steel rod
maintained at 1000 °C that reduces sulfate to
S02. The flow then passes through a PTFE filter
and into a trace-level S02fluorescence
analyzer.
.icmin
lsmm
M/A
™rt
M/A
™rt
N/A
N/A
Up to 25% lower
than filter S042~
and within 6%
ofR&P8400S,
PILS-IC and
AMS 20<21
SAMPLE DISSOLUTION FOLLOWED BY 1C ANALYSIS INSTRUMENTS
an Qoo/ 20
8°-98/0
Thermo Electron Model 5020 sulfate particulate
analyzer (TE-5020) 15 . N/A
The commercial version of CASM, with slight lsmm N/rt
changes in the sample flow path.
0.3 ug/m3
<10%<89 3Vg89 ,
0 5 ug/m
for 15-min
avgd
S042 toS02
conversion
efficiency
depends on
ambient
composition 89
-20% lower
than filter S04
88-90%89
Energy Research Center of the Netherlands
(ECN) IC-based ion analyzer
Entrains particles into water drops using the
steam jet aerosol collector. The drops are
collected using a cyclone and the combined
flow from collected droplets containing dis-
solved aerosol components and wall steam
condensate is directed to an anion 1C for
analysis of sulfate. Interfering gases are pre-
removed by a rotating wet annular denuder
system.
1 h
N/A
N/A
N/A
N/A
Within 15% of
filter and
continuous
S042"
See Weber etal.
82 for details.
100%
Texas Tech University (TT) ion analyzer
Particles in the sample stream, after being
processed through a cyclone and a parallel
plate wet denuder, are collected alternatively on
one of two 2.5 cm pre-washed glass fiber filters 30 min N/A
for a period of 1 5 min. The particles on the
freshly sampled filter are automatically
extracted for 6.5 min with water and analyzed
for sulfate by 1C.
Particle into Liquid Sampler-Ion
Chromatography (PILS-IC)
Ambient particles are mixed with saturated
water vapor to produce droplets collected by 1 h N/A
impaction. The resulting liquid stream is
analyzed with an 1C to quantify aerosol ionic
components.
N/A N/A N/A
n x , Consistent
10%-15%e niR,,n/m3 water quality is
0.18 ug/m essen^a|fj
good precision.
Within 15% of
filter and
continuous ,
S042" 100%7
See Weber etal.
for details.
Within 30% of
filter and other fi/- ,„„ 20,21
continuous DO-/U/O
S042" 20'21
December 2009
A-17
-------
Instrument and Measurement Principle
Averaging Analytical
Time Accuracy3
Precision MDL
Data
Interferences Comparability Comp|eteness
Dionex-IC
The gas-denuded air stream enters the annular
channel of a concentric nozzle, where deionized
water generates a spray that entrains the
particles. The flow is then drawn through a 0.5-
um pore size PTFE filter. The remaining
solution is aspirated by a peristaltic pump and
sent to 1C for ion analysis.
Consistent
11%- N/A ^rquaNtyis W^1Q%of
good precision.
N/A
Ambient Ion Monitor (AIM; Model 9000)
Air is drawn through a size-selective inlet into a
liquid diffusion denuder where interfering gases
are removed. The stream enters a super
saturation chamber where the resulting droplets
are collected through impaction. The collected
particles and a fraction of the condensed water
are accumulated until the particles can be
injected into 1C for hourly analysis.
1h
N/A
N/A forlTavg N/A
N/A
N/A
PARTICLE MASS SPECTROMETER
Aerosol Mass Spectrometer (AMS)
Airstream is drawn through an aerodynamic
lens and focused into a beam in a vacuum
chamber. This aerosol beam is chopped by a
mechanical chopper and the flight time of the
particles through a particle-sizing chamber is
determined by the time-resolved mass
spectrometer measurement. The particle
impacts onto a 600 °C heated plate where it
decomposes and is analyzed by a quadruple
mass spectrometer. The sulfate ion, along with
other ions, is detected by the mass
spectrometer.
A few
seconds
N/A
N/A
N/A
Subject to
interferences
from fragments
of other species Up to 30% lower
with mass to than filter S042~
charge ratio in and within 5%
the same range of R&P8400S,
as fragments of PILS-IC and
sulfate. Highly CASM 20'21
refractory
materials are
not detected.
93-98%2
3 Accuracy is the ability of analytical methods to quantify the observable of a standard reference material correctly; does not refer to measurement accuracy, since no standards available.
b Overall uncertainty estimated by error propagation.
c Co-located precision estimate based on regression slope (slope = 0.95, intercept = 0.01-0.2, R >0.98).
d Manufacturer specified measurement parameter.
s Uncertainty estimated from uncertainties in flow rates and calibrations; does not refer to co-located precision.
' Co-located precision with respect to PC-BOSS PM25 S042~.
N/A: Not available
'Chow (1995, 0770121:2Watson and Chow (2001,1571231:3 Watson et al. (1983, 0450841:4Fehsenfeld et al. (2004,1573601:5Solomon et al. (2001,1571931: BWatson et al. (2005,1571241:7Mikel (2001,
1567621: BWatson et al. (1999, 0209491:8Solomon and Sioutas (2006,1569951:1DGraney et al. (2004, 0537561: "Tanaka et al. (1998,1570411:12Pancras et al. (2005, 0981201: "John et al. (1988, 0459031:
"Hering and Cass (1999, 084958): 15Fitz et al. (1989, 077387): 1BHering et al. (1988, 036012): "Solomon et al. (2003,156994): 1BCabada et al. (2004,148859): "Fine et al. (2003,155775): 20Hogrefe et al.
(2004, 0990031:21Drewnick et al. (2003, 0991601:22Watson et al. (2005,1571251:23Ho et al. (2006,1565521:24Decesari et al. (2005,1445361:25 Mayol-Bracero et al. (2002, 0450101:2BYang et al. (2003,
1561671:27Tursic et al. (2006,1570631:2BMader et al.(2004,1567241:28Xiao and Liu (2004, 0568011:30Kiss et al. (2002,1566461:31Cornell and Jickells (1999,1563671:32 Zheng et al. (2002, 0261001:
33Fraser et al. (2002,140741): x Fraser et al. (2003, 042231) 35Schauer er al. (2000, 012225): 3BFine et al. (2004,141283): 37Yue et al. (2004,157169): 3BRinehart et al. (2006,115184): 38Wan and Yu (2006,
1571041:40Poore (2000, 0128391:41Fraser et al. (2003, 0402661:42Engling et al. (2006,1564221:43Yu et al. (2005,1571671:44Tran et al. (2000, 0130251:45Yao et al. (2004,1022131:4BLi and Yu (2005,
1566921: "Henning et al. (2003,1565391:4BZhang and Anastasio (2003,1571821:48Emmenegger et al. (2007,1564181:50 Watson et al, (1989,1571191:51Greaves et al. (1985,1564941:52Waterman et al.
(2000,157116): 53Waterman et al. (2001,157117): aFalkovich and Rudich (2001,156427): 55Chow et al. (2007,157209): 5BMiguel et al. (2004,123260): 57Crimmins and Baker (2006, 097008): 5BHo and Yu
(2004,1565511:58Jeon et al. (2001, 0166361: BDMazzoleni et al. (2007, 0980381: B1Poore (2002, 0514441:B2Butler et al. (2003,1563131: B3Chow et al. (2006,1466221: "Russell et al. (2004, 0824531: B5Grover
et al. (2006,1380801: BBGrover et al. (2005, 0900441: B7Schwab et al. (2006, 0984491: BBHauck et al. (2004,1565251: B8Jaques et al. (2004,1558781:7DRupprecht and Patashnick (2003,1572071:71Pang et al
(2002, 030353): 72Eatough et al. (2001, 010303): 73Lee et al. (2005,128139); MLee et al. (2005,156680): 75Babich et al. (2000,156239): 7BLee et al. (2005,155925): 77Lee et al. (2005,128139): 7BAnderson
and Ogren (1998,1562131:78Chung et al. (2001,1563571: BOKidwell and Ondov (2004,1558981: B1Lithgow et al. (2004,1266161: B2Weber et al. (2003,1571291: B3Harrison et al. (2004,1367871: MRattigan et
al. (2006,1158971: B5Wttig et al. (2004,1034131: BBVaughn et al. (2005,1570891: B7Chow et al. (2005, 0990301: BBWeber et al (2001, 0246401; B8Schwab et al. (2006, 0987851:8DLim et al. (2003, 0370371:
81Watson and Chow (2002, 037873): 82Venkatachari et al. (2006,105918): 83Bae et al. (2004,156243): 84Arhami et al. (2006,156224): 85Park et al. (2005,156843): ™Bae et al. (2004, 098680): 87Chow et al.
(2006,1563501:8BArnott et al. (2005,1562271: "Bond et al. (1999,1562811: "°Virkkula et al. (2005,1570971:101Petzold et al. (2002,1568631: "2Park et al. (2006, 0981041:103Arnott et al. (1999, 0206501:
"Veters et al. (2001, 0169251:105Pitchford et al. (1997,1568721: "BRees et al. (2004, 0971641:107Watson et al. (2000, 0103541: "BLee et al. (2005,1566801: "8Hering et al. (2004,1558371: ™Watson et al.
(1998,198805): 111Chakrabarti et al. (2004,157426): 112Mathai et al. (1990,156741): '"Kidwell and Ondov (2001,017092): '"Stanier et al. (2004, 095955): '"Khlystovet al. (2005,156635): '"Takahama et
al. (2004,1570381:117Chow et al. (2005,1563481:11BZhang et al. (2002,1571811: '"Subramanian et al. (2004, 0812031:120Chow et al. (2006,1552071: "1 Birch and Cary (1996, 0260041:122Birch (1998,
0249531: 123Birch and Cary (1996, 0023521: ™NIOSH (1996, 1568101:125NIOSH (1999, 1568111:12BChow et al. (1993, 0774591: 127Chow et al. (2007, 1563541: 12BEllis and Novakov (1982, 1564161:
128Peterson and Richards (2002,156861): 130Schauer et al. (2003, 037014): "iMiddlebrook et al. (2003, 042932): "2Wenzel et al. (2003,157139): "3Jimenez et al. (2003,156611): "Vhares et al. (2003,
1568661:135Qin and Prather (2006,1568951: "BZhang et al. (2005,1571851:137Bein et al. (2005,1562651: "BDrewnick et al. (2004,1557541: "8Drewnick et al. (2004,1557551: MOLake et al. (2003,1566691:
"'Lake etal. (2004, 088411)
Source: Chow etal. (2008,1563551
December 2009
A-18
-------
Table A-9. Measurement and analytical specifications for ions other than N03" and S04
2-
Instrument and Measurement Averaging Analytical „ • •
Principle Time Accuracy3 Precislon
MDL
Data
Interferences Comparability Comp|eteness
SAMPLE DISSOLUTION FOLLOWED BY 1C ANALYSIS INSTRUMENTS
N02 by Particle into Liquid Sampler-Ion
Chromatography (PILS-IC)
Ambient particles are mixed with saturated
water vapor to produce droplets collected
by impaction. The resulting liquid stream is
analyzed with an 1C to quantity aerosol
ionic components.
Consistent water
1h
N/A 10%b88 0.14ug/m
N/A
N/A
good precision
NH4+ by Particle into Liquid Sampler-Ion
Chromatography (PILS-IC)
Ambient particles are mixed with saturated
water vapor to produce droplets collected
by impaction. The resulting liquid stream is
analyzed with an 1C to quantity aerosol
ionic components.
1h
0,5ug/m'
Consistent water
good precision
CI", Na+, K+, Ca" by Particle into Liquid
Sampler-Ion Chromatography (PILS-IC)
Ambient particles are mixed with saturated
water vapor to produce droplets collected
by impaction. The resulting liquid stream is
analyzed with an 1C to quantity aerosol
ionic components.
1h
Consistent water
N/A 10%b88 0.1ug/m388 ^jjjfo, N/A
good precision
N/A
Cf, NO,", N03", P04^' S042", NH4+, Na+'
Mg++, K , Ca" by Ambient Ion Monitor
(AIM; Model 9000)
Air is drawn through a size-selective inlet
into a liquid diffusion denuder where
interfering gases are removed. The stream
enters a super saturation chamber where
the resulting droplets are collected through
impaction. The collected particles and a
fraction of the condensed water are
accumulated until the particles can be
injected into 1C for hourly analysis.
1 h
N/A N/A
0.1 ug/m3for
1-havgd N/A
N/A
N/A
3 Accuracy is the ability of analytical methods to quantify the observable of a standard reference material correctly; does not refer to measurement accuracy, since no standards are available.
b Uncertainty estimated from uncertainties in flow rates and calibrations; does not refer to co-located precision.
c All-sampler avg appears to include a combination of 10 integrated and 3 continuous samplers, although specific details are missing7. Performance evaluations at sites dominated by semi-volatile ammonium
nitrate are needed.
d Manufacturer specified measurement parameter
1Chow (1995, 077012): 2Watson and Chow (2001,157123):3 Watson et al. (1983, 045084): 'Fehsenfeld et al. (2004,157360): 5Solomon et al. (2001,157193): "Watson et al. (2005,157124): 7Mikel (2001,
1567621: "Watson et al. (1999, 0209491:8Solomon and Sioutas (2006,1569951:1DGraney et al. (2004, 0537561: "lanaka et al. (1998,1570411:12Pancras et al. (2005, 0981201: "John et al. (1988, 0459031:
"Hering and Cass (1999, 0849581:15Fitz et al. (1989, 0773871:1BHering et al. (1988, 0360121: "Solomon et al. (2003,1569941:1BCabada et al. (2004,1488591: "Fine et al. (2003,1557751:20Hogrefe et al.
(2004, 099003): 21Drewnick et al. (2003, 099160): 22Watson et al. (2005,157125): 23Ho et al. (2006,156552): 24Decesari et al. (2005,144536):25 Mayol-Bracero et al. (2002, 045010): 2BYang et al. (2003,
1561671:27Tursic et al. (2006,1570631:2BMader et al.(2004,1567241:28Xiao and Liu (2004, 0568011:30Kiss et al. (2002,1566461:31Cornell and Jickells (1999,1563671:32 Zheng et al. (2002, 0261001:
33Fraser et al. (2002,1407411: x Fraser et al. (2003, 042231135Schauer er al. (2000, 0122251:3BFine et al. (2004,1412831:37Yue et al. (2004,1571691:3BRinehart et al. (2006,1151841:38Wan and Yu (2006,
157104): "Poore (2000, 012839): "Fraser et al. (2003, 040266): 42Engling et al. (2006,156422): 43Yu et al. (2005,157167): "Iran et al. (2000, 013025): <5Yao et al. (2004,102213): 4BLi and Yu (2005,
1566921:47Henning et al. (2003,1565391:4BZhang and Anastasio (2003,1571821:48Emmenegger et al. (2007,1564181:50 Watson et al, (1989,1571191:51Greaves et al. (1985,1564941:52Waterman et al.
(2000,1571161:53Waterman et al. (2001,1571171: aFalkovich and Rudich (2001,1564271:55Chow et al. (2007,1572091:5BMiguel et al. (2004,1232601:57Crimmins and Baker (2006, 0970081:5BHo and Yu
(2004,156551): 58Jeon et al. (2001, 016636): BOMazzoleni et al. (2007, 098038): B1Poore (2002, 051444):B2Butler et al. (2003,156313): B3Chow et al. (2006,146622): MRussell et al. (2004, 082453): B5Grover
et al. (2006,1380801: BBGrover et al. (2005, 0900441: B7Schwab et al. (2006, 0984491: BBHauck et al. (2004,1565251: B8Jaques et al. (2004,1558781:7DRupprecht and Patashnick (2003,1572071:71Pang et al
(2002, 0303531:72Eatough et al. (2001, 0103031:73Lee et al. (2005,128139); "Lee et al. (2005,1566801:75Babich et al. (2000,1562391:7BLee et al. (2005,1559251:77Lee et al. (2005,1281391:7BAnderson
and Ogren (1998,156213): 78Chung et al. (2001,156357): BOKidwell and Ondov (2004,155898): B1Lithgow et al. (2004,126616): B2Weber et al. (2003,157129): B3Harrison et al. (2004,136787): "Rattigan et
al. (2006,115897): B5Wttig et al. (2004,103413): BBVaughn et al. (2005,1570891: B7Chow et al. (2005, 0990301: BBWeber et al (2001, 0246401; B8Schwab et al. (2006, 0987851: ™Lim et al. (2003, 0370371:
81Watson and Chow (2002, 0378731:82Venkatachari et al. (2006,1059181:83Bae et al. (2004,1562431:84Arhami et al. (2006,1562241:85Park et al. (2005,1568431: ™Bae et al. (2004, 0986801:87Chow et al.
(2006,156350): 8BArnott et al. (2005,156227): 88Bond et al. (1999,156281): "°Virkkula et al. (2005,157097): 101Petzold et al. (2002,156863): "2Park et al. (2006, 098104): 103Arnott et al. (1999, 020650):
""Peters et al. (2001, 016925): 105Pitchford et al. (1997,1568721: "BRees et al. (2004, 0971641:107Watson et al. (2000, 0103541: "BLee et al. (2005,1566801: "8Hering et al. (2004,1558371: ™Watson et al.
(1998,1988051:111Chakrabarti et al. (2004,1574261:112Mathai et al. (1990,1567411: '"Kidwell and Ondov (2001.0170921: ™Stanier et al. (2004, 0959551: '"Khlystovet al. (2005,1566351:11BTakahama et
al. (2004,157038): 117Chow et al. (2005,156348): '"Zhang et al. (2002,157181): '"Subramanian et al. (2004, 081203): "°Chow et al. (2006,155207): "'Birch and Cary (1996, 026004): "2Birch (1998,
0249531: 123Birch and Cary (1996, 0023521: ™NIOSH (1996, 1568101:125NIOSH (1999, 1568111: "BChow et al. (1993, 0774591: "7Chow et al. (2007, 1563541: 12BEllis and Novakov (1982, 1564161:
"8Peterson and Richards (2002,1568611:130Schauer et al. (2003, 0370141: "iMiddlebrook et al. (2003, 0429321: "2Wenzel et al. (2003,1571391: "3Jimenez et al. (2003,1566111: "Vhares et al. (2003,
156866): 135Qin and Prather (2006,156895): "BZhang et al. (2005,157185): 137Bein et al. (2005,156265): "BDrewnick et al. (2004,155754): "8Drewnick et al. (2004,155755): ""Lake et al. (2003,156669):
"'Lake etal. (2004, 088411)
Source: Chow etal. (2008,156355)
December 2009
A-19
-------
Table A-10. Measurement and analytical specifications for continuous carbon.
Instrument and Measurement Principle AvT[mge'ng Accuracy' Precision
MDL
Data
Interferences Comparability Comp|eteness
PARTICLE COLLECTION ON IMPACTOR FOLLOWED BY FLASH VOLATILIZATION INSTRUMENT
Aerosol Dynamic Inc. continuous carbon
analyzer (ADI-C)
Particle collection by impaction followed by
flash oxidation and detection of the evolved 10min N/A
gases by a non-dispersive infrared C02
analyzer. OC is estimated as twice the
oxidizable carbon. EC is not quantified.
OC:
2 ug/m3
EC.TC:
N/A applicable, N/A
since it
measures
only OC 90
15-22% lower
OC than that by KV, ?
R&P-5400and OJ/0
RU-OGI
PARTICLE COLLECTION ON FILTER / IMPACTOR FOLLOWED BY HEATING/ANALYSIS INSTRUMENTS
Rupprecht and Patashnick 5400 continuous
ambient carbon analyzer (R&P-5400)
Particles collected on an impactor, which is
heated to 275 °C to 350 °C, then to 700 °C
after sample collection is complete. Evolved 1 h N/A
C02 is measured by an infrared detector. OC
is defined as the carbon measured at the
lower temperature, and EC is the remaining
carbon measured at the higher temperature.
Rutgers University-Oregon Graduate Institute
(RU-OGI) in-situ thermal/optical transmittance
carbon analyzer.
Air is sampled through a quartz-fiber filter for
1 h and then analyzed by heating through ,n • .,,„
different temperature steps to determine OC JU ™ IN/M
and EC. Sample flow is pre-split into two
identical systems that alternate every hour
between sampling and analysis mode to
achieve continuous measurements.
Sunset semi-continuous realtime carbon
aerosol analysis instrument (Sunset OCEC)
Particles collected on a quartz-fiber filter are
subject to heating temperature ramps following
the NIOSH 5040 TOT protocol and the 1h N/A
resulting C02 is analyzed by nondispersive
infrared (NDIR) detector to quantify OC and
EC. Instrument is alternated between sampling
and analytical mode.
OC:
0.5 ug/m
N/A O^ug/m3 N/A
TC:
0.5 ug/m390
OC:
0.3 ug/m
3%b'? 0E5:ug/m3 N/A
TC:
0.4 ug/m390
OC:10%C Ir^N/A
EC:20%C if0,. N/A .,,„
TC-10%C TC' 3 N/A
JsV07" 0.4 ug/m3
(1-havg)95
20 to 60% lower
TC than filter TC cc cno/ 6,91
by TOR or 56-60%
TOT.91'92
8% higher OC
and 20% lower aRO/ ?
ECthanR&P- 86/0
5400 90
Within 7 to 25%
of filter OC and
EC and within
15%forTC.
Wide variation Qn ano/ 6-95
due to different- 8°-89/0
ces in tempera-
ture and
analysis
protocols. 92'95'96
LIGHT ABSORPTION INSTRUMENTS
Aethalometer (AE-1 6, AE-21 , AE-31)
Attenuation of light transmitted through a
quartz-fiber filter tape that continuously sam- ,- , .,,„
pies aerosol is measured and converted to a
BC mass concentration using oabs of 14625/A
(m2/g).
Subject to multi-
ple scattering
effects by parti-
cle and filter
matrix resulting
5t BCe. in absorption
lore- 0.1 ug/m39' |;— •
rections have
been proposed
98 that can cor-
rect for such
effects.
Within ± 25% of
RU-OGI, Sunset 7CQno/6
and filter EC by 75-90/0
TOR/TOT.90'92
December 2009
A-20
-------
Instrument and Measurement Principle AvT[mge'ng Accuracy' Precision
MDL
Data
Interferences Comparability Comp|eteness
Particle Soot Absorption Photometer (PSAP)
Attenuation of light transmitted through a
glass-fiber filter that continuously samples 1 min N/A
aerosol is measured to quantify light
absorption (babs).
Multi-Angle Absorption Photometer (MAAP)
Light transmittance at 0° and reflectance from
a glass-fiber filter at 130° and 165° from the . • .,,„
illumination direction are used in a radiative
transfer model to estimate babs and is
converted to BC using oabs of 6.6 m2/g.
DRI Photoacoustic Analyzer (DRI-PA)
Light absorption by particles in air results in a
heating of the surrounding air. The expansion
of the heated air produces an acoustic (sound
wave) signal which is detected by a micro- 5 s N/A
phone to determine babs, which is converted
to BC using oabs = 5 m2/g for the 1047 nm
instrument and oabS = 10 m2/g for the 532 nm
in^tn impnt
II loll Ul 1 ICl II.
6 to BCf:
8%"<100 0.1 ug/m390
BCh:
0.05 ug/m3
(or
babs = 0.33
Mm for
10-min
12%9'101 avg)
0.02 ug/m
(or
babs = 0.13
Mm for
30-min
avg)1"1
BC1:
0.04 ug/m
(or
M/A bate - 0.4
N/A Mm'1 for
10-min
avg) at
532 nm103
Instrument in-
cludes an em-
pirical correction
for scattering
and loading ef-
fects 99 and
adjustments
have been pro-
posed for the
three wave-
length model
100
The instrument
is designed to
minimize mul-
tiple scattering
and loading ef-
fects by mea-
suring both
transmittance
and reflectance
and using a
two-stream
approximation
radiative
transfer model
to calculate babs.
At 532 nm,
absorbance by
N02 interferes
with that by
particles. Ac-
counted by
either removing
N02 from sam-
ple line using
denuders or by
doing a periodic
background
(particle-free
air) subtraction.
-50% lower
thanAE-16,RU-
OGI and R&P-
5400 EC.90
Within 18% of
filter EC by
IMPROVE TOR
(R2 = 0.96) and
up to 40%
higher than
Sunset EC. 102
Good correlation
(R2 >0.80), but
more than 40%
lower than
aethalo meter,
MAAP and filter
IMPROVE_TOR
EC. Suggests
need for a
different oabs. 102
N/A
N/A
N/A
PHOTO-IONIZATION INSTRUMENTS
Photoionization monitor for polycyclic aromatic
hydrocarbons (PAS-PAH) The air stream is
exposed to UV radiation, which ionizes the
particle-bound PAH molecules. The charged
particles are collected on a filter element and
the piezoelectric current is proportional to the
particle-bound PAH.
5 min
N/A
N/A
~3ng/mJ1'K N/A
N/A
>91%6
December 2009
A-21
-------
Instrument and Measurement Principle
>lng tSSSSf Precision MDL Interferences Comparability Com°ae?eness
3 Accuracy is the ability of analytical methods to quantify the observable of a standard reference material correctly; does not refer to measurement accuracy, since no standards are available.
b No specific details on how the precision was estimated; appears to be based on replicate analysis, may not represent overall co-located measurement precision
c Co-located precision estimates based on variation in avg ratios of replicate analysis using laboratory instrument and regression slopes (Slopes for OC = 1.01, EC = 0.82, TC = 0.94; R2 = 0.97-0.99) of co-
located field measurements.
d Estimated using co-located AE-21 and AE-31 BC measurements at Fresno, CA.97
e While the default manufacturer recommended conversion factor (or mass absorption efficiency, oabs) is 16.6 m2/g at 880 nm, Lim et al. (2003, 037037) assumed a value of 12.6 m2/g.
f Assuming a oabs of 10 m2/g.
9 Co-located precision estimate based on the variability of the avg ratio (0.99 ± 0.12).
hAssuming a oabs of6.5 m2/g.
Assuming a oabs of 10m2/g at 532 nm and 5 m2/g at 1047 nm.
J Specified by manufacturer as "lower threshold"; needs to be calibrated with site-specific PAH. Typically used as a relative measure in terms of electrical output in femtoamps.
k Manufacturer specified measurement parameter
N/A: Not available.
1Chow (1995, 077012): 2Watson and Chow (2001,157123):3 Watson et al. (1983, 045084): 4Fehsenfeld et al. (2004,157360): 5Solomon et al. (2001,157193): BWatson et al. (2005,157124): 7Mikel (2001,
156762): "Watson et al. (1999, 020949): 8Solomon and Sioutas (2006,156995): 10Graney et al. (2004, 0537561: "lanaka et al. (1998,1570411:12Pancras et al. (2005, 0981201: "John et al. (1988, 0459031:
"Hering and Cass (1999, 0849581:15Fitz et al. (1989, 0773871:1BHering et al. (1988, 0360121: "Solomon et al. (2003,1569941:1BCabada et al. (2004,1488591: "Fine et al. (2003,1557751:2DHogrefe et al.
(2004, 0990031:21Drewnick et al. (2003, 099160): 22Watson et al. (2005,157125): 23Ho et al. (2006,156552): 24Decesari et al. (2005,144536):25 Mayol-Bracero et al. (2002, 045010): 2BYang et al. (2003,
156167): 27Tursic et al. (2006,1570631:2BMader et al.(2004,1567241:28Xiao and Liu (2004, 0568011:30Kiss et al. (2002,1566461:31Cornell and Jickells (1999,1563671:32 Zheng et al. (2002, 0261001:
33Fraser et al. (2002,1407411: M Fraser et al. (2003, 042231135Schauer er al. (2000, 0122251:3BFine et al. (2004,1412831:37Yue et al. (2004,1571691:3BRinehart et al. (2006,1151841:38Wan and Yu (2006,
157104): "Poore (2000, 012839): 41Fraser et al. (2003, 040266): 42Engling et al. (2006,156422): 43Yu et al. (2005,157167): 44Tran et al. (2000, 013025): 45Yao et al. (2004,102213): 4BLi and Yu (2005,
156692): "Henning et al. (2003,156539): 4BZhang and Anastasio (2003,1571821: <8Emmenegger et al. (2007,1564181:50 Watson et al, (1989,1571191:51Greaves et al. (1985,1564941:52Waterman et al.
(2000,1571161:53Waterman et al. (2001,1571171: aFalkovich and Rudich (2001,1564271:55Chow et al. (2007,1572091:5BMiguel et al. (2004,1232601:57Crimmins and Baker (2006, 0970081:5BHo and Yu
(2004,1565511:58Jeon et al. (2001, 0166361: BOMazzoleni et al. (2007, 098038): B1Poore (2002, 051444): B2Butler et al. (2003,156313): B3Chow et al. (2006,146622): MRussell et al. (2004, 082453): B5Grover
et al. (2006,1380801: BBGrover et al. (2005, 0900441: B7Schwab et al. (2006, 0984491: BBHauck et al. (2004,1565251: B8Jaques et al. (2004,1558781:70Rupprecht and Patashnick (2003,1572071:71Pang et al
(2002, 0303531:72Eatough et al. (2001, 0103031:73Lee et al. (2005,128139); MLee et al. (2005,1566801:75Babich et al. (2000,1562391:7BLee et al. (2005,1559251:77Lee et al. (2005,1281391:7BAnderson
and Ogren (1998,156213): 78Chung et al. (2001,156357): BOKidwell and Ondov (2004,155898): B1Lithgow et al. (2004,126616): B2Weber et al. (2003,157129): B3Harrison et al. (2004,136787): "Rattigan et
al. (2006,1158971: B5Wittig et al. (2004,1034131: BBVaughn et al. (2005,1570891: B7Chow et al. (2005, 0990301: BBWeber et al (2001, 0246401; B8Schwab et al. (2006, 0987851:80Lim et al. (2003, 0370371:
81Watson and Chow (2002, 0378731:82Venkatachari et al. (2006,1059181:83Bae et al. (2004,1562431:84Arhami et al. (2006,1562241:85Park et al. (2005,1568431:8BBae et al. (2004, 0986801:87Chow et al.
(2006,1563501:8BArnott et al. (2005,156227): 88Bond et al. (1999,156281): 1DDVirkkula et al. (2005,157097): 1D1Petzold et al. (2002,156863): 102Park et al. (2006, 098104): 103Arnott et al. (1999, 020650):
""Peters et al. (2001, 0169251:105Pitchford et al. (1997,1568721:10BRees et al. (2004, 0971641:107Watson et al. (2000, 0103541: ™Lee et al. (2005,1566801:108Hering et al. (2004,1558371: ™Watson et al.
(1998,1988051: '"Chakrabarti et al. (2004,1574261:112Mathai et al. (1990,1567411: '"Kidwell and Ondov (2001.0170921: ™Stanier et al. (2004, 0959551: '"Khlystovet al. (2005,1566351:11BTakahama et
al. (2004,1570381:117Chow et al. (2005,156348): '"Zhang et al. (2002,157181): '"Subramanian et al. (2004, 081203): 120Chow et al. (2006,155207): ™Birch and Cary (1996, 026004): 122Birch (1998,
024953): 123Birch and Cary (1996, 0023521: ™NIOSH (1996, 1568101:125NIOSH (1999, 1568111:12BChow et al. (1993, 0774591: 127Chow et al. (2007, 1563541: ™Ellis and Novakov (1982, 1564161:
128Peterson and Richards (2002,1568611:130Schauer et al. (2003, 0370141: "iMiddlebrook et al. (2003, 0429321: "2Wenzel et al. (2003,1571391: "3Jimenez et al. (2003,1566111: "Vhares et al. (2003,
156866): 135Qin and Prather (2006,156895): "BZhang et al. (2005,157185): 137Bein et al. (2005,156265): "BDrewnick et al. (2004,155754): "8Drewnick et al. (2004,155755): "°Lake et al. (2003,156669):
"'Lake etal. (2004, 088411)
Source: Chow etal. (2008,156355)
December 2009
A-22
-------
Table A-11. Summary of mass measurement comparisons.
Site / Period / Sampler / Configuration
1 . Birmingham, AL (11/04/96 To 11/23/96)
2. Denver-Adams City, CO (12/11/96 To 1/7/97)
3. Bakersfield, CA (1/21/97 To 3/19/97)
4. Denver-Welby, Co (12/12/96 To 12/21/96)
5. Phoenix, AZ (12/06/96 To 12/21/96)
6. Azusa, CA (3/25/97 To 5/1 9/97)
7. Research Triangle Park (RTP), NC (1/17/97 To 8/14/97)
8. Rubidoux, CA (1/6/99 To 2/26/99)
9. Atlanta, GA (8/3/99 To 8/31/99)
Sampler Flow Rate (L/Min) Filter Type"
RAAS2.5-100 ,R7
PM2.5FRM ID-'
RAAS2.5-300 ,R7
PM2.5FRM ID-'
RAAS2.5-200 ,R7
PM2.5FRM ID-'
R&P Partisol 2000 1R7
PM2.5FRM ID-'
R&P Partisol-plus 1R7
2025PM2.5FRM ID''
BGIPQ200PM25 1R7
FRM lb''
Sierra Instruments 1R 7
SA-244Dichot ID''
IMPROVE PM2.5 22.8
Harvard PM2.5 .,,
Impactor
Airmetrics battery
powered PM2.5 5
MiniVol
Teflon (N/A)
Teflon (N/A)
Teflon (N/A)
Teflon (N/A)
Teflon (N/A)
Teflon (N/A)
Teflon (N/A)
Teflon (N/A)
Teflon (N/A)
Teflon (N/A)
Denuder
None
None
None
None
None
None
None
None
None
None
Atlanta Supersite, GA: 8/3/99 to 9/1/99
Four km NW of downtown, within 200 m of a bus maintenance yard and several
warehouse facilities, representative of a mixed commercial-residential neighborhood.
Flow
Sampler Rate
(L/Min)
R&P-2000FRM 16.7
RAAS-100FRM 16.7
RAAS-400 24
SASS 6.7
MASS-400 16.7
R&P-2300 10
R&P-2025 Dichot:
PM2.5 15
PM10-15 1.67
URG-PCM 16.7
ARA-PCM 16.7
PC-BOSS ,n,
(operated by TVA) lu&
Filter Type3
Teflon (P)
Teflon (P)
Teflon (P)
Teflon (P)
Teflon (P)
Teflon (P)
Teflon (P)
Polycarbonate
Teflon (P)
Teflon (N/A)
Teflon (W)
Denuder1'
None
None
None
None
Na2C03
None
None
None
Na2C03/Citric
Acid
Na2C03/Citric acid
GIF
Summary of Findings
Peters et al. (2001, 017108)104: Pitchford et al. (1997, 156872)105
dataset
Co-located precision (CV) for the RAAS2.5-100 samplers ranged
from 1 .5% at Bakersfield to 6.2% at Birmingham.
In Birmingham, CV for two co-located Harvard Impactor was 1% and
for three Dichots was 6.2%. The IMPROVE samplers had greater
variability, with a CV of 11 .3% (Denver-Adam City) and 1 0.8%
(Bakersfield).
Partisol and RAAS showed the strongest pairwise comparison
(slope = 1 .0 ± 0.06, intercept = 0.26 ± 1 .81 , and correlation = 1 .0),
within the EPA equivalency criteria. Strong relationships (correlation
>0.96; slope = 0.9-1 .12, intercept < 3o) were observed for other
samplers in reference to the RAAS.
At Denver-Welby, 6 RAAS samplers were deployed (3 with and 3
without temperature compensation for flow control). The units with
temperature compensation had a positive bias relative to the non-
temperature compensated units.
Non-FRM samplers did not meet the EPA equivalency criteria,
despite strong linear relationships with the FRM sampler.
Peters et al. (2001 , 0169261104: RTP 97 dataset
CVwas 1 .7%, 2.3%, 3.4%, 6.4% for the PQ200, Partisol 2000,
RAAS2.5100, and Dichot, respectively. Dichot flows were valve
controlled and set visually by the operator using rotameters.
Good one-to-one correspondence was observed for FRM
comparisons. The FRM averages were within -1 .2% to 3.2%, within
the acceptable ±10% range
Peters etal. (2001, 016925)104: Rubidoux 99 and Atlanta 99
dataset
In Rubidoux, the precision for PQ200 was 6.1 %, higher than at RTP
97. In Atlanta, the grouped data from PQ200, RAAS2.5-300, and
Partisol yielded a precision of 1.7%.
Linear regression results met the EPA equivalency criteria for all
FRMs.
Solomon et al. (2003, 166994) 1T
PM2 5 mass from individual samplers was compared to all-sampler
avgs, called the filter relative reference (filter RR) value. Overall
agreements were within ± 20% of filter RR.
FRM samplers were within 3.5% of filter RR.
Avg mass measured by RAAS-400, SASS and URG-PCM were
within + 10% of filter RR. Avg mass measured by MASS-400, R&P-
2300 and R&P-2025 dichot were greater than filter RR but
within ± 20%. Avg mass measured by PC-BOSS (BYU) and ARA-
PCM were lower than filter RR within ±10%.
All samplers except PC-BOSS (TVA) had R2 >0.80, relative to filter
RR.
While avg mass for each sampler was within 20% , daily variability
was >50% of filter RR.
Glycerol in the Na2C03 denuder may have contaminated the filter in
the MASS-400 sampler resulting in higher PM2.5 values.
PC-BOSS samplers removed particles < 0.1 urn aerodynamic
diameter from PM2.5 measurements. Corrections were made using
sulfate (S042~) concentrations in the major flow or immediately after
the PM2 5 inlet, but before the flow split-up. This was insufficient to
bring PC-BOSS mass close to filter RR. PC-BOSS was also
equipped with upstream denuders ahead of the filters, which may
have enhanced loss of semi-volatile components, resulting in a lower
mass on the filter.
December 2009
A-23
-------
Site / Period / Sampler / Configuration
PC-BOSS ,,-n
(operated by BYU) lou
PIVhs Continuous pf
SamP'er (UMin)
TEOM 16.7
Teflon (W)
Inlet Temperature
30 °C
GIF
Dryer Other
Nafion PM2.5
Atlanta Supersite, GA: 11/21/01 to 12/23/01
Flow
PIVh.s Sampler Rate
(L/Min)
R&P-2025FRM 16.7
PIVhs Continuous pT
Sampler [»•„,
TEOM 16.7
SES-TEOM 16.7
CAMM 0.3
RAMS 16.7
Radiance .,,„
Research M903 ™rt
Radiance .,,„
Research M903 ™rt
PITTSBURGH SUPERSITE,
park on the top of a hill
Flow
Sampler Rate
(L/Min)
MOUDI-110 30
And-241 Dichot 16.7
R&P-2000PM25 1R7
FRM lb''
PM2.s Continuous p °w
Sampler [»•„,
SES-TEOM 16.7
DAASS N/A
Filter Type3
Teflon (N/A)
Inlet Temperature
30 °C
30 °C
N/A
30 °C
N/A
N/A
Denuded
None
Dryer Other
Nafion PM2.5
Nafion PM2.5
Nafion PM2.5
PM2.5
TEA & GIF
Nafion denuders With
particle
concentrator
Nafion bscat
None bscat
PA: 7/1/01 to 6/1/02 6 km east of downtown in a
Filter Type3
Teflon (P,d)
Teflon (P
Teflon (W)
Inlet Temperature
30 °C
30 °C
Denuder
None
None
None
Dryer Other
Nafion PM2.5
Nafion or D,,
None PM«
Summary of Findings
Butler etal. (2003, 156313)62
The sum of individual species accounted for ~78% of the RAAS-100
FRM PM25 mass concentration.
TEOM explained -82 to 92% of the species sum of RAAS with
R2=0.86.
Lee etal. (2005, 128139)73
RAMS PM2.5 adjusted using particle concentrator efficiency of 0.5.
Good correlation between SES-TEOM and Radiance Research
• M903s (R2 = 0.80), while medium correlation was found between
CAMM and Radiance Research M903 (R2 = 0.64) or RAMS and
Radiance Research M903 (R2 = 0.63).
CAMM = (0.75 ± 0.03) SES-TEOM + (2.51 ± 0.51); R2 = 0.78;
N = 196
RAMS = (0.85 ± 0.06) SES-TEOM + (5.34 ± 1 .04); R2 = 0.52; N = 96
RAMS = (0.91 ± 0.07) CAMM + (5.71 ± 1 .20); R2 = 0.43; N = 196
Semi-volatile material explains the difference between RAMS and
•SESTEOM.
CAMM = (0.75 ± 0.08) R&P-2025 FRM + (2.47 ± 1 .02); R2 = 0.76;
N = 31
RAMS = (0.97 ± 0.22) R&P-2025 FRM + (2.39 ± 3.42); R2 = 0.64;
N = 13
SES-TEOM = (1 .07 ± 0.05) R&P-2025 FRM + (-1 .34 ± 0.71);
R2=0.95;N = 26
CAMM vs. FRM yielded lower slopes (0.75) with high intercepts.
Cabada et al. (2004, 148869)18: Rees et al. (2004, 097164)106
• MOUDI PM10 = 0.80 Dichot PM10, R2 = 0.85
MOUDI PM2.5= 1 .03 Dichot PM2.5, R2 = 0.78
MOUDI PM25= 1.01 FRM PM25,R2 = 0.78
Dichot PM2 5 = 0.97 FRM PM2 5 + 0.02; R2 = 0.94
Good agreement for PM25FRM, Dichot, and MOUDI. Lower slope for
PMio suggests loss of coarse particles in the MOUDI sampler.
Ultrafine (< 100 nm) mass (PM0.io) measurements had high
uncertainties (-30%)
Ultrafine mass by MOUDI showed no correlation with ultrafine
volume (V0.10) by DAASS. Ratio of PM0.io/PM2.5 mass ratio showed
reasonable agreement with volume ratio (V0.10/V2.5, R2 = 0.55,
slope = 0.76). Bounce of large particles to smaller stages in MOUDI
was small, since mass ratio (PM0.io/PM2.5) did not exceed volume
ratio (V0.10/V2.5). Low correlation between ultrafine mass and
volume could be due to the ultrafine mass measurement uncertainty
or due to fundamental differences in the measurement methods
employed by MOUDI and DAASS. Ambient conditions and
characteristics of the aerosols (such as non-spherical shapes of fresh
particles) could also influence these estimates.
Rees et al. (2004, 0971641106
SES-TEOM PM2.5 = 1 .02 FRM PM2.5 + 0.65; R2 = 0.95
Volatilization did not affect SES-TEOM performance when PM2.5
mass >20-30 ug/m3. When ambient temperature was < -6 °C, and
when mass was low, SES-TEOM was lower (up to 50%) than FRM or
Dichot.
December 2009
A-24
-------
Site / Period / Sampler / Configuration
Summary of Findings
FRESNO SUPERSITE, CA and other CRPAQS sites; 12/2/99 to 2/3/01 . Some Chow et al. (2006, 146622)63
comparisons included data till 12/29/03 . Fresno Supersite was located 5.5 km _.. . . .. .. _.. . .... -.,. .
northeast of downtown in a mixed residential-commercial neighborhood. '» ^^^t/rllJlS Z^^TtS? °f
Sampler
RAAS-100PM25
FRM
RAAS-300PM25
FRM
R&P-2000PM25
FRM
R&P-2025PM25
FRM
RAAS-400 PM2.5
SASS PM2.5
And-246 Dichot
PM2.5
PM 10-15
DRI-SFS PM2.5
MiniVol PM2.5
MOUDI-100
And-hlVOLPM10
FRM
Flow
Rate
(L/Min)
16.7
16.7
16.7
16.7
24
6.7
15
1.67
113
5
30
1130
Filter Type3
Teflon (P)
Teflon (P)
Teflon (P)
Teflon (P)
Teflon (P)
Teflon (P)
Teflon (P)
Teflon (P)
Teflon (P)
Teflon (P)
FEPb Teflon (P)
Teflon (P)
Denuder
None
None
None
None
None
None
None
None
None
None
None
None
RAAS-300 FRM.
All the FRM samplers were within ±10% of each other.
All the tiller samplers were well correlated with each other (K*
>0.90).e
DRI-SFS (with HN03 denuder) and And-246 Dichot PM2.5were lower
(~5% and 7%, respectively, on avg) than FRM, possibly due to nitrate
(N03~ volatilization.
Poor correlation (R2) found between TEOM PM25 concentrations and
RAAS-100 FRM. TEOM PM25was lower than RAAS-100 FRM by
22%. Heating of TEOM inlet to 50 °C resulted in loss of semi-volatile
components such as ammonium nitrate (NH4N03) and possibly some
semi-volatile organic compounds.
TEOM PMm concentrations were 28% lower than the And-hlVOL10
FRM on avg, ranging from 13% in summer to 43% in winter.
TEOM was neither equivalente nor comparablee to the FRM sampler
for PM, 5 or PM10
BAM PM2 5 concentrations showed high correlation (R2 >0.90) with
the RAAS-100 and RAAS-300 FRM samplers, with slopes ranging
from 0 92 to 0 97 BAM PM2 5 was typically higher than FRM (1 7 to
30%) except at Bakersfield, CA, where it was 21% lower, suggesting
FRM concentration on avg (R >0.92).
Higher BAM measurements were attributed to water absorption by
hygroscopic particles. BAM PM2.5 and PMio deviations were larger for
Groveretal. (2006,138080)65
PC-BOSS PM25 = (0.88 ± 0.04) FDMS-TEOM + (6.7 ± 4.3); R2 = 95;
n = 29
PC-BOSS PM25 = (1.11 ± 0.07) D-TEOM + (7.5 ± 6.1); R2 = 0.90;
n = 29
TEOM50C PM25 = (0.80 ± 0.01) TEOm3OC + (1.1 ± 3.1); R2 = 0.91;
n = 507
TEOm3OC PM25 = (0.50 ±0.01) FDMS-TEOM -(1.7 ± 6.9); R2 = 0.68;
n = 516
Heated GRIMM PM concentrations were lower than FDMS-TEOM
and ambient temperature GRIMM, suggesting loss of semi-volatile
matter.
Data recovery was greater than 95% for all continuous instruments,
except for D-TEOM, which had 86% recovery.
Reasonable agreement was seen between FDMS-TEOM, D-TEOM,
BAM, and GRIMM PM25when semi-volatile matter was dominated by
NH4N03. However, the FDMS-TEOM was higher than the other
instruments during high concentration periods, associated with days
with a high fraction of semi-volatile organic compounds (SVOCs).
Possible differences in SVOCs may have contributed to the
differences between FDMS and other instruments.
Continuous Sampler
TEOM
BAM
Sampler
PC-BOSS PM2.5
Flow Rate (L/Min)
16.7
16.7
Flow Rate (L/Min)
150
Inlet
Temperature
50 °C
Ambient
Filter Type3
Teflon (W)
Dryer
None
None
Other
PM2.5andPM10
PM2.5andPM10
Denuder11
GIF
December 2009
A-25
-------
Site / Period / Sampler / Configuration
Continuous Sampler Flow Rate (L/Min) Temperature Dr^er
TEOM 16.7 50 °C None
TEOM 16.7 30 °C None
FDMSTEOM 16.7 30 °C Nafion
D-TEOM 16.7 30 °C Nafion
GRIMM1100 1.2 Ambient None
80 °C heater,
GRIMM1100 1.2 aerosol9 '" Heater
temperature
BAM 16.7 Ambient None
Other
PM2.5
PM2.5
PM15
PM2.5
bscat
bscat
PM2.5
HOUSTON SUPERSITE, TX; 1/1/00 to 2/28/02
The Houston Supersite included three sites located in southeast Texas including one on the grounds
of a municipal airport at the edge of a small community, one adjacent to the highly industrial ship
channel and one on the grounds of a middle school in a suburban community.
PM2.s Sampler Flow Rate (L/Min) Filter Type3
R&P-2025FRM 16.7 Teflon (N/A)
Continuous Sampler Flow Rate (L/Min) Temperature Or^er
TEOM 16.7 50 °C None
SES-TEOM 16.7 30 °C Nafion
CAMM 0.3 Ambient Nafion
RAMS 16.7 30 °C Nafion
Radiance Research m m Nafjon
Denuder
None
Other"
PM2.5
PM2.5
Aug-Sep '00
PM2.5
Aug-Sep '00
PM2.5
TEA & GIF
denuders; Aug-
Sep '00
Bscat Aug-Sep
'00
LOS ANGELES SUPERSITE, CA; 9/01 to 8/02
The Los Angeles Supersite consisted of multiple sampling locations in the South Coast Air Basin to
provide wide geographical and seasonal coverage, including urban "source" sites and downwind
"receptor" sites.
Sampler Flow Rate (L/Min) Filter Typea
R&P-2025 Dichot
PM2.5 15 Teflon (P)
PM ,0-2.5 16.7 N/A
MOUDI-110 30 Teflon (P
HEADS PM2.5 10 Teflon (N/A)
Continuous Sampler Flow Rate (L/Min) TemDerature Dryer
D-TEOM 16.7 30 °C Nafion
Nano-BAM
(BAM-1 020 with d50 16.7 Ambient None
148 ±10 nm inlet)
Denuder1'
None
None
None
NaHC03
Other
PM2.5
~150 nm cut-
point at 16.7
L/min
Summary of Findings
Russell et al. (2004, 082463)64; Lee et al. (2005,
166680)108
Good correlations between 24-h SES-TEOM PM25
. and R&P-2025 FRM mass.
CAMM = (0.93 + 0.03) RAMS + (3.14 + 0.74);
R2 = 0.81
SES-TEOM = (0.92 ± 0.03) RAMS + (1 .52 ± 0.77);
R2 = 0.80
SES-TEOM = (1 .01 ± 0.03) CAMM + (-1 .91 ± 0.79);
. R2 = 0.83
Correlation of Radiance Research M903 and SES-
TEOM was good (R = 0.95), while that of Radiance
Research M903 with CAMM or RAMS was poor (R2
-0.4).
RAMS >SES-TEOM at high temperature and low
RH (< 60%), suggesting loss of water and
particulate N03 from SES-TEOM.
CAMM = (1 .02 + 0.08) R&P-2025 + (1 .62 + 1 .35) ;
R2 = 0.89
RAMS = (1 .10 ± 0.08) R&P-2025 + (0.68 ± 1 .28);
R2 = 0.89
SES-TEOM = (1 .09 ± 0.07) R&P-2025 +
(0.21±1.27);R2 = 0.94
Integrated mass < Continuous PM2.5 mass.
Difference possibly related to loss of SVOCs and
N03 from integrated sampler
Jaques et al. (2004, 166878)69; Hering et al.
(2004, 155837)109
Dichot PM25 = 0.83 MOUDI + 1 .23; R2 = 0.83
. (n = 37)
Dichot PM2 5 showed higher N03~ loss than MOUDI,
consistent with anodized aluminum surfaces serving
as efficient denuders that remove volatilized N03~
2,110.
D-TEOM PM2 5 = 1 .1 8 MOUDI - 1 .28; R2 = 0.86
• (n = 20)
Over-estimation of D-TEOM may be due to particle
losses in the MOUDI.
PM2.5 by D-TEOM during ESP-off phase (net artifact
• effect) tracked well with the N03 concentrations.
N03~ vaporization from the TEOM was caused by
• the temperature of the TEOM filter (-30-50 °C)
rather than the pressure drop across the filter.
Vaporization from the TEOM had a time constant
between 1 0 and 1 00 min depending on ambient and
TEOM filter temperatures, the vapor pressure, and
December 2009
A-26
-------
Site / Period / Sampler / Configuration
Summary of Findings
SMPS-3936
0.3
Ambient
None
Number to
mass assu-
ming spherical
particles of 1.6
g/cc density
the extent ot vapor saturation upstream and
downstream of the TEOM filter. The mass measured
during 5-min periods (ESP-on and off cycle in D-
TEOM) provides an estimate of the dynamic
vaporization losses.
Chakrabarti et al. (2004,1674261111
Good agreement between MOUDI PM015 and Nano-
BAM PMois (MOUDI PM0.15 = 0.97 Nano-BAM
PMo.is + 0.60; R2 = 0.92; n = 24)
Nano-BAM captured peak PMo.is concentrations not
quantified by SMPS. Potential particle
agglomeration (with resulting high surface areas)
caused SMPS to include particles in the
accumulation- rather than ultrafine-mode, since
mobility diameter is a function of surface area.
RUBIDOUX, CA; 08/15/01 to 09/07/01, 07/01/03 to 07/31/03. Rubidoux is located in the eastern
section of the South Coast Air Basin (SoCAB) in the northwest corner of Riverside County, 78 km
downwind of the central Los Angeles metropolitan area and in the middle of the remaining agricultural
production area in SoCAB.
Sampler
Flow Rate (L/Min) Filter Type"
Denuded
PC-BOSS PM2
150
Teflon (W)
GIF
R&P-2025 PM2.5 FRM 16.7
Teflon (N/A)
None
Continuous Sampler Flow Rate (L/Min)
Inlet
Temperature
Dryer
Other
TEOM
16.7
50 °C
None
PM2.
FDMS-TEOM
16.7
30 °C
Nafion
PM,
D-TEOM
16.7
30 °C
Nafion
PM2
RAMS
16.7
30 °C
Nafion
PM2.5
Denuders used
CAMM
0.3
N/A
None
PM2
Radiance Research
M903
N/A
N/A
Nafion
bscat
Radiance Research
M903
N/A
N/A
None
bscat
Grover et al. (2005, 090044)'"' (2003
measurements):
D-TEOM = (0.98 ± 0.02) FDMS-TEOM +
- (-0.6 ± 5.3); R2=0.85; n = 426; excludes 38 data
points when FDMS-TEOM PM2 5 was higher than D-
- TEOM PM2.5 by -21 ug/m3.
- RAMS = (0.93 ± 0.02) FDMS-TEOM + (2.4 ± 8.2);
R2 = 0.81; n = 337
FDMS-TEOM = (0.96 ± 0.06) PC-BOSSconstructed
mass + (-0.3 ± 3.9); R2 = 0.90; n = 33
R&P-2025 FRM = (0.96 ± 0.06) FDMS-TEOM +
-(-9.3±3.9);R2 = 0.90;n = 29
- The R&P-2025 FRM PM2.5 was, on avg, -32% lower
than FDMSTEOM. Losses of NH4N03and organics
- can account for the difference.
TEOM @ 50 °C PM2 5 was consistently lower than
- FDMS-TEOM, DTEOM or RAMS and was, on avg,
- 50% lower than FDMS-TEOM. This difference is
" due to loss of semi-volatile NOj. and organics from
the heated TEOM.
- FDMS-TEOM and D-TEOM needed little attention
from site operators.
- Lee et al. (2005,155925)76 (2001 measurements)
D-TEOM PM2.5 and Radiance Research M903s light
scattering (with and without dryers) showed good
correlation.
D-TEOM = (3.69 ± 0.09) Radiance Research
M903no-dryer + (2.74 ± 0.89); R2 = 0.84; n = 299
D-TEOM = (3.79 ± 0.10) Radiance Research
M903dryed + (4.08 ± 0.84); R2 = 0.83; n = 312
Radiance Research M903no-dryer = (1.03 ± 0.01)
Radiance Research M903dryed + (0.34 ± 0.05);
R = 0.98; n = 513; absorbed water did not affect
relationship to PM25
CAMM and RAMS compared poorly (R2 = 0 to 0.25)
with D-TEOM, Radiance Research M903s and
among themselves.
RAMS correlated well with D-TEOM for PM2.5
>30 ug/m due to RAMS's efficient particle collection
of larger particle sizes (historically associated with
high mass loadings at this site) in the PM2.5 size
range.
D-TEOM PM2 5 correlated well with ADI-N sized N03
(R2 = 0.62) and OC by Sunset OCEC (R2 = 0.61)
suggesting that D-TEOM measured PM2.5 mass with
minimum loss of SVOCs. RAMS showed R2 of 0.20
(N03~) to 0.30 (OC), while CAMM showed no
correlation.
December 2009
A-27
-------
Site / Period / Sampler / Configuration
LINDON, UT; 01/29/03 to 02/12/03
Sampler
PC-BOSS PM2.5
CONTINUOUS
SAMPLER
TEOM
FDMS-TEOM
RAMS
Flow Rate (L/Min)
150
FLOW RATE (L/MIN)
16.7
16.7
16.7
Filter Type3
Teflon (W)
INLET
TEMPERATURE
30 °C
30 °C
30 °C
DRYER
None
Nafion
Nafion
Denuder"
GIF
OTHER
PM2.5
PM2.5
PM2.5 Denuder
used
PHILADELPHIA, PA; 07/02/01 to 08/01/01 At water treatment center in a grassy field surrounded by
mixed deciduous and pine trees on three sides and a river on the other. Within 0.5 km of Interstate I-
95 and within 30 km from downtown Philadelphia.
Sampler
Harvard Impactor PM2
Continuous Sampler
SES-TEOM
CAMM
RAMS
Radiance Research
M903
Radiance Research
M903
Flow Rate (L/Min)
.5 10
Flow Rate (L/Min)
16.7
0.3
16.7
N/A
N/A
Filter Type3
Teflon (N/A)
Inlet Temperature
35 °C
N/A
30 °C
N/A
N/A
Dryer
Nafion
Nafion
Nafion
Nafion
None
BALTIMORE SUPERSITE, MD; 05/17/01 to 06/11/01. Located near a freeway and
Sampler
RAAS-100PM2.5FRM
Continuous Sampler
SES-TEOM
CAMM
RAMS
Radiance Research
M903
Radiance Research
M903
Flow Rate (L/Min)
16.7
Flow Rate (L/Min)
16.7
0.3
16.7
N/A
N/A
Filter Type
Teflon
Inlet Temperature
35 °C
N/A
30 °C
N/A
N/A
Dryer
Nafion
Nafion
Nafion
Nafion
None
Denuder1'
N/A
Other
PM2.5
PM2.5
PM2.5TEA&
GIF denude rs
With particle
concentrator
bscat
bscat
bus yard.
Denuder
None
Other
PM2.5
PM2.5
PM15TEA&
CIFdenuders;
No particle
bscat
bscat
Summary of Findings
Grover et al. (2005, 090044)66
RAMS required regular maintenance.
RAMS = (0.92 ± 0.03) FDMS-TEOM + (1 .3 ± 3.9);
' R2 - 0.69; n- 332
- PC-BOSS constructed mass = (0.89 ± 0.21) FDMS-
TEOM + (1 .8 ± 2.8); R2 = 0.66; n = 11
- TEOM @ 30 °C PM2.5 was consistently lower than
FDMS-TEOM and the difference was consistent
with concentrations SVOCs and NH4N03 measured
by PC-BOSS.
Lee etal. (2005, 128139)73
Radiance Research M903dryer = (0.78 ± 0.01)
- Radiance Research M903no dryer + (0.30 + 0.03);
R2 = 0.95
Radiance Research M903s vs. CAMM, R2 = 0.78
Radiance Research M903s vs. RAMS, R2 = 0.63
Radiance Research M903s vs. SES-TEOM,
- R2 = 0.72
CAMM = (0 60 + 0 03) SES-TEOM + (2 0 + 0 42)'
R2 = 0.71; N = 185
RAMS = (0.71 ± 0.04) SES-TEOM + (2.51 ± 0.59);
R2 = 0.63; N = 185
RAMS = (0.93 ± 0.06) CAMM + (2.44 ± 0.68);
R2 = 0.55; N = 185
Both RAMS and CAMM under-measured ambient
PM2.5.
CAMM = (0.70 ± 0.06) HI + (0.16 ± 0.96); R2 = 0.87;
N = 22
SES-TEOM = (1 .0 ± 0.10) HI + (-0.68 ± 1 .74);
R2 = 0.89;N = 15
Lee etal. (2006, 1281 39)73
Radiance Research M903dryed = (0.65 ± 0.02)
- Radiance Research M903no dryer + (1 .80 ± 0.20);
R2 = 0.75, suggesting influence from particle-bound
water.
- High correlation (R2 = 0.75) between Radiance
Research M903s.
Poor correlation among the continuous instruments.
Radiance Research M903s did not follow PM2 5
concentrations measured by other continuous
instruments.
CAMM = (0.32 ± 0.07) SES-TEOM + (9.45 ±1.61);
R2 = 0.14; N = 120
- RAMS = (0.82 ± 0.10) SES-TEOM + (6.41 ± 2.09);
R2 = 0.38; N = 120
RAMS = (0.71 ± 0.12) CAMM + (11 .3 ± 2.23);
R2 = 0.21; N = 120
CAMM = (0.80 ±0.29) RAAS-100 FRM +
(-0.83±5.85);R2=0.60;N = 7
RAMS = (1 .05 ± 0.12) RAAS-100 FRM +
(4.80 ± 2.60) ; R2 = 0.90; N = 11
SES-TEOM = (0.86 ± 0.10) RAAS-100 FRM +
(2.96±1.99);R2 = 0.90;N = 10
December 2009
A-28
-------
Site / Period / Sampler / Configuration
SEATTLE, WA; 01/28/01 to 02/21/01
Urban area near major highway and interstate, 8 km southeast of downtown.
SAMPLER FLOW RATE (L/MIN) FILTER TYPE3
MASSPM2.5 16.7 Teflon (N/A)
Continuous Sampler Flow Rate (L/Min) Inlet Temperature Dryer
SES-TEOM 16.7 30 °C Nafion
CAMM 0.3 Ambient Nafion
RAMS 16.7 30 °C Nafion
Radiance Research m m Nafjon
Radiance Research m m None
DENUDER"
Na2C03
Other
PM2.5
PM2.5
PM2.5TEA&
GIF denude rs
bscat
bscat
Summary of Findings
Leeetal. (2006, 166680)108
Radiance Research M903dryed = 0.94 + 0.00
Radiance Research M903no dryer; R2 = 1 .0.
Correlation of Radiance Research M903 vs. SES-
TEOM R2~080 while that of Radiance Research
M903 with CAMM was R2 = 0.84 and with RAMS
CAMM ~ (1 07 + 0 05) RAMS + (1 03 + 0 55)'
R2 = 0.61
SES-TEOM = (0.95 ± 0.03) RAMS + (1 .24 ± 0.38);
R2 = 0.72
SES-TEOM = (0.87 ± 0.03) CAMM + (0.55 ± 0.37);
R2 = 0.74
SES-TEOM likely lost semi-volatile organic matter.
Continuous PM2 5 samplers were similar to filter
PM2.5 sampler. Number of samples was small (~7).
Some SES-TEOM mass values were less than
MASS filter values suggesting that loss of mass is
likely for a SES-TEOM at 30°C, particularly during
the cold season.
NEW YORK SUPERSITE, NY; 01/01/03 to 12/31/04
Urban site located at Queens College, NY, about 14 km west of Manhattan, within 2 km of freeways,
and within 12 km of international airports. A rural site was located at Pinnacle State Park surrounded
by golf course, picnic areas, undeveloped forest lands, and no major cities within 15 km.
Sampler
R&P-2025 PM2.5 FRM
R&P-2300 PM2.5
Continuous Sampler
TEOM
FDMS-TEOM
Flow Rate (L/Min)
16.7
16.7
Flow Rate (L/Min)
16.7
16.7
Filter Type3
Teflon (N/A)
Teflon (N/A)
Inlet Temperature
50 °C
30 °C
Dryer
None
Nafion
Denuded
None
None
Other
PM2.5
PM2.5
BAM
16.7
"smart" heater on @ RH >44%
PM2
Schwab et al. (2006,098449)"'
FDMS-TEOM had operational difficulties resulting in
low data capture (65% at urban site and 57% at
rural site).
BAM had data captures greater than 95% at both
sites.
Urban site:
BAM = (1.02 ± 0.02) FDMS-TEOM + 1.72;
R2 = 0.93; n = 244
FDMS-TEOM = (1.25 ± 0.02) FRM - (0.63 ± 0.26);
R2 = 0.95; n = 238
BAM = (1.28 ± 0.03) FRM + (1.27 ± 0.38);
R2 = 0.88; n = 320
Rural site:
FDMS-TEOM = (1.09 ± 0.02) FRM - (0.004 ± 0.18);
R2 = 0.95; n = 349
PM2.5 FDMS-TEOM >FRM >TEOM50°C, suggesting
that FRM captured a fraction, but not all, of the
volatile components. TEOM50°C volatilizes PM2.5,
particularly during winter.
December 2009
A-29
-------
Site / Period / Sampler / Configuration
Summary of Findings
3Filter Manufacturer in parentheses - W: Whatman, Clifton, NJ; P: Pall-Gelman, Ann Arbor, Ml; S: Schleicher & Schnell. Keene, NH; N/A: not available or not reported.
bNa2C03: Sodium carbonate; NaHC03: Sodium bicarbonate GIF: Charcoal Impregnated Filter; FEP: Fluorinated Ethylene Propylene copolymer; TEA: Triethanolamine; TSP: Total Suspended PM.
C37 mm filter.
d37 mm after-filter for stages smaller than 0.16 urn and 47-mm for higher stages.
Equivalence requires correlation coefficient (r) Ł 0.97, linear regression slope 1.0 ± 0.05 and an intercept 0 ± 1 ug/m3; Comparability requires r>0.9 and linear regression slope equal 1 within 3 standard errors
and intercept equal zero within 3 standard errors; Predictability requires r>0.9. 91,112
1Chow (1995, 077012): 2Watson and Chow (2001,157123):3 Watson et al. (1983, 045084): 'Fehsenfeld et al. (2004,157360): 5Solomon et al. (2001,157193): "Watson et al. (2005,157124): 7Mikel (2001,
1567621: BWatson et al. (1999, 0209491:9Solomon and Sioutas (2006,1569951:10Graney et al. (2004, 0537561: "Tanaka et al. (1998,1570411:12Pancras et al. (2005, 0981201: "John et al. (1988, 0459031:
"Hering and Cass (1999, 0849581:15Fitz et al. (1989, 0773871:1BHering et al. (1988, 0360121: "Solomon et al. (2003,1569941:1BCabada et al. (2004,1488591: "Fine et al. (2003,1557751:20Hogrefe et al.
(2004, 099003): 21Drewnick et al. (2003, 099160): 22Watson et al. (2005,157125): 23Ho et al. (2006,156552): 2
-------
Table A-12. Summary of element and liquid water content measurement comparisons.
SITE/PERIOD/SAMPLER
SUMMARY OF FINDINGS
College Park, MD; 11/18/1999 to 11/19/1999,11/22/1999
Adjacent to a parking lot in the University of Maryland
campus, influenced by motor vehicles, coal-fired power plants
and incinerators ~21 km southwest of site and regionally
transported material.
Concentrated Slurry/Graphite Furnace Atomic Absorption
Spectrometry (GFAAS) (collectively known as Semi-
Continuous Elements in Aerosol Sampler, SEAS)
Ambient air is pulled in at a flow rate of 170 L/min. Particles
are grown using steam injection to about 3 to 4 urn in
diameter, which are then concentrated and separated from
the air stream in the form of a slurry using impactors. The
slurry is collected in glass sample vials, which are
subsequently analyzed by GFAAS in the laboratory.
Kidwell and Ondov (2001, 017092: 2004,166898)
Overall collection efficiency (of the entire system) measured using latex particles was 40%
for particles initially 0.1 to 0.5 urn in diameter, increasing with size to 68% for particles 3 urn
in diameter. Major losses were in the virtual impactor major flow channel and in the
condensers.
Six elements were detected simultaneously, limited by spectral interference and the
minimum detectable limit (MDL). Twelve elements (Al, Cr, Mn, Fe, Ni, Cu,Zn,As, Se,Cd,
Sb, and Pb) were measured.
MDLs ranged from 3.2 picogram (pg = 10~12 gram) to 440 pg.
Comparison with NIST standards showed good agreement, except for Al, Grand Fe, due to
poor atomization. The method was valid for dissolved solutions, but not for large particles
(>10um).
Overall avg relative standard deviation (RSD) was 20 to 43% by error propagation, mainly
due to the collection and analytical efficiencies.
There were possible memory effects due to particle adhesion to impactor collection
surfaces.
Lower MDLs may be possible through redesign and introduction of a wash cycle between
samples. A2.5 urn inlet might improve analytical efficiency by removing coarse particles.
Pittsburgh Supersite, PA; 08/26/2002 to 09/02/2002
6 km east of downtown in a park on the top of a hill.
Laser Induced Breakdown Spectroscopy (LIBS)
Ambient air was concentrated using a PM2.s inlet and a virtual
impactor. The concentrated stream was transported through a
Teflon tube to the sample cell of the LIBS system. The sample
cell was excited using a Nd: YAG laser. The resulting plasma
was collected and focused into a spectrometer, generating
spectra characteristic of different elements.
Lithgow et al. (2004,126616)
Calibration was done by sampling particle-laden streams with known metal concentrations.
Good linear fits with correlation coefficients 0.97 to 0.99
Seven metals (Na, Mg, Al, Ca, Cr, Mn, and Cu) were analyzed.
The MDLs were in the order of femtograms (fg = 10~15 gram) per sample.
This system has the capability of identifying the components, quantifying them and also
giving a particle size distribution. Mass was underestimated because of missing small
particles.
Pittsburgh Supersite, PA; 07/01/2001 to 08/31/2001,
01/01/2002 to 07/01/2002.
6 km east of downtown in a park on the top of a hill.
Dry Ambient Aerosol Size Spectrometer (DAASS)
Measures the aerosol size distribution (using nano-SMPS,
SMPS and APS) alternatively, at ambient relative humidity
(RH) (ambient channel) and at low RH (18 ± 6%) (dry
channel). A comparison of the two size distributions provides
information on the water absorption and change in size due to
RH.
Stanier et al. (2004,096966): Khlystov et al. (2006,166636)
Measured water content ranging from less than 1 ug/m3 to 30 ug/m3, constituting < 5% to
100% of the dry aerosol mass.
Small differences between dry and ambient channels of the DAASS. Number concentrations
were within 5% of each other.
Additional sources of error are associated with temperature differences between measured
outdoor ambient temperature and the temperature at which the ambient measurement
channel was maintained. Although the measurement system was placed in a ventilated
enclosure, it was ~4 °C higher than ambient temperature during July 2001. During winter,
the system was maintained at a minimum temperature of 9 °C, while the outdoor
temperature dropped to -5 °C. This caused differences in RH sensed by the system in the
ambient channel versus the actual outdoor RH.
RH differences cause underestimation of the particle number at sizes < 200 nm and an
overestimation at sizes >200 nm. This causes the volume growth factor to be higher by 2 to
14%, with the highest bias occurring at high RH and low temperature (92% outside RH and -
5°C).
The difference in temperature might also lead to evaporation of semi-volatile components
such as NH4N03. For the winter period, it was estimated that, for the worst case, the volume
growth factor would be underestimated by about 10% for 60-90% RH.
Insufficient purging of dry air between the dry and ambient cycles (implying the need for
supplemental vacuum power during the vent stages) causes uncertainties in estimated
growth factors. Correction factors were between 0.97 and 1.03.
Water content estimated by DAASS can be used to evaluate the thermodynamic models.
For the Pittsburgh study, the models underestimated the water content by 37%.
Data from DAASS showed that the aerosol was wet even at ambient RH less than 30%.
Source: Chow et al. (2008,1563551
December 2009
A-31
-------
Table A-13. Summary of PM2.6 N03" measurement
comparisons.
SITE / PERIOD / SAMPLER / CON FIGURATION
ATLANTA SUPERSITE, GA: 8/3/99 to 9/1/99 Four km NW of downtown, within 200 m of a bus
maintenance yard and several warehouse facilities, representative of a mixed commercial-residential
neighborhood.
Sampler
R&P-2000 FRM
RAAS-400
SASS
MASS-400
MASS-450
R&P-2300
VAPS
URG-PCM
ARA-PCM
PC-BOSS (TVA)
PC-BOSS (BYU)
PC-BOSS (BYU)
MOUDI-100
Continuous Sampler
ADI-N
ARA-N
PILS-IC
ECN
TT
Flow Rate (L/Min)
16.7
24
6.7
16.7
16.7
10
15
16.7
16.7
105
150
150
30
Flow Rate (L/Min)
1
3
5
16.7
5
Filter Type3
Quartz (P)
Nylon (P)
Nylon (P)
Teflon (P)-Nylon (P
Quartz (P)
Nylon (P)
Polycarbonatec (front &
back-up)
Teflon (P)-Cellulose-fiber
(W
Teflon (N/A)-Nylon (N/A)
Teflon (W)-
Nylon (P)
Teflon (W)-
Nylon (P)
Quartz (P)-
CIF (S)
Teflon (N/A-
Quartz (N/A
Denuder1'
None
MgO
MgO
Na2C03
None
Na2C03
Na2C03
Na2C03
Na2C03/Citric acid
GIF
GIF
GIF
None
Denuder Analysis Method1'
Activated Carbon NOX
Potassium iodide
(Kl) and dual sodium NOX
chlorite (NaCI02)
Two URG annular
glass denuders in
series containing 1C
citric acid and
CaC03
Rotating annular wet ,c
denuder system
Wet parallel plate ,„
denuder
Chemiluminescence
Chemiluminescence
SUMMARY OF FINDINGS
Solomon et al.(2003, 166994) 1T
PM25 N03" from each sampler was compared to
reference (filter RR) value. Overall agreements
due to volatilized N03", differences in denuder
design and filter types, and low concentrations
(close to analytical uncertainty).
A small positive artifact (few tenths of ug/m3) might
be present when using Na2
C03 impregnated filters, due to possible collection
(and subsequent oxidation) of MONO and N02 on
carbonate-impregnated filters. In addition, glycerol
in Na2C03 coated denuders may contaminate the
filters downstream.
PM15 N03- R&P-2000 FRM and MOUDI-100
samplers are consistently lower than other
samplers.
Weber etal. (2003, 157129)82
Hourly PM2.5 N03-were compared to all-sampler
averages (continuous RR), similar to the approach
used for integrated filter samplers. Overall
agreements were within + 20-30% (or + 0.2 ug/m )
except for ARA-N.
Except for ARA-N, good correlations (R2= 0.70 to
0.90) were found during the second half of the
study. The poor performance of ARA-N was
probably due to an inefficient denuder (25-60%
efficient) resulting in high background.
Large discrepancies between continuous and filter
RR probably due to low ambient concentrations
(study avg = 0.5 ug/m3) near the detection limit
(-0.1 ug/m3, except for ARA-N, which had
0.5 ug/m3).
The ARA-N was within 13%, ADI-N, ECN and
PILS-IC within 1 8% and TT within 26% of filter RR
(all<0.2 ug/m3 difference).
Filter samples showed more variability (Relative
Standard Deviation, RSD = 22%) than continuous
measurements (RSD = 13%). This is probably due
to sampling artifacts in filter samples; N03-
be minimal due to shorter averaging times and
rapid stabilization in solutions.
December 2009
A-32
-------
SITE / PERIOD / SAMPLER / CON FIGURATION
SUMMARY OF FINDINGS
PITTSBURGH SUPERSITE, PA; 7/1/01 to 8/1/02 6km east of downtown in a park on the top of a I
Sampler
Flow Rate (L/Min) Filter Type"
Denuder"
MOUDI-110
30
Teflon (W)
Teflon (W)
None
CMU
16.7
Nylon (W)
MgO/Citric acid
R&P-2000 FRM
16.7
Teflon (W)
None
Cabada et al. (2004,148869) ; Takahama et al.
- (2004,1670381116
- More than 70% (-0.5 ug/m3) of N03 mass was lost
from MOUDI samplers during summer.
- MOUDI N03 = 0.27 CMU; R2 = 0.40; Summer
MOUDI N03 = 0.99 CMU; R2 = 0.49; winter
' Wittigetal. (2004,103413f
Avg conversion efficiency to NOX (tested using
NH4N03 solution) was 0.85 ± 0.08. Gas analyzer
efficiency was stable at 0.99 ± 0.04.
Corrections were made for instrument offset,
software calculation error, conversion efficiency,
gas analyzer efficiency, vacuum drift, and sample
flow drift. The overall avg correction was 8%,
ranging from-62% to 93%.
Data Recovery >80%. Data loss was associated
with vacuum pump failures and excessive flash
strip breakage.
R&P-8400N = 0.83 CMU
-0.20 ug/m3; R2= 0.84
Underestimation in the R&P-8400N could be due
to incomplete particle collection or incomplete
conversion of various forms of N03".
Used co-located filter measurements for final
calibration.
FRESNO SUPERSITE, CA and other CRPAQS sites; 12/2/99 to 2/3/01
Located 5.5 km northeast of downtown in a mixed residential-commercial neighborhood. 107
Sampler
DRI-SFS
RAAS-400
RAAS-400
RAAS-100 FRM
Continuous Sampler
R&P-8400N
Sampler
PC-BOSS
Continuous Sampler
R&P-8400N
Dionex-IC
Flow Rate (L/Min)
113
24
24
16.7
Flow Rate (L/Min)
5
Flow Rate (L/Min)
150
Flow Rate (L/Min)
5
5
Filter Type3
Quartz (Pellulose
Quartz (P)-Nylon (P)
Quartz (P)-Quartz (P)
Quartz (P)
Denuder
Activated Carbon
Filter Type3
Teflon (W)- Nylon (P)
Denuder
Activated Carbon
Parallel plate wet denuder
Denuder11
AI203
Na2C03
None
None
Analysis Method"
NOx
Chemiluminescence
Denuder11
GIF
Analysis Method"
NOx
Chemiluminescence
1C
Chow et al. (2005, 099030)87
Maximum N03- volatilization was observed during
summer (Jun-Aug), while the lowest volatilization
was observed during winter (Dec-Feb).
Seasonal avg volatilized N03- in particulate N03"
(PN03 , the sum of non-volatilized and volatilized
N03") ranged from less than 10% during winter to
more than 80% during summer.
Volatilized NH4N03 accounted for 44% of actual
PM2s mass (i.e., measured mass plus volatilized
NH4N03) in Fresno during summer.
Front-quartz non-volatilized N03- concentrations
were similar for DRISFS (0.52 ± 0.26 ug/m3) and
RAAS-100 FRM (0.81 +0.33 ug/m3) for warm
months (May-Sep). With preceding denuders, the
DRI-SFS PN03 concentration (3 + 19 ug/m3) was
much higher than the RAAS100 FRM N03",
gaseous nitric acid (HN03) resulting in N03-
volatilization. FRM Teflon-membrane filters are
subject to similar N03 losses.
Chow et al. (2005, 1 66348)
December 2009
A-33
-------
SITE / PERIOD / SAMPLER / CON FIGURATION
SUMMARY OF FINDINGS
High correlation (R >0.90) between 24-h avg
R&P-8400N N03 and SFS filter N03"
concentrations, but R&P-8400N N03- was 7 to
25% lower than filter N03,
Limited comparison (n < 15) with filter samples at
Bakersfield showed that the slopes were close to
unity during early morning hours, while they
decreased during the afternoon hours, indicating
possible loss of N03" by the R&P-8400N
instrument.
The R&P-8400N required substantial maintenance
and careful operation.
Grover et al. (2006,138080165
Dionex-IC N03 = (0.71 ± 0.04) PC-BOSS N03 +
(3.2±1.1);R2 = 0.91;n = 29
R&P-8400N = (1.10 ± 0.06) PC-BOSS N03 -
(0.8±1.8);R2 = 0.93;n = 29
R&P-8400N = (0.55 ± 0.01) Dionex-IC +
(1.4±1.8);R2 = 0.75;n = 493
R&P-8400N measured less than Dionex-IC,
particularly at high RH. R&P-8400N may suffer
incomplete flash vaporization under conditions of
high RH.
December 2009
A-34
-------
SITE / PERIOD / SAMPLER / CON FIGURATION
SUMMARY OF FINDINGS
BALTIMORE SUPERSITE, MD; 2/14/02 to 11/30/02
Adjacent to a parking lot in the University of Maryland campus, influenced by motor vehicles, coal-fired
power plants and incinerators -21 km southwest of site and regionally transported material.
Sampler
Flow Rate (L/Min) Filter Type3
Denuder
SASS
6.7
Nylon (N/A)
MgO
Continuous Sampler Flow Rate (L/Min) Denuder
Analysis Method"
R&P-8400N
Activated Carbon
NOx
Chemiluminescence
Harrison et al. (2004,136787) °J
Corrections were made to R&P-8400N data for
. software calculation error, conversion efficiency,
gas analyzer efficiency, vacuum drift and sample
- flow drift.
- The relative uncertainty of R&P-8400N
measurements averaged 8.7%, ranging from 6.3%
- to 23%.
Data capture >95%.
-by
R&P-8400N underestimated SASS filter N03'
-33%, attributed to variations in conversion
efficiency, matrix effects, and impaction efficiency.
This suggested a true conversion efficiency of
68% as compared to an avg conversion efficiency
of R&P-8400N to NOX (tested using potassium
nitrate solution) of 0.90 ± 0.04.
Large errors occurred when the concentrations
were near the detection limit, when the
temperature difference (between instrument and
ambient) was large, and when the ambientRH was
< 40%. Ridged flash strips produced lower
dissociation losses than flat strips.
Reliable measurements were obtained when the
instrument-outdoor temperature differences were
minimal and when grooved/ridged flash strips were
used. A co-located filter measurement was used
for final corrections.
NEW YORK SUPERSITE, NY; 06/29/01 to 08/05/01 and 07/09/02 to 08/07/02
Urban site located at Queens College, NY, about 14 km west of Manhattan, within 2 km of freeways,
and within 12 km of international airports. Rural site located at Whiteface mountain, 600 m above sea
level, in a clearing surrounded by deciduous and evergreen trees and no major cities within 20 km of
Hocirefeetal. (2004.099003)-°
Data completeness: 86-88% for R&P-8400N, 94 -
98% for AMS, and 65-70% for PILS-IC.
me sue.
Sampler
R&P-2300
Continuous Sampler
R&P-8400N
PILS-IC
AMS
Flow Rate (L/Min)
10
Flow Rate (L/Min)
5
5
0.1
Filter Type3
Nylon (N/A)
Denuder
Activated Carbon
Na2C03and citric
acid
None
Denuder"
Na2C03
Analysis Method"
NOx Chemiluminescence
1C
Mass Spectrometry
Some PILS measurements were invalidated owing
to larger aqueous flow caused by bigger tubing.
Larger aqueous flow and inconsistent water quality
affected N03~ concentrations.
R&P-8400N NOs. was lower than R&P-2300 filter
- N03~ PILS-IC was within 5% of R&P-2300 filter
NOs. concentrations.
At the urban site, AMS was within 10% of the filter
N03 concentration. At the rural site, AMS had a
NOs"
NEW YORK SUPERSITE, NY; 10/01 to 07/05 (urban), 07/02 to 07/05 (rural) Urban site located at a
school in South Bronx, NY in a residential area, within a few kilometers away from major highways and
a freight yard (experiencing significant truck traffic). Rural site located at Whiteface mountain, 600 m
above sea level, in a clearing surrounded by deciduous and evergreen trees and no major cities within
20 km of the site.
Sampler
R&P-2300
TEOM-ACCU
Continuous Sampler
Flow Rate (L/Min)
10
16.7
Flow Rate (L/Min)
Filter Type3
Nylon (N/A)
Zefluor
Denuder
Denuder"
Na2C03
None
Analysis Method"
R&P-8400N
Activated Carbon NOX Chemiluminescence
Rattigan et al. (2006,116897)"
Data capture was more than 94%.
Data were adjusted for span and zero drifts,
conversion efficiency, flow drift, and blanks.
R&P-8400N NOs'was systematically lower than
R&P-2300 filter N03over all concentration ranges,
except at <1 ug/m .
Urban: R&P-8400N = 0.59 R&P-2300 N03 + 0.28;
R2 = 0.88; n = 305
" Rural: R&P-8400N = 0.73 R&P-2300 N03+ 0.01;
R2 = 0.90; n~161; however concentrations were
low with 95% of data < 1 ug/m3
Required weekly or biweekly maintenance by
trained personnel.
LOS ANGELES SUPERSITE, CA; 7/13/01 to 9/15/01 (Rubidoux) and 9/15/01 to 2/10/02
(Claremont)
Multiple sampling locations in the South Coast Air Basin (SoCAB), including urban "source" sites and
downwind "receptor" sites.
Fine etal. (2003,155775)19
MOUDI = 0.68 HEADS; R2 = 0.88
. ADI-N Sized = 0.80 HEADS; R2 = 0.79
Sampler
Flow Rate (L/Min)
Filter Type3 Denuder"
December 2009
A-35
-------
SITE / PERIOD / SAMPLER / CON FIGURATION
SUMMARY OF FINDINGS
MOUDI
30
Teflon (P)
None
HEADS
10
Teflon (N/A)-GF- Carbona(e
Continuous Sampler Flow Rate (L/Min) Denuder
Analysis Method"
ADI-N Sized
0.9
Activated Carbon NOX Chemiluminescence
ADI-N Sized = 1.12 MOUDI; FT = 0.53
ADI-N NOs- showed better agreement with HEADS
at lower concentrations, the ADI-N deviated
- (biased low) from the HEADS concentrations at
higher N03 concentrations. This deviation was
- attributed to N03.vaporization, loss of N03~
associated with particles less than 0.1 urn not
" collected by the ADI-N sampler, or loss of particles
in the ADI-N inlet tubing.
The underestimation of N03. by MOUDI compared
to HEADS may be due to N03~volatilization from
MOUDI stages, since S042~ comparisons showed
MOUDI to explain 85% of HEADS S042~
ADI-N and MOUDI showed better correlation
(R2 = 0.67) for the 1 -2 urn size range N03 relative
to other size ranges (R2< = 0.56). This is possibly
due to N03 in the form of non-volatilized sodium
nitrate (NaN03) than volatilized NH4N03in the 1-
2 urn size range. Single particle analysis also
indicated this possibility of NaN03 in the 1-2 urn
range.
RUBIDOUX, CA; 07/01/03 to 07/31/03
Located in the eastern section of SoCAB in the northwest corner of Riverside County, 78 km downwind
of the central Los Angeles metropolitan area and in the middle of the remaining agricultural production
area in SoCAB.
Grover et al. (2005,090044)'"'
R&P-8400N = (0.65 ± 0.07) PC-BOSS +
(3.3±2.4);R2 = 0.73;n = 31
At higher concentrations (no numerical value
reported), R&P-8400N N03-was lower than PC-
BOSS N03., possibly due to incomplete
volatilization of NH4N03in R&P-8400N at higher
concentrations (and higher relative humidity).
At the urban site, the continuous instruments
correlated well with filter N03~. measurements and
among themselves (R2> 0.89). At the rural site, R2
ranged from 0.61 to 0.83, except for the AMS
versus R&P2300 comparison, with an R2 of 0.46.
Sampler
PC-BOSS
Continuous Sampler
R&P-8400N
R&P-8400N
PILS-IC
AMS
Flow Rate (L/Min)
150
Flow Rate (L/Min)
5
5
5
0.1
Filter Type3
Teflon (W)-Nylon (P)
Denuder
Activated Carbon
Activated Carbon
Na2C03 and Citric acid
None
Denuder1'
GIF
Analysis Method'1
NOX Chemiluminescence
NOX Chemiluminescence
1C
Mass Spectrometry
December 2009
A-36
-------
SITE / PERIOD / SAMPLER / CON FIGURATION
SUMMARY OF FINDINGS
3Filter Manufacturer in parenthesis - W: Whatman, Clifton, NJ; P: Pall-Gelman, Ann Arbor, Ml; S: Schleicher & Schnell. Keene, NH; N/A: not available or not reported.
bAI203: Aluminum oxide; GF: Na2C03 impregnated Glass Fiber Filters; 1C: Ion chromatography; MgO: Magnesium oxide; Na2C03: Sodium carbonate; NaHC03: Sodium bicarbonate NOX: Oxides of nitrogen;
GIF: Charcoal Impregnated Filter; FEP: Fluorinated Ethylene Propylene copolymer; TEA: Triethanolamine; TSP: Total Suspended PM.
cNa2C03 impregnated.
d37 mm filter.
1Chow (1995, 077012): 2Watson and Chow (2001,1571231:3 Watson et al. (1983, 045084): 4Fehsenfeld et al. (2004,1573601:5Solomon et al. (2001,1571931: BWatson et al. (2005,1571241:7Mikel (2001,
156762): "Watson et al. (1999, 020949): 'Solomon and Sioutas (2006,156995): 10Graney et al. (2004, 053756): "Tanaka et al. (1998,157041): 12Pancras et al. (2005, 098120): "John et al. (1988, 045903):
"Hering and Cass (1999, 0849581:15Fitz et al. (1989, 0773871:1BHering et al. (1988, 0360121: "Solomon et al. (2003,1569941:1BCabada et al. (2004,1488591: "Fine et al. (2003,1557751:2DHogrefe et al.
(2004, 0990031:21Drewnick et al. (2003, 0991601:22Watson et al. (2005,1571251:23Ho et al. (2006,1565521:24Decesari et al. (2005,1445361:25 Mayol-Bracero et al. (2002, 0450101:26Yang et al. (2003,
156167): 27Tursic et al. (2006,157063): 2BMader et al.(2004,156724): 28Xiao and Liu (2004, 056801): 30Kiss et al. (2002,156646): 31Cornell and Jickells (1999,156367):32 Zheng et al. (2002, 026100):
33Fraser et al. (2002,1407411: x Fraser et al. (2003, 042231135Schauer er al. (2000, 0122251:3BFine et al. (2004,1412831:37Yue et al. (2004,1571691:3BRinehart et al. (2006,1151841:39Wan and Yu (2006,
1571041: "Poore (2000, 0128391:41Fraser et al. (2003, 0402661:42Engling et al. (2006,1564221:43Yu et al. (2005,1571671:44Tran et al. (2000, 0130251:45Yao et al. (2004,1022131:4BLi and Yu (2005,
156692): "Henning et al. (2003,156539): 4BZhang and Anastasio (2003,157182): "Emmenegger et al. (2007,156418):50 Watson et al, (1989,157119): 51Greaves et al. (1985,156494): 52Waterman et al.
(2000,1571161:53Waterman et al. (2001,1571171: aFalkovich and Rudich (2001,1564271:55Chow et al. (2007,1572091:5BMiguel et al. (2004,1232601:57Crimmins and Baker (2006, 0970081:5BHo and Yu
(2004,1565511:58Jeon et al. (2001, 0166361: BOMazzoleni et al. (2007, 0980381: B1Poore (2002, 0514441: B2Butler et al. (2003,1563131: B3Chow et al. (2006,1466221: MRussell et al. (2004, 0824531: B5Grover
et al. (2006,138080): BBGrover et al. (2005, 090044): B7Schwab et al. (2006, 098449): BBHauck et al. (2004,156525): B8Jaques et al. (2004,155878): 70Rupprecht and Patashnick (2003,157207): 71Pang et al
(2002, 0303531:72Eatough et al. (2001, 0103031:73Lee et al. (2005,128139); MLee et al. (2005,1566801:75Babich et al. (2000,1562391:7BLee et al. (2005,1559251:77Lee et al. (2005,1281391:7BAnderson
and Ogren (1998,1562131:79Chung et al. (2001,1563571: BOKidwell and Ondov (2004,1558981: B1Lithgow et al. (2004,1266161: B2Weber et al. (2003,1571291: B3Harrison et al. (2004,1367871: "Rattigan et
al. (2006,115897): B5Wittig et al. (2004,103413): BBVaughn et al. (2005,157089): B7Chow et al. (2005, 099030): BBWeber et al (2001, 024640); B8Schwab et al. (2006, 098785): ™Lim et al. (2003, 037037):
"Watson and Chow (2002, 0378731:82Venkatachari et al. (2006,1059181:83Bae et al. (2004,1562431:84Arhami et al. (2006,1562241:85Park et al. (2005,1568431:8BBae et al. (2004, 0986801:87Chow et al.
(2006,1563501:8BArnott et al. (2005,1562271:88Bond et al. (1999,1562811:1DDVirkkula et al. (2005,1570971:1D1Petzold et al. (2002,1568631:102Park et al. (2006, 0981041:103Arnott et al. (1999, 0206501:
"Veters et al. (2001, 016925): 105Pitchford et al. (1997,156872): 10BRees et al. (2004, 097164): 107Watson et al. (2000, 010354): 10BLee et al. (2005,156680): 108Hering et al. (2004,155837): ™Watson et al.
(1998,1988051: '"Chakrabarti et al. (2004,1574261:112Mathai et al. (1990,1567411: '"Kidwell and Ondov (2001.0170921: '"stanier et al. (2004, 0959551: '"Khlystovet al. (2005,1566351:11BTakahama et
al. (2004,1570381:117Chow et al. (2005,1563481: '"Zhang et al. (2002,1571811: '"Subramanian et al. (2004, 0812031:120Chow et al. (2006,1552071: ™Birch and Cary (1996, 0260041:122Birch (1998,
024953): 123Birch and Cary (1996, 002352): ™NIOSH (1996, 156810): 125NIOSH (1999, 156811): 12BChow et al. (1993, 077459): 127Chow et al. (2007, 156354): ™Ellis and Novakov (1982, 156416):
128Peterson and Richards (2002,1568611:130Schauer et al. (2003, 0370141: "iMiddlebrook et al. (2003, 0429321: "2Wenzel et al. (2003,1571391: "3Jimenez et al. (2003,1566111: "Vhares et al. (2003,
1568661:135Qin and Prather (2006,1568951: "BZhang et al. (2005,1571851:137Bein et al. (2005,1562651: "BDrewnick et al. (2004,1557541: "8Drewnick et al. (2004,1557551: "°Lake et al. (2003,1566691:
"'Lake etal. (2004, 088411)
Source: Chow etal. (2008,156355)
December 2009
A-37
-------
Table A-14. Summary of PM2.6 S042" measurement comparisons
SITE/PERIOD/SAMPLER/CONFIGURATION
ATLANTA SUPERSITE, GA: 08/03/99 to 09/01/99
Four km NW of downtown, within 200 m of a bus maintenance yard and several warehouse facilities,
representative of a mixed commercial-residential neighborhood.
Sampler
R&P-2000 FRM
RAAS-400
SASS
MASS-450
R&P-2300
VAPS
URG-PCM
ARA-PCM
ARA-PCM
PC-BOSS (TVA)
PC-BOSS (TVA)
PC-BOSS (BYU)
PC-BOSS (BYU)
MOUDI-100
Continuous Sampler
ADI-S
PILS-IC
ECN
TT
PITTSBURGH SUPERSITE
6 km east of downtown in a
Sampler
MOUDI-110
CMU
Flow Rate (L/Min)
16.7
24
6.7
16.7
10
15
16.7
16.7
16.7
105
105
150
150
30
Flow Rate (L/Min)
2.7
5
16.7
5
, PA; 070/1/01 to 08/01/02
park on the top of a hill
Flow Rate (L/Min)
30
16.7
Filter Type3
Quartz (P)
Teflon (P)
Teflon (P)
Quartz (P)
Quartz (P)
Quartz (P)
Teflon (P)-Cellulose-fiber
(W)
Teflon (N/A)
Nylon (N/A)
Teflon (W)
Quartz (P)
Teflon (W)
Quartz (P)
Teflon (N/A)
Quartz (N/A)
Denuder
Activated Carbon
Two URG annular glass
denuders in series
containing citric acid &
CaC03
Rotating annular wet
denuder system
Wet parallel plate
denuder
Filter Type3
Teflon (W)
Teflon (W)
Denuder"
None
None
None
None
None
XAD-4
Na2C03/Citric acid
Na2C03/Citric acid
GIF
GIF
GIF
GIF
None
Analysis Method"
S02, UV
Fluorescence
1C
1C
1C
Denuder11
None
MgO/Citric acid
SUMMARY OF FINDINGS
Solomon et al. (2003, 166994) 1T
PM2 5 S042~ from each sampler was
compared to all-sampler averages, called the
filter relative reference (filter RR) value. The
samplers agreed to within 10% of filter RR,
except for the PC-BOSS (TVA) and MOUDI-
100.
While avg mass was within 10%, daily
correlated well (R2 >0.90) with daily filter RR.
PC-BOSS (TVA) had instrument leaks.
The R&P-2000 FRM, on avg, agreed within
1% of filter RR.
MOUDI-100 was -13% low compared to filter
RR
Weber etal (2003 1 571 29)82i Zhang etal
(2002, 167181)118 '
Hourly PM2.5 S042~ were compared to all-
the approach used for filter samplers . Overall
agreement was within 16% or 2 ug/m
Good correlations (R - 0.76 to 0.94) were
found during the second half of the study,
except for TT versus ADI.
Good correlation (R2 = 0.84) was found
between continuous and filter-based S042
Continuous RR = (1 .15 ± 0.15), Filter RR +
(0.41 ±1.73)
Variability among continuous S042~
instruments (RSD = 13%) was similar to that
for NOs. instruments. Filter sample variability
was low (RSD - 8%) indicating more
uniformity among samplers.
The ECN and TT instruments were within
15%, PILS-IC was within 20% and ADI-S was
within 26% of filter RR.
Cabadaetal. (2004, 148859)18; Takahama
etal. (2004,157038)™
MOUDI S042' 0.80 CMU; R2 = 0.95; Summer
MOUDI S042" 0.97 CMU; R2 = 0.48; winter
Wittig et al. (2004, 103413)85
December 2009
A-38
-------
SITE/PERIOD/SAMPLER/CONFIGURATION
SUMMARY OF FINDINGS
R&P-2000 FRM
16.7
Teflon (W)
None
Continuous Sampler
Flow Rate
(L/Min)
Denuder
Analysis Method"
R&P-8400S
Activated Carbon S02 UV Fluorescence
Avg conversion efficiency to S02 (tested
- using ammonium sulfate [(NH4)2S04]
solution) was 0.65 ± 0.07. Gas analyzer
- efficiency was stable at 0.99 ± 0.06.
Corrections were made for instrument offset,
- software calculation error, conversion
efficiency, gas analyzer efficiency, vacuum
drift, and sample flow drift. The overall
correction was, on avg, -1 % and ranged from
-90% to 100% for individual samples.
Data Recovery >90%. Data loss was
associated with vacuum pump failures or
excessive flash strip breakage.
R&P-8400S (S042~) = 0.71 CMU +
0.42 ug/m3; R2 = 0.83
Underestimation is attributed to incomplete
particle collection or incomplete conversion of
various forms of S042~.
Used co-located filter measurements for final
calibration.
LOS ANGELES SUPERSITE, CA; 07/13/01 to 09/15/01 (Rubidoux) and 09/15/01 to 02/10/02
(Claremont)
Multiple sampling locations in the South Coast Air Basin (SoCAB), including urban "source" sites and
downwind "receptor" sites.
Fine et al. (2003,166776)19
MOUDI explained 85% of HEADS S04:
(R2 = 0.89; n = 40)
Sampler
Flow Rate
(L/Min)
Filter Type3
Denuded
MOUDI
30
Teflon (P)
None
HEADS
10
Teflon (N/A) GF-GFC Carbonate
NEW YORK SUPERSITE, NY; 06/29/01 to 08/05/01 and 07/09/02 to 08/07/02
Urban site located at Queens College, NY, about 14 km west of Manhattan, within 2 km of freeways, and
within 12 km of international airports. Rural site located at Whiteface mountain, 60m above sea level, in a
clearing surrounded by deciduous and evergreen trees and no major cities within 20 km of the site.
Sampler
R&P-2300
SCS
TEOM-ACCU
Continuous Sampler
R&P-8400S
PILS-IC
AMS
CASM
Flow Rate
(L/Min)
10
42
16.7
Flow Rate
(L/Min)
5
5
0.1
5
Filter Type3
Nylon (N/A)
Zefluor (N/A)
Zefluor (N/A)
Denuder
Activated Carbon
Na2C03 and Citric acid
None
Na2C03 and Carbon
and a Nafion dryer
Denuder"
Na2C03
None
None
Analysis Method'1
S02 UV Fluorescence
1C
Mass Spectrometry
S02 UV Fluorescence
Drewnick et al. (2003. 0991 SO)'1: Hogrefe
et al. (2004, 099003)2"
Data completeness: 89-93% for R&P-8400S,
94-98% for AMS, 81-98% for CASM, and
65-70% for PILS-IC.
The urban site data showed good
correlations (R2 = 0.87 to 0.94) with slopes
ranging from 0.97 to 1.01. At the rural site,
the variability was large (R2 = 0.73 to 0.91)
with slopes ranging from 0.76 to 1.32. S04
from PILS-IC was overestimated by -25%
when compared to the AMS at the rural site.
Filter samples were within 5% of each other,
except for comparison of ACCU with R&P-
2300 at the rural site, with high correlations
(R2 = 0.97 to 1.0). ACCU underestimated
S042"by~15%.
Continuous versus 6-h SCS filter
comparisons showed high R2 (0.91 to 0.95) at
the urban site. Continuous instruments
consistently measured lower S042~
concentrations compared to the SCS filter
measurements (slopes 0.68 to 0.73)
On avg, 85% of the filter-based S042" was
measured by the continuous instruments with
consistent relationships. At the rural site,
PILS-IC overestimated S042~ concentrations
(slopes 1.11 to 1.15), AMS and R&P-8400S
showed slopes of 0.71 -0.74 against SCS and
ACCU, while it ranged from 0.53- 0.68
against R&P-2300.
Error estimates:
Sampling losses: 2-3% for AMS and PILS-IC,
5-10% for R&P-8400S and none for CASM.
Continuous instruments probably
experienced more inlet transport losses (~
December 2009
A-39
-------
SITE/PERIOD/SAMPLER/CONFIGURATION
SUMMARY OF FINDINGS
'Ł>%) than tiller samplers due to longer inlet
lines.
Small (< 2%) positive artifact was found in
filters.
NEWYORK SUPERSITE, NY; 10/01 to 07/05 (urban), 07/02 to 07/05 (rural)
Urban site located at a school in South Bronx, NY in a residential area, within a few kilometers from major
highways and a freight yard (experiencing significant truck traffic). Rural site located at Whiteface mountain,
600m above sea level, in a clearing surrounded by deciduous and evergreen trees and no major cities within
20 km of the site. The study by Schwab etal.89 was based at a rural site located at Pinnacle State Park
surrounded by golf course, picnic areas and undeveloped forest lands and no major cities within 15 km.
Integrated Sampler
R&P-2300
Flow Rate
(L/Min)
10
Filter Type3
Nylon (N/A)
Denuder"
Na2C03
TEOM-ACCU
16.7
Zefluor
None
Continuous Sampler
Flow Rate (L/Min)
Denuder
Analysis Method"
R&P-8400S
Activated
Carbon
S02 pulsed fluorescence
TE-5020
(07/14/04to11/01/04)
Na2C03
S02 pulsed fluorescence
Rattigan et al. (2006,116897)84
Data capture was above 85%. Data loss was
primarily due to frequent flash strip failures,
every 2 wk and without warning.
Data were adjusted for span and zero drifts,
measured conversion efficiency, flow drift,
and blanks.
Calibrations used aqueous standards of
(NH4)2S04and oxalic acid solution in 1:4
ratio. Lower fractions of oxalic acid showed
- lower conversion efficiencies.
- Urban South Bronx site:
. R&P-8400S = 0.82 TEOM-ACCU + 1.15;
R2 = 0.84;n = 513
. R&P-8400S = 0.74 R&P-2300 + 1.14;
R2 = 0.81; n = 322
Rural Whiteface mountain:
R&P-8400S = 0.75 TEOM-ACCU + 0.22;
R2 = 0.95; n = 207
R&P-8400S = 0.78 R&P-2300 + 0.17;
R2 = 0.85; n= 198
Required weekly or biweekly maintenance by
trained personnel
Schwab et al. (2006, 098786)89
TE-5020 = 0.78 ACCU -0.2; R2 = 0.94
Similar studies at St. Louis, MO, show slopes
near unity. This suggests that the instrument
is sensitive to aerosol composition.
Low maintenance and calibration
requirements for TE-5020 compared to PILS-
IC and R&P-8400S.
FRESNO SUPERSITE.CA; 12/01/03 to 12/23/03
Located 5.5 km northeast of downtown in a mixed residential-commercial neighborhood. Flow Sampler
(L/min) Filter Typea Denuderb
Groveretal. (2006.138080)6
Dionex-IC S042"(1.03 ± 0.03) PC-BOSS S04
Sampler
PC-BOSS
Flow Rate (L/Min)
150
Filter Type3
Teflon (W)-
Nylon (P)
Denuder"
GIF
R&P-8400S S042" (0 95 + 0 05) Dionex-IC
S04 + (0.3 ± 0.6); R2 = 0.68; n = 195
Continuous Sampler
R&P-8400S
Dionex-IC
Flow Rate (L/Min)
5
5
Denuder
Activated
Carbon
Parallel plate
wet denuder
Analysis Method"
S02 pulsed fluorescence
1C
December 2009
A-40
-------
SITE/PERIOD/SAMPLER/CONFIGURATION
SUMMARY OF FINDINGS
aFilter Manufacturer in parentheses - W: Whatman, Clifton, NJ; P: Pall-Gelman, Ann Arbor, Ml; S: Schleicher & Schnell. Keene, NH; N/A: not available.
bAI203: Aluminum oxide; 1C: Ion chromatography; GIF: Charcoal Impregnated Filter; FEP: Fluorinated Ethylene Propylene copolymer; MgO: Magnesium oxide; Na2C03: Sodium carbonate; NaHC03: Sodium
bicarbonate NOX: Oxides of nitrogen; S02: Sulfur dioxide; TEA: Triethanolamine; TSP: Total Suspended PM; UV: Ultraviolet; XAD-4: Hydrophobic, non-polar polyaromatic resin.
cNa2C03 impregnated.
d37 mm filter.
1Chow (1995, 077012): 2Watson and Chow (2001,1571231:3 Watson et al. (1983, 045084): 4Fehsenfeld et al. (2004,1573601:5Solomon et al. (2001,1571931: BWatson et al. (2005,1571241:7Mikel (2001,
156762): "Watson et al. (1999, 020949): 'Solomon and Sioutas (2006,156995): 10Graney et al. (2004, 053756): "Tanaka et al. (1998,157041): 12Pancras et al. (2005, 098120): "John et al. (1988, 045903):
"Hering and Cass (1999, 0849581:15Fitz et al. (1989, 0773871:1BHering et al. (1988, 0360121: "Solomon et al. (2003,1569941:1BCabada et al. (2004,1488591: "Fine et al. (2003,1557751:2DHogrefe et al.
(2004, 0990031:21Drewnick et al. (2003, 0991601:22Watson et al. (2005,1571251:23Ho et al. (2006,1565521:24Decesari et al. (2005,1445361:25 Mayol-Bracero et al. (2002, 0450101:26Yang et al. (2003,
156167): 27Tursic et al. (2006,157063): 2BMader et al.(2004,156724): 28Xiao and Liu (2004, 056801): 30Kiss et al. (2002,156646): 31Cornell and Jickells (1999,156367):32 Zheng et al. (2002, 026100):
33Fraser et al. (2002,1407411: x Fraser et al. (2003, 042231135Schauer er al. (2000, 0122251:3BFine et al. (2004,1412831:37Yue et al. (2004,1571691:3BRinehart et al. (2006,1151841:39Wan and Yu (2006,
1571041: "Poore (2000, 0128391:41Fraser et al. (2003, 0402661:42Engling et al. (2006,1564221:43Yu et al. (2005,1571671:44Tran et al. (2000, 0130251:45Yao et al. (2004,1022131:4BLi and Yu (2005,
156692): "Henning et al. (2003,156539): 4BZhang and Anastasio (2003,157182): "Emmenegger et al. (2007,156418):50 Watson et al, (1989,157119): 51Greaves et al. (1985,156494): 52Waterman et al.
(2000,1571161:53Waterman et al. (2001,1571171: aFalkovich and Rudich (2001,1564271:55Chow et al. (2007,1572091:5BMiguel et al. (2004,1232601:57Crimmins and Baker (2006, 0970081:5BHo and Yu
(2004,1565511:58Jeon et al. (2001, 0166361: BOMazzoleni et al. (2007, 0980381: B1Poore (2002, 0514441: B2Butler et al. (2003,1563131: B3Chow et al. (2006,1466221: MRussell et al. (2004, 0824531: B5Grover
et al. (2006,138080): BBGrover et al. (2005, 090044): B7Schwab et al. (2006, 098449): BBHauck et al. (2004,156525): B8Jaques et al. (2004,155878): 70Rupprecht and Patashnick (2003,157207): 71Pang et al
(2002, 0303531:72Eatough et al. (2001, 0103031:73Lee et al. (2005,128139); MLee et al. (2005,1566801:75Babich et al. (2000,1562391:7BLee et al. (2005,1559251:77Lee et al. (2005,1281391:7BAnderson
and Ogren (1998,1562131:79Chung et al. (2001,1563571: BOKidwell and Ondov (2004,1558981: B1Lithgow et al. (2004,1266161: B2Weber et al. (2003,1571291: B3Harrison et al. (2004,1367871: "Rattigan et
al. (2006,115897): B5Wittig et al. (2004,103413): BBVaughn et al. (2005,157089): B7Chow et al. (2005, 099030): BBWeber et al (2001, 024640); B8Schwab et al. (2006, 098785): ™Lim et al. (2003, 037037):
"Watson and Chow (2002, 0378731:82Venkatachari et al. (2006,1059181:83Bae et al. (2004,1562431:84Arhami et al. (2006,1562241:85Park et al. (2005,1568431:8BBae et al. (2004, 0986801:87Chow et al.
(2006,1563501:8BArnott et al. (2005,1562271:88Bond et al. (1999,1562811:1DDVirkkula et al. (2005,1570971:1D1Petzold et al. (2002,1568631:102Park et al. (2006, 0981041:103Arnott et al. (1999, 0206501:
"Veters et al. (2001, 016925): 105Pitchford et al. (1997,156872): 10BRees et al. (2004, 097164): 107Watson et al. (2000, 010354): 10BLee et al. (2005,156680): 108Hering et al. (2004,155837): ™Watson et al.
(1998,1988051: '"Chakrabarti et al. (2004,1574261:112Mathai et al. (1990,1567411: '"Kidwell and Ondov (2001.0170921: '"stanier et al. (2004, 0959551: '"Khlystovet al. (2005,1566351:11BTakahama et
al. (2004,1570381:117Chow et al. (2005,1563481: '"Zhang et al. (2002,1571811: '"Subramanian et al. (2004, 0812031:120Chow et al. (2006,1552071: ™Birch and Cary (1996, 0260041:122Birch (1998,
024953): 123Birch and Cary (1996, 002352): ™NIOSH (1996, 156810): 125NIOSH (1999, 156811): 12BChow et al. (1993, 077459): 127Chow et al. (2007, 156354): ™Ellis and Novakov (1982, 156416):
128Peterson and Richards (2002,1568611:130Schauer et al. (2003, 0370141: "iMiddlebrook et al. (2003, 0429321: "2Wenzel et al. (2003,1571391: "3Jimenez et al. (2003,1566111: "Vhares et al. (2003,
1568661:135Qin and Prather (2006,1568951: "BZhang et al. (2005,1571851:137Bein et al. (2005,1562651: "BDrewnick et al. (2004,1557541: "8Drewnick et al. (2004,1557551: "°Lake et al. (2003,1566691:
"'Lake etal. (2004, 088411)
Source: Chow etal. (2008,156355)
Table A-15. Summary of PM2.6 carbon measurement comparisons.
SITE/PERIOD/SAMPLER/CONFIGURATION
ATLANTA SUPERSITE, GA: 08/03/99 to 09/01/99
Four km NW of downtown, within 200 m of a bus maintenance yard and several warehouse facilities,
representative of a mixed commercial-residential neighborhood.
Sampler
R&P-2000 FRM
RAAS-400
SASS
MASS-450
R&P-2300
VAPS
URG-PCM
ARA-PCM
PC-BOSS (TVA)
PC-BOSS (BYU)
MOUDI-100
Continuous Sampler
Flow Rate
(L/Min)
16.7
24
6.7
16.7
10
15
16.7
16.7
150
150
30
Flow Rate
(L/Min)
Filter Type3
Quartz (P)
Quartz (P)
Quartz (P)-
Quartz (P)
Quartz (P)
Quartz (P)-
Quartz (P)
Quartz (P)
Quartz (P)-
Quartz (P)
Quartz (N/A)-
Quartz (N/A)
Quartz (P)-
CIF (N/A)
Quartz (P)-CIF
(S)
Al Foil-Quartz
(N/A)f
Denuder
Denuder1'
None
None
None
None
None
XAD-4
XAD-4
GIF
GIF
GIF
None
OC EC
Analysis Method0
NIOSH 5040-TOT
NIOSH 5040-TOT
NIOSH 5040-TOT
NIOSH 5040-TOT
NIOSH 5040-TOT
NIOSH 5040-TOT
Front: NIOSH 5040-TOT;
Backup: custom-TOTd
IMPROVE_TOR
Front: IMPROVE TOR;
Backup: TPV
TPB
Custom-TORtosuitAI8
Comments
SUMMARY OF FINDINGS
Solomon et al. (2003, 166994) 1T
Organic Carbon (OC);
PM2.s OC from each sampler was compared to
the all-sampler avg, called the relative
reference (RR) value The samplers agreed to
within 20 to 50% of RR. Only front filter OC is
35%) than RR, while non-denuded sampler
OC was higher (5 to 35%).
Among non-denuded samplers, as filter face
exception of R&P-2300.
EC1
PM2.5 EC from each sampler was compared to
reference (RR) value. The samplers agreed to
within 20 to 200% of RR.
TOT samples showed less EC than RR by 15
to 30%, while TOR samples showed more EC
than RR h\/ 40 tn 00% PPRO^^ fRYl h >RR
value by 140%. EC by TOR is -twice EC by
TOT.
Major difference in EC is due to the carbon
analysis protocol and optical monitoring
Lim etal. (2003, 037037)90
December 2009
A-41
-------
SITE/PERIOD/SAMPLER/CONFIGURATION
SUMMARY OF FINDINGS
ADI-C
2.7
Activated Carbon
Not
known
RU-OGI
16.1
None
700 in
He
R&P-5400
16.7
None
275 in
air
PSAP
1.26
None
Part of S04 instrument TC concentrations measured by the RU-OGI
w/C02 non-dispersive and R&P-5400 correlated reasonably well
N/A infrared (NDIR) analyzer; (R2 = 0.83), with a slope of 0.96. The ratio of
data corrected for avg field the mean RU-OGI to mean R&P-5400 TC was
blank; OC = 2 oxidized OC 1.02.
R&P-5400 OC was 8% lower than the RU-OGI
(R2 = 0.73), while the R&P-5400 EC was 20%
-higher than RU-OGI (R2 = 0.74).
OC measured by ADI-C was lower than R&P-
- 5400 and RUOGI by 15% and 22%,
respectively.
850 in 2%
02
TOT; Dynamic blank for
adsorption correction
750 in air No pyrolysis correction
babs@
565 nm
10m /g factor
AE-16
None
babs@
880 nm
12.6 m/g factor
- EC from PSAP and AE-16 correlated well
(R2 = 0.97). PSAP was lower by -50%,
compared with AE-16, R&P-5400 and RU-
OGI.
EC measured by AE-16 was ~12% higher
than RU-OGI. Calibration factors for the light
absorption instruments need to be adjusted for
better correlation.
Calibration factor might be non-linear over the
range of absorbance measured.
The mean OC from R&P-5400 and RU-OGI
were within 10% of filter RR values. Mean
ADI-C OC was 14% lower than filter RROC.
EC from continuous instruments was 2-2.5
times filter RR EC; continuous TC was also
greater than filter RR TC by 17% (R&P-400) to
27% (RU-OGI).
December 2009
A-42
-------
SITE/PERIOD/SAMPLER/CONFIGURATION
SUMMARY OF FINDINGS
PITTSBURGH SUPERSITE, PA; 06/01/01 to 07/31/02
Six km east of downtown in a park on the top of a hill.
Sampler
Flow
Filter Type/Pack3
Denuder
Analysis Method0
16.7
CMU Custom-1
Teflon
Non-denuded (P/W)- .,.„,,
sampte Quartz (P) None
(QBT)
NIOSH5040-TOT
16.7
sample
NIOSH5040-TOT
16.7
Denuded
sample
Denuder-
Quartz (P)- Activated Carbon NIOSH 5040-TOT
CIG (S)
16.7
Teflon
D^™c Kder
blank (DYN) ^f(rp).
CIG (S)
Activated Carbon NIOSH 5040-TOT
16.7 Non-denuded Teflon
blank (UDB) (P/W)-
Quartz (P)-
CIG (S)
None
NIOSH5040-TOT
CMU Custom-2
Subramanian etal. (2004,081203)""
. Particulate OC (POC) was estimated from
denuded sample (Quartz OC + CIG OC) after
- subtracting DYN POC.
Denuder efficiency (1-DYN POC/UDB POC)
was 94 ± 3%. No seasonal variability or
deterioration in denuder performance was
- observed.
Positive artifact due to denuder breakthrough
was 18.3 ±12.5% of the denuded sample
. POC.
Negative artifact (CIGsample-CIGDYN) was,
on avg, 6.3 ± 6.2% of POC.
" Positive artifact was 34 ± 10% from QBT, and
was 13 ± 5% from QBQ. QBT »QBQ.
QBT over-corrected the positive artifact by
20%. OC volatilization from the front Teflon
filter that subsequently adsorbed on the back-
" up quartz filter, resulted in an overestimation
of the positive artifact.
Non-denuded QBQ provided a more
representative estimate of the positive artifact
on the non-denuded front quartz filter for 24-h
samples. However, it was not suitable for 4- to
6-h samples, because the filters were not in
equilibrium with the air stream.
Positive artifact dominated when sampling
with a non-denuded quartz filter.
Comparison of 24-h avg non-denuded front
quartz OC versus denuded POC over the year
showed an intercept of 0.53 ug/m3, indicative
of a positive artifact on quartz filter samples.
The artifacts were higher in summer on an
absolute basis; however, they showed no
seasonal variation when expressed as a
fraction of POC.
ST. LOUIS SUPERSITE, IL, MO; 01/01/02 to 12/31/02
Three km east of St. Louis, MO City center, also impacted by industrial sources, and located in a mixed
residential light commercial neighborhood.
Bae et al. (2004,156243)93; Bae et al. (2004,
Sampler
University of
Wisconsin Custom-1
University of
Wisconsin Custom-2
Flow Rate
(Lmin)
24
24
Filter
Type/Pack3
Quartz (P)
Denuder-Quartz
(P)
Denuder-Quartz
(P)
Teflon (N/A)-
Denuder-Quartz
(P)
Denuder1'
None
GIF
GIF
GIF
Analysis Method0
ACE Asia TOT
ACE Asia TOT
ACE Asia TOT
ACE Asia TOT
Continuous Sampler
Denuder
OC
EC
Comments
Denuder breakthrough was 0.17 ± 0.15 ug/m3,
and constituted less than 5% of annual avg
OC concentration.
Non-denuded OC = (1.06 ± 0.02) x denuded
OC +(0.34 ±0.10)
Equivalence of OC intercept and denuder
breakthrough implies that the low-level artifact
is caused by denuder breakthrough.
Non-denuded EC = (1.04 ± 0.03) x denuded
EC + (0.07 ± 0.03), indicating negligible EC
artifact.
Results suggested higher summertime OC
artifact, on an absolute basis.
December 2009
A-43
-------
SITE/PERIOD/SAMPLER/CONFIGURATION
SUMMARY OF FINDINGS
Sunset OCEC
GIF
340,
500,615,
870°C in
100% He
550, 625,
700, 775,
850, 900
°C in 2%
02, 98%
He
ACE Asia TOT; CH4 Comparison of continuous Sunset TC and OC
FID detector with 24-h filter samples showed good
correlations (R2) of 0.89 and 0.90,
respectively.
Continuous Sunset TC
in ug/m3 = (0.97 ± 0.02) x filter TC +
(0.83 ± 0.11), indicating comparability with the
filter measurements.
Continuous Sunset OC = (0.93 ± 0.02) x filter
OC +(0.94 ±0.09)
Positive intercept was interpreted to be a
blank correction for the continuous
measurements.
EC comparison was poor with large scatter in
data (R2 = 0.60), probably due to low EC
concentrations (avg = 0.70 ug/m3), close to
the detection limit (0.5 ug/m3).
FRESNO SUPERSITE, CA and other CRPAQS sites; 12/02/99 to 02/03/01, 12/1/03 to 11/30/04
Fresno Supersite was located 5.5 km northeast of downtown in a mixed residential-commercial
neighborhood.
Sampler
DRI-SFS
RAAS-400
RAAS-400
RAAS-100 FRM
Continuous Sampler
R&P-5400
Sunset OCEC
MAAP
AE-16
AE-21
AE-31
DRI-PA
Sampler
Flow Rate
(Lmin)
113
24
24
16.7
Flow Rate
(L/Min)
16.7
8.5
16.7
6.8
6.8
6.8
3
Flow Rate
(Lmin)
Filter
Type/Pack3
Quartz (P)
Teflon (P)-
Quartz (P)
(QBT)
(P) (QBT)
Quartz (P)-
Quartz (P)
(QBQ)
Quartz (P)-
Quartz (P)
(QBQ)
Quartz (P)
Denuder
None
CIG
None
None
None
None
None
Filter
Type/Pack3
Denuder1'
None
None
None
XAD-4/CIF
None
OC EC
275°Cinair ?f°°C in
air
650, 750,
850,
940°C in
2%02in
He
babs@
670 nm
babs@
880 nm
icn cnn
650, 850°C babs @
in He 370, 880
nm
babs@
370, 470,
520, 590,
660, 880
and 950
nm
babs@
1047 nm
Denuder11
Analysis Method0
IMPROVE_TOR
IMPROVE_TOR
IMPROVE_TOR
IMPROVE_TOR
IMPROVE_TOR
Comments
No pyrolysis
correction
Transmittance
Transmittance
6.5 m2/g factor
Transmittance
14625/Am2/g factor,
where A is in nm
Absorption, 5 m2/g
factor
Analysis Method0
Watson and Chow (2002, 037873)91; Chow
et al. (2005, 166348)"7; Chow et al. (2006,
155207)120; Watson et al. (2005, 157124)6;
Non-denuded RAAS-400 and RAAS-100 FRM
400 and RAAS-100 FRM samplers showed
comparability for front filter TC, OC and EC
measurements.
Positive OC artifact was 1 .62 ± 0.58 ug/m3
(~24% of non-denuded front quartz OC) from
QBT, and 1 .12 ± 0.91 ug/m3 (-17% of non-
denuded front quartz OC) from QBQ. QBT
»QBQ
positive OC artifact of 34% (of the non-
denuded front quartz OC) from QBT and
17.5% (of the non-denuded front quartz OC)
from QBQ.
Positive artifact was higher during summer
than winter.
Negative artifact was, on avg,
0.61 ± 0.58 ug/m3 (-10% of POC) at Fresno.
Over all the CRPAQS sites, it ranged from
2.3% in winter to 11 % in summer, with an avg
of 4.9%.
Positive artifact is estimated to be 0.5 ug/m3.
No difference in denuded quartz backup OC
denude rs.
against filter samples showed poor correlation
(R2<0.55).
TC from R&P-5400 was 40-60% higher than
filter TC by TOR. None of the R&P-5400
versus TOR filter comparisons were
comparable or predictable, due to several
frequent instrument malfunctions during the
experiment and the small data set (-35 data
points).
IMPROVE_TOR EC was consistently 20-25%
higher than aethalometer BC.
IMPROVE TOR EC was comparable to
MAAP BC.
Comparison of light absorption (babs) from
DRI-PA (1047 nm), MAAP (670 nm), and AE
December 2009
A-44
-------
SITE/PERIOD/SAMPLER/CONFIGURATION SUMMARY OF FINDINGS
PC-BOSS
Continuous Sampler
R&P-5400
150
Flow Rate
(L/Min)
16.7
Quartz (P)-CIG
(S)t
Denuder
None
GIF
OC
375°C in air
EC
750°C in
air
TPV
Comments
No pyrolysis
(880 nm) analyzers witn tne tiller
IMPROVE_TOR EC, gave a oabs of 2.3, 5.5
and 10 m /g, differing from the default
conversion factors of 5, 6.5, and 16.6 mz/g
used for each instrument at the specified
wavelength
Grover et al. (2006, 138080)65
R&P-5400 TC = (0.50 ± 0.01) Sunset TC +
(3.6±1.5);R2=0.73;n = 480
„,_„,-„„ 650750 Sunset TC = [0.63 ±0.05) PC-BOSS TC +
250,500, o?nipi"' NIOSH 5040 TOT (4 1+3 2V R2= 086'n = 29
Sunset OCEC 8.0 CIG 650,850°C °5°Ł'" ~ l*-!--^,* u.oo.n
in He PRO/ HP NDIR c°2 detector R&R-5400 TC = (0.41 ± 0.02) PC-BOSS TC H
98/oHe (6.7±1.6);R2=0.91;n = 29
December 2009 A-45
-------
SITE/PERIOD/SAMPLER/CONFIGURATION
BALTIMORE SUPERSITE, MD; 02/1 5/2002 to 11/30/2002
East of downtown in an urban residential area. Within 91 m of bus maintenance facility.
Sampler
SASS
Continuous Sampler
Sunset OCEC
Flow Rate
(L/Min)
6.7
Flow Rate
(L/Min)
8
Filter
Type/Pack3
Quartz (P)-
Quartz (P)
Denuded
Carbon
Denuded
None
OC
600°C,
then
870°C in
He
EC
870°C in
2% 02 in He
Analysis Method0
STN_TOT
Comments
TOT; CH4 FID
detector; Denuder
breakthrough ~
0.5-1 ug C/m3; Used
0.5 to correct OC
concentrations
RUBIDOUX, CA; 07/13/03 to 07/26/03
Rubidoux is located in the eastern section of the South Coast Air Basin (SoCAB) in the northwest corner of
Riverside County, 78 km downwind of the central Los Angeles metropolitan area and in the middle of the
remaining agricultural production area in SoCAB.
Sampler
PC-BOSS
Continuous Sampler
Sunset OCEC
Sunset OCEC
Flow Rate
(L/Min)
150
Flow Rate
(L/Min)
8
8
Filter
Type/Pack3
Quartz (P)-CIG
(S)
Denuded
GIF
GIF
Denuded
GIF
OC
N/A
N/A
EC
N/A
Not meas-
ured
Analysis Method0
TPB (CIG heated to
450 °C in N2)
Comments
TOT; NDIR detector;
NIOSH 5040
protocol
TOT; has blank
quartz filter before
entering analyzer.
Used as "blank"
stream for
quantifying OC
artifacts;3-step
analysis only in He.
NEW YORK SUPERSITE, NY; 01/12/04 to 02/05/04
Urban site located at Queens College, NY, about 14 km west of Manhattan, within 2 km of freeways, and
within 12 km of international airports.
Integrated Sampler
R&P-2300
Continuous Sampler
R&P-5400
Sunset OCEC
AE-20
AMS
Flow Rate
(L/Min)
10
Flow Rate
(L/Min)
16.7
N/A
N/A
N/A
Filter
Type/Pack3
Quartz
Denuded
None
GIF
None
None
Denuded
None
OC
340 °C in
air
600, 870
°C in He
N/A
EC
750 °C in
air
870 °C at
10%02in
He
babs@370,
880 nm
N/A
Analysis Method0
STN_TOT
Comments
No pyrolysis
correction
Transmittance
Transmittance,
14625 A m2/g factor,
where A is in nm
~ 1 urn cut-point
SUMMARY OF FINDINGS
Parketal. (2005, 156843)95
Data capture 93.8%
Compared to SASS, Sunset underestimated
OC and EC by 22% and -11 .5%, respectively.
Higher OC in SASS was attributed to the
absence of a denuder (i.e., positive artifact by
gaseous adsorption) and to temperature
- differences between the STN_TOT and
Sunset TOT carbon analysis temperature
protocols.
EC discrepancy was probably related to the
differences in temperature protocol.
Grover et al. (2005, 090044)66
Sunset OCEC TC = (0.90 ± 0.06) PC-BOSS +
(2.0±2.1);R2=0.93;n = 21
Sunset TC was adjusted for carbon artifacts
measured by second (blank) instrument.
Venkatachari et al. (2006, 106918)92
Regression of OC from Sunset OCEC against
- PM2s mass concentration yielded an intercept
of 1 .14 ug/m3, which was used as a measure
of the positive artifact on the Sunset data. The
- Sunset OC data was corrected for this artifact.
- AE-20 BC concentrations were -86% of
Sunset EC and R&P2300 filter EC
- concentrations.
AE-20 versus R&P-5400 showed high scatter.
Sunset Optical EC = 0.58 ± 0.05 Sunset
Thermal EC; R2 = 0.86; n = 506
Sunset Optical EC = 0.62 ± 0.05 AE-20 BC;
- R2 = 0.96; n = 539
R&P-5400 TC tracked filter TC closely, but
differed widely for OC and EC.
Sunset OC = (0.75 ± 0.76) R&P-2300 OC +
(0.08 ± 0.36); R2 = 0.67; n = 16
Sunset OC = (0.98 + 0.11) R&P-5400 OC -
(0.47±0.17);R2 = 0.44;n = 327
December 2009
A-46
-------
SITE/PERIOD/SAMPLER/CONFIGURATION
SUMMARY OF FINDINGS
R&P-5400 OC = (0.60 ± 0.47) R&P-2300 OC
+ (0.58 ± 0.82); R2= 0.58; n = 17
Organic matter measurements byAMS
showed reasonable correlation (R2 = 0.76)
with filter (R&P-2300) OC, while being poorly
correlated with continuous OC by Sunset
(R2 = 0.32) and R&P-5400 (R2 = 0.36)
Sunset EC = (1.21 ± 0.44) R&P-2300 EC -
(0.03 ± 0.13); R2 = 0.94; n = 16
Sunset EC = (1.35 ± 0.12) R&P-5400 EC +
(0.06 ±0.04);R2 = 0.61; n = 327
R&P-5400 EC = (0.49 ± 0.46) R&P-2300 EC +
(0.09 ± 0.26); R2 = 0.77; n = 15
Sunset TC = (0.86 ± 0.39) R&P-2300 TC -
(0.06 ± 0.69); R2 = 0.77; n = 16
Sunset TC = (1.31 ±0.10) R&P-5400TC-
(1.15 ± 0.15); R2 = 0.59; n = 327
R&P-5400 TC = (0.77 ± 0.58) R&P-2300 TC +
(0.35±1.37);R2 = 0.83;n = 16
aFilter Manufacturer in parentheses - W: Whatman, Clifton, NJ; P: Pall-Gelman, Ann Arbor, Ml; S: Schleicher & Schnell. Keene, NH; N/A: not available. QBT: quartz backup filter behind Teflon front filter. QBQ:
quartz backup filter behind Quartz front filter.
bAI203: Aluminum oxide; 1C: Ion chromatography; GIF: Charcoal Impregnated Filter; CIG: Charcoal Impregnated Glass-Fiber Filter; FEP: Fluorinated Ethylene Propylene copolymer; MgO: Magnesium oxide;
Na2C03: Sodium carbonate; NaHC03: Sodium bicarbonate NOX: Oxides of nitrogen; S02: Sulfur dioxide; TEA: Triethanolamine; TSP: Total Suspended PM; UV: Ultraviolet; XAD-4: (hydrophobic, non-polar
polyaromatic resin.
CNIOSH 5040_TOT: National Institute of Occupational Safety and Health Method 5040 Thermal Optical Transmittance Protocol. "1122'123' ™' "5 OC: 250, 500, 650, 850 °C for OC1, OC2, OC3, and OC4
fractions, respectively, for 60, 60, 60, 90 sec respectively, in 100% He atmosphere. EC: 650, 750, 850, 940 °C for EC1, EC2, ECS, and EC4 fractions, respectively, 30, 30, 30, >120 sec respectively, in 98% He
and 2% 02 atmosphere. OPT: Pyrolysis correction by transmittance. IMPROVE_TOR: Interagency Monitoring of Protected Visual Environments Thermal Optical Reflectance Protocol 12B OC fractions: 120,
250, 450, 550 °C for OC1, OC2, OC3, and OC4 fractions, respectively, until a well defined peak has evolved at each step, with a time limit of min 80 sec and max of 580 sec, in 100% He atmosphere. EC
fractions: 550, 700, 800 °C for EC1, EC2, and ECS fractions, respectively, until a well defined peak has evolved at each step, with a time limit of min 80 sec and max of 580 sec, in 2% 02 and 98% He
atmosphere. OPR: Pyrolysis correction for pyrolyzed organic carbon (OP) by reflectance. OC = OC1+OC2+OC3+OC4+OP EC = EC1+EC2+EC3-OPTC = OC+EC. IMPROVE_ATOR: 127 Note that as of
May, 2007, the U.S. EPA is switching samples from the Speciation Trends Network thermal optical transmittance protocol to the IMPROVE_A protocol. OC: 140, 280, 480, 580 °C for OC1, OC2, OC3, and
OC4, fractions, respectively, until a well defined peak has evolved at each step, with a time limit of 80 sec and max of 580 sec, in 100% He atmosphere EC: 580, 740, 840 °C for EC1, EC2, and ECS fractions,
respectively, until a well defined peak has evolved at each step, with a time limit of min 80 sec and max of 580 sec, in 2% 02and 98% He atmosphere. OPR: Pyrolysis correction for pyrolyzed organic carbon
(OP) by reflectance. OPT: Pyrolysis correction by transmittance. TPV: Temperature Programmed Volatilization "' B112B For CIF Filters: Heated from 50 °C to 300 °C at a ramp rate of 10 °C/min in N2. For
Quartz filters: Heated from 50 °C to 800 °C at a ramp rate of 28 °C/min in 70% N2 and 30% 02; EC estimated from high temperature peak (>450 °C) on thermogram obtained from quartz-fiber filter analysis;
No pyrolysis correction. STN_TOT: Speciation Trends Network Thermal Optical Transmittance Protocol.129 OC: 310, 480, 615, 920 °C for 60, 60, 60, 90 sec respectively, in 100% He atmosphere. EC: 600,
675, 750, 825, 920 °C for 45, 45, 45, 45,120 sec respectively, in 98% He and 2% 02 atmosphere. ACE Asia TOT: Aerosol Characterization Experiments in Asia Thermal Optical Transmittance Protocol. 130
OC: 340, 500, 615, 870 °C for 60, 60, 60, 90 sec respectively, in 100% He atmosphere. EC: 550, 625, 700, 775, 850, 900 °C for45, 45, 45, 45, 45,120 sec respectively, in 98% He, 2% 02. Pyrolysis correction
by transmittance.
dCustom TOT: XAD-4 impregnated quartz, analyzed in He-only atmosphere with a maximum temperature 176 °C; EC is not measured.
eCustom TOR to suit Al substrate; details not reported.
'37 mm filter
'Chow (1995, 0770121:2Watson and Chow (2001,1571231:3 Watson et al. (1983, 0450841:4Fehsenfeld et al. (2004,1573601:5Solomon et al. (2001,1571931: BWatson et al. (2005,1571241:7Mikel (2001,
1567621: BWatson et al. (1999, 0209491:8Solomon and Sioutas (2006,1569951:1DGraney et al. (2004, 0537561: "Tanaka et al. (1998,1570411:12Pancras et al. (2005, 0981201: "John et al. (1988, 0459031:
"Hering and Cass (1999, 084958): 15Fitz et al. (1989, 077387): 1BHering et al. (1988, 036012): "Solomon et al. (2003,156994): 1BCabada et al. (2004,148859): "Fine et al. (2003,155775): 20Hogrefe et al.
(2004, 0990031:21Drewnick et al. (2003, 0991601:22Watson et al. (2005,1571251:23Ho et al. (2006,1565521:24Decesari et al. (2005,1445361:25 Mayol-Bracero et al. (2002, 0450101:2BYang et al. (2003,
1561671:27Tursic et al. (2006,1570631:2BMader et al.(2004,1567241:28Xiao and Liu (2004, 0568011:30Kiss et al. (2002,1566461:31Cornell and Jickells (1999,1563671:32 Zheng et al. (2002, 0261001:
33Fraser et al. (2002,140741): x Fraser et al. (2003, 042231) 35Schauer er al. (2000, 012225): 3BFine et al. (2004,141283): 37Yue et al. (2004,157169): 3BRinehart et al. (2006,115184): 38Wan and Yu (2006,
1571041:40Poore (2000, 0128391:41Fraser et al. (2003, 0402661:42Engling et al. (2006,1564221:43Yu et al. (2005,1571671:44Tran et al. (2000, 0130251:45Yao et al. (2004,1022131:4BLi and Yu (2005,
1566921: "Henning et al. (2003,1565391:4BZhang and Anastasio (2003,1571821:48Emmenegger et al. (2007,1564181:50 Watson et al, (1989,1571191:51Greaves et al. (1985,1564941:52Waterman et al.
(2000,157116): 53Waterman et al. (2001,157117): aFalkovich and Rudich (2001,156427): 55Chow et al. (2007,157209): 5BMiguel et al. (2004,123260): 57Crimmins and Baker (2006, 097008): 5BHo and Yu
(2004,1565511:58Jeon et al. (2001, 0166361: BDMazzoleni et al. (2007, 0980381: B1Poore (2002, 0514441:B2Butler et al. (2003,1563131: B3Chow et al. (2006,1466221: "Russell et al. (2004, 0824531: B5Grover
et al. (2006,1380801: BBGrover et al. (2005, 0900441: B7Schwab et al. (2006, 0984491: BBHauck et al. (2004,1565251: B8Jaques et al. (2004,1558781:7DRupprecht and Patashnick (2003,1572071:71Pang et al
(2002, 030353): 72Eatough et al. (2001, 010303): 73Lee et al. (2005,128139); MLee et al. (2005,156680): 75Babich et al. (2000,156239): 7BLee et al. (2005,155925): 77Lee et al. (2005,128139): 7BAnderson
and Ogren (1998,1562131:78Chung et al. (2001,1563571: BOKidwell and Ondov (2004,1558981: B1Lithgow et al. (2004,1266161: B2Weber et al. (2003,1571291: B3Harrison et al. (2004,1367871: MRattigan et
al. (2006,1158971: B5Wittig et al. (2004,1034131: BBVaughn et al. (2005,1570891: B7Chow et al. (2005, 0990301: BBWeber et al (2001, 0246401; B8Schwab et al. (2006, 0987851:8DLim et al. (2003, 0370371:
81Watson and Chow (2002, 037873): 82Venkatachari et al. (2006,105918): 83Bae et al. (2004,156243): 84Arhami et al. (2006,156224): 85Park et al. (2005,156843): ™Bae et al. (2004, 098680): 87Chow et al.
(2006,1563501:8BArnott et al. (2005,1562271: "Bond et al. (1999,1562811: "°Virkkula et al. (2005,1570971:101Petzold et al. (2002,1568631: "2Park et al. (2006, 0981041:103Arnott et al. (1999, 0206501:
"Veters et al. (2001, 0169251:105Pitchford et al. (1997,1568721: "BRees et al. (2004, 0971641:107Watson et al. (2000, 0103541: "BLee et al. (2005,1566801: "8Hering et al. (2004,1558371: ™Watson et al.
(1998,198805): 111Chakrabarti et al. (2004,157426): 112Mathai et al. (1990,156741): '"Kidwell and Ondov (2001,017092): '"Stanier et al. (2004, 095955): '"Khlystovet al. (2005,156635): '"Takahama et
al. (2004,1570381:117Chow et al. (2005,1563481:11BZhang et al. (2002,1571811: '"Subramanian et al. (2004, 0812031:120Chow et al. (2006,1552071: "1 Birch and Cary (1996, 0260041:122Birch (1998,
0249531: 123Birch and Cary (1996, 0023521: ™NIOSH (1996, 1568101:125NIOSH (1999, 1568111:12BChow et al. (1993, 0774591: 127Chow et al. (2007, 1563541: 12BEllis and Novakov (1982, 1564161:
128Peterson and Richards (2002,156861): 130Schauer et al. (2003, 037014): "iMiddlebrook et al. (2003, 042932): "2Wenzel et al. (2003,157139): "3Jimenez et al. (2003,156611): "Vhares et al. (2003,
1568661:135Qin and Prather (2006,1568951: "BZhang et al. (2005,1571851:137Bein et al. (2005,1562651: "BDrewnick et al. (2004,1557541: "8Drewnick et al. (2004,1557551: MOLake et al. (2003,1566691:
"'Lake etal. (2004, 088411)
Source: Chow etal. (2008,1563551
December 2009
A-47
-------
Table A-16. Summary of particle mass spectrometer measurement comparisons.
ATLANTA SUPERSITE, GA: 08/03/99 to 09/01/99
Four km NW of downtown, within 200 m of a bus maintenance yard and several warehouse facilities, representative of a mixed commercial-residential neighborhood.
Spectrometer
PALMS
ATOFMS
RSMS-II
AMS
Inlet Characteristics (Flow Rate
[L/Min]; Size Inlet; Dryer
Aerodynamic Diameter, (jrn;
Particle Sizing Method)
N/A
PM2.5 cyclone
Nafion (17 days) / None (4 days)
0.35-2.5
Light scattering
1
None
None
0.2-2.5
Aerosol TOP
N/A
Nonp
INUI 1C
Nafion
0.015-1.3
Aerodynamic focusing; Need to
pre-select sizes to be analyzed
N/A
PM2.5 cyclone
None
0.05-2.5
Aerosol TOP
Volatilization/
lonization Hit Rates
Method"
LDI,
ArF193nm 14 to 100%,
a „ overall 87%
2x10 to 5x10
W/cm2
LDI,Nd:YAG266 25.30o/0
nm laser occasionally
~1x108W/cm2 as low as 5%
LDI, Arf laser, 193
nm
8 8 N/A
1x108to2x108
W/cm2
T~550°C/ El N/A
Mass Spectrometer
Single TOP reflectron;
Ion polarity needs to
be pre-selected
Dual TOP reflectron;
Detects both positive
and negative ions
Single linear TOP; Ion
polarity needs to be
pre-selected
Quadrupole;
Mass weighted size
distributions on pre-
selected positive ions
onlv
Particle Analysis/
Classification
Peak ID/regression
tree analysis
Aerosol TOP
Peak ID/artificial
neural network
ID using standard
El ionization
databases
Other
Pure sulfuric
acid (H2S04),
(NH4)2S04,
and water
(H20)
have relatively
high ionization
thresholds (i.e.
difficult to
ionize).
Fraction of
molecules
ionized in the
particles is on
the order of
10"5to10"6.
Does not
detect/ analyze
highly
refractory
materials such
as metals, sea
salt, soot etc.
Fraction of
molecules
ionized in the
particles is on
the order of
10'6to10'7
Middlebrook et al. (2003, 042932)131; Wenzel et al. (2003,157139)132; Jimenez et al. (2003,1666111133
Particle sizing is approximate in PALMS, while ATOFMS, RSMS-II and AMS provide relatively accurate particle sizing.
Particle transmission in AMS is ~100% (i.e., it uses all particles in the sampled air) between 60 and 600 nm, while that for PALMS, ATOFMS and RSMS-II
range from 10-6 for submicron particles to 2% for supermicron (>0.8 urn) particles.
AMS has fewer matrix effects (due to separate volatilization and ionization steps) compared to single-step LDI instruments.
While four major particle classifications (organic/S042~, sodium/potassium sulfate, soot/hydrocarbon and mineral) were observed by all three laser
instruments, they differed in the classification frequencies. Differences in frequencies that are detected and grouped are related to the differences in the
laser ionization conditions (e.g., wavelength), particle transmission, sizing method and the way the spectra were classified.
Shorter ionization wavelengths are able to produce ions more easily than longer ones.
Low hit rates in ATOFMS corresponded to periods of high S042~ concentrations. Low hit rates in PALMS were related to a variety of factors including high
Sp42~ concentrations, differing laser fluence and laser position relative to particle beam. Use of a dryer in PALMS enhanced ionization of particles that were
difficult to ionize at high ambient RH.
The RSMS-II and ATOFMS were less sensitive to S042~ and hence may have fewer organic/S042~ particles (i.e., underestimate S042~, pure sulfuric acid
etc.).
The PALMS, ATOFMS and RSMS (laser based instruments) are qualitative, while the AMS can be quantitative. The relative ratio of ion intensities from the
laser instruments, however, may be indicative of relative concentrations, thus giving semi-quantitative information.
Comparison of the ratio of N03 to S04 peaks with the results from the semi continuous instruments showed better correlation with the AMS (R2 = 0.93) than
PALMS (R2 = 0.65 for non-dry particles to 0.70 for dry particles). While reasonable correlations between the PALMS and the composite semi-continuous
data indicate the possibility for calibration of laser-based data for certain ions, the calibration factors may vary depending on the particle matrix, water
content and laser ionization parameters, and averaging the spectra according to these factors may minimize these effects.
Comparison of AMS S04with PILS S04 showed good correlation (R2= 0.79), and the data uniformly scattered around a 1:1 line. N03 comparison was poor
(R2 = 0.49) because of the low signal to noise ratio at low concentrations
The continuum between particle classifications indicates that the particles were not adequately represented by non-overlapping classifications.
December 2009
A-48
-------
ATLANTA SUPERSITE, GA: 08/03/99 to 09/01/99
Four km NW of downtown, within 200 m of a bus maintenance yard and several warehouse facilities, representative of a mixed commercial-residential neighborhood.
HOUSTON SUPERSITE, TX; 08/23/00 to 09/18/00
Houston Regional Monitoring Site was located < 1.0 km north of the Houston ship channel, where chemical and other industries are present. The site was
located between a railway to the south and a chemical plant to the north. Major freeways were located just to the north and east of the sampling site.
Inlet Characteristics (Flow Rate
c»./.tr~~.t.r [L/Min]; Size Inlet; Dryer
Spectrometer Aerodynamic Diameter, Vm;
Particle Sizing Method)
N/A
IN/r\
None
PCMC M Nafion
Kbivib-ii 0.035-1.14
Aerodynamic focusing; Need to
pre-select sizes to be analyzed
Volatilization/ pa,+i,.i»
lonization HitRatesb Mass Spectrometer0 n™,
MothrwH' UaSSI
ivicuiuQ
LDI ArF laser Single linear TOF; lon Peak ID/a
193nm ' N/A pre-selecteddS '° ^ neural net
flcation'5' Other
At each size
point, aerosol
was sampled
in each cycle
for either 10
riif. . i min or until
™^' mass spectra
worK for 30 particles
per major
class were
collected,
whichever
came first.
Phares et al. (2003,166866V34
27,000 spectra were classified using a neural network into 15 particle types
Fifteen particle type mass spectra were presented along with their size distribution, avg time of day occurrence, and wind direction dependence
Major classes were a K+ dominant, Si/Silicon Oxide, Carbon, Sea Salt, Fe, Zn, Amines, Lime, Vanadium, Organic Mineral, Pb and K, Al, and a Pb salt
particle type.
FRESNO SUPERSITE, CA: 11/30/00 to 2/4/01
Urban location in a residential neighborhood.
Spectrometer
ATOFMS
Inlet Characteristics (Flow Rate
[L/Min]; Size Inlet; Dryer
Aerodynamic Diameter, urn;
Particle Sizing Method)
1
None
None
0.3-2.5
Aerodynamic
Volatilization/ D rt-
lonization HitRatesb Mass Spectrometer0 nfll,*^*™
Method3 Classification
miNn-YAfi Peak ID/artificial
ORR nm N/A Dual reflectron TOF
266 nm neural network
Other
ATOFMS
unsealed
detected
particles
tracked p
attenuation
monitor PM2.s
mass
concentration
Qin and Prather (2008,166986V35
Biomass burning particles reached a maximum at night and a minimum during the day. These particles were less than 1 urn in diameter and accounted for
more than 60% of the particles detected at night.
Another particle class characterized by high mass carbon fragments had a similar diurnal pattern. These particles were larger than 1 urn and were
interpreted as biomass particles that have undergone gas to particle conversion of semi-volatile species followed by dissolution in a water droplet.
PITTSBURGH SUPERSITE, PA; 09/07/02 TO 09/22/02 FORAMS; 09/20/01 to 09/26/02 for RSMS-III
6 km east of downtown in a park on the top of a hill
Spectrometer
AMS
Inlet Characteristics (Flow Rate [L/Min];
Size Inlet; Dryer Aerodynamic
Diameter, urn; Particle Sizing Method)
1 .4 cc/s
PM2.s cyclone
None
0.05-1.0
Aerosol TOF
Volatilization/
lonization
Method"
T-600°C/EI
Hit Rates
Quadrupole;
Mass weighted size distributions on
pre-selected positive ions only.
Mass Spectrometer0
Particle size-cut of ~1 urn
RSMS-II
N/A
None
Nafion
0.03-1.1
Aerodynamic focusing; Need to pre-
select sizes to be analyzed.
LDI.ArF laser, 193
nm
Dual TOF feflectron; Detects both
positive and negative ions
At each size point, aerosol
was sampled in each cycle for
either 10 min or until mass
spectra for 30 particles per
major class were collected,
whichever came first
December 2009
A-49
-------
ATLANTA SUPERSITE, GA: 08/03/99 to 09/01/99
Four km NW of downtown, within 200 m of a bus maintenance yard and several warehouse facilities, representative of a mixed commercial-residential neighborhood.
Zhang et al. (2005,1671861136: Bein et al. (2005,1662661137
The AMS observed 75% of the S042" measured by R&P-8400S (R2 = 0.69).
Collection efficiency (CE) of 0.5 used for S042~, N03and NH4+and 0.7 for organics to correct mass concentrations for incomplete detection. Use of a
constant CE irrespective of size and shape may overestimate accumulation mode (mostly, oxygenated) organics (true CE ~ 0.5) and underestimate smaller
mode (primary) organics (true CE ~ 1.0).
Comparison of AMS organics (organic matter, OM) with OC measured by a continuous Sunset OCEC instrument showed good correlation (R2 = 0.88) with
aslope of 1.69.A24-h avg comparison, showed a slope of 1.45. These values are in the typical range of 1.2 to 2.0 for OM/OC ratios.
AMS could be used along with the SMPS to estimate particle density. The AMS did not always agree with SMPS, probably due to non-spherical particles
(irregular) such as soot from fresh traffic emissions, whose mass may be overestimated by the SMPS.
Comparison of AMS mass with the MOUDI, showed differences for aerodynamic diameters >600 nm, probably due to the AMS transmission being less than
unity for particles larger than 600 nm.
For RSMS-III, 54% of the detected particles were assigned to one class (carbonaceous ammonium nitrate). This class was preferentially detected during
the colder months and was detected from many different wind directions.
The next largest RSMS-III class was EC/OC/K class at 11%, and is believed to be from biomass burning.
An unidentified organic carbon RSMS-III class (3.3% of all detected particles) was seen to be highly dependent on wind direction dependence and was
primarily detected during August and September of 2002. These particles likely originated from a landfill.
NEW YORK SUPERSITE; 06/30/01 to 08/05/01 (urban); 07/09/02 to 08/07/02: (rural)
Urban Site: Queens College, Queens, New York, located at the edge of a parking lot and within 1 km from expressways and highways in New York City
Metropolitan area.
Rural Site: Whiteface Mountain, New York, located in a cleared area surrounded by mix of deciduous and evergreen trees, ~2 km away from the closest
highway with no major cities within 20 km.
Inlet Characteristics (Flow
.„.„„._ Rate [L/Minl; Size Inlet; Dryer
Spectrometer Aerodynamic Diameter, um;
Particle Sizing Method)
0.1
PM2.5 cyclone
AMS None
0.02-2.5
Aerosol TOF
Volatilization/
lonization
Method9
T-700°C/EI
Hit Rates
Quadrupole;
Mass weighted size
distributions on pre-
selected positive ions
only.
Mass Spectrometer0
Data are 10-min
averages
Drewnick et al. (2004,1667641138: Drewnick et al. (2004,155755)139; Hogrefe et al. (2004,099003)20
Transport losses were 1.3% on avg.
Inlet losses (at the in let of AMS) were 1.9%, on avg, ranging from 11% for a 20 nm particle to 9% for a 2.5 um particle, with a minimum of 0.7% for a 350 nm
particle
Overall measurement uncertainty of particle diameter was ~11%.
The AMS was reliable with proper calibration, care, and maintenance. Valid 10 min averages were obtained for all components more than 93% of the time.
The mass to charge ratios (m/z) of fragments from different components may overlap (e.g., NH+, a fragment of NH4+ and CH3+, a fragment of organic
species, have m/z = 15) resulting in an interference (called as isobaric interference) Interfering signals were not used to calculate concentrations. This loss
in concentration was adjusted by applying a correction factor determined from laboratory studies.
Typical interferences were from fragments of organic species, water and oxygen.
With adjustments, the S042~, N03~ and ammonium concentrations measured by the AMS were consistently lower than that measured by other co-located
instruments, probably due to incomplete focusing of the (NH4)2S04and NH4N03 particles by the aerodynamic lens.
At the urban site, AMS N03 was within 10% of the filter N03 concentration. At the rural site, it had a slope of 0.51 and R2 of 0.46.
AMS S04 showed good agreement with R&P-8400S at both the rural and urban locations (R2 = 0.89 to 0.92, slope = 0.99, n = 407 to 695) and was within
70 to 85% of filter S042" concentration.
Comparison of the total non-refractory mass measured by the AMS with the PM2.s TEOM mass (operated at 50°C or with dryer) at the urban location,
showed good correlation (R2 = 0.91) with near zero intercept (0.22 ug/m3). On avg, the AMS observed 64% of the mass measured by the TEOM.
The unexplained mass (36%) was attributed to transport losses, transmission and optical losses, and refractory components in the aerosol sample (e.g.,
metals, EC). The mass closure was within the estimated uncertainty of the AMS mass measurements (5-10%).
December 2009
A-50
-------
ATLANTA SUPERSITE, GA: 08/03/99 to 09/01/99
Four km NW of downtown, within 200 m of a bus maintenance yard and several warehouse facilities, representative of a mixed commercial-residential neighborhood.
BALTIMORE SUPERSITE, MD; 04/01/02 to 11/30/02
East of downtown in an urban residential area. Within 91 mofa bus maintenance facility.
Inlet Characteristics (Flow W«I,*;M,,««»/
^rtromPtPr Rate [L/Min]; Size Inlet Dryer Vua±±n/
Spectrometer Aerodynamic Diameter, |im Sin"
Particle Sizing Method) Metnoa
0.2-18, based on particle size
chosen
None
RSMS-III 0 MW 3 LDI, ArF laser, 193 nm
Aerodynamic focusing; Need
topre-selectsizestobe
analyzed
Hit Rates Mass Spectrometer
At each size set point,
aerosol was sampled
in each cycle for either
TOP with dual ion polarity m,ln *ir un LJT1355
spectra trom ju
particles were
collected, whichever
came first.
Lake et al. (2003,166669)140, Lake et al. (2004,088411)141
Utilizing both positive and negative ion detection enables detection of more species. However, detection efficiencies of negative ions decreased for smaller
particles.
S04+ concentration (number or mass) was not accurately quantified.
RSMS-III was most efficient in 0.050 to 0.77 um range.
Particle compositions could be related to specific source categaories.
aEI: Electon Impact; LDI: Laser Desorption / lonization
bHitrate refers to the number of particles with a mass spectrum as a fraction of the number of particles detected. It does not apply to RSMS and AMS because there is no separate detection
cTOF: Time fo Flight
1Chow (1995, 0770121:2Watson and Chow (2001,1571231:3 Watson et al. (1983, 0450841:4Fehsenfeld et al. (2004,1573601:5Solomon et al. (2001,1571931: BWatson et al. (2005,1571241:7Mikel (2001,
1567621: "Watson et al. (1999, 0209491: 'Solomon and Sioutas (2006,1569951:10Graney et al. (2004, 0537561: "lanaka et al. (1998,1570411:12Pancras et al. (2005, 0981201: "John et al. (1988, 0459031:
"Hering and Cass (1999, 084958): 15Fitz et al. (1989, 077387): 1BHering et al. (1988, 036012): "Solomon et al. (2003,156994): 1BCabada et al. (2004,148859): "Fine et al. (2003,155775): 20Hogrefe et al.
(2004, 0990031:21Drewnick et al. (2003, 0991601:22Watson et al. (2005,1571251:23Ho et al. (2006,1565521:24Decesari et al. (2005,1445361:25 Mayol-Bracero et al. (2002, 0450101:2BYang et al. (2003,
1561671:27Tursic et al. (2006,1570631:2BMader et al.(2004,1567241:23Xiao and Liu (2004, 0568011:30Kiss et al. (2002,1566461:31Cornell and Jickells (1999,1563671:32 Zheng et al. (2002, 0261001:
33Fraser et al. (2002,140741): x Fraser et al. (2003, 042231) 35Schauer er al. (2000, 012225): 3BFine et al. (2004,141283): 37Yue et al. (2004,157169): 3BRinehart et al. (2006,115184): 38Wan and Yu (2006,
1571041: "Poore (2000, 0128391:41Fraser et al. (2003, 0402661:42Engling et al. (2006,1564221: J3Yu et al. (2005,1571671:44Tran et al. (2000, 0130251:45Yao et al. (2004,1022131:4BLi and Yu (2005,
1566921: "Henning et al. (2003,1565391:4BZhang and Anastasio (2003,1571821: "Emmenegger et al. (2007,1564181:50 Watson et al, (1989,1571191:51Greaves et al. (1985,1564941:52Waterman et al.
(2000,157116): 53Waterman et al. (2001,157117): 54Falkovich and Rudich (2001,156427): 55Chow et al. (2007,157209): 5BMiguel et al. (2004,123260): 57Crimmins and Baker (2006, 097008): 5BHo and Yu
(2004,1565511:58Jeon et al. (2001, 0166361: BOMazzoleni et al. (2007, 0980381: B1Poore (2002, 0514441:B2Butler et al. (2003,1563131: B3Chow et al. (2006,1466221: MRussell et al. (2004, 0824531: B5Grover
et al. (2006,1380801: BBGrover et al. (2005, 0900441: B7Schwab et al. (2006, 0984491: BBHauck et al. (2004,1565251: B8Jaques et al. (2004,1558781:70Rupprecht and Patashnick (2003,1572071:71Pang et al
(2002, 030353): 72Eatough et al. (2001, 010303): 73Lee et al. (2005,128139); 7\ee et al. (2005,156680): 75Babich et al. (2000,156239): 7BLee et al. (2005,155925): 77Lee et al. (2005,128139): 7BAnderson
and Ogren (1998,1562131:79Chung et al. (2001,1563571: BOKidwell and Ondov (2004,1558981: B1Lithgow et al. (2004,1266161: B2Weber et al. (2003,1571291: B3Harrison et al. (2004,1367871: "Rattigan et
al. (2006,1158971: B5Wttig et al. (2004,1034131: BBVaughn et al. (2005,1570891: B7Chow et al. (2005, 0990301: BBWeber et al (2001, 0246401; B8Schwab et al. (2006, 0987851: ™Lim et al. (2003, 0370371:
"Watson and Chow (2002, 037873): 82Venkatachari et al. (2006,105918): 33Bae et al. (2004,156243): MArhami et al. (2006,156224): 85Park et al. (2005,156843): %Bae et al. (2004, 098680): 37Chow et al.
(2006,1563501:8BArnott et al. (2005,1562271: "Bond et al. (1999,1562811: "°Virkkula et al. (2005,1570971:101Petzold et al. (2002,1568631:102Park et al. (2006, 0981041:103Arnott et al. (1999, 0206501:
""Peters et al. (2001, 0169251:105Pitchford et al. (1997,1568721:10BRees et al. (2004, 0971641:107Watson et al. (2000, 0103541: ™Lee et al. (2005,1566801: ™Hering et al. (2004,1558371: ™Watson et al.
(1998,198805): 111Chakrabarti et al. (2004,157426): 112Mathai et al. (1990,156741): '"Kidwell and Ondov (2001,017092): ™Stanier et al. (2004, 095955): '"Khlystovet al. (2005,156635): '"Takahama et
al. (2004,1570381:117Chow et al. (2005,1563481:11BZhang et al. (2002,1571811: ™Subramanian et al. (2004, 0812031: "°Chow et al. (2006,1552071: "'Birch and Cary (1996, 0260041:122Birch (1998,
0249531: 123Birch and Cary (1996, 0023521: ™NIOSH (1996, 1568101: "5NIOSH (1999, 1568111:12BChow et al. (1993, 0774591: 127Chow et al. (2007, 1563541: ™Ellis and Novakov (1982, 1564161:
"'Peterson and Richards (2002,156861): 130Schauer et al. (2003, 037014): "iMiddlebrook et al. (2003, 042932): "2Wenzel et al. (2003,157139): "3Jimenez et al. (2003,156611): "Vhares et al. (2003,
1568661:135Qin and Prather (2006,1568951: "BZhang et al. (2005,1571851:137Bein et al. (2005,1562651: "BDrewnick et al. (2004,1557541: "3Drewnick et al. (2004,1557551: "°Lake et al. (2003,1566691:
"'Lake etal. (2004, 088411)
Source: Chow etal. (2008,1563551
December 2009
A-51
-------
Table A-17. Summary of key parameters for TD-GC/MS and pyrolysis-GC/MS.
Reference
Sample Type
TD-GC/MS WITH RESISTIVELY HEATED EXTERNAL
Greaves etal. (1985, 156494;
1987, 156495) ;Veltkampet
al. (1996,081594)
Waterman et al. (2000,
157116)
Waterman et al. (2001 ,
157117)
Sidhu etal. (2001, 155202)
Aerosol sample and NISTSRM
1649
NISTSRM 1640a
NISTSRM 1649a
TDUnit
OVEN
A cylindrical aluminum block
containing a heating cartridge
connected to a thermocouple
External oven mounted on the top
of the GC/MS system
Same as above
Aerosol collected on glass fiber A stainless steel tube (0.635 cm
filters from combustion of alternative °-D-) laced m a GC oven
dieselfuel.
Analytical Instrument
HP5892AGC/MSinEI
mode
HP5890GC/Fisons
MD 800 MS, scan
range: 40-520 amu
HP5890GC/Fisons
MD 800 MS, scan
range: m/z 40 to 520
Two GCs and one MS. The
firstGCis used as the TE
unit. The second GC
separates the desorbent.
Total Analysis
Time
ambient sample:
55.5 min NIST
standard: 45.5
min
90 min
90 mins
Ua
Hays et al. (2003,156529: Aerosol collected from residential
2004.156530): Dong et al. wood combustion, residential oil
(2004,156409) furnace and fireplace appliance
A glass tube placed in an external
oven (TDS2 Gerstel Inc.)
Agilent 6890 GC/5793 MSD,
scan range: 50 to 500 amu 99 mjn
CURIE POINT TD-GC/MS
Jeon etal. (2001, 016636)
Neusussetal. (2000,
156804)
High-volume PM10 ambient samples
collected along the U.SVMexico Curie point pyrolyzer HP 5890 GC/5792 MSD Ua
border
Ambient aerosol collected during the
2nd Aerosol Characterization Curie point pyrolyzer Fisons Trio 1 000 35 min
Experiment
IN-INJECTION PORT TED-GC/MS
u i • t i
neimigeiai.
Aerosol samples collected on glass- GC injector port, with modified
fiber filters at a forest site septum cap
Carlo ErbaMega5160
GC/VG 250/70 SE MS, scan 47 min
range: 45-400 amu
Hall etal. (1999.156512)
NISTSRM 1649
Micro-scale sealed vessel placed
inside the injector port
HP5890GC/FisonsMD800
MS, scan range: 40-500 82.5 min
amu
ai.
Aerosol samples collected on
quartz-and-glass filters in Ontario
AGC injection port was added with
three minor components, including
a small T-connector, 3-way valve,
and needle valve
HP 5892A GC/5972A MS in
El mode
71 min
Falkovich and Rudich (2001,
156427); Falkovich et al.
(2004,156428); Graham et
al. (2004,156490)
NISTSRM 1649a; urban aerosols Direct Sample Introduction (DSI)
collected with an 8-stage impactor in device (ChromatoProbe, Varian
Tel-Aviv, Israel Co.)
Varian Saturn 3400 GC/MS 64.2 min
Ho and Yu (2004.156551):
Yang et al. (2005,102388)
Ambient aerosol samples collected
on Teflon-impregnated glass-fiber
filters in Hong Kong and on quartz
filters at Nanjing, China
Conventional GC injection port. No
modification of GC injector and liner
HP5890GC/5791 MSD,
scan range: 50-650 amu
TD-GCXGC-MS
Welthagen et al. (2003,
104056) ; Schnelle-Kreis etal.
(2005, 112944)
Hamilton et al. (2004,
156516)
Hamilton et al. (2005,
088173)
Ambient samples in Augsburg,
Germany
PM2.s aerosol collected in London
Secondary organic aerosol formed
during the photo-oxidation of
toluene with OH radicals
Injection port Optic III with
autoloader (ATAS-GL, Veldhoven,
NL)
Conventional GC injection port
The same as above
Agilent 6890 GC/LECO
Pegasus III TOF/MS with a
LECO Pegasus 4D GCxGC
modulator
The same as above, scan
range: 20-350 amu
The same above
175 min
93.7 min
102.5 min
December 2009
A-52
-------
Reference Sample Type TD Unit Analytical Instrument Total Analysis
IN SITU SEMI-CONTINUOUS AND CONTINUOUS TD SYSTEMS
IAI-II- » i ™inc -icc-ic7\ In situ aerosol samples collected in Collection-TE cell with conventional Agilent 6890 GC/5793 MSD, cn •
Williams etal. (2006,156151) Berkley, CA GC injection port scan range: 29-550 amu 59 mln
PYROLYSIS TD-GC/MS
Voorhees etal. (1991, PM0 6 and PM>0.45 collected on A tube furnace directly interfaced to Extrel Simulscan GC/MS, ,, , .
1571011 quartz fiber m pristine regions of anQC/MS scan range: 35-450 amu 317mln
Subbalakshmi et al. (2000, Ambient aerosol collected on glass- „ . . , Agilent 6890 GC/5973 MS, R,,- .
157023) fiber filters in Jakarta, Indonesia M pyromjecior scan range: 50-550 amu DJ'° mln
DM I, ( . i ,. f.,( A pyrolyzer directly connected to the Varian 3400 GC/Saturn II ion
Fabbri etal. (2002,156426) jnantSrialareaof Itahf GC injector port through an trap MS, scan range: 45-400 57 min
' interface heated at 250° C amu
PM2.6 collected on quartz-fiber filters
Blazso etal. (2003,156278) sSmpledlS^n ATfoill in A pyrolyzer Agilent 6890 GC/5973 MS 30.3 min
Brazil
Labbanetal. (2006,156665) ™q0Uafrtz3-torPfiltedrsd S°''C°"ected Curie point pyrolyzer HP5890 GC/5972 MS 25.5. min
3Total analysis time could not be determined because of insufficient experimental details
Source: Chow etal. (2007,157209)
December 2009 A-53
-------
A.1.2. Networks
Table A-18. Relevant Spatial Scales for PMio, PM2.6, and PMi0.2.6 Measurement
Spatial
Scales
PMio
PMZ5
PMlO-2.5
Microscale This scale would typify areas such as
downtown street canyons, traffic corridors, and
(-6-100 m) fence |jne stationary source monitoring
locations where the general public could be
exposed to maximum PMio concentrations.
Microscale PM sites should be located near
inhabited buildings or locations where the
general public can be expected to be exposed
to the concentration measured. Emissions from
stationary sources such as primary and
secondary smelters, power plants, and other
large industrial processes may, under certain
plume conditions, likewise result in high ground
level concentrations at the microscale. In the
latter case, the microscale would represent an
area impacted by the plume with dimensions
extending up to approximately 100 m. Data
collected at microscale sites provide
information for evaluating and developing hot
spot control measures.
This scale would typify areas such as downtown
street canyons and traffic corridors where the
general public would be exposed to maximum
concentrations from mobile sources. In some
circumstances, the microscale is appropriate for
particulate sites; community-oriented SLAMS sites
measured at the microscale level should,
however, be limited to urban sites that are
representative of long-term human exposure and
of many such microenvironments in the area. In
general, microscale PM sites should be located
near inhabited buildings or locations where the
general public can be expected to be exposed to
the concentration measured. Emissions from
stationary sources such as primary and secondary
smelters, power plants, and other large industrial
processes may, under certain plume conditions,
likewise result in high ground level concentrations
at the microscale. In the latter case, the
microscale would represent an area impacted by
the plume with dimensions extending up to
approximately 100 m. Data collected at
microscale sites provide information for evaluating
and developing hot spot control measures. Unless
these sites are indicative of population-oriented
monitoring, they may be more appropriately
classified as SPM.
This scale would typify relatively small areas
immediately adjacent to: industrial sources;
locations experiencing ongoing construction,
redevelopment, and soil disturbance; and
heavily traveled roadways. Data collected at
microscale stations would characterize
exposure over areas of limited spatial extent
and population exposure, and may provide
information useful for evaluating and
developing source-oriented control measures.
Middle Scale Much of the short-term public exposure to
coarse fraction particles (PM10) is on this scale
(-100-600 m) anc| on the neighborhood scale. People moving
through downtown areas or living near major
roadways or stationary sources, may
encounter particulate pollution that would be
adequately characterized by measurements of
this spatial scale. Middle scale PM10
measurements can be appropriate for the
evaluation of possible short-term exposure
public health effects. In many situations,
monitoring sites that are representative of
micro-scale or middle-scale impacts are not
unique and are representative of many similar
situations. This can occur along traffic corridors
or other locations in a residential district. In this
case, one location is representative of a
neighborhood of small scale sites and is
appropriate for evaluation of long-term or
chronic effects. This scale also includes the
characteristic concentrations for other areas
with dimensions of a few hundred meters such
as the parking lot and feeder streets
associated with shopping centers, stadia, and
office buildings. In the case of PM10, unpaved
or seldomly swept parking lots associated with
these sources could be an important source in
addition to the vehicular emissions themselves.
People moving through downtown areas, or living
near major roadways, encounter particle
concentrations that would be adequately
characterized by this spatial scale. Thus,
measurements of this type would be appropriate
for the evaluation of possible short-term exposure
public health effects of PM pollution. In many
situations, monitoring sites that are representative
of microscale or middle-scale impacts are not
unique and are representative of many similar
situations. This can occur along traffic corridors or
other locations in a residential district. In this case,
one location is representative of a number of
small scale sites and is appropriate for evaluation
of long-term or chronic effects. This scale also
includes the characteristic concentrations for other
areas with dimensions of a few hundred meters
such as the parking lot and feeder streets
associated with shopping centers, stadia, and
office buildings.
People living or working near major roadways
or industrial districts encounter particle
concentrations that would be adequately
characterized by this spatial scale. Thus,
measurements of this type would be
appropriate for the evaluation of public health
effects of PM10.25 exposure. Monitors located in
populated areas that are nearly adjacent to
large industrial point sources of PMio-2 5
provide suitable locations for assessing
maximum population exposure levels and
identifying areas of potentially poor air quality.
Similarly, monitors located in populated areas
that border dense networks of heavily-traveled
traffic are appropriate for assessing the
impacts of resuspended road dust. This scale
also includes the characteristic concentrations
for other areas with dimensions of a few
hundred meters such as school grounds and
parks that are nearly adjacent to major
roadways and industrial point sources,
locations exhibiting mixed residential and
commercial development, and downtown areas
featuring office buildings, shopping centers,
and stadiums.
December 2009
A-54
-------
Spatial
Scales
PMio
PMzs
PMlO-2.5
Neighborhood Measurements in this category represent
Scale conditions throughout some reasonably
homogeneous urban sub-region with
(-600 m-4 km) dimensions of a few kilometers and of
generally more regular shape than the middle
scale. Homogeneity refers to the PM
concentrations, as well as the land use and
land surface characteristics. In some cases, a
location carefully chosen to provide
neighborhood scale data would represent not
only the immediate neighborhood but also
neighborhoods of the same type in other parts
of the city. Neighborhood scale PMio sites
provide information about trends and
compliance with standards because they often
represent conditions in areas where people
commonly live and work for extended periods.
Neighborhood scale data could provide
valuable information for developing, testing,
and revising models that describe the larger-
scale concentration patterns, especially those
models relying on spatially smoothed emission
fields for inputs. The neighborhood scale
measurements could also be used for
neighborhood comparisons within or between
cities.
Measurements in this category would represent
conditions throughout some reasonably
homogeneous urban sub- region with dimensions
of a few kilometers and of generally more regular
shape than the middle scale. Homogeneity refers
to the PM concentrations, as well as the land use
and land surface characteristics. Much of the
PM25 exposures are expected to be associated
with this scale of measurement. In some cases, a
location carefully chosen to provide neighborhood
scale data would represent the immediate
neighborhood as well as neighborhoods of the
same type in other parts of the city. PM25 sites of
this kind provide good information about trends
and compliance with standards because they
often represent conditions in areas where people
commonly live and work for periods comparable to
those specified in the NAAQS. In general, most
PM25 monitoring in urban areas should have this
scale.
Measurements in this category would
represent conditions throughout some
reasonably homogeneous urban sub-region
with dimensions of a few kilometers and of
generally more regular shape than the middle
scale. Homogeneity refers to the PM
concentrations, as well as the land use and
land surface characteristics. This category
includes suburban neighborhoods dominated
by residences that are somewhat distant from
major roadways and industrial districts but still
impacted by urban sources, and areas of
diverse land use where residences are
interspersed with commercial and industrial
neighborhoods. In some cases, a location
carefully chosen to provide neighborhood scale
data would represent the immediate
neighborhood as well as neighborhoods of the
same type in other parts of the city. The
comparison of data from middle scale and
neighborhood scale sites would provide
valuable information for determining the
variation of PM10-2.5 levels across urban areas
and assessing the spatial extent of elevated
concentrations caused by major industrial point
sources and heavily traveled roadways.
Neighborhood scale sites would provide
concentration data that are relevant to
informing a large segment of the population of
their exposure levels on a given day.
Urban Scale
This class of measurement would be used to
characterize the PM concentration over an entire
metropolitan or rural area ranging in size from 4 to
50 kilometers. Such measurements would be
useful for assessing trends in area-wide air
quality, and hence, the effectiveness of large scale
air pollution control strategies. Community-
oriented PM25 sites may have this scale.
Regional
Scale
(-60-1 OOs km)
These measurements would characterize
conditions over areas with dimensions of as much
as hundreds of kilometers. As noted earlier, using
representative conditions for an area implies
some degree of homogeneity in that area. For this
reason, regional scale measurements would be
most applicable to sparsely populated areas. Data
characteristics of this scale would provide
information about larger scale processes of PM
emissions, losses and transport. PM25 transport
contributes to elevated particulate concentrations
and may affect multiple urban and State entities
with large populations such as in the eastern
United States. Development of effective pollution
control strategies requires an understanding at
regional geographical scales of the emission
sources and atmospheric processes that are
responsible for elevated PM25 levels and may also
be associated with elevated 03 and regional haze.
December 2009
A-55
-------
Table A-19. Major routine operating air monitoring networks9
Network Lead Number Initiated Measurement
Agency of Sites Parameters
Location of Information and/or Data
STATE /LOCAL I FEDERAL NETWORKS
NCoreb - National Core EPA 75 2008
Monitoring Network
SLAMS1 State and EPA ~3000 1978
Local Ambient
Monitoring Stations
STN-PM15 Speciation EPA 300 1999
Trends Network
PAMS— Photochemical EPA 75 1994
Assessment
Monitoring Network
IMPROVE NPS 110 plus 1988
Interagency Monitoring 67 protocol
of Protected Visual sites
Environments
CASTNet- Clean Air EPA 80+ 1987
Status and Trends
Network
GPMN-Gaseous NPS 33 1987
Pollutant Monitoring
Network
POMS-Portable NPS 14 2002
Ozone Monitoring
Stations
Passive Ozone NPS 43 1995
Sampler Monitoring
Program
NADP/NTN— National USGS 200+ 1978
Atmospheric
Deposition Program /
National Trends
Network
NADP/MDN— National None 90+ 1996
Atmospheric
Deposition Program /
Mercury Deposition
Network
03, NO/N02/NOV, S02,
CO, PM2.5/PM10-2.5,
PM25speciation, NH3,
HN03, surface
meteorology0
03, NOX/N02, S02,
PM2.5/PM10, CO, Pb
PM2.5, PM2.5
speciation, major
Ions, metals
03, NOx/NOy, CO,
speciated VOCs,
carbonyls, surface
meteorology SUpper
Air
PM2.5/PM10, major
ions, metals, light
extinction, scattering
coefficient
03, S02, major ions,
calculated dry
deposition, wet
deposition, total
deposition for
sulfur/nitrogen,
surface meteorology
03, N0x/N0/N02,
S02, CO, surface
meteorology, (plus
enhanced monitoring
of CO, NO, NOX,
NOY, and S02 plus
canister samples for
VOC at 3 sites)
03, surface
meteorology, with
CASTNet-protocol
filter pack (optional)
sulfate, nitrate,
ammonium, nitric
acid, sulfur dioxide
03 dose (weekly)
Major Ions from
precipitation
chemistry
Mercury from
precipitation
chemistry
http ://www.epa .gov/ttN/Amtic/monstratdoc .html
http://www.epa.gov/air/oaqps/qa/monproa.html
http://www.epa.gov/ttnamti1/specqen.html
http://www.epa.gov/air/oaqps/pams/
http://vista.cira.colostate.edu/IMPROVE/
http ://www.epa .gov/castnet/
http://www2.nature.nps.gov/air/Monitoring/network.cfmtfdata
http://www2.nature.nps.gov/air/studies/port03.cfm
http://www2.nature.nps.gov/air/Studies/Passives.cfm
http://nadp.sws.uiuc.edu/
http://nadp.sws.uiuc.edu/mdn/
December 2009
A-56
-------
Network Lead Number Initiated
Agency of Sites
AIRMoN— National NOAA 8 1992
Atmospheric
Deposition Program /
Atmospheric Integrated
Research Monitoring
Network
lADN-lntegrated EPA 20 1990
Atmospheric
Deposition Network
NAPS— National Air Canada 152+ 1969
Pollution Surveillance
Network
CAPMoN— Canadian Canada 29 2002
Air and Precipitation
Monitoring Network
Mexican Air Quality Mexico 52-62 Late
Network 1960s
Mexican City Ambient Mexico 49 Late
Air Quality Monitoring 1 960s
Network
Measurement
Parameters
Major Ions from
precipitation
chemistry
Note: some sites
began in 1976 as part
oftheDOEMAPSS
program; early data
are archived on
NADPandARL
servers.
PAHs, PCBs, and
organochlorine
compounds are
measured in air and
precipitation samples
S02, CO, 03, NO,
N02, NOX, VOCs,
SVOCs,PM10, PM25,
TSP, metals
03, NO, N02, NOY,
PAN,NH3,PM2.5,
PMio and coarse
fraction mass, PM2.5
speciation, major ions
for particles and trace
gases, precipitation
chemistry for major
ions
03, NOX, CO, S02,
PM10,TSP,VOC
03, NOx, CO, S02,
PM10,TSP,VOC
Location of Information and/or Data
http://nadp.sws.uiuc.edu/AIRMoN/
http://www.epa.aov/alnpo/monitoring/air/
http://www.etc-cte.ec.gc.ca/NAPS/index e.html
http://www.msc.ec.gc.ca/capmon/index e.cfm
http://www.ine.aob.mx/daicur/calaire/indicadores.html
http://www.ine.aob.mx/daicur/calaire/indicadores.html
AIR TOXICS MONITORING NETWORKS
NATTS— National Air EPA 23 2005
Toxics Trends Stations
State/Local Air Toxics EPA 250+ 1987
Monitoring
NDAMN-National EPA 34 1998-2005
Dioxin Air Monitoring
Network
VOCs, Carbonyls,
PM,o metalsd, Hg
VOCs, Carbonyls,
PM,o metalsd, Hg
CDDs, CDFs, dioxin-
like PCBs
http://www.epa.aov/ttN/Amtic/airtoxpg.html
http://www.epa.gov/ttN/Amtic/airtoxpg.html
http://cfpub.epa.gov/ncea/CFM/recordisplav.cfm?deid=54811
TRIBAL MONITORING NETWORKS
Tribal Monitoring' EPA 120+ 1995
03, NOX/N02, S02,
PM2.5/PM10, CO, Pb
http://www.epa.gov/air/tribal/airprogs.htmltfambmon
INDUSTRY/ RESEARCH NETWORKS
New Source Permit None variable variable
Monitoring
HRM Network None 9 1980
Houston Regional
Monitoring Network
ARIES /SEARCH— None 8 1992
Aerosol Research
Inhalation
Epidemiology Study /
Southeastern Aerosol
Research and
Characterization Study
experiment
03, NOX/N02, S02,
PM2.5/PM10, CO, Pb
03, NOx, PM25/PM10,
CO,S02, Pb.VOCs,
surface meteorology
03, NO/N02/NOY,
S02, CO, PM2.5/PM10,
PM2.5 speciation,
major Ions, NH3,
HN03, scattering
coefficient, surface
meteorology
Contact specific industrial facilities
http://hrm.radian.com/houston/how/index.htm
http://www.atmospheric-research.com/studies/SEARCH/index.html
December 2009
A-57
-------
Network Lead Number Initiated Measurement
Agency of Sites Parameters
SOS-SERON- EPA -40
Southern Oxidant
Study - Southeastern
Regional Oxidant
Networks
1990 03, NO, NOY, VOCs,
CO, surface
meteorology
Location of Information and/or Data
http://www.ncsu.edu/sos/pubs/sos3/State of SOS
3.pdf
NATIONAL/GLOBAL RADIATION NETWORKS
RadNet formerly EPA 200+
Environmental
Radiation Ambient
Monitoring System
(ERAMS)
SASP Surface Air DHS 41
Sampling Program
NEWNET DOE 26
Neighborhood
Environmental Watch
Network
1973 Radionuclides and
radiation
1963 89Sr, 90Sr, naturally
occurring
radionuclides, 7Be,
210Pb
1993 Ionizing gamma
radiation, surface
meteorology
http://www.epa.aov/enviro/html/erams/
http://www.eml.st.dhs.gov/databases/sasp/
http://newnet.lanl.gov/
SOLAR RADIATION NETWORKS
UV Index EPASunrise EPA
Program9
UV Net Ultraviolet EPA
Monitoring Program
NEUBrew (NOAA-EPA NOAA
Brewer
Spectrophotometer UV
and Ozone Network
UV-B Monitoring and USDA
Research Program
SURFRAD- Surface NOAA
Radiation Budget
Network
AERONET Aerosol NASA
RObotic NETwork co-
located
networks
MPLNET Micro-pulse
Lidar Network
PRIMENet-Park NPS
Research
& Intensive Monitoring of
Ecosystems NETworl?
-50 U.S. 2002 Calculated UV
cities radiation index
21 1995/2004 Ultraviolet solar
radiation (UV-B and
UV-A bands),
irradiance, ozone,
N02
6 2005 Ultraviolet solar
radiation (UV-B and
UV-A bands),
irradiance, ozone,
S02
35 1992 Ultraviolet-B radiation
7 1993 Solar and infrared
radiation, direct and
diffuse solar
radiation,
photosynthetically
active radiation, UVB,
spectral solar, and
meteorological
parameters
22 + other 1998 Aerosol spectral
participants optical depths,
aerosol size
distributions, and
precipitable water
8 2000 Aerosols and cloud
layer heights
14 1997 ozone, wet and dry
deposition, visibility,
surface meteorology,
and ultraviolet
radiation
http://www.epa.gov/sunwise/uvindex.html
http://www.epa.gov/uvnet/access.html
http://www.esrl.noaa.gov/gmd/grad/neubrew/
http://uvb.nrel.colostate.edu/UVB/index.isf
http://www.srrb.noaa.gov/surfrad/index.html
http://aeronet.gsfc.nasa.gov/index.html
http://mplnet.gsfc.nasa.gov/
http://www.cfc.umt.edu/primenet/Assets/Announcements/99PReport.pdf
3Some networks listed separately may also serve as subcomponents of other larger listed networks; as a result, some double counting of the number of individual monitors is likely.
bNCore is a network proposed to replace NAMS, as a component of SLAMS; NAMS are currently designated as national trends sites.
Surface meteorology includes wind direction and speed, temperature, precipitation, relative humidity, solar radiation (PAMS only).
dPM10 metals may include arsenic, beryllium, cadmium, chromium, lead, manganese, nickel, and others.
eThe number of sites indicated for tribal monitoring is actually the number of monitors, rather than sites. The number of sites with multiple monitors is probably <80.
f Sunrise program estimates UV exposure levels through modeling - does not include measurements.
9NEUBREW is a subset Original UV brewer network (UV Net); PRIMENET participated in UV Net program.
December 2009
A-58
-------
A.1.3. Monitor Distribution with Respect to Population Density
Atlanta Combined Statistical Area
: Kilometers
0 5 10 20 30 40 50
01
2005 Population Density
| Atlanta PMzs Monitors (15 km buffer)
Population per Sq Km
j^B °-89
j^l 90-177
178-886
887-1772
| 1773-4431
^H 4432-17722
0 15 30 60 90 120 150
: Kilometers
Figure A-1. PM2.s monitor distribution in comparison with population density, Atlanta, GA.
December 2009
A-59
-------
Atlanta Combined Statistical Area
l Kilometers
0 5 10 20 30 40 50
Q/\
r
2005 Population Density
| | Atlanta PMio Monitors (15 km buffer)
Population per Sq Km
^H °-89
^B 90-177
178-886
887-1772
1773-4431
• 4432-17722
0 15 30 60 90 120 150
: Kilometers
Figure A-2. PMio monitor distribution in comparison with population density, Atlanta, GA.
December 2009
A-60
-------
Birmingham Combined Statistical Area
0 5 10 20 30 40 50
01
] Kilometers
2005 Population Density
| Birmingham PIVb.s Monitors (15 km buffer)
Population per Sq Km
^^| 4-23
24-47
48 - 235
| 236-469
| 470-1173
I 1174-4692
0 15 30 60 90 120 150
] Kilometers
Figure A-3. PM2.s monitor distribution in comparison with population density, Birmingham,
AL
December 2009
A-61
-------
Birmingham Combined Statistical Area
0 5 10 20 30 40 50
01
] Kilometers
2005 Population Density
| Birmingham PMio Monitors (15 km buffer)
Population per Sq Km
^^| 4-23
24-47
48 - 235
| 236-469
| 470-1173
I 1174-4692
0 15 30 60 90 120 150
] Kilometers
Figure A-4. PM10 monitor distribution in comparison with population density, Birmingham,
AL
December 2009
A-62
-------
Boston Combined Statistical Area
] Kilometers
0 5 10 20 30 40 50
A
: Kilometers
0 15 30 60 90 120 150
2005 Population Density
| || Boston PM2.5 Monitors (15km buffer)
Population per Sq Km
^H 0-251
H252 - 5°2
503 - 2508
2509-5016
5017-12539
^B 12540-50155
Figure A-5. PM2.s monitor distribution in comparison with population density, Boston, MA.
December 2009
A-63
-------
Boston Combined Statistical Area
] Kilometers
0 5 10 20 30 40 50
A
01
2005 Population Density
| || Boston PM-io Monitors (15 km buffer)
Population per Sq Km
^H 0-251
252 - 502
503 - 2508
2509-5016
5017-12539
^B 12540-50155
: Kilometers
0 15 30 60 90 120 150
Figure A-6. PM10 monitor distribution in comparison with population density, Boston, MA.
December 2009
A-64
-------
Chicago Combined Statistical Area
] Kilometers
0 5 10 20 30 40 50
2005 Population Density
| Chicago PIVh.s Monitors (15 km buffer)
Population per Sq Km
^B ° - 22°
221 - 441
| 442 - 2204
2205 - 4407
4408-11019
I 11020-44074
0 15 30 60 90 120 150
] Kilometers
Figure A-7. PM2.s monitor distribution in comparison with population density, Chicago, IL.
December 2009
A-65
-------
Chicago Combined Statistical Area
] Kilometers
0 5 10 20 30 40 50
2005 Population Density
| Chicago PMio Monitors (15 km buffer)
Population per Sq Km
^B ° - 22°
^B221 -441
442 - 2204
2205 - 4407
4408-11019
I 11020-44074
0 15 30 60 90 120 150
] Kilometers
Figure A-8. PMio monitor distribution in comparison with population density, Chicago, IL.
December 2009
A-66
-------
Denver Combined Statistical Area
: Kilometers
0 5 10 20 30 40 50
2005 Population Density
^^ Denver PIVh.s Monitors (15 km buffer)
Population per Sq Km
^B °-67
^B 33-135
136-673
674-1347
| 1348-3364
I 3365-13456
: Kilometers
0 15 30 60 90 120 150
Figure A-9. PM2.s monitor distribution in comparison with population density, Denver, CO.
December 2009
A-67
-------
Denver Combined Statistical Area
: Kilometers
0 5 10 20 30 40 50
2005 Population Density
Denver PMio Monitors (15km buffer)
Population per Sq Km
^B °-67
^B 68-135
| 136-673
674-1347
| 1348-3364
I 3365-13456
: Kilometers
0 15 30 60 90 120 150
Figure A-10. PMio monitor distribution in comparison with population density, Denver, CO.
December 2009
A-68
-------
Detroit Combined Statistical Area
0 5 10 20 30 40 50
: Kilometers
Ql\
r
2005 Population Density
^^ Detroit PM2.5 Monitors (15 km buffer)
Population per Sq Km
^B °-164
^B 135-327
| 328-1637
1638-3274
| 3275-8185
I 8186-32741
0 15 30 60 90 120 150
: Kilometers
Figure A-11. PM2.s monitor distribution in comparison with population density, Detroit, Ml.
December 2009
A-69
-------
Detroit Combined Statistical Area
0 5 10 20 30 40 50
: Kilometers
Q/\
r
2005 Population Density
| Detroit PMio Monitors (15 km buffer)
Population per Sq Km
^B °-164
^B 165-327
328-1637
1638-3274
3275-8185
8186-32741
0 15 30 60 90 120 150
] Kilometers
Figure A-12. PMio monitor distribution in comparison with population density, Detroit, Ml.
December 2009
A-70
-------
Houston Combined Statistical Area
0 5 10 20 30 40 50
] Kilometers
Q/\
r
2005 Population Density
| Houston PMz5 Monitors (15 km buffer)
Population per Sq Km
|^B 0-132
JH 183-364
365-1822
1823-3644
3645-9109
I 9110-36435
] Kilometers
0 1530 60 90 120 150
Figure A-13. PM2.s monitor distribution in comparison with population density, Houston, TX.
December 2009
A-71
-------
Houston Combined Statistical Area
0 5 10 20 30 40 50
Q/\
r
] Kilometers
2005 Population Density
| Houston PMio Monitors (15 km buffer)
Population per Sq Km
|^B 0-132
JH 183-364
365-1822
1823-3644
3645-9109
I 9110-36435
] Kilometers
0 1530 60 90 120 150
Figure A-14. PMio monitor distribution in comparison with population density, Houston, TX.
December 2009
A-72
-------
Los Angeles Core Based Statistical Area
0 5 10
A
] Kilometers
30 40 50
2005 Population Density
|| Los Angeles PMzs Monitors (15 km buffer)
Population per Sq Km
^Bl °-271
^H 272 - 542
543-2711
2712-5422
5423-13556
13557-54222
0 15 30 60 90 120 150
] Kilometers
Figure A-15. PM2.s monitor distribution in comparison with population density, Los Angeles,
CA.
December 2009
A-73
-------
Los Angeles Core Based Statistical Area
0 5 10
] Kilometers
30 40 50
2005 Population Density
|| Los Angeles PMio Monitors (15km buffer)
Population per Sq Km
^Bl °-271
^H 272 - 542
543-2711
2712-5422
5423-13556
13557-54222
0 15 30 60 90 120 150
] Kilometers
Figure A-16. PM10 monitor distribution in comparison with population density, Los Angeles,
CA.
December 2009
A-74
-------
New York Combined Statistical Area
: Kilometers
0 5 10 20 30 40 50
2005 Population Density
|| New York PM2.5 Monitors (15 km buffer)
Population per Sq Km
^Bl ° -832
^H 833-1664
1665-8319
8320-16637
16638-41593
41594-166371
] Kilometers
0 1530 60 90 120 150
Figure A-17. PM2.s monitor distribution in comparison with population density, New York, NY.
December 2009
A-75
-------
New York Combined Statistical Area
: Kilometers
0 5 10 20 30 40 50
2005 Population Density
[| | New York PMio Monitors (15 km buffer)
Population per Sq Km
^B ° -S32
^B 833-1664
| 1665-8319
8320 -16637
^B 16638-41593
• 41594-166371
] Kilometers
0 1530 60 90 120 150
Figure A-18. PM10 monitor distribution in comparison with population density, New York, NY.
December 2009
A-76
-------
Philadelphia Combined Statistical Area
i Kilometers
0 5 10
20 30
40
50
2005 Population Density
| | Philadelphia PIVh.s Monitors (15 km buffer)
Population per Sq Km
^B 0-133
^B 184-366
367-1829
1830-3658
3659 - 9144
^H 9145-36577
0 15 30 60 90 120 150
] Kilometers
Figure A-19. PM2.s monitor distribution in comparison with population density, Philadelphia,
PA.
December 2009
A-77
-------
Philadelphia Combined Statistical Area
0 5 10 20 30 40 50
i Kilometers
2005 Population Density
| | Philadelphia PMio Monitors (15 km buffer)
Population per Sq Km
^B 0-133
^B 184-366
367-1829
1830-3658
3659 - 9144
^H 9145-36577
0 15 30 60 90 120 150
] Kilometers
Figure A-20. PM10 monitor distribution in comparison with population density, Philadelphia,
PA.
December 2009
A-78
-------
Phoenix Core Based Statistical Area
0 5 10 20 30 40 50
] Kilometers
2005 Population Density
| Phoenix PIVhs Monitors (15 km buffer)
Population per Sq Km
81 -159
160-795
796-1591
1592-3977
3978-15907
: Kilometers
0 15 30 60 90 120 150
Figure A-21. PM2.s monitor distribution in comparison with population density, Phoenix, AZ.
December 2009
A-79
-------
Phoenix Core Based Statistical Area
0 5 10 20 30 40 50
] Kilometers
2005 Population Density
| Phoenix PMio Monitors (15 km buffer)
Population per Sq Km
0-80
^B 81 ~159
| 160-795
796-1591
1592-3977
I 3978-15907
: Kilometers
0 15 30 60 90 120 150
Figure A-22. PM™ monitor distribution in comparison with population density, Phoenix, AZ.
December 2009
A-80
-------
Pittsburgh Combined Statistical Area
0 5 10 20 30 40 50
Ql\
r
] Kilometers
2005 Population Density
[ | Pittsburgh PIVh.s Monitors (15 km buffer)
Population per Sq Km
^B 6 -204
205 - 409
| 410-2045
2046 - 4090
4091 -10225
I 10226-40898
0 15 30 60 90 120 150
l Kilometers
Figure A-23. PM2.s monitor distribution in comparison with population density, Pittsburgh, PA.
December 2009
A-81
-------
Pittsburgh Combined Statistical Area
] Kilometers
0 5 10 20 30 40
50
Ql\
r
2005 Population Density
Pittsburgh PMio Monitors (15 km buffer)
Population per Sq Km
^B 6 -204
205 - 409
| 410-2045
2046 - 4090
4091 -10225
I 10226-40898
0 15 30 60 90 120 150
: Kilometers
Figure A-24. PM™ monitor distribution in comparison with population density, Pittsburgh, PA.
December 2009
A-82
-------
Riverside Core Based Statistical Area
] Kilometers
0510 20 30 40 50
] Kilometers
0 1530 60 90 120 150
2005 Population Density
| Riverside PMzs Monitors (15 km buffer)
Population per Sq Km
0-89
^B 90-178
| 179-892
893-1784
| 1785-4460
• 4461 -17838
Figure A-25. PM2.s monitor distribution in comparison with population density, Riverside, CA.
December 2009
A-83
-------
Riverside Core Based Statistical Area
] Kilometers
0510 20 30 40 50
: Kilometers
012.95 50 75 100 125
2005 Population Density
| Riverside PMio Monitors (15 km buffer)
Population per Sq Km
0-89
^B 90-178
| 179-892
893-1784
| 1785-4460
• 4461 -17838
Figure A-26. PM™ monitor distribution in comparison with population density, Riverside, CA.
December 2009
A-84
-------
Seattle Combined Statistical Area
: Kilometers
0 5 10 20 30 40 50
2005 Population Density
|j Seattle PIVh.s Monitors (15km buffer)
Population per Sq Km
0-120
121-240
241 -1201
1202-2402
2403 - 6006
6007 - 24022
0 15 30 60 90 120 150
l Kilometers
Figure A-27. PM2.s monitor distribution in comparison with population density, Seattle, WA.
December 2009
A-85
-------
Seattle Combined Statistical Area
: Kilometers
0 5 10 20 30 40 50
2005 Population Density
[| |] Seattle PMio Monitors (15 km buffer)
Population per Sq Km
^m 0-120
^B 121 ~24°
241 -1201
1202-2402
2403 - 6006
^H 6007 - 24022
0 15 30 60 90 120 150
l Kilometers
Figure A-28. PMio monitor distribution in comparison with population density, Seattle, WA.
December 2009
A-86
-------
St. Louis Combined Statistical Area
0 5 10 20 30 40 50
: Kilometers
2005 Population Density
| StLouis PM2.5 Monitors (15 km buffer)
Population per Sq Km
^B °-54
^B 55-109
110-544
545-1088
| 1089-2720
• 2721 -10878
0 15 30 60 90 120 150
: Kilometers
Figure A-29. PM2.s monitor distribution in comparison with population density, St. Louis, MO.
December 2009
A-87
-------
St. Louis Combined Statistical Area
0 5 10 20 30 40 50
l Kilometers
2005 Population Density
| StLouis PMio Monitors (15 km buffer)
Population per Sq Km
0-54
55-109
110-544
545-1088
1089-2720
2721 -10878
0 15 30 60 90 120 150
: Kilometers
Figure A-30. PM™ monitor distribution in comparison with population density, St. Louis, MO.
December 2009
A-88
-------
A.2. Ambient PM Concentration
A.2.1. Speciation Trends Network Site Data
Copper - Annual
copper(ug/m3): <=0.002 >0.002 to 0.004 >0.004 to 0.006 >0.006-0.008 >0.008
Copper - Spring
r.-;-ip-r(i -,'i.' ~i) =r nn; n •in; tr. n c J4
Copper - Summer
Copper -
O « «s
• Jj^r
Figure A-31. Three-yr avg of 24-h PM2.s Cu concentrations measured at CSN sites across the
U.S., 2005-2007.
December 2009
A-89
-------
Iron - Annual
ron(ug/m3): <=0.05 >0.05 to 0.075 >0.075 to 0.1 >0.1-0.125 >0.126
Iron ~~ SpnnQ
0 05 to 0.075 >0.075toO 1
•
>Q.1-0.125
Iron ~ Summer
r\
•
*
°-
V
•^ M^H ~* » -^' N. \
^f^
V
o • •
>0.075to0.1 >0.1-0.125 >0.125
Iron - Fall
Iran - Winter
ron(ug/m3): <=0.05 >0.05 to 0.075 >0.075 to 0.1 >0 1-0.125 >0.125
Figure A-32. Three-yr avg of 24-h PM2.s Fe concentrations measured at CSN sites across the
U.S., 2005-2007
December 2009
A-90
-------
Ni - Annual
Nickel(ug/m3): <=0.0005 >0.0005 to 0.001 >0.001 to 0.0015 >0.0015-0.002 >0.002
NI - Spring
Nickel(ug/m3): <=00005 >O.ODD5toD 001 >0.001 to 0.0015 >00015-0.002 >0 002
Nickd(ug/m3): <=0.0005 >0.0005 toO.001 >0001to00015 >0.0015-O.D02 >O.D02
NI -
NI - Winter
0
oo o •
Nlckd(ug/m3): <-0.0005 >0 0005 too.001 >0.001 to 0.0015 >O.D015-0002
.o C • •
Nlckel(og/m3). <-0.0005 >0 0005 to 0.001 >0.001 to 0.0015 >0.0015-0002 >0 002
Figure A-33. Three-yr avg of 24-h PM2.s Ni concentrations measured at CSN sites across the
U.S., 2005-2007
December 2009
A-91
-------
Lead - Annual
Lead(ug/m3): <=0.005 >0.005to0.01
O
>0.01 to 0.015
Lead - Spring
Leadfug/tTi3). <=0005 >O.OQ5 to 0.01 >Q.Q1 to 0.015 >0 015-0.02 sO.02
Lead - Summer
Lead(ug/m3). <=QB05 >O.Q05 to D 01 s001 to 0.015 >0015-0 02
Lead - Fall
Lead - Winter
Lead(ug/m3): <=0.005 >0.005to0.01 >OD1 to 0.015 >0.015-0.02
Lead(ug/m3). <=Q 005 >O.OD5 to 0.01 >0 01 to 0.015 >0.015-0 02
Figure A-34. Three-yr avg of 24-h PM2.s Pb concentrations measured at CSN sites across the
U.S., 2005-2007
December 2009
A-92
-------
Selenium - Annual
Selenium(ug/m3): <=0.001 >0.001 to 0.002 >0.002 to 0.003 >0.003-0.004 >0.004
Selenium - Spring
Selenium (ug/mS)1 <=O.OQ1 >0 001 to 0.002 >O.OD2 to D 003 >0 003-0 004 >0.004
Selenium - Fail
Selemum(ug/m3) <=D.DD1 >OOQ1toD.OD2 >D.DD2toOD03 SO 003-0 004 >O.OM
Selenium - Summer
Łide''ijrn(ug,'m3): <=0 001 >0 001 to 0.002 >0 002 to OJ303 >O.OD3-0.004 >O.D04
Selenium — Winter
(uq/m3): <=0.001 >0.001 to 0.002 >O.D02 to 0003 >0 003-0 004 >0.004
Figure A-35. Three-yr avg of 24-h PM2.s Se concentrations measured at CSN sites across the
U.S., 2005-2007
December 2009
A-93
-------
vanadium - Annual
Vanadium(ug/m3): <=0.004 »0.004 to 0.006 »0.006 to 0.008 >0.008-0.01 >0.01
vanadium - Spring
Vanadium (ug/in 3)- <=0.004 >OOQ4tO 0.006 >0.006to 0.008 >0 008-0.01 >0.01
Vanadium - Summer
V:-.na Jiurr, ^gym 3) <=Q.Q04 >Q.OQ4 to 0.006 >0 006 to 0 I
Vanadium - Fall
Vanadium - Winter
Vanadium(ug/m3): <=0.004 >0.004to 0.006 >0.006to 0.008 >O.OOS-0 01 >O.D1
3 o • •
>0.004tO 0.006 >0.006tO 0 008 >0.000-0.01 >0.01
Figure A-36. Three-yr avg of 24-h PM2.s V concentrations measured at CSN sites across the
U.S., 2005-2007
December 2009
A-94
-------
A.2.2. Intraurban Variability
The following figures and tables exemplify the intraurban variability among PM2.5, PMi0_2.5
and PMio measurements for select CSAs/CBSAs (2005-2007) including Atlanta, Birmingham,
Chicago, Denver, Detroit, Houston, New York City, Philadelphia, Phoenix, Riverside, Seattle and St.
Louis. Maps are included to show monitor locations relative to major roadways. Box plots show the
median and interquartile range of concentrations with whiskers extending to the 5th and 95th
percentiles at each site during (1) winter (December-February); (2) spring (March-May); (3) summer
(June-August); and (4) fall (September-November). Tables of inter-sampler comparison statistics and
scatter plots of inter-sampler correlation vs. distance illustrate variability present in each area.
December 2009 A-95
-------
Atlanta Combined Statistical Area
A
01
Atlanta CSA
• PM2 5 Monitors
Interstate Highways
- Major Highways
0 10 20 40 60 80
100
—i Kilometers
Figure A-37. PM2.s monitor distribution and major highways, Atlanta, GA.
December 2009
A-96
-------
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
SiteG
SiteH
Site I
13-063-0091
13-067-0003
13-067-0004
13-089-0002
13-089-2001
13-121-0032
13-135-0002
13-139-0003
13-223-0003
AB CDEFG H
Mean 16.2 16.2 15.4 15.3 15.2 15.7 15.2 13.9 14.4
Obs 351 352 339 1014 946 1036 221 336 344
SD 7.5 7.9 7.7 7.2 7.6 8.2 7.1 6.9 7.6
40 -
30-
20 -
10-
o -
Ol
c
o
4-*
ro
•*-*
C
u
c
o
1=winter
2=spring
3=summer
4=fall 1234 1234 1234 1234 1234 1234 1234 1234 1234
Figure A-38. Box plots illustrating the seasonal distribution of 24-h avg PM2.s concentrations
for Atlanta, GA.
December 2009
A-97
-------
Table A-20. Inter-sampler correlation statistics for each
Atlanta, GA.
A B C D
A 1.00 0.88 0.87 0.93
(0.0,0.00) (5.2,0.11) (6.2,0.12) (3.9,0.11)
351 330 310 330
B 1.00 0.96 0.89
(0.0,0.00) (4.1,0.08) (5.7,0.12)
352 309 327
C 1.00 0.87
(0.0,0.00) (5.2,0.12)
339 315
D 1.00
(0.0, 0.00)
1014
E
LEGEND
F R
(P90, COD)
N
G
H
I
E
0.89
(5.3,0.12)
315
0.88
(4.6,0.10)
314
0.86
(5.6,0.11)
304
0.89
(4.8,0.12)
883
1.00
(0.0, 0.00)
946
pair of PM2.6 monitors reporting to AQS for
F
0.91
(4.6,0.11)
334
0.91
(3.6, 0.08)
333
0.88
(4.4,0.10)
324
0.80
(3.7,0.11)
978
0.79
(3.8,0.11)
904
1.00
(0.0, 0.00)
1036
G
0.85
(6.9,0.15)
207
0.88
(5.6,0.13)
205
0.85
(5.8,0.13)
193
0.87
(5.8,0.13)
208
0.88
(5.3,0.12)
208
0.88
(5.3,0.12)
213
1.00
(0.0, 0.00)
221
H
0.72
(8.7,0.19)
319
0.78
(9.0,0.17)
313
0.79
(7.9,0.17)
298
0.74
(8.3,0.18)
314
0.74
(7.8,0.17)
305
0.70
(8.5,0.19)
321
0.73
(8.8,0.17)
195
1.00
(0.0, 0.00)
336
I
0.85
(7.2,0.15)
326
0.88
(6.5,0.13)
321
0.90
(4.5,0.11)
303
0.82
(7.3,0.15)
322
0.83
(6.4,0.14)
309
0.84
(6.3,0.14)
327
0.79
(7.4,0.15)
198
0.76
(8.7,0.17)
309
1.00
(0.0, 0.00)
344
December 2009 A-98
-------
1.00 -
0.80
0.60
o
13
o
O
0.40
0.20
•**•*<•**?* ^
0.00
10 20 30
40 50 60
Distance Between Samplers (km)
70 80 90 100
Figure A-39. PM2.s inter-sampler correlations as a function of distance between monitors for
Atlanta, GA.
December 2009
A-99
-------
Birmingham Combined Statistical Area
A
01
Birmingham CSA
• PM2.5 Monitors
Interstate Highways
— Major Highways
0 10 20 40 60 80
100
—i Kilometers
Figure A-40. PM2.s monitor distribution and major highways, Birmingham, AL.
December 2009
A-100
-------
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
SiteG
SiteH
Sitel
SiteJ
AQS Site ID
01-073-0023
01-073-1005
01-073-1009
01-073-1010
01-073-2003
01-073-2006
01-073-5002
01-073-5003
01-117-0006
01-127-0002
A
Mean 18.9
Obs 1087
SD 102
50 -
40 -
c
o
30 -
20 -
c
01
u
c
o
u
10 -
1=winter
2=spring
3=summer 0 -
4=fall
ABCDEFGH J
18.9 15.9 14.1 16.2 17.7 15.4 14.8 14.9 14.6 14.4
1087 363 359 182 1079 364 363 360 361 351
10.2 8.2 8.2 8.1 9.4 7.8 7.9 8.3 7.6 7.5
1234 1234 1234 1234 1234 1234 1234 1234 1234 1234
Figure A-41. Box plots illustrating the seasonal distribution of 24-h avg PM2.s concentrations
for Birmingham, AL
December 2009
A-101
-------
Table A-21 . Inter-sampler correlation statistics for each pair of PM2.6 monitors reporting to AQS for
Birmingham, AL
ABC
A 1.00 0.91 0.86
(0.0,0.00) (10.4,0.15) (13.7,0.21)
1087 360 356
B 1.00 0.93
(0.0,0.00) (5.3,0.12)
363 356
C 1.00
(0.0, 0.00)
359
D
E
LEGEND
F R
(P90, COD)
N
G
H
I
J
D E F
0.91 0.88 0.91
(9.7,0.13) (8.1,0.13) (10.8,0.15)
182 1072 361
0.93 0.85 0.96
(4.7,0.09) (8.3,0.15) (3.6,0.08)
181 359 358
0.93 0.81 0.93
(5.9,0.13) (10.1,0.20) (4.6,0.12)
180 355 354
1.00 0.88 0.96
(0.0,0.00) (7.9,0.12) (3.6,0.08)
182 179 179
1.00 0.87
(0.0,0.00) (8.1,0.15)
1079 360
1.00
(0.0, 0.00)
364
G
0.87
(12.6,0.18)
360
0.91
(5.4,0.11)
360
0.91
(4.3,0.12)
355
0.95
(3.8, 0.09)
181
0.85
(8.7,0.16)
359
0.95
(3.9, 0.09)
359
1.00
(0.0, 0.00)
363
H
0.88
(11.7,0.18)
357
0.93
(5.1,0.11)
355
0.94
(4.0,0.10)
350
0.95
(4.7,0.10)
179
0.85
(8.8,0.17)
356
0.95
(4.1,0.10)
354
0.96
(3.3, 0.08)
356
1.00
(0.0, 0.00)
360
I
0.88
(12.3,0.18)
358
0.93
(4.9,0.10)
358
0.90
(4.9,0.12)
353
0.93
(4.7,0.10)
180
0.86
(9.2,0.16)
357
0.95
(3.4, 0.09)
357
0.92
(4.5,0.10)
359
0.91
(5.0,0.11)
354
1.00
(0.0, 0.00)
361
J
0.84
(12.5,0.19)
348
0.89
(6.1,0.12)
348
0.90
(4.9,0.11)
343
0.89
(6.1,0.12)
174
0.81
(10.6,0.18)
347
0.90
(5.6,0.11)
348
0.89
(4.9,0.11)
350
0.93
(4.3, 0.09)
344
0.87
(5.8,0.12)
349
1.00
(0.0, 0.00)
351
December 2009 A-102
-------
1.00
0.80
* » * » »
A A • • 4
* * » » * »
0.60
o
13
o
O
0.40
0.20
0.00
10 20 30 40 50 60
Distance Between Samplers (km)
70 80 90 100
Figure A-42. PM2.s inter-sampler correlations as a function of distance between monitors for
Birmingham, AL
December 2009
A-103
-------
Boston Combined Statistical Area
01
r
Boston CSA
• PM2 5 Monitors
Interstate Highways
— Major Highways
0 10 20 40 60 80
100
—i Kilometers
Figure A-43. PM2.s monitor distribution and major highways, Boston, MA.
December 2009
A-104
-------
SiteA
SteB
SteC
SteD
SteE
SiteF
SiteG
SiteH
Shel
ShseJ
AQS Site ID
25-005-1004
25-009-2006
25-009-5005
25-009-6001
25-023-0004
25-025-0002
25-025-0027
25-025-0042
25-025-0043
25-027-0016
1 -winter
2-spring
3=summer
4=fall
SrtEK
SiteL
SiteM
SiteN
Sited
SrteP
SiteQ
SiteR
SteS
AQS Site ID
25-027-0023
33-001-2004
33-011-1015
33-013-1006
33-015-0014
44-007-0022
44-007-0026
44-007-0028
44-007-1010
A B C D E F G H J
Ivtean 9.1 9.1 8.9 9.4 9.6 11.7 11.6 10.5 12.1 10.7
Obs 341 350 342 355 357 349 398 349 1015 335
SO
40-
ST- 30~
Ol
a.
c
o
ro 20-
concenti
10-
o -
6.0 6.5 6.3 6.6 6.6 7.0 6.8 6.9 6.9 7.2
1 1 1
' ill i 'I' 1 I
1234 1234 1234 1234 1234 1234 1234 1234 1234 1234
4=fall
K L
Mean 11.4 7.2
Obs 346 183
SD
40-
__. 30-
c
o
4_t 7 n —
concentra
lo-
iter
ing
Timer
0 -
7.4 5.3
1
|
M N O P Q R S
10.0 9.7 8.9 10.1 11.9 10.5 9.7
361 362 362 1027 321 313 998
6.7 6.0 5.8 6.6 6.9 6.5 6.5
1
'1
1
!
1234 1234 1234 1234 1234 1234 1234 1234 1234
Figure A-44. Box plots illustrating the seasonal distribution of 24-h avg PM2.s concentrations
for Boston, MA.
December 2009
A-105
-------
Table A-22. Inter-sampler correlation statistics for each pair of PM2.6 monitors reporting to AQS for
Boston, MA.
Site
A
B
A
1.00
(0.0, 0.00) (6J
341
B
0.80
6,0.21)
326
1.00
(0.0, 0.00)
C
D
E
F
G
H
1
J
LEGEND
Pearson R
(P90, COD)
n
350
C D
0.77 0.71
(6.2, 0.22) (6.9, 0.23)
318 323
0.92 0.87
(4.1,0.17) (4.1,0.18)
328 331
1.00 0.90
(0.0,0.00) (3.5,0.17)
342 321
1.00
(0.0, 0.00)
355
E
0.84
(4.8,0.19)
329
0.87
(4.7,0.19)
339
0.85
(5.3,0.21)
331
0.80
(5.6, 0.20)
336
1.00
(0.0, 0.00)
357
F
0.79
(8.1,0.23)
318
0.90
(6.3,0.21)
326
0.90
(6.3, 0.23)
316
0.88
(5.8,0.21)
324
0.90
(5.9,0.19)
330
1.00
(0.0, 0.00)
349
G
0.78
(7.7, 0.24)
319
0.90
(6.2, 0.23)
323
0.89
(6.3, 0.24)
318
0.88
(5.8, 0.22)
329
0.90
(5.8,0.21)
333
0.94
(3.8,0.14)
324
1.00
(0.0, 0.00)
398
H
0.79
(6.8, 0.22)
325
0.90
(4.9,0.19)
333
0.90
(5.0, 0.20)
326
0.86
(4.6,0.19)
332
0.89
(5.0,0.19)
340
0.94
(3.5,0.15)
324
0.94
(4.0,0.16)
325
1.00
(0.0, 0.00)
349
1
0.79
(7.9, 0.25)
338
0.90
(7.1,0.26)
343
0.88
(6.8, 0.26)
336
0.86
(7.0, 0.26)
345
0.87
(6.9, 0.24)
350
0.92
(4.5,0.17)
339
0.94
(4.3,0.15)
338
0.93
(4.7,0.19)
342
1.00
(0.0, 0.00)
1015
J
0.77
(7.5, 0.24)
310
0.85
(5.5,0.21)
317
0.86
(6.2,0.21)
311
0.87
(5.8,0.19)
313
0.87
(5.4, 0.20)
322
0.92
(5.4,0.18)
310
0.89
(5.7, 0.20)
308
0.89
(5.0,0.17)
318
0.86
(6.9, 0.23)
330
1.00
(0.0, 0.00)
335
Site
A
B
C
D
E
F
G
H
1
J
K
L
M
N
0
P
K
077
(8.1,0.23)
320
0.86
(6.6,0.21)
329
0.86
(6.9,0.21)
321
0.88
(6.4,0.19)
325
0.87
(6.3, 0.20)
333
0.91
(4.7,0.17)
323
0.90
(5.0,0.19)
320
0.90
(4.4,0.17)
327
0.87
(6.1,0.20)
341
0.95
(3.0,0.14)
316
1.00
(0.0, 0.00)
346
L
0.61
(8.3, 0.29)
173
0.80
(6.2, 0.23)
175
0.89
(4.8, 0.23)
173
0.79
(5.7, 0.25)
174
0.72
(8.3, 0.27)
179
0.78
(9.6, 0.33)
168
0.77
(9.0, 0.33)
172
0.75
(9.4, 0.30)
175
0.75
(10.0,0.36)
181
0.73
(9.2, 0.28)
167
0.71
(10.3,0.31)
170
1.00
(0.0, 0.00)
183
LEGEND
Pearson R
(P90, COD)
n
M
0.71
(8.0, 0.23)
324
0.87
(5.3,0.19)
331
0.93
(4.4,0.17)
323
0.91
(3.5,0.16)
329
0.83
(5.8,0.17)
338
0.90
(5.3,0.18)
323
0.90
(5.3,0.19)
326
0.88
(4.9,0.18)
332
0.86
(6.7, 0.22)
352
0.87
(5.2,0.18)
314
0.88
(6.0,0.16)
326
0.89
(6.7, 0.24)
176
1.00
(0.0, 0.00)
361
N
0.68
(7.9, 0.23)
334
0.83
(6.0,0.21)
341
0.90
(4.6,0.19)
335
0.85
(4.7,0.19)
339
0.79
(6.3, 0.20)
347
0.85
(6.4, 0.20)
334
0.85
(6.3, 0.20)
335
0.83
(5.6,0.21)
341
0.82
(7.2, 0.23)
356
0.84
(5.9, 0.20)
326
0.85
(6.5,0.19)
337
0.91
(5.9, 0.23)
181
0.94
(3.8,0.13)
341
1.00
(0.0, 0.00)
362
0
0.73
(7.0, 0.22)
331
0.88
(4.7,0.18)
336
0.93
(3.8,0.18)
328
0.86
(4.2,0.18)
334
0.84
(4.8,0.18)
343
0.85
(7.5, 0.22)
330
0.87
(7.0, 0.22)
329
0.84
(6.8,0.21)
336
0.83
(8.2, 0.25)
357
0.80
(7.5, 0.22)
323
0.81
(8.2, 0.22)
332
0.90
(4.8,0.21)
177
0.90
(4.6,0.16)
336
0.90
(4.4,0.17)
346
1.00
(0.0, 0.00)
362
P
0.87
(5.3,0.18)
326
0.86
(5.6,0.19)
335
0.83
(5.9,0.21)
329
0.80
(6.2, 0.20)
342
0.91
(4.5,0.17)
343
0.89
(5.2,0.16)
336
0.88
(5.5,0.17)
383
0.89
(4.5,0.16)
335
0.88
(6.1,0.20)
957
0.90
(5.0,0.17)
321
0.89
(5.2,0.16)
331
0.68
(10.0,0.29)
181
0.83
(5.5,0.16)
345
0.77
(6.7,0.19)
347
0.80
(5.8,0.19)
348
1.00
(0.0, 0.00)
Q
0.81
(7.2, 0.23)
292
0.80
(7.9, 0.26)
300
0.79
(7.8, 0.26)
290
0.75
(7.8, 0.25)
300
0.86
(6.3, 0.22)
306
0.86
(6.0,0.16)
295
0.86
(5.3,0.17)
296
0.86
(6.0,0.19)
299
0.84
(6.0,0.16)
314
0.86
(5.9, 0.20)
283
0.86
(5.8,0.18)
296
0.63
(12.1,0.35)
153
0.81
(7.4, 0.20)
300
0.75
(8.1,0.22)
309
0.75
(8.8, 0.25)
304
0.95
(3.6,0.14)
R
0.85
(5.6, 0.20)
285
0.85
(5.7,0.21)
288
0.81
(6.2, 0.23)
281
0.79
(6.2,0.21)
287
0.88
(4.9,0.18)
295
0.88
(4.9,0.16)
281
0.87
(5.2,0.17)
282
0.87
(4.5,0.16)
289
0.85
(6.0,0.18)
306
0.87
(5.3,0.17)
272
0.87
(5.5,0.16)
286
0.72
(9.1,0.30)
149
0.82
(5.8,0.17)
288
0.78
(6.4, 0.20)
297
0.79
(6.8,0.21)
292
0.97
(2.0, 0.09)
S
0.86
(5.2,0.18)
306
0.85
(6.0,0.19)
314
0.82
(6.0,0.21)
309
0.80
(5.8, 0.20)
321
0.91
(3.9,0.17)
324
0.89
(5.5,0.17)
316
0.88
(5.7,0.19)
356
0.88
(5.1,0.17)
314
0.87
(6.3,0.21)
936
0.88
(5.2,0.18)
302
0.88
(5.5,0.18)
313
0.69
(9.8, 0.29)
164
0.84
(5.1,0.16)
326
0.78
(6.2,0.19)
327
0.80
(6.0,0.19)
330
0.97
(2.1,0.08)
December 2009
A-106
-------
1027
307
(0.0, 0.00)
299
(3.1,0.13)
943
0.94
(4.0,0.16)
268
290
0.94
(0.0, 0.00)
(2.7,0.12)
280
1.00
(0.0, 0.00)
998
•
»
** *
0.8
0.6
o
1
0.4
0.2
10 20 30 40 50 60
Distance Between Samplers (km)
70
90
100
Figure A-45. PM2.s inter-sampler correlations as a function of distance between monitors for
Boston, MA.
December 2009
A-107
-------
Chicago Combined Statistical Area
A
01
Chicago CSA
• PMz.s Monitors
Interstate Highways
— Major Highways
0 10 20 40 60 80
100
—i Kilometers
Figure A-46. PM2.s monitor distribution and major highways, Chicago, IL
December 2009
A-108
-------
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
SiteG
SiteH
Site I
SiteJ
SiteK
AQS Site ID
17-031-0022
17-031-0050
17-031-0052
17-031-0057
17-031-0076
17-031-1016
17-031-2001
17-031-3103
17-031-3301
17-031-4007
17-031-4201
1234 1234 1234 1234 1234 1234 1234 1234 1234 1234 1234
SiteL
SiteM
SiteN
SiteO
SiteP
SiteQ
SiteR
SiteS
Site!
SiteU
AQS Site ID L
17-031-6005 Mean 151
17-043-4002 ..
Obs 331
17-089-0003
17-089-0007 8J
17-097-1007 50 -
17-111-0001
17-197-1002
17-197-1011
18-089-0006 40 -
18-089-0022
Dl
C
O
4-»
ro
•*-*
c
u
c
o
1=winter
2=spring
3=summer
4=fall
30 -
20 -
10 -
0 -
LMNOPQRSTU
i 15.1 14.0 13.5 14.3 12.1 12.3 14.0 11.7 14.4 15.6
3 331 179 176 174 181 347 175 164 330 351
1 8.7 8.5 8.5 8.9 8.2 7.5 8.5 7.2 7.7 8.2
1234 1234 1234 1234 1234 1234 1234 1234 1234 1234
December 2009
A-109
-------
SiteV
SiteW
SiteX
SiteY
SiteZ
Site AA
Site AB
Site AC
Site AD
Site AE
AQS Site ID
18-089-0026
18-089-0027
18-089-1003
18-089-2004
18-089-2010
18-091-0011
18-091-0012
18-127-0020
18-127-0024
55-059-0019
1234 1234 1234 1234 1234 1234 1234 1234 1234 1234
Figure A-47. Box plots illustrating the seasonal distribution of 24-h avg PM2.s concentrations
for Chicago, IL
December 2009
A-110
-------
Table A-23. Inter-sampler correlation statistics for each pair of PM2.6 monitors reporting to AQS for
Chicago, IL.
A B C D
A 1.00 0.98 0.93 0.94
(0.0,0.00) (3.1,0.08) (5.5,0.12) (4.7,0.11)
178 156 176 149
B 1.00 0.94 0.95
(0.0,0.00) (4.6,0.11) (3.6,0.10)
343 320 276
C 1.00 0.96
(0.0,0.00) (4.4,0.11)
984 313
D 1.00
(0.0, 0.00)
333
E
F
G
H
I
J
K
L
M
N
0
E
0.97
(3.9, 0.09)
154
0.97
(3.3, 0.08)
300
0.92
(5.7,0.11)
325
0.94
(3.8,0.10)
286
1.00
(0.0, 0.00)
351
F
0.95
(5.7,0.13)
154
0.95
(5.2,0.13)
296
0.91
(4.8,0.11)
318
0.93
(4.2,0.12)
280
0.95
(5.0,0.11)
306
1.00
(0.0, 0.00)
345
LEGEND
R
(P90, COD)
G
0.97
(3.9, 0.09)
151
0.97
(2.7, 0.07)
296
0.90
(6.0,0.12)
324
0.94
(3.8,0.10)
283
0.98
(2.4, 0.06)
304
0.95
(5.1,0.12)
301
1.00
(0.0, 0.00)
350
H
0.94
(4.6,0.12)
156
0.95
(4.3,0.11)
289
0.94
(4.3,0.11)
312
0.95
(4.1,0.13)
270
0.95
(4.5,0.11)
292
0.95
(4.5,0.12)
294
0.95
(4.9,0.12)
284
1.00
(0.0, 0.00)
335
I
0.96
(4.2,0.11)
164
0.96
(3.4, 0.09)
312
0.92
(5.5,0.11)
336
0.94
(3.3,0.10)
299
0.98
(2.6, 0.07)
320
0.96
(4.5,0.10)
322
0.97
(3.0, 0.07)
315
0.95
(4.3,0.11)
311
1.00
(0.0, 0.00)
361
J
0.91
(6.8,0.16)
163
0.93
(6.3,0.16)
315
0.90
(8.8,0.18)
332
0.93
(6.2,0.15)
296
0.92
(5.8,0.16)
321
0.89
(8.5, 0.20)
323
0.90
(6.3,0.15)
318
0.91
(7.4,0.19)
309
0.90
(6.7,0.17)
341
1.00
(0.0, 0.00)
356
K
0.95
(5.8,0.14)
166
0.93
(6.5,0.15)
306
0.91
(7.2,0.17)
337
0.93
(5.2,0.14)
289
0.92
(5.7,0.15)
313
0.91
(7.9,0.19)
311
0.91
(5.8,0.14)
309
0.92
(6.4,0.18)
302
0.92
(5.9,0.16)
328
0.92
(4.7,0.13)
330
1.00
(0.0, 0.00)
361
L
0.95
(4.6,0.12)
141
0.95
(4.0,0.10)
288
0.92
(4.5,0.12)
311
0.92
(3.6,0.10)
273
0.95
(4.4,0.10)
286
0.94
(5.7,0.12)
285
0.94
(4.7,0.10)
287
0.94
(4.4,0.13)
275
0.96
(3.9,0.10)
304
0.90
(7.0,0.17)
304
0.93
(5.9,0.15)
292
1.00
(0.0, 0.00)
331
M
0.91
(5.7,0.15)
165
0.92
(5.1,0.15)
157
0.88
(7.5,0.16)
178
0.89
(5.3,0.14)
151
0.95
(4.8,0.11)
159
0.94
(7.0,0.15)
161
0.95
(4.2,0.11)
154
0.93
(6.4,0.16)
164
0.96
(4.6,0.12)
173
0.91
(5.7,0.14)
171
0.94
(5.2,0.13)
173
0.94
(6.4,0.13)
147
1.00
(0.0, 0.00)
179
N
0.92
(6.6,0.15)
152
0.93
(5.8, 0.14)
152
0.92
(7.9,0.16)
175
0.96
(5.1,0.13)
146
0.94
(5.0,0.11)
154
0.94
(7.9,0.17)
157
0.95
(5.0,0.12)
149
0.94
(7.1,0.16)
157
0.95
(5.3,0.13)
169
0.94
(4.4,0.12)
165
0.96
(4.0,0.10)
166
0.95
(5.9,0.13)
142
0.97
(3.9, 0.09)
160
1.00
(0.0, 0.00)
176
0
0.89
(6.0,0.16)
156
0.90
(5.2,0.15)
150
0.86
(7.5,0.17)
173
0.88
(4.5,0.15)
145
0.92
(4.6,0.13)
152
0.94
(7.9,0.16)
154
0.95
(4.4,0.12)
148
0.91
(5.9,0.17)
156
0.93
(4.6,0.14)
166
0.89
(5.4,0.16)
164
0.92
(4.9,0.15)
167
0.92
(6.0,0.14)
142
0.95
(2.7,0.11)
165
0.95
(3.8,0.11)
152
1.00
(0.0, 0.00)
174
December 2009 A-111
-------
p
A 0.90
(8.0,0.19)
166
B 0.90
(8.0, 0.20)
159
C 0.89
(10.2,0.22)
180
D 0.90
(7.6,0.19)
153
E 0.91
(8.3,0.18)
160
F 0.91
(10.5,0.23)
163
G 0.91
(7.9,0.19)
156
H 0.91
(9.3, 0.23)
165
I 0.91
(8.2,0.21)
175
J 0.92
(5.6,0.14)
173
K 0.94
(5.2,0.12)
176
L 0.90
(9.3, 0.20)
151
M 0.92
(6.2,0.16)
175
N 0.92
(5.4,0.13)
162
0 0.88
(7.5,0.18)
166
P 1.00
(0.0, 0.00)
181
Q
R
S
T
U
V
W
X
Y
Z
AA
AB
AC
AD
AE
Q
0.89
(7.5,0.19)
151
0.90
(6.7,0.17)
290
0.91
(8.3,0.19)
324
0.94
(6.5,0.16)
278
0.92
(5.6,0.16)
294
0.92
(8.6, 0.20)
295
0.91
(5.9,0.16)
292
0.93
(7.1,0.20)
284
0.92
(6.1,0.17)
314
0.91
(5.1,0.14)
313
0.93
(4.2,0.12)
298
0.92
(6.7,0.17)
285
0.95
(4.5,0.14)
157
0.98
(2.8, 0.08)
151
0.92
(4.9,0.15)
152
0.92
(5.2,0.13)
159
1.00
(0.0, 0.00)
347
R
0.91
(5.6,0.16)
157
0.91
(5.5,0.16)
153
0.86
(7.1,0.17)
172
0.89
(5.5,0.14)
147
0.94
(4.9,0.12)
155
0.92
(8.5,0.17)
159
0.95
(4.1,0.12)
154
0.92
(6.6,0.17)
158
0.95
(4.7,0.12)
168
0.90
(6.2,0.16)
167
0.93
(6.1,0.16)
169
0.93
(6.7,0.14)
144
0.96
(3.4, 0.09)
165
0.94
(4.1,0.12)
153
0.93
(3.8,0.13)
157
0.90
(7.1,0.17)
166
0.92
(5.4,0.16)
154
1.00
(0.0, 0.00)
175
LEC
(P90
s
0.83
(8.0, 0.24)
145
0.83
(8.0, 0.24)
143
0.78
(10.4,0.25)
164
0.80
(8.6, 0.22)
135
0.87
(7.1,0.20)
142
0.87
(10.0,0.25)
144
0.88
(7.1,0.21)
140
0.82
(9.6, 0.26)
145
0.87
(7.7, 0.22)
154
0.85
(6.7,0.21)
153
0.86
(7.2, 0.20)
155
0.86
(8.9,0.21)
132
0.88
(6.3,0.19)
152
0.89
(5.8,0.17)
140
0.86
(6.5, 0.20)
145
0.84
(7.2, 0.20)
152
0.85
(7.2,0.19)
139
0.91
(5.8,0.18)
143
1.00
(0.0, 0.00)
164
SEND
R
, COD)
N
T
0.96
(4.4,0.11)
154
0.95
(3.9,0.10)
292
0.90
(6.9,0.13)
309
0.92
(5.0,0.12)
280
0.94
(4.1,0.10)
300
0.90
(6.9,0.14)
302
0.94
(3.9,0.10)
293
0.92
(5.7,0.13)
287
0.93
(4.2,0.10)
318
0.89
(6.1,0.17)
319
0.91
(5.1,0.16)
310
0.93
(5.5,0.12)
285
0.91
(5.9,0.14)
162
0.92
(6.2,0.13)
156
0.89
(5.9,0.15)
155
0.88
(8.5, 0.20)
164
0.88
(6.1,0.18)
290
0.93
(5.1,0.13)
157
0.83
(8.5, 0.22)
144
1.00
(0.0, 0.00)
330
U
0.83
(7.2,0.16)
162
0.81
(6.7,0.15)
310
0.76
(8.5,0.18)
327
0.74
(7.8,0.19)
294
0.77
(7.5,0.17)
320
0.74
(9.2,0.19)
320
0.76
(7.5,0.17)
315
0.78
(7.5,0.17)
307
0.78
(7.1,0.17)
338
0.73
(8.6, 0.22)
341
0.78
(8.0,0.21)
327
0.75
(8.2,0.19)
301
0.74
(9.0, 0.22)
171
0.81
(7.9, 0.20)
165
0.75
(8.8, 0.22)
166
0.73
(12.0,0.26)
174
0.71
(9.3, 0.24)
309
0.76
(8.6, 0.22)
167
0.66
(11.3,0.28)
153
0.81
(5.9,0.15)
318
1.00
(0.0, 0.00)
351
V
0.93
(5.9,0.13)
159
0.94
(5.2,0.11)
300
0.89
(6.4,0.13)
315
0.91
(5.9,0.13)
283
0.94
(5.6,0.12)
310
0.93
(5.4,0.11)
308
0.95
(4.7,0.10)
303
0.92
(6.1,0.13)
297
0.94
(5.0,0.10)
327
0.89
(8.7, 0.20)
329
0.89
(8.4,0.19)
319
0.92
(6.2,0.13)
290
0.90
(8.0,0.17)
166
0.91
(8.3,0.17)
160
0.89
(7.6,0.17)
161
0.87
(10.9,0.24)
168
0.89
(9.1,0.21)
296
0.90
(7.5,0.17)
161
0.81
(11.6,0.26)
148
0.93
(6.2,0.12)
307
0.77
(7.6,0.17)
327
1.00
(0.0, 0.00)
339
W
0.95
(4.7,0.12)
149
0.95
(5.0,0.11)
289
0.90
(7.7,0.15)
305
0.94
(5.4,0.13)
273
0.96
(4.3,0.10)
299
0.92
(8.2,0.16)
297
0.97
(3.7, 0.09)
293
0.91
(6.8,0.15)
288
0.95
(4.6,0.11)
316
0.89
(6.0,0.16)
317
0.91
(5.2,0.15)
304
0.93
(5.7,0.13)
282
0.94
(4.7,0.12)
159
0.92
(4.9,0.12)
157
0.92
(5.1,0.13)
154
0.89
(6.7,0.18)
160
0.90
(5.5,0.16)
289
0.94
(4.4,0.11)
153
0.86
(6.7, 0.20)
143
0.95
(3.4,0.10)
297
0.79
(6.6,0.17)
319
0.96
(5.9,0.11)
306
1.00
(0.0, 0.00)
328
X
0.96
(4.4,0.10)
156
0.96
(4.0,0.10)
292
0.90
(7.1,0.14)
311
0.93
(5.4,0.13)
282
0.95
(4.4,0.11)
303
0.92
(7.2,0.15)
305
0.96
(3.6, 0.09)
296
0.92
(5.9,0.15)
292
0.94
(4.8,0.11)
322
0.89
(6.3,0.16)
327
0.90
(5.5,0.15)
313
0.93
(5.8,0.13)
286
0.92
(5.2,0.14)
164
0.91
(5.6,0.14)
158
0.90
(5.8,0.15)
158
0.89
(7.4,0.18)
166
0.88
(6.5,0.16)
294
0.93
(5.0,0.13)
160
0.83
(8.0,0.21)
146
0.97
(3.2, 0.09)
302
0.81
(6.0,0.15)
322
0.97
(4.8,0.10)
314
0.98
(2.8, 0.06)
299
1.00
(0.0, 0.00)
334
Y
0.95
(4.5,0.12)
160
0.96
(4.2,0.11)
300
0.90
(7.9,0.15)
317
0.93
(5.8,0.13)
284
0.96
(3.9, 0.09)
310
0.92
(7.6,0.15)
311
0.97
(3.5, 0.08)
303
0.92
(6.7,0.14)
299
0.95
(4.7,0.10)
328
0.89
(6.2,0.16)
329
0.91
(5.2,0.15)
319
0.92
(6.2,0.13)
293
0.93
(5.0,0.13)
168
0.93
(4.6,0.13)
162
0.91
(5.5,0.14)
161
0.89
(6.9,0.18)
169
0.90
(5.5,0.16)
302
0.94
(4.0,0.12)
161
0.87
(7.2,0.19)
148
0.96
(2.9, 0.08)
305
0.81
(6.3,0.16)
326
0.95
(5.8,0.12)
316
0.98
(2.5, 0.07)
306
0.98
(2.3, 0.07)
311
1.00
(0.0, 0.00)
340
Z
0.98
(3.4,0.10)
160
0.98
(2.9, 0.09)
309
0.93
(6.7,0.15)
323
0.95
(4.8,0.13)
292
0.96
(3.9,0.11)
317
0.93
(6.3,0.16)
316
0.96
(3.4,0.11)
312
0.93
(5.9,0.15)
303
0.95
(4.2,0.13)
335
0.91
(5.9,0.16)
337
0.92
(5.4, 0.14)
325
0.95
(5.0,0.13)
299
0.92
(5.8,0.16)
169
0.93
(4.9,0.15)
165
0.90
(5.6,0.16)
162
0.90
(7.6,0.19)
171
0.90
(6.3,0.16)
306
0.91
(5.8,0.17)
164
0.83
(7.3, 0.22)
151
0.97
(3.2,0.12)
315
0.81
(6.4,0.17)
336
0.95
(5.7, 0.14)
325
0.96
(3.6,0.11)
312
0.97
(3.3,0.10)
318
0.97
(3.6,0.11)
322
1.00
(0.0, 0.00)
347
AA
0.94
(5.8,0.17)
154
0.94
(5.9,0.17)
288
0.87
(8.6, 0.20)
491
0.92
(6.4,0.17)
274
0.93
(5.8,0.17)
292
0.91
(8.5, 0.22)
292
0.92
(5.7,0.17)
288
0.90
(7.7, 0.22)
281
0.92
(6.5,0.18)
306
0.89
(5.6,0.16)
307
0.92
(5.2,0.15)
301
0.92
(7.2,0.17)
277
0.91
(6.4,0.17)
158
0.91
(5.4,0.15)
153
0.89
(6.1,0.17)
151
0.89
(6.1,0.16)
158
0.89
(5.3,0.16)
292
0.92
(6.2,0.17)
153
0.84
(6.1,0.21)
141
0.92
(5.2,0.17)
284
0.78
(8.1,0.22)
305
0.93
(7.7, 0.20)
292
0.95
(4.5,0.15)
281
0.95
(4.6,0.14)
286
0.95
(4.7,0.16)
296
0.95
(4.6,0.15)
305
1.00
(0.0, 0.00)
532
AB
0.93
(6.8,0.17)
159
0.92
(6.8,0.17)
308
0.88
(8.9, 0.20)
323
0.90
(6.9,0.18)
291
0.91
(6.9,0.17)
314
0.90
(9.4,0.21)
317
0.92
(6.8,0.16)
311
0.89
(8.1,0.22)
301
0.91
(6.8,0.18)
334
0.88
(5.8,0.16)
335
0.91
(4.9,0.15)
325
0.92
(7.6,0.17)
299
0.91
(5.5,0.15)
168
0.91
(4.9,0.14)
162
0.90
(5.7,0.16)
161
0.90
(5.7,0.15)
170
0.88
(5.3,0.16)
303
0.92
(5.6,0.16)
164
0.84
(7.4, 0.20)
151
0.91
(5.5,0.16)
312
0.76
(8.4, 0.22)
334
0.91
(8.6, 0.20)
323
0.93
(4.8,0.15)
310
0.93
(4.9,0.14)
319
0.93
(5.0,0.15)
322
0.93
(5.3,0.15)
331
0.98
(2.4, 0.07)
305
1.00
(0.0, 0.00)
346
AC
0.95
(6.0,0.16)
158
0.93
(6.5,0.18)
299
0.88
(9.6,0.21)
313
0.90
(7.0,0.19)
287
0.90
(6.8,0.18)
304
0.87
(9.1,0.23)
306
0.90
(6.8,0.18)
300
0.88
(8.3, 0.22)
293
0.89
(7.4,0.19)
323
0.87
(6.3,0.17)
327
0.90
(5.3,0.16)
315
0.90
(7.4,0.18)
286
0.88
(6.9,0.19)
165
0.88
(6.5,0.17)
158
0.87
(7.0,0.19)
159
0.89
(6.3,0.17)
167
0.86
(6.3,0.18)
293
0.88
(7.1,0.19)
160
0.81
(7.8, 0.23)
149
0.92
(5.4,0.18)
311
0.79
(7.2,0.21)
324
0.90
(8.3,0.21)
314
0.92
(5.4,0.16)
306
0.94
(4.9,0.15)
305
0.92
(5.3,0.17)
311
0.94
(4.9,0.15)
321
0.97
(2.9, 0.08)
287
0.96
(3.1,0.09)
317
1.00
(0.0, 0.00)
336
AD
0.94
(5.4,0.15)
159
0.87
(5.3,0.16)
305
0.84
(8.7,0.18)
323
0.85
(6.2,0.17)
290
0.86
(6.3,0.16)
313
0.83
(8.3, 0.20)
317
0.87
(5.9,0.15)
308
0.83
(8.1,0.20)
301
0.85
(6.6,0.17)
334
0.82
(6.3,0.17)
336
0.85
(5.3,0.16)
323
0.85
(7.3,0.16)
294
0.89
(6.2,0.17)
168
0.89
(5.4,0.16)
161
0.87
(6.2,0.17)
162
0.89
(6.4,0.17)
170
0.82
(5.9,0.17)
303
0.89
(6.4,0.17)
163
0.82
(7.1,0.22)
148
0.87
(4.9,0.15)
313
0.74
(7.0,0.19)
333
0.88
(6.9,0.17)
321
0.89
(3.9,0.13)
311
0.91
(3.6,0.11)
316
0.89
(4.4,0.14)
321
0.89
(4.1,0.14)
328
0.89
(3.2,0.11)
300
0.89
(3.7,0.11)
328
0.91
(2.8,0.10)
320
1.00
(0.0, 0.00)
346
AE
0.88
(7.1,0.17)
162
0.87
(7.2,0.17)
311
0.79
(8.5, 0.20)
333
0.87
(6.9,0.17)
297
0.87
(7.3,0.17)
318
0.84
(9.3,0.21)
322
0.85
(7.5,0.17)
314
0.88
(7.6, 0.20)
307
0.86
(7.0,0.18)
339
0.87
(6.4,0.15)
340
0.91
(5.1,0.13)
328
0.89
(7.1,0.17)
301
0.89
(6.9,0.17)
171
0.90
(6.0,0.14)
165
0.87
(7.0,0.18)
166
0.92
(5.7,0.13)
173
0.91
(5.5,0.14)
310
0.89
(7.1,0.17)
166
0.80
(9.0, 0.22)
153
0.85
(6.6,0.18)
319
0.69
(10.0,0.23)
338
0.83
(9.3, 0.22)
325
0.85
(6.9,0.17)
316
0.85
(6.8,0.17)
321
0.85
(6.7,0.18)
326
0.86
(6.8,0.17)
335
0.88
(5.9,0.17)
304
0.86
(6.5,0.17)
333
0.85
(6.7,0.17)
322
0.79
(7.2,0.18)
332
1.00
December 2009 A-112
-------
P QR S T U VWXYZAAABACADAE
(0.0, 0.00)
355
0.8
0.6
» « « »•
»* *
o
O
0.4
0.2
10 20 30
40 50 60
Distance Between Samplers (km)
70 80 90 100
Figure A -48. PM2.s inter-sampler correlations as a function of distance between monitors for
Chicago, IL
December 2009
A-113
-------
Denver Combined Statistical Area
A
01
0 10 20 40
| | Denver CSA
• PMz.s Monitors
Interstate Highways
— Major Highways
60 80 100
H^^^H^Z^ZI Kilometers
Figure A-49. PM2.s monitor distribution and major highways, Denver, CO.
December 2009
A-114
-------
AQS Site ID A
Site A 08-001-0006
SiteB 08-005-0005 "
SiteC 08-013-0003 Obs 369
SiteD 08-013-0012 SD
SiteE 08-031-0002 4Q .
SiteF 08-031-0023
SiteG 08-123-0006
SiteH 08-123-0008
" 30 -
Ol
c
o
2 20-
c
u
c
o
10 -
1=winter
2=spring
3=summer "
5.9
B
363
5.3
I
"
C
8.2
361
4.7
•I
f
D
7.0
354
3.9
II
1
E
F
9.2 9.7
1046 1006
5.5
5.6
G
H
8.3 9.1
359 334
5.3
5.8
4=fall 1234 1234 1234 1234 1234 1234 1234 1234
Figure A-50. Box plots illustrating the seasonal distribution of 24-h avg PM2.s concentrations
for Denver, CO.
December 2009
A-115
-------
Table A-24.
A
A 1.00
(0.0, 0.00)
369
B
C
D
E
F
G
H
Inter-sampler correlation statistics for each pair of PM2.6 monitors reporting to AQS for
Denver, CO.
B C D E
0.74 0.84 0.68 0.86
(6.0,0.21) (5.4,0.17) (7.9,0.26) (4.1,0.14)
353 347 332 362
1.00 0.58 0.76 0.92
(0.0,0.00) (5.7,0.19) (3.9,0.17) (3.2,0.13)
363 344 328 356
1.00 0.74 0.71
(0.0,0.00) (4.4,0.19) (4.5,0.17)
361 326 354
1.00 0.82
(0.0,0.00) (5.6,0.21)
354 347
1.00
LEGEND (0.0, 0.00)
R 1046
(P90, COD)
N
F
0.91
(3.0,0.11)
339
0.84
(4.4,0.17)
336
0.75
(5.4,0.18)
336
0.77
(6.0, 0.24)
332
0.94
(2.3, 0.09)
969
1.00
(0.0, 0.00)
1006
G
0.76
(5.9,0.19)
341
0.50
(7.8, 0.23)
337
0.83
(3.5,0.14)
333
0.54
(7.2, 0.24)
318
0.64
(7.1,0.21)
353
0.68
(6.6,0.21)
333
1.00
(0.0, 0.00)
359
H
0.83
(4.6,0.14)
325
0.49
(6.6,0.21)
323
0.88
(3.7,0.13)
320
0.57
(6.4, 0.24)
305
0.60
(5.6,0.18)
330
0.69
(5.9,0.17)
317
0.88
(3.4,0.13)
313
1.00
(0.0, 0.00)
334
December 2009 A-116
-------
0.8
0.6
o
o
0.4
0.2
10 20 30 40 50 60
Distance Between Samplers (km)
70 80 90 100
Figure A-51. PM2.s inter-sampler correlations as a function of distance between monitors for
Denver, CO.
December 2009
A-117
-------
Detroit Combined Statistical Area
A
01
Detroit CSA
• PMz.s Monitors
Interstate Highways
— Major Highways
0 10 20 40 60 80
100
—i Kilometers
Figure A-52. PM2.s monitor distribution and major highways, Detroit, Ml.
December 2009
A-118
-------
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
SiteG
AQS Site ID
26-049-0021
26-099-0009
26-115-0005
26-125-0001
26-147-0005
26-161-0008
26-163-0001
) A B C D E F G
1 Mean 11.6 12.9 13.8 13.9 13.5 13.7 14.0
Q
Obs 356 306 342 308 303 350 1049
1 SD 7.5 9.4 9.1 9.5 9.9 8.9 8.9
5 50 -
8
^
40 -
IE
Dl
3 30 -
c
-------
Table A-25. Inter-sampler correlation statistics for each pair of PM2 5 monitors reporting to AQS for
Detroit, Ml.
A
A 1.00
(0.0, 0.00)
356
B
C
D
E
F
G
H
I
J
K
L
M
B C D E F
0.91 0.86 0.91 0.89 0.90
(5.9,0.17) (7.8,0.19) (6.7,0.17) (7.6,0.18) (5.9,0.18)
299 333 301 296 341
1.00 0.90 0.94 0.92 0.92
(0.0,0.00) (6.8,0.17) (5.3,0.14) (5.9,0.16) (5.8,0.17)
306 286 296 290 294
1.00 0.90 0.87 0.91
(0.0,0.00) (7.0,0.16) (8.8,0.20) (5.5,0.15)
342 289 284 326
1.00 0.93 0.94
(0.0,0.00) (6.3,0.15) (4.5,0.14)
308 292 296
1.00 0.90
(0.0,0.00) (7.5,0.18)
303 291
1.00
(0.0, 0.00)
350
LEGEND
R
(P90, COD)
N
G
0.89
(8.1,0.20)
349
0.93
(6.2,0.18)
300
0.93
(5.9,0.14)
335
0.96
(4.3,0.13)
303
0.90
(7.3, 0.20)
297
0.95
(4.5,0.13)
343
1.00
(0.0, 0.00)
1049
H
0.88
(8.3, 0.22)
334
0.90
(7.5,0.21)
288
0.90
(7.2,0.17)
320
0.92
(5.8,0.16)
291
0.89
(8.2, 0.22)
286
0.90
(6.2,0.17)
326
0.94
(5.1,0.14)
336
1.00
(0.0, 0.00)
342
I
0.89
(8.0,0.19)
284
0.92
(5.8,0.18)
277
0.91
(6.3,0.16)
273
0.94
(4.5,0.12)
281
0.90
(7.0,0.19)
276
0.92
(5.7,0.15)
280
0.95
(4.9,0.12)
549
0.93
(4.8,0.15)
273
1.00
(0.0, 0.00)
572
J
0.91
(7.3,0.17)
301
0.91
(4.9,0.16)
297
0.90
(6.2,0.14)
286
0.94
(3.8,0.11)
297
0.90
(6.4,0.18)
292
0.92
(5.2,0.14)
297
0.92
(4.5,0.14)
302
0.91
(5.4,0.15)
290
0.92
(4.4,0.13)
279
1.00
(0.0, 0.00)
308
K
0.92
(5.5,0.16)
293
0.92
(5.4,0.17)
286
0.89
(6.2,0.16)
279
0.94
(3.6,0.13)
291
0.90
(6.9,0.18)
284
0.95
(3.9,0.12)
288
0.93
(5.6,0.16)
295
0.89
(6.9,0.18)
288
0.90
(6.1,0.14)
271
0.91
(5.3,0.15)
288
1.00
(0.0, 0.00)
301
L
0.87
(11.0,0.26)
336
0.89
(10.2,0.24)
292
0.87
(10.4,0.20)
321
0.91
(8.2,0.18)
290
0.87
(10.7,0.25)
288
0.89
(9.8,0.21)
329
0.90
(8.2,0.18)
337
0.91
(7.6,0.16)
321
0.92
(7.9,0.18)
274
0.90
(8.1,0.17)
291
0.88
(9.5,0.21)
281
1.00
(0.0, 0.00)
344
M
0.88
(7.8,
333
0.91
(6.1,
288
0.93
(4.9,
319
0.92
(6.2,
290
0.87
(7.7,
288
0.93
(5.7,
326
0.95
(4.7,
335
0.91
(6.1,
319
0.93
(5.8,
274
0.91
(5.6,
291
0.91
(6.3,
283
0.91
(8.5,
322
1.00
0.21)
0.19)
0.13)
0.15)
0.21)
0.15)
0.12)
0.15)
0.14)
0.13)
0.16)
0.17)
(0.0, 0.00)
342
December 2009 A-120
-------
** •
0.8
0.6
o
13
o
O
0.4
0.2
10 20 30 40 50 60 70
Distance Between Samplers (km)
80 90 100
Figure A-54. PM2.s inter-sampler correlations as a function of distance between monitors for
Detroit, Ml.
December 2009
A-121
-------
Houston Combined Statistical Area
A
01
Houston CSA
• PM2.5 Monitors
Interstate Highways
— Major Highways
0 10 20 40 60 80
100
—i Kilometers
Figure A-55. PM2.s monitor distribution and major highways, Houston, TX.
December 2009
A-122
-------
AQS Site ID
Site A 48-201-0058
SiteB 48-201-1035
Mean
Obs
SD
A
11.3
326
5.5
B
15.8
1016
6.2
40 -
30 -
IE
Ol
IL
0 20 -
ro
concenti
O
1=winter
2=spring
3=summer 0 -
4=fall
1234 1234
Figure A-56. Box plots illustrating the seasonal distribution of 24-h avg PM2.s concentrations
for Houston, TX.
Table A-26. Inter-sampler correlation statistics for each pair of PM2.6 monitors reporting to AQS for
Houston, TX.
A B
A
1.00 0.66
(0.0,0.00) (10.0,0.24)
326 310
B
1.00
(0.0,0.00)
1016
LEGEND
R
(P90, COD)
N
December 2009
A-123
-------
0.8
0.6
HI
5
o
0.4
0.2
10 20 30 40 50 60
Distance Between Samplers (km)
70 80 90 100
Figure A-57. PM2.s inter-sampler correlations as a function of distance between monitors for
Houston, TX.
December 2009
A-124
-------
Los Angeles Core Based Statistical Area
01
r
Los Angeles CBSA
• PM2.5 Monitors
Interstate Highways
— Major Highways
0 10 20 40 60 80
100
—i Kilometers
Figure A-58. PM2.s monitor distribution and major highways, Los Angeles, CA.
December 2009
A-125
-------
SteA
SteB
SiteC
SteD
SiteE
SiteF
SiteG
SiteH
She I
SiteJ
SteK
AQSSKelD
06-037-0002
06-037-1002
06-037-1103
06-037-1201
06-037-1301
06-037-2005
06-037^1002
06-037-1O04
06-037-9033
06-059-O007
06-059-2022
1=winter
2=spring
3=summer
4=fall
ABCDEFGH1JK
Mean 16.1 17.0 16.7 13.3 16.7 14.3 4.7 4.2 8.2 14.4 10.9
Obs 862 308 1004 291 327 334 946 990 221 999 318
SD 10.8 10.2 9.8 7.5 9.3 8.9 8.4 7.7 3.8 8.5 6.4
60 -
50 -
IE
c
o
TO 30 -
c=
QJ
020-
1 0-
o -
1 | 1 1
1
1
II
I
1
'!"
i
I
1'
1234 1234 1234 1234 1234 1234 1234 1234 1234 1234 1234
Figure A-59. Box plots illustrating the seasonal distribution of 24-h avg PM2.s concentrations
for Los Angeles, CA.
December 2009
A-126
-------
Table A-27. Inter-sampler correlation statistics for each pair of PM2.6 monitors reporting to AQS for
Los Angeles, CA.
A B C D E
A 1.00 0.86 0.87 0.81 0.80
(0.0,0.00) (9.0,0.18) (7.7,0.16) (9.0,0.19) (9.7,0.21)
862 252 803 238 262
B 1.00 0.92 0.87 0.83
(0.0,0.00) (5.5,0.11) (9.1,0.19) (9.0,0.15)
308 293 250 278
C 1.00 0.80 0.89
(0.0,0.00) (9.6,0.20) (5.8,0.11)
1004 274 315
D 1.00 0.69
(0.0,0.00) (10.9,0.23)
291 263
E 1.00
(0.0, 0.00)
327
F
G
LEGEND
Pearcon R
(P90 COD)
H
I
J
K
F G
0.88 0.68
(5.8,0.14) (11.5,0.22)
269 761
0.88 0.77
(7.6,0.15) (9.8,0.17)
279 268
0.92 0.84
(6.4,0.13) (9.0,0.15)
319 880
0.77 0.63
(7.4,0.18) (11.3,0.22)
263 256
0.79 0.95
(9.1,0.19) (5.9,0.11)
301 289
1.00 0.70
(0.0,0.00) (10.5,0.18)
334 290
1.00
(0.0, 0.00)
946
H
0.64
(12.4,0.23)
793
0.73
(11.6,0.18)
282
0.79
(10.0,0.17)
913
0.60
(11.1,0.22)
268
0.92
(7.6,0.13)
301
0.70
(9.2,0.19)
302
0.96
(4.0, 0.09)
859
1.00
(0.0, 0.00)
990
1
0.30
(18.0,0.36)
179
0.31
(24.1,0.38)
177
0.29
(18.6,0.38)
213
0.41
(14.8,0.31)
164
0.34
(19.7,0.39)
192
0.33
(14.8,0.34)
184
0.23
(17.0,0.35)
194
0.26
(15.3,0.34)
208
1.00
(0.0, 0.00)
221
J
0.70
(10.5,0.21)
804
0.74
(11.9,0.19)
292
0.82
(9.4,0.16)
920
0.64
(9.6,0.21)
274
0.88
(8.2,0.15)
307
0.69
(9.8,0.19)
311
0.92
(5.4,0.12)
882
0.91
(5.9,0.12)
914
0.21
(18.3,0.35)
205
1.00
(0.0, 0.00)
999
K
0.82
(11.4,0.23)
259
0.71
(15.0,0.27)
277
0.78
(13.2,0.25)
305
0.60
(11.6,0.23)
261
0.76
(13.7,0.27)
291
0.72
(9.9,0.21)
293
0.78
(11.0,0.21)
277
0.78
(9.5,0.21)
294
0.31
(9.7, 0.28)
180
0.84
(9.8,0.19)
298
1.00
(0.0, 0.00)
318
0 8 -
c
o
1
I
n -
*• * \
• •' '
» «
«
~ * *
4
• *
• •
«*
*
«*. .
10 20 30 40 50 60 70
Distance Between Samplers (km)
80
90
100
Figure A-60. PM2.s inter-sampler correlations as a function of distance between monitors for
Los Angeles, CA.
December 2009
A-127
-------
New York Combined Statistical Area
5/1
| | New York CSA
• PM2.5 Monitors
Interstate Highways
— Major Highways
0 10 20 40 60 80 100
i Kilometers
Figure A-61. PM2.s monitor distribution and major highways, New York City, NY.
December 2009
A-128
-------
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
SiteG
SiteH
Sitel
SiteJ
SiteK
SiteL
SiteM
SiteN
SiteO
SiteP
SiteQ
SiteR
SiteS
Site!
AQS Site ID A
09-001-0010 .,
Mean 132
09-001-1123
09-001-3005 Obs 349
09-001-9003 SD 8.7
09-005-0005 50 -
09-009-0026
09-009-0027
09-009-1123
09-009-2008 40 -
09-009-2123
"E
-- 30 -
Ol
c
o
4-1
TO
- 20-
c
u
c
o
10 -
1=winter
2=spring
3=summer 0 -
4=fall 1234
AQS Site ID
K
34-003-0003
34-013-0015 Mean 13'2
34-017-1003 Obs 345
34-021-0008 SD 89
34-021-8001 50 -
34-023-0006
34-027-0004
34-027-3001
34-029-2002 4Q _
34-031-0005
"E
-- 30 -
Ol
IL
C
O
S 20 -
e
01
u
c
o
u
10 -
1=winter
2=spring
3=summer
A fi\\ ^
A B C D
i 13.2 12.5 12.3 11.2
s 349 339 332 565
3 8.7 8.3 8.1 7.6
E F G H J
8.0 12.1 12.4 12.8 11.1 12.6
341 341 992 338 344 352
6.8 8.0 8.1 8.6 7.5 8.2
1234 1234 1234 1234 1234 1234 1234 1234 1234 1234
K L M N 0 P
13.2 13.3 13.7 12.3 10.8 12.1
345 334 559 545 313 336
8.9 8.8 8.5 7.7 7.0 7.7
Q R S T
11.3 10.0 10.6 12.9
336 357 550 330
7.7 7.4 6.8 8.7
1234 1234 1234 1234 1234 1234 1234 1234 1234 1234
December 2009
A-129
-------
AQS Site ID
SiteU 34-039-0004 Mea|1
Site V 34-039-0006
SiteW 34-039-2003
Site X 36-005-0080
SiteY 36-005-0083
SiteZ 36-005-0110
SiteAA 36-047-0122
SiteAB 36-059-0008
Site AC 36-061-0056
Site AD 36-061-0079
l=winter
2=spring
3=summei
4=fall
U V W X Y Z AA AB AC AD
Mean 14.5 13.2 12.9 15.5 13.0 13.0 14.0 11.4 15.9 13.5
Obs 1017 352 332 349 359 1059 342 337 357 363
SD 8.7 8.4 8.4 9.1 8.2 8.2 8.4 7.5 8.9 8.4
50 -
40 -
30 -
20 -
10 -
r 0 -
1234 1234 1234 1234 1234 1234 1234 1234 1234 1234
SiteAE 36-061-0128 AE AF AG AH AI AJ
SiteAF 36-071-0002 Mean 15.3 10.8 US 13.3 11. 4 11.7
SiteAG 36-081-0124 Obs 341 342 951 337 335 355
Site AH 36-085-0055 SD 8.8 7.6 7.5 s.o 7.3 7.8
SiteAl 36-085-0067 50 -
SiteAJ 36-119-1002
40 -
IE
Dl
a 30-
c
o
•*-»
ro
•*-*
Ł 20 -
u
c
o
u
10 -
1=winter
2=spring
3=summer 0 -
Figure A-62. Box plots illustrating the seasonal distribution of 24-h avg PM2.s concentrations
for New York, NY.
December 2009
A-130
-------
Table A-28. Inter-sampler correlation statistics for each pair of PM2 5 monitors reporting to AQS for
New York, NY.
Site A B
A 1.00 0.89
(0.0,0.00) (5.3,0.15)
349 322
B 1.00
(0.0, 0.00)
339
C
D
E
F
G
H
I
J
K
L
M
N
0
P
Q
R
C D E F
0.97 0.97 0.82 0.96
(3.6,0.09) (4.8,0.11) (11.8,0.33) (3.8,0.11)
316 322 325 328
0.93 0.91 0.78 0.91
(4.5,0.13) (5.3,0.14) (10.4,0.32) (4.7,0.13)
312 315 319 316
1.00 0.98 0.82 0.96
(0.0,0.00) (3.4,0.08) (10.8,0.32) (3.9,0.10)
332 314 309 310
1.00 0.85 0.96
(0.0,0.00) (8.4,0.29) (3.4,0.11)
565 314 316
1.00 0.82
(0.0,0.00) (10.0,0.31)
341 321
1.00
(0.0, 0.00)
341
1 PftPND
R
(rUU, (sUU)
N
G
0.96
(4.0,0.11)
321
0.92
(4.6,0.13)
313
0.95
(4.1,0.11)
308
0.96
(3.8,0.11)
532
0.82
(10.7,0.33)
313
0.99
(2.1,0.07)
314
1.00
(0.0, 0.00)
992
H
0.96
(3.4,0.10)
324
0.91
(4.6,0.14)
313
0.96
(3.6,0.10)
307
0.94
(5.0,0.13)
315
0.79
(11.4,0.33)
317
0.98
(2.9, 0.09)
319
0.96
(2.9,0.10)
315
1.00
(0.0, 0.00)
338
I
0.96
(4.6,0.12)
326
0.91
(5.0, 0.14)
315
0.97
(4.0,0.11)
310
0.96
(3.0,0.10)
313
0.83
(8.8, 0.28)
319
0.98
(2.8, 0.09)
321
0.98
(3.6,0.11)
319
0.98
(3.7,0.10)
320
1.00
(0.0, 0.00)
344
J
0.93
(5.1,0.12)
335
0.92
(4.5,0.13)
330
0.94
(4.8,0.11)
319
0.92
(5.5,0.13)
325
0.81
(10.3,0.32)
330
0.94
(4.7,0.11)
328
0.93
(5.2,0.12)
326
0.94
(3.7,0.10)
324
0.95
(4.1,0.11)
327
1.00
(0.0, 0.00)
352
K
0.91
(5.8,0.12)
329
0.83
(7.3,0.17)
319
0.91
(5.7,0.13)
314
0.90
(7.1,0.15)
319
0.80
(12.5,0.34)
322
0.88
(6.7, 0.14)
321
0.88
(7.1,0.15)
319
0.88
(7.1,0.14)
318
0.89
(7.0,0.16)
324
0.87
(7.0,0.16)
332
1.00
(0.0, 0.00)
345
L
0.91
(5.7,0.12)
316
0.84
(7.1,0.17)
305
0.91
(5.8,0.14)
299
0.89
(6.9,0.15)
308
0.77
(13.0,0.34)
305
0.89
(6.8,0.15)
308
0.89
(6.7,0.15)
309
0.89
(7.1,0.14)
303
0.90
(7.0,0.16)
307
0.87
(7.2,0.16)
316
0.95
(3.4, 0.09)
314
1.00
(0.0, 0.00)
334
M
0.92
(5.5,0.13)
331
0.85
(7.8,0.19)
321
0.91
(6.5,0.15)
316
0.91
(6.7,0.18)
517
0.76
(13.8,0.39)
323
0.89
(6.8,0.16)
323
0.89
(6.9,0.16)
526
0.89
(6.6,0.16)
321
0.89
(7.7, 0.20)
323
0.87
(8.5,0.17)
334
0.93
(4.5,0.12)
330
0.97
(4.1,0.10)
321
1.00
(0.0, 0.00)
559
N
0.88
(6.6,0.16)
301
0.82
(7.2,0.19)
291
0.89
(5.4,0.15)
287
0.88
(6.3,0.17)
506
0.76
(11.6,0.35)
294
0.86
(6.4,0.17)
293
0.84
(6.9,0.18)
513
0.84
(6.7,0.18)
292
0.87
(6.4,0.18)
296
0.84
(6.9,0.18)
303
0.88
(6.4,0.15)
301
0.91
(6.4, 0.14)
289
0.91
(5.5, 0.14)
499
1.00
(0.0, 0.00)
545
0
0.84
(9.1,0.19)
296
0.79
(7.7, 0.20)
292
0.84
(6.9,0.17)
289
0.87
(6.5,0.16)
288
0.79
(9.1,0.30)
291
0.85
(6.8,0.18)
295
0.84
(8.0,0.19)
286
0.82
(8.1,0.20)
285
0.85
(6.6,0.17)
291
0.79
(7.9, 0.20)
299
0.86
(7.5,0.17)
296
0.86
(8.0,0.18)
288
0.86
(8.4,0.21)
300
0.93
(4.7, 0.14)
270
1.00
(0.0, 0.00)
313
P
0.87
(8.3,0.16)
321
0.82
(7.6,0.18)
310
0.88
(6.3, 0.14)
307
0.89
(6.0,0.15)
311
0.78
(10.4,0.32)
316
0.88
(6.1,0.15)
312
0.86
(7.6,0.16)
310
0.85
(7.8,0.17)
310
0.87
(6.5,0.16)
313
0.82
(8.1,0.18)
321
0.90
(5.7,0.13)
317
0.94
(5.2,0.12)
309
0.93
(6.7,0.15)
326
0.95
(4.1,0.11)
293
0.93
(4.3,0.12)
294
1.00
(0.0, 0.00)
336
Q
0.89
(7.6,0.16)
318
0.82
(6.6,0.18)
307
0.89
(6.2,0.15)
305
0.90
(5.5,0.14)
309
0.87
(7.9, 0.28)
311
0.87
(7.3,0.16)
310
0.87
(8.1,0.17)
306
0.85
(7.5,0.17)
307
0.88
(6.5,0.15)
313
0.84
(7.5,0.17)
322
0.92
(5.8,0.14)
319
0.93
(5.9,0.13)
303
0.92
(7.5,0.18)
318
0.91
(5.8,0.15)
292
0.91
(4.9, 0.14)
287
0.94
(4.9,0.12)
308
1.00
(0.0, 0.00)
336
R
0.84
(9.3,0.21)
316
0.78
(8.4, 0.22)
305
0.84
(8.2, 0.20)
297
0.86
(6.6,0.18)
330
0.87
(7.3, 0.24)
305
0.83
(7.5,0.21)
308
0.82
(8.4, 0.23)
327
0.79
(9.2, 0.23)
304
0.83
(7.6,0.19)
310
0.79
(9.0, 0.22)
316
0.86
(8.7, 0.20)
312
0.87
(8.3, 0.20)
301
0.85
(9.7, 0.25)
337
0.88
(7.2, 0.20)
316
0.94
(4.3,0.14)
279
0.91
(5.5,0.16)
303
0.95
(3.8,0.13)
307
1.00
(0.0, 0.00)
357
December 2009 A-131
-------
s
A 075
(10.4,0.21)
323
B 0.68
(10.8,0.23)
314
C 0.76
(8.5, 0.20)
307
D 0.80
(7.7,0.19)
509
E 0.67
(9.8, 0.32)
315
F 0.79
(7.9,0.19)
316
G 0.77
(8.7,0.21)
513
H 0.74
(9.6, 0.22)
314
I 0.76
(8.1,0.20)
315
J 0.67
(11.1,0.22)
327
K 0.74
(10.9,0.21)
320
L 0.78
(9.8, 0.20)
313
M 0.80
(9.9, 0.22)
504
N 0.88
(6.4,0.17)
492
0 0.87
(5.6,0.16)
295
P 0.86
(6.2,0.15)
312
Q 0.79
(8.1,0.19)
313
R 0.82
(6.5, 0.20)
330
S 1.00
(0.0, 0.00)
550
T
U
V
W
X
Y
Z
AA
AB
AC
AD
AE
T
0.89
(6.1,0.13)
315
0.84
(5.9,0.16)
307
0.89
(6.1,0.14)
304
0.88
(7.3,0.16)
306
0.79
(11.3,0.34)
306
0.87
(6.7,0.15)
306
0.87
(6.3,0.15)
304
0.86
(6.6,0.15)
304
0.88
(7.1,0.17)
308
0.84
(6.6,0.16)
316
0.94
(3.9,0.11)
317
0.94
(3.9,0.11)
303
0.93
(5.4,0.13)
318
0.86
(6.5,0.16)
287
0.82
(7.2,0.18)
289
0.89
(6.2,0.14)
307
0.92
(5.0,0.14)
303
0.86
(7.6,0.21)
296
0.69
(10.4,0.22)
306
1.00
(0.0, 0.00)
330
U
0.90
(7.1,0.15)
337
0.83
(8.6, 0.20)
328
0.88
(7.8,0.18)
321
0.88
(8.1,0.20)
537
0.74
(14.9, 0.40)
329
0.86
(8.5,0.19)
329
0.87
(7.8,0.18)
928
0.86
(8.4,0.18)
326
0.87
(8.7,0.21)
332
0.85
(9.0,0.19)
343
0.92
(5.7,0.14)
336
0.97
(4.5,0.12)
325
0.97
(3.8, 0.09)
534
0.90
(6.5,0.15)
519
0.86
(9.9, 0.22)
302
0.92
(7.4,0.17)
325
0.91
(8.2, 0.20)
327
0.84
(10.9,0.26)
347
0.77
(10.5,0.24)
525
0.92
(6.0,0.15)
319
1.00
(0.0, 0.00)
1017
V
0.90
(6.0,0.13)
299
0.83
(6.5,0.18)
290
0.89
(6.4,0.15)
283
0.87
(7.1,0.17)
326
0.76
(11.7,0.36)
290
0.86
(6.8,0.16)
293
0.86
(7.0,0.16)
327
0.87
(7.1,0.16)
289
0.88
(7.4,0.17)
293
0.86
(6.7,0.16)
301
0.95
(3.4,0.10)
302
0.98
(2.9, 0.08)
292
0.97
(3.5, 0.09)
341
0.91
(5.7,0.13)
313
0.87
(7.3,0.18)
280
0.94
(5.0,0.12)
296
0.93
(6.2,0.14)
287
0.85
(8.2,0.21)
291
0.75
(10.5,0.21)
324
0.93
(4.5,0.12)
293
0.98
(3.9,0.10)
341
1.00
(0.0, 0.00)
352
LEC
(P90,
w
0.88
(7.2,0.15)
316
0.84
(6.8,0.18)
305
0.89
(6.0,0.15)
297
0.88
(6.9,0.17)
304
0.75
(12.1,0.36)
307
0.87
(6.6,0.16)
309
0.88
(6.3,0.15)
303
0.87
(6.9,0.16)
306
0.88
(6.9,0.17)
309
0.86
(6.8,0.17)
318
0.92
(4.3,0.12)
317
0.95
(4.0,0.10)
303
0.96
(4.7,0.11)
319
0.91
(4.5,0.13)
290
0.88
(6.4,0.16)
284
0.95
(4.0,0.11)
305
0.92
(5.4,0.15)
304
0.87
(7.0, 0.20)
304
0.78
(9.2,0.19)
306
0.93
(4.8,0.13)
301
0.97
(5.0,0.12)
325
0.97
(2.8, 0.09)
288
1.00
(0.0, 0.00)
332
SEND
R
, COD)
N
X
0.92
(7.2,0.16)
332
0.84
(9.0,0.21)
325
0.92
(7.8,0.18)
317
0.89
(9.7,0.21)
324
0.75
(15.2,0.41)
324
0.89
(8.2,0.19)
325
0.89
(8.3,0.17)
319
0.90
(7.5,0.17)
322
0.90
(9.4, 0.22)
326
0.88
(8.8,0.19)
337
0.92
(6.0,0.15)
330
0.94
(6.3,0.15)
314
0.95
(4.9,0.11)
331
0.86
(8.2,0.18)
301
0.84
(11.1,0.24)
299
0.88
(8.9,0.19)
319
0.89
(9.7, 0.22)
321
0.82
(11.6,0.28)
314
0.72
(12.5,0.25)
325
0.92
(6.7,0.16)
313
0.92
(5.4,0.12)
337
0.93
(6.1,0.14)
300
0.90
(7.0,0.16)
316
1.00
(0.0, 0.00)
349
Y
0.94
(4.0,0.11)
342
0.88
(5.9,0.16)
334
0.95
(4.4,0.11)
326
0.94
(5.6, 0.14)
332
0.80
(11.5,0.34)
334
0.93
(5.0,0.12)
335
0.92
(5.4,0.13)
329
0.92
(5.2,0.13)
331
0.93
(5.7,0.15)
334
0.90
(6.1,0.14)
345
0.93
(3.8,0.12)
339
0.93
(4.5,0.12)
323
0.95
(4.7,0.12)
342
0.90
(5.9, 0.14)
309
0.87
(6.7,0.18)
308
0.90
(5.9, 0.14)
329
0.90
(6.3,0.16)
328
0.87
(8.6,0.21)
323
0.78
(8.6,0.19)
336
0.91
(5.2,0.14)
323
0.91
(6.9,0.15)
347
0.92
(5.0,0.13)
307
0.91
(5.5,0.13)
325
0.96
(5.8,0.13)
344
1.00
(0.0, 0.00)
359
Z
0.93
(4.7,0.12)
348
0.85
(6.8,0.17)
338
0.93
(5.4,0.13)
331
0.92
(6.2,0.15)
548
0.78
(13.1,0.36)
340
0.91
(6.4,0.14)
340
0.90
(5.7,0.14)
958
0.89
(5.6,0.14)
337
0.92
(6.5,0.16)
343
0.88
(7.1,0.16)
351
0.92
(4.2,0.12)
344
0.94
(4.2,0.11)
333
0.96
(3.5,0.10)
545
0.89
(5.4,0.15)
529
0.86
(8.6,0.19)
312
0.89
(6.8,0.15)
335
0.90
(7.3,0.17)
335
0.84
(10.0,0.23)
355
0.79
(9.3,0.21)
536
0.91
(4.9,0.13)
329
0.93
(5.2,0.13)
987
0.92
(4.4,0.12)
351
0.91
(5.3,0.13)
331
0.97
(4.4,0.11)
349
0.97
(3.2, 0.08)
359
1.00
(0.0, 0.00)
1059
AA
0.93
(5.5,0.13)
325
0.84
(7.3,0.18)
317
0.92
(5.6,0.15)
311
0.91
(7.0,0.17)
315
0.76
(13.9,0.38)
319
0.90
(6.7,0.15)
320
0.90
(7.1,0.15)
315
0.90
(6.4,0.15)
315
0.91
(7.2,0.18)
323
0.86
(7.3,0.17)
330
0.93
(4.3,0.12)
324
0.95
(4.1,0.11)
306
0.97
(3.4, 0.09)
326
0.91
(5.3,0.14)
297
0.88
(8.2, 0.20)
292
0.92
(6.4,0.14)
312
0.90
(7.3,0.18)
314
0.86
(9.1,0.24)
309
0.82
(9.4, 0.20)
319
0.90
(5.6,0.14)
308
0.94
(4.9,0.12)
332
0.94
(4.2,0.12)
294
0.92
(4.8,0.13)
310
0.95
(5.0,0.11)
328
0.96
(3.9, 0.09)
338
0.97
(2.9, 0.09)
342
1.00
(0.0, 0.00)
342
AB
0.88
(7.6,0.18)
320
0.81
(7.9, 0.20)
313
0.89
(6.1,0.16)
306
0.89
(5.9,0.16)
313
0.73
(10.1,0.33)
314
0.90
(5.6,0.16)
317
0.88
(7.5,0.17)
308
0.88
(7.3,0.19)
310
0.87
(6.2,0.17)
313
0.81
(8.2,0.19)
324
0.84
(8.5,0.19)
318
0.86
(8.1,0.18)
305
0.88
(8.3, 0.20)
320
0.89
(5.6,0.17)
289
0.89
(5.2,0.15)
295
0.89
(5.3,0.14)
309
0.83
(6.9,0.19)
306
0.83
(7.0,0.21)
301
0.88
(5.0,0.16)
314
0.80
(8.4, 0.20)
303
0.84
(9.9, 0.22)
326
0.83
(8.2,0.19)
290
0.85
(7.0,0.18)
309
0.86
(10.0,0.23)
324
0.90
(6.5,0.16)
333
0.90
(7.2,0.17)
337
0.92
(7.1,0.18)
317
1.00
(0.0, 0.00)
337
AC
0.89
(7.3,0.18)
340
0.81
(8.4, 0.22)
331
0.88
(7.5, 0.20)
325
0.87
(9.6, 0.23)
330
0.74
(15.7,0.43)
332
0.87
(8.4, 0.20)
334
0.86
(8.1,0.19)
327
0.87
(7.9,0.19)
329
0.87
(9.6, 0.24)
332
0.85
(9.0,0.21)
343
0.90
(6.2,0.17)
338
0.91
(6.3,0.17)
322
0.94
(4.5,0.12)
339
0.85
(8.1,0.18)
308
0.82
(10.3,0.25)
307
0.87
(8.9,0.21)
327
0.90
(9.9, 0.24)
329
0.81
(11.2,0.30)
324
0.74
(11.6,0.26)
333
0.87
(6.9,0.18)
321
0.91
(5.0,0.12)
346
0.90
(6.6,0.16)
305
0.89
(6.8,0.17)
323
0.94
(3.3, 0.09)
342
0.93
(5.4,0.15)
352
0.94
(4.4,0.13)
357
0.94
(3.8,0.11)
336
0.85
(9.2, 0.24)
330
1.00
(0.0, 0.00)
357
AD
0.94
(4.4,0.11)
346
0.86
(7.0,0.17)
336
0.93
(5.3,0.12)
330
0.92
(6.6,0.15)
336
0.79
(13.1,0.35)
338
0.91
(6.3,0.14)
339
0.91
(6.4,0.14)
333
0.91
(5.7,0.13)
335
0.92
(6.3,0.16)
338
0.89
(6.9,0.15)
349
0.93
(3.8,0.11)
343
0.95
(4.2,0.10)
327
0.96
(3.5,0.10)
345
0.90
(5.3,0.14)
313
0.88
(8.4,0.18)
311
0.91
(6.4,0.14)
333
0.91
(6.9,0.16)
332
0.86
(9.5, 0.22)
327
0.78
(9.9, 0.20)
339
0.93
(4.9,0.12)
327
0.94
(5.2,0.12)
351
0.94
(4.3,0.11)
311
0.93
(5.0,0.12)
328
0.97
(4.5,0.11)
348
0.98
(2.8, 0.08)
358
0.98
(1.8,0.07)
363
0.98
(2.9, 0.07)
341
0.89
(7.2,0.17)
337
0.95
(4.4,0.13)
356
1.00
(0.0, 0.00)
363
AE
0.88
(7.2,0.19)
326
0.81
(8.8, 0.23)
315
0.88
(7.4, 0.20)
308
0.85
(9.2, 0.23)
315
0.70
(15.0,0.42)
316
0.86
(8.0,0.21)
319
0.85
(8.2, 0.20)
314
0.86
(7.3, 0.20)
313
0.86
(9.2, 0.24)
318
0.84
(8.9, 0.22)
327
0.88
(6.2,0.18)
321
0.91
(5.9,0.17)
309
0.93
(4.5,0.13)
326
0.85
(7.7,0.18)
294
0.81
(10.6,0.25)
290
0.87
(8.3, 0.20)
313
0.87
(9.5, 0.24)
311
0.78
(11.0,0.30)
309
0.73
(11.5,0.25)
322
0.85
(7.2, 0.20)
306
0.90
(5.9,0.14)
330
0.91
(6.1,0.16)
294
0.88
(6.9,0.18)
308
0.93
(4.1,0.11)
326
0.93
(5.4,0.15)
337
0.92
(4.6,0.14)
341
0.93
(4.1,0.11)
319
0.86
(9.0, 0.23)
316
0.98
(3.0, 0.08)
334
0.93
(4.6, 0.14)
341
1.00
AF
0.89
(6.7,0.16)
323
0.92
(5.8,0.16)
316
0.89
(6.7,0.15)
312
0.90
(5.4, 0.14)
313
0.88
(7.6, 0.26)
316
0.89
(6.4,0.16)
317
0.88
(6.7,0.17)
311
0.88
(6.9,0.17)
313
0.90
(5.5,0.13)
319
0.90
(6.4,0.16)
329
0.89
(7.4,0.17)
321
0.89
(6.8,0.17)
306
0.88
(8.5,0.21)
323
0.85
(7.5, 0.20)
292
0.84
(7.0,0.18)
290
0.87
(6.7,0.16)
311
0.91
(4.8, 0.14)
312
0.88
(6.2,0.17)
308
0.69
(10.3,0.22)
316
0.89
(6.3,0.17)
308
0.86
(9.6, 0.23)
330
0.88
(7.0,0.18)
290
0.87
(6.8,0.18)
310
0.88
(9.8, 0.24)
326
0.89
(6.5,0.18)
335
0.88
(7.8,0.19)
342
0.87
(8.1,0.20)
319
0.79
(8.1,0.20)
313
0.84
(10.4,0.26)
336
0.89
(7.1,0.18)
339
0.82
AG
0.93
(5.1,0.12)
299
0.84
(6.6,0.17)
292
0.93
(4.7,0.11)
282
0.92
(4.8,0.12)
496
0.77
(11.3,0.32)
294
0.93
(4.6,0.12)
290
0.90
(5.2,0.13)
856
0.90
(5.6,0.14)
290
0.93
(5.1,0.14)
296
0.88
(6.4,0.15)
301
0.90
(5.9,0.13)
294
0.92
(5.4,0.12)
283
0.95
(5.0,0.14)
484
0.91
(4.9,0.14)
477
0.88
(6.1,0.16)
269
0.91
(5.5,0.13)
285
0.89
(6.0,0.15)
287
0.86
(7.6, 0.20)
304
0.85
(7.2,0.17)
478
0.88
(6.2, 0.14)
281
0.90
(6.9,0.17)
878
0.89
(5.8,0.15)
301
0.89
(5.1,0.14)
281
0.93
(6.9,0.17)
301
0.97
(3.3, 0.09)
308
0.95
(4.0,0.10)
919
0.97
(4.0,0.11)
292
0.95
(4.1,0.13)
291
0.96
(6.6,0.18)
304
0.97
(4.0,0.10)
311
0.94
AH
0.90
(6.2,0.15)
317
0.87
(6.8,0.18)
311
0.91
(5.7,0.14)
305
0.90
(6.5,0.16)
308
0.75
(12.5,0.36)
310
0.89
(5.7,0.15)
312
0.89
(6.1,0.15)
312
0.88
(6.8,0.16)
308
0.89
(6.7,0.18)
310
0.87
(7.5,0.16)
321
0.91
(5.0,0.13)
318
0.96
(4.0,0.11)
305
0.96
(4.3,0.10)
319
0.92
(4.6,0.13)
292
0.89
(6.1,0.18)
290
0.94
(4.7,0.12)
307
0.90
(6.3,0.17)
303
0.87
(7.8, 0.22)
298
0.82
(8.5,0.18)
310
0.90
(5.8,0.14)
306
0.96
(5.8,0.12)
325
0.96
(4.0,0.10)
291
0.96
(3.7,0.10)
304
0.92
(6.5,0.14)
319
0.93
(4.9,0.12)
328
0.93
(4.7,0.11)
337
0.95
(4.3,0.10)
313
0.90
(6.1,0.16)
310
0.91
(6.6,0.15)
326
0.95
(4.4,0.10)
333
0.92
Al
0.89
(7.5,0.16)
318
0.86
(7.1,0.18)
309
0.89
(6.0,0.15)
302
0.91
(5.3,0.14)
310
0.80
(9.4,0.31)
309
0.89
(5.3,0.15)
310
0.88
(6.1,0.16)
309
0.86
(6.4,0.17)
308
0.89
(5.8,0.16)
311
0.84
(7.7,0.18)
320
0.89
(6.5,0.15)
314
0.92
(6.4,0.14)
299
0.93
(6.8,0.16)
318
0.92
(5.3,0.14)
293
0.92
(4.7,0.14)
283
0.95
(3.5,0.10)
306
0.91
(5.3,0.14)
302
0.90
(5.7,0.17)
302
0.88
(5.5,0.14)
312
0.86
(7.3,0.17)
298
0.91
(8.4,0.18)
323
0.91
(6.4,0.15)
287
0.92
(4.9,0.13)
303
0.88
(9.2, 0.20)
319
0.92
(5.3,0.13)
329
0.91
(6.1,0.14)
335
0.94
(6.6,0.15)
312
0.93
(3.8,0.12)
310
0.89
(9.3,0.21)
326
0.93
(6.0,0.13)
333
0.89
AJ
0.96
(4.4,0.12)
338
0.88
(5.5,0.15)
329
0.96
(3.5,0.10)
322
0.96
(3.7,0.10)
328
0.84
(9.8, 0.29)
331
0.94
(4.1,0.12)
332
0.93
(5.0,0.14)
325
0.93
(5.0,0.13)
327
0.94
(4.1,0.12)
330
0.91
(5.6,0.14)
341
0.92
(4.8,0.13)
335
0.92
(5.4,0.13)
319
0.93
(5.7,0.17)
338
0.90
(5.4,0.16)
306
0.88
(5.4,0.16)
304
0.91
(5.0,0.14)
326
0.92
(5.5,0.13)
324
0.88
(6.3,0.17)
320
0.79
(8.1,0.19)
331
0.91
(5.9,0.14)
319
0.90
(7.5,0.19)
343
0.91
(5.3,0.15)
304
0.90
(5.0,0.15)
320
0.93
(8.2,0.19)
340
0.97
(3.5,0.11)
350
0.95
(4.9,0.14)
355
0.95
(5.1,0.15)
335
0.90
(5.5,0.16)
329
0.91
(7.4, 0.22)
348
0.96
(4.6,0.13)
354
0.89
December 2009 A-132
-------
s
AF
AG
AH
Al
AJ
T U V W X Y Z AA AB AC AD AE AF
(0.0,0.00) (10.0,0.26)
341 319
1.00
(0.0, 0.00)
342
AG
(6.2,0.18)
290
0.86
(7.0,0.16)
289
1.00
(0.0, 0.00)
951
AH
(5.6,0.15)
313
0.87
(7.1,0.18)
310
0.93
(4.8,0.12)
289
1.00
(0.0, 0.00)
337
Al
(8.4, 0.20)
314
0.87
(6.4,0.16)
313
0.94
(4.5,0.11)
283
0.97
(4.1,0.10)
307
1.00
(0.0, 0.00)
335
AJ
(8.0, 0.22)
332
0.91
(5.5,0.14)
331
0.96
(3.7,0.11)
304
0.92
(4.9,0.15)
327
0.92
(4.8,0.14)
324
1.00
(0.0, 0.00)
355
*#
0.8
0.6
o
13
o
O
0.4
0.2
10 20 30 40 50 60 70
Distance Between Samplers (km)
80
90
100
Figure A-63 PM2.s inter-sampler correlations as a function of distance between monitors for
New York, NY.
December 2009
A-133
-------
Philadelphia Combined Statistical Area
A
01
\ Philadelphia CSA
• PMz.s Monitors
Interstate Highways
— Major Highways
0 10 20 40 60 80
100
^Kilometers
Figure A-64. PM2.s monitor distribution and major highways, Philadelphia, PA.
December 2009
A-134
-------
AQSSitelD
Site A 10-003-1003
SiteB 10-003-1007
SiteC 10-003-1012
SiteD 10-003-2004
SiteE 24-015-0003
Site F 34-007-0003
SiteG 34-007-1007
SiteH 42-017-0012
Site I
SiteJ
SiteK
SiteL
SiteM
SiteN
SiteO
2=spring
3-sur
4=fall
ID
}°3 Mear
307 Ob.
X sc
303 50 -
303
307
312
40 -
Ol
— 30 -
_0
ra
concent
CD
10 -
ter
ng
nmer o -
ABCDEFGH
13.4 12,6 13.5 14,7 12.5 13.5 13.5 13.2
335 346 331 999 348 539 340 317
7.7 7.4 7.7 8.3 7.4 8.0 8.5 8.2
1234 1234 1234 1234 1234 1234 1234 1234
AQS Site ID
J K L M N 0
42-045-0002 Mean R2 15° 12'6 13'8 12'8 14'9 13'4
42-091-0013 Obs 277 331 307 890 805 596 780
42-101-0004 SD 83 3.1 7.7 8.4 8.0 8.3 7.7
42-101-0024 50 _
42-101-0047
42-101-0136
40 -
1
O1
3 30-
c
o
4->
ro
4->
C
-------
Table A-29. Inter-sampler correlation statistics for each pair of PM2.6 monitors reporting to AQS for
Philadelphia, PA.
A B
A 1.00 0.94
(0.0,0.00) (4.7,0.12)
335 305
B 1.00
(0.0, 0.00)
346
C
D
E
F
G
H
I
J
K
L
M
N
0
C D
0.96 0.98
(3.1,0.08) (3.2,0.08)
282 318
0.95 0.93
(4.3,0.12) (6.4,0.15)
288 329
1.00 0.96
(0.0, 0.00) (4.3, 0.09)
331 312
1.00
(0.0, 0.00)
999
LEGEND
R
(P90 COD)
E
0.92
(4.8,0.12)
311
0.94
(3.4,0.11)
318
0.95
(3.5,0.11)
289
0.91
(6.5,0.15)
325
1.00
(0.0, 0.00)
348
F
0.96
(3.5,0.10)
312
0.92
(5.2, 0.14)
313
0.94
(4.7,0.12)
292
0.94
(4.9,0.12)
490
0.91
(5.6,0.14)
320
1.00
(0.0, 0.00)
539
G
0.93
(4.2,0.11)
308
0.88
(6.0,0.15)
315
0.88
(5.3,0.14)
286
0.92
(5.0,0.14)
317
0.87
(6.1,0.15)
321
0.95
(3.4, 0.09)
317
1.00
(0.0, 0.00)
340
H
0.89
(5.3,0.13)
289
0.83
(6.8,0.17)
293
0.88
(6.0,0.14)
270
0.88
(6.3,0.15)
297
0.83
(6.7,0.16)
301
0.90
(5.3,0.13)
296
0.90
(4.8,0.14)
295
1.00
(0.0, 0.00)
317
I
0.95
(4.2,0.12)
247
0.90
(6.7,0.17)
253
0.93
(3.5,0.12)
242
0.94
(4.1,0.12)
257
0.90
(6.6,0.16)
255
0.92
(5.4,0.14)
261
0.90
(5.9,0.16)
258
0.84
(5.7,0.16)
240
1.00
(0.0, 0.00)
277
J
0.92
(4.6, 0.14)
298
0.87
(6.5,0.18)
302
0.88
(6.6,0.16)
278
0.90
(5.3,0.14)
312
0.86
(7.1,0.19)
310
0.89
(5.9,0.16)
309
0.87
(6.2,0.17)
305
0.83
(8.0,0.19)
288
0.87
(5.5,0.17)
248
1.00
(0.0, 0.00)
331
K
0.86
(4.7,0.15)
277
0.81
(5.9,0.18)
285
0.84
(5.5,0.17)
261
0.85
(5.8,0.18)
287
0.86
(5.7,0.15)
287
0.87
(4.4,0.15)
284
0.85
(4.7,0.16)
289
0.89
(4.4,0.13)
275
0.81
(5.7,0.17)
228
0.79
(7.4,0.21)
278
1.00
(0.0, 0.00)
307
L
0.96
(3.5, 0.08)
283
0.91
(6.5, 0.14)
293
0.93
(5.0,0.12)
281
0.95
(4.3,0.11)
801
0.88
(6.8,0.15)
296
0.96
(3.7,0.10)
466
0.93
(3.7, 0.09)
288
0.90
(5.0,0.13)
273
0.91
(4.9,0.14)
235
0.89
(5.8,0.15)
282
0.87
(4.7,0.15)
268
1.00
(0.0, 0.00)
890
M
0.96
(3.7,0.10)
243
0.92
(5.0,0.14)
253
0.93
(4.8,0.13)
245
0.93
(5.6,0.14)
732
0.90
(5.3,0.13)
255
0.96
(3.6,0.10)
437
0.97
(3.1,0.09)
251
0.94
(4.0,0.12)
234
0.92
(5.4,0.15)
215
0.89
(6.4,0.17)
246
0.95
(3.7,0.13)
230
0.98
(3.1,0.09)
672
1.00
(0.0, 0.00)
805
N
0.95
(4.5,0.12)
236
0.88
(7.3,0.17)
238
0.91
(6.0, 0.14)
225
0.93
(4.2,0.10)
540
0.87
(7.0,0.18)
242
0.95
(4.5,0.13)
414
0.92
(5.7,0.13)
235
0.87
(5.9,0.17)
215
0.90
(5.2,0.16)
196
0.89
(5.7,0.13)
237
0.84
(6.8, 0.20)
211
0.95
(3.7,0.11)
512
0.95
(4.7,0.14)
495
1.00
(0.0, 0.00)
596
0
0.97
(3.2, 0.08)
236
0.89
(5.9,0.13)
246
0.93
(4.6,0.11)
237
0.95
(4.5,0.11)
704
0.89
(5.7,0.13)
254
0.96
(3.4, 0.09)
396
0.96
(3.5, 0.08)
240
0.89
(4.8,0.13)
227
0.92
(5.1,0.14)
195
0.91
(5.0,0.14)
231
0.86
(4.3,0.13)
212
0.97
(3.4, 0.07)
630
0.96
(3.2, 0.09)
563
0.97
(3.5,0.10)
447
1.00
(0.0, 0.00)
780
December 2009 A-136
-------
0.8
0.6
o
13
o
O
0.4
0.2
10 20 30 40 50 60 70
Distance Between Samplers (km)
80 90 100
Figure A-66. PM2.s inter-sampler correlations as a function of distance between monitors for
Philadelphia, PA.
December 2009
A-137
-------
Phoenix Core Based Statistical Area
A
01
\ | Phoenix CBSA
• PM2.5 Monitors
Interstate Highways
— Major Highways
0 1020 40 60 80 100
i Kilometers
Figure A-67. PM2.s monitor distribution and major highways, Phoenix, AZ.
December 2009
A-138
-------
Site A
SiteB
SiteC
SiteD
SiteE
AQSSitelD
04-013-0019
04-013-4003
04-013-9997
04-021-0001
04-021-3002
en
c
o
u
c
o
u
1=winter
2=spring
3=summe
4=fall
A B
Mean 12.3 12.6
Obs 370 352
SD 7.6 7.3
50 -
40 -
30 -
20 -
10 -
r 0 -
II
c
9.8
360
5.5
II
i
D E
8.9 5.9
227 325
4.4 2.8
1
\
1234 1234 1234 1234 1234
Figure A-68. Box plots illustrating the seasonal distribution of 24-h avg PM2.6 concentrations
for Phoenix, AZ.
December 2009
A-139
-------
Table A-30. Inter-sampler correlation statistics for each pair of PM2.6 monitors reporting to AQS for
Phoenix, AZ.
A B C D E
1.00
0.87
0.92
0.50
0.12
(0.0,0.00)
(6.4,0.15)
(6.5,0.16)
(10.4,0.25)
(14.4,0.40)
370
345
355
222
321
1.00
0.89
0.54
0.23
(0.0, 0.00)
(6.8,0.17)
(9.6, 0.25)
(13.2,0.40)
352
338
212
307
1.00
0.54
0.18
(0.0, 0.00)
(7.2, 0.20)
(9.3,0.33)
360
216
315
LEGEND
R
(P90, COD)
N
1.00
(0.0, 0.00)
227
0.51
(7.8, 0.27)
200
1.00
(0.0, 0.00)
325
December 2009
A-140
-------
0.8
0.6
o
o
0.4
0.2
10 20 30
40 50 60
Distance Between Samplers (km)
70 80 90 100
Figure A-69. PM2.s inter-sampler correlations as a function of distance between monitors for
Phoenix, AZ.
December 2009
A-141
-------
01
r
Pittsburgh Combined Statistical Area
0 10 20
40
60
Pittsburgh CSA
• PM2.5 Monitors
Interstate Highways
Major Highways
80
100
—i Kilometers
Figure A-70. PM2.s monitor distribution and major highways, Pittsburgh, PA.
December 2009
A-142
-------
Mean 15.1 19.8 13.2 13.6 15.1 16.4
Obs 1063 1066 306 165 332 337
SD 8.9 14.7 8.0 8.5 9.0 9.5
70-
AQSSitelD
Site A 42-003-0008
SteB 42-003-0054 60~
SiteC 42-003-0067
SiteD 42-003-0095 ~ 50_
SiteE 42-003-1008 -t
SiteF 42-003-1301 3
C 40 -
g
Ł
OJ
u
c.
o
u
20-
1 0 -
1 -winter
2=spring
3=summer
4=fall 0 -
II
1
234 234 234 1234 1234 1234
G H J K L
Mean 15.3 16.4 15.5 14.8 13.4 15.4
Ofcs 171 328 354 345 966 350
SD 8.3 9.3 8.6 8.0 8.6 8.7
70 -
AQS Site ID
SiteG 42-003-3007
SiteH 42-007-0014 60 "
Site! 42-125-0005
SiteJ 42-125-0200 ~ 5Q_
SiteK 42-125-5001 -5
en
SiteL 42-129-0008 3
c 40-
O
c 30-
OJ
u
c
0
u
20-
1 =winter
2=spring
3=summer
4=fall o -
1234 1234 1234 1234 1234
Figure A-71. Box plots illustrating the seasonal distribution of 24-h avg PM2.s concentrations
for Pitsburgh, PA.
December 2009
A-143
-------
Table A-31. Inter-sampler correlation statistics for each pair of PM2 5 monitors reporting to AQS for
Pittsburgh, PA.
ABCDEFGH I JKL
A 1.00 079 0.95 0.92 0.93 0.95 0.95 0.85 0.90 0.93 0.91 0.88
(0.0,0.00) (15.9,0.19) (5.6,0.13) (47,0.11) (47,0.11) (4.9,0.10) (3.8,0.10) (6.4,0.13) (6.4,0.13) (5.0,0.12) (6.0,0.13) (5.6,0.12)
1063 1035 298 164 323 329 170 319 344 337 934 340
_B UK) 071 0.65 0.80 0.85 076 0.69 071 0.68 0.68 0.67
(16.9,0.24) (17.4,0.25) (14.4,0.19) (12.5,0.14) (157,0.20) (17.0,0.19) (157,0.21) (17.8,0.23) (19.3,0.25) (15.9,0.21)
303 165 329 335 171 324 350 341 938 346
1.00 0.93 0.90 0.91 0.94 0.80 0.93 0.96 0.95 0.91
(0.0,0.00) (2.8,0.09) (6.6,0.16) (87,0.17) (6.0,0.14) (9.4,0.19) (67,0.15) (4.6,0.12) (4.5,0.10) (6.5,0.15)
306 144 282 282 148 268 290 286 270 286
100 0.84 0.87 0.91 0.79 0.89 0.91 0.97 0.85
(0.0,0.00) (6.4,0.15) (8.5,0.16) (5.8,0.13) (9.2,0.17) (5.9,0.13) (4.6,0.11) (3.1,0.08) (6.5,0.15)
165 153 161 158 156 158 155 146 157
yra 0.90 0.90 OJH 0.55 g.86 g.88 o.83
(6.4,0.13) (6.5,0.13) (6.8,0.14) (8.3,0.16)
LEGEND 332 313 157 295 320 315 290 318
Pearson R 1.00 OJM 0.82 0.88 0.88 0.89 0.86
(0.0,0.00) (67,0.13) (7.4,0.14) (7.1,0.15) (7.9,0.15) (8.8,0.17) (7.0,0.14)
337 167 302 327 319 296 322
100 0.78 0.94 0.93 0.90 0.91
(0.0,0.00) (7.3,0.16) (4.0,0.10) (5.0,0.11) (6.6,0.15) (5.0,0.13)
171 159 163 159 149 161
yra 0.80 g78 0.82 0.70
(0.0,0.00) (8.4,0.15) (8.2,0.17) (9.0,0.18) (9.2,0.18)
328 317 3JK 288 314
1.00 0.93 0.89 0.88
354 334 310 339
1.00 0.93
(5.5,0.12) (5.9,0.13)
345 302 331
0.86
(0.0,0.00) (6.9,0.15)
966 306
(0.0, 0.00)
350
December 2009 A-144
-------
* • ** * •
• • - »* • •
* * **«»
0.8 -
0.6 -
o
1
g
5
0.4 -
0.2 -
10 20 30
40 50 60 70
Distance Between Samplers (km)
80 90 100
Figure A-72. PM2.s inter-sampler correlations as a function of distance between monitors for
Pittsburgh, PA.
December 2009
A-145
-------
Riverside Core Based Statistical Area
A
01
\ | Riverside CBSA
• PM2.5 Monitors
Interstate Highways
— Major Highways
0 1020 40 60 80 100
i Kilometers
Figure A-73. PM2.s monitor distribution and major highways, Riverside, CA.
December 2009
A-146
-------
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
SiteG
AQS Site ID
06-065-1003
06-065-2002
06-065-8001
06-071-0025
06-071-0306
06-071-2002
06-071-9004
en
c
o
4->
TO
•>->
C.
OJ
u
c
o
u
1 = winter
2=spring
3=summer g
4=fall
A B
Mean 17.7 9.9
Obs 314 310
SD 11.6 4.9
70 -
60 -
50 -
40 -
30 -
20 -
10 -
0 -
I*
C D E
19.7 18.4 9.9
934 319 236
12.7 10.9 4.4
I
IPI
F G
18.4 17.7
328 310
11.9 12.2
'
1234 1234 1234 1234 1234 1234 1234
Figure A-74. Box plots illustrating the seasonal distribution of 24-h avg PM2.6 concentrations
for Riverside, CA.
December 2009
A-147
-------
Table A-32. Inter-sampler correlation statistics for each pair of PM2.6 monitors reporting to AQS for
Riverside, CA.
ABC
A 1.00 0.45 0.96
(0.0,0.00) (20.6,0.32) (5.0,0.10)
314 269 297
B 1.00 0.49
(0.0, 0.00) (22.7, 0.35)
310 289
C 1.00
(0.0, 0.00)
934
D
LEGEND
E R
(P90, COD)
N
F
G
D
0.92
(7.2,0.13)
282
0.49
(20.9, 0.34)
270
0.91
(8.2,0.14)
300
1.00
(0.0, 0.00)
319
E
0.36
(22.1,0.35)
191
0.42
(8.2, 0.25)
203
0.37
(26.6, 0.37)
227
0.36
(20.1,0.35)
195
1.00
(0.0, 0.00)
236
F
0.94
(6.0,0.12)
281
0.49
(19.7,0.33)
285
0.92
(6.9,0.12)
302
0.93
(6.7,0.14)
289
0.40
(21.1,0.36)
201
1.00
(0.0, 0.00)
328
G
0.90
(5.7,0.13)
273
0.50
(18.8,0.31)
266
0.91
(7.6,0.12)
287
0.82
(9.6,0.17)
274
0.41
(21.6,0.34)
190
0.90
(6.7,0.12)
276
1.00
(0.0, 0.00)
310
December 2009 A-148
-------
1.2
0.8
0.6
o
o
0.4
0.2
* «» *
10 20 30 40 50 60 70
Distance Between Samplers (km)
80 90 100
Figure A-75. PM2.s inter-sampler correlations as a function of distance between monitors for
Riverside CA.
December 2009
A-149
-------
Seattle Combined Statistical Area
01
r
\ Seattle CSA
• PM2.5 Monitors
Interstate Highways
— Major Highways
0 10 20 40 60 80
100
—i Kilometers
Figure A-76. PM2.s monitor distribution and major highways, Seattle, WA.
December 2009
A-150
-------
AQS Site ID
Site A 53-033-0024
Site B 53-053-0029
SiteC 53-061-1007
c
o
ro
c
Ol
u
c
o
u
1=winter
2=spring
3=surr
4=fall
ABC
Mean 8.9 10.2 9.2
Obs 352 354 591
SD 7.3 10.1 7.9
60 -
50 -
40 -
30 -
10 -
;r 0 -
ll
ll
I
1234 1234 1234
Figure A-77. Box plots illustrating the seasonal distribution of 24-h avg PM2.s concentrations
for Seattle, WA.
December 2009
A-151
-------
Table A-33. Inter-sampler correlation statistics for each pair of PM2 5 monitors reporting to AQS for
Seattle, WA.
1.00
0.89
352
LEGEND
R
(P90, COD)
N
337
1.00
354
0.86
(0.0,0.00) (6.3,0.16) (4.5,0.14)
331
0.80
(0.0, 0.00) (7.8, 0.20)
335
1.00
(0.0, 0.00)
591
0.8
0.6
o
13
o
O
0.4
0.2
10 20 30 40 50 60 70
Distance Between Samplers (km)
80
90
100
Figure A-78. PM2.s inter-sampler correlations as a function of distance between monitors for
Seattle, WA.
December 2009
A-152
-------
St. Louis Combined Statistical Area
01
r
St. Louis CSA
• PM2.5 Monitors
Interstate Highways
— Major Highways
0 10 20 40 60 80
100
—i Kilometers
Figure A-79. PM2.s monitor distribution and major highways, St. Louis, MO.
December 2009
A-153
-------
AQSSitelD A B C D E F
Site A 17-083-1001 Megn 132 16g R6 144 158 142
SiteB 17-119-1007
SiteC 17-119-2009 Obs 173 329 163 349 166 349
SiteD 17-119-3007 SD 7.9 8.2 7.7 7.5 7.6 7.1
SiteE 17-163-0010 50-
SiteF 17-163-4001
40 -
li
1 30-
c
o
ro
I 2°-
— ' \ i
° III
10 -
1=winter
2=spring
3=summer 0 -
4=fall 1234 1234 1234 1234 1234 1234
AQSSitelD G H J K L
SiteG 29-099-0012 Mean 13g 132 135 144 144 146
SiteH 29-183-1002
Site I 29-189-2003 °bs 104° 566 619 1049 1038 1046
SiteJ 29-510-0007 SD 7.4 7.4 7.3 7.3 7.5 7.5
SiteK 29-510-0085 50 -
SiteL 29-510-0087
40 -
^
cn
5 30 -
c
0
TO
C
-------
Table A-34.
Site A
A 1.00
(0.0,0.00)
173
B
C
D
E
F
G
H
I
J
K
L
Inter-sampler correlation statistics for each pair
St. Louis, MO.
B C D E F
0.85 0.93 0.89 0.88 0.86
(10.5,0.23) (4.7,0.17) (5.0,0.17) (7.3,0.20) (6.2,0.18)
156 129 162 146 156
1.00 0.89 0.86 0.85 0.82
(0.0,0.00) (8.6,0.16) (7.4,0.16) (7.7,0.16) (8.6,0.17)
329 135 301 156 306
1.00 0.94 0.91 0.88
(0.0,0.00) (4.0,0.11) (6.4,0.13) (5.7,0.13)
163 139 124 133
1.00 0.89 0.84
(0.0,0.00) (5.7,0.13) (6.0,0.15)
349 156 314
1.00 0.90
(0.0,0.00) (5.5,0.12)
166 152
1.00
(0.0,0.00)
LEGEND 349
R
(P90, COD)
N
G
0.85
(4.8,0.17)
167
0.88
(7.8,0.17)
312
0.90
(5.5,0.13)
158
0.89
(4.9,0.12)
331
0.91
(6.2,0.13)
159
0.89
(5.4,0.12)
333
1.00
(0.0, 0.00)
1040
Of PM2.6
H
0.93
(4.1,0.13)
158
0.89
(8.2,0.18)
305
0.96
(3.9,0.11)
141
0.94
(4.3,0.12)
315
0.90
(5.8,0.16)
153
0.86
(6.1,0.16)
317
0.93
(4.3,0.10)
533
1.00
(0.0, 0.00)
566
monitors reporting to AQS for
i
0.86
(4.4,0.16)
162
0.88
(7.9,0.17)
318
0.94
(5.3,0.11)
144
0.92
(4.5,0.11)
326
0.91
(5.3,0.14)
157
0.88
(5.4,0.13)
332
0.94
(3.3, 0.08)
586
0.96
(3.0, 0.08)
550
1.00
(0.0, 0.00)
619
J
0.84
(6.0,0.18)
168
0.86
(7.7,0.17)
316
0.90
(5.7,0.13)
158
0.89
(4.7,0.13)
335
0.93
(5.1,0.13)
160
0.88
(5.3,0.14)
337
0.96
(2.9, 0.08)
994
0.95
(4.1,0.12)
552
0.96
(3.1,0.09)
605
1.00
(0.0, 0.00)
1049
K
0.84
(5.7,0.19)
169
0.87
(7.5,0.16)
316
0.89
(5.6,0.14)
160
0.88
(4.6,0.12)
332
0.91
(4.9,0.13)
163
0.85
(5.6,0.14)
332
0.93
(3.9,0.10)
987
0.95
(3.8,0.12)
546
0.95
(3.1,0.10)
599
0.96
(2.5, 0.09)
1001
1.00
(0.0, 0.00)
1038
0.88
(5.3,
166
0.89
(6.8,
315
0.94
(4.4,
156
0.92
(3.9,
336
0.95
(3.7,
160
0.88
(5.4,
334
0.94
(3.8,
992
0.96
(4.0,
544
0.96
(3.4,
598
0.97
(2.5,
1007
0.97
(1.9,
991
1.00
(0.0,
L
0,
0,
0,
o.
0,
0,
0,
0,
.17)
.14)
.11)
.11)
.10)
.13)
.10)
.11)
0.09)
0.08)
0.07)
0.00)
1046
December 2009 A-155
-------
*••*..>
• ». •«*.
«. * • **
0.2
0 10 20 30
40 50 60 70
Distance Between Samplers (km)
90 100
Figure A-81 PM2.s inter-sampler correlations as a function of distance between monitors for
St. Louis, MO.
December 2009
A-156
-------
Atlanta Combined Statistical Area
Atlanta CSA
• PMio Monitors
Interstate Highways
Major Highways
100
Kilometers
Figure A-82. PM™ monitor distribution and major highways, Atlanta, GA.
December 2009
A-157
-------
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
AQS Site ID
13-089-2001
13-097-0003
13-121-0001
13-121-0032
13-121-0048
13-255-0002
ID Mean 24.7
31 Obs 172
33 SD
31 70 -
32
18
32 60 -
^ 50 -
en
c 40 -
0
4-*
ro
c 30 -
CD
U
O
U 20 -
10 -
1=winter
2=spring
3=summer^ "
4=fall
13.0
1234 '
B C D E
21.4 23.4 26.6 25.0
178 171 174 995
9.3 9.5 11.8 11.5
234 1234 1234 1234
F
21.6
178
9.7
1234
Figure A-83. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations
for Atlanta, GA.
December 2009
A-158
-------
Table A-35. Inter-sampler correlation statistics for each pair of PMio monitors reporting to AQS for
Atlanta, GA.
Site
1.00
0.69
0.74
0.78
0.70
0.59
(0.0,0.00)
(18.0,0.22)
(15.0,0.20)
(13.0,0.20)
(16.0,0.22)
(20.0, 0.24)
172
169
162
165
158
164
1.00
0.88
0.79
0.71
0.82
(0.0, 0.00)
(6.0,0.12)
(14.5,0.17)
(16.0,0.18)
(10.0,0.14)
178
167
170
162
169
1.00
0.88
0.84
0.82
(0.0, 0.00)
(9.0,0.13)
(10.0,0.13)
(9.0,0.15)
171
162
155
161
LEGEND
R
(P90, COD)
N
1.00
0.75
(0.0,0.00)
(12.0,0.15)
174
158
1.00
0.74
(15.0,0.20)
166
0.67
(0.0,0.00)
(17.0,0.19)
995
163
1.00
(0.0, 0.00)
178
December 2009
A-159
-------
0.8
0.6
o
o
0.4
0.2
10 20 30
40 50 60
Distance Between Samplers (km)
70 80 90 100
Figure A-84. PM™ inter-sampler correlations as a function of distance between monitors for
Atlanta, GA.
December 2009
A-160
-------
Birmingham Combined Statistical Area
A
01
Birmingham CSA
• PMio Monitors
Interstate Highways
— Major Highways
0 10 20 40 60 80
100
—i Kilometers
Figure A-85. PM™ monitor distribution and major highways, Birmingham, AL.
December 2009
A-161
-------
AQS Site ID
SiteA 01-073-0002
SiteB 01-073-0023
SiteC 01-073-0034
SiteD 01-073-1003
SiteE 01-073-1008
SiteF 01-073-1010
SiteG 01-073-2003
SiteH 01-073-6002
Site I 01-073-6003
SiteJ 01-073-6004
c
O
concen
1=winter
2=spring
3=summer
4=fall
A B C D E F
Mean 29.1 32.9 26.5 27.1 27.6 24.7
Obs 180 1095 224 183 179 179
3D 12.6 20.0 11.4 11.8 14.3 11.0
150 -
140 -
130 -
120 -
110 -
100 -
90 -
80 -
60 -
50 -
40 -
30 -
20 -
0 -
III
I
H
i
i
i
I
il
i il I
T
ll
i
"
i
G H J
26.0 28.8 38.0 48.2
1090 181 1087 1080
14.0 12.6 29.5 30.0
I
\
I
'
1
1
1234 1234 1234 1234 1234 1234 1234 1234 1234 1234
Figure A-86. Box plots illustrating the seasonal distribution of 24-h avg PMio concentrations
for Birmingham, AL
December 2009
A-162
-------
Table A-36. Inter-sampler correlation statistics for each pair
Birmingham, AL
ABC
A 1.00 0.80 0.88
(0.0,0.00) (23.0,0.16) (11.0,0.11)
180 180 174
B 1.00 0.82
(0.0,0.00) (23.0,0.17)
1095 224
C 1.00
(0.0, 0.00)
224
D
E
F
LEGEND
R
G (P90, COD)
N
H
I
J
D E F
0.86 0.78 0.84
(12.0,0.13) (12.0,0.14) (13.0,0.13)
180 176 171
0.74 0.61 0.73
(25.0,0.21) (26.0,0.20) (26.0,0.19)
183 179 179
0.84 0.66 0.78
(10.0,0.12) (15.0,0.16) (12.0,0.14)
175 171 168
1.00 0.67 0.79
(0.0,0.00) (15.0,0.17) (12.0,0.15)
183 178 173
1.00 0.67
(0.0,0.00) (16.0,0.15)
179 169
1.00
(0.0, 0.00)
179
of PMio monitors reporting to AQS for
G
0.77
(15.0,0.18)
180
0.75
(25.0, 0.20)
1090
0.74
(14.0,0.17)
224
0.76
(14.0,0.17)
183
0.64
(18.0,0.18)
179
0.75
(14.0,0.16)
179
1.00
(0.0, 0.00)
1090
H
0.78
(14.0,0.15)
178
0.71
(25.0, 0.22)
181
0.80
(13.0,0.15)
173
0.84
(11.0,0.12)
180
0.56
(19.0,0.20)
176
0.74
(15.0,0.17)
171
0.76
(15.0,0.19)
181
1.00
(0.0, 0.00)
181
I
0.41
(41.0,0.30)
179
0.26
(51.0,0.33)
1087
0.33
(43.0, 0.32)
222
0.45
(42.0, 0.30)
182
0.33
(45.0, 0.32)
178
0.36
(43.0, 0.32)
178
0.59
(43.0, 0.27)
1083
0.58
(38.0, 0.27)
180
1.00
(0.0, 0.00)
1087
J
0.29
(68.0, 0.34)
177
0.23
(57.0, 0.36)
1080
0.41
(62.0, 0.34)
221
0.41
(65.5, 0.34)
180
0.12
(71.0,0.39)
176
0.21
(71.0,0.38)
177
0.15
(63.0, 0.39)
1075
0.50
(59.0,0.31)
178
0.05
(72.0, 0.40)
1072
1.00
(0.0, 0.00)
1080
December 2009 A-163
-------
0.8
0.6
* • * *
* 1 «
o
o
0.4
0.2
10 20 30
40 50 60
Distance Between Samplers (km)
70 80 90 100
Figure A-87 PM™ inter-sampler correlations as a function of distance between monitors for
Birmingham, AL
December 2009
A-164
-------
Boston Combined Statistical Area
A
01
Boston CSA
• PMio Monitors
Interstate Highways
Major Highways
0 10 20 40 60 80
100
^Kilometers
Figure A-88. PM™ monitor distribution and major highways, Boston, MA.
December 2009
A-165
-------
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
SiteG
SiteH
AQS Site ID
25-025-0042
25-027-0023
33-01 1-0020
33-015-0014
44-003-0002
44-007-0022
44-007-0026
44-007-0027
1 =winter
2=spring
3=summer
4=fall
ABCDEFGH
Mean 16.4 21.6 16.5 14.8 10.7 17.0 21.3 18.8
Ote 191 174 182 175 171 182 169 168
SD
60 -
50-
-I 40-
Ol
3.
C
O
tXJ
concent
0
1
10 -
o -
7.7 11.9 9.2 8.2 7.0 8.7 10.8 9.4
\ [
\
1234 1234 12341234 1234 1234 1234 1234
Figure A-89. Box plots illustrating the seasonal distribution of 24-h avg PMi0 concentrations
for Boston, MA.
Table A-37. Inter-sampler correlation statistics for each pair of PMio monitors reporting to AQS for
Boston, MA.
Site A B
A 1.00 0.69
(0.0,0.00) (15.0,0.22)
191 169
B 1.00
(0.0, 0.00)
174
C
D
LEGEND
/pqn (*oru
F n
G
C
0.69
(12.0,0.20)
179
0.66
(17.0,0.24)
167
1.00
(0.0, 0.00)
182
D
0.73
(10.0,0.22)
173
0.56
(19.0,0.28)
161
0.72
(10.0,0.22)
170
1.00
(0.0, 0.00)
175
E
0.71
(13.0,0.30)
171
0.45
(24.0, 0.39)
158
0.47
(17.0,0.33)
168
0.63
(11.0,0.29)
163
1.00
(0.0, 0.00)
171
F
0.84
(8.0,0.14)
182
0.69
(15.0,0.21)
169
0.62
(12.0,0.21)
179
0.68
(10.0,0.23)
173
0.84
(13.0,0.29)
171
1.00
(0.0, 0.00)
182
G
0.70
(15.0,0.20)
169
0.77
(12.0,0.17)
156
0.64
(16.0,0.26)
166
0.59
(19.0,0.30)
161
0.58
(22.0, 0.38)
161
0.81
(11.0,0.16)
169
1.00
(0.0, 0.00)
H
0.79
(10.0,0.17)
167
0.65
(16.0,0.20)
154
0.59
(16.0,0.24)
164
0.69
(13.0,0.26)
158
0.80
(15.0,0.33)
157
0.95
(5.0,0.11)
167
0.79
(10.0,0.13)
169 154
H
1.00
(0.0, 0.00)
168
December 2009
A-166
-------
0.8
0.6
o
u
0.4 -
0.2
10 20 30 40 50 60 70
Distance Between Samplers (km)
90 100
Figure A-90 PM™ inter-sampler correlations as a function of distance between monitors for
Boston, MA.
December 2009
A-167
-------
Chicago Combined Statistical Area
01
r
J Chicago CSA
• PMio Monitors
Interstate Highways
Major Highways
0 10 20 40 60 80
100
—i Kilometers
Figure A-91. PM™ monitor distribution and major highways, Chicago, IL.
December 2009
A-168
-------
Site A
SiteB
SteC
SiteD
SiteE
SiteF
SteG
SiteH
Site ID ABCDEFGH
1-0001 Mean 22.4 24.7 30.4 32.4 24.5 27.2 26,9 21.6
1-0022 Obs 179 1077 176 1058 174 176 174 176
1-0060 3D
1-1016 8° -
1-1901
1-2001 ?0 _
1 iini
I -OoU I
7-1002
60 -
-- 50 -
CTi
a.
O 40 -
m
ai 30 -
u
c
o
u 20 -
10 -
1 -winter
2=spring _
3=summer
4=fall
10.9 13.1 15.1 173 11.9 13.4 12,3 12.3
1234 1234 12341234
1234 1234 1234 234
AQSSitelD
Site I 18-089-0006
SiteJ 18-089-0022
SiteK 18-089-0023
SiteL 18-089-2010
SiteM 18-127-0022
SiteN 18-127-0023
SiteO 18-127-0024
1 =winter
2=spring
3=sum
4=fall
ID J K L M N 0
06 Mean 28.2 27.9 26.4 16.6 17.2 24.1 18.3
22 Obs 182 1059 172 169 173 1051 169
SD 13.8 18.1 9.5 9.8 11.2 15.1 10.7
10 8°-
22
23 70-
24
" 60 -
_E
en
3 50-
c
o
H 40 -
c
a;
u
c 30 -
o
u
20 -
10 -
ter
ng
mer o -
1234 1234 1234 1234 1234 1234 1234
Figure A-92. Box plots illustrating the seasonal distribution of 24-h avg PMi0 concentrations
for Chicago, IL
December 2009
A-169
-------
Table A-38. Inter-sampler correlation statistics for each pair of PMio monitors reporting to AQS for
Chicago, IL.
Site A B
A 1.00 0.78
(0.0,0.00) (15.0,0.18)
179 176
B 1.00
(0.0, 0.00)
1077
C
D
E
F
G
H
I
J
K
L
M
N
0
C
0.68
(23.0, 0.24)
173
0.66
(23.0, 0.23)
173
1.00
(0.0, 0.00)
176
LEGEND
R
(P90, COD)
N
D E
0.83 0.93
(25.0,0.22) (8.0,0.10)
174 171
0.74 0.76
(23.0,0.21) (14.0,0.17)
1040 171
0.63 0.72
(26.0,0.23) (21.0,0.21)
171 169
1.00 0.79
(0.0,0.00) (27.0,0.21)
1058 169
1.00
(0.0, 0.00)
174
F
0.92
(11.0,0.13)
173
0.84
(12.0,0.15)
173
0.74
(18.5,0.19)
170
0.85
(19.0,0.17)
171
0.93
(9.0,0.10)
168
1.00
(0.0, 0.00)
176
G
0.86
(12.0,0.17)
171
0.79
(13.0,0.18)
171
0.64
(19.0,0.21)
168
0.79
(23.0,0.19)
169
0.84
(13.0,0.16)
166
0.84
(12.0,0.15)
169
1.00
(0.0, 0.00)
174
H
0.79
(12.0,0.18)
167
0.74
(17.0,0.23)
173
0.62
(22.0, 0.27)
164
0.74
(27.0, 0.28)
171
0.86
(10.0,0.16)
163
0.86
(13.0,0.19)
165
0.77
(15.0,0.22)
162
1.00
(0.0, 0.00)
176
I
0.75
(13.0,0.18)
179
0.68
(16.0,0.19)
179
0.62
(23.0, 0.20)
176
0.70
(20.0,0.19)
177
0.74
(13.0,0.16)
174
0.77
(12.0,0.14)
176
0.69
(14.0,0.18)
174
0.71
(16.0,0.23)
170
1.00
(0.0, 0.00)
182
J
0.14
(22.0, 0.28)
173
0.36
(22.0, 0.24)
1041
0.19
(26.5, 0.28)
170
0.23
(32.0, 0.29)
1022
0.17
(22.0, 0.26)
168
0.21
(23.0, 0.25)
170
0.28
(23.0, 0.26)
168
0.18
(27.0, 0.30)
169
0.24
(22.0, 0.24)
176
1.00
(0.0, 0.00)
1059
K
0.69
(15.0,0.21)
169
0.73
(16.0,0.19)
169
0.49
(24.0, 0.23)
166
0.69
(24.0, 0.23)
168
0.70
(15.0,0.19)
164
0.75
(16.0,0.17)
166
0.74
(14.0,0.18)
165
0.66
(18.0,0.25)
161
0.69
(12.0,0.15)
172
0.49
(15.0,0.20)
166
1.00
(0.0, 0.00)
172
L
0.89
(13.0,0.22)
166
0.81
(18.0,0.27)
166
0.66
(29.0, 0.37)
163
0.82
(31.0,0.36)
166
0.89
(15.0,0.25)
161
0.89
(18.0,0.28)
163
0.86
(19.0,0.31)
161
0.83
(13.0,0.23)
157
0.75
(20.0, 0.32)
169
0.38
(25.0, 0.34)
163
0.80
(17.0,0.32)
161
1.00
(0.0, 0.00)
169
M
0.55
(21.0,0.30)
170
0.66
(23.0,0.31)
170
0.39
(33.0, 0.40)
167
0.61
(36.0, 0.39)
168
0.53
(22.0, 0.33)
166
0.62
(25.0, 0.34)
167
0.52
(24.0, 0.36)
165
0.59
(19.0,0.29)
161
0.50
(26.0, 0.37)
173
0.22
(28.0, 0.36)
168
0.54
(24.0, 0.35)
165
0.60
(15.0,0.26)
161
1.00
(0.0, 0.00)
173
N
0.27
(16.0,0.24)
171
0.33
(19.0,0.25)
1033
0.27
(26.0, 0.26)
168
0.29
(31.0,0.29)
1020
0.34
(17.0,0.22)
166
0.32
(20.0, 0.23)
168
0.33
(19.0,0.24)
166
0.36
(17.0,0.25)
168
0.39
(16.0,0.21)
174
0.48
(22.0,0.21)
1018
0.49
(14.0,0.19)
164
0.33
(19.0,0.31)
161
0.24
(21.0,0.35)
165
1.00
(0.0, 0.00)
1051
0
0.75
(15.0,0.23)
166
0.77
(20.0, 0.26)
166
0.61
(31.0,0.35)
163
0.76
(31.0,0.33)
164
0.73
(18.0,0.25)
163
0.80
(20.0, 0.27)
163
0.70
(22.0, 0.30)
163
0.76
(14.0,0.22)
157
0.68
(21.0,0.30)
169
0.22
(27.0, 0.33)
164
0.65
(21.0,0.31)
162
0.78
(10.0,0.20)
158
0.84
(8.0,0.16)
161
0.31
(19.0,0.29)
161
1.00
(0.0, 0.00)
169
December 2009 A-170
-------
0.8
0.6
« «
* * •
*»* » » »
o
o
0.4
0.2
10 20 30
40 50 60
Distance Between Samplers (km)
70 80 90 100
Figure A-93. PM10 inter-sampler correlations as a function of distance between monitors for
Chicago, IL
December 2009
A-171
-------
Denver Combined Statistical Area
01
r
\ Denver CSA
• PMio Monitors
Interstate Highways
Major Highways
0 10 20 40 60 80
100
m Kilometers
Figure A-94. PM™ monitor distribution and major highways, Denver, CO.
December 2009
A-172
-------
AQS Site ID
Site A 08-001-0006
SiteB 08-001-3001
SiteC 08-013-0012
SiteD 08-031-0002
SiteE 08-031-0017
SiteF 08-123-0006
1=winter
2=spring
3=summer
4=fall
ABC
Mean 36.0 28.2 19.8
Obs 1043 1074 169
SD 18.3 13.2 9.7
90 -
80 -
70 -
•s,
; '60 -
T\
)l
: so -
>
i 40-
J
J
: 30 -
5
J
20 -
r
er 0 -
D E F
24.2 25.8 22.2
1039 1028 353
10.6 11.5 11.2
1234 1234 1234 1234 1234 1234
Figure A-95. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations
for Denver, CO.
December 2009
A-173
-------
Table A-39. Inter-sampler correlation statistics for each pair of PMio monitors reporting to AQS for
Denver, CO.
Site
1.00
0.84
0.43
0.70
0.72
0.67
(0.0,0.00)
(20.0,0.16)
(36.0,0.34)
(29.0, 0.24)
(26.0,0.21)
(27.0, 0.28)
1043
1022
164
987
980
339
1.00
0.57
0.72
0.74
0.72
(0.0, 0.00)
(28.0, 0.27)
(17.0,0.18)
(15.0,0.16)
(18.0,0.22)
1074
169
1019
1007
348
1.00
0.75
0.72
0.51
(0.0, 0.00)
(17.0,0.23)
(16.0,0.23)
(16.0,0.23)
LEGEND
D R
(P90, COD)
N
169
169
156
1.00
0.89
(0.0,0.00)
(9.0,0.13)
1039
976
164
0.52
(17.0,0.22)
341
1.00
0.58
(0.0, 0.00)
(17.0,0.23)
1028
330
1.00
(0.0, 0.00)
353
December 2009
A-174
-------
0.8
0.6
o
o
0.4
0.2
10 20 30 40 50 60
Distance Between Samplers (km)
70 80 90 100
Figure A-96. PM10 inter-sampler correlations as a function of distance between monitors for
Denver, CO.
December 2009
A-175
-------
Detroit Combined Statistical Area
A
01
Detroit CSA
• PMio Monitors
Interstate Highways
— Major Highways
0 10 20 40 60
80
100
m Kilometers
Figure A-97. PM™ monitor distribution and major highways, Detroit, Ml.
December 2009
A-176
-------
Site A 26-163-0001
SiteB 26-163-0015
SiteC 26-163-0033
itelD
0001 Mean
0015 Obs
0033 SD
80 -
70 -
60 -
m
<: 50 -
cn
0 40 -
Ł 3°-
u
c
0
u 20 -
10 -
1 =winter
2=spring
3=summer 0 -
4=fall
1
ABC
22.5 26.4 32.0
174 176 1057
11.8 14.9 17.9
234 1234 1234
Figure A-98. Box plots illustrating the seasonal distribution of 24-h avg PMi0 concentrations
for Detroit, Ml.
December 2009
A-177
-------
Table A-40. Inter-sampler correlation statistics for each pair of PMio monitors reporting to AQS for
Detroit, Ml.
Site
1.00
0.77
(0.0, 0.00)
174
169
1.00
0.74
(14.0,0.18) (28.0,0.26)
172
0.79
LEGEND (0.0,0.00) (21.0,0.21)
R 176 174
(P90,COD) 1-00
N (0.0,0.00)
1057
0.8
0.6
o
o
0.4
0.2
10 20 30 40 50 60 70
Distance Between Samplers (km)
80
90
100
Figure A-99. PM10 inter-sampler correlations as a function of distance between monitors for
Detroit, Ml.
December 2009
A-178
-------
Houston Combined Statistical Area
01
r
J Houston CSA
• PMio Monitors
Interstate Highways
Major Highways
0 10 20 40 60 80
100
—i Kilometers
Figure A-100. PM™ monitor distribution and major highways, Houston, TX.
December 2009
A-179
-------
AQS Site ID
Site A 48-201-0024
SiteB 48-201-0047
SiteC 48-201-0062
SiteD 48-201-0066
SiteE 48-201-0071
SiteF 48-201-1035
SiteG 48-201-1039
01
c
o
-4— »
fD
C
o>
u
c
o
1=winter
2=spring
3=summer
4=fall
A
Mean 24.5
Obs 174
SD 9.7
170 -
160 -
150 -
140 -
130 -
120 -
110 -
100 -
90 -
80 -
70 -
60 -
50 -
40 -
30 -
20 -
10 -
0 -
I 1
r
i
BCD
24.0 23.4 22.3
178 174 175
9.1 9.7 10.0
i
li
''"
I
nil
it
I
E F G
22.7 54.8 15.8
174 359 163
9.3 35.5 7.7
I I
j flf
I
--•''
1234 1234 1234 1234 1234 1234 1234
Figure A-101. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations
for Houston, TX.
December 2009
A-180
-------
Table A-41. Inter-sampler correlation statistics for each pair of PMio monitors reporting to AQS for
Houston, TX.
SITE ABC
A 1.00 0.84 0.78
(0.0,0.00) (9.0,0.12) (11.0,0.16)
174 163 158
B 1.00 0.86
(0.0,0.00) (9.0,0.11)
178 156
C 1.00
(0.0, 0.00)
174
D
E
LEGEND
R
F (P90, COD)
N
G
D
0.76
(12.0,0.16)
165
0.86
(9.0,0.12)
160
0.83
(10.0,0.14)
156
1.00
(0.0, 0.00)
175
E
0.43
(15.0,0.20)
167
0.38
(15.0,0.19)
163
0.41
(17.0,0.19)
159
0.32
(18.0,0.20)
163
1.00
(0.0, 0.00)
174
F
0.56
(77.0, 0.37)
159
0.52
(74.0, 0.39)
158
0.38
(74.0, 0.40)
151
0.43
(81.0,0.43)
155
0.15
(78.0, 0.43)
158
1.00
(0.0, 0.00)
359
G
0.75
(17.0,0.28)
156
0.79
(16.0,0.26)
152
0.85
(14.5,0.25)
150
0.76
(16.0,0.23)
154
0.38
(20.0, 0.28)
157
0.37
(92.0, 0.54)
149
1.00
(0.0, 0.00)
163
December 2009 A-181
-------
Houston
•• • . *
0.8
0.6
0.4
0.2
0 10 20 30
40 50 60 70
Distance Between Samplers (km)
80 90 100
Figure A-102. PMio inter-sampler correlations as a function of distance between monitors for
Houston, TX.
December 2009
A-182
-------
Los Angeles Core Based Statistical Area
01
r
Los Angeles CBSA
• PMio Monitors
Interstate Highways
Major Highways
0 10 20 40 60 80
100
—i Kilometers
Figure A-103. PM™ monitor distribution and major highways, Los Angeles, CA.
December 2009
A-183
-------
AQSSitelD
Site A 06-037-0002
SiteB 06-037-1103
SiteC 06-037-4002
SiteD 06-037-6012
Site E 06-037-9033
Site F 06-059-0007
SiteG 06-059-2022
1=winter
2=spring
3=summer
4=fall
A B C D E F G
/lean 35.3 31.1 31.5 27.3 23.7 33.5 21.6
Obs 169 175 178 176 985 175 162
SD 19.8 13.3 19.6 18.1 12.1 37.6 9.4
90 -
80 -
70-
60-
50 -
40-
30-
20 -
1 0 -
o -
1234 1234 1234 1234 1234 12
34 1234
Figure A-104. Box plots illustrating the seasonal distribution of 24-h avg PMi0 concentrations
for Los Angeles, CA.
December 2009
A-184
-------
Table A-42. Inter-sampler correlation statistics for each pair of PMio monitors reporting to AQS for
Los Angeles, CA.
Site A B C
A 1.00 0.73 0.44
(0.0,0.00) (17.0,0.17) (27.0,0.24)
169 153 154
B 1.00 0.61
(0.0,0.00) (14.0,0.14)
175 159
C 1.00
LEGEND (O.o.o.oo)
Pearson R ^
(1 BU.tUUU)
E
F
G
D
0.73
(24.0, 0.22)
157
0.57
(21.0,0.24)
159
0.65
(27.0, 0.28)
158
1.00
(0.0, 0.00)
176
E
0.47
(28.0, 0.26)
169
0.52
(23.0, 0.23)
173
0.43
(22.0, 0.24)
176
0.70
(16.0,0.20)
175
1.00
(0.0, 0.00)
985
F
0.41
(29.0, 0.24)
155
0.42
(15.0,0.16)
162
0.93
(11.0,0.11)
159
0.65
(26.0, 0.28)
161
0.29
(26.0, 0.25)
173
1.00
(0.0, 0.00)
175
G
0.65
(30.0, 0.28)
143
0.73
(20.0, 0.23)
149
0.73
(21.0,0.22)
148
0.57
(19.5,0.24)
150
0.38
(20.0, 0.24)
159
0.65
(21.5,0.22)
150
1.00
(0.0, 0.00)
162
0.8
0.6
0.4
0.2
10 20 30 40 50 60 70
Distance Between Samplers (km)
80
90
100
Figure A-105. PMio inter-sampler correlations as a function of distance between monitors for
Los Angeles, CA.
December 2009
A-185
-------
New York Combined Statistical Area
A
01
r
New York CSA
• PMio Monitors
Interstate Highways
— Major Highways
0 10 20 40 60 80 100
•Z«Z^^^H^Z^Z^^^B^Z^ZI Kilometers
Figure A-106. PM™ monitor distribution and major highways, New York, NY.
December 2009
A-186
-------
60-
50-
40-
I 30 H
Ł
"c
CD
O
O 20 H
O
10-
0-
Figure A-107. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations
for New York, NY.
December 2009
A-187
-------
Table A-43. Inter-sampler correlation statistics for each pair of PMio monitors reporting to AQS for
New York, NY.
Site
A
B
C
A
1.00
(0.0, 0.00)
167
LEGEND
R
(P90, COD)
N
B
0.88
(11.0,0.20)
156
1.00
(0.0, 0.00)
169
C
0.82
(12.0,0.16)
164
0.74
(18.0,0.25)
166
1.00
(0.0, 0.00)
178
1 t
0.8
0.6
o
1
HI
0.4
0.2
10 20 30 40 50 60
Distance Between Samplers (km)
70
80
90
100
Figure A-108. PM10 inter-sampler correlations as a function of distance between monitors for
New York, NY.
December 2009
A-188
-------
Philadelphia Combined Statistical Area
A
01
Philadelphia CSA
• PMio Monitors
Interstate Highways
— Major Highways
0 10 20 40 60 80
100
—i Kilometers
Figure A-109. PM™ monitor distribution and major highways, Philadelphia, PA.
December 2009
A-189
-------
AQS Site ID
Site A 10-003-2004
SiteB 42-017-0012
SiteC 42-045-0002
SiteD 42-091-0013
2=spring
3=surr
4=fail
A B C D
Mean 22.8 17.1 19.9 17.6
Obs 1059 1040 1059 1049
SD 11.7 9.3 9.4 9.8
60 -
50 -
"Ł 40 -
01
c
.2 30 -
ro
4-"
C
01
Ł 20-
O
u
10 -
er
ig
mer -
1234 1234 12341234
Figure A-110. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations
for Philadelphia, PA.
December 2009
A-190
-------
Table A-44. Inter-sampler correlation statistics for each pair of PMio monitors reporting to AQS for
Philadelphia, PA.
Site
A
B
C
D
A
1.00
(0.0, 0.00)
1059
LEGEND
R
(P90, COD)
N
B
0.81
(13.0,0.21)
1005
1.00
(0.0, 0.00)
1040
C
0.64
(14.0,0.19)
1025
0.71
(11.0,0.20)
1006
1.00
(0.0, 0.00)
1059
D
0.84
(12.0,0.20)
1013
0.93
(6.0,0.12)
994
0.73
(11.0,0.19)
1014
1.00
(0.0, 0.00)
1049
0.8
0.6
01
8
0.4
0.2
10 20 30 40 50 60
Distance Between Samplers (km)
70
80
90
100
Figure A-111. PMio inter-sampler correlations as a function of distance between monitors for
Philadelphia, PA.
December 2009
A-191
-------
Phoenix Core Based Statistical Area
A
01
\ | Phoenix CBSA
• PMio Monitors
Interstate Highways
— Major Highways
0 1020 40 60 80 100
i Kilometers
Figure A-112. PM™ monitor distribution and major highways, Phoenix, AZ.
December 2009
A-192
-------
AQS Site ID
Site A 04-013-0019
SiteB 04-013-1003
SiteC 04-013-1004
SiteD 04-013-3002
SiteE 04-013-3003
SiteF 04-013-3010
SiteG 04-013-4003
SiteH 04-013-4004
2=spring
3=sum
4=fall
AQS Site ID
04-013-4006
04-013-4009
04-013-4010
04-013-4011
SiteM 04-013-8006
SiteN 04-013-9812
SiteO 04-013-9997
SiteP 04-021-0001
Site I
SiteJ
SiteK
SiteL
1= winter
2=spring
3=sum
4=fall
ID ABCDEFGH
19 Mean 486 30.9 32.6 40.8 325 51.5 56.6 34.7
D3 Obs 790 179 182 1084 182 780 336 181
34 SD 23.0 14.5 14.6 20.0 15.2 23.1 25.8 17.0
32 26°"
33 240 -
10
D3 22° '
34 200 -
Ł- 180 -
E:
^ 140 -
O
'ro 12°-
= 100 -
OJ
u
5 80 -
O
u
60 -
40 -
ter 20 -
ng
imer g -
III
I 1 l
1 1
1 i 1
1 1
| ,
1 1
'III !i 1
i
1
1
1 1 '
1234 1234 1234 234 1234 1234 1234 1234
,n JKLMNOP
' IU Mean 55.6 75.6 32.5 53.0 58.4 65.5 34.3 49.7
306 Obs 1073 1083 178 1090 174 1086 1067 407
309 SD 30.6 39.5 16.1 27.8 30.9 34.9 21.3 54.2
310 260-
511 240-
306
312 220-
397
301
^ .180 -
E
^ 160 -
c 140 -
O
ro 120 -
•1— '
v 100 -
u
c
O 80 -
u
60 -
40 -
ter
ng 20-
mer o-
Illl
Ill
1234 1234 1234 1234 1234 1234 1234 1234
December 2009
A-193
-------
AQS Site ID
SiteQ 04-021-3002
SiteR 04-021-3004
SiteS 04-021-3006
Site! 04-021-3007
SiteU 04-021-3008
SiteV 04-021-3011
SiteW 04-021-3012
SiteX 04-021-7004
1=winter
2=spring
3=sum
4=fall
ID
32 Mean
™ °soS
06 260-
07
08 240 -
11
12 22° "
04 200 -
ST 180 -
E:
Ol
a.
^ 140 -
O
c 100 -
cu
u
c 80 -
O
60 -
40 -
ter 20 -
ng
mer 0 -
QRSTUVWX
20.8 38.8 13.7 27.0 80.6 74.7 21.5 54.3
172 171 172 175 476 475 170 322
10.0 21.4 6.8 17.1 72.5 137.5 12.5 38.3
!'='
:••-••
I
III
ill'
1234 1234 1234 1234 1234 1234 1234 1234
Figure A-113. Box plots illustrating the seasonal distribution of 24-h avg PMi0 concentrations
for Phoenix, AZ.
December 2009
A-194
-------
Table A -45. Inter-sampler correlation statistics for each pair of PMio monitors reporting to AQS for
Phoenix, AZ.
Site AB C DEFGHI
A 1.00 0.71 0.85 0.85 0.67 0.94 0.86 0.77 0.73
(0.0,0.00) (38.0,0.25) (33.0,0.21) (21.0,0.12) (38.0,0.23) (14.0,0.09) (22.0,0.13) (34.0,0.21) (35.0,0.18)
790 178 181 788 181 779 335 180 772
B 1.00 0.84 0.82 0.85 0.67 0.74 0.81 0.67
(0.0,0.00) (13.0,0.12) (23.0,0.19) (11.0,0.11) (37.0,0.29) (47.0,0.30) (13.0,0.13) (49.0,0.30)
179 179 177 179 175 179 178 175
C 1.00 0.88 0.81 0.78 0.80 0.81 0.70
(0.0,0.00) (20.0,0.16) (12.0,0.11) (38.0,0.27) (44.0,0.28) (13.0,0.13) (48.0,0.29)
182 180 182 178 182 181 178
D 1.00 0.76 0.88 0.81 0.82 0.76
(0.0,0.00) (23.0,0.17) (22.0,0.14) (29.0,0.16) (18.0,0.17) (39.0,0.20)
1084 180 778 334 179 1062
E 1.00 0.64 0.68 0.74 0.66
(0.0,0.00) (40.0,0.27) (47.0,0.29) (16.0,0.14) (48.0,0.29)
182 178 182 181 178
F 1.00 0.83 0.76 0.75
(0.0,0.00) (22.0,0.13) (36.0,0.25) (32.0,0.17)
780 331 177 762
G 1.00 0.77 0.65
(0.0,0.00) (44.0,0.26) (38.0,0.19)
336 181 326
H 1.00 0.79
(0.0, 0.00) (47.0, 0.26)
R 181 m
i (P90COD) 100
(0.0, 0.00)
1073
J
K
L
M
J
0.83
(59.0, 0.24)
781
0.68
(84.0, 0.43)
176
0.73
(84.0,0.41)
179
0.78
(71.0,0.31)
1072
0.59
(88.0, 0.42)
179
0.86
(54.0,0.21)
772
0.78
(48.0,0.19)
333
0.81
(79.0, 0.39)
178
0.79
(52.0, 0.22)
1061
1.00
(0.0, 0.00)
1083
K
0.77
(34.0, 0.24)
177
0.75
(16.0,0.15)
175
0.81
(15.0,0.14)
178
0.79
(22.0,0.19)
176
0.67
(15.0,0.15)
178
0.74
(41.0,0.28)
175
0.71
(46.0, 0.30)
178
0.79
(16.0,0.14)
177
0.76
(48.0, 0.29)
174
0.78
(83.0, 0.42)
175
1.00
(0.0, 0.00)
178
L
0.70
(30.0,0.17)
789
0.60
(51.0,0.31)
178
0.63
(49.0, 0.29)
181
0.65
(35.0, 0.20)
1080
0.51
(49.0, 0.30)
181
0.69
(30.0,0.17)
779
0.65
(36.0,0.19)
335
0.69
(43.0, 0.27)
180
0.69
(33.0,0.17)
1068
0.73
(57.0, 0.23)
1078
0.72
(45.0, 0.29)
177
1.00
(0.0, 0.00)
1090
M
0.87
(28.5,0.16)
170
0.63
(56.0, 0.32)
164
0.75
(55.0, 0.30)
167
0.83
(42.0,0.21)
172
0.61
(58.0,0.31)
167
0.87
(25.0,0.15)
167
0.80
(33.0,0.16)
169
0.72
(53.0, 0.29)
167
0.68
(38.0, 0.20)
171
0.80
(51.0,0.22)
171
0.68
(56.0, 0.32)
164
0.63
(42.0, 0.20)
173
1.00
(0.0, 0.00)
174
December 2009 A-195
-------
A
B
C
D
E
H
G
H
1
J
K
L
M
N
0
P
Q
R
S
T
U
V
W
X
N
0.87
(39.0,0.18)
784
0.59
(67.0, 0.37)
178
0.70
(69.0, 0.35)
181
0.78
(57.0, 0.25)
1075
0.60
(67.0, 0.35)
181
0.91
(35.0, 0.14)
774
0.77
(35.0,0.16)
332
0.70
(66.0, 0.33)
180
0.76
(42.0,0.18)
1064
0.91
(29.0,0.12)
1074
0.69
(73.0, 0.36)
177
0.68
(48.0, 0.20)
1081
0.86
(32.0,0.16)
173
1.00
(0.0, 0.00)
1086
0
0.68
(28.0,0.17)
783
0.75
(15.0,0.15)
179
0.87
(11.0,0.12)
182
0.86
(15.0,0.12)
1056
0.73
(14.0,0.14)
182
0.68
(31.0,0.21)
773
0.57
(41.0,0.25)
336
0.75
(15.0,0.14)
181
0.61
(49.0, 0.27)
1045
0.58
(83.0, 0.38)
1055
0.71
(16.0,0.16)
178
0.55
(44.0, 0.26)
1063
0.81
(53.0, 0.29)
174
0.58
(66.0, 0.32)
1059
1.00
(0.0, 0.00)
1067
P
0.47
(29.0,0.19)
406
0.75
(22.0,0.17)
175
0.80
(19.0,0.15)
178
0.73
(30.0,0.19)
405
0.68
(21.0,0.17)
178
0.46
(30.0, 0.22)
403
0.47
(36.5, 0.24)
330
0.82
(18.0,0.15)
177
0.52
(39.0, 0.22)
397
0.41
(68.0,0.31)
404
0.75
(19.0,0.18)
174
0.51
(37.0, 0.22)
406
0.75
(47.0, 0.30)
165
0.41
(51.0,0.27)
403
0.90
(35.0, 0.22)
407
1.00
(0.0, 0.00)
407
Q
0.53
(49.0, 0.42)
171
0.73
(23.0, 0.27)
169
0.70
(24.0, 0.28)
172
0.63
(38.0, 0.38)
170
0.72
(21.0,0.28)
172
0.48
(60.0, 0.46)
169
0.55
(61.0,0.47)
172
0.63
(29.0,0.31)
171
0.57
(66.0, 0.47)
169
0.48
(103.0,0.58)
169
0.52
(28.0, 0.29)
168
0.47
(66.0, 0.47)
171
0.48
(74.0, 0.48)
157
0.48
(88.0, 0.53)
171
0.61
(28.0,0.31)
172
0.67
(32.0, 0.29)
169
1.00
(0.0, 0.00)
172
LEGEND
R
(P90, COD)
N
R
0.68
(34.0, 0.27)
171
0.63
(30.0, 0.25)
168
0.71
(26.0, 0.24)
171
0.68
(27.0, 0.25)
169
0.64
(27.0, 0.24)
171
0.63
(37.0, 0.30)
167
0.65
(41.0,0.30)
171
0.74
(24.5, 0.22)
170
0.71
(41.0,0.27)
168
0.65
(75.0, 0.40)
168
0.64
(27.0, 0.23)
167
0.57
(44.5, 0.29)
170
0.64
(51.0,0.32)
158
0.67
(62.5, 0.35)
170
0.64
(25.0, 0.24)
171
0.81
(22.0,0.19)
170
0.72
(40.0, 0.33)
162
1.00
(0.0, 0.00)
171
S
0.40
(64.0, 0.57)
171
0.55
(32.0, 0.43)
169
0.48
(36.0, 0.44)
172
0.49
(46.0, 0.53)
170
0.43
(33.0, 0.44)
172
0.38
(68.0, 0.60)
168
0.46
(73.0,0.61)
172
0.55
(37.0, 0.46)
171
0.51
(77.0, 0.60)
168
0.48
(115.0,0.69)
169
0.52
(34.0, 0.44)
168
0.48
(71.0,0.60)
171
0.37
(80.0,0.61)
158
0.42
(98.0, 0.65)
171
0.39
(38.0, 0.47)
172
0.58
(44.0, 0.45)
169
0.65
(15.0,0.28)
163
0.66
(55.0, 0.48)
162
1.00
(0.0, 0.00)
172
T
0.69
(40.0, 0.34)
174
0.59
(21.0,0.24)
172
0.64
(22.0, 0.24)
175
0.65
(31.0,0.31)
173
0.48
(21.0,0.25)
175
0.63
(45.0, 0.39)
172
0.62
(58.0,0.41)
175
0.62
(24.0, 0.25)
174
0.58
(60.0, 0.40)
171
0.65
(92.0,0.51)
172
0.62
(22.0, 0.24)
171
0.49
(62.0, 0.40)
174
0.62
(58.5,0.41)
160
0.63
(75.0, 0.46)
174
0.60
(22.0, 0.26)
175
0.78
(21.0,0.21)
172
0.57
(23.0, 0.24)
167
0.68
(32.0, 0.27)
165
0.60
(28.0, 0.35)
167
1.00
(0.0, 0.00)
175
U
0.50
(82.0,0.31)
475
0.53
(94.0,0.41)
172
0.56
(91.0,0.40)
175
0.66
(87.0, 0.34)
474
0.42
(93.0,0.41)
175
0.47
(80.0,0.31)
470
0.49
(78.0, 0.28)
329
0.60
(84.0, 0.38)
174
0.59
(72.0, 0.27)
461
0.51
(69.0, 0.26)
473
0.71
(89.0, 0.40)
171
0.59
(75.0, 0.27)
475
0.46
(62.0,0.31)
165
0.42
(71.0,0.29)
470
0.72
(94.0, 0.39)
475
0.82
(80.0, 0.30)
400
0.36
(104.0,0.53)
165
0.53
(75.0, 0.35)
164
0.46
(115.0,0.65)
165
0.56
(94.0, 0.47)
169
1.00
(0.0, 0.00)
476
V
0.27
(49.0, 0.27)
474
0.66
(62.0, 0.34)
177
0.71
(59.0, 0.32)
180
0.45
(59.0, 0.30)
473
0.69
(63.0, 0.32)
180
0.28
(50.0, 0.27)
469
0.44
(45.0, 0.24)
334
0.76
(58.0, 0.29)
179
0.37
(46.0, 0.23)
461
0.28
(59.0, 0.27)
472
0.68
(59.0, 0.33)
176
0.33
(53.0, 0.24)
474
0.65
(48.0, 0.26)
168
0.26
(55.0, 0.27)
469
0.59
(69.0, 0.35)
473
0.64
(52.0, 0.23)
404
0.58
(78.0, 0.46)
171
0.82
(47.0, 0.25)
171
0.59
(86.0, 0.59)
171
0.66
(71.0,0.39)
174
0.54
(66.0, 0.24)
464
1.00
(0.0, 0.00)
475
W
0.56
(48.0, 0.43)
169
0.65
(24.0, 0.30)
167
0.62
(28.0,0.31)
170
0.58
(38.0, 0.39)
168
0.51
(25.0, 0.32)
170
0.42
(57.0, 0.47)
166
0.57
(59.0, 0.48)
170
0.64
(30.0, 0.33)
169
0.51
(63.0, 0.47)
167
0.46
(101.0,0.58)
167
0.55
(28.0, 0.32)
166
0.50
(67.0, 0.48)
169
0.44
(68.0, 0.49)
156
0.40
(88.0, 0.54)
169
0.55
(29.0, 0.33)
170
0.71
(32.0,0.31)
167
0.68
(15.0,0.22)
161
0.68
(40.0, 0.34)
160
0.72
(19.0,0.28)
162
0.68
(18.0,0.24)
165
0.52
(101.0,0.53)
165
0.60
(78.0, 0.47)
169
1.00
(0.0, 0.00)
170
X
0.65
(31.0,0.20)
262
0.64
(46.0, 0.29)
155
0.60
(43.0, 0.28)
157
0.70
(32.0,0.21)
318
0.52
(46.0, 0.28)
157
0.66
(34.0, 0.22)
259
0.64
(32.0, 0.22)
185
0.76
(39.0, 0.25)
156
0.80
(30.0,0.16)
314
0.74
(62.0, 0.27)
319
0.68
(44.0, 0.29)
153
0.68
(29.0,0.18)
321
0.59
(42.0, 0.24)
145
0.60
(48.0, 0.24)
319
0.64
(44.0, 0.26)
317
0.67
(39.0, 0.24)
197
0.47
(62.0, 0.43)
148
0.68
(39.0, 0.24)
148
0.52
(74.0, 0.58)
149
0.61
(51.5,0.37)
150
0.71
(61.0,0.25)
204
0.64
(35.0, 0.20)
206
0.56
(63.0, 0.44)
145
1.00
(0.0, 0.00)
322
December 2009
A-196
-------
0.8
• * •••
• ++
v~ ** **r /;*•.* * * ^
** **; '*: **<**/*«* *******
* K* * ** * » 4 * * ^
* **»•»*•»« ^*.***» «.**«^
» » » * **»% *^ ***** ~ **
0.6
***** *
•s
oi
5
o
0.4
0.2
*
*
*
10 20 30 40 50 60
Distance Between Samplers (km)
70
80
90
100
Figure A-114. PM10 inter-sampler correlations as a function of distance between monitors for
Phoenix, AZ.
December 2009
A-197
-------
A
01
Pittsburgh Combined Statistical Area
Pittsburgh CSA
PMio Monitors
Interstate Highways
Major Highways
0 10 20 40 60
80
100
m Kilometers
Figure A-115. PM™ monitor distribution and major highways, Pittsburgh, PA.
December 2009
A-198
-------
Obs
SD
90-
AQSStelD
SiteA 42-003-0002 80-
SiteB 42-003«)21
SiteC 42-003-0031 70-
SiteD 42-003-0064 j-
SiteE 42-003-0092 ^ 60"
Site F 42-003-0095 2
SiteG 42-003-0116 "^ 50"
SieH 42-003-1301 -
03 40-
§ 30-
c
o
u 20-
10-
1=winter
2=spring 0 "
3=summer
Wall _10.
1077 1019 1083 1087
12.9 11.4 12.3 20.3
f
176 179 978 182 1022
11. 9.9 11.7 16.8 19.3
' f
1234 12341234 1234 1234 1234 12341234 1234
Stel 42-003-3006
SteJ 42-003-3007
Ste K 42-003-7004
Site L 42-007-0014
SiteM 42-073-0015
SteN 42-125-0005
SiteO 42-125-5001
SiKP 42-129-0007
SiteQ 42-12*0008
J K L M N O P Q
Wtan 20.1 36.5 26.3 26.4 21.7 19.7 27.3 21.1
Obs 177 1061 1051 1069 1092 167 178 1079
SD
UO-
ItelD no.
006
007 10°"
004 „ 90-
1014 "c
1015 Tj, 80-
005 ""
c 70-
001 o
1007 U 60-
KX8 Z
Ł 50-
0 40-
u
30-
20-
1 -winter
2-spring 1 0 -
3~summer
4=fall 0 -
10.3 26.7 16.3 15.0 12.4 11.0 12.0 11.4
)
I
\
1234 1234 1234 1234 1234 1234 1234 1234
Figure A-116. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations
for Pittsburgh, PA.
December 2009
A-199
-------
Table A-46. Inter-sampler correlation statistics for each pair of PMio monitors reporting to AQS for
Pittsburgh, PA.
Site
A
t
R
C
D
E
E
G
H
I
A
1 nn
inn oooi in
1077
in
LEGEND
Pearson R
(P90, COD)
n
B C
Hen Hen
0 01S) (80 014)
1009 106S
1 00 0 96
0 000) (80 015)
1019 1007
1 00
(00 000)
1083
D
nan
(930 091)
1070
080
(990 094)
1019
081
(930 090)
1075
100
(000 00)
1087
E
na9
(80 019)
175
091
(11 0 090)
163
094
(60 011)
173
079
(91 0 0 90)
176
100
(000 00)
176
F
n 80
(140 018)
178
099
(60 016)
166
093
(70 019)
176
066
(96 0 0 94)
179
090
(100 014)
173
1 00
(00 000)
179
G
noT
(800 14)
960
097
(50 010)
911
094
(80 013)
966
076
(970 094)
970
090
(100 017)
154
094
(70 019)
157
1 00
(000 00)
978
H
n?Q
(160 017)
181
081
(95 0 0 99)
169
077
(91 0 0 99)
179
083
(140 018)
189
078
(90 0 0 90)
175
070
(95 0 0 97)
178
070
(99 0 0 98)
160
1 00
(00 000)
189
1
n RR
(180 018)
1005
089
(99 0 0 90)
954
087
(190 017)
1010
088
(160 014)
1014
077
(900 019)
166
074
(95 0 0 99)
168
087
(900 019)
910
076
(170 090)
171
1 00
in n n nm
1099
A
R
0
D
E
F
G
H
I
.1
K
L
M
N
0
P
Q
J
flM
(14 n n 9m
176
093
(7 n mm
164
nnn
(80 0131
174
n?3
(94 n n 99)
177
086
MOO 016)
171
am
(70 0191
174
099
(70 0131
156
074
(93 0 0 96)
176
079
(99 0 0 90)
166
1 00
(000 00)
177
K
H7fi
(400 030)
1044
076
(43 0 0 36)
986
075
(390 030)
1049
084
(94 0 0 99)
1055
065
(360 099)
169
057
(41 0 0 34)
179
073
(450 035)
955
068
(96 0 0 99)
175
083
(300 095)
999
066
(44 5 0 33)
170
1 00
(00 0 00)
1061
LEGEND
Pearson F
(P90, COD
n
L
nss
(150 018)
1033
088
(190 093)
989
088
(140 017)
1039
080
(900 018)
1043
083
(160 016)
169
089
(900 090)
179
087
(180 091)
938
077
(150 018)
175
089
(160 017)
978
079
(180 090)
170
074
(310 096)
1017
100
(00 0 00)
1051
t
I)
M
ns5
(160 019)
1059
081
(90 0 0 96)
994
083
(150 019)
1057
078
(90 0 0 90)
1061
080
(140 017)
179
075
(190 099)
175
078
(190 094)
959
078
(170 018)
178
078
(180 090)
998
079
(180 099)
173
075
(33 0 0 94)
1035
087
(130 016)
1095
100
(00 0 00)
1069
N
nsfi
(11 0 016)
1074
091
(100 016)
1016
089
(90 019)
1080
076
(95 0 0 90)
1084
084
(190 014)
176
086
(11 0 014)
179
089
(90 015)
975
074
(91 0 0 99)
189
081
(900 017)
1019
088
(80 013)
177
070
(40 0 0 30)
1058
085
(160 017)
1048
074
(180 091)
1067
100
(00 0 00)
1099
0
nn
(160 099)
166
076
(190 019)
157
078
(190 018)
164
057
(980 096)
167
077
(140 019)
161
083
(90 015)
164
081
(11 0 017)
146
060
(970 099)
167
066
(960 094)
158
078
(11 0 017)
163
047
(44 0 0 36)
160
070
(99 0 0 94)
160
064
(190 096)
163
0.79
(130 018)
167
100
(00 0 00)
167
P
H7R
(150 019)
177
083
(1800 98)
165
088
(130 019)
175
064
(90 0 0 95)
178
084
(130 016)
179
084
(160 099)
175
084
(170 096)
157
065
(190 099)
177
069
(91 0 0 95)
167
086
(160 091)
173
058
(34 0 0 30)
171
074
(170 091)
171
067
(170 099)
174
086
(140 090)
178
075
(180 095)
163
100
(00 0 00)
178
Q
nsfi
(11 0 015)
1061
088
(100 018)
1003
090
(90 019)
1067
074
(960 091)
1071
085
(11 0 015)
174
086
(90 014)
177
086
(100 016)
967
076
(91 5 0 94)
180
078
(990 019)
1009
086
(80 015)
175
068
(43 0 0 30)
1048
080
(180 019)
1035
077
(180 019)
1053
086
(100 014)
1076
069
(140 019)
165
084
(150 091)
176
100
(00 000)
1079
December 2009
A-200
-------
0.8
* •• * * V * *
* 4 » « *^ » * * *
* »' »*** * ** **• *»
\ * » V .«*«••*
» ^ » » ^ »*
-»^« *^
0.6
0.4 -
0.2
0 10 20 30
40 50 60 70
Distance Between Samplers (km)
90 100
Figure A-117. PM10 inter-sampler correlations as a function of distance between monitors for
Pittsburgh, PA.
December 2009
A-201
-------
Riverside Core Based Statistical Area
A
01
\ Riverside CBSA
• PMio Monitors
Interstate Highways
— Major Highways
0 15 30 60 90 120
150
—i Kilometers
Figure A-118. PM™ monitor distribution and major highways, Riverside, CA.
December 2009
A-202
-------
Site A
SiteB
SiteC
SiteD
SiteE
SiteF
AQS Site ID
06-065-0003
06-065-2002
06-065-5001
06-065-6001
06-065-8001
06-071-0013
A
Mean 37.6
Obs 174
SD 27.9
130 -"
120 -
B C D E F G
52.3 28.0 53.6 55.1 24.6 43.5
315 170 173 358 177 181
27.4 21.1 91.9 35.4 20.9 24.8
SiteG 06-071-0025 110 -
100 -
, 90 -
0> 80 -
c
o
1=winter
2=spring
3=summer
4=fall
AQS Site ID
SiteH 06-071-0306
Site I 06-071-1234
06-071-2002
06-071-4001
06-071-4003
70
60
50
40
30
20 -
10 -
SiteJ
SiteK
SiteL
20.6
1015
15.4
SiteM 06-071-9004
E:
01
c
QJ
U
1=winter
2=spring
3=summer
4=fall
1234 1234 1
H I
Mean 31.4
Obs 1060
SD 20.2
130 -
120 -
110 -
100 -
90 -
70 -
50 -
40 -
30 -
20 -
0 -
234 1
J
54.4
178
234 1
K
29.6
173
13.6
234
1234
L
36.5
178
20.2
1234
M
47.7
175
1234 1234 1234 1234 1234 1234
Figure A-119. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations
for Riverside, CA.
December 2009
A-203
-------
Table A-47. Inter-sampler correlation statistics for each pair of PMio
Riverside, CA.
A
B
C
D
E
F
G
H
I
J
K
L
M
A B C D E F G
1.00 0.09 0.15 0.90 0.94 0.25 0.94
(0.0,0.00) (50.0,0.31) (36.0,0.32) (33.0,0.19) (37.0,0.24) (41.0,0.38) (16.0,0.13)
174 170 155 165 172 169 171
1.00 0.86 0.07 0.13 0.31 0.12
(0.0,0.00) (48.0,0.37) (47.0,0.28) (45.0,0.27) (57.0,0.47) (49.0,0.26)
315 161 167 298 173 176
1.00 0.13 0.21 0.36 0.20
(0.0,0.00) (49.0,0.37) (58.0,0.42) (24.0,0.31) (40.0,0.35)
170 151 162 156 160
1.00 0.93 0.19 0.83
(0.0,0.00) (29.0,0.17) (52.0,0.43) (23.0,0.17)
173 169 167 168
1.00 0.23 0.93
(0.0,0.00) (63.0,0.48) (27.0,0.17)
358 174 179
1.00 0.27
(0.0,0.00) (44.0,0.41)
177 173
1.00
(0.0, 0.00)
181
LEGEND
R
(P90, COD)
N
H
0.24
(25.0, 0.22)
174
0.32
(48.0, 0.33)
309
0.34
(27.0, 0.28)
170
0.11
(38.0, 0.27)
173
0.26
(46.0, 0.33)
351
0.73
(28.0, 0.33)
177
0.27
(30.0, 0.25)
181
1.00
(0.0, 0.00)
1060
monitors reporting to AQS for
i
0.12
(40.0, 0.39)
173
0.29
(55.0, 0.49)
302
0.36
(24.0, 0.30)
168
0.05
(52.0, 0.46)
172
0.16
(63.5,0.51)
340
0.32
(27.0, 0.32)
176
0.20
(46.5, 0.45)
180
0.26
(27.0, 0.33)
983
1.00
(0.0, 0.00)
1015
J
0.83
(38.5, 0.24)
160
0.13
(51.0,0.25)
172
0.23
(57.5, 0.41)
150
0.73
(26.0,0.18)
157
0.86
(18.0,0.13)
175
0.35
(57.0, 0.46)
162
0.90
(25.0,0.16)
165
0.47
(45.0, 0.32)
178
0.20
(62.0,0.51)
177
1.00
(0.0, 0.00)
178
K
0.27
(30.0, 0.23)
158
0.31
(49.0, 0.35)
163
0.38
(24.0, 0.27)
147
0.13
(43.0, 0.30)
155
0.27
(54.0, 0.36)
165
0.43
(24.5, 0.32)
160
0.35
(34.0, 0.27)
163
0.48
(18.0,0.18)
172
0.45
(25.0, 0.32)
172
0.42
(49.0, 0.35)
155
1.00
(0.0,0.00)
173
0.46
(32.0
169
0.35
(51.0
173
0.50
(30.0
159
0.38
(40.0
165
0.57
(40.0
175
0.44
(35.0
170
0.58
(29.0
174
0.40
(29.0
178
0.38
(41.0
177
0.70
(37.0
163
0.49
(30.0
162
1.00
(0.0,
178
L
, 0.25)
,0.31)
, 0.25)
, 0.26)
, 0.28)
, 0.35)
, 0.24)
, 0.25)
, 0.39)
, 0.27)
, 0.26)
0.00)
M
0.78
(33.0,
164
0.29
(44.0,
168
0.40
(41 .0,
154
0.69
(24.5,
160
0.82
0.21)
0.24)
0.34)
0.16)
(26.0,0.15)
171
0.48
(46.0, 0.43)
164
0.85
(24.0,
168
0.44
(34.0,
175
0.35
(48.0,
173
0.85
(20.0,
157
0.48
(38.0,
157
0.84
(24.0,
167
1.00
0.15)
0.26)
0.46)
0.15)
0.29)
0.20)
(0.0, 0.00)
175
December 2009 A-204
-------
0.8
0.6
O
O
0.4
0.2
10 20 30
40 50 60
Distance Between Samplers (km)
70 80 90 100
Figure A-120. PM™ inter-sampler correlations as a function of distance between monitors for
Riverside, CA.
December 2009
A-205
-------
Seattle Combined Statistical Area
A
01
0 10 20 40
| | Seattle CSA
• PMio Monitors
Interstate Highways
- Major Highways
60 80 100
H^^^B^Z^ZI Kilometers
Figure A-121. PM™ monitor distribution and major highways, Seattle, WA.
December 2009
A-206
-------
Site A 53-033-0057
SiteB 53-033-2004
elD A B
057 Mean 21.9 15.7
004 Obs 1059 1077
SD 9.9 8.6
60 -
50 -
"g 40 -
c
° 30 -
ru
C
01
Ł 20 -
o
u
10 -
1=winter
2=spring
3=summer g _
Figure A-122. Box plots illustrating the seasonal distribution of 24-h avg PMi0 concentrations
for Seattle, WA.
Table A-48. Inter-sampler correlation statistics for each pair of PMio monitors reporting to AQS for
Seattle, WA.
A
B
A
1.00
(0.0, 0.00)
1059
LEGEND
R
(P90, COD)
B
0.77
(14.0,0.24)
1041
1.00
(0.0, 0.00)
1077
N
December 2009
A-207
-------
0.8
0.6
O
O
0.4
0.2
10 20 30 40 50 60
Distance Between Samplers (km)
70 80 90 100
Figure A-123. PM10 inter-sampler correlations as a function of distance between monitors for
Seattle, WA.
December 2009
A-208
-------
St. Louis Combined Statistical Area
A
01
SLLouis CSA
• PMio Monitors
Interstate Highways
Major Highways
0 10 20 40 60 80
100
zzi Kilometers
Figure A-124. PM™ monitor distribution and major highways, St. Louis, MO.
December 2009
A-209
-------
AQSSitelD
Site A 17-117-0002
SiteB 17-119-0010
SiteC 17-119-3007
SiteD 17-163-0010
SiteE 29-189-5001
SiteF 29-510-0085
SiteG 29-510-0086
SiteH 29-510-0087
Site I 29-510-0088
en
c
o
c
cu
u
c
o
u
1=winter
2=spring
3=summer
4=fall
A B C D E F G
Mean 22.8 35.4 28.3 34.0 20.5 28.2 22.7
Obs 171 173 174 176 185 176 179
SD 9.4 15.5 12.2 14.7 14.4 11.7 10.4
130 -
120 -
110 -
100 -
90 -
80 -
70 -
60 -
50 -
40 -
30 -
20 -
10 -
0 -
I I
.!l
'
"
H
29.5 40.3
180 1050
12.8 28.5
1234 1234 1234 1234 1234 1234 1234 1234 1234
Figure A-125. Box plots illustrating the seasonal distribution of 24-h avg PM10 concentrations
for St. Louis, MO.
December 2009
A-210
-------
Table A-49.
A
A 1.00
(0.0, 0.00)
171
B
C
D
E
F
G
H
I
Inter-sampler correlation
St. Louis, MO.
B C
0.50 0.75 0.67
(30.0,0.28) (14.0,0.17) (23.0
161 158 156
1.00 0.65 0.63
(0.0,0.00) (20.0,0.21) (20.0
173 161 158
1.00 0.75
(0.0,0.00) (17.0
174 157
1.00
(0.0,
176
(P90 COD)
N
statistics for each
D E
0.47
,0.24) (16.0,0.29)
158
0.46
,0.19) (37.0,0.42)
160
0.57
,0.17) (23.0,0.33)
158
0.44
0.00) (30.0, 0.40)
157
1.00
(0.0, 0.00)
185
pair of PMio
F
0.65
(16.0,0.18)
163
0.68
(23.0, 0.20)
167
0.80
(12.0,0.13)
165
0.82
(16.0,0.15)
163
0.53
(22.0, 0.34)
164
1.00
(0.0, 0.00)
176
monitors
G
0.67
(13.0,0.17)
166
0.68
(28.0, 0.28)
169
0.76
(13.0,0.18)
169
0.81
(21.0,0.24)
165
0.62
(17.0,0.26)
166
0.89
(11.0,0.16)
173
1.00
(0.0, 0.00)
179
reporting
H
0.73
(18.0,0.19)
168
0.64
(22.0, 0.20)
170
0.82
(12.0,0.13)
169
0.80
(14.0,0.15)
166
0.56
(25.0, 0.35)
167
0.86
(12.0,0.11)
174
0.83
(16.0,0.19)
177
1.00
(0.0, 0.00)
180
to AQS for
i
0.55
(52.0, 0.33)
164
0.52
(36.0, 0.28)
166
0.65
(41.0,0.27)
168
0.59
(36.0, 0.27)
169
0.34
(55.0, 0.42)
179
0.67
(41.0,0.27)
169
0.65
(47.0, 0.32)
173
0.64
(41.0,0.27)
173
1.00
(0.0, 0.00)
1050
December 2009 A-211
-------
0.8
0.6
o
13
oi
5
o
0.4
0.2
10 20 30 40 50 60 70
Distance Between Samplers (km)
80
90
100
Figure A-126. PM10 inter-sampler correlations as a function of distance between monitors for St.
Louis, MO.
Table A-50. Correlation coefficients of hourly and daily average particle number, surface and volume
concentrations in selected particle size ranges.
Size range
(nm)
3-10
10-30
30-50
50-100
100-500
500-800
10-100
10-800
Total number
Total surface
Total volume
Hourly averages
All days (N = 5481)
0.40
0.35
0.38
0.46
0.55
0.73
0.31
0.55
0.30
0.57
0.66
Sundays (N = 701)
0.24
0.22
0.42
0.56
0.65
0.75
0.28
0.65
0.24
0.63
0.69
Weekdays (N = 3227)
0.42
0.31
0.29
0.39
0.49
0.70
0.24
0.49
0.24
0.51
0.62
Event days (N = 577)
0.73
0.57
0.56
0.57
0.62
0.76
0.52
0.62
0.58
0.65
0.73
No events (N
0.37
0.33
0.36
0.45
0.55
0.72
0.29
0.55
0.28
0.56
0.65
Daily avg
= 4904) All days (N
0.32
0.27
0.36
0.46
0.55
0.71
0.24
0.55
0.20
0.57
0.67
= 263)
Source: Tuch et al. (2006)
December 2009
A-212
-------
A.2.3. Speciation
Atlanta, GA
Atlanta, GA
tOJS :f? atioti • Spnno
Atlanta, GA
FRW PM 2 S apsciatiftn • Sunimei
Atlanta, GA
1=1 OCM ^H Crustal
SANDWICH sulfate am
Atlanta. GA
=Z1 Sullate "^ Ntt
=) OCM ^
SflNOWlCH sulftte and nltralo rclud
Figure A-127. Seasonally averaged PM2.s speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in Atlanta,
GA.
December 2009
A-213
-------
Birmingham, AL FRMPII,25
(=1 Sutfate ^H Nitrate ^H E>
i 1 COM ^H Crustal
SANDWICH sutfate and iWratelndiMe ammonium and
Birmingham, AL
FRMPW2.S special Spring
Birmingham, AL
:ia!ion • Summer
SANDWICH stifateafid niraie inciude ammoniim srtd
i 1 OCM ^H Crustal
SANDWICH sutfate and nitrate Include ammonium and
Birmingham, AL
SANDWICH stjf.-,te .-,rvj mt-.re IT uae
Birmingham, AL
FRM PWiSsnee.aliort • W
SANDWICH sul&te »r\a rflrit* Include irnmonlum sna
Figure A-128. Seasonally averaged PM2.s speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in
Birmingham, AL.
December 2009
A-214
-------
Boston, MA/NH FRMPM25spsciaBon-<-Se9SO"Avg
I 1 Sulfate ^H Nitrate
i 1 OCM VHl Crustal
SANDWICH sulfale and nitrate
Boston, MA/NH
FRMPM25 specatlon - Spring
[=1 Sulfate ^M Nitrate ^H EC
i 1 OCM mmm Crustal
SANDWICH Sulfate and ntr;
Boston, MA/NH
FRM PMiSspecialton - J
SANDWICH sulfate ai
] Sulfate ^H Nitrate ^^ EC
] OCM ^H Crustal
inium and wsler
Boston, MA/NH FB"P"'-S"••"""-T-M
i i Sulfate I^B Nitrate •
i 1 OCM I^B Crustal
SANDWICH suifate and ntrate include
Boston, MA/NH
I I Sulfate ^^ Nitrate •
i 1 OCM ^H Crustal
SANDWICH sulfate and nitrate include ammonium and
Figure A-129. Seasonally averaged PM2.s speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in Boston,
MA.
December 2009
A-215
-------
Chicago, IL/IN
FRM PM2.5spetiaiion - <-Season A
Chicago, IL/IN
FR*IPM2.5!psciirtkin - Spring
IL/IN
- Sulfate ••• itiate
J OCM ••• CrusEal
SAPJDWICH iiirale anU rirate include amrronium and wsla
[ 3 Suifaie ••• itrae B^H EC
' 1 OCM ••• Crui al
SANDWICH sulfaie aril nlfate include atnrwtKum and
Chicago, IL/IN
FRM PM2.5 spenaion - T_Faii
Chicago, IL/IN
FRMPM25speciallon - Winter
i 1 OCM ••• Crustal
SANDWICH sJfate
1=1 Sulfate ••• Nitrate •"• EC
i 1 OCM ••• Crus al
SANDWICH sulfale anfl nitrate"IncludT
Figure A-130. Seasonally averaged PM2.s speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in Chicago,
IL
December 2009
A-216
-------
Denver, CO
i Sulfate ^™ Nitt
i - 1 OCM
SSNDWICH sul
Deliver. CO FHMPMJ 5 gpaclaticr -sprng
Denver, CO
^ Sulfate ^™ Nitra
D OCM MM Cms
./
/
SANDWICH sjlftte and nilrae include smrtioniiim
(5
i 1 OCM ^m Crustal
SANDWICH sulfae snii itlt'f.e inr luoe ^r
CO FRM PM2.S speciatio-- T_Fatl
Denver, CO
(5
~?
Figure A-131. Seasonally averaged PM2.s speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter, derived using the SANDWICH method in Denver,
CO.
December 2009
A-217
-------
it, M I
FRO! PM2 Sspeelation - 4-SeasBfiAvg
i. -Kte ^H Nitrate ••
CM ^« Crustal
H sullate and niraie mciuae ammonium and
Detroit, Ml
FRMPMJSspeciaiiwi • S
Detroit. M I
FRM PHlSwwtttan - Summer
1 Sulfate
J OCM
usu
1-"'1!. h iiC.^I'- vnli.:t^|Mi'iili.,i- HI, • if.rniir.1 AIO
1 Sulfate ^^ Nitrate
1 OCM ^m Crusta
Cft'WCH suitale ana ntraie ncTirie amnomum and
Detroit, Ml
1 Sulfate ^™ Nitrate
l 1 OCM ^H CrusUt
SANDWd>i suliate arm nitiate mrlnrff* sir^innwrn and Walw"
Detroit, Ml
I Sulfate ^^ Nitrate •• EC
1 OCM <^m CrusUI
DWCH suiraie ana mtrase nciuie annmoniji-n
Figure A-132. Seasonally averaged PM2.s speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in Detroit,
Ml.
December 2009
A-218
-------
HOUStOn, I A. FRM PM!-5speciatJor) - 4-SeasOR
: ._-,:;.!-M: •• '•! r:o: ^H EC
: OCM •• Crustal
:!..--M>/'
-------
Los Angeles, CA
I I Sutfate •• Nitrate ^
E 1 OCM ^H Crustal
SANDWICH suiraie and nrr/aie nclude
LOS
Los Angeles, CA
r 1 Sulfate •• Kitrate
i 1 OCM ^m Crustal
SANDWICH ajlfaie and ni
I I Sulfate •• I ii -?,r= ^B EC
i 1 OCM ^m Crustal
SANDWICH sullaie and nitrate Hi
Los Angeles, CA
Los Angeles, CA
I 1 Sulfate ^H Wrate ^B EC
I 1 OCM ^m Crustil
SAIJDWICH stilus arid rierat
I I Sulfate ^H N^rate ^^ EC
I 1 OCM ^M Crujtal
SANOWICK iulrale dlid rnuale include
Figure A-134. Seasonally averaged PM2.s speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in Los
Angeles, CA.
December 2009
A-220
-------
New York, NY/NJ/OT" «>!„».-«. -«—,»,
G
' ' OCM ^H Criwtal
SANDWICH sfclfale anil mlrsle m'uc
New York. NY/NJ/CT"«p»»»«=«•» -**»
New York, NY/NJ/CT™«"«
G
!=l OCM ^» Crustal
SANDWICH aJfateani] nirale incude anrmDnum ang w
= OCM ^H Crustal
m-IDWICH sulfate and n.liale nouae atr
-------
Philadelphia,
i 1 OCM IHI Cruttil
SANDWICH sulfaie and niliale hciude ai
wrium and water
Philadelphia, PA/NJ FRMP«.SM«M«,-SP«,
Philadelphia, PA/NJ FR""is
OCM i^H Crustal
Hatr am: r.-;i3ie hi-h.de a
nomum antjv&tet
Philadelphia, PA/NJ
Philadelphia, PA/NJ FRM«»»««
=i OCM ^^ Crustal
3AMDWICH sJlateandiMirateirxiudearwroriurn snd w
i 1 OCM ^M Crustal
SANDWICH sunaie and rnnie hciLdeancemum and water
Figure A-136. Seasonally averaged PM2.s speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in
Philadelphia, PA.
December 2009
A-222
-------
Phoenix, AZ
3 • i.-,r- H*i Mr MI-- I^B EC
•J OCM I^H Coital
Q Q
•:.-. '[.'••!: I - .l-a:e a vj Hr.li: in; j'Jo -.iti,-r nnjm
Phoenix, AZ
FRtf PM2.5 spscBtlcn • Spi
: S fate ^Hl Nitrate ^B EC
: OCM ^M Crystal
Q Q
Phoenix, AZ
---' .-.-..1 nr-.v.-- i-.-i i-r -i-riir»-,-iiini wtd water
MBS
C=I1 O
:>'l,i|),\ni; 11 r.iili.itr- -,rri riltir.- ]-iiijrt:- vTrinii i-- .-inr ,,,-j-.--
ti:f- "• 'iiti -,;:- ^^ EC
^^ CrysUI
Q
Phoenix, AZ
S fate W Nitrate I^M EC
OCM •• Crustsl
Q Q
' fulnre a-nJ nr-.^e ITI j-je iJtirmvujrn and \v3lei
Phoenix, AZ
Itate ^B> Nitrate ^^ EC
M ^^ Crystal
Q Q
v^i: 11 s.,if,Htf an.! mraie nciude anroortum .-1,1,7 >
Figure A-137. Seasonally averaged PM2.s speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in Phoenix,
AZ.
December 2009
A-223
-------
Pittsburgh, PA
Pittsburgh, PA
Pittsburgh, PA FRHPUi$wecai&> - Summer
- OCM ^H Crustal
iNDWICH Su1
rata ntujde arnrnortiim anflwatei
Pittsburgh, PA
Pittsburgh, PA
lude airmurijum and wiw
i 1 OCM
SANDWICH Su
Figure A-138. Seasonally averaged PM2.s speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in
Pittsburgh, PA.
December 2009
A-224
-------
Riverside, CA
I i Sulfate ^H Nitrate
i 1 OCM ^m Crustal
SANDWICH sullate and rntrale TIC'L
Riverside, CA
FRK PW2.Sspeca1mn - Spring
Riverside, CA
I i Sulfate ^K Nitrate "• EC
t 1 OCM ^^ Crusts!
SANDWICH sulfate and nlrale include ammonium aid water
I 1 Sulfate ••§ Nitrate ••• EC
i 1 OCM i^m Crustal
SANDWICH suKale and rilrste rc-lude
Riverside, CA rmfl»ji^Mn.-,J
Riverside, CA FRI
-------
Seattle, WA
FRM PM2 5speaati«m - 4-SeasanAvg
I 1 OCM ^Bl Crystal
SANDWICH sutfst? a-id -iviie r.ci j:ie anminjum and
Seattle, WA
J Sulfate ^^ l'J:!(.rc
: OCM ^M Crustal
Seattle, WA
FRKPM 2 5 specimen - Summei
>JvNr.'i.'V'l[-H = 'fit- i-,, rii'V- .:-,-! jrifismmopwm Ud w
1 Suifate ^^ Nitrate
1 OCM ^M Crustal
SANDWICH SUHateanOrJIrateiiflule ammonium and
Seattle, WA
FRM PUŁ5speciaiion - Wnti
i > Sulfats ^^f Nitrate
i 1 OCM ^^ Crustal
3AND\^ICHsuHa» ana pirate ntluaearmwrtum and water
Figure A-140. Seasonally averaged PM2.s speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in Seattle,
WA.
December 2009
A-226
-------
St. Louis, MO/IL
^ Sulfate
=1 OCM
SANDWICH sul
i Nitrate ^H EC
•ustal
V J\f N\
!^OX
St. Louis, MO/IL
St. Louis, MO/IL
'T tulfatc ^^M Nil'
1 OCM •• Crustal
jfale and nitrate include ammnnurn anij wattt
St. Louis, MO/IL
I I Sutfate ^B Mitral ^^ EC
[ ! OCM ^H Crus al
Sftr«W1CH Sul^lt an-: Mliile uu.ljUy r,rin:i.(:ij[[i
St. Louis, MO/IL FRMPW25
spetiabon - Winter
Figure A-141. Seasonally averaged PM2.s speciation data for 2005-2007 for a) annual, b) spring,
c) summer, d) fall and e) winter derived using the SANDWICH method in St. Louis,
MO.
December 2009
A-227
-------
Figure A-142. Seasonal patterns in PM2.s chemical composition from city-wide monthly average
values for Atlanta, GA, 2005-2007. The gray line represents the difference in OCM
calculated using material balance and blank corrected OCx1.4.
Figure A-143. Seasonal patterns in PM2.s chemical composition from city-wide monthly average
values for Birmingham, AL, 2005-2007. The gray line represents the difference in
OCM calculated using material balance and blank corrected OCx1.4.
December 2009
A-228
-------
E 20-1
a
1254567
Figure A-144. Seasonal patterns in PM2.s chemical composition from city-wide monthly average
values for Boston, MA, 2005-2007. The gray line represents the difference in OCM
calculated using material balance and blank corrected OCx1.4.
Figure A-145. Seasonal patterns in PM2.s chemical composition from city-wide monthly average
values for Chicago, IL, 2005-2007. The gray line represents the difference in OCM
calculated using material balance and blank corrected OCx1.4.
December 2009
A-229
-------
Figure A-146. Seasonal patterns in PM2.s chemical composition from city-wide monthly average
values for Denver, CO, 2005-2007. The gray line represents the difference in OCM
calculated using material balance and blank corrected OCx1.4.
Figure A-147. Seasonal patterns in PM2.s chemical composition from city-wide monthly average
values for Detroit, Ml, 2005-2007. The gray line represents the difference in OCM
calculated using material balance and blank corrected OC x 1.4.
December 2009
A-230
-------
Figure A-148. Seasonal patterns in PM2.s chemical composition from city-wide monthly average
values for Houston, TX, 2005-2007. The gray line represents the difference in
OCM calculated using material balance and blank corrected OCx1.4.
Figure A-149. Seasonal patterns in PM2.s chemical composition from city-wide monthly average
values for Los Angeles, CA, 2005-2007. The gray line represents the difference in
OCM calculated using material balance and blank corrected OCx1.4.
December 2009
A-231
-------
Figure A-150. Seasonal patterns in PM2.s chemical composition from city-wide monthly average
values for New York, NY, 2005-2007. The gray line represents the difference in
OCM calculated using material balance and blank corrected OCx1.4.
Figure A-151. Seasonal patterns in PM2.s chemical composition from city-wide monthly average
values for Philadelphia, PA, 2005-2007. The gray line represents the difference in
OCM calculated using material balance and blank corrected OCx1.4.
December 2009
A-232
-------
Figure A-152. Seasonal patterns in PM2.s chemical composition from city-wide monthly average
values for Phoenix, AZ, 2005-2007. The gray line represents the difference in OCM
calculated using material balance and blank corrected OCx1.4.
Figure A-153. Seasonal patterns in PM2.s chemical composition from city-wide monthly average
values for Pittsburgh, PA, 2005-2007. The gray line represents the difference in
OCM calculated using material balance and blank corrected OCx1.4.
December 2009
A-233
-------
Figure A-154. Seasonal patterns in PM2.s chemical composition from city-wide monthly average
values for Riverside, CA, 2005-2007. The gray line represents the difference in
OCM calculated using material balance and blank corrected OCx1.4.
"Ł 20-
a
Figure A-155. Seasonal patterns in PM2.s chemical composition from city-wide monthly average
values for Seattle, WA, 2005-2007. The gray line represents the difference in OCM
calculated using material balance and blank corrected OCx1.4.
December 2009
A-234
-------
1254567
Figure A-156. Seasonal patterns in PM2.s chemical composition from city-wide monthly average
values for St. Louis, MO, 2005-2007. The gray line represents the difference in
OCM calculated using material balance and blank corrected OCx1.4.
A.2.4. Diel Trends
in
c\i
77 n
58 -
38 -
19 -
Weekday (N = 5156)
77 -,
58 -
38 -
19 -
12 18 24
0
Weekend (N = 2086)
• Median
Mean
90'th&10'th
95'th & 5'th
1
12 18 24
Figure A-157. Diel plots generated from all available hourly FRM-like PM2.s data, stratified by
weekday (left) and weekend (right), in Atlanta, GA. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for
each hour.
December 2009
A-235
-------
Weekday(N =1975)
Weekend (N = 797)
c\
77 -,
58 -
38 -
19 -
0 -i
6 12 18 2<
77 -,
58 -
38
19
•Median
Mean
90'th & 10'th
•••••• 95'th & 5'th
12 18 24
Figure A-158. Diel plots generated from all available hourly FRM-like PM2.s data, stratified by
weekday (left) and weekend (right), in Chicago, IL. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for
each hour.
c\
77 -i
58 -
38 -
19 -
Weekday (N = 7032)
77 -i
58 -
38 -
19 -
Weekend(N = 2854)
•Median
Mean
90'th&10'th
•••••• 95'th & 5'th
1
12 18 24
1
12 18 24
Figure A-159. Diel plots generated from all available hourly FRM-like PM2.s data, stratified by
weekday (left) and weekend (right), in Houston, TX. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for
each hour.
December 2009
A-236
-------
Weekday(N = 3348)
up
cxi
77 i
58 -
38 -
19 -
77 -,
58 -
38
19
12 18 24
0
Weekend (N = 1364)
• Median
Mean
90'th & 10'th
•••••• 95'th & 5'th
12 18 24
Figure A-160. Diel plots generated from all available hourly FRM-like PM2.s data, stratified by
weekday (left) and weekend (right), in New York, NY. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for
each hour.
in
csi
77 T-
58 -
38 -j
19 -j
Weekday(N =981)
12 18 24
Weekend (N = 407)
77 ->
58 -
38 -j
19 -j
12 18 24
• Median
Mean
90'th & 10'th
••-••• 95'th & 5'th
Figure A-161. Diel plots generated from all available hourly FRM-like PM2.s data, stratified by
weekday (left) and weekend (right), in Pittsburgh, PA. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for
each hour.
December 2009
A-237
-------
Weekday(N = 5775)
un
CN
>
Q_
/ / -
58 -
38 -
19 -
0 -I
•-".." I'"- ----':;_• •.,.••;,"---
77 -j
58 -
38
19
12 18 24
0
Weekend(N = 2332)
• Median
Mean
90'th&10'th
•••••• 95'th & 5'th
12 18 24
Figure A-162. Diel plots generated from all available hourly FRM-like PM2.s data, stratified by
weekday (left) and weekend (right), in Seattle, WA. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for
each hour.
un
CN
77 n
58 -
38 -
19 -
Weekday(N =1727)
12 18
24
77 -j
58 -
38 -
19 -
0
1
Weekend (N = 692)
• Median
Mean
90'th&10'th
•••••• 95'th & 5'th
12 18
24
Figure A-163. Diel plots generated from all available hourly FRM-like PM2.s data, stratified by
weekday (left) and weekend (right), in St. Louis, MO. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for
each hour.
December 2009
A-238
-------
Weekday (N = 715)
218 -
145 -
73 -
12 18 24
Weekend (N = 293)
218 -
145 i
73 ^
1
12 18 24
•Median
Mean
90'th & 10'th
•••••• 95'th & 5'th
Figure A-164. Diel plot generated from all available hourly FRM/FEM PMi0 data, stratified by
weekday (left) and weekend (right), in Atlanta, GA. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for
each hour.
218 -
145 -
73 -
1
Weekday(N =1971)
12
18
24
218 -
145 -
73 -
Q
1
Weekend (N = 793)
• Median
Mean
90'th & 10'th
•••••• 95'th & 5'th
12
18
24
Figure A-165. Diel plot generated from all available hourly FRM/FEM PMi0 data, stratified by
weekday (left) and weekend (right), in Chicago, IL. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for
each hour.
December 2009
A-239
-------
Weekday (N = 1310)
291 i
218 -
145
73 -
Weekend (N = 529)
291 i
218 -
145
73 -
12 18 24
• Median
Mean
90'th & 10'th
95'th & 5'th
12 18 24
Figure A-166. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by
weekday (left) and weekend (right), in Denver, CO. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for
each hour.
291 n
218
145 -
73 -
Weekday (N = 756)
1
12 18
24
291 n
218
145 -
73 -
1
Weekend (N = 302)
•Median
Mean
90'th & 10'th
95'th & 5'th
12 18
24
Figure A-167. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by
weekday (left) and weekend (right), in Detroit, Ml. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for
each hour.
December 2009
A-240
-------
Weekday(N = 692)
Weekend (N = 277)
218 -
145 -
73 -
0 -I
218 -
145 -
;'**'-. ... 73 "
_ - — i^*1^ ^**-^ ' ^ i - -• • g ' ~"* w ' • •_
r"'""i ""i'"-'""r; i ••••"•"•• i o -1
• Median
Mean
90'th&10'th
•••••• 95'th & 5'th
12 18 24
12 18 24
Figure A-168. Diel plot generated from all available hourly FRM/FEM PMi0 data, stratified by
weekday (left) and weekend (right), in Los Angeles, CA. Included are the number
of monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for
each hour.
291 n
218 -
145 -
73 -
Weekday (N = 742)
12 18
24
291 -j
218 -
145 -
73 -
0
1
Weekend (N = 303)
12 18
24
• Median
Mean
90'th&10'th
-••••• 95'th & 5'th
Figure A-169. Diel plot generated from all available hourly FRM/FEM PMi0 data, stratified by
weekday (left) and weekend (right), in Philadelphia, PA. Included are the number
of monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for
each hour.
December 2009
A-241
-------
Weekday(N =1532)
Weekend (N = 618)
291 ->
218 -
145 -
73 -
291 ->
218 -
145 -
12
18 24
73 -
• Median
Mean
90'th&10'th
••-••• 95'th & 5'th
12
18 24
Figure A-170. Diel plot generated from all available hourly FRM/FEM PM10 data, stratified by
weekday (left) and weekend (right), in Phoenix, AZ. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for
each hour.
Weekday (N = 6385)
Weekend (N = 2600)
/ia i -
218 -
145 -
73 -
U \
/ja i -
218 -
145 -
•••••••••'„'-'',;••. ..... 73 -
*• " ^ x ^' • • • • '•_"
••"•""P '^r— 1 ^ o -i
F — - — i '" "i"" 1 1
• Median
Mean
90th&10'th
••-•-• 95'th & 5'th
12 18 24
12 18 24
Figure A-171. Diel plot generated from all available hourly FRM/FEM PMi0 data, stratified by
weekday (left) and weekend (right), in Pittsburgh, PA. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for
each hour.
December 2009
A-242
-------
Weekday (N = 1660)
291 ->
218 -
145
73 -
0 -P
Weekend (N = 673)
291 i
218
145
73 -
• Median
Mean
90'th&10'th
•••••- 95th & 5th
12 18 24
16 12 18
24
Figure A-172. Diel plot generated from all available hourly FRM/FEM PMi0 data, stratified by
weekday (left) and weekend (right), in Riverside, CA. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for
each hour.
291 -,
218 -
145 -
73 -
Weekday(N = 752)
1
Weekend(N = 309)
218 -i
145 -
73 -
12
18 24
0
1
12 18
24
• Median
Mean
90'th & 10th
•••••• 95th & 5th
Figure A-173. Diel plot generated from all available hourly FRM/FEM PMi0 data, stratified by
weekday (left) and weekend (right), in Seattle, WA. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for
each hour.
December 2009
A-243
-------
Weekday(N =1741)
Weekend (N = 706)
291 ->
218 -
145 -
73 - ---•
12 18 24
291 -I
218 -
145 -
73 -
0 -I
^
• Median
Mean
90'th&10'th
••-••• 95'th & 5'th
12 18 24
Figure A-174. Diel plot generated from all available hourly FRM/FEM PMi0 data, stratified by
weekday (left) and weekend (right), in St. Louis, MO. Included are the number of
monitor days (N) and the median, mean, 5th, 10th, 90th and 95th percentiles for
each hour.
December 2009
A-244
-------
A.2.5. Copollutant Measurements
Winter
Spring
PM10 (daity avg).
PM10-2.5 (daity avg).
SO2 (daiV avg)-
NO2 (daiV avg)-
CO (daily avg)-
O3 (daily max 8-hr)-
oo
o
oo
o o
O O
O
O O
O O
ooo
^ ^ Q°> ^ Ł <$> Q-i <$> & ^
Summer Fall
PM10 (daiV avg)-
PM10-2.5 (daiV avg).
SO2 (daity avg)-
NO2 (daity avg)-
CO (daily avg)-
O3 (daily max 8-hr)-
(
c
o o
)
o
o
<32>
O
o
o o
o o
ooo
Figure A-175. Correlations between 24-h PM2.s and co-located 24-h avg PMio, PMi0-2.5, S02, N02
and CO and daily maximum 8-h avg 03 for Atlanta, GA, stratified by season (2005-
2007). One point is included for each available monitor pair.
December 2009
A-245
-------
Winter
Spring
PM10 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
c
©
)
Summer
PM10 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)-
O3 (daily max 8-hr)-
GD
-------
Winter
Spring
PM10 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily av<
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)'
OO OO
oasis)
O(D>
QGS)
O
C2DOOO
OOOD
OO O
OQOO
PM10 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)'
NO2 (daily avg)<
CO (daily avg)<
O3 (daily max 8-hr)'
Summer
Fall
(
O C
O(
am
) O
>ODO O
CX3D>O
X) O
OQ)
(
(330
(D)
)<0> OO
O®
OO O
OO OS) O
Figure A-177. Correlations between 24-h PM2.s and co-located 24-h avg PMio, PMi0-2.5, S02, N02
and CO and daily maximum 8-h avg 03 for Boston, MA, stratified by season
(2005-2007). One point is included for each available monitor pair.
December 2009
A-247
-------
Winter
Spring
PM10 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
O
O CD
O Q88HD
OO
OOO
O
O
O (2XSE)
O
(D) O
00 O
OO
OOKDO O
PM10 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)'
NO2 (daily avg)<
CO (daily avg)<
O3 (daily max 8-hr)'
Summer
Fall
i
OTE
O
D OO
OO 000
OO
(J3GDO
O
(
O
OOODO
) O OO
(DO
OO
OHD ooo
Figure A-178. Correlations between 24-h PM2.s and co-located 24-h avg PMio, PMi0-2.5, S02, N02
and CO and daily maximum 8-h avg 03 for Chicago, IL, stratified by season (2005-
2007). One point is included for each available monitor pair.
December 2009
A-248
-------
Winter
Spring
PM10 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
o
am)
o
o
o
o
Summer
PM10 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)-
O3 (daily max 8-hr)-
(SD>
o
o
o
o
D
o
Fall
O
<0)GD
O
O
O
O
-------
Winter
Spring
PM10 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
o
oo
ooo
o
oo
Summer
PM10 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)-
O3 (daily max 8-hr)-
oo
G2D OO
OO
O O®
G08D
(
Fall
O OO
ooo o
GD
OOO
)O OCE»OD
OO
(M»
O O
OOO
0 GOBDO
Figure A-180. Correlations between 24-h PM2.s and co-located 24-h avg PMio, PMi0-2.5, S02, N02
and CO and daily maximum 8-h avg Osfor Detroit, Ml, stratified by season
(2005-2007). One point is included for each available monitor pair.
December 2009
A-250
-------
Winter
Spring
PM10 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
o
o
o
o
o
Summer
PM10 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)-
O3 (daily max 8-hr)-
o
o
o
o
o
I
c
I
Fall
o
D
)
o
D
O
O
O
O
O
Figure A-181. Correlations between 24-h PM2.s and co-located 24-h avg PMio, PMi0-2.5, S02, N02
and CO and daily maximum 8-h avg 03 for Houston, TX, stratified by season
(2005-2007). One point is included for each available monitor pair.
December 2009
A-251
-------
Winter
Spring
PM10 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily av<
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr>
GOOGE)
o am
OD O
vUUL)
ODD GO
1
c
ooooo
D O O
X03OODOO
mo a
Summer
Fall
PM10 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)-
O3 (daily max 8-hr)-
O OD O
GO)
GGBQO
SSEBE)
GDOCDCD
(
o oo a
OOQDO) O
Figure A-182. Correlations between 24-h PM2.s and co-located 24-h avg PMio, PMi0-2.5, S02, N02
and CO and daily maximum 8-h avg 03 for Los Angeles, CA, stratified by season
(2005-2007). One point is included for each available monitor pair.
December 2009
A-252
-------
Winter
Spring
PM10 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily av<
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr>
OO O
OO
000
OO
GfflBBO
OOGDO
OQD
CD O3D
Summer
Fall
PM10 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)-
O3 (daily max 8-hr)-
(5
GOO
)
030OQD)
O OdD
OO OOO
-------
Winter
Spring
PM10 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
<3SD
OB2)
OODO
(23Q>
OOSKD
Summer
PM10 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)-
O3 (daily max 8-hr)-
O (
(SOO
OD OODO
OOOO O
3 (O>
(3HQD
O
Fall
GO
QfloxjoQaju
(OS3D
OOOO
O ©OO
00)
CX3OD
OS3D
(33$) O
(OD» 02DO(D)
Figure A-184. Correlations between 24-h PM2.s and co-located 24-h avg PMio, PMi0-2.5, S02, N02
and CO and daily maximum 8-h avg Os for Philadelphia, PA, stratified by season
(2005-2007). One point is included for each available monitor pair.
December 2009
A-254
-------
Winter
Spring
PM10 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
OS)
<3EO
O
O
3D
OO
Summer
PM10 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)-
O3 (daily max 8-hr)-
O ®OO
O
O
OO
OO
GO
Fall
ooo
O OO
O
O
OO
ooo
OGD
O GO)
O
O
O
OO
Figure A-185. Correlations between 24-h PM2.s and co-located 24-h avg PMio, PMi0-2.5, S02, N02
and CO and daily maximum 8-h avg 03 for Phoenix, AZ, stratified by season
(2005-2007). One point is included for each available monitor pair.
December 2009
A-255
-------
Winter
Spring
PM10 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
o
O O (SEC
OGBOO O
GED
OO O
Summer
PM10 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)-
O3 (daily max 8-hr)-
o
c
O (SHB
OdE>GD
D (2BD
OO
-------
Winter
Spring
PM10 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
(
OOOD
OO
-------
Winter
Spring
PM10 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
o
02X30
O O O
O O
O
Summer
PM10 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)-
O3 (daily max 8-hr)-
o
OGD
O O
O O
O O
(GSQSD
O(
(
Fall
OOGD
) O
GO
OO
)QO)O
OOO
OO
(D)
O O
O OOGD
Figure A-188. Correlations between 24-h PM2.s and co-located 24-h avg PMio, PMi0-2.5, S02, N02
and CO and daily maximum 8-h avg 03 for St. Louis, MO, stratified by season
(2005-2007). One point is included for each available monitor pair.
December 2009
A-258
-------
Winter
Spring
P M2. 5 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
oo
o
o
o
Summer
P M2. 5 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)-
O3 (daily max 8-hr)-
o o
o
o
o
Fall
o o
o
o
o
o
o
o
Figure A-189. Correlations between 24-h PMi0 and co-located 24-h avg PM2.6, PMi0-2.5, S02, N02
and CO and daily maximum 8-h avg 03 for Atlanta, GA, stratified by season (2005-
2007). One point is included for each available monitor pair.
December 2009
A-259
-------
Winter
Spring
P M2. 5 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
©
o
o o
o
Summer
P M2. 5 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)-
O3 (daily max 8-hr)-
GD
O
oo
CD)
Fall
<32>
O
O O
OO O
(22)
O
OO
O (SO
Figure A-190. Correlations between 24-h PMi0 and co-located 24-h avg PM2.6, PMi0-2.5, S02, N02
and CO and daily maximum 8-h avg 03 for Birmingham, AL, stratified by season
(2005-2007). One point is included for each available monitor pair.
December 2009
A-260
-------
Winter
Spring
P M2. 5 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
OOODO
OO
O ODO
O OD
O O O
O
Summer
P M2. 5 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)-
O3 (daily max 8-hr)-
(
am
o o
©o o
oooo
) OO
GD
Fall
(
OCE)
OO
OOO O
OS)
OS)
o© o
(330
O
D (30
-------
Winter
Spring
P M2. 5 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
o
O Q88HD
O
O
O
Summer
P M2. 5 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)-
O3 (daily max 8-hr)-
OTE
O
O
O
O OO
Fall
O (2XSE)
O
O
O
o e
OOODO
o
o
o
o o o
Figure A-192. Correlations between 24-h PMi0 and co-located 24-h avg PM2.6, PMi0-2.5, S02, N02
and CO and daily maximum 8-h avg 03 for Chicago, IL, stratified by season (2005-
2007). One point is included for each available monitor pair.
December 2009
A-262
-------
Winter
Spring
P M2. 5 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
o
am)
o
o o
oo
oo
Summer
P M2. 5 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)-
O3 (daily max 8-hr)-
c
(SD>
c
o o
oo
o
) O
o
Fall
o o
©GD
C
OD
O O
©
O
-------
Winter
Spring
P M2. 5 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
o
oo
o
o
Summer
P M2. 5 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)-
O3 (daily max 8-hr)-
oo
o
o
o o
Fall
o oo
o
o
o o
oo
o
o
o o
Figure A-194. Correlations between 24-h PMi0 and co-located 24-h avg PM2.6, PMi0-2.5, S02, N02
and CO and daily maximum 8-h avg 03 for Detroit, Ml, stratified by season (2005-
2007). One point is included for each available monitor pair.
December 2009
A-264
-------
Winter
Spring
P M2. 5 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
o<
o
o o
O
Figure A-195. Correlations between 24-h PMi0 and co-located 24-h avg PM2.6, PMi0-2.5, S02, N02
and CO and daily maximum 8-h avg 03 for Houston, TX, stratified by season
(2005-2007). One point is included for each available monitor pair.
December 2009
A-265
-------
Winter
Spring
P M2. 5 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
CD(DK2)
o am
GD
OGDO
-------
Winter
Spring
P M2. 5 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
ooo
©
o
o
Summer
P M2. 5 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)-
O3 (daily max 8-hr)-
GOO
OO
O O
o
OO
o
Fall
000
OO
o o
o
o
o
OO
00
o o
o
GO
o
Figure A-197. Correlations between 24-h PMi0 and co-located 24-h avg PM2.6, PMi0-2.5, S02, N02
and CO and daily maximum 8-h avg Os for New York, NY, stratified by season
(2005-2007). One point is included for each available monitor pair.
December 2009
A-267
-------
Winter
Spring
P M2. 5 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
OB2)
O OS)
O02D
O O
Summer
P M2. 5 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)-
O3 (daily max 8-hr)-
(
(SOO
O OOO
O O
3DO
O 00
C
Fall
GO
OD OD
OOO
) OO
O O
00)
oooo
O
-------
Winter
Spring
P M2. 5 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
G3DO
<3»0>
O
OO O
O (3SSS)
O
(
(
O OO
C
O
ms)
008D O GOO
3SHED
P M2.5 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)'
NO2 (daily avg)<
CO (daily avg)<
O3 (daily max 8-hr)'
Summer
Fall
O
C
C
I
O GBOO
C
D
)
-------
Winter
Spring
P M2. 5 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
O O (SEC
GHOO O
O CD
OO OO
Summer
P M2. 5 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)-
O3 (daily max 8-hr)-
c
o
O (&8D
oss>o
• OOD O
ooo o
<3XS)
(
Fall
o
ozm
OO3DO O
O (JD O
O OO
GOOO O
oooo
Figure A-200. Correlations between 24-h PMi0 and co-located 24-h avg PM2.6, PMi0-2.5, S02, N02
and CO and daily maximum 8-h avg 03 for Pittsburgh, PA, stratified by season
(2005-2007). One point is included for each available monitor pair.
December 2009
A-270
-------
Winter
Spring
PM2.5 (daiV avg)-
PM10-2.5 (daiV avg)-
SO2 (daity avg)-
NO2 (daity avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
I
(OEDCSI
OO <00
D (D)
CE) O
O OO
i
C
c
I
OD OSD
DOO
> O ODO
O O
DODO)
P M2.5 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)'
NO2 (daily avg)<
CO (daily avg)<
O3 (daily max 8-hr)'
Summer
Fall
(
O
I
OGDO (D)
)© O
O O OO
OD O
OXODQD O
C
O
GDO
Figure A-201. Correlations between 24-h PMi0 and co-located 24-h avg PM2.6, PMi0-2.5, S02, N02
and CO and daily maximum 8-h avg 03 for Riverside, CA, stratified by season
(2005-2007). One point is included for each available monitor pair.
December 2009
A-271
-------
Winter
Spring
P M2. 5 (daily avg)-
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)'
O3 (daily max 8-hr)-
(
02X30
OO O
o
OGD
DOO
Summer
P M2. 5 (daily avg)'
PM10-2.5 (daily avg)-
SO2 (daily avg)-
NO2 (daily avg)-
CO (daily avg)-
O3 (daily max 8-hr)-
c
OGD
O O O
O O
OOO
OOD
Fall
OOGD
OO O
OO
O OO
00 OOO
OOO
GOD
OO
GDO
(BDO
Figure A-202. Correlations between 24-h PMi0 and co-located 24-h avg PM2.6, PMi0-2.5, S02, N02
and CO and daily maximum 8-h avg 03 for St. Louis, MO, stratified by season
(2005-2007). One point is included for each available monitor pair.
December 2009
A-272
-------
A.3. Source Apportionment
A.3.1. Type of Receptor Models
Table A-51. Different receptor models used in the Supersite source apportionment studies: chemical
mass balance.
Receptor Model
Description
Strengths and
Weaknesses
Effective Variance CMB42'1'1
(Note that all models based on
eq 1 or 2 are CMB equations.
The term CMB used here
reflects the historical solution in
which source profiles are
explicitly used as model input
and a single sample effective
variance solution is reported.)
CMB software is currently
distributed by EPA. The most
recent version is the CMB 8.2,
which is run in the Microsoft
Windows system.
Principle
Ambient chemical concentrations are expressed as the sum of products of species
abundances in source emissions and source contributions (Equations A-1 or A-2). These
equations are solved for the source contribution estimates when ambient concentrations
and source profiles are input. The single-sample effective variance least squares is the
most commonly used solution method because it incorporates uncertainties of ambient
concentrations and source profiles in the estimate of source contributions and their
uncertainties. This reduced to the tracer solution when it is assumed that there is one
unique species for each source. Choices of source profiles should avoid collinearity, which
occurs when chemical compositions of various source emissions are not sufficiently
different.121
r
iklmn
^
~> rj-\ C1
ijm ijklmn iklmn
for i = 1 to I
Strengths
Software available providing
a good user interface.
Provides quantitative
uncertainties on source
contribution estimates based
on input concentrations,
measurement uncertainties,
and collinearity of source
profiles.
Quantifies contributions from
source types with single
particle and organic
compound measurements.
Weaknesses
Data Needs
CMB requires source profiles, which are the mass fractions of particulate or gas species in
source emissions. The species and particle size fraction measured in source emissions
should match those in ambient samples to be apportioned. Several sampling and analysis
methods provide time-integrated speciation of PM2 5 and volatile organic compounds
(VOCs) for CMB. Source profiles are preferably obtained in the same geographical region
as the ambient samples, although using source profiles from different regions is commonly
practiced in the literature. The practitioner needs to decide the source profiles and species
being included in the model, on the basis of the conceptual model and model performance
measures.
Output
Effective variance CMB determines, if converged, source contributions to each sample in
terms of PM or VOC mass. CMB also generates various model performance measures,
including correlation R2, deviation X2, residue/ uncertainty ratio, and MPIN matrix that are
useful for refining the model inputs to obtain the best and most meaningful source
apportionment resolution.
Equation A-1 Completely compatible
source and receptor
measurements are not
commonly available.
Assumes all observed mass
is due to the sources
selected in advance, which
involves some subjectivity.
Chemically similar sources
may result in collinearity
without more specific
chemical markers.
Equation A-2
Typically does not apportion
secondary particle
constituents to sources. Must
be combined with profile
aging model to estimate
secondary PM.
12 Hidy and Friedlander (1972,156546)
121 Vtetson et al. (1997,1571211122(1984, 045693)
Source: Watson et al. (2008,1571281
December 2009
A-273
-------
Table A-52. Different receptor models used in the Supersites source apportionment studies: factor
analysis.
Receptor Model
Description
Strengths and
Weaknesses
PMF
PMFx(PMF2andPMF3)
software is available from
Dr. Pentti Paatero at the
University of Helsinki,
Finland. This software is a
Microsoft DOS application.
EPA distributes EPA PMF76
version 1.1 as a Microsoft
Windows application with
better user interface.
Principle
PMFx contains PMF2 and PMF3. PMF2 solves the CMB equations (i.e., Equations A-2 and A-
3) using an iterative minimization algorithm. Source profiles Fj and contribution Sjt are solved
simultaneously. The non-negativity constraint is implemented in the algorithm to decrease
the number of possible solutions (local minimums) in the PMF analyses, because both
source profile and contribution should not contain negative values. There is rotational
ambiguity in all two-way factor analyses (i.e.,Fjtand Sjt matrices may be rotated and still fit
the data). PMF2 allows using the FPEAK parameter to control the rotation. A positive FPEAK
value forces the program to search such solutions where there are many zeros and large
values but few intermediate values in the source matrix Fjt.Fkey can further bind individual
elements in FjiO zero On the basis of a similar algorithm, PMF3 solves a three-way problem.
Strengths
Software available.
Can handle missing or below-
detection-limit data.
Weights species
concentrations by their
analytical precisions.
Downweight outliers in the
robust mode.
PMFx and UNMIX estimate Fj and Sjt by minimizing:
0 or x2= 22 f»vf= 22 [(
Derives source profiles from
ambient measurements as
they would appear at the
receptor (does not require
Equation A-3 source measurements).
Where the weighing factor, oil, represents the magnitude of Eit, PMFx limits solutions of
Equation A-2 to non-negative Fj and Sjt.
Data Needs
A large number of ambient samples (usually much more than the number of factors in the
model) are required to produce a meaningful solution. Species commonly used in PMF are
also those in CMB. Weighting factors associated with each measurement need to be
assigned before analysis. The practitioner also needs to decide the number of factors,
FPEAK, and Fkey in the model.
Output
PMFx reports all the elements in Fs and Sj matrices (PMF2). It also calculates model
performance measures such as deviation X2 and standard deviation of each matrix element.
The practitioner needs to interpret the results linking them to source profiles and source
contributions.
Weaknesses
Requires large (>100) ambient
datasets.
Need to determine the number
of retaining factors.
Requires knowledge of source
profiles or existing profiles to
verify the representativeness
of calculated factor profiles
and uncertainties of factor
contributions.
Relies on many
parameters/initial conditions
adjustable to model input;
sensitive to the preset
parameters.
ME2125
ME2 code is available from
Dr. Pentti Paatero at the
University of Helsinki,
Finland as a Microsoft DOS
application.
Principle
The PMFx algorithm is derived from ME2. Unlike PMFx that is limited to questions in the
form of Equation A-1 or A-2, ME2 solves all models in which the data values are fitted by
sums of products of unknown (and known) factor elements. The first part of the algorithm
interprets instructions from the user and generates a table that specifies the model. The
second part solves the model using an iterative minimization approach. Additional constraints
could be programmed into the model to reduce the ambiguity in source apportionment.
These constraints may include known source profiles and/or contributions (e.g., contributions
are known to be zero in some cases).
Data Needs
Data needs are similar to those of PMFx but are more flexible. In theory, any measured or
unknown variables may be included in the model as long as they satisfy linear relationships.
The users need to specify the model structure, the input, and the output.
Output
ME2 calculates and reports all unknown variables in the model.
Strengths
Software available.
Can handle user-specified
models.
Possibility to include all
measured variables into the
model, such as speciated
concentration over different
time scales, size distributions,
meteorological variables, and
noise parameters.
Weaknesses
Require substantial training to
access the full feature of the
software and develop a model.
Generally requires large
ambient datasets.
Need to assume linear
relationships between all
variables.
Relies on many
parameters/initial conditions
adjustable to model input;
sensitive to the preset
parameters.
December 2009
A-274
-------
Receptor Model
Description
Strengths and
Weaknesses
UNMIX ^<12b
UNMIX code is available
from Dr. Ron Henry at the
University of Southern
California as an MatLab
application. A stand-alone
version (UNMIX version 6) is
also available from EPA.
Principle
UNMIX views each sample as a data point in a multidimensional space with each dimension
representing a measured species. UNMIX solves Equations A-2 and A-3 by using a principle
component analysis (PCA) approach to reduce the number of dimensions in the space to the
number of factors that produce the data, followed by an unique "edge detection" technique to
identify "edges" defined by the data points in the space of reduced dimension (e.g Figures 1
and 3). The number of factors is estimated by the NUMFACT algorithm in advance , which
reports the R and signal-to-noise (S/N) ratio associated with the first N principle components
(PCs) in the data matrix. The number of factors should coincide with the number of PCs with
S/N ratio >2. Once the data are plotted on the reduced space, an edge is actually a
hyperplan that signifies missing or small contribution from one or more factors. Therefore,
UNMIX searches all the edges and uses them to calculate the vertices of the simplex, which
are then converted back to source composition and contributions. Geometrical concepts of
self-modeling curve resolution are used to ensure that the results obey (to within error) non-
negativity constraints on source compositions and contributions.
Data Needs
A large number of ambient samples (usually much more than the number of factors in the
model) are required to achieve a meaningful solution. Species commonly used in UNMIX are
also those in CMB. The measurement precision is not required. The practitioner needs to
specify the number of factors on the basis of the NUMFACT results.
Output
UNMIX determines all the elements in the factor (Fs) and contribution (Sj) matrices. It also
calculates the uncertainty associated with the factor elements and model performance
measures including: (1) R2, (2) S/N ratio, and (3) strength.
Strengths
Software available with
graphical user interface.
Does not require source
measurements.
Provide graphical problem
diagnostic tools (e.g., species
scatter plot).
Provide evaluation tools (e.g.,
R2, S/N ratio).
Weaknesses
Requires large (>100) ambient
datasets.
Need to assume or
predetermine number of
retained factors.
Does not make explicit use of
errors or uncertainties in
ambient measurements.
Cannot use samples
containing missing data in any
species.
Limited to a maximum of 7 or
14 (UNMIXversion 6) factors.
Can report multiple or no
solutions.
Requires knowledge of
existing source profiles to
evaluate the solutions.
December 2009
A-275
-------
Receptor Model Description Strengths and
Weaknesses
PDRM97 Principle Strengths
PDRM was developed under PDRM estimates contributions from selected stationary sources for a receptor site using high Explicitly include
the Supersites Program and time-resolution measurements and meteorological data. In PDRM, Equation A-2 is modified meteorological information and
requires MatLab or to: stack configuration of
equivalent software to stationary sources into the
perform the calculation. ,, ,v, model.
4=^0^1 +Ł>
j J1 Equation A-4 Do not require source
measurements.
where ŁRU is interpreted as the emission rate of species i from stationary source j and (X/Q,t .
is the meteorological dispersion factor averaged over the time interval t. Equation A-4 is Uo n. n™ lnte/Pl'et tne
solved for ERij and (X/Q,t simultaneously by a nonlinear fit minimizing the objective function, relatlons between factors and
Commercial software (e.g.,
.••v'.™" f MatLab) available for
(oJ -c, performing nonlinear fit.
Equation A-5 Suitab|e for hjgh time-
Because the number of solutions for a product of unknowns is infinite, additional constraints resolutlon measurement.
are set up for (X/Q,t on the basis of the Gaussian plume model, thus: Weaknesses
», Can only handle stationary
IB\Q) sources but not area or mobile
11 sources.
1 —l 1Al—I 1iz~Tl Need to assume that only
2»vve»l-2,If/|""| 2' ,r, /j stationary sources are
- exp
1 ./+)l,!,, considered in the model
2\
contribute significantly for a
Equations A-6 & A-7 mteeasurement at the receP"»
Equations A-6 and A-7 limit the solution of Equation A-5 within the lower (LB) and upper (UB) Do not account for uncertainty
bound of those predicted by the Gaussian plume model using different parameterizations. jn (ne measurement
Data Needs Meteorological data may not
PDRM requires speciated measurements at a higher time-resolution than typical CMB or be alwaVs avallable or
PMF applications because of the fast-changing meteorological parameters. PDRM also accurate.
requires data for Equation A-7: transport speed (u), lateral and vertical dispersion parameters Gaussian plume model may
(oy and oz), and stack height (h). not be representative of the
0UtpUt actual atmospheric dispersion.
PDRM determines emission rates and contributions from each point source considered in Sensitive to the imposed
the model at the same time resolution as the measurement. constraints (UB and LB).
December 2009 A-276
-------
Receptor Model
Description
Strengths and
Weaknesses
PLSL
Principle
PLS examines the relationships between a set of predictor (independent) and response
(dependent) variables. It assumes that the predictor and response variables are controlled
by independent "latent variables" less in number than either the predictor or the response
variables. In recent applications,96 PM chemical composition and size distribution are used
as predictor (X) and response (Y) variables, respectively. Equation A- 2 is modified to:
Equation A-8
Equation A-9
where T and U are matrices of so-called "latent variables," and P and C are loading matrices.
If X and Y are correlated to some degree, T and U would show some similarity. Equations
A-8 and A-9 are solved by an iterative algorithm "NIPALS," which attempts to minimize E,D,
and the difference between T and U simultaneously. If T and U end up being close enough,
the X and Y variables can be explained by the same latent variables. These latent variables
may then be interpreted as source or source categories.
Data Needs
Typical applications of PLS require both chemical speciated and size-segregated
measurements. The practitioner needs to decide the number of latent variables on the basis
of the correlation of resulting T and U matrices.
Output
PLS calculates latent variables, which are common factors best explaining the predictor and
response variables, and the residues from fitting. Rx and Ry,
Strengths
Fit two types of measurements
(e.g., chemistry and size) with
common factors. Provide more
information to identify sources.
Analyze strongly collinear and
noisy dataset.
Do not require source
measurements.
Weaknesses
Requires large (>100) ambient
datasets.
Difficult to relate latent
variables to any physical
quantities.
Do not provide quantitative
source contribution estimates.
Need to decide the number of
latent variables.
Do not explicitly make use of
measurement uncertainties.
Can result in no solution.
R, = 1 - var(ŁVvar(Jf)
Equation A-10
Equation A-11
indicate the degree to which variables X and Y are explained by the latent variables.
Henry (1997. 0209411
Lewis etal. (2003, 088413)
Ogulei et al. (2006,1199751
Park etal. (2005,156844)
Paatero (1997. 087001)
Paatero et al. (2002,1568361
Paatero (1999.1568351
Henry (2003,1565401
Source: Watson et al. (2008,1571281
December 2009
A-277
-------
Table A-53. Different receptor models used in the Supersites source apportionment studies: tracer-
based methods.
Receptor Model
Description
Strengths and Weaknesses
The EF method may use a MLR
algorithm, which is available in
most statistical and spreadsheet
software
Principle
A tracer (or marker) for a particular source or source category is a species
enriched heavily in the source emission against other species and other sources.
Using EFs-, concentration of the ith pollutant at a receptor site at time t (i.e., Ci,t)
can be expressed as:
Strengths
No special software needed.
Indicate presence or absence of
particular emitters.
Provides evidence of secondary PM
formation and changes in source
impacts by changes in ambient
composition.
Equation A-12 Could use a large (>100) dataset or a
small (e.g., < 10) dataset.
where the enrichment factor EFi,pj is the ratio of emission rate of the pollutant of
interest (Fj) and tracer species (Fpj) from source j. Cpjil is the concentration of
tracer species for source j at time t, and Zi,t represents contributions from all other
sources (including the background level). The solution for eq 12 is situation-
dependent. EFjipj is usually unknown but may be estimated from source profiles,
edges of a two-way scatter plot or the ratio of Cit to Cpjl for a particular period
when it is believed that a single source is dominant. In cases where ZM is a
constant, EFiipj may be derived from MLR.
Data Needs
The minimum data needs include concentrations of all primary tracers at the
receptor site. Known EFs or background levels are helpful.
Output
The EF method determines contributions to species i from each source
considered in the model.
Weaknesses
Semiquantitative method, not specific
especially when the EFs are
unknown in advance.
Limited to sources with unique
markers.
Tracer species must be exclusively
from the sources or source
categories examined.
Provide very limited error estimates.
More useful for source/process
identification than for quantification.
NNLS
The MatLab Optimization Toolbox
provides a function "Isqnonneg"
for performing the NNLS
calculation.
Principle
NNLS also solves the EF equation (Equation A-12 or equivalent) with known
target species and tracer concentrations. Conventional MLR solutions to eq 12
may lead to negative EFs due to the uncertainty in measurements or colinearity in
source contributions. This is avoided in the NNLS approach since additional non-
negative constraints are built into the algorithm, i.e.:
Strengths
Implemented by many statistical
software packages.
Generate only non-negative EFs or
regression coefficients.
Do not require source
measurements.
Possible to include meteorological or
Equation A-13 °'ner (besides chemistry) data into
the model.
Utilizing orthogonal decomposition, a NNLS problem can be reduced to the more
familiar least-distance programming and solved by a set of iterative subroutines
developed and tested by Lawson and Hanson.131 In a more general sense,
NNLS linearly relates a response variable to a set of independent variables with
only non-negative coefficients.
Data Needs
When applied to EF or MLR problems, NNLS requires the concentration of target
(response) and tracer (independent) species.
Output
NNLS generates non-negative regression coefficients for an EF/MLR problem and
these coefficients can be related to the source contributions.
Weaknesses
Require a large (>100) set of ambient
measurements.
Semiquantitative method, not
specific.
Do not explicitly consider
measurement uncertainties.
Tracer species must be exclusively
from the sources or source
categories examined.
Non-negative constraints may not be
appropriate in some cases.
December 2009
A-278
-------
Receptor Model
Description
Strengths and Weaknesses
FAC
Principle
FAC provides a simple mean of estimating the SOA production rate using the
emission inventories of primary precursor VOCs. FAC is actually a source-
oriented modeling technique but it does not take into account all the atmospheric
processes. FAC is defined as the fraction of SOA that would result from the
reactions of a particular VOC:
[SQA] = ^ FACi x ip/OGi], x Fraction of VOC /reacted!
Equation A-14
where [VOCjo is the emission rate of VOC: and [SOA] is the formation rate of
SOA. Equation A-14 can be viewed as an extension of Equation -12 but
concentrations are replaced with emission rates and EFs are replaced with FACs.
FAC and the fraction of VOC reacted under typical ambient conditions have been
developed for a large number of hydrocarbons >C6 11. The most significant SOA
precursors are aromatic compounds (especially toluene, xylene, and
trimethylbenzenes) and terpenes. In most applications, these FACs are used
directly to estimate SOA.
Data Needs
FAC requires the VOC emission inventory in the region of interest. The knowledge
of 03 and radiation intensity is also helpful for slight modifications of the FACs.
Output
FAC method estimates the total production rate of SOA.
Strengths
Link SOA to primary VOC emissions
so that SOA can also be treated as
primary particles in the PM modeling.
Simple and inexpensive.
Weaknesses
Ignore the influence of aerosol
concentration and temperature-
dependent gas-particle partitioning on
SOA yield.
Limited by the accuracy of VOC
emission inventory.
Do not directly infer the contribution
of each source to ambient SOA
concentration.
Difficult to verify.
Grosjean and Seinfeld (1989, 0456431
Darns etal. (1970,156379)
Reimann and De Capital (2000, 0132691
Lawson and Hanson (1974,1566731
Wang and Hopke (1989,1571051
Source: Watson et al. (2008, 157128)
December 2009
A-279
-------
Table A-54. Different receptor models used in the Supersites source apportionment studies:
meteorology-based methods.
Receptor Model
Description
Strengths and
Weaknesses
CPF1
Principle
CPF estimates the probability that a given source contribution from a given wind direction will
exceed a predetermined threshold criterion (e.g., upper 25th percentile of the fractional contribution
from the source of interest). The calculation of CPF uses source contributions (i.e., 03 in Equation
A-2) determined for the receptor site and local wind direction data matching each of the source
contributions in time. These data are then segregated to several sectors according to wind direction
and the desired resolution (usually 36 sectors at a 10° resolution). Data with very low wind speed
(e.g., < 0.1 m/sec) are usually excluded from analysis because of the uncertain wind direction. CPF
is then determined by:
.,
CPRH) = -
"is
Equation A-15
where mA6 is the number of occurrences in the direction sector 9 -> 9 + A6 that exceeds the
specified threshold, and nA6 is the total number of wind occurrences in that sector. Because wind
direction is changing rapidly, high-time resolution measurements (e.g., minutes to hours) are
preferred for a CPF analysis. If the calculated source contributions represent long-term averages,
wind direction needs to be averaged over the same duration. In addition to source contribution,
CPF can be applied directly to pollutant concentration measurements at a receptor site.
Data Needs
CPF requires the time series of source contributions at a receptor site, which is usually determined
by CMB or factor analysis methods using speciated measurements at the site. CPF also requires
wind direction and wind speed data averaged over the same time resolution as the sampling
duration.
Output
CPF reports the probability of "high" contribution from a particular source or factor occurring within
each wind direction sector. The results are often presented in a wind rose plot.
Strengths
Infer the direction of
sources or factors
relative to the
receptor site.
Provide verification
for the source
identification made
by factor analysis
method.
Easy to implement.
Weaknesses
Criterion for the
threshold is
subjective.
Absolute source
contribution (or
fractional
contribution) may be
influenced by other
factors besides wind
direction (e.g., wind
speed, mixing
height).
Local and near-
surface wind
direction only has a
limited implication for
long-range transport.
Easy to be biased by
a small number of
wind occurrences in
a particular sector.
Work better for
stationary sources
than area or mobile
sources.
December 2009
A-280
-------
Receptor Model
Description
Strengths and
Weaknesses
NPR1
Principle
NPR calculates the expected (averaged) source contribution as a function of wind direction
following:
where Wi is the wind direction for the ith sample and Si is the contribution from a specific source to
that sample, determined from measurements at the receptor site. K is a weighting function called
the kernel estimator. There are many possible choices for K. Henry et al.136 recommend either
Gaussian or Epanechnikov functions. The most important decision in NPR is the choice of the
smoothing parameter A6. If A6 is too large, S(6) will be too smooth and meaningful peaks could be
lost. If it is too small, S(6) will have too many small, meaningless peaks. A6 needs to be chosen
according to the project-specific spatial distribution of sources. NPR also estimates the confidence
intervals of S(6) based on the asymptotic normal distribution of the kernel estimates, thus:
Strengths
Infer the direction of
sources or factors
relative to the
receptor site.
Provide verification
for the source
identification made
by factor analysis
method.
Require no
Equation A-16 assumption about
the function form of
the relationship
between wind
direction and source
contribution.
Data Needs
NPR requires the same data as the CPF method, including the time series of source/factor
contributions (or fractional contributions) at the receptor site and local wind direction data matching
the sampling duration in time.
Output
NPR reports the distribution of source contribution as a function of wind direction and the
confidence level associated with it.
Provide uncertainty
estimates.
Easy to implement.
Weaknesses
Choices for the
kernel estimator and
smoothing factor are
subjective.
Absolute source
contribution (or
Equation A-17 fractional
contribution) may be
influenced by other
factors besides wind
direction (e.g., wind
speed, mixing
height).
Local and near-
surface wind
direction only has a
limited implication for
long-range transport.
Easy to be biased by
a small number of
wind occurrences in
a particular sector.
Work better for
stationary sources
than area or mobile
sources.
December 2009
A-281
-------
Receptor Model
Description
Strengths and
Weaknesses
TSA1J"
ISA requires the calculation of
air parcel back trajectory, which
is often accomplished using the
HY-SPLIT model.115'139 HY-
SPLIT version 4.5 is available
at http://www.arl.noaa.gov/-
ready/hysplit4.html.
Principle
Similar to CPF, ISA clusters the measured pollutant concentration or calculated source contribution
according to the wind pattern. However, air parcel back trajectory, rather than local wind direction,
is used. A back trajectory traces the air parcel backward in time from a receptor. The initial height is
often between 200 and 1000 m above ground level where the wind direction could differ from the
surface wind direction substantively. For each sample i, ISA obtains one or more trajectories and
calculates their total residence time in the jth directional sector (Ty, i.e., the total number of 1-h
trajectory end points that fall into the sector). The pollutant concentration or source contribution in
the sample, Si, is then linearly apportioned into each directional sector according to T,J and
averaged over all samples to produce the directional dependent pollutant concentration/source
contribution for the period of interest:
Equation A-18
where N is the number of samples. Compared with CPF and NPR, ISA considers the entire air
mass history rather than just the wind direction at the receptor.
Data Needs
ISA requires the time series of pollutant concentration or source contribution at the receptor site,
and back trajectories initiated over the site during the sampling duration. Trajectory is usually
calculated once every hour so ISA is more suitable for analyzing measurements of >1-h resolution.
Output
ISA reports the avg pollutant concentration or source contribution as a function of wind direction
based on back trajectory calculations.
Strengths
Infer the direction of
sources or factors
relative to the
sampling site.
Provide verification
for the source
identification made
by factor analysis
method.
Account for air mass
transport over
hundreds to
thousands of
kilometers and on
the order of several
days.
Can represent plume
spread from vertical
wind shear at
different hours of day
by adjusting the
initial height of back
trajectories.
Weaknesses
Need to generate
and analyze the back
trajectory data.
Uncertainty in back
trajectory calculation
increases with its
length in time.
Source contribution
depends on not only
trajectory residence
time but also
entrainment
efficiency, dispersion,
and deposition.
Difficult to resolve
the direction of more
localized sources.
December 2009
A-282
-------
Receptor Model
Description
Strengths and
Weaknesses
PSCF ™
PSCF requires the calculation
of air parcel back trajectory,
which is often accomplished
using the HY-SPLIT
model.115'139 HY-SPLIT version
4.5 is available at
http://www.arl.noaa.gov/-
ready/hysplit4.html.
Principle
Ensemble air parcel trajectory analysis refers to the statistical analysis on a group of trajectories to
retrieve useful patterns regarding the spatial distribution of sources. Uncertainties associated with
individual trajectory calculations largely cancel out for a sufficient number of trajectories or
trajectory segments. As a popular ensemble back trajectory analysis, PSCF estimates the
probability that an upwind area contributes to high pollutant concentration or source contribution.
Back trajectories are first calculated for each sample at the receptor site. To determine the PSCF, a
study domain containing the receptor site is divided into an array of grid cells. Trajectory residence
time (the time it spends) in each grid cell is calculated for all back trajectories and for a subset of
trajectories corresponding to "high" pollutant concentration or source contribution at the site. PSCF
in cell (i,j) is then defined as:
PSCF,, =
Sum of "high" residence time in cell (/, ft
Sum of all residence time in cell it/i
Equation A-19
The criterion for high pollutant concentration or source contribution is critical for the PSCF
calculation. The 75th or 90th percentileofthe concentration or factor is often used. ' '
Residence time can be represented by the number of trajectory end points in a cell.
Data Needs
Similar to TSA, PSCF calculation requires the time series of pollutant concentration or source
contribution at the receptor site, and back trajectories initiated over the site during the sampling
period. Trajectories should be calculated with 1 -to 3-h segment to reduce the uncertainty from
interpolation (if needed).
Output
PSCF reports the probability that an upwind area contributes to high pollutant concentrations or
source contribution at the downwind receptor site. The results are often presented as a contour plot
on the map. A high probability usually suggests potential source region.
Strengths
Infer the location of
sources or factors
relative to the
sampling site.
Provide verification
for the source
identification made
by factor analysis
method
Account for air mass
transport over
hundreds to
thousands of
kilometers and on
the order of several
days.
Resolve the spatial
distribution of source
strength
(qualitatively).
Weaknesses
Need to generate
and analyze the back
trajectory data.
Need to correct for
the central tendency
(residence time
always increases
toward the receptor
site regardless of
source contribution).
Uncertainty in back
trajectory calculation
increases with its
length in time.
Source contribution
depends on not only
trajectory residence
time but also
entrainment
efficiency, dispersion,
and deposition.
Difficult to resolve
the location of more
localized sources.
December 2009
A-283
-------
Receptor Model
Description
Strengths and
Weaknesses
SQTBA 117 143
SQTBA requires the calculation
of air parcel back trajectory,
which is often accomplished
using the HY-SPLIT
model.115'139 HY-SPLIT version
4.5 is available at
http://www.arl.noaa.gov/-
ready/hysplM.html.
Principle
SQTBA is another type of ensemble air parcel trajectory analysis. The concept of SQTBA is to
estimate the "transport field" for each trajectory ignoring the effects of chemical reactions and
deposition. Back trajectories are first calculated for each sample at the receptor site, and a study
domain containing the receptor site is divided into an array of grid cells. SQTBA assumes that the
transition probability that an air parcel at (x',y',f), where x' and y' are spatial coordinates and t'
means time, will reach a receptor site at (x,y,t) is approximately normally distributed along the
trajectory with a standard deviation that increases linearly with time upwind ' ,thus:
Strengths
Imply the location of
sources or factors
relative to the
sampling site.
Account for air mass
transport over
hundreds to
thousands of
kilometers and on
Equation A-20
where (X,Y) is the coordinate of the grid center, a is the dispersion speed, and x'(t') and x' (t')
represent the trajectory. The probability field, Q, for a given trajectory is then integrated over the
upwind period, T, to produce a two-dimensional "natural" (nonweighted) transport field:
,,.._
4 (Jf. y \x ', y > =
f°
Qi.x, y, t\x', y', z'}
Equation A-21
After the transport field for each trajectory is established, they are weighted by the corresponding
pollutant concentration or source contribution at the receptor site and summed to yield the overall
""Needs'
Resolve the spatial
distribution of source
strength
(qualitatively).
Weaknesses
and analyze the back
trajectory data.
Need to correct for
the central tendency
(res idence time
always increases
toward the receptor
site regardless of
source contribution).
Need to estimate
dispersion velocity.
inuoiue comnlicated
involve complicated
unclear.
SQTBA requires the time series of pollutant concentration or source contribution at the receptor
site, and back trajectories initiated over the site during the sampling period. Trajectories should be n-ffi ,, , ,
calculated with 1to 3-h segment to reduce the uncertainty from interpolation (if needed). the locationof rrore
Output localized sources.
SQTBA put more weight on trajectories associated higher pollutant concentration or source
contribution and therefore the resulting field may imply the major transport path.
December 2009
A-284
-------
Receptor Model
Description
Strengths and
Weaknesses
RTWC 146
RTWC requires the calculation
of air parcel back trajectory,
which is often accomplished
using the HY-SPLIT
model.115'139 HY-SPLIT version
4.5 is available at
http://www.arl.noaa.gov/
ready/hys plit4.html
Principle Strengths
As an ensemble air parcel trajectory analysis, RTWC requires back trajectories calculated for each Imply the location of
sample at the receptor site, and a study domain containing the receptor site divided into an array of sources or factors
grid cells. RTWC assumes that no major pollutant sources are located along "clean" (associated relative to the
with low pollutant concentrations) trajectories and that "polluted" trajectories picked up emissions sampling site.
along their paths. In practice, RTWC distributes pollutant concentrations at the receptor to upwind
grid cells along the back trajectories according to the trajectory residence times in those cells.117'14
resident time In cell i
'"'average residence time in each cell
where Sk is the pollutant concentration or source contribution determined upon the arrival of
trajectory k and Siik is the redistributed pollutant concentration or source contribution for cell i
upwind.
RTWC is known for the problem of "tailing effect," i.e., spurious source areas can be identified
when cells are crossed by a very small number of trajectories. Although some corrections were
proposed! 47 these approaches are purely empirical.
Account for air mass
transport over
hundreds to
thousands of
kilometers and on
the order of several
Equation A-22 days.
Resolve the spatial
distribution of source
strength
(qualitatively).
Weaknesses
Need to generate
and analyze the back
trajectory data.
Need to correct for
the central tendency
and tailing effect.
The amount of
emission
entrainment should
not be proportional to
the residence time of
trajectories (so there
is no linear
relationship between
RTWC field and
source strength).
Physical meaning of
the RTWC field is
unclear.
Difficult to resolve
the location of more
localized sources.
(Pekneyetal., 2006,086115)
' (Zhou etal., 2004,157190)
'(Ashbaugh, 1983,156229)
' (Ashbaugh et al., 1984, 0451481
'(Henry etal, 2002,136097)
' (Yu etal. ,2004, 101779)
3(Parekh and Husain, 1981,156840)
](Hopke etal, 1995,156566)
3(Keeler and Samson, 1989,156633)
'(Sam son.1978.188974)
5 (Samson. 1980. 0730101
3 (Stohl, 1996,157014)
'(Cheng etal, 1993,052294)
Source: (Watson et al., 2008,1571281
December 2009
A-285
-------
A.3.2. Source Profiles
Table A-55. Source Profiles: Part 1
Element
Aluminum
Antimony
Arsenic
Barium
Cadmium
Calcium
Chloride ion
Chromium
Cobalt
Copper
Total carbon
Gallium
Gold
Indium
Iron
Lanthanum
Lead
Magnesium
Manganese
Mercury
Molybdenum
Nickel
Nitrate
Organic
carbon
Palladium
Phosphorus
Potassium
Rubidium
Selenium
Silicon
Silver
Sodium
Strontium
Sulfate
Sulfur
Motor Vehicle Exhaust -
Symbol Gasoline
Al
Sb
As
Ba
Cd
Ca
Cl-
Cr
Co
Cu
TC
Ga
Au
In
Fe
La
Pb
Mg
Mn
Hg
Mo
Ni
N03"
OC
Pd
P
K
Rb
Se
Si
Ag
Na
Sr
S04"
S
Weight %
0.1
0.01
0.01
0.42
0.39
0.01
0.02
0
1.27
0
0.08
0.14
0.01
0
0.01
0.06
59.37
0.27
0.01
1.61
0.01
0.37
Uncertainty
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
Coal Combustion
Weight %
5.968
0
0
1.3315
0
3.4536
0.0176
0
0.0179
4.2763
0.014
0
2.916
0
0.068
0.0284
0
0
0.0072
0
0
0
0.9372
0.4644
0.0053
0.0406
9.0112
0
0.1964
10.1716
2.948
Uncertainty
0.5247
0.0625
0.0164
1 .0801
0.0341
1.0411
0.0041
0.0432
0.0112
4.2579
0.014
0.0404
0.3827
0.2462
0.0336
0.0139
0.0154
0.0134
0.0019
0.2116
2.9263
0.0263
0.6322
0.0602
0.0043
0.0407
0.5675
0.0312
0.0686
8.9405
2.729
Highway
Weight %
5.729
0
0
0.1377
0
2.5657
0.0271
0
0.0219
14.3927
0
0
4.5713
0
0.067
0.087
0
0
0.0081
0
12.7127
0
0
2.7161
0.0184
0
17.596
0
0.0395
1.1604
0.598
Road Dust
Uncertainty
0.4058
0.0335
0.0123
0.1027
0.019
0.1388
0.0023
0.0668
0.0101
2.3449
0.005
0.022
0.2661
0.1341
0.0074
0.009
0.0083
0.0071
0.0015
0.094
2.1296
0.0151
0.0324
0.3069
0.0023
0.0024
1.4183
0.0175
0.0078
0.2003
0.0509
Unpaved Road Dust
Weight %
7.4822
0
0
0
0
2.163
0.0312
0
0.0474
4.2671
0
0
5.5128
0
0.0288
0.1372
0
0
0.0091
0
4.2671
0
0.1603
2.8299
0.0184
0
24.2969
0
0.0313
0.8688
0.2808
Uncertainty
0.9315
0.1601
0.0226
0.5473
0.0881
1.0444
0.0161
0.0869
0.0307
3.7193
0.0233
0.1041
2.1152
0.6521
0.0284
0.0509
0.0383
0.0331
0.0057
0.6371
2.2637
0.0701
0.044
0.4949
0.0093
0.0108
4.0089
0.083
0.0112
1.3788
0.3884
Refinery
Weight %
8.4853
0
0
0
0
0.1236
0.0443
0
0.0299
0
0
0
1.4708
0
0.0097
0.016
0
0.0079
0.04
0
0
0
0.0689
0.0825
0
0
17.9733
0
0.0094
2.3243
0.6304
Uncertainty
2.3478
0.0285
0.0045
0.0979
0.0155
0.056
0.0127
0.0218
0.0082
1.6175
0.0059
0.0183
0.2216
0.1146
0.0063
0.002
0.0073
0.0088
0.0065
0.0772
1 .5288
0.0127
0.0144
0.0234
0.002
0.0021
5.1834
0.0151
0.0031
3.4523
0.9627
December 2009
A-286
-------
Thallium
Tin
Titanium
Uranium
Vanadium
Yttrium
Zinc
Zirconium
Ammonium
Sodium ion
Carbonate
Organic
carbon II
Organic
carbon III
Organic
carbon IV
ECI
Chlorine
atom
EC III
EC
Bromine
Atom
Organic
carbon 1
ECU
Sulfur
dioxide
Potassium
ion
Tl
Sn
Ti
U
V
Y
Zn
Zr
NH4+
Na+
C03"
OC2
OC3
OC4
EC1
Cl-
EC3
EC
Br
OC1
EC2
S02
K+
Motor Vehicle Exhaust- Coa| Combustion
0 0.0527
0.4315 0.0651
0 0.0734
0 0.006
0.49 N/A 0.0797 0.0341
0.0247 0.0043
0.34 N/A 0.3476 0.1352
0.0629 0.0221
16.44 N/A 4.2763 3.0931
0.0147 0.0154
7262.6687 7677.5681
0.1109 0.0571
Highway Road Dust
0 0.0298
0.3612 0.0313
0.0288 0.0074
0.0046 0.0012
0.0932 0.0256
0.0128 0.0025
0 0.025
3.4403 0.5505
1.68 0.9817
0.0037 0.0011
0.2295 0.1046
Unpaved Road Dust Refinery
0 0.1464 0 0.0254
0.5258 0.1289 0.6178 0.0711
0 0.0646 0.0432 0.0084
0 0.0146 0 0.0029
0.0502 0.021 0.0166 0.003
0.0219 0.0168 0.0166 0.0022
0 0.1317 0.3281 0.5565
0.1519 0.0755 0.0186 0.0074
0 2.9512 0 0.5283
0 0.0078 0 0.0017
0.1263 0.0744 0.0115 0.0059
Source: USA EPA Specials database http://www.epa.gov/ttnchie1/software/speciate/index.html
December 2009
A-287
-------
Part II
Element
Aluminum
Antimony
Arsenic
Barium
Cadmium
Calcium
Chloride ion
Chromium
Cobalt
Copper
Total carbon
Gallium
Gold
Indium
Iron
Lanthanum
Lead
Magnesium
Manganese
Mercury
Molybdenum
Nickel
Nitrate
Organic carbon
Palladium
Phosphorus
Potassium
Rubidium
Selenium
Silicon
Silver
Sodium
Strontium
Sulfate
Sulfur
Thallium
Tin
Titanium
Uranium
Symbc
Al
Sb
As
Ba
Cd
Ca
Cl-
Cr
Co
Cu
TC
Ga
Au
In
Fe
La
Pb
Mg
Mn
Hg
Mo
Ni
N03"
OC
Pd
P
K
Rb
Se
Si
Ag
Na
Sr
S042"
S
Tl
Sn
Ti
U
Residential Wood
Burning
Weight %
0.0034
0.0002
0.0003
0.0093
0.0013
0.0664
0.0028
0.0003
0.0005
0.0002
70.6416
0
0.0021
0.0038
0.0086
0.0031
0.003
0.0004
0
0.0002
0.2025
49.4961
0.0006
0
0.6346
0.0007
0.0001
0.0443
0.0023
0.0006
0.4553
0.1533
0.0006
0.001
Uncertainty
0.0103
0.0108
0.0016
0.0369
0.0058
0.0165
0.0004
0.0012
0.0005
0.0007
7.1435
0.0016
0.0069
0.0017
0.0431
0.0018
0.0013
0.0027
0.0024
0.0005
0.0156
5.481
0.0047
0.0051
0.1008
0.0007
0.0008
0.0167
0.0054
0.0009
0.0359
0.0173
0.0092
0.012
Oil Combustion
Weight
%
0
0
0.02
0
0
0
0.01
0.05
0.01
3.55
0.01
0
0.68
0
0
0
0
0
2.36
0
1.71
0
0
0
0
0.03
0
0
0
25.29
16.48
0
0.01
Uncertainty
0.05
0.01
0
0.03
0.01
0.04
0.01
0.01
0.01
1 .0855
0
0.01
0.1
0.04
0
0
0
0
0.23
0
0.56
0
0.65
0
0
0
0.09
0
0
5.62
1.62
0.01
0.01
Weight
%
0
0
0
0.01
0
0.01
0
0
0
98.94
0
0
0
0.02
0
0
0
0
0
0.06
90.8
0
0.01
0
0
0
0.01
0
0
0.53
0.59
0
0
DE
Uncertainty
0.01
0.01
0
0.04
0.01
0.01
0
0
0
17.859
0
0.01
0
0.05
0
0
0
0
0
0.01
14.79
0
0.02
0
0
0
0.01
0.01
0
0.07
0.21
0.01
0.01
Fly Ash
Weight
%
1.5708
0.007
0.001
0.0303
0
10.1398
17.5498
0.0054
0.0015
0.017
1 .4329
0.0013
0.0008
0
0.8306
0.0046
0.0031
0.4455
0.0426
0.0008
0.0041
0.0028
0
1 .4329
0
0.5808
24.4341
0.0351
0.0018
4.0201
0
2.8137
0.0406
8.0717
2.6349
0.0011
0.0067
0.058
0.0021
Uncertainty
0.4755
0.0218
0.0023
0.0655
0.0154
1 .7825
1.5419
0.001
0.0128
0.0013
0.2009
0.0018
0.0033
0.0164
0.059
0.0868
0.0031
0.0465
0.0033
0.0025
0.001
0.0004
0.2192
0.1592
0.0126
0.2447
5.0076
0.0026
0.0003
1.2886
0.0143
0.2174
0.0029
0.6409
0.1873
0.0025
0.0198
0.0093
0.0052
Incinerator
Weight
%
1.15
0.01
0
0.14
0.01
2.37
0.02
0
0.08
55.79
0
0.01
1.72
8.43
14.56
0.04
27.63
0.01
0.01
5.5
37.21
0.02
0.05
1.28
0
0.01
4.42
0.02
0.02
10.46
3.16
0.04
0.11
Uncertainty
0.83
0.15
0.04
0.55
0.08
0.62
0.02
0.03
0.1
27.5948
0.02
0.1
0.31
61.15
11.69
0.01
47.27
0.04
0
4.55
18.03
0.07
0.16
0.86
0.02
0.01
1.82
0.08
0.01
2.6
0.63
0.14
0.17
December 2009
A-288
-------
Residua, Wood Oi| Combustion
Vanadium V 0.0007 0.005 0.4
Yttrium Y 0.0001 0.0011 0
Zinc Zn 0.0762 0.0054 0.01
Zirconium Zr 0 0.0014 0
Ammonium NH4+ 0.1132 0.014 0.84
Sodium ion Na+ 0.11
Carbonate C03~ 0
0.04
0
0
0
0.24
0.02
0.0214
0
0
0.02
0
0.03
0
0.2577
DE Fly Ash
0.01 0.0038 0.011
0 0.0013 0.0021
0.02 0.031 0.0023
0 0.0039 0.0008
0.01 0.0234 0.022
0.01 4.7518 0.3438
0.4463
Incinerator
0.01 0.07
0 0.02
0.57 0.39
0 0.02
7.41 7.81
1.81 2.63
Organic carbon QC2 ?513 „ 66?5
Organic carbon QC3 8 962? 1 m5
Organic carbon QC4 2 ?683 1 1919
EC I EC1 20.342 2.9324
Chlorine atom CI" 0.2874 0.0404 0.05
0.01
0.03
0.01 27.5797 8.1193
6.35 10.46
EC III ECS 2.2878 0.4252
EC EC 21.1455 4.5813 1.84
Bromine Atom Br 0.0029 0.0011 0
0.93
0
8.14
0
10.01 0 0.1227
0 0.0441 0.0032
18.58 20.89
0.19 0.3
Organic carbon I OC1 25.1452 4.6648
ECU EC2 2.9362 1.2422
Sulfur dioxide S02
Potassium ion K+ 0.5208 0.0795 0.01
A.3.3. Receptor Model Results
0.01
0
0.01 14.5473 1.3393
1.01 0.42
Source: U.S. EPA SPECIATE database http://www.epa.gov/ttnchie1/software/speciate/index.html
Table A-56. PM2.6 receptor model results (pg/m3)
o«m«iin«o;*« Measured PM2 5 Vegetative
Sampling Site Concentration Burning
Albany, NY 2000-2001 20.9 5.5
Birmingham, AL, 2000-2001 16.2 3.3
Houston, TX, 2000-2001 12.4 3.1
Long Beach, CA, 2000-2001 30.0 4.6
Las Vegas, NV, 2000-2001 2.5 1.0
El Paso, TX, 2000-2001 5.5 0.7
Wfestbury, NY, 2000-2001 11.5 1.7
Road
Dust,
Soil
1.9
1.4
2.6
1.3
2.0
2.8
0.7
(NH4)2S04
2.4
3.7
1.6
2.1
0.5
0.7
5.2
NH4N03 Tailpipe ^
4.6 2.9 0.0
2.4 5.7 0.0
2.2 2.6 0.0
16.3 4.1 0.4
0.3 1.5 0.0
0.3 2.0 0.3
2.2 5.3 0.0
Source: Abu-Allaban et al. (2007, 0985751
December 2009
A-289
-------
Table A-57. PMio receptor model results (mass
Sampling Site
Apline.CA, 1994-1995
Apline.CA, 1995
Apline.CA, 1995
Atascadero, CA, 1994-1995
Atascadero, CA, 1995
Atascadero, CA, 1995
Lake Arrowhead, CA, 1994-1995
Lake Arrowhead, CA, 1995
Lake Arrowhead, CA, 1995
Lake Elsinore, CA, 1994-1995
Lake Elsinore, CA, 1995
Lake Elsinore, CA, 19952
Lancaster, CA, 1994-1995
Lancaster, CA, 1995
Lancaster, CA, 1995
Lompoc.CA, 1994-1995
Lompoc.CA, 1995
Lompoc.CA, 1995
Long Beach, CA, 1994-1995
Long Beach, CA, 1995
Long Beach, CA, 1995
MiraLoma.CA, 1994-1995
MiraLoma.CA, 1995
MiraLoma.CA, 1995
Riverside, CA, 1994-1995
Riverside, CA, 1995
Riverside, CA, 1995
San Dimas.CA, 1995
San Dimas.CA, 1995
Santa Maria, CA, 1994-1995
Santa Maria, CA, 1995
Santa Maria, CA, 1995
Upland, CA, 1994-1 995
Upland, CA, 1995
Upland, CA, 1995
Wood
Smoke
15.00
9.92
10.97
44.22
21.36
73.45
6.86
4.85
9.91
12.72
17.13
6.84
22.49
3.69
34.89
13.09
10.12
2.38
14.32
4.68
5.20
27.97
14.14
6.20
25.28
7.62
22.01
18.66
12.94
12.24
20.33
7.33
28.10
Diesel
33.19
58.78
65.64
22.16
38.99
18.11
46.55
65.20
38.90
44.01
74.72
38.48
43.14
46.18
37.30
18.16
51.27
79.42
43.24
70.25
56.80
48.87
53.72
41.88
46.67
52.15
47.65
71.35
61.34
23.99
52.57
48.13
46.39
68.69
46.52
Gasoline
Vehicles
46.46
11.47
10.81
26.44
12.41
33.92
7.40
46.70
18.61
10.85
20.56
12.66
7.33
49.65
14.73
10.19
16.49
5.47
6.15
18.10
6.65
8.87
12.03
7.93
4.87
4.48
22.03
11.87
10.79
14.08
3.50
4.90
percent)
Natural Gas
Combustion
2.73
4.95
0.79
0.26
0.21
0.45
0.20
0.61
0.13
0.86
0.72
0.16
0.15
0.23
0.27
0.47
0.17
0.33
Vegetative
Detritus
5.31
19.63
12.66
17.89
3.14
9.85
17.65
3.66
4.21
7.81
15.55
3.73
8.21
7.78
5.89
20.73
10.87
3.97
6.79
5.34
8.82
18.79
11.50
6.83
14.54
6.91
8.35
3.70
5.58
9.63
18.04
4.49
9.19
10.30
Tire Wear
Debris
6.91
9.43
5.31
20.42
28.01
9.78
29.17
11.93
26.38
26.00
14.11
16.61
19.52
15.71
9.85
20.31
19.06
20.17
7.85
8.15
12.78
15.05
14.70
11.25
9.81
Source: Manchester-Neesvig et al. (2003, 098102'
December 2009
A-290
-------
A.4. Exposure Assessment
A.4.1. Exposure Assessment Study Findings
Table A-58. Exposure Assessment Study Summaries
Adar et al. (2007, 098635)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Periods
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Cohort
March 2002-June 2002
St. Louis, Missouri
Senior citizens exposed to traffic-related PM
60
NR
Samples of FeNO were collected between 8:00 and 9:00 a.m. on the mornings before and after each trip. In the hours
surrounding these samples, group-level measurements of particle concentrations also were collected using several continuous
instruments installed on two portable carts. These carts were first positioned in a central location inside the participants' living
facilities 24-h before each trip. The carts remained at the facilities until it was time for the trips, at which point they followed the
participants from the health testing room, onto the bus, to the group activity, and to lunch. After the trip home aboard the bus,
the carts were returned to the central location in the living facility where they remained until the conclusion of the health testing
on the following morning. Continuous measurements of ambient particles and gases also were collected from a central
monitoring station in East St. Louis, Illinois. Two portable carts containing continuous air pollution monitors were used to
measure group-level micro-environmental exposures to traffic related pollutants, including PM2.5, BC, and size-specific particle
counts. PM2.s concentrations were measured continuously using a DustTrak aerosol monitor model 8520 with a Nafion diffusion
dryer. Integrated samples of PM2.5 mass also were collected using a Harvard Impactor for daily calibration of the trip and
facility.
Continuous BC concentrations were measured using a portable aethalometer with a 2.5-um impaction inlet. Particle counts
were measured using a model CI500 optical particle counter with a modified flow rate of 0.1 cubic feet per minute.
NR
PM2.5
PM2.5, PM10
BC, pollen and mold also assessed
PM2.5 exposures resulted in increased levels of FeNO in elderly adults, suggestive of increased airway inflammation. These
associations were best assessed by microenvironmental exposure measurements during periods of high personal particle
exposures. In pre-trip samples, both microenvironmental and ambient exposures to PM25 were positively associated with
FeNO. For example, an interquartile increase of 4 ug/m3 in the daily microenvironmental PM2.5 concentration was associated
with a 13% [95% Cl: 2-24) increase in FeNO. After the trips, however, FeNO concentrations were associated predominantly
with microenvironmental exposures, with significant associations for concentrations measured throughout the whole day.
Associations with exposures during the trip also were strong and statistically significant with a 24% (95% Cl: 15-34) increase in
FeNO predicted per interquartile increase of 9 ug/m3 in PM2 5. Although pre-trip findings were generally robust and the post-trip
findings were generally robust, the post-trip findings were sensitive to several influential days.
Adgate et al. (2002, 030676)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Comparison of outdoor, indoor and personal PM25 in three communities.
April-June, June-August, September-November, 1999
Battle Creek, East St. Paul, and Phillips, Minnesota, constituting the Minneapolis-St. Paul metropolitan area.
Adults in urban areas
Mean age 42 ± 10, range 24-64 yr
No
Inertial impactors (PEM) in a foam-insulated bag with shoulder strap with the inlet mounted on the front.
PM2.5
PM2.5
PM2.5
NR
The relative level of concentrations report in other studies was duplicated. Outdoor < indoor < personal. On days with paired
samples (n = 29), outdoor concentrations were significantly lower (mean difference 2.9 ug/m3, p = 0.026) than outdoor at home.
December 2009
A-291
-------
Adgate et al. (2007,156196)
Study Design
Period
Location
Population
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
NR
1999-; April 26-June 20, June 21-August 11, September 23-November 21
Minneapolis-St. Paul metropolitan area
NR
Cigarette smoke, resuspension of house dust from carpets, furniture and clothes, and emissions from stoves and kerosene
heaters (Leaderer et al., 1993; Ferro et al., 2004).
Personal monitoring was conducted for two consecutive days, and was conducted so that the two 24-h averages matched
indoor (I) and personal (P) measurements were collected in concert with outdoor (0) samples in each community. Gravimetric
concentrations for P and I were collected using inertial impactor environmental monitoring inlets and air sampling pumps. To
obtain I measurements, monitors were placed inside each residence in a room where the participants reported spending the
most waking hours. P measurements were obtained by carrying personal pumps in small bags. 0 samples were collected near
the approximate geographic center of each neighborhood and monitors ran from midnight to midnight for two consecutive 24-h
periods, followed by a day to change filters. Gravimetric 0 PM2.5 concentrations were obtained using a federal reference
method sampler.
PM2.5
PM2.5
PM2.5
Ag, Al, Ca, Cd, Co, Cr, Cs, Cu, Fe, K, La, Mg, Mn, Na, Ni, Pb, S, Sb, Sc, Ti, Tl, V, Zn
The relationships among P, I, and 0 concentrations varied across trace elements (TE). Unadjusted mixed-model results
demonstrated that 0 monitors are more likely to underestimate than overestimate exposure to many of the TEs that are
suspected to play a role in the causation of air pollution related health effects. These data also support the conclusion that TE
exposures are more likely to be underestimated in a lower income and centrally located community than in a comparatively
higher income community. Within the limits of statistical power for this sample size, the adjusted models indicated clear
seasonal and community related effects that should be incorporated in long-term exposure estimates for this population.
Adgate et al. (2003, 040341)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Time-series epidemiologic study
April-November 1999; spring: 26 April-20 June; summer: 21 June-11 August;fall: 23 September-21 November
Minneapolis-St. Paul, Minnesota
Healthy non-smoking results
24-64 yr (mean age 42 ± 10)
NR
Personal and indoor gravimetric PM concentrations were collected using PM2.5 inertial impactor environmental monitoring inlets
and air sampling pumps. Monitors were placed inside each participant's residence in the room where he/she reported spending
the majority of their waking hours to obtain I measurements. Participants also carried personal pumps in small bags to obtain P
measurements. Start times for indoor and personal monitors were always within a few minutes of each other. Gravimetric 0
and central site PM2 5 concentrations were obtained using a federal reference method sampler and EPA site requirements for
ambient sampling. Gravometric samples were collected near the approximate geographic center of each neighborhood, and
monitors ran from midnight to midnight for 2 consecutive 24-h periods, followed by a day to change filters.
NR
NR
NR
NR
PM2.5 concentrations were higher than I concentrations, which were higher than 0 concentrations. In healthy non-smoking
adults, moderate median for correlation between P and I; modest median for correlation between I and 0; and minimal median
correlation between P and 0 longitudinal were observed for PM25 measurements. A sensitivity analysis indicated that
correlations did not increase if the days with exposures to environmental tobacco smoke or occupational exposures were
excluded. In the sample population neither P nor I monitors provided a highly correlated estimate of exposure to 0 PM25 over
time. These results suggest that the studies showing relatively strong longitudinal correlation coefficients between P and 0
PM2.5 for individuals sensitive to air pollution health effects do not necessarily predict exposure to PM2.5 in the general
population.
Allen etal. (2003. 053578)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Use of continuous light scattering data to separate indoor PM into indoor- and outdoor-generated components to enhance
knowledge of the outdoor contribution to total indoor and personal PM exposures.
November 1999-May 2001
Seattle, WA
Elderly people and children spending most of their time (up to 70%) indoors. The study included healthy elderly subjects, elderly
with COPD and coronary heart disease (CHD), and child subjects with asthma.
Age n; 0-29 25; 30-59 36; >60 22; unknown 2
Suggested (not identified)
NR. Indoor and outdoor sampling conducted
NR
PM2.5
December 2009
A-292
-------
Ambient Size PM2.5
Component(s) S
Primary Findings A recursive mass balance model can be successfully used to attribute indoor PM to its outdoor and indoor components and to
estimate an avg Penetration, air exchange rate, deposition rate, and NH4+for each residence.
Allen etal. (2007.154226)
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Heating seasonOctober-February; Non-heating season March-September
Seattle, WA
NR
NR
NR
Indoor and outdoor PM2.5was measured using a 10-l/min single-stage Harvard Impactor (HI) with 37-mm Teflon filters. The
relationship between particle mass concentration and light scattering coefficient (bsp) was also measured on a continuous
basis indoors and outdoors using nephelometers (model 902 and 903).
NR
PM2.5
PM2.5
S (measured by XRF)
The authors showed that RM can reliably estimate Finf. Simulation results suggest that the RM Finf estimates are minimally
impacted by measurement error. In addition, the average light scattering response per unit mass concentration was greater
indoors than outdoors. Results show that the RM method is unable to provide satisfactory estimates of the individual
components of Finf. Individual homes vary in their infiltration efficiencies, thereby contributing to exposure misclassification in
epidemiologic studies that assign exposures using ambient monitoring data. This variation across homes indicates the need for
home-specific estimation methods, such as RM or S, instead of techniques that give average estimates of infiltration across
homes.
Annesi-Maesano et al. (2007,093180)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Population based
March 1999 to October 2000
Bordeaux, France; Clermont-Ferrand, France; Creteil, France; Marseille, France; Strasbourg, France; Reims, France
School children
10.4±0.7yr
NR
PM2.5 was monitored simultaneously in both schoolyards (proximity level) and fixed-site monitoring stations (city level) using
4L/min battery operated pumps attached to polyethylene filter sampling cartridges.
NR
NR
PM2.5
NR
Results show an increased risk for EIB and flexural dermatitis at the period of the survey, past year atopic asthma and SPT
positivity to indoor allergens in children exposed to high levels of traffic-related air pollution (PM2 5 concentrations exceeding
10 ug/m3). Population based findings are also consistent with experimental data that have demonstrated that inhalation of
traffic-related air pollutants either individually or in combination, can enhance the immune responses and airway response to
inhaled allergens, such as pollens or house dust mites, in atopic subjects.
Balasubramanian and Lee (2007,156248)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Primary Findings
Case study of 3 rooms of 1 flat on the 8th floor, and "outside the home."
May 12-23, 2004
Singapore
Residents of an urban area in a densely populated country.
NR
Time-activity logs identified tobacco smoking, cooking, household cleaning and general resident movements.
NR
NR
PM2.5
PM2.5
I/O suggest that chemicals such as CI", Na+, Al, Co, Cu, Fe, Mn, Ti, V, Zn, and EC were derived from the migration of outdoor
particles (I/O <1 or~1).
December 2009
A-293
-------
Barn et al. (2008,156252)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Measure indoor Fu of PM2.5 from forest fires/wood smoke, effectiveness of high-efficiency particulate air (HEPA) filter air
cleaners in reducing indoor PM2.5, and to analyze the home determinants of Finf and air cleaner effectiveness (ACE).
2004-2005 (summer 2004 and 2005, winter 2004)
British Columbia, Canada
Homes affected by either forest fire smoke or residential wood smoke
NR
NR
Personal Data RAM for ambient air sampling
Indoor home PM25
NR
Outdoor home PM25
NR
Use of HEPA filter air cleaners can dramatically reduce indoor PM25 concentrations. Number of windows and season predict Finf
(p< 0.001).
Baxter et al. (2007, 092726)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Part of a prospective birth cohort study performed by the Asthma Coalition for Community, Environment, and Social Stress
(ACCESS)
2003-2005. Non-heating season: May to October; Heating season: December to March
Boston (urban)
Lower socio-economic status (SES) households
NR
NR
PM2.5 samples were collected with Harvard personal environmental monitors (PEM). NO concentrations were measured using
Yanagisawa passive filter badges.
NR
PM2.5
PM2.5
EC
The authors' regression models indicated that PM2.5 was influenced less by local traffic but had significant indoor sources, while
EC was associated with local traffic and N02 was associated with both traffic and indoor sources. However, local traffic was
found to be a larger contributor to indoor N02 where traffic density is high and windows are opened, whereas indoor sources
are a larger contributor when traffic density is low or windows are closed. Similarly, traffic contributed up to 0.2 |ig/m3 to indoor
EC for homes with open windows, with an insignificant contribution for homes where windows were closed.; Comparing models
based on p-values and using a Bayesian approach yielded similar results, with traffic density volume within a 50 m buffer of a
home and distance from a designated truck route as important contributors to indoor levels of N02 and EC, respectively.
However, results from the Bayesian approach also suggested a high degree of uncertainty in selecting the best model. The
authors concluded that by utilizing public databases and focused questionnaire data they could identify important predictors of
indoor concentrations for multiple air pollutants in a high-risk population.
Baxter et al. (2007, 092725)
Study Design
Period
Location
Population
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Copollutant(s)
Primary Findings
Simultaneous indoor and outdoor samples taken in 43 low SES homes in heating and non-heating seasons. Homes were
selected from a prospective birth cohort study of asthma etiology (n = 25). Non-cohort homes were in similar neighborhoods
(n = 18).
2003-2005
Boston, Massachusetts
Lower SES populations in urban areas
Home type, year built, tobacco smoke, opening windows, time spent cooking, use of candles or air freshener, cleaning activities,
air conditioner use.
NR
NR
PM2.5
NR
EC (m-1 x 10"5); Ca (ng/m3); Fe (ng/m3); K (ng/m3); Si (ng/m3); Na (ng/m3); Cl (ng/m3); Zn (ng/m3); S (ng/m3); V (ng/m3)
N02
The effect of indoor sources may be more pronounced in high-density multi-unit dwellings. Cooking times, gas stoves, occupant
density and humidifiers contributed to indoor pollutants.
December 2009
A-294
-------
BeruBe et al. (2004, 007894)
Study Design
Period
Location
Population
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
6 homes in Wales and Cornwall were monitored four times per year, inside samples in the living areas and outside the home.
NR but < 2003
Wales and Cornwall, UK
Urban, suburban, and rural homes
ETS, pets, cleaning, traffic load
NR
NR
PM10
NR
NR
There are greater masses of PMio indoors, and that the composition of the indoor PM10 is controlled by outdoor sources and to
a lesser extent by indoor anthropogenic activities, except in the presence of tobacco smokers. The indoor and outdoor PMio
collected was characterized as being a heterogeneous mixture of particles (soot, fibers, sea salt, smelter, gypsum, pollen and
fungal spores).
Branis et al. (2005,156290)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Human exposure assessment in a university lecture hall
Oct. 8, 2001-Nov. 11,2001
Prague, Czech Republic
University students
NR
Presence of people identified as a source of coarse particles ; outdoor air identified as a source of indoor fine particles (PMi 0
and PM2.5)
Harvard impactors (HI) with membrane Teflon filters
PM,, PM15, PM10
PMi, PM2.5, PM10
PM^
NR
Presence of people is an important source of coarse particles indoors ; Outdoor air may be an important source of fine indoor
particles.
Brunekreef et al. (2005, 090486)
Study Design Exposure assessment
Period Winter and spring 1 998-1 999
Location Amsterdam and Helsinki
Population Elderly
Age Groups 50-84 yr
Indoor Source NR
Personal Method Amsterdam Gillian with made to fit bags with belt with GK2.05 cyclone samplers 4L/min; Helsinki BGI with shoulder strap or
backpack with GK2.05 cyclone samplers 4 L/min.
Personal Size PM2.5
Microenvironment Size PM2.5
Ambient Size PM2.5
Component(s) S042"
Primary Findings In both cities, personal and indoor PM2.5 were highly correlated with outdoor concentrations.
Chillrud et al. (2004, 054799)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Repeated measures on a cohort of high school students in New York City
Summer and winter of 1999 (eight weeks each)
Manhattan, Bronx, Queens, Brooklyn, NY
Persons traveling the subway
14-1 Syr
No
Sampling packs carried by subjects
PM2.5
PM2.5 (home indoor and home outdoor)
PM2 5. Urban fixed-site and upwind fixed site operated for three consecutive 48-h periods each week.
Elemental Fe, Mn,and Crare reported in this study out of 28 elements sampled.
Personal samples had significantly higher concentration of Fe, Mn, and Cr than home indoor and ambient samples. The ratios
of Fe (ng/ug of PM2.5) vs Mn (pg/ug PM2.5) showed personal samples to be twice the ratio for crustal material. Similarly for the
Cr/Mn ratio. The ratios and strong correlations between pairs of elements suggested steel dust as the source. Time-activity
data suggested subways as a source of the elevated personal metal levels.
December 2009
A-295
-------
Conner and Williams (2004,156364)
Study Design This is part of the EPA Baltimore PM Study of the Elderly.
Period July-August, 1998
Location Towson, Maryland
Population 65+ adults
Age Groups 65+ yr
Indoor Source Personal sampling devices (PEM)
Personal Method PM2.5
Personal Size PM2.5
Microenvironment Size NR
Ambient Size NR
Primary Finding(s) A greater variety of particles was observed in the personal samples compared to the fixed-location apartment samples.
Cortez-Lugo et al. (2008,156368)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Cohort
Feb-Nov 2000
Mexico City, Mexico
Ambulatory adults with moderate to severe COPD, active smokers excluded
Adults
carpeting, aerosol sprays used, boiler use and location, animals, mold, tobacco smoking, windows closed
Personal pumps with 37-mm Teflon filters, flow rate 4 l/min in a bag with shoulder strap. The impactor was near the breathing
zone
PM2.5
PM25,PMio
PM25,PM10
NR
Indoor PM25 concentrations explained 40% of the variability of personal exposure. The best predictors of personal exposure
were indoor contact with animals (12%), mold (27%), being present during cooking (27%), and aerosol use (17%).
Crist et al. (2008,156372)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Indoor, outdoor, and personal monitoring
January 1999-August 2000
Ohio
Fourth & fifth-grade children
9-11 yr old
Filter, portable pump
Filter, PM2.5
Indoor school; Filter, PM2.s
Outdoor school; Filter, PM25
PM2.5
NR
Higher correlation was observed between P and I compared with the correlation between either P and ambient (A) or I and A.
Delfino et al. (2004, 056897)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Copollutant(s)
Primary Findings
Panel study with repeated measures
Sep-Oct 1999 or Apr-Jun 2000
Alpine, California
Children
9-17 yr
No
Personal dataRAM (pDR) carried at waist level using a fanny pack, shoulder harness, or vest.
0.1-10 urn
PM10 and PM2 5; measured immediately outside the house and in the living room of the home.
PM10
03 and N02 measured at central site
Percent predicted FEV, was inversely associated with personal exposure to fine particles. Also with indoor, outdoor and central
site gravimetric PM2.5, PM10, and with hourly TEOM PM10.
December 2009
A-296
-------
Delfino et al. (2006, 090745)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Cohort. Measured daily expired NO (FeNO)
Aug-Dec 2003
Riverside and Whittier, California
Children with asthma exacerbations in previous 12 months, non-smokers, non-smoking households
9-1 Syr
No
Wore a backpack during waking hours for PM2.5, EC and OC, N02, temperature, and relative humidity. Exhaled air collected in
Mylar bags to analyze for NO.
24-h PM25; 1-h max PM25; 8-h max PM25; 24-h N02
NR
24-h PM25; 24-h PM10; 8-h max 03; 8-h max N02; 24-h N02; 8-h max CO
24-h PM15 EC; 24-h PM2.5OC
The strongest positive associations were between FeNO and 2-day average pollutant concentrations. Per IQR increases 1.1
ppb FeNO/24 ug/m3 personal PM2.5.; 0.7 ppb FeNO/0.6 ug/m3 personal EC; 1.6 ppb FeNO /17 ppb personal N02 Ambient
PM2.5 and personal and ambient EC were significant only when subjects were taking inhaled corticosteroids. Subjects taking
both inhaled steroids and antileukotrienes had no significant associations. Distributed lag models showed personal PM25 in the
preceding 5 h was associated with FeNO.
Diapouli et al. (2007,156397)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Exposure assessment. Sampling of schools, residence, private vehicle
Schools-11/2003-02/2004 and 10/2004-12/2004.; Residence-10/2004; Vehicle-10/204-12/2004
Athens, Greece
Primary school children
NR
NR
Handheld portable Condensation Particle Counters (TSI, Model 3007) were used for all sampling locations. Primary schools
indoor measurements were primarily conducted inside classrooms, at table height. However, at three of the schools, rooms of
different uses were selected. These included a teachers' office (where smoking was permitted), a computer day lab (used by
students only part of the day), and a library and gymnasium (where intense activity took place almost all day long). Outdoor
measurements took place in the yard of each school. Residence samples were taken in a bedroom at breathing height and on
the terrace, for indoor and outdoor samples, respectively. In-vehicle samples were taken by placing the CPC 3700 on the
passenger seat while the vehicle drove along predetermined routes.
NR
0.01-1 urn
0.01-1 urn
NR
The results showed that children attending primary schools in the Athens area are exposed to significant PM concentration
levels, both indoors and outdoors. Vehicular emissions seem to be a major contributor to the measured outdoor concentration
levels at the studied sites. Indoor PM concentrations appeared to be influenced by both vehicular emissions and indoor
sources including cleaning activities, smoking, a high number of people in relation to room volume and furniture material (i.e.,
carpet). UFPs concentrations diurnal variation, both outside the schools and the residence, supports the close relation of UFPs
levels with traffic density. Indoor concentrations within schools exhibited variability during the school day only when there were
significant changes in room occupancy. 24-h variation of indoor concentrations at the residence were well correlated with the
outdoor concentration (R2 = 0.89).
Diapouli et al. (2008,190893)
Study Design Indoor, outdoor air monitoring of PM. To determine children exposure in school environment. To evaluate relationship between
indoor and outdoor levels.
Athens, Greece
Primary schools
NR
Indoor PMi, PM2.5, PM10, presence of children and activities of children in classrooms, infiltrated vehicular exhaust
Harvard PEM, Teflon filters Dust Trak Condensation particle counter
NR
PM^PMzs, PM10
Period
Location
Population
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
N03", S042"
High levels of PMio and PM2.5 measured indoors and outdoors. PMio more variable spatially than PM2.5. I/O ratio for PMio and
PM2.5 close to 1 at almost all sites. Ratio of PMi smaller than 1 in all cases. Vehicular traffic presumed to be the main source of
PM,. Indoor PM2.5 and PM10 levels dependent on the amount of activity in classroom and outdoor levels. Indoor S042~
concentrations strongly associated with outdoor levels. Result suggests that S042~ can be used as a proper surrogate for
indoor PM of outdoor origin.
December 2009
A-297
-------
Ebeltetal. (2005. 056907)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Personal exposure assessment related to health outcomes for a sensitive sub-population
Summer 1998
Vancouver, British Columbia, Canada
16 persons who had COPD
Mean subject age 74 yr, Range 54 to 86
Separated total personal exposure into "ambient" and "non-ambient" based on sulfate results and modeling.
24-h integrated filter sample
PM2.5
PM2.5, PM10, PM10.2.5
PM25, PM10, PM10.25
S042'
Ambient exposures and (to a lesser extent) ambient concentrations were associated with health outcomes. Total and
nonambient particle exposures were not.
Farmer et al. (2003, 089017)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Case control molecular epidemiology studies of carcinogenic environmental pollutants, particularly PAHs
12 months
Prague, Czech Republic (2 sites); Kosice, Slovak republic; Sofia, Bulgaria
Policeman and busdrivers usually working through busy streets in 8-10 h shifts and a control population.
Variable, range not stated
NR
Personal Monitoring Devices; Blood and Urine Samples; Stationary Versatile Air Pollution Samplers (VAPS)
PM10
NR
PM10; PM2.5 (not reported)
Extractable organic matter (EOM), B[a]P, c-PAHs
EOM per PM^was at least 2-fold higher in winter than in summer, and c-PAHs over 10-fold higher in winter than in summer.
Personal exposure to B[a]P and to total c-PAHs in Prague ca. was 2-fold higher in the exposed group compared to the control
group, in Kosice ca. 3-fold higher, and in Sofia ca. 2.5-fold higher.
Ferro et al. (2004, 055387)
Study Design
Period
Location
Population
Age Groups
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Case study, 1 home
Redwood City, California
NR
NR
NR
Co-located real-time particle counters and integrated filter samplers (Met-One Model 237B) were used to measure personal
(PEM), indoor (SIM) and outdoor (SAM) PM concentrations. The PEM was attached to a backpack frame and worn by the
investigator while performing prescribed activities. The SIM was attached to a six foot step-ladder with the intake at breathing
height. The SAM was located under a two-sided roofed shed in the backyard of the home with the filter samplers supported by
a metal stand and the real-time particle counters sitting on a table.
PM5
PM25;PM5
PM25;PM5
NR
The results of this study indicate that house dust resuspended from a range of human activities increases personal PM
concentrations and this resuspension effect significantly contributes to the personal cloud. The results of this study also
suggest that normal human activities that resuspend house dust may contribute significantly to the strong correlations found
between personal exposure and indoor PM concentrations in previous studies. The PEM/SIM ratios for human activity
presented in this paper are also in the range of those reported by previous studies.
December 2009
A-298
-------
Gadkari and Pervez (2007,156459)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Evaluation of relative source contribution estimates of various routes of personal RPM in different urban residential
environments.
Summer 2004 (March 15-June 15)
Chattisgarh, India
All likely. Not specified
21-61 yr, average age 40 ± 15 yr
No
Personal respirable dust samplers (RDS) with GFF
RPM
NR
RPM
Fe, Ca, Mg, Na K, Cd, Hg, Ni, Cr, Zn, As, Pb, Mn and Li
Authors concluded that "(1) indoor activities and poor ventilation qualities are responsible for major portion of high level of
indoor RPM, (2) majority of personal RPM is greatly correlated with residential indoor RPM, (3) time-activity diary of individuals
has much impact on relationship investigations of their personal RPM with their respective indoor and ambient-outdoor RPM
levels; as reported in earlier reports and (4) residential indoors, local road-traffic and soil-borne RPMs are the dominating
routes of personal exposure compared to ambient outdoor RPM levels."
Gauvin et al. (2002, 034893)
Study Design
Period
Location
Population
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Fine particle exposure assessment for children in French urban environments, part of VESTA study
March 1998-December2000
Paris, Grenoble, Toulouse, France
Children aged 8-14 yr
ETS from mother, rodents at home.
SKC pump 4 Lpm with PM2 5 inlet and 37 mm, 2 micron Teflon filter
PM2.5
NR
PM10
NR
The final model explains 36% of the between subjects variance in PM25 exposure, with ETS contributing more than a third to
this.
Graney et al. (2004, 053756)
Study Design
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
The study was designed to assess the trace metal quantification abilities of several analytical methods to measure the total as
well as soluble amounts of metals with PM2.5 collected from indoor and PM samples. (X-ray fluorescence and instrumental
neutron activation analysis)
Retirement facility in Towson, Maryland
Retirement facility with subjects who spent 94% of their time indoors
Mean age = 84 yr
NR
Measured using personal exposure monitors (MSP Inc) with nozzle to remove particles > 4 urn
PM2.5
PM2.5
NR
42 elements were analyzed for in the PM2 5 samples collected from personal and well as indoor samples
Most of the extractable components of the metals were in a water-soluble form suggesting a high potential for bioavailability of
elements from respiratory exposure to PM25. Based on comparison of trace metals in central I site vs. P samples, resident
activities result in exposure to higher concentration of soluble trace metals.
December 2009
A-299
-------
Haverinen-Shaughnessy et al. (2007,156526)
Study Design
Period
Location
Population
Age Groups
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Cross-sectional
Winter, year not reported
Eastern Sweden
Elementary school teachers
NR
Button inhalable aerosol samplers
Particle mass
Particle mass
NR
Absorbance coefficient/m x 10"5; Total fungi (spores/m3); Total bacteria (cells/m3); Viable fungi MEA (CFU/m3); Viable fungi
DG18 (CFU/m3); Viable bacteria (CFU/nf)
The recall period of 7 days provided the most reliable data for health effect assessment. Both personal exposure and
concentrations of pollutants at home were more frequently associated with health symptoms than work exposures.
Ho et al. (2004, 056804)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Human exposure assessment
25 Sept. 2002 to 8 March 2003
Hong Kong
Occupied buildings located near major roadways
NR
Yes. Regression of indoor versus outdoor concentrations of OC and EC revealed an indoor source of OC not present for EC,
presumably due to such activities of cooking, smoking, and cleaning.
Co-located mini-volume samplers (flow rate 5 L/min) and Partisol model 2000 sampler with 2.5 |im inlet. All samples on 47 mm
Whatman quartz microfiber filters, weighed on an electronic microbalance. Analyzed for OC and EC using DRI Model 2001
Thermal/Optical Carbon Analyzer.
PM2.5
NR
PM2.5
OC,EC,OM,TC
The major source of indoor EC, OC, and PM2.5 appears to be penetration of outdoor air, with a much greater attenuation in
mechanically ventilated buildings.
Hoek et al. (2008,156554)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Exposure assessment, characterizing indoor/outdoor particle relationships
October 2002-March 2004
4 European cities Amsterdam, Athens, Birmingham, Helsinki
Urban populations
NR
Smoking, candle burning, cooking/frying
No personal exposure assessment was conducted
NR
PM10, PM25, PM10-25, Ultrafine (UFP)
PM10, PM2.5, PM10-2.5, UFP
soot, sulfate
Correlation between 24-h average central site and indoor concentrations was lower for UFP than for PM2.5, soot, or S042",
probably related to greater losses during infiltration due to smaller particle size. Infiltration factors for UFP and PM2.5 were low.
Hopke et al. (2003, 095544)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Exposure assessment
26 July to 22 August 1998
Retirement facility in Towson, MD
Elderly residents
Mean age of 84
Ammonium sulfate and ammonium nitrate, secondary sulfate, OC, and motor vehicle exhaust
Inertial impactor PEM in the breathing zone of the subjects
PM2.5
PM2.5
PM2.5
S042'
Personal exposures were influenced by a combination of indoor and outdoor factors. Indoor factors included gypsum, personal
grooming products, and an unknown indoor source. Outdoor factor included S042", soil, and an unknown factor. Outdoor
factors accounted for 63% of personal exposure, and S042- was the largest ambient contributor to personal exposure (48%).
December 2009
A-300
-------
Jacquemin et al. (2007,156600)
Study Design Assessment of relationship between outdoor and personal concentrations of PM2 5 absorbance and sulfur among post-
myocardial infarction patients
Period January 2004-June 2004
Location Barcelona, Spain
Survivors of a myocardial infarction exposed to ETS
n = 38, including 32 and 15 over age 64.
ETS
Personal Method Personal samplers (BGI GK2.05 cyclones and battery operated BGIAFC400S pumps)
Personal Size PM25
NA
PM2.5
S
Primary Findings Ambient measurements of light extinction and S can be used as surrogates to personal PM2.5 exposure, especially for those
exposed to ETS.
Population
Age Groups
Indoor Source
Microenvironment Size
Ambient Size
Component^
Janssen et al. (2005, 088692)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Panel Study
Amsterdam 11/2/1998-6/18/1999; Helsinki 11/1/1998-4/30/1999
Amsterdam, The Netherlands; Helsinki, Finland
Elderly Cardiovascular Patients
50-84 yr
No
Personal PM25 GK2.05; cyclones; indoor & outdoor Harvard Impactors; Reflectance EEL 43 reflectometers; Elemental
Composition Tracer Spectrace 5000 ED-XRF system
PM2.5
PM2.5
PM2.5
Estimated EC, elemental composition of a subset of personal, indoor and outdoor samples
For most elements, personal and indoor; concentrations were lower than and highly correlated with outdoor concentrations. The
highest correlations (median r = 0.9) were found for sulfur and particle absorbance (EC), which both represent fine; mode
particles from outdoor origin. Low correlations were observed for elements that represent the coarser part of the PM25 particles
(Ca,Cu,Si, CI").
Jedrychowski et al. (2006,156606)
Study Design Prospective cohort
Period 11/2000-3/2003
Location Krakow, Poland
Population Non-smoking pregnant women
Age Groups Yes
Personal Method Personal Exposure Monitor Sampler (PEMS, Harvard; School of Public Health)
PM2.5
NR
PM10
NR
Primary Findings The contribution of the background ambient PM10 level was a very strong determinant of the total personal exposure to PM2.5,
and it explained about 31% of variance between the subjects.
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Johannesson et al. (2007,156614)
Study Design Cohort
Period Spring and fall seasons of 2002 and 2003
Location Gothenburg, Sweden
Population General adult population
Age Groups 23-51 yr
Indoor Source NR
Personal Method Fine particles were measured for 24 h using both personal and stationary monitoring equipment. Personal monitoring of PM2 5
and PMi was carried out simultaneously with parallel measurements of PM2.5 and PMi indoors in living rooms and outside the
house on a balcony, porch, etc. In addition, urban background PM25 levels were measured. Personal monitoring was
performed in two ways. The 20 randomly selected subjects carried personal monitoring equipment for PM2.5 only, while the 10
staff members carried two pieces of personal monitoring equipment at the same time. On the first measuring occasion, the staff
members carried one PM2.5 cyclone and one PMi cyclone. On the second occasion, duplicate monitors for PM2.5 were used.
For personal and residential monitoring, the BGI Personal Sampling Pump was used together with the GK2.05 cyclone for
PM2.5 sampling and the Triplex cyclone SCC1.062 for PMisampling. The personal sampling pump was placed in a small
December 2009
A-301
-------
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
shoulder bag and the cyclone attached to the shoulder strap near the subject's breathing zone. The personal monitoring
equipment was carried by the subject during awake time. During the night, it was placed in the living room. For indoor
monitoring in living rooms, cyclones (PM2.5 and PMi) were placed at about 1.5 m above the floor. The same setup was used for
residential outdoor monitoring. The urban background monitor was placed on top of a roof somewhat south of the city center
but not near any major highway.
PM^PM,
PMjslPM,
PM25IPM,
BS
Personal exposure of PM2.5 correlated well with indoor levels, and the associations with residential outdoor and urban
background concentrations were also acceptable. Statistically significantly higher personal exposure compared with residential
outdoor levels of PM2.5 was found for nonsmokers. PMi made up a considerable proportion (about 70-80%) of PM2.5. For BS,
significantly higher levels were found outdoors compared with indoors, and levels were higher outdoors during the fall than
during spring. There were relatively low correlations between particle mass and BS. The urban background station provided a
good estimate of the residential outdoor concentrations of both PM2.5 and BS2.5 within the city. The air mass origin affected the
outdoor levels of both PM2.5 and BS2.5; however, no effect was seen on personal exposure or indoor levels.
Kaur et al. (2005, 086504)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Exposure assessment, evaluation of exposures between modes of transport, routes, timing
April 28-May 23, 2003
Street canyon intersection in Central London, UK
Users of an urban street canyon intersection
NR
NR
PM2.5 measured usnig high-flow gravimetric personal samplers (PM2.5) operating at a flow rate of 16 l/min carried in a backpack
with sampling head positioned in personal breathing zone. UFP measured using TSI P-TRAK particle counters in which
isopropyl alcohol condenses to form droplets that can be easily counted by a photodetector as they pass through a laser beam.
PM25, UFP (0.02-1 .Oum)
PM25, UFP (0.02-1 .Oum)
PM2.5
NR
Personal exposures to PM2.5 while walking were significantly lower then while riding in a car or taxi, likely a function of greater
distance to roadside. No significant differences in PM2 5 were observed between exposures on the high traffic road compared
with the backroad. Personal exposure levels were lowest during midday measurements for PM2.5 and highest in the early
evening. Personal exposures to ultrafine particles were lowest while walking and highest while riding the bus. Exposures to
ultrafine particles were also significantly higher on the high traffic road and during morning measurements. Exposure to
ultrafine particles were highest in the morning, likely the result of peak traffic density in the morning. Exposure assessment also
revealed that the background and curbside monitoring stations were not representative of the personal exposure of individuals
to PM2.5 and CO at and around a street canyon intersection.
Kaur et al. (2005, 088175)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Personal exposure assessment of pedestrians walking along high-traffic urban road
April 19, 2004-June 11, 2004
Central London, UK
Pedestrians
NR
NR
PM2 5 gravimetric filter measurement, UFP (0.02-1 urn) P-TRAK device, reflectance reflectometer measurement of PM25 filter
PM25, UFP (0.02-1 urn)
NR
PM2.5, UFP (0.02-1 urn)
Absorbance of PM2.5 filter
PM2.s pedestrian exposure was well correlated with and above background fixed-site monitoring levels. PM pedestrian exposure
was influenced by proximity to curbside and the side of the road walked on.
December 2009
A-302
-------
Kim et al. (2005,156640)
Study Design Panel study
Period 8/1999-11/2001
Location Toronto, Canada
Population Cardiac-compromised patients
Age Groups Mean age 64 yr
Indoor Source Gas range (68%); indoor grill (11%); outdoor barbeque (30%); Gas heating fuel (68%); Oil heating fuel (7%)
Personal Method Rupprecht and Patashnick ChemPass Personal Sampling System
Personal Size PM2.5
Microenvironment Size NR
Ambient Size PM2.5
Component(s) NR
Primary Findings Personal PM2.5 exposures were higher than outdoor ambient levels. Personal PM2.5 exposures levels were correlated with
ambient levels, mean r = 0.58
Koistinenetal. (2004.156655)
Study Design Representative Population-based study
Period Oct1996-Dec1997
Location Helsinki, Finland
Population Non-smoking adults not exposed to environmental tobacco smoke.
Age Groups Adults 25-55 yr
Indoor Source Soil from outdoors, cooking, smoking, aerosol cleaners, sea salt, combustion sources
Personal Method Integrated 24-h filter sample
Personal Size PM2.5
Microenvironment Size PM2.5
Ambient Size PM2.5
Component(s) BS
Primary Findings Population exposure assessment of PM25, based on outdoor fixed-site monitoring, overestimates exposures to outdoor sources
like traffic and long-range transport and does not account for the contribution of significant indoor sources.
Kousaetal. (2001. 025270)
Study Design Population based exposure assessment
Period October 1996 to June 1998
Location Helsinki, Finland; Basel, Switzerland; Prague, Czech Republic; Athens, Greece
Population Adult urban populations
Age Groups 25-55 yr
Indoor Source Sometimes ETS
Personal Method Integrated 48-h filter sample
Personal Size PM2.5
Microenvironment Size PM2.5
Ambient Size PM2.5, PM10
Component(s) NR
Primary Findings Throughout the study, the highest correlations were those between personal exposures and indoor concentrations, which
suggests that indoor sources were important. Correlations were generally lower between ambient concentrations and personal
exposures.
December 2009
A-303
-------
Koutrakis et al. (2005, 095800)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Panel study
Baltimore 6/28/98-8/22/98 (summer), 2/1/99-3/16/99 (winter); Boston 6/13/99-7/23/99 (summer), 2/1/00-3/12/00 (winter)
Baltimore, MD Boston, MA
Healthy older adults, children, adults with COPD
Children 9-13y/o; Seniors 65+y/o
NR
Personal exposure samples of PM25; were collected using a specially designed multipollutant sampler (Demokritou et al. 2001).
PM2.s was collected using personal environmental monitors (PEMs) and 37-mm; Teflon filters (Teflo, Gelman Sciences, Ann
Arbor Ml).
PM2.5
NR
PM2.5
EC, S042"
Ambient PM25 and S042~ are strong predictors of respective personal exposures. Ambient S042~ is a strong predictor of
personal exposure to PM25. Because PM25 has substantial indoor sources and S042~ does not, the investigators; concluded
that personal exposure to S042~ accurately reflects exposure to ambient PM25 and therefore the ambient component of
personal exposure to PM2.5 as well.
Lai et al. (2004, 056811)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Personal exposure study
December 1998-February 2000
Oxford, UK
Adults
25-55 yr(avg = 41)
Cooking, active smoking, passive smoking heating by gas heater
Integrated 48-h filter samples
PM2.5
PM2.5
PM2.5
Ag, Cr, Mn, Si, Al, Cu, Na, Sm, As, Fe, Ni, Sn, Ba, Ga, P, Sr, Br, Ge, Pb, Ti, Ca, Hg, Rb, Tl, Cd, I, S, Tm, Cl, K, Sb, V, Co, Mg,
Se.Zn.Zr
Personal exposures were influenced by both indoor and ambient sources, and indoor levels exceeded ambient levels for PM25
as well as for VOCs and eight other compounds. Correlation between personal and indoor PM2.5 was 0.60 (p < 0.001).
Larson et al. (2004,098145)
Study Design
Period
Location
Population
Age Groups
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Time-series epidemiologic study
Sep26,2000-May25,2001
Seattle, Washington
"Susceptible Populations"
Time-activity diary
Harvard Personal Environmental Monitor
PM2.5
PM2.5 outside subject's residence, and inside residence
PM2.5 at Central outdoor site (downtown Seattle)
Light absorbing carbon (LAC) and trace elements
Five sources of PM2 5 identified vegetative burning, mobile emissions, secondary sulfate, a source rich in chlorine, and crustal-
derived material. The burning of vegetation (in homes) contributed more PM25 mass on average than any other sources in all
microenvironments.
December 2009
A-304
-------
Li et al. (2003,047845)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Primary Findings
Concurrent 10-min avg indoor and outdoor concentrations of PM10 and PM2.5 were recorded for 2 days each in 10 homes with
swamp coolers
Summer 2001
El Paso, Texas
Cooking, cleaning, walking
NR
NR
PM2.5 and PMi0; indoor and outdoor; tapered element oscillating microbalance (TEOM) instruments. 2 days were monitored for
PM25, and2forPM10.
NR
NR
Evaporative coolers were found to act as PM filters, creating indoor concentrations approximately 40% of outdoor PMio and
35% of outdoor PM2.5, regardless of cooler type.
Liu et al. (2003, 073841)
Study Design
Period
Location
Population
Age Groups
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Primary Findings
Comprehensive exposure assessment
1999-2001
Seattle, WA
High-risk sub populations
Children 6-13 yr, elderly 65-90 yr (one person was below 65, but his/her age was not specified)
Harvard Personal Environmental Monitor for PM2 5 (HPEM2 5)
PM25,PM10
PM25,PM10
PM2.5, PM10
Average personal PM2.5 exposure was similar to ambient PM2.5 concentrations but much higher than average indoor
concentrations. Personal, indoor, and outdoor PM2.5 and PM10, as well as the ratio PM2.5/PM10, were all significantly higher
during the winter. Personal PM2.5 and PM^ exposures were highest for the children in the study.
Lung et al. (2007,156719)
Period
Location
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Weekdays between Nov 1998 and Feb 1999
6 communities in Taiwan, China 2 in Taipei, 2 in Taichung, and 2 in Kaohsiung. Sites are industrial, commercial, residential and
mixed.
18to>70
Being in kitchen, park, major boulevard, stadium, incense burning, household work, factory, environmental tobacco smoke,
traffic, ventilation conditions
Personal Environmental Monitor with a SKC personal pump at 2 L/min, 37 mm Teflon filters
PM10
PM10
PM10
None
Outdoor rather than indoor levels contributed significantly to personal exposure. Important factors include time spend outdoors
and on transportation, riding a motorcycle, passing by factories, cooking or being in the kitchen, incense burning at home.
Meng et al.. (2005, 081194)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Evaluation of the use of central-site PM, rather than actual exposure, in PM epidemiology
Summer 1999-spring 2001
3 cities:Houston (TX), Los Angeles County (CA), and Elizabeth (NJ)
NR
NR
NR
MSP monitors on the front strap of the sampling bag near the breathing zone. Pump, battery, and motion sensor were on the
hip or back.
PM2.5
PM2.5
PM2.5
EC,OC,S,Si
Use of central-site PM2 5 as an exposure surrogate underestimates the bandwidth of the distribution of exposures to PM of
ambient origin.
December 2009
A-305
-------
Meng et al. (2005, 058595)
Study Design
Period
Location
Population
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
RIOPA study matched indoor home & outdoor exposure assessment
May-October (hot); November-April (cool); (1999-2001)
Los Angeles County, CA; Elizabeth, NJ; Houston, TX
Non-smoking homes
Combustion (primary); atmospheric (secondary); sulfate, organics, nitrates; mechanically (abrasion) generated.
Filter (not specified)
NR
Indoor home.; PM2.5
PM2.5, outdoor home
Organic and elemental carbon; 24 elements (metals).
The median contribution of ambient sources to indoor PM2.5 using the mass balance approach was 56% for all study homes,
63% for California, 52% for New Jersey, and 33% for Texas.
Molnar et al. (2005,156772)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Indoor/outdoor exposure assessment related to domestic wood burning
10 February to 12 March 2003
Hagfors, Sweden
Adult residents of Hagfors
NR
NR
Integrated filter samples with a dichotomous virtual impactor to separate PM10.25 from PM25
PM2.5
PMio-2.s, PM2.5
PM10-25, PM25
BS, S, Cl, K, Ca, Mn, Fe, Cu, In, Br, Rb, Pb
Wood burning made statistically significant contributions to personal exposure to K, Ca, and Zn. Cl, Mn, Cu, Rb, Pb, and BS
were found to be potential personal exposures from wood smoke, but their association was not always statistically significant.
S had no significant association with personal exposure to wood smoke.
Molnar et al. (2006,156773)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Cross-sectional
Autumn and spring in 2002 and 2003
Goteborg, Sweden,
Persons living in urban settings
20 subjects 20-50 yr randomly selected from the population and 10 from departmental colleagues.
NR
Integrated filter samples with cyclones for PM25 and PM, cut points
PM25andPM1
NR
NR
S, Cl, K, Ca, Ti, V, Mn, Fe, Ni, Cu, Zn, Br, Pb
Personal exposure to Cl, K, Ca, Ti, Fe, and Cu in PM2.5 were significantly higher than outdoor and central site ambient
concentrations, and personal exposure to Cl, Ca, Ti, Fe, and Br were also significantly higher than indoor levels. In most
cases, indoor concentrations were not higher than outdoor concentrations.
Na and Cocker (2005,156790)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Human exposure assessment
Sept. 2001-January 2002
Mira Loma, CA
Residential homes and a high school
NR
Indoor EC (elemental carbon) concentrations primarily of outside origin; Indoor PM2 5 significantly influenced by indoor OC
(organic carbon) sources, including indoor smoking.
Integrated filter samples for PM25
PM2.5
NR
PM2.5
EC.OC
Indoor PM25 was significant influenced by indoor OC sources. Indoor EC sources were predominantly of outdoor origin.
December 2009
A-306
-------
Naumova et al. (2003, 089213)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
RIOPAStudy-PAH partitioning indoor and outdoor pollutants to evaluate the hypothesis that outdoor air pollution contributed
strongly to indoor air pollution.
July 1999-June 2000
Los Angeles, CA, Houston, TX, Elizabeth, NJ
Houses
NR
NR
Modified MSP Samplers, 37 mm quartz filter
PM2.5
PM2.5
PM2.5
OC, EC
Both EC and OC were associated with gas/particle partitioning of PAHs, with EC being a better predictor. High correlation
between EC and OC suggests that PAHs adsorb onto PM containing EC during combustion.
Nerriere et al. (2005, 0894811
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Copollutant(s)
Primary Findings
Exposure assessment with stratified sampling of children and adults in 3 environments: high traffic emissions, local industrial
sources, and urban background.
"Hot" season May-June and "cold" season Feb-Mar. Grenoble in 2001, Paris in 2002, Rouen in 2002-2003, Strasbourg 2003.
Grenoble, Paris, Rouen, and Strasbourg, France
Persons living, working, or going to school in 3 urban areas one highly exposed to traffic emissions, one influenced by local
industrial sources, and a background urban environment. Industrial sources of pollution were present in each city.
6-13 yr and 20-71 yr. All non-smokers and not exposed to environmental tobacco smoke or industrial air pollution.
NR
Rucksack with Harvard ChemPass
PM25,PM10
NR
PM25,PM10
N02
The difference between ambient air concentrations and average total exposure is pollutant specific. PM2.5 and PM10
concentrations underestimate population exposures across almost all cities, season, and age groups, but the opposite is true
for N02.
Noullett et al (2006,155999)
Study Design Cohort
Period 5 February to 16 March 2001
Location Prince George, British Columbia
Population Children
Age Groups 10-12 yr
Indoor Source NR
Personal Method PM2.5 Harvard Personal Environment Monitors (HPEM2.5)
Personal Size PM2.5
Microenvironment Size NR
Ambient Size PM2.5
Component(s) S042", ABS (light absorbing carbon)
Primary Findings Thermal inversions were associated with personal exposures as well as ambient PM2 5 concentrations and likely caused
observed spatial variability. However, ambient sampling locations were correlated in time. Similar observations were made for
S042' and ABS.
December 2009
A-307
-------
Rojas-Bracho et al. (2004, 054772)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Cohort study with repeated measures.
Winter or summer of 1996-1997
Boston, Massachusetts
COPD patients
Adult
Housecleaning, cooking, transport in motor vehicles, low-effort home activities, moderate-effort home activities, activities in
public places, and resting or sleeping.
PEM
PM2.5, PM10, and PM10.2.5
PM25, PMlo, & PM10-25
NR
NR
During both seasons, personal exposures were higher than indoor or outdoor means, except during the winter when indoor
concentrations were higher than the personal or outdoor.
Rotko et al. (2002, 037240)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Copollutant(s)
Primary Findings
European multi-city air pollution study
Athens, Greece:26 January 1997-4 June 1998
Basel, Switzerland 3 February 1997-23 January 1998
Milan, Italy 10 March 1997-23 May 1998
Oxford, UK November 1998-7 October 1999
Prague, Czech Republic 3 June 1997-4 June 1998
Helsinki, Finland 26 September 1996-10 December 1997
Athens, Greece; Basel, Switzerland; Milan, Italy; Oxford, UK; Prague, Czech Republic; Helsinki, Finland
Adults
25-55 yr
NR
Integrated 48-h PM2 5 filter samples
PM2.5
PM2.5
PM2.5
N02
Personal PM2.5 and N02 levels were associated with subjects' level of annoyance. Highest annoyance levels occurred while in
traffic.
Sanderson and Farant (2004,156942)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Indoor and outdoor air monitoring of PAH. Investigate the relationship between indoor and outdoor PAH.
NR
Canada
Residential homes in neighborhoods around aluminum smelting plant
NR
NR
Indoor quartz filter sample
PM2.5
NR
NR
4-6 ring PAHs on indoor particle
Indoor concentration of 4-6-ring PAH were linked to outdoor industrial sources in residences without any major indoor source,
but with industrial facility as the main outdoor source. This study suggests that simultaneous measurements of indoor and
outdoor concentrations of PAH >4 rings predominantly associated with fine PM could provide useful estimates of particle
infiltration efficiency.
December 2009
A-308
-------
Sarnat et al. (2006, 089166)
Study Design
Period
Location
Population
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Outdoor-indoor pollutant infiltration, occupied residences
July28,2001-February25,2002
Los Angeles, CA
NR
Yes; cleaning, cooking, home ventilation (open windows/doors), kitchen fans, air conditioner/heating usage, number of
occupants, nearby roadways
NR
NR
PM2.5, Particle number
PM2.5
BC (nonvolatile component) ; N03 (volatile component)
Infiltration rate for PM2.5 was intermediate, while BC was highest and N03 lowest. Infiltration rate varied with particle size, air
exchange rate, outdoor N03. PM2.5 infiltration was lowest for volatile components. Outdoor volatile PM25 components may be
less representative of indoor exposure to volatile PM25 of ambient origin. Outdoor nonvolatile PM25 components may be more
representative of indoor exposure to nonvolatile PM25 of ambient origin.
Sarnat et al. (2006, 090489)
Study Design Personal and ambient exposure assessment
Period June 14-August 18 (summer); Sep 24-Dec 15 (fall), 2000
Location Steubenville, OH
Population Non-smoking, older adults
Age Groups NR
Personal Method Integrated filter gravimetric measurement
Personal Size PM2.5
Microenvironment Size NR
Ambient Size PM2.5
Component(s) S042"; EC
Primary Findings 24-h ambient measurements are more representative of personal particle exposure than gases, and ventilation is an important
exposure modifier.
Sarnat et al. (2005. 087531)
Study Design Time-series epidemiologic study
Period Summer 1999 and winter 2000
Location Boston, MA. Comparisons to a previous study in Baltimore are also made.
Population School children and seniors
Age Groups NR
Indoor Source PM2.5
Personal Method NR
Personal Size PM2.5
Microenvironment Size NR
Ambient Size PM2.5
Component(s) S04,
Copollutant(s) 03, N02, S02
Primary Findings Substantial correlations between ambient PM25 concentrations and corresponding personal exposures. Summertime gaseous
pollutant concentrations may be better surrogates of personal PM25 exposures (especially personal exposures to PM25 of
ambient origin) than they are surrogates of personal exposures to the gases themselves.
Shalatetal. (2007. 156971)
Study Design Indoor home exposure assessment; sampling technology demonstration
Period Winter heating season
Location Residential home
Population Children
Age Groups Pre-toddler (6- to 1 2-month-old) children
Indoor Source NR
Personal Method Integrated filter and real-time nephelometer at floor height and at a height of 11 0 cm
Personal Size TSP, inhalable PM
Microenvironment Size NR
Ambient Size NR
Copollutant(s) NR
Primary Findings The study results suggest that young children are exposed to more inhalable PM and TSP because PM becomes resuspended
from the floor with motion.
December 2009
A-309
-------
Shao et al. (2007,156973)
Study Design Exposure assessment
Period July and Winter 2003
Location Beijing, China
Population General population
Age Groups NR
Indoor Source NR
Personal Method PMio measured with integrated filter samples
Personal Size PMio
Microenvironment Size PMio
Ambient Size PM10
Component(s) NR
Primary Findings Plasmid scission assay, coupled with the image analysis, can be used to evaluate the relationship between particle physico-
chemistry and toxicity.
Shilton et al. (2002. 049602)
Study Design Respirable particulates inside and outside of a building were collected and compared
Period 24-h sampling from 12:45 pm Mondays to Fridays between 9/19/00 to 5/01/01
Location Wolverhampton city center, University of Wolverhampton, UK
Population NR
Indoor Source Mn,AI, N03, CI" (wind-blown dust), Cu and Zn"
Personal Method Active sampling using Casella sampler (filter)-
Personal Size Respirable PM
Microenvironment Size Respirable PM
Ambient Size Respirable PM
Component(s) N03", metals (Zn, Cu, Mn, Al), S042", CI"
Primary Findings The indoor particulate concentration was driven by ambient concentration, and meteorological-induced changes in ambient PM
were detected indoors.
Strand et al. (2007,157018)
Study Design
Period
Location
Population
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Cohort
Winter of 1999-2000 and winter of 2000-2001
Denver, Colorado
Asthmatic children
NR
Modeling/extrapolation from fixed-site ambient monitoring (multiple methods)
NR
NR
PM2.5
NR
Using modeled or extrapolated personal ambient PM exposure results in a deattenuation of decrements in FEV, associated with
PM exposure, relative to use of fixed-site ambient monitoring PM levels. Associations between FEV1 decrements and the
various estimation procedures (modeling and extrapolation) were similar to each other.
Tang et al. (2007, 091269)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Cohort Study
12/2003-2/2005
Sin-Chung City, Taiwan
Asthmatic children
6-12 yr
No
Portable particle monitor; DUSTcheck Portable Dust Monitor, model 1.108, GRIMM Labortechnik Ltd., Germany
PM10,PM25, PM^PM^s, PM25-i
NR
PM10,PM25, PMio-25-
NR
Results of linear mixed-effect model analysis suggested that personal PM data was more suitable for the assessment of change
in children's PEFRthan ambient monitoring data.
December 2009
A-310
-------
Thornburg et al. (2004,157052)
Study Design PM exposure studies
Period RTP: Summer 2000-spring 2001
Tampa: October-November 2002
Location Research Triangle Park (RTP), NC and Tampa, FL
Population Residential home occupants
Age Groups NR
Indoor Source Resuspension of PM10 from a carpet and cooking
Personal Method Harvard impactors and PEMs, ME pdMOOO nepholometer
Personal Size PM2.5, PM10
Microenvironment Size NR
Ambient Size PM2.5, PM10
Component(s) NR
Primary Findings The association of duty cycle with indoor-outdoor (I/O) ratio was confounded by the short time span of ventilation system
operation and the presence of strong indoor sources.
Toivola et al. (2002, 0265711
Study Design Random sample of teachers
Period Nov 1998-Mar 1999 and Nov-Dec 1999
Location 2 cities in eastern Finland
Age Groups Adult
Indoor Source Fungi, bacteria
Population Elementary school teachers
Personal Method Button inhalable aerosol sampler
Personal Size Particle Mass; BS
Microenvironment Size Particle Mass; BS
Ambient Size NR
Component(s) Total fungi, total bacteria, viable fungi, viable bacteria
Primary Findings Personal BS exposure correlated with both home and work BS exposures. BS concentrations explained best the variation of
particle mass in personal and home concentrations.
Trenga et al. (2006,155209)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Panel study with repeated measures
3 sampling periods Oct 1999-Aug 2000, Oct 2000-May 2001, Oct 2001-Feb 2002
Seattle, Washington
Adults with and without COPD and children with asthma
adults ages 56-89 and children ages 6-13
NR
Carrying personal monitor (Harvard Personal Environmental Monitor for PM25)
PM2.5
PM2.5
PM10-2s, PM25for residential outdoor, PM25 for central site
NR
FEVi decrements associated with 1 -day lagged central site PM2.5 in adult subjects with COPD. Associations between PM and
lung function decrements were significant only in asthmatic children not receiving anti-inflammatory medication.
Turpin et al. (2007,157062)
Study Design
Period
Location
Population
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
RIOPAStudy 24-h integrated indoor, outdoor, and personal samples collected in 3 cites.
Summer 1991-spring 2001
Elizabeth, NJ, Houston, TX, and Los Angeles County, CA
309 adults and 118 children (89-18)
NR
PEM on the front strap of a harness near the breathing zone. The bag on the hip contained the pump, battery pack, and motion
sensor
PM2.5
PM2.5, in the main living area (not kitchen)
PM2.5, in the front or backyard
18 volatile organics, 17 carbonyl, PM2.5 mass and >23 PM2.5 species, organic carbon, elemental carbon, and PAHs
The best estimate of the mean contribution of outdoor to indoor PM2 5 was 73% and the outdoor contribution to personal was
26%.
December 2009
A-311
-------
Vallejo et al. (2006,157081)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Panel study
4/2002-8/2002
Mexico City, Mexico
Health young, non-smoking adults
Mean age 27 yr
NR
pDR nephelometric method
PM2.5
NR
NR
NR
The descriptive analysis showed that overall outdoor median concentration of PM2.5 was higher than the indoor concentration.
In the indoor microenvironment, the highest concentrations occurred in the subway followed by the school, and the lowest was
at home. The outdoor microenvironment with the highest concentrations was the public transportation (bus), while the
automobile had the lowest. It was found that PM2.5 concentration levels had a circadian-like behavior probably related to an
increase in the population daily activities during the morning hours, which decrease in the evening, especially at indoor
microenvironments. The Center city area was found to have the highest concentrations of PM2.5.; Multivariate analysis
corroborated that PM25 concentrations are mainly determined by geographical locations and hour of the day, but not by the
type of microenvironment.
van Roosbroeck et al. (2006, 090773)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Personal exposure assessment, effect of traffic-related pollutants
March-June 2003
Amsterdam, The Netherlands
Schoolchildren
9-12 yr
ETS, cooking
Integrated filter gravimetric measurement. Light absorbance.
PM2.5
NR
PM2.5
Absorbance
Children living near busy roads had 35% higher personal exposure to 'soot' than children living in urban background locations.
Vinzents et al. (2005, 087482)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Primary Findings
Panel study
3/2003-6/2003
Copenhagen, Denmark
Healthy young adults
Mean age = 25 yr
No
Condensation particle counters
UFP(10-100nm)
UFP(10-100nm)
PM10
UFP exposure predicted oxidative DNA damage.
Wallace and Williams (2005, 057485)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Cohort
2000-2001
Raleigh, North Carolina
African-American persons with elevated risk from exposure to particles.
NR
NR
PEM PM2 5 monitor
PM2.5
Indoors PM2.5
Outdoors near residence PM25 PM25
S
Using outdoor particles to determine the effect on health is not accurate. The infiltration factor is a good estimator for personal
exposure. Indoor and outdoor measurements of sulfur could be used in the absence of personal exposure measurement to
estimate the contribution of outdoor fine particles to personal exposures.
December 2009
A-312
-------
Wallace et al. (2006, 088211)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Time series continuous monitoring of subjects with controlled hypertension or implanted defibrillators were monitored for 7
consecutive days in 4 seasons.
2000-2001
North Carolina
Health-compromised adults, non-smokers
Adults
Cooking, cleaning, personal care, smoking
PEM
PM2.5
PM2 5; Indoor and outdoor
NR
NR
Use of continuous particle measuring instruments allowed more precise identification of sources, frequency and magnitude of
short-term peaks, and more accurate calculation of individual personal clouds.
Wang et al. (2006,157108)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Exposure assessment, identification of sources of outdoor and indoor PM and trace elements
Aug4-Sep10,2004
Guangzhou, China
4 hospitals
NR
NR
No personal exposure assessment was conducted.
NR
PM10,PM25
PM10,PM25
Na, Al, Ca, Fe, Mg, Mn, Ti, K, V, Cr, Ni, Cu, Zn, Cd, Sn, Pb, As, Se
High correlation between PM2.5 and PMio suggest that they came from similar emission sources. Outdoor infiltration could lead
to direct transportation of PM indoors. Human activities and ventilation types could also influence indoor PM. levels.
Ward et al. (2007,157112)
Study Design
Period
Location
Population
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Indoor air sampling to determine size fractionated concentrations of PM, OC, EC, and total carbon
Jan-Mar 2005
Libby, Montana
Children exposed to wood-burning stoves in elementary and middle schools
Burning wood in stoves for heating
NR
NR
PM >2.5,1.0-2.5, 0.5-1.0, 0.25-0.5, and < 0.25 urn
PM >2.5,1.0-2.5, 0.5-1.0, 0.25-0.5, and < 0.25 urn
OC and EC
Total measured PM mass concentrations were much higher inside the elementary schools, with particle size fraction (>2.5, 0.5-
1.0, 0.25-0.5, and < 0.25 mm) concentrations between 2 and 5 times higher when compared to the middle school. The 1.0-2.5
mm fraction had the largest difference between the two sites, with elementary school concentrations nearly 10 times higher
than the; middle school values.
Weisel et al. (2005,1571311
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Matched indoor, outdoor, and personal concentrations in proximity to pollution sources.
May 1999-Feb 2001
Elizabeth, NJ, Houston, TX, and Los Angeles County, CA
urban children and adults
Children and adults (6-89 yr)
Age of house, recent renovations (< 1 yr), type of home (single, multiple family), attached garage, carpet indoors, local pollution
sources.
PEM on a harness with inlet near breathing zone.
PM2.5
PM2.5
PM2.5
NR
Personal PM2.5 was significantly higher than indoor and outdoor PM2.5 concentrations.
December 2009
A-313
-------
Wichmann et al. (2005, 086240)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Exposure assessment
November 29, 1993-March 30, 1994; October 17, 1994-December 22, 1994
Amsterdam, The Netherlands
Adults and schoolchildren living near high-traffic or low-traffic roads
Adults (50-70 yr), schoolchildren (10-1 2 yr)
NR
Personal impactor
PM10
Absorbance coefficient measurements
Found tentative support for using type of road as a proxy for indoor and personal exposure to traffic-related absorbance PM.
Williams et al. (2003. 053338)
Study Design Cohort study, longitudinal
Period Summer 2000, fall 2000, winter 2001 , and spring 2001
Location Raleigh and Chapel Hill, North Carolina
Population Elderly persons
Age Groups > 50 yr
Indoor Source Occasional ETS
Personal Method Integrated filter samples
Personal Size PM2.5
Microenvironment Size PM2.5; PMi0; PM^-is
Ambient Size PM2.5; PM10; PM10-2.5
Component(s) NR
Primary Findings When comparing cohorts, there was no statistically significant difference between PM2.5 exposure. Little spatial variability was
observed regarding PM2.5 concentrations; this was observed to a lesser extent for PM^ as well.
Wilson and Brauer (2006, 088933)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Exposure assessment
April-September 1998
Vancouver, Canada
Subjects with physician-diagnosed COPD
54-86-years-old
No
Personal integrated filter gravimetric measurement; TEOM outdoor ambient
PM2.5
NR
NR
S042'
It was observed that ambient PM2.5 exposure, estimated with the S042" method, accounted for 71 % of measured ambient
concentration and 44% of measured total personal exposure. No correlation between nonambient exposure and ambient
concentration was observed.
Wu et al. (2006,179950)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Panel study
9/3/2002-11/1/2002
Pullman, WA
Asthmatic adults
18-52yr
No
Co-located Harvard Personal Environmental Monitors (HPEM2.5; Harvard School of Public Health, Boston, MA)\
PM2.5
PM2.5
PM2.5
Levoglucosan (LG); Elemental Carbon (EC); Organic Carbon (OC)
The authors observed significant variability between subjects for burning and nonburning episodes. The authors postulated that
activity patterns contribute to this variability and that central-site measurements of LG might not be a good surrogate for
biomass combustion smoke exposureforthis reason.
December 2009
A-314
-------
Wu et al. (2005, 086397)
Study Design Panel study
Period 1999-2000
Location Alpine, CA
Population Asthmatic children
Age Groups 9-17 yr
NR
pDR
PM2.5
PM2.5
PM2.5
NR
Primary Findings Personal exposure was higher than those affixed sites. Subjects received only 45.0% of their exposure indoors at, although
they spent more than 60% of their time there. In contrast, 29.2% of their exposure was received at school where they spent
only 16.4% of their time. Thus, exposures in microenvironments with high PM levels where less time is spent can make
significant contributions to the total exposure.
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Yeh and Small (2002, 040077)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Comparative assessment of AME and IES models
1997 (364 days) spring March-May, summer June-August, Fall September-November, winter December-February
Los Angeles County, CA
General population; ETS and non-ETS Homes
NR
Indoor Cooking, ETS, Other sources and unexplained particulates that maybe generated with engaging in various activities
NR
PM10PM25
NR
PM10PM25
NR
Adjusting from outdoor concentrations to personal exposures and correcting dose-response bias produce nearly equal results.
Roughly the same premature mortalities associated with short-term exposure to both ambient PM2.5 and PM10 are predicted by
both models
Yip et al. (2004,157166)
Study Design
Period
Location
Population
Age Groups
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
A panel study with repeated measures with personal & home monitoring for 8 2-week Periods. Children were stratified into
smoking and non-smoking households.
2000-2001
Detroit, Michigan
School-age children with asthma
7-11 yr
PEM in a backpack
PM10
PMi0; indoor at home & indoor at school
PM10
NR
Personal PM concentrations were significantly correlated with home environment (r = 0.38 to 0.70), with the strongest
relationships in home with non-smokers.
December 2009
A-315
-------
Zhao et al. (2006,156181)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Aerosol source apportionment under four environments (personal, residential indoor, residential outdoor and ambient) to
evaluate the relationship between different environments through exposure analysis, and to demonstrate the utility of the
combined receptor model on air quality studies of various environments.
June 2000 to May 2001
Raleigh and Chapel Hill, NC
NR. People with respiratory ailments most likely.
NR
4 main sources to residential indoor PM Cu-factor mixed with indoor soil, secondary sulfate, Personal care and activity, ETS and
its mixture
PEMandHI
NR
NR
NR
S042", OC, EC, and trace elements
Secondary S042" and vehicle emissions were significant contributors of personal PM exposure and residential indoor PM
concentrations.
Zhao et al. (2007,156182)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Component(s)
Primary Findings
Comprehensive analysis of the sources of PMi5 exposure on children with moderate to severe asthma in urban-poor settings.
Two winter periods (October 2002-March 2003 and October 2003-March 2004)
Elementary school for children with significant asthma, Denver, CO
Schoolchildren in urban-poor settings suffering from moderate to severe asthma
6-13 yr (60% in the range 10-13 yr, rest in the range 6-9 yr)
Yes, House cleaning compounds, and smoking were identified as primary internal sources.
PEM
PM2.5
PM2.5
PM2.5
EC.CI.Si, N03
Four external sources and three internal sources were resolved in this study. Secondary nitrate and motor vehicle were two
major outdoor PM2.5 sources. Cooking was the largest contributor to the personal and indoor samples. Indoor environmental
tobacco smoking also has an important impact on the composition of the personal exposure samples.
Zhu et al. (2005,157191)
Study Design
Period
Location
Population
Age Groups
Indoor Source
Personal Method
Personal Size
Microenvironment Size
Component(s)
Primary Findings
4 apartments near the freeway were monitored at 2 times for 6 consecutive days, 24 h per day. Subjects did not enter the
bedrooms where the samplers were, no cooking, cleaning, children, or pets.
Oct. 2003-Dec. 2003 and Dec. 2003-Jan. 2004
Los Angles, CA
Urban Populations near major freeways.
NR
NR
NR
Indoor and Outdoor ultrafine particles (6-220 nm)
NR
CO
The size distributions of indoor aerosols showed less variability than the adjacent outdoor aerosols. Indoor to outdoor ratios for
ultrafine particle concentrations depended strongly on particle size. I/O ratios were dependent on the indoor ventilation
mechanisms applied. Size-dependent particle penetration factors and deposition rates were predicted from data by fitting a
dynamic mass balance model.
December 2009
A-316
-------
Zollner et al. (2007,157192)
Study Design
Period
Location
Population
Age Groups
Personal Method
Personal Size
Microenvironment Size
Ambient Size
Primary Findings
Exposure assessment
Winter Period of 2005 and 2006
Baden-Wuerttemberg, Germany
School children
NR
NR
NR
They only reported concentrations for PM2.5. PM ranging in size from 0.02 to >20 |im were collected and analyzed but only
PM2.s concentration were reported.
They only reported concentrations for PM2.5. PM ranging in size from 0.02 to >20 |im were collected and analyzed but only
PM2.s concentration were reported.
The impaction of PM was strongly influenced by specific weather conditions. Time resolution of measurements in classrooms
showed variation in particle concentration depending on the type of building and indoor activities. E'[Concentrations of very
small particles indoors and in ambient air measured by condensation particle counter were influenced by traffic emissions.
Table A-59. Examples of studies showing developments with UFP sampling methods since the 2004
PMAQCD.
Reference
Biswas et al.
(2005, 150694)
Feldpausch et
al. (2006,
155773)
Heringetal.
(2005, 155838)
Hermann etal.
(2007, 155840)
Kinsey et al.
(2006, 130654)
Kulmala et al.
(2007, 097838)
Kulmala et al.
(2007, 155911)
PM Size PM
Ranges Constituents
20-100 Carbonaceous
nm aerosols
3-40 nm Ag, NaCI
10nm-5 nF
urn
i in n™ Atmospheric
2-20nm aerosol, Ag
Instruments
CPC (water)
DS with CPC, compared with
DMA
CPC (water)
CPC (water and butanol)
TEOM, SMPS, CPC, DustTrak, E-
BAM.ELPI, integrated filter
samples
CPC
Battery of CPCs (water, butanol,
n-butanol)
Study Description
Water-based CPC performance evaluation.
The DS with CPC compared fairly well with the DMA for particle sizes
up to 40 nm with 20-40% underestimation depending on discharge
frequency settings. The DS sampling period is 3-5 s in comparison with
the 1 min scanning time of the DMA.
Water-based CPC performance evaluation.
Roughly 95% collection efficiency for d >5 nm for TSI models 3776 and
3786, 95% efficiency for d >20 nm for model 3775, near 90% efficiency
for d>20 nm for model 3785, near 90% efficiency for d >25 nm for
model 3772.
TEOM compared best with gravimetric filter among mass concentration
analyzers, ELPI and SMPS comparable for differential number
distribution but ELPI not useful for gravimetric analysis because mass is
not significant at small end of distribution.
Changing temperature difference between saturator and condenser
within CPC allowed for differences in cut-off diameters.
Used the battery to discriminate between water-soluble, water-insoluble,
butanol-soluble, and butanol-insoluble nucleation-mode particles.
Ntziachristos
and Samaras
(2006,116722)
7 nm-1
urn
Automobile
exhaust
5 instruments used
simultaneously to reduce
uncertainty: Teflon-coated filter
downstream of constant volume
sampling, ELPI with
thermodenuder, CPC, SMPS,
diffusion charger
Use of four reduced variables combining output from all instruments
(ratio of particle number concentration from CPC and ELPI, estimated
mean geometric mobility diameter from signal of diffusion charger and
number concentration from CPC, ratio of signal of diffusion charger to
constant volume sampler mass, ratio of constant volume sampler mass
to volume collected by ELPI) resulted in identification of clear outliers
and factors related to driving and fuel properties rather than
measurement errors.
Olfert et al.
(2008,156004)
30-100
NaCI, ambient FIMS (compared with SMPS)
Particle number concentrations reported by the FIMS were 8-23%
higher than the SMPS using an inversion technique designed to correct
for particle residence time in the FIMS, which operates at 0.1 s
resolution.
Petaja et al.
(2006,156021)
CPC (water)
Water-based CPC performance evaluation.
Winkleretal.
(2008,156160)
1.5-4 nm Tungsten oxide CPC (n-Propanol)
Authors remove excess charge on particles with ion trap to detect
particles down to ~ 1 nm (by eliminating electrostatic attraction to
agglomerate).
December 2009
A-317
-------
Table A-60. Summary of in-vehicle studies of exposure assessment.
Reference
Briggs et al.
(2008, 156294)
Diapoulietal.
(2007, 156397)
Fruin et al. (2008,
097183):
Westerdahletal.
(2005, 086502)
[Note: same data
presented.]
Gomez-Perales et
al. (2004,
054418)
Study Design Mode of Transport
UFP (P-Trak), PM,o, PM25, and PM, Car
(OSIRIS light scatter) were operated in a
car while driving or walking on one of 48 Walking
routes in London. Trips ranged 1.5-15 min
by car and were repeated up to 5 times to
improve statistics.
Study Period: Weekdays in May and June
2005.
UFP (CPC) concentrations were Car
measured at school, residential, and in-
vehicle environments in Athens, Greece.
Study Period: school hours, Nov
2003-Feb 2004 and Oct-Dec 2004
On-road zero emissions vehicle driven on Car
33-mi arterial road and 75-mi freeway
measured UFP (CPCs, SMPS, BAD), BC
(aethalometer), NOX (chemiluminescence),
PM-bound PAHs (UV-photoionization), and
CO (Q-Trak). DVD analysis of traffic
density and car speed.
Study Period: Feb-Apr 2003 for 2- to 4-h
periods.
PM2.s (personal filter pump), CO (T15 Bus
electrochemical cell), and benzene
(canister) were measured on transit Minibus
routes, and PM2 5 filters were analyzed .. ,
for mass, OC/EC, S042", N03-, and trace lvlelro
metals.
Exposures
Units: PMrPMmfug/m0),
UFP (pern"3)
Avg Car Exposure:
PM105.87 (3.09)
PM25 3.01 (1.10)
PM,1.82(1.10)
UFP 21639 (14379)
Avg Walking Exposure:
PM1027.56(13.16)
PM25 6.59(3.12)
PM,3.37 (3.40)
UFP 30334 (17245)
15-min median
(1000p/cm3):
School indoor 13.6
School outdoor 16.6
Residence indoor11.2
Residenceoutdoor24.0
In-vehicle 78.0
Arterial range of
medians:
UFP (1000p/cm3) 13-43
PM2.5(ug/m3)7.9-45
BC(ug/m3) 0.74-3.3
Freeway range of
medians:
UFP (1000p/cm3) 47-190
PM25(ug/m3)25-110
BC(ug/m3)2.4-13
PM2.5(ug/m3):
Bus 68
Minibus71
Metro 61
Primary Findings
I n-car concentrations of PM2 5, PMi, and UFP
correlated well with walking concentrations (R = 0.806,
0.800, 0.799 respectively). Avg walking concentrations
were 1.4-4.7 times higher than average in-car
concentrations. Cumulative walking exposures (not
shown here) were 4.4-15.2 times higher than those in
cars, likely resulting from longer transit times for
walking.
In-vehicle UFP concentrations were roughly 3.5-7
times higher than school or residence
concentrations. Indoor concentration diel patterns
were also shown to follow outdoor levels, which
suggests that indoor levels are of outdoor origin.
Measurements of freeway UFP, BC, PM-bound PAH,
and NOK concentrations were roughly one order of
magnitude higher than ambient measurements.
Multiple regression analysis suggests these
concentrations were a function of truck density and
total truck count. (Only PM measurements reported
here).
Generally, PM2.5 concentration was higher in the
morning than evening rush hour, but variability was
higher for minibuses than other modes of transport.
Wind speed was found to be associated with PM2.5
concentration on minibuses.
Study period: 3-h morning and evening
rush hour May-June 2002
December 2009
A-318
-------
Reference
Gomez-Perales et
al. (2007,138816)
Gulliver and
Briggs (2004,
053238)
Gulliver and
Briggs (2007,
155814)
Study Design Mode of Transport
PM2.5 (personal filter pump), CO (T15 Bus
electrochemical cell), and benzene
(canister) were measured on transit Minibus
routes, and PM2 5 filters were analyzed ., ,
for mass, OC/EC, S042", N03-, and trace lvlelro
metals.
Study period: 3-h morning and evening
rush hour Jan-March 2003
PMio, PM2.5, and PMi sampled (OSIRIS Car
light-scatter devices) in a car while
driving or walking on northern corridor of ™lk
Northhampton UK.
Study Period: 1-h interval of morning and
evening rush hour during Winter
1999-2000.
TSP, PM10, PM2.5, and PM, sampled Car
(OSIRIS light-scatter devices) in a car
while driving or walking on one of 48 ™lk
routes in London. Trips ranged 1 .5-15
min by car and were repeated up to 4
times to improve statistics.
Exposures
Units: PM2.5 mass
(ug/m ), components (%
of mass)
Bus:
PM25 20-58
(NH4035-8
(NH4)2S0410-18
OC 17-39
EC 8-20
Crustal15-18
Non-crustal 2-3
Unknown 6-24
Minibus:
PM25 25-55
(NH4034-13
(NH4)2S047-22
OC22-37
EC 9-1 9
Crustal12-13
Non-crustal 3-3
Unknown 4-26
Metro:
PM25 24-41
(NH4035-8
(NH4)2S0410-21
OC 35-42
EC 9-1 3
Crustal10-16
Non-crustal 2-4
Unknown 5-20
Walking concentrations,
Units: ug/rn
Walk, Car, Background
PM,0 38.2,43.2, 26.6
PM2515.1,15.5
PM, 7.1, 7.0
Mean concentrations,
Units: ug/rn
Walk, Car, Background
TSP-PM10 19.1 (19.8)
18.2(18.0)4.9(5.1)
Primary Findings
Buses and minibuses had similar concentration
levels for PM2.5 mass, and metro exposures were
lower. CO and benzene concentrations were higher
on minibuses than buses. OC was the largest PM
constituent for all modes of transport. Measured
concentrations were higher in the morning than in
the evening rush hour periods. Maximum historical
wind speeds (1995-2003) appeared to be inversely
associated with measured concentration.
In-car PM^ concentrations were elevated
compared with walking and background. PM2.5 and
PM, concentrations were comparable for walking
and background. Periods of elevated PM25
compared with PM10 generally corresponded to
times when S042~ levels were also high.
Walking exposures larger than car and
background, and car exposures were generally
larger than background except for PMi. Peak
exposures during walking were significantly higher
than peak in-car exposures.
Study Period: Jan-Mar 2005.
PM10-2522.1 (22.8)15.1
(14.2)10.0(9.0)
PM25 -1 10.9(10.4)8.3
(8.4)7.6(7.1)
PM, 4.8 (3.4)2.9 (2.6)4.2
(2.4)
Rossneretal. Measured PM25 exposure of 50 city bus Bus
(2008,156927) drivers and 50 controls in Prague, Czech
Republic using personal samplers (type
not specified) and VOCs using passive
samplers. PM2.5 filters analyzed for c-
PAHs. Focus of study is oxidative stress
biomarkers in drivers.
Study period: winter 2005, summer
2006, winter 2006.
Units: ng/m
Winter 2005: Bus Control
C-PAH7.1 (3.7)9.4 (5.5)
B[a]P1.3(0.7)1.8(1.0)
Smmer 2006: BusControl
C-PAH1.8 (0.5)2.0 (0.8)
B[a]P0.2 (0.1)0.3 (0.2)
Wnter 2006: Bus Control
C-PAH5.4 (3.5)4.1 (1.7)
B[a]P1.0(0.5)0.8 (0.4)
c-PAH and B[a]P exposure to bus drivers was
significantly higher in Winter 2006, but control
exposure was significantly higher in Winter 2005
for c-PAH and B[a]P and in summer 2006 for c-
PAH. No significant difference in VOC exposure
between bus drivers and controls was observed.
Oxidative stress markers were significantly higher
in bus drivers than controls for all seasons.
December 2009
A-319
-------
Reference
Sabinetal.(2005,
088300)
Study Design
BC (aethalometer), particle-bound PAH
(UV-photoionization), and NO (luminol
reaction) were measured on 3 diesel
school buses, 1 diesel school bus with a
particle trap, and one compressed gas
bus during before- and after-school
commutes.
Study Period: May-June 2002.
Mode of Transport
School bus (diesel, diesel
with particle trap (TO),
compressed gas (CNG))
Exposures
In-bus mean
concentration
Units: BC(ug/m3)
PAH (ng/m5)
Windows closed:
Dp DflU
bU rAH
Ambient: 2.5,27
CNG:2.3,57
10:7.1,190
Diesel:11,290
Windows open:
BC PAH
Ambient: 1.9,26
CNG:1.5,43
10:2.3,42
Diesel: 3.9,58
Primary Findings
Mean concentrations on diesel buses without
newer emissions control technologies were 2-4.4
times higher than background. On buses with
particle traps, concentrations were 1.2-2.5 times
higher than background, while concentrations on
compressed gas-fueled school buses were actually
lower than background.
Table A-61. Summary of personal PM exposure studies with no indoor source during 2002-2008.
Reference / Location
Personal
Micro
Ambient
SOUTHWEST
Delfino et al. (2004, 056897)
Alpine, California
Method: pDR, Units = |ig/m°
Last 2-hPM2 5 34.4 (33.7)
Diurnal PM2 5 55.7(31.6)
Nocturnal PM25 22.3 (13.6)
1-h max PM25151.0(120.3)
4-hmaxPM2587.5(55.3)
8-hmaxPM2567.6(39.0)
24-hPM25 37.9 (19.9)
Method: HI, Units =
lndoor24-hPM1030.3(11.9)
Indoor 24-hPM2 5 12.1 (5.4)
Outdoor 24-hPM1025.9 (10.4)
Outdoor 24-hPM2511.0(5.4)
Method: TEOM, Units = |ig/m°
Diurnal PM1035.1 (11.3)
Nocturnal PM10 23.3 (8.4)
1-h max PM10 54.4 (13.8)
4-h max PM10 44.5 (12.4)
8-hmaxPM1039.8(11.2)
24-hPM10 23.6 (9.1)
24-h PM25 10.3 (5.6)
Delfino et al. (2006, 090745)
Riverside and Whittier, California
Method: PEM, Units = ng/nf
Riverside:
n13
24-hPM2532.78(21.84)
1-h max PM25 97.94 (70.29)
8-h max PM2.5 47.21 (30.0)
Whittier:
n32
24-hPM2536.2(21.84)
1-hmaxPM2593.63(75.19)
8-h max PMis 51.75 (36.88)
Method:FRM, Units = |ig/m°
Riverside:
24-hPM25 36.63 (23.46)
24-h PM10 70.82 (29.36)
Whittier:
24-h PM25 18.0 (12.14)
24-h PM10 35.73 (16.6)
Turpin et al. (2007,157062)
Los Angeles County, CA (and Elizabeth,
NJ, Houston, TX)
Method: PEM, Units = |ig/m0
Avgof48-h PM25
Child 40.2
Adult 29.2
Method: HI, Units = |ig/m0
Avgof48-hPM2.5:16.2
Method: HI, Units = |ig/m°
Avgof48-hPM2.5:19.2
Wu et al. (2005,157155)
Alpine, CA
Method :pDR, Units = |ig/m°
n11
Avgof24-hPM25 11.4(7.8)
Method :pDR, Units = ng/m°
n14
Avgof24-hPM2.55.6(2.9)
Method: HI
n14
Avgof24-hPM259.8(2.5)
Method :pDR, Units = ng/m°
n8
Avgof24-hPM2.5 14.0(11.4)
Method: HI
n8
Avgof24-hPM2.5 14.3(7.8)
December 2009
A-320
-------
Reference / Location
Personal
Micro
Ambient
NORTHWEST
Jansenetal. (2005. 082236)
Seattle, Washington, USA
NR
Method: HI, Units = ng/m°
Indoor home:
PM1011.93
PM2.5 7.29
Outdoor home:
PM1013.47
PM;.5 10.47
Method: HI, Units = |ig/m°
PM1018.0
PM25 14.0
Koenig et al. (2003,156653)
Seattle, WA
13.4 ±3.2 ug/m°
Inside homes = 11.1 ±4.9
Outside homes = 13.3 ±1.4
3 Central-sites = 10.1 ±5.7
Liu Setal. (2003.073841)
Seattle, WA
Summary of PM concentrations
(ug/m3) between October 1999 and
May 2001 by study group.
Group Mean ± SD Personal PM25
COPD 10.5 ± 7.2 Healthy 9.3 ± 8.4
Asthmatic 13.3 ± 8.2 CHD 10.8 ± 8.4
Summary of PM concentrations (ug/m")
between October 1999 and May 2001 by
study group.
Group Mean ±SD
Indoor
PM2.5
COPD 8.5 ±5.1
Healthy 7.4 ±4.8
Asthmatic 9.2 ±6.0
CHD 9.5 ±6.8
PM10
COPD 14.1 ±6.6
Healthy 12.7 ±7.8
Asthmatic 19.4 ±11.1
CHD 16.2 ±11.3
Summary of PM concentrations (ug/m")
between October 1999 and May 2001 by
study group.
Location Pollutant
Group Mean ±SD
Outdoor PM2 5
COPD 9.2 ±5.1
Healthy 9.0 ±4.6
Asthmatic 11.3 ±6.4
CHD 12.7 ±7.9
PM10
COPD 14.3 ±6.8
Healthy 14.5 ±7.0
Asthmatic 16.4 ±7.4
CHD 18.0 ±9.0
Mar etal. (2005.087566)
Seattle, WA USA
Method: HI, Units = |ig/m°
PM2.5:
Healthy: 9.3 (8.4)
CVD: 10.8(8.4)
COPD: 10.5 (7.2)
Method: HI, Units = |ig/m0
PM2.5:
Healthy: 7.4 (4.8)
CVD: 9.5 (6.8)
COPD: 8.5 (5.1)
PM10:
Healthy: 12.7(7.8)
CVD: 16.2(11.3)
COPD: 14.1 (6.6)
Method: HI, Units = |ig/m°
PM2.5:
Healthy: 9.0 (4.6)
CVD: 12.7 (7.9)
COPD: 9.2 (5.1)
PM10:
Healthy: 14.5(7.0)
CVD: 18.0(9.0)
COPD: 14.3 (6.8)
Trenga et al. (2006,155209)
Seattle, Washington
Method: PEM, Units = |ig/m°
Median PM25
Child 11.3
Adult 8.5
Method: HI, Units = |ig/m°
Median PM25
Child 7.5
Adult 7.6
Method: HI, Units = |ig/m°
Residential Outdoor
Median PM2 5
Child 9.6
Adult 8.6
Residential Outdoor
Median PMcoarse
Child 4.7
Adult 5.0
Residential Outdoor
Median PM2 5 central site
Child 11.2
Adult 10.3
Wu et al. (2006,179950)
Pullman, WA
During non-burning times: 13.8 (11.1)
During burning episodes: 19.0 (11.8)
SOUTHCENTRAL
Turpin et al. (2007,157062)
Houston (and Elizabeth, NJ, and Los
Angeles County, CA)
Houston, Units = ng/m°
Child: 36.6
Adult: 37.2
avg)
Houston: 17.1
Houston: 14.7
December 2009
A-321
-------
Reference / Location
Personal
Micro
Ambient
MIDWEST
Adgate et al. (2002, 030676)
Battle Creek, East St. Paul, and Phillips,
Minnesota, constituting the Minneapolis-
St. Paul metropolitan area.
PM2.5, Units = |ig/m
Battle Creek
All Seasons: 118,22.7, (25.7), 16.2
(2.2)
Spring: 41, 26.3 (25.7), 19.4 (2.1)
summer: 31,28.5 (36.1), 20.3 (2.1)
Fall 46,15.5 (13.4), 11.9(2.1)
E. St. Paul
All Seasons: 107,30.5 (38.7), 20.6
(2.3)
Spring: 44, 33.9 (34.4), 23.9 (2.3)
summer:25,20.5 (15.0),17.2 (1.8)
Fall: 38, 33.1 (51.9), 19.5(2.5)
Phillips
All Seasons: 107,26.5 (24.3), 20.9
(2.0)
Spring: 28, 37.5 (37.6), 30.0 (1.8)
summer:40,22.7 (15.3),19.2 (1.7)
Fall: 39, 22.7 (16.7),. 17.6 (2.1)
PM2.5, Units = |ig/m
Battle Creek
All Seasons: 108,10.6 (6.6), 9.0 (1.8)
Spring: 25,12.7 (7.7), 11.0(1.7)
summer: 36,8.9 (3.8), 8.1 (1.5)
Fall: 47,10.9 (7.4), 8.8 (2.0)
E. St. Paul
All Seasons: 97,17.4 (20.3), 12.2 (2.2)
Spring: 30, 20.7 (26.4), 13.6 (2.4)
summer: 26,15.8 (11.4), 13.7 (1.6)
Fall 41 16.019.610.42.4
Phillips
All Seasons: 89,14.2 (13.0), 11.3 (1.9)
Spring: 15,16.9 (14.2), 13.0 (2.1)
summer: 36,13.2 (6.4), 11.4 (1.7)
Fall: 38,14.4 (16.7), 10.6(2.0)
PM2.5, Units = |ig/m
Battle Creek
All Seasons: 88 9.4 (6.2), 7.8 (1.8)
Spring: 36,10.5 (7.1), 8.5 (2.0)
summer: 22, 8.7 (4.4), 7.8 (1.6)
Fall: 30,8.4 (6.2), 7.1 (1.7)
E. St. Paul
All Seasons: 95,10.8 (6.6), 9.3 (1.8)
Spring: 36,12.0(7.3), 10.1 (1.9)
summer: 25, 8.5 (3.2), 7.8 (1.6)
Fall: 34,11.3 (7.5), 9.6 (1.8)
Phillips
All Seasons: 88,10.0 (5.8), 8.7, (1.7)
Spring: 30 (12.1), 7.2 (10.5)
summer: 30, 8.6 (3.8), 7.8 (1.6)
Fall: 28,9.3 (5.5), 8.1 (1.7)
Crist et al. (2008,156372)
Ohio River Valley near Columbus
PM2s, Units = |ig/m°
Athens (rural): 17.61 (17.81)
Koebel (urban): 14.59 (13.05)
New Albany (suburb): 13.93 (12.25)
PM2.5, Units = |ig/m°
Indoor
Athens (rural): 17.20 (13.56)
Koebel (urban): 14.98 (12.30)
New Albany (suburb): 16.52 (13.53)
PM25, Units = |ig/m°
Athens (rural): 13.66 (8.91)
Koebel (urban): 13.89 (9.29)
New Albany (suburb): 12.72 (8.86)
Sarnatetal. (2006, 089784)
Steubenville, OH
Mean (SD):PM15, Units = |ig/m°
Summer
n = 169
mean(SD) = 19.9(9.4)
Fall
mean (SD) = 20.1 (11.6)
Mean (SD):PM15, Units = |ig/m°
Summer
n = 65
mean (SD) = 20.1 (9.3)
Fall
mean (SD) = 19.3 (12.2)
SOUTHEAST
Wallace and Williams (2005, 057485)
Raleigh, North Carolina
Units = |ig/m°
PM25pers = 23.0(16.4)
PM2.5 pers/PM15 out =1.31 (0.99)
Units = |ig/m°
PM25 in =19.4 (16.5)
PM2.5 in/PM15 out = 1.08 (1.05)
Units = |ig/m°
PM2.5 out = 19.5 (8.6) 18.1 (8.1)
Williams etal. (2003.053338)
SE Raleigh, North Carolina
Chapel Hill, North Carolina
Pooled PM mass concentrations
(ug/m) across all subjects,
residences, seasons, and cohorts
Variable N Geo mean Mean RSD(a)
Personal PM2.5(b) 712 19.2 23.0 70.1
(a) Relative standard deviation of the
presented arithmetic mean.
(b) measured using PEMs.
Pooled PM mass concentrations (ug/m°)
across all subjects, residences, seasons,
and cohorts
Variable N Geo mean Mean RSD(a)
Indoor PM25 (c) 761,15.3,19.1,80.1
Outdoor PM25 (c) 761,17.5,19.3,43.7
Indoor PM10 (b) 761,23.2,27.7,70.6
Outdoor PM10 (b) 761,27.5,30.4,46.4
Indoor PM10.25(d) 761,6.3,8.6,111.8
Outdoor PM10.2.5 (d) 761, 8.5,11.1, 86.9
(a) Relative standard deviation of the
presented arithmetic mean.
(b) measured using PEMs.
(c) measured using HI samplers.
(d) measured by difference in PEM PMio
monitor and co-located HI PM2.5 mass
concentrations.
Pooled PM mass concentrations (ug/m°)
across all subjects, residences, seasons,
and cohorts
Variable N Geo mean Mean RSD(a)
Ambient PM25 (c) 746,17.3,19.2, 44.9
Ambient PM10 (b) 752, 27.9,31.4,51.5
Ambient PM10.2.5(d) 210,8.6,10.0,62.3
(a) Relative standard deviation of the
presented arithmetic mean.
(b) measured using PEMs.
(c) measured using HI samplers.
(d) measured by difference in PEM PM^
monitor and co-located HI PM2.5 mass
concentrations.
December 2009
A-322
-------
Reference / Location
Personal
Micro
Ambient
NORTHEAST
Koutrakis et al. (2005, 095800)
Baltimore, MD Boston, MA
PM2.5, Units = ng/m°: NR
(Baltimore, Boston)
Winter: Seniors: 15.1 (14.6), 14.1 (6.0)
Children: 24.0 (21.8), 18.5(12.8)
COPD: 16.4 (12.7), NR
Summer: Seniors: 22.1 (10.1), 18.8
(9.7)
Children: 18.6 (8.1), 30.3 (14.2)
COPD:NR, NREC:
(Baltimore, Boston)
Winter: Seniors: NR, 1.4(0.9)
Children: 2.8 (1.8), 1.6(1.6)
COPD: 2.0(1.2), NR
Summer: Seniors: NR, NR
Children: NR.NR
COPD:NR, NRS04:
(Baltimore, Boston)
Winter: Seniors: 1.9 (1.1), 1.9(1.2)
Children: NR, 2.3(1.7)
COPD: 1.5 (0.8), NR
Summer: Seniors: 5.7 (3.5), 2.9 (1.9)
Children: NR, NR
COPD:NR,NR
PM2.s, Units = |ig/m :
(Baltimore, Boston)
Winter:
All: 20.1 (9.4), 11.6 (6.8)
summer:
Seniors: 25.2 (11.5), 12.7(5.4)
Children: 23.2 (14.0), 17.0 (11.5)
COPD: NR, NREC:
(Baltimore, Boston)
Winter:
All: 1.2 (0.6)
summer: NR, NRS04:
(Baltimore, Boston)
Winter:
All: 4.0 (1.7), 3.1 (1.8)
summer:
Seniors: 10.5 (7.1), 3.1 (1.8)
Children: NR, 6.5 (6.0)
Sarnat etal. (2005. 087531)
Boston, Massachusetts. Comparisons to a
previous study in Baltimore are made.
Units = |ig/rrf:
Winter-Children:
PM25:17.4-25.8
S04:1.6-3.3
Winter-Seniors:
PM25:10.8-16.2
S04:1.6-2.6
Summer-Children
PM25:25.4-32.8
S04:2.7-3.3
Summer-Seniors
PM25:17.8-20.5
S04:2.7-3.3
NR
Units =|ig/rrf:
Winter:
PM25:6.5-15.5
S04:1.7-4.2
Summer:
PM25:11.9-21.4
S04:3.6-9.0
Turpin etal. (2007,157062)
Elizabeth, NJ, (and Houston, TX, and Los
Angeles County, CA+
48-h avg PM25, Units = |ig/m :
Elizabeth
Child: 54.0
Adult: 44.8
Elizabeth: 20.1
Elizabeth: 20.4
December 2009
A-323
-------
Table A-62. Summary of PM species exposure studies.
Reference Particle Sizes Measured Component Results
Adgate et al. (2007, 156196) Personal, Micro, and Ag.AI, Ca, Cd.Co, Cr, Cs, Median, units :ng/m3:
Ambient: PM25 - broken down Cu, Fe, K, La, Mg, Mn, Na,
intoTE Ni.Pb, S,Sb, Sc.Ti.TI, V,Zn Outdoor, Indoor, Personal
5334.4,272.1,351.6
Ca 232.2,85.0, 174.1
Al 96.3,23.3, 58.6
Na 11 1 9(1 fi 11 Q-
INd OO. I , ŁU.U, O I .3,
Fe12.6, 43.1, 78.6
Mg 10.9, 16.3, 27.5
K ? 9 ?8 4 47 5
l\ O.t, OO.t, *T( .O
Ti7 n n A 14
1 10. U, U.O, 1 .*f
7n9 7 fi 5 Q fi
Z-ii 1. / , o.vj, y .u
PM 9 4 1 ^ 4 Q
OU Ł .*T, I .O, *T.J?
MIMA fl 1 1 A
INIINM -U.I, I .0
me 9 4 Q 9
.vJ, t .*T, O.t
Mn n fi 1 5 9 ?
I VI 1 1 U.U, I .U, Ł.O
Sb 0.08,0.21, 0.30
Cd 0.05, 0.12, 0.14
V0.05, 0.12, 0.16
LaO.02, 0.05, 0.11
Cs 0.00, 0.00,0.00
Th 0.00, 0.00, 0.00
Sc 0.00, 0.00, 0.01
Ag 0.00, 0.07, 0.08
Co NA0.02, 0.07
Cr -0.09, 1.2, 2.6
Brunekreefetal.(2005, Personal, Micro & Ambient: N03" Mean (SD), units = ng/m3:
090486) PM2.5 „ , J
Primary Findings
The relationships among P, I,
and 0 concentrations varied
across TEs. Unadjusted mixed-
model results demonstrated
that ambient monitors are more
likely to underestimate than
overestimate exposure to many
of the TEs that are suspected
to play a role in the causation
of air pollution related health
effects. These data also
support the conclusion that TE
exposures are more likely to be
underestimated in the lower
income and centrally located
PHI community than in the
comparitively higher income BC
K community. Within the limits
of statistical power for this
sample size, the adjusted
models indicated clear
seasonal and community
related effects that should be
incorporated in long-term
exposure estimates for this
population.
In both cities, personal and
indoor PM2 5 were lower than
Brunekreefetal.(2005,
090486)
Personal, Micro, and
Ambient: PM2.5
S(V~, N03"
Amsterdam:
Personal 1389(1965)
Indoor 1348(1843)
outdoor 4063(4435)
Helsinki:
Personal 161 (202)
Indoor 267 (215)
Outdoor 1276(1181)
Mean, units = ug/m3:
S042":
P, I, 0
Amsterdam 4.6 4.7 5.9
Helsinki 2.7 3.0 5.0
N03":
P, I, 0
Amsterdam 1.4 1.4 4.0
Helsinki 0.2 0.3 1.3
highly correlated with outdoor
concentrations. For most
elements, personal and indoor
concentrations were also highly
correlated with outdoor
concentrations.
In both cities personal and
indoor PM2.5 were lower than
highly correlated with outdoor
concentrations. For most
elements, personal and
indoor concentrations were
also highly correlated with
outdoor concentrations.
Chillrud et al. (2004, 054799) Personal: PM2.
Micro :PM2.5
Home indoor and home
outdoor
Ambient: Urban fixed-site and
upwind fixed site operated for
three consecutive 48-h
periods each week.
Elemental iron, manganese,
and chromium are reported in
this study out of 28 elements
sampled.
Mean of duplicate samples:
PM2.5:62ug/m3
Fe:26 ug/n?
Mn: 240 ng/m3
Cr: 84 ng/m3
Variability: 1-15%
Personal samples had
significantly higher
concentration of iron,
manganese, and chromium
than home indoor and
ambient samples. The ratios
of Fe (ng/ ug of PM2.5) vs Mn
(pg/ ug PM2.5) showed
personal samples to be twice
the ratio for crustal material.
Similarly for the Cr/Mn ratio.
The ratios and strong
correlations between pairs of
elements suggested steel
dust as the source. Time-
activity data suggested
subways as a source of the
elevated personal metal
levels.
December 2009
A-324
-------
Reference
Particle Sizes Measured
Component
Results
Primary Findings
Ebeltetal. (2005,056907) Personal: PM2.5
Micro: "ambient exposure":
PM2.5,PM10, PM2.5-io;
"non-ambient exposure:"
PM2.5
Ambient: PM2.5, PM10, PM2.5.
Ambient SO/,
Ambient non-sulfate,
Personal sulfate,
personal ambient non-sulfate
Mean(SD),Units ug/m°
Ambient sulfate: 2.0 (1.1),
Ambient non-sulfate: 9.3
(3.7),
Personal sulfate: 1.5 (0.9),
personal ambient non-sulfate:
6.5 (3.0)
Ambient exposures and (to a
lesser extent) ambient
concentrations were
associated with health
outcomes; total and
nonambient particle
exposures were not.
Farmer etal. (2003,089017) Personal: PM10
Micro: NR
A h' t- DM
AmDieni. riviio
Extractable organic material
/Cl"ll\/h
(tUM)
BtelP
LJ[QJI
cPAHs
Benzo[a]pyrene (B[a]P)
Carcinogenic polycyclic
aromatic hydrocarbons
(cPAHs)
Units :ng/m3
Exposed, controls:
Prague:
cPAHs = 12.04(11. 10), 6.17
(3.48)
B[a]P= 1.79 (1.67), 0.84
(0.60)
Personal exposure to B[a]P
and to total carcinogenic
PAHs in Prague was two fold
higher in the exposed group
compared to controls, in
Kosice three fold higher, and
in Sofia 2.5 fold higher.
Kosice:
cPAHs = 21.72(3.12),6.39
(1.56)
B[a]P= 2.94 (1.44), 1.07
(0.66)
Sofia:
cPAHs = 93.84 (55.0) police,
94.74 (120.34) bus drivers,
41.65(33.36)
B[a]P=4.31 (2.6) police, 5.4
(3.18) bus drivers, 1.96 (1.53)
Farmer etal. (2003,089017) Personal: PM10
Micro: NR
Ambient: PM10
PM2.5 (not reported)
PM10
EOM
EOM2
B[a]P
c-PAHsb
Prague-SMWinter Summer
EOM (ug/m3) 14.93 4.96
EOM2 (%)23.9 13.4
B[a]P (ug/m3)3.5 0.28
c-PAHsb (ug/m3) 24.69 2.29
Prague-LB Winter Summer
EOM (ug/m3) 10.86 3.72
EOM2 (%)27.9 14.1
B[a]P (ug/m3)2.9 0.17
c-PAHsb (ug/m3) 20.36 1.32
Kosice Winter Summer
EOM (ug/m3) 15.3 1.67
FDM9 Wn19fid fifl
Extractable organic matter
(EOM) per PM10 was at least
2-fold higher in winter than in
summer, and c-PAHs over
10-fold higher in winter than
in summer Personal
exposure to B[a]P and to total
c-PAHs in Prague ca. was 2-
fold higher in the exposed
group compared to the
control group, in Kosice ca. 3-
fold higher, and in Sofia ca.
O C fnlrl hinhor
Ao-Toid nigner.
B[a]P (ug/m3) 1.37 0.15
c-PAHsb (ug/m3) 11.87 1.2
Sofia Winter Summer
EOM (ug/m3) 24.6 3.95
EOM2 (%)27.37 13.3
B[a]P (ug/m3)4.84 0.36
c-PAHsb (ug/m3) 36.44 2.43
Gadkari et al. (2007,156459) Personal: Respirable PM
(RPM)
Micro: NR
Ambient: RPM
Fe.Ca, Mg, NaK.Cd.Hg,
Ni,Cr,Zn,As,Pb, Mn and Li
Source contributions varied
widely among 12 sites.
Indoor: 0-95%
Ambient: 0-26%
Road: 0-94%
Soil: 0-75%
Authors conclude that personal
exposure to ambient RPM is
related to local traffic and soil
resuspension. They felt that
indoor activities or ventilation
determined indoor levels of
RPM.
December 2009
A-325
-------
Reference
Particle Sizes Measured
Component
Results
Primary Findings
Gevhetal. (2005. 186949)
Hanninen et al. (2004,
056812)
Ho et al. (2004, 056804)
Jacquemin et al. (2007,
192372)
Personal: TD, PM10, PM2.5 EC
Micro:NR VOC also assessed
Ambient: TD, PM10, PM2.5
Personal: PM25 PM25 -bound S
Micro: NR
Ambient: PM15
Personal: PM25 OC
EC
Micro: NR QM
Ambient: PM2.5 TCA
Personal: PM15 S
Micro :NA
Mean (SD), units = ug/m3:
Summary Statistics by Area
Location
October 2001 :
Albany and West
EC 5.9 (NA) OC 36 (NA)
Liberty and Greenwich
EC 5.3 (59) OC 30 (56)
Park Place and Greenwich
EC 1 4.5 (5. 4) OC 72 (26)
Church and Dey
EC 7. 9 (3. 3) OC 48 (15)
April 2002:
Liberty and West
EC 4.2 (2.1) OC 26 (13)
Barclay and Greenwich
EC 4.0 (2.6) OC 18 (14)
Church and Dey
EC4.5(1.9)OC27(15)
Middle of the Pile
EC6.7(1.0)OC40(25)
IndoorOutdoor
Athens5.3 (2.0)7.6 (5.1)
Basel2.6 (1.6)3.3 (1.6)
Helsinki 1.6 (1.3) 2.2 (1.5)
Prague 3.1 (1.3)4.0(1.5)
Mean, Unit = ug/m3
Indoors:
OM = 18.1; TCA =22.9
Outdoors:
OM = 20.1; TCA =26.5
Mean, units = ug/m3:
Personal: 1.3 outdoor: 1.2
Comparison of recorded EC
and OC values from October
2001 and April 2002 with
previous studies suggests
that the primary source of
exposure to EC for the WTC
truck drivers was emissions
from their own vehicles.
Associated with indoor
concentration: wooden
building material, city, building
age, floor of residence (i.e.
ground, 1st, etc.), and use of
stove other than electric.
The major source of indoor
EC, OC, and PM2.5 appears
to be penetration of outdoor
air, with a much greater
attenuation in mechanically
ventilated buildings.
Authors suggest that "outdoor
measurements of absorbance
and sulnhur can be used to
Jansen et al. (2005. 082236)
Personal, Micro, and
Ambient: PM2.5
Estimated Elemental Carbon
Elemental composition of a
subset of personal, indoor
and outdoor samples
Mean (SD), units = ug/m°:
Amsterdam, Helsinki
P,0,P,0
PM2.5 14.5,15.7,9.4,11.4
Zn13.2,18.3,11.7,18.6
Fe57.0, 71.3, 41.6,79.2
K 87.4,70.3,103.1,93.9
Ca 72.9, 40.2, 68.5, 36.4
Cu 5.4,2.5,4.3,1.8
Si 29.7, 13.7,79.5,93.9
CI40.8, 72.7, 9.8,44.2
estimate both the daily
variation and levels of
personal exposures also in
Southern European
countries, especially when
exposure to ETS has been
taken into account. For PM2.5,
indoor sources need to be
carefully considered."
For most elements, personal
and indoor concentrations
were lower than and highly
correlated with outdoor
concentrations. The highest
correlations (median r.0.9)
were found for sulfur and
particle absorbance (EC),
which both represent fine
mode particles from outdoor
origin. Low correlations were
observed for elements that
represent the coarser part of
the PM25 particles (Ca.Cu,
Si, Cl).
December 2009
A-326
-------
Reference
Johannesson et al. (2007,
156614)
Particle Sizes Measured
Personal, Micro, and BS
Ambient: PM2.5, PM,
Component
Results
BS2.5 Mean SD
Personal 0.65 0.47
Primary Findings
Personal exposure of PM2.5
correlated well with indoor
Exclusively smokers 0.62
0.47
Residential indoor 0.56 0.47
Exclusively smokers 0.52
0.46
Residential outdoor 0.68 0.51
Exclusively smokers 0.71
0.54
Urban background 0.63 0.37
All measurements 0.68 0.40
PM,/BS1
Personal 0.55 0.20
Residential indoor 0.54 0.45
Exclusively smokers 0.49
0.43
Residential outdoor 0.66 0.51
Exclusively smokers 0.68
with residential outdoor and
urban background
concentrations were also
acceptable. Statistically
significantly higher personal
exposure compared with
residential outdoor levels of
PM2.s was found for
nonsmokers. PM, made up a
considerable proportion
(about 70-80%) of PM2.5. For
BS, significantly higher levels
were found outdoors
compared with indoors, and
levels were higher outdoors
during the fall than during
spring. There were relatively
low correlations between
particle mass and BS. The
urban background station
provided a good estimate of
the residential outdoor
concentrations of both PM2.5
and BS2.5 within the city. The
air mass origin affected the
outdoor levels of both PM2.5
and BS2.5; however, no effect
was seen on personal
exposure or indoor levels.
Kim etal. (2005,156640) Personal: PM15
Micro: NR
Ambient: PM15
Koistinen et al. (2004, Personal, Micro, and
156655) Ambient: PM15
S042", EC, Ca2*, Mn2*, K, Na* Mean (SD), Units = ug/m3:
S042": 2.7 (3.2)
Ca2+:0.12(0.12)
Mg2+: 0.02 (0.01)
K: 0.07 (0.08)
Na+: 0.09 (0.20)
EC: 0.60 (0.54)
Black smoke, S042", N03-, % contribution to PM25
NhV, Al, Ca, Cl, Cu, K, Mg, P, Outdoor - Indoor - Work -
S,Si,Zn Personal
CoPM*35,28,32,33
Secondary** 46, 36, 37,31
Soil 16, 27, 27, 27
Detergents 0, 6, 2, 6
Sea Salt 3, 2, 1,2
* PnDM io tha Hiffaran^a
Traffic- related combustion,
regional, and local crustal
materials were found to
contribute 19% ±17%,
52% ±22%, and 10% ±7%,
respectively.
Among participants that spent
considerable time indoors,
exposure to outdoor PM2.5
includes a greater relative
contribution from combustion
sources, compared with
outdoor (ambient) PM2.5
measurements.
Population exposure
assessment of PM2.5, based
on outdoor fixed-site
monitoring, overestimates
exposures to outdoor sources
like traffic and long-range
transport and does not
account for the contribution of
significant indoor sources.
between total mass and other
identified components; i.e.,
primary combustion particles,
nonvolatile primary and
secondary organic particles,
and particles from tire wear,
water, etc. ** Secondary
particles are the sum of
sulfate, nitrate, and
ammonium. 4 factors were
identified for each exposure
type (residential indoor,
residential outdoor, workplace
indoor, and personal). The
factors contained the
elements AI,Ca,CI, Cu, K,
Mg, P, S, Si, Zn, and black
smoke, (insert in cell to left
after consolidating PM size)
December 2009
A-327
-------
Reference Particle Sizes Measured Component
Koutrakis et al. (2005, Personal: PM25 Elemental Carbon (EC),
0958001 S042-
Micro:NR
Ambient: PM2.5
Kulkarni and Patil (2003, Personal: PM5 Pb
156664)
Micro: NR Ni
Ambient: PM5 Cd
Cu
Cr
Fe
Mn
Results
Mean (SD) data are provided
for Baltimore and Boston,
Units = |jg/m3:
EC:
(Baltimore, Boston)
Winter:
Seniors: NR, 1.4(0.9)
Children: 2.8 (1.8), 1.6(1.6)
COPD:2.0(1.2), NR
en 2~-
GW4
(Baltimore, Boston)
Winter:
Seniors: 1.9 (1.1), 1.9 (1.2)
Children: NR, 2.3 (1.7)
COPD:1.5(0.8), NR
Summer:
Seniors: 5.7 (3.5), 2.9 (1.9)
Personal samples,
Units = pg/m3:
Mean + SD
Type
Pb
Occupational
4.384 ± 7.766 pg/m3
Residential
4.093 ± 5.925 pg/m3
24-h integrated
4.205 + 1.523 pg/m3
PH
UQ
Occupational
0.201 ±0.158 pg/m3
Rssidsntisl
0.111 ±0.165 pg/m3
24-h integrated
0.1 34 ±0.1 40 pg/m3
Mn
IvIM
Occupational
1.979 ±7.842 pg/m3
Residential
0.1 80 ±0.261 pg/m3
24-h integrated
1.983 ±6.824 pg/m3
Occupational
3.473 ± 4.691 pg/m3
Residential
4.589 ± 4.619 pg/m3
24-h integrated Check
Primary Findings
Ambient PM2.5 and S042~ are
strong predictors of
respective personal
exposures. Ambient S04 is
a strong predictor of personal
exposure to PM2.5. Because
PM2.5 has substantial indoor
sources and S042~ does not,
the investigators concluded
that personal exposure to
S04 accurately reflects
exposure to ambient PM2.5
and therefore the ambient
component of personal
exposure to PM2.5 as well.
All listed metals were
detected in the ambient air
where as only Lead,
Cadmium, Manganese, and
Potassium were detected in
personal exposures. Mean
daily exposure to lead
exceeds the Indian NAAQS
by a factor of 4.2. However,
ambient concentration of lead
conforms to this standard.
There is a rising trend in the
personal exposures and
ambient levels of cadmium.
However, they are low and do
not pose any major health
risk as yet. Personal
exposures to toxic metals
exceed the corresponding
ambient levels by a large
factor ranging from 6.1 to
13.2. Thus, ambient
concentrations may
underestimate health risk due
to personal exposure of toxic
metals. Outdoor exposure to
toxic metals is greater than
the indoor (ratios ranging
from 2.3 to 1.1) except for
potassium (ratio 0.77).
However, there is no
significant correlation
between these two.
December 2009
A-328
-------
Reference
Particle Sizes Measured
Component
Results
Primary Findings
Lai et al. (2004, 056811)
Personal, Micro, and
Ambient: PM2.5
Ag Cr Mn Si
Al Cu Na Sm
As Fe Ni Sn
BaGaPSr
BrGePbTi
CaHgRbTI
Cd I S Tm
CIKSbV
Co Mg Se Zn
Zr
GM(GSD), Units: ng/mj
P,RI,RO,WI, I/O
Al 280 (7.0), 67 (7 .2), 22
(2.9), 110 (7.5), 1.4
As 4.7 (1.6), 3.7(1.8), 2.6
(2.7), 6(-),1.4
Br 4.7 (2.2), 3.9(2.0), 2.4
(2.5), 6.2(2.5), 1.6
Ca 260 (2.0), 120(2.1), 30
(1.6), 280(2.9), 3.3
Cd23(1.4), 19(1.8), 7 (-),
43 (2.2), -
Cl 400 (3.0), 270 (3.9), 220
(5. 2), 380 (3. 9), 1.0
Cu 120 (1.3), 88 (1.7), 2.3
(2.8), 230 (2.1), 37.1
Fe 59(2.3),30(3.8),19(3.5),
85 (2. 9), 1.6
Ga 0.9 (2.1), 0.6 (2.2), 0.2
(2.2), 2.0 (3.4), 2.4
K 250 (2. 4), 180 (2. 7), 93
(2.0),130(4.0),1.7
Mg 260(2.1),130(3.1),140
(2.9), 120 (2.8), 0.7
Mn 2.1 (2. 6), 1.8 (2. 4) ,2. 2
(1.5), 3.5 (3.0) ,0.8
Na 21 00 (1.6) ,1800 (1.7),
11 00 (3 .2), 2700 (1.9), 1.6
NM1 (2.2), 8.6 (2.5), 1 8 (— ),
23(2.9),—
P 110 (2.1), 70 (2.2), 27(1.8),
86 (2. 4), 2. 5
Pb 26 (1.7), 19(1.8), 9.4(2.8),
32(2.0),1.9
S 1200 (1.9),1200 (2.0), 890
(4.8), 1.2
Se 8.4 (1.5) ,6.8 (1.7) ,2.3
(1.8), 16 (2. 2) ,2. 8
Si 740 (3.4), 360 (2.9) ,95
(2.2), 570 (3.8), 2.6
Sn35(1.5),27(1.8),0(-),68
(2.6),-
Ti 6.2 (1.7) ,2.8 (2.2), 1.1
(2.0), 6.1 (3.2) ,2.3
Both the indoor and outdoor
environments have sources
that elevated the indoor
concentrations in a different
extent, in turn led to higher
personal exposures to
various pollutants.
Geometric mean (GM) of
personal and home indoor
levels of PM25,14 elements,
total VOC (TVOC) and 8
individual compounds were
over 20% higher than their
GM outdoor levels. Those of
N02,5 aromatic VOCs, and 5
other elements were close to
their GM outdoor levels. For
PM2.s and TVOC, personal
exposures and residential
indoor levels (in GM) were
about 2 times higher among
the tobacco-smoke exposed
group compared to the non-
smoke exposed group,
suggesting that smoking is an
important determinant ofthese
exposures. Determinants for
CO were visualised by real-
time monitoring, and the
authors showed that the peak
levels of personal exposure
to CO were associated with
smoking, cooking and
transportation activities.
Moderate to good
correlations were only found
between the personal
exposures and residential
indoor levels for both PM25
(r = 0:60; p< 0:001) and N02
(r = 0:47; p= 0:003).
Zn 18(2.4),15(2.2),13
(2. 5) ,23 (2. 4), 0.9
December 2009
A-329
-------
Reference Particle Sizes Measured Component
Larson et al. (2004, 098145) Personal: PM2.5 Light absorbing carbon (LAC)
andAI, As, Br, Ca, Cl, Cr, Cu,
Micro: PM2.5 outside subject's Fe K Mn Ni Pb Sj, s, Ti, V
residence, and inside
residence
Ambient: PM2.5 at Central
outdoor site (downtown
Seattle)
Maitre et al. (2002, 156726) Personal: PM4 PAH, benxene-toluene-
Micro: NR ^nes (BTX), aldehydes,
Ambient: PM4 a^ldehydf^6'
Meng et al. (2005, 081194) Personal: PM2.5 EC, OC, S, Si
Micro :NA
Ambient: NR
Results
Personal, Rl, RO, Central
Mass
10,50010,25012,69311,970
AI32, 19,21,31
As1, 1,2,2
LAC* 1439, 1105, 1830, 1741
Br 3, 2, 3, 3
Ca 72, 46, 36, 50
0248,173,75,78
Cr2,2, 1,2
Cu 3, 4, 2, 3
Fe 63, 35, 61, 95
K57, 54, 78, 67
Mn2, 2, 3, 6
NiO.O, 1, 1
Pb 2, 2, 5, 5
Si 109, 65, 66, 62
S 289, 289, 468, 492
Ti 4, 3, 3, 6
VO, 1,2,3
Median
Personal Ambient
Resp ug/m 124, 124 (mean)
BaPng/m3 0.28,0.1 4
PAHcng/m31.19,1.56
PAH ng/m3 13.14,12.26
Benzene ug/m323.5, 17
Toluene ug/m3 94.5, 52
Xylene ug/m3 74, 39
BTX ug/m3 192, 108
Formaldehyde ug/m3 21 ,17.5
Acetaldehyde ua/m3 17, 10.5
Aldehyde ug/nf 38, 28
Mean (SD), units = ng/m3:
Indoor:
EC: 11 65.9 (2081.0)
OC: 7725.5 (9359.3)
5:902.3(602.2)
51:124.0(79.0)
Outdoor:
EC: 1144.1 (968.1)
OC: 3777.7 (2520.1)
5:1232.3(633.2)
51:141.1(171.3)
Primary Findings
Five sources of PM2.5
identified: vegetative burning,
mobile emissions, secondary
sulfate, a source rich in
chlorine, and crustal-derived
material. The burning of
vegetation (in homes)
contributed more PM2.5 mass
on avg than any other
sources in all
microenvironments.
The occupational exposure of
policemen does not exceed
any currently applicable
occupational or medical
exposure limits. Individual
particulate levels should
preferably be monitored in
Grenoble in winter to avoid
underestimations.
Use of central-site PM2.5 as
an exposure surrogate
underestimates the
bandwidth of the distribution
of exposures to PM of
ambient origin.
December 2009
A-330
-------
Reference Particle Sizes Measured Component
Molnaretal. (2005, 156772) Personal: PM25 BS
S
Micro and Ambient: PM10-2.s ci
and PM2.5 K
Ca
Mn
Fe
Cu
Zn
Br
Rb
Pb
Molnaretal. (2006, 156773) Personal: PM15 and PM, S
Micro and Ambient: NR CI
Ca
Ti
V
Mn
FP
re
Ni
Cu
Zn
Br
Pb
Results
Median, unit = ng/m3
Wood burners
Ref 1 -sided p-value
BS0.97, 0.74, 0.053
5880,650,0.500
CI 200, 160,0.036
K 240, 140, 0.024
Ca 76, 43, 0.033
Mn 4.8, 3.5, 0.250
Fe 64, 49,0.139
Cu 8.9, 2.4, 0.016
Zn 38, 22, 0.033
Urban background PM25
Mean, median, range (ng/m3)
3620,320,95-1900
CI 97, 54, 25-460
K55, 50, 32-130
Ca21, 17,6.6-6.2
Ti 2.1, 1.9, 1.3-3.8
V3.4, 2.4, 1.0-13
Mn 1 6 1 4 0 67- 3 8
Fe36, 33, 7.1-100
NI1.6, 1.2,0.33-5.7
Cu2.1, 1.4,0.33-11
Zn14, 11,2.8-38
BM.7, 1.4,0.47-44.3
Pb 3.3, 2.1,0.94-11
Personal PM25
Mean, median, range (ug/m )
S- < 470, 270-1400
CI 270, 170, 60-920
K140, 96, 39-690
Ca110, 80, 27-670
TM 1,9.5, 3.7-27
V4.7, 4.0, 2.7-9.4
Mn---
Fe68, 69, 23-150
Ni 4.2, 2.6, 0.89-46
Cu10, 6.6, 1.1-81
Zn 21, 16, 6.6-70
Br2.0, 1.3,0.91-14
Pb 2.9, 2.6, 0.92-8.3
Primary Findings
Statistically significant
contributions of wood burning
to personal exposure and
indoor concentrations have
been shown for K, Ca, and Zn.
Increases of 66-80% were
found for these elements,
which seem to be good wood-
smoke markers. In addition, CI,
Mn, Cu, Rb, Pb, and BS were
found to be possible wood-
smoke markers, though not
always to a statistically
significant degree for personal
exposure and indoor
concentrations. For some of
these elements, subgroups of
wood burners had clearly
higher levels which could not
be explained by the information
available.
Sulfur, one of the more typical
elements mentioned as a
wood-smoke marker, showed
no relation to wood smoke in
this study due to the large
variations in outdoor
concentrations from LOT air
pollution. This was also the
case for PM25 mass. Personal
exposures and indoor levels
correlated well among the
subjects for all investigated
species, and personal
exposures were generally
higher than indoor levels.
PM2 5 personal exposures were
significantly higher than both
outdoor and urban background
for the elements CI, K, Ca, Ti,
Fe, and Cu. Personal exposure
was also higher t an indoor
levels of CI, Ca, Ti, Fe, and Br,
but lower than outdoor Pb.
Residential outdoor levels were
significantly higher than the
corresponding indoor levels for
Br and Pb, but lower for Ti and
Cu. The residential levels were
also significantly higher than
the urban background for most
elements.
Personal PM,
Mean, median, range (ug/m3)
S-,< 470, 240-1200
CI-, < 110,54-160
K 80, 82, 50-130
Ca 32, 23, 8.4-87
Ti 6.5,6.3,3.7-11
V-,< 4.2,2.8-8.9
December 2009
A-331
-------
Reference
Particle Sizes Measured
Component
Results
Primary Findings
Mn---
Fe 28, 25, 7.6-68
Ni 8.2,1.2,0.83-58
Cu 5.0, 4.4,1.6-14
Zn15,14, 7.6-37
Br 1.6,1.5,0.83-4.4
Pb 3.6, 2.8,1.1-11
Residential Outdoor PM2.5
Mean, median, range
5640,460,190-1800
016.3,140,57-840
K 200,78,32-200
Ca 82, 28, 4.6-85
Ti 34, 5.2, 3.3-21
V6.3,3.9,2.1-14
Mn~
Fe 5.5, 31, 8.8-200
Ni 45, < 1.6, 0.65-5.5
Cu 2.6,1.3,0.65-17
Zn 22,15,5.5-85
Br 2.0, >450,0.91-51
Pb4.6, 2.6, 0.90-20
Residential Outdoor PMi
S-, 1.3,24-2000
CI-, < 110,44-170
K 76, 68, 34-170
Ca- < 12, 5.1-78
Ti-,< 5.0, 2.2-9.5
V5.6,4.47,2.2-14
Mn~
Fe 23,14,3.7-140
Ni 3.3,1.4,0.73-28
Cu-< 1.1, 0.73-12
Zn 15,14,5.2-30
Br 1.5,1.4,0.78-4.3
Pb 4.1,1.5,1.0-17
Na and Cocker (2005,
156790)
Personal: PM2.5 EC, OC
Micro: NR
Ambient: PM2.5
Mean (SD), units = ug/m3
Residential homes: EC 2.0
(NR)
OC14.8(NR)
Hinh school (FCV
Indoor PM2.5was significant
influenced by indoor OC
sources.
Indoor EC sources were
predominantly of outdoor
origin.
Weekday samples 1.1 (0.9)
Weekend samples 1.0 (0.5)
High school (OC):
Weekday samples 8.8 (4.7)
Weekend samples 7.4 (2.4)
Noulett et al. (2006,155999) Personal: PM15
Micro: NR
Ambient: PM15
ABS (light absorbing carbon)
Measurement Mean s.d.
Ambient so42" 2.72*3.11
Ambient ABS 1.4** 1.0
Personal S042" 1.33* 1.47
Personal ABS 1.0** 1.7
* Mean S042" values reported
in ug/m3
** Mean ABS values reported
in10-5/m"1
S042" and light absorbing
carbon concentrations had
higher personal-ambient
correlations and less
variability. This indicates that
S042" and ABS were of
outdoor origin, while PM2.5
mass was of varied indoor
and outdoor origin.
December 2009
A-332
-------
Reference
Salmaetal. (2007. 113852)
Sarnatetal.(2005RMID
9171) (2005.087531)
Sarnatetal. (2006. 089784)
Shilton et al. (2002, 049602)
Particle Sizes Measured
Personal: PMi0-2.o and PM2.0
Micro :NA
Ambient: NR
Personal: PM15
Micro :N/A
Ambient: PM15
Personal: PM25
Micro: NR
Ambient: PM2.5
Personal, Micro, and
Ambient: Respirable PM
Component Results
30 elements (Na, Mg.AI.Si, Units: ng/m3:
P, S,CI, K, Ca,Ti,V,Cr, Mn,
Fe, Ni, Cu, Zn, Ga, Ge, As, PM10-2.0; PM2.0
Se.Br, Rb, Sr.Y.Zr, Nb, Mo, Mg 296 130
Ba.andPb) AI53193
' Si 2.09 442
S 978 828
Cl 305 104
K318127
Ca 2.57 41 3
Ti 47 25
Cr3515
Mn 310 148
Fe 33.5 15.5
NI298
Cu 496 190
Zn11850
Br13DL
Ba145DL
Pb 47 21
PM 83.6 33.0
S04, 03, N02, S02 Correlations between
personal PM2.5 and ambient
gas
03 correlated in summer.
Spearman's R=0.4,
Anti-correlated in winter,
R=0.3-0.1.
NOX somewhat correlated in
summer. R=0.3
Winter, R*0.2-0.4
S02 not well correlated in
summer or winter. R=0-0.1 .
CO somewhat correlated in
summer. R=0. 1-0.3.
Correlated in winter R=0.2-
0.3.
No results were significant.
S042" Mean (SD), units = ug/m3:
EC Personal
Ambient
S042"
Summer
5.9 (4.2)
7.7 (4.8)
Fall
4.4 (3.3)
6.2 (4.7)
EC
Summer
1.1 (0.6)
1.1 (0.5)
Fall
1.2(0.7)
1.1 (0.7)
Respirable PM, metals (Zn, Cu, IndoorOutdoor
Mn, Al), S042", N03", and Cl" 3
Cu (ng/m3) 43.3, 24.99
Mn(ng/m3) 15.6,4.18
Al (ng/m3) 305.2, 52.90
S042- (ng/m3) 4.72, 3.47
Cl (ng/m'*) 1.08, 0.1 5
N03'(ng/m3).35, 1.08
Primary Findings
The concentrations observed
in the Astoria underground
station were clearly lower (by
several orders of magnitude)
than the corresponding
workplace limits.
Substantial correlations
between ambient PM2.5
concentrations and
corresponding personal
exposures.
Summertime gaseous
pollutant concentrations may
be better surrogates of
personal PM2.5 exposures
(especially personal
exposures to PM2.5 of
ambient origin) than they are
surrogates of personal
exposures to the gases
themselves.
High association between
personal and ambient S042~
and EC, especially for S04
for which there is no
significant indoor source.
The indoor particulate cone
was driven by ambient cone;
meteorological-induced
changes in ambient PM were
detected indoors;
December 2009
A-333
-------
Reference
Particle Sizes Measured
Component
Results
Primary Findings
Smith etal. (2006,156990) Personal: PM2.5
Micro :PM2.s
Area samplers in the offices,
freight dock, or shop.
Ambient: PM2.5
Samplers were located in the
yard upwind of the terminal.
EC
OC
Work Area EC, OC, EC/TC
Office 0.31 (3.72), 11.29
(1.63)
Dock0.53 (3.24), 5.01 (1.76),
3% (3.10)
YardO.73 (2.89), 7.77 (1.65),
9% (2.49)
Shop 1.54 (3.52), 10.37
(2.00), 8% (2.21)
Non-smokers on-site: 12%
(2.13)
Clerk 0.09 (9.98), 15.97
(1.31)
Dock worker 0.76(2.13),
13.89(1.45),!% (10.19)
Mechanic 2.00 (3.82), 16.89
(1.64), 5% (1.96)
Hostler 0.88 (3.04), 14.89
(1.86), 10% (2.71)
Non-smokers off-site 5%
(2.09)
Pickup/deliver driver 1.09
(2.46), 12.40 (1.54)
Long haul driver 1.12(1.91),
19.26 (2.30),8% (2.13)
Smokers On-Site 7% (1.82)
Clerk 1.19 (1.70), 32.25
(1.70),NR
Dock worker 0.98(1.93),
24.02(1.87)
Mechanic 2.41 (2.27), 24.35
(1.78)
Hostler 1.74 (2.21), 43.92
(2.03)
Smokers off-site
Pickup & Delivery drivers
1.33 (3.84), 24.24 (2.14)
Long haul drivers 1.37
(2.40),32.81 (3.23)
December 2009
A-334
-------
Reference
Particle Sizes Measured
Component
Results
Primary Findings
Ssrensen et al. (2003, Personal: PM25 BS Units: 10"6/m
157000)
Micro: NR n Median Q25-Q75
Ambient: PM2.5 All 177 6.8 (5.0-13.2)
Autumn 42 7.1 (6.5-17.2)
Winter 46 8.2 (5.1 -13.3)
Spring 46 12.6 (5.4-10.4)
Summer 47 8.1 (3.4-9.0)
Personal PM2.5 exposure was
found to be a predictor of 8-
oxodG in lymphocyte DMA.
No other associations
between exposure markers
and biomarkers could be
distinguished. ETS was not a
predictor of any biomarker in
the present study. The current
study suggests that exposure
to PM2.s at modest levels can
induce oxidative DMA
damage and that the
association to oxidative DMA
damage was confined to the
personal exposure, whereas
the ambient background
concentrations showed no
significant association.
For most of the biomarkers
and external exposure
markers, significant
differences between the
seasons were found.
Similarly, season was a
significant predictor of SBs
and PAH adducts, with avg
outdoor temperature as an
additional significant
predictor.
Sorenson et al. (2005,
089428)
Personal: PM25 and BS
Micro: PM25 and BS
Ambient: Street monitoring
station and roof of a campus
building PM25 and BS
BS
Mean, IQR, Units = ug/m
Personal:
Cold Season: 10.2 (5.6-14.8)
Warm Season: 7.1 (5.5-11.4)
Micro:
Cold Season
Home Indoor: 6.2 (5.5-11.4)
Home front door: 10.8 (7.4-
16.3)
Warm Season
Home Indoor: 6.1 (3.7-7.6)
Home front door:
8.8(5.6-11.54)
Ambient:
Cold Season: Street Station:
31.6 (27.5-34.0)
Urban Background:
7.7(5.9-11.0)
Indoor sources of PM and BS
were shown to be greatly
influenced by indoor sources.
Srametal. (2007, 192084) Personal: PMm, PM,S c-PAHs, BFalP
Micro: NR
Ambient: PM10, PM2.5
Warm Season:
Street Station:
30.6 (24.7-36.0)
Urban Background:
6.8 (4.6-8.6)
B[a]P:
Exposed 1 .6 ng/m
Control 0.8 ng/m3
c-PAHs:
Exposed 9.7 ng/m
Control 5.8 ng/m3
Ambient air exposure to c-
PAHs increased fluorescent
in situ hybridization (FISH)
cytogenetic parameters in
non-smoking policemen
exposed to ambient PM
December 2009
A-335
-------
Reference Particle Sizes Measured Component Results
Turpinetal. (2007. 157062) Personal: PM2.5 18 volatile organics, 17 For LosAngeles
carbonyl, PM25 mass and .
Micro :PM2.5, in the main >23 PM25 species organic Carbon (|jgC/m )
living area (not kitchen) carbon, elemental carbon, EC 1.4
Ambient: PM15, in the front or and PAHs Elements (ng/m3)
backyard
AgO.5
Al 24.7
As 0.5
Ba 22.9
Br5.3
Ca 80.9
CdO.4
Cl 62.0
CoND
CrO.6
Cu5.5
Fe 162.9
GaO.1
GeO.1
HgO.1
In 0.3
K74.1
La 2.3
Mn2.9
Mo 0.4
NI2.0
Pb4.7
PdO.3
P0.1
RbO.1
S 1022.9
Sb2.1
Se1.4
Si 128.9
Sn7.9
Sr1.8
TI10.4
V5.3
Y0.1
Zn16.4
ZrO.5
Primary Findings
The best estimate of the
mean contribution of outdoor
to indoor PM2.5 was 73% and
the outdoor contribution to
personal was 26%.
Wallace and Williams (2005, Personal: PM25
057485)
Indoor Micro: PM2.5
Outdoor Micro: PM2.
Mean (SD), units = ng/m :
Personal: 1046 (633)
Indoor: 1098 (652)
Outdoor: 1951 (1137)
Generally, Finf provides a
reliable estimate of personal
exposure. S can be used in lieu
of personal exposure to PM
because it is generally derived
from outdoors.
Wu et al. (2006, 179950) Personal: PM25 LG Mean personal exposure
EC (pg/m3):
Micro:PM25
OC LG: 0.01 8 (0.024)
Amhipnt' PM
r\l 1 lUlcl 11 . rlvl25
EC: 0.4 (0.5)
OC: 8.5 (2.7).
Ambient: check component
During non-burning times:
0.026 (0.030)
During burning episodes:
0.010(0.012)
Authors "found a significant
between-subject variation
between episodes and non-
episodes in both the
Exposure during agricultural
burning estimates and
subjects' activity patterns.
This suggests that the LG
measurements at the central
site may not always represent
individual exposures to
agricultural burning smoke
"Evidence of "Hawthorne
Effect": During declared
episodes (i.e. real and sham),
subjects spent less time
indoors at home and more
time in transit or indoors
away from home than during
non-declared episode
periods. The differences
remained even when limited
to weekdays only.
December 2009
A-336
-------
Reference Particle Sizes Measured Component Results
Zhao et al. (2007, 1561821 Personal, Micro, and Ambient: EC, Cl, Si, N03 Units = pg/m3:
PM25
Personal:
EC: 1.64
N03: 0.135
Si: 0.176
Cl: 0.116
i H
indoor.
K. -1 Q-IQ
. i .0 iy
Nfv nni?
INW3. U.U I O
Si: 0.051
Q1 n roA.
. U.UZ4
Outdoor:
EC: 1.876
N03: 0.292
Si: 0.115
Cl: 0.013
Primary Findings
Four external sources and
three internal sources were
resolved in this study.
Secondary N03" and motor
vehicle exhaust were two major
outdoor PM25 sources. Cooking
was the largest contributor to
the personal and indoor
samples. Indoor environmental
tobacco smoking also has an
important impact on the
composition of the personal
exposure samples.
Table A-63. Summary of personal PM exposure source apportionment studies.
Reference Study Design Results Primary Findings
Source apportionment of
personal and indoor central and
apartment and outdoor PM2 5.
Hopke et al. (2003, 095544) Ba|(imore re(irement home ^
10 elderly subjects.
July-Aug 1998.
Source apportionment of
personal and residences and
central outdoor PM25 around
Seattle with 10 elderly subjects
and 10 asthmatic children. The
Larson etal (2004 098145) purpose of the article was to
compare PMF2 and PMF3
methods.
Seattle
Sep 2000 and May 2001
Source apportionment of
personal and residential indoor
and residential outdoor and
central outdoor PM2 5,
Zhao etal. (2006. 156181)
Raleigh and Chapel Hill NCwith
38 subjects.
Summer 2000 and Spring 2001.
% control P,l, C,0
External
Secondary
SO,1"
Unknown
Soil
46.3,
13.6,
2.8,
64.0,
14.5,
3.1,
79.0,
17.4,
3.6,
64.0
14.5
3.1
nternal
Gypsum
Activity
Personal
care
0.7,
36.2,
0.4,
0.4,
17.8,
0.3,
0.0,
0.0,
0.0,
0.0
0.0
0.0
PMF2:
% control P,l,0
Veg burn
Mobile
Fuel oil
S, Mn, Fe
Secondary
Cl-rich
Crustal
Crustal 2
28.8, 47.6, 56.7
0.0, 3.6, 7.5
0.0, 0.0, 6.7
8.1, 0.0, 0.0
0.0, 34.5, 20.9
9.9,3.6, 3.7
25.2, 10.7, 4.5
27.9, 0.0, 0.0
PMF3:
% control P,l, 0
Veg burn
Mobile
Secondary
Crustal
41.0, 57.4, 71.3
7.2,4.3, 8.2
19.3, 13.8, 18.0
32.5, 24,5 2.5
% control P, 1, R,0
Motor
vehicle
Soil
Secondary
S0<2~
Secondary
NO,"
ETS
Personal
care and
activity
CU-factor
mix with
indoor soil
Cooking
10.0,
3.5,
15.9,
4.4
7.0,
8.0,
0.4,
52.5,
9.4,
3.7,
22.5,
,4.7,
10.0,
19.1,
1.2,
53.6,
17.2,
9.3,
59.3,
61.9
7.6,
0.0,
0.0,
0.0,
0.0,
19.4
8.5
7.8
0.0
0.0
0.0
0.0
63% of personal exposure
could be attributed to outdoor
sources (with 46% from S042),
and resuspension of indoor PM
during vacuuming, cleaning, or
other activities contributed 36%
of personal exposure.
Results showed that vegetative
burning was the largest
contributor to personal
exposure and that was related
to outdoor combustion. Crustal
exposures were related to
indoor activities.
Secondary sulfate was the
largest ambient source and the
largest ambient contribution to
personal exposure. Cooking
produced the largest
contribution to personal and
indoor concentrations. Note that
sums over 100% because
multiple sources obscured PMF
resolution.
December 2009
A-337
-------
Reference
Study Design
Results
Primary Findings
Meng et al. (2007,194618)
Source apportioned infiltration for
personal and residential indoor
and residential outdoor and
central outdoor PM2.5.
Los Angeles, Houston, and
Elizabeth, NJ with 100 non-
smoking residences and
residents in each city.
In each season between summer
1999 and spring 2001 (RIOPA).
% contrOutdoor Indoor
(Outdoor Origin)
Mechanically generated 2,17
Primary Combustion43,43
Secondary Formation*55,40
'excludes nitrates
Differential infiltration of the
PM2.5 resulted in a reduction of
secondary formation products
relative to outdoors.
Reffetal. (2007.156045)
Strand et al. (2006, 089203)
Functional group distinction for
personal and residential indoor
and residential outdoor and
central outdoor PM2 5, PM25
samples from 219 homes were
used for this analysis.
Los Angeles, Houston, and
Elizabeth, NJ with 100 non-
smoking residences and
residents in each city.
In each season between summer
1999 and spring 2001 (RIOPA).
S04'~:
R,0,1, P
01.0
10.54-0.761.0
P 0.54-0.73 0.84-0.901.0
C = 0:
R,0,1, P
01.0
10.12-0.611.0
P-0.13-0.69 0.07-0.771.0
CH:
R,0,1, P
01.0
1-0.08-0.351.0
P-0.07-0.190.41-0.851.0
The main finding was that
indoor and personal levels of
CH in organic carbons were
found to be substantially higher
than outdoors. This reduced the
polarity of indoor and personal
organic carbons
Source apportionment of
personal and indoor school and
outdoor school PM25.
Zhao et al. (2007, 156182) Denver with 56 asthmatic
children.
Oct 2002-March 2003 and Oct
2003-March 2004.
% contrP 10
Secondary
SO,7"
Soil
Secondary
NOT
Motor
vehicle
Cl-based
cleaning
Cooking
ETS
4.3
6.6
9.4
13.3
2.8
54.8
9.2
8.9
4.2
2.8
26.5
0.4
30.2
2.1
9.6
12.4
40.8
26.5
0.0
0.0
0.0
The largest personal exposure
was from cooking (54.8%), but
motor vehicle emissions were
the largest outdoor contributor
(13.3%) to personal exposure.
Secondary nitrate comprised
the largest outdoor source but
accounted for only 9.4% of
personal exposure.
Using positive matrix factorization
and an extrapolation method to
estimate PM25 based on S042~Fe
components.
Denver.
Winter 1999-2000 and
2000-2001.
Estimation method, Mean (SD, range):
PMF:7.42 (1.93, 3.43-12.89)
Extrapolation Method:
Using S042": 6.38 (1.60, 3.20-10.97)
Using S042"and Fe:6.50 (1.36, 3.54-10.12)
Using S042 and Fe, temperature adjusted: 7.02
(1.48,3.79-11.02)
Using S042" (no gamma): 8.23 (2.06, 4.12-14.14)
Similar results were found with
each technique.
December 2009
A-338
-------
Table A-64. Summary of PM infiltration studies.
Reference
Study Design
Hnf
I/O
Allen et al. (2003, 053578)
Objective: Enhance knowledge of the
outdoor contribution to total indoor
and personal PM exposures.
Methods: Continuous light scattering
monitoring.
Subjects: Elderly and children
spending most of their time indoors.
Healthy individuals, elderly with
COPD or CHD and children with
asthma. 44 residences measured for
55 10-day sessions. Seattle, WA.
PM2.5avg-0.65 ±0.21
Non-heating season- 0.79 ± 0.18
Heating season-0.53 ±0.16
Open windows (mean)- 0.69
Closed windows (mean)- 0.58
All days (mean)- 0.65
Light scattering (whole peak): 0.75 ±
0.25
Light scattering (uncensored data):
0.77 ± 0.24
Sulfur concentration (slope): 0.65 ±
0.01
Arhami et al. (2009,190096)
Objective:To examine associations
between size-segregated PM, their
particle components, and gaseous
co pollutants.
Methods: Data analyzed with linear
mixed-effect models.
Subjects: Four different retirement
communities in San Gabriel Valley,
CAand Riverside, CA. 2005-2007.
PM2.5:0.38-0.57
EC: 0.64-0.82
OC: 0.60-0.98
N/A
Balasubramanian et al. (2007, Objective: PM monitoring and
156248) assessment based on analysis of
chemical and physical characteristics of
indoor and outdoor particles.
Methods : Particle number and mass
concentrations measured using real-
time particle counter and low-volume
particulate sampler.
Subjects: 3 residential indoor and 1
residential outdoor environments in
Choa Chu Kang, Singapore. May 12-
May23,2004.
Barn et al. (2008, 156252) Objective: Measure infiltration factor
from PM2.s from forest fires and
determine effectiveness of HEPA
filter.
Methods: pDR for ambient air
sampling.
Subjects: Homes affected by forest
fire or residential wood smoke. British
Columbia, Canada. 38 homes
sampled (valid samples : 1 9 winter, 1 3
summer).
N/A
PM25 (mean)
Summer:
HEPA: 0.19 ±0.20
Unfiltered: 0.61 ± 0.27
Winter:
HEPA: 0.10 ±0.08
Unfiltered: 0.28 ±0.18
Both:
HEPA: 0.13 ±0.14
Unfiltered: 0.42 ± 0.27
PM2.5: 0.93-1 .90
Chemical Species:
CI": 0.35-0.45
N02": 2.50-4.13
N03":1.41-5.41
S042": 1.21 -1.70
Na+: 0.43-0.74
NH4+: 1.43-2.39
EC: 0.75-0.96
OC: 1.04-1 .92
AM. 04-1 .92
Co: 0.86-1 .32
Cr: 1.35-2.90
Cu: 0.50-0.69
Fe: 0.30-0 .42
Mn: 0.23-0.42
Pb: 0.40-2.47
Zn: 0.59-0.81
Cd: 0.74-1 .75
Ni: 0.71-1 .32
Ti: 0.73-0.78
V: 1.01-1 .05
Mean:
Summer:
HEPA: 0.43
Unfiltered: 0.77
Winter:
HEPA: 0.21
Unfiltered: 0.36
Both:
HEPA: 0.25
Unfiltered: 0.47
December 2009
A-339
-------
Reference
Baxter et al. (2007, 092726)
Baxter et al. (2007, 092725)
Brown etal. (2008.190894)
Cao et al. (2005, 156321)
Study Design
Objective: To develop predictive
models of residential indoor air
pollutant concentrations for lower
SES, urban households. Part of
ACCESS cohort study of asthma
etiolooy
Methods: Regression analysis; mass
balance model; Finffrom slope in
univariate regression analyses.
Subjects: Lower SES populations. 43
homes, 23 homes monitored in both
seasons, 15 in the non-heating
season (May-Oct) only, 5 in heating
season (Dec-Mar ) only; 2003-2005.
Objective: To predict residential
indoor concentrations of traffic-
related air pollutants in lower SES
urban households. Part of ACCESS
cohort study of asthma etiology.
Methods: Regression modeling,
Bayesian variable selection I/O is
slope from multivariate model
Subjects: Lower statuses, urban
households in Boston, MA. 43 sites
among 39 households, 66 sampling
sessions, nonheating (May-Oct) and
heating (Dec-Mar) 2003-2005
Objective: To examine if ambient,
home outdoor, and home indoor
particle concentrations can be used
as proxies of corresponding personal
exposure.
Methods : Associations characterized
using univariate mixed effects models
that included a random subject term.
Subjects: 15 participants in Boston,
MA in winter (Nov. 1999-Jan. 2000)
and summer (June-July 2000).
Objective: To determine relationships
and distributions of indoor and outdoor
PM25, OC, and EC. To determine
indoor/outdoor sources of indoor
carbonaceous aerosol.
Finf
PM25:0.91±0.23
EC: 0.72 ±0.49
Ca: 0.56 ±0.30
Fe: 0.38 ±0.26
K: 0.83 ±0.52
Si: 0.02 ±0.00
Na: 0.46 ±0.43
Cl: 0.40 ±0.12
Zn: 0.85 ±0.28
S:0.95 ±0.78
V: 0.60 ± 0.77
N/A
N/A
N/A
I/O
PM25 (mean, coefficient of variation
(CV)): 1.14 (0.71)
EC: 0.89 (0.64)
Ca: 1.16 (1.90)
Fe: 0.69 (1.40)
K: 1.10 (0.95)
Si: 1.04 (1.31)
Na: 1.05 (1.84)
0:3.18(3.79)
Zn: 0.83 ±(1.13)
S: 0.76 ± (0.32)
V: 0.76 ± (0.46)
PM25:
Open Wndows: 0.98
Closed Wndows: 0.64
EC: 0.38
PM25:
Wnter: Median: 1.2, Range: 0.8-1.8
Summer: Median: 0.9, Range: 0.6-1.2
EC:
Wnter: Median: 1.1, Range: 0.7-4.5
Summer: Median- 1.0, Range: 0.9-1.3
S042':
Wnter: Median: 0.5, Range: 0.3-0.8
Summer: Median: 0.8, Range: 0.4-1.0
20minPM25:
Roadside: 0.7-4.0
Urban: 0.9-6.7
Rural: 0.5-1. 7
Methods: Gravimetric analysis to
determine PM2.? concentrations. OC
and EC determined by TOR following
IMPROVE protocol.
Subjects: 6 residences in Hong Kong
(2 roadside, 2 urban, 2 rural). March
6-ApriM8,2004.
Roadside: 0.8-1.4
Urban: 1.2-2.0
Rural: 1.0-1.8
OC (average and range):
Roadside: 1.9 (1.1-2.3)
Urban: 2.3 (1.5-4.0)
Rural: 1.3 (1.2-2.2)
EC (average and range):
Roadside: 1.0 (0.9-1.1)
Urban: 1.1 (0.8-1.3)
Rural: 1.1 (0.9-1.8)
December 2009
A-340
-------
Reference
Study Design
Hnf
I/O
Cortez-Lugo et al. (2008,156368)
Objective:To determine personal N/A
PM2.5 and its relationship with outdoor
and indoor PM2.5 and PM10.
Methods: Linear regression model
used to compare personal and indoor
PM2.5. I/O variation studied using
analysis of variance and predictors
determined by generalized estimating
equation models. I/O PM2.5 ratio
transformed into natural logarithm.
Subjects: 38 nonsmoking long-time
Mexico residents with COPD. Mexico
City, Mexico. Feb-Nov 2000.
PM25:
Average: 1.2
Range: 0.05-6.1
Diapoulietal. (2008,190893)
Objective: To characterize the PM10f
PM2.5, UFP concentrations at primary
schools. To examine the relationship
between indoor and outdoor
concentrations.
Methods: Chemical analysis of
collected filters. Regressions to
examine correlations between indoor
and outdoor concentrations.
Subjects: 7 primary schools with
different characterizations of
urbanization and traffic density in
Athens, Greece. No ventilation
system. Nov. 2003-Feb. 2004 and
Oct.-Dec. 2004.
N/A
PM10:0.54-2.46
PM2.5- 0.67-2.77
UFP- 0.33-0.74
Dimitroulopoulou et al. (2006,
090302)
Objective: To develop a probabilistic N/A
indoor air model (INDAIR).
Methods: INDAIR predicts frequency
distributions of concentrations of up to 4
pollutants simultaneously (N02, CO,
PMio, PM25). 3 scenarios: no source,
gas cooking, smoking.
Subjects: 5 UK sites- Harwell (rural),
Birmingham East (urban
background), Bradford (urban
center), Bloomsbury (urban center),
Marylebone Road (roadside). Winter
(October 1-March 31), summer (April
1-September 30), 1997-1999.
No source:
PM10:0.5-0.65;
PM2.5:0.6-0.7
Gas cooking:
PM10:0.6-0.9 (bedroom), 1.0-2.0
(lounge), 1.6-4.3 (kitchen);
PM25:0.74-0.9 (bedroom), 0.9-1.6
(lounge), 1.6-2.9 (kitchen)
Smoking:
PM10: 0.7-1.1 (bedroom), 1.1-2.7
(lounge), 1.1-2.5 (kitchen);
PM25:0.8-1.3 (bedroom), 1.3-2.8
(lounge), 1.4-2.6 (kitchen)
Frommeetal. (2008,155147)
Objective: To characterize the
chemical and morphological
properties of PM in classrooms and in
corresponding outdoor air.
Methods: PM Finf derived from sulfate
Finf and a correction factor that
results from division of BPM
(increase of indoor PM per outdoor
PM, linear relationship) by Bsulf
(increase of indoor sulfate per
outdoor sulfate, linear relationship). If
no indoor source, the sulfate Finfis
equal to the sulfate I/O.
Subjects: Primary school in northern
Munich. Densely populated
residential area 160m away from a
very busy street. Classrooms had 21-
23 students. Sampling during
teaching hours. Oct.-Nov. 2005.
N/A
PM,,:
S04?": 0.3,
N03':0.1,
CI": 0.6,
Na2*: 0.9,
NH4*:0.1,
Mg: 0.6,
Ca2*: 1.4,
EC: 0.7,
OC:1.1
PM,5:
SO?': 0.4,
N03":0.2,
CF: 0.5,
Na2*: 0.6,
NH4+:0.3,
Mg: 0.5,
Ca2*: 1.6
December 2009
A-341
-------
Reference
Study Design
Hnf
I/O
Guoetal. (2004.156506)1
Objective:To investigate pollutant
concentrations at air-conditioned and
non-air-conditioned markets. To
compare indoor air quality with the
Hong Kong standard.
Methods: PMio concentrations
measured by Hi-Vol sampler
correlated with corresponding levels
measured by Dust-Trak monitor.
Subjects: 3 non-air-conditioned and 2
air-conditioned markets in Hong
Kong. Sept. 2001-Jan. 2002.
N/A
PM10:
Non-air-conditioned: -0.7, Air-
conditioned: -0.98
Hanninen et al. (2004,056812)2
Objective: To assess indoor PM2.5 by
origin and potential determinants.
Methods: Part of EXPOLIS study.
Pump and filter with gravimetric
analysis. Univariate single and
stepwise-multiple regression
analyses.
Subjects: Residential homes in
Athens, Greece; Basle, Switzerland;
Helsinki, Finland; Prague, Czech
Republic. Homes by city: Athens 50,
Basle 50, Helsinki 189, Prague 49.
PM25(mean):
Athens-0.70 ±0.12
Basle-0.63 ±0.15
Helsinki-0.59 ±0.17
Prague-0.61 ±0.14
S (mean):
Athens-0.82 ±0.14
Basle-0.80 ±0.19
Helsinki-0.70 ±0.20
Prague-0.72 ±0.16
PM25:
Athens: -0.84
Basle:-1.37
Helsinki:-1.30
Prague:-1.33
S:
Athens:~0.70
Basle: -0.80
Helsinki: -0.74
Prague: -0.77
Ho et al. (2004, 056804)3
Objective: PM2.5, OC, and EC
exposure assessment of occupied
buildings located near major
roadways under natural ventilation
(NV) and mechanical ventilation
(MV).
Methods: Co-located mini-volume
samplers and Partisol model 2000
sampler with 2.5 micron inlet.
IMPROVE TOR carbon analysis.
Subjects: Occupants of MV (1
classroom and office) and NV (3
residences) buildings located within
10m of major roadway; Hong Kong,
China. Sep. 2002-Feb. 2003.
PM25:0.42
EC: MV: 0.42, NV: 0.76
OC: MV: 0.66, NV: 0.71
PM25 (average): 0.2-1.6
MV (average): <0.7
NV (average): 0.6-1.6
EC: Range: 0.5+0.1-1.1+0.4
OC: Range: 0.6+0.2-1.5+1.0
Hoek et al. (2008,156554)
Objective: Exposure assessmentof
indoor/outdoor particle
relationships.RUPIOH study.
Methods: Sampling by condensation
particle counters and Harvard
impactors. Gravimetric analysis and
reflectance. Calculations performed
for 24h avg concentrations. Finf
estimated by linear regression
analysis.
Subjects: 4 European cities (Helsinki,
Finland; Athens, Greece; Amsterdam,
The Netherlands; Birmingham,
England). Urban populations. >35yrs.
Asthma or COPD. Non-smoking
households. Work <16h/wk outside
home. 153 homes sampled Oct.
2002-Mar. 2004.
Regression slope for indoor vs. central N/A
site outdoor:
PM25: 0.30-0.51
PM10:0.17-0.41
PM10-2.5:0.01-0.17
S042": 0.59-0.78
Soot: 0.43-0.87
Regression slope for indoor vs.
residential outdoor:
PM2.5:0.34-0.48
PM10:0.26-0.44
PM10-2.5:0.11-0.16
Soot: 0.63-0.84
December 2009
A-342
-------
Reference
Study Design
Hnf
I/O
Hopke et al. (2003, 095544)
Objective: To use advanced factor
analysis models to identify and
quantify PM sources. 1998 BPMEES
data.
Methods: PEM, outdoor and indoor
sampling of unoccupied apartment in
retirement facility. PMF used to derive
source contributions. Multilinear
Engine used to derive joint factors.
Subjects: 10 non-smoking elderly
subjects of mean age 84 who did not
cook.Towson, MD. July 26-Aug. 22,
1998.
N03"-SO/:0.03
S042': 0.38
OC: 0.77
MV Exhaust: 0.32
N/A
Hystad et al. (2008,190890)
Objective: To explore the feas ibility of Seattle:
modeling residential PM2s Finf for
occupied residences using data Mean (all>: 0.59 ± 0.21
readily available for most of North
America.
N/A
Mean (detached residences): 0.60 +
0.20
Methods: Finf calculated by recursive
mass balance model where Finf is a
function of penetration efficiency,
particle removal rate, and air
exchange.
Subjects: 46 residences in Seattle,
WA1999-2003. 38 nonsmoking
residences in Victoria, British
Columbia, Canada 2006. Heating
(Oct.-Feb.) and nonheating (March-
Sept.).
Victoria:
Mean (all): 0.62 ± 0.22
Mean (detached residences): 0.59 +
0.22
Klinmalee et al. (2008.190888)
Objective: To monitor indoor and N/A
outdoor pollution in an university
campus and shopping center.
Methods: PM measured by PEM and
quartz filters. Analyzed for mass,
water soluble ions by ion
chromatography, and black carbon by
a smokestain reflectometer. I/O
calculated for each sample pair then
average taken.
Subjects: University campus and
shopping center in northern suburb of
Bangkok, Thailand. Dec. 2005-Feb.
2006.
PM2.5:
University:
Weekdays: 0.6, Weekends: 0.5
Shopping center:
Weekdays: 1.5, Weekends: 2.0
BC in PM2.5:
University: 0.9
Shopping center: 0.67
December 2009
A-343
-------
Reference
Study Design
mf
I/O
Koistinen et al. (2004, 156655)
Objective:To identify PM2.5 sources in
personal exposures with principal
component analysis of the elemental
compositions in residential outdoor,
indoor, and workplace indoor
microenvironments. Part of EXPOLIS
study.
Methods: Principal component
analysis to identify sources of
microenvironmental and personal
PM2.5 exposure. Specific mass
contributions of sources calculated by
source reconstruction.
Subjects: Non-smoking, 25-55yrs.
Helsinki, Finland. Oct. 1996-Dec.
1997.
N/A
Median seasonal:
PM25:
Wnter: 0.77, Spring: 1.03, Summer:
0.95, Fall: 0.92, Total: 0.92
Pb:
Wnter: 0.67, Spring: 0.56, Summer:
0.86, Fall: 0.69, Total: 0.67
S:
Wnter: 0.60, Spring: 0.63, Summer:
0.90, Fall: 0.75, Total: 0.69
Wnter: 0.57, Spring: 0.72, Summer:
0.98, Fall: 0.89, Total: 0.77
BS:
Wnter: 0.65, Spring: 0.67, Summer:
0.91, Fall: 0.88, Total: 0.79
In:
Wnter: 0.58, Spring: 0.75, Summer:
0.66, Fall: 0.75, Total: 0.68
Fe:
Wnter: 0.52, Spring: 0.96, Summer:
0.90, Fall: 0.95, Total: 0.83
K:
Wnter: 0.95, Spring: 1.05, Summer:
1.01, Fall: 1.08, Total: 1.05
Cl:
Wnter: 1.01, Spring: 1.24, Summer:
1.37, Fall: 1.74, Total: 1.24
Al:
Wnter: 1.19, Spring: 1.08, Summer:
1.41, Fall: 2.20, Total: 1.27
Objective: To establish effects of
evaporative coolers on indoor PM
concentrations.
Methods: Concurrent 10min avg
indoor and outdoor concentrations
recorded for 2 days.I/O determined
by equation based on mass
conversation principles.
Subjects: 10 homes with evaporative
coolers. El Paso, TX. June 22-Aug.
23,2001.
N/A
PM10:
All: 0.60
Cooler On: 0.57
Cooler Off: 0.66
PM25:
All: 0.65
Cooler On: 0.63
Cooler Off: 0.73
Lundenetal. (2008,155949)
Objective: To investigate the
physiochemical processes that
influence the transport and fate of
outdoor particles to the indoor
environment.
Methods: I/O calculated from
measurements of aerosols collected
on quartz filters.
Subjects: 3-bedroom single-story
unoccupied house in Clovis, CA. 3
periods: Oct. 9-23, 2000; Dec. 1-19,
2000; Jan. 12-23,2001.
N/A
PM25: Oct.: 0.46 ±0.2, Dec.: 0.39 ±0.2,
Jan.: 0.38 ±0.3, All periods: 0.41 ±0.2
Carbon: Oct.: 0.50 ±0.1, Dec.: 0.46 ±
0.1, Jan.: 0.52 ±0.2, All periods: 0.50 ±
0.2
OC: Oct.: 0.48 ±0.1, Dec.: 0.44 ±0.1,
Jan.: 0.50 ± 0.2, All periods: 0.47 ± 0.2
Black carbon: Oct.: 0.60 ± 0.2, Dec.:
0.60 ± 0.2, Jan.: 0.65 ± 0.2, All periods:
0.61 ±0.2
Macintosh et al. (2009,190887)
Objective: To estimate the potential
for residential air cleaning systems to
mitigate exposure to fine particles of
ambient origin.
Methods: Multi-zone indoor air quality
model to examine annual, 24h avg
and diurnal concentrations of outdoor
PM2.5 in residential indoor air.
Subjects: Homes in Cincinnati,
Cleveland, and Columbus, OH that
have natural ventilation, forced air
heating and cooling with conventional
in-duct filtration, or forces air heating
and cooling with high-efficiency in-
duct air cleaning. 2005.
N/A
PM2.5 (range):
Natural ventilation: 0.23-0.97
Forced air-conventional filtration:
0.13-0.94
Forced air-high-efficiency
electrostatic: 0.02-0.80
December 2009
A-344
-------
Reference
Study Design
Hnf
I/O
Martuzevicius et al. (2008,190886)
Objective: To determine the N/A
contribution of traffic-related PM to
the indoor aerosols.
Methods: Receptor modeling based
on a PARAFAC model.
Subjects: 6 houses 30-300mfrom a
highway, with conventional windows,
central HVAC, and with smoking and
cooking allowed. Spring: Mar. 30-May
14, 2004. Fall: Sept. 13-Oct. 22,
2004. Cincinnati, OH.
Range- PM25: Spring: 0.5 ± 0.2-2.9 ±
1.2; Fall: 0.7 ±0.1-4.7 ±6.9
EC:
Spring: 0.3 ±0.1-2.2 ±1.7;
Winter: 0.6 ±0.1-1.3 ±0.7
OC:
Spring: 1.0+ 0.7-6.9+ 3.9;
Winter: 1.2 ±0.1-7.6 ±10
Si:
Spring: 0.4+ 0.1-5.1 +3.9;
Winter: 0.5+ 0.1-5.3+ 4.5
S:
Spring: 0.4 ±0.1-0.7 + 0.1;
Winter: 0.5+ 0.1-0.9+ 0.4
Mn:
Spring: 0.3+ 0.2-0.8+ 0.6;
Winter: 0.3+ 0.2-1.0 + 0.2
Fe:
Spring: 0.3+ 0.0-1.3+ 0.8;
Winter: 0.4+ 0.1-0.9+ 0.6
In:
Spring: 0.3 ±0.1-0.7 + 0.6;
Winter: 0.6+ 0.1-1.1 +0.8
Br:
Spring: 0.3+ 0.1-1.0 + 0.5;
Winter: 0.2+ 0.1-0.9+ 0.6
Pb:
Spring: 0.3+ 0.3-0.9+ 0.6;
Winter: 0.2+ 0.2-1.9+ 2.3
Meng et al. (2005, 058595)
Objective: Analyses of RIOPAdata, PM25- 0.46
which investigated relationships
between indoor, outdoor, and
personal exposure for several air
pollutants.
Methods: PM measured on Teflon
filters collected by PEMs for 48h. The
mass balance model and RCS
statistical model used to estimate
indoor and and personal PM
concentrations.
Subjects: 212 nonsmoking homes
sampled. Houston, TX; Los Angeles
County, CA; Elizabeth, NJ. Summer
1999-spring 2001, all 4 seasons.
Los Angeles:
PM25: Mean: 0.84, Median: 0.90;
EC: Mean: 0.93, Median: 0.92;
OC: Mean: 1.32, Median: 1.31
Elizabeth:
PM25: Mean: 0.99, Median: 0.86;
EC: Mean: 1.0, Median: 0.85;
OC: Mean: 2.4, Median: 1.8
Houston:
PM25: Mean: 1.16, Median: 1.02;
EC: Mean: 1.0, Median: 0.71;
OC: Mean: 2.25, Median: 2.35
Molnaretal. (2007,156774)
Objective: To characterize and
compare indoor and outdoor PM2.5
trace element concentrations in
difference microenvironments related
to children.
Methods: Elemental concentrations
analyzed using X-ray fluorescence
spectroscopy.
Subjects: 40 sampling sites (10
classrooms in 5 schools, 10
preschools, 20 non-smoking homes).
3 communities in Stockholm,
Sweden. Sampled once during spring
and once during winter. Dec. 1, 2003-
July 1,2004.
PM25 (containing S or Pb): 0.4-0.9
S (median):
Both seasons: 0.61 (homes), 0.53
(schools), 0.69 (preschools);
Winter: 0.47 (homes), 0.36 (schools),
0.63 (preschools);
Spring: 0.63 (homes), 0.55 (schools),
0.90 (preschools)
Pb (median):
Both seasons: 0.70 (homes), 0.59
(schools), 0.70 (preschools);
Winter: 0.62 (homes), 0.43 (schools),
0.63 (preschools);
Spring: 0.70 (homes), 0.64 (schools),
0.75 (preschools)
December 2009
A-345
-------
Reference
Study Design
Hnf
I/O
Ngetal. (2005,155996)
Objective:To estimate PM exposures N/A
following the September 11,2001
attack in NYC.
Methods: Outdoor PM25 interpolated
and used in a deterministic micro-
environmental model (INTAIR) to
simulate analytically concentrations in
indoor micro-environments. Linear
regression equations used.
Subjects: Lower Manhattan residents
divided into representative individuals
- home-maker, office/shop-worker,
student/child. Estimates Sept. 14-31.
Mean I/O in home simulated with
INTAIR:
No Source: 0.6
Smoking: 1.9
Cooking: 1.3
Smoking and Cooking: 2.3
I/O of micro-environments simulated
by analytical and empirical methods
(no indoor source):
Office/Shop: 0.4
Classroom: 0.9:
Transport Area: 1.9
Store: 1.2
Pascholdetal. (2003,156847)
Objective: To identify PM sources
inside homes with evaporative
coolers.
Methods: PM element composition
analysis by ICP-MS.
Subjects: 10 residences. El Paso, TX.
Summer 2001.
N/A
PM10:
Na: 0.33, Mg: 0.43, Al: 0.50, K: 0.48,
Ca: 0.40, Ti: 0.52, Mn: 0.48, Fe: 0.46,
Cu: 0.74, Zn: 0.52, Ba: 0.54, Pb: 0.76
PM2.5:
Na: 0.20, Mg: 0.29, Al: 0.34, K:0.30,
Ca: 0.52, Ti: 0.40, Mn: 0.35, Fe: 0.30,
Cu: 0.67, Zn: 0.34, Ba: 0.47, Pb: 0.51
Polidorietal. (2007,156877)
Objective: To investigate the
relationships of indoor and outdoor
PM2.s, its components, seasonal
variations, and gaseous copollutants.
PM2.
Only l/0's<1 considered
July 6-Aug. 20:0.71 ±0.10; Aug. 24-
Oct. 15:0.60 ± 0.05; Oct. 19-Dec. 10:
0.59 ±0.07; Jan. 4-Feb. 18:0.45 ±
Methods: Finf estimated by analysis of 0.06
l/0's and a recursive model
technique. OC:
Subjects: 2 retirement facilities in Los
Angeles, CA. July 6-Aug. 20, 2005.
Aug. 24-Oct. 15, 2005. Oct. 19-Dec.
10, 2005. Jan. 4-Feb. 18, 2006.
July 6-Aug. 20:0.86 ± 0.05; Aug. 24-
Oct. 15:0.77 ± 0.09; Oct. 19-Dec. 10:
0.82 ±0.07; Jan. 4-Feb. 18:0.64 ±
0.10
EC:
July 6-Aug. 20:0.73 ± 0.07; Aug. 24-
Oct. 15:0.71 ±0.05; Oct. 19-Dec. 10:
0.77 ±0.06; Jan. 4-Feb. 18:0.64 ±
0.10
Ramachandran et al. (2003,195017)
Objective: To examine variability in N/A
measurements of 24h avg and 15min
avg PM2.5 concentrations.
Methods: Linear regression of
gravimetric measurements.
Subjects: 3 urban residential
neighborhoods in Minneaopolis-St.
Paul, MN. 9-10 nonsmoking
residences. Spring (April 26-June 2),
summer (June 20-Aug. 10), fall (Sept.
23-Nov.20)of1999.
24h avg:
Mean: 1.7, Median: 1.3, Standard
deviation: 1.6
15minavg:
Mean: 2.7, Median: 1.2, Standard
deviation: 8.7
Rojas-Bracho et al. (2004, 054772)
Objective:To examine determinants
of personal exposure to PM2 5, PMio,
PM2.5-10.
Methods: 2 sets of mixed models.
Personal exposures modeled as
dependent variables. Subject
variability modeled using random
effects. Explanatory variables and
season modeled as fixed effects.
Subjects: 18 COPD subjects in
nonsmoking households. Boston,
MA. Winters of 1996 and 1997,
summer of 1996.
N/A
PM2.5:
Winter: Mean: 1.58, Median: 2.11;
Summer: Mean: 1.08, Median: 0.88
PM10:
Winter: Mean: 2.02, Median: 3.77;
Summer: Mean: 1.14, Median: 1.05
PM2.5-10:
Winter: Mean: 2.65, Median: 3.59;
Summer: Mean: 1.26, Median: 1.39
December 2009
A-346
-------
Reference
Study Design
Hnf
I/O
Sarnatetal. (2006. 089166)
Objective: To assess the ability of
outdoor PM2.5 its volatile and
nonvolatile components and particle
sizes to infiltrate indoors.
Methods: PM2.5 mass contributions
estimated by the mean concentration
ratio between each component and
PM2.5. Indoor and outdoor particle
concentrations relationships
examined by Spearman correlation
coefficient. I/O concentration ratios
used during overnight (nonsource)
period to estimate fraction of ambient
particles remaining airborne indoors
Subjects: 17 occupied, nonsmoking
Los Angeles, CA residences. July 28,
2001 -Feb. 25, 2002.
PM2.5:
Median: 0.48, Interquartile range:
0.39-0.57
BC:
Median: 0.84, Interquartile range:
0.70-0.96
UFP (0.02-0.03 urn):
Median: 0.50, Interquartile range:
0.39-0.60
UFP (0.08-0.3 urn):
Median:-0.75
Coarse particles (5-10 urn):
Median: <0.17
PM2.5:
Overnight: 0.40-0.57, Morning: 0.43-
0.74, Afternoon: 0.45-0.90, Evening:
0.42-0.82
BC:
Overnight: 0.70-0.97, Morning: 0.67-
0.98, Afternoon: 0.77-1.04, Evening:
0.70-1.01
Stranger et al. (2008,190884)
Objective: To assess indoor air
quality by determining indoor and
outdoor PM2.5 mass concentrations,
elemental composition, and gaseous
compounds.
Methods: PM mass concentrations
determined gravimetrically.
Subjects: 27 primary schools in city
center and suburbs of Antwerp,
Belgium. Dec. 2002 and June 2003.
N/A
PM2.5:
Urban: Range: 0.3-6.9, Average: 1.3;
Suburban: Range: 0.2-8.8, Average:
2.3
V, Ni.Zn, Pb, Br, Mn:<1
CI,Ca,AI,Si,K,Ti, Fe:>1
BS: Urban: Average: Dec.- 0.7 + 0.1,
June-1.1 ±0.3; Suburban: Dec.-0.8 +
0.2, June-1.0 + 0.4
Stranger et al. (2009,190883)
Objective: To assess indoor air N/A
quality in residences by quantifying
various gaseous pollutants, and PM
mass concentrations, elemental
composition, and water-soluble ionic
content.
Methods: PM mass concentrations
gravimetrically determined. Elemental
bulk analysis on filters.
Subjects: 19 residential homes in
Antwerp, Belgium that were a subset
of participants in the ECRHS II study.
PM,: Houses 1-15: Average: 2.0,
Range: 0.3-9.6; Smokers: Average:
3.9, Range: 1 .2-9.7; Non-smokers:
Average: 0.8, Range: 0.3-14
PM2.5: Houses 1-15: Average: 1.5,
Range: 0.4-5.4, Smokers average:
2.5, Smokers range: 1 .2-5.4, Non-
smokers average: 0.8, Non-smokers
range : 0.4-1 .3 ; Houses 16-19:
Average: 2.6, Range: 0.3-3.9
PM10: Houses 1-15: Average: 1.3,
Range: 0.4-4.1 , Smokers average:
2.1, Smokers range: 1.1-4.1, Non-
smokers average: 0.8, Non-smokers
range: 0.4-1 .2
Ca,Ti,V,Cr, Mn.Fe, Ni.Zn, Pb.Si,
K,Cu,Br,AI:>1
Turpin et al. (2007,157062)
Objective: To characterize and
compare outdoor, indoor, personal
PM2.5 exposure. Identify indoor and
personal PM25 sources. Estimate
outdoor PM2.5 effect on indoor and
personal PM2.5. RIOPA study.
Methods :Finf calculated in three
ways: RCS model used to obtain
constant Finf. Mass balance model
shows Finf varying with AER. Robust
regression uses major PM2.5 species
for home-specific Finf.
Subjects: 309 nonsmoking adults and
118 children with no preexisting
conditions. 219 homes sampled.
Elizabeth NJ, Houston TX, and Los
Angeles County CA.
PM2.5:
RCS model: 0.46
Least-Trimmed Squared Regression:
Mean: 0.69, Median: 0.70,SD:0.23
Mass Balance Model: ~0.08-~0.85
Los Angeles:
PM25: Mean: 0.84, Median: 0.90
EC: Mean: 0.93, Median: 0.92
OC: Mean: 1.32, Median: 1.31
Elizabeth:
PM2.5: Mean: 0.99, Median: 0.86
EC: Mean: 1.0, Median: 0.85
OC: Mean: 2.4, Median: 1.8
Houston:
PM25: Mean: 1.16, Median: 1.02
EC: Mean: 1.0, Median: 0.71
OC: Mean: 2.25, Median: 2.35
December 2009
A-347
-------
Reference
Study Design
Hnf
I/O
Wallace and Williams (2005, 057485)
Objective: To estimate the
contribution of outdoor PM2.5 to
personal exposure in high-risk
subpopulations.
Methods: Longitudinal regressions of
estimated indoor and outdoor PM2.5
for Finf.
Subjects: 29 African-Americans with
hypertension and 8 with implanted
cardiac defibrillators. Measured
7d/season, 4 seasons in 2000-2001.
Raleigh, NC.
Range: 0.35-0.87
PM2.5:
Mean: 1.08 ± 1.05, Median: 0.75,
Range: 0.24-9.48
S:
Mean: 0.59 ± 0.16, Median: 0.58,
Range: 0.17-1.06
Williams etal. (2003,053338)
Objective:To estimate ambient PM2.5
contributions to personal and indoor
residential PM mass concentrations.
Methods: Finf estimated from least
squares, regression analysis, and
mixed model slope.
Subjects: Nonsmoking, ambulatory, >
SOyrs. 2 cohorts: mostly Caucasian
with implanted cardiac defibrillators in
Chapel, NC; 30 African-Americans
with controlled hypertension in low-to-
moderate SES neighborhoods in
Raleigh, NC. 7d/season, 4 seasons in
2000-2001.
Least squares estimate of indoor
filtration factors:
Mean: 0.42 ± 0.38, Range: -0.55 to
1.62
Regression analysis: 0.43 ± 0.06
Mixed model slope: Mean- 0.45 ±
0.21, Range-0.05-0.94
N/A
Williams etal. (2008.191201)
Objective: To examine the spatial
variability of PM2.5 and PMiq-2.s and
their components to determine the
suitability of conducting health
outcome studies using a central site
monitor in a metropolitan area having
multiple source impacts.
Methods: Gravimetric analysis of PM
mass concentrations. ED-XRF
analysis of PM elements.
Subjects: Non-smoking, ambulatory,
and living in detached homes and
non-smoking households. Detroit, Ml.
PM2.5:
Range: 0.16-6.45, Mean: 0.7 ± 0.33,
Median: 0.70 (indicate indoor sulfur
source when Finf>1)
N/A
Wilson and Brauer (2006. 088933) Objective: To provide additional SO/: 0.72
insight into factors affecting exposure
to airborne PM and the resultant
health effects.
Methods: Finf estimated by mass
balance equation.
Subjects: 16 nonsmoking subjects
with COPD. 54-86yrs. Vancouver,
British Columbia. April-Sept. 1998.
N/A
Wu et al. (2006,179950)
Objective: To assess personal PM2.5
exposures from ambient sources and
agriculture burning smoke.
Methods: Finf estimated by PCS model.
Application of Robust regression
algorithm.
Subjects: 33 adult asthmatics. 18-
52yrs. Pullman, WA. Sept. 3,2002-
Nov. 1,2002.
Range: 0.25-0.94
N/A
December 2009
A-348
-------
Reference
Study Design
Hnf
I/O
Yang et al. (2009,190885)
Objective: To characterize the N/A
concentrations of different indoor air
pollutants.
Methods: PM10 collected on pall flex
membrane filer using MiniVol portable
air samplers. Arithmetic and
geometric means calculated for
indoor concentrations. Differences in
concentrations measured by Kruskal-
Wallis test.
Subjects: 55 schools in 6
metropolitan areas in Korea. Samples
from a classroom, laboratory, and
computer classroom. 3 seasons,
July-Dec. 2004.
PM10:
Classroom: Summer: 1.98, Autumn:
2.25, Winter: 2.07, Total: 2.06
Laboratory: Summer: 1.33, Autumn:
1.32, Winter: 1.72, Total: 1.46
Computer classroom: Summer- 0.77,
Autumn: 1.43, Winter: 2.08, Total:
1.43
Zhuetal. (2005.190081)
Objective:To determine penetration
behavior of outdoor ultrafine particles
into indoor environments in areas
close to freeways.
Methods: Dynamic mass balance
model.
Subjects: 4 2-bedroom apartments
within 60m from the center of the 405
Freeway in Los Angeles, CA. Non-
smoking tenants. 2 sampling periods
(non-cooking, non-cleaning): Oct.-
Dec. 2003 and Dec. 2003-Jan. 2004.
N/A
Highest (largest ultrafine particles-
70-100nm): 0.6-0.9
Lowest (smallest ultrafine particles-
10-20nm): 0.1-0.4
. I/O estimated from Figure 8 in study.
I/O calculated from indoor and outdoor concentrations in Table 1 in study.
Fjnf measured by coefficient of determination, R .
4. RIOPA calculated l/0's.
. I/O calculated from mean and median indoor and outdoor concentrations listed in Table 1 of study.
6. l/0's estimated from Figure 3 in study.
Mean and median I/O concentrations calculated from all residences in study.
. Finf estimated from Figure 2 in study.
. Finf presented in box plot (Figure 8), however data is difficult to deduce. No numeric values reported.
December 2009
A-349
-------
Table A-65. Summary of PM - copollutant exposure studies.
Reference PM metric Copollutant metric Association between PM and copollutant Primary findings
Fruin et al. (2008, In-vehicle UFP, BC, PM- In-vehicle NOX, CO
0971831 bound PAH
Schwartz et al. (2007, Ambient and personal Ambient and personal 03
0902201 PM2 .5 data from the and N02 data from the
Baltimore panel study Baltimore panel study.
Tolbert et al. (2007, Ambient PM10, PMi0-25, Ambient 03, N02, CO,
0903161 PM25, EC, OC.TC, S042" S02
, water-soluble metals,
oxygenated
hydrocarbons
R
UFP
PM25
NO
BC
CO
C02
UFP
1
PM,,
071
1
NO
0.97
0.69
1
BC
0.95
0.89
0.91
1
CO CO,
0.63 0.72
0.66 0.68
0.78 0.85
0.65 0.74
1 0.94
1
\lote that these correlations are computed from data
presented by Fruin et al. (2008, 0971831 for mean
concentrations at different loc ations.
Median p for regressions:
Personal
PM25
Personal
PM25of
ambient origin
Personal
so/-
Personal 03
Personal N02
Ambient Ambient 03
PM25
0.0143 -0.0016
0.0183 -0.0037
0.0051 0.0035
0.0014 0.0010
0.0015 0.0009
Amb ent
N02
0.0115
0.0124
0.0006
0.0009
0.0010
PM10
03
N02
CO
SO,
PMc
PM25
SO/
EC
OC
TC
Metals
OHC
sor
SO^
EC
OC
TC
Metals
OHC
PM10
1.0
0.6
0.5
0.5
0.2
07
0.8
07
0.6
07
07
07
0.5
EC
1.0
0.3
0.3
0.3
07
0.5
03
1.0
0.4
0.3
0.2
0.4
0.6
0.6
0.4
0.5
0.5
0.4
0.4
OC
1.0
0.8
0.9
0.5
0.4
N02
1.0
0.7
0.4
0.5
0.6
0.1
0.6
0.6
0.7
0.3
0.2
TC
1.0
1.0
0.5
0.4
CO
1.0
0.3
0.4
0.4
0.1
0.7
0.6
0.6
0.4
0.3
Metals
1.0
0.5
0.4
S02 PMc
1.0
0.2 1.0
0.2 0.5
0.1 0.3
0.2 0.5
0.2 0.5
0.2 0.5
0.1 0.5
0.1 0.4
OHC
1.0
0.5 1.0
Brook et al. (2007, Anbient PM10, PM10-2 5, Ambient N02, NO Sn '^nrwi m^ nm '
091153) PM25,S042-,andtrace ^100(100,100)
metals in 10 Canadian KK&)
PM10-2 5 0.31 (0.04, 0.50)
PM,n 0.50 (0.23 0.70)
S04 0.33(0.10, 0.48
Fe 0.44 (0.29, 0.56)
Zn 0.39 (0.28, 0.52)
NiO.20 (0.06, 0.40)
Mn 0.51 (0.37 0.62)
As 0.21 (0.07, 0.39)
AID. 07 (-0.1 7, 0.18)
Cu 0.03 (-0.07 0.15)
PbO.28 (0.16, 0.39)
310.19(0.00,0.32)
Se 0.1 4 (-0.04 0.35)
Measurements of
freeway UFP, BC, PM-
bound PAH, and NO
concentrations were
roughly one order of
magnitude higher than
ambient
measurements.
Multiple regression
analysis suggests
these concentrations
were a function of truck
density and total truck
count.
Results suggest that
ambient 03 exposure
may be related to
personal S042~
exposure but not to
personal PM25
exposure on the whole.
Ambient N02 exposure
was associated with
personal PM25
exposure, possibly
because both have
traffic sources.
PM consliluenls.
uomponents were
used in a multi-
pollutant model to
^ oredict emeroencv
07 lo be the mosl
0.7 significant predictor ot
0.5 cardiovascular disease
visits in one-, two-, and
lespiialuiy disease
visits i none-, two-, and
three-pollutant models.
N02 showed the
strongest association
with mortality, but it is
unclear if this
association is due to
health effects of N02 or
health effects of
copollutant PM.
December 2009
A-350
-------
Reference
PM metric
Copollutant metric
Association between PM and copollutant
Primary findings
Ito et al. (2007,
1565941
Ambient PM2
Ambient 03, N02,
CO
S02, Shown in figure format only.
Authors tested
relationship between
meteorological
variables and
copollutants to
determine if multi-
pollutant models are
impacted by spatial or
temporal variation or by
meteorological
conditions.
Multicollinearity varied
by pollutant and
Kauretal. (2005,
0865041
Fixed-site and personal
PM25, personal UFP
Fixed site and personal
CO
Personal R:
PM25UFPCO
PM2510.5 0.2
UFPO.510.7
CO 0.2 0.71
Fairly low correlation
was observed between
PM25andCOand
between PM25 and
UFP, stronger
correlations between
UFP and CO.
Kaur et al. (2005, Fixed-site and personal Fixed site and personal
0881751 PM2 5 analyzed post- CO
sample for light
absorbance (as indicator
for carbonaceous
aerosol), personal UFP
Personal R:
RPM25AbsCOUFP
PM2510.3-0.1 0.0
Abs 0.310.20.7
CO -0.10.210.1
UFP 0.00.70.11
Strongest correlation
observed between UFP
and absorption, which
is reasonable given
that much absorptive
carbonaceous aerosol
is in the ultrafine range.
S0renson et al. Personal, indoor Personal, indoor
(2005, 089428) residential, and outdoor residential, and outdoor
residential PM25 and BC residential N02
Sabin et al. (2005, BC, particle-bound PAH N02 on a school bus.
087728) on a school bus.
^ersonal exposure regression coefficients to:
Bedroom
Front door
Background
BC
PB-PAH
N02
PM,,
072
0.46
0.29
BC
1
BC
0.47
0.61
0.03
NO,
0.70
0.60
0.56
PB-PAH
0.94
1
N02
0.49
0.37
1
Note that these correlations are computed from dat
jresented by Sabin et al. for mean concentrations \
est bus travelled behind different vehicles.
Personal N02
concentration is more
strongly influenced by
background than PM2 5
orBC.
Less correlation was
observed between N02
and PM species. This
study was aimed more
a, .. at fuel choices and
when the control technologies for
children's exposures on
school buses.
Lai et al. (2004,
0568111
Microenvironmental and
personal PM25 and trace
elements for personal
exposure (P), residential
indoor (Rl), residential
outdoor (RO), and
workplace (Wl)
measurements.
Microenvironmental and
personal VOCs, N02, and
CO.
R(PM,,)
TVOC
N02
CO
P
0.21
-0.1
-0.07
Rl
0.21
-0.02
NR
RO
0.41
-0.16
NR
Wl
-0.32
0.09
NR
The EXPOLIS Oxford
study was more
focused on the indoor-
outdoor exposure
relationship, but the
correlation results
showed no important
relationships between
the pollutants shown.
Gomez-Perales et al. Microenvironmental PM25 Microenvironmental CO. Ratio of Cone PM25 CO Benzene
(2004, 054418: 2007, with S04 , N0r, EC,
1388161 OC.
Minibus/Bus 1.04 1.54 2.01
1.20 1.40 1.33
Minibus/Metro 1.70 2.02 3.20
1.433.033.10
Morning and evening
measurements of PM2 5
were on avg higher and
more variable than for
benzene and CO (in
order). Benzene and
CO had higher and
more variable
concentrations for
minibuses than for
buses and metros,
respectively, while
PM25 concentrations
were not substantially
different for buses and
minibuses.
December 2009
A-351
-------
Reference
PM metric
Copollutant metric
Association between PM and copollutant
Primary findings
Sarnat et al. (2001 , Fixed site and personal Ambient 03, N02, S02,
019401) PM25 monitors. and CO
R
PM25
03
N02
S02
CO
PM,,
1
-0.72
075
-0.17
0.69
0,
0.67
1
-0.71
0.41
-0.67
NO,
0.37
0.02
1
-0.17
0.76
SO,
...
1
-0.12
CO
0.15
-0.06
0.75
-0.32
1
Strong association
between ambient N02
and personal PM25
suggests that ambient
gas may be a suitable
surrogate for personal
exposure.
December 2009
A-352
-------
Annex A References
Abu-Allaban M; Gillies JA; Gertler AW; Clayton R; Proffitt D. (2007). Motor vehicle contributions to ambient
PM10 and PM2.5 at selected urban areas in the USA. Environ Monit Assess, 132: 155-63. 098575
Adar SD; Adamkiewicz G; Gold DR; Schwartz J; Coull BA; Suh H. (2007). Ambient and microenvironmental particles and
exhaled nitric oxide before and after a group bus trip. Environ Health Perspect, 115: 507-512. 098635
Adgate JL; Mongin SJ; Pratt GC; Zhang J; Field MP; Ramachandran G; Sexton K. (2007). Relationships between personal,
indoor, and outdoor exposures to trace elements in PM(2.5). Sci Total Environ, 386: 21-32. 156196
Adgate JL; Ramachandran G; Pratt GC; Waller LA; Sexton K. (2002). Spatial and temporal variability in outdoor, indoor,
and personal PM25 exposure. Atmos Environ, 36: 3255-3265. 030676
Adgate JL; Ramachandran G; Pratt GC; Waller LA; Sexton K. (2003). Longitudinal variability in outdoor, indoor, and
personal PM25 exposure in healthy non-smoking adults. Atmos Environ, 37: 993-1002. 040341
Allen R; Larson T; Sheppard L; Wallace L; Liu L-JS. (2003). Use of real-time light scattering data to estimate the
contribution of infiltrated and indoor-generated particles to indoor air. Environ Sci Technol, 37: 3484-3492. 053578
Allen R; Wallace L; Larson T; Sheppard L; Liu L-JS. (2007). Evaluation of the recursive model approach for estimating
particulate matter infiltration efficiencies using continuous light scattering data. J Expo Sci Environ Epidemiol, 17:
468-477. 154226
Anderson TL; Ogren JA. (1998). Determining aerosol radiative properties using the TSI 3563 integrating nephelometer.
Aerosol Sci Technol, 29: 57-69. 156213
Annesi-Maesano I; Moreau D; Caillaud D; Lavaud F; Le Moullec Y; Taytard A; Pauli G; Charpin D. (2007). Residential
proximity fine particles related to allergic sensitisation and asthma in primary school children. Respir Med, 101:
1721-1729.093180
Arhami M; Kuhn T; Fine PM; Delfino RJ; Sioutas C. (2006). Effects of Sampling Artifacts and Operating Parameters on
the Performance of a Semicontinuous Particulate Elemental Carbon/Organic Carbon Monitor. Environ Sci Technol,
40: 945-954. 156224
Arhami M; Polidori A; Delfino RJ; Tjoa T; Sioutas C. (2009). Associations between personal, indoor, and residential
outdoor pollutant concentrations: implications for exposure assessment to size-fractionated particulate matter. J Air
Waste Manag Assoc, 59: 392-404. 190096
Arnott WP; Hamasha K; Moosmilller H; Sheridan PJ; Ogren JA. (2005). Towards Aerosol Light-Absorption Measurements
with a 7-Wavelength Aethalometer: Evaluation with a Photoacoustic Instrument and 3-Wavelength Nephelometer.
Aerosol Sci Technol, 39: 17-29. 156227
Arnott WP; Moosmuller H; Rogers CF; Jin T; Bruch R. (1999). Photoacoustic spectrometer for measuring light absorption
by aerosol; instrument description. Atmos Environ, 33: 2845-2852. 020650
Ashbaugh LL. (1983). A statistical trajectory technique for determining air pollution source regions. J Air Waste Manag
Assoc, 33: 1096-1098. 156229
Ashbaugh LL; Myrup LO; Flocchini RG (1984). A principal component analysis of sulfur concentrations in the western
United States. Atmos Environ, 18: 783-791. 045148
Babich P; Wang PY; Allen G; Sioutas C; Koutrakis P. (2000). Development and Evaluation of a Continuous Ambient
PM2.5 Mass Monitor. Aerosol Sci Technol, 32: 309-324. 156239
Bae M-S; Schauer JJ; DeMinter JT; Turner JR; Smith D; Gary RA. (2004). Validation of a semi-continuous instrument for
elemental carbon and organic carbon using a thermal-optical method. Atmos Environ, 38: 2885-2893. 098680
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS)
December 2009 A-353
-------
Bae MS; Schauer JJ; Deminter JT; Turner JR. (2004). Hourly and daily patterns of particle-phase organic and elemental
carbon concentrations in the urban atmosphere. J Air Waste Manag Assoc, 54: 823-833. 156243
Balasubramanian R; Lee SS. (2007). Characteristics of indoor aerosols in residential homes in urban locations: a case study
in Singapore. J Air Waste Manag Assoc, 57: 981-990. 156248
Barn P; Larson T; Noullett M; Kennedy S; Copes R; Brauer M. (2008). Infiltration of forest fire and residential wood
smoke: an evaluation of air cleaner effectiveness. JExpo Sci Environ Epidemiol, 18: 503-511. 156252
Baxter LK; Clougherty JE; Laden F; Levy JI. (2007). Predictors of concentrations of nitrogen dioxide, fine particulate
matter, and particle constituents inside of lower socioeconomic status urban homes. J Expo Sci Environ Epidemiol,
17: 433-444. 092726
Baxter LK; Clougherty JE; Paciorek CJ; Wright RJ; Levy JI. (2007). Predicting residential indoor concentrations of
nitrogen dioxide, fine particulate matter, and elemental carbon using questionnaire and geographic information
system based data. Atmos Environ, 41: 6561-6571. 092725
Bein KJ; Zhao Y; Wexler AS; Johnston MV. (2005). Speciation of size-resolved individual ultrafine particles in Pittsburgh,
Pennsylvania. J Geophys Res, 110: D07S05. 156265
BeruBe KA; Sexton KJ; Jones TP; Moreno T; Anderson S; Richards RJ. (2004). The spatial and temporal variations in
PM10 mass from six UK homes. Sci Total Environ, 324: 41-53. 007894
BirchME. (1998). Analysis of carbonaceous aerosols: interlaboratory comparison. Analyst, 123: 851-857. 024953
Birch ME; Gary RA. (1996). Elemental carbon-based method for monitoring occupational exposures to particulate diesel
exhaust. Aerosol Sci Technol, 25: 221-241. 026004
Birch ME; Gary RA. (1996). Elemental carbon-based method for occupational monitoring of particulate diesel exhaust:
methodology and exposure issues. Analyst, 121: 1183-1190. 002352
Biswas S; Fine PM; Geller MD; Hering SV; Sioutas C. (2005). Performance evaluation of a recently developed water-
based condensation particle counter. Aerosol Sci Technol, 39: 419-427. 150694
Blanchard P; Hopper JF. (1997). Concentrations and distributions of PAHs and n-alkanes in atmospheric aerosols at non-
urban sites in Ontario, Canada. Presented at 6th International Conference on Carbonaceous Particles in the
Atmosphere; Vienna, Austria. 157195
Blanchard P; Brook JR; Brazal P. (2002). Chemical characterization of the organic fraction of atmospheric aerosol at two
sites in Ontario, Canada. J Geophys Res, 107: 8348. 189737
Blazso M; Janitsek S; Gelencser A; Artaxo P; Graham B; Andreae MO. (2003). Study of tropical organic aerosol by
thermally assisted alkylation-gas chromatography mass spectrometry. J Anal Appl Pyrol, 68: 351-369. 156278
Bond TC; Anderson TL; CampbellD. (1999). Calibration and Intercomparison of Filter-Based Measurements of Visible
Light Absorption by Aerosols. Aerosol Sci Technol, 30: 582-600. 156281
Branis M; Rez cov P; Domasov M. (2005). The effect of outdoor air and indoor human activity on mass concentrations of
PM10, PM2.5, and PM1 in a classroom. Environ Res, 99: 143-149. 156290
Briggs DJ; de Hoogh K; Morris C; Gulliver J. (2008). Effects of travel mode on exposures to particulate air pollution.
Environ Int, 34: 12-22. 156294
Brook JR; Burnett RT; Dann TF; Cakmak S; Goldberg MS; Fan X; Wheeler AJ. (2007). Further interpretation of the acute
effect of nitrogen dioxide observed in Canadian time-series studies. J Expo Sci Environ Epidemiol, 17 : S36-S44.
091153
Brown K; Sarnat J; Suh H; Coull B; Spengler J; Koutrakis P. (2008). Ambient site, home outdoor and home indoor
particulate concentrations as proxies of personal exposures. J Environ Monit, 10: 1041-1051. 190894
Brunekreef B; JanssenN; De Hartog J; Oldenwening M; Meliefste K; Hoek G; Lanki T; Timonen K; Vallius M; Pekkanen
J; Van Grieken R. (2005). Personal, indoor, and outdoor exposures to PM2.5 and its components for groups of
cardiovascular patients in Amsterdam and Helsinki. Health Effects Institute. Montepelier, VT. 090486
Butler AJ; Andrew MS; Russell AG (2003). Daily sampling Of PM2. 5 in Atlanta: Results of the first year of the
Assessment of Spatial Aerosol Composition. J Geophys Res, 108: 8415. 156313
December 2009 A-354
-------
Cabada JC; Rees S; Takahama S; Khlystov A; Pandis SN; Davidson CI; Robinson AL. (2004). Mass size distributions and
size resolved chemical composition of fine participate matter at the Pittsburgh supersite. Atmos Environ, 38: 3127-
3141. 148859
Cao JJ; Lee SC; Chow JC; Cheng Y; Ho KF; Fung K; Liu SX; Watson JG. (2005). Indoor/outdoor relationships for PM2.5
and associated carbonaceous pollutants at residential homes in Hong Kong - case study. Indoor Air, 15: 197-204.
156321
Chakrabarti B; Singh M; Sioutas C. (2004). Development of a Near-Continuous Monitor for Measurment of the Sub-150
nm PM Mass Concentration. Aerosol Sci Technol, 38: 239-252. 157426
Cheng M-D; Hopke PK; Zeng Y. (1993). A receptor-oriented methodology for determining source regions of particulate
sulfate observed at Dorset, Ontario. J Geophys Res, 98: 16,839-16,849. 052294
Chillrud SN; Epstein D; Ross JM; Sax SN; Pederson D; Spengler JD; Kinney PL. (2004). Elevated airborne exposures of
teenagers to manganese, chromium, and iron from steel dust and New York City's subway system. Environ Sci
Technol, 38: 732-737. 054799
Chow JC. (1995). Measurement methods to determine compliance with ambient air quality standards for suspended
particles. J Air Waste Manag Assoc, 45: 320-382. 077012
Chow JC. (2007). The application of thermal methods for determining chemical composition of carbonaceous aerosols: A
review. J Environ Sci Health A Tox Hazard Subst Environ Eng, 42: 1521-1541. 157209
Chow JC; Chen L-WA; Lowenthal DH; Doraiswamy P; Park K; Kohl S; Trimble DL; Watson JG. (2005). California
Regional PM10/PM2.5 Air Quality Study (CRPAQS) - Initial data analysis of field program measurements. Desert
Research Institute. Reno, NV. 156348
Chow JC; Doraiswamy P; Watson JG; Chen LW; Ho SS; Sodeman DA. (2008). Advances in integrated and continuous
measurements for particle mass and chemical composition. J Air Waste Manag Assoc, 58: 141-163. 156355
Chow JC; Watson JG; Chen LW; Chang MC; Robinson NF; Trimble D; Kohl S. (2007). The IMPROVE_A temperature
protocol for thermal/optical carbon analysis: maintaining consistency with a long-term database. J Air Waste Manag
Assoc, 57: 1014-1023. 156354
Chow JC; Watson JG; Doraiswamy P; Chen L-WA; Sodeman DA; Ho SSH; Tropp RJ; Kohl SD; Trimble DL; Fung KK.
(2006). Climate Change - Characterization of black carbon and organic carbon air pollution emissions and
evaluation of measurement methods, Phase I. Desert Research Institute. Reno, NV. 156350
Chow JC; Watson JG; Lowenthal DH; Chen LWA; Magliano KL. (2006). Particulate carbon measurements in California's
San Joaquin Valley. Chemosphere, 62: 337-348. 155207
Chow JC; Watson JG; Lowenthal DH; Chen LWA; Tropp RJ; Park K; Magliano KA. (2006). PM25 and PM10 Mass
Measurements in California's San Joaquin Valley. Aerosol Sci Technol, 40: 796-810. 146622
Chow JC; Watson JG; Lowenthal DH; Magliano KL. (2005). Loss of PM2.5 nitrate from filter samples in central
California. J Air Waste Manag Assoc, 55: 1158-1168. 099030
Chow JC; Watson JG; Pritchett LC; Pierson WR; Frazier CA; Purcell RG (1993). The DRI thermal/optical reflectance
carbon analysis system: description, evaluation and applications in US air quality studies. Atmos Environ, 27: 1185-
1201.077459
Chung A; Chang DP; Kleeman MJ; Perry KD; Cahill TA; Dutcher D; McDougall EM; Stroud K. (2001). Comparison of
real-time instruments used to monitor airborne particulate matter. J Air Waste Manag Assoc, 51: 109-120. 156357
Conner TL; Williams RWE. (2004). Identification of possible sources of particulate matter in the personal cloud using
SEM/EDX. Atmos Environ, 38: 5305-5310. 156364
Cornell SE; Jickells TD. (1999). Water-soluble organic nitrogen in atmospheric aerosol: a comparison of UV and persulfate
oxidation methods. Atmos Environ, 33: 833-840. 156367
Cortez-Lugo M; Moreno-Macias H; Holguin-Molina F; Chow JC; Watson JG; Gutierrez-Avedoy V; Mandujano F;
Hernandez-Avila M; Romieu I. (2008). Relationship between indoor, outdoor, and personal fine particle
concentrations for individuals with COPD and predictors of indoor-outdoor ratio in Mexico city. J Expo Sci
Environ Epidemiol, 18: 109-115. 156368
December 2009 A-355
-------
Crimmins BS; Baker JE. (2006). Improved GC/MS methods for measuring hourly PAH and nitro-PAH concentrations in
urban particulate matter. Atmos Environ, 40: 6764-6779. 097008
Crist KC; Liu B; Kim M; Deshpande SR; John K. (2008). Characterization of fine particulate matter in Ohio: Indoor,
outdoor, and personal exposures. Environ Res, 106: 62-71. 156372
Dams R; Rahn KA; Robbins JA; Nifong GD; Winchester JW. (1970). Multi-Element Analysis of Air Pollution Particulates
by Nondestructive Neutron Activation. Presented at Proceedings of the Second International Clean Air Congress,
New York. 156379
Decesari S; Facchini MC; Fuzzi S; McFiggans GB; Coe H; Bower KN. (2005). The water-soluble organic component of
size-segregated aerosol, cloud water and wet depositions from Jeju Island during ACE-Asia. Atmos Environ, 39:
211-222. 144536
Delfino RJ; Quintana PJE; Floro J; Gastanaga VM; Samimi BS; Kleinman MT; Liu L-JS; Bufalino C; Wu C-F; McLaren
CE. (2004). Association of FEV1 in asthmatic children with personal and microenvironmental exposure to airborne
particulate matter. Environ Health Perspect, 112: 932-941. 056897
Delfino RJ; Staimer N; Gillen D; Tjoa T; Sioutas C; Fung K; George SC; Kleinman MT. (2006). Personal and ambient air
pollution is associated with increased exhaled nitric oxide in children with asthma. Environ Health Perspect, 114:
1736-1743. 090745
Diapouli E; Chaloulakou A; Mihalopoulos N; Spyrellis N. (2008). Indoor and outdoor PM mass and number concentrations
at schools in the Athens area. Environ Monit Assess, 136: 13-20. 190893
Diapouli E; Chaloulakou A; Spyrellis N. (2007). Levels of ultrafine particles in different microenvironments—implications
to children exposure. Sci Total Environ, 388: 128-136. 156397
Dimitroulopoulou C; Ashmore MR; Hill MTR; Byrne MA; Kinnersley R. (2006). INDAIR: a probabilistic model of indoor
air pollution in UK homes. Atmos Environ, 40: 6362-6379. 090302
Dong Y; Hays MD; Dean Smith N; Kinsey JS. (2004). Inverting cascade impactor data for size-resolved characterization of
fine particulate source emissions. J Aerosol Sci, 35: 1497-1512. 156409
Drewnick F; Jayne JT; Canagaratna M; Worsnop DR; Demerjian KL. (2004). Measurement of Ambient Aerosol
Composition During the PMTACS-NY 2001 Using an Aerosol Mass Spectrometer. Part II: Chemically Speciated
Mass Distributions. Aerosol Sci Technol, 38: 104-117. 155755
Drewnick F; Schwab J; Jayne J; Canagaratna M; Worsnop D; Demerjian K. (2004). Measurement of Ambient Aerosol
Composition During the PMTACS-NY 2001 Using an Aerosol Mass Spectrometer. Part I: Mass Concentrations.
Aerosol Sci Technol, 38: 92-103. 155754
Drewnick F; Schwab JJ; Hogrefe O; Peters S; Husain L; Diamond D; Weber R; Demerjian KL. (2003). Intercomparison
and evaluation of four semi-continuous PM2.5 sulfate instruments. Atmos Environ, 37: 3335-3350. 099160
Eatough DJ; Eatough NL; Obeidi F; Pang Y; Modey W; Long R. (2001). Continuous determination of PM25 mass,
including semi-volatile species. Aerosol Sci Technol, 34: 1-8. 010303
Ebelt ST; Wilson WE; Brauer M. (2005). Exposure to ambient and nonambient components of particulate matter: a
comparison of health effects. Epidemiology, 16: 396-405. 056907
Ellis EC; Novakov T. (1982). Application of thermal analysis to the characterization of organic aerosol particles. Sci Total
Environ, 23: 227-238. 156416
Emmenegger C; Reinhardt A; Hueglin C; Zenobi R; Kalberer M. (2007). Evaporative light scattering: A novel detection
method for the quantitative analysis of humic-like substances in aerosols. Environ Sci Technol, 41: 2473-2478.
156418
Engling G; Carrico CM; Kreidenweis SM; Collett JL; Day DE; Malm WC; Lincoln E; Min Hao W; linuma Y; Herrmann
H. (2006). Determination of levoglucosan in biomass combustion aerosol by high-performance anion-exchange
chromatography with pulsed amperometric detection. Atmos Environ, 40: 299-311. 156422
Fabbri D; Prati S; Vassura I. (2002). Molecular characterisation of organic material in air fine particles (PM10) using
conventional and reactive pyrolysis-gas chromatography-mass spectrometry. J Environ Monit, 4: 210-215. 156426
Falkovich AH; Rudich Y. (2001). Analysis of semivolatile organic compounds in atmospheric aerosols by direct sample
introduction thermal desorption GC/MS. Environ Sci Technol, 35: 2326-2333. 156427
December 2009 A-356
-------
Falkovich AH; Schkolnik G; Ganor E; Rudich Y. (2004). Adsorption of organic compounds pertinent to urban
environments onto mineral dust particles. J Geophys Res, 109: D02208. 156428
Farmer PB; Singh R; Kaur B; Sram RJ; Binkova B; Kalina I; Popov TA; Garte S; Taioli E; Gabelova A; Cebulska-
Wasilewska A. (2003). Molecular epidemiology studies of carcinogenic environmental pollutants: effects of
poly cyclic aromatic hydrocarbons (PAHs) in environmental pollution on exogenous and oxidative DNA damage.
MutatRes, 544: 397-402. 089017
Fehsenfeld FC;FŁastie D;Chow JC;Solomon PA. (2004). Particle and gas measurements. In McMurry, P. H.Shepherd, M.
F.Vickery, J. S. (Ed.),Particulate Matter Science for Policy Makers: ANARSTO Assessment (pp. 159-189).
Cambridge, MA: NARSTO. 157360
Feldpausch P; Fiebig M; Fritzsche L; Petzold A. (2006). Measurement of ultrafine aerosol size distributions by a
combination of diffusion screen separators and condensation particle counters. J Aerosol Sci, 37: 577-597. 155773
Ferro AR; Kopperud RJ; Hildemann LM. (2004). Source strengths for indoor human activities that resuspend particulate
matter. Environ Sci Technol, 38: 1759-1764. 055387
Fine PM; Chakrabarti B; Krudysz M; Schauer JJ; Sioutas C. (2004). Diurnal Variations of Individual Organic Compound
Constituents of Ultrafine and Accumulation Mode Particulate Matter in the Los Angeles Basin. Environ Sci
Technol, 38: 1296-1304. 141283
Fine PM; Jaques PA; Hering SV; Sioutas C. (2003). Performance Evaluation and Use of a Continuous Monitor for
Measuring Size-Fractionated PM 2.5 Nitrate. Aerosol Sci Technol, 37: 342 - 354. 155775
Fitz D; Chan M; Cass G; Lawson D; Ashbaugh L. (1989). A multi-component size-classifying aerosol and gas sampler for
ambient air monitoring. Presented at Presented at: 82nd annual meeting of the Air & Waste Management
Association; June; Anaheim, C A. Pittsburgh, PA: Air & Waste Management Association; paper no. 89-140.1.
077387
Fraser MP; Cass GR; Simoneit BRT. (2003). Air quality model evaluation data for organics 6 C3-C24 organic acids.
Environ Sci Technol, 37: 446-453. 040266
Fraser MP; Yue Z W; Buzcu B. (2003). Source apportionment of fine particulate matter in Houston, TX, using organic
molecular markers. Atmos Environ, 37: 2117-2123. 042231
Fraser MP; Yue ZW; Tropp RJ; Kohl SD; Chow JC. (2002). Molecular composition of organic fine particulate matter in
Houston, TX. Atmos Environ, 36: 5751-5758. 140741
Fromme H; Diemer J; Dietrich S; Cyrys J; Heinrich J; Lang W; Kiranoglu M; Twardella D. (2008). Chemical and
morphological properties of particulate matter (PM10, PM25) in school classrooms and outdoor air. Atmos Environ,
42: 6597-6605. 155147
Fruin S; Westerdahl D; Sax T; Sioutas C; Fine PM. (2008). Measurements and predictors of on-road ultrafine particle
concentrations and associated pollutants in Los Angeles. Atmos Environ, 42: 207-219. 097183
Gadkari NM; Pervez S. (2007). Source investigation of personal particulates in relation to identify major routes of exposure
among urban residentials. Atmos Environ, 41: 7951-7963. 156459
Gauvin S; Reungoat P; Cassadou S; Dechenaux J; Momas I; Just J; Zmirou D. (2002). Contribution of indoor and outdoor
environments to PM2.5 personal exposure of children—VESTA study. Sci Total Environ, 297: 175-181. 034893
Geyh A; Chillrud S; Wiliams D; Herbstman J; Symons J; Rees K; Ross J; Kim S; Lim H; Turpin B; Breysse P. (2005).
Assessing truck driver exposure at the World Trade Center disaster site: personal and area monitoring for particulate
matter and volatile organic compounds during October 2001 and April 2002. J Occup Environ Hyg, 2: 179-193.
186949
Gomez-Perales JE; Colvile RN; Fernandez-Bremauntz AA; Gutierrez-Avedoy V; Paramo-Figueroa VH; Blanco-Jimenez S;
Bueno-Lopez E; Bernabe-Cabanillas R; Mandujano F; Hidalgo-Navarro M; Nieuwenhuijsen MJ. (2007). Bus,
minibus, metro inter-comparison of commuters' exposure to air pollution in Mexico City. Atmos Environ, 41: 890-
901. 138816
Gomez-Perales JE; Colvile RN; NieuwenhuijsenMJ; Fernandez-Bremauntz A; Gutierrez-Avedoy VJ; Paramo-Figueroa
VH; Blanco-Jimenez S; Bueno-Lopez E; Mandujano F; Bernabe-Cabanillas R. (2004). Commuters' exposure to
PM25, CO, and benzene in public transport in the metropolitan area of Mexico City. Atmos Environ, 38: 1219-
1229.054418
December 2009 A-357
-------
Graham B; Falkovich AH; Rudich Y; Maenhaut W; Guy on P; Andreae MO. (2004). Local and regional contributions to the
atmospheric aerosol over Tel Aviv, Israel: a case study using elemental, ionic and organic tracers. Atmos Environ,
38: 1593-1604. 156490
Graney JR; Landis MS; Norris GA. (2004). Concentrations and solubility of metals from indoor and personal exposure
PM25 samples. Atmos Environ, 38: 237-247. 053756
Greaves RC; Barkley RM; Sievers RE. (1985). Rapid sampling and analysis of volatile constituents of airborne particulate
matter. Anal Chem, 57: 2807-2815. 156494
Greaves RC; Barkley RM; Sievers RE; Meglen RR. (1987). Covariations in the concentrations of organic compounds
associated with springtime atmospheric aerosols. Atmos Environ, 21: 2549-2561. 156495
Grosjean D; Seinfeld JH. (1989). Parameterization of the formation potential of secondary organic aerosols. Atmos
Environ, 23: 1733-1747. 045643
Grover BD; Eatough NL; Eatough DJ; Chow JC; Watson JG; Ambs JL; Meyer MB; Hopke PK; Al-Horr R; Later DW;
Wilson WE. (2006). Measurement of Both Nonvolatile and Semi-Volatile Fractions of Fine Particulate Matter in
Fresno, CA. Aerosol Sci Technol, 40: 811-826. 138080
Grover BD; Kleinman M; Eatough NL; Eatough DJ; Hopke PK; Long RW; Wilson WE; Meyer MB; Ambs JL. (2005).
Measurement of total PM25 mass (nonvolatile plus semivolatile) with the Filter Dynamic Measurement System
tapered element oscillating microbalance monitor. J Geophys Res, 110: D07S03. 090044
Gulliver J; Briggs DJ. (2004). Personal exposure to particulate air pollution in transport microenvironments. Atmos
Environ, 38: 1-8. 053238
Gulliver J; Briggs DJ. (2007). Journey-time exposure to particulate air pollution. Atmos Environ, 41: 7195-7207. 155814
Guo H; Lee SC; Chan LY. (2004). Indoor air quality investigation at air-conditioned and non-air-conditioned markets in
Hong Kong. Sci Total Environ, 323: 87-98. 156506
Hall PA; Watson APR; Garner GV; Hall K; Smith S; Waterman D; Horsfield B. (1999). An investigation of micro-scale
sealed vessel thermal extraction-gas chromatography-mass spectrometry (MSSV-GC-MS#) and micro-scale sealed
vessel pyrolysis-gas chromatography-mass spectrometry applied to a standard reference material of an urban dust.
Sci Total Environ, 235: 269-276. 156512
Hamilton JF; Webb PJ; Lewis AC; Hopkins JR; Smith S; Davy P. (2004). Partially oxidised organic components in urban
aerosol using GCXGC-TOF/MS. Atmos Chem Phys, 4: 1279-1290. 156516
Hamilton JF; Webb PJ; Lewis AC; Reviejo MM. (2005). Quantifying small molecules in secondary organic aerosol formed
during the photo-oxidation of toluene with hydroxyl radicals. Atmos Environ, 39: 7263-7275. 088173
Hanninen OO; Lebret E; Ilacqua V; Katsouyanni K; Kunzli N; Sram RJ; Jantunen M. (2004). Infiltration of ambient PM2.5
and levels of indoor generated non-ETS PM2.5 in residences of four European cities. Atmos Environ, 38: 6411-
6423. 056812
Harrison D; Shik Park S; Ondov J; Buckley T; Roul Kim S; Jayanty RKM. (2004). Highly time resolved fine particle
nitrate measurements at the Baltimore Supersite. Atmos Environ, 38: 5321-5332. 136787
Hauck H; Stopper S; Puxbaum H; Kundi M; Preining O; Berner A; Gomiscek B. (2004). On the equivalence of gravimetric
PM data with TEOM and beta-attenuation measurements. JAerosol Sci, 35: 1135-1149. 156525
Haverinen-Shaughnessy U; Toivola M; Aim S; Putus T; Nevalainen A. (2007). Personal and microenvironmental
concentrations of particles and microbial aerosol in relation to health symptoms among teachers. J Expo Sci
Environ Epidemiol, 17: 182-190. 156526
Hays MD; Smith ND; Dong Y. (2004). Nature of unresolved complex mixture in size-distributed emissions from residential
wood combustion as measured by thermal desorption-gas chromatography-mass spectrometry. J Geophys Res, 109:
D16S04. 156530
Hays MD; Smith ND; Kinsey J; Dong Y; Kariher P. (2003). Poly cyclic aromatic hydrocarbon size distributions in aerosols
from appliances of residential wood combustion as determined by direct thermal desorption—GC/MS. JAerosol
Sci, 34: 1061-1084. 156529
Helmig D; Bauer A; Mueller J; Klein W. (1990). Analysis of particulate organics in a forest atmosphere by
thermodesorption GC/MS. Atmos Environ, 24: 179-184. 156536
December 2009 A-358
-------
Henning S; Weingartner E; Schwikowski M; Gaggeler HW; Gehrig R; Hinz KP; Trimborn A; Spengler B; Baltensperger U.
(2003). Seasonal variation of water-soluble ions of the aerosol at the high-alpine site Jungfraujoch (3580 m asl). J
Geophys Res, 108: 4030. 156539
Henry RC. (1997). History and fundamentals of multivariate air quality receptor models. Chemometr Intell Lab Syst, 37:
37-42. 020941
Henry RC. (2003). Multivariate receptor modeling by N-dimensional edge detection. Chemometr Intell Lab Syst, 65: 179-
189. 156540
Henry RC; Chang YS; Spiegelman CH. (2002). Locating nearby sources of air pollution by nonparametric regression of
atmospheric concentrations on wind direction. Atmos Environ, 36: 2237-2244. 136097
Hering S; Cass G (1999). The magnitude of bias in the measurement of PM2.5 arising from volatilization of particulate
nitrate from Teflon filters. J Air Waste Manag Assoc, 49: 725-733. 084958
Hering S; Fine PM; Sioutas C; Jaques PA; Ambs JL; Hogrefe O; Demerjian KL. (2004). Field assessment of the dynamics
of particulate nitrate vaporization using differential TEOM® and automated nitrate monitors. Atmos Environ, 38:
5183-5192. 155837
Hering SV; Lawson DR; Allegrini I; Febo A; Perrino C; Possanzini M; Sickles JE II; Anlauf KG; Wiebe A; Appel BR; John
W; Ondo J; Wall S; Braman RS; Sutton R; Cass GR; Solomon PA; Eatough DJ; Eatough NL; Ellis EC; Grosjean D;
Hicks BB; Womack JD; Horrocks J; Knapp KT; Ellestad TG; Paur RJ; Mitchell WJ; Pleasant M; Peake E; MacLean
A; Pierson WR; Brachaczek W; Schiff HI; Mackay GI; Spicer CW;. (1988). The nitric acid shootout: field
comparison of measurement methods. Atmos Environ, 22: 1519-1539. 036012
Hering SV; Stolzenburg MR; Quant FR; Oberreit DR; Keady PB. (2005). A Laminar-Flow, Water-Based Condensation
Particle Counter (WCPC). Aerosol Sci Technol, 39: 659-672. 155838
Hermann M; Wehner B; Bischof O; Han HS; Krinke T; Liu W; Zerrath A; Wiedensohler A. (2007). Particle counting
efficiencies of newTSI condensation particle counters. JAerosol Sci, 38: 674-682. 155840
Hidy GM; Friedlander SK. (1972). The nature of the Los Angeles aerosol. Presented at Second International Clean Air
Congress, New York. 156546
Ho KF; Cao JJ; Harrison RM; Lee SC. (2004). Indoor/outdoor relationships of organic carbon (OC) and elemental carbon
(EC) in PM2.5 in roadside environment of Hong Kong. Atmos Environ, 38: 6327-6335. 056804
Ho KF; Chow JC; Watson JG; Fung K; Lee SC; Cao JJ; Li YS. (2006). Variability of organic and elemental carbon, water
soluble organic carbon, and isotopes in Hong Kong. Atmos Chem Phys, 6: 4579-4600. 156552
Ho SSH; Yu JZ. (2004). In-injection port thermal desorption and subsequent gas chromatography-mass spectrometric
analysis of poly cyclic aromatic hydrocarbons and n-alkanes in atmospheric aerosol samples. J Chromatogr, 1059:
121-129. 156551
Hoek G; de Hartog J; Meliefste K; ten Brink H; Katsouyanni K; Karakatsani A; Lianou M; Kotronarou A; Kavouras I;
Pekkanen J; Vallius M; Kulmala M; Puustinen A; Thomas S; Meddings C; Ayres J; van Wrjnen J; Hameri K; Kos G;
Harrison R. (2008). Indoor-outdoor relationships of particle number and mass in four European cities. Atmos
Environ, 42: 156-169. 156554
Hogrefe O; Schwab JJ; Drewnick F; Lala GG; Peters S; Demerjian KL; Rhoads K; Felton HD; Rattigan OV; Husain L;
Dutkiewicz VA. (2004). Semicontinuous PM2.5 sulfate and nitrate measurements at an urban and a rural location in
New York: PMTACS-NY summer 2001 and 2002 campaigns. J Air Waste Manag Assoc, 54: 1040-1060. 099003
Hopke PK; Li CL; Ciszek W; Landsberger S. (1995). The use of bootstrapping to estimate conditional probability fields for
source locations of airborne pollutants. Chemometr Intell Lab Syst, 30: 69-79. 156566
Hopke PK; Ramadan Z; Paatero P; Norris GA; Landis MS; Williams RW; Lewis CW. (2003). Receptor modeling of
ambient and personal exposure samples: 1998 Baltimore Particulate Matter Epidemiology-Exposure Study. Atmos
Environ, 37: 3289-3302. 095544
Hystad P; Setton E; Allen R; Keller P; Brauer M. (2008). Modeling residential fine particulate matter infiltration for
exposure assessment. J Expo Sci Environ Epidemiol, 19: 570-579. 190890
Ito K; Thurston GD; Silverman RA. (2007). Characterization of PM2.5, gaseous pollutants, and meteorological interactions
in the context of time-series health effects models. J Expo Sci Environ Epidemiol, 17 Suppl 2: S45-S60. 156594
December 2009 A-359
-------
Jacquemin B; Cabrera L; Querol X; Bellander T; Moreno N; Peters A; Pey J; Pekkanen J; Lanki T; Sunyer J. (2007). Levels
of outdoor PM2.5, absorbance and sulphur as surrogates for personal exposures among post-myocardial infarction
patients in Barcelona, Spain. Atmos Environ, 41: 1539-1549. 156600
Jacquemin B; Sunyer J; Forsberg B; Gotschi T; Bayer-Oglesby L; Ackermann-Liebrich U; de Marco R; Heinrich J; Jarvis
D; Toren K; Kilnzli N. (2007). Annoyance due to air pollution in Europe. Int J Epidemiol, 10: 1-12. 192372
Jansen KL; Larson TV; Koenig JQ; Mar TF; Fields C; Stewart J; Lippmann M. (2005). Associations between health effects
and particulate matter and black carbon in subjects with respiratory disease. Environ Health Perspect, 113: 1741-
1746. 082236
JanssenNA; Lanki T; Hoek G; Vallius M; De Hartog JJ; Van Grieken R; PekkanenJ; Brunekreef B. (2005). Associations
between ambient, personal, and indoor exposure to fine particulate matter constituents in Dutch and Finnish panels
of cardiovascular patients. Occup Environ Med, 62: 868-877. 088692
Jaques PA; Ambs JL; Grant WL; Sioutas C. (2004). Field evaluation of the differential TEOM monitor for continuous PM
2.5 mass concentrations. Aerosol Sci Technol, 38: 49-59. 155878
Jedrychowski WA; Perera FP; Pac A; Jacek R; Whyatt RM; Spengler JD; Dumyahn TS; Sochacka-Tatara E. (2006).
Variability of total exposure to PM2.5 related to indoor and outdoor pollution sources Krakow study in pregnant
women. Sci Total Environ, 366: 47-54. 156606
Jeon SJ; Meuzelaar HLC; Sheya SAN; Lighty JS; Jarman WM; Kasteler C; Sarofim AF; Simoneit BRT. (2001).
Exploratory studies of PM10 receptor and source profiling by GC/MS and principal component analysis of
temporally and spatially resolved ambient samples. J Air Waste Manag Assoc, 51: 766-784. 016636
Jimenez JL; Jayne JT; Shi Q; Kolb CE; Worsnop DR; Yourshaw I; Seinfeld JH; Flagan RC; Zhang X; Smith KA. (2003).
Ambient aerosol sampling using the Aerodyne Aerosol Mass Spectrometer. J Geophys Res, 108: 8425. 156611
Johannesson S; Gustafson P; Molnar P; Barregard L; Sallsten G. (2007). Exposure to fine particles (PM2.5 and PM1) and
black smoke in the general population: personal, indoor, and outdoor levels. J Expo Sci Environ Epidemiol, 17:
613-624. 156614
John W; Wall SM; Ondo JL. (1988). A new method for nitric acid and nitrate aerosol measurement using the dichotomous
sampler. Atmos Environ, 22: 1627-1635. 045903
Kaur S; Nieuwenhuijsen M; Colvile R. (2005). Personal exposure of street canyon intersection users to PM2.5, ultrafine
particle counts and carbon monoxide in central London, UK. Atmos Environ, 39: 3629-3641. 086504
Kaur S; Nieuwenhuijsen MJ; Colvile RN. (2005). Pedestrian exposure to air pollution along a major road in Central
London, UK. Atmos Environ, 39: 7307-7320. 088175
Keeler GJ; Samson PJ. (1989). Spatial representativeness of trace element ratios. Environ Sci Technol, 23: 1358-1364.
156633
Khlystov A; Stanier CO; Takahama S; Pandis SN. (2005). Water content of ambient aerosol during the Pittsburgh Air
Quality Study. J Geophys Res, 110: D07S10. 156635
Kidwell CB; Ondov JM. (2001). Development and evaluation of a prototype system for collecting sub-hourly ambient
aerosol for chemical analysis. Aerosol Sci Technol, 35: 596-601. 017092
Kidwell CB; Ondov JM. (2004). Elemental Analysis of Sub-Hourly Ambient Aerosol Collections. Aerosol Sci Technol, 38:
205-218. 155898
Kim D; Sass-Kortsak A; Purdham JT; Dales RE; Brook JR. (2005). Sources of personal exposure to fine particles in
Toronto, Ontario, Canada. J Air Waste Manag Assoc, 55: 1134-1146. 156640
Kinsey JS; Mitchell WA; Squier WC; Linna K; King FG; Logan R; Dong Y; Thompson GJ; Clark NN. (2006). Evaluation
of methods for the determination of diesel-generated fine particulate matter: Physical characterization results. J
Aerosol Sci, 37: 63-87. 130654
Kiss G; Varga B; Galambos I; Ganszky I. (2002). Characterization of water-soluble organic matter isolated from
atmospheric fine aerosol. J Geophys Res, 107: 8339. 156646
Klinmalee A; Srimongkol K; Kim Oanh NT. (2008). Indoor air pollution levels in public buildings in Thailand and
exposure assessment. Environ Monit Assess, 156: 581-594. 190888
December 2009 A-360
-------
Koenig JQ; Jansen K; Mar TF; Lumley T; Kaufman J; Trenga CA; Sullivan J; Liu LJ; Shapiro GG; Larson TV. (2003).
Measurement of offline exhaled nitric oxide in a study of community exposure to air pollution. Environ Health
Perspect, 111: 1625-1629. 156653
Koistinen KJ; Edwards RD; Mathys P; Ruuskanen J; Kunzli N; Jantunen MJ. (2004). Sources of fine particulate matter in
personal exposures and residential indoor, residential outdoor and workplace microenvironments in the Helsinki
phase of the EXPOLIS study. Scand J Work Environ Health, 30 Suppl 2: 36-46. 156655
Kousa A; Monn C; Rotko T; Aim S; Oblesby L; Jantunen MJ. (2001). Personal exposures to NO2 in the EXPOLIS-study:
relation to residential indoor, outdoor and workplace concentrations in Basel, Helsinki and Prague. Atmos Environ,
35: 3405-3412. 025270
Koutrakis P; Suh HH; Sarnat JA; Brown KW; Coull BA; Schwartz J. (2005). Characterization of particulate and gas
exposures of sensitive subpopulations living in Baltimore and Boston. Health Effects Institute. Boston, MA. 131.
http://pubs.healtheffects.org/view.php?id=91. 095800
Kulkarni MM; Patil RS. (2003). Personal exposure to toxic metals in an Indian metropolitan region. J Inst Eng, 84: 23-29.
156664
Kulmala M; Mordas G; Petaja T; Gronholm T; Aalto PP; Vehkamaki H; Hienola AI; Herrmann E; Sipila M; Riipinen I.
(2007). The condensation particle counter battery (CPCB): A new tool to investigate the activation properties of
nanoparticles. J Aerosol Sci, 38: 289-304. 155911
Kulmala M; Riipinen I; Sipila M; Manninen HE; Petaja T; Junninen H; Maso MD; Mordas G; Mirme A; Vana M; Hirsikko
A; Laakso L; Harrison RM; Hanson I; Leung C; Lehtinen KE; Kerminen VM. (2007). Toward direct measurement
of atmospheric nucleation. Science, 318: 89-92. 097838
Labban R; Veranth JM; Watson JG; Chow JC. (2006). Feasibility of soil dust source apportionment by the pyrolysis-gas
chromatography/mass spectrometry method. J Air Waste Manag Assoc, 56: 1230-1242. 156665
Lai HK; Kendall M; Ferrier H; Lindup I; Aim S; Hanninen O; Jantunen M; Mathys P; Colvile R; Ashmore MR; Cullinan P;
NieuwenhuijsenMJ. (2004). Personal exposures and microenvironment concentrations of PM25, VOC, NO2 and
CO in Oxford, UK. Atmos Environ, 38: 6399-6410. 056811
Lake DA; Tolocka MP; Johnston MV; Wexler AS. (2003). Mass Spectrometry of Individual Particles between 50 and 750
nm in Diameter at the Baltimore Supersite. Environ Sci Technol, 37: 3268-3274. 156669
Lake DA; Tolocka MP; Johnston MV; Wexler AS. (2004). The character of single particle sulfate in Baltimore. Atmos
Environ, 38: 5311-5320. 088411
Larson T; Gould T; Simpson C; Liu LJ; Claiborn C; Lewtas J. (2004). Source apportionment of indoor, outdoor, and
personal PM2.5 in Seattle, Washington, using positive matrix factorization. J Air Waste Manag Assoc, 54: 1175-87.
098145
Lawson CL; Hanson RJ. (1974). Solving least squares problems. Englewood Cliffs: Prentice Hall. 156673
Lee JH; Hopke PK; Holsen TM; Lee DW; AJaques P; Sioutas C; Ambs JL. (2005). Performance evaluation of continuous
PM2.5 mass concentration monitors. J Aerosol Sci, 36: 95-109. 155925
Lee JH; Hopke PK; Holsen TM; Polissar AV (2005). Evaluation of Continuous and Filter-Based Methods for Measuring
PM 2.5 Mass Concentration. Aerosol Sci Technol, 39: 290-303. 156680
Lee JH; Hopke PK; Holsen TM; Polissar AV; Lee DW; Edgerton ES; Ondov JM; Allen G. (2005). Measurements of fine
particle mass concentrations using continuous and integrated monitors in Eastern US Cities. Aerosol Sci Technol,
39: 261-275. 128139
Lewis CW; Norris GA; Conner TL; Henry RC. (2003). Source apportionment of Phoenix PM2.5 aerosol with the Unmix
Receptor model. J Air Waste Manag Assoc, 53: 325-338. 088413
Li W-W; Paschold H; Morales H; Chianelli J. (2003). Correlations between short-term indoor and outdoor PM
concentrations at residences with evaporative coolers. Atmos Environ, 37: 2691-2703. 047845
Li YC; Yu JZ. (2005). Simultaneous Determination of Mono-and Dicarboxylic Acids, o-Oxo-carboxylic Acids, Midchain
Ketocarboxylic Acids, and Aldehydes in Atmospheric Aerosol Samples. Environ Sci Technol, 39: 7616-7624.
156692
December 2009 A-361
-------
Lim H-J; Turpin BJ; Edgerton E; Hering SV; Allen G; Maring H; Solomon P. (2003). Semi-continuous aerosol carbon
measurements: Comparison of Atlanta supersite measurements. J Geophys Res, 108: 8419. 037037
Lithgow GA; Robinson AL; Buckley SG (2004). Ambient measurements of metal-containing PM25 in an urban
environment using laser-induced breakdown spectroscopy. Atmos Environ, 38: 3319-3328. 126616
Liu L-J; Box M; Kalman D; Kaufman J; Koenig J; Larson T; Lumley T; Sheppard L; Wallace L. (2003). Exposure
assessment of particulate matter for susceptible populations in Seattle. Environ Health Perspect, 11: 909-918.
073841
Lunden MM; Kirchstetter TW; Thatcher TL; Hering SV; Brown NJ. (2008). Factors affecting the indoor concentrations of
carbonaceous aerosols of outdoor origin. Atmos Environ, 42: 5660-5671. 155949
Lung SC; Mao IF; Liu LJ. (2007). Residents' particle exposures in six different communities in Taiwan. Sci Total Environ,
377: 81-92. 156719
Macintosh D; Minegishi T; Kaufman M; Baker B; Allen J; Levy J; Myatt T. (2009). The benefits of whole-house in-duct air
cleaning in reducing exposures to fine particulate matter of outdoor origin: A modeling analysis. J Expo Sci
Environ Epidemiol, TBD: TBD. 190887
Mader BT; Yu JZ; Xu JH; Li QF; Wu WS; Flagan RC; Seinfeld JH. (2004). Molecular composition of the water-soluble
fraction of atmospheric carbonaceous aerosols collected during ACE-Asia. J Geophys Res, 109: D06206. 156724
Maitre A; Soulat JM; Masclet P; Stoklov M; Marques M; de Gaudemaris R. (2002). Exposure to carcinogenic air pollutants
among policemen working close to traffic in an urban area. Scand J Work Environ Health, 28: 402-410. 156726
Manchester-Neesvig JB; Schauer JJ; Cass GR. (2003). The distribution of particle-phase organic compounds in the
atmosphere and their use for source apportionment during the Southern California Children's Health Study. J Air
Waste Manag Assoc, 53: 1065-79. 098102
Mar TF; Koenig JQ; Jansen K; Sullivan J; Kaufman J; Trenga CA; Siahpush SH; Liu L-JS; Neas L. (2005). Fine particulate
air pollution and cardiorespiratory effects in the elderly. Epidemiology, 16: 681-687. 087566
Martuzevicius D; Grinshpun S; Lee T; Hu S; Biswas P; Reponen T; LeMasters G. (2008). Traffic-related PM2. 5 aerosol in
residential houses located near major highways: Indoor versus outdoor concentrations. Atmos Environ, 27: 6575-
6585.190886
Mathai CV; Watson Jr JG; Rogers CF; Chow JC; Tombach I; Zwicker JO; Cahill T; Feeney P; Eldred R. (1990).
Intercomparison of ambient aerosol samplers used in western visibility and air quality studies. Environ Sci Technol,
24: 1090-1099. 156741
Mayol-Bracero OL; Guyon P; Graham B; Roberts G; Andreae MO; Decesari S; Facchini MC; Fuzzi S; Artaxo P. (2002).
Water-soluble organic compounds in biomass burning aerosols over Amazonia: 2. Apportionment of the chemical
composition and importance of the polyacidic fraction. J Geophys Res, 107: 8091. 045010
Mazzoleni LR; Zielinska B; Moosmuller H. (2007). Emissions of Levoglucosan, Methoxy Phenols, and Organic Acids
from Prescribed Burns, Laboratory Combustion of Wildland Fuels, and Residential Wood Combustion. Environ Sci
Technol, 41: 2115-2122. 098038
Meng QY; Turpin BJ; Korn L; Weisel CP; Morandi M; Colome S; Zhang J; Stock T; Spektor D; Winer A; Zhang L; Lee
JH; Giovanetti R; Cui W; Kwon J; Alimokhtari S; Shendell D; Jones J; Farrar C; Maberti S. (2005). Influence of
ambient (outdoor) sources on residential indoor and personal PM2.5 concentrations: analyses of RIOPA data. J
Expo Sci Environ Epidemiol, 15: 17-28. 058595
Meng QY; Turpin BJ; Lee JH; Polidori A; Weisel CP; Morandi M; Colome S; Zhang JF; Stock T; Winer A. (2007). How
does infiltration behavior modify the composition of ambient PM2.5 in indoor spaces? An analysis of RIOPA data.
Environ Sci Tech, 41: 7315-7321. 194618
Meng QY; Turpin BJ; Polidori A; Lee JH; Weisel C; Morandi M; Colome S; Stock T; Winer A; Zhang J. (2005). PM2.5 of
ambient origin: Estimates and exposure errors relevant to PM epidemiology. Environ Sci Technol, 39: 5105-5112.
081194
Middlebrook AM; Murphy DM; Lee S-H; Thomson DS; Prather KA; Wenzel RJ; Liu D-Y; Phares DJ; Rhoads KP; Wexler
AS; Johnston MV; Jimenez JL; Jayne JT; Worsnop DR; Yourshaw I; Seinfeld JH; Flagan RC. (2003). A comparison
of particle mass spectrometers during the 1999 Atlanta Supersites project. J Geophys Res, 108: 8424. 042932
December 2009 A-362
-------
Miguel AH; Eiguren-Fernandez A; Jaques PA; Froines JR; Grant BL; Mayo PR; Sioutas C. (2004). Seasonal variation of
the particle size distribution of poly cyclic aromatic hydrocarbons and of major aerosol species in Claremont,
California. Atmos Environ, 38: 3241-3251. 123260
Mikel DK. (2001). Quality assurance final report for the Southern Oxidant Study Atlanta Supersite Field Experiment
August 3 - September 1, 1999. 156762
Mo mar P; Bellander T; Sallsten G; Boman J. (2007). Indoor and outdoor concentrations of PM2.5 trace elements at homes,
preschools and schools in Stockholm, Sweden. J Environ Monit, 9: 348-357. 156774
Molnar P; Gustafson P; Johannesson S; Boman J; Barregard L; Sallsten G. (2005). Domestic wood burning and PM
sub(2.5) trace elements: Personal exposures, indoor and outdoor levels. Atmos Environ, 39: 2643-2653. 156772
Molnar P; Johannesson S; Boman J; Barregard L; Sallsten G. (2006). Personal exposures and indoor, residential outdoor,
and urban background levels of fine particle trace elements in the general population. J Environ Monit, 8: 543-551.
156773
Na K; Cocker DR. (2005). Organic and elemental carbon concentrations in fine particulate matter in residences,
schoolrooms, and outdoor air in Mira Loma, California. Atmos Environ, 39: 3325-3333. 156790
Naumova YY; Offenberg JH; Eisenreich SJ; Meng QY; Polidori A; Turpin BJ; Weisel CP; Morandi MT; Colome SD; Stock
TH; Winer AM; Alimokhtari S; Kwon J; Maberti S; Shendell D; Jones J; Farrar C. (2003). Gas/particle distribution
of poly cyclic aromatic hydrocarbons in coupled outdoor/indoor atmospheres. Atmos Environ, 37: 703-719. 089213
Nerriere E; Zmirou-Navier D; Blanchard O; Momas I; Ladner J; Le Moullec Y; Personnaz M-B; Lameloise P; Delmas V;
Target A; Desqueyroux H. (2005). Can we use fixed ambient air monitors to estimate population long-term
exposure to air pollutants? The case of spatial variability in the Genotox ER study. Environ Res, 97: 32-42. 089481
Neususs C; Weise D; Birmili W; Wex H; Wiedensohler A; Covert DS. (2000). Size-segregated chemical, gravimetric and
number distribution-derived mass closure of the aerosol in Sagres, Portugal during ACE-2. Tellus B Chem Phys
Meteorol, 52: 169-184. 156804
Ng SP; Kendall M; Dimitroulopoulou C; Grossinho A; Chen LC. (2005). PM2.5 exposure assessment of the population in
Lower Manhattan area of New York City after the World Trade Center disaster. Atmos Environ, 39: 1979-1992.
155996
NIOSH. (1996). NIOSH Method 5040 Issue 1 (Interim): Elemental Carbon (diesel exhaust). National Institute for
Occupational Safety and Health. Cincinnati, OH.http://www.cdc.gov/niosh/docs/2003-154/method-e.html. 156810
NIOSH. (1999). NIOSH Method 5040 Issue 3 (Interim): Elemental Carbon (diesel exhaust). National Institute for
Occupational Safety and Health. Cincinnati, OH.http://www.cdc.gov/niosh/docs/2003-154/pdfs/5040.pdf. 156811
Noullett M; Jackson PL; Brauer M. (2006). Winter measurements of children's personal exposure and ambient fine particle
mass, sulphate and light absorbing components in a northern community. Atmos Environ, 40: 1971-1990. 155999
Ntziachristos L; Samaras Z. (2006). Combination of aerosol instrument data into reduced variables to study the consistency
of vehicle exhaust particle measurements. Atmos Environ, 40: 6032-6042. 116722
Ogulei D; Hopke PK; Chalupa DC; Utell MJ. (2006). Modeling Source Contributions to Submicron Particle Number
Concentrations Measured in Rochester, New York. Aerosol Sci Technol, 41: 179-201. 119975
Olfert JS; Kulkarni P; Wang J. (2008). Measuring aerosol size distributions with the fast integrated mobility spectrometer. J
Aerosol Sci, 39: 940-956. 156004
Paatero P. (1997). Least squares foumulation of robust non-negative factor analysis. Chemometr Intell Lab Syst, 37: 23-35.
087001
Paatero P. (1999). The Multilinear Engine: A Table-Driven, Least Squares Program for Solving Multilinear Problems,
including the n-Way Parallel Factor Analysis Model. J Comput Graph Stat, 8: 854-888. 156835
Paatero P; Hopke PK; Song XH; Ramadan Z. (2002). Understanding and controlling rotations in factor analytic models.
Chemometr Intell Lab Syst, 60: 253-264. 156836
Pancras JP; Ondov JM; Zeisler R. (2005). Multi-element electrothermal AAS determination of 11 marker elements in fine
ambient aerosol slurry samples collected with SEAS-II. Anal Chim Acta, 538: 303-312. 098120
Pang Y; Eatough NL; Modey WK; Eatough DJ. (2002). Evaluation of the RAMS continuous monitor for determination of
PM2.5 mass including semi-volatile material in Philadelphia, PA. J Air Waste Manag Assoc, 52: 563-572. 030353
December 2009 A-363
-------
ParekhPP; HusainL. (1981). Trace element concentrations in summer aerosols at rural sites in New York state and their
possible sources. Atmos Environ, 15: 1717-1725. 156840
Park K; Chow JC; Watson JG; Trimble DL; Doraiswamy P; Arnott WP; Stroud KR; Bowers K; Bode R; Petzold A; Hansen
AD. (2006). Comparison of continuous and filter-based carbon measurements at the Fresno supersite. J Air Waste
Manag Assoc, 56: 474-91. 098104
Park SS; Harrison D; Pancras JP; Ondov JM. (2005). Highly Time-Resolved Organic and Elemental Carbon Measurements
at the Baltimore Supersite in 2002. J Geophys Res, 110: D07S06. 156843
Park SS; Pancras JP; Ondov J; Poor N. (2005). Anew pseudodeterministic multivariate receptor model for individual
source apportionment using highly time-resolved ambient concentration measurements. J Geophys Res, 110:
D07S15. 156844
Paschold H; Maciejewska B; Li WW; Morales H; Pingitore NE. (2003). Elemental analysis of airborne particulate matter
and cooling water in west Texas residences. Atmos Environ, 37: 2681-2690. 156847
Pekney NJ; Davidson CI; Bein KJ; Wexler AS; Johnston MV. (2006). Identification of sources of atmospheric PM at the
Pittsburgh supersite, part I: single particle analysis and filter-based positive matrix factorization. Atmos Environ, 2:
411-423.086115
Petaja T; Mordas G; Manninen H; Aalto PP; Hmeri K; Kulmala M. (2006). Detection efficiency of a water-based TSI
condensation particle counter 3785. Aerosol Sci Technol, 40: 1090-1097. 156021
Peters TM; Gussman RA; Kenny LC; Vanderpool RW. (2001). Evaluation of PM25 size selectors used in speciation
samplers. Aerosol Sci Technol, 34: 422-429. 017108
Peters TM; Norris GA; Vanderpool RW; Gemmill DB; Wiener RW; Murdoch RW; McElroy FF; Pitchford M. (2001). Field
performance of PM25 federal reference method samplers. Aerosol Sci Technol, 34: 433-443. 016925
Peterson MR; Richards MH. (2002). Thermal-optical-transmittance analysis for organic, elemental, carbonate, total carbon,
and OCX2 in PM2.5 by the EPA/NIOSH method. Presented at Symposium on Air Quality Measurement Methods
and Technology, Pittsburgh. 156861
Petzold A; Kramer H; Schonlinner M. (2002). Continuous Measurement of Atmospheric Black Carbon Using a Multi-angle
Absorption Photometer. Environ Sci Pollut Res Int, 4: 78-82. 156863
Phares DJ; Rhoads KP; Johnston MV; Wexler AS. (2003). Size-resolved ultrafine particle composition analysis 2. Houston.
J Geophys Res, 108: 8420. 156866
Pitchford ML; Chow JC; Watson JG; Moore CT; Campbell DE; Eldred RA; Vanderpool RW; Ouchida P; Hering SV; Frank
NH. (1997). Prototype PM2.5 federal reference method field studies report - An EPA staff report. 156872
Polidori A; Arhami M; Sioutas C; Delfino RJ; Allen R. (2007). Indoor/Outdoor relationships, trends, and carbonaceous
content of fine particulate matter in retirement homes of the Los Angeles Basin. J Air Waste Manag Assoc, 57: 366-
379. 156877
PooreMW. (2000). Oxalic acid in PM2.5 particulate matter in California. J Air Waste Manag Assoc, 50: 1874-1875.
012839
Poore MW (2002). Levoglucosan in PM2.5 at the Fresno supersite. J Air Waste Manag Assoc, 52: 3-4. 051444
Qin X; Prather KA. (2006). Impact of biomass emissions on particle chemistry during the California Regional Particulate
Air Quality Study. Int J Mass Spectrom, 258: 142-150. 156895
Ramachandran G; Adgate JL; Pratt GC; Sexton K. (2003). Characterizing indoor and outdoor 15 minute average PM2.5
concentrations in urban neighborhoods. Aerosol Sci Technol, 37: 33-45. 195017
Rattigan OV; Hogrefe O; Felton HD; Schwab JJ; Roychowdhury UK; Husain L; Dutkiewicz VA; Demerjian KL. (2006).
Multi-year urban and rural semi-continuous PM25 sulfate and nitrate measurements in New York state: Evaluation
and comparison with filter based measurements. Atmos Environ, 40: 192-205. 115897
Rees SL; Robinson AL; Khlystov A; Stanier CO; Pandis SNSN. (2004). Mass balance closure and the Federal Reference
Method for PM2.5 in Pittsburgh, Pennsylvania. Atmos Environ, 38: 3305-3318. 097164
Reff A; Weisel CP; Zhang J; Morandi M; Stock T; Colome S; Winer A; Turpin BJ; Offenberg JH. (2007). A functional
group characterization of organic PM2.5 exposure: Results from the RIOPA study. Atmos Environ, 41: 4585-4598.
156045
December 2009 A-364
-------
Reimann C; De Caritat P. (2000). Intrinsic flaws of element enrichment factors (EFs) in environmental geochemistry.
Environ Sci Technol, 34: 5084-5091.013269
Rinehart LR; Fujita EM; Chow JC; Magliano K; Zielinska B. (2006). Spatial distribution of PM 2.5 associated organic
compounds in central California. Atmos Environ, 40: 290-303. 115184
Rojas-Bracho L; Suh HH; Catalano PJ; Koutrakis P. (2004). Personal exposures to particles and their relationships with
personal activities for chronic obstructive pulmonary disease patients living in Boston. J Air Waste Manag Assoc,
54: 54:207-217. 054772
Rossner P; Svecova V; Milcova A; Lnenickova Z; Solansky N; Sram RJ. (2008). Seasonal variability of oxidative stress
markers in city bus drivers - Part I. Oxidative damage to DNA. Mutat Res Fund Mol Mech Mutagen, 642: 14-20.
156927
Rotko T; Oglesby L; Kunzli N; Carrer P; Nieuwenhuijsen MJ; Jantunen M. (2002). Determinants of perceived air pollution
annoyance and association between annoyance scores and air pollution (PM25, NO2) concentrations in the
European EXPOLIS study. Atmos Environ, 36: 4593-4602. 037240
Rupprecht & Patashnick Company. (2003). Innovative instrument for an ambient air particulate mass measurement
standard. Rupprecht & Patashnick Co., Inc.. Albany, NY. Final Report 03-07. 157207
Russell M; Allen DT; Collins DR; Fraser MP. (2004). Daily, seasonal and spatial trends in PM25 mass and composition in
southeast Texas. Aerosol Sci Technol, 1: 14-26. 082453
Sabin LD; Kozawa K; Behrentz E; Winer AM; Fitz DR; Pankratz DV; Colome SD; Fruin SA. (2005). Analysis of real-time
variables affecting children's exposure to diesel-related pollutants during school bus commutes in Los Angeles.
Atmos Environ, 39: 5243-5254. 087728
Sabin LD; Lim JH; Stolzenbach KD; Schiff KC. (2005). Contribution of trace metals from atmospheric deposition to
stormwater runoff in a small impervious urban catchment. Water Res, 39: 3929-3937. 088300
Salma I; Ocskay R; Chi X; Maenhaut W. (2007). Sampling artefacts, concentration and chemical composition of fine
water-soluble organic carbon and hutnic-like substances in a continental urban atmospheric environment. Atmos
Environ, 41: 4106-4118. 113852
Samson PJ. (1978). Ensemble trajectory analysis of summertime sulfate concentrations in New York State. Atmos Environ,
12: 1889-1893. 188974
Samson PJ. (1980). Trajectory analysis of summertime sulfate concentrations in the northeastern United States. J Appl
Meteorol, 19: 1382-1394. 073010
Sanderson EG; Farant JP. (2004). Indoor and outdoor polycyclic aromatic hydrocarbons in residences surrounding a
Soderberg aluminum smelter in Canada. Environ Sci Technol, 38: 5350-5356. 156942
Sarnat JA; Brown KW; Schwartz J; Coull BA; Koutrakis P. (2005). Ambient gas concentrations and personal particulate
matter exposures: implications for studying the health effects of particles. Epidemiology, 16: 385-395. 087531
Sarnat JA; Schwartz J; Catalano PJ; Suh HH. (2001). Gaseous pollutants in particulate matter epidemiology: Confounders
or surrogates?. Environ Health Perspect, 109: 1053-1061. 019401
Sarnat SE; Coull BA; Ruiz PA; Koutrakis P; Suh HH. (2006). The influences of ambient particle composition and size on
particle infiltration in Los Angeles, CA residences. J Air Waste Manag Assoc, 56: 186-196. 089166
Sarnat SE; Coull BA; Schwartz J; Gold DR; Suh HH. (2006). Factors affecting the association between ambient
concentrations and personal exposures to particles and gases. Environ Health Perspect, 114: 649-654. 089784
Sarnat SE; Suh HH; Coull BA; Schwartz J; Stone PH; Gold DR. (2006). Ambient particulate air pollution and cardiac
arrhythmia in a panel of older adults in Steubenville, Ohio. Occup Environ Med, 63: 700-706. 090489
Schauer JJ; Cass GR. (2000). Source apportionment of wintertime gas-phase and particle-phase air pollutants using organic
compounds as tracers. Environ Sci Technol, 34: 1821-1832. 012225
Schauer JJ; Mader BT; Deminter JT; Heidemann G; Bae MS; Seinfeld JH; Flagan RC; Gary RA; Smith D; Huebert BJ;
Bertram T; Howell S; Kline JT; Quinn P; Bates T; Turpin B; Lim H-J; Yu JZ; Yang H; Keywood MD. (2003). ACE-
Asia intercomparison of a thermal-optical method for the determination of particle-phase organic and elemental
carbon. Environ Sci Technol, 37: 993-1001.037014
December 2009 A-365
-------
Schnelle-Kreis J; Sklorz M; Peters A; Cyrys J; Zimmermann R. (2005). Analysis of particle-associated semi-volatile
aromatic and aliphatic hydrocarbons in urban particulate matter on a daily basis. Atmos Environ, 39: 7702-7714.
112944
Schwab JJ; Felton HD; Rattigan OV; Demerjian KL. (2006). New York State urban and rural measurements of continuous
PM2.5 mass by FDMS, TEOM, and BAM: Evaluations and Comparisons with the FRM . J Air Waste Manag
Assoc, 56: 372-83. 098449
Schwab JJ; Hogrefe O; Demerjian KL; Dutkiewicz VA; Husain L; Rattigan OV; Felton HD. (2006). Field and Laboratory
Evaluation of the Thermo Electron 5020 Sulfate Particulate Analyzer. Aerosol Sci Technol, 40: 744 - 752. 098785
Schwartz J; Sarnat JA; Coull BA; Wilson WE. (2007). Effects of exposure measurement error on particle matter
epidemiology: a simulation using data from a panel study in Baltimore, MD. J Expo Sci Environ Epidemiol, 17: S2-
S10. 090220
Shalat SL; Lioy PJ; Schmeelck K; Mainelis G. (2007). Improving estimation of indoor exposure to inhalable particles for
children in the first year of life. J Air Waste Manag Assoc, 57: 934-939. 156971
Shao L; Yang S; Li H; Li W; Jones T; Sexton K; B,ruB, K; Li J; Zhao H. (2007). Associations between particle
physicochemical characteristics and oxidative capacity: An indoor PM10 study in Beijing, China. Atmos Environ,
41: 5316-5326. 156973
Shilton V; Giess P; Mitchell D; Williams C. (2002). The relationships between indoor and outdoor respirable particulate
matter: meteorology, chemistry and personal exposure. Indoor Built Environ, 11: 266-274. 049602
Sidhu S; Graham J; Striebich R. (2001). Semi-volatile and particulate emissions from the combustion of alternative diesel
fuels. Chemosphere, 42: 681-90. 155202
Slama R; Darrow L; Parker J; Woodruff TJ; Strickland M; Nieuwenhuijsen M; Glinianaia S; Hoggatt KJ; Kannan S; Hurley
F; Kalinka J; Sram R; Brauer M; Wilhelm M; Heinrich J; Ritz B. (2008). Meeting report: atmospheric pollution and
human reproduction. Environ Health Perspect, 116: 791-798. 156985
Smith TJ; Davis ME; Reaser P; Natkin J; Hart JE; Laden F; Heff A; Garshick E. (2006). Overview of particulate exposures
in the US trucking industry. J Environ Monit, 8: 711-720. 156990
Solomon P; Baumann K; Edgerton E; Tanner R; Eatough D; Modey W; Marin H; Savoie D; Natarajan S; Meyer MB.
(2003). Comparison of integrated samplers for mass and composition during the 1999 Atlanta supersites project. J
Geophys Res, 108: 8423. 156994
Solomon PA; Norris G; Landis M; Tolocka M. (2001). Chemical analysis methods for atmospheric aerosol components. In
Baron PA;Willeke K (Ed.),Aerosol measurement: principles, techniques, and applications (pp. 261-293). New York:
John Wiley & Sons Ltd. 157193
Solomon PA; Sioutas C. (2006). Continuous and semi-continuous methods for PM mass and composition. EM, x: 17-23.
156995
Sram RJ; Beskid O; Rossnerova A; Rossner P; Lnenickova Z; Milcova A; Solansky I; Binkova B. (2007). Environmental
exposure to carcinogenic polycyclic aromatic hydrocarbons—the interpretation of cytogenetic analysis by FISH.
ToxicolLett, 172: 12-20. 192084
Stanier CO; Khly stov AY; Chan WR; Mandiro M; Pandis SN. (2004). A Method for the In Situ Measurement of Fine
Aerosol Water Content of Ambient Aerosols: The Dry-Ambient Aerosol Size Spectrometer (DAASS). Aerosol Sci
Technol, 38: 215 - 228. 095955
Stohl A. (1996). Trajectory statistics-A new method to establish source-receptor relationships of air pollutants and its
application to the transport of particulate sulfate in Europe. Atmos Environ, 30: 579-587. 157014
Strand M; Hopke PK; Zhao W; Vedal S; Gelfand E; Rabinovitch N. (2007). A study of health effect estimates using
competing methods to model personal exposures to ambient PM2.5. J Expo Sci Environ Epidemiol, 17: 549-558.
157018
Strand M; Vedal S; Rodes C; Dutton SJ; Gelfand EW; Rabinovitch N. (2006). Estimating effects of ambient PM2.5
exposure on health using PM2.5 component measurements and regression calibration. J Expo Sci Environ
Epidemiol, 16: 30-38. 089203
Stranger M; Potgieter-Vermaak S; Van Grieken R. (2008). Characterization of indoor air quality in primary schools in
Antwerp, Belgium. Indoor Air, 18: 454-463. 190884
December 2009 A-366
-------
Stranger M; Potgieter-Vermaak S; Van Grieken R. (2009). Particulate matter and gaseous pollutants in residences in
Antwerp, Belgium. Sci Total Environ, 1182-1192: 407. 190883
Subbalakshmi Y; Patti AF; Lee GSH; Hooper MA. (2000). Structural characterisation of macromolecular organic material
in air particulate matter using Py-GC-MS and solid state 13 C-NMR. J Environ Monit, 2: 561-565. 157023
Subramanian R; Khlystov AY; Cabada JC; Robinson AL. (2004). Positive and negative artifacts in particulate organic
carbon measurements with denuded and undenuded sampler configurations. Aerosol Sci Technol, 1: 27-48. 081203
S0rensen M; Daneshvar B; Hansen M; Dragsted LO; Hertel O; Knudsen L; Loft S. (2003). Personal PM2.5 exposure and
markers of oxidative stress in blood. Environ Health Perspect, 111: 161-166. 157000
S0rensen M; Loft S; Andersen HV; Raaschou-Nielsen O; Skovgaard LT; Knudsen LE; Nielsen IV; Hertel O. (2005).
Personal exposure to PM25, black smoke and NO2 in Copenhagen: relationship to bedroom and outdoor
concentrations covering seasonal variation. JExpo Sci Environ Epidemiol, 15: 413-422. 089428
Takahama S; Wittig AE; Vayenas DV; Davidson CI; Pandis SN. (2004). Modeling the diurnal variation of nitrate during the
Pittsburgh Air Quality Study. J Geophys Res, 109: D16S06. 157038
Tanaka S; Yasushi N; Sato N; Fukasawa T; Santosa SJ; Yamanaka K; Ootoshi T. (1998). Rapid and simultaneous multi-
element analysis of atmospheric particulate matter using inductively coupled plasma mass spectrometry with laser
ablation sample introduction. J Anal At Spectrom, 13: 135-140. 157041
Tang C-S; Chang L-T; Lee H-C; Chan C-C. (2007). Effects of personal particulate matter on peak expiratory flow rate of
asthmatic children. Sci Total Environ, 382: 43-51. 091269
Thornburg JW; Stevens CD; Williams RW; Rodes CE; Lawless PA. (2004). A pilot study of the influence of residential
HAC duty cycle on indoor air quality. Atmos Environ, 38: 1567-1577. 157052
Toivola M; Aim S; Reponen T; Kolari S; Nevalainen A. (2002). Personal exposures and microenvironmental concentrations
of particles and bioaerosols. J Environ Monit, 4: 166-174. 026571
Tolbert PE; Klein M; Peel JL; Sarnat SE; Sarnat JA. (2007). Multipollutant modeling issues in a study of ambient air
quality and emergency department visits in Atlanta. J Expo Sci Environ Epidemiol, 17: S29-S35. 090316
Tran NK; Steinberg SM; Johnson BJ. (2000). Volatile aromatic hydrocarbons and dicarboxylic acid concentrations in air at
an urban site in the Southwestern US. Atmos Environ, 34: 1845-1852. 013025
Trenga CA; Sullivan JH; Schildcrout JS; Shepherd KP; Shapiro GG; Liu LJ; Kaufman JD; Koenig JQ. (2006). Effect of
particulate air pollution on lung function in adult and pediatric subjects in a Seattle panel study. Chest, 129: 1614-
1622. 155209
TurpinBJ; Weisel CP; Morandi M; Colome S; Stock T; Eisenreich S; Buckley B. (2007). Relationships of Indoor, Outdoor,
and Personal Air (RIOPA): Part II. Analyses of concentrations of particulate matter species. 157062
Tursic J; Podkrajsek B; Grgic I; Ctyroky P; Berner A; Dusek U; Hitzenberger R. (2006). Chemical composition and
hygroscopic properties of size-segregated aerosol particles collected at the Adriatic coast of Slovenia.
Chemosphere, 63: 1193-1202. 157063
Vallejo M; Ruiz S; Hermosillo AG; Borja-Aburto VH; Cardenas M. (2006). Ambient fine particles modify heart rate
variability in young healthy adults. J Expo Sci Environ Epidemiol, 16: 125-130. 157081
Van Roosbroeck S; Wichmann J; JanssenNAH; Hoek G; Van Wijnen JH; Lebret E; Brunekreef B. (2006). Long-term
personal exposure to traffic-related air pollution among school children, a validation study. Sci Total Environ, 368:
565-573. 090773
Vaughn D; O'Brien T; Roberts PT; Rice J. (2005). Comparison of integrated filter and semi-continuous measurements of
PM2.5 nitrate, sulfate, and carbon aerosols in the Speciation Trends Network (STN). 157089
Veltkamp PR; Hansen KJ; Barkley RM; Sievers RE. (1996). Principal component analysis of summertime organic aerosols
at Niwot Ridge, Colorado. J Geophys Res, 101: 19,495-19,504. 081594
Venkatachari P; Zhou L; Hopke P; Schwab J; Demerjian K; Weimer S; Hogrefe O; Felton D; Rattigan O. (2006). An
intercomparison of measurement methods for carbonaceous aerosol in the ambient air in New York City. Aerosol
Sci Technol, 40: 788-795. 105918
Vinzents PS; Moller P; Sorensen M; Knudsen LE; Herte LQ; Jensen FP; Schibye B; Loft S. (2005). Personal exposure to
ultrafine particles and oxidative DNA damage. Environ Health Perspect, 113: 1485-1490. 087482
December 2009 A-367
-------
Virkkula A; Ahlquist NC; Covert DS; Arnott WP; Sheridan PJ; Quinn PK; Coffman DJ. (2005). Modification, Calibration
and aFieldTest of an Instrument for Measuring Light Absorption by Particles. Aerosol Sci Technol, 39: 68-83.
157097
Voorhees KJ; Schulz WD; Kunen SM; Hendricks LJ; Currie LA; Klouda G. (1991). Analysis of Insoluble Carbonaceous
Materials from Airborne Particles Collected in Pristine Regions of Colorado. J Anal Appl Pyrol, 18:189-205.
157101
Wallace L; Williams R. (2005). Use of personal-indoor-outdoor sulfur concentrations to estimate the infiltration factor and
outdoor exposure factor for individual homes and persons. Environ Sci Technol, 39: 1707-1714. 057485
Wallace L; Williams R; Rea A; Croghan C. (2006). Continuous weeklong measurements of personal exposures and indoor
concentrations of fine particles for 37 health-impaired North Carolina residents for up to four seasons. Atmos
Environ, 40: 399-414. 088211
Wan ECH; Yu JZ. (2006). Determination of sugar compounds in atmospheric aerosols by liquid chromatography combined
with positive electrospray ionization mass spectrometry. J Chromatogr, 1107: 175-181. 157104
Wang D; Hopke PK. (1989). The use of constrained least-squares to solve the chemical mass balance problem. Atmos
Environ, 23: 2143-2150. 157105
Wang X; Fu J; Bi X; Sheng G. (2006). Hospital indoor PM10/PM2.5 and associated trace elements in Guangzhou, China.
Sci Total Environ, 366: 124-135. 157108
Ward TJ; Noonan CW; Hooper K. (2007). Results of an indoor size fractionated PM school sampling program in Libby,
Montana. Environ Monit Assess, 130: 163-171. 157112
Waterman D; Horsfield B; Leistner F; Hall K; Smith S. (2000). Quantification of poly cyclic aromatic hydrocarbons in the
NIST Standard Reference Material (SRM1649A) urban dust using thermal Desorption GC/MS. Anal Chem, 72:
3563-3567. 157116
Waterman D; Smith S; Green D; Horsfield B; Hall K. (2001). The application of a thermal desorption GCMS technique for
the organic analysis of airborne particulate matter. Adv Mass Spectrom, 14: 887-888. 157117
Watson JG; Chen LW; Chow JC; Doraiswamy P; Lowenthal DH. (2008). Source apportionment: findings from the U.S.
Supersites Program. J Air Waste Manag Assoc, 58: 265-288. 157128
Watson JG; Chow JC. (2001). Ambient air sampling. In Baron PA; Willeke K (Ed.),Aerosol measurement: principles,
techniques, and applications (pp. 821-844). New York: John Wiley & Sons Ltd. 157123
Watson JG; Chow JC. (2002). Comparison and evaluation of in situ and filter carbon measurements at the Fresno Supersite.
J Geophys Res, 107: 8341. 037873
Watson JG; Chow JC; Bowen JL; Lowenthal DH; Hering S; Ouchida P; Oslund W. (2000). Air quality measurements from
the Fresno supersite. J Air Waste Manag Assoc, 50: 1321-1334. 010354
Watson JG; Chow JC; Chen LWA. (2005). Summary of organic and elemental carbon/black carbon analysis methods and
intercomparisons. Aerosol Air Qual Res, 5: 69-102. 157125
Watson JG; Chow JC; Doraiswamy P; Chen LWA; Lowenthal DH; Trimble D; Park K. (2005). Quality Assurance Final
Report for the Fresno Supersite. 157124
Watson JG; Chow JC; Frazier CA. (1999). X-ray fluorescence analysis of ambient air samples. In Landsberger, S.;
Creatchman, M. (Ed.),Elemental analysis of airborne particles (pp. 67-96). Amsterdam, The Netherlands: Gordon
and Breach Science Publishers. 020949
Watson JG; Chow JC; Moosmilller H; Green MC; Frank NH; Pitchford ML. (1998). Guidance for using continuous
monitors in PM2.5 monitoring networks. U.S. Environmental Protection Agency. Research Triangle Park, NC.
EPA-454/R-98-012. http://www.epa.gov/ttn/amtic/pmpolgud.html. 198805
Watson JG; Chow JC; Shah JJ; Pace TG (1983). The effect of sampling inlets on the PM-10 and PM-15 to TSP
concentration ratios. J Air Waste Manag Assoc, 33: 114-119. 045084
Watson JG; Cooper JA; Huntzicker JJ. (1984). The effective variance weighting for least squares calculations applied to the
mass balance receptor model. Atmos Environ, 18: 1347-1355. 045693
December 2009 A-368
-------
Watson JG; Robinson NF; Lewis CW; Coulter CT; Chow JC; Fujita EM; Lowenthal DH; Corner TL; Henry RC; Willis RD.
(1997). Chemical mass balance receptor model version 8 (CMB) users manual. Washington, D.C.: U.S.
Environmental Protection Agency. 157121
Watson JG; Turpin BJ; Chow JC. (1989). The measurement process: precision, accuracy, and validity . In BS Cohen; CSJ
McCammon (Ed.),Air Sampling Instruments for Evaluation of Atmospheric Contaminants, Ninth Edition (pp. 201-
216). Cincinnati, OH: American Conference of Governmental Industrial Hygienists. 157119
Weber RJ; Orsini D; Daun Y; Lee Y-N; Kotz PJ; Brechtel F. (2001). A particle-into-liquid collector for rapid measurement
of aerosol bulk chemical composition. Aerosol Sci Technol, 35: 718-727. 024640
Weber RJ; Orsini D; Duan Y; Baumann K; Kiang CS; Chameides W; Lee YN; Brechtel F; Klotz P; Jongejan P. (2003).
Intercomparison of near real time monitors of PM2. 5 nitrate and sulfate at the US Environmental Protection
Agency Atlanta Supersite. J Geophys Res, 108: 8421. 157129
Weisel CP; Zhang J; Turpin BJ; Morandi MT; Colome S; Stock TH; Spektor DM; Korn L; Winer AM; Kwon J; Meng QY;
Zhang L; Harrington R; Liu W; Reff A; Lee JH; Alimokhtari S; Mohan K; Shendell D; Jones J; Farrar L; Maberti S;
Fan T. (2005). Relationships of Indoor, Outdoor, and Personal Air (RIOPA): Part I. Collection methods and
descriptive analyses. Res Rep Health Eff Inst, 130: 1-107. 157131
Welthagen W; Schnelle-Kreis J; Zimmermann R. (2003). Search criteria and rules for comprehensive two-dimensional gas
chromatography-time-of-flight mass spectrometry analysis of airborne particulate matter. J Chromatogr, 1019: 233-
249. 104056
Wenzel RJ; Liu DY; Edgerton ES; Prather KA. (2003). Aerosol time-of-flight mass spectrometry during the Atlanta
Supersite Experiment: 2. Scaling procedures. J Geophys Res, 108: 8426. 157139
Westerdahl D; Fruin S; Sax T; Fine PM; Sioutas C. (2005). Mobile platform measurements of ultrafine particles and
associated pollutant concentrations on freeways and residential streets in Los Angeles. Atmos Environ, 39: 3597-
3610. 086502
Wichmann J; JanssenNAH; Van Der Zee S; Brunekreff B. (2005). Traffic-related differences in indoor and personal
absorption coefficient measurements in Amsterdam, the Netherlands. Atmos Environ, 39: 7384-7392. 086240
Williams B; Goldstein A; Kreisberg N; Hering S. (2006). An In-Situ Instrument for Speciated Organic Composition of
Atmospheric Aerosols: Thermal Desorption Aerosol GC/MS-FID (TAG). Aerosol Sci Technol, 40: 627-638.
156157
Williams R; Rea A; Vette A; Croghan C; Whitaker D; Stevens C; McDow S; Fortmann R; Sheldon L; Wilson H; Thornburg
J; Phillips M; Lawless P; Rodes C; Daughtrey H. (2008). The design and field implementation of the Detroit
exposure and aerosol research study. J Expo Sci Environ Epidemiol, 19: 643-659. 191201
Williams R; Suggs J; Rea A; Sheldon L; Rodes C; Thornburg J. (2003). The Research Triangle Park particulate matter
panel study: modeling ambient source contribution to personal and residential PM mass concentrations. Atmos
Environ, 37: 5365-5378. 053338
Wilson WE; Brauer M. (2006). Estimation of ambient and non-ambient components of particulate matter exposure from a
personal monitoring panel study. J Expo Sci Environ Epidemiol, 16: 264-274. 088933
Winkler PM; Steiner G; Vrtala A; Vehkamaki H; Noppel M; Lehtinen KEJ; Reischl GP; Wagner PE; Kulmala M. (2008).
Heterogeneous Nucleation Experiments Bridging the Scale from Molecular Ion Clusters to Nanoparticles. Science,
319: 1374-1377. 156160
Wittig AE; Takahama S; Khlystov AY; Pandis SN; Hering S; Kirby B; Davidson C. (2004). Semi-continuous PM25
inorganic composition measurements during the Pittsburgh Air Quality Study. Atmos Environ, 38: 3201-3213.
103413
Wu C-F; Delfino RJ; Floro JN; Quintana;. (2005). Exposure assessment and modeling of particulate matter for asthmatic
children using personal nephelometers. Atmos Environ, 39: 3457-3469. 086397
Wu CF; Delfino RJ; Floro JN; Samimi BS; Quintana PJ; Kleinman MT; Liu LJ. (2005). Evaluation and quality control of
personal nephelometers in indoor, outdoor and personal environments. J Expo Sci Environ Epidemiol, 15: 99-110.
157155
Wu CF; Jimenez J; Claiborn C; Gould T; Simpson CD; Larson T; Liu LJS. (2006). Agricultural burning smoke in Eastern
Washington: Part II. Exposure assessment. Atmos Environ, 40: 5379-5392. 179950
December 2009 A-369
-------
Xiao H-Y; Liu C-Q. (2004). Chemical characteristics of water-soluble components in TSP over Guiyang, SW China, 2003.
Atmos Environ, 38: 6297-6306. 056801
Yang H; Jian ZY; Hang Ho SS; Xu J; Wu WS; Chun HW; Wang X; Wang L. (2005). The chemical composition of
inorganic and carbonaceous materials inPM25 in Nanjing, China. Atmos Environ, 39: 3735-3749. 102388
Yang H; Li Q; Yu JZ. (2003). Comparison of two methods for the determination of water-soluble organic carbon in
atmospheric particles. Atmos Environ, 37: 865-870. 156167
Yang W; Sohn J; Kim J; Son B; Park J. (2009). Indoor air quality investigation according to age of the school buildings in
Korea. J Environ Manage, 90: 348-354. 190885
Yao X; Fang M; Chan CK; Ho KF; Lee SC. (2004). Characterization of dicarboxylic acids in PM25 in Hong Kong. Atmos
Environ, 38: 963-970. 102213
Yeh S; Small MJ. (2002). Incorporating exposure models in probabilistic assessment of the risks of premature mortality
from particulate matter. J Expo Sci Environ Epidemiol, 12: 389-403. 040077
Yip FY; Robins TG; Parker EA; Israel BA; Brakefield-Caldwell W; Keeler GJ; Dvonch JTE. (2004). Personal exposures to
particulate matter among children with asthma in Detroit, Michigan. Atmos Environ, 38: 5227-5236. 157166
Yu KN; Cheung YP; Cheung T; Henry RC. (2004). Identifying the impact of large urban airports on local air quality by
nonparametric regression. Atmos Environ, 38: 4501-4507. 101779
Yu LE; Shulman ML; Kopperud R; Hildemann LM. (2005). Characterization of Organic Compounds Collected during
Southeastern Aerosol and Visibility Study: Water-Soluble Organic Species. Environ Sci Technol, 39: 707-715.
157167
Yue Z; Fraser M. (2004). Characterization of Nonpolar Organic Fine Particulate Matter in Houston, Texas. Aerosol Sci
Technol, 38: 60-67. 157169
Zhang J; Chameides WL; Weber R; Cass G; Orsini D; Edgerton E; Jongejan P; Slanina J. (2002). An evaluation of the
thermodynamic equilibrium assumption for fine particulate composition: Nitrate and ammonium during the 1999
Atlanta Supersite Experiment. J Geophys Res, 108: 8414. 157181
Zhang Q; Anastasio C. (2003). Free and combined amino compounds in atmospheric fine particles (PM2. 5) and fog waters
from Northern California. Atmos Environ, 37: 2247-2258. 157182
Zhang Q; Jimenez JL; Canagaratna MR; Jayne JT; Worsnop DR. (2005). Time- and size-resolved chemical composition of
submicron particles in Pittsburgh: Implications for aerosol sources and processes. J Geophys Res, 110: 1-19.
157185
Zhao W; Hopke PK; Norris G; Williams R; Paatero P. (2006). Source apportionment and analysis on ambient and personal
exposure samples with a combined receptor model and an adaptive blank estimation strategy. Atmos Environ, 40:
3788-3801. 156181
Zhao W; Rabinovitch N; Hopke PK; Gelfand EW (2007). Use of an expanded receptor model for personal exposure
analysis in schoolchildren with asthma. Atmos Environ, 41: 4084-4096. 156182
Zheng M; Cass GR; Schauer JJ; Edgerton ES. (2002). Source apportionment of PM2.5 in the southeastern United States
using solvent-extractable organic compounds as tracers. Environ Sci Technol, 36: 2361-2371. 026100
Zhou L; Hopke PK; Liu W. (2004). Comparison of two trajectory based models for locating particle sources for two rural
New York sites. Atmos Environ, 38: 1955-1963. 157190
Zhu Y; Hinds WC; Krudysz M; Kuhn T; Froines J; Sioutas C. (2005). Penetration of freeway ultrafine particles into indoor
environments. J Aerosol Sci, 36: 303-322. 190081
Zhu Y; Kuhn T; Mayo P; Hinds WC. (2005). Comparison of Daytime and Nighttime Concentration Profiles and Size
Distributions of Ultrafine Particles near a Major Highway. Environ Sci Technol, 39: 2531-2536. 157191
Zollner I. (2007). Concentrations of Particulate Matter in Schools in Southwest Germany. Inhal Toxicol, 19: 245-249.
157192
December 2009 A-370
-------
Annex B. Dosimetry
B.1. Ultrafine Disposition
Table B-1. Ultrafine disposition in humans.
Reference Study Group Aerosol
Study Protocol
Observations
Mills et
al.(2006,
0887701
Holler etal.
(2008,
1567711
Wiebert et al.
(2006,
1561541
Wiebert et al.
(2006,
157146)
Healthy
nonsmokers
(5 M, 5 F;
21-24yr)
Healthy
nonsmokers
(n = 9;
50 ± 11 yr)
QrYinkPrQ
Ol I lUrvCI o
/n - -in1
V ' "l
^1 + 1D \ir}
j i z iu yi i
COPD patients
(n - 7'
([!-/,
RQ 4- 10 \/r\
oy I iu y[j
Subjects having
varied health
status (9M, 6F;
46-74 yr)
6 healthy
5 asthmatic
4 smokers
Healthy
subjects (4M,
5F; 56 + 9 yr)
Asthmatics (2M,
3F; 59 + 6 yr)
Pnntrnl M M1
OUI ILIUI ^ 1 IVI,
Łn wr\
ou yi)
Carbon -
99mTc
108nm
CMD
(08 = 2.2)
Technegas
Generator
Carbon -
99mTc
~100nm
CMD
Tprhnpn3Q
icui M icyao
Generator
Carbon -
99mTc
87nmCMD
(eg = 1.7)
Tprhnpn3Q
icui M icyao
Qpnprgfor
Carbon -
99mTc
34nmCMD
(eg = 1.5)
T h
lecnnegas
oenerator
Lung activity in the lung was measured at 0, 1,
and 6 h post aerosol inhalation.
On two separate occasions, subjects inhaled
100 ml aerosol boli to target front depths of
150 and 800 ml into the lungs to target the
airways and alveoli, respectively. Retention
measured at 10min, 1.5, 5.5, 24 and 48 h
post inhalation. Isotope (99mTc) leaching from
particles assessed via filters in saline, blood,
and urine. 81 mKr utilized to assess ventilation.
Technegas system was modified to reduce
leaching of 99mTc radiolabel from particles.
The avg tidal volume during aerosol inhalation
was 1.8 L (range 0.8-3.3). Activity in chest
region measured at 0, 2, 24, 46, and 70 h after
inhalation. Leaching assessed in vitro and via
urine collection.
Slow deep aerosol inhalations with 10s breath
hold. Mean inhalation time of 6 min. Control
subject inhaled aerosol with loosely bound
radiolabel. Retention scans at 10 min, 60 min,
100 min, and 24 h post inhalation. Leaching
assessed in vitro and via collection of blood
and urine.
On avg, lung activity decreased 3.2 + 0.7% during the first h and
1.2 + 1.7% over the next 5 h. With 95.6% of the particles in the lungs
at 6 h post inhalation and no accumulation of radioactivity detected
over the liver or spleen, findings did not support rapid translocation
from the lungs into systemic circulation.
Shallow airways boli-Total deposition in airways (shallow boli) similar
between groups. Pattern of deposition was significantly more central
in the healthy subjects which was thought due to non-uniform
ventilation distribution in smokers and COPD patients as visualized
by gamma-camera scans. Airway retention after 1.5 h was
significantly lower in healthy subjects (89 + 6%) than smokers (97 +
3%) or COPD patients (96 + 6%). At 24 and 48 h, retention
significantly remained higher in COPD patients (86 + 6% and 82 +
6%) than healthy subjects (75 + 10% and 70 + 9%).
Deep alveolar boli - Total deposition in alveoli (deep boli) significantly
greater in smokers (64 + 11%) and COPD patients (62 + 5%) than
healthy subjects (50 + 8%). Alveolar retention of particles similar at all
times between groups. For example, at 48 h, 97 + 3% in healthy
subject, 96 + 3% in smokers, and 96 + 2% in COPD patients.
Retention at 24 and 48 correlated with isotope leaching, suggesting
that the small amount of clearance primarily reflected the
disassociation of 99mTc from the particles with little transport of
particles from the lungs.
Lung function not significantly different between healthy and affected
lungs. The aerosol deposition fraction was 41 + 10%. Lung retention
was 99 + 3%, 99 + 5%, and 99 + 10% at 24, 46, and 70 h post
inhalation. Cumulative in vitro leaching by 70 h was 2.6 + 0.96%.
Except for radiotracer leaching from particles (1.0 + 0.6% of initially
deposited activity in urine by 24 h), there was not significant
clearance from the lungs by 70 h. Individual leaching was not
correlated with individual retention.
Avg deposition fraction of 60 + 17% which was correlated with tidal
volume during aerosol inhalation (p = 0.01). Activity excreted in urine
over 24-h post inhalation was 51 % in the control subject (high 99mTc
disassociation) and 3.6 + 0.9% of deposited activity. In the blood of
the control subject, activity was 30%, 31%, and 5% of the deposited
activity at 20 min, 80 min, and 24-h (respectively), whereas it was
only 0.9 + 0.6%, 1.1 + 0.4%, and 1.5 + 0.5% the other 13 subjects at
these times. Lung retention in the control subject was 30% at 1-h and
18%at24h. In the remainder of subjects, lung retention was
approximately 100% through 24 h.
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009
B-1
-------
Table B- 2. Ultrafine disposition in animals.
Reference Study Group
Aerosol
Study Protocol
Observations
Bermudez et
al.(2004,
0567071
Fischer 344 rats,
females (6 wk)
B3C3F1 mice,
females (6 wk)
Hamsters,
females (6 wk)
Ti02:1.29-1.44pm
MMAD
(og = 2.46-3.65), 21
nm primary particles
Animals exposed 6 h/day, 5 day/wk,
foMSwk to 0.5, 2and10mg/m3.
Control animals exposed to filtered
air. Animals sacrificed at 0, 4,13,
26, and 56 (49 for hamsters) post-
exposure. Groups of 25 animals per
species and time point.
Ti02 pulmonary retention half-times for the low-, mid-, and high-
exposure groups, respectively: 63,132, and 365 days in rats; 48, 40,
and 319 days in mice; and 33, 37, and 39 days in hamsters.
Burden of Ti02 in lymph nodes increase with time postexposure in
mid- and high-dosed rats; in high-dosed mice; but was unaffected in
hamsters at any time or dosage group. In high-exposure groups of
mice, epithelial permeability remained elevated (~2x control groups)
out to 52 wk without signs of recovery. Epithelial permeability was
3-4* control in high exposed rats through 4 wk post exposure, but
approached control by 13 wk. Epithelial permeability was unaffected
in all groups of hamsters.
Chen et al.
(2006,
0879471
Sprague-Dawley
male rats
(220 ± 20 g)
Polystyrene
1251 radiolabel
Ultrafine: 56.4 nm
Fine: 202 nm
Intratracheal instillation of particles
in healthy rats or those pretreated
with IPS (12 h before particle
instillation). Healthy rats sacrificed
between 0.5-2 h and at 24 or 48 h
post-instillation. IPS treated rats
were sacrificed 0.5-2 h
post-instillation.
In healthy rats, there were no marked differences in lung retention or
systemic distribution between the Ultrafine and fine particles. Results
for healthy animals focused on Ultrafine particles which were
primarily retained in lungs (72 + 10% at 0.5-2 h; 65 + 1% at 1 day;
62 ± 5% at 5 days). Initially, there was rapid particle movement into
the blood (2 ± 1% at 0.5-2 h; 0.1 ± 0.1% at 5 days) and liver (3 ± 2%
at 0.5-2 h; 1 + 0.1% at 5 days). At 1 day post-instillation, about 13%
of the particles where in the urine or feces. Following IPS treatment,
Ultrafine accessed the blood (5 vs. 2%) and liver (11 vs. 4%) to a
significantly greater extent than fine particles.
Geiser et al. Wistar rats 20
(2005, adult males
0873621
(250+10 g)
Also included
in in vitro
studies
Ti02 (22 nm CMD,
1.7 og)
Spark generated
0.11 mg/m3
7.3x106
particles/cm3
Rats exposed 1-h via endotracheal
tube while anesthetized and
ventilated at constant rate. Lungs
fixed at 1 or 24-h postexposure.
Distributions of particles among lung compartments followed the
volume distribution of compartments and did not differ significantly
between 1 and 24-h post-inhalation. On avg, 79.3 + 7.6% of
particles were on the luminal side of the airway surfaces, 4.6 + 2.6%
in epithelial or endothelial cells, 4.8 + 4.5% in connective tissues,
and 11.3 + 3.9% within capillaries. Particles within cells were not
membrane-bound.
Kapp et al.
(2004,
1566241
Charles River
rats
5 young adult
male
(250+10 g)
Ti02 (22 nm CMD,
1.7 og)
Spark generated
Rats exposed 1-h via endotracheal
tube while anesthetized and
ventilated at constant rate. Lungs
fixed immediately postexposure.
Of particles in tissues, 72% were aggregates of 2 or more particles;
93% of aggregates were in round or oval shape aggregates, 7%
were needle-like. The size distribution of particles in lung tissues (29
nm CMD, 1.7 og) was remarkably similar to the aerosol; the small
discrepancy may have been due to differences sizing techniques. A
large 350 nm aggregate was found in a type II pneumocyte, a 37 nm
particle in a capillary close to the endothelial cells, and a 106 nm
particle within the surface-lining layer close to the alveolar
epithelium.
December 2009
B-2
-------
Table B- 3. In vitro studies of ultrafine disposition.
Reference
Animal
Particles
Study Protocol
Observations
Edetsberger Human cervix
et al. (2005, carcinoma cells
1557591 (HeLa cells)
Polystyrene spheres
(0.020 urn)
Cells incubated with polystyrene particles
having negative surface charges. Cell
cultures were naive or treated with
Genistein or Cytochalasin B (CytB) prior
to particle application. Genistein inhibits
endocytotic processes, expecially
caveolae internalization. CytB inhibits
actin polymerization and phagocytosis.
Particles translocated into cells by first measurement (~1 min
after particle application) independent of treatment group. In
naive cells, agglomerates of 88-117 nm were seen by 15-20
min and of 253-675 nm by 50-60 min after particle
application. Intracellular aggregates thought to be result from
particle incorporation into endosomes or similar structures. In
treated cells, onlya small number of agglomerates (161-308
nm) were found and only by 50-60 min. At 50-60 min, 90%
and 98% of particles were in the 20-40 nm range in naive and
treated cells, respectively. Particles did not translocate into
dead cells, rather they attached to outside of the cell
membrane.
Geiser et al. Porcine lung
(2005, macrophages(106
0873621 cell/mLuman red
blood cells (RBC; 8
Also included x i06cells/mL)
inhalation
study
Fluorescent
polystyrene spheres
(0.078, 0.2, and
1pm)
Gold shheres
(0.025 urn)
Cells cultured for 4 h with each sized
polystyrene spheres. RBC were
employed as a model of nonphagocytic
cells. Some macrophages cultures were
treated with cytochalasin D (cytD) to
inhibit phagocytosis. In addition, RBC
were also cultured with gold particles.
Of the non-cytD treated macrophages, 77 + 15%, 21 + 11%,
and 56 + 30% contained 0.078, 0.2, and 1 pm particles,
respectively. CytD treatment of macrophages effectively
blocked the phagocytosis of 1 pm particles, but did not alter
the uptake of the 0.078 and 0.2 pm particles. Human RBC
were found to contain 0.078 and 0.2 pm polystyrene spheres
as well as the 0.025 pm gold particles, which were not
membrane bound. In contrast, the RBC did not contain the
larger 1 pm polystyrene spheres. Results suggest that
ultrafine and fine (0.078 and 0.2 pm diameter) particles cross
cellular membranes by a non-endocytic (i.e. not involving
vesicle formation) mechanisms such as adhesive interactions
and diffusion.
Geys et al.
(2006,
1557891
Human alveolar
(A549) and
bronchial (Calu-3)
epithelial cellsRat
primary type II
pneumocytes
Amine- and
carboxyl-modified
fluorescent
polystyrene (46 nm)
Cells cultured in clear polyester
transwells with 0.4 or 3 pm pores.
Monolayer considered "tight" when <1%
sodium fluorescein moved from apical to
basolateral compartment. Particle
translocation assessed in transwells with
and without cells. Cells incubated with
particles for 14-16 h to assess
translocation from apical to basolateral
compartment.
Without cells, 13.5% of carboxyl-modified particles passed
through the 0.4 pm pores (n = 7) and 67.5% through 3 pm
pores (n = 3). Movement of the amine-modified particles was
4.2% through 0.4 pm pores (n = 7) and 52.7% through 3 pm
pores (n = 3). The integrity of the monolayerwas insufficient
for translocation studies using the A549 cells (0.4 and 4 pm
pore size) and rat pneumocytes (0.3 pm pore). Using 0.4 pm
pores, there was no detectable translocation through either
Calu-3 or rat pneumocyte monolayers. Using 3 pm pores,
~6% of both particle types passed through the Calu-3
monolayer; however, results were highly variable with no
translocation in 2 (of 5) and 3 (of 6) trials with carboxyl- or
amine-modified particles, respectively.
B.2. Olfactory Translocation
Table B-4. Olfactory particle translocation.
Reference Study Group
Aerosol
Study Protocol
Observations
DeLorenzo Squirrel
(1970,
156391)
monkeys
Young males
(1kg)
Silver-coated colloidal Intranasal instillation of 1 mL particle
gold (50 nm) suspension. Animals sacrificed at 0.25,
0.5,1, and 24-h after instillation.
Rapid movement (30-60 min) into olfactory bulbs. Within
30 min of being placed on nasal mucosa, particle
aggregates were seen in axoplasm of the fila olfactoria
Within 1 h, particles were in olfactory glomerulus.
Particles in the olfactory bulb were located preferentially
in mitochondria and not free in the cytoplasm.
Dormanet Crl:CDrats
al (2001
055433) Males (6 wk old)
Soluble and insoluble
Mn particle types;
MMAD= 1.3-2.1 urn;
GSD<2
Whole body exposure (6 h/day, 14
consecutive days) to 0, 0.03, 0.3, and
3 mg Mn/m . Tissues analyzed in six
animals per concentration exposed to
soluble (MnS04) or insoluble (Mn304)
aerosols.
Increased Mn levels in olfactory bulb observed following
MnS04 of > 0.3 mg Mn/m3 and following Mn304 of 3 mg
Mn/m . At 3 mg Mn/m , Mn levels were significantly
greater in olfactory bulb (1.4-fold) and striatum (2.7-fold)
following soluble Mn04 than insoluble Mn304. Mn levels
in the cerebellum were unaffected following all
exposures.
December 2009
B-3
-------
Reference
Dorman et
al. (2004,
155752)
Elder etal.
(2006,
089253)
Oberdb'rster
etal. (2004,
055639)
Persson et
al. (2003,
051846)
Study Group
Crl:CD rats
Males (6 wk old)
Fisher 344 rats
Males
(200-250 g)
Fisher 344 rats
Males(14wk;
284 ± 9 g)
Sprague-Dawley
male rats (150g)
Freshwater Pike
female (3 kg)
Aerosol
Soluble and insoluble
Mn particle types;
MMAD= 1.5-2 urn;
GSD=1.4-1.6
Mn oxide (~30 nm
equivalent sphere with
3-8 nm primary
particles)
Spark generated
0.5 mg/m3
18 x 106 particles/cm3
13C(36nmCMD, 1.7
°g)
Spark generated
65ZnCI2 dissolved in
0.1 M HCI
Study Protocol
Whole body exposure (6 h/day,
5 days/wk, 13 wk) to MnS04 at 0, 0.01 ,
0.1 , and 0.5 mg Mn/m3. Compared to
Mn phosphate (as hureaulite)
exposure of 0.1 mg Mn/m3. Brain Mn
levels assessed immediately following
90 days of exposure or 45 days
postexposure.
Whole body inhalation exposure to
either filtered air or Mn oxide for 1 2
days (6 h/day, 5 days/wk) with both
nares open or Mn oxide for 2 days (6
h/day) with right nostril blocked.
Intranasal instillation in left nostril of
Mn oxide particles or soluble MnCI2
suspended in 30 uL saline. Analyzed
Mn in the lung, liver, olfactory bulb, and
other brain regions.
Rats (n = 12 3 per time point) exposed
to 1 60 ug/m for 6 h in whole-body
chamber and sacrificed at 1 , 3, 5, and
7 day postexposure. Lung, olfactory
bulb, cerebrum, and cerebellum
removed for 13C analysis. Tissue
ISC-levels were determined by isotope
ratio mass spectroscopy and
background corrected for 13C levels in
unexposed controls (n = 3).
Rats: intransal (0.03 pg Zn in 10 pL) or
intraperitoneally (0.03 pg Zn in 100 pL);
autoradiography and y spec at 1 day or
1, 3, or 6 wk postexposure.
Pike: instilled (0.1 2 pgZn in 10|jL)in
right or both olfactory chambers, assayed
2wk postexposure
Observations
Relative to air, the insoluble hureaulite was significantly
increased at 90 days of exposure in the olfactory bulb,
but not striatum or cerebellum. The soluble Mn
phosphate showed a dose dependent increase in
olfactory bulb Mn levels at 90 days. At 0.1 mg Mn/m3, Mn
levels following Mn phosphate were significantly
increased in the olfactory bulb and striatum relative to
hureaulite and air exposures. At 45 days postexposure,
relative to air, olfactory bulb Mn levels only remained
increased Mn phosphate group at 0.5 mg Mn/m .
After 12 day exposure via both nostrils, Mn in the
olfactory bulb increased 3.5-fold, whereas in the lung Mn
concentrations doubled; there were also increases in the
striatum, frontal cortex, and cerebellum. After the 2 days
exposure with the right nostril blocked, Mn accumulated
in the mainly in the left olfactory bulb (~2.4-fold increase)
in to a lesser extent in the right olfactory bulb (1 .2-fold
increase). At 24-h post instillation, the left olfactory bulb
contained similar amounts of the poorly soluble Mn oxide
(8.2 ± 0.7%) and soluble MnCI2 (8.2 ± 3.6%) as a
percent of the amount instilled.
At 1 day postexposure, the lungs of rats exposed to ultrafme
13C particles contained 1.34 + 0.22 pgof 13C
(1.39 pg/g-lung) following background corrected. By 7 days
postexposure, the 13C concentration had decreased to
0.59 ug/g-lung. There was a significant and persistent
increase in 13C in the olfactory bulb of 0.35 pg/g on day 1,
which increased to 0.43 pg/g by day 7. Day 1 concentrations
of 13C in the cerebrum and cerebellum were also
significantly increased but the increase was inconsistent,
possibly reflecting translocation of particles from the blood
across the blood-brain barrier into brain regions.
Zn uptake in olfactory epithelium and transport along
olfactory neurons to olfactory bulb. Zn continued into
interior of olfactory bulb and in rat went into anterior
olfactory cortex. Zn found bound to both cellular
constituents and cytosolic components. Some Zn bound
to metallothionein in olfactory mucosa and olfactory bulb.
Wang et al. CD-1 (ICR) mice Rutile Ti02
(2007,
155147) 21 and 80 nm
Anatase Ti02155 nm
Twenty mice (n = 5 per group)
exposed 0 or 0.01 g-Ti02 per mL Dl.
Instilled 25 uL each day for 5 days,
then inhaled 10 uL every other day.
Mice sacrificed after 1 mo.
Rutile particles were observed to be column/fiber
shaped, whereas anatase was octahedral. Ti02 particles
taken up by olfactory bulb via the olfactory nerve layer,
olfactory ventricle, and granular cell layer of the olfactory
bulb. Fine Ti02 showed greater entry into the olfactory
bulb presumably due to aggregation of smaller rutile
particles that was not seen for the fine anatase particles.
Yu et al.
(2003,
Sprague-Dawley
male rats, 6 wk
Stainless steel
welding-fume
«,5um
Whole body exposure 2 h/day for 1,
15, 30, or 60 days
Low: 64 ± 4 mg/m3 (1.6 mg/m3 Mn)
High: 107 ± 6 mg/m3 (3.5 mg/m3 Mn)
Significant increases in cerebellum Mn at 15-30 days of
exposure.
Slight increases in Mn in substantia nigra, basal ganglia,
temporal cortex, and frontal cortex after 60 days.
Significant increase at 30 days in basal ganglia at low
dose. Authors suggested that pharmacokinetics and
distribution of welding fume Mn differs from pure Mn.
December 2009
B-4
-------
B.3. Clearance and Age
Table B-5. Studies of respiratory tract mucosal and macrophage clearance as a function of age.
Reference Animal
Particles
Study Protocol
Observed Effect(s)
NASAL AND TRACHEAL CLEARANCE
Hoetal. Human, Not applicable
(2001, males and
156549) females
Ninety subjects (47 M, 43 F; 52 ± 23 yr)
between 11 and 90 yr of age were recruited
to measure nasal saccharine clearance and
ciliary beat frequency.
Ciliary beat frequency (n = 90; r = -0.48, p
<0.0001) and nasal mucociliary clearance time
(n = 43; r = 0.64, p O.001) were correlated with
subject age. Nasal clearance times were
significantly (p <0.001) faster in individuals
under 40 yr of age (9.3 + 5.2 min) versus older
subjects (15.4 + 5.0 min). Results similar
between males and females.
Goodman et Humans, Radiolabed Teflon disks (1 mm
al. (1978, males and diameter, 0.8 mm thick)
071130) females
Tracheal mucus velocity following delivery via
bronchoscope to the tracheal mucosa. Ten
young (2 M, 8 F; 23 ± 3 yr) and ten elderly (2
M, 5 F; 63 ± 5 yr) nonsmokers served as
control subjects. Measurements were also
made in young smokers, ex-smokers, and
individuals with chronic bronchitis.
Young nonsmokers had a tracheal mucus
velocity of 10.1 ±3.5 mm/min which was
significantly faster than the velocity of
5.8 ± 2.6 observed in the elderly
nonsmokers.
Whaley et al. Beagle Macroaggregated albumin99mTc
(1987, dogs, labelled
156153) males and
females
Intratracheal instillation of 10- pi droplet of
labelled albumin in saline. Tracheal clearance
followed 25 min. Longitudinal measure
measurements in 5 males and 3 females when
young adults (2.8-3 yr), middle-aged (6.7-6.9 yr),
and mature (9.6-9.8 yr). Additional 5 females
and 3 males comprised immature group (9-10
mo) and 4 males and 4 females used as aged
group (13-16 yr).
Tracheal mucus velocity significantly
(p <0.05) greater in young (9.7 ± 0.6 [SE]
mm/min) and middle-aged (6.9 ± 0.5) groups
than in immature (3.6 ± 0.4), mature
(3.5 ± 0.8), and aged (2.9 ± 0.5) dogs.
Yeates et al. Humans, Radioaerosols 99mTc labelled
(1981, males and
095391) females
Tracheal mucus velocities compiled for 74
healthy non-smoking subjects (60 M, 14 F;
10-65 yr, mean 30 yr) from prior studies.
Forty-two (32 M, 10 F) inhaled albumin in
saline droplets (6.2-6.5 urn MMAD), Yeates et
al. (1975); twenty-two (21 M, 1 F) inhaled iron
oxide (4.2 urn MMAD), Yeates et al. (1981 b);
and ten (7 M, 3 F) inhaled monodisperse iron
oxide aerosol (7.5 urn MMAD), Leikauf et al.
(1981). Inhalations were via a mouthpiece
with an inspiratory flow of ~1 liter/sec.
Alognormal distribution of tracheal mucus
velocities was reported. Age did not appear
to affect velocities, e.g., 4.7 ± 2.5 mm/min in
18-24yr olds vs. 4.6 ± 3.2 mm/min in
individuals >30 yr of age. However, it should
be noted that only 2 subjects were greater
than 45 yr of age and that the data was
complied from three studies using differing
experimental techniques. Rather similar
tracheal mucus velocities in males (4.7 ± 3.0
mm/min) and females (4.9 ± 2.4 mm/min).
BRONCHI AND BRONCHIOLES CLEARANCE
Puchelle et
al. (1979,
006863)
Human,
males
7.4 urn MMAD99mTc labelled
resin
Mucociliary clearance measured for 1 h post
aerosol inhalation in 19 healthy non-smoking
males (21-69yrof age). Clearance measure
on two occasions in 16 individuals.
Negative correlation (r = -0.472, p <0.05)
between mucociliary clearance and age.
Younger subjects (n = 9;21-37yr) had1-h
clearance of 34 ± 14% which was
significantly greater than the 22 ± 8% found
in the older subjects (n = 5;>54yr).
Separated by 5.4 wk (on avg), there was a
good correlation between repeated
clearance measurements (r= 0.65, p
<0.001)
Svartengren Humans, 6 urn MMAD111 In labelled Teflon
etal.(2005, males and
157034) females
Small airway clearance measured in five age
groups (< 24 yr, n = 13; 25-29 yr, n = 8;
30-49yr, n = 7; 50-64, n = 9; >65 yr, n = 9) of
healthy subjects. Aerosol inhaled via
mouthpiece at extremely slow rate of 0.05
L/s. Activity in lungs measured at 1 day, 2
days, and 1, 2, and 3 wk post-exposure.
Under the presumption that most large airway
clearance was complete by 24 h, retention at
24 h was normalized to 100%.
Large and small airway clearance slowed
with increasing age. Clearance correlated
with age at all times (r = -0.46 to -0.50, -0.55,
-0.66, and -0.70 at 1 day, 2 days, 1 wk, 2 wk,
and 3 wk, respectively). Based on linear
regression, the clearance from 1 to 21 days
post-exposure was 47% in a 20 yr-old
versus 23% in an 80 yr-old. Lung function
was not a significant predictor of clearance
when age considered.
December 2009
B-5
-------
Reference Animal
Particles
Study Protocol
Observed Effect(s)
Vastag et al. Humans, Monodisperseerythrocytes99mTc
(1985, males and labelled
157088) females
Clearance measured for 1-h post-inhalation
in eighty healthy (59 M, 21 F; 43 ± 17 yr)
subjects who had never smoked. Smokers
and ex-smokers also studied. Aerosol
inhalation not described.
Clearance significantly associated with age.
Based on linear regression, total mucociliary
clearance at 1-h post-exposure was 46% in
a 20 yr old versus 23% in an 80 yr old.
Similar results for males and females.
ALVEOLAR CLEARANCE
Muhle et al. Fischer 3.5 urn MMAD 85Sr labelled
(1990, 344 rats polystyrene latex
006853)
Control animals compared across several Typical alveolar clearance half-time of 45
studies. Aerosol inhaled by short-term nose days in 5-mo-old rats compared to 74 days
only exposure. Alveolar clearance determined in 23-mo-old rats. Statistical significance of
by exponential fit to thoracic activity findings not proved.
measured over 75-100 days excluding the
first 15 days post-exposure.
December 2009
B-6
-------
Annex B References
Bermudez E; Mangum JB; Wong BA; Asgharian B; Hext PM; Warheit DB; Everitt JI. (2004). Pulmonary responses of
mice, rats, and hamsters to subchronic inhalation of ultrafine titanium dioxide particles. Toxicol Sci, 77: 347-357.
056707
Chen C-H; Xirasagar S; Lin H-C. (2006). Seasonality in adult asthma admissions, air pollutant levels, and climate: a
population-based study. J Asthma, 43: 287-292. 087947
DeLorenzo AJD. (1970). The olfactory neuron and the blood-brain barrier. In Taste and Smell in Vertebrates (pp. 151-175).
London: Churchill Livingstone. 156391
Dorman DC; McManus BE; Parkinson CU; Manuel CA; McElveen AM; Everitt JI. (2004). Nasal Toxicity of Manganese
Sulfate and Manganese Phosphate in Young Male Rats Following Subchronic (13-Week) Inhalation Exposure. Inhal
Toxicol, 16: 481-488. 155752
Dorman DC; Struve MF; James RA; Marshall MW; Parkinson CU; Wong BA. (2001). Influence of particle solubility on
the delivery of inhaled manganese to the rat brain: manganese sulfate and manganese tetroxide pharmacokinetics
following repeated (14-day) exposure. Toxicol Appl Pharmacol, 170: 79-87. 055433
Edetsberger M; Gaubitzer E; Valic E; Waigmann E; Kohler G. (2005). Detection of nanometer-sized particles in living cells
using modern fluorescence fluctuation methods. Biochem Biophys Res Commun, 332: 109-116. 155759
Elder A; Gelein R; Silva V; Feikert T; Opanashuk L; Carter J; Potter R; Maynard A; Ito Y; Finkelstein J; Oberdorster G.
(2006). Translocation of inhaled ultrafine manganese oxide particles to the central nervous system. Environ Health
Perspect, 114: 1172-1178. 089253
Geiser M; Rothen-Rutishauser B; Kapp N; Schurch S; Kreyling W; Schulz H; Semmler M; Im Hof V; Heyder J; Gehr P.
(2005). Ultrafine particles cross cellular membranes by nonphagocytic mechanisms in lungs and in cultured cells.
Environ Health Perspect, 113: 1555-1560. 087362
Geys J; Coenegrachts L; Vercammen J; Engelborghs Y; Nemmar A; Nemery B; Hoet PHM. (2006). In vitro study of the
pulmonary translocation of nanoparticles A preliminary study. Toxicol Lett, 160: 218-226. 155789
Goodman RM; Yergin BM; Landa JF; Golinvaux MH; Sackner MA. (1978). Relationship of smoking history and
pulmonary function tests to tracheal mucous velocity in nonsmokers, young smokers, ex-smokers, and patients with
chronic bronchitis. Am Rev Respir Dis, 117: 205-214. 071130
Ho JC; Chan KN; Hu WH; Lam WK; Zheng L; Tipoe GL; Sun J; Leung R; Tsang KW. (2001). The effect of aging on nasal
mucociliary clearance, beat frequency, and ultrastructure of respiratory cilia. Am J Respir Crit Care Med, 163: 983-
988. 156549
Kapp N; Kreyling W; Schulz H; Im Hof V; Gehr P; Semmler M; Geiser M. (2004). Electron energy loss spectroscopy for
analysis of inhaled ultrafine particles in rat lungs. Microsc Res Tech, 63: 298-305. 156624
Mills NL; Amin N; Robinson SD; Anand A; Davies J; Patel D; de la Fuente JM; Cassee FR; Boon NA; Macnee W; Millar
AM; Donaldson K; Newby DE. (2006). Do inhaled carbon nanoparticles translocate directly into the circulation in
humans?. Am J Respir Crit Care Med, 173: 426-431. 088770
Moller W; Felten K; Sommerer K; Scheuch G; Meyer G; Meyer P; Haussinger K; Kreyling WG (2008). Deposition,
retention, and translocation of ultrafine particles from the central airways and lung periphery. Am J Respir Crit Care
Med, 177: 426-432. 156771
Muhle H; Creutzenberg O; BellmannB; HeinrichU; Mermelstein R. (1990). Dust overloading of lungs: investigations of
various materials, species differences, and irreversibility of effects. J Aerosol Med, 1: S111-S128. 006853
Oberdorster G; Sharp Z; Atudorei V; Elder A; Gelein R; Kreyling W; Cox C. (2004). Translocation of inhaled ultrafine
particles to the brain. Inhal Toxicol, 16: 437-445. 055639
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009 B-7
-------
Persson E; Henriksson J; Tallkvist J; Rouleau C; Tjalve H. (2003). Transport and subcellular distribution of intranasally
administered zinc in the olfactory system of rats and pikes. Toxicology, 191: 97-108. 051846
Puchelle E; Zahm J-M; Bertrand A. (1979). Influence of age on bronchial mucociliary transport. Scand J Respir Dis, 60:
307-313.006863
Svartengren M; Falk R; Philipson K. (2005). Long-term clearance from small airways decreases with age. Eur Respir J, 26:
609-615. 157034
Vastag E; Matthys H; Kohler D; Gronbeck L; Daikeler G. (1985). Mucociliary clearance and airways obstruction in
smokers, ex-smokers and normal subjects who never smoked. Eur J Respir Dis, 139: 93-100. 157088
Wang C. (2007). Impact of direct radiative forcing of black carbon aerosols on tropical convective precipitation. Geophys
Res Lett, 34: 5709. 156147
Whaley SL; Muggenburg BA; Seiler FA; Wolff RK. (1987). Effect of aging on tracheal mucociliary clearance in beagle
dogs. JApplPhysiol,62: 1331-1334. 156153
Wiebert P; Sanchez-Crespo A; Falk R; Philipson K; Lundin A; Larsson S; Moller W; Kreyling W; Svartengren M. (2006).
No Significant Translocation of Inhaled 35-nm Carbon Particles to the Circulation in Humans. Inhal Toxicol, 18:
741-747. 156154
Wiebert P; Sanchez-Crespo A; Seitz J; Falk R; Philipson K; Kreyling WG; Moller W; Sommerer K; Larsson S; Svartengren
M. (2006). Negligible clearance of ultrafine particles retained in healthy and affected human lungs. Eur Respir J,
28: 286-290. 157146
YeatesDB; Gerrity TR; Garrard CS. (1981). Particle deposition and clearance in the bronchial tree. Ann BiomedEng, 9:
577-592. 095391
Yu IJ; Park JD; Park ES; Song KS; Han KT; Han JH; Chung YH; Choi BS; Chung KH; Cho MH. (2003). Manganese
distribution in brains of Sprague-Dawley rats after 60 days of stainless steel welding-fume exposure.
Neurotoxicology, 24: 777-785. 156171
December 2009 B-8
-------
Annex C.
Controlled Human Exposure Studies
Table C-1. Cardiovascular effects.
Study
Pollutant
Exposure
Findings
Reference: Barregard et al. Wood smoke
(2006, 091381)
Particle Size:
Subjects: 13 healthy adults Session 1: GMD 42 nm;
Gender: 6 M/7 F
Age: 20-56 yr
Session 2: GMD 112 nm
Particle Number/Count:
Session 1:180,000/cm3;
Session 2:95,000/cm3
Concentration:
Session 1: median:
279 ug/m3;
Session 2: median
243 ug/m3
Subjects exposed in two groups for 4 h to
filtered air, followed a wk later by a 4-h
exposure to wood smoke. Exposures
conducted with two 25-min periods of light
exercise. Other measured combustion
products:
Session 1: N02 (0.08 ppm), CO (13 ppm),
formaldehyde (114 ug/m3), acetaldehyde
(75 ug/m ), benzene (30 ug/m ),
1,3-butadiene (6.3 ug/m3);
Session 2: N02 (0.09 ppm), CO (9.1 ppm),
formaldehyde (64 ug/nv), acetaldehyde
(40 ug/m3), benzene (20 ug/m3),
1,3-butadiene (3.9 ug/m3).
Time to analysis: Immediately following
exposure as well as 3 and 20 h post-
exposure.
Statistically significant increase in plasma
factor VIII 20 h post wood smoke exposure
relative to filtered air. The factor Vlll/von
Willebrand ratio in plasma was increased with
wood smoke relative to filtered air at 0, 3, and
20 h post-exposure. Wood smoke exposure
increased the urinary excretion of free
8-iso-prostaglandin2a relative to clean air 20 h
post-exposure (n = 9). These findings were
more pronounced in session 1 than session 2
(similar mass concentration but higher number
concentration in Session 1).
Reference: Beckett et al. Ultrafine and fine zinc oxide
(2005,156261)
v Particle Size: UF:
Subjects: 12 healthy adults <0.1 urn; Fine: 0.1-1.0 urn
Gender: 6 M/6 F
Age: 23-52 yr
Particle Number/Count:
UF: 4.6 xio7/cm3; Fine: 1.9
x105/cm3
Concentration: 500 ug/m3
Subjects exposed via mouthpiece for 2 h
during rest to filtered air, ultrafine, and fine
zinc oxide in a randomized crossover study
design. Exposures were separated by at least
3wk.
Time to analysis: Immediately following
exposure and 3, 6,11,23, and 24 h after
exposure.
Exposure to ultrafine and fine zinc oxide did
not affect HRV (time and frequency domain
parameters) relative to clean air immediately
following exposure, or at 3, 6,11, and 23 h
post-exposure. Exposure did not affect blood
pressure through 24 h post-exposure. No
effects of exposure to either fine or ultrafine
zinc oxide observed on factor VII, von
Willebrand factor (vWf), tissue plasminogen
activator (t-PA), or fibrinogen. No effect of
exposure observed on peripheral blood cell
counts or levels of pro-inflammatory cytokines.
Reference: Blomberg et al. DE
(2005,191991)
Subjects: 15 older adults
(former smokers) with COPD
Age: 56-72 yr
Concentration: 300 ug/m3
Subjects exposed for 1 h with intermittent DE was not observed to affect blood levels of
exercise to DE and filtered air in a randomized C-reactive protein, fibrinogen, D-Dimer,
crossover study design. prothrombin factor 1-2, or von Willebrand
factor activity at 6 and 24 h post-exposure.
Time to analysis: 6 and 24 h post-exposure.
Reference: Brauner et al.
(2007, 091152)
Subjects: 29 healthy adults
Gender: 20 M/9 F
Age: 20-40 yr
Urban traffic particles
Particle Number/Count:
6-700nm:10,067/cm3
Concentration: PIvl;:
9.7 pg/m3; PMio-25-
12.6|jg/m3
Subjects exposed to urban traffic particles and
filtered air for 24 h with and without two
90-min periods of light exercise in a
randomized crossover study design.
Concentrations of NOX and NO were low and
did not differ between filtered and unfiltered
exposures. CO concentrations were higher
with filtered air (0.35 and 0.41 ppm), while 03
concentrations were lower with filtered air
(12.08 and 4.29 ppb).
Time to analysis: 6 and 24 h after the start of
exposure.
An increase in DNA strand breaks and
formamidopyrimidine-DNAglycosylase sites in
peripheral blood mononuclear cells were
observed after 6 and 24 h of exposure to urban
particulates. The particle concentration at the
57nm mode was shown to be the major
contributor to these effects.
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009
C-1
-------
Study
Pollutant
Exposure
Findings
Reference: Brauner et al.
(2008,156293)
Subjects: 42 healthy older
adults (21 couples)
Age: 60-75 yr
Indoor air particles
Particle Number/Count:
10-700 nm:10,016/cm3
Concentration: Coarse:
9.4 ug/m3; Fine: 12.6 ug/m3
Exposures consisted of two 48 h periods in
the home of each subject with or without the
use of a HEPAfilter (randomized crossover
design). HEPA filters reduced coarse
concentration from 9.4 to 4.6 ug/m3, and fine
concentration from 12.6 to 4.7 ug/m3.
Concentrations of N02 did not differ between
the 2 sessions (20 ppb).
Time to analysis: After the completion of
each 48 h session.
The use of HEPA filters significantly improved
microvascular function (p = 0.04) after 48 h
(reactive hyperemia-peripheral arterial
tonometry). Microvascular function was
assessed using a scoring system representing
the extent of reactive hyperemia. The
reduction in PM concentration through the use
of HEPA filters did not significantly affect blood
pressure following the 48-h exposures.
Lowering PM concentration did not significantly
affect inflammatory response markers in
peripheral venous blood (IL-6,TNF-a,
C-reactive protein, plasma amyloid A).
Reference: Brauner et al.
(2008,191966)
Subjects: 29 healthy adults
Gender: 20 M, 9 F
Age: M avg 27 yr, F avg
26 yr
Urban traffic particles
Particle Number/Count:
11,600/cm3
Concentration: PM25:
10.5|jg/m3;PMio.25:
13.8|jg/m3
Subjects exposed to urban traffic particles and
filtered air for 24 h with and without two
90-min periods of light exercise in a
randomized crossover study design.
Concentrations of NOX and NO were low and
did not differ between filtered and unfiltered
exposures. CO concentrations were higher
with filtered air, while 03 concentrations were
lower with filtered air.
Time to analysis: 6 and 24 h after the start of
exposure.
Exposure to urban traffic particles was not
observed to affect microvascular function
(digital peripheral artery tone) at 6 or 24 h after
the start of exposure. No difference in various
blood markers of coagulation, inflammation, or
protein oxidation (e.g., fibrinogen, platelet
count, CRP, IL-6, TNF- a) were demonstrated
between particle and filtered air exposure.
Reference: Carlsten et al.
(2007,155714)
Subjects: 13 healthy adults
Gender: 11 M/2 F
Age: 20-42 yr
DE
2002 Cummins B-series
diesel engine (6BT5.9G6,
5.9 L) operating at load
Concentration: Fine PM:
100, 200 ug/m3
Subjects exposed for 2 h at rest to filtered air
and each of the two DEPs concentrations in a
randomized crossover study design.
Exposures were separated by at least 2 wk.
Other diesel emissions measured: N02 (10-35
ppb), CO (0.7-1.8 ppm).
Time to analysis: 3, 6, and 22 h after the
start of exposure.
No statistically significant changes in
plasminogen activator inhibitor-1 (PAI-1), vWf,
D-dimer, or platelet count observed 3, 6, or
22 h following exposure to DE relative to
filtered air. Non-statistically significant
increases in D-dimer, vWf, and platelet count
were observed at 6 h following the start of
exposure (4 h post-exposure). No
diesel-induced increase in C-reactive protein
observed relative to filtered air in peripheral
venous blood at 1 or 20 h post-exposure.
Reference: Carlsten et al.
(2008,156323)
Subjects: 16 adults with
metabolic syndrome
Gender: 10M/6F
Age: 25-48 yr
DE
2002 Cummins B-series
diesel engine (6BT5.9G6,
5.9 L)
Concentration: Fine PM:
100, 200 ug/m3
Subjects exposed for 2 h at rest to filtered air
and each of the two DE particle
concentrations in a randomized crossover
study design. Exposures were separated by at
least 2 wk. Other diesel emissions measured:
N02 (30 ppb), NO (1.69 ppm), CO (0.65 ppm).
Time to analysis: 3, 7, and 22 h after the
start of exposure.
At 5 h after the end of diesel exposure (fine
particulate concentration 200 ug/nv), the
authors observed a significant decrease in vWf
in peripheral venous blood. No other changes
in thrombotic markers (vWf, D-dimer, PAI-1)
were observed at either concentration between
1 and 20 h post-exposure.
Reference: Danielsen et al.
(2008,156382)
Subjects: 13 healthy adults
Gender: 6 M/7 F
Age: 20-56 yr
Wood smoke
Particle Size:
Session 1: GMD 42 nm;
Session 2: GMD 112 nm
Particle Number/Count:
Session 1:180,000/cm3;
Session 2:95,000/cm3
Concentration:
Session 1: median:
279 ug/m3;
Session 2: median
243 ug/m3
Subjects exposed in two groups for 4 h to
filtered air, followed a wk later by a 4-h
exposure to wood smoke. Exposures
conducted with two 25-min periods of light
exercise. Other measured combustion
products:
Session 1: N02 (0.08 ppm), CO (13 ppm),
formaldehyde (114 ug/m), acetaldehyde
(75 ug/m3), benzene (30 ug/m3),
1,3-butadiene (6.3 ug/m3);
Session 2: N02 (0.09 ppm), CO (9.1 ppm),
formaldehyde (64 ug/nv), acetaldehyde
(40 ug/m3), benzene (20 ug/m3),
1,3-butadiene (3.9 ug/m3).
Time to analysis: 3 and 20 h post-exposure.
Exposure to wood smoke increased the mRNA
levels of hOGG1 in PBMCs relative to filtered
air 20 h after exposure. DNA strand breaks
were shown to decrease in PBMCs 20 h after
wood smoke exposure.
Reference: Devlin et al.
(2003, 087348)
Subjects: 10 healthy older
adults
Gender: 7 M/3 F
Age: Avg 66.9 yr
Fine CAPs (Chapel Hill, NC) Exposures conducted for 2 h at rest to filtered
air and CAPs in a randomized crossover
study design.
Concentration: Mean:
40.5 ug/m3, Range:
21.2-80.3 ug/m3
Time to analysis: Immediately following
exposure and 24 h post-exposure.
CAPs exposure resulted in statistically
significant reductions (p <0.05) in time domain
(PNN50) and frequency domain (HF power)
parameters relative to clean air immediately
following exposure. These relative decreases
were still apparent 24 h after exposure (p
<0.08).
December 2009
C-2
-------
Study
Pollutant
Exposure
Findings
Reference: Fakhri et al.
(2009,191914)
Subjects: 50 adults (40
healthy, 10 asthmatic)
Gender: 24 M/26 F
Age: 19-48yr
Fine CAPs (Toronto)
Concentration: 127±
62 ug/m3 with and without
co-exposure to 03 (114 ±
ppb)
Exposures conducted through a facemask
which covered the subject's nose and mouth.
Subjects were exposed to CAPs, 03, CAPs +
03 and filtered air for 2 h at rest in a
randomized crossover study design.
Time to analysis: Every 30 min during
exposure, with the final measurement made
immediately prior to the end of the exposure.
Exposure to CAPs or 03, alone or in
combination, resulted in no significant changes
in HRV or blood pressure relative to filtered air.
However, a negative concentration response
relationship was reported between CAPs
concentration with co-exposure to 03 and
SDNN, rMSSD, HF power and LF power
(statistically significant for LF power). Diastolic
blood pressure was observed to increase with
exposure to CAPs + 03, but not with either
pollutant alone. There was no difference in
response between asthmatics and healthy
subjects.
Reference: Frampton et al.
(2006, 088665)
Subjects: 16 asthmatic
adults, 40 healthy adults
Gender: Asthmatics:
8 M/8 F, Healthy: 20 M/20 F
Age: 18-40yr
Ultrafine EC
Particle Size: CMD -25 nm
Particle Number/Count:
10 ug/m3:-2.0 xioW;
25ug/m3:~7.0xio6/cm3;
50ug/m3:~10.8xio6/cm3
Concentration: 10,25, and
50 ug/m3
Study conducted using a randomized
crossover design with 2-h exposures.
Asthmatics (n = 16) exposed to filtered air and
10 ug/m3.12 healthy adults exposed to filtered
air and 10 ug/m3 at rest; 12 healthy adults
exposed to filtered air, 10 and 25 ug/m3 with
intermittent exercise; 16 healthy adults
exposed to filtered air and 50 ug/m with
intermittent exercise. Exposures were
conducted via mouthpiece.
Time to analysis: Immediately following
exposure as well as 3.5,21, and 45 h post-
exposure.
No effect of ultrafine particle exposure on
leukocyte counts or leukocyte expression of
adhesion molecules observed in healthy
subjects exposed at rest to 10 ug/m3. Among
healthy adults exposed to ultrafine carbon
during exercise, monocyte expression of
adhesion molecules CD54 and CD18
decreased relative to filtered air immediately
following exposure. An ultrafine
particle-induced decrease in PMN expression
of CD18 was also observed 0-21 h
post-exposure. Expression ofCDHbon
monocytes and eosinophils was reduced
following exposure to ultrafine particles in
exercising asthmatics 0-21 h post-exposure. A
decrease in total leukocyte count was
observed following ultrafine particle exposure
in exercising healthy and asthmatic subjects.
Reference: Gong et al.
(2004, 087964)
Subjects: 13 older adults
with COPD, 6 healthy older
adults
Gender: COPD: 5 M/8 F,
Healthy: 2 M/4 F
Age: COPD:avg 68 yr,
Healthy: avg 73 yr
Fine CAPs (Los Angeles)
Particle Size: 85% of mass
between 0.1 and 2.5 urn
Concentration: Mean:
194 ug/m3, Range:
135-229 ug/m3
Exposures to CAPs and filtered air
(randomized crossover) for 2 h with
intermittent light exercise (four 15-min
periods). Exposures were separated by at
least 2 wk.
Time to analysis: Immediately following
exposure as well as 4 and 22 h post-
exposure.
SDNN shown to decrease following CAPs
exposure relative to filtered air in healthy older
adults (4-22 h post-exposure). No
CAPs-induced changes in HRV were observed
in older adults with COPD. Ectopic heart beats
were observed to increase slightly with CAPs
relative to filtered air among healthy subjects,
but decreased among subjects with COPD.
Exposure to CAPs did not affect platelet or
white blood cell count, or levels of fibrinogen,
vWF, or factor VII.
Reference: Gong et al.
(2004, 055628)
Subjects: 12 adult
asthmatics, 4 healthy adults
Gender: Asthmatics:
4 M/8 F, Healthy: 2 M/2 F
Age: Asthmatics: avg 38 yr,
Healthy: avg 32 yr
Coarse CAPs (Los Angeles)
Particle Size: 80% of mass
between 2.5 and 10 urn,
20% of mass <2.5 urn
Concentration: Mean:
157 ug/m , Range:
56-218 ug/m3
Exposures to CAPs and filtered air
(randomized crossover) for 2 h with
intermittent light exercise (four 15-min
periods). Exposures were separated by at
least 2 wk.
Time to analysis: Immediately following
exposure as well as 4 and 22 h post-
exposure.
SDNN shown to decrease following CAPs
exposure relative to filtered air in healthy
adults (4-22 h post-exposure). Decrease in
PNN50 also observed in healthy adults at 4 h
post-exposure. No CAPs-induced decreases in
HRV demonstrated in asthmatics.
Reference: Gong et al.
(2008,156483)
Ultrafine CAPs (Los
Angeles)
Subjects: 14 adult Particle Number/Count:
asthmatics, 17 healthy adults 145,000/cm3, Range
39,000-312,000/cm3
Gender: Asthmatics:
9 M/5 F, Healthy: 5 M/12 F Concentration: Mean-
100 ug/m , Range-
Age: Asthmatics: 34 ± 12 yr, -| 3.277 uq/m3
Healthy: 24 ± 8 yr Ha
Subjects exposed for 2 h during intermittent
exercise (15-min periods) to both CAPs and
filtered air in random order. The first 7
subjects underwent whole body exposure,
while the remaining subjects were exposed
through a facemask. Facemask exposures
had higher particle counts but lower particle
mass than whole body exposures. Exposures
were separated by at least 2 wk.
Time to analysis: Immediately following
exposure as well as 4 and 22 h post-
exposure.
Relative to filtered air, exposure to ultrafine
CAPs resulted in a transient decrease in LF
power 4 h post-exposure. This effect of CAPs
on HRV was not influenced by health status.
CAPs exposure was not observed to affect any
other measures of HRV, blood pressure, or
blood markers of inflammation or coagulation.
There were no differences in response
observed between facemask and whole body
exposures.
December 2009
C-3
-------
Study
Pollutant
Exposure
Findings
Reference: Graff etal.
(2009,191981)
Subjects: 14 healthy adults
Gender: 8 M/6 F
Age: 20-34 yr
Coarse CAPs (Chapel Hill,
NC)
Concentration: 89
±49.5 ug/m3 (estimated
inhaled dose * 67% of
measured particle mass)
Subjects exposed for 2 h with intermittent
exercise (15-min periods) to coarse CAPs and
filtered air in a randomized crossover design.
Exposures were separated by at least 2 mos.
Time to analysis: 0-1 and 20 h post-
exposure.
At 20 h post-exposure, tPAwas observed to
decrease by 32.9% from baseline (pre-
exposure) per 10 ug/m3 increase in CAPs
concentration (p = 0.01). D-dimer
concentration decreased 11.3% per 10 ug/m3,
a change of marginal statistical significance (p
= 0.07). No other coarse CAPs-induced
changes in blood biomarkers of coagulation
(e.g.,vWF, factor VII, plasminogen, fibrinogen,
or PAI-1) or inflammation (e.g., CRP) were
observed. At 20 h post-exposure, overall HRV
(SDNN) was shown to decrease by 14.4%
relative to pre-exposure measurements per
10 ug/m3 increase in CAPs concentration. No
other changes in HRV were observed following
exposure to coarse CAPs.
Reference: Huang et al.
(2003, 087377)
Subjects: 38 healthy adults
Gender: 36 M/2 F
Age: Avg 26.2 ±0.7 yr
Fine CAPs (Chapel Hill, NC) Subjects exposed to CAPs (n = 30) or filtered The increase in blood fibrinogen following
air (n = 8) for 2 h with intermittent exercise exposure to fine CAPs reported by Ghio et al.
(subjects did not serve as their own controls). (2000, 012140) was shown to be associated
Component data of CAPs was available for 37 with copper, zinc, and vanadium content in the
of the 38 subjects. CAPs.
Time to analysis: 18 h after exposure.
Concentration:
23.1-311.1 ug/m3
Reference: Larsson et al.
(2007, 091375)
Subjects: 16 healthy adults
Gender: 10 M/6 F
Age: 19-59yr
Traffic particles (road
tunnel)
Particle Size: PM25, PM10;
PM25 mass constituted
~36%ofPM10mass
Particle Number/Count:
20-1,000 nm: 1.1 xio5/cm3,
< 100 nm: 0.85 xio5/cm3
Concentration: PM25-
46-81 ug/m3; PM10-
130-206 ug/m3
DE
Protocol 1 (n=8): idling
Deutz diesel engine
(F3M2011,2.2 1,500 rpm)
using gas oil
Protocol 2 (n=12): idling
Volvo diesel engine (TD45,
4.5 L, 4 cylinders, 680 rpm)
using Gasoil E10
Particle Number/Count:
Protocol!: 1.2 xio6/cm3;
Protocol 2:1.26 xio6/cm3
Concentration: Protocol 1:
348 ug/m3, Protocol 2:
330 ug/m3
Exposures were conducted for 2 h with
intermittent exercise in a room adjacent to a
busy road tunnel. Study used a randomized
crossover design with each subject also
exposed to normal air (control). Exposures
were separated by 3-10 wk. No exposures to
filtered air were conducted. Other traffic
emissions measured: NO (874 ug/m3), N02
(230 ug/m), CO (5.8 ug/m reported, likely
5.8 mg/m3).
Time to analysis: 14 h post-exposure.
No change in plasma levels of fibrinogen or
PAI-1 observed 14 h post-exposure.
Reference: Lucking et al.
(2008,191993)
Subjects: 20 healthy adults
Gender: M
Age: 21-44yr
In both protocols, exposures were conducted
with intermittent exercise (15-min periods) to
DE and filtered air in a randomized crossover
design with exposures separated by at least
one wk.
Protocol 1 (n=8): Exposures conducted for 2
h. Other diesel emissions measured: NOX
(0.58 ppm), N02 (0.23 ppm), NO (0.36 ppm)
CO (3.54 ppm), total hydrocarbon (2.8 ug/m ).
Time to analysis: 6 h post-exposure.
Protocol 2 (n= 12): Exposures conducted for
1h. Other diesel emissions measured: NOX
(2.78 ppm), N02 (0.62 ppm), NO (2.15 ppm),
CO (3.08 ppm), total hydrocarbon
(1.58 ug/m).
Time to analysis: 2 and 6 h post-exposure.
Thrombus formation was observed to increase
with diesel 2 and 6 h post-exposure using an
ex vivo perfusion chamber. Both platelet-
neutrophil and platelet-monocyte aggregates
increased relative to filtered air 2 h following
exposure to diesel (only evaluated in Protocol
2). Plasma concentrations of soluble CD40L
were also observed to increase with diesel.
Exposure to diesel was not shown to affect
total leukocyte, monocyte, or platelet counts.
Reference: Lund et al.
(2009,180257)
Subjects: 10 healthy adults
Gender: 4 M/6 F
Age: 18-40yr
DE
Idling Cummins diesel
engine (5.9 L) using
commercial No. 2 fuel
Particle Size: MMAD
0.10 urn
Concentration: 100 ug/m3
Subjects exposed for 2 h with intermittent
exercise (15-min periods) to DE and filtered
air in a randomized crossover study design.
Other diesel emissions measured: NOX (4.7
ppm), N02 (0.8 ppm), CO (2.8 ppm), total
hydrocarbons (2.4 ppm).
Time to analysis: 30 min and 24 h post-
exposure.
Exposure to diesel resulted in an increase
in MMP-9 plasma concentration and activity as
well as an increase in endothelin-1 plasma
concentration at both 30 min and 24 h post-
exposure.
December 2009
C-4
-------
Study
Pollutant
Exposure
Findings
Reference: Lundback et al.
(2009,191967)
Subjects: 12 healthy adults
Gender: M
Age: 21-30 yr
DE
Idling Volvo diesel engine
(TD45,4.5 L, 4 cylinders,
680 rpm) using Gasoil E10
Particle Number/Count:
1.26xio6/cm3
Concentration: 330 ug/m3
Subjects exposed for 1 h with intermittent
exercise (15-min periods) to DE and filtered
air in a randomized crossover study design.
Exposures were separated by at least one wk.
Other diesel emissions measured: NOX (2.78
ppm), N02 (0.62 ppm), NO (2.15 ppm), CO
(3.08 ppm), total hydrocarbon (1.58 ug/m ).
Time to analysis: 10,20,30, and 40 min
post-exposure.
Diesel-induced increase in arterial stiffness
(increases in augmentation pressure and
augmentation index, as well as decrease in
time to wave reflection) observed at 10 and 20
min post-exposure using radial artery pulse
wave analysis. No effect of diesel observed on
carotid-femoral pulse wave velocity which was
assessed 40 min post-exposure, but not at
earlier time points. No effect of diesel observed
on blood pressure 10-30 min post-exposure.
Reference: Mills et al.
(2005, 095757)
Subjects: 30 healthy adults
Gender: M
Age: 20-38 yr
DE
Idling 1991 Volvo diesel
engine (TD45, 4.5 L, 4
cylinders, 680 rpm)
Particle Number/Count:
1.2xio6/cm3
Concentration: 300 ug/m3
Subjects exposed for 1 h with intermittent
exercise (15-min periods) to DE and filtered
air in a randomized crossover study design.
Exposures were separated by two wk. Other
diesel emissions measured: N02 (1.6 ppm),
NO (4.5 ppm), CO (7.5 ppm), total
hydrocarbon (4.3 ppm), formaldehyde
(0.26 ug/m).
Time to analysis: 2-4 h post-exposure for 15
subject; 6-8 h post-exposure for the other 15
subjects.
Forearm blood flow increase (induced by
bradykinin, acetylcholine, and sodium
nitroprusside) was attenuated by DE 2 and 6 h
post-exposure. A6 mmHg increase in diastolic
blood pressure (p = 0.08) 2 h following
exposure to DE was observed relative to
filtered air control. Bradykinin-induced release
of t-PA was attenuated by diesel exposure 6 h
post-exposure. DE did not affect the release of
t-PA 2 h post-exposure. No diesel-induced
changes in serum IL-6 or TNF-a observed 6 h
post-exposure.
Reference: Mills et al.
(2007, 091206)
Subjects: 20 older adults
with prior myocardial
infarction
Gender: M
Age: 60 ± 1 yr
DE
Idling 1991 Volvo diesel
engine (TD45, 4.5 L, 4
cylinders, 680 rpm) using
low sulfur gas-oil E10
Particle Size: Median
particle diameter 54 nm,
Range 20-120 nm
Particle Number/Count:
1.26xio6/cm3
Concentration: 300 ug/m3
Subjects exposed for 1 h with intermittent
exercise (15-min periods) to DE and filtered
air in a randomized crossover study design.
Exposures were separated by at least two wk.
Other diesel emissions measured: NOX (4.45
ppm), N02 (1.01 ppm), NO (3.45 ppm), CO
(2.9 ppm), total hydrocarbon (2.8 ppm).
Time to analysis: During exposure and 6-8 h
post-exposure.
A greater increase in exercise induced
ST-segment depression and ischemic burden
was observed during exposure to DE than
clean air. No diesel-induced effects on
vasomotor dysfunction observed 6 h
post-exposure. Bradykinin-induced release of
t-PA was attenuated by diesel exposure
relative to filtered air 6 h post-exposure. Effect
of diesel on t-PA release was not evaluated at
earlier times post-exposure. No diesel-induced
changes in blood leukocyte counts or serum
C-reactive protein 6 h post-exposure.
Reference: Mills et al.
(2008,156766)
Subjects: 12 adults with
coronary heart disease, 12
healthy adults
Gender: M
Age:CHD:59±2yr,
Healthy: 54 ± 2 yr
Fine CAPs (Edinburgh,
Scotland, UK)
Particle Size: Mean
1.23 urn
Particle Number/Count:
99,400/cm3
Concentration:
190 ±37 ug/m3
Exposures conducted for 2 h with intermittent
exercise. Subjects exposed to CAPs and
filtered air using a randomized crossover
design with exposures separated by at least 2
wk.
Time to analysis: 2,6-8, and 24 h post-
exposure.
CAPs exposure had no significant effect on
vascular function in healthy adults or adults
with coronary heart disease 6-8 h
post-exposure (i.e., no change in forearm
blood flow as assessed using venous
occlusion plethysmographyj.The authors
attributed this lack of response to a low
concentration of combustion-derived particles.
Small increase in blood platelet and monocyte
concentration observed following CAPs
exposure. Exposure to CAPs did not affect
serum CRP concentration or total leukocyte or
neutrophil count.
Reference: Peretz et al.
(2007,156853)
Subjects: 5 healthy adults
Gender: M
Age: 20-31 yr
DE
2002 Cummins B-series
diesel engine (6BT5.9G6,
5.9 L); operating at 75% of
rated capacity
Concentration: Fine PM
50,100, 200 ug/m3
Subjects exposed for 2 h at rest to filtered air
and each of the three DE particle
concentrations in a randomized crossover
study design. Exposures were separated by at
least 2 wk. Other diesel emissions measured,
200 ug/m3 exposure: N02 (23 ppb), NO (1.75
ppm), CO (1.58 ppm).
Time to analysis: 6 and 22 h after the start of
exposure.
PBMC expression of 10 genes involved in the
inflammatory response were observed to be
significantly affected by exposure to DE at the
highest concentration tested (8 upregulated, 2
downregulated) 6 h after the start of exposure.
The expression of 4 genes (1 upregulated, 3
downregulated) associated with the
inflammatory response showed significant
changes 22 h after diesel exposure. PBMC
expression of 5 genes involved in the oxidative
stress pathways showed significant changes at
6 h after the start of diesel exposure at the
highest concentration tested (4 upregulated, 1
downregulated). 7 genes involved in the
oxidative stress pathways showed significant
changes at 22 h following exposure (4
upregulated, 3 downregulated).
December 2009
C-5
-------
Study
Pollutant
Exposure
Findings
Reference: Peretz et al.
(2008,156854)
Subjects: 17 adults with
metabolic syndrome, 10
healthy adults
Gender: Metabolic
syndrome: 11 M/6 F,
Healthy: 8 M/2 F
Age: Metabolic syndrome:
20-48 yr, Healthy: 20-42 yr
DE
2002 Cummins B-series
diesel engine (6BT5.9G6,
5.9 L) using No. 2 undyed
on-highway fuel; operating
at 75% of rated capacity
Particle Size: Median
particle diameter 1.04 urn
Concentration: Fine PM
100, 200 ug/m3
Subjects exposed for 2 h at rest to both
concentrations of DE as well as filtered air in a
randomized crossover design. Exposures
were separated by at least 2 wk. Other diesel
emissions measured, 100 ug/m3 exposure:
N02 (16.5 ppb), NO (0.96 ppm), CO (0.51
ppm); 200 ug/m exposure: N02 (24.7 ppb),
NO (1.54 ppm), CO (0.89 ppm).
Time to analysis: Immediately following
exposure (within 30 min post-exposure) and
3 h from the start of exposure.
Exposure to 200 ug/m elicited a statistically
significant decrease in brachial artery diameter
relative to filtered air immediately following
exposure. A smaller decrease in brachial artery
diameter was also observed following
exposure to DE at 100 ug/m3. Plasma levels of
endothelin-1 were observed to increase
following DE exposure (200 ug/m3). The
observed effects were more pronounced in
healthy subjects than in subjects with
metabolic syndrome. DE did not affect
endothelium-dependent flow-mediated
dilatation. No effect of DE on blood pressure
was demonstrated immediately following
exposure.
Reference: Peretz et al.
(2008,156855)
Subjects: 13 adults with
metabolic syndrome, 3
healthy adults
Gender: Metabolic
syndrome: 8 M/5 F, Healthy:
3M/OF
Age: Metabolic syndrome:
31-48 yr, Healthy: 24-39 yr
DE
2002 Cummins B-series
diesel engine (6BT5.9G6,
5.9 L) using No. 2 undyed
on-highway fuel; operating
at 75% of rated capacity
Concentration: Fine PM
100, 200 ug/m3
Subjects exposed for 2 h at rest to both
concentrations of DE as well as filtered air in a
randomized crossover design. Exposures
were separated by at least 2 wk. Other diesel
emissions measured, 100 ug/m3 exposure:
N02 (20.6 ppb), NO (0.95 ppm), CO (0.47
ppm); 200 ug/m3 exposure: N02 (28.3 ppb),
NO (1.63 ppm), CO (0.74 ppm).
Time to analysis: 1, 3, 6, and 22 h from the
start of exposure.
Exposure to 200 ug/m increased HF power
and decreased the LF/HF ratio 1h
post-exposure; however, this effect was not
consistent across subjects. No effect of DE
was observed at later time points. Subjects
with metabolic syndrome did not experience
greater changes in HRVthan healthy subjects.
Reference: Power et al.
(2008,191982)
Carbon and ammonium
nitrate particles
Subjects: 5 adults with mild- Concentration:
to-moderate allergic asthma
Gender: 1 M/4 F
Age: 28-51 yr
With co-exposure to 0.2ppm
03:255 ug/m3,
Without co-exposure to
0.2ppm03:313ug/m3
Subjects exposed for 4 h with intermittent
exercise (30-min periods) to filtered air,
particles, and particles + 03 in a crossover
study design. Exposures were separated by at
least 3 wk.
Time to analysis: 3 h 40 min from the start of
exposure.
Time and frequency domain HRV parameters
were not affected by particle exposure relative
to filtered air. However, exposure to particles
with 03 resulted in a significant decrease in
SDNN as well as changes to both high and low
frequency power normalized to the difference
between total and very low frequency power.
Reference: Routledge et al.
(2006, 088674)
Subjects: 20 older adults
with coronary artery disease,
20 healthy older adults
Gender: CAD: 17 M/3F,
Healthy: 10M/1 OF
Age: CAD: 52-74yr,
Healthy: 56-75 yr
Ultrafine carbon
Particle Size: <10-300 nm;
mode at 20-30 nm
Concentration: Ultrafine
carbon: 50 ug/m3; S02:200
ppb
Exposures conducted (head dome system) to
filtered air, Ultrafine carbon, S02, and Ultrafine
carbon + S02 for 1 h at rest using a
randomized crossover study design.
Time to analysis: Immediately following
exposure as well as 3 and 23 h post-
exposure.
No PM-induced changes in HRV observed
among subjects with coronary artery disease.
Among healthy subjects, small increase in
HRV (RR, SDNN, rMSSD, and LF power)
demonstrated immediately post-carbon
exposure. Relative to filtered air control,
exposure to Ultrafine carbon did not
significantly affect blood pressure in healthy
adults or adults with coronary artery disease
0-3 h post-exposure. Exposure to Ultrafine
carbon, either alone or with S02, did not affect
plasma levels of fibrinogen or D-dimer at 3 or
23 h post-exposure. Exposure to Ultrafine
carbon did not affect peripheral blood
leukocyte count or C-reactive protein levels 3
or 23 h post-exposure.
Reference: Rundell and Gasoline emissions
Caviston (2008,191986)
2.5 hp gasoline engine
Subjects: 15 healthy college running 10 s each min
athletes during exposure and in the
min prior to exposure
Gender: M
Age: Avg 19.5yr
Particle Size: PM1.0
Particle Number/Count:
Trial!: 336,730 ±
149,206/cm3;
Trial 2:396,200 ±
82,564/cm3
Subjects were exposed twice to both clean air
and dilute gasoline exhaust during 6-min
periods of maximal exercise on a cycle
ergometer. Clean air exposures occurred first
and were separated by 3 days. Gasoline
exhaust exposures were also separated by 3
days, with the first occurring 7 days after the
second clean air exposure. Other emissions
measured: CO (6.3 ± 3.4 ppm).
Time to analysis: 6 min
There was no difference in total work done (kJ)
between the clean air exposures or between
the clean air exposures and the first exposure
to gasoline exhaust. However, the second
gasoline exhaust exposure was demonstrated
to significantly decrease work accumulated
over the 6min exercise period compared with
either of the other exposure conditions (p <
0.01).
December 2009
C-6
-------
Study
Pollutant
Exposure
Findings
Reference: Samet et al.
(2007,156940)
Subjects: Ultrafine CAPs:
20 healthy adults, Coarse
CAPs: 14 healthy adults
Gender: Ultrafine CAPs:
11 M/9F, Coarse CAPs:
8M/6F
Age: Ultrafine CAPs:
18-35yr, Coarse CAPs:
18-35yr
CAPs (Chapel Hill, NC)
Particle Size: Ultrafine
0.049 ± 0.009 urn; Coarse
3.59 ± 0.58 urn
Concentration: Ultrafine
47.0 ±20.2 ug/m3; Coarse
89.0 ±49.5 ug/m3
Preliminary report comparing effects of
controlled exposures to coarse, fine, and
Ultrafine CAPs among healthy adults (3
separate studies). A randomized crossover
design was used in evaluating effects of
coarse CAPs (n=14) and Ultrafine CAPs
(n=20) relative to filtered air following 2-h
exposures with intermittent exercise. Results
compared with previous study of controlled
exposure to fine CAPs (Chapel Hill, NC)
where subjects did not serve as their own
controls (Ghio et al., 2000, 012140).
Time to analysis: 0-20 h post-exposure.
Statistically significant decrease in SDNN
observed 20 h following exposure to coarse
CAPs relative to filtered air. Subjects in the
high Ultrafine CAPs group experienced a
decrease in SDNN based on an analysis of
24 h ambulatory Holler monitoring relative to
filtered air. Fine CAPs did not significantly
affect HRV. Increased levels of D-dimer
observed 18 h following exposure to Ultrafine
CAPs. No CAPs-induced changes in plasma
factor VII, plasminogen, fibrinogen, PAI-1, vWf,
or t-PA. No CAPs-induced changes in
C-reactive protein levels were observed.
Reference: Samet et al.
(2009,191913)
Subjects: 19 healthy adults
Gender: 10M/9F
Age: 18-35 yr
Ultrafine CAPs (Chapel Hill,
NC)
Particle Size: < 0.16 urn
Particle Number/Count:
120,662 ±48,232
particles/cm
Concentration: 49.8 ±
20 ug/m3
Subjects exposed for 2 h with intermittent 15
periods of exercise to UF CAPs and filtered
air using a randomized crossover study
design.
Time to analysis: Immediately following
exposure and 1 and 18 h post-exposure.
UF CAPs exposure resulted in an increase in
plasma concentrations of D-dimer both
immediately following exposure (20.6%
increase per 105 particles/cm3) as well as 18 h
post-exposure (18.2% increase per 105
particles/crtv). Plasma concentration of PAH
also increased with UF CAPs, although this
increase was not statistically significant (24%
increase, p = 0.1). No UF CAPs-induced
changes observed in plasma concentrations of
tPA, vWF, CRP, fibrinogen, plasminogen,
or Factor VII. HF and LF power were both
observed to increase with UF CAPs exposure
relative to filtered air at 18 h post-exposure
(41.8% and 36%, respectively, per 10
particles/cm3 increase in UF CAPs). UF CAPs
expressed as mass concentration was not
observed have a statistically significant effect
in HF total power. UF CAPs was not observed
to affect time domain measures of HRV over
24 h. The QT interval was shown to decrease
both immediately following and at 18 h post
exposure (not statistically significant
immediately following exposure).
Reference: Shah et al. Ultrafine EC
(2008, 156970)
Particle Number/Count:
Subjects: 16 healthy adults 10.8 ± 1 .7 x 106 /cm3
Age: 26.9 ± 6.9 yr Concentration: 50 ug/m3
Exposures conducted via mouthpiece for 2 h
with intermittent exercise to filtered air and
Ultrafine carbon in a randomized crossover
study design.
Time to analysis: Immediately following
exposure as well as 3.5, 21 , and 45 h post-
exposure.
Exposure to Ultrafine carbon attenuated peak
forearm blood flow after ischemia relative to
filtered air 3.5 h post-exposure. Venous nitrate
levels were significantly lower at 21 h following
exposure to UF carbon compared with filtered
air exposure. PM exposure was not observed
to affect blood pressure relative to filtered air at
times 0-45 h post-exposure.
Reference: Tornqvist et al.
(2007, 091279)
Subjects: 15 healthy adults
Gender: M
Age: 18-38yr
DE
Idling 1991 Volvo diesel
engine (TD45, 4.5 L, 4
cylinders, 680 rpm)
Concentration: 300 ug/m3
Subjects exposed for 1 h with intermittent
exercise (15-min periods) to DE and filtered
air in a randomized crossover study design.
Exposures were separated by at least two wk.
Other diesel emissions measured: NOX (4.44
ppm), N02 (0.82 ppm), NO (3.62 ppm), total
hydrocarbon (2.21 ppm).
Time to analysis: 24 h post-exposure.
DE was observed to significantly attenuate
endothelium-dependentvasodilation 24 h
post-exposure. Endothelium-independent
vasodilation was not affected by diesel
exposure. Exposure to DE did not affect blood
pressure relative to filtered air 24 h after
exposure. DE significantly increased plasma
levels of IL-6 and TNF-a 24 h following
exposure. Exposure to diesel resulted in an
increase in total antioxidant capacity of plasma
relative to filtered air 24 h post-exposure.
Reference: Urch et al.
(2004, 055629)
Subjects: 24 healthy adults
Gender:! 4 M/1 OF
Age: 35±10yr
Fine CAPs (Toronto)
Concentration: 150 ug/m3
(range 101-257 ug/m) with
120ppb03
Exposures conducted through a facemask
which covered the subject's nose and mouth.
Subjects were exposed to CAPs + 03 and
filtered air for 2 h at rest in a randomized
crossover study design. Exposures were
separated by at least 2 days.
Time to analysis: Immediately following
exposure.
CAPs + 03 exposure resulted in a significant
decrease in brachial artery diameter
immediately post-exposure (Brook et al., 2002,
024987). which was demonstrated to be
associated with both the organic and EC
fractions of the CAPs.
December 2009
C-7
-------
Study
Pollutant
Exposure
Findings
Reference: Urch et al.
(2005, 081080)
Subjects: 23 healthy adults
Gender:! 3 M/1 OF
Age: 32±10yr
Fine CAPs (Toronto);
Concentration: 150 ug/m3
(range 102-214 ug/m) with
120ppb03
Exposures conducted through a facemask
which covered the subject's nose and mouth.
Subjects were exposed to CAPs + 03 and
filtered air for 2 h at rest in a randomized
crossover study design.
Time to analysis: Every 30 min during
exposure, with the final measurement made
immediately prior to the end of the exposure.
An increase in diastolic blood pressure of 6
mmHg was observed at the end of CAPs + 03
exposure, which was statistically different from
the change in blood pressure experienced
during exposure to filtered air (1 mmHg). This
effect was associated with the organic fraction
ofPM2.5.
Reference: Zarebaetal.
(2009,190101)
Subjects: 24 healthy adults
Gender: 12 M/12 F
Age: 18-40yr
Ultrafine EC
Particle Size: Count
median diameter 25 nm
Particle Number/Count:
2x106/cm3(10ug/m3),
7x106/cm3(25ug/m3)
Concentration: 10 ug/m3;
25 ug/m3
Protocol 1 (n=12, 6 M/6 F): Subjects exposed
to 10 ug/m UF carbon and filtered air for 2 h
at rest in a randomized crossover design.
Exposures were separated by at least 2 wk.
Protocol 2 (n=12, 6 M/6 F): Subjects exposed
to 10 ug/m , 25 ug/m3, and filtered air for 2 h
with intermittent exercise (15-min periods) in a
restricted randomized crossover design (all
subjects exposed to 10 ug/m3 before
25 ug/m3). Exposures were separated by at
least 2 wk.
Time to analysis (both protocols):
Immediately following exposure and 3.5 and
21 h post-exposure.
Exposure to 10 ug/m at rest resulted in no
change in HRV frequency domain parameters
relative to filtered air exposure. Time domain
parameters were observed to increase slightly
with UF carbon exposure (10 ug/m at rest),
however, only the increase in rMSSD was
statistically significant (p = 0.032). Some
trends toward less shortening of QT interval,
increase in ST segment, and increase in
variability of repolarization (variability of T
wave complexity) were observed with
exposure to 10 ug/m3 at rest, none of which
were statistically significant.
In Protocol 2, exposure to 10 ug/m3 UF carbon
was observed to slightly increase HRV time
domain parameters as was demonstrated in
Protocol 1. However, this was not observed at
the higher concentration (25 ug/m3). As with
exposure at rest, exposure to UF carbon
during exercise was observed to affect
repolarization (reduction in QT duration and
increase in T-wave amplitude), although this
effect was not statistically significant.
December 2009
C-8
-------
Table C-2. Respiratory effects.
Reference
Pollutant
Exposure
Findings
Reference: Alexis et al.
(2006, 1543231
Subjects: 9 healthy adults
Gender: 3 M/6 F
Age: 18-35 yr
Coarse fraction particles
(Chapel Hill, NC)
Heat-treated (biologically
inactive) and non-heated
particles
Particle Size: MMAD 5 urn
Concentration: 0.65 mg per
subject
Subjects were administered heat-treated
PM10.2.5, non-heated PM^.s, and 0.9% saline
(control) via nebulization in a randomized
crossover study design. Exposures were
separated by at least 1 wk.
Time to analysis: 2-3 h post-inhalation.
Both heat-treated and non-heated coarse PM
were observed to increase neutrophil counts in
induced sputum 2-3 h post-inhalation. Biologically
active PM (non-heated) induced an increase
expression of macrophage TNF-a mRNA, eotaxin,
and immune surface phenotypes on
macrophages (mCD14, CD11b/CR3, and
HLA-DR).
Reference: Barregard et al.
(2008, 1556751
Subjects: 13 healthy adults
Gender: 6 M/7 F
Age: 20-56 yr
Wood smoke
Particle Size:
Session 1: geometric mean
diameter 42 nm, Session 2:
geometric mean diameter 112
nm
Particle Number/Count:
Session 1:180,000/cm3;
Session 2: 95,000/cm3
Concentration: Session 1:
median 279 pg/m3; Session 2:
median 243 pg/m3
Subjects exposed in two groups for 4 h to filtered
air, followed a wk later by a 4-h exposure to
wood smoke. Exposures conducted with two
25-min periods of light exercise. Other measured
combustion products:
Session 1: N02 (0.08 ppm), CO (13 ppm),
formaldehyde (114 pg/m ), acetaldehyde
(75 pg/m3!, benzene (30 pg/m3), 1,3-butadiene
(6.3 pg/m3);
Session 2: N02 (0.09 ppm), CO (9.1 ppm),
formaldehyde (64 pg/m3), acetaldehyde
(40 pg/m3!, benzene (20 pg/m3), 1,3-butadiene
(3.9 |jg/m3).
Time to analysis: Immediately following
exposure as well as 3 and 20 h post-exposure.
Relative to filtered air, exposure to wood smoke
was observed to increase levels of eNO 3 h
post-exposure. Serum Clara cell protein
increased 20 h after wood smoke exposure.
Wood smoke was observed to increase levels of
malondialdehyde in breath condensate
immediately after as well as 20 h post-exposure.
Effects of wood smoke on eNO and
malondialdehyde levels were similar between the
two sessions of wood smoke exposure. However,
serum Clara cell protein was significantly
increased with wood smoke in session 1 (higher
particle count) but not in session 2.
Reference: Bastain et al.
(2003, 0986901
Subjects: 18 nonsmoking
adults with positive allergy
skin test to short ragweed
Gender: 7 M/11 F
Age: 18-38 yr
DEP
Isuzu diesel engine, 4 cylinder,
4JB1
Concentration: 0.3 mg in
200 pi saline
Subjects underwent nasal provocation challenge
(intranasal spray) with allergen and either DEP or
placebo (saline) in a randomized crossover study
design. Challenges were separated by 30 days.
This protocol was then repeated 30 days after
the last exposure.
Time to analysis: 24 h post-exposure and 4 and
8 days after exposure.
DEP significantly increased allergic responses to
short ragweed. Relative to allergen + placebo,
allergen + DEP increased allergen specific IgE
4days following exposure, and increased IL-4 1
day post-exposure. The enhancement of allergic
response with DEP was observed to be
reproducible within subjects.
Reference: Beckett et al.
(2005, 1562611
Subjects: 12 healthy adults
Gender: 6 M/6 F
Age: 23-52 yr
Ultrafine and fine zinc oxide
Particle Size: UF:
O.1 pm; Fine: 0.1-1.0 urn
Subjects exposed via mouthpiece for 2 h during
rest to filtered air, ultrafine, and fine zinc oxide in
a randomized crossover study design.
Exposures were separated by at least 3 wk.
No changes observed in neutrophil count in
induced sputum. No PM (zinc oxide)-induced
changes in respiratory symptoms observed
0-24 h post-exposure.
Particle Number/Count: UF: 71™ to analysis: 11 and 24 h after exposure.
4.6 x 107cm .Fine: 1.9 x
4.6xl07<
105/cm3
Concentration: 500 pg/m3
Reference: Behndig et al.
(2006, 0882861
Subjects: 15 healthy adults
Gender: 8 M/7 F
Age: 21-27 yr
DE
Idling 1991 Volvo diesel
engine (TD45, 4.5 L, 4
cylinders, 680 rpm)
Particle Size: PM10; majority
ofPM mass made up of
particles < 1 pm in diameter
Concentration: 100 pg/m3
Exposures conducted for 2 h with intermittent
exercise to both DE and filtered air in a
randomized crossover design. Exposures were
separated by at least 3 wk. Other diesel
emissions measured: NOX (1.8 ppm), N02 (0.4
ppm), NO (1.3 ppm), CO (10.4 ppm), total
hydrocarbons (1.3 ppm).
Time to analysis: 18 h post-exposure.
Exposure to DE increased neutrophil and mast
cell numbers in bronchial mucosa at 18 h
post-exposure. Neutrophils, IL-8, and
myeloperoxidase observed to increase in
bronchial lavage fluid following exposure relative
to filtered air. No inflammatory response observed
in the alveolar compartment. Exposure to DE
increased urate and reduced glutathione
bronchoalveolar lavage at 18 h post-exposure.
Reference: Blomberg et al.
(2005, 1919911
Subjects: 15 older adults
(former smokers) with COPD
Age: 56-72 yr
DE
Concentration: 300 pg/m3
Subjects exposed for 1 h with intermittent
exercise to DE and filtered air in a randomized
crossover study design.
Time to analysis: 6 and 24 h post-exposure.
DE was not observed to affect levels of Clara cell
protein in peripheral blood at 6 and 24 h post-
exposure.
December 2009
C-9
-------
Reference
Pollutant
Exposure
Findings
Reference: Bosson et al.
(2007, 1562861
Subjects: 16 healthy adults
Gender: 7 M/9 F
Age: 20-28 yr
DE
Idling Volvo diesel engine
Concentration: PM
300 pg/m followed by
exposure to filtered air or 0.2
ppm03
Subjects exposed to DE for 1 h followed 5 h later
by a 2-h exposure to either filtered air or 03 (0.2
ppm) using a randomized crossover study
design. All exposures were conducted with
subjects engaged in intermittent exercise.
Time to analysis: 18 h after second exposure
(filtered air or 03).
The percentage of neutrophils and concentration
of myeloperoxidase in induced sputum (18 h
post-Os/air exposure) was significantly higher
following diesel + 03 than diesel + air.
Reference: Bosson et al.
(2008, 1966591
Subjects: 14 healthy adults
Gender: 9 M/5 F
Age: 21-29 yr
DE
Idling 1991 Volvo diesel
engine (TD45, 4.5 L, 4
cylinders)
Concentration: PM
300 pg/m or filtered air
followed by exposure to 0.2
ppm03
Subjects exposed to DE or filtered air for 1 h
followed 5 h later by a 2-h exposure to 03 (0.2
ppm) using a randomized crossover study
design. All exposures were conducted with
subjects engaged in intermittent exercise. Other
diesel emissions measured: N02 (0.51 ppm), NO
(1.65 ppm), total hydrocarbons (1.18 ppm).
Time to analysis: 24 h after the start of the initial
exposure.
Neutrophil and macrophage numbers in bronchial
wash were significantly increased 16 h following
03 exposure when preceded by exposure to
diesel, compared to 03 exposure preceded by
exposure to filtered air.
Reference: Brauner et al.
(2009, 1902441
Subjects: 29 healthy adults
Gender: 20 M, 9 F
Age: M avg 27 yr, F avg 26 yr
Urban traffic particles
Particle Size: PM2.5, PM10.2.5
Particle Number/Count: 6-
700nm:10,067/cm3
Concentration: PM25:
9.7 pg/m3, PMio.25:12.6 pg/m3
Subjects exposed to urban traffic particles and
filtered air for 24 h with and without two 90-min
periods of light exercise in a randomized
crossover study design. Concentrations of NOX
and NO were low and did not differ between
filtered and unfiltered exposures. CO
concentrations were higher with filtered air (0.35
and 0.41 ppm), while 03 concentrations were
lower with filtered air (12.08 and 4.29 ppb).
Time to analysis: 2.5, 6, and 24 h after the start
of exposure.
Epithelial membrane integrity and blood-gas
barrier permeability, assessed using pulmonary
clearance of 99mTc-labeled diethylenetriamine
pentaacetic acid (DTPA), was observed to
increase with exercise, but was not affected by
exposure to urban particles (2.5 h of exposure).
Exposure to urban particles was not shown to
affect plasma or urine concentration of Clara cell
16 protein at 6 and 24 h after the start of
exposure. No relationship between exposure and
pulmonary function was observed at 2.5 h.
Reference: Gilliland et al.
(2004, 1564711
Subjects: 19 adults with
allergic rhinitis and positive
skin test to ragweed, GSTM1
(14 null, 5 present); GSTT1 (9
null, 10 present); GSTP1
codon 105 variants (13 I/I, 6
IA/, OVA/)
Gender:7M/12F
Age: 20-34 yr
DEP
Isuzu diesel engine, 4 cylinder,
4JB1
Concentration: 0.3 mg DEP
in 300 pL saline
Subjects were challenged intranasally with
allergen and placebo (saline) as well as allergen
plus DEP in saline in a randomized crossover
design. Challenges were separated by at least 6
wk.
Time to analysis: 10 min, 24 h, and 72 h post-
challenge.
Subjects who were GSTM1 null or homozygous
for GSTP1 1105 wild-type allele experienced
significantly greater increase in nasal IgE and
histamine with diesel plus allergen compared to
subjects with functional GSTM1 or who were
heterozygous for GSTP1 IA/(105).
Reference: Gong et al.
(2004, 0879641
Subjects: 13 older adults with
COPD, 6 healthy older adults
Gender: COPD: 5 M/8 F,
Healthy: 2 M/4 F
Age: COPD: avg 68 yr,
Healthy: avg 73 yr
Fine CAPs (Los Angeles)
Particle Size: 85% of mass
between 0.1 and 2.5pm
Concentration: Mean:
194 ug/m , Range:
135-229 ug/m3
Exposures to CAPs and filtered air (randomized
crossover) for 2 h with intermittent light exercise
(four15-min periods). Exposures were separated
by at least 2 wk.
Time to analysis: Immediately following
exposure as well as 4 and 22 h post-exposure.
No CAPs-induced respiratory symptoms
observed in healthy older adults or older adults
with COPD at 0, 4, or 22 h post-exposure.
Exposure to CAPs did not significantly affect FVC
or FEVi. CAPs exposure caused a decrease in
arterial oxygen saturation immediately following
exposure which was more pronounced in healthy
older adults than in older adults with COPD.
Exposure to CAPs was not observed to affect the
levels of white blood cells, columnar epithelial
cells, IL-6, or IL-8 in induced sputum.
Reference: Gong et al.
(2004, 0556281
Subjects: 12 adult
asthmatics, 4 healthy adults
Gender: Asthmatic: 4 M/8 F,
Healthy: 2 M/2 F
Age: Asthmatic: avg 38 yr,
Healthy: avg 32 yr
Coarse CAPs (Los Angeles)
Particle Size: 80% of mass
between 2.5 and 10 pm, 20%
of mass<2.5|jm
Concentration: Mean:
157 ug/m3; Range:
56-218 ug/m3
Exposures to CAPs and filtered air (randomized
crossover) for 2 h with intermittent light exercise
(four15-min periods). Exposures were separated
by at least 2 wk.
Time to analysis: Immediately following
exposure as well as 4 and 22 h post-exposure.
No effect of CAPs exposure on spirometry or
arterial oxygen saturation was observed 0, 4, or
22 h post-exposure. No respiratory symptoms
reported 0-22 h post-exposure in either healthy or
asthmatic adults. Sputum cell counts at 22 h post-
exposure did not differ between CAPs and filtered
air.
December 2009
C-10
-------
Reference
Pollutant
Exposure
Findings
Reference: Gong et al.
(2005, 0879211
Subjects: 18 older adults with
COPD, 6 healthy older adults
Gender: COPD: 9 M/9 F,
Healthy: 2 M/4 F
Age: COPD: avg 72 yr,
Healthy: avg 68 yr
Fine CAPs (Los Angeles)
Concentration: CAPs:
200 pg/m3; N02: 0.4 ppm
Each subject was exposed to CAPs, N02, CAPs
+ N02, and filtered air for 2 h with intermittent
exercise. Exposure order was not fully
counterbalanced as N02 exposures were
conducted after the majority of the CAPs and
filtered air exposures had been completed.
Exposures were separated by at least 2 wk.
Time to analysis: Immediately following
exposure as well as 4 and 22 h post-exposure.
Exposure to CAPs was observed to decrease
maximal mid-expiratory flow and arterial oxygen
saturation relative to filtered air 4-22 h
post-exposure. This response was more
pronounced in healthy older adults than in older
adults with COPD. Concomitant exposure to N02
did not enhance the response. No other
respiratory responses (symptoms, spirometry,
sputum cell counts) were affected by exposure to
CAPs.
Reference: Gong et al.
(2008, 1564831
Subjects: 14 adult
asthmatics, 17 healthy adults
Gender: Asthmatics: 9 M/5 F,
Healthy:5M/12F
Age: Asthmatics: 34 + 12yr,
Healthy:24 ± 8 yr
Ultrafine CAPs (Los Angeles)
Particle Number/Count:
145,000/cm3, Range 39,000-
312,000/cm3
Concentration: Mean:
100 ug/m3, Range:
13-277 ug/m3
Subjects exposed for 2 h during intermittent
exercise (15-min periods) to both CAPs and
filtered air in random order. The first 7 subjects
underwent whole body exposure, while the
remaining subjects were exposed through a
facemask. Facemask exposures had higher
particle counts but lower particle mass than
whole body exposures. Exposures were
separated by at least 2 wk.
Time to analysis: Immediately following
exposure as well as 4 and 22 h post-exposure.
No significant differences in respiratory symptoms
observed between filtered air and ultrafine CAPs
exposures. Individuals exposed to higher particle
counts tended to experience greater symptoms
with CAPs than with filtered air. An ultrafine
CAPs-induced decrease in arterial oxygen
saturation (0.5%) was observed at 0, 4, and 22 h
post-exposure. A decrease in FEVi (2%) was also
observed 22 h post-exposure relative to filtered
air. Responses were not significantly different
between healthy and asthmatic adults. CAPs
exposure was not observed to affect total sputum
cell counts or cytokine levels. There were no
differences in response observed between
facemask and whole body exposures.
Reference: Graff et al. (2009, Coarse CAPs (Chapel Hill,
1919811 NC)
Subjects: 14 healthy adults Concentration: 89
± 49.5 pg/m3 (estimated
Gender: 8 M/6 F jnnaied dose = 67% of
Age: 20-34 yr measured particle mass)
Subjects exposed for 2 h with intermittent
exercise (15-min periods) to coarse CAPs and
filtered air in a randomized crossover design.
Exposures were separated by at least 2 mos.
Time to analysis: 0-1 and 20 h post-exposure.
Pulmonary function (FVC, FEVi, and carbon
monoxide diffusing capacity) was not affected by
exposure to coarse CAPs either immediately
following exposure or 20 h post-exposure. A
significant increase in percent PMNs (10.7%
increase per 10 pg/m3 coarse CAPs) was
observed in BAL fluid 20 h post-exposure.
Percent monocytes in BL fluid were slightly
decreased at 20 h post-exposure (2.0% decrease
per 10 [jg/m CAPs; p = 0.05). Total protein in
BAL fluid was also observed to decrease
following CAPs exposure (1.8% decrease per
10 pg/m CAPs). Markers of inflammation in BAL
and BL fluids including IL-6, IL-8, and PGE2 were
not affected by exposure to coarse CAPs.
Reference: Huang et al.
(2003, 0873771
Subjects: 38 healthy adults
Gender: 36 M/2 F
Age: Avg 26.2 + 0.7 yr
Fine CAPs (Chapel Hill, NC)
Concentration:
23.1-311.1 pg/m3
Subjects exposed to CAPs (n = 30) or filtered air
(n = 8) for 2 h with intermittent exercise (subjects
did not serve as their own controls). Component
data of CAPs was available for 37 of the 38
subjects.
Time to analysis: 18 h after exposure.
The increase in bronchoalveolar lavage fluid
neutrophils observed by Ghio et al. (2000,
0121401 following exposure to fine CAPs was
shown to be associated with iron, selenium, and
sulfate content of the CAPs.
Reference: Kongerud et al.
(2006, 1566561
Subjects: 17 asthmatic
adults, 46 healthy adults
Gender: Asthmatics-
6M/11 F, Healthy-24 M/22 F
Age: Asthmatics: avg 23 yr,
Healthy: avg 26 yr
DEP
N 1ST 1650, heavy duty diesel
engine
Concentration: Untreated
and treated with 0.1 ppm 03
(48 h); 300 pg per nostril
DEP (with and without 03 pre-treatment) were
intranasally instilled, using the saline vehicle as
control. Subjects did not serve as their own
controls (not a crossover design).
Time to analysis: 4 and 96 h post-instillation.
Exposure to DEP was not observed to alter
markers of inflammation in nasal lavage fluid
(e.g., cell counts, IL-8, IL-6) at 4 or 96 h
post-instillation.
Reference: Larsson et al.
(2007, 0913751
Subjects: 16 healthy adults
Gender: 10 M/6 F
Age: 19-59 yr
Traffic particles (road tunnel)
Particle Size: PM2 5, PM,0;
PM25 mass constituted -36%
ofPM10mass
Particle Number/Count: 20-
1,000 nm: 1.1 xioW, <
100 nm: 0.85 xl05/cm3
Concentration: PM25-
46-81 pg/m3; PM10-
130-206 pg/m3
Exposures were conducted for 2 h with
intermittent exercise in a room adjacent to a busy
road tunnel. Study used a randomized crossover
design with each subject also exposed to normal
air (control). Exposures were separated by
3-10wks. No exposures to filtered air were
conducted. Other traffic emissions measured:
NO (874 pg/m3), N02 (230 pg/m3), CO (5.8 pg/m3
reported, likely 5.8 mg/m3).
Time to analysis: 14 h post-exposure.
An increase in bronchoalveolar lavage fluid cell
number, lymphocytes, and alveolar macrophages
were observed 14 h after road tunnel exposure
relative to control. Traffic particulate exposure
was not shown to effect cytokine or adhesion
molecule expression in bronchial tissues.
Respiratory symptoms were reported to increase
during exposure to road tunnel air relative to
pre-exposure symptom ratings. Exposure to road
tunnel air was not shown to affect lung function.
December 2009
C-11
-------
Reference
Pollutant
Exposure
Findings
Reference: Mudway et al.
(2004,180208)
Subjects: 25 healthy adults
Gender: 16 M/9 F
Age: 19-42 yr
DE
Idling 1991 Volvo diesel
engine (TD45, 4.5 L, 4
cylinders, 680 rpm)
Concentration: PM10
100 pg/m3
Subjects exposed to DE and filtered air for 2 h
with intermittent exercise (15-min periods) in a
randomized crossover design. Exposures were
separated by at least 3 wk. Other diesel
emissions measured: N02 (0.2 ppm), NO (0.6
ppm), CO (1.7 ppm), total hydrocarbons (1.4
ppm), formaldehyde (43.5 pg/m ).
Time to analysis: 1 h after the start of exposure,
immediately following exposure, and 6 h post-
exposure.
DE caused mild throat irritation in some subjects
and a significant increase in airways resistance
(Raw) during or immediately following exposure.
No changes in FEVi or FVC were observed
following exposure to diesel. Neutrophil numbers
in the bronchial airways tended to increase
following exposure to DE; however, this increase
was highly variable between subjects and did not
reach statistical significance. Exposure to DE did
not affect levels of SOD or malondialdehyde in
the airways. An increase in levels of ascorbate
and GSH in nasal lavage fluid was observed 6 h
following exposure to DE.
Reference: Pietropaoli et al.
(2004, 1560251
Subjects: 16 asthmatic
adults, 40 healthy adults
Gender: Asthmatic: 8 M/8 F,
Healthy: 20 M/20 F
Age: 18-40 yr
Ultrafine EC
Particle Size: CMD -25 nm
Particle Number/Count:
10|jg/m3:~2.0xl06/cm3;
25|jg/m3:~7.0xl06/cm3;
50|jg/m3:~10.8xl06/cm3
Concentration: 10, 25, and
50 pg/m3
Study conducted using a randomized crossover
design with 2-h exposures. Asthmatics (n = 16)
exposed to filtered air and 10 pg/m3.12 healthy
adults exposed to filtered air and 10 pg/m3 at
rest; 12 healthy adults exposed to filtered air, 10
and 25 pg/m3 with intermittent exercise; 16
healthy adults exposed to filtered air and
50 pg/m with intermittent exercise. Exposures
were conducted via mouthpiece.
Time to analysis: Immediately following
exposure as well as 3.5, 21, and 45 h post-
exposure.
No PM-induced changes in eNO or cell counts,
IL-6, or IL-8 in induced sputum were observed in
any of the protocols 21 h following exposure.
Ultrafine carbon was not observed to increase
respiratory symptoms in any of the study
protocols. Healthy adults experienced an Ultrafine
PM-induced reduction in maximal mid-expiratory
flow and CO diffusing capacity relative to filtered
air 21 h following exposure.
Reference: Pourazar et al.
(2005, 0883051
Subjects: 15 healthy adults
Gender: 11 M/4 F
Age: 21-28 yr
DE
Idling Volvo diesel engine
Particle Number/Count: 4.3
x106/cm3
Concentration: PM10
300 pg/m3
Subjects exposed to DE and filtered air for 1 h
with intermittent exercise (randomized crossover
study design). Other diesel emissions measured:
N02 (1.6 ppm), NO (4.5 ppm), CO (7.5 ppm),
total hydrocarbons (4.3 ppm), formaldehyde
(0.26 mg/m3).
Time to analysis: 6 h post-exposure.
Exposure to DE significantly increased nuclear
translocation of NF-KB, AP-1, phosphorylated
p38, and phosphorylated JNK in bronchial
epithelium 6 h post-exposure.
Reference: Pourazar et al.
(2008, 1568841
Subjects: 15 healthy adults
Gender: 11 M/4 F
Age: 21-28 yr
DE
Idling Volvo diesel engine
Particle Number/Count: 4.3
x106/cm3
Concentration: PM10
300 pg/m3
Subjects exposed to DE and filtered air for 1 h
with intermittent exercise (randomized crossover
study design). Other diesel emissions measured:
N02 (1.6 ppm), NO (4.5 ppm), CO (7.5 ppm),
total hydrocarbons (4.3 ppm), formaldehyde
(0.26 mg/m3).
Time to analysis: 6 h post-exposure.
Exposure to DE observed to enhance epidermal
growth factor receptor (EGFR) expression in
bronchial epithelium 6 h post-exposure.
Reference: Riechelmann
et al. (2004, 1801201
Subjects: 30 healthy adults
Gender: 11 M/19 F
Age: 22-32 yr
Urban dust
NISTSRM1649a
Concentration: 150,
500 pg/m3
Subjects exposed to both concentrations of
urban dust (nose only exposure system) as well
as filtered air for 3h at rest in a randomized
crossover design. Exposures were separated by
at least 1 wk.
Time to analysis: 30 min, 8 h, and 24 h post-
exposure.
An increase in nasal secretion (nasal cytology) of
IL-6 and IL-8 were observed 24 h after exposure
to 500 pg/m3 urban dust.
Reference: Samet et al.
(2007, 1569401
Subjects: Ultrafine CAPs: 20
healthy adults, Coarse CAPs:
14 healthy adults
Gender: Ultrafine CAPs:
11 M/9 F, Coarse CAPs:
8M/6F
Age: 18-35 yr
CAPs (Chapel Hill, NC)
Particle Size: Ultrafine:
0.049 + 0.009 pm, Coarse:
3.59 +0.58pm
Concentration: Ultrafine:
47.0 + 20.2 pg/m3, Coarse:
89.0 ± 49.5 pg/m3
Preliminary report comparing effects of controlled
exposures to coarse, fine, and Ultrafine CAPs
among healthy adults (3 separate studies). A
randomized crossover design was used in
evaluating effects of coarse CAPs (n=14) and
Ultrafine CAPs (n=20) relative to filtered air
following of 2-h exposures with intermittent
exercise. Results compared with previous study
of controlled exposure to fine CAPs (Chapel Hill,
NC) where subjects did not serve as their own
controls (Ghio et al., 2000, 0121401
Time to analysis: 0-20 h post-exposure.
As was shown with fine CAPs, exposure to
coarse CAPs increased the percentage of
neutrophils in bronchoalveolar lavage fluid 20 h
following exposure. Unlike fine CAPs, coarse
CAPs did not increase the percent of monocytes
in bronchoalveolar lavage fluid. Ultrafine CAPs
were not shown to affect any markers of
pulmonary inflammation in bronchoalveolar
lavage fluid 18 h after exposure. No
CAPs-induced changes in lung function were
observed.
Reference: Samet et al.
(2009,1919131
Subjects: 19 healthy adults
Gender: 10 M/9 F
Age: 18-35 yr
Ultrafine CAPs (Chapel Hill,
NC)
Particle Size: < 0.16pm
Particle Number/Count:
120,662 ± 48,232
particles/cm3
Concentration: 49.8 +
20 pg/m3
Subjects exposed for 2 h with intermittent 15
periods of exercise to UF CAPs and filtered air
using a randomized crossover study design.
Time to analysis: Immediately following
exposure and 1 and 18 h post-exposure.
No effect of UF CAPs observed on pulmonary
function immediately following exposure or 18 h
post-exposure. IL-8 in BAL fluid was observed to
increase with UF CAPs 18 h post-exposure. UF
CAPs was not shown to alter any other
inflammatory markers in BAL fluid.
December 2009
C-12
-------
Reference
Pollutant
Exposure
Findings
Reference: Schaumann et al.
(2004, 0879661
Subjects: 12 healthy adults
Gender: 4 M/8 F
Age: Avg27±2.5yr
FinePM
Collected (filter) from
industrialized and
non-industrialized areas in
Germany
Concentration: 100 |jg per
subject
Bronchoscopic instillation of particles collected
from both areas was conducted in contralateral
lung segments for each subject.
Time to analysis: 24 h post-instillation.
Particles collected from the industrialized area
(transition metal-rich) increased the percentage of
monocytes and oxidant radical generation in
bronchoalveolar lavage fluid 24 h after exposure
compared with PM containing less metal.
Reference: Stenfors et al.
(2004, 1570091
Subjects: 15 asthmatic
adults, 25 healthy adults
Gender: Asthmatic: 10 M/5 F,
Healthy:! 6 M/9F
Age: Asthmatic: 22-52 yr,
Healthy:! 9-42 yr
DE
Volvo diesel engine
Concentration: PM1t
108 pg/m3
Subjects were exposed for 2 h with intermittent
exercise to DE and filtered air using a
randomized crossover study design. Other diesel
emissions measured: N02 (0.7 ppm).
Time to analysis: 1 h after the start of exposure,
immediately following exposure, and 6 h post-
exposure.
DE was observed to increase neutrophilia and
IL-8 in bronchial lavage fluid among healthy
subjects 6 h after exposure. Among asthmatic
subjects, exposure to DE did not cause an
increase in inflammatory markers. No diesel-
induced change in pulmonary function was
observed during or immediately following
exposure.
Reference: Tunnicliffe et al.
(2003, 0887441
Subjects: 12 asthmatic
adults, 12 healthy adults
Gender: Asthmatics: 7 M/5 F,
Healthy: 5 M/7 F
Age: Asthmatics: avg 35.7 yr,
Healthy: avg 34.5 yr
Aerosols of ammonium
bisulfate and sulfuric acid
Particle Size: HMD 0.3 pm
Concentration: 200,
2,000 pg/m3
Subjects were exposed for 1 h at rest to
ammonium bisulfate (low and high
concentrations), sulfuric acid (low and high
concentrations) and filtered air in a randomized
crossover design. Exposures were separated by
at least 2 wk and were conducted using a head
dome exposure system.
Time to analysis: Immediately following
exposure as well as 5.5-6 h post-exposure.
Neither ammonium bisulfate nor aerosolized
sulfuric acid were observed to affect lung function
or respiratory systems following exposures to 200
or 2,000 pg/m3 among healthy or asthmatic
adults. Exposures to ammonium bisulfate at both
concentrations resulted in a significant increase in
eNO in the asthmatic subjects.
Table C- 3. Central nervous system effects.
Reference
Pollutant
Exposure
Findings
Reference: Cruts et al.
(2008,156374)
Subjects: 10 healthy
adults
Gender: M
Age: 18-39yr
DE
Idling Volvo diesel engine
(TD45,4.5 L, 4 cylinders, 680
rpm)
Particle Number/Count: 1.2
x106/cm3
Concentration: 300 ug/nr
Subjects were exposed to DE and filtered air for
1 h at rest in a randomized crossover study
design. Exposures were separated by 2-4 days.
Other diesel emissions measured: N02 (1.6 ppm),
NO (4.5 ppm), CO (7.5 ppm), total hydrocarbons
(4.3 ppm).
Time to analysis: From onset of exposure until
1 h post-exposure.
Exposure to DE was observed to significantly
increase the median power frequency (MPF) in
the frontal cortex during exposure, as well as in
the hour following the completion of the
exposure.
December 2009
C-13
-------
Annex C References
Alexis NE; Lay JC; Zeman K; Bennett WE; Peden DB; Soukup JM; Devlin RB; Becker S. (2006). Biological material on
inhaled coarse fraction particulate matter activates airway phagocytes in vivo in healthy volunteers. J Allergy Clin
Immunol, 117: 1396-1403. 154323
Barregard L; Sallsten G; Andersson L; Almstrand AC; Gustafson P; Andersson M; Olin AC. (2008). Experimental exposure
to wood smoke: effects on airway inflammation and oxidative stress. Occup Environ Med, 65: 319-324. 155675
Barregard L; Sallsten G; Gustafson P; Andersson L; Johansson L; Basu S; Stigendal L. (2006). Experimental exposure to
wood-smoke particles in health humans: effects on markers of inflammation, coagulation, and lipid peroxidation.
Inhal Toxicol, 18: 845-853. 091381
BastainTM; Gilliland FD; Li Y-F; Saxon A; Diaz-Sanchez D. (2003). Intraindividual reproducibility of nasal allergic
responses to diesel exhaust particles indicates a susceptible phenotype. Clin Immunol, 109: 130-136. 098690
Beckett WS; Chalupa DF; Pauly-Brown A; Speers DM; Stewart JC; Frampton MW; Utell MJ; Huang LS; Cox C; Zareba
W; Oberdorster G. (2005). Comparing inhaled ultrafine versus fine zinc oxide particles in healthy adults - A human
inhalation study. Am J Respir Grit Care Med, 171: 1129-1135. 156261
Behndig AF; Mudway IS; Brown JL; Stenfors N Helleday R Duggan ST; Wilson SJ; Boman C Cassee FR; Frew AJ; Kelly
FJ; Sandstrom T Blomberg A. (2006). Airway antioxidant and inflammatory responses to diesel exhaust exposure in
healthy humans. Eur Respir J, 27: 359-365. 088286
Blomberg A; Tornqvist H; Desmyter L; Deneys V; Hermans C. (2005). Exposure to diesel exhaust nanoparticles does not
induce blood hypercoagulability in an at-risk population. J Thromb Haemost, 3: 2103-2105. 191991
Bosson J; Barath S; Pourazar J; Behndig AF; Sandstrom T; Blomberg A; Adelroth E. (2008). Diesel exhaust exposure
enhances the ozone-induced airway inflammation in healthy humans. Eur Respir J, 31: 1234-40. 196659
Bosson J; Pourazar J; Forsberg B; Adelroth E; Sandstrom T; Blomberg A. (2007). Ozone enhances the airway inflammation
initiated by diesel exhaust. Respir Med, 101: 1140-1146. 156286
Brauner EV; Forchhammer L; Moller P; Barregard L; Gunnarsen L; Afshari A; Wahlin P; Glasius M; Dragsted LO; Basu S;
Raaschou-Nielsen O; Loft S. (2008). Indoor particles affect vascular function in the aged: an air filtration-based
intervention study. Am J Respir Grit Care Med, 177: 419-425. 156293
Brauner EV; Forchhammer L; Moller P; Simonsen J; Glasius M; Wahlin P; Raaschou-Nielsen O; Loft S. (2007). Exposure
to ultrafine particles from ambient air and oxidative stress-induced DNA damage. Environ Health Perspect, 115:
1177-82.091152
Brauner EV; Mortensen J; Moller P; Bernard A; Vinzents P; Wahlin P; Glasius M; Loft S. (2009). Effects of ambient air
particulate exposure on blood-gas barrier permeability and lung function. Inhal Toxicol, 21: 38-47. 190244
Brauner EV; M011er P; Barregard L; Dragsted LO; Glasius M; Wahlin P; Vinzents P; Raaschou-Nielsen O; Loft S. (2008).
Exposure to ambient concentrations of particulate air pollution does not influence vascular function or
inflammatory pathways in young healthy individuals. Part Fibre Toxicol, 5:13. 191966
Brook RD; Brook JR; Urch B; Vincent R; Rajagopalan S; Silverman F. (2002). Inhalation of fine particulate air pollution
and ozone causes acute arterial vasoconstriction in healthy adults. Circulation, 105: 1534-1536. 024987
Carlsten C; Kaufman JD; Trenga CA; Allen J; Peretz A; Sullivan JH. (2008). Thrombotic markers in metabolic syndrome
subjects exposed to diesel exhaust. Inhal Toxicol, 20: 917-921. 156323
Carlsten C; Kaufman Joel D; Peretz A; Trenga Carol A; Sheppard L; Sullivan Jeffrey H. (2007). Coagulation markers in
healthy human subjects exposed to diesel exhaust. Thromb Res Suppl, 120: 849-855. 155714
Cruts B; van Etten L; Tornqvist H; Blomberg A; Sandstrom T; Mills NL; Borm PJ. (2008). Exposure to diesel exhaust
induces changes in EEG in human volunteers. Part Fibre Toxicol, 5: 4. 156374
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009 C-14
-------
Danielsen PH; Brauner EV; Barregard L; Sallsten G; Wallin M; Olinski R; Rozalski R; Moller P; Loft S. (2008).
Oxidatively damaged DNA and its repair after experimental exposure to wood smoke in healthy humans. Mutat Res
Fund Mol Mech Mutagen, 642: 37-42. 156382
Devlin RB; Ohio AJ; Kehrl H; Sanders G; Cascio W. (2003). Elderly humans exposed to concentrated air pollution
particles have decreased heart rate variability. Eur Respir J, 40: 76S-80S. 087348
Fakhri AA; Ilic LM; Wellenius GA; Urch B; Silverman F; Gold DR; Mittleman MA. (2009). Autonomic effects of
controlled fine particulate exposure in young healthy adults: Effect modification by ozone. Environ Health
Perspect, 117: 1287-1292. 191914
Frampton MW; Stewart JC; Oberdorster G; Morrow PE; Chalupa D; Pietropaoli AP; Frasier LM; Speers DM; Cox C;
Huang LS; Utell MJ. (2006). Inhalation of ultrafine particles alters blood leukocyte expression of adhesion
molecules in humans. Environ Health Perspect, 114: 51-58. 088665
Ghio AJ; Kim C; Devlin RB. (2000). Concentrated ambient air particles induce mild pulmonary inflammation in healthy
human volunteers. Am J Respir Grit Care Med, 162: 981-988. 012140
Gilliland FD; Li YF; Saxon A; Diaz-Sanchez D. (2004). Effect of glutathione-S-transferase Ml and PI genotypes on
xenobiotic enhancement of allergic responses: randomised, placebo-controlled crossover study. Lancet, 363: 119-
125. 156471
Gong H Jr; Linn WS; Clark KW; Anderson KR; Geller MD; Sioutas C. (2005). Respiratory responses to exposures with
fine particulates and nitrogen dioxide in the elderly with and without COPD. Inhal Toxicol, 17: 123-132. 087921
Gong H Jr; Linn WS; Clark KW; Anderson KR; Sioutas C; Alexis NE; Cascio WE; Devlin RB. (2008). Exposures of
healthy and asthmatic volunteers to concentrated ambient ultrafine particles in Los Angeles. Inhal Toxicol, 20: 533-
545. 156483
Gong H Jr; Linn WS; Terrell SL; Anderson KR; Clark KW; Sioutas C; Cascio WE; Alexis N; Devlin RB. (2004).
Exposures of elderly volunteers with and without chronic obstructive pulmonary disease (COPD) to concentrated
ambient fine particulate pollution. Inhal Toxicol, 16: 731-744. 087964
Gong H Jr; Linn WS; Terrell SL; Clark KW; Geller MD; Anderson KR; Cascio WE; Sioutas C. (2004). Altered heart-rate
variability in asthmatic and healthy volunteers exposed to concentrated ambient coarse particles. Inhal Toxicol, 16:
335-343. 055628
Graff D; Cascio W; Rappold A; Zhou H; Huang Y; Devlin R. (2009). Exposure to concentrated coarse air pollution
particles causes mild cardiopulmonary effects in healthy young adults . Environ Health Perspect, 117: 1089-1094.
191981
Huang Y-CT; Ghio AJ; Stonehuerner J; McGee J; Carter JD; Grambow SC; Devlin RB. (2003). The role of soluble
components in ambient fine particles-induced changes in human lungs and blood. Inhal Toxicol, 15: 327-342.
087377
Kongerud J; Madden MC; Hazucha M; Peden D. (2006). Nasal responses in asthmatic and nonasthmatic subjects following
exposure to diesel exhaust particles. Inhal Toxicol, 18: 589-594. 156656
Larsson B-M; Sehistedt M; Grunewald J; Skold CM; Lundin A; Blomberg A; Sandstrom T; Eklund A; Svartengren M.
(2007). Road tunnel air pollution induces bronchoalveolar inflammation in healthy subjects. Eur Respir J, 29: 699-
705.091375
Lucking A; Lundback M; Mills N; Faratian D; Barath S; Pourazar J; Cassee F; Donaldson K; Boon N; Badimon J;
Sandstorm T; Blomberg A; Newby D. (2008). Diesel exhaust inhalation increases thrombus formation in man. Eur
Heart J, 29: 3043-3051. 191993
Lund AK; Lucero J; Lucas S; Madden MC; McDonald JD; Seagrave JC; Knuckles TL; Campen MJ. (2009). Vehicular
emissions induce vascular MMP-9 expression and activity associated with endothelin-1 mediated pathways.
Arterioscler Thromb Vase Biol, 29: 511-517. 180257
Lundback M; Mills NL; Lucking A; Barath S; Donaldson K; Newby DE; Sandstrom T; Blomberg A. (2009). Experimental
exposure to diesel exhaust increases arterial stiffness in man. Part Fibre Toxicol, 6: 7. 191967
December 2009 C-15
-------
Mills NL; Robinson SD; Fokkens PH; Leseman DL; Miller MR; Anderson D; Freney EJ; Heal MR; Donovan RJ;
Blomberg A; Sandstrom T; MacNee W; Boon NA; Donaldson K; Newby DE; Cassee FR. (2008). Exposure to
concentrated ambient particles does not affect vascular function in patients with coronary heart disease. Environ
Health Perspect, 116: 709-715. 156766
Mills NL; Tornqvist H; Gonzalez MC; Vink E; Robinson SD; Soderberg S; Boon NA; Donaldson K; Sandstrom T;
Blomberg A; Newby DE. (2007). Ischemic and thrombotic effects of dilute diesel-exhaust inhalation in men with
coronary heart disease. N Engl J Med, 357: 1075-1082. 091206
Mills NL; Tornqvist H; Robinson SD; Gonzalez M; Darnley K; MacNee W; Boon NA; Donaldson K; Blomberg A;
Sandstrom T; Newby DE. (2005). Diesel exhaust inhalation causes vascular dysfunction and impaired endogenous
fibrinolysis. Circulation, 112: 3930-3936. 095757
Mudway IS; Stenfors N; Duggan ST; Roxborough H; Zielinski H; Marklund SL; Blomberg A; Frew AJ; Sandstrom T;
Kelly F J. (2004). An in vitro and in vivo investigation of the effects of diesel exhaust on human airway lining fluid
antioxidants. ArchBiochem Biophys, 423: 200-212. 180208
Peretz A; Kaufman JD; Trenga CA; Allen J; Carlsten C; Aulet MR; Adar SD; Sullivan JH. (2008). Effects of diesel exhaust
inhalation on heart rate variability in human volunteers. Environ Res, 107: 178-184. 156855
Peretz A; Peck EC; Bammler TK; Beyer RP; Sullivan JH; Trenga CA; Srinouanprachnah S; Farin FM; Kaufman JD.
(2007). Diesel exhaust inhalation and assessment of peripheral blood mononuclear cell gene transcription effects:
an exploratory study of healthy human volunteers. Inhal Toxicol, 19: 1107-1119. 156853
Peretz A; Sullivan JH; Leotta DF; Trenga CA; Sands FN; Allen J; Carlsten C; Wilkinson CW; Gill EA; Kaufman JD.
(2008). Diesel exhaust inhalation elicits acute vasoconstriction in vivo. Environ Health Perspect, 116: 937-942.
156854
Pietropaoli AP; Frampton MW; Hyde RW; Morrow PE; Oberdorster G; Cox C; Speers DM; Frasier LM; Chalupa DC;
Huang LS; Utell MJ. (2004). Pulmonary function, diffusing capacity, and inflammation in healthy and asthmatic
subjects exposed to ultrafine particles. Inhal Toxicol, 16: 59-72. 156025
Pourazar J; Blomberg A; Kelly FJ; Davies DE; Wilson SJ; Holgate ST; Sandstrom T. (2008). Diesel exhaust increases
EGFR and phosphorylated C-terminal Tyr 1173 in the bronchial epithelium. Part Fibre Toxicol, 5: 8. 156884
Pourazar J; Mudway IS; Samet JM; Helleday R; Blomberg A; Wilson SJ; Frew AJ; Kelly FJ; Sandstrom T. (2005). Diesel
exhaust activates redox-sensitive transcription factors and kinases in human airways. Am J Physiol, 289: L724-
L730. 088305
Power K; Balmes J; Solomon C. (2008). Controlled exposure to combined particles and ozone decreases heart rate
variability. J Occup Environ Med, 50: 1253-1260. 191982
Riechelmann H; Rettinger G; Lautebach S; Schmittinger S; Deutschle T. (2004). Short-term exposure to urban dust alters
the mediator release of human nasal mucosa. J Occup Environ Med, 46: 316-322. 180120
Routledge HC; Manney S; Harrison RM; Ayres JG; Townend JN. (2006). Effect of inhaled sulphur dioxide and carbon
particles on heart rate variability and markers of inflammation and coagulation in human subjects. Heart, 92: 220-
227. 088674
Rundell KW; Caviston R. (2008). Ultrafine and fine particulate matter inhalation decreases exercise performance in healthy
subjects. J Strength Cond Res, 22: 2-5. 191986
Samet JM; Graff D; Berntsen J; Ohio AJ; Huang YC; Devlin RB. (2007). A comparison of studies on the effects of
controlled exposure to fine, coarse and ultrafine ambient particulate matter from a single location. Inhal Toxicol, 19
Suppl 1:29-32. 156940
Samet JM; Rappold A; Graff D; Cascio WE; Berntsen JH; Huang YC; Herbst M; Bassett M; Montilla T; Hazucha MJ;
Bromberg PA; Devlin RB. (2009). Concentrated ambient ultrafine particle exposure induces cardiac changes in
young healthy volunteers. Am J Respir Grit Care Med, 179: 1034-1042. 191913
Schaumann F; Borm PJA; Herbrich A; Knoch J; Pitz M; Schins RPF; Luettig B; Hohlfeld JM; Heinrich J; Krug N. (2004).
Metal-rich ambient particles (particulate matter 2.5) cause airway inflammation in healthy subjects. Am J Respir
Grit Care Med, 170: 898-903. 087966
December 2009 C-16
-------
Shah AP; Pietropaoli AP; Frasier LM; Speers DM; Chalupa DC; Delehanty JM; Huang LS; Utell MJ; Frampton MW.
(2008). Effect of inhaled carbon ultrafine particles on reactive hyperemia in healthy human subjects. Environ
Health Perspect, 116: 375-380. 156970
Stenfors N; Nordenhall C; Salvi SS; Mudway I; Soderberg M; Blomberg A; Helleday R; Levin JO; Holgate ST; Kelly FJ;
Frew AJ; Sandstrom T. (2004). Different airway inflammatory responses in asthmatic and healthy humans exposed
to diesel. EurRespir J, 23: 82-86. 157009
Tornqvist H; Mills NL; Gonzalez M; Miller MR; Robinson SD; Megson IL; MacNee W; Donaldson K; Soderberg S;
Newby DE; Sandstrom T; Blomberg A. (2007). Persistent endothelial dysfunction in humans after diesel exhaust
inhalation. Am J Respir Grit Care Med, 176: 395-400. 091279
Tunnicliffe WS; Harrison RM; Kelly FJ; Dunster C; Ayres JG. (2003). The effect of sulphurous air pollutant exposures on
symptoms, lung function, exhaled nitric oxide, and nasal epithelial lining fluid antioxidant concentrations in normal
and asthmatic adults. Occup Environ Med, 60: 1-7. 088744
Urch B; Brook JR; Wasserstein D; Brook RD; Rajagopalan S; Corey P; Silverman F. (2004). Relative contributions of
PM2.5 chemical constituents to acute arterial vasoconstriction in humans. Inhal Toxicol, 16: 345-352. 055629
Urch B; Silverman F; Corey P; Brook JR; Lukic KZ; Rajagopalan S; Brook RD. (2005). Acute blood pressure responses in
healthy adults during controlled air pollution exposures. Environ Health Perspect, 113: 1052-1055. 081080
Zareba W; Couderc JP; Oberdorster G; Chalupa D; Cox C; Huang LS; Peters A; Utell MJ; Frampton MW. (2009). EGG
parameters and exposure to carbon ultrafine particles in young healthy subjects. Inhal Toxicol, 21: 223-233. 190101
December 2009 C-17
-------
Annex D. Toxicological Studies
Table D-1. Cardiovascular effects.
Study
Pollutant
Exposure
Effects
Reference: Anselme et DE: monocylinder Diesel engine using
al. (2007, 0970841 Euro 4 ELF 85A reference gasoline
Route: Whole-body Inhalation
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Age: Adult
Weight: 200-225g
Particle Size: DE: 10-650 nm (85 nm
mean mobility diameter)
; Other
Dose/Concentration: DE: 0.5 mg/m
emissions measured: non-methane
hydrocarbons (7.7 ppm), N02 (1.1 ppm), CO
(4.3 ppm)
Time to Analysis: Experiments started 3 mo
after L coronary artery ligation. ECG started at
tO and the DE exposure at t30 min for a 3-h
period; ventricular premature beats (VPBs) and
RMSSD calculated every 30 min during clean
room air exhaust and PE periods. Early (1210-
300 min) and late ((480-540 min) PE were
analyzed.
Immediate decrease in RMSSD was observed in
both healthy and CHF rats PE. Immediate
increase in VPBs observed in CHF rats only;
which lasted 4-5 h after exposure ceased.
Whereas HRV progressively returned to baseline
values within 2.5 h post-exposure (PE), the
proarrhythmic effect persisted as late as 5 h PE
termination in CHF rats
Reference: Bagate et IPS and EHC-93 (PM): Urban Air
al. (2004, 0556381 collected at the Health Effects Institute
Ottawa, Canada
Species: Rat
Particle Size: EHC-93: 0.8-0.4 pm
Gender: Male (mean) (range: <3 pm)
Strain: SH
Age: 13-15 wk
Route: IT Instillation
Dose/Concentration: PM: 10 mg/kg; IPS- 350
EU/animal
Time to Analysis: Sacrificed 4 or 24 h post-
instillation
PM and IPS elicited a significant increase in
receptor-dependent vasorelaxation of the aorta
compared to saline-instilled rats.
Reference: Bagate et
al. (2004, 0556381
Species: Rat
Gender: Male
Strain: SH
Age: 13-15 wk
EHC-93 (PM), CB-V or CB-Fe, IPS Route: Aortic Suspension Fluid
Particle Size: EHC-93: 0.8-0.4 pm
(mean) (range: <3 pm)
Dose/Concentration: Cumulative
concentrations of EHC-93, CB-V and CB-Fe
(10,25,50,75, 100|jg/mL)
CB 1.5-2.0 nm (mean) (range <5 pm)
Time to Analysis: Immediately post-exposure
of aortic rings to cumulative concentrations of
EHC-93, CB-V, CB-Fe and IPS.
CB-V particles induced more relaxation than CB-
Fe particles or EHC-93 in a dose-dependent
manner. PM and IPS had an acute transient
effect on the receptor dependent vasorelaxation.
PM and IPS attenuated ACh-elicited
vasocontraction in denuded aortic rings (DARs).
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009
D-1
-------
Study
Pollutant
Exposure
Effects
Reference: Bagate et
al. (2004, 0556381
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Age: 13-15 wk
EHC-93 (PM): Urban Air collected at
the Health Effects Institute Ottawa,
Canada.
EHC-93 filtrate (PMF)
Zn2* and Cu2* particles (10,000 and
845 pg PM respectively)
Particle Size: PM: 4.6 pm (GSD = 3.2)
Route: In Vitro
Dose/Concentration: PM Suspensions: 10-100
pg/mL; CuS04/ZnS041-100 pmol; Phe 2 pm;
arbacol: 10pm
Time to Analysis: Measured immediately after
maximum response for each cumulative dose
was achieved.
PM-lnduced Contraction: No effect of
suspension or filtrate seen on resting tension of
aorta and SMRA.
PM- and Metal-Induced Vasorelaxation:
Cumulative concentrations (10-100 pg/mL) of PM
suspension and its water soluble components
(PMF) elicited dose-dependent relaxation in
aorta. Relaxation induced by particle suspension
was higher than relaxation induced by free
filtrate. The difference was significant at 100
pg/mL In SMRA, vasorelaxation similar to aorta's
was observed, and the activity of the particle sus-
pension was stronger than the filtrate, with the
difference being significant starting at 30 pg/mL
Both Zn and Cu in sulfate salts (10-100 pmol)
induced relaxation in pre-contracted aortic rings,
with Cu2* having a greater effect than Zn2* at the
same concentration. Ions didn't affect ACh
relaxation.
Effect of PM on a-Adrenergic Contraction:
Phenylephrine-induced dose-response
contraction, starting at 1pM with max at 100
pmol. Pre-treatment of SMRA did not change the
phenylephrine-induced contraction.
Reference: Bagate et
al. (2006, 0976081
Species: Rat
Gender: Male
Strain: Wistar Kyoto
andSH
Age: 13-15 wk
EHC-93 (PM) Route: In Vitro
EHC-93 (Filtrate) Dose/Concentration: PM and PMF
2t 2t Suspensions: 10-100 pg/mL; CuS04 or
Cu andZn solutions ZnS04:10-100|jmol; Phenylephrine: 2 pm;
Particle Size: PM: 4.6 pm (GSD = 3.2) Carbacol: 10 ^m
Time to Analysis: Responses evaluated at
maximum of each dose-response.
PM and its soluble components elicited
endothelium-independent vasodilation in rat aorta
rings. This response is a result of the activation of
sGC since its inhibition by NS2028 practically
eliminated relaxation. PM suspensions stimulated
cGMP production in purified isolated sGC.
Neither receptor nor their signaling pathways
played a significant role in the direct relaxation by
PM or metals. Vasodilation responses were
significantly higher in SH than WKY control rats.
Reference: Bagate et
al. (2006, 0961571
Species: Rat
Gender: Male
Strain: SH/NHsd
Age: 11-12 wk
Weight: 250-350 g
EHC-93 (PM): Urban Air collected at
the Health Effects Institute Ottawa,
Canada.
EHC-93 (Filtrate),
Zinc (in PM), IPS
Particle Size: PM: 4.6 pm (GSD = 3.2)
Route: IT Instillation
Dose/Concentration: PM: 10 mg/kg; IPS: 350
EU/animal (0.5 ml)
Time to Analysis: 4 h post-exposure
Effect of Pretreatment on Baseline
Parameters of Isolated Perfused Heart: After
PM exposure a slight increase of baseline coro-
nary flow (CF) and heart rate (HR) was noted. In
contrast, a significant decrease of left developing
ventricular pressure (LDVP) was observed in SH.
IPS also elicited a non-significant decrease in
LVDP
Effect of Pretreatment and Ischemia on
Cardiac Function: When SH rats were
pretreated with PM or IPS the isolated heart had
a reduced ability to recover to baseline levels
after occlusion, in comparison with saline treated
rats. After occlusion was released CF went back
to baseline values. Saline and IPS treated rats,
showed a gradual decrease in CF noted during
the reperfusion period. Isolated hearts from PM-
exposed SH showed a complete restoration of
CF and no gradual decrease. The increase of
Zn2+ elicited a rapid decrease of LDVP and HR.
The impairment of cardiac function measured by
LDVP and HR started immediately upon Zn2+
infusion and remained the same during the
perfusion period (no Zn2+ was present in the
perfusate).
Reference: Bagate et
al. (2006, 0961571
Species: Rat
Strain: H9c2 (EACC),
cardiomyocyte cells
EHC-93 (PM) Filtrate: Urban Air col-
lected at the Health Effects Institute
Ottawa, Canada,
ZnS04
Particle Size: PM:4.6|jm
(GSD = 3.2); Carbon Particles: 44 nm
Route: In Vitro
Dose/Concentration: PM: 1, 50,100 pg/mL;
ZnS04: 50 pmol
Time to Analysis: 30 min incubation
Effect of EHC-93 filtrate on Ca2* Uptake in
Cardiomyocytes: Both PMF and Zn * inhibited
ATP or ionophore-stimulated Ca2* influx in
cardiomyocytes.
December 2009
D-2
-------
Study
Pollutant
Exposure
Effects
Reference: Bartoli et
al. (2009, 1562561
Species: Dog
Gender: Female
Strain: Mixed breed
Age: 2-12 yr
Weight:
Average: 15.7 kg,
Range: 13.6-18.2 kg
CAPs (Boston; Harvard Ambient
Particle Concentrator)
Particle Size: Diameter: 0.15-2.5 pm
Route: Permanent Tracheostomy
Dose/Concentration: Concentration range and
mean: CAPs: 94.1-1557(358.1 ± 306.7) pg/m3,
BC: 1.3-32(7.5 + 6.1) pg/m3, Particle count:
3000-69300(18230+13.151) particles/cm3
Time to Analysis: Preanesthetized.
Tracheostomy. 5 h exposures separated by
minimum 1wk. Prazosin administered in 8 of 13
dogs 30-60 min before exposure. 55 exposure
days.
CAPs significantly increased SBP, DBP, mean
arterial pressure, HR and rate-pressure product.
Prazosin (a-adrenergic antagonist) decreased
these CAPs-induced effects. CAPs mass, BC,
particle number concentrations were positively
and significantly associated with each of the
cardiovascular parameters except for pulse
pressure.
Reference: Bartoli et
al. (2009, 1799041
Species: Dog
Gender: Female
Strain: Mixed breed
Age: Adult
Weight: 14-18 kg
CAPs (Boston; Harvard Ambient
Particle Concentration)
Particle Size: Diameter: Ł2.5 pm
Route: Permanent Tracheostomy
Dose/Concentration: Concentration range and
mean: CAPs: 94.1-1556.8 (349 + 282.6) pg/m3,
BC: 1.3-32 (7.5 + 5.6) pg/m3, Particle number:
3000-69300 (20381 + 13075) particles/cm3
Time to Analysis: Tracheostomy. Minimum 3
wk recovery. Acclimatized. Exposed 5 h. 2 5
min occlusions of LAD coronary artery
separated by 20 min rest. Exposure days
separated by 1wk minimum.
During coronary artery occlusion, CAPs exposure
reduced myocardial blood flow and increased
coronary vascular resistance, SBP and DBP
CAPs effects were greater in ischemic tissue
than nonischemic. Increases in CAPs mass,
particle number and BC concentrations were
significantly associated with decreased
myocardial blood flow and increased coronary
vascular resistance.
Reference: Campen et High Whole DE (HVVDE); Low Whole
al. (2005, 0839771 DE (LWDE); High PM Filtered (HPMF);
Low PM Filtered (LPMF)
Species: Mouse
Particle Size: NR
Gender: Male
Strain: C57BL/6J and
Apo E"'"
Age: 10-12 wk
Route: Whole-body Inhalation and Ex-vivo
Exposures (isolated, pressurized septal
coronary arteries)
Dose/Concentration: HWDE: PM = 3.6 mg/m3;
N0x=102ppm
LWDE: PM = 0.512 mg/m3; NOX = 19 ppm;
PM = 0.770 mg/m3; NOX= 105 ppm
LPMF: PM = 0.006 mg/m3; NOX = 26 ppm
Time to Analysis: Whole-body Exposures: DE
or PFDE for 6 h/day for 3 days, euthanized at
the end of last exposure.
Coronary Vessels Exposure: PSS bubbled
with DE to expose coronary vessels to the
soluble contents of DE. Analysis occurred
immediately post exposure.
Whole-body Exposure on ApoE : During DE
exposure, ApoE" mice HR consistently
decreased during high concentration exposures,
compared to the C57BL/6J strain.
Coronary Vascular Effects on ApoE : DE had
no significant effects on the resting myogenic
tone of isolated septal coronary arteries. Control
coronary arteries showed constrictive responses
to ET-1 and dilatory responses to SNP. DE
exposed PSS vessels responses to ET-1
enhanced compared to control. SNP-induced
dilation blunted in vessels resting in diesel-
exposed saline.
Reference: Campen et DE: generated by either of two Route: Whole-body exposure
al.(2003, 0556261 Cummins (2000 model) 5.9-L ISB turbo
engines fueled by Number 2 Diesel Dose/Concentration: 0, 30,100, 300,1000
Certification Fuel. M9/m
Species: Rat
Gender: Male and
Female
Strain: SH
Age: 4 mo
Particle Size: 0.1-0.2 pm aerodynamic Time to Analysis: 6 h/day for 7 days; ECG
diameter measurements taken 4 days post-exposure.
HR: Significantly higher in exposed animals and
not concentration-dependent. More substantial
results seen in male rats.
ECG: The PQ interval was significantly prolonged
among exposed animals in a concentration-
dependent manner.
Reference: Campen et Road dust from paved surfaces (Reno, Route: Whole-body inhalation
al. (2006, 0968791
Species: Mouse
Gender: Male
Strain: ApoE"'"
Age: 10 wk
NV)
Gasoline engine emissions, containing
PM, NOX, COandHC
Particle Size: Road dust: 1.6 pm
(Standard Deviation 2.0)
Gasoline engine emissions: Average
particle diameter of 15 nm
Dose/Concentration: Road dust: 0.5 and 3.5
mg/m3
Gasoline engine emissions: 5 to 60 pg/m3 (at
dilutions of 10:1,15:1, and 90:1)
Mean concentrations of PM: 61 pg/m3; NOX:
18.8 ppm; CO: 80 ppm.
Time to Analysis: 6 h/days for 3 days.
Sacrificed 18 h post-exposure.
ET-1: Gasoline exhaust significantly upregulated
ET-1 in a dose-dependent manner. ET-1 in-
creased levels in the PM filtered group and
decreased in the low levels of road dust.
ECG: HR consistently decreased from beginning
to end of exposure in all groups. No significant
HR effects on road dust or gasoline exposure
was observed. No significant effects on P-wave,
PQ-interval, QRS-interval, or QT-interval were
observed in either treatment.
T-wave: Mice exposed to whole gasoline exhaust
displayed significant increases in T-wave
morphology from the beginning of exposures; this
effect was consistent on all exposure days.
December 2009
D-3
-------
Study
Pollutant
Exposure
Effects
Reference: Cascio et UFPM: Ultra fine PM, EPA Chapel Hill, Route: IT Instillation
al. (1987, 0075831
Species: Mouse
Gender: Male
Strain: ICR
Age:6-10wk
NC
Particle Size: O.1 pm
Dose/Concentration: 100 |jg in 100 pi
Time to Analysis: 24 h post-exposure (single
exposure)
UFPM exposure double the size of myocardial in-
farction attendant to an episode of ischemia and
reperfusion while increasing post ischemic
oxidant stress. UFPM alters endothelium-
dependent/independent regulation of systemic
vascular tone; increases platelet number, plasma
fibrinogen, and soluble P-selectin levels; reduces
bleeding time.
Reference: Chang et UfCB: Ultra fine carbon black Ferric
al. (2007, 1557201
Species: Rat
Gender: Male
Strain: SH
Age: 60 days
sulfate Fe2(S04)3
Nickel sulfate NiS04
Particle Size: UfCB
Route: IT Instillation
Dose/Concentration: UfCB: 415 and 830 pg
Ferric Sulfate: 105 and 210 pg
Nickel Sulfate: 263 and 526 pg
Combined UfCB and ferric sulfate: 830 pg UfCB
+ 105 pg ferric sulfate
Combined UfCB with Nickel Sulfate: 830 pg
UfCB + 263 pg Nickel Sulfate
Time to Analysis: Single dose, radiotelemetry
readings recorded for 72 h post- exposure.
Both high/low-dose UfCB decreased ANN
(normal-to-normal intervals) slightly around the
30th h, concurrent increases of LnSDNN.
LnRMSSD returned to baseline levels after small
initial increases. Minor effects observed after low-
dose Fe and Ni instillation; biphasic changes
occurred after high-dose instillations. Combined
exposures of UfCB and either Fe or Ni resulted in
HRV trends different from values estimated from
individual-component effects.
Reference: Chang et CAPs: collected during a dust storm
al. (2007,1557201 from Chung-Li, Taipei
Species: Rat
Gender: Male
Strain: SH
Age: 10 wk
Particle Size: PM25
Route: Nose-only Inhalation
Dose/Concentration: 315.55 pg/m3
Time to Analysis: 6 h
A linear mixed-effects model revealed sigmoid
increases in HR and a sigmoid decrease of QAI
during exposure, after an initial incubation period.
Reference: Chang et
al. (2004, 0556371
Species: Rat
Gender: Male
Strain: SH
Age: 60 days
CAPs collected in Chung-Li, Taipei
(spring and summer periods)
Particle Size: PM25
Route: Nose-only Inhalation
Dose/Concentration: Spring exposure: 202.0 +
68.8 pg/m ; Mean number concentration: 2.30 *
105 particles/cm3 (range: 7.12 xio3 - 8.26 xio5)
Summer exposure: 141.0 + 54.9 pg/m3; Mean
number concentration: 2.78 xio5 particles/cm3
(range: 7.76 xio3-8.87 xio5)
Time to Analysis: 4 days of spring exposure
and days of summer exposure for 5 h each
exposure. Parameters measured throughout
duration of exposures.
During spring exposures, the maximum increase
of heart rate (HR) and blood pressure (BP) were
51.6 bpm and 8.5 mmHg respectively. The
maximum decrease of QAI (measures cardiac
contractility) noted at the same time was 1.6ms.
Similar pattern was observed during summer
exposure, however, the responses were less
prominent.
Reference: Chang, et CAPs collected in Chung-Li, Taipei
al. (2005, 0977761
v ' Particle Size: PM25 (0.1-2.5 pm)
Species: Rat
Gender: Male
Strain: SH
Weight: 200 g
Route: Nose-only Inhalation
Dose/Concentration: 202.0 + 68.8 pg/m3
Time to Analysis: 5 h/days for 4 days
During the inhalation stage, crude effects of both
LnSDNN and LnRMSSD for exposure and
control groups decreased from the baseline
values. Immediately after the experiments, both
LnSDNN and LNRMSSD decreased due to
stresses produced by release from the exposure
system, then returned to the baseline values.
Reference: Chauhan
et al. (2005, 1557221
SRM-1879(Si02)andSRM-154b
(Ti02) from the NIST
Tumor Cell Line: A549 EHC-93 from Ontario, Canada
derived from alveolar
type II epithelial cells
(EHCsol, EHCinsol)
Particle Size: EHC-93 median
physical diameter: 0.4 pm; Ti02 and
Si02 particle size distribution: 0.3-0.6
pm
Route: Cell Culture
Dose/Concentration: 0,1, 4, and 8 mg EHC
total equivalent per 5 mL
Time to Analysis: Culture medium was
removed from flasks and replaced w/ 5 mL of
the particle suspension media. Plates were
incubated for 24 h. After 24 h cell culture
supernatants were collected and analyzed.
The decreased expression of preproET-1 in A549
cells suggests that epithelial cells may not be the
source of higher pulmonary ET-1 spillover in the
circulation measured in vivo in response to
inhaled urban particles. However, higher levels
ECE-1 in A549 post-exposure to particles sug-
gests an increased ability to process bigET-1 into
mature ET-1 peptide, while increased receptor
expression implies responsiveness. The
increased release of II-8 and VEGF by epithelial
cells in response to particles could possibly up
regulate ET-1 production in the adjacent pulmo-
nary capillary endothelial cells, with concomitant
increased ET-1 spillover in the systemic
circulation.
December 2009
D-4
-------
Study
Pollutant
Exposure
Effects
Reference: Chen et al. CAPs (NYU, NY)
(2005, 0872181
Species: Mouse
Strain: C57 and ApoE"'"
Particle Size: PM25
Route: Whole-body Inhalation
Dose/Concentration: 19.7 pg/m3 average
concentration over 5 mo (daily average expo-
sure concentration was 110 pg/m3)
Time to Analysis: 6 h/days, 5 days/wk, for 5
mo. Parameters measured continuously
throughout.
Significant decreasing patterns of HR, body
temperature, and physical activity for ApoE"'"
mice, with nonsignificant changes for C57 mice.
SDNN and RMSSD in the late afternoon and
overnight for ApoE'" mice showed a gradual
increase for the first 6 wk, a decline for about 12
more weeks, and a slight turn upward at the end
of the study period. For C57 mice, there were no
chronic effect changes in SDNN or RNSSD in the
late afternoon, and a slight increase after 6 wk for
the overnight period.
Reference: Chen and
Nadziejkov(2005,
087219)(Chen and
Nadziejko, 2006,
087219)
Species: Mouse
Strain: C57, ApoE"'"
Age: 26-28 wk(C57),
39-41 wk(ApoE"'"), and
18-20wk(LDLr"'"[DK])
CAPs (NYU, NY)
Particle Size: PM25
Route: Whole-body Inhalation
Dose/Concentration: Mean exposure
concentration: 110|jg/m3
Time to Analysis: 6 h/days, 5 days/wk for up to [.' "r'f wc;c V VT °
5mo. Sacrificed 3-6 days after last exposure "Pld content AP°E
All DK mice developed extensive lesions in the
aortic sinus regions. In male DK mice, the lesion
areas appeared to be enhanced by CAPs
exposure. Plaque cellularity was increased, but
Ł ,
prominent areas of severe atherosclerosis.
Quantitative measurements showed that CAPs
increased the percentage of aortic intimal surface
covered by grossly discernible atherosclerotic
lesion.
Reference: Corey LM
et al. (2006, 1563661
(2006,166366)(Corev
et al, 2006, 156366)
Species: Mouse
Gender: Male
Strain: ApoE"'"
Age: 11-12 mo
Weight: 32.84 g (avg)
PM collected November- March
between 1996-1999 (Seattle, WA)
Silica (U.S. Silica Company, Berkeley
Springs, WV)
Particle Size: PM25
Route: Nasal Instillation
Dose/Concentration: PM: 1.5 mg/kg; Saline:
50 ul; Silica: Min-u-Sil 5 in 50 pi saline
Time to Analysis: Mice monitored for 1 day
baseline prior to and for 4 days following
exposure.
After an initial increase in both HR and activity in
all groups, there was delayed bradycardia with no
change in activity of the animals in the PM and
silica exposed groups. In addition, with PM and
silica exposure, there was a decrease in HRV
parameters.
Reference: Cozzi et al.
(2006, 0913801
Species: Mouse
Strain: ICR
Age:6-10wk
Ultrafine PM (collected continuously
over 7-day periods in Oct 2002 in
Chapel Hill, NC)
Particle Size: <150nm
Route: IT Instillation
Dose/Concentration: 100 pg of PM in vehicle
Time to Analysis: 24 h post-exposure
Ischemia-Reperfusion: PM exposure doubled
the relative size of myocardial infarction com-
pared with the vehicle control. No difference was
observed in the percentage of the vehicle at the
risk of ischemia. PM exposure increased the
level of oxidative stress in the myocardium after I-
R. The density of neutrophils in the reperfused
myocardium was increased by PM exposure, but
differences in the numbers of blood leukocytes,
expression of adhesion molecules on circulating
neutrophils, and activation state of circulating
neutrophils, 24 h after PM exposure, could not be
correlated to the increase I-R injury observed.
Isolated Aortas: Aortas isolated from PM-
exposed animals exhibited a reduced
endothelium-dependent relaxation response to
ACh.
Reference: Dvonch JT CAPs, Detroit, Ml
et al. (2004, 0557411
Particle Size: PM25
Species: Rat
Gender: Male
Strain: Brown Norway
Route: Whole-body Inhalation
Dose/Concentration: Average concentration
354 pg/m3
Time to Analysis: 8 h/days for 3 consecutive
days; plasma samples collected 24 h post-
exposure.
Plasma concentrations of asymmetric
dimethylarginine (ADMA) were significantly
elevated in rats exposed to CAPs versus filtered
December 2009
D-5
-------
Study
Pollutant
Exposure
Effects
Reference: Elder et al.
(2004, 0556421
Species: Rat
Gender: Male
Strain: Fischer 344
andSH
Age: 23 mo (Fischer);
11-14mo(SH)
Weight: NR
UFP - Ultrafine carbon particles;
IPS (Sigma)
Particle Size: UFP: 36 nm (median
size)
Route: Whole-body inhalation; intraperitoneal
injection (ip) for saline and IPS
Dose/Concentration: Particles: 150 pg/m3;
IPS: 2 mg/kg bw
Time to Analysis: Single 6-h exposure to
particles. Sacrificed 24 h after ip IPS exposure.
BAL Fluid Cells: Neither inhaled UFP nor ip IPS
cause a significant increase in BAL fluid total
cells or the percentage of neutrophils in either rat
strain. No significant exposure-related alteration
in total protein concentration or the activities of
LDH and b-glucuronidase.
Peripheral Blood: In both rat strains ip IPS
induced significant increase in the number and
percentage of circulating PMNs. When combined
with inhaled UFP, PMNs decreased, significantly
for F-344 rats. Plasma fibrinogen increased with
ip IPS in both rat strains with the magnitude of
change greater in SH rats. UFP alone decreased
plasma fibrinogen in SH rats. Combined UFP and
IPS response was blunted, but significantly
higher than controls. Hematocrit was not altered
in either rat strain by any treatment.
TAT Complexes: Wth all exposure groups
averaged, plasma TAT complexes in SH rats
were 6.5 times higher than in F-344 rats. IPS
caused an overall increase in TAT complexes for
F-344 rats that was further augmented by inhaled
UFP. UFP alone decreased response. In SH rats,
UFP alone significant increased response and
IPS decreased response.
ROS in BAL Cells: In F-344 rats both UFP and
IPS have independent and significant effects on
DCFD oxidation. Effects were in opposite
directions; particles decreased ROS, IPS
increased ROS.
Reference: Finnerty et
al. (2007, 1564341
Species: Mouse
Gender: Male
Strain: C57BL/6
Age: 9 wk
Weight: 22-2 7g
Coal Fly Ash (U.S. EPA), Analysis:
(PM25 samples) low unburned carbon
(0.53 wt%), moderate levels of
transition metals, including (in pg/g):
Fe(30, 400), Mg (31, 200), Ti (6, 180),
Mn (907), and V (108).
Particle Size: 1.8 and 2.5pm
Route: IT Instillation
Dose/Concentration: PM: 200 pg; 200 pg
PM+10 pg IPS; 200 pg PM+100 pg IPS
Time to Analysis: 18 h after IT instillation
Plasma: TNF-a significantly increased in both
PM+LPS10 and PM+LPS100 treatments. For
plasma IL-6, all groups tended to rise with a
significant increase in the PM+LPS100 group.
Reference: Floyd et al.
(2009, 1903501
Species: Mouse
Gender: Male
Strain: ApoE"'"
Age: 20 wk
CAPs: PM2 5 concentrated from
Tuxedo, NY (April-Sept 2003)
Particle Size: NR
Route: Whole-body Inhalation
Dose/Concentration: Avg 120 pg/m3
(n=6/group)
Time to Analysis: 6 h/days x 5 days/wk x 5 mo
Gene Expression: Microarry gene expression
identified 395 genes downregulated and 216
genes upregulated in the aortic plaques.
Ontologic analysis identified a list of functional
cellular movement, cellular growth and
proliferation, hematological system development
and function, lipid metabolism, cardiovascular
system function, cellular assembly and
organization, and cell death.
Reference: Folkmann
et al. (2007, 0973441
Species: Mouse
Gender: Female
Strain: Wld type and
ApoE'-
Age: 11-13 wk
Weight: 21 g (avg)
DEP: SRM2975 (particulate fraction of
exhaust from a filtering system
designed for diesel-powered forklifts).
Particle Size: DEP: NR
Route: Intraperitoneal linjection
Dose/Concentration: 0, 50, 500, 5,000 pg
DEP/kg
Time to Analysis: 6 or 24 h post-ip injection
The expression of inducible nitric oxide synthase
(iNOS) mRNA was increased in the liver 6 h post-
ip injection. The level of oxidized purine bases,
determined byformamidopyrimidine DNA
glycosylase sites increased significantly in the
liver after 24 h in mice injected w/ 50|jg/kg. There
was no indication of systemic inflammation
determined as the serum concentration of nitric
oxide and iNOS expression, and DNA damage
was not increased in the aorta.
December 2009
D-6
-------
Study
Pollutant
Exposure
Effects
Reference: Furuyama
et al. (2006, 0970561
Species: Rat
Cell Type: Heart Micro
vessel Endothelial
(RHMVE) Cells
OE-DEP, OE-UFP (from Urawa,
Saitama, Japan)
OE = Organic Extracts
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 0, 5, 10, 25 pg/mLof
OE-DEP or OE-UFP
Time to Analysis: Exposed for 0, 6,12, 24, or
36 h
The cell monolayer exposed to 10 pg/mL OE-
UFP produced a larger amount of HO-1 than
cells exposed to 10 pg/mL OE-UFP. OE-DEP and
OE-UFP exposure reduced PAI-1 production by
the cells but did not affect the production of
thrombomodulin, tissue-type PA, or urokinase-
type PA. Increased PAI-1 synthesis in response
to treatment with 1 ng/mLTNF-aor0.5ng/mL
TGF-|31 was reduced by OE-DEP exposure.
Suppression of PAI-1 production by OE-DEP
exposure was mediated through oxidative stress
and was independent of HO-1 activity.
Reference: Gerlofs-
Nijland et al. (2009,
1903531
Species: Rat
Gender: Male
Strain: SH
Age: 12 wk
Weight: 200-300 g
PM (Prague, Czech Republic; Route: IT Instillation
Duisburg, Germany; Barcelona, Spain)
(Prague and Barcelona coarse PM Dose/Concentration: 7 mg /kg
Time to Analysis: DTPA added to some PM
samples preinstillation. Instilled with PM.
Necropsy 24 h post-exposure.
organic extracts)
Particle Size: Coarse: 2.5-10 pm,
Fine: 0.2-2.5 pm
Inflammation (LDH, protein, albumin), cytotoxicity
(NAG, MPO, TNF-a), and fibrinogen were
increased by PM, and were greatest in the
coarse PM fraction. Metal-rich PM had greater
inflammatory and cytotoxic effects, and increased
fibrinogen and vWF and decreased ACE. PAH
content influenced greater inflammation
(including neutrophils), cytotoxicity, and
fibrinogen. Generally, whole PM and coarse PM
were more potent than organic extracts and fine
PM, respectively.
Reference: Gerlofs-
Nijland et al. (2007,
0978401
Species: Rat
Gender: Male
Strain: SH/NHsd
Age: 13 wk
Weight: 250-350 g
PM samples collected from:
1. MOB high traffic density
2. HIA high traffic density
3. ROM high traffic density
4. DOR moderate traffic density
5 MGH low traffic density
6 LYC low traffic density
Particle Size: Coarse: 2.5-10 pm;
Fine: 0.1-2.5pm
Route: IT Instillation
Dose/Concentration: 3,10 mg/kg
Time to Analysis: 24 h
Hematology: Fibrinogen responses of SH rats
increased significantly at the high dose of both
fractions of all PM samples, except fine PM from
DOR.
Location-Related Differences: Coarse PM from
LYC lowered fibrinogen values more than PM
from location MOB, HIA, and MGH. Fine PM
showed less differences among the various sites.
Reference: Gerlofs-
Nijland et al. (2005,
0886521
Species: Rat
Gender: Male
Strain: SH/NHsd
Age: 11-12 wk
Weight: 250-350 g
RTD: road tunnel dust (obtained from a Route: IT Instillation
Motorway tunnel in Hendrik-ldo-
Ambacht, Netherlands) Dose/Concentration: 0.3,1, 3,10 mg/kg;
EHC-93:10 mg/kg
EHC-93
(Ottawa, Canada) Time to Analysis: 4, 24, 48 h
Particle Size: Coarse: 2.5-10 pm; fine:
0.1-2.5 pm
Hematology: No significant changes in plasma
for bigET-1 or von Willebrand factor were
observed. At the highest dose, fibrinogen levels
significantly increased at 24 and 4 h for both PM
types.
Reference: Ghelfi et al.
(2008, 1564681
Species: Rat
Strain: SD
Age: Adult
CAPs
CPZ (Capsazepine) (Axxora LLC, San
Diego, CA)
Particle Size: PM25
Route: CAPs: Whole-body Inhalation; CPZ: IP
Injection or Aerosol
Dose/Concentration: CAPs: mean mass
concentration: 218 + 23 pg/m ; CPZ: 10 mg/kg
(ip), 500 pmol (aerosol)
Time to Analysis: Experiment 1: CPZ ip or 20
min aerosol pretreatment immediately prior to
CAPs exposure. Single CAPs exposure for 5 h.
Parameters measured immediately following
exposure.
Experiment 2: CPZ ip pretreatment prior to
CAPs exposure. Exposed to CAPs for 5 h/day
for 4 mo. Parameters measured throughout
duration of experiment.
CPZ (ip or aerosol) decreased CAPs-induced
chemiluminescence (CL), lipid thiobarbituric acid
reactive substances (TBARS), and edema in the
heart, indicating that blocking TRP receptors,
systemically or locally, decreases heart CL. CAPs
exposure led to significant decreases in HR and
in the length of QT, RT, Pdur and Tpe intervals.
These changes were observed immediately upon
exposure, and were maintained throughout the
5-h period of CAPs inhalation. Changes in
cardiac rhythm and ECG morphology were pre-
vented by CPZ.
December 2009
D-7
-------
Study
Pollutant
Exposure
Effects
Reference: Gilmour et
al. (2004, 0541751
Species: Rat
Gender: Male
Strain: Wistar
ufCB (Printex 90 from Frankfurt,
Germany)
fCB(Huber990fromUK)
Particle Size: ufCB: 114 nm (MMAD);
fCB: 268 nm (MMAD).
Route: Whole-body Inhalation
Dose/Concentration: ufCB: 1.66 mg/m3; fCB:
1.40mg/m3
Time to Analysis: Single exposure for 7 h.
Sacrificed and samples taken at 0,16, and 48 h
post-exposure.
Exposure to ultrafine, but not fine, CB particles
was also associated with significant increases in
the total number of blood leukocytes. Plasma
fibrinogen factor VIII and vWFwere unaffected by
particle treatments as was plasma Trolox
equivalent antioxidant status.
Reference: Gilmour et
al. (2005, 0874101
Species: Human
Cell Types: Primary
Human Monocyte
Derived Macrophages
(MP); Human Umbilical
Vein Endothelial Cells
(HUVEC);A549 cells;
16HBE
PM10: (Carbon Black from Degussa
Ltd, Frankfurt, Germany)
Particle Size: PM10
Route: Cell Culture
Dose/Concentration: PM10: 50 and 100 pg/mL
Time to Analysis: 6 and 20 h
The culture media from MPsand 16HBE cells but
not A549 cells, exposed to PM10 had an
enhanced ability to cause clotting. H202 also
increased clotting activity. Apoptosis was
significantly increased in MPs exposed to PM10
and IPS as shown by annexin V binding. TF
gene expression was enhanced in MPs exposed
to PM10 and HUVEC tissue factor. tPA gene and
protein expression were inhibited.
Reference: Gilmour et
al. (2006, 1564721
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Age: 12-14 wk
Weight: 280-340 g
Zinc Sulfate
(ZnS04 in saline solution)
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 131 pg/kg (2 pmol/kg)
Time to Analysis: 1,4, 24, 48 h
Zinc levels in plasma and tissue: At 1-24 h
post-exposure, zinc plasma levels increased to
nearly 20% above baseline.
mRNA expression: Cardiac tissues demon-
strated similar temporal increases in expressions
of TF, PAI-1 and thrombomodulin mRNA,
following pulmonary instillation of Zn.
Cardiac histopathology: Mild and focal acute,
myocardial lesions developed in a few Zn ex-
posed rats. No changes in fibrin deposition or
troponin disappearance were observed. At 24
and 48h PE to Zn, increases occurred in levels of
systemic fibrinogen an the activated partial
thromboplastin time.
Reference: Gong et al.
(2007, 0911551
Species: Mouse
Cell Type: Human
Microvascular
Endothelial Cells
(HMEC)
Strain: C57BL/6J
Gender: Male
Age: 2 mo
Organic DEP extract: collected from
exhaust in a 4JB1-type LD, 2.74 liter,
4-cylinder Isuzu diesel engine
(provided by Masaru Sagai, Tsukuba,
Japan)
ox-PAPC: (provided by Judith Berliner,
UCLA, CA)
In vivo validation: Ultrafine (ufp) and
fine (fp) particulate matter
Particle Size: DEP <1 pm (diameter)
Route: Cell culture; In vivo validation via
Whole-body inhalation
Dose/Concentration: ox-PAPC: 10, 20, and 40
pg/mL; DEP: 5,15, and 25 pg/mL; DEP (5
pg/mL)+ox-PAPC: 10 or 20 pg/mL
In Vivo Validation: Ufp: 3.2 4x105/cm3; fp: 2.7
x105/cm3
In vivo validation: Ufp: <0.18|jm; fp: <2.5|jm
Time to Analysis: 4 h
In vivo validation: Exposed to CAPs for 5 h/day,
3 days/wk for 8 wk. Sacrificed 24 h after last
CAPs exposure.
Gene-expression profiling showed that both DEP
extract and ox-PAPC co-regulated a large
number of genes. Network analysis to identify co-
expressed gene modules, led to the discovery of
three modules that were highly enriched in genes
that were differentially regulated by the stimuli.
These modules were also enriched in
synergistically co-regulated genes and pathways
relevant to vascular inflammation.
In vivo validation: Results were validated by
demonstrating that hypercholesterolemic mice
exposed to ambient ultrafine particles inhibited
significant upregulation of the module genes in
the liver.
December 2009
D-8
-------
Study
Pollutant
Exposure
Effects
Reference: Goto et al.
(2004, 0881001
Species: Rabbit
Gender: Female
Strains: New Zealand
White
Age: NR
Weight: 2.3 kg
EHC-93 (Ottawa, ON.Canada)
CC: Coilloidal Carbon (obtained from
Hamburg, Germany)
Particle Size: EHC-93: PM,0; CC: <1
pm
Route: Intrabronchial Instillation
Dose/Concentration: AMs incubated with
EHC-93 or CC: 0.6 ml/kg
EHC-93 alone: 1 ml (500 pg/ ml)
CC alone: 1 ml (1%CC)
Time to Analysis: WBC counts measured
4-168 h after BrdU injection. Sacrificed /days
post instillation.
Lung Distribution of PMi0: PM-containing AMs
were distributed diffusely. PM-containing AMs
were more prevalent in the PM exposed animals.
There was noAM-containing particle difference
between the CC-exposed and EHC-93-exposed
groups.
Monocyte Release from Bone Marrow: EHC
exposure increased WBC and band cell counts
from 12 h after instillation. Monocyte count was
not affected. Labeled monocytes peaked more
quickly after DEP exposure (12 vs 16 h for
control). There was no observed change in BM
monocyte pool.
Cytokine Release: EHC stimulation increased
the release of GM-CSF, IL-6, IL-lp, TNF-a, IL-8
and MCP-1. No effect on m-CSF and MIP-1p. CC
particles induced increases in IL-6 and TNF-a;
other cytokine levels did not differ from control.
Supernatant Instillation: AM+EHC increased
circulating WBC and band cell counts. Circulating
monocyte counts were unaffected. AM+EHC
showed a major increase in fraction and amount
of monocyte released as well as faster clearance
when compared to control. The BM monocyte
pool was similar in all groups.
Monocyte Transit Time Through BM: Exposure
to EHC, CC only shortened the transit time of
monocytes as compared to controls. AM+EHC
also shortened monocyte transit time whereas
AM+CC had a nonsignificant effect.
Reference: Gottipolu
et al. (2009, 1903601
Species: Rat
Gender: Male
Strain: Wstar Kyoto,
SH
Age: 14-16 wk
Weight: NR
DE (30-kW (40hp) 4-cylinder indirect Route: Inhalation
injection Deutz diesel engine) (02-
20%, CO-1.3-4.8 ppm, NO- <2.5-5.9
ppm, N02- <0.25-1.2ppm, S02" 0.2-0.3
ppm, OC/EC-0.3+ 0.03)
Dose/Concentration: Low: 507 + 4 pg/m3:
High: 2201 + 14 pg/m3
Particle Size: Number Median
Diameter: Low- 83 + 2 nm, High- 88. 2
nm; Volume Median Diameter: Low-
207 +2 nm, High-225 +2 nm
Time to Analysis: Exposed 4 h/days, 5
days/wk, 4 wk. Necropsied 1day post-exposure.
DE dose-dependently inhibited mitochondrial
aconitase activity. DE caused 377 genes to be
differentially expressed within WKY rats, most of
which were downregulated, but none in SH rats.
However, WKY rats had an expression pattern
shift that mimicked baseline expression of SH
rats without DE. These genes regulated
compensatory response, matrix metabolism,
mitochondrial function, and oxidative stress
response.
Reference: Graff etal.
(2005, 0879561
Species: Rat
Cell Type: Ventricular
Myocytes
Zn;V
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 0, 6.25,12.5, 25, or 50
pm
Beat Rate: There were statistically significant
reductions in spontaneous beat rate 4 and 24 h
post-exposure (greater reductions were observed
with Zn).
Time to Analysis: Toxicity: 24 h post- exposure inflammation: Exposure to Zn or V (6.25-50 pm)
Beat Rate- 0 5 1 2 4 and 24 h PE for 6 h Produced significant increases in IL-6, IL-
Beat Rate. 0.5, 1 , 2, 4, and 24 h PE ^ hegt shocR protein ^ gnd connexin 43
PCR: 6 and 24 h PE
Impulse Conduction: 24 h post-exposure, Zn
induced significant changes in the gene ex-
pression of Kv4.2 and KvQLt, a-1 subunit of L-
type Ca channel, Cx43, IL-6, and IL-1a. V pro-
duced a greater effect on Cx43 and affected only
KvLQTl
Reference: Gunnison
and Chen (2005,
0879561
Species: Mouse
Gender: Male
Strain: F2 generation
DK (ApoE'~, LDLr'l
Age: 18-20 wk
CAPs (Tuxedo, NY)
Copollutants measured: 03 and N02.
Particle Size: 389 + 2 nm
Route: Whole-body Inhalation
Dose/Concentration: CAPs: 131 + 99 pg/m3
(range 13-441 pg/m3)
03:10 ppb
N02: 4.4 ppb
Time to Analysis: 6 h/days, 5 days/wk for
approximately 4 mo. Tissue collection was
performed 3-4 days after the last day of
exposure.
Gene Expression: In CAPs-exposed heart
tissue, the expression of Limdl and Rex3 were
the most consistently affected genes among the
exposed mice. Limdl was down regulated by
1.5-fold or greater from moderate baseline
expression. Rex3 showed a relatively small
increase in absolute expression.
December 2009
D-9
-------
Study
Pollutant
Exposure
Effects
Reference: Gurgueira
et al. (2002, 0365351
Species: Rat
Gender: Male
Strain: SD
Weight: 250-300 g
CAPs;
Carbon Black (CB from Fisher
Scientific, Pittsburgh, PA): C (85.9 ±
0.2%); 0(13 + 0.2%); S (1.17 +
0.02%)
ROFA: obtained from an oil-fired power
plant (Boston, MA)
Particle Size: CAPs size range: 0.1-
2.5 fjm;CB and ROFA (PM2.5)
Route: Whole-body Inhalation
Dose/Concentration: CAPs: average mass
concentration: 300 + 60 pg/m3; ROFA: 1.7
mg/m3;CB: 170|jg/m3
Time to Analysis: CAPs: 1, 3, and 5 h; ROFA:
30 min; CB: 5 h
Oxidative Stress: Rats breathing CAPs aerosols
for 5 h showed significant oxidative stress,
determined as in situ chemiluminescence (CL) in
the lung, heart, but not in the liver. ROFA also
triggered increases in oxidant levels but not
particle-free air or CB. Increases in CL showed
strong associations with the CAPs content of Fe,
Al, Si and Ti in the heart. The oxidant stress
imposed by 5 h exposure to CAPs was
associated with slight, but significant increases in
the lung and heart water content, with increased
serum levels of lactate dehydrogenase, indicating
mild damage to tissues. CAPs inhalation also led
to tissue-specific increases in the activities of
SOD and catalase.
Reference: Gursinsky
etal. (1976. 0156071
Species: Rat
Cell Type: Fibroblasts
isolated from adult
male Wistar rats hearts
Fly ash (TAF98)
Particle Size: NR
Route: In Vitro
Dose/Concentration: TAF98: 0,1, 2 3,10, 25,
50, 100, 200|jg/mL
Time to Analysis: 0, 5,10, 30, 60,120 min
Brief treatment of fibroblasts with fly ash triggered
the immediate formation of ROS. Using
phosphospecific antibodies the activation of p38
MAP kinase, p44/42 MAP kinase (ERK1/2) and
p70S6 kinase. Prolonged incubation with fly ash
increased the expression of collagen 1 and TGF-
P1, but decreased mRNA levels of MMP9 and
TNF-a. Cell proliferation was inhibited at high
concentrations of fly ash. An increase in the level
of advanced glycation end product modification
of various cellular proteins was observed.
Reference: Hansen et
al. (2007, 0907031
Species: Mouse
Gender: Female
Strain: ApoE"'" and
C57BL/6J ApoE*'*
Age: 11-13 wk
DEP: SRM-2975 (NIST)
Particle Size: DEP: 215 nm
(geometric mean diameter)
Route: Intraperitoneal Injection
Dose/Concentration: DEP: 0, 0.5 and 5 mg/kg;
Aorta segments incubated with 0,10 and 100
pg DEP/mL
Time to Analysis: Sacrificed 1 h after ip
injection.
Exposure to 0.5 mg/kg DEP caused a decrease
in the endothelium-dependent Ach elicited
vasorelaxation in ApoE"'" mice, whereas the re-
sponse was enhanced in ApoE*'* mice. No
significant changes were observed after
administration of 5 mg/kg DEP. K* or
phenylephrine induced constriction was not af-
fected.
Reference: Hansen et
al. (2007, 0907031
Species: Mouse
Gender: Female
Strain: ApoE"'" and
C57BL/6J ApoE*'*
Use: Aorta rings used
for in-vitro studies
DEP: SRM-2975 (NIST)
Particle Size: DEP: 215 nm
(geometric mean diameter)
Route: Cell Culture
Dose/Concentration: 0,10 and 100 |jg
DEP/mL
Time to Analysis: Basal tone measured at 5
different points throughout experiment.
Exposure to 100 pg DEP/mL enhanced ACh-
induced relaxation and attenuated
phenylephrine-induced constriction.
Vasodilatation induced by sodium nitroprusside
was not affected by any DEP exposure.
Reference: Harder et
al. (2005, 0873711
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Age: 12-15 wk
Carbon UFPs ((generated by Electric
Spark Generator GFG 1000; Palas,
Karlsruhe, Germany)
Particle Size: 37.6 + 0.7 nm (mean)
Route: Whole-body Inhalation
Dose/Concentration: 180 pg/m3
Time to Analysis: Days 1-3: baseline reading,
Day 4: exposure to UFPs or filtered air for 4 or
24 h then sacrificed immediately following
exposure period OR
Sacrificed following 1-3 days recovery period.
Cardiovascular Performance: Mild but
consistent increase in HR, which was associated
with a significant decrease in HR variability
during exposure (particle-induced alteration of
cardiac autonomic balance, mediated by a
pulmonary receptor activation).
Lung Inflammation and Acute-Phase Re-
sponse: BALF revealed significant but low-grade
pulmonary inflammation.
Effects on Blood: There was no evidence of an
inflammation-mediated increase in blood
coagulability; no changes in plasma fibrinogen or
factor Vila.
Pulmonary and Cardiac Histopathology: Spo-
radic accumulation of particle-laden
macrophages found in the alveolar region. No
signs of cardiac inflammation or cardiomyopathy.
mRNA Expression Levels: No significant
changes in the lung or heart.
December 2009
D-10
-------
Study
Pollutant
Exposure
Effects
Reference: Hirano et Organic Extracts of DEP (DEP) and Route: Cells Culture
al. (2003, 0973451 Organic Extracts of Ultra Fine Particles
Species: Rat (UFP)'
(Urawa City, Saitama, Japan)
Cell Types: Heart
Microvessel Endothelial Particle Size: DEP and UFP: <2.0 pm
Cells (RHMVE)
Dose/Concentration: MAC effects on viability:
DEP: 25 fjg/ml; UFP: 50 pg/ml
mRNA levels for DEP and UFP: 0,1,3,10 pg/ml
cell monolayer exposed to DEP and UFP:
1,10,100 fjg/ml
Time to Analysis: mRNA levels measured after
6 h incubation with DEP or UFP. Other
parameters measured after 24 h.
Cytotoxicity and Oxidative Stress: LC50
values were 17 and 34 pg/mL for DEP and UFP
respectively. The viability of DEP and UFP
exposed cells was ameliorated by N-acetyl-L-
cysteine (MAC).
mRNA Levels: mRNA levels increased dose-
dependently with DEP and HO-1 mRNA showed
the most marked response to DEP. mRNA levels
of antioxidant enzymes and heat shock protein
72 (HSP72) in DEP-exposed cells were higher
than UFP exposed cells at the same
concentration. The transcription levels of HO-1
and HSP72 in DEP and UFP-exposed cells were
also reduced by NAC.
Reference: Hwang et
al. (2005, 0894541
Species: Mouse
Strain: C57 and ApoE"'"
CAPs (Tuxedo, NY)
Particle Size: 389 + 2 nm
Route: Whole-body Inhalation
Dose/Concentration: CAPs Range: 5-627
pg/m3. Mean CAPs Concentration: 133|jg/m3.
Mean Concentrations of 03 and N02 in CAPs:
10 and 4.4 ppb respectively.
Time to Analysis: 6 h/day, 5 days/wk for 5 mo.
Long-term Analysis: Significant decreasing
patterns of HR, body temperature, and physical
activity in ApoE"'" mice. Nonsignificant changes
for C57 mice. The chronic effect changes for
ApoE" mice were maximal in the last three wk.
Short-term Analysis: Dose-dependent relation-
ship for HR variations in ApoE"'" mice.
Heart Rate Fluctuation: HR fluctuations in
ApoE"'" mice during the period of 3-6 h increased
by 1.35 fold at the end of the exposure and
during a 15 min period increases by 0.7 fold at
the end of the exposure.
Reference: Inoueetal.
(2006,1901421
Species: Mouse
Gender: Male
Strain: ICR
Age: 6- 7 wk
DEP (obtained from a 4Jb1-type light-
duty, 4-cylinder, 2.74-L Isuzu diesel
engine)
Washed DEP (carbonaceous nuclei of
DEP after extraction) and DEP-OC
(organic chemicals in DEP extracted
with CH2CI2); Washed DEP+LPS and
DEP-OC+LPS
Particle Size: PM25
Route: IT Instillation
Dose/Concentration: Washed DEP: 4 mg/kg.
DEP-OC: 4 mg/kg. IPS: 2.5 mg/kg. Washed
DEP+LPS and DEP-OC+LPS: respective addi-
tions of LPS to each component prior-
experimentation.
Time to Analysis: Sacrificed 24 h post single
dose instillation.
Both DEP components exacerbated vascular
permeability. The increased fibrinogen and E-
selectin levels induced by LPS. This exacerbation
was more prominent with washed DEP than with
DEP-OC. Washed DEP+LPS significantly
decreased protein Cand antithrombin-lll and
elevated circulatory levels of IL-6, KC and LPs
without significance.
Reference: Inoue et al. DEP (derived from 4 cyl, 2. 74! light
(2006, 0978151 duty diesel engine)
Species: Mouse Particle Size: NR
Gender: Male
Strains: C3H/HeJ
(TLR-4 point mutant)
and C3H/HeN (Control)
Route: IT Instillation
Dose/Concentration: 12 mg/kg
Time to Analysis: 24 h
Hematology: DEP increased plasma fibrinogen
in both strains but with a greater increase in the
knockout mice than the wild type.
Age: 6 wk
Reference: Ito et al.
(2008, 0968231
Species: Rat
Gender: Male
Strain: Wstar Kyoto
(Specific pathogen-
free)
Age: 13-14 wk
CAPs (f-PM), Yokohama City, Japan.
Particle Size: 0.1-2.5pm
Route: Whole-body Inhalation
Dose/Concentration: 0.6-1.5 mg/m3
Time to Analysis: Three groups exposed to:
(1) filtered air for 4 days, (2) filtered air for 3
days and CAPs for 1 day or (3) CAPs for 4
days. All groups exposed for a maximum of 4.5
h/days for 4 consecutive days.
mRNA Expression and Cardiovascular
Function: In samples of heart tissue, the mRNA
of cytochrome P450 (CYP) 1B1, heme
oxygenase-1 (HO-1), and endothelin A (ETA)
receptor were up-regulated by CAPs; their levels
were significantly correlated with the cumulative
weight of CAPs in the exposure chamber. The
up-regulation of ETA receptor mRNA was signifi-
cantly correlated with the increase in HO-1
mRNA and weakly with the increase in MBP
December 2009
D-11
-------
Study
Pollutant
Exposure
Effects
Reference: Khandoga
Aetal. (2004, 0879281
Species: Mouse
Gender: Female
Strain: C57B1/6
Age: 5-7 wk
UFPs: Ultra fine carbon black particles Route: Aortic Infusion
(Printex 90) 7 7
Dose/Concentration: 1 xi O7 and 5 xi o7 total
Particle Size: 14 nm diameter (60%
<100nm)
particles infused
300 m2/g surface area
Time to Analysis: Single exposure, analysis
2 h post-exposure
Platelet Effects: Application of UFPs caused
significantly enhanced platelet accumulation on
endothelium of postsinusodal venules and
sinusoids in healthy mice. UFP-induced platelet
adhesion was not preceded by platelet rolling but
was strongly associated with fibrin deposition and
an increase in vWF expression on the endothelial
surface.
Inflammatory Effects: In contrast, inflammatory
parameters such as the number of
rolling/adherent leukocytes, P-selectin expres-
sion/translocation, and the number of apoptotic
cells were not elevated. UFPs did not affect
sinusoidal perfusion and Kupffer cell function.
Reference: Knuckles ROFA-L: Leachate
et al. (2007, 1566521
Particle Size: <0.2 pm
Species: Rat
Gender: Female
(Pregnant, purchased
atGD19)
Strain SD
Route: Cell Culture
Dose/Concentration: 3.5 |jg/mL
Time to Analysis: 1 h
ROFA-L Induced Alterations to the RCM
Transcriptosome: 38 genes were suppressed
and 44 genes were induced PE. Genomic
alterations in pathways related to IGF-1, VEGF,
IL-2, PI3/AKT, CVD, and free radical scavenging
were detected. Global gene expression was
altered in a manner consistent with cardiac myo-
cyte electrophysiological remodeling, cellular
oxidative stress and apoptosis.
Age: 60-90 days
Weight: 300 g
Use: RMCs were
harvest from 1 day-old
neonatal pups
ROFA-L Induced Alterations to the RCM
Transcription Factor Proteome: ROFA-L
altered the transcription factor proteome by sup-
pressing activity of 24 and activating 40 trans-
cription factors out of 149.
Reference: Knuckles
et al. (2008, 1919871
Species: Mouse
Gender: Male
Strain: C57BL/6
Age:8-10wk
Weight: NR
DE (single cylinder Yanmar diesel
generator burning #2 certified diesel
fuel (Chevron-Phillips, Borger, TX)
under 100% load)
Particle Size: PM25
Route: Whole-body Inhalation. Ex Vivo.
Dose/Concentration: In vivo: 350 ug/m3; Ex
vivo: PM2 5 concentration 2-3 mg/m flow rate
500 mL/min
Time to Analysis: Exposed 4 h. Ex vivo
assays.
Veins: DE increased vascular reactivity to ET-1.
Ex vivo exposed vessels had greater
vasoconstriction. L-NAME (an arginine blocker)
did not promote constriction in DE-exposed rats
but did so in controls.
Arteries: DE did not significantly alter vascular
reactivity. Carbonyls or alkanes alone or with DE
did not alter vasoconstriction.
Reference: Kodavanti
et al. (2008, 1559071
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Age: 12-14 wk
G1: saline (control); G2: Mount Saint
Helen's ash (SH); G3: whole suspen-
sion of oil combustion PM at high
concentration (PM-HD); G4: whole
suspension of oil combustion PM at
low concentration (PM-LD); G5: saline-
leachable fraction of PM high-
concentration suspension; G6: zinc
sulfate
Particle Size: PM25
Route: IT Instillation
Dose/Concentration: Doses (mg/kg/wk) are for
8 and 16 wk (PM-solid and soluble Zn)
respectively. G1: 0.00-0.00 and 0.00-0.00; G2:
4.60-0.00 and 2.30-0.00; G3: 4.60-66.8 and
2.30-33.4; G4: 2.30-33.4 and 1.15-16.7; G5:
0.00-66.8 and 0.00-33.4; G6: 0.00-66.8 and
0.00-33.4
Time to Analysis: 1 x/wk for 8 or 16 wk;
analyzed 48 h after last instillation.
DMA Damage (left ventricular tissue): All
groups except MSH caused varying degrees of
damage relative to control. Total cardiac
aconitase activity was inhibited in rats receiving
soluble Zn. Analysis of heart tissue revealed mo-
dest changes in mRNA for genes involved in sig-
naling, ion channels function, oxidative stress,
mitochondria! fatty acid metabolism, and cell
cycle regulation in Zn, but not MSH-exposed rats.
Reference: Kooter et
al. (2006, 0975471
Species: Rat
Gender: Male
Strain: SH
Ana- •]r) '^A u/l/
Age. IZ- 14 WK
CAP-F = fine (Site 1)
CAP-UF = fine + ultrafme (Site II)
(Netherlands)
Some measured components: Ammo-
nium, nitrate, sulfate ions: 56 ± 16%
CAP-F mass, 17 ± 6% CAP-UF mass
Particle Size: 0.15
-------
Study
Pollutant
Exposure
Effects
Reference: Kyoso et
al. (2005, 1869981
Species: Rat
Gender: NR
Strain: NR
Age: 15 mo
DE
PM and NOX exposures
Particle Size: NR
Route: Whole-body Inhalation
Dose/Concentration: PM (mg/m3): 0.01, 0.109,
0.54, 1.09, 0.01 (from 1.09 concentration w/o
PM)
NOX (ppm): 0.19, 0.59, 2.60, 5.53, 5.47 (w/o
PM)
Time to Analysis: Exposed 16 h/days (from
All of the resting R-R intervals before exposure
were lower at night than during the day, but few
changes were found after exposure.
5pm-9am) for 7 mo
Reference: Lei et al.
(2004, 0878841
Species: Rat
Gender: Male
Strain: SD
Weight: 300-350 g
CAPs from Asian dust storm (Taiwan) Route: Nose-only Inhalation
Measured Components: Si, Al, S, Ca,
K, Mg, Fe.As, Ni,W,V, OC, EC, S02,
N02, nitrate, sulfate
Particle Size: 0.01- 2.5pm
Dose/Concentration: 315.6 pg/m3 (Low)
or 684.5 pg/m3 (High)
Time to Analysis: Low: Exposed for 6 h.
Sacrificed 36 h post-exposure
High: Exposed for 4.5 h. Sacrificed 36 h post-
exposure
Pulmonary hypertension induced 2 wk pre-
exposure.
Hematology: PM induced a dose-dependent
increase in WBCs. No change was seen in
RBCs. Platelet results were highly variable.
Reference: Lei et al.
(2005, 0886601
Species: Rat
Gender: Male
Strain: SD
Weight: 200-250 g
Use: ip STZ (60 mg/kg)
dissolved in citric acid
buffer administered to 8
rats to induce diabetes;
ip citric acid buffer
administered to 8 non-
diabetic rats
CAPs: Hsin-Chuang, Taipei
Particle Size: PM: 0.01-2.5 pm
Route: IT Instillation
Dose/Concentration: PM25: 200 pg in 0.5 mL
saline. Components (pg/m): (9.8-SD2.4)), EC
(3.6-SD 3.2), Sulfate (4.8-SD 1.2), Nitrate (6.3-
SD 3.4)
Time to Analysis: Single dose. Animals
sacrificed 24 h post instillation.
Effects of Diabetes: Body weight (bw) of
diabetic (D) rats (397.5 g) was lower than non-
diabetic (ND) rats (483.1 g). Mean plasma
glucose level was 163 mg/daysL in ND rats and
448.2 mg/daysL in D rats. D rats had significant
greater levels of 8-OHdG in plasma compared to
ND rats. D rats had significantly increased levels
of plasma [nitrate+nitrite]. No observable
changes in TNF-a for D and ND rats.
Effects of PM Exposure ND Rats: Increase in
plasma levels of 8-OHdG and plasma IL-6, TNF-
a, and serum CRP Significant reduction of
plasma [nitrate+nitrite]. No significant effect on
plasma ET-1.
Effects of PM Exposure STZ-D Rats:
Significant elevation of plasma ET-1. Decrease in
plasma [nitrate+nitrite] Plasma 8-OHdG and
TNF-a significantly increased. No significant
alterations in IL-6 and CRP.
Reference: Lemos et
al. (2006, 0885941
Species: Mouse
Gender: NR
Strain: BALB/c
Age: 1day (neonatal)
n:10
Weight: 4-6 g
PM10, CO, N02, and S02 from
Universidade de Sao Paulo, Brazil.
Particle Size: PIvl-;
Route: Whole-body Inhalation
Dose/Concentration: Mean (+ SD)
concentrations were: C02: 2.06 + 0.08 ppm (8h
mean); N02:104.75 + 42.62 pg/m3 (24 h mean);
S02:11.07 + 5.32 pg/m3 (24 h mean);
PM,0: 35.52 + 12.84 pg/m3 (24 h mean)
Time to Analysis: 24 h/days, 7 days/wk for 4
mo
Morphometric measurements of the ratio
between the lumen and the wall (L/W) areas
were performed on transverse sections of renal,
pulmonary and coronary arteries. A significant
decrease of L/W with exposure to air pollution
was detected in pulmonary and coronary arteries,
whereas no effects of air pollution were observed
in renal vessels.
Reference: Li et al.
(2005, 0886471
Species: Rat
Strain: SD
Tissues/Cell Types:
Cultured HPAECs;
Pulmonary Artery Rings
(PARs)
Urban Particles (UPsSRM 1648)
Major Constituents (mass fraction in
%):AI(3.4), Fe(3.9), K(1.1).
Minor Constituents (mass fraction in
%): Na (0.43), Pb (0.66), Zn (0.48).
Trace Constituents (ng/mg): As (115),
Cd (75), Cr (403), Cu (609), Mn (786),
Ni (82), Se (27), U (5.5), V (127).
Particle Size: NR
Route: PARs: In vitro organ model HPAECs:
grown to 80% confluence
Dose/Concentration: PARs and HPAECs: 1 to
100 pg/mL; Losartan treatment: 0.2 pmol
Captopril treatment: 100 pmol
Time to Analysis: PARs were exposed to
increasing doses of UPs from 1 to 100 pg/mL
Maximum tension was recorded within 5 min
after each UPs dose. HPAECs: exposed to UPs
from 1 to 100 pg/mL for up to 2 min
Effects of UPs on the constriction of isolated rat
pulmonary PARs and the activation of
extracellular signal-regulated kinases 1 and 2
(ERK1/2) and p38 mitogen-activated protein
kinases (MAPKs) in HPAECs with or without
Losartan at 1-100 pg/mL induced acute
vasoconstriction. UPs also produced a time- and
dose-dependent increase in phosphorylation of
ERK1/2 and p38 MAPK. Losartan pre-treatment
inhibited both vasoconstriction and activation of
ERK1/2 and p38. The water soluble fraction of
UPs was sufficient for inducing ERK1/2 and p38
phosphorylation, which was also inhibited by
Losartan. Cu (CuS04) and V (VOS04), induced
pulmonary vasoconstriction and phosphorylation
of ERK1/2 and p38, but only phosphorylation of
p38 was inhibited by Losartan. UPs induced
activation of ERK1/2 and p38 was attenuated by
Captopril.
December 2009
D-13
-------
Study
Pollutant
Exposure
Effects
Reference: Li et al.
(2006, 1566931
Species: Rat, Rabbit,
and Human
Tissues/Cell Types:
Pulmonary Artery Rings
(PARs) (rat); isolated
buffer-infused lungs
(rabbits) and cultured
HPAECs
Strain: SD Rats, New
Zealand White Rabbits
Weight: Rat: 200-350
g; Rabbit: 2.5-3.0 kg
Urban Particles (UPsSRM 1648).
Major Constituents (mass fraction in
%):AI(3.4), Fe(3.9), K(1.1).
Minor Constituents (mass fraction in
%): Na (0.43), Pb (0.66), Zn (0.48).
Trace Constituents (ng/mg): As (115),
Cd (76), Cr (403), Cu (609), Mg (786),
Ni (82), Se (27), U (5.5), V (127).
Particle Size: NR
Route: In Vitro
Dose/Concentration: PARs and HPAECs: 1 to
100 pg/mL
Time to Analysis: PARs: treatment given 15
min prior to exposure. Exposed to increasing
doses of UPs from 1 to 100 pg/mL. Maximum
tension was recorded within 5 min after each
UPs dose. HPAECs: exposed to UPs from 1 to
100|jg/mLfor20and120min.
Effects of UP on H202 Release: Within minutes
after UPs treatment, HPAEC increased H202
production that could be inhibited by DPI, APO,
and NaN3. The water soluble fraction of UPs as
well as its two transition metal components Cu
and V, also stimulated H202 production. NaN3
inhibited H202 production stimulated by Cu and
V, whereas DPI and APO inhibited only Cu-stimu-
lated H202 production. Inhibitors of other H202-
producing enzymes, including N-methyl-L-
arginine, indomethacin, allopurinol, cimetidine,
rotenone, and antimycin, had no effects.
Effects of UP-induced H202 on MARK
Activation: DPI but not NaN3 attenuated UPs-
induced pulmonary vasoconstriction and
phosphorylation of ERK1/2 and p38 MAPKs.
Knockdown of p47phox gene expression by
small interfering RNA attenuated UPs-induced
H202 production and phosphorylation of ERK1/2
and p38 MAPKs.
Reference: Lippmann
et al. (2005, 0874531
Species: Mouse
Strain: C57 and ApoE"'"
(March-September 2003). Chemical
Composition: regional secondary
sulfate (SS) characterized by high S,
Si, and organic C; resuspended soil
(RS) characterized by high
concentrations of Ca, Fe, Al, and Si;
RO-fired powered emissions of the
Eastern U.S. identified by the presence
ofV, Ni, and Se; and motor vehicle
(MV) traffic and other sources.
Contributors to Average Mass: SS
56.1%), RS(11.7%), RO combustion
1.4%), MV traffic and other sources
(30.9%)
Particle Size: PM25
Route: Whole-body Inhalation
Dose/Concentration: PM25 concentrated ten-
fold, producing an average of 113 pg/m3
Time to Analysis: 6 h/days, 5 days/wk for 5
mo. Parameters measured daily: during
exposure, the afternoon after exposure, and
late at night
Associations Between Sources and Short-
term Heart Rate Changes: There were no
significant associations between SS, RS, RO,
and MV factors and HR in C57 mice at any of the
three intervals. There were significant asso-
ciations between PM2 5 and the RS source factor
and decreases in HR for the ApoE"'" mice during
the daily CAPs exposures but no associations
with the other factors. There was no residual
association of HR with PM2 5 or the RS factor
later in the afternoon or late at night. In the
afternoon, there was a significant association
between decreases in HR and the SS factor for
the ApoE" mice that had not been present during
exposure and did not persist into the night time
period. MV traffic and others were not
significantly associated with HR during any of
these three time periods. For the C57 mice, there
were no significant associations of HR with PM2 5
or any of its components during any of the three
daily time periods.
Associations Between Sources and Short-
term HRV Changes: Signal noise during
exposures did not permit reliable analyses of
HRV changes during the hours of CAP exposure.
Reference: Lippmann CAPs (Sterling Forest, spring-summer
et al. (2005, 0874531 2003)
Species: Mouse
Gender: NR
Strain: ApoE"'", ApoE"'"
LDLr"'", C57BL/6
Age: NR
Weight: NR
Particle Size: PM2
Route: Inhalation
Dose/Concentration: PM25 average
concentration: 110 pg/m3, Long-term average:
19.7 pg/m3
Time to Analysis: Exposed 6 h/days, 5
days/wk, 5 or 6 mo. Semicontinuous EKG
recordings.
HR increased in ApoE" mice but not C57 mice.
HRF increased over the duration of the
experiment. Atherosclerotic plaque deposits and
coronary artery disease lesions occurred in both
CAPs-exposed mice and controls, but invasive
lesions were only present in CAPs-exposed
mice. A gene affecting circadian rhythm was
upregulated in double knockout mice. CAPs
activated NF-KB. No inflammation occurred in the
pulmonary system.
Reference: Lippman M CAPs from Tuxedo, NY. Component of Route: Whole-body Inhalation
et al. (2006, 091165) interest: Ni.
Species: Mouse
Gender: Male
Strain: ApoE"'"
Age: 6 wk
Particle Size: PM25
Dose/Concentration: Average daily CAPs:
85.6 pgr3: Average daily Ni: 43 ng/m3
Time to Analysis: 6 h/day, 5 days/wk, for 6 mo
(July 2004-January 2005). 10-sECG, HR,
activity, and body temperature data were
sampled every 5 min for the duration of the
experiment.
For the CAPs-exposed mice, on 14 days there
were Ni peaks at approximately 175 ng/m and
usually low CAPs and V For those days back-
trajectory analysis identified a remote Ni point
source. ECG measurements on CAPs-exposed
and sham-exposed mice showed Ni to be
significantly associated with acute changes in HR
and HRV.
December 2009
D-14
-------
Study
Pollutant
Exposure
Effects
Reference: Lund et al.
(2007, 1257411
Species: Mouse
Gender: Male
Strain: ApoE"'"
Age: 10 wk
Use: Mice were placed
on a high fat at the
beginning of the
exposure.
Varying dilutions of gasoline emis-
sions: (generated using two 1996
model 4.3L General Motors V-6
engines, fueled with conventional,
unleaded, non-oxygenated gasoline,
equipped with stock exhaust systems).
Composition for Hi, Med, and Lo
dilutions:
PM, NOX, CO, and Total Hydrocarbons
(THC)
Particle Size: NR
Route: Whole-body Inhalation
Dose/Concentration: FA: PM (2 pg/m3), NOX
(0 ppm), CO (0.1 ppm), HC (0.1 ppm);
Low (1: 90 dilution of exhaust): PM (8 pg/m3),
NOX (2 ppm), CO (9 ppm), HC (0.9 ppm);
Mid (1: 20): PM (39 pg/m3), NOX (12 ppm), CO
(50 ppm), HC (8.4 ppm);
High (1:12): PM (61 pg/m3), N0x(19ppm), CO
(80 ppm), HC(12ppm);
High-filtered (1:12): PM (2 pg/m3), NOX (18
ppm), CO (80 ppm), HC (12.7 ppm).
Time to Analysis: 6 h/day, 7 days/wk for 7 wk.
Mice were sacrificed within 16 h PE. During the
study period all animals concurrently exposed
to the following: FA: 8 pg/m3 and 40 pg/m3; PM
Whole Exhaust: 60 pg/m ; or Filtered Exhaust
w/ gases matching the 60 pg/m3 concentration.
Inhalation exposure to gasoline engine emissions
resulted in increased aortic mRNA expression of
matrix metalloproteinase-3 (MMP-3), MMP-7, and
MMP-9, tissue inhibitor of MMP-2, ET-1 and HO-
1 in ApoE"'" mice; increased aortic MMP-9 protein
levels were confirmed through immunochemistry
Elevated ROS were also observed in arteries
from exposed animals, despite absence of
plasma markers. Similar findings were also
observed in the aortas ApoE" mice exposed to
particle filtered atmosphere, implicating the
gaseous components of the whole exhaust in
mediating the expression of markers associated
with vasculopathy.
Reference: Lund et al.
(2007, 1257411
Species: Mouse
Gender: Male
Strain: ApoE"'"
Age: 10 wk
Weight: NR
GEE (conventional unleaded,
nonoxygenated, nonreformulated
gasoline- ChevronPhillips Specialty
Fuels Division)
Particle Size: 0.150 pm(MMAD)
Route: Inhalation
Dose/Concentration: PM: 60 pg/m3, N02: 2
ppm, NO: 16 ppm, CO: 80 ppm, THC: 12.7 ppm
Time to Analysis: Mice fed high-fat diet 30
days before exposure. Exposed 6 h/day, 1 or 7
days. Some groups dosed with Tempol or BQ-
123. Killed within 18 h of last exposure.
Aorta gelatinase activity increased with GEE
exposure time. MMP-2/9 activity spread
throughout the vasculature by day 7. 7 day GEE
exposure significantly increased the aorta protein
expression of MMP-9, MMP-2, TIMP-2, and
plasma MMP-9. Generally, in GEE-exposed
mice, Tempol decreased TBARS and vascular
ET-1, and BQ-123 decreased vascular ROS, ET-
1, MMP-9, and gelatinase activity.
Reference: McQueen
et al. (2007, 0962661
Species: Rat
Gender: Male
Strain: Wstar Kyoto
Weight: 228-500 g
DEP: SRM 2975 (NIST)
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 0.5 mL/rat of 1 mg/mL;
1-2.2 mg/kg
Time to Analysis: 6 h.
Pre-exposure: Vagotomy (sectioning of vagus
nerve) or atropine, 1 mg/kg i.p. administered 30
min prior, 2 and 4 h post.
Cardiovascular Response: Blood pressure and
heart rate were unaffected. Average arterial 02
increased after DEP, but not when compared for
each animal. C02 and pH were not affected
Reference: Medeiros
et al. (2004, 0960121
Species: Mouse
Gender: Male
Strain: BALB/c
Age: 60 days
Weight: 20-30 g
CP: Carbon particles
PSA: ROFA (solid waste incinerator
hospital Sao Paulo, Brazil)
PSB: electric precipitator, steel plant,
Brazil)
PSA/PSB Characteristics: Generally,
PSB had greater component
concentrations than PSA: Br(100+x),
Cr (3x), Fe (10+x), Mn (2x), Rb (60+x),
Se (7x), Zn (4x). PMA>PMB: Ce (3x),
Co (10+x), La(100x), Sb(15x),V
(50x).
Particle Size: CP: 1.7 + 2.5 pm
(78%<2.5 fjm)
PMA:1.2±2.2|jm(98%<2.5|jm)
PMB:1.2±2.2|jm(98%<2.5|jm)
Reference: Intranasal Instillation
Dose/Concentration: CP: 10 pg/mouse;
0.5mg/kg
PSA: 0.1,1 or 10 pg/mouse; 0.005, 0.05, 0.5
mg/kg
PSB: 0.1,1 or 10 pg/mouse; 0.005, 0.05, 0.5
mg/kg
Time to Analysis: Single, 24 h
Hematology: PSA and PSB decreased
leukocyte count (all 3 doses) and platelet count
(2 high doses). No effect on hemoglobin,
erythrocytes and reticulocytes was observed.
Fibrinogen levels increased for both PSB and
PSA with PSB seeing a higher increase. None of
the effects were dose-dependent.
Bone Marrow: Erythroblasts increased for PSA
at all dose levels and PSB at mid and high dose
levels (high variability).
December 2009
D-15
-------
Study
Pollutant
Exposure
Effects
Reference: Montiel-
Davalosetal.(2007,
1567781
Species: Human
Cell Types: HUVEC
(from primary human
endothelial cells) and
U937 (human leukemia
pro-monocytic) cell
cultures.
PM25 and PMi0 from Mexico City
Particle Size: PM25, PM10
Route: In Vitro
Dose/Concentration: HUVEC TNF-a (10
ng/mLl, and a PM range of 5,10, 20, and 40
pg/cm concentrations.
Time to Analysis: 6 or 24 h (early and late
adhesion molecules respectively)
Results showed that both PM25 and PM10
induced the adhesion of U937 cells to HUVEC,
and their maximal effect was observed at 20
pg/cm . This adhesion was associated with an in-
crease in the expression of all adhesion
molecules evaluated for PM10, and E-selectin, P-
selectin, and ICAM-1 for PM25. In general the
maximum expression of adhesion molecules in-
duced by PM25 and PM10 was obtained with 20
pg/cm ; however PMio-induced expression was
observed from 5 pg/cm2. E-selectin and ICAM-1
had the strongest expression in response to
particles.
Reference: Moyer et In phosphide (InP), Co sulfate hep- Route: Inhalation
al. (2002, 0522221
Species: Mouse
Gender: Male and
Female
Strain: B6C3F1
tahydrate (CoS04 7H20), Vanadium
pentoxide(V205) Gallium arsenide
(GaAs), Ni oxide (NiO), Ni subsulfide
Ni3S2), Ni sulfate hexahydrate
NiS04 • 6H20), talc, and Mo trioxide
(Mo03)
Particle Size: MMAD particle size
(pm): InP (1.1-1.3), CoS047H20 (IS-
IS), V205: (1.0), GaAs: (1.0)
Dose/Concentration: High-Dose Con-
centration in Chronic Studies, Male (pg/m ):
InP: 0.3, CoS047H20: 3.0, V205: 4.0, GaAs:
1.0
High-Dose Concentration in Sub-Chronic
Studies, Male or Female (pg/m3): InP: 100,
CoS047H20: 30, V205:16, GaAs: 75
Time to Analysis: Phase One: Evaluation of
heart, kidney and lung tissues from all control
and high dose male B6C3F1 mice exposed by
inhalation to 9 particulate compounds for a 2yr
period. Phase Two: evaluated heart, lung,
kidney and mesentery tissues of control and
high dose male and female B6C3F1 mice from
the 90-day studies of the 4-compounds demon-
strating arteritis after a 2-yr period.
Phase One: High-dose males developed signifi-
cantly increased incidences of arteritis over
controls in 2 of the 9 studies (InP and CoS04
7H20), while marginal increases of arteritis were
detected in 2 additional studies (V205 and GaAs).
In contrast, arteritis of the muscular arteries of
the lung was not observed. Morphological
features of arteritis in these studies included an
influx of mixed inflammatory cells including
neutrophils, lymphocytes, and macrophages.
Partial and complete effacement of the normal
vascular wall architecture, often with the exten-
sion of the inflammatory process into the
periarterial connective tissue, was observed.
Phase Two: Results showed that arteritis did not
develop in the 90-day studies, suggesting that
long-term chronic exposure to lower-dose
metallic PM may be necessary to induce or
exacerbate arteritis.
Reference: Mutlu et al.
(2007,1214411
Species: Mouse
Gender: Male
Strain: 57BL/6 (IL-6*'*
and IL-ff'l
Age: 6-8 wk
Weight: 20-25 g
PMio from ambient air in Dusseldorf,
Germany
Particle Size: PM10
Route: IT Instillation
Dose/Concentration: PMi0:10 pg; Clodronate:
120mg
Time to Analysis: For alveolar macrophage
depletion, clodronate instilled into mice lungs
following endotracheal intubation 48 h prior to
instillation of PM. Parameters measured 24 h
post-exposure.
Mice treated with PMio exhibited a shortened
bleeding time, decreased prothrombin and partial
thromboplastin times (decreased plasma clothing
times), increased levels of fibrinogen, and
increased activity of factors II, VIII, and X. This
prothrombotic tendency was associated with
increased generation of intravascularthrombin,
an acceleration of arterial thrombosis, and an
increase in BALF concentration of prothrombotic
IL-6. IL-6" mice were protected against PM-
induced intravascular thrombin formation and the
acceleration of arterial thrombosis. Depletion of
macrophages by the IT administration of
liposomal clodronate attenuated PM-induced IL-6
production and the resultant prothrombotic
tendency.
Reference: Nadziejko
et al. (2002, 0874601
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Age: 16 wk
CAPs (PM25) from Tuxedo, NY. (S02,
N02 03 and NH3 were removed prior to
exposure).
H2S04 (fine and ultrafme)
Particle Size: Ultrafme H2S04 50-75
nm (MMAD)
Route: Nose-only Inhalation
Dose/Concentration: CAPs: 80 and 66 pg/m3
(avg 73); Fine H2S04: 299, 280,119, and 203
pg/m3 (avg 225); Ultrafme H2S04:140, 565,
416, 750 pg/m3 (avg 468)
Time to Analysis: 4 h/exposure
Exposure to CAPs caused a striking decrease in
respiratory rate that was apparent soon after the
start of exposure and stopped when exposure to
CAPs ceased. The decrease in respiratory rate
was accompanied by a decrease in HR.
Exposure of the same animals to fine-particle-
size H2S04 aerosol also caused a significant
decrease in respiratory rate similar to the effect of
CAPs. Ultrafine H2S04 had the opposite effect on
respiratory rate compared to CAPs.
Reference: Nadziejko
et al. (2004, 0556321
Species: Rat
Gender: Male
Strain: F344
Age: 18 mo
PM/CAPs (Tuxedo, NY)
UFC (lab generated)
S02
Particle Size: PM (Size Range): 0.5-
2.5pm; UFC (MMAD): 30-50 nm
Route: Nose-only Inhalation
Dose/Concentration: PM (pg/m3): 161-200,
avg. 180; UFC (pg/m3): 500-1280, avg. 890;
S02(ppm):1.2,1.2, avg. 1.2
Time to Analysis: A total of 8 exposures were
performed: 2 exposures to CAPs, 2 exposures
to UFC, 4 exposures to S02. All three pollutants
were tested w/ a crossover design so that each
group alternated exposure to air and to
pollutant. Exposures lasted 4 h and were
performed at least 1wk apart. Parameters
measured throughout duration of experiment.
Old F344 rats had many spontaneous
arrhythmias. There was a significant increase in
the frequency of irregular and delayed beats after
exposure to CAPs. The same rats were
subsequently exposed to UFC, S02 or air with
repeated crossover design. In these experiments
there was no significant change in the frequency
of any category of spontaneous arrhythmia
following exposure to UFC or S02.
December 2009
D-16
-------
Study
Pollutant
Exposure
Effects
Reference: Nemmar et DEP (SRM 2975)
al. (2008, 0965661
Particle Size: <1 |jm
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Weight: 440+14 g
Route: Intravenous via the tail vein
Dose/Concentration: DEP: 0.02mg orO.lmg
DEP/kg (corresponding to about 8 pg or 44 pg
DEP/rat)
Time to Analysis: 48 h following systemic
administration of saline or DEP
Intravenous administration of DEP (0.1 mg/kg)
triggered systemic inflammation characterized by
an increase in monocyte an granulocyte
numbers. Both doses of DEP caused a reduction
of RBC numbers and hemoglobin concentration.
TEM analysis of RBCs after in vitro incubation (5
pg/mL) or in vivo administration of DEP, revealed
the presence of ultrafme-sized aggregates of
DEP within the RBC. Larger aggregates were
also taken up by the RBC. The myocardial mor-
phology and capillary bed were not affected by
DEP exposure.
Reference: Nemmar et DEP (SRM 2975)
,.(2007,156800) Partjc|e sjze: NR
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Age: 16 wk
Weight: 424 ± 8 g
Route: Tail Vein Injection
Dose/Concentration: 8, 42, or 212 pg DEP/rat
(150|jl of 0.02, 0.1, or 0.5 mg/kg)
Time to Analysis: 24 h
Effect of DEP on Blood Pressure: Significant
decrease on BP in DEP-exposed rats at doses of
0.02 mg/kg, compared with mean BP observed in
controls.
Effect of DEP on HR: Doses of 0.02, 0.1, and
0.5 mg/kg in rats, resulted in significant reduction
of HR compared to controls.
Effect of DEP on Tail Bleeding Time:
Shortening of tail bleeding time in rats exposed to
0.02, 0.1, and 0.5 mg/kg. The shortening was
significant at the dose of 0.02 and 0.5 mg/kg
compared w/ controls. Platelet counts in blood
did not significantly increased post-DEP
administration.
Effect of DEP on WBC and RBC Numbers: No
significant effect of DEP at doses of 0.02, 0.1 and
0.5 mg/kg on the numbers of granulocytes,
monocytes, or lymphocytes compared with
control.
Reference: Nemmar et DEP (SRM 1650)
al. (2003, 0965671
Particle Size: NR
Species: Hamster
Gender: Male and
Female
Weight: 100-110 g
Route: IT Instillation
Dose/Concentration: 120 \i\ (5, 50, or 500
pg/animal)
Time to Analysis: In-vivo: formation and
embolization of thrombus were continuously
monitored for 40 min. Ex-vivo: animals were
ITIy instilled w/ DEPs (0 or 50 pg per animal),
and blood was collected 5,15, 30, and 60 min
post-instillation. In-vitro: Saline or saline-
containing DEPs (0.1, 0.5,1, and 5 pg/mL) was
added to venous blood from untreated
hamsters, and closure time was measured in
thePFA-100after5min/animal.
Doses of 5-500 pg enhanced experimental
arterial and venous platelet-rich thrombus
formation in-vivo. Blood samples taken from
hamsters 30 and 60 min after instillation of 50 pg
of DEPs yielded accelerated aperture closure
(platelet activation) ex-vivo, when analyzed in the
PFA-100. The direct addition of as little as 0.5
pg/mL DEPs to untreated hamster blood
significantly shortened closure time in vitro.
Reference: Nemmar et DEP (SRM 1650); Positively Charged
al. (2004, 0879591 Polystyrene Particles (PCPSP)
Species: Hamster
Gender: Male and
Female
Weight: 100-110 g
Particle Size: PCPSP: 400 nm; DEP:
NR
Route: IT Instillation
Dose/Concentration: DEP: 50 pg/animal, or
PCPSP: 500 pg/animal
Time to Analysis: Pretreatment Phase:
Hamsters were pretreated w/ Dexametasone IP
(5 mg/kg) or IT (0.1 or 0.5 mg/kg) or Sodium
Cromoglycate given IP (40 mg/kg), 1 h before
DEP or vehicle instillation. Thrombosis: In-vivo
thrombogenesis assessed 24 h post-instillation
of DEP or vehicle.
DEP increased thrombosis without elevating
plasma vWF The IT instillation of PCPSP equally
produced histamine release and enhanced
thrombosis. Histamine in plasma resulted from
basophil activation. IP pretreatment with
Dexametasone abolished the DEP-induced
histamine increase in BALF and plasma and
abrogated airway inflammation and
thrombogenicity The IT pretreatment with
Dexametasone showed a partial but parallel
inhibition of all these parameters. Pretreatment
with Sodium Cromoglycate strongly inhibited
thrombogenicity, and histamine release.
Reference: Nemmar et Ultrafine Particles: Unmodified
al. (2003, 0974871 Polystyrene Particles (UPSPs); Nega-
tively Charged Carboxylate-Modified
Species: Hamster Polystyrene Particles (NCC-MPSPs);
Positively-Charged Amine Modified
Polystyrene Particles (PCA-MPSPs)
Route: IT Instillation Unmodified and negative UFPs did not modify
thrombosis. Positive UFPs increased thrombosis
Dose/Concentration: 5, 50, and 500 pg/animal at 500 and 50 pg/animal, but not at 5 pg/animal.
Gender: Male and
Female
in 120 pi saline
Time to Analysis: 1 h post-instillation
Weight: 100-110 g
Particle Size: UPSPs: 60 nm; NCC-
MPSPs: 60 nm; PCA-MPSPs: 60 or
400 nm
Positive 400 nm particles (500 pg/animal) did not
affect thrombosis. PFA-100 analysis showed that
platelets were activated by the in-vitro addition of
positive UFPs and 400 nm particles to blood.
December 2009
D-17
-------
Study
Pollutant
Exposure
Effects
Reference: Nemmar et DEP (SRM 1650)
al. (2003, 0879311
Species: Hamster
Weight: 100-110 g
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 50 pg/animal in 120 |jl
saline
Time to Analysis: 1, 3, 6, and 24 h
At 1, 6, and 24 h after instillation of 50 pg DEPs,
the mean size of in-vivo induced and quantified
venous thrombosis was increased by 480, 770,
and 460%, respectively. Platelets activation in
blood was confirmed by a shortened closure time
in the PFA-100 analyzer. In plasma, histamine
was increased only at 6 and 24 h. Pre-treatment
with a H1 receptor antagonist (diphenhydramine,
30 mg/kg intraperitoneally) did not affect DEP-
induced thrombosis or platelet activation at 1 h;
however both were markedly reduced at 6 and
24 h.
Reference: Niwa et al.
(2007, 0913091
Species: Mouse
Gender: Male
Strain: LDLr/KO
Age,:6wk(n = 20)
Use: IT CB dispersion;
10-14 wk acute effect
of CB dispersion on
circulating CRP
Carbon Black
Particle Size: 23-470 nm (mean size
120.7 nm)
Route: IT Dispersion
Dose/Concentration: IT CB Dispersion Study:
1 mg per animal/wk; Acute Effect of CB
Dispersion on Circulating CRP Study:
1mg/animal (single administration)
Time to Analysis: IT CB Dispersion Study:
1x/wkfor 10wk
Acute Effect of CB Dispersion on Circulating
CRP Study: Single CB administration, blood
samples collected 24 h post-administration
IT CB Dispersion Study: Although no difference
in body weight (bw) between the four groups was
observed at baseline, and all mice experienced
an increase in bw with advancing age, the mice
treated with CB tended to be smaller than those
treated with vehicle (air). No significant
differences were observed in cholesterol and TG
levels among the for groups. Development of
aortic lipid-rich lesions occurred in mice under a
0.51 % cholesterol diet with or without CB
infusion, but not in the mice fed a 0% cholesterol
diet.
Acute Effect of CB Dispersion on Circulating
CRP Study: Circulating levels of CRP were
significantly higher in mice exposed to CB versus
those exposed to air, indicating an acute
inflammatory response. Although the presence of
CB in pulmonary macrophage-like cells in CB
treated mice under 0.51% cholesterol diet was
confirmed, CB was not detected in aortas, livers,
kidneys, or spleens.
Reference: Niwa et al.
(2007, 0913091
Species: Mouse
Cell Types: RAW264.7
Carbon Black (CB); Water-Soluble
Fullerene
(C60(OH)24); Fluoresbrite Carboxylate
Microspheres; Ox-LDL; Acetylated-LDL
Particle Size: Carbon Black and
C60(OH)24:7.1 nm(SD2.4);
Fluoresbrite Carboxylate
Microspheres: 6 nm
Route: Cell Culture
Dose/Concentration: CB: 1,10,100 pg/mL;
C6o(OH)24: 20,100ng/mL
Time to Analysis: RAW264.7+CB
for 24 h, 13 days, and 50 days;
RAW264.7+ C60(OH)24for 24 h or 10 days;
RAW264.7+ C60(OH)24for 8 days, then co-
treated w/ Ox-LDL for an additional 48 h;
RAW264.7+Ox-LDL for 5 days, and then co-
cultured w/ C60(OH)24 for an additional 48 h;
RAW264.7+ 6 nm beads: 3 days, the Ox-LDL or
acetylated-LDL added for 24 h
CB alone had no significant effects on RAW264.7
cell growth. C60(OH)24 alone or CB and C60(OH)24
together w/ Ox-LDL induced cytotoxic
morphological changes, such as Ox-LDL uptake-
induced foam cell-like formation and decreased
cell growth, in a dose-dependent manner.
C60(OH)24 induced LOX-1 protein expression,
pro-matrix metalloprotease-9 protein secretion,
and tissue factor mRNA expression in lipid-laden
macrophages. Although CB or C60(OH)24 alone
did not induce platelet aggregation, C60(OH)24
facilitated ADP-induced platelet aggregation.
C60(OH)24 also acted as a competitive inhibitor of
ADP receptor antagonists in ADP-mediated
platelet aggregation.
Reference: Niwa et al. CB from Kyoto, Japan
(2008, 1568121
Species: Rat
Strain: SD
Age: 6 wk
Route: Whole-body Inhalation
Particle Size: Mean size (nm) + SD Dose/Concentration: 15.6 + 3.5 mg/m3
determined at 1, 8,15, 22, and 29 day
post-exposure was 1181+24 1191+ Timeto Analysis: 6 h/day, 5 days/wk, for a
Although the presence of CB was confirmed in
pulmonary macrophages, electron microscopic
survey did not detect CB in other tissues
including, liver, spleen and aorta. CB exposure
+ 3.6 respectively
cuff plethysmography at 1,14,
-exposure. Sacrificed At 1, 7,14, 28, and 30
day post-exposure
inflammatory marker proteins, including
monocyte chemo attractant protein-1, IL-6, and
CRP, were higher in the CB treated groups than
in control groups.
Reference: Nurkiewicz ROFA (from Everett, MA). Major metal
et al. (2004, 0879681 contaminants are: Fe, Al, V, Ni, Ca,
and Z. Main soluble metals are: Al, Ni,
Species: Rat and Ca.
Gender: Male
Strain: SD
Age: 7-8 wk
Particle Size: 2.2 pm (ROFA mean
count diameter)
Route: IT Instillation
Dose/Concentration: ROFAgroup: 0.1, 0.25,
1, or 2 mg/rat. Vehicle control group: 300 pi
saline. Particle control group: Ti02 0.25 mg/rat.
Saline Treated Rats: A23187 dilated arterioles
up to 72 + 7% max.
ROFA and Ti02 Exposed Rats: A23187-induced
dilation was significantly attenuated.
Time to Analysis: After single IT instillation of a Sensitivity of Arteriolar Smooth Muscle to
particular dose, all rats recovered for 24 h.
NO: Similar in saline treated and ROFA exposed
rats.
Other: Significant increase in venular leukocyte-
adhesion and rolling observed in ROFA exposed
rats.
December 2009
D-18
-------
Study
Pollutant
Exposure
Effects
Reference: Nurkiewicz
et al. (2006, 0886111
Species: Rat
Gender: Male
Strain: SD
Age: 7-8 wk
ROFA from Everett, MA
Particle Size: ROFA mean count
diameter: 2.2 pm; Ti02 mean diameter:
1.0 pm
Route: IT Instillation
Dose/Concentration: ROFA group: 0.1 or 0.25
mg/rat. Vehicle control group: 300 pi saline.
Particle control group: Ti02 0.1 or 0.25 mg/rat.
Time to Analysis: After single IT instillation of a
particular dose, all rats recovered for 24 h.
ROFA or TiO; Exposure and Arteriolar
Dilation: Exposure caused a dose-dependent
impairment of endothelium-dependent arteriolar
dilation.
ROFA or TiO Exposure and Arteriolar
Constriction: Exposure did not affect
microvascular constriction in response to PHE.
ROFA and TiO and Leukocyte Rolling and
Adhesion: Exposure significantly increased
leukocyte rolling and adhesion in aired venules,
and these cells were identified as PMN
leukocytes.
ROFA and Ti02 and MPO: MPO was found in
PMN leukocytes, adhering to the systemic mi-
crovascular wall. Evidence suggests that some of
this MPO had been deposited in the
microvascular wall. There was also evidence of
oxidative stress in the microvascular wall.
Reference: Nurkiewicz Ti02 (DeGussa, Sigma-Aldrich)
et al. (2008, 1568161
Particle Size: Fine-1 pm, UF-21 nm
Species: Rat
Gender: Male
Strain: SD
Age: 6-7wk
Weight: NR
Route: Whole-body Inhalation
Dose/Concentration: Concentrations: Fine- 3-
16 mg/m3; UF-1.5-12 mg/m3; Dose: Fine-8, 20,
36, 67, 90|jg;UF-4, 6, 10, 19, 30|jg
Time to Analysis: Acclimated 5 days. Exposed
4-12 h. Sacrificed 24 h post-exposure.
Particle accumulation within AMs, anuclear
macrophages, particle-laden AMs intimately
associated with the alveolar wall were all present
in exposed rats. Calcium ionophore impaired
arteriolar dilation in a dose-dependent manner in
UF and fine exposed rats. UF produced greater
systemic microvascular dysfunction.
Microvascular dysfunction was the same for
three groups of rats exposed to 30 pg UF Ti02
under different conditions.
Reference: Nurkiewicz
et al. (2009, 1919611
Species: Rat
Gender: Male
Strain: SD
Age: 7-8 wk
Weight: NR
Fine Ti02 (Sigma-Aldrich, St. Louis,
MO) (-99% rutile)
Ti02 nanoparticles (DeGussa-
Aeroxide Ti02 P25, Parsippany, NJ)
(80% anatase, 20% rutile)
Particle Size: Fine Ti02- MMAD: 402
nm, Primary size: <5 pm, ,CMD: 710
nm; Nano-Ti02- MMAD: 138 nm,
Primary size: 21 nm,, CMD: 100 nm
Route: Aerosol Inhalation
Dose/Concentration: 1.5-16mg/m3
Time to Analysis: Acclimated 5 days. Exposed
240-720 min. Anesthetized 24 h post-exposure.
Intravital microscopy, NO measurement,
microvascular oxidative stress measurement,
nitrotyrosine staining.
Arteriolar Dilation: Nano-Ti02 significantly
impaired endothelium-dependent arteriolar
dilation. Equivalent levels of arteriolar dysfunction
were found in fine and nano-Ti02. Arteriolar
dilation in response to abluminal
microiontophretic application of SNP was not
different from the controls or between the
exposure groups. Arteriolar dilation was partially
restored by radical scavenging with TEMPOL and
catalase, NADPH oxidase with apocynin, and
MPO inhibition with ABAH.
Microcirculation: ROS increased in both
groups. Nano-Ti02 significantly increased the
area of tissue containing nitrotyrosine in the lung
and spinotrapezius microcirculation.
NO: Fine and nano-Ti02 significantly and dose-
dependently decreased stimulated NO
production in isolated microvessels. NO
production was increased by radical scavenging
with TEMPOL and catalase or NADPH oxidase
with apocynin, and was largest in the fine Ti02
group.
December 2009
D-19
-------
Study
Pollutant
Exposure
Effects
Reference: Okayama
et al. (2006, 1568241
Species: Rat
Cell Type: Ventricular
Cardiac Myocytes from
Wistar Rats,
approximately 3 days
old
DEP (Tsukuba, Japan)
DEPE: 5g of DEP in 5 mL PBS
containing 0.05% Tween 80.
Others: Catalase, LDH, MPG and
SOD.
Particle Size: DEP mass median
diameter: 0.34 pm.
Route: In Vitro
Dose/Concentration: DEPE: 0-100 pg/mL;
MPG: 0-1 mM; SOD: 800 U/mL; Catalase: 500
U/mL
Time to Analysis: cells were incubated for 1, 2,
4, or 8, 24 or 48 h.
LDH Activity of Supernatant: 24 h post-DEPE
exposure.
SOD, Catalase, MPG on DEPE-induced
Toxicity: SOD, catalase or MPG was added to
cells w/ or w/o DEPE & incubated for 4 or 24 h.
Medium then replaced w/serum-free & cells
incubated for another 24 h to analysis.
Cytotoxic Effects of DEPE on Cardiac
Myocytes: DEPE above 20 pg/mL damaged
cardiac myocytes in a time and concentration-
dependent manner in both long- and short-term
exposure conditions. However damage was
greater after long-term exposure. LDH activity
showed a concentration-dependent increase at
higher levels of exposure (greater than 20
|jg/mL).
Effects of ROS Scavenging Enzymes and
Antioxidant on DEPE-induced Cell Damage:
SOD or catalase attenuated 50 pg/mL DEPE-
induced cell damage compared with DEPE-
treated groups lacking antioxidant enzymes. Co-
incubation with SOD and catalase showed more
protective effects towards DEPE-induced cell
damage, although these effects were not
statistically significant from cells treated with
SOD only. MPG attenuated 50 pg/mL DEPE-
induced cell damage in a concentration-
dependent manner in both long and short-term
exposure conditions. Especially in long-term
exposure MPG showed strong protective effects
against DEPE-induced cell damage. Cell viability
was not affected by SOD, catalase, or MPG.
Reference: Proctor et
al. (2006, 0884801
Species: Rat
Gender: Male
Age: 12 wk
Use: Thoracic Aorta
from cp/cp and +/?
Male Rats
cp/cp = homozygous
for cp gene. Prone to
obesity and insulin
resistant.
+/? = heterozygous for
either +/cp or +/+. Lean
and metabolically
normal.
ROFA from Birmingham, AL
Particle Size: 1.95 +0.18 fjm
(aerodynamic diameter)
Route: Protocol 1: Used two aorta rings per
each experimental treatment group (4 groups
total). Protocol 2: Used four rings.
Dose/Concentration: Protocol 1: exposed to
12.5 pg/mL ROFA-L (at 10 mg/mL).
Protocol 2: exposed to 1.56, 3.25, 6.26,12.5
pg/mL ROFA-L (at 10 mg/mL).
Time to Analysis: Protocol 1: Cells treated with
12.5 pg/mL ROFA-L and/or 104mol/L L-NAME
for20min
Protocol 2: Parameters measured after ROFA-L
only treatment
Contractile response to phenylephrine (PE) was
measured
ROFA-L (12.5 pg/mL) increased PE-mediated
contraction in obese, but not in lean rat aortae.
Effect was exacerbated by L-NAME, and it
reduced ACh-mediated relaxation in obese and
lean aortae. Initial exposure of aortae to ROFA-L
caused a small contractile response, which was
markedly greater on second exposure in the
obese aortae but marginal in lean.
Reference: Radomski
et al. (2005, 0913771
Species: Rat
Strain: Wistar Kyoto
Carbon Nano Particles (CNPs)
(purchased from SES Research,
Houston, TX): Multiplewall Nanotubes
MWNT); Single wall Nanotubes
SWNT); C60 Fullerenes (C60CS);
Mixed Carbon Nanoparticles (MCN)
PM:(SRM1648)(NIST)
Particle Size: CNPs: NR; PM: 1.4 pm
average size
Route: Simultaneous single PM injection into
femoral vein as FeCI3 injected to induce carotid
thrombosis
Dose/Concentration: 0.5 mL suspension of 50
|jg/mL of PM in 0.9% NaCI solution.
Time to Analysis: Blood flow continuously
monitored for 900 s.
Vascular Thrombosis: FeCI3 induced carotid ar-
tery thrombosis and MCN had an amplifying
effect in the development of thrombosis.
Infusions of MCN, SWNT, and MWNT signifi-
cantly accelerated the time and rate of
development of carotid artery thrombosis in rats.
SRM1648 was less effective than CNPs in
inducing thrombosis, while C60CS exerted no
significant effect on the development of vascular
thrombosis.
Reference: Radomski
et al. (2005, 0913771
Species: Human
Cell Types: Platelets
Use: Human platelet
aggregation
Carbon Nano Particles (CNPs)
(purchased from SES Research,
Houston, TX): Multiplewall Nanotubes
(MWNT); Singlewall Nanotubes
(SWNT); C60 Fullerenes (C60CS);
Mixed Carbon Nanoparticles (MCN);
PM (SRM1648)
Particle Size: CNPs: NR; PM: 1.4 pm
average size
Route: Cell Culture (2.5x108 platelets/mL)
Dose/Concentration: CNPs: 0.2-300 pg/mL;
PM: 5-300 pg/mL
Time to Analysis: Prostacyclin (PGI2), S-ni-
troso-glutathione (GSNO), aspirin, 2-methylthio-
AMP, phenanthroline, EDTA and Go6976 were
pre-incubated w/ platelets for 1 min before
particle addition. Particles added to platelets
and platelet aggregation studied for 8min.
Platelet Aggregation: All CNPs, except C60CS,
stimulated platelet aggregation (MCN 2
SWNT>MWNT>SRM1648). All particles resulted
in upregulation of GPIIb/llla in platelets. In
contrast, particles differentially affected the
release of platelet granules, as well as the
activity of thromboxane-, ADP, matrix
metalloproteinase- and protein kinase C-de-
pendent pathways of aggregation. Particle-
induced aggregation was inhibited by
prostacyclin and GSNO, but not by aspirin.
December 2009
D-20
-------
Study
Pollutant
Exposure
Effects
Reference: Reed et al.
(2006, 1560431
Species: Rat, Mouse
Gender: Male and
Female
Strain: CDF
(F344)/CrlBR (rat), SH
(rat), A/J (mouse), and
C57BL/6 (mouse)
Age: 6-12 wk
HWS (burned mix of hardwood in
noncertified wood stove using a
Pineridge model 27000, Heating and
Energy Systems, Inc. Clackamas, OR)
Measured Components: EC, OM, N03,
S04, NH4, metals
Particle Size: ~0.25 pm
Route: Whole-body Inhalation
Dose/Concentration: Low: 30 pg/m3
Mid-low: 100|jg/m3
Mid-high: 300 pg/m3
High: 1000 pg/m3
Time to Analysis: 6 h/day, /days /wk for 1 wk
or 6 mo. Immediate post-exposure analysis.
Organ Weights: Liver declined in rats of both
genders at 1 wk and female rats at 6 mo. Lung
volume increased and lung weight decreased in
female rats at 6 mo. Spleen weight increased in
female mice and rats at 1wk. Thymus weight
decreased in male rats at 1wk.
Clinical Chemistry: Cholesterol decreased at
the high dose for male rats at 1wk and 6 mo and
increased at mid-low and mid-high doses for
female rats at 6 mo. ALP decreased for rats of
both genders at 1wk and 6 mo for mid-low, mid-
high and high dose levels (14-38%). AST
decreased by 24% in male rats at Iwkwith high
dose. No effect on females. Creatinine serum
levels decreased in males at 1wkat mid-high and
high dose by 13%. No effect observed at 6 mo.
BUN/Cre ratio decreased in females at 1wk
(25%) and both genders at 6 mo at mid-high and
high dose (18-19%).
Hematology: Hemoglobin and hematocrit
increased in 6 mo male rats. Bilirubin increased
in female rats at 6 mo at high dose. Platelets
increased for male and female rats at 1wk (21%,
19% respectively). No effect observed at 6m.
WBC increased in males at 1wk.
Reference: Reed et al.
(2004, 0556251
Species: Rat, Mouse
Gender: Male and
Female
Strain: CDF
F344)/CrlBR (rat), A/J
mouse)
e:12wk
DE: generated from two 2000 model
5.9 L Cummins ISM turbo diesel
engines
Co-exposure to 8 gas and 8 solid
exhaust components measured
Particle Size: 0.10-0.15pm
Route: Whole-body Inhalation
Dose/Concentration: Low: 30 pg/m3
Mid-low: 100|jg/m3
Mid-high: 300 pg/m3
High: 1000 pg/m3
Time to Analysis: 6 h/day, 7 days/wk for 1wk
or 6 mo. Analyzed 1 day post-exposure.
Organ Weights: Kidney weight increased after
6m for both males and female rats at the high
dose. Kidney and liver weight increased for
female mice at all dose levels at 6 mo. Lung
weight increased at high dose at 6mo for female
mice and male rats. Spleen weight decreased in
male mice at the low and mid-high levels.
Clinical Chemistry: There was a massive
decrease in cholesterol (24%) for rats of both
genders after 1 wk and a smaller decrease for
male rats at 6 mo. GGT significantly increased at
6 mo for male and female rats at the mid-high
and high dose. ALP increased in male rats at 1
wk by 10%. AST decreased at mid-high (15%)
and high dose in female rats at 6 mo. BUN and
BUN/Creatine declined (19%, 17%) in female rats
at mid-high and high doses after 6 mo. BUN
increased by 21% at mid-low, mid-high and high
doses in male rats at 1wk.
Hematology: WBC decreased in high females
after 6 mo. Factor VII (blood clotting) decreased
in MH and HR males after 1wk and male and
female HR after 6 mo. Thrombin-antithrombin
complex declined massively but only in males
after 1wk.
Reference: Reed et al.
(2008, 1569031
Species: Rat
Gender: Male, Female
Strain: CDF
(F344)/CrlBR, SH
Age: NR
Weight: NR
GEE (two 1996 General Motors 4.3-L
V-6 engines; regular, unleaded, non-
oxygenated, non-reformulated
Chevron-Phillips gasoline, U.S.
average consumption for summer
2001 and winter 2001-2002)
Particle Size: 150 nm(MMAD)
Route: Whole-body Inhalation
Dose/Concentration: PM: Low- 6.6 + 3.7
pg/m3, Medium-30.3 + 11.8|jg/m3, High-59.1 +
28.3 pg/m3
Time to Analysis: 2 wk quarantine period in
chamber. Exposed 6 h/day, 7 days/wk, 3 day-6
mo. SH- surgery to implant telemeter in
peritoneal cavity. 4 wk recovery. ECG data
obtained every 15 min beginning 3 day pre-
exposure, 7 day exposure, 4 day post-
exposure.
Organ Weight: At 6 mo exposure, the heart
weights of male and female rats increased and
male rats' seminal vesicle weight decreased.
Clinical Chemistry: Serum alanine
aminotransferase, aspartate aminotransferase,
and phosphorus decreased in medium and high-
exposure females.
Hematology: Hematocrit, red blood cell count,
and hemoglobin dose-dependently increased for
both genders at both time points. Plasma
fibrinogen increased at 1wk in males.
CV Effects in SH Rat: Lipid peroxides were
significantly increased in males in the high
exposure group. TAT complexes decreased in
females in the high exposure group.
Removal of Emission PM: The removal of
emission PM strongly linked PM to increased
seminal vesicle weight, red blood cell counts,
LDH, lipid peroxides, and methylation.
December 2009
D-21
-------
Study
Pollutant
Exposure
Effects
Reference: Rhoden et
al. (2005, 0878781
Species: Rat
Gender: Male
Strain: SD
Age: Adult
Weight: 300 g
Urban Ambient Particles (UAPs): SRM- Route: UAPs: IT Instillation. CAPs: Inhalation
1649; CAPs (Boston, MA) „ ln tt. ,,,„,„
Dose/Concentration: UAPs: 750 pg
Particle Size: NR suspended in 300 pi saline; CAPs: 700 + 180
pg/m3
Time to Analysis: UAPs: 30 min post-instil-
lation. CAPs: immediately after 5 h exposure
period
Oxidative Stress and HR Function: UAPs
instillation led to significant increases in heart
oxidants. HR increased immediately after
exposure and returned to basal levels over the
next 30 min. SDNN was unchanged immediately
after exposure, but significantly increased during
the recovery phase.
Role of ROS in Cardiac Response: Rats were
treated with 50 mg/kg MAC 1 h prior to UAPs in-
stillation or CAPs inhalation. MAC prevented
changes in heart rate and SDNN in UAPs-
exposed rats.
Role of the Autonomic Nervous System in
PM-induced Oxidative Stress: Rats were given
5 mg/kg atenolol, 0.30 mg/kg glycopyrrolate, or
saline immediately before CAPs exposure. Both
atenolol and glycopyrrolate effectively prevented
CAPS-induced cardiac oxidative stress.
Reference: Rivero et
al. (2005, 0886531
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Age: 3 mo
Weight: -250 g
PM2 5, collected from heavy traffic area
in Sao Paulo, Brazil. PM25 Composi-
tion (%): S (3.05), As (0.30), Br (0.21),
Cl (2.09), Co (2.65), Fe (2.67), La
(5.42), Mn (0.64), Sb (0.21), Sc (3.25),
Th(8.14)
Particle Size: PM25
Route: IT Instillation
Dose/Concentration: 100 or 500 pg of PM25.
Time to Analysis: 24 h post-instillation
Hematology: Total reticulocytes significantly
increased at both PM25 doses, while hematocrit
levels increased in the 500 pg group. Quan-
tification of segmented neutrophils and fibrinogen
levels showed a significant decrease, while
lymphocytes counting increased with 100 pg of
PM25.
Pulmonary Vasculature: Significant dose-de-
pendent decrease of intra-acinar pulmonary
arteriole lumen/wall ratio was observed in both
PM25 groups.
Wet-to Dry Weight Ratio: Significant increase in
heart wet-to-dry weight ratio was observed in the
500 pg group.
Reference: Rivero et
al. (2005, 0886591
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Age: 3 mo
Weight: -250 g
PM2 5, collected from heavy traffic area
in Sao Paulo, Brazil. PM25 Composi-
tion (%): S (3.05), As (0.30), Br (0.21),
Cl (2.09), Co (2.65), Fe (2.67),
La (5.42), Mn (0.64), Sb (0.21),
Sc (3.25), Th (8.14)
Particle Size: PM25
Route: IT Instillation
Dose/Concentration: 50 and 100 pg of PM25.
Time to Analysis: HR and SDNN were
assessed immediately b
60 min post-instillation.
< 3°and
HR decreased significantly with time, but no
significant effect of treatment or interaction
between time and treatment was observed. In
contrast, there was a significant SDNN
interaction between time and treatment. The
PM25 concentration of 50 and 100 pg.
Reference: Seagrave
et al. (2008, 1919901
Species: Rat
Gender: Male
Strain: SD
Age: 10-12 wk
Weight: 250-300 g
GEE (2 1996 General Motors 4.3-LV6
gasoline engines; conventional
Chevron Phillips gasoline, U.S.
average composition) (CO, NO, N02,
S02, THC) (PM25 composition- EC,
OC, S04, NH4, N03)
Simulated downwind coal emission
atmospheres (SDCAs) (fly ash, gas-
phase pollutants, sulfate aerosols, NO,
N02, S02)
Paved Road Dust (RD) (Los Angeles,
CA; New York City, NY; Atlanta, GA)
Particle Size: GEE: 150 nm (MMAD),
RD: 2.6+1.7 pm.SDCA: 0.1-1.Opm
Route: Nose-only Inhalation
Dose/Concentration: GEE: 60 pg/m3, SDCAs:
317-1072 pg/m3, RD: 306-954 pg/m3; GEE: 00-
104 ppm, NO-16.7 ppm, N02-1.1 ppm, S02-
1.0 ppm, THC- 12ppm; SDCAs: CO- <1 ppm,
NO- 0.19-0.62 ppm, N02- 0.10-0.37 ppm, S02-
0.07-0.24 ppm, THC- <1 ppm
Time to Analysis: 6 h exposure. Cannula
ligated into trachea and connected to rodent
ventilator. Thorax and abdomen opened.
GEE produced CL in the lungs, heart, and liver.
RD produced a significant effect in the heart at
the low dose. SDCAs had no effect on CL. RD
significantly increased the heart's oxidative
stress, as demonstrated by the TBARS levels..
Reference:
Simkhovich et al.
(2007, 0965941
Species: Rat
Gender: Female
Strain: Fischer 344 x
Brown Norway hybrid
Age: 4, 26 mo
Ultra Fine Particles (UFPs) isolated
from industrial diesel reference PM
2975
Particle Size: UFPs < 0.1 pm
Route: Heart Perfusion (ex-vivo)
Dose/Concentration: UFPs 12.5, 25, and 37.5
mg.
Time to Analysis: Hearts perfused w/ UFPs for
30 min and analysis conducted every 10 min.
Young adult and old hearts demonstrated equal
functional deterioration in response to direct
infusion of UFPs. Developed pressure in young
adult UFPs-treated hearts fell from 101 ± 4 to 68
+ 8 mmHg. In the old UFPs-treated hearts
developed pressure fell by 35%. Positive dP/dt
was equally affected in the young adult and old
UFPs-treated hearts and was decreased by 28%
in both groups.
December 2009
D-22
-------
Study
Pollutant
Exposure
Effects
Reference: Smith et al.
(2006,1108641
Species: Rat
Gender: Male
Strain: SD
Age: 8 wk
Weight: 260-270 g
CFA: Coal Fly Ash (400 MW, Wasatch
Plateau, Utah) (aerodynamic
separation)
Particle Size: 0.4-2.5pm
Route: Nose-only Inhalation
Dose/Concentration: 1400 pg/m3 PM25
including 600 pg/m3 PM,
Time to Analysis: 4 h/day for 3 consecutive
days. Parameters measured 18 or 36 h post-
exposure.
Hematology: Plasma protein increased at 18h.
Lymphocyte and hematocrit percentage
decreased at 36 h.
Reference: Stinn et al.
(2005, 0883071
Species: Rat
Gender: Male and
Female
Strain: Crl: (WIU BR
Age: 40 day
DE (generated from 1.6 L VWdiesel
under USFTP 72)
CO: 10, 37 ppm
C02: 2170, 6540 ppm
NO: 7.0, 22.8 ppm
NOX: 8.6, 28.3 ppm
S02: 0.83, 3.09 ppm
NH4: ND
Measured Major Components: NO,
S02,1-nitropyrene, Zi. 50% by DE
weight is EC.
Particle Size: 0.19-0.21 pm (MMAD)
Route: Nose-only Inhalation
Dose/Concentration: 3 and 10 mg/m3
Time to Analysis: 6 h/day, 7 day/wk for 24 mo;
6 mo post-exposure
Hematology: Erythrocyteswere unaffected (12,
24, 30) except in high dose females at 24 and 30
mo. Hemoglobin and hematocrit increased dose-
dependently with no gender differences.
Leukocytes increased in a dose- and time-
dependent manner.
Reference: Sun et al.
(2005, 0879521
Species: Mouse
Gender: Male
Strain: ApoE"'"
Age: 16 wk
CAPs: PM25 from Tuxedo, NY.
HFCD: High Fat Chow Diet
NCD: Normal Chow Diet
Particle Size: PM25
'm3; Daily
Route: Whole-body Inhalation
Dose/Concentration: PM25:
concentration: 10.6 (SD3.4) pg/m1 (mean)
Average exposure over 6 mo period: 15.2
pg/m3.
Time to Analysis: Study diets fed for at least
10 wk prior to exposure to PM2 5 or FA. Exposed
for 6 h/day, 5 days/wk for 6 mo. Sacrificed 15-
47 days after exposure.
Vasomotor Function: Mice fed HFCD and
exposed to PM25 demonstrated an increase in
the half-maximal dose for dilation to ACh with no
changes in peak relaxation compared to the mice
exposed to FA and fed HFCD and NCD.
Atherosclerosis Burden with PM25: In vivo MRI
imaging of atherosclerosis burden in the
abdominal aorta revealed significantly increased
plaque burden in the mice fed HFCD compared
with the mice fed NCD. Mean (SD) plaque areas
in the mice exposed to PM2 5 and fed HFCD vs.
mice exposed to FA and fed HFCD were 33 (10)
vs. 27 (13) units, respectively.
Plfe and Vascular Inflammation: A 2.6-fold
higher inducible NOS content was apparent in
the mice exposed to PM2 5 and fed HFCD com-
pared with the mice exposed to FA and fed HFCD
chow and a 4-fold increase in the mice exposed
to PM2 5 and fed NCD compared with the mice
exposed to FA and fed NCD.
Reference: Sun et al.
(2008, 1570331
Species: Mouse
Gender: Male
Strain: ApoE"'"
Age: 6 wk
CAPs PM25
Collected from Sterling Forest State
Park, Tuxedo NY (40 miles NW of
Manhattan)
Particle Size: PM25
Route: Whole-body Inhalation
Dose/Concentration: Average Concentration
of: 85 pg/m3 CAPs in chamber.
Average exposure over 6 mo was 15.2 pg/m3.
Time to Analysis: 6 h/day, 5 day/wk for 6 mo.
Mice received two different diets, high-fat chow
and normal-chow.
Macrophage and Tissue Factor Expression in
Aortic Segments: Tissue Factor (TF) expression
was noted predominantly in the extracellular
matrix surrounding macrophages, foam cell-rich
areas and around smooth muscle cells.
1. High-Fat Diet: Increased TF and increased
macrophage infiltration was observed in the
plaques of high-fat chow mice exposed to PM
compared to mice exposed to air and high fat
diet.
2. Normal Diet: PM-exposed mice saw an
increase in CD68 expression compared to air-
exposed. However TF expression was not
significantly different in PM exposed normal diet
mice compared to control normal diet mice.
December 2009
D-23
-------
Study
Pollutant
Exposure
Effects
Reference: Sun et al.
(2008, 1570331
Species: Human
Cell Lines: BEAS-2B;
Vascular Smooth
Muscle Cells (hSMCs);
and Monocytes (THP-
1)
Ambient Particles collected from
Sterling Forest State Park, Tuxedo, NY
(24h/dayfor4wk)
Particle Size: Particle size ranges: 1.
<0.18|jm
2.1.8-2.5pm or
3. 2.5-10pm
Route: In vitro
Dose/Concentration: 10-300 pg/ml
Time to Analysis: Doses were tested for
durations up to 24 h.
Dose durations tested for up to 24-h did not
indicate detectable effects on cell viability.
Effect of PM on TF Expression and Activity in
hSMCs: In the PM size range of 1-3 pm,
significant increases in TF expression was
observed at doses of 100 and 300 pm /ml. In the
<0.18 pm size range, significant increase in TF
expression was observed at all doses. The
particles with sizes 0.18 -1.0 pm did not induce
significant change in TF expression.
Effect of PM on TF Expression and Activity in
Monocyte Cells: TF protein expression
increased with O.18 pm and the 1-3 pm range
particles. Expression was increased in the 0.18-
1.0 pm particle range but it was limited compared
to the other PM size ranges. In general TF
expression was higher in monocytes than in
hSMCs cells, but not significantly.
Effect on TF Expression and Activity in
Bronchial Epithelial Cells: 100 pg/mLofthe 1-3
pm and O.18 pm particles significantly increased
TF expression.
TF mRNA Expression: TF mRNA was increased
rapidly within the first hour in response to SRM-
1694a PM. The lowest dose of SRM PM,0 pg/mL
induced highest levels of mRNA in hSMCs, no
further increase was observed at higher
concentrations.
Reference: Sun et al.
(2008, 1570321
Species: Rat
Gender: Male
Strain: SD
Age: 500-650 g
PM25orUFP
Particle Size: PM25; UFP: <0.1 pm
Route: Whole-body Inhalation
Dose/Concentration: Mean PM25
concentration: 79.1 ± 7.4 pg/m . Normalized
PM25 over 10wk period: 14.1 pg/m3.
Time to Analysis: 6 h/day, 5 day/wk random
exposure to PM2 5, UFP, or FA for a total of 10
wk. At the end of wk 9 exposure, rats were
infused w/ 0.75 mg/kg/day of All for 7 days.
PM25, UFP, or FA, continued during All infusion
period.
All = angiotensin II
Mean Arterial Pressure (MAP): After All
infusion, MAP was significantly higher in PM2 5 -
All vs. FA-AII group. Aortic Vasoconstriction to
PE was potentiated with exaggerated relaxation
to the Rho-kinase (ROCK) inhibitor Y-27632 and
increase in ROCK-1 mRNA levels in the PM25 -
All group. Superoxide production in the aorta was
increased in the PM25. All group compared to FA-
AII group, inhabitable by apocynin and L-NAME
with coordinate upregulation of NAD(P)H oxidase
subunits p22phox and p47phox and depletion of
tetrahydrobipterin.
Reference: Sun et al.
(2008, 1570321
Species: Rat
Gender: Male
Strain: SD
Age: 500-650 g
Cell Line: Primary Rat
Aortic Smooth Muscle
Cells (RASMCs)
PM25orUFP
Particle Size: PM25; UFP: <0.1 pm
Route: Cell Culture
Dose/Concentration: UFP, PM25:10 or 50
pg/mL;AII: 100nmol/L
Time to Analysis: Exposed to UFP or PM25
and parameters measured at 0,1, 3, 6, and 15
All = angiotensin I
Exposure to UFPs and PM25 was associated with
an increase in ROCK activity, phosphorylation of
myosin light chain, and MYPT1. Pretreatment
with N-Acetylcysteine and the Rho kinase
inhibitors (Fasudil and Y-27632) prevented MLC
and MYPT-1 phosphorylation by UFPs sug-
gesting a superoxide-mediated mechanism for
PM2 5 and UFPs effects.
December 2009
D-24
-------
Study
Pollutant
Exposure
Effects
Reference: Sun et al.
(2009, 1904871
Species: Mouse
Gender: Male
Strain: C57BL/6, c-
fmsYFP (transgenic,
yellow fluorescent
protein under
monocyte-specific
promoter)
Age: 8,10wk
Weight: NR
PM (concentrated- northeastern
regional background; Tuxedo Park,
NY)
Particle Size: 2.5 pm (diameter)
Route: Whole-body Inhalation. IT Instillation.
Dose/Concentration: Exposure chamber
(mean): 72.7 pg/m3, IT: 1.6mg/kg
Metabolic Impairment: PM induced insulin,
homeostasis model assessment indexes,
elevated glucose, and abnormalities in lipid
profile consistent with the IR phenotype.
Time to Analysis: C57BL/6 mice, fed high-fat Vascular Endothelium: PM decreased peak
chow 10wk. Exposed in vivo 6 h/day, 5 day, 128 relaxation and ED50 to ACH and peak relaxation
days. fmsYFP rendered diabetic or fed normal to insulin Lower levels of NO release were seen
chow 10 wk. IT instilled with PM 2 times/wk for
10 wk. Insulin Signaling: PM reduced the
phosphorylation of Akt in intact aorta. PKC-|311
was the only PKC isoform to increase.
Adipose Inflammation, Visceral Adiposity: PM
significantly increased TNF-a, IL-6, E-selectin,
ICAM-1, plasminogen activator inhibitor-1, and
restin. PM increased visceral and mesenteric fat
mass. F4/80+ macrophages in fat tissue and
adipocyte size increased. PM downregulated IL-
10 and glactose-N-acetylgalactosamine-specific
lectin.
YFP Cell Adhesion and Infiltration: PM
increased YFP cells in the adipose tissue, YFP
cell infiltration in the mesenteric fat, and YFP cell
adhesion to endothelium.
Reference: Tamagawa PM10 (urban; Ottawa, Canada)
et al. (2008, 1919881
Particle Size:: 0.8 + 0.4 pm (mean
Species: Rabbit diameter)
Gender: Female
Strain: New Zealand
White
Age: 12 wk
Weight: Acute
(average)-2.4 ±0.2 kg,
Chronic (average)- 2.7
± 0.3 kg
Route: Intrapharyngeal Instillation
Dose/Concentration: Acute- 2.6mg/kg,
Chronic- 2mg/kg
Inflammation: PM10 induced more
macrophages, AMs, positive and activated AMs,
and fewer tissue macrophages. NO, WBCand
PMN were only significantly higher in the first two
wk and IL-6 in the first wk.
Time to Analysis: Acute animals exposed days
1, 3, 5. Chronic animals exposed 2 times/wk for Vascular endothelial function: PM10
4wk.
significantly reduced Ach-stimulated relaxation
and did not alter SNP-stimulated relaxation. A
significant inverse relationship between IL-6 and
Ach-induced relaxation occurred at wk 1 in the
acute model and wks 1 and 2 in the chronic
model.
AMs: The chronic model had a significant
correlation between IL-6 and both positive and
activated AMs at wk 1. A significant inverse
relationship occurred between Ach and both the
volume fraction of positive and activated AMs.
Reference: Tankersley Carbon black (CB) (Wright dust feed
et al. (2008,1570431 particle generator-BGI, Waltham, MA)
Species: Mouse
Gender: Male
Strain: C57BL/6,
C3H/HeJ, B6C3F1
Age: 18, 28 mo
Weight: NR
Particle Size: 0.1-1. Opm
Route: Whole-body Inhalation
Dose/Concentration: Average PM25
concentration- 401 ± 46 pg/m , Average PMi0
concentration- 553 + 49 pg/m3
Time to Analysis: 3 h/day, 4 days
Hemodynamics: CB significantly elevated right
atrial and ventricular pressures, pulmonary
arterial pressure and vascular resistance, all of
which were more pronounced in the 28 mo-old
mice. RV contractility (specifically, the ejection
fraction and maximum change in pressure over
time) reduced in CB-exposed 28 mo-old mice.
Heart Tissue: CB significantly declined Ca2+-
dependent NOS activity and was more
pronounced in 28 mo-old mice, who also had
NOS2 upregulated. CB enhanced ROS
generation and NOS-uncoupling and was
greatest in 28 mo-old mice. CB also increased
MMP-2, MMP-9, ANP, BNP, which were greatest
in 28 mo-old mice. CB also reduced PKG-1 in 28
mo-old mice.
December 2009
D-25
-------
Study
Pollutant
Exposure
Effects
Reference: Tankersley Carbon Black (CB)
etal. (2007,091910) partic|eSjze: c& 2 4pm (MMAD)
Species: Mouse (GSD 2.75 pm).
Gender: Male
Strain: C3H/HeJ and
C57BL/6J
Age: 10 wk
Weight: 22-26 g
Route: CB: Whole-body Inhalation;
Sympathetic (S) & Parasympathetic (PS)
blockade: IP Injection
Dose/Concentration: CB: 159 + 12 pg/m3; PS
(atropine): 0.5 mg/kg; S(propanolol): 1 mg/kg
Time to Analysis: Successive 3 h CB and FA
Exposures: conducted from 9 a.m. to 1 p.m., or
at least 3 h after dark-to-light transition
(exposure period selected based on the nadir in
circadian pattern in HR responses).
Subgroups of both strains exposed to PS S S
blockade.
FA Exposure with Saline: A significantly greater
3 h average response occurred in C3 compared
with B6 mice.
PS Blockade: No evident strain difference
between C3 and B6 was observed.
S Blockade: 3 h average HR responses for C3
mice were significantly reduced compared with
saline.
CB Exposure: HR responses were significantly
elevated in C3 compared with B6 mice, but these
HR responses were not different relative to FA
exposure.
S Blockade: HR was significantly elevated in B6
mice during CB relative to FA, but was not
changed in C3 mice.
Reference: Tankersley Carbon Black (CB) and Filtered Air
et al. (2004, 0943781 (FA)
Species: Mice
Strain: AKR/J
Age:-180 days
Particle Size: CB: 0.1 to 1 urn.
Route: Whole-body Inhalation
Dose/Concentration: CB average
concentration: 160±22|jg/m3
Time to Analysis: FA exposure on day 1, CB
exposure 3 h/day for 3 consecutive days (days
2-4)
On day 1, HR was significantly depressed during
FA in terminally senescent mice. By day 4, HR
had significantly slowed due to the effects of 3
days CB exposure. The combined effects of
terminal senescence and CB exposure acted to
depress HR to an average (+ SEM) 445 + 40
bpm, ~ 80 bpm lower compared to healthy HR
responses. The change in rMSSD was
significantly greater on day 1 and day 4 in
terminally senescent mice, compared to healthy
mice. LF/HF ratio was significantly depressed in
terminally senescent mice on day 1. By day 4,
significant increases in LF/HF were evident in
healthy mice during CB exposure. Terminally
senescent mice modulated a lower HR without
change in the LH/HF ratio during CB exposure.
Reference: Thomson
et al. (2005, 0875541
Species: Rat
Gender: Male
Strain: F344
Weight: 200-250 g
Urban Ambient Particles (EHC-93)
from Ottawa, Canada; 03
Particle Size: Respirable Modes
(aerodynamic diameter): 1.3 and 3.6
pm. Non-respirable Mode
(aerodynamic diameter): 15 pm
Route: Nose-only Inhalation
Dose/Concentration: EHC-93: 0, 5, 50 mg/m3;
03: 0, 0.4, 0.8 ppm
Time to Analysis: 4 h to particles, 03, or
combination of particles and 03.
Both pollutants individually increased preproET-1,
ET-1 and endothelial NOS mRNA levels in the
lungs shortly after exposure, consistent w/ the
concomitant increase in plasma of ET-1 [1-21].
Prepro-ET1 mRNA remained elevated 24 h post-
exposure to particles but no after 03. Both
pollutants transiently increased ET-B receptor
mRNA expression, while 03 decreased ET-A
receptor mRNA levels. Coexposure to particles
plus 03 increased lung preproET-1 mRNA but not
plasma ET-1[1-21], suggesting alternative proc-
essing or degradations of endothelins. This
coincided w/ an increase of MMP-2 in the lungs
(this enzyme cleaves bigET-1 to ET-1 [1-32]).
Reference: Thomson
et al. (2006, 0974831
Species: Rat
Gender: Male
Strain: F344
Weight: 200-250 g
Urban Ambient Particles (EHC-93)
from Ottawa, Canada; 03
Particle Size: NR
Route: Nose-only Inhalation
Dose/Concentration: EHC-93: 0, 50 mg/m3;
03:0, 0.8 ppm
Time to Analysis: 4 h to particles, 03, or
combination of particles and 03. Sacrificed
immediately following exposure or following
24 h recovery.
Circulating levels of both ET-1[1-21] and ET-3[1-
21] were increased immediately after exposure to
PM and 03. While expression of preproET-1
mRNA in the lungs increased, expression of
preproET-3 mRNA decreased immediately after
exposure. PreproET-2 mRNA was not detected in
the lungs, and exposure to either pollutant did not
affect plasma ET-2 levels. Coexposure to 03 and
particles, while altering lung preproET-1 and
preproET-3 mRNA levels in a fashion similar to
03 alone, did not cause changes in the circulating
levels of the two corresponding peptides.
December 2009
D-26
-------
Study
Pollutant
Exposure
Effects
Reference: Totlandsdal
et al. (2008, 1570561
Species: Rat
Gender: Male
Strain: WKY/NCrl and
Crl: Wl (Han)
Age: Adult
Weight: Crl/WI, 250-
300 g
Use: Isolation of Rat
Ventricular
Cardiomyocytes and
Cardiofibroblasts
(RVCMsandRVCFBs)
Pigment Black Printex 90 (Frankfurt,
Germany); PM:SRM 1648
Particle Size: Printex 90:12-17 nm;
PM:NR
Route: Cell Culture
Dose/Concentration: Printex 90: 0, 50,100,
200 or 400 pg/mL; PM: 0, 200 pg/mL
Time to Analysis: 20 h
Cardiac Cell Cultures: IL-6 release was strongly
enhanced upon exposure to conditioned media,
and markedly exceeded the response to direct
particle exposure. IL-1, but not TNF-a, seemed
necessary, but not sufficient, for this enhanced
IL-6 release. The role of IL-1 was demonstrated
by use the use of an IL-1 receptor antagonist that
partially reduced the effect of the conditioned
media, and by a stimulating effect on the cardiac
cell release of IL-6 by exogenous addition of IL-1
a and IL-113.
Reference: Tzeng et
al. (2007, 0978831
Species: Rat
Strain: Wistar Kyoto
Cell Type: Primary
Vascular Smooth
Muscle Cell Culture
(VSMCs): isolated from
thoracic aortas from
200-250 g rats.
Motorcycle Exhaust Particulate Extract Route: In vitro
(MEPE) collected from a Yamaha
motorcycle with a 50 cm3 two-stroke
engine using 95% octane unleaded
gasoline.
Particle Size: PM,, PM25, PM10
Dose/Concentration: 10-100 pg/mL
Time to Analysis: 3 days
Exposure of VSMCs to MEPE (10-100 pg/mL),
enhanced serum-induced VSMC proliferation.
The expression of proliferating cell antinuclear
antigen was also enhanced in the presence of
MEPE. VSCMs treated with MEPE induced
increase COX-2 mRNA, protein expression, and
PGE2 production, whereas the level of COX-1
protein was unchanged. MEPE increased the
production of ROS in VSMCs, in a dose-
dependent manner. MEPE triggered time-
dependent ERK1/2 phosphorylation in VSMCs
which was attenuate by antioxidants (MAC,
PDTC). The level of translocation of NF-KB-p65
in the nuclei of VSMCs was also increased
during MEPE exposure. The potentiating effect of
MEPE in serum-induced VSMC proliferation was
abolished by COX-2 selective inhibitor NS-398,
specific ERK inhibitor PD98059, and antioxidants
(MAC, PTDC).
Reference: Tzeng et
al. (2003, 0972471
Species: Rat
Strain: Wistar Kyoto
Cell Type: Primary
Vascular Smooth
Muscle Cell Culture
(VSMCs)
Motorcycle Exhaust Particulate Extract Route: In vitro
(MEPE) collected from a Yamaha
motorcycle with a 50 cm3 two-stroke
engine using 95% octane unleaded
gasoline.
Particle Size: NR
Dose/Concentration: MEPE: 10 pg/mL;
Nifedipine: lOpmol; Manganese Acetate: 100
pmol; Staurosporine: 1-2 nM; Chelerythrine: 1
pm
Time to Analysis: 18 h
MEPE induced a concentration-dependent
enhancement of vasoconstriction elicited by
phenylephrine in the organ cultures of intact and
endothelium-denuded aortas for 18h. Nifedipine,
manganese acetate, and Staurosporine, but not
Chelerythrine, inhibited the enhancement of
vasoconstriction by MEPE. ML-9 inhibited the
enhancement of vasoconstriction by MEPE.
MEPE enhanced the phosphorylation of 20k-Da
in rat vascular smooth muscle cells. N-
acetylcysteine significantly inhibited the
enhancement of vasoconstriction by MEPE. A
time-dependent increase in ROS production by
MEPE was also detected in primary cultures of
VSMCs.
Reference: Upadhay
et al. (2008, 1593451
Species: Rat
Gender: Male
Strain: SH
Age: 6 mo
Weight: NR
Ultrafine Carbon Particles (UFCP)
Particle Size: Size- 31 + 0.3 nm,
MMAD- 46 nm, Surface area
concentration-0.139 m particles/m ,
Mass specific surface area- 807m2/g
Route: Whole-body Inhalation
Dose/Concentration: 172 pg/m3
Time to Analysis: Acclimatized 2 day. 1 day
baseline. 24 h exposure. 4 recovery. Sacrificed
1st or 3rd day of recovery.
Cardiophysiology: The mean arterial BP and
HR increased but returned to baseline levels by
the 4th recovery day. SDNN and HRV decreased.
RMSSD and LF/HF decreased but were not
significant.
Pulmonary Inflammation: UFCP did not cause
pulmonary inflammation.
Pulmonary and Cardiac Tissue: HO-1, ET-1,
ETA, ETB, TF, PAI-1 significantly increased in the
lung on the 3rd recovery day. HO-1 was
repressed in the heart, but the other markers had
slight, nonsignificant increases.
Systemic Responses: Neutrophil and
lymphocyte cell differentials significantly
increased on the 1st recovery day. Other blood
parameters were unaffected. The plasma renin
concentration increased on the first 2 recovery
days. Ang I and II concentrations increased on
the 1st recovery day but was not significant.
December 2009
D-27
-------
Study
Pollutant
Exposure
Effects
Reference: Wallenborn Zinc Sulfate (ZnS04, aerosolized)
et al. (2008,1911711
v Particle Size: NR
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Age: 13 wk
Weight: NR
Route: Nose-only Inhalation
Dose/Concentration: 9.0 + 2.1 pg/m3, 35 + 8.1
pg/m3,123.2 + 29.6 pg/m3
Time to Analysis: Exposed 5 h/day, 3 days/wk,
16wk. Half of the rats used for plasma/serum
analysis, other half for isolation of cardiac
mitochondria.
A trend toward increased BALF protein was
seen. Cardiac mitochondrial ferritin had a small,
significant increase. Mitochondrial succinate
dehydrogenase and glutathione peroxidase had
small, significant decreases. Subchronic
exposure to 100 pg/m3 caused expression
changes of cardiac genes involved with cell
signaling events, ion channels regulation, and
coagulation. No pulmonary-related effects were
Reference: Wallenborn PM: precipitator unit power plant
et al. (2007,1561441 residual oil combustion
Species: Rat
Gender: Male
Strain: WKY, SH, and
stroke-prone SH
(SHRSP)
Age: 12-15 wk
Particle Size: PM: 3.76pm (bulk) ±
2.15
Route: IT Instillation
Dose/Concentration: WKY vs SHRSP: 1.11,
3.33, 8.33 mg/kg
SH vs SHRSP: 3.33, 8.33 mg/kg
Time to Analysis: Single, 24 h
Note: 4 h post-exposure study done on WKY vs
SHRSP but not published.
Oxidative Stress - Cardiac: SOD increased in
the SHRSP vs WKY experiment only. Only
SHRSP at 8.33 mg/kg showed a significant
increase when compared to the control.
GPx: No action but SHRSP levels were similar to
SHR and, in the WKY vs SHRSP experiment,
SHRSP exhibited higher activity level than WKY.
Ferritin: Equivocal results were observed. Levels
decreased at the high dose for WKY and SHRSP
but increased at medium doses for SH and
SHRSP.
ICDH: Levels increased for WKY and decreased
for SHRSP.
Reference: Wellenius CAPs
et al. (2003, 055691
Species: Dog
Gender: Female
Strain: Mixed mongrel
Age: NR
Weight: 14-17 kg
Particle Size: 0.26 + 0.04 pm
Route: Permanent Tracheostomy
Dose/Concentration: Median: 285.7 pg/m3,
Range: 161.3-957.3 pg/m3
Time to Analysis: Thoracotomy and
tracheostomy performed. 5-13 wk recovery.
Pairs of subjects: exposed 6 h/day either 2nd or
3rd exposure time and filtered air other days. 5
min preconditioning occlusion. 20 min rest
interval. 5 min experimental occlusion. Some
dogs exposed 6 h/d, 4 days (consecutive),
filtered air on day 4.
CAPs increased the ST-segment elevation and
remained elevated 24 h after exposure. This
increase was seen in precordial leads V4 and V5.
Multivariate regression analyses showed that the
mass concentration of Si was significantly
associated with the peak ST-segment elevation
and integrated ST-segment change. Univariate
regression analyses showed Pb to also be
significantly associated with these measures.
CAPs had no effect on peak heart rate during
occlusion or the maximum occlusion-induced
increase in heart rate.
Reference: Wellenius CAPs (Boston, MA); exposures during
et al. (2004, 0878741 the period of 07/2000 and 01/2003.
Species: Rat
Gender: Male
Strain: SD
Age: Adult
Weight: -250 g
Use: Rat Model for
Acute Myocardial
Infarction (AMI): Left-
ventricular Ml induced.
Animals allowed to
recover for at least 12 h
after surgery.
CO
Particle Size: PM25
Route: Whole-body Inhalation
Dose/Concentration: CO: 35ppm; CAPs
(median concentration): 350.5 pg.m3;
CAPs+CO: (CAPs median concentration):
318.2 pg/m3
Time to Analysis: 1 h exposure to CAPs or
CAPs+CO for 1 h. Exposure to pollutants was
preceded and followed by 1 h exposure to FA.
CO exposure reduced the ventricular premature
beat (VPB) frequency by 60.4% during the
exposure time compared to controls. This effect
was modified by both infarct type and the number
of pre-exposure VPBs, and was mediated
through changes in HR. Overall, CAPs exposure
increased VPB frequency during the exposure
period, but this did not reach statistical
significance. This effect was modified by the
number of pre-exposure VPBs. In rats with a high
number of pre-exposure VPB, CAPS exposure
significantly decreased VPB frequency (67.1%).
Overall, neither CAPs nor CO had any effect on
HR, but CAPs increased HR in specific
subgroups. No significant interactions were
observed between the effects of CO and CAPs.
December 2009
D-28
-------
Study
Pollutant
Exposure
Effects
Reference: Wellenius
et al. (2006, 1561521
Species: Rat
Gender: Male
Strain: SD
Age: Adult
Weight: -250 g
Use: Rat Model for
Acute Myocardial
Infarction (AMI): Left-
ventricular Ml induced.
Animals allowed to
recover for at least 12 h
after surgery.
CAPs: (Boston, MA)
Particle Size: PM25
Route: Whole-body Inhalation
Dose/Concentration: CO: 35 ppm; CAPs
(median concentration): 645.7 pg.m3;
CAPs+CO: 37.9 ppm
Time to Analysis: CAPs or CAPs+CO
exposure for 1 h. Exposure to pollutants was
preceded and followed by 1 h exposure to FA.
Among rats in the CAPs group, the probability of
observing supraventricular arrhythmias (SVA) de-
creased from the baseline to exposure and post-
exposure periods. The pattern was significantly
different than that observed for the FA group
during the exposure period. In the subset with
one or more SVA during the baseline period, the
change in SVA rate from baseline to exposure
period was significantly lower in the CAPs and
CO groups only, when compared to the FA group.
No significant effects were observed in the group
simultaneously exposed to CAPs and CO.
Reference: Wichers et
al. (2004, 0556361
Species: Rat
Gender: Male
Strain: SH
Age: 75 day
HP-12 (oil-combustion derived PM
obtained from inside wall of a Boston
power plant stack burning residual oil
number 6).
Water-leachable constituents (pg/mg):
S04 (217.3); Zn (11.4); Ni (6.9); Fe
(0.0);V(1.3);Cu(0.2);Pb(0.0)
1M HCI-leachable constituents
(fjg/mg): S04 (220.6); Zn (15.5); Ni
14.8); Fe (15.6); V (32.9); Cu(1.1);Pb
1.7)
Particle Size: 3.76 pm (MMAD) (GSD
2.16)
Route: IT Instillation
Dose/Concentration: HP-12 (mg/kg): 0.00
saline control), 0.83 (low), 3.33 (mid), 8.33
high)
Time to Analysis: 96 h or 192 h post-
instillation.
Exposures to mid and high-dose HP-12 induced
large decreases in HR, BP, and body
temperature. The decreases in HR and BP were
most pronounced at night and did not return to
pre-instillation values until 72 h (HR) and 48 h
(BP) after dosing. ECG abnormalities (rhythm
disturbances, bundle branch block) were
observed primarily in the high dose group.
Reference: Wold et al.
(2006, 0970281
Species: Rat
Gender: Female
Strain: SD
Use: Left jugular vein
and right carotid artery
were cannulated.
UFPs from either ambient air (UFAAs)
or diesel engine exhaust (UFDGs);
UFIDs from industrial forklift exhaust
and soluble fraction UFID suspension,
particle free (SF-UFID)
Particle Size: UFAAs diameter < 150
nm; UFDGs diameters 100 nm
Route: IV Infusion
Dose/Concentration: UFDG (50 pg/m)
Time to Analysis: Infused w/UFAA or UFDG
Monitored continuously for 1 h then sacrificed.
Infusion of UFDGs caused ventricular premature
beats (VPBs) in 2 out of 3 rats. Ejection fraction
increased slightly in rats receiving UFAA and was
unchanged in the UFDG and saline groups.
Reference: Wold et al.
(2006, 0970281
Species: Rat
Gender: Female
Strain: SD
UFPs from either ambient air (UFAAs)
or diesel engine exhaust (UFDGs);
UFIDs from industrial forklift exhaust
and soluble fraction UFID suspension,
particle free (SF-UFID)
Particle Size: UFAAs diameter <150
nm; UFDGs diameters 100 nm
Route: Lagendorff Heart Perfusion
Dose/Concentration: UFDG (100 pg/2ml);
UFID (12.5 pg/l in perfusate); SF-UFID (12.5
Time to Analysis: Lagendorff 1: Treated
w/UFDG Lagendorff 2: Treated with UFID &
SFUFID. Both experiments were monitored
continuously for 1 h after injection.
UFDGs caused a marked increase in left-
ventricular and end-diastolic pressure (LVEDP)
after 30 min of exposure. UFIDs caused a
significant decrease in left-ventricular systolic
pressure (LVSP) at SOmin after the start of
infusion. This effect was absent when SF-UFID
was studied.
Reference: Yatera et
al. (2008, 1571621
Species: Rabbit
Gender: Female
Strain: WHHL
Age: 42 wk
Weight: 3.2 ±0.1 kg
(avg)
EHC-93 from Ottawa, Canada
Particle Size: NR
Route: IT Instillation
Dose/Concentration: PM10 suspension: 5 mg
EHC-93 in 1 ml saline
Time to Analysis: Exposed 2 times/wk for 4
wk. Acute effects observed at 0.5,1, 2, 4, 8,12,
and 24 h after initial instillation. Subchronic
effects observed once/wk for 4 wk.
Exposure to PM10 caused progression of
atherosclerotic lesions in thoracic and abdominal
aorta. It also decreased circulating monocytes
expressing high levels of CD31 and CD49day,
and increased expression of CD54 (ICAM-1) and
CD106 (VCAM-1) in plaques. Exposure to PM,0
increased the number of BrdU-labeled (*) mono-
cytes into plaques and into smooth muscle
underneath plaques.
December 2009
D-29
-------
Study
Pollutant
Exposure
Effects
Reference: Ying et al.
(2009,1901111
Species: Mice
Gender: Male
Strain: ApoE"'"
Age: 16 wk
CAPs:,New York City (Manhattan), NY; Route: Whole-body Inhalation
May-Sept 2007
Dose/Concentration: 138.4 + 83.7 pg/m
Particle Size: PM25
Time to Analysis: 6 h/day, 5 day wk, 4 mo
Vascular Tone: Significant decrease in PE-
induced maximum contraction of aortic rings in
CAPs-exposed mice. No difference in sensitivity
to PE between groups. Treatment with the
soluble guanylate cyclase inhibitor ODQ restored
the response to PE in CAPs aortic rings. No
significant differences in relaxation induced by
ACh. CAPs abolished the relaxation induced by
Ca ionophore A23187. CAPs exposure slightly
(but significantly) decreased maximum relaxation
induced by SNP
Protein Expression: iNOS mRNA expression
was increased in the aortas of CAPs-exposed
mice. eNOS and GTPCH levels were unchanged.
Distribution of inOS protein expression was
limited to plaque in air-exposed mice and was
found in the plaque and media for CAPs-exposed
Superoxide Production: Superoxide levels in
CAPs-exposed mice were increased in the aorta
compared to air-exposed mice. The addition of L-
NAME significantly increased superoxide
production. Extensive protein nitration in aortas of
CAPs mice. NADPH subunits Pad and p47 phox
mRNA expression was increased in aortas of
mice exposed to CAPs.
Atherosclerosis: Significant increase in plaque
area of CAPs-exposed mice. Higher levels of
macrophage infiltration, collagen deposition, and
lipid composition of plaques from CAPs-exposed
mice.
Reference: Yokota et
al. (2004, 0965161
Species: Rat
Gender: Male
Strain: SD
Weight: 345-498.2 g
DEP (obtained from the Japan
Automobile Research Institute)
Particle Size: NR
Route: IT Instillation
Dose/Concentration: Group 1: DEP: 1 mg/0.1
ml; Group 2: DEP: 0.2 ml (10,12.5 or 25
mg/ml); Group 3: DEP 2.5 or 5 mg/0.2 ml
Time to Analysis: DEP pre-treatment 24-72 h
before ischemia/reperfusion.
DEP Effects on Mmyocardial
Ischemia/Reperfusion-induced Arrhythmia:
An increased mortality was observed in the DEP
group compared to the vehicle-treated group.
46% of the animals in DEP died during the first 3
min reperfusion period. The animals of other
groups were intratracheally instilled with DEP at
the beginning of ischemia/reperfusion experi-
ment, or were pretreated with polyethylene
glycol-conjugated SOD (1000 lU/kg, iv). In these
animals, incidences of both arrhythmia and
mortality were similar to those in the animals
treated with the vehicle.
DEP Rffects on the Biochemical and
Hematological Parameters: Neutrophil count
was elevated by a higher dose (5 mg) of DEP at
24 h after the IT instillation, and oxygen radical
production, which was induced by 12-0-
tetradecanoylphorbol 13-acetate, was enhanced
at 72 h.
Reference: Yokota et
al. (2005, 0960031
Species: Rat
Gender: Male
Strain: SD
Weight: 303-472.2 g
DEP from Japan
Particle Size: NR
Route: IT Instillation
Dose/Concentration: DEP: 5 mg/animal
Time to Analysis: Single exposure 0.5,1, 2, 3,
6, 12, 24, 48 h.
At 12 and 24 h post-instillation, circulatory
neutrophil counts in the 5 mg DEP group were
significantly elevated, and were 2.1-fold (12 h)
and 2.3 fold (24 h) in vehicle treated animals. 1
mg DEP caused an increase of approximately
0.4-fold in CNC at 6 h. 12-0-
tetradecanoylphorbol 13-acetate induced
oxyradical production (ORP) in the isolated
neutrophil was enhanced at 12 and 24 h after
instillation with 5 mg DEP. In Serum, a marked
elevation of CINC-1 and a slight elevation of MIP-
2 were also observed, while TNF-a was not
detected. GM-CSFwas not detected in serum 24
h post-instillation.
December 2009
D-30
-------
Study
Pollutant
Exposure
Effects
Reference: Yokota et
al. (2008, 1901091
Species: Mouse
Gender: Male
Strain: ddy
Age: NR
Weight: 39.6-46.0 g
DEP (DMSC (dichloromethane
soluble-component), RPC (residual
particle-component))
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 5 mg/kg, 10 mg/kg
Time to Analysis: DMSC and RPC extracted
from DEP. Mice acclimatized 7 day
. DEP, DMSC, or RPC instilled. BALFand blood
obtained and G-CSF, GM-CSF, IL-6 measured
2,4,12, 24 h post-instillation.
Inflammation: At 5 mg/kg DEP increased the
total cell and macrophage count. DEP or RPC
increased neutrophils at 5 and 10 mg/kg. 10
mg/kg DEP or RPC increased macrophages at
4h and decreased at 12 h.
RPC or DEP caused sustained increases in RBC,
WBC, and neutrophils.
Cytokines: 5 mg/kg RPC markedly increased G-
CSF and IL-6. Other cytokine increases at this
dose were transient. 10 mg/kg DEP increased IL-
6 at 4 h, and DEP or RPC increased G-CSF and
IL-6 at 12 h. DEP or RPC also increased IL-1p.
Myocardium: Myocardial MPO activity
significantly increased in 5 mg/kg RPC at 12 and
24 h. Myocardial MIP-2 increased the most in 5
mg/kg RPC, while LIX tended to be lowered by
RPC.
Table D-2. Respiratory effects: in vitro studies.
Study
Pollutant
Exposure
Effects
Reference: Aam and
Fonnum (2007,
1551231
Species: Human, Rat
Tissues/Cell Types:
Human-Neutrophil
Granulocytes (NG);
Rat-AM
DEP: SRM 1975
Particle Size: NR
Route: Cell Culture
Dose/Concentration: NG: 8.8 - 280 pg/mL
AM: 140, 280|jg/mL
Vitamin E = 5 pM
Time to Analysis: 1 h
ROS of NG: Formation of ROS in NG decreased
with increased doses of DEP. Lucigenin
chemiluminescence of ROS formation diminished
25% at 8.8 pg/mL DEP and luminol
chemiluminescence 32% with 17.5 pg/mL DEP. DCF
fluorescence required much higher doses of DEP.
Controls without PMA stimulation had highly reduced
lucigenin and luminol with DEP dose of 140 pg/mL
while DCF increased 116%.
ROS of AM: 280 pg/mL of DEP decreased ROS
level by 19% with DCF. DEP with PMA-unstimulated
cells increased 24% with DCF.
Necrosis: NG cell death was DEP dose-dependent.
At 280 fjg/mL, cell death increased 5.4% as
compared to control. LDH concentration increased
1.6% with 70 pg/mL DEP and 3.9% with 280 pg/mL
after 1 h.
Reference: Agopyan
et al. (2003, 0560651
Species: Human
Tissues/Cell Types:
BEAS-2B, NHBE,
SAEC
PC: synthetic carboxylate-modified
particles
Particle Size: 2,10pm
Route: Cell Culture
Dose/Concentration:
PC2 = 0.83 g/mL or 3.4x109 particles/mL
PC10 = 0.8 g/mL or 3x106 particles/mL
Time to Analysis:
PC2 = 12, 24, 8 h
PC10 = 2, 6, 12, 24 h
Calcium Imaging: PC10 induced increase of Ca2*
concentration in all capsaicin-sensitive cells 100%.
Similar reaction observed in cells exposed to PC2.
However, more than 3-PC2s were required to induce
a Ca increase unlike PC10. CPZ (10um) and
amiloride could fully block PC-induced response.
cAMP: Post 6 h, a dose-dependent increase in
cAMP was observed. Again, CPZ blocked increase
by 70-90% depending on cell type: SAEC >NHBE ~
BEAS-2B.
Apoptosis: PC10 and PC2 induced apoptosis time-
dependently. PC2 was slower in induction than
PC10. Post 48 h, 80-95% cells were apoptotic in all
cell types. Noncapaisin-sensitive cells (which did not
bind to particles) did not exhibit apoptosis. CPZ
reduced apoptosis by 97% BEAS-2B, 96% NHBE
and 98% SAEC. Amiloride did not block apoptosis.
Necrosis: Induction of necrosis by PC2 and PC 10
was negligible. A slight increase from 1 % to 2% was
observed at 24-48 h in NHBE and SAEC. BEAS-2B
showed slight decrease from 3% to 4% in same time
period.
December 2009
D-31
-------
Study
Pollutant
Exposure
Effects
Reference: Agopyan
et al. (2004, 1561981
Species: Human,
Mouse
Tissues/Cell Types:
Human-NHBE, SAEC;
Mouse-Wildtype and
TRPV1(-/-) Terminal
Ganglion Neurons
(TG)
ROFA
MSHA:MtSt Helen Ash
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 100 pg/mL ROFA or
MSHA
Time to Analysis: ROFA/MSHA in NHBE
and SAEC = 2, 6, 24, 48 h
ROFA/ MSHA in TG = 24 h
Calcium Imaging in NHBE and SAEC: In 100% of
reactive cells, ROFA/MSHA induced an increase in
Ca2*. Levels remained elevated as long as PM
bound to plasma membrane. Washing and disjoining
PM from membrane caused Ca2* to slowly decline to
baseline. CPZ (orCPZand amiloride) reversibly
inhibited PM-induced rises in Ca *.
Calcium Imaging in TRPV1(+/+) and (-/-) mice
cAMP measurements with NHBE and SAEC sensory neurons: All sensitive neurons in
exposed to ROFA/MSHA = 6 h TRPV1(+/+) increased Ca in response to ROFA.
K No effect of ROFA in TRPV1 (-/-).
cAMP: ROFA and MSHA induced increases in Ca2*
in NHBE and SAEC cells, which was completely
blocked by cAM P.
Apoptosis: ROFA or MSHA induced time-
dependent apoptosis, peaking at 24 h. CPZ again
inhibited this response. Neurons bound to PM
(<25um) induced apoptosis in TRPV1 (+/+). Cells
without bound PM or bound with PM (>25 pm)
showed no effect. No apoptosis occurred in the
absence of Ca *.
Necrosis: Necrosis for any of the cell types was
negligible.
PKA: Inhibition of PKA resulted in 90+% apoptosis
in NHBE and SAEC. Again, no apoptosis was
observed in a Ca2* free environment.
Reference: Ahn et al.
(2008, 1561991
Species: Human
Tissues/Cell Types:
A549
DEP: (6 cyl, 11L, turbo-charged, heavy- Route: Cell Culture
duty diesel engine, South Korea)
Dose/Concentrations: 0, 1, 5,10, 50 and
100|jg/mLofDEP
Dex: anti-inflammatory (Sigma, St.
Louis, MO)
Particle Size: NR
Some cells pre-treated with 10, 20, 40, 50
pg/mL of Dex.
Time to Analysis: 24 h
COX-2 Expression: Cells expressed dose-
dependent increases in COX-2 expression after
treatment with 10-100 pg/mL of DEP. Treatment of
50 |jg/mL for 24 h induced statistically significant
COX-2 expression in both mRNA and protein levels.
Pre-treatment with Dex significantly reduced
expression of COX-2 mRNA and protein. Dex
treatment induced dose-dependent suppression of
DEP-induced protein levels.
PGE2 Levels: Levels of the inflammatory mediator,
PGE2, increased when were cells exposed to 50
|jg/mL of DEP. Pre-treatment with 50 pg /mL Dex
completely inhibited DEP-induced release of PGE2.
Reference: Ahsan
(2005, 1562001
Species: Human
Tissues/Cell Types:
Trx-1-transfected
Clone of Murine L-929
cells; Control Clone (L-
929-Neo1);A549
DEP: provided by Dr. Masaru Sagai,
University of Health and Wfelfare,
Aomori, Japan
Particle Size: NR
Route: Cell Culture
Dose/Concentration:
DEP: 50 pg/mL
hTrx-1- or L-929-Neo1: 40 pg/mL
Pretreatment: rhTrx-1 (10 pg/mL) or DM-
rhTrx-1 (NR)
Time to Analysis: Pretreatment for 1 h.
Parameters measured 3 h post exposure.
ROS: DEP induced significant increases of ROS in
L929-Neo1 cells. hTRx-1 cells showed no affect. RT-
PCR revealed hTrx-1 mRNA expression in
transfected cells but not control L929-Neo1 cells.
Endogenous murine Trx-1 mRNA expression
increased in control cells, but not in hTrx-1 cells.
A549 cells had increased ROS levels but these
levels were suppressed with rhTrx-1 pretreatment.
Pre-treatment with DM-rhTrx-1 increased ROS
levels more.
Akt (antiapoptotic molecule): Phosphorylated Akt
prevents apoptosis. DEP induced phosphorylation of
Akt in control cells after 3 h and dephosphorylation
after 5 h. In hTrx-1 cells, Akt remained
phosphorylated after 5 h. In A549 cells, Akt
phosphorylated at 3 h and slowly turned off at 12-
24 h. Pre-treatment with rhTrx-1 blocked
dephosphorylation. This suggests that Trx-1
preserves active form of Akt and thereby protects
against cytotoxicity from DEP.
December 2009
D-32
-------
Study
Pollutant
Exposure
Effects
Reference: Alfaro-
Moreno et al. (2002,
1562041
Species: Human,
Mouse, Rat
Strain: Human-A549;
Mouse-J774A.1,
BALB-c
Tissues/Cell Types:
HUVEC, Mouse
Fibroblasts, Rat Lung
Fibroblasts (RLF)
PMi0: Collected from 3 zones in Mexico Route: Cell Culture
City: North (industrial), Center
(business) and South (residential)
15000 cells/cm2 except:
Particle Size: PM1(
Cytotoxicity: Confluent Cultures 180,000
cells/cm .
DMA Breakage: 20,000 cells/well.
Cytokine Assays: 180,000 cells/cm2
Dose/Concentration: Cytotoxicity: 10, 20,
40, 80, 160 pg/cm2
Apoptosis: 160|jg/cm2
DMA Breakage: 2.5, 5,10, 20, 40 pg/cm2
Cytokine Assays: 10, 20, 40, 80 pg/cm2
E-Selectin Expression: 40 pg/cm2
Time to Analysis: Cytotoxicity: 24, 48, 72 h;
Apoptosis: 24 h; DMA Breakage: 72 h;
Cytokine Assays: 24 h
Cytotoxicity: Cytotoxic effect exhibited dose-
dependency after 72 h in proliferating cells of
J774A.1, BALB-c and RLF cell lines.
Proliferating Cells: Northern particles induced a
statistically larger effect than central or southern
particles. J774A.1 was more susceptible while
BALB-c was less susceptible. A549 was most
resistant to decreased viability during exposure. No
significant variation in viability was observed when
compared to the control. Particles were not cytotoxic
among confluent cell growth for any cell lines when
exposed to 20-160 pg/cm .
Apoptosis: Overall, particles induced low rates of
cell death via apoptosis. J774A.1 depicted similar
levels of apoptosis when exposed to three PM
zones, -15% apoptotic cells measured. BALB-c was
not reported. Results for the A549 measured
apoptotic cells were: South- 4%, Central-11% and
North-15%. HUVEC cells indicated an increase in
apoptosis with northern particles.
DMA Breakage: PMi0 from all zones induced DNA
breakage. A dose-dependent relationship was
established with PM2s particles at concentrations of
10 pg/cm2. The Southern zone required a higher
dose of PM (10 pg/cm2) to produce the same effect
as other zones (2.5 pg/cm2).
Cytokines: Particles induced TNF-a and IL-6
secretion in J774A.1 cells dose-dependently IL-6
increased significantly with central particles. PGE2
secretion in RLF cells induced by exposure to PM
showed dose-dependent responses. PM from the
central zone induced the most PGE2 secretion. Max
secretion was observed at doses of 40 pg/cm2 from
all three PM zones.
E-Selectin Expression: HUVEC cells showed a
25% increase in E-selectin expression after
exposure to 40 pg/cm2 of PM.
Reference: Amakawa DEP (obtained from a 4JB1, Isuzu, Route: Cell Culture
Species: Mouse,
Human
Strain: Mouse-ICR
Tissues/Cell Types:
AMs
Gender: Male
Age: Mouse 6-7 wk;
Human 20-24 yr
1500rpm, 4cyl diesel engine)
DEPE = DEP Extract (methanol)
CB = Charcoal (Sigma)
Particle Size: DEP- 0.4 pm, CB- 0.7
pm
cells/ml
Dose/Concentration: DEP = 1 or 10 pg/mL;
DEPE=1or10pg/mL;CB=1,10,100
pg/mL
Time to Analysis: Human cells pre-treated
with LPS 1 pg/mL. Murine cells pre-treated
with SOD 300 lU/mL. Parameters measured
24 h post exposure.
Cells: For mice, more than 90% of the cells were
macrophages and over 90% were viable. For
humans, 96% of the cells were macrophages, 3%
lymphocytes and 1% neutrophils; over 95% of the
human cells were viable.
DEP Cytotoxicity: None observed
Cytokines: DEP (10 pg/mL) suppressed release of
TNF-a and IL-6 for both mice and humans in a dose-
dependent manner. Murine cells pre-treated with
LPS or IFN-y released even less TNF-a and IL-6. IL-
10 was unaffected. Human macrophages pre-
treated with LPS also released lower levels of TNF-
a, IL-6 and IL-8.
ROS: Pre-treatment of SOD on murine cells partially
attenuated the suppressive effect of DEP as well as
decreased the production of ROS generated by DEP
(10pg/mL).
Carbon: Carbon particles did not suppress TNF-a or
IL-6 release from murine AMs; however,100 pg/mL
of CB stimulated TNF-a production.
Methanol: No Cytotoxicity nor cytokine release
effects were observed.
DEPE: DEPE suppressed TNF-a and IL-6 release in
a similar way as DEP.
December 2009
D-33
-------
Study
Pollutant
Exposure
Effects
Reference: Amara et DEP = SRM 2975
al. (2007, 156212) ^ . u , _, t
CSC = cigarette smoke condensates
Species: Human (collected from Kentucky standard
~.... ,^n „,„, cigarettes, 2R4F; University of
Cell Lines: A549, NCI- Klntucky)
H292
DC = DEP + CSC
CB (Degussa, Frankfurt, Germany)
Particle Size: CB: 95 nm; DEP: NR
Route: Cell Culture
Dose/Concentration: DEP = 5-10 pg/cm2
CB=10|jg/cm2
CSC=10|jg/cm2
Time to Analysis: 6 or 24 h
Inflammatory Markers: LDH of A549 was
unaffected at either time point with DEP or CB. LDH
increased with CSC at concentrations high than 10
pg/mL at both time points. DC had no effect.
Proteases: MMP-1 mRNA expression showed a
dose dependent increase with DEP in A549 cells.
DEP also increased MMP-1 in NCI-H292 cells. CB
and CSC had no effect. MMP-1 mRNA expressions
were inhibited by N-acetylcysteine antioxidant.
Similar inhibition was observed with NOX4 oxidase.
DC induced a similar effect to DEP. MMP-1 protein
expression increased post 24 h with DEP. MMP-2,
TIMP-1, TIMP-2 mRNA expression was unaffected.
TGF: TGF-|3 mRNA expression was unaffected.
ROS: DEP and DC increased ROS formation after 1
h. DEP effect was inhibited by N-acetylcysteine
antioxidant pre-treatment.
MAP-Kinase: DEP induced MMP-1 expression
increased ERK1/2 phosphorylation after 10 min,
peaking at 30 min, and returning to normal levels at
60 min. Treatment with CBPs did not increase
ERK1/2 phosphorylation whereas treatment with
CSC resulted in phosphorylation. Only inhibitors of
ERK1/2 reduced DEP induced MMP-1 activity. P38
and JNK inhibitors had no effect.
Reference: Anseth et
al. (2005, 0886461
Species: Human
Cell Lines: A549;
A549-pO (lacking
mitochondria)
s-ROFA: soluble portion
Particle Size: 1.95 +018 fjm
Route: Cell Culture (3X105 cells/ml)
Dose/Concentration: 100 |jg/mL
Time to Analysis: Experiments conducted
by spreading monolayer of Infasurf (calf lung
surfactant extract on PBS, PBS+ROFA or
conditioned media from A549 AEC.
Parameters measured after one 6-h
incubation period.
Lung Surfactant Gelation: ROFA alone and A549
conditioned media alone did not significantly alter
Infasurf rheology However, conditioned media from
A549 AEC at 16 h induced a significant increase in
elastic storage and viscous loss moduli. Inhibiting
ROS production lowered effect, indicating s-ROFA
gelation mediated through ROS.
ROS: ROS mediated through mitochondria as
evidenced by the effect of ROFA-AEC on surfactant
gelation in the presence of mitochondria ROS
inhibitors as well as A549-pO cells.
Reference: Auger et
al. (2006, 1562351
Species: Human
Tissue/Cell Type:
Nasal Epithelial Cells
DEP:SRM1650
PM25: obtained from a highway in
Paris, France
Particle Size: DEP: 400 nm (mean
diameter); PM25
Route: Cell Culture (2-3.5x104 cells/cm2)
Dose/Concentration: 10-80 pg/cm2
Time to Analysis: Cells treated on apical
side. Parameters measured 24 h following
treatment.
Cytotoxicity (LDH): No cytotoxicity for DEP or
PM25 (80 pg/cm2).
Cytokines: In non-stimulated All cultures, IL-8 was
the most abundantly secreted cytokine, followed by
GM-CSF, TNF-a, and IL-6 in decreasing levels of
production. Amphiregulin was moderately, but
consistently, secreted. After treatment, both DEP
and PM25 induced IL-8 and amphiregulin release in
a dose-dependent manner through the basolateral
surface. PM25 stimulated IL-6 and GM-CSF release
through the apical surface.
ICAM-1 expression: No effect from DEP or PM25.
ROS: DEP and PM25 both increased ROS
production in a dose-dependent manner.
December 2009
D-34
-------
Study
Pollutant
Exposure
Effects
Reference: Bachoual PMi0 from two Paris, France subway
et al. (2007,1556671 sites: PER and Metro
Species: Mouse CB (Frankfurt, Germany)
Cell Type: RAW 264.7 Ti02 (Calais, France)
DEP:SRM1650(NIST)
Particle Size: CB: 95 nm; Ti02:150 nm;
DEP: NR
RER PM10:79% <0.5 \im, 20% 0.5-1
pm;
Metro PM,0: 88% <0.5 \im, 11% 0.5-1
pm.
Route: Cell Culture (40,000 cells/ml)
Dose/Concentration: All particles: 0.01,
0.1,1,10|jg/cm2
Time to Analysis: 3, 8, 24 h
Cell Viability: No effects from any particulate at
concentrations up to 10 pg/cm2 for 24 h.
Inflammatory Effect: Exposure of cells to 10
pg/cm2 of RER or Metro induced time-dependent
increase in TNF-a and MIP-2 protein release. This
effect was similar to both locations. No effect was
observed at low concentrations of PM10. No effect of
CB, Ti02 or DEP was observed.
GM-CSF or KC production: RER and Metro PM,0
did not induce any effect at any concentration.
Effect on Protease mRNA Expression: Exposure
of cells to 10 pg/cm2 RER or Metro PM10 did not
modify mRNA expression of MMP-2 or -9 or their
inhibitors TIMP-1 and-2. MMP-12 expression
significantly increased after exposure to RER or
MetroPMi0for8h.
Effects on HO-1 Protein Expression: Exposure to
10 pg/cm2 of RER or Metro PM,0 for 24 h induced
positive cytoplasmic staining for HO-1.
Reference: Baulig et WUB: Winter Urban Background
al. (2007,1517331 Particles (obtained from Vitry-sur-
Seine, suburb of Paris, France)
Species: Human
SUB: summer Urban Background
Cell Line: 16-HBE14o- Particles Vitry-sur-Seine)
WC: Winter Curbside Particles,
SRM1648 (obtained from Porte-
d'Auteuil, ring road of Paris, France)
SC: Summer Curbside Particles, SRM
1648(Porte-d'Auteuil)
DEP: SRM 1650a (NIST)
DPI (control)
Particle Size: WUB, SUB: PM25; WC,
SC, DEP: NR
Route: Cell Culture (20,000 cells/cm2)
Dose/Concentration: 10 pg/cm2
Time to Analysis: 18 or 24 h
EOF: All native PM25 induced similarAR secretion
by bronchial epithelial cells (in decreasing order WC,
WUB, SC, SUB), but this release was significantly
greater than the release induced by DEP. (3-cellulin
increased with SC, WUB and WC. No data was
available for SUB or DEP.
Interleukins: IL-1a increased significantly with
WUB, WC.SC, DEP, DPI (in decreasing order). No
data was available for SUB. Exposure to WUB
caused IL-113 to increase to induction factor of over
2. IL-11 R a decreased significantly with SUB.
Cytokines: Exposure to WUB caused G-CSF to
increase with an induction factor of over 2.Though
not statistically significant, TNF-R1 also increased.
Proteases: TIMP-2 decreased with WUB but
significantly increased with SUB. Overall, SUB
downregulated integrins and interleukins seen with
other particles while upregulating neurotrophic
factors, chemokine receptors and adhesion
molecules. MMPs were not measured.
Chemokines: CCR-3 significantly increased with
SUB. GRO-v and GRO-a increased with WC at both
18 and 24 h. DEP had no effect with GRO-a.
Removal of metal from particles lowered response of
GRO-a.
December 2009
D-35
-------
Study
Pollutant
Exposure
Effects
Reference: Bayram et DEP: (obtained from a 4JB1-type, light- Route: Cell Culture
al. (2006, 0884391 duty, 4 cyl, 2.74-L Isuzu diesel engine)
Species: Human DEP-FCS: DEP + PCS
Cell Type: A549 DEP-NAC: DEP + N-acetylcystine,
antioxidant
DEP-A: DEP + AEOL10113, catalytic
antioxidant
DEP-S: DEP + SP600125, inhibitor of
JNK
DEP-N: DEP + SN50, inhibitor of NF-
kB
Particle Size: DEP: 0.4 pm (mean
diameter)
Dose/Concentration: DEP: 0, 5,10, 50,
100, 200 pg/mL
Time to Analysis: 24, 48, 72 h
Cell Growth: With 10% PCS (as a positive control),
A549 cells exhibited time dependent growth. A
mixture of PCS and DEP did not affect cell growth
for up to 48 h. With DEP alone, cell growth was
prevented from cell number reduction due to
removal of serum at 48 and 72 h. A dose of 10
pg/mL induced a maximum proliferation effect.
Cell Cycle: DEP increased the percentage of
serum-starved cells in S phase at 48 h. DEP
decreased the percentage in GO/1 phase and G2/M
phase.
Apoptosis: DEP prevented the increase in
apoptotic, serum-starved cells.
Protein Expression: p21CIP1/WAF1 expression
increased at 48 h. DEP dose-dependently
decreased this expression.
MAC: MAC alone, at 33 mM, induced an increase in
cell numbers. DEP-NAC inhibited cell numbers at 48
h. DEP-NAC inhibited cell numbers in S phase; thus,
cells in GO/1 phase increased. DEP-NAC induced a
further decrease of cells in G2/M phase.
AEOL10113: DEP-A caused a dose-dependent
decrease in cell numbers.
SP600125: Alone, SP600125 increased cell
numbers at 33 mM. DEP-S decreased cell numbers.
Reference: Becher et SPM = suspended PM SRM-1648
al. (2007, 0971251
Particle Size: 6-8 pm
Species: Rat
Strain: Crl/WKY
Cell Type: AM,
Alveolar Type II
Gender: Male
Weight: 200 g
Route: Cell Culture (1.5x106 cells/well AM;
6x106 cells/well Type II)
Dose: 200 pg/mL = 20 pg/cm2
Time to Analysis: 20 h
Cytokines in Macrophages: SPM increased TNF-a
and MIP-2. NADPH inhibitor DPI reduced MIP-2
response, whereas iNOS inhibitor 1400Wdid not
affect either.
Cytokines in Type 2 Cells: SPM increased IL-6 and
MIP-2 significantly This SPM effect was inhibited by
DPI, whereas!400W reduced the IL-6 response
significantly.
ROS in Type 2 Cells: SPM significantly increased
ROS formation. DPI largely blocked this SPM effect.
ROS in Macrophages: No significant increases
were observed.
Reference: Becker et PM (Coarse, Fine, Ultrafine): Chapel
al. (2005, 0885901 Hill, NC
Species: Human
Gender: Male and
Female
Age: 18-35 yr
Cell Types: Alveolar
Macrophages, NHBE
Particle Size: PM-C: PM25; PM-
F:PM01;PMUF:<0.1|jm
Route: Cell Culture (0.5-1 xio5 cells/well
NHBE; 2-3x105/mL AM)
Dose/Concentration: NH BE: 25, 50,100,
250 pg/mL of PM; AMs: 50 pg/mL of DEP or
10ng/mLofLPS
Time to Analysis: 18h for NHBE; overnight
forAMs
Cytokines: All 3 fractions induced dose-dependent
increases in IL-8 secretion with PM-c, PM-F, PM-UF
(in order of decreasing effects). TLR-2 antibody
blocked these particle induced IL-8 effects.
Inhibitors of Endotoxin effects and TLR-4
activation: No effects were observed in NHBE, but
all 3 fractions repressed the IL-6 release in AMs.
TLR mRNA Expression: PM did not affect TLR-2
mRNA in NHBEs. PM-C and PM-F induced a slight
increase in TLR-4 mRNA in NHBEs while PM-UF
induced a substantial increase. PM-C increased
TLR-2 mRNA in AMs and decreased TLR-4 mRNA
in AMs.
Induction of Hsp70: PM-C and PM-F induced
Hsp70 in NHBE dose-dependently Hsp70 was not
induced in AM following particle stimulator.
December 2009
D-36
-------
Study
Pollutant
Exposure
Effects
Reference: Becker et PM (Coarse, Fine, Ultrafine): Chapel Route: Cell Culture (3-5x10 cells/well
al. (2005, 0885921
Species: Human
Gender: Male
Age: 18-35 yr
Cell Types: AM,
NHBE
ROFA
Fe, Si, Cr Components
Oct2001, Jan 2002, April 2002, July
2002
Particle Size: PM-C: 2.5-10 pm; PM-F:
<0.1 pm; PM-UF: <0.1 pm
NHBE; 2-3x1 Ob cells/mLAM)
Dose/Concentration: NHBE: 11 pg/mLof
PM;AM:50|jg/mLofPM
Time to Analysis: 18-24 h NHBE; 18 hAM
IL-8 Release in NHBE: PM-C and PM-UF induced
effects. No effects from PM-F (all 4 dates).
IL-6 Release in AM: All 3 fractions induced increase
with later dates having generally lower effects.
ROS (DCF): NHBE, at lower exposures, were
observed to be more responsive to PM than AMs.
AM exhibited highly variable results over time.
ROS (DHR): NHBE cells were observed to be more
responsive to PM than AMs. AM responsiveness to
PM increased over 4 time periods; this was not
observed in NHBE.
Seasonal Variability: Coarse particles were more
potent than F and UF regardless of the month, and
the potency for PM to induce IL-6/IL-8 production
varied significantly. Coarse particles induced a 5-25
fold change in IL-6 release for AMs and a 3-6 fold
change in IL-8 release for NHBEs.
Metal Correlation to IL-6/8 induction: Fe and Si
were positively associated with IL-6 release in AMs
incubated with the coarse fraction. Crwas positively
associated with IL-8 release in NHBE cells
incubated with F or UF.
Reference: Beck-
Speier et al. (2005,
1562621
Species: Human,
Canine (Beagle)
Cell Types: Human
AMs, Canine AM
(CAM)
DEP = SRM 1650a (NIST)
EC = Ultrafine EC (spark discharge)
P90 = Printex 90 (Carbon Black,
Degussa)
PG = Printex G (Carbon Black,
Degussa)
Particle Size: DEP: 20-40 nm; EC: 5-
10nm;P90:14nm;PG:51 nm
Route: Cell Culture (1x1 Ob cells/mLAM)
Dose/Concentration: All particles: 1 (EC
only), 3.2,10,32,100|jg/mL
Time to Analysis: 60 min
Phagocytosis: All particles were phagocytosed by
CAM within 60 min.
Oxidative Potential: EC showed a very high effect.
DEP, P90 and PG had no effect
Formation of Lipid Mediators: DEP, EC P90 and
PG increased arachidonic acid and PGE2/TXB2 in
CAM in a dose-dependent manner. Only EC
increased LTB4 and 8-isoprostane.
ROS Activation: All particles increased activity in
canine macrophages with EC, P90 and PG
increasing activity in a dose-dependent manner.
DEP increased activity in canine macrophages.
Similar results were observed human alveolar
macrophages but only EC and P90 were tested.
Particle Mass vs Particle Surface Area:
PGE2/TXB2 effects were highly correlated with
particle surface area.
Reference: Bitterle et
al. (2006, 1562761
Species: Human
Cell Type: A549
C-UFP = Ultrafine carbonaceous
particles (obtained from a spark
discharge aerosol generator GFG
1000, Palas, Karlsruhe, Germany)
Particle Size: 90 nm (count median
mobility diameter)
Route: Cell Culture (3x107 cells)
Dose/Concentration: 44 + 4 ng/cm2; 87 +
23 ng/cm2; 230 + 70 ng/cm2
Time to Analysis: 6 h
Cell Viability: Exposure to clean air resulted in a
93.7 ±9.1% viability. Exposure to low, mid and high
doses of C-UFP resulted in a 94.9 + 9.5% viability.
Thus C-UFP had no effect on cell viability.
Interleukins: Clean air controls induced a 2-3 fold
increase in IL-6 and IL-8 production vs submersed
control. U-CFP exposures induced a similar effect
on IL-8 and IL-6 levels.
Antioxidant enzyme HO-1: The mid dose
increased transcription of HO-1 by 2.7 fold. There
was no observed effect at the high dose level which
indicates possible cytotoxicity.
December 2009
D-37
-------
Study
Pollutant
Exposure
Effects
Reference: Blanche!
et al. (2004, 0879821
Species: Human
Cell Type: 16H BE
PM25
(Vitry-sur-Seine, Paris, France)
DEP = SRM 1650a
CB = Carbon Black (Degussa)
TI02 (Huntsman)
Particle Size: CB: 95 nm; TI02:150 nm
Route: Cell Culture (45,000 cells/cm')
Dose/Concentration: All particles: 0.1,1,
10, 30 pg/cm2
Time to Analysis: 6,18, 24, 30 h
Amphiregulin Expression: DEP and PM25 both
increased AR mRNA expression from 6 to 30 h, with
PM25 inducing higher expression levels than DEP.
Both DEP and PM25 increased AR protein secretion.
No observed effect for CB and Ti02. PM25 induced
protein secretion dose-dependently.
Signal Pathways in AR Secretion: MAP kinase
and tyrosine kinase inhibitors reduced effects of
DEP and PM2 5 but p38MAP kinase inhibitor did not.
Role of Oxidative Stress: N-Acetylcysteine blocked
AR secretion following PM25. Antioxidant enzyme
catalase had no effect.
Cytokines: DEP induced a significantly high release
ofGM-CSF, higher than PM25. EGFR antibody
reduced GM-CSF release at 0.25 pg/mL dose.
Reference: Bonvallot DEP: SRM 1650
etal. (2001, 156283'
Route: Cell Culture (3x1 Ob cells)
OE-DEP: dichloromethane extract (2x) Dose/Concentration: DEP, sDEP, nDEP
and CB = 10 pg/cm2
Species: Human of DEP
Cell Type: 16HBE14o- nDEP: native DEP
sDEP: nDEP - OE-DEP
CB: Carbon Black FR103 (Degussa)
BaP: Benzo[a]pyrene
CB: 95 nm
NR
Particle Size: CB: 95 nm; DEP: NR
OE-DEP = 15 pg/mL
BaP = 0.25, 50 and 250 pg/mL
Time to Analysis: 24 h
Proinflammatory Response: At 10 pg/cm , nDEP
induced GM-CSF release by 4.7 fold. OE-DEP
increased GM-CSF by 3.7 fold. BaP and sDEP also
induced increases of CN-CSF but had smaller effect.
CB had no effect.
NF-KB Activation: nDEP and OE-DEP induced
enhanced degradation of 1KB at 2-4 h and 1 h
respectively. NF-KB DNA binding was enhanced by
OE-DEP (15 pg/mL, peak <1 h) and nDEP (10
pg/cm2, peak at 2-h with plateau till 4 h). Both OE-
and nDEP enhanced NF-kB DNA binding levels
were higher than BaP enhanced binding levels.
CYP1A1 mRNA: The CYP1A1 mRNA level was
markedly increased in nDEP and OE-DEP treated
cells in comparison with their respective controls.
Radical Scavengers (decreased ROS in situ):
Increases of GM-CSF and NF-KB DNA binding by
nDEP and OE-DEP was attenuated by radical
scavengers.
MAPK Activation: Increases by nDEP and OE-DEP
of GM-CSF was inhibited by Erk1/2 inhibitor but not
by p38 inhibitors. Both nDEP and OE-DEP triggered
Erk1/2 and p38 phosphorylation. sDEP affected p38
phosphorylation only.
Reference: Brown et
al. (2007, 1563001
Species: Human,
Mouse
Cell Type: PBMC,
A549 (Human);
J774A.1 (Mouse)
PM10 (London, England)
CM from PM10-treated human
monocytes
Particle Size: PM10
Route: Cell Culture (1 xio6 cells/ml
J774A.1; 5x106 cells/ml PBMC; 5x105
cells/well A549)
Dose/Concentration: PMi0:75 pi (10
pg/mL);CM:250pl;tBHP:12.5pm(in
J774);TNF:0, 500 pg, 1 ng, 10 ng
Time to Analysis: tBHP:1, 2, 4 h; PM: 4 h;
TNF:18h
Cytokines: PM10 induced release TNF-a protein
from PBMCs at 10 pg/mL for 4 h. Further inhibited
by verapamil and BAPTA-AM. Calmodulin inhibitor
W-7 had no effect. CM increased IL-8 from A549
cells 3 fold. Verapamil, BAPTA-AM and W-7
significantly inhibited IL-8 release induced by CM.
ICAM-1: A549 cells treated with TNF-a showed
dose-dependently effect of TNF-a on ICAM-1
upregulation at 18 h. CM also induced upregulation.
Verapamil, BAPTA-AM and W-7 fully inhibited CM-
induced upregulation.
December 2009
D-38
-------
Study
Pollutant
Exposure
Effects
Reference: Calcabrini PM25(Rome, Italy)
et al. (2004, 0968651
v Particle Size: PM25
Species: Human
Cell Type: A549
Route: Cell Culture (5^104 cells/well) Particle Characterization: Components measured
, include C-rich particles, Ca sulfates, silica, silicates,
Dose/Concentration: 30, 60 pg/cm (aliquot Fe.rich particles, metals. Carbonaceous particles
ofO.1 fjg/fjl) made up majority of PM.
Time to Analysis: 5, 24, 48, 72 h
Cell Surface Changes: PM deposited on the cell
surface showed dose and time-dependent increases
in microvilli rearrangement and cell shape alterations
without affecting apoptotic markers for up to 72 h.
PM internalization: At 24 h with the low dose,
aggregates of PM in cytoplasm or surrounded by
membrane was observed. With the high dose, large
particle aggregates often close to nuclear envelopes
were observed.
Cytoskeleton: At 72 h PM induced dose-dependent
alterations from rearrangement/interweaving of
microtubules to bundling of microtubules with some
shortening/disruption.
Cell Growth: PM decreased cell growth in a dose
and time-dependent manner
ROS: PM increased ROS at the high dose for 5 h
but not at 24 h or with the low dose.
Cytokines: PM induced TNF-a peaked at 5 h at
high dose and 48 h at low dose, both ND at 72 h.
PM induced IL-6 starting at 24 h thru 72 h in time
and dose dependent manner.
Reference: Cao et al.
(2007, 1563221
Species: Human
Cell Type: HAEC
NIST-DEP: collected using a diesel
forklift and hot bag filter system. (NIST,
Minneapolis, MN)
C-DEP: obtained from a 30-kw (40 hp)
four-cylinder Deutz BF4M1008 diesel
engine (U.S. EPA)
Organic extract fraction of particles
NIST-DEP 2%
C-DEP 20 %
Particle Size: NR
Route: Cell Culture (5x105 cells)
Dose/Concentration: NIST-DEP, C-DEP: 0,
12.5,25,50, 100, 200|jg/mL
Time to Analysis: 1-4h
Cell Viability: DEP had no effect.
StatS: Both DEPs induced time-dependent
phosphorylization of StatS in cytoplasm. NIST-DEP
induced phosphorylization dose-dependently from
12.5 to 50 fjg/mL but stayed level at 100 and 200
pg/mL p-Stat3 induction was inhibited by antioxidant
BHA though it was reactivated with exposure to
H202. Reaction induced by H202 was similar to that
of DEP.
pStatS Nuclear Transport: NIST-DEP induced
cytoplasmic pStatS to move from cytoplasm into
nucleus.
pEGFR Dephosphorylation: After 4 h of NIST-DEP
exposure, dephosphorylation was inhibited for up to
90 min.
Reference: Chang et
al. (2005, 0977761
Species: Human
Cell Type: A540, THP-
1
UfCB (Printex 90, Degussa)
Particle Size: 14 nm
Route: Cell Culture (7x105 cells)
Dose/Concentration: 100 |jg/mL
Time to Analysis: 4 h
ROS in THP-1 and A649: UFCB increased ROS.
NAC pretreatment blocked most of the UFCB-
induced ROS production.
VEGF in THP-1: UFCB increased VEG. NAC
decreased the UFCB effects below those of the
control.
VEGF in A649: Produced similar, but less marked,
results as with THP-1.
December 2009
D-39
-------
Study
Pollutant
Exposure
Effects
Reference: Chauhan
et al. (2004, 0966821
Species: Mouse
Strain: BALB/c
Cell Type: RAW
264.7;J774A.1;
WR19M.1
EHC-T: total EHC-93 (Env Health Ctr,
Ottawa, Canada)
EHC-I: insoluble EHC
EHC-S: soluble EHC
SRM1648: urban particulate St. Louis
(NIST)
SRM1649: urban dust/organics
Washington (NIST)
VERP: fine PM25 (Vermillion, Ohio)
Cristobalite: SRM 1879 (NIST)
Ti02:SRM154b(NIST)
Particle Size: EHC-93: 0.5 pm (median
diameter); Cristobalite, SRM 1648,
SRM 1649, Ti02,:NR; VERP: PM25
Route: Cell Culture (15000 cells/well)
Dose/Concentration: Particle suspensions:
20, 50, 100 fjg/well
IPS: 0-5 pg/mL
IFN-y:0-1000U/mL
Time to Analysis: Particles added to culture
at Oh, IPS and IFN-v added at 2 h.
Parameters measured after 22 h incubation
period.
Stimulation with LPS/IFN-y: IPS and IFN-y each
induced NO release. Combination of IPS and IFN-y
produced larger effect in all cell lines. L-NMMA,
NOS inhibitor, suppressed most of the NO
production with 100 nmol/L
Cellular Viability and Cytotoxicity: Exposure of
cells to particulates did not result in overt cytotoxicity
or excessive loss of cellular material. There was no
correlation between the cytotoxicity of the particles
in the surviving cells and the loss of protein mass in
monolayers.
Nitrite Production: EHC-T, EH-93-I, SRM1648 and
SRM 1649 produced dose-dependent decreases.
Cristobalite only decreased at higher doses. No
effect from EHC-S, VERP or Ti02.
iNOS: EHC-I, EHC-T, Crisobalite and SRM1648
inhibited iNOS expression. Ti02 had no effect. EHC
sol, SRM 1649 and VERP were not tested.
Reference: Chauhan
et al. (2005, 1557221
Species: Human
Cell Type: A549
EHC-T: total EHC-93
EHC-I insoluble EHC
EHC-S: soluble EHC
Cristobalite (Si02): SRM-1879
Ti02:SRM-154b
Particle Size: EHC-93: 0.4 pm (median
physical diameter); Ti02, Si02: 0.3-0.6
pm
Route: Cell Culture (150000 cells/flask)
Dose/Concentration: All particles: 0,1, 4,
mg/5ml
Time to Analysis: 24 h
Cellular Viability: Decreased after exposure to
EHC-T, EHC-I and Cristobalite. Rate of reduction
was not consistent across doses. EHC-S and Ti02
had no effect on viability.
ET-1: Release of ET-1 peptide decreased dose-
dependently for EHC-T, -S and -I. Fractions of EHC-
S and EHC-I were more potent than EHC-T. Ti02
and Cristobalite also reduced ET-1 secretion
although this was not consistent across the dose
range.
Cytokines: Results showed no detectable amounts
ofGM-CSF, IL-1P or TNF-a in cell culture
supernatants. IL-8 increased dose-dependently with
EHC-T, EHC-I and Cristobalite.
VEGF: VEGF significantly increased dose-
dependently with EHC-T, EHC-S and Cristobalite.
EHC-S induced a significant decrease in VEGF.
Gene Expression: mRNA levels for preproET-1
reduced at 24 h for all particle types. EHC-S induced
a significant decrease in ET-1 expression at this high
dose. ECE-1 mRNA expression increased with EHC-
T and EHC-I. Other particles had no effect. ETaR
mRNA increased with EHC-T, EHC-S, and Ti02 in
biphasic manner where the highest expression of
mRNA was seen at the middle dose levels. EHC-S
had no effect. ETbR mRNA increased with a low
dose of EHC-T and decreased with a high dose of
EHC-T. EHC-S, EHC-I and Cristobalite induced an
increase of ETbR. Ti02 induced a significant
decrease.
Proteases: mRNA levels for MMP-2 reacted
similarly to preproET-1. mRNA levels for TIMP-2 was
significantly induced with EHC-I. EHC-T and EHC-S
induced small effects.
Reference: Cheng et
al. (2003, 1563371
Species: Human
Cell Type: A549
DEP-h: DEP with high sulfur
DEP-LS:DEP with low sulfur
Route: In Vitro Cellular Exposure (Exhaust
flow-through cell culture with air-cell-
GEP: gasoline engine exhaust particles interface, exhaust diluted 10-15x with 8^10
Primed cells pretreated with TNF-a cells/mL)
Particle Size: DEP-h: 15.9 nm; DEP-
LS: 17.7 nm;GEP: 8.3 nm
Dose/Concentration: DEP (total): 1.5-
3.5x106 particles/cm3; GEP (total): 1-2x106
particles/cm3; TNF-y: 5ml (25 ng/ml)
Time to Analysis: 60-360 min
IL-8: DEP-h induced a 3 fold increase in IL-8 than
that of the control. DEP-LS also induced increases.
Primed cell cases had higher levels (10x) than
unprimed when exposed to DEP-LS. DEP-h induced
higher levels of IL-8 than DEP-LS. This response
lasted for up to 6 h. GEP induced a statistically
insignificant increase of IL-8 in unprimed cells. Wth
primed cells, GEP induced levels of IL-8 that
exceeded those of DEP-h and DEP-LS. This
response lasted for 1-2 h.
December 2009
D-40
-------
Study
Reference: Chin et al.
(2003, 1563401
Species: Rat, Human
Cell Line/Type: RAW
264.7, MHS (Alveolar
Macrophage Cell
Line), A549
Pollutant
CB: (N339, with benzo[a]pyrene
absorbed on surface. Manufactured in
Cabot, Boston, MA)
BaP
Benzo [a] pyrene 1, 6-quinone: BP-1,6-
Q (obtained from NCI, Kansas City,
MO)
Particle Size: CB 0.1 pm (mean
diameter)
Exposure
Route: Cell Culture
Dose/Concentration :
CB: 1, 2, 4|jg/mL
BaP: 2 pg/mL
BP-1,6-Q:1 pM
Time to Analysis: 1 -24 h
Effects
HO-1 mRNA Expression: In RAW2647, HO-1
mRNA levels increased with 2 and 4 pg/mL at 2 h.
Increases continued to 8 h and declined by 24 h.
BaP had no effect. BP-1.6-Q increased HO-1 mRNA
after 1 h and was maintained until 8 h. In A549 and
MHS, HO-1 mRNA increased after 1 h, peaking at 8
hinA549and4hinMHS.
HO-1 Protein Expression: An increase of protein
was observed from 4-8 h in RAW2647.
AP-1 : Increases in binding activity were observed in
RAW 264.7 cells at 2 h.
Reference: Churg et
al. (2005, 0882811
Species: Rat
Strain: SD
Weight: 250 g
Cell Type: Epithelial
Cells of Tracheal
Explants
EHC93 (Ottawa Urban Air Particles) Route: Cell Culture
TiFe = Iron-loaded fine Ti02 (obtained
from Aldrich Chemicals, Milwaukee,
Particle Size: EHC-93: 3-4 pm
(MMAD); TiFe: 0.12 +1.4 pm
(geometric mean diameter)
Dose/Concentration: EHC-93, TiFe: 500
pg/cm2
Time to Analysis: 1, 24 h. Some
experiments (referred to as 2 h) explants
transferred to different dish and incubated
for additional hour. Pre-treated with
Inhibitors/Chelatorsfor2h.
Activation of NF-KB: Both particle types increased
nuclear translocation of NF-KB. TiFe and EHC-93
increased NF-KB 1.5 fold at 1 h. TiFe increased NF-
KB 3.5 fold at 2 h. EHC-93 increased NF-KB more
than 2 fold. Ti02 by itself did not increase NF-KB at
any exposure duration.
Morphological changes in tracheal epithelial
cells: No evidence of dust particles was observed
(EHC-93 or Ti02) in the epithelial cell cytoplasm at 2
h. No evidence of morphologic cell damage from
particles was observed.
Colchicine: Treatment with colchicine did not
prevent NF-KB activation.
Inhibitors/Activators: Tetramethylthiourea (TMTU)
(membrane-permeable active oxygen scavenger),
Deferoxamine (redox-inactive metal chelator), PPS
(Src inhibitor) AG1478 (epidermal growth factor
receptor inhibitor) prevented NF-KB activation in
both EHC93 and TiFe exposed-cells. Iron-containing
citrate extract of both dusts increased NF-KB
activation in both EHC93 and TiFe exposed-cells.
Reference: Courtois et PM (SRM 1648)
al. (2008, 1563691
(63% in, 4-7% , mass fraction >1 %: Si,
Species: Rat S, Al, Fe, K, Na)
Strain: Wstar
UF carbon black (FW2, P60)
Route: Cell Culture
Dose/Concentration: 100, 200 pg/mL
Time to Analysis: 24 h incubation
Cell Line: Dissected Particle Size: SRM 1648 mean
intrapulmonary arteries diameter 0.4 pm; ultrafme carbon black:
from rats used in FW2-13 nm, P60- 21 nm
corresponding in vivo
experiments
NO: Generally, Ach-induced relaxation in
intrapulmonary arteries decreased, Ach-induced
cGMP accumulation decreased, and relaxation by
SNP or DEA-NO also decreased. UF carbon black
did not affect NO responsiveness.
Oxidative Stress, Inflammatory: Dexamethasone
prevented SRM 1648-induced impairment of the Ach
relaxation response but antioxidants did not. TNF-a,
MIP2, IL-8 increased. ROS was not affected.
Reference: Dagher et
al., (2007, 0975661
Species: Human
Cell Type: L132
(Normal Lung
Epithelial Cells)
LC10, LC50 = PM25
(collected Jan-Sept in Dunkerque,
France)
Particle Size: cumulative frequency:
0.5 pm: 34%; 1 pm: 64%; 1.5 pm: 79%;
2 pm: 87%; 2.5 pm: 92%; 5 pm: 98%;
10|jm:100%
Route: Cell Culture (3x106,1.5x106,
0.75x106cells/20mL)
Dose/Concentration: LC10:19 pg/mL;
LC50: 75 pg/mL
Time to Analysis: 24, 48 or 72 h
p66 Protein: Phosphorylation of p65 increased in
PM-exposed L132 cells in dose-dependent manner.
kBo Protein: Phosphorylated iKBa protein
concentrations increased in cytoplasm with both
particle types at all time points.
p66 and p60 DMA: p65 DNA binding increased at
24 h with LC10 and LC50, at 48 h with LC10, and at
72 h with LC10 and LC50. p50 DNA binding
increased at all time points with LC10 and LC50.
December 2009
D-41
-------
Study
Pollutant
Exposure
Effects
Reference: Dai et al.
(2003, 0879441
Species: Rat
Strain: SD
Weight: 250 g
Cell Type: Tracheal
Explants
EHC-93 (Environmental Health Center,
Ottawa)
DEP: SRM 1650a (NIST)
Particle Size: EHC-93: 3-4 pm
(MMAD); DEP 1.55 ± 0.04 pm (CMD)
Route: Cell Culture
Dose/Concentration: ECH, DEP: 500
pg/cm2
Time to Analysis: Exposed for 1 h.
Parameters measured following a 7 day
incubation period.
Hydroxyproline: EHC93 induced an almost 3 fold
increase in explant hydroxyproline. DEP increased
tissue hydroxyproline 2.5 fold.
Procollagen: EHC-93 doubled gene expression of
procollagen. Procollagen gene expression could be
fully inhibited by SN50, TMTU or treatment of the
PM with DFX. Treatment of explants with p38 or
ERK (inhibitors) had no effect on procollagen
expression. DEP induced an increase in procollagen
gene expression but this increase was completely
prevented by SN50 and MAP kinase inhibitors
(SB203580 and PD98059). Neither TMTU or DFX
has any effect.
TGF01: Treatment of explant with EHC93
approximately doubled gene expression for TGF|31.
Treatment with SN50, TMTU and fetuin (TGFfS
antagonist) blocked increase. DFX, MAP kinase
inhibitors (SB203580 and PD98059) had no effect.
DEP roughly doubled TGF|31 expression.SN50 and
MAP kinase inhibitors (SB203580 and PD98059)
fully blocked this effect. TMTU and DFX had no
effect.
Reference: Doherty et Ratios of: V: Fe; Al: Fe; I
al. (2007, 0965321
Species: Rat
Strain: NR8383
Cell Types: AMs
V = sodium vanadate (NaV03)
Al = aluminum chloride hexahydrate
(AICI3)
Mn = manganese chloride tetrahydrate
(MnCI2)
Fe = ferric chloride hexahydrate (FeCI3)
Ratios based on PM25 measurements
from NYC, LA and Seattle
Particle Size: Metals from PM25
samples
: Fe Route: Cell Culture (2x105 cells/ml)
Dose/Concentration: Fe = 16 pmol
(equivalent to urban NYC 500 pg PM25); V
and Mn tested in molar rations of 0.02 to 0.4
relative to Fe; Al tested in molar ratios of
0.125 to 8 relative to Fe.
Time to Analysis: 20 h
IRP: Addition of V increased IRP activity 5 to 9 fold.
Though there was no seeming dose responsivity,
IRP activity remained strongly elevated over the
range of V:Fe ratios tested. Addition of Mn only
resulted in an effect at 0.1 molar ratio (two-fold), not
at higher or lower ratios. Al resulted in peak
increases of 5 fold at molar ratios 2 while declining
to 2 fold at molar ratios 4 and 8.
Cytotoxicity: Al was cytotoxic at molar ratios of 4
and 8. All other Al, V, Mn ratios had no effect.
Mixtures: The combination of metals tested at NYC
PM ratios and V drove all the Fe transport activity.
Combinations of V+Mn and V+AI increased activity
more than V:Fe alone.
Reference: Doornaert, DEP: SRM 1650 (NIST)
et al. (2003, 1564101
CB: (Sigma, France)
Species: Human
Cell Line/Type:
16HBE140-; P-HBE
DPC: Dipalmitoyl phosphatidylcholine
(positive control)
0.5 um
Particle Size: NR
Route: Cell Culture
Dose/Concentration: DEP and CB: 1-100
pg/mL
Time to Analysis: Parameters measured
24, 48, 72 h post exposure. 1-HBE Cell
Deadhesion Capacity: 24 h, evaluation of
detachment performed every 5min for 40
min after. Cell V\found Repair Capacity: 24 h,
repair evaluated 3.5, 7, 24 h after.
Cytotoxicity: DEP was cytotoxic at 100 pg/mL at all
time points in a time-dependent manner. CB and
DPC Cytotoxicity was substantially lower but
significant at 72 h.
Phagocytosis: 1-HBE cell levels that were in
contact with DEP or CB or have phagocytized those
particles increased in a dose-dependent manner.
DEP induced greater levels of cell contact and
phagocytosis than CB.
F-actin: Only DEPs were engulfed by F-actin
stained cell fragments.
Actin CSK Stiffness: DEP (5, 20,100 pg/mL)
induced net dose-dependent decrease in
cytoskeleton stiffness and a dose-dependent
decrease in actin cytoskeleton stiffness. CB
produced no significant decrease.
Adhesion Molecules: DEP induced a concomitant
reduction of both CD49 (a3) and CD29 (Ł1) integrin
subunits and a decrease in level of CD44 (HBE cell-
cell and cell-matrix adhesion molecule) at both 20
and100|jg/mL
Proteases: DEP also induced an isolated decrease
in MMP-1 expression without change in tissue
inhibitor of TIMP-1 or TIMP-2 at 100 fjg/mLCB
produced no change or insignificant results.
1-HBE Cell Deadhesion Capacity: DEP exposure
induced a dose-dependent amplification of cell
detachment at 5 min of incubation and onward.
Cell Wound Repair Capacity: DEP inhibited wound
repair/wound closure in a dose-dependent manner.
December 2009
D-42
-------
Study
Pollutant
Exposure
Effects
Reference: Dostert et Asbestos
al. (2008, 1557531
Species: Human
Cell Line/Type: THP1,
monocyte-derived
macrophages (MM)
Silica
DEP
CSE: cigarette smoke extract
MSU: monosodium urate crystals
Particle Size: NR
Route: Cell Culture
Dose/Concentration: Asbestos: 0.1, 0.2
mg/mL; Silica: 0.1, 0.2, 0.25, 0.5 mg/mL;
DEP: 0.2, 0.25, 0.5 mg/mL; CSE: 5%, 10°/<
in solution mg/mL; MSU: 0.1, 0.2 mg/mL
Time to Analysis: 1, 3, 6h
IL-10: Increased levels of IL-1|3 with asbestos and
silica were observed in THP1 at 6 h. CSE and DEP
had no effect. MM also had increased levels with
asbestos, silica and MSU at high dose levels only.
Caspase-1: Asbestos increased caspase-1 activity.
ROS: Asbestos doses in THP1 exhibited an
increase in ROS formation.
Reference: Doyle, et
al. (2004, 0884041
Species: Human
Cell Type: A549 from
non-smoking adults
BD: 1,3-butadiene, known carcinogen
Acrolein: photochemical and NO
product of BD in atmosphere
Acetaldehyde: photochemical and NO
product of BD in atmosphere
Formaldehyde: photochemical and NO
product of BD and ISO in atmosphere
ISO: isoprene, 2-methyl analog of BD
Methacrolein: photochemical and NO
product of ISO in atmosphere
Methyl vinyl ketone: photochemical and
NO product of ISO in atmosphere
Particle Size: NR
Route: Environmental Irradiation (smog)
Chambers
Cytotoxicity: ISO+NO and BD+NO induced small
increases of LDH inA549. However, ISO+NO+light
and BD+NO+light increased LDH levels 4-6 fold
Dose/Concentration: 50 ppb NO; 200 ppbV indicating photochemical products of ISO and BD
are highly cytotoxic. LDH levels of each combination
were equivocal.
ISO, BD
Time to Analysis: Exposed to gases for 5 h.
Analysis 9 h post exposure.
IL-8 Protein: Methacrolein, methyl vinyl ketone and
formaldehyde (products of ISO) increased IL-8
protein levels significantly. ISO+NO had no effect.
BD photochemical products (acrolein, acetaldehyde
and formaldehyde) also increased IL-8 protein, more
than doubling the photochemical products induced
by ISO. BD+NO had no effect.
IL-8 mRNA: IL-8 mRNA expression also increased
with photochemical products of ISO and BD but did
not reach a statistically significant level.
Reference: Duvall et
al. (2008, 0979691
Species: Human
Cell Type: Airway
Epithelial Cells
PM-F, -C, -UF
Particles collected from: Seattle, WA
(PM-S); Salt Lake City, UT (PM-SL);
Phoenix, AZ (PM-P); South Bronx, NY
(PM-SB); Hunter College, NY (PM-
Sterling Forest, NY (PM-SF)
Particle Size: Coarse: >2.5 pm; Fine:
<2.5|jm;UFP:<0.1 pm
Route: Cell Culture (100,000 cells/cm')
Dose/Concentration: 5 mg/ml
Time to Analysis: 1, 24 h post exposure
Particle Characterization: PM-HR, PM-SL and PM-
S contained the highest UF, F, and C concentrations.
PM-SB and PM-HR had similar F and C
concentrations. Sulfate was highest in PM-F for all
sites except in PM-SB and PM-HR. V\food
combustion was highest in PM-SL, PM-S, PM-P. Soil
dust was highest in PM-SL and PM-S.
IL-8: PM-UF induced a greater increase in IL-8 than
other types of PM except PM-P. PM-UF is
associated with vanadium, lead, copper, sulfate. PM-
F-HR caused the greatest increase followed by PM-
SB. PM-F-SF and PM-F-P was least effective. PM-C
also caused an increase in IL-8 levels and was
associated with vanadium and EC.
COX-2: PM-F-S induced the greatest increase in
COX-2 expression. Other PM-F sites induced similar
increases. UF PM had no effect. PM-C, associated
with EC, induced increases.
HO-1: PM-F-SF induced the greatest increase in
HO-1. PM-F-SL was the least effective. UF PM had
no effect. PM-C, associated with copper, barium and
EC, caused an increase.
Reference: Dybdahl et DEP: SRM 1650 (NIST)
al. (2004, 0890131
Particle Size: 90 nm (MMAD)
Species: Human
Cell Type: A549
Route: Cell Culture (105cells/mL)
Dose/Concentration: 0,10, 50,100, 500
pg/mL
Time to Analysis: 2, 5, or 24 h
Cytokines: DEP induced dose-dependent increases
of IL-1a, IL-6, IL-8 and TNF-a at 24 h. Cytokines
increased between 4 and 18 fold at the highest DEP
dose as compared to controlled cells. DEP also
increased IL-6 mRNA expression levels in a dose
and time-dependent manner.lL-6 mRNA levels
increased 14 fold at 24 h, 8 fold at 5 h, and 2 fold at
2h.
Cell Viability: DEP exposure did not decrease cell
viability at any dose tested.
December 2009
D-43
-------
Study
Pollutant
Exposure
Effects
Reference: Fritsch et
al. (2006, 1564521
Species: Mouse
Cell Type: RAW 264.7
MAF02: incinerator fly ash (collected by Route: Cell Culture (1 xio cells/well)
electrostatic precipitation in commercial
municipal waste incinerator facility)
composition representing 12% of total
mass (mg/g):
Fe (9.1); Pb (23.3); Zn (75.7); C (7.5)
Particle Size: 165 nm (modal value)
Dose/Concentration: 6.3-188 pg/cm for
Toxicity; 2.6, 6.5,13.2 pg/cm2 for
ArachidonicAcid; 13.2 pg/cm2 for MARK
Pathway; Other doses noted in Effect of
Particles
Time to Analysis: 1,2.5, 5, 24 h
Toxicity: Viability decreased from 99% to 18% at
62.5-188 pg/cm2. Lower doses had no effect.
Arachidonic Acid: At 2.5 h, AA level increased 2
fold for 6.5 pg/cm2 and 6 fold for 13.2 pg/cm2. No
increase was observed after 5 h.
MAPKs: Cells pretreated with PD98059, an inhibitor
of MEK-1, inhibited AA liberation due to MAF02
treatment of 13.2 pg/cm2
COX-2: A time-dependent increase of COX-2 protein
expression was exhibited at 2.5 and 5 h.
ROS: A dose-dependent increase in ROS formation
was observed at concentrations greater than 31.3
pg/cm after 3 h.
GSH: There was an observed increase of production
at 20 h. Doses greater than 60 pg/cm2 reduced total
glutathione.
HO-1: There was an observed dose-dependent
increase in expression at 4 h.
Reference: Fujii et al. PMi0: EHC-93 (Ottawa, Canada)
(2002,036478)
Species: Human
Cell Type: HBEC
(from current
smokers), AMs, Co-
Culture: AMs+HBEC
Age: HBEC: 48-70 yr
Route: Cell Culture (HBEC: 2.5-3x 106
cells/well); (AMs: 1.0x107 total)
Dose/Concentration: 100, 500 pg/mL
Time to Analysis: 2, 8, 24 h
Viability: Over 90% of HBEC were viable after a 24
h exposure of up to 500 pg/mL of PM. AMs
incubated with and without 100 pg/mL saw no
significant difference in viability.
Cytokine mRNA: TNF-a, GM-CSF, IL-lp, IL-6, LIF,
OSM and IL-8 mRNA expression increased in co-
culture with 100 pg/mL at 2 and 8 h. In AMs, TNF-a,
IL-1|3, IL-6 mRNA expression increased with 100
pg/mLat2h. INHBECs, IL-lp and LIF increased
with 100 pg/mL at 2 h. HBECs added to AMs
exposed to PM10, further increase in mRNA of IL-1|3,
LIF and IL-8.
Cytokine Protein: In co-culture and AMs, significant
increase in protein production of GM-CSF, IL-8, IL-
lp, IL-6 and TNF-a in dose-dependent manner. GM-
CSF and IL-6 production significantly higher in co-
culture then AM or HBEC alone.
Bone Marrow: Co-culture instillation of
supernatants increased circulating band cell counts
at 6 and 24 h with 100pg/mL
Reference: Fujii et al.. PM10:EHC93 (Ottawa, Canada)99%
(2001, 1564551 <3.0um
Species: Human
Cell Type: HBEC from
current smokers
Age: 48-70 yr
Particle Size: PM10( 99% < 3.0 pm)
Route: Cell Culture (2.5-3x1 Ob cells/dish)
Dose/Concentration: 10,100, 500 pg/mL
Time to Analysis: 2, 8, 24 h
Phagocytosis: 18.6% of cells engulfed particles
when exposed to 100 pg/mL. Over 90% remained
viable.
Cytokine mRNA: LIF mRNA increased dose-
dependently at 2 h but declined at 8 and 24 h. GM-
CSF increased dose-dependently at 8h and peaked
at 24 h. IL-1a increased at 2 h, increased dose-
dependently at 8 h and peaked at 24 h. M-CSF,
MCP-1, IL-8 were unaffected.
Cytokine Protein: LIF, GM-CSF, IL-1|3 and IL-8
increased dose-dependently. Soluble fraction of 100
pg/mL PMio did not affect cytokine production.
December 2009
D-44
-------
Study
Pollutant
Exposure
Effects
Reference: Garcon et
al. (2006, 0966331
Species: Human
Cell Type: L132
PM25 (collected in Dunkerque, France
for 9mo, Jan-Sept)
Particle Size: PM25- 0-0.5 pm
(33.63%), 0.5-1.Opm (30.61%), 1.0-1.5
pm (14.33%), 1.5-2.0 pm (8.69%), 2.0-
2.5 pm (4.89%), >2.5 pm (7.87%)
Route: Cell Culture: 3x10b cells/20ml (24 h);
1.5x106cells/20ml(48h); 0.75x106
cells/20ml (72 h)
Dose/Concentration: 18.84, 37.68, 56.52,
75.36,150.72 pg/mL; LC10-18.84 pg/mL;
LC50- 75.36 pg/mL
Time to Analysis: 24, 48 or 72 h
Cytotoxicity: PM induced dose-dependent
(R2 =.9907) cytotoxic effect in proliferating L132
cells.
LDH: Increase at 72 h with 56.52 and 75.36 pg/mL.
Oxidative Stress: A decrease in MDF activity was
observed at all exposure levels at 24, 48, and 72 h
(72-h <5 % of control). MDA levels showed increase
concentration after 72 h, both LC10 and LC50. LC10
and LC50 saw an increase in SOD activity at 24 h;
LC50 saw a decrease in activity after 48 and 72 h. 8-
OHdG and PARP exhibited increases at all time
points with LC10 and LC50.
Inflammatory Response: Increases of TNF-a
concentration was exhibited at 24 h at LC50, and at
48 h and 72 h at LC10 and LC50. iNOS activity
increase at all time points at LC10 and LC50.NO
concentration exhibited increases at all time points
after exposure to LC10 and LC50.
Reference: Geng et
al. (2005, 0966891
Species: Rat
Strain: Wistar Kyoto
Tissue/Cell Type:
Lung macrophages
BPM: Blowing PM25; PM collected from Route: Cell Culture
Wuwei City, Gansu Province, China
(Blowing days correspond to desert
storm days)
NPM: Non-blowing (normal) PM25
Particle Size: PM25
Dose/Concentration: 0, 33,100, 300 pg/mL
Time to Analysis: 4 h
Cytotoxicity: Dosages greater than 150 pg/mL
decreased cell viability.
Plasma Membrane Fluidity: Dose-dependent
decrease had no effect on membrane lipid
hydrophilic region.
Plasma Membrane Permeability: LDH enzyme
activity and extracellular AP activity increased dose-
dependently, indicating increased membrane
permeability, but this was only statistically significant
at 300 pg/mL dose. NPM may affect some
parameters at 100 pg/mL Overall, NPM induced a
slightly higher increase than BPM.
Intracellular Ca2+: A dose-dependent increase was
observed.
Lipid Peroxidation (TBA): An increase was
observed only at 300 pg/mL
Antioxidant (GSH): A decrease was observed only
at 300 pg/mL
Reference: Geng et
al. (2006, 097026)
Species: Rat
Strain: Wistar Kyoto
Tissue/Cell Type:
Lung macrophages
DPM: dust storm samples; PM
collected from Baotou City, Inner
Mongolia, China in March 2004
NPM: normal PM
Particle Size: PM25
Route: Cell Culture
Dose/Concentration: 0, 33, 100, 300 pg/mL
Time to Analysis: 4 h
Cytotoxicity: MTT reduction assay revealed a
significant decrease in cell viability at 150 pg/mL and
300 pg/mL LDH enzyme activity significantly
increased at 150 and 300 pg/mL
GSH levels: Significant decreases were seen in
cellular GSH levels and increases in TBARS levels
in both groups with a 300 pg/mL dose.
Plasma Membrane Activity: In the plasma
membrane, Na, K-ATPase were significantly
inhibited. Ca *Mg *-ATPase were unaffected.
Plasma Membrane Lipid Fluidity: Results indicate
that DPM could increase the surface fluidity of
membrane lipid.
Intracellular Ca2+: A dose-dependent increase in
free intracellular Ca2+ levels was observed.
Reference: Ghio et al.
(2005, 0882721
Species: Human
Cell/Tissue Type:
BEAS-2B
FAC: ferric ammonium citrate
(component of ROFA)
VOS04: vanadyl sulfate
(component of ROFA)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 100 pM FAC -
preexposed before metal compounds or oil
fly ash
50 pM VOS04 - preexposed before metal
compounds or oil fly ash
100|jg/mLROFA
Time to Analysis: 0-1 h, 4 h
IRE DMT1: FAC increased mRNA and protein
expression for -IRE DMT1. VOS04 decreased
mRNA and protein expression for-IRE DMT1. +IRE
DMT1 unaffected by any treatment.
Metal transport: Uptake of iron increased after pre-
exposure to FAC and decreased after pre-exposure
to VOS04. Pre-exposure to FAC again increase the
uptake of both iron and vandium. VOS04 induced
opposite effect, decreasing Fe uptake.
ROS: Increased acetaldehyde, indicating increased
oxidative stress. ROS decreased with FAC
pretreatment. ROS increased with VOS04
pretreatment.
December 2009
D-45
-------
Study
Pollutant
Exposure
Effects
Species: Rat
Strain: SD
Reference: Gilmour et Coal Fly Ash
al. (2004, 0574201 MU = Montana Ultrafine
MF = Montana Fine
MC = Montana Coarse
KF = V\fest Kentucky Fine
KC = V\fest Kentucky Coarse
Cell/Tissue Type: AM Cog| combus(ion usjng g |aboratory.
scale down-fired furnace rated at
50kW. Montana subbituminous coal
and western Kentucky bituminous coal
Particle Size: Coarse: >2.5 pm; Fine:
<2.5|jm;UFP:<0.2|jm
Route: Cell Culture (2x1 Ob cells/ml)
Dose/Concentration: 125 pg/mLor250
pg/mL
Time to Analysis: 4 or 24 h
LDH: Mid and high doses of Montana ultrafine
particles showed significant increase after 4 h
exposure vs control. Other particle types had no
effect. After 24 h, LDH level was not statistically
significant between particles tested and control.
Cytokines: Treatment with Montana ultrafine
particles resulted in a significant production increase
of TNF-a. MIP-2 showed increases in all the fine and
ultrafine treatments, with Montana ultrafine and W.
Kentucky fine PM showing the highest increases. IL-
6 increased with Montana ultrafine particles although
there was some variability and the increases were
not statistically significant.
Reference: Gilmour et PM10: Collected from the Marylebone
al. (2005, 0874101 and Bloomsbury monitoring sites in
London, UK
Species: Human
Cell/Tissue Types:
monocyte derived
macrophages,
HUVECs, A549,
16HBE
Particle Size: PIvl-
Route: Cell Culture
Dose/Concentration: 50 |jg/mL
Time to Analysis: 4 h, 6 h, 20 h
IL-8: PM10 at 50 pg/mL induced a significant
increase in IL-8mRNA and protein expression in
PMM and 16HBE at 6 and 20h. A less substantial
increase was also observed in A549.
Procoagulant Activity: PMi0 induced a significant
decrease in macrophage mediated clotting time in
16HBE. Other cell types were unaffected.
Annexin V Binding: At 100 pg/mL, PM10 induced a
significant increase in binding macrophages at 4 and
20 h. There was no effect at 50 pg/mL
Tissue Factor mRNA Expression: Expression was
increased in macrophages at 6 h only.
tPA Expression: mRNA expression decreased at
6 h. Protein expression decreased at 4 h and 20 h in
a dose-dependent manner.
TF Expression: TF mRNA expression increased in
a dose-dependent manner at 6 h in HUVECs.
Protein levels also increased at 4 h but declined to
basal levels by 20 h.
Reference: Gilmour et PM10: Collected from the Marylebone
al. (2003, 0969591 and Bloomsbury monitoring sites in
London, UK
Species: Human
Cell/Tissue Type:
A549
TSA
H202
NAC
Mannitol
Provided by Sigma Chemical, Poole,
UKorGIBCO-BRL, Paisley, UK
Particle Size: PM10
Route: Cell Culture
Dose/Concentration: PM10:100 pg/mL;
TSA: 100 ng/mL; H202: 200 pM; NAC and
Mannitol: 5mM
Time to Analysis: 24 h
IL-8: PMio, TSA and H202 treatment induced an
increase of IL-8. Concomitant exposure of TSA with
PM10 or H202 significantly increased IL-8 release
when compared to PM10 or H202 alone. IL-8 mRNA
expression with PMio or H202 exposure and TSA
coincubation caused significant increases. Silver
staining of PCR products indicated that the IL-8
gene promoter was associated with acetylated H4
following TSA, PM10 and TNF treatment.
H4: PMio exposure significantly increased
acetylation levels of H4 over controls. Increased
acetylated H4 was mediated by PM10 in a dose-
dependent manner. Treatment with PMio and H202
increased HAT activity associated with H4 by 245%
and 166% respectively. Significant increases in
acetylation of H4 following treatment of cells with
TSA, PM,o and H202 for 24 h was observed.
PM10induced HAT activity was significantly
decreased in the presence of NAC and mannitol.
Nuclear presence of HDAC2 protein was
significantly reduced by exposure to both HDAC
inhibitor and PMio. There was a decreasing trend in
HDAC2 gene expression following TSA and PM10
treatment.
NF-KB: The activation of the transcription factor NF-
KB was enhanced following the inhibition of HDAC
with TSA and by treatment with
December 2009
D-46
-------
Study
Pollutant
Exposure
Effects
Reference: Graffetal. PM
(2007, 1564881
Species: Human
Cell/Tissue Type:
HAEC
-UF: ultrafine
-F: fine
-C: coarse
Particles collected from Seattle, WA (-
S), Salt Lake City, UT (-SL), Phoenix,
AZ (-P), South Bronx, NY (-SB), Hunter
College, NY (-H), Sterling Forest, NY (-
SF)
Particle Size: UF:<0.1 pm; F: 0.1-2.5
pm;C: 2.5-10 pm
Route: Cell Culture
Dose/Concentration: 250 |jg/mL
Time to Analysis: 6 h, 24 h
Gene Expression: PM-UF, PM-F, and PM-C both
upregulated and downregulated genes in the HAECs
though downregulation was far more common for all
the three PM fractions. PM-F affected the greatest
number of transcripts, followed by the UF and C
fractions.
IL-8: mRNAexpression increased, with PM-F-S
having the greatest impact. Aluminum, strontium,
manganese and potassium were highly associated
with expression. Wood combustion was moderately
associated.
HOX-1: mRNA expression increased, with PM-F-SF
having the greatest impact. Potassium, manganese,
strontium and wood combustion were highly
associated with expression. Aluminum and
vanadium were moderately associated.
Reference: Gualtieri et TD: Tire debris extracted in methanol,
al. (2005, 0978411 constituent of PM,0
Species: Human
Cell Type: A549
(generated by spinning a new
automotive tire against abrasive
surface)
Particle Size: 10-80 pm
Route: Cell Culture
Dose/Concentration: 10, 50, 60, 75 pg/mL
Time to Analysis: 24, 48, 72 h
Cytotoxic Effect: Treated cells presented inhibitory
effect on reduction of MTT which appeared to be
dose and time-dependent. A statistically significant
reduction was observed at 48 and 72 h. Trypan blue
showed a significant PM lethality as well as a dose-
dependent increase in mortality.
DMA Damage: At 24 and 72 h, DNA damage
increased dose dependently in damaged and ghost
cells.
Cell Cycle Analysis: At 24 h, TD extract-treated
cells presented a significant increase in the
percentage of cells in G1 phase when com pared
with control. This increase was associated with a
decrease in the percentage of cells in S phase. At 48
and 72 h, the increase in percentage of cells in G1
was associated with a decrease in the percentage of
cells in both S and G2/M phases. Cells exposed to
TD extracts presented changed morphology.
Modifications most obvious at 72 h. The highest dost
produced increased vacuolization in cytoplasm and
apoptotic nuclear images.
Reference: Hetland et
al. (2005, 0878871
Species: Rat
Gender: Male
Strain: Crl/Wky
Cell Type: AMs
PMC = Coarse
PMF = Fine
-A = Amsterdam
-L=Lodz
-R = Rome
-0 = Oslo
Coexposures PAH, Fe, Al, Zn, Cu, V
Particle Size: PMC: 2.5-10 pm; PMF:
0.2-2.5 pm
Route: Cell Culture (1.5x10b cells/well)
Dose/Concentration: 50,100 pg/mL PM
Time to Analysis: 20 h
IL-6: PMC from all cities exhibited increases in IL-6
release with spring and summer roughly equal and
both inducing higher levels than the winter PMC. For
the Spring and Summer samples, PMC-L exhibited
the highest IL-6 releases (440% and 460%
respectively) followed by Rome, Adam/Oslo, and
Oslo/Adam. For the winter samples, Rome and
Amsterdam induced higher IL-6 levels (340% and
300% respectively) than Lodz and Oslo (165% and
160%). The fine fractions did not induce any
significant cytokine release.
TNF-a: PMC from all cities increased TNF-a release
with 50 pg/mL generally inducing a slightly higher
increase than 100 pg/mL
Constituent Correlation: Levels of Fe, Al, Zn, Cu
and V as well as PAH (total and fractions) showed
no correlation with IL-6 release.
Endotoxin Correlation with IL-6 release: A
confirmatory test revealed no correlation.
December 2009
D-47
-------
Study
Pollutant
Exposure
Effects
Reference: Hetland et
al. (2004, 0975351
Species: Rat, Human
Cell Type: Alveolar
Macrophages (Rat),
A549
Strain: Wky/NHsd
Gender: Male
Weight: 180-230 g
AMC = Ambient Coarse
AMF = Ambient Fine
AMUF = Ambient Ultrafine
(AM samples taken at a suburban site,
without a dominating PM source, near
Utrecht, Netherlands)
Road PM: PM10, (collected in a road
tunnel with predominating road
abrasion due to use of studded tires in
Trondheim, Norway)
Particle Size: AMC: 2.5-10 pm; AMUF:
<0.1 pm
Route: Cell Culture (1 xlOb cells/well)
Dose/Concentration: 0,100, 200, 400, 600,
800,1000pg/mL
Time to Analysis: 20h (Type 2 cells); 40h
(A549 cells)
IL-8: All 3 AM fractions showed dose-dependent
increases in A549 cells until 600 pg/mL; at that
concentration, levels declined. AMC showed the
most pronounced decline which correlates with
decreased viability. Road PM showed a near linear
response until 1000 pg/mL, whereas DEP plateaued
at600|jg/mLinA549.
MIP-2: AMC and AMUF had no effect on Type 2
cells. DEP induced increases at 200 pg/mL,
whereas Road PM induced the strongest increase,
peaking at 600 pg/mL in Type 2 cells.
IL-6: AMC induced increases at 100 pg/mL in Type
2, but levels declined below normal at 200 pg/mL.
AMUF induced a decline of IL-6 levels. Road PM
induced significant increases in Type 2. DEP had a
slight effect. AM fractions induced increases in A549
cells, peaking at 600 pg/mL with AMF DEP and
Road PM induced a dose-dependent increase.
Cell Survival: AMC showed major effects at 200
pg/mL in Type 2. AMUF showed effects at 400
pg/mL Road PM and DEP showed a gradual
decline from 75% to 50% at 800 pg/mL in Type 2. All
AM fractions induced a decrease in viability after
600 pg/mL in A549 with AMC inducing a larger
decrease than AMUF and AMF; AMUF and AMF
induced similar levels. Road PM and DEP had no
effect on A549.
Apoptosis: AMC elicited a marked induction of
apoptosis 200 pg/mL in Type 2 cells. AMF showed a
dose-dependent increase in A549. Other AM
fractions showed some slight increases in both cell
types. Statistical significance was reached for all
particles except for Road PM.
Reference: Holder et
al. (2008, 0933221
Species: Human
Cell Type: 16HBE14o
DEP: generated from a single cylinder
diesel engine using , commercial
certified #2 diesel fuel
Copollutants: NOX7 ppm, C020.1%
Particle Size: Suspension: 223 nm
(mean diameter); All: 122 nm (mean
diameter)
Route: Suspension (1 xio cells/cm 1, Air
Liquid Interface (All, 1xl05 cells/cm)
Dose/Concentration: Suspension: 0.13,
0.24,1.88, 2.5, and 12.5 pg/cm2; All: 1
g/cm3 (total number of particles: 2.3x107
particles/cm2)
Time to Analysis: Exposure for 6 h.
Parameters measured 20 h post-exposure.
All vs Tracheal Bronchial (TB) Deposition: The
TB region deposition is 1.5 nominally x All, but
particle diameter deposited in the TB was 62 nm
(geometric mean diameter) as compared to the
particle deposition in the All, measuring 260 nm.
Inflammatory Response: Suspended DEP
decreased viability at concentrations of 2.5 pg/cm or
higher. IL-8 release (corrected for viability) increased
at concentrations of 1.88 pg/cm2 or higher in a dose-
dependent manner. IL-8 exhibited intermediate
levels of secretion between in vitro levels of 0.25
and 1.88 pg/cm2. No statistically significant results
were observed in ALL Viability for ALI was near
100%(75%uncorrected).
Reference: Huang et
al. (2003, 0873761
Species: Human,
Mouse
Cell Type: BEAS-2B,
RAW 264.7
PMC: PM coarse
PMSM: PM submicron
Collected between September-
December 2000 from 4 ambient
monitoring stations in Taiwan that
represented background, urban, traffic,
and industrial sites
Particle Size: PMC: 2.5-10 pm; PMF:
1-2.5pm; PMSM: <1 pm
Route: Cell Culture (5x1 Ob cells/mL)
Dose/Concentration: All PM: 50, 70,100
pg/mL
Time to Analysis: BEAS-2B: 8h; RAW
264.7:16 h
Viability: None of the PM fractions affected cell
viability.
IL-8: Only PMSM induced a significant IL-8 increase
in BEAS-2B. IL-8 response was associated with a
combination of Mn and Cr (R2 = 0.28). Response
was also correlated with nitrate, although
significance disappeared when 1 extreme nitrate
value was removed.
Lipid Peroxidation: Only PMSM enhanced lipid
peroxidation in BEAS-2B, correlating with both
elemental and.
TNF-a: In RAW264.7, PMSM increased TNF-a
production. Polymixin pretreatment significantly
reduced TNF-a levels for all 3 PMs which indicates
an endotoxin role in macrophage response. TNF-a
production (after polymixin pretreatment only) was
associated with Cr and Fe content.
December 2009
D-48
-------
Study
Pollutant
Exposure
Effects
Reference: Hutchison PMi0: Samples collected for 7 day
et al. (2005, 0977501 during closure (-C) and reopening of
steel plant (-R)
Species: Mouse
CellLine:J774.1A
PMT: PM total (aqueous sonicate)
PMS: PM soluble aqueous
PMI: PM insoluble aqueous
Particle Size: PIvl-;
Route: Suspension
Dose/Concentration: 500 |jl (estimated
concentrations of 112, 143,156,180, 233,
255 pg/1 ml water)
Time to Analysis: 4 h
Particle Characterization: Reopening of the plant
showed a significant increase in the total and acid
extractable metal content of PM. Aqueous
extractable metal content did not change. Soluble
zinc, copper and manganese also increased
significantly post reopening. Iron was the most
abundant in acid extractable metals and increased
greatly at the reopening.
TNF-a: PMT-R and PMT-C induced a statistically
significant increase. Treatment with chelation agent
reduced effect to control levels.
Reference: Imrich et
al. (2007, 1558591
Species: Rat
Gender: Female
Age: 12-14 wk
Cell Type: AM
UAP:SRM 1649 (positive control)
Ti02: Particle control
CAPs (Boston, MA)
All cells primed with IPS
Coexposure with MAC,
dimethylthiourea (DMTU), H202 or
catalase
Particle Size: CAPs: <2.5 pm; UAP:
PM25; Ti02: ~1 pm
Route: Cell Culture (2x105 cells/well)
Dose/Concentration: Caps 100 pg/mL;
UAP: 50 or 100 pg/mL; IPS 250 ng/mL;
MAC, DMTU: 2,10, 20 mM; Catalase: 1, 5,
10 mM; H202 0-50 pm/hr
Time to Analysis: 18-20h
TNF-a: DMTU at 20 mM reduced TNF in LPS-
primed cells in control and UAP-treated groups.
MAC at 20 mM reduced TNF release but this was
not statistically significant. Catalase significantly
inhibited TNF in control and UAP-treated groups.
CAPs (especially the insoluble portion) significantly
increased TNF unless co-exposed with NAC, DMTU
or catalase. All three reduced levels back to around
basal levels. DMTU was particularly effective at
diminishing TNF release. H202 increased TNF
release in CAPs-exposed cells. Ti02 had no
increased ability to induced cytokine release when
mixed with H202.
Cell Death: Viability decreased substantially when
exposed to H202 + CAPs. The soluble fraction of
CAPs showed to be more effective with H202 than
the insoluble portion. Ti02 had no significant effect.
NO: Some CAPs induced slight increases when
mixed with H202. No difference was observed
between soluble and insoluble portions of CAPs.
DFO: DFO at 0.05 mM completely inhibited
oxidation induced with soluble CAPs + H202.
Insoluble CAPs + H202 was also DFO-sensitive.
DFO was ineffective against the insoluble CAPs
induction of TNF and MIP-2.
Reference: Ishii et al.
(2004, 0881031
Species: Human
Cell Type: A549
(collected from 6
lobectomy or
pneumonectomy
smokers), HBEC
EHC-93:PM10 (obtained from
Environmental Health Directorate,
Ottawa, Ontario, Canada)
Particle Size: PM10
Route: Cell Culture (1x!07 cells)
Dose/Concentration: 100 |jg/mL
Time to Analysis: 3, 6, 24 h
Cytokines: TNF-a, IL-lp, GM-CSF, IL-6, and IL-8
levels were significantly increased in A549 cells.
mRNA Expression: MCP-1, ICAM-1 and IL-8
mRNA expression increased in untreated AM
supernatants at 3 h. Only the MCP-1 levels were
statistically significant at 3 h. Levels declined by 6 h.
When A549 cells were exposed to PM10 exposed
AM, levels of RANTES, TNF-a, ICAM-1, IL-1|3, and
LIF increased. Except for RANTES mRNA, these
differences were less in the 6 h samples. VEGF
increased as well, but this increase was not
statistically significant.
TNF-a and IL-1p-neutralizing Antibodies: L-1p
antibody alone or in combination with TNF-a
significantly reduced expression of all eight mRNAs.
Combinations for some mRNAs reduced expression
by up to 1/2. This effect was not observed when
A549 was treated with the control AM.
Transcription Factor Binding Activity: Binding of
AP-1 and Sp1 increased when A549 treated with
supernatants from PMio-exposed AM, but not from
control AM.
December 2009
D-49
-------
Study
Pollutant
Exposure
Effects
Reference: Ishii et al.
(2005, 0961381
Species: Human
Cell Type: AMs
(obtained from 10
EHC-93: PM10 (obtained from
Environmental Health Directorate,
Ontario, Canada)
Particle Size: PM10
Route: Cell Culture (HBEC: 2.5-3.0x1 06
cells; AM: 1 xio7 cells; co-culture of
AM/HBEC: 5x1 06 cells)
Dose/Concentration: 100 |jg/mL
Ti me to Analysis: 2, 24 h
mRNA Expression After 2 h Exposure: AM or
HBEC exhibited no effect. In contrast, co-culture
increased expression of MIP-lp, GM-CSF, M-CSF,
IL-6, MCP-1 and ICAM-1-mRNA.
mRNA Expression After 24 h Exposure: AMs
exhibited no effect. HBEC increased levels of GM-
smokers who stopped
smoking 6 wk prior),
HBEC
CSf, LIF and ICAM-1. Co-culture, on the other hand,
increased expression of MIP-lp, GM-CSF, M-CSF
and ICAM-1 mRNA.
Protein Levels: AM and HBEC both increased GM-
CSF, IL-6 and MIP-1 p release into the supernatant.
Co-culture effect was not additive but synergistic
(i.e., higher than expected). MCP-1 levels did not
increase significantly. Co-culture appeared to
decrease protein levels for both the control and PM
values. M-CSF levels increased for co-culture only.
Surface Expression of ICAM-1: Upon 24 h
exposure to PM, HBEC exhibited an increase in
expression. Expression in AMs were not affected by
2 h PM stimulation.
ICAM-1 Inhibitors: IgG or anti-CD! 1b antibody was
unaffected in co-culture.
Reference: Jalava et
al. (2005, 0886481
UPM: SRM1649a (Washington, DC)
DEP:SRM1650(NIST)
Species: Mouse
EHC-93: Ottawa dust (Environmental
Cell Type: RAW 264.7 Health Center, Ottawa, Canada)
HFP-00: Pooled ambient air PM25
sample from Helsinki, Finland
M-UPM: methanol extract of UPM
Particle Size: SRM 1649a, SRM 1650,
EHC-93: NR; HFP-00: PM25
Route: Cell Culture (5x105 cells/ml)
Dose/Concentration: 150 |jg/mL
Time to Analysis: Methanol treatment of
PM samples: 24 h; Exposure to ambient PM
samples: 2, 4, 8,16, or24h.
TNF-a: All the PM samples increased TNF-a.
Cell Viability: SRM1649a exhibited the most
cytotoxicity, followed by HFP-00 and EHC-93.
Methanol significantly affected cytotoxicity of the
EHC-93 sample only.
Cytokines: TNF-a concentrations in the cell culture
medium significantly increased at all time points
between 2 and 24 h. The highest increase was seen
in EHC-93. IL-6 production also increased at
different levels with the highest increase observed in
EHC-93. No response was observed for IL-10.
Cell Viability: Duration of exposures had no
significant effect on any of the samples. A 2 h
exposure time was sufficient to induce the typical
reductions in cell viability.
Reference: Jalava et
al. (2006, 1558721
Species: Mouse
Cell Type: RAW 264.7
PM: Collected east of Helsinki, Finland
between Aug 23 and Sept 23, 2002
Divided in 12 groups (4 sizes by 3
exposure types):
-S: seasonal average
-W: wildfire
-M: mixed
-B: blank
Particle Size: PMio-25;PM25-1; PMi.02;
PMo.2
Route: Cell Culture (5x106 cells/mL)
Dose/Concentration: 15, 50,150 and 300
pg/mL
Time to Analysis: 24 h
Particulate Mass Concentrations in HVCL Size
Ranges: The largest increase of PM concentrations
was observed in PM^.
NO: All 12 samples increased NO production when
compared to corresponding unexposed controls.
Peaks were observed at 150 pg/mL, except in PMi.
0.2'
Cytokines: All 12 samples increased TNF-a and IL-
6 production. PM10.2.5 and PM25.i produced a much
larger response than PM^and PM02. IL-6
production for PM02 was not measured. MIP-2
production also increased with similar trends.
Cytotoxicity: All 12 samples induced dose-
dependent decreases in cell viability. PM10-25 were
the least active inducers of apoptosis while PMO.2
showed the highest activity (4-17% of apoptotic
cells).
December 2009
D-50
-------
Study
Pollutant
Exposure
Effects
Reference: Jalava et
al (2007 096950)
'v --
Species: Mouse
Cell Type: RAW2647
Urban background PM
Route: Cell Culture (5x10 cells/ml)
PM10, PM2 5, and PM0.2 collected from Dose/Concentration: 15, 50, 150, 300
6 European cities during different times pg/mL
e year from October 2002 to July
-D: Duisburg (Fall)
-P: Prague (Winter)
-A: Amsterdam (Winter)
-HR: Helsinki (spring),
-B: Barcelona (spring)
-AT: Athens (summer)
Particle Size: PMi0: 2.5-10 pm; PM25:
0.2-2.5 pm; PM02: <0.2 pm
PM Characterizations: The highest mass
concentrations of PM10 and PM0 2 were measured in
Athens. Prague had the highest PM25
concentrations.
NO: All PM fractions induced statistically significant
NO production in macrophages. PM25 -P and PM25 -
AT produced significantly larger responses, though
all samples at 150 and 200 pg/mL induced
statistically significant production. When compared
to the other PM0 2 samples, -P and -HR produced
significantly larger responses.
Cytokines: PM10 showed average cytokine
production to be 7.8 fold and 83 fold for TNF-a, and
4.4 fold and 530 fold for MIP-2 when compared to
PM25 and PM02 respectively. PM10 induced
statistically significant increases in production of
TNF-a, MIP-2 and IL-6. PM25, with exception of
Prague, caused significant increases in cytokines.
PM02-A and -AT showed small yet statistically
significant increases in TNF-a. An increase in MIP-2
was observed with -P and -HR. IL-6 increased
significantly with PMi0 and slightly with PM2 5. In the
PMo.2 range, only the -A and -AT samples caused a
small, statistically significant TNF- a production.
MIP-2 production was only detected from the -P and
-HR samples. PM0 2 effects on IL-6 response were
negligible.
Cytotoxicity: The average cytotoxicity of PM10 and
PM25 were roughly equal, but PM02 were less
cytotoxic with the exception of -P. The dose-
response trends for most of the samples were
linearly declining, with PM10 and PM2 5 exhibiting
statistically significant declines in viability.
Reference: Jimenez et PMi0: Collected from London and
Edinburgh air particulate monitoring
stations.
al. (2002, 1566101
Species: Human
Cell Type/Line: A549,
THP-1, Mono Mac 6
(DSMZ)
Ti02:Tioxide Europe (London, UK) and
Degussa-Huls (Cheshire, UK)
UFTi02: Tioxide Europe (London, UK)
and Degussa-Huls (Cheshire, UK)
Particle Size: PM10, Ti02: 200 nm;
UFTi02: 20 nm
Route: Cell Culture (110,625 cells/well)
Dose/Concentration: PM10, Ti02, UFTi02:
100 pg/mL; TNF-a: 10ng/mL
Time to Analysis: 4 h
NF-KB and AP-1 DMA Binding: NF-KB DNA binding
increased in PM10 and TNF-a exposed macrophages
by 9.5 and 12 fold. NF-KB activity remained
unaltered in Ti02 and UFTi02 exposed
macrophages.
IL-8: Cells treated with PMio conditioned media
increased transcription binding of NF-kB to IL-8
promoter sites. Increases were observed in gene
expression after exposure to TNF-a and PMi0. Ti02
or UFTi02 had no effect. Increases observed in IL-8
production with PM10.
IL-8 Promoter CAT Activity: PM10 media increased
CAT expression by 65% over control. No differences
observed with Ti02 or UFTi02 media.
Neutrophil Chemotaxis: PMio conditioned media
induced a 2.3 fold increase compared to control.
TNF-a and IL-1p Production: PM1? media
increased TNF-a and IL-1|3 production. No increases
were observed in Ti02 and UFTi02 media.
December 2009
D-51
-------
Study
Pollutant
Exposure
Effects
Reference: Jung et al. Soot Particles: Generated using a co- Route: Surrogate Lung Fluid
(2006, 1324211
Species: N/A
Type: Surrogate Lung
Fluid
flow, laminar, diffusion flame system
CB (Degussa)
PM25: Collected using IMPROVE air
pollution samplers
Particle Size: Soot: 185 nm; CB: 25
nm, PM25
Dose/Concentration: Soot: 0-30 mg; CB: 5-
10 mg; PM25: 50 or 100|jg/mL
Time to Analysis: Parameters measured
continuously over 2 h.
OH Radical Formation: Formation occurred with
linear dependence on soot mass. Average response
was 0.89 nmol OH produced per mg of soot.
Formation also occurred with soot + hydrogen
peroxide. Hydrogen peroxide alone did not form OH
radicals.
Fe: Average Fe concentration in soot particles was
305 + 172 nM. Observed negative correlation
between amount of Fe and amount of OH radical
formation. DSF inhibited iron-induced increase in
OH radical formation.
Carbon Black: OH radical generation by carbon
black was significantly less than soot. OH generation
by CB was observed to be linearly proportional to
PM mass, but CB was much less efficient at
generating the OH radical.
PMis: A high variability in the increase of OH
radicals was observed with PM25. Pretreatment with
DSF partially blocked OH radical production, but a
significant level remained. This may be due to PM2 5
containing high levels of Fe and Cu.
Reference: Kafoury
and Madden (2005,
1566171
Species: Mouse
Cell Type: RAW 264.7
DEP: SRM 1975 (purchased from
NIST, Rockville, MD)
BAY11 -7082, NF-KB inhibitor
(coexposure)
IL-1B: obtained from Santa Cruz
Biotechnology (Santa Cruz, CA)
Particle Size: DEP: 0.3 pm (mean
diameter)
Route: Cell Culture (3-4x1 Ob cells)
Dose/Concentration: DEP 25,100, or 250
fjg/mL;IL-1IJ:100ng/mL
Time to Analysis: DEP: 4 h pre-treated with
BAY11-7082for1.5h;IL-1|3:4h
TNF-a: DEP induced a significant release of TNF-a
at 100 and 250 pg/mL dose-dependently. Exposure
at 25 |jg/mL had no effect. IL-113 containing PM
samples at 100 pg/mL also resulted in a significant
release of TNF-a.
NF-KB Binding Activity: Treatment of RAW264.7
with BAY11-7082 significantly inhibited IL-1|3-
induced TNF-a release. Similar effects observed
with DEP-induced TNF-a release.
Apoptosis: Inhibition of NF-kB binding activity by
BAY11-7082 resulted in DEP-induced apoptotic
response. Without BAY11-7082, apoptosiswas not
induced even at the DEP dose of 250/|jg/mL for 4 h.
The control, U937 cells with campothecin, induced
apoptosis.
Reference: Karlsson
et al. (2006, 1566251
Species: Human
Cell Type: A549,
Monocytes (isolated
from heparinized
whole blood)
PM
(W1: wood burning in old-type boiler;
W2: wood burning in modern boiler;
P: wood pellets burning in pellets
burner; T1: PM10 tire debris with
studded tires and ABT pavement;
T2a: PM10 tire debris with studded tires
and ABS pavement; T2b: PM25 tire
debris with studded tires and ABS
pavement; T3: PMi0 tire debris with
friction tires and ABS pavement;
St: PM10 from busy street in Stockholm,
Sweden; Su: PMi0 from platform of
subway station in Stockholm)
Particle Size: W: NR, T1, T2a, T3, St,
Su: PM10, T2b: PM25
Route: Cell Culture
Dose/Concentration: Suspension: 40
pg/cm2; Culture: 100 pg/cm2 (1 ml/well)
Time to Analysis: Suspension: 4 h; Culture:
18 h
PM Characterization: Boiler emitting PM-W1 led to
4 times higher emission of particles when compared
to PM-W2 and 8 times higher emissions when
compared to PM-P. Total concentration and CO was
substantially higher in the old-type wood boiler.
Effects with Filter Fibers: No increase of DNA
damage was observed compared to the water
control. Filter fibers led to the induction of cytokines
in human macrophages.
Genotoxicity: All particulate samples induced DNA
damage in A549 cells. PM-Su exhibited the most
genotoxicity and induced 4-5 times more DNA
damage than others.
Cytokines on Glass Fiber Filters: PM-W2 induced
a significant increase in IL-8. PM-St induced the
highest increases of IL-6, IL-8, and TNF-a.
Cytokines on Teflon Filters: PM-2a and PM-2b
samples caused significant increases of IL-6, IL-8,
and TNF-a.
December 2009
D-52
-------
Study
Pollutant
Exposure
Effects
Reference: Katterman PM: Oils: OAAF, Oil Q, Oil III, NF2
et al. (2007, 0963581
PM: Coal Germany and Ohio
Species: Rat
Diesel particulates: ZODDA (doped with
Cell Type/Line: RLE- Zn), ZSDDA (doped with Zn and S): S:
6TN (Alveolar PMs washed in solution; F: Fresh
Epithelial Cell Line) samples; L: Leached
AI203, Fe203, Si02, Ti02, ZnO also
tested
Particle Size: NR
Route: Cell Culture (cytotoxicity: 50,000
cells/well; SEM: 25,000 cells)
Dose/Concentration: Oils 0.2 mg/mL;
Coals 0.7 mg/mL; Diesel 0.01 mg/mL; AI203
0.5 mg/mL; Fe203 0.7 mg/mL; Si02 0.7
mg/mL; Ti02 0.7 mg/mL; ZnO 0.05 mg/mL
Time to Analysis: 24 h
Metabolic Activity: For oils comprised of 3/4 fresh
and 1/4 leached, metabolism decreased. Coals
(fresh and leached) had no effect. ZODDA-F and
ZSDDA-F both induced decreases in activity.
ZSDDA-L had no effect.
Cellular Morphology: PM-S had a minimal effect.
PM-F induced widespread cell damage.
Constituent Differences between PM-F and PM-
L: In oil samples Cu, Ti and Ca salts were removed
upon washing. Fe, Al, Si remained constant.
Grinding Effects: Coal toxicity increased upon
grinding, whereas diesel PM toxicity decreased upon
grinding.
Metal Oxide Effects: Only Si02, and ZnO (much
higher at lower concentrations than other metal
oxides) decreased metabolic activity. Fresh, washed
and sonicated samples exhibited similar results.
Grinding only affected Ti02 (increase) and ZnO
(decrease).
Reference: Kendall et
al. (2004, 1566341
Species: Human
Tissue Type: BALF
(obtained by
bronchoscope from 6
nonsmokers and 3
smokers)
PM25 sample sites; 2 schools in Bronx,
NY, 6 background urban, 6 urban
roadside. Sampling occurred 24 h/day
for 12 days.
Particle Surface Chemistry: 79-87%
carbonaceous material (Ch, COO, C-
0,N)), 10-17% 0 (01s), 1.5-4% N
NH4, N-C, N032'), 0.6-1 %S, and 0.3-2
% Si.
Only N03 - higher in roadside samples.
NH4 and N03 - correlated with NO and
NOX in air but not N02.
Particle Size: PM25
Route: BALF interaction
Dose/Concentration: 5-10 ml of 0.5 M NaCI
or BALF
Time to Analysis: Filters treated with BALF
for4h
Saline Washing: Removed particles and decreased
NH4, N03, 0 and S relative to C1.
BALF treatment (XPS): PM2$ surfaces interacted
strongly with BALF within hours of contact. Specific
surface components of PM2 5 immersed in BALF
were desorbed while biomolecules from BALF were
adsorbed to particles. N-C on the PM surface
increased 3 fold for smokers and 4 fold for
nonsmokers (range 1.4-7.4). This is most likely
related to protein-like adsorption on PM. Treatment
also induced a slight increase in COO and
decreases in NH4, N03, 0 and S.
ToF-SIMS - Organics: Particle loading and surface
hydrocarbons showed a linear correlation. Loss of
hydrocarbons from PM25 surface averaged 55% (10-
75) after undergoing saline and BALF washes. In
only 3/12 samples BALF removed less hydrocarbon.
BALF treatment increased the amino acid and
phospholipid content of the PM25 surface.
ToF-SIMS - Inorganic: Saline washing appeared to
increase Al and Si but with extreme variability; this
increase was not statistically significant. Both saline
and BALF washing decreased NH4 and Na levels to
a similar extent. BALF washing did not affect Al or
Si.
Reference: Kim et al.
(2005, 0884541
Species: Human
Cell Type/Line:
BEAS-2B
Zn2*
Particle Size: NA
Route: Cell Culture
Dose/Concentration: 15, 50,100 pmol
Time to Analysis: 1-20h
Cell Viability: At 50 pM for 20 h, no apoptosis was
induced.
IL-8: At 12 h, IL-8 increased in dose-dependent
manner. At 15 or 50 pM, Zn2* increased protein 1.6
and 4.6 fold respectively. IL-8 mRNA expression
increased dose-dependently, reaching statistical
significance at 2 h and continuing until 4 h.
EGFP (adenoviral IL-8 promoter): Levels
increased 2.4 fold with 50 pM Zn2*.
Proteases: With 50 pM Zn2*, phosphorylation of
MAPKs ERK, JNKand p38 increased by 15 min and
continued increasing up to 2 h. Pre-exposure of
inhibitors of MEK, JNK, before Zn2* exposure
caused inhibition ofZn-induced IL-8 mRNA and
protein production. Inhibitor of p38 had no effect.
Dephosphorylation of ERK and JNK was partially
inhibited with exposure to Zn *.
December 2009
D-53
-------
Study
Pollutant
Exposure
Effects
Reference: Kleinman UF1: Utrecht 1 Fine (urban freeway)
et al. (2003, 0879381
v UC1: Utrecht 1 Coarse
Species: Rat
Strain: Wistar Kyoto,
F344
Age: 22-24 mo, 10wk
Cell Type/Line: AM
UF2: Utrecht 2 Fine (urban, freeway,
light industrial)
UC2: Utrecht 2 Coarse
SRM 1650
SRM 1648
Particle Size: UF1: 0.2-2.5 pm; UC1:
2.5-10 pm;UF2: 0.2-2.5 pm; UC2: 2.5-
10pm
Route: Cell Culture (10b cells/well at 10b
cells/ml)
Dose/Concentration: 1.2 to 1200 ng/106
cells
Time to Analysis: 4,18 h
Macrophage PMA-stimulated respiratory burst
activity: SRM 1648 and 1650 induced dose-
dependent decreases approaching 0 at 50 -100
pg/10 cells. Large dose-dependent decreases from
old rat AMs exposed to fine PM exposure were
followed by young rat AMs exposed to fine PM.
However, no age-related effects were statistically
significant.
Free radical production: All coarse particles
depressed free radical production in a semi-dose-
dependent manner, with UC2 exhibiting more
potency than UC1. Both fine particles also showed
dose-dependent responses but UF1 and UF2
responses were greater than the control at 3 pg/106
cells.
PM Characterization: Ratios between coarse and
fine PM were similar for metals tested (Al, Fe, Mn,
Zn). Al was higher in coarse samples and Zn higher
in fine PM, although large variability was observed.
Fe and Mn results were roughly equivalent for all
samples.
Reference: Kocbach PMW: Wood smoke particles Route: Cell Culture (1 xio cells/ml)
et al (2008 1988741
Collected from conventional Norwegian Dose/Concentration: 30-280 pg/mL
Species: Human wood stove burning birch
Time to Analysis: 2, 5,12 h
Cell Type/Line: THP-1 PMT+: Traffic-derived particles;
collected from road tunnel in winter
when studded tires were used
PMT-: Traffic-derived particles;
collected from road tunnel in summer
without studded tires
DEP: SRM2975
Porphyr: fine grain syenite porphyry
(prepared bySINTEF, Trondheim,
Norway)
Polymyxin B Sulphate (endotoxin
inhibitor)
Particle Size: PMVV PMT, DEP: NR;
Porphyr 8 pm (mean)
Particle Characterization: PMT+ contained a high
mineral particle content. PMT- contained carbon
aggregates, and polycyclic aromatic hydrocarbons
(PAH). PMW and DEP contained carbon
aggregates. PAH content of PMW was greater than
DEP. Porphyr was not included in the analysis.
Cytokines: PMT+ induced releases of TNF-a, IL-1|3,
and IL-8 with 30 or 70 pg/mL PMW similarly
induced TNF-a and IL-8. DEP induced IL-lp and IL-
8. Porphyr induced IL-8 increases. IL-4, IL-6 and IL-
10 were unaffected. Overall, the order of effective
cytokine induction from most to least effective was
PMT+, PMW, DEP, and Porphyr. mRNA expression
of TNF-a, IL-1P, IL-8, and IL-10 increased with 140
pg/mL of PMT+ and slightly for PMW
LDH: PMT + induced small but statistically
significant increases at low doses. DEP increased
LDH at 280 pg/mL only.
Polymyxin B Sulphate: The endotoxin inhibitor
significantly inhibited LPS-induced cytokine release
by 80-90% and reduced PMT+ induction by 50-60%.
Organic Extraction: PMT+ washed and native
particles showed equivocal induction of cytokine
release. PMT+ organic extract had no effect. PMT-
and PMW organic extracts significantly increased
TNF-a and IL-8. Washed particles induced less
significant increases of IL-8. DEP organic extract
had no effect.
December 2009
D-54
-------
Study
Pollutant
Exposure
Effects
Reference: Kristovich
et al. (2004, 0879631
Species: Human
Cell Type/Line:
HUVEC, HPAEC,
HPMVEC, HPBMC
CP: carbon particle (carbonaceous
negative image of zeolite)
CFE: C/Fe particulate (synthesized)
CFE+: C-Fe/F-AI-Si particulate
(synthesized)
CFA: Coal Fly Ash (Coal-fired power
plant, NOS)
DEP: (exhaust pipe of diesel powered
truck)
CP, CFE, CFE+ approx 1 pm
(resembling zeolite)
Particle Characterization (Surface
chemistry): CP = 88% C, 1% Si, 10%
0, 1% N. CFE = 80% C, 2% Fe, 2% Si,
16% 0. CFE+ = 20% C, 6% Al, 3% Si,
50% F, 6%0, 11% N, 4%Na.
CFA = 25% C, 3% Fe, 13% Al, 17% Si,
41 %0, 1%N. DEP = 70%C, 3%Fe,
24% 0, 1%N, 2%S.
Particle Size: CP, CFE, CFE+:
approximately 1 pm (resembling
zeolite); CFA: <2 pm; DEP: 150 nm
Route: Cell Culture (4x1 Ob cells/well)
Dose/Concentration: CP: 5-50 pg/cm2;
CFE: 2.5-25 pg/cm2' CFE+: 2.5-25 pg/cm2;
CFA: 10-100 fjg/cm , DEP: 2.5-25 pg/cm2
Time to Analysis: 4, 8, or 24 h
Cytotoxicity: CP exhibited no effects. DEP and
CFE exhibited intermediate toxicities in the range of
50-70 pg/cm2. No toxicity was apparent when
treated with CFA (up to 200 pg/ cm ) or synthesized
C particulates.
Endothelial Activation: ICAM-1, VCAM-1, and E-
selectin were activated dose-dependently by DEP,
CFE, and CFE+. No effects observed for CFA or CP.
These effects were not the result of endotoxin
release.
Individual Variability: Donors (humans) showed
variability in responses especially for CFA. 3/9 had a
medium response negated by ND responses in 6/9.
Reference: Kubatova
et al. (2006, 1988351
Species: Rat, Human burning hardwoods"
PMW: Wood Smoke
Collected from airtight wood stove
Route: Cell Culture (RAW264.7:10b
cells/ml; BEAS-2B: 105 cells/ml)
Dose/Concentration: 50,100, 200 pg/mL
Cell Type/Line: RAW -P: Polar (fraction extracted from 25-50 Time to Analysis: 12 h
264.7, BEAS-2B C)
-MP: Mid Polar (fraction extracted from
100-150 C)
-NP: Nonpolar (fraction extracted from
200-300 C)
-C: P + MP + NP
Particle Size: NR
GSH: PMW-MP and PMW-NP induced GSH
depletion substantially in a dose dependent manner
starting at 50 pg/mL in both cell types. DMSO had
no effect.
Cytotoxicity: PMW-MP and PMW-NP increased
cytotoxicity at 200 pg/mL in RAW264.7. BEAS-2B
was unaffected.
Particle Characterization: PMW-MP contained
higher concentrations of oxy-PAHs, disyringyls,
syringylguaiacyls and PAHs. oxy-PAHs include 9-
fluorenone, 1-phenalenone, 9,10-anthraquinoneand
hydroxycadalene. PAHs included phenanthrene,
fluoranthene and pyrene.
Effects of Individual Components of PMW-MP on
GSH: 1,8-dihydroxy-9-10anthraquinone and 9,10-
phenanthraquinone depleted GSH. 9,10-
anthraquione, anthrone, 1-hydroxypyrene increased
GSH. Phenanthrene, 1-methylpyrene, 9-fluorenone
and xanthone had no effect.
Reference: Kubatova
et al. (2004, 0879861
Species: Monkey
Cell Type/Line:
African green monkey
kidney cells
designated COS-1
(CV-1 cells with origin -
defective mutants of
SV40), E coli PQ 37
(SOS Chromotest)
DEP: Obtained from diesel bus
PMW: Wood smoke particulates
obtained from airtight wood stove
burning hardwood
HSF: Hot pressure fractionation
-C: P + MP + NP
-P: Polar
-MP: Mid Polar
-NP: Nonpolar
OE: Organic Extraction
-HNP: n-hexane nonpolar
-MEP: methanol polar
Particle Size: NR
Route: Cell Culture (10,000 cells/180 \i\)
Dose/Concentration: 0, 50,100,150, 200,
250, 300 pg/mL
Time to Analysis: Cytotoxicity: 24 h;
Chomotest: 2 h SOS
Cytotoxicity: PMW induced cytotoxicity in a dose-
dependent manner. PMW-HNP induced low
cytotoxicity, followed by PMW-C (intermediate) and
PMW-MEP (highest). Levels above 25 pg/mL were
cytotoxic. DEP-HNP induced cytotoxicity but was not
dose-dependent. Results similar for all 3 fractions
(highly variable). All fractions with concentrations
higher than 100 pg/mL were cytotoxic.
Extraction Water Temperature Effect: PMW was
cytotoxic at temperatures over 50 C. DEP was
cytotoxic at temperatures higher than 200° C. At
250", cytotoxicity between DEP and PMW was
similar. At 300° C, PMW cytotoxicity declined and
DEP stayed high, resulting in DEP inducing higher
cytotoxicity than PMW.
SOS Chromotest: (3-Galactosidase formation
increased, peaked at 200° C with DEP and declined
to control at 300° C. Individual fractions showed
linear dose response from 25-200 pg/mL with 150° C
and 200° C extracts significantly higher.
December 2009
D-55
-------
Study Pollutant
Reference: Lee et al. MEP: Motorcycle Exhaust Particles
(2005, 1566821 (Yamaha Cabin engine, 95 octane
Species: Human
~,,T „- ,c,n MEPE: MEP Particle Free
Cell Type/Line: A549
Exposure
Route: Cell Culture (1 xio5 cells/well)
Dose/Concentration: MEP 0.02, 0.2, 0.2, 2,
20 |jg/mL; MEPE 20 pg/mL
Time to Analysis: 24 h
Effects
IL-8: MEP induced IL-8 at concentrations greater
than 0.2 pg/mL Levels increased 2fold at 24 h with
pg/mL Induction of IL-8 mRNA expression was
dose-dependent with MEP and MEPE.
Particle Size: MEP 0.5 pm; MEPEO.2
pm
Cytotoxicity: Exposure to particles did not affect
cytotoxicity.
NF-KB: MEP (20 pg/l) induced time-dependent
activation for 2 h and continued at same level for up
to 6 h. Pretreatment of PDTC (1 mM) fully inhibited
MEP induction.
MAP Kinase: MEP induced time-dependent
activation up to 30 min and stayed elevated for at
least 60 min.
ROI: MEP treatment induced a time-dependent
increase in ROI for up to 1 h and then continued the
at same level for up to 6 h.
Reference: Lee and
Kang (2002, 1988641
Species: Mouse
Cell Type/Line:
Peritnoeal
Macrophages, RAW
264.7
MEP Yamaha 2-stroke engine using
unleaded gas)
MEPE(particle-free MEP)
Particle Size: Q.5|jm
Route: Cell Culture (5x1 Ob cells/mL
(Cytotoxicity), 3x105 cells/mL (Apoptosis),
2x106 cells (MMP and ROI), 1 xio7 cells
(GSH)
Dose/Concentration: 5,10, 50,100, 300,
1000 |jg/mL
Time to Analysis: 6,12, 18, 24 h
Cytotoxicity: Viability decreased dose and time-
dependently in all cell types at 24 h.
Apoptosis: subG1 significantly and dose-
dependently increased at the 300 MEP pg/mL dose
in all cell types, indicating increased apoptosis.
MEPE induced similar results. Inhibition was
successful against MEP-induced apoptosis by
calcium chelators EGTA, BAPTA-AM, cyclosporin A
and antioxidants MAC, GSH, catalase and SOD.
Ca2+: MEP and MEPE increased Ca2* at 300
L. BAPTA-AM completely inhibited induction.
ROI: MEP increased ROI in a time-dependent
manner. Calcium chelators and antioxidants
substantially attenuated induction.
GSH: MEP significantly decreased GSH.
MMP: Mitochondria membrane potential decreased
dose-dependently with MEP 100 (jg/mL and 300
pg/mL Calcium chelators and antioxidants partially
inhibited reduction.
Reference: Li et al.
(2002, 0420801
Species: Mouse
VACES (Biosampler PM10 in Downey,
CA-DEP concentrate in water)
DEPM (DEP methanol extract)
Cell Line: RAW264.7, DEPME (DEP methylene chloride
THP-1 extracts)
DEPAL (DEPME aliphatic (hexane))
DEPAR (DEPME aromatic
(hexane/methlene chloride))
DEPPO (DEPME polar (methylene
chloride/methanolj)
Particle Size: NR
Route: Cell Culture (2x106 cells/well Mouse
RAW264.7 and THP-1 ;0.67x106 cells/well
Murine RAW 264.7)
Dose/Concentration: 10-200 pg/mL
JNKActivation and IL-8 Production: THP-1
cells- 0, 10, 25, 50, 100 pg/mL DEPM; THP-
1 cells-0,10, 25, 50,100|jg/mLof DEP;
RAW264.7 cells-10-100 DEP pg/mL
Cytotoxicity: 1,10, 25 (THP-1 cells only), 50,
100, 200 pg/mL
GHS/GSSG:0, 10, 25, 50, 100|jg/mL
HO-1 Expression: 0, 25, 50,100, 200 pg/mL
Time to Analysis: GHS/GSSG: DEPM,
whole DEP (RAW264.7 only) 8 h.
HO-1, MnSOD Expression: RAW264.7,
THP-1 7h. RAW 264.7 cells exposed to
whole DEP 16 h.
JNKActivation, IL-8 Production: THP-1 cells
30 min, 16 h. RAW264.7 cells 90 min.
Cytotoxicity: RAW264.7, THP-1 18 h.
GSH/GSSG Ratio: DEPM induced dose-dependent
decrease in GSH/GSSG ratios in both cell lines.
DEP induced decreases at comparable doses to
DEPM.
HO-1 Expression: Cells exhibited dose-dependent
increases in HO-1 expression.
HO-1 Expression in Murine RAW 264.7: VACES-F
consistently induced HO-1 expression over a 9m
period, whereas VACES-C was effective in inducing
HO-1 during fall and winter. HO-1 induction
positively correlated to higher OC and PAHsthat
were represented in VACES-F, but also seen with a
rise in PAHs in VACES-C during winter months.
MnSOD: At doses of 2.5 pg/mL, DEPM increased
MnSOD in THP-1 cells.
JNKActivation: DEPM dose-dependently
increased JNK phosphorylation but did so without a
change in the JNK expression level. DEP-exposed
mouse RAW264.7 cells exhibited similar increases
in JNK phosphorylation but without increasing JNK
expression.
IL-8: Exposure to DEPM elicited dose-dependent
increase in IL-8 levels of THP-1 cells.
December 2009
D-56
-------
Study
Pollutant
Exposure
Effects
Reference: Li et al.
(2002, 0874511
Species: Human
Cell Line: BEAS-2B,
NHBE, THP-1
macrophages
DEPM (DEP methanol extract)
DEPME (DEP methylene chloride
extracts)
Route: Cell Culture (10bcells/mL)
Dose/Concentration: 0,10, 25, 50,100
pg/mL
DEPAL (DEPME aliphatic (hexane)) Time to Analysis: 30, 60,120 min
DEPAR (DEPME aromatic
(hexane/methlene chloride))
DEPPO (DEPME polar (methylene
chloride/methanolj)
Particle Size: 0.05-1 pm
ROS: BEAS-2B cells demonstrated increased HE
fluorescence, indicating increased ROS formation.
THP-1 cells were unaffected.
GSH/GSSG Ratio: DEPM dose-dependently
decreased GSH/GSSG in THP-1 and BEAS-2B
cells. Similar changes occurred with NHBE cells.
THP-1 cells maintained a higher ratio of GSH/GSSG
than BEAS-2B and NHBE cells.
MAC on GSH/GSSG Ratio: Exposure to DEPM in
the presence of NAC did not affect the GSH/GSSG
ratio in BEAS-2B and NHBE cells. In THP-1 cells,
NAC prevented a decline in the GSH/GSSG ratio.
MnSOD and HO-1: THP-1, BEAS-2B and NHBE
cells showed constitutive MnSOD expression and
dose-dependent expression of HO-1 protein and
mRNA. No change occurred in the expression of |3-
actin.
DEPAL, DEPAR, DEPPO, CoPP on HO-1
Expression: DEPPO was more potent than DEPAR.
DEPAL lacked activity for THP-1 and BEAS-2B cells.
The potency of DEPPO was sufficient to affect
cellular viability and HO-1. CoPP induction of HO-1
failed in THP-1 cells, but succeeded in BEAS-2B
cells. However, it did not protect against the
oxidizing effects of DEPM.
JNK: JNK activation increased in DEP-exposed
THP-1 and BEAS-2B cells. JNK isoforms were
observed at doses of > 25 pg/mL. In BEAS-2B cells
a high rate of cell death diminished this response at
100 pg/mL NHBE also showed increased JNK
phosphorylation at doses 50-100 pg/mL
NAC on JNK: NAC led to inhibition of JNK
activation.
IL-8: THP-1 cells showed dose-dependent increases
of IL-8. NHBE cells showed incremental increases
followed by rapid decline at 100 pg/mL attributed to
apoptosis. BEAS-2B cells responded to 10 pg/mL
with increased IL-8, but cellular toxicity and cell
death led to a drop in IL-8 production at higher
doses.
Cytotoxicity: Comparing cytotoxicity at 25 pg/mL
DEP, BEAS-2B cells had a higher rate of cell death
than THP-1 cells. BEAS-2B cells showed a
significant rise in cell death at doses larger than 10
pg/mL In THP-1 cells, it took doses of 25 pg/mL or
more before significant increases occurred.
In BEAS-2B, cell death began at 2 h. In THP-1,
increases in cell death prolonged for 8h or longer.
NHBE cells also showed increase rates of
cytotoxicity compared to macrophages. NAC in
THP-1 interfered with a generation of cytotoxicity,
but NAC did not have any decreasing effect on cell
death in BEAS-2B or NHBE cells.
December 2009
D-57
-------
Study
Pollutant
Exposure
Effects
Reference: Lindbom
et al. (2007, 1559341
Species: Mouse
Cell Line/Type: RAW
264.7
Route: Cell Culture (130,000 cells/cm')
Dose/Concentration: 1,10 or 100 pg/mL
Tlme to Analysis: 18, 24 h
PM10:
-ST: Street
-S: Subway
-G: Granite Analysis of Arachidonic Release (AA): Cells
pre-incubated w/1 pCI tritium marked for AA
-Q: Quartzite and wasned exposed to 10, 50,100 and 250
(-G and -Q generated by road simulator ^ m
at Swedish National Road and
Transport Research Institute)
Particle Size: PM10; Bimodal with
peaks around 4-5 urn and 7-8 urn.
Cellular Viability: Viability was not influenced by
any particle types and in all cases exhibited 90% or
higher viability, except for the combination of subway
particles and MAC where viability dropped to 20%.
Cytokines: All particles induced TNF-a secretion in
a dose-dependent fashion. PM-S was most potent at
1 fjg/mL PM-G and PM-ST induced effects at 10
pg/mL PM-Q induced increase of TNF-a at 100
|jg/mLPM-ST induced IL-6 release at 10 pg/mL.
PM-G, PM-Q, PM-S induced IL-6 secretion at 100
pg/mL DFX inhibited TNF-a in cells exposed to PM-
S and PM-ST. DFX induced increase of TNF-a with
PM-Q. For all PM types (except PM-ST) DFX
inhibited induced IL-6 secretion.
NO: PM-ST and PM-G induced a significant release
of NO, with PM-ST inducing a higher NO release
than PM-G.
MAC: NAC treatment significantly inhibited both
TNF-a and IL-6 secretion with all PM particles.
L-NAME: L-NAME caused a decrease in NO
secretion at 100 pg/mL of PM-ST. L-NAME did not
have an effect on granite-induced NO secretion at
100 pg/mL
Cytokine Gene Expression: TNF-a mRNA showed
a trend to increase for -ST, but this did not reach
significance. IL-6 gene expression increased for PM-
Q, PM-ST, PM-S but not for PM-Q.
AA Release: PM-S exposure at 100 and 250 pg/mL
was the only PM to induce AA release.
Lipid Peroxidation: All particle types induced lipid
peroxidation. PM-S and PM-ST induced significantly
higher lipid peroxidation as compared to PM-Q and
PM-G.
ROS: All particle types induced ROS formation. PM-
S and PM-ST induced significantly higher formation
at 10 pg/mL PM-Q and PM-G induced small but
significant decreases in absorption at 100|jg/mL.
Both PM-ST and PM-S had significant dose
responses for all concentrations tested. No
difference was observed between PM-G and PM-Q.
PM-S and PM-ST pretreated with DFX had a lower
ability to induce ROS formation.
Endotoxin Content: Only PM-ST showed positive
results for endotoxin content.
December 2009
D-58
-------
Study
Pollutant
Exposure
Effects
Reference: Liu et al.
(2005, 0883041
Species: Human
Cell Type: HPAECs
SE: V\food Smoke Extract; generated
using a stainless steel receptacle
containing 100g of dry wood dust
Particle Size: NA
Route: Cell Culture
Dose/Concentration: 40 |jg/mL
Time to Analysis: 0-4 h; Mitochondria!
Membrane Destabilization: 0-60 min; DMA
Defragmentation: 0-6 h; Cytotoxicity: 24 h
Viability: SE exposure reduced cell viability dose-
dependently Reduction reached ~38% of control.
Effect on Oxidative Stress/Antioxidant Enzymes:
SE caused an increase in ROS levels, in particular
02- and H202 in a time-dependent manner.
Exposure to SE for up to 4 h caused a decrease in
GSH levels in a time-dependent manner. Increased
expression of Cu/Zn SOD mRNAand HO-1 mRNA
was observed. Catalase or GPx mRNA expression
was unaffected. Upregulation of Cu/Zn SOD and
HO-1 occurred in a time-dependent manner
Mitochondrial Translocation/ Capsase-
Independent Apoptosis/DNA fragmentation:
Exposure for up to 60 min caused an increase in the
percentage of annexin V-FITC-pos cells but not Pl-
pos cells. At 4 h, FDA-pos cells was unaffected. SE
exposure caused a loss of mitochondrial membrane
potential (indicated by the change in JC-1
fluorescence). Cytosolic bax levels increased after
exposure for 1 or 2 h and returned to basal level at
4 h after exposure. Levels of procaspase-3 and
caspase-9 were unaltered by SE exposure after 4 h.
Procaspase-3 increased and caspase-9 decreased
by H202 exposure. SE exposure increased levels of
AIF and EndoG (exposure up to 4 h). At 6 h,
increased DNAdefragmentation was observed. Pre-
treatment with caspase inhibitors (CMK and Z-VAD-
FMK) failed to suppress SE-induced apoptosis.
MAC: Treatment with MAC prevented ROS increase
in cells exposed to SE for 60 min. MAC addition
prevented the reduction of GSH by SE. MAC
decreased nuclear levels of AIF and EndoG and
completely reduced DNA-fragmentation. MAC
alleviated the SE-induced reduced viability. GSH
and DMA fragmentation were unaffected by MAC.
Reference: Long et al.
(2005, 0874541
Species: Human
Cell Types: Human,
Peripheral blood
mononuclear cells
(PBMCs) differentiated
into MDMs (90-95 %
CD14+)andT
lymphocytes
Synthetic C and C/Fe particles (phenol
and paraformaldehyde polymers on a
zeolite template)
C/Fe analysis Al 1.38 %, Si 0.33 %, Fe
0.46%
Particle Size: 1 |jm
Route: Cell Culture (5x106 cells, 2 mL /well
MDMs) Dose/Concentration: 5 pg/cm
Time to Analysis: 2-24 h
ROS release: Oxidative burst form C/Fe maxes out
at 20 min with no effect from C particles.
Cellular participate actions: C particulates were
present within lysosomes with small clumps forming
after 24 h outside of lysosomes with no evidence of
organelle lysis and/or agglomeration. C/Fe
particulates showed similar initial effects progressing
at 24-h total organelle lysis extending to the outer
cell membrane.
T cell effects: No effects from C or C/Fe particles
Medium Effect: Particle agglomeration appears to be
a direct result of serum present within a cell free
medium
Hydroxyl radical formation: C/Fe particles showed
an order of magnitude of higher hydroxyl formation
as compared to C particles
December 2009
D-59
-------
Study
Pollutant
Exposure
Effects
Reference: Ma et al.
(2004, 0884171
Species: Mouse
Cell Line: JB6P+
(Epidermal Cell Line)
DEP: SRM 1975
Particle Size: 0.5pm
Route: Cell Culture
Dose/Concentration: Non-cytotoxic: 5,10,
20 pg/mL; Cytotoxic: 0,10, 20, 40, 80,100,
160 pg/mL
Tlme to Analysis: 24, 48 h;
NF-KBandAP-1:12h
Phosphorylation of Akt: 5-120 min.
Effect of LY294002 on DEP: Cells pretreated
with LY294002 (0 or 10 pM) for 30 min and
then exposed to DEP for 0-60 min.
Viability: Below 20 pg/mL, DEP had no effect. At
concentrations greater than 20 pg/mL, DEP caused
apoptosis.
NF-KB and AP-1: DEP stimulated NF-KB activity at
5 and 10 pg/mL. At 20 pg/mL, NF-KB activity
decreased, but was still greater than the control.
DEP had no effect on AP-1 activity.
PI3K/Akt Signaling Pathway: DEP induced
phosphorylation of Akt on both Thr-308 and Ser-473.
LY294002 (an inhibitor of P13K) blocked
phosphorylation of Akt, p70/p85 s6 kinase and GSK
3b. LY294002 eliminated DEP-mediated
phosphorylation of Akt. Inhibition of P13K by
expressing p85 also blocked DEP-induced Akt
phosphorylation. DEP induced phosphorylation on
GSH-3B on Ser-9 without affecting tyrosine
phosphorylation and enhanced phosphorylation of
p70/p85 S6 kinase on Thr-389. DEP had no effect
on phosphorylation of FKHR.
SAPK/JNK Pathway: DEP slightly activated the
pathway. Increased transient activation of MKK4 (a
signal component of the SAPK/JNK pathway) and
thus enhanced phosphorylation of SAPK/JNK. DEP
promoted phosphorylation of c-Jun and ATF-2. DEP
did not affect p38 MAPK or ERK phosphorylation.
LY294002: Treatment with LY294002 (P13K
inhibitor) eliminated DEP-induced NF-KB activity. A
similar effect was observed with the use of another
P13K inhibitor, wortmannin. TDZD-8 (GSK-3B
inhibitor), D-JNKI(a JNK inhibitor), SB202190
(inhibitor for p38 MAPK) or PD98059 (inhibitor for
MEK1) had little effect on DEP-mediated NF-KB
activation.
Reference:
Maciejczyk and Chen
(2005, 0874561
Species: Human
Cell Type: BEAS-2B
CAPs: PM25
Collected via cyclone inlet on side of
building in Tuxedo, NY. Weekdays 9-3
March 4 to September 5, 2003
Mass contributions of the Regional
Sulfate, Soil, Oil- Combustions and
Unknown/other categories to CAPs are:
Regional Sulfate- 65%, Soil- 20%,
Unknown/Other-13% and Oil
Combustion- 2%.
Composition:
* Regional Sulfate characterized by
high concentrations of S, Si and .
* Soil characterized by high
concentrations of Ca, Fe, Al and Si.
* Oil-Combustion characterized by high
concentrations of V, Ni and Se.
Particle Size: PM25
Route: Cell Exposure (subchronic
exposures); Cell Culture (NF-KB) (9x104
cells/well)
Dose/Concentration: CAPS 109 + 178
pg/m3 (air exposure); 300 pg PM/ml (culture)
Time to Analysis: 24 h
NF-KB: NF-KB response most notably correlated
with V and Ni - elements associated with oil
combustion source category (oil combustion makes
up the group that is the smallest percentage of CAP
mass).
December 2009
D-60
-------
Study
Pollutant
Exposure
Effects
Reference: Madden et
al. (2003, 1988771
Species: Human
Cell Type: NHBE
DEP(SRM 2975)
Route: Cell Culture
Obtained from Caterpillar diesel engine,
4 cycl, 4 stroke, model 3304
Diesel Exhaust Extracts from a High Dose/Concentration: 0,10, 50,100, 250,
load (HL~75% engine load) or Low load 500 pg/well
(LLO% engine load): t . , - „, t, „ -i. ,
Time to Analysis: 24 h (after 2 h of
treatment, 0.5 ml of BEGM added to each
well and cells incubated for an additional 22
Particle Characterization: LL extract
has greater amount of low-molecular-
weight carbonyls (2-5 carbons). HL had
more intermediate size carbonyls (6-9
carbons). Largest carbonyls analyzed
(11-12 carbons) found in similar ratios
in the two types of extract (number of
carbons is indicative of differences in
boiling points).
Particle Size: NR
Cytotoxicity: LL, HL and SRM had no effect on
LDH release.
61Cr: Incubation of cells with LLorSRM (10 to 500
pg/well) had no effect. 500 pg/well of HL induced a
significant increase in 51 Cr release.
IL-8: HL induced a 5-fold increase in IL-8 at 500
pg/well. A decrease was observed at the highest
dose of LL extract. SRM did not significantly alter IL-
8 production.
PGE2: Production of PGE2 (inflammatory/immune
mediator) increased in cells treated with HL extract
at 500 pg/well. LL had no effect. Stimulation with
melittin caused LL extract to have inhibitory effect on
PGE2 at 500 pg/well. SRM had no effect.
Reference: Matsuo et
al. (2003, 1988791
Species: Human
Cell Type: NHBE,
NHPAE, TIG-1.TIG-7
(normal human lung
embryonic fibroblasts)
DEP: prepared at National Institute for
Environmental Studies (Tsukuba,
Japan)
RDEP: residual DEP (after sequential
extraction with hexane (NOS),
benzene, dichloromethane, methanol,
1N ammonium hydroxide)
Particle Size: 0.4 pm(MMAD)
Route: Cell Culture (NHBE: 5x104 cells/cm ,
NHPAE: 3x103 cells/cm2; TIG-1 and TIG -7:
3x103 cells/cm2. Apoptosis: 2x105 cells/cm2;
ROS/NO: 2x10 cells/cm2; Cytotoxicity
Modulating Agent: 3x104 cells/cm2; GSH:
3x104 cells/cm2)
Dose/Concentration: 25, 50,100, 200, 300,
400, 500 pg/mL
Time to Analysis: 1 h
Cytotoxicity in NHBE: Both DEP and RDEP
exhibited dose-dependent Cytotoxicity at
concentrations beginning from 50 pg/mL and higher.
RDEP was less cytotoxic than DEP. DEP exposure
resulted in necrosis, not apoptosis.
Comparative Cytotoxicity: The order of LC50
values (50% lethal concentration) was: NHBE (118
fjg/ml), NHPAE (137 pg/ml), TIG (270 fjg/ml).
NHBE's susceptibility was higher than the
susceptibility of NHPAE and TIG cells.
ROS/NO: DEP induced dose-dependent increases
at25and50|jg/mL
Reduced Glutathione: DEP induced dose-
dependent decreases. At 200 or 300 pg/mL, GSH
levels decreased by 55.2 or 97.3%, respectively.
Antioxidant Effects: Various antioxidants either
decreased DEP Cytotoxicity (PMC, Ebselen, EUK-8)
or had no effect on DEP Cytotoxicity (SOD, catalase,
GSH, a-tocopherol)
Chelating Agents: DEP became less cytotoxic
when lon-chelating agents were preincubated for
24 h. No effect on DEP Cytotoxicity was observed
when chelating agents were administered to cells
immediately after sonication.
Endocytosis inhibitors: Decreased DEP toxicity
was observed in a dose-dependent manner.
Reference: Matsuzaki
et al. (2006, 1995171
Species: Human
Cell Type: Peripheral
neutrophils
Gender: Male and
Female
Age: 20-40 yrs
DEP: generated from a 4JB1-type, 4
cyl Isuzu diesel engine
me-DEP: methanol extract of DEP (40
% of DEP by dry weight)
Particle Size: 04 |jm
Route: Cell Suspensions (5x10 cells/mL)
Dose/Concentration: all me-DEP
f-actin: 1, 5,10|jg/mL
CD11b:5,10, 30|jg/mL
IL-8: 5,10, 30|jg/mL
H202:5,10, 30, 60|jg/mL
MMP-9, LTB-4:5, 10, 30, 60|jg/mL
Time to Analysis: f-Actin: 15 min
CD11b:2h
IL-8:2or24h
H202: 30 min
MMP-9, LTB-4:2or24h
F-Actin: Treatment with me-DEP showed a dose-
dependent increase in the f-actin content of
neutrophils and this increase was significantly higher
at 5 and 10|jg/mL
CD-11b: Treatment increased CD-11b expression
two-fold at 30 pg/mL
IL-8: Minimal response was observed after 2 h. A
significant increase was observed (243%) at 24 h
with 30 fjg/mL.
LTB-4: At 2 h, LTB4 increased to 115% and 119%
with 30 and 60 pg/mL me-DEP respectively. At 24 h
with 60 |jg/mL me-DEP, LTB-4 increased to153%.
H&f. Exposure to 30 and 60 pg/mL of me-DEP
induced large dose-dependent increases of 563%
and 1220%, respectively.
MMP-9: A significant increase at 2 and 24 h were
observed. In both exposure periods, 30 pg/mL
induced larger increases than 60 pg/mL
December 2009
D-61
-------
Study
Pollutant
Exposure
Effects
Reference: Molinelli et PMH: PMi0 extracts in hexane
al. (2006, 198949) nllA nll t t .
PMA = PM10 extracts in acetone of
residue after hexane extraction
Species: Human
Cell Type: NHBE,
BEAS-2B
Route: Cell Culture (3x10' cells/well)
Dose/Concentration: NHBE exposed to 0-
-G: Guaynabo(Urban) and
-F: Fajardo (Preservation Area)
Particle Size: PM10
BEAS-2B exposed to 10,100, 250 pg/mL of
PM10
Time to Analysis: 48 h
Metal analysis: Hexane extracts Cu, V, Ni all higher
in winter than summer. For hexane extracts within
the same season, metal concentrations were higher
in the Fajardo extracts. On the other hand, the
acetone extracts from Guaynabo generally had
higher metal concentrations than Fajardo.
Cytotoxicity NHBE: The order of most to least toxic
for PM extracted with hexane is: winter-G, winter-F,
summer -G , summer-F The order of most to least
toxic for PM extracted with acetone is: summer-G,
summer-F, winter-g.
Cytotoxicity BEAS-2: For PM extracted with
hexane, the Cytotoxicity order is: winter-G, winter-F,
summer-G, summer-F. The order for acetone
extracted PM is: summer-G , summer-F, winter-F,
winter-G. Effects trend similar to metal levels (no
analysis). Summer extracts showed linear dose-
response curves. Winter extracts exhibited more
equivocal results, especially for Fajardo. Results
suggest that NHBE cells are more sensitive than the
BEAS-2B cells to PM extracts.
Reference: Holler et
al. (2002, 0365891
Species: Canine,
Mouse
fTi02 (origin NR)
ufTi02 (origin NR)
ufP-G: carbon black (Printex-G,
Degussa, Frankfurt, Germany)
Cell Type: Beagle-Dog
Alveolar Macrophages ufP90: carbon black (Printex90,
— — • Degussa, Frankfurt, Germany)
Route: Cell Suspension
Dose/Concentration: 10, 32,100, 320
pg/mL
Time to Analysis: 24 h
(BD-AM), J774A.1
ufEC90: EC (produced by electrical
spark generator under standardized
conditions with low impurities)
DEP(SRM1650)
UrbD: Urban Dust (SRM 1649a)
Particle Size: (in diameter) Ti02: 220
nm; ufTi02: 20 nm; ufP-G: 51 nm;
ufP90:12 nm; ufEC90: 90 nm;
DEP: 120 nm; UrbD: NR
Cytoskeleton of J774A.1: At doses of 32 pg/mL or
less, the particles did not significantly influence
relaxation and stiffness. fTi02and ufP90 had no
effect at any dose. ufTi02 at 320 pg/mL induced
retarded relaxation and significant stiffening. uEC90
induced dose-dependent retardation of relaxation
and increased stiffening. DEP and UrbD induced
similar results.
Cytoskeleton of BD-AM: ufTi02 and fTi02 both
induced some retarded relaxation and increased
stiffening at 100 pg/mL dose. ufTi02 appears to
increase stiffening in a dose-dependent manner.
ufEC90 induced dose-dependent acceleration of
relaxation due to the carbon content of uEC90. DEP
also induced acceleration of relaxation as well as a
decrease in stiffness.
Phagocytosis: At 24 h, ufTi02 and fTi02
significantly reduced phagocytic ability in J774A.1
but not in BD-AM. All carbonaceous particles
induced significant impairment in J774A.1. All
ultrafine carbon particles inhibited BD-AMs.
Cell Proliferation: ufTi02 significantly inhibited
proliferation compared to the control and fTi02 at
100 pg/mL in J774A.1. ufEC90 and ufP90 inhibited
proliferation slightly with uEC90 inducing slightly
greater inhibition than ufP90. UrbD and DEP also
significantly reduced proliferating.
Apoptosis: All particles induced decreased viability
at 100 |jg/mL in both cell types. With ufTi02 inducing
greater apoptosis than fTi02, uEC90 than ufP90
and ufEC90 than ufP-G
Reference: Mutlu et
al. (2006, 1559941
Species: Human, Rat
Cell Type: A549
PM10
(Collected by baghouse from ambient
air in Dusseldorf, Germany)
Particle Size: PM10
Route: Cell Culture
Dose/Concentration: 0.05, 0.5, 5. 50
pg/cm2
Time to Analysis: 24 h
Na, K-ATPase Plasma Membrane Protein: PM10
induced a decrease of protein in the plasma
membrane of A549 cells. Total Na,K-ATPase levels
were unaffected.
ROS: Pretreatment with EUK-134, superoxide
dismutase and catalase mimetic, inhibited the
decrease of GSH. Furthermore, it attenuated the
decrease of NA,K-ATPase in A549 cells.
NA, K-ATPase Activity: PMi0 induced a dose-
dependent decrease in ouabain-sensitive liberation
of32Pfrom AT32P in primary rat alveolar type II
cells. This effect was inhibited with pretreatment with
EUK-134.
December 2009
D-62
-------
Study
Pollutant
Exposure
Effects
Reference: Nam et al.
(2004, 1988871
Species: Human
Cell Type: A549
PM25
Collected from hospital rooftop, Seoul,
South Korea
Particle Size: PM25
Route: Cell Culture
Dose/Concentration: 0.5,1,10, 25, 50
pg/cm2
Time to Analysis: 6, 24 h
NF-KB/kBa: 50 pg/cm DEP induced iKBa
degradation which peaked at 2 h and recovered
after 4 h. Treatment with increasing amount of PM2 5
resulted in a dose-dependent decrease in IKBa.
PM25 increased NF-KB in a dose-dependent manner
up to 10 pg/cm2. NF-KB induction peaked at 12 h.
IL-8: PM25 treatment increased protein level more
than 3 fold with 100 pg/cm2 PM25. mRNA levels also
increased.
iNOS Inhibitor: PM25 induced IL-8 elevation was
completely blocked by iNOS inhibitor. iNOS inhibitor
also negated PM25 induction of NF-KB activity.
Antioxidants and iNOS inhibitor reduced PM-induced
IKBa degradation.
Reference: Nozaki et
al. (2007, 0978621
Species: Mouse
Cell Line: LA-4
(Alveolar Epithelial
Cells)
PM: Rooftop of 5 story building, urban,
Japan
PME: dichloromethane extract of PM
filtered
P90: Printex 90 (carbon black)
(Degussa)
Particle Size: PM: 0.22pm; PME: 2.5
pm; P90:14 nm
Route: Cell Culture (1.4x104 cells/cm2)
Dose/Concentration: 1.1 pg/cm2
Time to Analysis: 24, 28, 72 h
Cytotoxicity: P90 had no effect. PM and PME were
cytotoxic at similar levels.
Protein Expression: All particles affected protein
expression (no specific protein- 2D gel
electrophoresis).
Reference: Obot et al.
(2002, 0423701
Species: Mouse
Cell Line: BALB/c
Cell Type: AM
PM:SRM1648
PM-100:PM heated to 100° C
PM-500:PM heated to 500° C
PM-PH: PM acid digestion
PMAC: Acetone extraction
PMCH: Cyclohexane extraction
PMH20: Water extraction
All extract fraction used as residual
particles
Particle Size: NR
Route: Cell Culture (5x105 cells/mL)
Dose/Concentration: PM: 200 pg/mL; PM-
100:188 pg/mL; PM-500:130 mg/l; PM-
PH: 94 pg/mL; PMAC: 173 pg/mL;
PMCH: 171 pg/mL; PMH20:188 pg/mL
Fraction doses adjusted for mass loss during
fraction treatment
Time to Analysis: 4 h
Cytotoxicity: All 7 fractions had cytotoxic effects.
PM had highest Cytotoxicity. PM-500, PM-PH, PMAC
less toxic than PM.
Apoptosis: All 7 fractions significantly increased
apoptosis. The PM fractions that induced the
greatest apoptosis in descending order are: PM,
PMH20, PM-100, PM-500, PMAC, PMCH and PM-
PH. PM-induced apoptosis (only PM, PM-500 and
PMAC tested) was blocked by poly I or 2F8 antibody
(scavenger receptors).
Particle Characterization: Untreated PM and PM-
100 did not have measurable amounts of transition
metals on its surface. Measured components include
carbon, 02, N, S, Si, Ca, Al, P, Cl. PM-PH mostly
contained 02 and Si. PM-500 had increased 02, Si
compared to PM and measurable amounts of Na, K,
Zn, Co, Pb, Fe. Included increased surface density
of S, P, Al. PMCH lacked nonpolar organic
compounds.
Reference: Obot et al.
(2004, 0959381
Species: Mouse (7-
9wk), Human
Cell Line: Mouse-
BALB/c
Cell Type: AM
PM: SRM 1648 (collected by bag-
house in St. Louis, MO).
PM-100: PM heated to 100° C
PM-500: PM heated to 500'C
PM-PH: PM acid digestion
PMAC: Acetone extraction
PMCH: Cyclohexane extraction
PMH20:V\Mer extraction
All of the 6 extract fractions from
PM1648
PM25: Collected in Houston, TX
Particle Size: PM1648: NR; PM25
Route: Cell Culture (5x105 cells/mL)
Dose/Concentration: PM: 200 pg/mL; PM-
100: 188 pg/mL; PM-500:130 mg/l; PM-
PH: 94 pg/mL; PMAC: 173 pg/mL;
PMCH: 171 pg/mL; PMH20:188 pg/mL
Human AM Viability: Only PM, PM-100, PMAC and
PMH20 decreased viability.
Human AM Apoptosis: PM, PM-100 and PMH20
increased apoptosis. PM induced greater apoptosis
than PM-100 and PMH20.
Fraction doses adjusted for mass loss during Regression Analysis Mouse vs Human: Although
individual fractions differed somewhat, cell viability
and apoptosis of all 7 fractions showed linear
regression
fraction treatment
PM25 = 50, 100, 150, 200pg/mL
Time to Analysis: Mouse-4 h; Human-24 h.
Human and Mouse AM Viability with PM2.6:
Nearly identical dose-dependent decrease was
exhibited starting at 50 pg/mL
Human and Mouse AM Apoptosis with PM25:
Nearly identical dose-dependent increases were
exhibited with human AM responses peaking at 150
pg Art and declining at 200 pg/mL (no mouse data
for200pg/mL).
Regression Analysis with PM2.6: Excellent
correlation of mouse and human responses for
viability and apoptosis was exhibited.
December 2009
D-63
-------
Study
Pollutant
Exposure
Effects
Reference: Okeson et CG: Coal ash, Germany
al. (2003, 0422921
Species: Rat
Cell Type: RLE-6TN
(Type 11 Alveolar
Epithelial Cells)
CU: Coal ash, USA
5C:PM# 5 Oil fly ash coarse
5F:PM#50ilflyashfme
6MSC: PM #6 Oil med sulfur fly ash
coarse
6HSC:PM# 6 Oil high sulfur fly ash
coarse
6HSF:PM# 6 Oil high sulfur fly ash
fine
Particle Size: CG, CU: NR; 5C, 6MSC,
6HSC: >2.5 pm; 5F, 6HSF: <2.5 pm
Route: Cell Culture
Dose/Concentration: Coal Fly Ash 12.5,
25, 50, 125, 250 pg/mL
OilFlyAsh-100|jg/mL
Time to Analysis: 24 h
Oil PM Characterization: Generally, the fine
fractions had higher metal levels than the coarse
fractions except forZn. High sulfur had a higher
metal content than med sulfur. Carbon percent
weight was stable across all 5 fractions.
Coal Ash Cytotoxicity: CG treatment exhibited
similar cytotoxic results as CU. Cytotoxic effects
were exhibited at concentrations of 12.5 pg/mL and
above. Effects remained steady at concentrations
above 50 pg/mL
Oil Ash Cytotoxicity: Cytotoxic effects were
induced by all. The order of PM fractions inducing
the most Cytotoxicity to the least is the following: 5F,
6HSF, 6HSC, 5C, 6MSC.
Correlation of Metal Content and Cytotoxicity:
Fe, V showed a reasonable correlation. Zn had no
correlation.
Cell Metabolism: An inhibitory effect was observed
with 100 |jg/mL coal ash after 6 h. After 12 h of
exposure, CU, unlike CG, does not continue to
inhibit cell metabolism. Oil ash was generally less
effective than coal ash. The order of PM fractions
inhibiting metabolism the most to the least is the
following: 5F, 6HSC, 5C, 6MSC. 6HSF not tested.
Reference: Okeson et Zn, V, Fe chloride as salts (valence
al. (2004, 0879611 state not reported)
Species: Rat
Cell Type: RLE-6TN
(Type 11 Alveolar
Epithelial Cells)
Particle Size: NR
Route: Cell Culture (50000 cells/well)
Dose/Concentration: 0.001, 0.01, 0.1,1.0,
10 mM
Time to Analysis: 24 h
Cytotoxicity: All metals cytotoxic at concentrations
greater than 0.1 mM. V is 5 times less cytotoxic than
Zn, and Fe is 7 times less cytotoxic than Zn with a
EC50 of 3mM and 4mM, respectively. At 10 mM of
each metal, no surviving cells were present.
NCS: Incubation with NCS (5 or 10 %) decreased
toxicity of Zn, especially at 0.1 mM, but had no effect
on Fe or V toxicity.
Albumin: BSA decreased Zn toxicity at equivalent
concentrations but to a lesser extent than NCS.
Reference: Osornio-
Vargas et al. (2003,
0524171
Species: Mouse
Cell Line/Type:
J774A.1, L929
(Mesenchymal Cells)
PM10
PM25
-N = Northern (industrial)
-SE = Southeastern (lake basin dust)
sites, both heavy vehicular traffic,
Mexico City, Mexico
Particle Size: PM10;PM2 5
Route: Cell Culture (J774A.1:15000
cells/cm2; L929: 30,000 cells/well)
Dose/Concentration: 20, 40, 80 pg/cm2
Time to Analysis: 24-72 h
PM Characterization: Elements similar in particle
types with elements in PM10 more abundant.
Northern particles contained more Co, Zn, Ni, Pb.
Endotoxin: All PM samples had detectable amounts
of endotoxin. PM25 -N had 22 EU/mg. PM,0-N had
30 EU/mg. PM25 -SE had 12 EU/mg. PM,0-SE had
59 EU/mg.
Cytotoxicity (J774A.1): The two northern samples,
PM25 and PMio, both induced similar cytotoxic
effects at 40% survival. PM10-SE and PM25 -SE
induced dose-dependent responses. In general, the
northern samples had a higher cytotoxic effect than
the southern samples.
Apoptosis (J774A.1): Northern samples induced
more apoptosis than did the southeastern samples.
There was no difference between PM10 and PM25
induced apoptosis.
TNF-a and IL-6 (J774A.1): TNF-a and IL-6 induced
dose-dependent increases. At 80 pg/cm2, PM10 -SE
induced the most production of IL-6 followed by
PM25 -SE, PM,o-N , and PM25 -N.
J774A.1 Supernatant Toxicity (L929): Conditioned
medium from J774A.1 pre-exposed to each PM type
reduced cell viability in L929 cells. This was
correlated with TNF-a level in supernatants.
December 2009
D-64
-------
Study
Pollutant
Exposure
Effects
Reference: Penn et al.
(2005, 0882571
Species: Human
Cell Type: BEAS-2B
BDS: Butadiene soot (created on-site
by passing BD through a back-flash
protected stainless steel two-stage
regulator to a stainless steel Bunsen
burner)
-P1:<2.5|jm
-P2: 2.5-1 Opm
-P3:>10pm
BDS-W: solvent washed
Graphite
Composition: <2.5 pm = 92%, 2.5-10
pm = 5%, >10pm = 3%
Particle Size: BDS-P1: <2.5 pm; BDS-
P2: 2.5-10 pm; BDS-P3: >10 pm
Route: Cell Culture (1-1.5x1 Ob cells)
Dose/Concentration: 3 mg BDS
Time to Analysis: 5 min-72 h
Particle Characterization: By weight, EC makes up
94% of BDS, hydrogen 2%, nitrogen and sulfur 1%,
and oxygen less than 0.1%.
PAH Components of BDS: 13 prominent PAHs:
acenaphthylene, fluorene, anthracene,
cyclopentaphenanthrene, fluoranthene,
acephenanthrylene, pyrene, benzofluorenes,
acepyrene, chrysene, benzopyrenes, perylene,
benzoperyiene.
BDS Activity: At 60-120 min, BDS was observed in
the cells. At 4 h, fluorescence observed in
cytoplasmic vesicles and increased during the first
24 h then plateaued for the next 72 h. BDS-W
appeared in vesicles sooner than BDS.
Reference: Pozzi et
al. (2005, 0886101
Species: Mouse
Cell Type: RAW 264.7
PM: Collected continuously for 15 days,
8-10 m from street, Sept 1999, Rome,
Italy
-F = Fine particulate
-C = Coarse particulate
CB (Degussa Huber NG90)
Particle Size: PM-F: 0.4-2.5 pm; PM-
C: 2.5-10 pm;CB: 200-250 nm
Route: Cell Culture (1.3x105 cells/well)
Dose/Concentration: 30 pg/mL; 14 pg/cm2
120pg/mL;54pg/cm2
Time to Analysis: 5, 24 h
Cytotoxicity: For 24 h, lower levels of PM-F, PM-C,
and CB had no effect on cell viability. Higher levels
of PM-C and CB induced a significant release of
LDH.
Arachidonic Acid (AA): Both fractions of PM
increased AA release in a dose-dependent manner
at 5 h. CB increased a release only at the higher
concentrations although, in terms of magnitude, the
CB-induced release was much less than the ambient
PM-induced release. Pretreatment with
deferoxamine was not effective in decreasing AA
release.
TNF-a: TNF-a levels increased significantly for both
concentrations and time periods for PM. PM-C at
24 h was significantly lower than at 5 h for both
concentrations. PM-C at 30 pg/mL induced a much
greater TNF-a release than PM-F at 5 h.
IL-6: PM-F significantly increased at 5 h for both
concentrations. Elevated IL-6 levels were exhibited
at both PM-C doses at 24 h. At 5 h, only the high
dose elevated IL-6 levels. CB was devoid of an
effect on IL-6. LPS-induced IL-6 response was
similar to coarse PM at the high dose, with the
response being greater at 24 h than at 5 h.
Reference: Prophete
et al. (2006, 1568881
Species: Rat
Cell Type: NR8383
AMs
Ambient PM25
NYC: 1 stand26St, NYC
LA: San Gabriel foothills, Claremont,
CA
SEA: 15th Ave S and S. Charleston,
Seattle, WA
V, Mn.AI, Fe levels in PM
added metals to cells
V: Na3V04
AI:AICI3-6H20
Mn: MnCI2-4 h20
Fe: FeCI3-6H20
Particle Size: PM25
Route: Cell Culture (2x105 cells/ml)
Dose/Concentration: Fe(lll) 16 pmol
V, Mn, and Fe(lll) mixtures with V or Mn in
molar ratios 0.02, 0.08, 0.2 and 0.4 x Fe(lll)
Al and Fe(lll) mixtures with Al in molar ratios
0.37, 0.75, 2,7.5 xFe(lll)
Time to Analysis: 20 h
Particle Characterization: Fe and metal to F ratios
based on ratios observed in PM2 5 from LA, SEA and
NYC sites. V: Fe ratios remarkably similar among
sites. Fe levels fixed at NYC level of 16 pm
(highest).
IRP: Coexposure with 3 metals increased IRP
binding activity relative to Fe(lll) alone, by up to 3.5
fold for Al (1.5-3 ratio), 2 fold for Mn (0.08-0.2 ratio)
and 7 fold for V (0.2 ratio). IRP activity dropped at
higher ratios. A drop in IRP activity at higher ratios
may be result of cytotoxicity forAI, but not for V and
Mn.
iNOS: Al induced iNOS expression dose-
dependently There was no observed effect for Mn
andV.
Induction of Hypoxia-inducible Factor (HIF-1a):
Only V and Al induced HIF-1a.
Activation of ERK1 and -2: V and Al induced
pERK1, but only V induced pERK2. Mn had no
increasing effects, but data indicated a decreasing
induction.
December 2009
D-65
-------
Study
Pollutant
Exposure
Effects
Reference: Ramage
and Guy (2004,
0556401
Species: Human
Cell Type: A549
PM10: Collected in Wolverhampton, UK Route: Cell Culture
ufCB: Ultrafine Carbon Particles Dose/Concentration: 80 pg/mL
(Origin not reported) Time to Analysis: 0, 0.5, 3, 6,18 h
Particle Size: PMi0, ufCB: <100 nm
(diameter)
CRP: Treatment with ufCB or PMio produced an
increase in CRP expression with similar effects
noted after 6 h. PM10 induced greater increases than
ufCB. Both the cytoplasm and nucleus contained
CRP.
Hsp70: PMio and ufCB induced increased levels at
all time points with ufCB inducing greater levels than
PMi0. Hsp70 expression was observed in the
cytoplasm and nucleus.
Antioxidants of CRP and Hsp70: Coincubation of
ufCB with Nacystelin and Trolox caused a small
reduction in CRP and Hsp 70.
Reference: Rao et al.
(2005, 0957561
Species: Rat
Strain1 9D
wllallK OLJ
Cell Type: AMs and
cultured lung
fibroblasts
Reference: Reibman
et al. (2003, 1569051
Species: Human
Cell Type: HBEC,
BEAS-2B
DEP: SRM 2975 (NIST)
Particle Size: 05 |jm
UFPM: Ultrafine PM
FPM:FinePM
IPM: Intermediate PM
CPM: Coarse PM
CB: Carbon black
Route: Cell Culture
Dose/Concentration: 200 |jg/mL
Time to Analysis: 4 h
Route: Cell Culture
Dose/Concentration: 11 pg/cm2; 100 pg/mL
Time to Analysis: 6, 18 h
mRNA Expression: No change in IL-1 13 or iNOS
were observed. Data suggests that the lung
fibroblasts is the main source of IL-6 and MCP-1 in
BAL fluid because of their comparatively high
message levels. Due to the extreme variability in
results, the cause of an increase on co-culture with
AMs and/or DEPs was not assessed.
Cytotoxicity: After treatment, cells were more than
90% viable. UFPM and FPM caused no gross
alterations in cell morphology or adhesion.
MIP3o/CCL20 mRNA (6 h): Stimulation of mRNA
released by HBEC upon exposure to UFPM
appeared similar to that provided by TNF-a (5
|jg/mL) and IL-1 13 (10 mg/mL).
All PM collected 8th floor, 26th St and
IstAve, New York City, NY
Particle Size: UFPM: <0.18 pm; FPM:
0.18 -1.0 pm; IPM: 1.0-3.2 pm; CPM:
>3.2 pm
MIP3|3/CCL20 protein in HBEC (18 h): TNF-a and
IL-113 induced a dose-dependent increase in
MIP3o/CCL20 protein (0-10 ng/mL), whereas II-4
and IL-13 induced MIP3a/CCL20 protein release
that reached maximum levels at 1 ng/mL. No
release of MIP1a/CCL3 nor RANTES/CCL-5 was
observed upon stimulation with cytokines.
Secretion of MIP3o/CCL20 in response to PM (18
h): All PM fractions less than 2.5 pm resulted in the
release of MIP3a/CCL20 protein in HBEC roughly
equivalent amounts. CB similar in size to UF/fme PM
did not result in the release of MIP3o/CCL20, nor did
LPS (0.01-1.0 |jg/mL). No release of MIP1o/CCL3
nor RANTES/CCL 5 was observed upon stimulation
by PM fractions.
Activation of MAPK (ERK1/2 and p38): ERK1/2
and p38 was activated by TNF-a, IL-lp, IL-4and IL-
13 within 15 min and was sustained for at least 60
min. Erk1/2 and p38 inhibitors reduced
MIP3o/CCL20 release in BEAS-2B cells in response
to cytokines.
December 2009
D-66
-------
Study
Pollutant
Exposure
Effects
Reference: Riley et al. Zn: ZnCI2
(2003, 0532371
Species: Rat
Cell Type: RLE-6TN
(Type 11 Alveolar
Epithelial Cells)
Cu: CuCI2
Fe: FeCI2
V: VCI4
Ni:NiCI2
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 0.01, 0.1,1.0,10 mM
Time to Analysis: 2, 4, 24, 72 h
Cytotoxicity (SDH): All particles were cytotoxic in a
dose-dependent manner. Zn and V were cytotoxic at
0.05 mM, Cu at 0.5 mM, Ni at 0.8 mM and Fe at 2
mM. ForZn, cell death (LDH) had a biphasic
response: a slow logslope until approx 0.1 mM at
which point it rapidly accelerated to a peak at 5 mM
with a small decline at 10 mM. Most of Zn
cytotoxicity was not due to apoptosis. IPS did not
affect either Zn or Cu cytotoxicity.
Metabolism Inhibition Time Course Response
(Cu and Zn only): At high (1 mM) concentrations,
Zn toxicity peaked at 36-48 h followed by a 2-fold
recovery by 72 h. Cu showed a faster, steady
decline plateauing after 36 h. At low concentrations
(0.1 mM), Cu showed a steady slow decline. At 48 h,
Zn decreased faster to max activity and returned to
control by 72 h.
IL-6 Secretion: Zn and Cu both decreased IL-6
secretion. Decreases were very similar for both
metals and concentrations when expressed as
secretion per viable cell ratio except for Zn at 1.0
mM.
Metal Combinations: Zn and Cu gave variable
results. Zn protected against V cytotoxicity. Zn and
Cu had an additive response. Zn did not affect Fe
toxicity.
Reference: Riley et al. Fe: FeCI2
(2005, 0964521
v ' Ni: NiCI2
Species: Rat, Human
Cu: CuCI2
Cell Type: RLE-6TN,
NR8383 Alveolar V:VC|2
Macrophages, A549
Particle Size: NR
Route: Cell Culture (5x104 cells/well
Alveolar Cells; 1.2x105 cells/well NR8383)
Dose/Concentration: AMs: 0.02, 0.05,
0.07, 0.08 mM; RLE-6TN: 0.1, 0.2, 0.6, 1.0,
6.0mM;A549:0.5, 0.8, 4.4, 4.8 mM
Time to Analysis: 2-48 h
Relative Sensitivity of Cell Strains to Metal
Chloride: NR8383 was more sensitive than RLE-
6TN and A549 except for V where
NR8383 and RLE-6TN were both more sensitive
thanA549.
Relative sensitivity of Cell Strains to Metal
Chloride vs Sulfate: With the exception of Cr,
sulfate was generally more cytotoxic than chloride
(note V valence state).
A649 Cytotoxicity Time Course: Zn cytotoxicity
takes 24 h to develop whereas Cu cytotoxicity
develops within 2 h. LDH release for Cu, however,
develops in 24 h.
RLE Cytotoxicity Time Course: Zn starts at 2 h
and develops until 24 h. Cu develops within 2 h and
continues until 24 h where it is less toxic than Zn.
Both release equivalent amounts of LDH after 24 h.
NR8383 Cytotoxicity Time Course: Both Zn and
Cu exhibit time dependent toxicity beginning as early
as 4 h. LDH release maximizes at 12 h and either
remains steady or declines.
Reference: Ritz et al.
(2007,1989011
Species: Human
Cell Type: BEAS-2B,
NHBEC
DX: Extract of DEP (generated from a
light duty four-cylinder diesel engine
4JB1 type Isuzu Automobile)
Particle Size: <1 pm (diameter)
Route: Cell Culture
Dose/Concentration: 0, 20, 50,100 pg/mL
Time to Analysis: 24 h
NQ01 (Sentinel Phase II Enzyme): Cells
transfected with NQ01 reduced induction of IL-8 by
DX exposure.
Sulfurophane: Increased gene expression of phase
II enzymes, particularly NQ01, was observed in both
cell types. Gene expression in BEAS-2B was greater
than that of NHBEC.
Sulfurophane did not upregulate GSTM1 in BEAS-
2B but induced a 2-fold increase in NHBEC.
Pretreatment also inhibited DX-induction of IL-8 in
both cell types.
Cytokines: DX induced significant increase of IL-8
in both cell types at concentrations of 10 pg/mL or
higher. GM-CSFand IL-8 remained unaffected in
BEAS-2B. GM-CSFand IL-8 Increased in NHBECs
and reached statistical significance at 25 pg/mL
December 2009
D-67
-------
Study
Pollutant
Exposure
Effects
Reference: Rosas
Perez et al. (2007,
0979671
Species: Mouse
Cell Type: J774A.1
PM10
Collected in Mexico City, Mexico from
January-June, 2002
North: Iztacala, manufacturing industry;
Center: Merced, heavy traffic;
South: Ciudad Universitaria, residential
Principal Component Analysis of Air
Pollution Data:
Group 1:S/K/Ca/Ti/Mn/Fe/Zn/Pb (43%
of variance);
Group2:CI/Cr/Ni/Cu(16%);
Group 3: Endotoxins/OC/EC (14%).
For all 3 sites: Averages of Group 1 is
statistically different among the center,
north and south sites with the central
site producing the highest values.
Group 2 is similar among the sites and,
for Group 3, the north had a lower
average than the center and south
sites.
Particle Size: PM10
Route: Cell Culture (1.5x104 cells/cm2)
Dose/Concentration: 20, 40 or 80 pg/cm2
Time to Analysis: 72 h
Cytotoxicity: Responses were dose-dependent;
there was no observed site interaction. Cytotoxicity
seems to be a result of the following components:
S/K/Ca/Ti/Mn/Fe/Zn/Pb.
IL-6: Only the center site at 40 pg/cm2 induced an
increase. Induction of higher IL-6 levels seems to be
related to high values of S/K/Ca/Ti/Mn/Fe/Zn/Pb and
endotoxins/OC/EC.
TNF-a: Production was induced by all samples in a
dose-dependent manner. Similar to IL-6, induction of
higher TNF-a levels seems to be a result of high
values of S/K/Ca/Ti/Mn/Fe/Zn/Pb and
endotoxins/OC/EC.
p63: Only south PM had effect. Induction of p54
seems to depend on high levels of CI/Cr/Ni/Cu and
low levels of S/K/Ca/Ti/Mn/Fe/Zn/Pb.
Reference: Sakamoto PM,0: EHC-93 (Obtained from Health
etal. (2007, 0962821 Canada, Canata)
Species: Human
Age: 58-82 yr
(Smokers)
Cell Type: HBEC
Particle Size: PM10
Route: Cell Culture
Dose/Concentration: 100, 300 and 500
pg/mL
Time to Analysis: Calcium responses: up to
60 min; cytokines: 6 or 24 h
Intracellular [Ca2+]: [Ca] concentration slowly
increased, elevating after 10 and 30 min for 500 and
300 mg/mL, respectively. The response plateaued at
35 min for 500 pg/mL.
Extracellular [Ca2+]: Starting at 20 min, the
removal of extracellular Ca decreased the PM10
response significantly. Calcium channel blocker
(10|jM or 1mM) LaCIS and (5mM) NICI2 significantly
blocked the PM-induced intracellular Ca. Lacl2
administration (1mM) inhibited the PM-induced
Ca2+ response in a dose-dependent manner.
Mode of Action: Intracellular Ca induced by ATP
declined more slowly in the cells exposed by PM10.
This indicates that PMio blocks Ca clearance via the
calcium pumps.
Cytokines: PMio induced a dose-dependent
increase in cytokine mRNA levels and cytokines IL-
1P, LIF, IL-8 and GM-CSF. Cytokine expression was
unaffected by the reduction of extracellular Ca2+ .
Preincubation with the calcium chelator reduced
responses for IL-lp and IL-8 but not LIF or GM-CSF.
Reference: Salnikow
et al. (2004, 0874691
Species: Human
Cell Line: 1 hAEo-
FeS04
FeCI3
NiS04
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 0.25 and 0.5 mM
Fe exposures also contained 60 pg/mL
apotransferrin
Time to Analysis: 24 h
Cytotoxicity: Both Fe had no effect. NiS04 caused
marginal Cytotoxicity (75%).
Hypoxic Stress: At 20 h, NiS04 (at concentrations
of 0.25 or 0.5 mM) induced NDRG-1/Cap43 protein
production indicating hypoxic stress. DFX and
DMOG induced a similar effect.
IL-8: NiS04 induced IL-8 time-dependently for up to
48 h. At 48 h, the increase was 6+ fold.
Coexposure (Ni + Fe) on Fe uptake: Fe(lll) uptake
was greater than Fe(ll) uptake. NiS04 had no effect.
Ni uptake was greater than Fe uptake but was
decreased by coexposure to Fe. Coexposure also
did not effect hypoxic stress. Coexposure with Fe
did reduce Ni-induced IL-8 production.
December 2009
D-68
-------
Study
Pollutant
Exposure
Effects
Reference: Salonen et PMi0 (urban traffic) Finland Route: (2^10 cells/well)
al (2004 1870531
' Pooled as winter (W), spring I (SI), or Dose/Concentration: 15, 50,150, 500,
Species: Mouse spring II (Sll) based on component/time 1000|jg/mL
considerations
Cell Type: RAW 264.7 Time to Analysis: 0, 24 h
Particle Size: PM10: 0.12-10 pm
Air quality parameters: Winter and spring I did not
differ. Sll much lower PM25
Metal data equivocal as well as highly variable
resuspension rates.
Total PAHs: W=303; Sl=233; SI 1=204 ng/mg
Inflammation (IL-6, TNF-o, N0)/Cytotoxicity: A
dose-dependent increase was observed for TNF-a,
IL-6 and NO except for SI. The IL-6 levels, of those
particles exposed to SI, decreased at 1000 pg/mL
TNF-a, IL-6: SI = SI l»W>control.
NO production: W>SI>SII
Cell Viability: W=SI=SII toxic at 500 and 1000
pg/mL
Water-soluble vs Insoluble: TNF-a and IL-6 were
nearly entirely the result of insoluble components of
PM10. Cytotoxicity was driven by both soluble and
insoluble components.
Metal Chelation: The addition of metal chelators did
not modify IL-6, TNF-a or cytotoxicity
IPS inhibitor: Treatment with the LPS inhibitor
eliminated the IL-6 response and, perhaps, slightly
reduced the TNF-a response but not cytotoxicity
Hydroxyl radicals: A dose-dependent induction of
hydroxyl radicals and induction of hydroxyl radical
lesions (at 500 and 1000 pg/m ) in the calf thymus
DNA were observed.
Reference: Samet et As: NaAS03
al. (2003, 1137821
Species: Human
Cell Type: A431
(Epidermoid Cells)
V: VOS04
Zn: ZnS04
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 5QQ|j[vl
Time to Analysis: 20, 30 or 90 min
EGFR Dimerization: Zn, V or As did not induce
EGFR dimerization in a cell free system i.e., no
direct crosslinking. Zn did not induce dimerization in
whole cells either.
Phosphorylation of EGFR: Zn induced
phosphorylation at 3 sites similar to EGF As and V
had no effect.
EGFR Kinase Inhibitor: While EGF action was
blocked, Zn continued to induce phosphorylation
and was independent of EGFR kinase activity.
c-Src: Blocking of c-Src tyrosine kinase
(transactivator of phosphorylation) negated all Zn-
induced phosphorylation but only had a slight effect
on EGF stimulated cells.
ERK1/2 Phosphorylation: Zn increased levels of
ERK1/2. Pretreatment with EFGR kinase inhibitor
reduced both Zn and EGF effect. This effect was not
blocked by the c-Src blocker.
Reference: Santini et DEP: Collected adjacent to moderate
al. (2004, 0878791 traffic in Rome, Italy
Species: Mouse Particle Size: PM25
Cell Type: RAW 264.7
Route: Cell Culture (2.5x105 cells/ml)
Dose/Concentration: 0.01, 0.1,1.0 pg/mL
Time to Analysis: 24 h
600 MHz Results (no 1 ug/mL): DEP induced a
dose-dependent increase in choline compounds, a-
and pgamma- glutamine/glutamate (0.01 >0.1
pg/mL), lactate, and CH2, CH3 moieties of fatty
acids. DEP decreased inositol and
phosphoreatinine.
700 MHz Results (no 1 ug/mL): DEP induced
similar results, except a-, pgamma-glutamine were
dose-dependent. Inositol showed no effect. Taurine
slightly increased. Results were confirmed after
eliminating biological interferences via perchloric
acid.
Growth Curves/Cell Cycle Analyses/Cell
Morphology: DEP had no effect.
Cytokines: IL-6 levels increased at 0.1 and 1
pg/mL TNF-a was unaffected.
December 2009
D-69
-------
Study
Pollutant
Exposure
Effects
Reference: Saxena et DEP: SRM 1650
al. (2003, 0969861
Species: Mouse
Cell Type: RAW 264.7
CO: Crude Organic Extract of DEP
Fractionated into
asphaltene (pentane/hexane),
saturated hydrocarbon,
less polar (aromatic) hydrocarbon,
more polar (aromatic) hydrocarbon,
resins, residual (resins)
Particle Size: NR
Route: Cell Culture (2.5*104 cells/ml) Cytotoxicity: No cytotoxic effects were observed.
NO: DEP alone induced NO in a dose-dependent
manner which peaked after 1 day and plateaued for
days 2 and 3. IFN-y + DEP showed dose- and time-
dependency. IPS + DEP showed no effect at 1 day,
but dose-dependently reduced NO production on
days 2 and 3. Addition of Bacillus Calmette-Guerin
(BCG) eliminated the effect of DEP at 2 days but
showed a dose-dependent decrease at 3 days.
Effectiveness of Particulate Components: The
carbonaceous core of DEP did not affect BC G-
stimulated NO production. CO significantly inhibited
BCG-stimulated NO production. Study indicated that
the extract of aromatic hydrocarbons and resins
caused an inhibitory effect in a dose-dependent
Dose/Concentration: DEP, CO 5,10,15,
20, 25 pg/mL
IFN-y: 10 ng/mL
IPS: 1 mg/mL
Time to Analysis: 1-3 days
Reference: Seagrave
et al. (2007, 0975491
Species: Human
Gender: Male (3
donors)
Age: 16, 23 yr
Cell Type: A549
DE: Generated by DE 5500 watt
generator using #2 certification oil
performed under SOOOw load.
Emissions diluted to 3 mg/m3 total
particulate matter.
Particle Size: 0.14-0.5pm
Route: Air Liquid Interface
Dose/Concentration: 8.33 mL/min/well
Time to Analysis: 3 h exposure; 1 or 21 h
post-exposure
Particle Deposition: 140 and 500 nm microspheres
demonstrated uniform deposition of approx. 10%.
Transepithelial Electric Resistance: No effect of
DE; rather, more effect was observed from air
controls.
Macromolecular permeability: DE caused an
increase 1 h but returned to control at 21 h.
LDH/Cytotoxicity: DE had a highly variablefdonor
specific) effect at 1 h and returned to control levels
at 21 h
Mitochondria! activity (WST): DE reduced activity
at 1 h and possibly increased activity at 21 h (high
donor-to-donor variability)
Mucus Like Substance Excretion: There was high
donor to donor variability; no overall effects were
observed.
Alkaline Phosphatase (AP): DE decreased at 1 h
and perhaps increased at 21 h
Glutathione: DE caused a large decrease at 1 h but
returned to normal at 21 h.
HO-1: After DE exposure, levels increased but were
still lower than air exposed controls
Cytokines: No differences for IL- 8 or 12, TNF-a,
GM-GSF, IL-1a, or IFN-y were observed. IL-4 and -6
were decreased upon DE exposure.
Reference: Seagrave
et al. (2004, 0874701
Species: Human
Cell Type: A549
DPM: SRM2975 (NIST)
DPM-0: DPM organic extract
(acetone/DCM)
CB: Carbon Black (Elftex-12, Cabot)
Particle Size: CB: 37 nm; Stokes
diameter! 98 nm
Route: Cell Culture (1 xlOb cells/well)
Dose/Concentration: 0.03 -1,000 pg/cm2
Time to Analysis: 0,18 h
IL-8 release: DPM increased semi dose-
dependently (perhaps steady based on error range)
up to 1 pg/cm after which IL-8 declined dose
dependently to zero (control = 100%) at 300 and
1000 pg/cm2. LDH release was steady which
indicates no cytotoxicity.
DPM interaction with IL-8: DPM depletes IL-8 from
solution in a dose-dependent manner (cell free).
BSA preincubation reduced the slope of the dose
response but not the final result. CB has no effect.
DPM-0 residuals act identical to DPM. Increasing
NaCI concentrations reduced DPMs depletion of IL-8
Neutrophil responses: DPM and bound IL-8
together caused a marked aggregation of cells
resulting in spindle shapes. DEM or IL-8 alone did
not cause this aggregation although DEP did recruit
neutrophils
December 2009
D-70
-------
Study
Pollutant
Exposure
Effects
Reference: Seagrave PM filter collection
et al. (2003, 0549791
Collected from diesel or gasoline
Species: Human, Rat powered vehicles as follows:
Cell Line: F344/&I BR BG: BS Gasoline
G30: Normal Emitter gasoline (30F)
Age: 11 wk (mouse)
G: Normal emitter gasoline (72F)
HD: High Emitter Diesel
D30: current technology diesel (30F)
D: current technology diesel (72F)
WG: White Smoke Gasoline
Particle Size: NR
Route: Cell Culture (1 xlOb cells/well)
Dose/Concentration: 0.03-10,000 pg/cm2
Time to Analysis: 16-18 h
Weight: 250 g
Cell Type: A549, AMs
Cytotoxicity: LDH activity increased in A549 cells.
The types of pollutants that are most toxic, in
decreasing order of cytotoxicity, are the following:
BG, G30, and G which are significantly different from
HD, D30, D, WG which are also significantly
different from DS. LDH activity also increased in rat
macrophages. G, G30, and BG were the most toxic.
HD and D30 were intermediately toxic and D, WG,
and DS were the least toxic. In both cell types,
gasoline was more cytotoxic than diesel.
Cytokines: All particle types except DS increased
IL-8 levels in A549 though not all increases were
statistically significant. Also, many particle samples
at high concentrations produced an apparent
suppression of IL-8 release.
Alkaline Phosphatase: G30 and G were more
potent than the other particle samples in A549. WG
and D30 induced no significant effects. For A549
cells, activity increased at low concentrations and
was suppressed at higher concentrations.
Macrophage Peroxide Production: In rat AMs,
peroxide production was often the highest at the
lowest concentrations and the lowest production
caused by the highest concentrations. D30 followed
this trend and induced the highest production as well
as the greatest suppression. Using two different
statistical methods, D30 >6 others which in turn
>DS. Using the second method D30 and D >all other
6. Order of potency between two methods
completely different. Authors noted that in vitro
potency quite different from in vivo potency
(previous paper).
Reference: Seaton. et PM25 from London Route: Cell Culture
al (2005 1989041
PM10 from Manchester (positive control) Dose/Concentration: 1-100 pg/mL
Species: Human
Cell Type: A549
PM from Holland Park, Hampstead and Time to Analysis: Cytotoxicity: 24 h; IL-8:
Oxford Circus stations (HP, HR and h; Generation of hydroxyl radicals: 8 h
OC)
Particle Size: PM25,PM10, Holland
Park, Hampstead and Oxford Circus
PM had a median diameter of 0.4 pm.
80% of the particles had a diameter
less than 1 pm.
Cytotoxicity: Dust from all three tunnels (Holland
Park, Hampstead and Oxford Circus) were able to
cause cell death (LDH). The release of LDH
indicated a dose-dependent relationship. The
highest dose of Holland Park PM induced the -17%
release of LDH, Hampstead triggered ~ 13% and
Oxford Circus -3% (no different than control). PM10
from Manchester caused a 7% LDH release at the
highest dose. The negative control (Ti02) caused no
response (2% release at highest dose).
IL-8: All three tunnel PMs induced a dose-
dependent release of IL-8. At the highest dose, all
three tunnel dusts induced IL-8 stimulation more so
than the control site PM25. HP induced a 3 fold
increase. Also, the highest Ti02 concentration
caused the least IL-8 stimulation.
Hydroxyl Radical Generation/ DMA Plasmid
assay: The plasmid assay indicated that the tunnel
dusts induce more free radical activity than the
Manchester PMi0 and Ti02.
HP nearly doubled the percentage of DNA damage
with intermediate results for HR and OC. Results for
PM10, Ti02 and control were identical
December 2009
D-71
-------
Study
Pollutant
Exposure
Effects
Reference: Singal and AE2:Aerosil 200, amorphous silica Route: Cell Culture (5^10 cells/well)
Finkelstein (2005,
1989051
Species: Human,
Mouse
Cell Type/Line:
A549Lud lung
adenocarcinoma
epithelial cell line
(human), MLE15Lud
and MLE15Luc2
(mouse)
All cells contain human
cytokine IL-8
controlling firefly
luciferase
(Degussa)
Cl: Carbon iron particles (25% Fe)
Dose/Concentration: 18 pg/mL, 36 pg/mL,
72 pg/mLall in 1 ml/well
Particle Size: AE2:12 nm surface area Time to Analysis: 24 h
~200±25m2/g;CI:~40nm
Luc Activity: Luciferase enzyme activity is
significantly less in MLE15Luc2 cells than in
MLE15Lud cells. For both cells, luciferase activity is
time- and dose-dependent peaking at 4-8 h.
Aerosil 200: AE2 induced dose- and time-
dependent Luc response which peaked at 3 h and
decreased thereafter in a similar way as TNF-a.
Contrary to TNF-a, AE2 induced much cytotoxicity
starting at 6 h.
Effect of Proteasomal Inhibitors (MG-132):
Inhibitor reduced AE2 Luc activity to near control
levels. Similarly, LDH-cytotoxicity was halved
A649 Human Cell Response: AE2 acted similarly
to the MLE response. Cl particles showed slightly
less activity without peaks. AE2 increased
cytotoxicity after 12 h, whereas Cl had no effect.
Contrary to MLE mouse, MGJ32 did not affect Luc
activity but PD98059 (selective noncompetitive
inhibitor of the MAP pathway) and SN50 (NF-KB
inhibitor) reduced AE2 and Cl-induced activity.
Reference: Song et al.
(2008, 1560931
Species: Rat
Cell Type: RAW 264.7
DEP collected from a 4JB1-type, light-
duty (2740 cc), four-cylinder diesel
engine operated using standard diesel
fuel at speeds of 1500 rpm under a
load of 10 torque.
Particle Size: 0.4 pm (mean diameter)
Route: Cell Culture (5x10 cells seeded on
a 24-well plate)
Dose/Concentration: 50 |jg/mL
Time to Analysis: 72 h
Nitrite Production: 50 pg/mL of DEP induced
production when compared to the control. Over the
72 h period, a general trend was not observed, but
maximal induction of nitrite occurred at 4 h after
stimulation.
Reference: PMC: PM Coarse
Steerenberg et al.
(2006, 088249) PMF: PM "ne
Species: Rat, Human Ambient air samples collected from
Route: Cell Culture
Dose/Concentration: NR
Time to Analysis: 20 h
Crustal material (metals and endotoxin but not Ti,
As, Cd, Zn, V, Ni, Se) were positively associated
with AM IL-6 and TNF-a and Type 2 MIP-2 and IL-6.
Sea spray (Na and Cl) was also correlated with AM
IL-6.
Cell Type: AM (rat),
Type 2 cells (rat), A549
Poland; Amsterdam, the Netherlands;
De Zilk, the Netherlands.
Particle Size: PMC: 2.35-8.5 pm; PMF:
0.12-2.35 pm
Reference: Tal et al.
(2006, 1085881
Species: Human
Cell Type: HAEC
100 mM Zn(ll) orV(IV) stock solutions
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 500 pmol
Time to Analysis: 5, 20 min
Zn-mediated EGFR Phosphorylation: EGFR
kinase activity was required but not EFGR ligand
binding. EGFR Kinase inhibition reduced Zn
mediated EGFR activation, (authors NOTE:
complete reverse of results in B82L and A431 cells).
Src Kinase is not required. Zn inhibiting Src kinase
was nearly total after 20 min.
EGFR-Specific Protein Tyrase Phosphatase
(PTP): Zn inhibited PTPs, similar to V(IV) resulting in
a decrease of exogenous EGFR dephosphorylation
Reference: Tamaoki et UFCB: Ultrafine Carbon Black - (Tokai Route: Cell Culture (104 cells/well)
al. (2004,1570401 Carbon, Japan) .,^,,0,^c
Dose/Concentration: 6.1,12.3,18.4, 24.5,
Species: Human FCB: Fine Carbon Black (Tokai Carbon, 30.7 pg/cm2
Jaoanl
Cell Type: HBEC Time to Analysis: Up to 72 h
Particle Size: UFCB: 11+0.5 nm
(mean diameter)
FCB: 250+16 nm (mean diameter)
DMA Synthesis/ Protein Synthesis: Synthesis
increased by UFCB (30.7) for up to 72 h and
flattened after 48 h. FCB had no effect. UFCB also
showed a dose-dependent response beginning at
12.3 pg/cm2 up to 24.5 after which the response
plateaued. The addition of Cu/Zn Super oxide
dismutase (SOD) or a NADPH oxidase inhibitor
completely inhibited the UFCB effects. Similarly, two
different EGFR tyrosine kinase inhibitors, and a Me
inhibitor all reduced UFCB response to control
levels.
ERK activation: UFCB caused phosphorylation of
ERK beginning at 2 min, peaking at 5 min and
decreasing at 10 min. ERK activation was inhibited
by EGFR tyrosine kinase inhibitor Cu/Zn SOD and
neutralizing body for HB-EGF but not by PDGF-R
kinase inhibitor.
HE (polyclonal heparin binding)-EGF release:
UFCB induced rapid cell surface loss with recovery
after 20 min and nearly full recovery at 360 min.
Metalloproteinase inhibitor and Cu/Zn SOD both
prevented HB-EGF release.
December 2009
D-72
-------
Study
Pollutant
Exposure
Effects
Reference: Tao and UAP: Urban Air Particles (SRM 1649)
Kobzik (2002, 1570441
Species: Rat
Cell Type: RLE-6TN
(Alveolar Type II
Epithelial Cells), Fetal
Lung Fibroblasts
(RFL), AMs
Ti02
Si02
ROFA
Particle Size: Ti02: ~1 pm; Si02: ~1
pm; ROFA: NR
Route: Cell Culture (1x1 Ob cells AM
1.4x105 cells RLE/RFL)
Dose/Concentration: 1-50 pg/mL
Time to Analysis: 24 h
Cytokines: TNF-a and MIP-2 in RLE was
unaffected by any particle samples. TNF-a and MIP-
2 in AM significantly increased with 25 pg/mL UAP.
TNF-a and MIP-2 in the co-culture of AM + RLE
increased with each particle. The order of particles
in decreasing order are as follows: Si02 at 25pg/mL,
UAP at 12.5 fjg/mL, ROFA at 25 pg/mL, and Ti02 at
50 pg/mL Except for Si02, the blocking of effects
caused by LPS absorbed on the particles did not
affect the cytokine response. For Si02, the response
was reduced but still above the control.
Co-culture: Physically separating AM and RLE cells
and adding PM completely negated the co-culture's
response to PMs. This indicates that cell to cell
contact is required for co-culture potentiation of PM
effects.
Inhibitors: Various inhibitors of cell adhesion
molecules (heparin, p -1, 2 or 3 integrin) had no
effect on UAP-induced cytokine release.
Reference: Veranth et Artificial particles and PMs
al. (2007, 0903461
N-AI: nano alumina AI203
Species: Human
Cell Type: BEAS-2B,
A549, NHBE
M-AI: Micro AI203
N-Ce: nano Ce02
M-Ce: micro Ce02
N-Fe: nano Fe203
M-Fe: micro Fe203
N-Ni: nanoNiO
M-Ni: micro NiO
N-Si: nano Si02
M-Si: micro Si02
N-Ti: nanoTi02
M-Ti: micro Ti02
KLN: kaolin
MUS: Min-U-Sil (ground crystalline
silica)
DD: desert rural soil Utah PM25
JE: Juarez, urban street PM25
MNC: Mancos, rural Utah PM25
LPS: lipopolysaccharide
V:VOS04 (soluble) (19 pg/mL)
Particle Size: (Surface mean diameter)
N-AI: 6 nm (261 nf/gl
M-AI: 210 nm (7.7 nf/g)
N-Ce:14nm(71 m2/g)
M-Ce:1500nm(0.6m2/g)
N-Fe: 5 nm (221 m2/g)
M-Fe:100nm(12nf/g)
N-Ni:6nm(145m2/g)
M-Ni: 16nm(57nf/g)
N-Si: 19 nm (127(11%)
M-Si: 440 nm (5.4 m7g)
N-Ti: 6 nm (242 m2/g)
M-Ti: 410 nm (3.5 nf/g)
KLN: 100 nm (24.3 nf/g)
MUS: (NOS <5 pml
DD: 400 nm (6.2 m /g)
JE: (NOS <3 fjm)
MNC: 200 nm (13.0 nf/g)
Route: Cell Culture (35,000 cells/cm2 BEAS;
2500 cells/cm2 NHBE; 20,000 cells/cm2
A549)
Dose/Concentration: 0.53, 5.3 and 53
|jg/cnf(=1,10,100|jg/mL)
Time to Analysis: 24 h
Cell Viability: Except for Ni and V no cytotoxicity
was observed at the highest concentration.
IL-6 Secretion in BEAS-2 B Cells: Nano and micro
sizes of the same metal showed no differences in
response (high experiment to experiment variability).
In general, the soil-derived dusts (JE, DD, MNC)
were more potent than the metal and ceramic oxide
particles. In KGM media, BEAS-2B cells are more
responsive to vanadium and other soluble metals
and less responsive to LPS, but this relationship is
reversed in LHC-9 media.
IL-8 Secretion in BEAS/LHC vs NHBE in BEGM
Cells: Levels were much higher in NHBE cells than
BEAS-2B cells. For BEAS-2B, the nano size Si and
both sizes of Ni induced levels statistically greater
than the control. For NHBE, only Si and Ni (for both
sizes) were statistically greater than control.
IL-6 in NHBE: The nano and micro sized particles of
Al, Ce, Fe and nano sized Si all induced statistically
significant increases. Control levels of IL-6 were
much higher in NHBE cells than in BEAS-2B cells.
Secretion induced by pure oxide particles was small
for both the mid and high concentration levels (5.3
and 53 pg/cm ).
BSA/ Bovine Serum Addition Effect: In a fixed
solution nano-Ni, nano-Ti and KLN all reduced the
measured IL-6 by 60+ percent. Addition of BSA or
bovine serum dose dependently reduced the action
of the particles to near control levels.
PM Effects (without added protein) on IL-6 In
Solution: Increasing metal concentration did not
affect a fixed IL-6 concentration until the 100 or 316
pg/mL levels.
December 2009
D-73
-------
Study
Pollutant
Exposure
Effects
Reference: Veranth et
al. (2007, 0903461
Species: Human,
mouse, rat
Cell Type: A549,
BEAS-2B (types E and
U), RAW 264.7,
Primary macrophages
S: desert dust (collected from unpaved
desert road in Utah, PM25 enriched)
V: vanadium soluble (prepared from
VOS04, Alfa Aesar, Ward Hill, MA)
C: Coal fly ash (PM2 5 enriched and
derived from commercial power plant
burning Utah bituminous coal)
D: Diesel PM (tail-pipe particles
collected from high emitting BSr on-
road light duty truck)
L: Lipopolysaccharide
T: Titanium dioxide (Alfa Aesar)
K: Kaolin (purchased from Capitol
Ceramics, UT)
Particle Size: BET surface (m2/g)
S: 6.2 (PM25 enriched)
V:NA
C: 5.4 (PM2.5 enriched)
D:NR
LIMA
T: 3.5 (1-2 pm)
K: 24 (<200 mesh = 74 \im)
Route: Cell Culture
Dose/Concentrations: Maximum
concentrations:
S= 100|jg/cm2
V = 100 pg/cm2
C=100|jg/cm2
D = 32 pg/cm2
L=1000EU/mL
T = 100 pg/cm2
K= 100|jg/cm2
Time to Analysis: 24 h
Viability: Generally, cell viability was greater than
75% of the control post treatment. Vanadium, at the
highest concentration, induced less than 50% of
control viability whereas kaolin, also at the highest
concentration, induced cell death.
IL-6: BEAS-2B E or U in LHC-9 showed a response
to S and L. BEAS-2B (U) was in LHC-9 medium with
added serum (FBS). This resulted in a doubling of
response coupled with at least an 8 fold increase in
control levels. BEAS-2B (E) showed response for S
and V but not L. A549 showed response to S and K.
RAW 264.7 and Rat macrophages showed
responses to Sfvery low) and L. In general, the IL-6
responses in A549 and RAW 264.7 were similar and
significantly lower than the responses in rat
macrophages or BEAS-2B.
Effect of Culture Media Composition (BEAS-2B):
Varying ratios of LHC-9 and KGM media resulted in
a near 10 fold increase in control rate once LHC was
33% or more of the media. Upon Soil Dust (NOS)
exposure IL-6 increased linearly with % LHC-9 in
culture/exposure media. Addition of calf serum (0.1-
10 %) raised control IL-6 levels at least 40 fold. At a
steady PM concentration, the addition of serum
resulted in a log-linear increase in IL-6 release which
blocked any PM effect.
Reversibility of Media Effect: Changing media with
every passage showed that media effects do not
persist once media are changed.
Culture Well Size: Going from a 6 well to 96 well
plate (decreasing well size) increased IL-6 control
values about ten fold, while the positive control
(TNF) response increased 3 fold. Hence the
sensitivity of the test (i.e., positive/control response)
declined from 11 fold to 3 fold with increasing well
number / decreasing well size. Because cell seeding
density and the like were held constant, these
changes suggest that edge effects are the cause of
the IL-6 changes.
Reference: Veranth et PM2.s samples from 28 samples from 8
locations in Utah, New Mexico and
Texas (rural, industrial, road side,
military)
al. (2006, 0874791
Species: Human
Cell Type: BEAS-2B
2 coal fly ash samples (a product of
combustion using Utah bituminous coal
and New Mexico bituminous coal)
Ti02
kaolin clay
Particle Size: PM25; Ti02:1-2 pm
Route: Cell Culture (35,000 cells/ cm2)
Dose/Concentration: 10, 20, 40, 80 pg/cm2
Time to Analysis: 24 h
Cell Assays: In sample soils viability declined dose
dependency while IL-6 increased dose-dependently
IL-8 was highly variable (peak at 20 pg/l, dose-
dependent increase or flat response.)
IL-6 Assays for All Soil PMs: Soils ranged across
an order of magnitude greater than LPS, coal fly
ash, Ti02 or kaolin samples. One soil even
exceeded the pos V control at equal concentrations
Correlation with Cell Viability: Correlation was
strong for Mn (p<0.001) and weak for ECS, K, Se,
andHg(0.01
-------
Study
Pollutant
Exposure
Effects
Reference: Veranth et PM25 enriched soil samples
al. (2004, 0874801 ^ _, ^ _, t _,
DD: desert dust, unpaved road, Utah
Species: Human
Cell Type: BEAS-2B
WM: West Mesa, sandy grazing site,
NM
R40: Range 40 gravel soil, TX
UN :Uinta, sandy soil, UT
Particle Size: 0.4-3 pm
Route: Cell Culture (20,000/crri)
Dose/Concentration: 10, 20, 40, 80,160
pg/cm2
Time to Analysis: 24 h
Elemental Analysis of PM: Major differences UN
generally lower in major minerals but high Fe
content and high EC. High Mn. Low Pb and Zn
Cytotoxicity: UN and WM were the most cytotoxic
at all dose levels, followed by R40 and DD. All
particles showed a dose-dependent cytotoxic
response.
IL-6 Release: DD and R40 (up to the 160 pg/cm2)
showed dose-dependent responses and induced an
8-fold increase at the highest concentration levels.
WM peaked at 40 pg/cm2 and UN induced similar
responses above 10 pg/cm2.
IL- 8 Release: DD induced a dose-dependent
response. WM peaked at 10 pg/cm2. Release
induced by DD and WM seemed to be limited by
toxicity There was no treatment with R40.
TNF-a: DD, WM and UN induced release was not
detected at the 40 or 80 pg/cm2 concentrations.
IPS: IPS was the primary factor in inducing IL-6
release when exposed to LPS-containing mixtures.
LPS alone induced lesser responses than treatment
to the environmental dust particles. TiLPS induced a
less than two-fold increase in IL-6 versus the over
seven-fold increase induced by soil dust positive
control. LPS treatments were less cytotoxic than DD.
Limited IL-6 and IL-8 responses were observed at
2000 EU/mL compared with DD at 80 pg/cm
Endotoxin: Inverse relationship between endotoxin
content and IL-6 release was observed.
Viability vs Physical Modification of Dust Sample
(no UN): Only leaching in a variety of water based
vehicles increased viability minimally (generally <25
%). Heat treatment (150-, 300, 550° F) and
methanol extraction had no effect
IL-6 Release vs Physical Modification of Dust
Sample (no UN): One hour thermal treatment at
150° F had no effect on IL-6 response. All other
treatments reduced IL-6 release (heat 350°, 500°
and extractions).
Reference: Veronesi
et al. (2002, 0245991
Species: Human
Cell Type: BEAS-2B
Ambient PM
- St. Louis: Urban particulates
- Ottawa: Urban particulates
-MSH: Volcanic dust from Washington
state's Mt. St. Helen
-Woodstove: Woodstove particles from
conventional fireplace burner
-CFA: Coal fly ash from western U.S.
power plant
-OFA: Oil fly ash from Niagara, NY
-A: Total Fractions
- B: Soluble Fractions
- C: Washed Fractions
Particle Size: PM >2.5 pm; PM: 2-10
pm; PM >10 pm
Route: Cell Culture
Dose/Concentration: 50 pg/mL; 30 pg/cm2
100pg/mL; 60 pg/cm2
Time to Analysis: 4,16 h
Ca: Calcium increased significantly with all particles
types.
IL-6: At 50 and 100 pg/mL, IL-6 increased with all
particle types at 4 and 16 h. Overall, fraction -A was
the most potent.
Surface charge: Surface charge correlated strongly
with increases in both Ca2* and IL-6 levels. OFA,
however, was unmeasurable due to technical
difficulties.
December 2009
D-75
-------
Study
Pollutant
Exposure
Effects
Reference: Vogel et
al. (2005, 0878911
Species: Human
Cell Type: U937
ATCC) monocytes
macrophage
differentiation)
UDP:SRM 1649 (MIST)
UDP-OE: DCM extract of SRM-1649,
0.45 pm filter
sUDP: stripped particles UDP
DEP: SRM 2975 (NIST)
DEP-OE: DCM extract of SRM-2975,
0.45 pm filter
sDEP: stripped particles DEP
CB95: Carbon Black (Degussa)
Particle Size: UDP, DEP: NR;
CB95: 95 nm
Route: Cell Culture (2x10b - 2x10b cells/ml)
Dose/Concentration: DEP, UDP: 2.5,10 or
40 pg/cm2
(eq to 12.5, 40, 200|jg/mL)
DEP-OE, UDP- OE: 10 pg/cm2 (particle
equivalent)
Time to Analysis: 24 h
Effect On mRNA Expression (COX-2, TNF-o, IL-6,
IL-8, C/EBPp, CRP, CYP1a1): All DEP and UDP
induced dose-dependent increases. IL-6 tended to
plateau at 10 pg/cm . Generally, with the exception
of COX-2, UDP effects on genes were stronger than
DEP.
Cytotoxicity: Both DEP and UDP were cytotoxic at
40 pg/cm2
Fractionation and mRNA Expression: For COX-2,
TNF-a, IL-8 mRNA fractions were much more active
than parent particles and consequently stripped
particles were much less active than parent
particles. CB95 had no effect. The reverse effect
occurred for IL-6 and CRP mRNA expression. The
particles that induced mRNA expression in
decreasing order are: sUDP, UDP, UDP-OE.
Inhibition Of mRNA Expression: CRP:
pretreatment with IgG and wortmannin (Fey receptor
binding and ingestion dependent inhibitors resp)
blocked the effects of DEP, UDP and sDEP and
sUDP Luteolin (AhR inhibitor) had no effect.
COX-2: Only luteolin inhibited COX-2 expression for
DEP, DEP-OE, UDP, and UDP-OE.
CYP1a1: Luteolin also inhibited OE-DEP and OE-
DUP effects (only those two particles tested).
Cholesterol Accumulation: DEP, UDP and UDP-
OE and DEP-OE at 10 pg/cm2 all increased
cholesterol accumulation by at least 2 fold
Reference: Wang et
al. (2003, 1571061
Species: Rat
Cell Type: Lung
Myofibroblasts
V205: (Aldrich Chemical Co.,
Wisconsin)
Particle Size: NR
Route: Cell Culture (1 xio5 cells/100 mm
dish; 3.2x104 cells/cm2)
Dose/Concentration: 400 |jm
Time to Analysis: 0.5,1,4, 24 h
H202 Drives STAT-1 Activation: Pretreatment with
NAC or catalase reduced V205-induced STAT
activation by more than 90% and completely
abolished H202-induced STAT activation. Within 5
min of V20s treatment, H202 was significantly
decreased in the supernatants of cultured
myofibroblasts and suppression of H202 levels
continued for up to 24 h post V205 treatment. This
supports the findings that myofibroblast-generated
H202 is required for V205-induced STAT activation.
Temporal STAT-1 Activation: H202 induced rapid
activation within minutes whereas activation by V205
occurred more slowly (beginning 8h post treatment).
p38, ERK, EGFR: p38 and EGFR are required for
H202- or VA-induced STAT-1 activation whereas
ERK is not required
Reference: Whitekus
et al. (2002, 1571421
Species: Mouse
Cell Line: RAW264.7
DEP (light-duty, four-cylinder engine-
4JB1 type, Isuzu Automobile, Japan;
standard diesel fuel) (extracts)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 50 pg/mL
Time to Analysis: 5 h
DEP significantly reduced the GSH:GSSG ratio. This
effect was prevented by adding thiol antioxidants
NAC or BUG. DEP increased lipid peroxide levels,
but the addition of all antioxidants decreased these
levels. DEP increased carbonyl groups. NAC, BUG,
and luteolin reduced HO-1 expression.
December 2009
D-76
-------
Study
Pollutant
Exposure
Effects
Reference: Wilson et CB: Carbon Black, Printex 90
al. (2007, 0972681 (Degussa)
Species: Mouse
Cell Type: J774
FeCI3
ZnCI2
Particle Size: CB: 14nm
Route: Cell Culture (4x1 Ob cells/ml at
1 ml/well)
Dose/Concentration: CB 1.9 -31 pg/mL;
FeCI3,ZnCI2 0.01-100 pmol
Time to Analysis: 4 h
ROS Production in Cells: CB alone increased
ROS. Coexposure with ZnCI2 did not affect ROS.
ROS Production - Cell Free: CB induced a
significant increase in ROS. ZnCI2 had no effect.
Coexposure CB/Zn also had no effect.
TNF-o Production (Fe -In 0.01-100 umol):
Coexposure of CB over a range of metals gave no
change over CB alone for Fe. For Zn, only at the
concentration of 100 pmol was there a small
interaction between Zn and CB.
Similar results were seen at metal concentrations
between 20 -100 pmol. Synergism was observed
between Zn and CB and no observed effect of Fe.
Macrophage Cytoskeleton: CB resulted in black
vacuoles. Co-treatment of cells with Zn and CB
increased the severity of Zn effects. Fe exhibited no
synergism.
Apoptosis /Necrosis: No synergism of CB with
either Fe or Zn.
Phagocytosis: Only at 31 pmol CB and 50 pmol Zn
did a synergistic effect occur; it resulted in a 4-fold
reduction.
Reference: Wottrich et Fe: hematite a-Fe203
al. (2004, 0945181
Si60: silicasol (Si02, amorphous silica)
Species: Human
SMOO: silicasol
Cell Type: A549, THP-
1, Mono Mac 6 Q: crystalline quartz DQ12
Particle Size: Fe: 50-90 nm; Si60: 60
nm; SMOO: 80-110 nm;Q<5|jm
Route: Cell Culture (2x104 cells/well. Co-
culture: 2x104A549 and 2x103
Macrophages)
Dose/Concentration: A549 light
microscopy hematite 100|jg/mL (23 pg/cm2)
TEM hematite 50 pg/mL (16 pg/cm2)
Cytotoxicity 10, 50,100 and 200 pg/mL (6.1,
30, 61 and 121 pg/cm2)
Cytokines 50 and 200 pg/mL
Time to Analysis: 24 h
Particle Uptake: Hematite agglomeration was
observed in all 3 cell lines. TEM confirmed cytosol
aggregates as well as single particles, which
includes particles transported intracellularly to
basolateral membrane of epithelial cells.
Cytotoxicity: LDH increased significantly in A549.
In decreasing order, Q, Fe, S60, and S100 (which
exhibited levels similar to controls) all induced
Cytotoxicity. THP-1 cells appeared the most sensitive
with Q, Fe, S60, S100, control inducing Cytotoxicity
in decreasing order. Mono Mac 6 cells were the least
sensitive with Fe, S60, Q, S100.
Cytokines: IL-6 and IL-8 released from A549 cells
upon exposure to all particles. No response was
observed in Mono Mac 6 or in THP-1 cells.
Co-cultures: Mix of A549 with either Mono Mac 6 or
THP-1 led to a large (ten fold) increase in response
to particles. Ten fold increases were observed in IL-6
and IL-8 levels with the Mono Mac 6 co-culture and
the THP-1 co-culture, respectively.
December 2009
D-77
-------
Study
Pollutant
Exposure
Effects
Reference: Wu et al.
(2007, 0984121
Species: Human
Cell Line: B82L
Cell Type: B82L- par
(parental fibroblasts),
B82L-wt (wild type
EGFR), B82L-K721M
(kinase defective
EGFR), B82L-c'958
(COOH-terminally
truncated EGFR at
Tyr-958)
ZnS04 (Sigma)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: Zn: 500 pmol
EGF:100ng/mL
Time to Analysis: 20 min
EGFR Mutations: EGFR-wt has a
functional tyrosine kinase domain, intact Src
phosphorylation (Tyr 845) and 5 tyrosine
autophosphorylation sites. EGFR-c'958 lacks all 5
tyrosine autophosphorylation sites. EGFR-
K721M lacks tyrosine kinase (ATP binding). EGFR-
Y845F lacks Src autophosphorylation (Tyr 845) and,
instead, has a receptor at Tyr 845 that is
phosphorylated by nonreceptor Tyrosine kinase Src.
Zn Induced Ras (MARK signaling protein): No
effect was observed in B82L-par cells. Zn had an
effect in -wt, -c'958, and -K721M which confirms the
need for EGFR. This indicates that neither tyrosine
kinase nor autophosphorylation sites were required
for Zn effects. No observed increase for Y845F
indicated that EGFR tyrosine 845 (phosphorylated
by c-Src) is required for Zn effects. However, it was
not required for EGF effects.
Src Kinase Requirement: Using a Src blocker
drastically reduced Zn effect but not the EGF effect.
Src activation occurred independent of EGFR Tyr-
845.
Zn Induced Association of EGFR with Src: Zn
induced a physical association in all 4 mutants; EGF
did not.
Zn Induced Phosphorylation of EGFR at Tyr-846:
Zn induced phosporylation of EGFR at Tyr-845 in
B82L-wt,-c'958 and -K721 M. EGF exhibited similar
effects. Src blockers significantly reduced
phosphorylation induced byZn but not for EGF.
Neither Zn or EGF induced phosphorylation in B82L-
Y845F cells.
Reference: Wu et al.
(2003, 1997491
Species: Human
Cell Type: BEAS-2B
Zinc Ion: Zn *
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 10, 25, 50 pmol
Time to Analysis: 0-8 h
Cytotoxicity: Exposure to 50 pmol Zn2* for 8 h did
not result in significant alterations in cell viability.
PTEN Protein Levels: 50 pmol Zn2* for 4 and 8 h
significantly decreased levels in a dose-dependent
manner. Exposure to 50 pM vanadyl sulfate
(tyrosine phosphatase inhibitor) had minimal effects
on PTEN. 100 ng/mL of non-specified EGF receptor
ligand for 1-8 h did not exhibit any significant effects
on PTEN levels.
P13K/Akt: Zinc induced Akt activation in a dose-
and time- dependent fashion. Active Akt levels were
the highest at 1 h post exposure to Zn2*,
corresponding with the time period when there was
a minimal effect on PTEN protein level. When
treated with LY294002 (inhibitor of P13K activity),
Akt phosphorylation was significantly inhibited.
PTEN mRNA Levels: Decreased PTEN mRNA
expression was observed in cells exposed to 50
pmol Zn2* for 8 h whereas PTEN protein levels
declined as early as 4 h.
Proteasome-mediated PTEN Degradation: Use of
MG132 (proteasome inhibitor) had no significant
effect on Zn2* induced PTEN mRNA expression.
Therefore mRNA expression may not play a critical
role in PTEN protein reduction. Instead data
suggested that 26 S proteasome played a vital role
in Zn2* induced PTEN degradation. PI3K inhibitor
blocked Zn-induced PTEN degradation, but failed to
prevent significant Zn-induced down-regulation of
PTEN mRNA.
December 2009
D-78
-------
Study
Pollutant
Exposure
Effects
Reference: Wu et al.
(2004, 0969491
Species: Human
Cell Type: NHBE
Zinc Ion: Zn *
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 100 pmol
Time to Analysis: 2 h
Cell Viability: After 2 h of exposure, Zn * induced
effects in NHBE cells at 100 and 200 pmol levels
(but not 50 pmol). Continuing exposure to 100 pmol
Zn * for 4 and 6 h did not significantly alter cell
viability. Thus, in all subsequent studies, NHBE cells
were treated with 100 pmol Zn2*.
Induced EGFR Phosphorylation: Exposure to
100pM Zn2* for 1-4 h induced phosphorylation of
EGFR in NHBE cells. EGFR kinase inhibitor
PD153035 (to determine if phosphorylation of EGFR
was the result of autophosphorylation of activated
EGFR tyrosine kinase activity) caused Zn * -induced
phosphorylation to subside. Zn2* activity requires
tyrosine kinase activity.
EGFR Phosphorylation Pathway: To test whether
Zn * exposure results in ligand release, which in turn
can activate phosphorylation, NHBE cells were
pretreated with LA1 blocking antibody Results
showed significant suppression of Zn * induced
phosphorylation, therefore Zn2* phosphorylation
might be initiated by the release of EGFR ligands.
HB-EGF, TGF-a, EOF: To examine the involvement
of specific ligands (HB-EGF. TGF-a and EGF) in the
phosphorylation pathway, cells were exposed to
anti-HB-EGF, anti-TGF-a and anti-EGF. Results
showed that anti-HB-EGF reduced Zn2* induced
phosphorylation significantly, anti-TGF-a produced
partial inhibition and anti-EGF had no inhibitory
effect. Exposure with blocking antibody LA1 was
tested to determine if it caused an increase in
soluble HB-EGF. HB-EGF mRNA expression was
also elevated in cells exposed to Zn . Previous
studies indicate metalloproteinase (MMP)
involvement in cleaving ligand precursors. It was
found that MMP-3 inhibitor partially blocks Zn2*
induced HB-EGF release. (MMP-2 and MMP-9 did
not show similar inhibition patterns) Zn2* exposure
increased the release of MMP-3 from HNBE cells.
December 2009
D-79
-------
Study
Pollutant
Exposure
Effects
Reference: Wu et al.
(2005, 0973501
Species: Human
Cell Line: Subclone
S6
Cell Type: BEAS-2B
Zinc Ion: Zn *
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 50 pmol
Time to Analysis: 4 or 8 h; EGFR
phosphorylation: 30, 60,120, 240 min
Cell Viability: Exposure to 50 pmol Zn * for 8 h did
not result in significant alterations in cell viability
(assessed by LDH release).
P13K/AM Signaling Pathway: To evaluate P13K's
on COX-2 Zn induced expression, LY-294002 (a
P13 inhibitor) and another unnamed P13 inhibitor
were used. Exposed cells indicated suppressed
levels of Zn2* induced COX-2. To determine Akt role,
ad-DN-Akt (AAA) was used. Infected cells indicated
over-expression of Akt and significant reduction of
Zn2* induced GSK-3o/|3 phosphorylation. Over
expression of DN-Akt(AAA) blocked Zn2* induced
COX-2 expression.
PTEN's Role in Blocking Zn2+ Induced COX-2
mRNA Expression: PTEN is an antagonist of
P13/AM pathway. Overexpression of wildtype PTEN
blocked Zn2*-induced mRNA COX-2 expression,
suggesting PTEN inhibits PIPS signal transduction to
Akt.
Analysis of the Src/EGFR Signaling Pathway:
Zn2* induced a time-dependent increase in Src and
EGFR phosphorylation in cells. Blockage of Src
activity via PP2 (Src inhibitor) decreased Zn2*
induced EGFR phosphorylation. The EGFR tyrosine
inhibitor completely blocked Zn2*-induced EGFR
phosphorylation. EGF (a ligand of EGFR signaling)
induced COX-2 expression, suggesting that EGFR
regulated Zn2* -induced COX-2 expression.
p-38 and EGFR Kinase Activity: Use of PD-
153035 (EGFR inhibitor) and PP2 (Src inhibitor) and
SB-203580 (p38 inhibitor) all blocked Zn2*-induced
Akt phosphorylation of Src., EGFR and p38. It is
thought that p38 is a critical kinase in regulation of
Zn *-induced COX-2 protein expression.
Reference: Yacobi et
al. (2007, 1561661
Species: Rat
Cell Type: L2 (Lung
epithelial cells)
PNP: Polystyrene nanoparticles,
negatively charged (Molecular Probes,
Eugene, OR)
PNPA:Amidine modified PNP,
positively charged
SWCNT: Single-wall carbon nanotubes
(Carbon Nanotech, Houston, TX)
QDC: Chitosan coated (CdSe/ZnS)
Quantum dots, positively charged
(made)
QDA: Alginate coated QD, negatively
charged
UAPS: Ultrafme Ambient particulate
suspensions (VACES) (48 % OC)
Particle Size: PNP20: 20 nm;
PNP100:100 pm; SWCNT: 0.8-1.2 nm
(diameter); SWCNT: 100-1000 nm;
QD:30nm;UAPS:<150nm
Route: Cell Culture (1.2x106 cells/cm2)
Dose/Concentration: PNP up to 706 pg/mL
QDupto176|jg/mL
SWCNT up to 88 pg/ ml
UAPS up to 36 pg/mL
Time to Analysis: on days 4, 5 or 6 by
replacing monolayer apical fluid with PM in
suspension for up to 1440 min.
Intermediate measurements at 15, 30, 60,
120, 240 and 1440 min.
UAPS and Rt (transmonolayer resistance): Rt
declined up to 60% within 1 h at 36 pg/mL Rt
plateaued (or exhibited a very slight upgradient) for
up to 24 h (last measurement). No cytotoxicity was
observed. Replacement of apical fluid with fresh
media after 2 h of exposure restored Rt to near
control values within 24 h.
UAPS and Leq (short-circuit current): Peak
decline of 30% after 4 h followed by gradual
recovery over 24 h. Replacing media after 2 h
exposure returned leq to control values within 24 h.
UAPS and Apparent Permeability: Permeability
measured via C14 mannitol and inulin showed no
effect of UAPS.
QD and Rt: QD depressed Rt by nearly 55% at 4 h
for positively charged and 30% for negatively
charged QDs. Recovery towards control values
started at 4 h and was near complete at 24 h
SWCNT and Rt: SWCNT depressed Rt by ~ 40% at
1 h (same for 22, 44, and 88 pg/mL). Recovery was
near complete at 4 h and complete at 24 h.
PNP and Rt: No statistically significant effects were
observed.
December 2009
D-80
-------
Study
Pollutant
Exposure
Effects
Reference: Yun et al. DEP: Collected using a 6 cyl 11L, Route: Cell Culture (3x104 cells/well)
(2005, 0883021 heavy duty (2001 yr) bus engine (South
v Korea) Dose/Concentration: 1,10,100,250,500
Species: Human and 1000 pg/mL; main testing 250 pg/mL
Particle Size: NR
Cell Type: A549 Time to Analysis: 12 h
NF-KB Transcription Activation: DEP induced
dose-dependent activity up to 250 pg/mL After
peaking at 250 pg/mL, concentrations above 250
induced dose- dependent declines. Activity peaked
at 12 h for 250 pg/mL and declined to control at 24
or 48 h. The mechanism of DEP action was the
degradation of iKBa which is an intracellular inhibitor
of nuclear translocation of NF-KB.
TAK1 and NIK Required for NF-KB Activation by
DEP: Dominant negative mutants of TAK1 and NIK
reduced DEP induced response to basal level. TAK1
was phosphorylated after DEP exposure and was
sustained for at least 90 min.
Reference: Zhang et
al. (2007, 1561791
PM25: Collected by baghouse from
Dusseldorf, Germany
Species: Human, Rat Particle Characterization: Carbon 20%,
~,,T Ac.nn, ^ Hydrogen 1.4%, Nitrogen <0.5%,
Cell Type: A549, RLE- oxygen 14.1%, Sulfur 2.1%, Ash
6TN 63*2%.
Particle Size: PM25
Route: Cell Culture
Dose/Concentration: 100 |jg/cm;
Time to Analysis: 24 h
Apoptosis: At 100 pg/mL for 24 h, PM induced a
2.5 fold increase in apoptosis in A549.
Mitochondria! Membrane Potential: A significant
reduction in AEC mitochondrial membrane potential
was observed.
Caspase -3 & -9: Increased activity of both
enzymes in both cell types was observed. More
specifically, a 2- to 2.5-fold increase of caspase -3
and -9 in A549 and an 8-fold increase of caspase-9
and 4-fold increase of caspase-3 in RLE-6TN were
observed.
BIM: Downregulation of BIM by RNA interference
inhibited PM-induced apoptosis. An inhibited
decrease in mitochondrial membrane potential and
activation of both caspases were observed.
Reference: Zhang et
al. (2004, 1571831
Species: Mice
Cell Line/Type: C10
(alveolar Type I Hike
epithelial cell line)
DEP: SRM 1650a
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 5 or 25 pg/mL
Time to Analysis: 30-360 min
fra Expression: DEP induces fra-1 but not fra-2
expression. mRNA induction peaks around 180 min
DEP affects fra-1 mRNA expression at the
transcriptional level.
ERK/JNK/p38 MARK signaling pathways: 3
inhibitors (PD-98059, SB-202190 or SP-600125) all
reduced DEP stimulated fra-1 induction to near
control levels. DEP stimulated phosphorylation of
the MAPKs which peaks at 60 min but stays
elevated at 180 min.
MMP-9 promoter activity: fra-1 upregulation may
play a role in DEP induced increases in MMP-9
promoter activity as fra-1 appears to bind at the -79
TRE sequence of the MMP-9 promoter.
Table D-3. Respiratory effects: in vivo studies.
Reference
Reference:
Adamson et al.
(2003, 0879431
Species: Rat
uGndGT! M3l6
Strain: SD
Weight: 150g
Pollutant
PM,0: EHC-93W (whole dust)
EHC-93S (soluble)
EHC-93L (leached)
EHC-2KW, -S, -L
Measured components Zn, Mg,
Pb, Fe, Cu, Al
Particle Size: EHC-93VV -93S, -
93L, -2KW, -2KS, -2KL: PM,0
Exposure
Route: IT Instillation
Dose/Concentration: 5 mg/rat; 33.3 mg/kg
Time to Analysis: 4 h, 1 day, 3 day, 7 days, 14
days
Effects
BALF Cells: The greatest increase in cell numbers
was observed with EHC-93W Activity peaked at 1 day
with a return to normal levels by 7 days. EHC-93L also
induced an increase in cell numbers, more so than
EHC-93S, but both particles induced statistically
significant increases. However, these increases were
mostly attributable to an increase in the AM and PMN
populations.
BALF Inflammatory/Injury Markers: Metallo-
proteinase (MMP) 2 and 9 both increased, peaking at 1
day and 4 h respectively. MMP2 activity appears
related to the soluble fraction whereas MMP-9 activity
appears to be related to the leachable fraction.
December 2009
D-81
-------
Reference
Pollutant
Exposure
Effects
Reference: Ahn et
al. (2008, 1561991
Species: Mouse
Gender: Male
Strain: BALB/C1
Age: 6 wk
Weight: 19-24 g
DEP: Collected using a turbo- Route: Oropharyngeal Aspiration
charged, intercooler, 6-cylinder,
heavy-duty, diesel engine (model Dose/Concentration: 1,10, 25 mg/kg per day;
— Those receiving 25 mg/kg DEP also received pre-
treatment of Dex (1, 5 mg/kg) 1 h prior
year 2000
DPBS: control
Particle Size: NR
Time to Analysis: 5 consecutive days; 72 h post
final exposure
BALF Inflammatory/Injury Markers: Lung injury was
more severe in mice exposed to 25 mg/kg of DEP than
when compared to mice exposed to 1 mg/kg DEP.
However, lung injury caused by exposure to 25 mg/kg
DEP could be completely prevented with pre-treatment
of 5mg/kg Dex. Treatment with 1 mg/kg Dex prior to
exposure to 25 mg/kg DEP depicted partial reduction in
lung injury.
BALF Cells: Treatment with DEP over a 5 day period
caused an increase in total number of cells
(macrophages, neutrophilsand lymphocytes) when
compared to control.
Total Cells: Control - 5.33 ± 0.44 cells
1 mg/kg DEP-6.26 ±0.87 cells
10 mg/kg DEP-14.40+ 1.90 cells
25 mg/kg DEP - 47.20 ± 3.40 cells
COX-2 Expression: Exposure to DEP lead to a dose-
dependent increase in COX-2 levels; specifically,
treatment with 25 mg/kg significantly increased COX-2
levels. This effect was completely reduced by treatment
with 5mg/kg of Dex.
Reference: Ahsan DEP: Obtained from Dr. Masaru Route: IT Instillation
et al. (2005,1562001 Sagai (Amori, Japan)
Particle Size: NR
Species: Mouse
Gender: Male and
Female
Strains: hTrx-1-
transgenic and
C57BL/6 (control)
Age: 8-8.5 wk
Dose/Concentration: Lung Damage: 0.1
mg/mouse; Survival Analysis: 0.2 mg/mouse; ESR:
0.05 mg/mouse
Time to Analysis: 24 h
ESR: hTrx-1 induced 0.05 mg generation of hydroxyl
radicals in the lungs (mid thorax ESR spectra)
compared to control.
BAL Inflammatory/Injury Markers: hTrx-1 attenuated
cellular damage from 0.1 mg DEP. Control mice showed
massive edema with neutrophilic infiltration,
hemorrhagic alveolar damage and collapsed air
spaces. hTrx-1 mice showed mild/moderate edema
with clear demarcation of air spaces.
Viability: After 4,12 and 24 h, survival was 32, 24 and
12% respectively as compared to 80, 52 and 40% for
hTrx-1 mice.
Reference: Andre et UFCP: Ultra Fine Carbon
al. (2006, 0913761 Particles (electric spark
generator, Model GFG 1000;
Species: Mouse pa|aS] Karlsruhe, Germany)
Gender: Female
Strain: BALB/cJ
Age: 10-12 wk
Measured Component:
UFCP>96% EC
Particle Size: 49 nm
Route: Whole-body Inhalation
Dose/Concentration: 380 |jg/m;
Time to Analysis: 4 and 24 h; 0 and 24 h post-
exposure
BALF Cells: A small increase in PMN number
suggests a minor inflammatory response after 24 h
exposure. Number of macrophages did not increase.
BAL Inflammatory/Injury Markers: Total protein
concentration significantly increased post 24 h
inhalation. Post 4 h, heat shock proteins were induced.
Post 24 h, immunomodulatory proteins (osteopontin,
galectin-3 and lipocalin-2) significantly increased in
alveolar macrophages and septal cells. 236 (1.9%)
genes was increased and 307 (2.5%) genes were
decreased with upregulated genes being primarily
related to the inflammatory process.
Reference:
Antonini etal.
(2004, 0971991
Species: Rats
Gender: Male
Strain: SD
Weight: -250 g
ROFA-P: Precipitator
-S: Soluble (0.22 pm filter),
Components: Fe, Al, Ni, Ca, Me
Zn
-I: insoluble, Components: Fe,
Al, Ni, Ca, Mg, Zn, V
-T: total
ROFA-AH: Air Heater
-S: Soluble (0.22 pm filter),
Components: Fe, V, Ni, AL
-I: Insoluble, Components: Fe, V,
Ni.AL
-T: Total
Particle Size: < 3 pm (mean
diameter)
Route: IT Instillation
Dose/Concentration: 1mg/100g bw in 300 pi
saline; 60 mg/kg
Time to Analysis: 24 h; Clearance Experiment:
two single exposures day 0 and 3 observed at day
6, 8 and 10
ESR: Only ROFA-P contained free radicals, primarily in
ROFA-P-S.
BALF Cells: No effects on alveolar macrophages were
observed, but all ROFA-P fractions increased lung
neutrophils. ROFA-P-S and ROFA-P-I effects combined
roughly equaled ROFA-P-T.
BAL Inflammatory/Injury Markers: ROFA-AH-T and
ROFA-AH-I increased LDH. ROFA-P and -AH
increased albumin for T and I fractions.
Pulmonary Clearance (Listeria Monocytogenes):
ROFA-P-T and ROFA-P-S significantly slowed bacteria
clearance from lungs. ROFA-AH and ROFA-P-I had no
effect.
December 2009
D-82
-------
Reference
Pollutant
Exposure
Effects
Reference: Arimoto
et al. (2007, 0979731
Species: Mouse
Strain: ICR
Gender: Male
Age: 6 wk
Weight: 29-33 g
Reference:
Bachoual et al.
(2007, 1556671
Species: Mouse
Strain: C5B1 7
Gender: Male
Age: 7 wk
Weight: 22.3 ±
073 g
DEP (collected using a 4JB1 4-
cyl, 2.74L Isuzu diesel engine)
DEP-OC: organic chemical
extracts
IPS
DL=DEP + LPS
DDL = DEP-OC + IPS
Particle Size: 0.4pm
PER: PM10
Paris, France subway
CB
Ti02
DEP
Particle Size: PER: 79% < 0.5
pm; 20%: 0.5-1 pm
PR' qc nm
WD. y\j MI M
Tif~l ' 1 ^D i im
1 1^2' ' ^^ M
DEP: NR
Route: IT Instillation
Dose/Concentration: DEP or DEP-OC: 4 mg/kg;
LPS:2.5mg/kg;DLorDOL:NR
Time to Analysis: 24 h
Route: IT Instillation
Dose/Concentration: 5, 50, 100 pg/mouse, 0.22,
2.2, 4.5 mg/kg
Time to Analysis: 8 or 24 h
Cytokines: DEP-OC or DEP alone did not change
levels of MIP-1a, MCP-1 orMIP-2. DL induced
significant increases in MIP-1, MIP-2 and MCP-1.
IPS: IPS and DDL induced increases in MCP-1
though the increase induced by DLwas greater. No
effect on MIP-1a or MIP-2 was observed.
BALF Cells: 100 pg RER and 100 pg DEP increased
total cell count and neutrophil influx after 8 h and
returned to normal by 24 h. Smaller doses of RER and
DEP induced no effect. CB induced no effect.
BAL Inflammatory/Injury Markers: 100 pg RER
increased BALF protein after 8 h. No effect was
observed after 24 h nor with smaller doses of PM. RER
significantly increased MMP-12 mRNA level after 8 h
and HO-1 total lung mRNA content. No effects on
MMP-2 or -9 or TIMP-1 or -2 expression were
observed. No effects from CB or DEP were observed.
f^wtnLinec- 100 I in PPP inrroacoH RAI TMF.rf anH
MIP-2 protein content after 8 h.
Reference: Batalha
et al. (2002, 0881091
Species: Rat
Gender: Male
Strain: SD
Age: NR
Weight: 200-250 g
CAPs (Harvard Ambient Particle
Concentrator)
Particle Size: Mean: 2.7pm
Route: Whole-body Inhalation
Dose/Concentration: Range: 73.5-733 pg/m3
Time to Analysis: CAPs exposure 5 h/day, 3 days
(consecutive). S02 exposure to induce CB 5 h/day,
5 days/wk, 6 wk. Killed 24 h postexposure.
Histopathology: CAPs slightly increased the wall
thickness of small pulmonary arteries and edema in the
adventitia and hyperplasia of the terminal bronchiole
and alveolar ducts epithelium.
L/W ratio: The L/W ratio decreased in CAPs-exposed
rats as particle mass, Si, Pb, S042-, EC and OC
increased. Univariate analyses showed significant
negative correlations between the L/W ratio and Si and
S042- in normal rats and Si and OC in CB rats.
Multivariate analysis showed only Si to be significant in
both groups.
Reference: Becher
et al. (2007, 0971251
Species: Mouse
Strain: Crl/Wky
(iNOS(-/-)) and
C57BI/6
Gender: Male
Age: 8-14 wk
Weight: 25 g
Suspended PM:SRM-1648
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 1.6 pg/lung; 64 mg/kg
Time to Analysis: 20 h
Cytokines: In both wild and KO strains, all particles
caused increases of IL-6, MIP-2 and TNF-a levels.
NADPH-oxidase KO mice showed significantly lower
levels of IL-6 and MIP-2 responses to SPM compara-
tively to wildtype. iNOS KO mice showed significantly
reduced IL-6, TNF-a, MIP-2 responses to SPM
comparatively to wildtype.
Free Radicals: SPM induced significant increases in
free radical formation in alveolar type 2 cells but could
be inhibited by DPI.
Reference:
Bhattacharyya et al.
(2004, 0880951
Species: Mouse
Strain: SD
Weight: 200-250 g
Douglas Fir Wood Smoke
(generated by burning wood at
400°C in crucible oven)
Particle Size: NR
Route: Nose-only Inhalation
Dose/Concentration: 25 g/mouse
Time to Analysis: Various exposure periods (0, 5,
1 0,15, 20 min). Parameters measured after 24 h
recovery period.
Biochemical Parameters: Lipid peroxidation
increased after 20 min of wood smoke inhalation as did
Myeloperoxidase at 20 min. No effects were observed
at other times or for total antioxidant status, reduced or
oxidized glutathione.
Antioxidant Enzyme Activities: No effect was
observed.
Histology: Dose-dependent damage progressing from
loss of cilia (5 min), degeneration of mucosal
epithelium, loss of mucosal epithelium to disrupted
mucosal epithelium with submucosal edema and
inflammation. Changes persisted for up to 4 days.
December 2009
D-83
-------
Reference
Pollutant
Exposure
Effects
Reference: Cao et
al. (2007, 0974911
Species: Rat
Strain: SH and
WKY
Age: 12 wk
PM25 (Shanghai, China)
Route: IT Instillation
Components: As, Cd, Cr, Cu, Fe, Dose/Concentration: 1.6, 8.0 and 40 mg/kg
Ni, Pb, Zn, V, Ba, Se, Mg, Co,
Mn Time to Analysis: Exposed 1/day for 3 days,
sacrificed 24 h following last exposure
Particle Size: PM2 5
BALF Cells: PM decreased macrophages and
increased neutrophils and lymphocytes in a dose-
dependent manner. For the same exposed dose, WKY
rats had a higher percentage than SH but a smaller
percentage of neutrophils and lymphocytes.
BAL Inflammatory/Injury Markers: LDH activity and
TBARs increased a in dose-dependent manner.
Notably, activity in SH rats was much higher than WKY
at the same dose exposed for each dose level.
Cytokines: PM induced pro-inflammatory cytokine
release (IL-lp, TNF-a, CD44, MIP-2, TLR-4, OPN).
Again, SH cytokine level was greater than WKY at all
dose levels. PM induced anti-inflammatory cytokines
CC16 and HO-1 in a similar manner but at much lower
rate.
Reference: Carter
et al. (2006, 0959361
Species: Rat,
Mouse, Hamster
Gender: Female
(all)
Strain: F-344 (rat),
B6C3F1 (mouse),
Syrian Golden
(hamster)
Age:7-10wk
CB: Printex 90
Particle Size: primary size: 17
nm; 1.2-1.6pm (aerosol
aerodynamic diameter)
Route: Whole-body Inhalation
Dose/Concentration: 1, 7, 50 mg/m3
Time to Analysis: 6 h/day for 5 days/wk for 13 wk;
1 day, 3 m, 11 m post-exposure
Superoxide: Levels rose in all species at 50 mg dose.
Hamsters had no increase at 7 and 1 mg doses. Mice
also increased at 7 mg. Rats significantly increased at
all dose levels. Rats maintained elevation except for
the 50 mg dose at 11 mo postexposure; it declined but
was still higher than control. Mice maintained elevation
at 50 mg while 7 mg returned to control levels by 3 mo
postexposure.
HJQf. At 50 mg, increased levels in all species, with the
highest in rat, were observed. At 7 mg, increased levels
in rats and mice were initially seen but levels returned
to baseline by 11 mo. Hamster levels were not
significant. At 1 mg, no significant changes were
observed.
NO: Induced similar reactions as H202. Rat response
continued through the study while mice and hamsters
returned to baseline by 11 mo postexposure. Rats
produced significantly higher levels at all times than
other species.
BALF Cells: CB induced significant increases in
neutrophils at 7 and 50 mg for all species. Rats had the
highest and most prolonged PMN response. Mice and
hamsters had very similar reactions.
Cytokines: TNF-a, MIP-2 and IL-10 increased in a
dose-dependent manner in rats and mice. Hamsters
increased for IL-10 only. MIP-2 levels were highest in
rats. TNF-a level were similar in all three species at 50
mg, but hamsters started with a markedly higher basal
level.
Glutathione Peroxidase: Hamsters were the most
responsive with significant increases at all levels. Rats
and mice increased at 50mg and continued to increase
for up to 11 mo. Hamster levels declined with time but
continued to be higher than control.
Glutathione Reductase: Rats increased only at 50mg
and remained elevated for up to 11 mo. Mice increased
at 7 and 50mg and remained elevated for up to 11 mo.
Hamsters increased at all levels at 11 mo, but at 50mg,
levels only increased post 1 day.
Superoxide Dismutase: All species reacted in a dose-
dependent manner. Rats were the least responsive.
Rat SOD activity increased over time while rat and
mouse activity decreased at 50mg. Data were
consistent with cytokine data.
Summary: Rats appear to produce proinflammatory
responses while mice and hamsters produce
antiinflammatory responses.
December 2009
D-84
-------
Reference
Pollutant
Exposure
Effects
Reference: Cassee
et al. (2005, 0879621
Species: Rat
Gender: Male
Strain: Wistar Kyoto
and SH/NHsd
e:7wkand8-12
wk
CAPs: PM25
Netherland suburban, industrial
and freeway tunnel site
collections
Wistar rats pre-exposed to 03
S04, N03 and NH4 ions: 54 ± 4%
suburban, 53 + 7% industrial and
35 + 5% freeway site cone, of
total CAPS mass
Particle Size: PM2 5
(0.15
-------
Reference
Pollutant
Exposure
Effects
Reference: Cho et
al. (2005, 1563441
Species: Mouse
Gender: Male
Strains: DBA/2J,
129P3/J, C57BL/6J,
BALB/cJ, A/J,
C3H/HeJ,
C3H/HeOuJ
Age: 6-8 wk
ROFA: Obtained from Power unit Route: IT Instillation BALF Cells: Significant genetic effects on number of
4, Boston, MA macrophages and PMNs after ROFA challenge. For
Dose/Concentration: 6 mg/kg bw (150 pg in 50pl/ PMNS] DBA/2j, C57BL/6J, BALB/cJ, and 129P3/J all
Absent of LPS 25 g) induced increases significantly higher than C3H/HeJ.
Particle Size: NR
o
mice, single. 1.5, 3 and 6 h (compare TLR-
mediated molecular events)
^ pM d macrphages ay increased with
^0uJ< .^ {mxeas^^nm^ different from
HeJ.
BAL Inflammatory/Injury Markers: Significant genetic
effect on mean total protein concentration was
observed. In decreasing order, DBA/2J, 129P3/J and
C57BL/6J all induced increases significantly higher
than C3H/HeJ.
TLR4 mRNA Expression: A significant decrease was
observed in TLR4 transcript level in HeJ- ROFA
exposed mice post 1 .5 h. Post 6 h, TLR4 levels were
greater than the control levels. OuJ expression
increased beginning 1.5 h post exposure.
TLR4 Protein Level: Protein level of OuJ mice
significantly exceeded (~2-3 fold) HeJ mice at 1.5, 3
and 6 h.
Activation of Downstream Signal Molecules:
Greater activation of MYD88, TRAF6, IRAK-1, NF-KB,
MAPK, and AP-1 was observed in OuJ mice than in
HeJ mice before the development of ROFA- induced
pulmonary injury.
Cytokines: IL-lp, LT-p, IL-1a, IL-7, IL-13, IL-16
increased in both strains (OuJ and HeJ). Levels of all
cytokines above were significantly higher in OuJ than in
HeJ.
Reference: Churg et
al. (2003, 0878991
Species: Human
Gender: Female
(Mexico City); Male,
Female (Vancouver)
Age: 66 + 9yr
(Mexico City); 76 +
11yr (Vancouver)
Weight: NR
PM (Mexico City-high PM
region, Vancouver- low PM
region)
Particle Size: Geometric mean
size of individual particles in
tissue: 0.040-0.067 pm;
Aggregates in tissue: 0.34-0.54
pm; Mexico City: 2.5,10 pm
Route: Ambient Air Exposure. Autopsy Tissue.
Dose/Concentration: 10 - >1000x106 g dry
tissue; Mexico City: PM10: 66 pg/m3, Vancouver:
PMi0:25|jg, PM25:15 pg
Time to Analysis: Lung samples taken from
deceased lifelong Mexico City residents and
Vancouver residents >20 yr. Subjects were never-
smokers, did not work in dust occupations or cook
with biomass fuels.
The lungs from Mexico City residents showed
increased muscle and fibrous tissue in the
membranous bronchioles and respiratory bronchioles
compared to the Vancouver residents. Pigmented dust,
lumental distortion and carbonaceous aggregates of
UFPs were present in the Mexico City lungs.
Reference: Costa et
al. (2006, 0884381
Species: Rat
Gender: Male
Strain: SD
Age: 60 day
ROFA
FPSLplant#6oil, 1% sulfur
Particle Size: ~1.95pm
Route: IT Instillation, Nose-only Inhalation (IH)
Dose/Concentration: IT instillation = 110 pg/rat
IH = 12mg/m3
Time to Analysis: IT instillation: single; IH:6h
24, 48, 96 h (histopathology 24 and 48 only)
ROFA distribution: IH and IT instillation resulted in
equivocal distribution (pg/g lung tissue) in 5 different
lung lobes.
Airway Hyperactivity: IT instillation resulted in
doubled airway hyperreactivity at 24 h which was
sustained for 96 h. IH hyperreactivity did not reach
statistically significant level.
BALF Cells: Neutrophils peaked at 24 h and slowly
declined at 48 and 96 h.
BAL Inflammatory/Injury Markers: IH and IT
instillation showed very similar responses (R2 = 0.98).
Time-dependent increases were observed for protein
and LDH.
Lung Pathology: IT instillation showed more alveolitis,
bronchial inflammatory and fibrinous fluid infiltrate. IH
showed relatively more congestion of small airways
and alveolar hemorrhage.
December 2009
D-86
-------
Reference
Reference: Courtois
et al. (2008, 1563691
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Age: 12-14 wk
\A/AiMkti MD
weight: NK
Reference: Dick et
al. (2003, 0366051
Species: Mouse
Gender: Female
Strain: CD1
Age:8-10wk
Weight: 20-25 g
Pollutant
PM(SRM1648;63%inOC, 4-
7% OC, >1% mass fraction- Si,
S, Al, Fe, K, Na)
Carbon black (FVV P60)
UF, fine Ti02
Particle Size: PM mean
diameter: 0.4 pm; Carbon black:
FW-13nm, P60-21 nm;Ti02
mean diameter: 0.1 4pm
CO: PM Coarse
Fl: PM Fine
FU: PM ultrafine
PM collected in RTP, NC
Particle Size: CO: 3.5-20 pm;
FI:1.7-3.5|jm;FU:<17|jm
Exposure
Route: IT Instillation
Dose/Concentration: 5 mg PM or Ti02
Time to Analysis: 6-72 h
Route: IT Instillation
Dose/Concentration: 10 |jg 50 |jg 100
pg/mouse; 0.5, 2.5, 5.0 mg/kg
Time to Analysis: DMTU 500 mg/kg bw 30 min
pre-exposure for some mice. Parameters
measured 18 h post-exposure.
Effects
Particles were present in lung parenchyma that was
removed 12 and 72 h post-instillation.
Particle Characteristics: S increased (00-33.20
pg/mg, Fl- 49.44 pg/mg FU- 122.79 pg/mg) with
decreasing particle size (mostly in the water-soluble
fraction). Fe and Cu higher in coarse and fine fractions
(mostly present in the insoluble). CO PM contained
more nickel (in both soluble and insoluble) than Fl or
FU particles. Also, endotoxin levels similar in CO and
Fl; much lower in FU (0.165 EU/mg).
BALF Cells: PMN increased with exposure for all 3
fractions ex cept 1 00 pgFI.
BAL Inflammatory/Injury Markers: Albumin increased
only at 100 pg Fl. No differences in NAG or LDH
observed.
Cytokines: IL-6 increased at 100 pg dose for all 3
fractions with similar responses. TNF-a increased a
100 pg dose of fine PM vs control.
Effect of PM After Pre-treatment w/DMTU: Systemic
administration of DMTU alone depicted a two-fold
increase in total antioxidant capacity.
DMTU halved neutrophil response observed with
PMs alone: No fractions were increased over DMTU
alone which was at least two-fold saline control. IL-6
concentrations were drastically reduced in the DMTU
group for the mice exposed to coarse particles (all
fractions were reduced but only coarse had a
significant response). TNF-a levels were decreased
after treatment with particles and DMTU but treatment
with particles and saline (control) produced similar
results.
Reference: Dybdahl DEP: SRM 1650 (NIST)
et al. (2004, 0890131
Particle Size: DEP: NR; Control:
PM 0.13pm diameter
Species: Mouse
Gender: Female
Strain: BALB/CJ or
trans-genie
(MutaMouse)
Age:9-10wk
Weight: -20 g
Route: Nose-only Inhalation
Dose/Concentration: I: 20, 80 mg/m3
II: 5, 20mg/m3
Time to Analysis: I: single exposure 90 min; II: 90
min/day for 4 days; I S II: parameters measured 1,
3, or 22 h post exposure
Cytokines: A single 90 min DEP exposure increased
IL-6 gene level dose-dependently in the lung. For 80
mg/m3 DEP, significantly higher IL-6 gene level was
observed, both 1 and 22 h post exposure. For 20
mg/m DEP, a significantly higher IL-6 level was
observed at 1 h post exposure but normalized at 3 h.
BALF Cells: Inhalation of DEP did not decrease
viability of BALF cells. For mice exposed to 20 mg/m3
DEP, at 1 h post exposure in BAL fluid there was 3 fold
increase in total cell number.
DMA Damage: Level of 8-oxodG increased post single
exposure with 80 mg/m3 inducing levels significantly
higher than controls. Repeated exposures were
associated with significantly higher DNA strand breaks.
Reference: Elder et UFP: argon-filled chamber with
al. (2004, 0556421 electric arc discharge (TSI, Inc.,
St. Paul, MN)
Species: Rat
Gender: Male
Strain: F344, SH
Age: 23 m (Fisher),
11-14 m(SH)
Particle Size: 36 nm
Route: Whole-body Inhalation.
Dose/Concentration: UFP: 150 pg/m3 bw; LPS: 2
mg/kg
Time to Analysis: 6 h, 18 h
BALF Cells: Neither inhaled UFP nor LPS cause a
significant increase in BALF total cells or percentage
of neutrophils in either rat strain. No significant
exposure-related alteration in total protein
concentration was observed. In both rat strains LPS
induced a significant increase in the amount of
circulating PMNs. When combined with inhaled UFP,
PMNs decreased; for F-344 rats, this decrease was
significant.
ROS in BALF: In F-344 rats, both UFP and LPS have
independent and significant effects on DCFD oxidation.
Effects were in opposite directions; particles decreased
ROS whereas LPS increased ROS.
December 2009
D-87
-------
Reference
Pollutant
Exposure
Effects
Reference: Elder et
al. (2004, 0873541
Species: Rat
Gender: Male
Strain: F344
Age: 21 mo
Freshly generated vehicle
exhaust emissions from I-90
between Rochester and Buffalo,
NY
Particle Size: NR
Route: Whole-body Inhalation; IT Instillation
(Influenza)
Dose/Concentration: Vehicle exhaust: 0.95-
3.13x105 particles/cm3
Endotoxin: 84 EU
Influenza (IV): 10, 000 EID 50 in 250 \i\
Time to Analysis: 1 *6 h, 3^6 h or both.
Parameters measured 18 h post-exposure. 48 h
prior toon-road exposures, instilled intratracheally
with IV Immediate pre-exposure of priming agent
endotoxin.
EXPERIMENTS
1:LPS+PM6h
2:LPS+PM6h, 3x6h
3:IV+PM6h
4:IV+PM6h, 3x6h
No departures from normal baseline cellular or
biochemical values were observed, suggesting that on-
road exposures were well tolerated by the rats.
BAL Inflammatory/Injury Markers: Increase in total
protein concentration, LDH and B-glucuronidase
activities were observed.
Specific results according to groups 1-4 are as
follows:
Experiment 1: No endpoints revealed significant
differences between groups of rats exposed to gas
phase only versus the gas-phase/particle mixture.
Experiment 2: Combination of endotoxin and particles
produced greater inflammatory responses than those
treated with saline and particles post 1 day. After 3
days, no statistically significant changes were noted.
Experiment 3: Influenza virus significantly increased
ROS release in BALF cells.
Experiment 4: Influenza virus significantly increased
both percentage of PMNs in BALF and BALF cell ROS
release.
Reference: Elder et
al. (2005, 0881941
Species: Rat,
Mouse, Syrian
Golden Hamster
Gender: Female
Strain: F-344,
B6C3F1FIB
HSCb: Printex-90 high surface
area carbon black, Deguss-
Huels (Trostberg, Germany).
LSCb: Sterling V, low surface
area carbon black, Cabot
(Boston, MA)
Particle Size: HSCb = 14 nm,
LSCb = 70 nm
Reference: Whole-body Inhalation
Dose/Concentration: 0,1, 7, 50 mg/m3 HSCb; 50
mg/m3 LSCb (rats only)
Time to Analysis: 6 h/day, 5 daus/wk for 13 wk.
Parameters measured 1 day, 3 mo, 11 mo post-
exposure
Body Weight: Environmental changes pre and post-
exposure affected test subjects' life spans, particularly
hamsters. Hamsters also experienced significant loss
of body weight when exposed to high doses of HSCb.
Effects of Carbon Black: In rats, lung weight of the
high dose HSCb doubled. After 11 mo, analysis of all
lungs showed no significant difference. Mice had the
highest relative lung burdens at the end of exposure
time but also cleared particles faster at high doses than
rats. However, clearance slowed over the 11 mo
recovery period, especially in high dose mice.
Hamsters showed significant elevations in lung carbon
black burden for all exposures at all time points.
Hamsters exposed to high dose HSCb exhibited
impaired clearance.
BALF Cells: Presence of PMNs was limited to the mid
and high dose groups. Overall maximal response was
reached in mice and hamsters, but not in rats with
increasing mass dose of HSCb.
Reference: Evans
et al. (2006, 0970661
Species: Rat
Gender: Male
Strain: SD
DEP: collected under dry,
outdoor, ambient conditions from
tractor exhaust pipe (1985,
Japanese ISEK11500 cc tractor)
burning Esso 2000 diesel and
20/30 mixture of Esso light
engine oil.
10% UF, 90% fine
Cabosil: amorphous silicon
dioxide
16%UF, 84% fine
Particle Size: DEP: 30 nm;
Cabosil: 7 nm
Route: IT Instillation
Dose/Concentration: 1 mg/rat DEP; 1 mg/rat
Cabosil
Time to Analysis: Pretreatment with 0.5 unit of
bleomycin; IT 3 or 7 days
after pre-treatment; 1wk post-IT
Lung permeability: In bleomycin-treated group,
obvious inflammatory status and edema within the lung
was observed. This was shown by significant increases
in acellular protein and free cells.
Changes in lung: Body weight ratio, lung surface
protein content, free cell counts, and apical surface
protein of rat type I cells were only altered by
bleomycin treatment and not particle exposure.
December 2009
D-88
-------
Reference
Pollutant
Exposure
Effects
Reference: Finnerty Coal Fly Ash (generated at U.S Route: IT Instillation
et al. (2007,1564341 EPA National Risk Management
Research Laboratory by burning Dose/Concentration: PM: 200 mg/mouse; 9.1
Species: Mouse Montana subbituminous coal m9/k9
under conditions simulating full- PM+LPS10: 200 mg PM+10 mg IPS
scale utility boiler conditions) PM+LPS100: 200 mg PM+100 mg IPS
.... ~C7D, „,, IPS: 100 ug
Strain. C57BL/61 Trgnsition metg|s Qf Cog| R
Age:12wk Ash: Fe, Mg, Ti, Mn, V Time to Analysis: 18 h
Weight: 24.3 ± 0.3 g Particle Size: >PM25
Gender: Male
BALF Cells: No significant differences in platelet
concentration or white blood cell count in any groups
were observed. The percentage of neutrophils in-
creased significantly with PM+LPS100. PMN rose in
PM groups and increased further with IPS treatment.
Increases in PM+LPS were groups statistically
significant. More leukocytes were present in the
alveolar space in PM+LPS10 compared to the PM
group. The most severe response was in the
PM+LPS100 group.
Cytokines: Plasma TNF-a and IL-6 significantly
increased for the PM+LPS100 group. An additive effect
of IPS and PM for IL-6 was observed. For saline and
PM groups, pulmonary TNF-a was below detection
range. A synergistic effect for TNF-a was observed. A
less than additive effect for IL-6 was observed.
Pulmonary TNF-a significantly increased in the PM+
LPS100 group. Pulmonary IL-6 significantly increased
in both PM+LPS groups.
Reference: Fujimaki DEP: collected from a 4-cylinder, Route: Whole-body Inhalation
et al. (2006, 0966011 2.74 L, Isuzu diesel engine
Species: Mouse Particle Size: 0.4 |jm
Gender: Male
Strain: IL-6(-/-) and
WT: B6J129SV
(control)
Age: 5-6 wk
Dose/Concentration: 1.0, 3.0 mg/m
Time to Analysis: 12 h/dayfor4wk. Parameters
measured 1 day post-exposure
BALF Cells: Treatment significantly increased BAL
cells from WT mice at both dose levels. The increase of
macrophages and neutrophils were dose-dependent.
An increase in lymphocytes were present in WT mice
with the low dose. No significant increase in cells were
observed from IL-6 (-/-).
Cytokines: TNF-a largely increased in IL-6(-/-) mice
exposed to 3 mg/m3 compared to WT mice. IL-6
production increased in WT mice exposed to 3 mg/m3.
CCL3 increased in both WT and IL-6(-/-) at high dose.
IL-1p remained at the control level.
Reference: Gerlofs- RTD: road tunnel dust (obtained Route: IT Instillation
Nijland et al. (2005,
0886521
Species: Rat
Gender: Male
Strain: SH/NHsd
Age: 11-12 wk
Weight: 250-350 g
from a Motorway tunnel in
Hendrik-ldo-Ambacht, Nether-
lands)
EHC-93
(Ottawa, Canada)
Particle Size: Coarse: 2.5-10
pm; fine: 0.1-2.5pm
Dose/Concentration: 0.3,1, 3,10 mg/kg; EHC-
93: 10mg/kg
Time to Analysis: 4, 24, 48 h
BALF Cells: PMN significantly increased in RTD (3
and 10 mg/kg dose) and EHC-93 exposed animals at
24 h and decreased by 48 h but remained statistically
significant. AM numbers decreased for 3 mg/kg RTD
group at 4 h.
BAL Inflammatory/Injury Markers: Myeloperoxidase
(measured at 24 h in 1, 3,10 mg/kg RTD groups) was
elevated in a dose-dependent manner. RTD induced
time-dependent increases in LDH activity at 24 and 48
h, although these increases were less than EHC-93
values at the same time points. Alkaline phosphatase
increased dose-dependently for RTD at 48 h. GSH
decreased at 24 h to approximately the same levels in
0.3,1, and 3 mg/kg RTD dose groups. Uric acid only
decreased in 1 mg/kg RTD group at 24 h.
Cytokines: IL-6 levels were elevated only at 10 mg/kg
for RTD and EHC-93 at 4 and 24 h; it remained
elevated for EHC-93 at 48 h. A dose-dependent
increase in TNF-a at 4 h for RTD was observed. TNF-a
levels remained elevated only for the 10 mg/kg groups
at 24 h and returned to control levels by 48 h. A dose-
dependent increase in MIP-2 for all RTD dose groups
were observed and remained elevated through 48 h for
both PM types (although values were returning to
control levels).
Pulmonary Histopathology: A dose-dependent
increase in the number of inflammatory foci at 24 and
48 h for 3 and 10 mg/kg RTD groups was observed.
The response was even greater for the EHC-93
exposed group at similar time points.
December 2009
D-89
-------
Reference
Pollutant
Exposure
Effects
Reference: Gerlofs-
Nijland et al. (2007,
0978401
Species: Rat
Gender: Male
Strain: SH/NHsd
Age: 13 wk
Weight: 250-350 g
PM samples collected from:
1. MOB high traffic density
2. HIA high traffic density
3. ROM high traffic density
4. DOR moderate traffic density
5 MGH low traffic density
6 LYC low traffic density
Particle Size: Coarse: 2.5-10
pm; Fine: 0.1 -2.5pm
Route: IT Instillation
Dose/Concentration: 3,10 mg/kg
Time to Analysis: 24 h
Age: 12 wk
Weight: 200-300 g
BALF Cells: Pulmonary inflammation was induced in a
significant and dose-dependent manner for both dose
levels. Inflammation in the BALF included airway
neutrophilia, increased macrophage numbers and mild
lymphocytosis. Both coarse and fine PM caused dose-
dependent alveolitis. Fine PM from LYC (10 mg/kg
dose) also caused some bronchiolitis.
BAL Inflammatory/Injury Markers: LDH was
significantly increased for all doses of coarse PM and
for the high dose of fine PM. BALF protein
concentration was observed predominantly at the high
dose of coarse PM. Location ROM had evidence of
attenuated responses with fine PM. Ascorbate concen-
trations were reduced but were only significant for rats
exposed to the highest dose of coarse PM fractions
from the locations MOD, HIA, and LYC.
Cytokines: TNF-a concentrations increased for all
coarse samples with the exception of DOR and LYC.
Fine PM induced similar responses for all sites. MIIP-2
concentrations increased only at certain sites for
coarse but not fine PM.
Location-related Differences: Coarse PM from MOB,
HIA and MGH induced higher LDH responses than
other locations. Coarse PM from HIA produced BALF
protein concentrations higher than LYC and ROM.
MGH induced greater amounts of BALF protein than
ROM. Coarse PM from LYC lowered fibrinogen values
more than PM from location MOB, HIA, and MGH. Fine
PM showed less differences among the various sites.
Particle Correlation: Fine PM exhibited significant
correlation between zinc content and BALF cytotoxicity
markers protein and LDH - mainly from HIA. Fine PM
also exhibited positive correlations with copper and
barium. Coarse PM showed positive correlation with
barium and copper, mainly from MOB.
Reference: Gerlofs- PM (Prague, Czech Republic;
Nijland et al. (2009, Duisburg, Germany; Barcelona,
1903531 Spain) (Prague and Barcelona
coarse PM organic extracts)
Species: Rat
Particle Size: Coarse: 2.5-10
Gender: Male ^ Fine: 0.2-2.5 pm
Strain: SH
Route: IT Instillation
Dose/Concentration: 7mg/kg
Time to Analysis: 24 h
Cytotoxicity (LDH, protein, albumin) and inflammation
(NAG, MPO, TNF-a were increased by PM, and were
greatest in the coarse PM fraction. Metal-rich PM had
greater inflammatory and cytotoxic effects. PAH content
influenced greater inflammation (including neutrophils),
and cytotoxicity. Generally, whole PM and coarse PM
were more potent than organic extracts and fine PM,
respectively.
Reference: Ghio et
al. (2005, 0882721
Species: Rat
Gender: Male
Strain: N8 b/b
Belgrade rats and
N8+ Ib Belgrade
controls
Oil Fly Ash (Southern Research
Institute, Birmingham, AL)
Particle Size: 1.95 +0.18 fjm
(MM AD)
Route: IT Instillation
Dose/Concentration: 500 pg/rat; 2 mg/kg
Time to Analysis: 24 h
BALF Cells: Homozygous Belgrade with mutation
G185R had higher levels of Fe and V 24 h post-
exposure. This may demonstrate a decreased ability to
remove Fe and V from the lower respiratory tract than
heterozygous +lb littermates. This also indicates that
DMT1 is normally responsible for at least some Fe and
V uptake; thus, a defective DMT1 transports less.
BAL Inflammatory/Injury Markers: Increased protein
and LDH concentrations in the homozygous strain were
observed when compared to control
Reference: Ghio et
al. (2005, 0882751
Species: Rat
Gender: Male
Strain: SD
Age: 60 day
Weight: 250-300 g
Ferric ammonium citrate (FAC)
Vanadyl sulfate (VOS04)
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 0.5 mL 100 pm FAC/rat; 0.5
mL 10 pm VOS04/rat; 500 pg oil fly ash; 2 mg/kg
Time to Analysis: Single or double exposure with
24 h rest period. Parameters measured 15, 30, 60
min, 24 h post-exposure.
DMT1 Immunohistochemistry and Lung Injury: FAC
increased and VOS04 decreased -IRE DMT1 staining.
Same exposures had no effect on +IRE DMT1. -IRE
DMT1 expression in macrophages, airway and alveolar
epithelial cells increased with increased Fe exposure.
Vanadium nearly eliminated staining except in alveolar
macrophages. Increased metal clearance with pre-
exposure to FAC. Less metal clearance with pre-
exposure to VOS04. Pre-exposure to iron diminished
lung injury whereas pre-exposure to vandium increased
lung injury after oil fly ash instillation. Lung injury
measured by concentration of protein and LDH in BAL.
December 2009
D-90
-------
Reference
Pollutant
Exposure
Effects
Reference: Gilmour
et al. (2007, 0964331
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 10-12 wk
Weight: 20-22 g
PM-CO, Fl, UF
(obtained from U.S. Seattle (S),
Salt Lake City (SL), South Bronx
(SB), Sterling Forest (SF))
SB: included 35% sulfate, 22%
gasoline, diesel and brake wear.
SF: 48% sulfate.
SL: 34% wood combustion and
28% sulfate
S: 39% wood combustion and
29% sulfate
Residual oil combustion and soil
dust less than 5% for all sites.
Particle Size: CO: 2.5-10 pm;
FI:<2.5|jm;UF:<0.1 pm
Route: Oropharyngeal Aspiration BALF Cells: PMN increased with the high dose of CO
samples from SB, SL, S, but not SF. No significant
Dose/Concentration: 25 pg or 100 pg PM; 1.25 or increases from Fl were observed, though the high dose
5 m9'k9 induced increased PMN. UF from SL caused a highly
Time to Analysis: 18 h variable response.
BAL Inflammatory/Injury Markers: Seattle CO
fractions showed no dose-dependent effect on protein
concentration. Results for other locations were
distinctly higher with 100 pg dose than 25 pg and
saline doses. SL CO high dose induced the most
significant increase. LDH response was weakly dose-
related. Only SB showed a statistically significant
increase for LDH with the high dose UF.
Cytokines: MIP-2 was similar to PMN response. SB
CO induced the most significant response. SL UF was
highly variable.
Particle Characteristics: LPS was higher in S (CO, Fl,
UF) and SL (CO, Fl, UF). Zn levels were highest in SB
(CO, Fl, UF). Fe was higher in all CO and Fl samples
with SB CO inducing the highest.
Reference:
Gilmour et al. (2004,
0574201
Species: Mouse
Gender: Female
Strain: CD1
Age:8-10wk
Weight: 20-25 g
Coal Fly Ash
MU: Montana Ultrafme
MF: Montana Fine
MC: Montana Coarse
KF: W. Kentucky Fine
KC: W. Kentucky Coarse
Particle Characteristics:
Montana Sulfur 0.83%, Ash
11.72%. Trace amounts of Ba, P,
Sr, V, Nb, Cd, Se, Ga, Cu.
Depleted in Si, Al, Fe, Mg, Ti.
Kentucky Sulfur 3.11%, Ash
8.07%
Particle Size: Coarse: >2.5 pm;
Fine: <2.5|jm;
Ultrafme: <0.2 pm
Route: Oropharyngeal Aspiration
Dose/Concentration: 25 ug or 100 pg/mouse
Time to Analysis: 18 h
BALF Cells: PMN highly increased for MU at both
doses. The level was comparable to the positive
control. PMN also increased with KF at high dose.
Coarse particles caused no significant increase in
PMN. Number of macrophages did not change, but
NAG increased significantly with MU for both dose
levels and with KF and MF at high dose level.
BAL Inflammatory/Injury Markers: Total protein and
LDH was not significantly elevated. Albumin
concentration increased significantly after treatment
with the fine high dose of both particle types.
Cytokines: MU particles caused a significant increase
in TNF-a. MIP-2 increased in all fine and ultrafme PM-
instilled animals with the highest in the MU and KF at
both doses. IL-6 was detectable only in the BALF of
MU and KF with substantial variability. The IL-6 levels
were not significant.
Reference: Gilmour
et al. (2004, 0879481
Species: Rat
Gender: Male
Strain: SH/NQIBR,
WKY
Age: 12 wk
Weight: 280-340 g
PM (collected from precipitator
unit of an oil burning power plant
in Boston)
Measured Components of PM:
S, Zn, Ni, V,AI, Cu, Pb, Fe, Ca,
Na, K, Mg, Endotoxin
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 0.0, 0.83, 3.3, and 8.3
mg/kg in SH rats; 0.0 or 3.3 mg/kg in WKY and SH
rats
Time to Analysis: 24 h
BALF Cells: No increase in macrophage number was
observed in either rat strain following saline or PM
exposure at 24 h.
BAL Inflammatory/Injury Markers: LDH activity
increased in a dose-related manner; this was observed
in SH rats after exposure to 0.83, 3.33 and 8.3 mg/kg
PM. SH rats showed greater lung permeability following
PM exposure than WKY rats. SH rats showed acute
lung inflammatory response after exposure to PM when
com pared to WKY rats.
Cytokines: MIP-2 mRNA expression increased
significantly in SH PM exposure group only. No
significant differences in TNF-a RNA expression in
either WKY, SH rats or control treatment groups were
observed.
CD14: A significant increase in lung CD14 protein was
observed only in SH rats exposed to PM.
TLR4: A significant increase in TLR4 protein in SH rats
exposed to PM was observed.
NF-KB: A significant increase in NF-KB binding protein
in the nuclei of SH rats exposed to PM was observed.
This effect was not observed in the control of PM-
exposed WKY rats.
December 2009
D-91
-------
Reference
Reference: Gilmour
et al. (2004, 0541751
Species: Rat
Gender: Male
Strain: Wstar Kyoto
Age: 12 wk
Reference:
Godleski et al.
(2002, 1564781
Species: Rat
Gender: Male
Strain: SD
Age: NR
Weight: 200-250 g
Reference:
Gottipolu et al.
(2009, 1903601
Species: Rat
Gender: Male
-
Strain: Wstar
Kyoto, SH
Age: 14-16 wk
Weight: NR
Reference:
Gunnison and Chen
(2005, 0879561
Species: Mouse
Gender: Male
Strain: DK (ApoE"'",
LDLr'')
Age: 18-20 wk
Reference:
Gurgueira et al.
(2002, 0365351
Species: Rat
Gender: Male
Strain: SD
Weight: 250-300 g
Pollutant
ufCB: Ultrafme carbon black
(Printex 90 (Degussa)
CB:(Huber990, HR. Haeffner
and Co)
Particle Size: ufCB: 14nm;CB:
260 nm (primary particle
diameter)
CAPs (Boston; Harvard Ambient
Particle Concentrator)
Particle Size: 0.27 + 2. 3pm
(diameter)
DE(30-kW(40hp)4-cylinder
indirect injection Deutz diesel
engine) (0,- 20%, CO- 1.3-4.8
ppm, NO- <2.5-5.9 ppm, N02-
<0.25-1. 2 ppm, S02- 0.2-0.3
nnm DP/FP H ? + D CfV\
|J|JIM, \J\slC\s~ U.O ± U.UOJ
Particle Size: Number Median
Diameter: Low- 83 + 2 nm, High-
88.2 nm; Volume Median
Diameter: Low- 207 + 2 nm,
High- 225 + 2 nm
CAPS
(Northeastern regional back-
ground)
Ambient air copollutants
measured 03, N02
Particle Size: 389 + 2 nm
CAPs
(Harvard Ambient Particle
Concentrator)
CB
(C1 98 Fischer Scientific, Pitts-
burg, PA USA)
Composed of 85.9+ 0.2%
Carbon, 13.0+ 0.2% 02, 1.17 +
0.2% Sulfur
ROFA
(Boston, MA USA oil-fired power
plant)
Particle Size: CAPs: 1-2.5 pm;
CB: <2.5 pm; ROFA: <2.5 pm
Exposure
Route: Whole-body Inhalation
Dose/Concentration: ufCB: 1.66 mg/m3
fCB:1.40mg/m3
Number concentrations
ufCB: 52380 particles/cm3
fCB: 3800 particles/cm3
Time to Analysis: Exposed for 7 h. Sacrificed 0,
16or48h post-exposure.
Route: Whole-body Inhalation
Dose/Concentration: 73.5-733 pg/m3
Time to Analysis: Exposed 5 h/days, 3 days
(consecutive). BAL 24 h post-exposure
Route: Inhalation
Dose/Concentration: Low- 507 + 4 pg/m3, High-
2201 + 14|jg/m3
Time to Analysis: Exposed 4 h/day, 5 days/wk, 4
wk. Necropsied 1 day post-exposure.
Route: Whole-body Inhalation
Dose/Concentration: CAPS =131+99 pg/m3
including 03 = 10 ppb and N02 = 4.4 ppb
Time to Analysis: 6 h/day, 5 days/wk for 4 mo
(5/12/03-9/5/03). Sacrificed 3-4 days post-
exposure.
Route: Whole-body Inhalation
Dose/Concentration: 300 + 60 pg/m3
Time to Analysis: 1, 3, 5 h CAPs Exposure
followed by immediate post-exposure analysis.
5 h CB, immediate analysis.
SOmin ROFA, Immediate analysis.
Effects
BALF Cells: Total number of cells increased
significantly in UfCB-exposed rats at 0 and 16 h.
Recruitment of cells did not occur in response to CB
exposure. PMNs increased significantly in the BALF of
ufCB-exposed rats at 16 h. Leukocytes remained
unchanged following CB exposure but increased
significantly at 0 and 48 h post exposure to ufCB.
Cytokine mRNA: A significant increase in BALF MIP-2
mRNA expression was observed at 48 h. No
differences in MIP-2 mRNA levels were observed in the
whole lung tissue.
BALF Cells and Inflammatory Markers: PMNs
significantly increased with CAPs exposure and also in
relation to CAPs mass, Br, S042", EC, OC and Pb. An
overall increase in pro-inflammatory mediators and
decrease in immune enhancer and evidence of
vascular endothelial responses occurred with CAPs
exposure.
DE increased neutrophils in a concentration-dependent
manner, and GGT activity at the high dose. Particle-
laden macrophages were found in DE-exposed rats.
Microarray Data: 13 genes in the heart tissue and 47
genes in the lung tissue were identified as possibly
affected. Strict standards (1.5 fold response, 10% false
discovery rate) resulted in responses by only 1/13
genes (Rex3 - no known heart physiology) in the heart
tissue and 0/47 genes in the lung tissue. Using more
liberal response (nonstatistical) standards (1.5 fold
only) and comparison of each CAPS animal with all 3
control animals (3x3 array) resulted in possible effects
on 7 additional genes in the heart tissue and 37 genes
in the lung tissue.
In situ Chemiluminescence(CL): Data show a
significant increase in lung and heart CL at 5 h. Lung
CL increased linearly with time of exposure.
Oxidants: CAPs-initiated oxidative stress was not
detectable in those rats allowed to recover in room air
after the simulated "peak' in particulate air pollution.
Rats breathing particle-free filtered air for 3 days had
significantly lower levels of oxidants. Exposure to inert
CB did not exert oxidant effects on the heart and lung.
BAL Inflammatory/Injury Markers: The water content
of the lung and heart increased significantly upon
exposure to CAPs but not to filtered air and increased
as a function of length of exposure. Rats breathing
CAPs also showed increases in LDH and CPK as a
function of length of exposure.
Antioxidant Enzymes: Data showed an increase in
SOD and catalase activities in both the lung and heart.
The pattern of increase was tissue specific.
December 2009
D-92
-------
Reference
Pollutant
Exposure
Effects
Reference: Hamoir
et al. (2003, 0966641
Species: Rabbit
Strain: New
Zealand
Age: 12-16 wk
Weight: 2.8 +0.5kg
PSC: Polystyrene particles, Car-
boxylate modified, 3 types
PSA: Polystyrene particles,
Amine modified, 1 type
Particle Size: PSC: 24,110 or
190nm(PSC24, PSC110,
PSC190);PSA:190nm
Route: IT Instillation
Dose/Concentration: PSC24: 0.04 or 4 mg/rabbit
PSC110, PSC190, PSA190: 4 mg/rabbit
Time to Analysis: 0, 30, 60, 90,120 min
Capillary Filtration Coefficient: A time-dependent
increase correlating to total number of particles/surface
area, not particle size, was observed. PSA induced a
significant increase in microvascular permeability as
compared to PSC. This suggests that the number of
particles exposed should be considered an important
parameter for measuring air quality rather than total
particle surface area.
Reference: Happo
et al. (2007, 0966301
Species: Mouse
Gender: Male
Strain: C57BL/6J
Weight: 19-30 g
Age: 10-11 wk
PMC (Coarse)
PMF (Fine)
PMUF(Ultrafme)
Collected in 6 European cities:
Duisburg, Prague, Amsterdam,
Helsinki, Barcelona, Athens
Particle Size: PMC: PM10-2.5;
PMF: PM25-0.2; PMUF: PM0.2
Route: IT Instillation
Dose/Concentration: 1, 3,10 mg/kg
Time course: 10 mg/kg
Time to Analysis: 1. Dose-Response study:
parameters measured 24 h post exposure. 2. Time
course study: parameters measured 4,12, 24 h
post single exposure (at 10 mg/kg).
BALF Cells: 1. For the dose-response study, all the
PMC samples exhibited dose-dependent increases of
total cell numbers. The 3 and 10 mg/kg doses of PMC
induced statistically significant increases. At 10 mg/kg,
only 2/6 samples induced statistically significant
increases. No PMUF samples induced effects at any
dose. 2. For the time-response study, no increases in
cell numbers were shown at 4 h. Though the levels
induced by PMC at 24 h were lower than at 12 h, both
levels were statistically significant. PMF induced
statistically significant increases only at 12 h for 4/6
samples. PMUF induced only 1 significant increase at
12 h; the 24 h time point was not tested.
BAL Injury Markers: 1. The lower doses of 1 and 3
mg/kg did not induce significant increases in any of the
PM samples, except for PMUF-Athens. All 6 samples of
PMC, at 10 mg/kg, induced significant increases. At 10
mg/kg, 4/6 PMF samples induced significant increases.
2. At 4 h, none of the samples increased protein
concentration. The PMC samples, excluding Prague,
induced significantly higher concentrations at 12 h. At
24 h, only 3/6 PMC samples induced significant
increases. Only 2 PMF samples induced significant
increases at 12 and 24 h. At 12 h, effects induced by
PMUF were minimal and inconsistent; the 24 h time
point was not tested.
Cytokines: 1. Only PMC induced dose-dependent
responses that reached statistical significance at 10
mg/kg. PMF and PMUF induced minimal and
inconsistent responses. 2. TNF-a levels increased
significantly at 4 and 12 h by PMC. At 24 h, TNF-a
levels returned to near control levels. PMF, at 4 h,
induced statistically significant increases for 3/6
samples and significant increases in 2/6 samples at 12
h. No PMUF samples significantly increased TNF-a
levels. PMC induced the highest IL-6 levels at 4 h.
Levels at 12 and 24 h were reduced with 6/6 and 3/6
samples showing statistically significant increases,
respectively. PMF showed a similar trend with 4 h
inducing the highest levels that were reduced at 12 and
24 h. Of the PMUF samples, only the Helsinki and
Duisburg samples induced statistically significant
results at 4 and 12 h. Generally, the PMUF responses
were negligible when compared to PMC and PMF. 2.
All PMC samples induced the highest levels of KC
production at 4 h. At 12 and 24 h, levels were reduced
but 4/6 samples induced statistically significant levels.
PMF showed a similar trend- the highest levels were
induced at 4 h (in 3/6 samples). PMUF at 4 h showed
small, though not significant, increases. At 12 h, only 2
samples showed statistically significant differences
from the control; the 24 h time point was not tested.
Reference: Harder
et al. (2005, 0873711
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Age: 14-17 wk
Weight: NR
Carbon UFP
Particle Size: 37.6 +0.7 nm
(diameter)
Route: Inhalation
Dose/Concentration: 180 pg/m3
Time to Analysis: 24 h exposure. 3 day recovery.
UFP induced mild pulmonary inflammation, significantly
increased PMN, and increased the total protein and
albumin concentrations. Particle-laden macrophages
sporadically accumulated in the alveolar region.
December 2009
D-93
-------
Reference
Pollutant
Exposure
Effects
Reference:
Harkema et al.
(2004, 0568421
Species: Rat
Gender: Male
Strain: F344, BN
Age: 10-12 wk
Weight: NR
CAPs (Detroit; July-Sept. 2000; Route: Inhalation, IT Instillation.
Harvard Ambient Fine Particle
Concentrator) Dose/Concentration: 4 day concentration: 676 ±
288 pg/m3, 5 day concentration: 313+119 pg/m3,
Particle Size: 2.5 pm (diameter) July concentration: 16-185 pg/m3, September
concentration: 81-755 pg/m3; IT Instillation- 200 pL
(soluble and insoluble)
Time to Analysis: F344 rats sensitized to
endotoxin, BN rats to OVA. Exposed 10 h/day 1, 4,
5 days (consecutive). Another group of rats IT
instilled. Both groups killed 24 h post-exposure.
The retention of PM in the airways was enhanced by
allergic sensitization. Recovery of anthropogenic trace
elements was greatest for CAPs-exposed rats.
Temporal increases in these elements were associated
with eosinophil influx, BALF protein content and
increased airway mucosubstances. A mild pulmonary
neutrophilic inflammation was observed in rats instilled
with the insoluble fraction but instillation of total, soluble
or insoluble PM25 in allergic rats did not result in
differential effects.
Reference:
Hiramatsu et al.
(2003, 1558461
Species: Mouse
Gender: Female
Strain: BALB/c and
C57BL/6
Age: 8 wk
Weight: 17-22 g
DE: generated by 2369-cc diesel Route: Whole-body Inhalation
engine (Isuzu) at 1050 rpm and
80% load with commercial light
Dose/Concentration: DEP: 100 pg/m3 or 3
Particle Size: NR
mg/m ; S02<0.01 ppm; N02 2.2 ±0.3 or 15 ± 1.5
ppm;C03.5 ± 0.1 or 9.5 ±0.6 ppm
Time to Analysis: 7h/d, 5 days/wk for 4 or 12 wk,
Immediate
BALF Cells: Alveolar macrophages (AMs) increased
dose-dependently at 30 and 90 day. High DE exposure
resulted in bronchus-associated lymphoid tissue
(BALT) around DEP-AMs; this was less conspicuous in
C57BL/6 than in BALB/c mice. B- and T-cell
populations were found in the BALT with no significant
differences observed between the strains.
Lymphocytes and neutrophils increased time- and
dose-dependently with a greater increase in BALB/c
than C57BL/6 observed. No eosinophils or basophils
were observed. Mac-1-positive cells exposed to high
DE levels increased in both strains at 1 month (33.8%)
and 3 mo (20.3%) vs. low dose group (5.3 and 7%
respectively).
Cytokines: At 30 days, TNF-a, IL-12p40, IL-4 and IL-
10 mRNA increased, IL1b and iNOS decreased. IFN-y
increased in BALB/c but decreased in C57BL/C. IL-6
mRNA was not affected. At 90 day, IL-4 and IL-10
mRNA similarly increased in C57BL/6 mice exposed to
low DE level but decreased at high DE level.
Reference:
Hollingsworth et al.
(2004, 0978161
Species: Mouse
Gender: Male
Strains:
C57BL//6TLR*'*,
C57BL//6TLR"'"
Age: 8-9 wk
ROFA
Particle Size: NR
Route: Oropharyngeal Aspiration
Dose/Concentration: 50 pi of 1pg/mL suspension
per mouse
Time to Analysis: Parameters measured post
single exposure of 6 and 24 h.
Methacholine sensitivity: No ROFA effect was
observed in wild type or knockout mice.
BALF Cells: ROFA increased total cell number. Total
number of neutrophils with lavage fluid increased 24 h
post-exposure in both strains.
Reference:
Hutchison et al.
(2005, 0977501
Species: Rat
Gender: Male
PM10
United Kingdom
samples collected before (-B),
during closure (-C) and
reopening of steel plant (-R)
Route: IT Instillation
Dose/Concentration: 112 to 180 pg PM in 500 pi;
0.44-0.72 mg/kg
Time to Analysis: 18 h
PMT = PM total (aqueous
sonicate)
Strain- Wistar Kvoto PMS = PM aclueous supernatant
btram. vvistar Kyoto pM| = pM inso|ub|e pe||et
Age:3m Particle Size: PM10
Weight: 250-300 g
BALF Cells: PMT-R neutrophil cell number and
percentage were significantly higher than PMT-Cor
control. PMS-R and PMI-R were also higher than their
respective controls. The neutrophil cell numbers
induced by PMI-R were greater than PMI-C and the
control. Total cell count unchanged.
BALF Inflammatory/Injury Markers: Only albumin
increased after PMT-R. Upon exposure, total protein
and LDH did not increase.
Cytokine mRNA expression: Only PMT-R increased
IL-1B mRNA expression. No effects on TNF-a and
TGF-B expression levels were observed. IL-6, MIP2,
and GM-CSF mRNA was not detected in BAL cell
extracts from either the control or treated groups.
Reference: Inoue et DEP (derived from 4 cyl, 2.74!
al. (2006, 0978151 light duty diesel engine)
Species: Mouse
Gender: Male
Strains: C3H/HeJ
(TLR-4 point mutant)
and C3H/HeN
(Control)
Age: 6 wk
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 12 mg/kg
Time to Analysis: 24 h
BALF Cells: DEP induced an increase in total cells,
neutrophils, and mononuclear cells. TLR4 knockout
mice (C3H/HeJ) showed a much lower response.
Cytokines: DEP induced a massive increase in MIP-
1x, IL-1P and KC. However, levels of MIP-1x were
significantly less in the knockout than the wild type
while levels of IL-113 and KC were significantly higher in
knockouts than the wild type.
December 2009
D-94
-------
Reference
Pollutant
Exposure
Effects
Reference: Inoue et DEP (derived from 4 cyl, 2.741
al. (2005, 0974811 light duty diesel)
Species: Mouse
Gender: Male
Strain: NC/Nga
Age: 10 wk
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 100 pg/mouse
Time to Analysis: 1/wkfor6wk. Parameters
measured 24 h after last administration
BALF Cells: DEP significantly increased total cells,
neutrophils and mononuclear cells but did not induce
an effect on eosinophils.
Cytokines: DEP increased IL-4, KC and MIP-1. The
increase in IL-5 was not statistically significant.
Reference: Ishihara
et al. (2003, 0964041
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Age: 5 wk
DE
(from 2 engines, produced on
site)
-L = low level DE
-M = medium level
-MG = DE w/o particulates
-HR = high level
Measured Components: N02,
S04, S02, CO, C02, NOX, NO,
HTHC, HCHO, 02
Particle Size: L: 0.33-0.50 pm
M: 0.35-0.40 pm
HR: 0.42-0.45 pm
Route: Whole-body Inhalation
Dose/Concentration: L: 0.18 - 0.21 mg/m3
M: 0.92-1.18 mg/m3
MG: 0.01 mg/m
HR: 2.57-2.94 mg/m3
Time to Analysis: 16h/day, 6days/wk, for 6,12,
18 & 24 mo. Parameters measured immediately
following last exposure.
Morbidity and Mortality: Weight gain in HR group was
less than other groups at 18 and 24 mo. This indicates
a significant difference between the HR and C group.
Mortality during the study was frequent. C group
experienced an 8% mortality rate, L group 12%, M
group 15%, MG group 12% and HR group 23%.
BALF Cells: The HR group showed a significant
increase in total cell count from 6 to 18 mo. The
percentage of PMN increased at 6mo in M, MG and HR
group. M group lymphocytes significantly increased at
6,12, and 24 mo of exposure. Macrophages decreased
at 6 mo for the M and HR groups.
BAL Inflammatory/Injury Markers: Significant
differences were seen among groups with respect to
number of total cells and percentages of cell
differential, total protein, fucose, sialic acid,
phospholipid and prostoglandin E2. Total protein
increase was observed in both M and HR dose groups
with the HR group increasing time-dependently.
Mucus and Surfactant: The HR group showed a
significant increase from 12 to 18 mo.
Reference: Jones et ASP: Amorphous silica particles
(Hypersil)
al. (2005, 1988831
Species: Rabbit
Strain: New
Zealand
Weight: 2.5- 3.5 kg
MCSP: Microcrystalline silica
particles
Particle Size: ASP: 5pm;
MCSP: 5 pm
Route: Intrapulmonary Instillation (Right upper
lobe of lung)
Dose/Concentration: 50mg in 0.5 ml saline
Time to Analysis: Parameters measured at
varying times from 6 h to 91 days post treatment.
MCSP: At 6 h, neutrophils increased. Macrophages
increased 3 fold. At 60 h, neutrophils were pyknotic and
the lungs displayed a thickened interstitium containing
silica particles. At 5 days, collagen deposition
appeared. At 8 days, fibroblastic activity and necrosis
were observed. At 15 days, aggregation of silica
particles and necrotic debris were apparent. At 8 wk,
fibroblasts were still present. At 13 wk, active scarring
and raised neutrophil macrophage counts were still
present.
ASP: At 15 h, neutrophils increased. Macrophages
tripled and remained increased for 3wk. At 4 day,
macrophages bore particles. At 13 day, neutrophils
decreased significantly. By 25 day, silica spheres were
gradually removed from lungs.
PET Scanning: 18F-fluoroproline showed increased
activity beginning at 14 days and peaking at 41-54
days.
Microautoradiography: 3 h-proline at 13 wk showed
radiolabel localization to fibroblasts in the challenged
lung.
December 2009
D-95
-------
Reference
Pollutant
Exposure
Effects
Reference: Kato
and Kagawa (2003,
Species: Rat
Gender: Male
Strain: Jcl Wistar
Age: 5 wk
Roadside air
(Prefectural Tokyo-Danishi-
Yokohama highway, Yokohama-
Haneda Airport Metropolitan
expressway and Satsukibashi-
Mizuecho city road, Japan)
Particle Size: NR
Route: Whole-body Inhalation
Dose/Concentration: Exposed group: 62.7 pg/m3
PM, 557ppbN02,;
Control group: 14.3 pg/m3 PM, 5.1 ppb N02
Time to Analysis: Exposed for 24, 48, 60 wk.
Parameters measured immediately following
exposure.
Respiratory Tissue: Post 24 wk, the lung surface was
light gray with some BC particle deposits. Post 48-60
wk, however, the surface was scattered with particle
deposits in addition to its light gray color.
Airway Changes: After 60 wk, no remarkable changes
seen in the epithelium. The structure of the airways
remained normal.
Cells: No proliferation or ectopic growth of goblet cells
were noted. Mast cells increased in epithelial
intercellular space. No mast cell degranulation was
observed. Lysosomes increased in ciliated cells post
48 wk. Clara cells were unaffected.
Lymph Nodes: Deposition of carbon particles were
noted in the trachea and bronchiole-associated lymph
nodes post 24 wk.
Alveolar Changes: No changes in morphology of
broncho-alveolar junctions were noted. Anthracosis
observed within alveolar walls and pleura post 24 wk
and became progressively marked with increased
exposure. No change in the number of alveolar holes
between exposure and control groups were observed.
Reference: Kato et
al. (2003, 1988821
Species: Rat
Gender: Male
Strain: SD
Age: 7 wk
Weight: 190-220 g
Polystyrene latex suspension of Route: IT Instillation with nebulizer
latex beads (Japan Synthetic
Rubber Co) uncoated or coated Dose/Concentration: 5 ml of 0.2% suspension
with lecithin' administered over 20 min at flow rate of 0.25
ml/min
Particle Size: 240 nm
Alveolar Macrophages: Following treatment, AMs
appeared undamaged. AMs ingested more uncoated
than coated beads, but both were ingested. Ingestion
of beads differed as coated beads were engulfed
individually while uncoated beads were engulfed
Time to Analysis: Exposed for 20 min.
Parameters measured 30 min following treatment.
individually or in aggregates.
Epithelial Cells: Tyf
Type I cells incorporated coated
beads within a layer of cytoplasm. Type II cells
incorporated beads in lamellar bodies. Uncoated beads
were not incorporated.
Other: Neither type of beads were incorporated into
endothelial cells, fibroblasts or interstitium of alveolar
wall
Monocytes: Only the coated beads were incorporated
by the monocytes. They were found inside and outside
phagosomes and lysosomes of monocytes. PMNs did
not incorporate any beads.
Reference:
Kleinman et al.
(2003, 0535351
Species: Rat
Gender: NR
Strain: F344n-NIA
Age: 22-24 m
03
Route: Nose-only Inhalation
CCL: 03 + Ammonium bisulfate Dose/Concentration: 03: 0.2 ppm
(ABS) + Elemental Carbon (EC)
CCH:03 + ABS+EC
Purified Air (control)
Particle Size: CCL: 0.30 ± 2.5
|jm;CCH: 0.29 + 2.3pm
CCL: 50 pg/m3 EC + 70 pg/m3 ABS + 0.2 ppm 03
CCH: 100 pg/m3 EC + 140 pg/m3 ABS + 0.2 ppm
03
Time to Analysis: 4 h/days, 3 consecutive
days/wk for 4 wk
BALF Cells: CCL and CCH induced macrophage
respiratory burst activity. The effect induced by 03 was
not significant.
BAL Inflammatory/Injury Markers: Total protein,
mucus glycoprotein and albumin were somewhat
elevated in all exposure groups but only reached
statistically significance for CCL and protein (very high
variability). CCL and CCH both depressed Fc receptor
side binding. No effect for 03 was observed.
DMA Replication: 03 caused a slight effect of 20-40%
increase. CCL and CCH caused between 250 - 340%
increase for interstitial and epithelial cells. CCL induced
greater reactions than the high dose.
Reference:
Kleinman and
Phalen (2006,
0885961
Species: Rat
Gender: Male
Strain: SD
Age: 6 wk
Weight: 200 g
L03: Low 03
H03: High 03
LS: Low H2S04
HS: High H2S04
LOLS: Low 03 + Low H2S04
LOHS: Low 03 + High H2S04
HOLS: High 03 + low H2S04
HOHS: High 03+ high H2S04
Particle Size: LS = 0.23 pm +
2 3
HS = 0.28 |jm + 2.1
LOLS = 0.23 pm ±2.3
LOHS = 0.28 pm ± 2.1
HOLS = 0.23 pm ± 2.3
HOHS = 0.28 pm ± 2.1
Route: Nose-only Inhalation
Dose/Concentration: L03 = 0.30 ppm
H03 = 0.61 ppm
LS = 0.48mg/m3
HS=1.00mg/m3
LOLS = 0.31 ppm + 0.41 mg/m3
LOHS = 0.31 ppm +1.04 mg/m3
HOLS = 0.60 ppm + 0.52 mg/m3
HOHS = 0.60 ppm + 0.86 mg/m
Time to Analysis: Exposed for 4 h. Parameters
measured 42 h post-exposure.
Inflammatory Lesions in Lung Parenchyma: Neither
Type 1 or 2 lung lesions were affected by sulfuric acid
alone. H03 doubled Type 1 lesions and increased Type
2 lesions 25-fold. Additions of H2S04 to 03 appeared to
have a dose-dependent protective effect for both types
OT issions.
DMA Synthesis in Nasal, Tracheal and Lung Tissue:
Increased DNA synthesis was observed at all high 03
exposures but was not affected by coexposure to
H2S04.
Macrophage FcR binding: No effects were observed
(no data for L03 and H03).
Macrophage Phagocytosis: All levels of exposure (no
data for L03 and H03) decreased phagocytosis.
December 2009
D-96
-------
Reference
Pollutant
Exposure
Effects
Reference:
Kodavanti et al.
(2005, 0879461
Species: Rat
Gender: Male
Strain: WKY and
SH/NCrlBR
Age: 11-14 wk
CAPs (EPA, NC)
Measured components included
Al, Be, Ba, Co, Cu, Zn, Pb, Mn,
Ni, Ag, Ti, As.
Particle Size: 1 day: 1.07-1.19
|jm;2days: 1.27-1.48pm
Route: Whole-body Inhalation
Dose/Concentration: 1 day study: 1138-1765
pg/m3
2 day study: 144-2758 pg/m3
Time to Analysis: 4 hr (SH only); 4 hr/day, 2 day
(WKY and SH)
Post-exposure: 1 day: 3 h except study #4,18-20
h;2day:18-20h
Breathing Parameters: In a paired analysis of control
SH and treated SH, treated SH showed an increase in
expiratory and inspiratory time due to CAPs. The
treated and control groups of WKY rats did not show
significant differences.
BALF Cells: In the 2 day study, WKY rats showed
decreases in total cells; this decrease was associated
with decreased macrophages. WKY showed an
increase in neutrophils.
BAL Inflammatory/Injury Markers: Total protein and
albumin in WKY rats decreased whereas SH rats
maintained the same approximate level. LDH activity
lowered slightly in both strains.
Cell Membrane Integrity: SH rats showed increased
GGT (membrane bound enzyme) activity and plasma
fibrinogen for 5/7 exposures but these increases did
not appear to be dose-dependent.
Cytokines: Levels were undetermined in SH rats.
WKY showed slight increases in IL-6, TNF-a, and MIP-
2 but these increases were not statistically significant.
Reference: Kooter CAP-F = fine (Site I)
et al. (2006, 0975471 CAP-UF = fine + ultrafine (Site
Species: Rat
Gender: Male
Strain: SH
Age: 12-14 wk
(Netherlands)
Some measured components:
Ammonium, nitrate, sulfate ions:
56+ 16% CAP-F mass, 17 +
6% CAP-UF mass
Particle Size: 0.15
-------
Reference
Pollutant
Exposure
Effects
Reference: Lei et al.
(2004, 0878841
Species: Rat
Gender: Male
Strain: SD
Weight: 300-350 g
CAPs from Asian dust storm
(Taiwan)
Measured Components: Si, Al,
S, Ca, K, Mg, Fe, As, Ni, W, V,
OC, EC, S02, N02, nitrate,
sulfate
Particle Size: 0.01- 2.5pm
Route: Nose-only Inhalation
Dose/Concentration: 315.6 pg/m3 (Low) or 684.5
pg/m3 (High)
Time to Analysis: Low: Exposed for 6 h.
Sacrificed 36 h post-exposure
High: Exposed for 4.5 h. Sacrificed 36 h post-
exposure
Pulmonary hypertension induced 2 wk pre-
exposure.
BALF Cells: PM induced dose-dependent increases in
total cells and percentage of neutrophils. No change in
macrophages, lymphocytes or eosinophils occurred.
Basophils were highly variable.
BALF Inflammatory/Injury Markers: Dose-dependent
increases were observed for total protein and LDH.
Cytokines: IL-6 increased dose-dependently (control:
33.5 ± 7.5, low 165.1 ± 117.2, 273.6 ± 62.8 pg/mL).
Reference: Li et al.
(2007, 1559291
Species: Mouse
Gender: Female
Strain: BALB/c,
C57BL/6
Age: 9 wk
Weight: NR
DEP (2369-cc diesel engine
manufactured by Isuzu Motor,
operated at 1050 rpm, 80% load,
commercial light oil)
Particle Size: NR
Route: Inhalation
Dose/Concentration: DEP: 103.1 + 9.2 pg/m3,
CO: 3.5 + 0.1 ppm, N02: 2.2 + 0.3 ppm, S02:
<0.01 ppm
Airway Hyperresponsiveness: Penh values
increased in BALB/c mice compared to the control at
day 0, but no significant changes occurred after this
time. Penh values increased in C57BL/6 mice at 1 wk
compared to the control but returned to control levels at
Time to Analysis: Protocol 1: Exposed 7h/day, 5 8 wk
days/wk. Sacrificed at day 0, week 1, 4, 8. Protocol BA|_F: Compared to the other strain, the total number
of cells and macrophages increased significantly at 1
wk in C57BL/6 mice and at 8 wk in BALB/c mice.
Neutrophils, lymphocytes, MCP-1, IL-12, IL-10, IL-4, IL-
13 increased significantly for both strains. No
eosinophils were found. IL-1|3and IFN-y increased
significantly in BALB/c mice compared to C57BL/6
2: DE alone or DE+NAC 7h/day, 1-5 days.
HO-1 mRNAand Protein: HO-1 mRNA was more
marked in BALB/c mice at 1 wk and C57BL/6 mice at 4
and 8 wk. HO-1 protein percentage changes from the
control were greater in BALB/c mice at 1 wk and
C57BL/C mice at 8 wk.
MAC: NAC inhibited the increased Penh values, total
number of cells and macrophages in C57BL/6 mice at
1 wk and neutrophils and lymphocytes in both strains.
Reference: Liu et al.
(2008, 1567091
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 11 wk
Weight: NR
DEP (5500-watt single-cylinder
diesel engine generator
(Yanmar, Model YDG 5500E),
406 cc displacement air-cooled
engine, Number 2 Diesel
Certification Fuel, 40 weight
motor oil)
Particle Size: -0.1 pm (MMAD)
Route: Intranasal
Dose/Concentration: Average particle
concentration: 1.28 mg/m
Time to Analysis: Four groups: saline+air control,
saline+DEP, A. fumigatus+air, A.fumigatus+DEP. A.
fumigatus exposure every 4 day for 6 doses. DEP
exposure 5 h/day for 3 wk concurrent with A.
fumigatus exposure.
A.fumigatus+DEP increased IgE, the mean BAL
eosinophil percentage, goblet cell hyperplasia, and
eosinophilic and mononuclear cell inflammatory
infiltrate around the airways and blood vessels
compared to the A. fumigatus or DEP treatments.
A.fumigatus+DEP also caused methylation at the IFN-y
promoter sites CpG-53, CpG-45, and CpG-205.
Reference: Lopes
et al. (2009, 1904301
Species: Mouse
Gender: Male
Strain: BALB/c
Age: 6-8 wk
Weight: NR
PM (high density traffic; winter
2004; Sao Paulo, Brazil) (N02,
CO, S02)
Particle Size: 10 pm (diameter)
Route: Open-Top Exposure Chamber
Dose/Concentration: 33.86 + 2.09 pg/m3
Time to Analysis: Some rats pretreated with
papain. Exposed to UAP or filtered air 24 h/day, 7
days/wk, 2 mo.
The papain+UAP treatment increased Lm values,
collagen fibers, and decreased the density of elastin
fibers over the papain+filtered air treatment. The
papin+UAP treatment increased 8-isoprotane more
than any other group.
December 2009
D-98
-------
Reference
Pollutant
Exposure
Effects
Reference:
Mangum et al.
(2004, 0973261
Species: Rat
Gender: Female
Strain: CDF
(F344)/CrlBR
Age: 7 wk
Ti02 (DuPont)
Particle Size: NR
Route: Whole-body Inhalation
Dose/Concentration: 10, 50 or 250 mg/m3
Time to Analysis: 6 h/day, 5 days/wk, 13 wk.
Parameters measured 0, 4,13, 26, 52 wk post-
exposure.
OPN (osteopontin) Expression: At 0 wk, OPN mRNA
expression exhibited a dose-dependent increase. Low
dose induced a 2-fold increase while the high dose
induced an almost 100 -fold increase. At 4 wk, the mid-
dose and high-dose elevated OPN mRNA levels. At 13
wk, the high dose elevated OPN mRNA levels. No
significant elevation with mid dose level was observed.
At 26 wk, the mid and high dose induced elevated OPN
mRNA levels. At 52 wk, rats in the low, mid and high
dose groups all indicated elevated levels of OPN
mRNA. Specifically, the low, mid and high doses
induced a 3-fold increase, 7-fold increase and 400-fold
increase, respectively.
OPN Protein in BALF: Data was not reported at 0 and
4 wk. At 13 wk, protein increased 9-fold (~800 pg/mL
OPN) at mid dose and 100-fold (-8000 pg/mLOPn) at
high dose. At 26 wk, the mid and high dose groups
remained elevated. At 52 wk, protein increased by 2.5
fold in low dose, 7-fold in mid dose and 166-fold in high
dose group.
Histopathology: At 52 wk, slight OPN
immunoreactivity was observed in control and low dose
group (immunostaining mostly limited to intraalveolar
MACSJ.Trichrome-stained lung sections from control
and low dose groups showed no increase in collagen.
Rats exposed to mid or high dose groups showed
areas of lesions.
Reference: Martin UAP-BA: Urban Air particles
et al. (2007, 0963661 (Buenos Airs, Argentina)
Species: Mouse
Gender: Male
Strain: BALB/c
Age: 1-2 mo
Particle Size: <2.5pm
Route: Intranasal Installation
Dose/Concentration: 0.17 mg/kg
Time to Analysis: 3*day, 3 days/wk, 2 days apart
(1, 4, 7 day). Parameters measured 1 h post-
exposure.
Particle Characteristics: 3 types, ultrafmes <0.2 um
(inorganics ND), bunched agglomerates of ultrafmes
and <40 um with aluminum silicates, ions and trace
metals.
BALF Cells: Increased amount of phagocytes in
alveolar area, reducing airspace percentage (control
52.9% ± 1.39, UAP-BA24.7% ± 2.87). Increased
number of PAS positive cells.
Morphometry: Induced focal inflammatory lesions.
Accumulation of refractile material in upper and lower
respiratory tract. PM in phagocytes of bronchiolar
lumen and alveolar space. No evidence of fibrosis
and/or collagen changes.
Reference: Mauad
et al. (2008, 1567431
Species: Mouse
Gender: Male,
Female
Strain: BALB/c
Age: 10 day
Weight: Parental:
21.4 + 4.0-26.3 +
2.8 g; 15 day-old
offspring: 7.8+ 1.1 -
9.0+ 1.0 g; 90 day-
old offspring: 20.3 +
2.3-27.4+1.8 q
PM (busy traffic street Sao
Paulo, Brazil; Aug. 2005-April
2006) (N02, S02, CO)
Particle Size: 2.5,10pm
(diameter)
Route: Open-Top Chamber
Dose/Concentration: PM25: filtered chamber- 2.9
+ 3.0 pg/m3, nonfiltered chamber-16.9 + 8.3
pg/m3; Outdoor concentration: PM10-36.3+ 15.8
pg/m3, CO- 1.7 + 0.7ppm, NO-89.4+ 31.9 pg/m3,
S02-8.1 + 4.8 pg/m3
Time to Analysis: Nonfiltered exposure 24 h/day
for 4 mo. Mated at 120 days exposure. After birth,
30 females and offspring transferred to filtered or
nonfiltered chamber. Killed 15 or 90 day of age.
Mild foci of macrophage accumulations containing
black dots of carbon pigment occurred in the alveolar
areas on 90 day-old mice. Surface-to-volume ratio
decreased from 15 to 90 days of age and was higher in
mice exposed to air pollution. PM exposure reduced
inspiratory and expiratory volumes at higher levels of
transpulmonary pressure.
Reference:
McDonald et al.
(2004, 0874591
Species: Mouse
Strain: C57BL/6
Age:8-10wk
DEE: high load, No 2, No cat
(620:1 dilution)
DEE-ER (Control): Emissions
Reduced (high load, low sulfur
ECD1) (same dilution)
(Yanmar diesel generator, 406
cc, 5500 watt load)
Particle Size: DEE: 110 nm;
DEE-ER: NR
Route: Whole-body inhalation
Dose/Concentration: DEE PM: 236 pg/m3
DEE-ER PM: 7 pg/m3
Time to Analysis: DEE: 6 h/day for 7 days.
DEE-ER: 6 h/day for 7 days. RSV administered
post-exposure for some: single, 4 days. Those not
infected with RSV sacrificed immediately upon last
exposure.
Differences in Exposure Conditions: CO, PM, EC,
OC, nitrate, alkyne, c2-c212 alkenes, phenanthrenes,
total particle PAHs, total Oxy-PAHs, benzene, pyrene,
benzojayrene, zinc were reduced by 90-100% in the
emissions reduction case. Most other components
were reduced by around 60%.
DEE vs. DEE-ER Effects: DEE increased viral
retention and lung histopathology DEE-ER increases
were not statistically significant.
Cytokines: DEE increased TNF-a, IL-6, IFN-y and HO-
1. DEE-ER responses were not statistically significant
(significantly higher variability in DEE-ER controls vs.
DEE controls).
December 2009
D-99
-------
Reference Pollutant
Reference: DEP: SRM 2975 (NIST)
McQueen etal.
(2007, 0962661 Particle Size: NR
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Weight: 228-500 g
Reference: CP: Carbon particles
Medeiros et al.
(2004, 096012) PSA: ROFA (solld waste
incinerator hospital Sao Paulo,
Species: Mouse Brazil)
Gender: Male PSB: electric precipitator, steel
~ • „„,„, plant, Brazil)
Strain: BALB/c
PSA/PSB Characteristics:
Age: 60 days Generally, PSB had greater
Weight: 20-30 g «^™*™*ff*
(10+x), Mn (2x), Rb (60+x), Se
(7x , Zn (4x). PMA>PMB: Ce
(3x, Co (10+x), La(100x), Sb
(15x),V(50x).
Particle Size: CP: 1.7 + 2.5pm
(78%<2.5|jm);PMA:1.2±2.2
pm(98%<2.5pm);PMB:1.2 +
2.2 pm (98%<2.5 pm)
Reference: Mutlu et PM10
al. (2006, 1559941 Collected by baghouse from
Dusseldorf, Germany
Species: Mouse
Particle Size: NR
Strain: C57BL/6
Age: 6-8 wk
Weight: 20-25 g
Reference: CAPs: produced at Tuxedo, NY
Nadziejko, et al. laboratory using centrifugal
(2002, 0874601 aerosol concentrator
Species: Rat FA: Fine Particle SulfuricAcid
„ ., ., , Aerosol
Gender: Male
UFA: Ultra-Fine Particle Sulfuric
Strain: SH Acid Aerosol
Age: 16 wk Particle Size: CAPs: PM25;
FA: 160 nm; UFA: 50-75 nm
Reference: Nemmar DEP: SRM 2975
et al. (2007, 1568001
Particle Size: <1 pm
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Age: 16 wk
Weight: 424 ± 8g
Exposure
Route: IT Instillation
Dose/Concentration: 0.5 mL/rat of 1 mg/mL; 1-
2.2 mg/kg
Time to Analysis: 6 h.
Pre-exposure: Vagotomy (sectioning of vagus
nerve) or atropine, 1 mg/kg i.p. administered 30
min prior, 2 and 4 h post.
Reference: Intranasal Instillation
Dose/Concentration: CP: 10 pg/mouse; 0.5
mg/kg
PSA: 0.1, 1 or 10 pg/mouse; 0.005, 0.05, 0.5
mg/kg
PSB: 0.1, 1 or 10 pg/mouse; 0.005, 0.05, 0.5
mg/kg
Time to Analysis: Single, 24 h
Route: IT Instillation
Dose/Concentration: 100 ng/mouse; 1 pg/mouse;
10 pg/mouse; 100 pg/mouse
Time to Analysis: 1-7 days
Route: Nose-only Inhalation
Dose/Concentration: CAPS 80, 66 pg/m3; avg 73
pg/m3
FA: 299, 280, 119, 203 pg/m3; avg 225 pg/m3
UFA: 140, 565, 416, 750 pg/m3; avg 468 pg/m3
Time to Analysis: 1 0 exposures of 4 h each, each
exposure at least 1 wk apart.
(2 exposures to CAPs, 4 to FA and 4 to UFA)
Route: Intravenous Injection
Dose/Concentration: 0.02, 0.1 or 0.5 mg/kg
Time to Analysis: single, 24 h
Effects
BALF Cells: A 9-fold increase in neutrophils with high
individual variability in response was observed.
Bilateral vagotomy prior to DEP reduced neutrophil
increase to 3 fold. Vagotomy with saline instillation had
no effect. Atropine reduced neutrophils to levels similar
to saline response. No differences were observed
between DEP response in anesthetized when
compared to conscious animals. Macrophages,
eosinophilsand lymphocytes remain unchanged.
Respiratory Response: RMV increased post DEP.
Vagatomy reduced response by one-third. Atropine pre-
treatment did not have effect.
BALF Cells: No change in BAL cell count was seen.
Quantitative cellular counts increased for perivascular
area for both groups at all dose levels. Inflammatory
cells in alveolar septum area only increased for PSA.
Alveolar Fluid Clearance: At 100 pg/mouse,
decreased clearance peaked at 24 h and recovered at
7 days.
Histology: Evidence of mild lung injury at doses of 100
pg/mouse or more was seen.
BALF Cells: Significant increase in total cell number
was observed. Neutrophils increased but this was not
statistically significant.
Wet/Dry Ratio: Exposure did not induce any effects.
Na, K-ATPase: At 100 pg/mouse, decreased activity of
Na, K-ATPase in basolateral membranes was
observed.
Respiratory Rate: CAPs decreased the respiratory
rate as did FA at all dose levels. However, the FA-
induced respiratory rate was not statistically significant
unless the data was combined. UFA increased this rate
significantly.
BALF Cells: Marked cellular influx at all dose levels
Was observed. Macrophages increased at the high
dose, but this was not statistically significant. PMN
increased significantly at all dose levels.
Wet/Dry Ratio: All dose levels induced increases.
December 2009
D-100
-------
Reference
Pollutant
Exposure
Effects
Reference: Nemmar PS: Polystyrene particles
et al. (2003, 0879311
PSC: Polystyrene particles, Car-
Species: Hamster boxylate modified
Gender: Male and
Female
Weight: 100-110 g
PSA: Polystyrene particles,
Amine modified
Particle Size: PS, PSC, PSA-
60: 60 nm; PSA-400: 400 nm
Route: IT Instillation
Dose/Concentration: 5, 50 or 500 pg/animal;
0.05, 0.5, 5 mg/kg
Time to Analysis: Single, 10 min post-exposure
Rose Bengal administered to induce thrombosis,
immediate study thereafter
BALF Cells: Both PSA-60 and PSA-400 (PSA-
60>PSA-400) induced a massive influx of PMNs. PSA-
60 effect may exhibit some dose-dependency
BALF Inflammatory/Injury Markers: Small increases
in total protein were seen at 500 pg level for both PSA-
60 and PSA-400. LDH was increased at all PSA-60
levels but not for 500 ug PSA-400. Histamine increased
for all PSA-60 levels and PSA-400 but due to high
variability only the effect at 500 pg PSA-60 was
statistically significant.
Reference: Nemmar DEP: SRM 1650
et al. (2003, 0974871
Species: Hamster
Gender: NR
Weight: 100-110 g
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 50 pg/animal
Time to Analysis: Single exposure, parameters
measured 1, 3, 6 or 24 h post- exposure.
BALF Cells: DEP led to a significant PMN flux at 1 h
(13% of total cell number), 6 h (22%) and 24 h (37%).
Histamine: Concentrations in BALF were consistently
elevated starting at 1 h. Plasma histamine did not
increase until 6 h.
Pretreatment with Histamine Receptor Antagonist:
A major decrease in DEP induced PMN infiltration was
seen. No effect on histamine in BALF or plasma was
observed.
Reference: Pereira Ambient Particles
et al. (2007,1560191 (Porto Allegre, Brazil)
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Age: 3 m
Particle Size: <10|jm
Route: Whole-body Inhalation
Dose/Concentration: P-6: 34, 22 or 225 pg/m3
P-20:139or112|jg/m3
P-l:99|jg/m3
Time to Analysis: P-6: single/continuous for 6 h
P-20: single/continuous for 20h
P-l: intermittent (5 h) periods per day for 4 days
consecutively
Parameters measured 0 or 24 h post-exposure
BAL Inflammatory/Injury Markers: An increase in
lipid peroxidation was statistically significant only for
the 20 h continuously exposed group. Leukocytes also
increased at P-20. No change at P-6. Total protein
remained unaffected at all dose levels.
Wet to Dry Ratio (Oh): No effect was observed.
Reference:
Pinkerton et al.
(2004, 0874651
Species: Rat
Gender: Female
(pregnant),
Offspring- NR
Strain: SD
Age: 10 days
(pups), Pregnant
females- 10-1 4 days
of gestation
Weight: NR
Reference:
Pinkerton et al.
(2002, 0876451
Species: Rat
Gender: Male,
Female
PM (Fe and soot from
combustion of acetylene and
ethylene in a laminar diffusion
flame system)
Particle Size: Median diameter:
72-74 nm; size range: 10-50 nm
PM (Fe, Soot) (ethylene, iron
pentacarbonyl, acetylene
combined; Fe203; soot: 60% EC,
40% OC) (CO, NOX)
Particle Size: Fe (diameter) 40
nm; Soot (primary particles,
diameter) 20-40 nm
Route: Inhalation
Dose/Concentration: Mean mass concentration:
243 ± 34 pg/m ; Average Fe concentration: 96
pg/m3
Time to Analysis: Exposed 10 days postnatal
age, 6 h/day, 3 days (consecutive).
Route: Whole-body Inhalation
Dose/Concentration: Adult males: Fe- 57, 90
pg/m3, Soot- 250 pg/m3, Fe+Soot- Fe: 45 pg/m3,
Total PM: 250 pg/m ; Neonates: Fe+Soot- Low:
Fe- 30 pg/m3, Total PM: 250 pg/m3, High: Fe- 100
pg/m3, Total PM: 250 pg/m3
Time tn Analwcic1 AHiilt malpc pYnncprl tn Fp
A significant reduction of cell proliferation occurred only
within the proximal alveolar region of exposed animals
compared to controls. There were no significant
differences between the groups for alveolar formation
and separation within the proximal alveolar region.
Fe: Only the high dose had significant effects. This
dose increased total protein in the lavage fluid,
decreased total antioxidant power, induced GST
activity, and induced a non-significant, increasing trend
of GSH and GSSG. IL-1|3, intracellular ferritin, and NF-
KB increased.
Fe+Soot, Soot: Fe+Soot significantly reduced the total
Strain: SD
Age: 11-13 wk (adult
male), 10-12 days
(neonatal)
Weight: NR
soot, Fe+Soot, or filtered air. Exposed 6 h/d, 3
days (consecutive). BAL, 2 h postexposure, lung
tissue, 24 h postexposure. Neonatal rats exposed
to Fe+Soot 10-12 day-old and 23-25 day-old.
antioxidant power in BALF and supernatant from lung
tissue homogenate. Fe+Soot significantly increased
GSSG, IL-1P, NF-KB, CYP1A1, and CYP2E1. CYP2B1
increased but was not significant. Soot alone was not
significant for anything.
Neonates: The high-dose significantly decreased cell
viability, increased LDH activity, and increased IL-lp
and ferritin. Both doses significantly increased GSSG,
GRR, and GST, and decreased total antioxidant power.
December 2009
D-101
-------
Reference
Pollutant
Exposure
Effects
Reference: Pires-
Neto et al. (2006,
0967341
Species: Mouse
Gender: Male
Strain: Swiss
Age: 6 days
Ambient Air:
PM25, N02andCB
(Sao Paulo, Brazil)
Particle Size: PM2 5
Route: Whole-body Inhalation
Dose/Concentration: PM25: 46.49 ug/m3
Control: 18.62 ug/m3
I\I02' 59 52 ug/m3
Control: 37.08 ug/m3
CB:12.52|jg/m3
Control: 0 ug/m3
Time to Analysis: 24 h/day, 7 days/wk for 5 mi
Nasal Cavity: Increased total mucus and acidic mucus
at proximal and medial areas of cavity. Nonsecretory
epithelium declined. No significant changes in amount
of neutral mucus, volume proportion of neutral mucus,
volume proportion of total mucus, thickness of
epithelium, volume proportion of nonsecretory
epithelium or ratio between neutral and acidic mucus
were observed.
Types of Acidic Mucus Cells: Proximal and medium
D cells increased. Effects on distal cells were equivocal.
(weaned at 21 days into exposure, mothers
removed)
Reference: DEP: generated from idling
Pourazar et al. Volvo diesel engine
(2005, 088305) 3
DEP 300 ug/m comprised of:
N021.6ppm
Species: Human NO 4.5 ppm
CO 7.5 ppm
Gender: Male and Hydrocarbons 4.3 ppm
Female (nonatopic & Formaldehyde 0.26 mg/m3
nonsmokers) Suspended particulates
Age: 21-28 yr 43x1°W
Route: Whole-body Inhalation
Dose/Concentration: DEP 300 pg/m3
Time to Analysis: Single exposure for 1 h.
Parameters measured 6 h post exposure.
Transcription Factors: Exposure induced increased
cytoplasmic and nuclear immunoreactivity of phos-
phorylated p38 MAPK in bronchial epithelium.
Increased nuclear translocation of phosphorylated p38
and JNK, MAPK as well as increased nuclear
phosphorylated tyrosine immunoreactivity were
observed. No change in total or nuclear c-fos
immunoreactivity was seen. Exposure induced
increased nuclear translocation of phosphorylated JNK
significantly associated with phosphorylation of nuclear
c-jun and also resulted in an increase in nuclear p65.
Particle Size: <10|jm
Cytokines: Expression of IL-8 was positively
associated with nuclear phosphorylated p38 post-
exposure.
Reference: Pradhan RSPM: Respirable Suspended
et al. (2005, 0961281 PM
(Lucknow, India)
Species: Rat
Quartz dust (positive control)
Gender: Female
Particle Size: < 5 urn
Strain: Wstar Albino
Weight: 120-180 g
Route: IT Instillation
Dose/Concentration: 2.5, 5.0, or 10.0 mg/0.05
ml; 20, 42, 83 mg/kg
Time to Analysis: 15 days.
Relative Lung Weight: A dose-dependent increase in
total lung weight of RSPM-instilled animals was
observed.
BALF Cells: Exposure induced a dose-dependent
increase in total cells dose-dependent with the low and
mid dose levels. PMNs increased massively at all dose
levels with RSPM inducing less of an increase than
Quartz. Exposure at low dose levels resulted in an
influx of inflammatory cells (predominantly
macrophages into lumen of alveolar ducts and alveoli).
Reaction at the high dose was more intense than that
seen in mid dose-exposed lungs.
BAL Inflammatory/Injury Markers: A significant dose-
dependent increase in LDH and NO was observed, but
the Quartz-induced increase was greater than the
RSPM-induced increase. An increase in protein was
significant at the mid dose level for RSPM and
significant at the high dose level for both RSPM and
Quartz.
Lung Biochemistry: An increase in lipid peroxidation
was dose-dependent. Superoxide dismutase (SOD)
enzyme levels showed a dose-dependent decrease.
Reference: Ramos WS (Pine wood) (00(<80ppm), Route: Whole-body Inhalation
et al. (2009, 1901161 C02 (0.35%), 02 (20.1 %), PM2 5,
. . „ . PM10)
Species: Guinea
Pig Particle Size: PM2 5, PM10
Dose/Concentration: WS: 60 g, PM25:363 ± 23
ug/m , PM10: 502 + 34 ug/m
Gender: NR
Strain: NR
Age: NR
Weight: 330-370 g
Time to Analysis: Exposed 3 h, 5 days/wk for 1,
2, 3, 4, 6, 7 mo.
WS significantly decreased body weight between 4 and
7 m exposure. The concentration of blood
carboxyhemoglobin increased. Recovered BALF cells
were higher in WS-exposed pigs. Macrophages and
neutrophils increased. Inflammation in the lungs was
seen. Pulmonary arterial hypertension and
emphysematous lesions were observed. Macrophage
and lung tissue homogenate elastolysis increased.
Collagenolysis increased. Generally, MMP-2, MMP-9,
and MMP-1 increased. BAL macrophage apoptosis
increased with time.
December 2009
D-102
-------
Reference
Pollutant
Exposure
Effects
Reference: Rao et
al. (2005, 0957561
Species: Rat
Strain: SD
Weight: 175 g
DEP: SRM 2975 Route: IT Instillation
Particle Size: 0.5 pm Dose/Concentration: 5, 35, 50 mg/kg bw
Time to Analysis: Sacrificed 1, 7, 30 days post
single exposure. Cytokines measured after 24 h
incubation (in vitro).
BALF Cells: Macrophages unaffected. Increased
PMNs at 1 day for all dose levels, sustained elevation
at 7days for mid and high dose and at 30 days for all
dose levels.
BAL Inflammatory/Injury Markers: Increased albumin
at 1 and 30 days at all dose levels. Increased LDH
except at low dose at 7 days.
Cytokines: The high dose induced a significant
increase of mRNA expression for IL-113, iNOS, MCP-1,
and MIP-2 in BAL cells. MCP-1 mRNA sustained high
levels at 7 days for mid and high dose and at 30 days
for all dose levels. mRNA expression of IL-6, IL-10,
TGF-|31, TNF-awere unaffected. However, IL-6 and
MCP-1 proteins increased significantly in BALF at 1
day for mid and high dose, returning to basal levels at 7
days. MIP-2 increased for all dose levels at all time
points. NO level unaffected.
Reference: Reed et
al. (2006, 1560431
Species: Rat,
Mouse
Gender: Male and
Female
Strain: CDF
(F344)/CrlBR (rat),
SH (rat), A/J
(mouse), and
C57BL/6 (mouse)
Aae: 6-1 2 wk
HWS (burned mix of hardwood
in noncertified wood stove using
a Pineridge model 27000,
Heating and Energy Systems,
Inc. Clackamas, OR)
Measured Components: EC,
OM, N03, S04, NH4, metals
Particle Size: -0.25 pm
Route: Whole-body Inhalation
Dose/Concentration: Low: 30 pg/m
Mid-low: 100|jg/m3
Mid-high: 300 pg/m3
High: 1000 pg/m3
Time to Analysis: 6 hr/day, 7 days/wk for 1wk or 6
mo. Immediate post-exposure analysis.
Organ Weights: Liver declined in rats of both genders
at 1 wkand female rats at 6 m. Lung volume increased
and lung weight decreased in female rats at 6 m.
Spleen weight increased in female mice and rats at 1
wk. Thymus weight decreased in male rats at 1 wk.
Cells: Eosinophils decreased and lymphocytes
increased in males at 6m. Neutrophils decreased at 6m
in both genders. Minimal increases in alveolar
macrophages and sparse brown-appearing
macrophages in all species.
Bacterial Clearance: Mice instilled with bacteria were
mostly unaffected by exposure, except for a decline in
histopathology summary score after 6m.
Tumorigenesis: No values for exposed groups differed
significantly from controls. There was no evidence of
progressive exposure related trend.
Reference: Reed et DE: generated from two 2000 Route: Whole-body Inhalation
al. (2004, 055625) model 5.9 L Cummins ISM turbo 3
diesel engines Dose/Concentration: Low: 30 pg/m
Species: Rat, ,
Mouse Co-exposure to 8 gas and 8 Mid-low: 1 00 pg/m
solid exhaust components . ., . .... ,nn , ,,3
Gender: Male and measured Mid-high. 300 pg/m
Female uiinkv innn,,n;m3
Organ Weights: Kidney weight increased after 6m for
both males and female rats at the high dose. Kidney
and liver weight increased for female mice at all dose
levels at 6 mo. Lung weight increased at high dose at
6m for female mice and male rats. Spleen weight
decreased in male mice at the low and mid-high levels.
Cells: Minimal increases in alveolar macroohaaes and
Particle Size: 0.10-0. 15pm
(F344)/CrlBR (rat),
A/J (mouse)
Age: 12 wk
Time to Analysis: 6 h/day, 7 days/wk for 1wk or 6
mo. Analyzed 1 day post-exposure.
PM within the macrophages were seen.
Cytokines: TNF-a decreased in female rats after 6m.
Tumorigenesis: No significant effect was observed.
Reference: Reed et
al. (2008, 1569031
Species: Mouse
Gender: Male,
Female
Strain: C57BL/6,
A/J, BALB/c
Age: NR
Weight: NR
GEE (2 1996 General Motors
4.3-LV-6 engines; unleaded
gasoline)
Particle Size: 150 nm(MMAD)
Route: Whole-body Inhalation
Dose/Concentration: Control: 2.5 ± 2.9, Low-
exposure: 6.6 + 3.7, Mid-exposure: 30.3 + 11.8,
High-exposure: 59.1 + 28.3, High filtered exposure:
2.3±2.6|jg/m3
Time to Analysis: Exposed 6 h/day, 7 days/wk, 3
days-6 mo.
Body and Organ Weight and Histopathology in A/J:
Kidney weight decreased, but no effects pertaining to
weight were significant. No visible inflammatory
changes were seen.
Lung Damage in A/J: No significant effect was seen,
but hypomethylation was seen in females at 1wk, and
methyiation was reduced in all exposed female groups.
Bacteria in Lungs of C67BL/6: Exposure did not
affect the clearance of bacteria from the lung.
Respiratory Allergic Response in BALB/c: Exposure
had little effect, but serum total IgE increased
significantly for the high-exposure group. Increasing
trends were seen in OVA-specific serum IgE and lgG1,
as well as neutrophils and eosinophils.
December 2009
D-103
-------
Reference
Pollutant
Exposure
Effects
Reference: Reed et
al. (2008, 1569031
Species: Mouse
Gender: Male,
Female (only
BALB/c)
Strain: C57BL/6,
A/J, BALB/c
Age: NR
Weight: NR
GEE (two 1996 General Motors
4.3-LV-6 engines; regular,
unleaded, non-oxygenated, non-
reformulated gasoline blended to
US average consumption for
summer 2001 and winter 2001-
2002- Chevron-Phillips)
Particle Size: 150 nm(MMAD)
Route: Whole-body Inhalation
Dose/Concentration: PM: Low- 6.6 + 3.7 pg/m3,
Medium- 30.3 ± 11.8 pg/m3, High- 59.1 ± 28.3
pg/m3
Time to Analysis: A/J - exposed 6 h/days, 7
days/wk, 3 days-6 mo. C57BL/6- 1wk exposure.
Instillation of P. aeruginosa. Killed 18 h
postinstillation. BALB/c- Conditioned to exposure
chambers and mated. Pregnant females exposed
GD 1 and throughout gestation. Offspring
exposures continued until 4 wk-old. Half of
offspring sensitized to OVA. Tested for airway
reactivity by methacholine challenge 48 h post-
instillation and euthanized.
The kidney weight of female A/J mice decreased at 6m
and was strongly related to PM by the removal of
emission PM. PM-containing macrophages increased
by 6 mo. Hypomethylation occurred in females at 1 wk.
The clearance of P. aeruginosa was unaffected by
exposure. Serum total IgE significantly and dose-
dependently increased. OVA-specific IgE and lgG1
gave slight exposure-related evidence but were not
significant.
Reference: Reed et
al. (2008, 1569031
Species: Rat
Gender: Male,
Female
Strain: CDF
(F344)/CrlBR, SHR
Age: NR
Weight: NR
GEE (two 1996 General Motors
4.3-LV-6 engines; regular,
unleaded, non-oxygenated, non-
reformulated Chevron-Phillips
gasoline, U.S. average
consumption for summer 2001
and winter 2001-2002)
Particle Size: 150 nm(MMAD)
Route: Whole-body Inhalation
Dose/Concentration: PM: Low- 6.6 ± 3.7 pg/m3,
Medium- 30.3 ± 11.8 pg/m3, High- 59.1 ± 28.3
pg/m3
Time to Analysis: 6 h/day, 7 days/wk, 3 days-6
mo.
Organ Weight: At 6 mo. exposure, the heart weights of
male and female rats increased and male rats' seminal
vesicle weight decreased.
Histopathology: PM-containing macrophages
increased by 6 mo.
Lung DMA Damage: Hypermethylation occurred in
medium- and high-exposure male rats at 6 mo.
BAL: For both genders in the high-exposure group,
LDH and MIP-2 significantly increased at 6 mo. ROS
decreased at Iwkand 6 mo. Generally, the production
of hydrogen peroxide and superoxide decreased in the
high-exposure group and medium- and high-exposure
groups, respectively.
Removal of Emission PM: The removal of emission
PM strongly linked PM to increased seminal vesicle
weight, red blood cell counts, LDH, lipid peroxides, and
methylation.
Reference:
Rengasamy et al.
(2003, 1569071
Species: Rat
Gender: Male
Strain: SD
Weight: -200 g
DEP:SRM1650
CBEIftex-12 furnace black,
Cabot, Boston, MA
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 5,15, or 35 pg/kg
Time to Analysis: Single; 1, 3, 5, 7 days post
exposure
CYP1A1: DEP at all doses significantly increased
CYP1A1 protein, was maximal at 1 day, and
normalized at 5 days. CB had no effect.
CYP2B1: DEP and CB at 15 and 35 mg/kg inhibited
activity at 1 day
Protein level significantly decreased at 1 day with 5,15
and 35 mg/kg DEP and at 15 and 35 mg/kg CB. A time
dependent decrease was shown at 35 mg/kg for both
DEP and CB.
Reference: Renwick
et al. (2004, 0560671
Species: Rat
Gender: Male
Strain: Wstar Kyoto
Weight: 370-470 g
FCB: Fine Carbon Black (Huber
990)
UCB: Ultrafme Carbon Black
(Printex 90, Degussa)
FTO: Fine Titanium Dioxide
(Tioxide)
UTO: Ultrafine Titanium dioxide
(Degussa)
Particle Size: FCB: 260 nm;
UCB:14nm;FTO:250nm;
UTO: 29 nm
Route: IT Instillation
Dose/Concentration: 125 or 500 pg/rat
Time to Analysis: Single, 24 h
BALF Cells: UTO and UCB induced a large dose-
dependent increase in percent neutrophils (only
statistically significant at 500 pg for UTO).
BAL Inflammatory/Injury Markers: UTO and UCB
also increased total protein content only at the 500 pg
dose. UCB induced LDH release at 125 and 500 pg,
UTO and CB at 500 pg. UTO and UCB induced large
dose-dependent increases in GGT activity (only
statistically significant at 500 pg for UTO).
Phagocytosis: All 4 particles decreased but only at the
500 pg level.
Chemotaxis: Only UTO and UCB at 500 pg/l
increased chemotactic migration.
December 2009
D-104
-------
Reference
Pollutant
Exposure
Effects
Reference: Rhoden CAPs
et al. (2004, 0879691 (Boston, MA)
Species: Rat
Gender: Male
Strain: SD
Weight: 250-300 g
Particle Size: CAPS: 0.1-2.5 pm
Route: Whole-body Inhalation
Dose/Concentration: 1060 + 300 pg/m3
Time to Analysis: Single exposure for 5 h.
Analyzed 24 h post-exposure.
(CAPS-MAC = CAPS with 50 mg/kg bw MAS (N-
acetylcysteine) pretreatment)
Particle Characteristics: Major components did not
appear to show any correlation to total particle mass.
Included Na, Mg, Al, Si, S, Cl, K, Ca, Ti, V, Cr, Mn, Fe,
Ni, Cu, Zn, Br, Ba, Pb. Metals Al, Si and Fe (somewhat
less for Pb, Cu, K) correlated with TBARS.
BALF Cells: CAPS increased PMN 4 fold. MAS
treatment reduced this increase to control levels.
BAL Inflammatory/Injury Markers: LDH and total
protein not affected. Histology confirms slight
inflammation with CAPS and no inflammation with
CAPs-NAC.
Oxidative Stress: CAPS increased TBARS and
oxidized protein by 2+ fold. MAS fully prevented the
increase in TBARS and partially prevented an increase
in protein carbonyl.
Tissue Damage: Wet/dry ratio increased with CAPS
but significantly decreased with MAC.
Reference: Rhoden
et al. (2008, 1904751
Species: Rat
Gender: Male
Strain: SD
Age: NR
Weight: 300 g
Urban Air Particles (UAP) (SRM
1649)
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 1mg in 100 |JL saline
Time to Analysis: Instilled with UAP. CL
analysis:15 min post-exposure. BAL
measurements: 4 h post-exposure.
Some rats pre-treated with MnTBAP 2 h prior to
UAP exposure.
UAP significantly increased the total cell number, PMN,
MPO activity, and protein levels. MnTBAP prevented
UAP-induced lung inflammation. UAP increased
oxidants in lung CL, which was prevented by MnTBAP.
Reference: Rivero
et al. (2005, 0886531
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Age: 3 mo.
Weight: 250 g
Ambient air (Sao Paulo, Brazil)
Particle Size: <2.5pm
Route: IT Instillation
Dose/Concentration: 100 or 500 pg/rat; 0.4 or 2
mg/kg
Time to Analysis: Single, 24 h
Histopathology: At both doses, acute alveolar
inflammation was observed and was more pronounced
in the 500 pg group.
Lung Morphometry: Lumen wall ratio values show a
dose-dependent increase in peribronchial as well as
intra-acinar pulmonary arterioles. No effect in
myocardial arterioles were observed.
Tissue Damage: Lung wet/dry ratios were unaffected.
Reference: Roberts
et al. (2004, 1989031
Species: Rat
Gender: Male
Strain: SD
Age: 60-90 days
Weight: 300-350 g
ROFA: SRI (cyclone power
plant)
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 0.5 mg/rat; 1.67 mg/kg
Time to Analysis: Single, 6 and 24 h
Technology: Laser capture microdissection of airway
cells were used to analyze results.
Protein: pERK1/2: ERK1/2 ratio increased by 60% at
6 h and 80% at 24 h. NF-KB activity increased at 6 h
but was not statistically significant.
Reference: Saber et
al. (2005, 0978651
Species: Mouse
Gender: Female
Strain: TNF(-/-) (B6,
129S-Tnftm1Gk1),
C57/BL
Age:9-10wk
DEP: SRM 2975
CB: Printex 90 (Degussa)
Particle Size: DEP: 215 nm;
CB: 90 nm
Route: Nose-only Inhalation
Dose/Concentration: DEP: 20 mg/m3; CB: 20
mg/m3
Time to Analysis: 90 min/day for 4 days
consecutively, 1 h
BALF Cells: Neutrophils increased significantly to 15%
when compared to control (4%) with DEP exposure. No
response difference was observed between TNF (+/+)
and TNF(-/-). CB did not induce any changes in
neutrophil numbers.
Cytokines: IL-6 increased 2-3 fold in DEP and CB
exposure in both normal and knockout mice. IL-1|3was
unaffected.
mRNA: In TNF (+/+) mice, DEP and CB increased
expression of TNF mRNA 2-fold. IL-6 mRNA
expression was high in DEP-exposed knockout mice
when compared to normal mice.
DMA: DNA strand breaks increased in both strains.
Knockout mice showed a higher response to CB and
DEP exposure. For normal mice, only CB induced a
statistically significant effect.
December 2009
D-105
-------
Reference
Pollutant
Exposure
Effects
Reference: Schins
etal. (2004,
0541731
Species: Rat
Gender: Female
Strain: Wistar Kyoto
Weight: 350-550 g
Soluble fractions
PMC:PM10.25
PMF: PM25
-B: Borken, Germany (rural)
-D: Duisburg, Germany (industri-
alized)
Particle Size: PMio-2.5, PM2.5
Route: IT Instillation
Dose/Concentration: 0.32 + 0.01 mg/rat; 0.91+
0.58 mg/kg
Time to Analysis: Single, 18 h
BALF cells: Both PMC showed a massive increase in
neutrophils. PMC-B induced the greatest increase
followed by PMC-D. Both PMF did not induce a
significant increase.
BAL Inflammatory/Injury Markers: PMC from both
sites induced markedly higher endotoxin concentration
vs PMF as follows in decreasing order: PMC-B, PMC-
D, PMF-B, PMF-D, control. Glutathione decreased only
for PMC-B. LDH and total protein were unaffected.
Cytokines: TNF-a and IL-8 increased with PMC from
both sites. PMF induced a slight increase in IL-8 but did
not induce an increase in TNF-a.
Radical Formation: Formation of hydroxyl radicals
increased with exposure. Relative intensity was: PMC-
D, PMF-D, PMC-B, PMF-B, and control.
Reference:
Seagrave etal.
(2005, 0880001
Species: Rat
Gender: Male
Strain: F344/DCrl
BR
Age: 11 +1 wk
PM from 3 sources:
Route: IT Instillation
NT: New Technology bus, Detroit Dose/Concentration: 0.25-2.2 mg/rat in 0.5ml
Diesel 50G, exhaust oxidation saline
catalyst, 216 miles, 2002 model -
jn use Time to Analysis: Single, 24 h
NE: Normal emitter bus, Detroit
Diesel 50G, no catalyst, 134259
miles, 1997 model - in use
HE: High Emitter bus, Cummins
L10G, no catalyst, >250, 000
miles, 1992, retired
Fuel composition very similar for
3 vehicles: methane (96-96.8%),
ethane (1.6-1.9%), carbon
dioxide (0.9-1.1%), nitrogen (0.6-
0.8%), traces of other gases
Particle Size: NR
Engine Specific Emission data: HE had significantly
higher PM and SVOC recovered emission rates than
NEandNT.
Organic mass in PM: The following PM sources are
listed in decreasing order of percent of total mass: HE,
NE, NT.
Total PAH: The following PM sources are listed in
decreasing order of total mass: HE, NT, Control, NE.
Nitro PAH: The following PM sources are also listed in
decreasing order of total mass: NE, HE, Control, NT.
Authors note confounding technical issues (mostly
technique related) with mostly mild effects.
BAL Inflammatory/Injury Markers: LDH showed
dose-dependent increases with HE inducing higher
increases than NT and NE. Total protein exhibited
dose-dependent increases with HE, NT and the
positive control SRM2975 inducing higher levels than
NE.
Potency Factors Cytotoxicity and Inflammation: HE
was significantly more potent than NT and NE, with NT
also showing significant potency.
Lung Toxicity: The results were highly variable but the
general toxicity levels in increasing order is the
following: NE, NT, HE, Normal gasoline, diesels, and
high gasolines, though individual factors may differ
greatly.
Reference:
Seagrave et al.
(2006, 0912911
Species: Rat
Gender: Male
Strain: F344/Crl BR,
Age: 11 +1 wk
PM25 sources: BHM: Birming-
ham, Alabama; urban
JST: Jefferson Street, Atlanta,
Georgia; urban
PNS: Pensacola, Florida; urban/
residential
CTR: Centreville, Alabama; rural
"smoke" = downwind of forest
fires/burns (NR)
Particle Size: PM2 5
Route: IT Instillation
Dose/Concentration: 0.75,1.5, 3 mg/rat
Time to Analysis: Single, 24 h
BALF PMN: In general, the winter samples induced
greater increases in potency than the summer samples
except for PNS. For the winter samples, the samples
that induced the greatest increases, in descending
order, are: JST, BHM, CTR, PNS and Smoke. For the
summer, the samples that induced increases, in
descending order, are: BHM, JST, PNS, and CTR.
BALF Macrophages: For the winter, the BHM and JST
samples significantly increased potency whereas the
PNS sample induced significantly negative potency.
For the summer, only the BHM sample significantly
induced potency.
BALF Lymphocytes: Only the JST-Wand BHM-W
significantly increased potency. The BHM-S, CTR-S
and PNS-S also significantly increased potency.
Histopathology: All the winter and summer samples,
excepting PNS, significantly induced inflammation.
Lung weight/body Weight Ratio: In general, for all
end points, JST-S was significantly less potent than
JST-W. The summer samples of BHM and CTR were
also generally more potent than their winter
counterparts.
December 2009
D-106
-------
Reference
Reference:
Seagrave et al.
(2005, 0880001
Species: Rat
Gender: Male,
Female
Strain: CDF(F-
344)/CrlBR
Age: 10-12 wk
Reference:
Seagrave et al.
(2008, 1919901
Species: Rat
Gender: Male
Strain: SD
Pollutant
DE:
(Two 6 cyl Cummins ISB turbO)
HWS = hardwood smoke
(mixed black/white oak,
uncertified conventional wood
stove)
DE:
,
EC = 557 pg/m
OC = 269 pg/m3
NO = 45 ppm
N02 = 4 ppm
CO = 30 ppm
THV = 2 ppm
HWS:
EC = 43 pg/m3
OC = 908 pg/m3
NO or N02 = 0 ppm
00= 13 ppm
THV = 3 ppm
Particle Size: DE: 0.14 ±1.8
pm; HWS: 0.36 ±2.1 pm
GEE (2 1996 General Motors
4.3-L V6 gasoline engines;
conventional Chevron Phillips
gasoline, U.S. average
composition) (CO, NO, N02,
S02, THC) (PM2 5 composition-
EC, OC, S04, NH4, N03)
Exposure
Route: Whole-body Inhalation
Dose/Concentration: 30, 100, 300, 1000 pg/m3
TPM
Time to Analysis: 6 h/day, 7 days/wk for 6 mo. 1
day post-exposure
Route: Nose-only Inhalation
Dose/Concentration: GEE: 60 pg/m3, SDCAs:
317-1072 pg/m3, RD: 306-954 pg/m3; GEE: 00-
104 ppm, NO- 16.7 ppm, N02- 1.1 ppm, S02-
LOppm, THC- 12 ppm; SDCAs: CO- <1 ppm, NO-
0.19-0.62 ppm, N02- 0.10-0.37 ppm, S02- 0.07-
0.24 ppm, THC- <1 ppm
Effects
Particle Characteristics: Major differences K:
HWS»DE; Ca DE»HWS; Zn: DE»HWS.
BALF Cells: No effects were observed except for an
increase in macrophages at 30 pg/m3 for HWS males
exposed to HWS.
Cytokines: IL-1 13 was unaffected by DE or HWS. MIP-
2 decreased for both genders at 1000 HWS. TNF-a
decreased in females with DE exposure. No TNF-a
effects for HWS were observed.
BAL Inflammatory/Injury Markers: LDH was
unaffected by DE. Exposure to HWS induced an
increase for males only at 100 and 300 but not at 1000
pg/m3. Protein was unaffected by DE. HWS exposure
showed male-only effects at 100 and 300 pg/m3 but not
at 1000. AP was unaffected by DE or HWS except for
slight decline induced by HWS at 1000 pg/m3 for both
genders.
Other: |3-glucose was unaffected by DE. HWS-
exposed females showed decreased (3-glucose at 100
and 300 but not at 1000 pg/m3.
BALF GSH to (GSH+GSSG): No effects for DE were
observed. HWS significantly decreased the ratio in both
males and females at 1000 pg/m3. The effect for
females was greater than the male effect.
GEE produced CL in the lungs, heart, and liver. RD
produced a significant effect in the heart at the low
dose. SDCAs had no effect on CL. GEE did not affect
the amount of macrophages or PMN. SDCAs increased
macrophages. The RD low dose increased
macrophages and PMN. SDCAs increased Penh
values and tidal volumes.
Age: 10-12 wk
Weight: 250-300 g
Simulated downwind coal
emission atmospheres (SDCAs]
(fly ash, gas-phase pollutants,
sulfate aerosols, NO, N02, S02)
Paved Road Dust (RD) (Los
Angeles, CA; New York City, NY;
Atlanta, GA)
Particle Size: GEE: MMAD-150
nm, RD: 2.6 ± 1.7pm, SDCA:
0.1-1.Opm
Time to Analysis: 6 h exposure, immediately
post-exposure
Reference: Singh et A-DEP (4cyl light duty 2.7I Isuzu Route: Oropharyngeal Aspiration
al. (2004, 0874721 diesel at 6 kg/m)
Dose/Concentration: 25 or 100 pg/mouse
DEP: SRM 2975 T L „ , . . , t ^
Species- Mouse Tin* to Analysis: single, 4 h
Particle Size: A-DEP >50 pm
Gender: Female (18 h post-exposure measurements taken but NR
due to similar results)
Strain: CD-1
Age: 6-8 wk
Particle Characteristics: DEP had 60% EC vs 9% in
A-DEP. A-DEP had 50% OC vs 5% in DEP.
Phenanthrene and Fluoranthene fractions were much
more prevalent in PAH from DEP than A-DEP.
BALF Cells: PMNs significantly increased dose-
dependently with DEP and remained elevated at 18h.
Endotoxin induced the greatest increases of PMNs.
Macrophages increased with A-DEP and were
unaffected by DEP.
Cytokines: Endotoxin induced massive responses for
IL-6, MIP-2 and TNF-a but no response from IL-5. A-
DEP increased all 4 cytokines but only at the 100 pg
dose level. Similarly, DEP only increased IL-6 at the
100 pg dose level.
BAL Inflammatory/Injury Markers: Microalbumin
increased for both pollutants except DEP induced
increases only at 100 pg. Endotoxin increased microal-
bumin. NAG increased with 100 pg A-DEP.
December 2009
D-107
-------
Reference
Pollutant
Exposure
Effects
Reference: Smith et CAPs
al. (2003, 0421071
Species: Rat
Gender: Male
Strain: SD
Age: 11-12 wk
(Fresno, CA)
Particle Size: <2. 5pm
Route: Whole-body Inhalation
Dose/Concentration: 6 exp in 2 sets of 3:
Fall! = 847 pg/m3
Fall2 = 260 pg/m3
FallS = 369 pg/m3
Winterl = 815 pg/m3
Wmter2 = 190 pg/m3
Winters = 371 pg/m3
Time to Analysis: 4 h/days for 3 consecutive
days. Parameters measured immediately following
last exposure.
Particle Characteristics: Nitrate showed the highest
variability near 10 fold, followed by Si, S and EC. OC
concentration was relatively consistent. Metals
otherwise appeared proportionate to the
concentrations.
BALF Cells: Total cells increased at wkl Percent of
macrophages reduced in wk2 with CAPs. Number of
neutrophils increased with CAPs, but only achieved
statistical significance during wk1 of the fall and winter.
Lymphocytes increased but were not statistically
significant.
BAL cell permeability: Upon CAPs exposure, the
proportion of nonviable cells were increased up to
242% when compared to controls. The fall of wk2
induced the highest significant increases followed by
fall wkl fall wk3, and winter wk3.
Reference: Smith et
al. (2006, 1108641
Species: Rat
Gender: Male
Strain: SD
Age: 8 wk
Weight: 260-270 g
CFA: Coal Fly Ash (400 MW,
Wasatch Plateau, Utah)
(aerodynamic separation)
Particle Size: 0.4-2.5 pm
Route: Nose-only Inhalation
Dose/Concentration: 1400 pg/m3 PM25 including
600 pg/m3 PM,
Time to Analysis: 4 h/days for 3 consecutive
days. Parameters measured 18 or 36 h post-
exposure.
BALF Cells: Percent and total number of neutrophils in
BALF and blood increased significantly at both 18 and
36 h. Percent of macrophages decreased slightly while
number of macrophages increased in bronchiole-
alveolar duct regions at both time periods.
Cytokines: MIP-2 and transferrin increased at 18 h. IL-
1|3 increased at 36 h.
Other: Gamma glutamyl transferase decreased at
36 h. Lung antioxidant increased at 18 h.
Reference: Song et
al. (2008, 1560931
Species: Mouse,
Gender: Female
Strain: BALB/c
Age: 5-6 wk
DEP collected from a 4JB1-type,
light-duty (2740 cc), four-cylinder
diesel engine operated using
standard diesel fuel at speeds of
1500 rpm under a load of 10
torque.
Particle Size: 0.4 pm (mean
diameter)
Route: Intranasal Instillation (days 1-5), Whole-
body Inhalation (days 6-8)
Dose/Concentration: 0.6 mg/mL in 50 pL of saline
(days 1-5), 6mg/m3 for 1 h/day for 3 days (days 6-
Time to Analysis: Enhanced Pause (Penh),
measured on day 9. BAL and lung tissues
collected on day 10.
Airway Hyperresponsiveness: Intranasal exposure
plus aerosolized DEP caused a significant increase in
methacholine-induced Penh over the control.
BAL Analysis: There was no significant increase in
IFN-y in the BAL fluid following DEP treatment but
there was a significant increase in IL-4 levels compared
to the control. (IL-4 increase could indicate that DEP
modulates Th-2 cytokines in the mouse model). DEP
also induced an increase in total neutrophils and
lymphocytes in the BAL when compared to the control.
The nitrite concentration in BAL (indicating NO
generation) was significantly greater in the DEP
exposed group than the control.
Histology: Peribronchial and perivascular infiltrates
were more common in the group exposed to DEP than
the control.
Ym1 and Ym2 Expression: (see explanation in
comments section) Ym1 and Ym2 transcripts were
upregulated in response to DEP exposure in mice.
Reference:
Steerenberg et al.
(2006, 0882491
Species: Rat
Strain: Crl/WKY
Age: 6-8 wk
Ambient air samples
PMC, PMF:
-1: Rome, Italy
-N: Oslo, Norway
-PL: Lodz, Poland
-NL: Amsterdam, Netherlands
Measured Components: Li, Be,
Route: IT Instillation
Dose/Concentration: 1 and 2.5 mg/animal
Time to Analysis: Single, 24 h
Particle Characteristics: Concentrations of metals
were highest in Rome. Amsterdam was noted for high
Mg and V. Lodz was noted for high Pb, Zn, PAH. More
of PMC was composed of Fe, Mn, Al, Cr, Cu. More of
PMF, on the other hand, was composed of Zn, Pb, Ni,
y
BALF Cells: PMNs increased.
B, Na, Mg, Al, K, Ca, Ti, V, Cr,
Mn, Fe, Co, Ni, Cu, Zn, As, Se,
Sr, Mo, Cd, Sn, Sb, Ba, Ce, Nd,
Sm.Au, Hg, TI, Pb, Bi, U, Si,
Endotoxins, Cl, NO-, S04
Particle Size: PMC: 2.35-8.5
pm; PMF: 0.12-2.35 pm
Cytokines: MIP-2 increased dose-dependently. TNF-a
also increased.
BAL Inflammatory/Injury Markers: CC16 decreased
substantially. Crustal material (endotoxin, Na, Cl and
metals but not Ti, As, Cd, Zn, V, Ni, Se) was positively
associated with short term CC16. Albumin increased.
December 2009
D-108
-------
Reference
Pollutant
Exposure
Effects
Reference: Stinn et
al. (2005, 0883071
Species: Rat
Gender: Male and
Female
Strain: Crl: (WIU BR
Age: 40 days
DE (generated from 1. 6 LVW
diesel under USFTP 72)
CO: 10, 37ppm
C02: 2170, 6540 ppm
NO: 7.0, 22.8 ppm
NOX: 8.6, 28.3 ppm
S02: 0.83, 3.09 ppm
NH4: ND
Measured Major Components:
NO, S02, 1-nitropyrene, Zi. 50%
by DE weight is EC.
Route: Nose-only Inhalation
Dose/Concentration: 3 and 10 mg/m3
Time to Analysis: 6 h/day, 7 days/wk for 24 mo; 6
mo. post-exposure
Body Weight: Mean weight increased substantially
during the first few weeks in all groups. Food
consumption decreased in 1-24 mo but was recovered
in 24-30 mo. Body weight decreased at 23 mo in all
categories, but recovered except in high dose males at
30 mo.
Organ Weight: Absolute weight of lungs, larynx and
trachea increased from 0 to 12 to 24 mo and stayed
elevated at 30 mo: Low
-------
Reference
Reference:
Sureshkumar et al.
(2005, 0883061
Species: Mouse
Gender: Male
Strain: Swiss
An-, -in 19 wV
MQd IU~ I Z Wl\
Weight: 20-25 g
Pollutant
GE: Gasoline Exhaust
(Honda generator EBK 1200,
four stroke one cyl)
Including: S02 = 0.11 mg/m3
NOX = 0.49 mg/m3
C0= 18.7 ppm
Particle Size: GE
>4|jm = 34.1%
3-4 pm = 15.8%
2-3 pm = 15.8%
1.5-2|jm = 10.6%
0.5-1. 5pm = 5.3%
<0.5pm = 18.4%
Exposure
Route: Nose-only Inhalation
Dose/Concentration: 0.635 mg/m3
Time to Analysis: 15min/day7, 14 or 21 days; ;
<1 h post-exposure
Effects
BALF Cells: Neutrophils (%) increased at 7, 14 and 21
days (stable). Total cell count, macrophages and
eosinophils were unaffected. Leukocytes and
lymphocytes increased, though not significantly.
Cytokines: GE caused time-dependent increases in
TNF-aand IL-6. IL-10and IL-1|3 were unaffected.
BAL Inflammatory/Injury Markers: y-GGT, ALP and
LDH increased after 2 wk of GE exposure and stayed
stable at 21 days. Total protein slightly increased on 14
and 21 days, though these increases were not
statistically significant.
Histopathology: Minor changes at 7 days, mild edema
in alveolar region at 14 days and sloughing of epithelial
cells in bronchiolar region and focal accumulation of
inflammatory cells in alveolar region at 21 days were
observed in a time-dependent manner.
Reference:
Tesfaigzi et al.
(2002, 0255751
Species: Rat
Gender: NR
Strain: Brown-
Norway
Age: 7-8 wk
Weight: 310-330 g
WS (wood stove- Vogelzang
Boxwood Stove, Model BX-42E,
wood-Pinusedulis)(CO, NO,
NOX, total hydrocarbon)
Particle Size: Smaller size
fraction: 0.405-0.496 pm, larger
size fraction: 6.7-11.7 pm
Route: Whole-body Inhalation
Dose/Concentration: Target concentration (low,
high exposure): 1,10 mg/m3; CO-15-106.4 ppm,
NO- 2.2-18.9 ppm, NOX- 2.4-19.7 ppm, total
hydrocarbon-3.5-13.8 ppm
Time to Analysis: 3 h/day, 5 days/wk, 4 or 12 wk.
Respiratory Function: Total pulmonary resistance
increased for exposure groups and was significant for
the low-exposure group. In exposed groups, forced
expiratory flows and quasistatic compliance were lower
and dynamic lung compliance higher, the latter being
significant for the high-exposure group. For the high-
exposure group, vital capacity slightly decreased,
residual volume slightly increased, and CO-diffusing
capacity had a slight, significant decrease.
BALF Cells: Macrophages decreased significantly in
the high-exposure group. Particle-laden macrophages
increased with concentration. Lymphocytes and
neutrophils slightly increased in the high-exposure
group.
Cytokines: LDH increased slightly and protein levels
decreased slightly in the high-exposure group.
Cytokines were below detectable levels.
Histopathology: WS caused minimal to mild chronic
inflammation in the epiglottis of the larynx. PAS-positive
cells increased in the 30 day high-exposure group. AMs
increased with time and concentration. Particle-laden
macrophages were seen after 90 days. AB- and PAS-
positive epithelial cells increased for the 90 day low
exposure group.
Reference: Tin-Tin-
Wn-Shwe et al.
(2006, 0884151
Species: Mouse
Gender: Male
Strain: BALB/c
Age: 7wk
CB14: Printex 90 (Degussa) Route: IT Instillation
CB90: Flammruss101
(Degussa)
Particle Size: CB14:14 nm
CB95: 95 nm
Dose/Concentration: 25, 125, 625 pg/mouse;
approx. 1, 5, 25 mg/kg
Time to Analysis: 1/wk for 4wk. 4 h post-
exposure
Body Weight, Thymus, Spleen, Splenic Cell Count:
No effects were observed.
BALF Cells: Increased total cell numbers were
observed for 125, 625 pg CB14 (dose-dependent) and
625 pg CB95. Total cell count was twice as high for
CB14 at 125 and 625 pg compared to CB95. AM
numbers exhibited a dose-dependent response for both
CB14 and CB95 for all doses except 125 pg. Lympho-
cyte numbers increased at 125 and 625 pg for CB14
and 625 pg for CB95. PMN numbers increased at 125
and 625 pg for CB14 and CB95, but the response was
greater with CB14. PMN numbers were proportional to
dose surface area for both PM sizes.
BAL Cytokines: CB14 and CB95 induced dose-
dependent increases in IL-1J3. TNF-a increased at 125
and 625 pg dose in CB14 with the 125 dose inducing a
slightly greater increase. CB14 and CB95 induced
CCL-3 increases 125 and 625 pg.
Chemokine mRNA in lung and lymph nodes: CCL-3
mRNA increased for CB14 but not CB95 4 h following
the last exposure. CCL-2 was unchanged.
Mediastinal lymph nodes: The number of CB-laden
phagocytes increased in a dose-dependent manner for
CB14 and CB95. CB14 had higher numbers at all
doses compared to CB95.
December 2009
D-110
-------
Reference
Pollutant
Exposure
Effects
Reference: long et
al. (2006, 0976991
Species: Mouse
Gender: Male
Strain: KP600 CD-1
Weight: 22-26 g
PM25 (collected from stacked
filter air sampler in Shanghai,
China)
Fe: FeS04
Zn: ZnS04
PMF:PM25+FeS04
PMFZ: PM25 + FeS04 + ZnS04
Major Measured Components:
Fe 26 ppm, Zn 9 ppm, S 61 ppm
Particle Size: PM2 5
Route: IT Instillation
Dose/Concentration: PM: 25 mg/mL, 1mg/mouse
Fe: 15mg/mL, 0.6 mg/mouse
Zn: 15mg/mL, 0.6 mg/mouse
PMF: PM 25 mg/mL + Fe 15 mg/mL,
1.6 mg/mouse
PMFZ: PM 25 mg/mL + Fe 15 mg/mL,
1.6 mg/mouse
Time to Analysis: Instilled twice at 0 and 24 h.
Parameters measured 24 h following last exposure
(at 48 h).
Synchrotron X-ray imaging: PMFZ showed the
greatest increase in alveolar changes. Fe induced
more hemorrhagic changes, whereas Zn induced more
nonuniformity of lung texture. This suggests that Zn
induces PBMC in a dose-dependent manner which
releases IL-1, IL-6, TNF-a, and IFN-y.
Histopathology: PMFZ induced the most severe
changes including serious inflammation/pus in bronchia
and bronchial epidermal cell hyperplasia. For Fe or
PMF hemorrhagic changes predominated but were less
severe than PMFZ.
Reference:
Upadhyay et al.
(2008, 1593451
Species: Rat
Gender: Male
Strain: SH
Age: 6 mo.
Weight: NR
Reference:
Wallenborn et al.
(2007, 1561441
Species: Rat
Gender: Male
Strain: WKY SH
and stroke-prone SH
(SHRSP)
Age: 12-15 wk
Ultrafme Carbon Particles
(UFCP)
Particle Size: Size- 31 + 0.3 nm,
MMAD- 46 nm, Surface area
concentration- 0.1 39 m
particles/m3, Mass specific
surface area- 807 m /g
PM: precipitator unit power plant
residual oil combustion
Particle Size: PM: 3.76 pm
(bulk) ±2. 15
Route: Whole-body Inhalation
Dose/Concentration: 172 pg/m3
Time to Analysis: 24 h exposure. 4 days
recovery. Sacrificed 1st or 3rd day of recovery.
Route: IT Instillation
Dose/Concentration: WKY vs SHRSP: 1.11, 3.33,
8.33 mg/kg
SH vs SHRSP: 3.33, 8.33 mg/kg
Time to Analysis: Single, 24 h
Note: 4 h post-exposure study done on WKY vs
SHRSP but not published.
Pulmonary Inflammation: UFCP did not cause
pulmonary inflammation.
Pulmonary and Cardiac Tissue: HO-1, ET-1, ETA,
ETB, TF, PAI-1 significantly increased in the lung on
the 3rd recovery day HO-1 was repressed in the heart,
but the other markers had slight, nonsignificant
increases.
BALF Cells: A dose-dependent increase in total cells
and neutrophils was observed. Equal response for all 3
strains except for SH, for both concentrations was
observed.
BAL inflammation/Injury Markers: LDH exhibited a
dose-dependent increase in equal response for all 3
strains. WKY had higher baseline levels of NAG activity
but, upon PM exposure, SHRSP induced higher
increases than WKY. GGT exhibited a dose-dependent
response for all 3 strains. SHRSP showed the highest
increase followed by WKY and SH. Protein levels
increased at the high dose level with SHRSP exhibiting
the highest increases followed by SH and WKY.
Albumin levels were inconsistent between experiments.
Oxidative Stress - Lung: (WKY vs SHRSP only): SOD
decreased following increased exposure levels with
SHRSP levels generally higher than WKY. Ferritin
levels declined only in SHRSP.
GPx: No action but SHRSP levels were similar to SHR
and, in the WKY vs SHRSP experiment, SHRSP
exhibited higher activity level than WKY.
Ferritin: Equivocal results were observed. Levels
decreased at the high dose for WKY and SHRSP but
increased at medium doses for SH and SHRSP.
ICDH: Levels increased for WKY and decreased for
SHRSP.
Reference: Zinc Sulfate (ZnS04,
Wallenborn et al. aerosolized)
(2008, 191171)
Particle Size: NR
Species: Rat
Gender: Male
Strain: Wstar Kyoto
Age: 13 wk
Weight: NR
Route: Nose-only Inhalation
Dose/Concentration: 9.0 + 2.1 pg/m3, 35 + 8.1
pg/m3, 123.2 ± 29.6 pg/m3
Time to Analysis: Exposed 5 h/days, 3 days/wk,
16wk. Half of the rats used for plasma/serum
analysis, other half for isolation of cardiac
mitochondria.
A trend toward increased BALF protein was seen. No
pulmonary-related effects were seen.
December 2009
D-111
-------
Reference
Reference:
Wegesser and Last
(2008, 1905061
Species: Mouse
fipnripr1 Mfllp
tjciiuci. iviaic
Strain: BALB/c
Age:8-10wk
Reference:
Whitekus etal.
(2002, 1571421
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 6-8 wk
Weight: NR
Pollutant
Ambient PM2.5-io
Collected from San Joaquin
Valley, CA
Particle Size: PM10.2 5
DEP (light-duty, four-cylinder
engine- 4JB1 type, Isuzu
Automobile, Japan; standard
diesel fuel) (extracts)
Particle Size: 0. 5-4 pm
Exposure
Route: IT Instillation
Dose/Concentration: 25-50 pg/mouse
Time to Analysis: 3, 6, 18, 24, 48, 72 h post IT
instiiistion.
Route: Inhalation
Dose/Concentration: 200, 600, 2000 pg/m3
Time to Analysis: Exposed 1 h/day,10 days.
Animals receiving OVA had 20 min OVA exposure
after DEP exposure.
Effects
BALF Cells: Increased amount of viable cells found in
PM-exposed mice with dose-response relationship
between dose of PM and number of total cells
recovered in BALF. At 6 h, increased numbers of
macrophages at both 25 and 50 pg/mouse. Increased
percentage of neutrophils observed with 50 pg/mouse
PM only. Furthermore, both macrophages and
neutrophils increased with longer time period from
instillation, peaking at 24 h. At 50 pg/mouse, MIP-2
concentrations increased, peaking at 3 h, though not
statistically significant and returned to basal levels by
6 h. Positive correlation observed between MIP-2
concentration and increased neutrophil counts. No
correlation found between MIP-2 and macrophages.
DEP+OVA dose-dependently increased IgEand lgG1,
being more effective than the OVA-alone treatment.
This effect was significantly suppressed by thiol
antioxidants MAC or BUG. DEP+OVA increased
carbonyl protein and lipid peroxide over OVA. MAC or
BUG suppressed lipid peroxide and protein oxidation.
No general markers for inflammation were observed.
Reference: Wichers PM (HP-12): inside wall of stack
et al. (2004, 0556361 of Boston, MA power plant
burning #6 oil.
Species: Rat
Particle Size: PM: 3.76 urn +
Gender: Male 2 15
Strain: SH
Age: 75 days
Route: IT Instillation
Dose/Concentration: 0.83, 3.33 or 8.33 mg/kg
Time to Analysis: single, 6 h for Whole-body
plethysmographs (WBP) and repeated daily for 4-7
days,
96 or 192 h post-exposure
non-WBP animals: single,
24,96,192 h post-exposure
Tidal Volume: A dose-dependent decrease in tidal
volume (45 % at high dose) was sustained for 1 day
with very slow recovery over 7 days.
Breathing Frequency: Dose-dependent increase (100
% at high dose) with recovery at 7 days was observed.
Minute Ventilation: Small dose-dependent increases
were observed with a return to normal ventilation in 2
days.
Penh (enhanced pause): Equivocal results in all
groups were observed (due to major control variation).
BALF Cells: Dose-dependent increases in total cells at
24 h, with declined, but still elevated, levels at 192 h.
Neutrophils increased significantly (10 fold) at 24 h in
the mid and high dose groups and showed declined,
but still elevated, levels at 192 h. Macrophages slowly
increased in a dose-dependent manner at 192 h.
BAL Inflammatory/Injury Markers: Protein and
albumin increased at 24 h, returned to relative basal
level at 192 h at the mid and high dose levels. NAG
exhibited dose-dependent increases at 24 h and
sustained these levels through 192 h.
Reference: Wichers PM (HP-12): inside wall of stack
et al. (2006,1038061 of Boston, MA power plant
burning #6 oil.
Species: Rat
Gender: Male
Strain: SH
Age: 71-73 days
Weight: 255-278 g
Particle Size: 1.95pm ±3.49
Route: Whole-body Inhalation
Dose/Concentration: 13 mg/m3
Time to Analysis: Phase 1:1st day, filtered air,
2nd day, 6 h of PM
Phase II: 1st day filtered air, 4 days of 6 h PM each
Immediate post-exposure
Body/ Lung Weight: No effects on Phase I rats were
observed. HP-12 exposure increased body weight, left
lung, right intercostal, and right diaphragmatic lobes in
Phase II rats. However, results appeared due to normal
growth in juvenile rats over 4 days.
Lung lobe to Body Weight Ratio: No effects at 1 or 4
days were observed.
Deposition calculations: V and Co were used to
estimate deposition rates (good correlation between
two metals at R2 = 0.94). Total HP-12 deposition using
Co was 26 and 99 pg (for 1 day and 4 day
experiments) and using V was 31 and 116 pg.
Modeling information estimated HP-12 deposition at
43% in conducting airways and 57% in alveolar region.
Breathing parameters: No changes were observed for
1 or 4 days studies except for a possible decrease in
frequency for the 1 day study.
December 2009
D-112
-------
Reference
Pollutant
Exposure
Effects
Reference: Witten DEP (heavy-duty Cummins N14 Route: Nose-only Inhalation
et al. (2005, 0874851 research engine operated at
75% throttle)
Species: Rat
Gender: Female
Strain: F344
Age: 8 wk
Weight: ~175 g
Dose/Concentration: Low- 35.3 ± 4.9 pg/m3,
Particle Size: 7.234-294.27 nm
High-632.9 ±47.61
Time to Analysis: Exposed 4 h/day, 5 days/wk,
3 wk. Pretreated with saline or capsaicin.
There were no differences for substance P. The low-
exposure group had significantly less NK1. DEP
reduced NEP activity. Plasma extraversion dose-
dependently increased and was greatest in capsaicin
animals. Respiratory permeability dose-dependently
increased. IL-113 was significantly higher for the low-
exposure group. IL-12 was significantly lower in the
capsaicin high-exposure group. TNF-a increased in the
high-exposure group and capsaicin low-exposure
group. High exposure induced particle-laden AMs in the
lungs, perivascular cuffing consisting of mononuclear
cells, alveolar edema and increased mast cell number.
Neutrophil and eosinophil influx was not seen.
Reference: Wong et
al. (2003, 0977071
Species: Rat
Gender: Female
Strain: F344/NH
Age: ~4 wk
Weight: ~175 g
DEP (Cummins N14 research Route: Nose-only Inhalation
engine at 75% throttle) (EC- ,
34 93-601 67 ug/m3 OC-1 90- Dose/Concentration: Low- 35.3 ± 4.9 pg/m ,
11.25 pg/m3, Sulfates 0.94-17.96 High- 669.3 ± 47.6 pg/m
ng/m', Fe- 3.17-6.44, Cr- 0.68-
1.31 ng/m3, Mn-0.11-0.22 ng/m3,
Pb-0.97-1.24 ng/m3)
Particle Size: 7.5-294.3 nm
3 wk. Pretreated with saline or capsaicin.
DEP dose-dependently increased plasma extraversion,
which was further increased by capsaicin. In the high-
exposure group, particle-laden AMs (which were
reduced by capsaicin), inflammatory cell margination,
perivascular cuffing with subsequent mononuclear cell
migration and dispersal, increased mast cells, and
decreased substance P were all seen. NK-1R was
downregulated in the low-exposure group and
upregulated in the capsaicin-pretreated high-exposure
group. NEP decreased significantly for both groups.
Reference: Wu et
al. (2003, 1997491
Species: Rat
Gender: Male
Strain: SD
Age: 60 days
Zn2*
Particle Size: NA
Route: IT Instillation
Dose/Concentration: 50 pm/rat
Time to Analysis: Single, 24 h
Cells: Decreased number of airway epithelial cells
shown with PTEN protein immunostaining.
Macrophages were unaffected.
Reference:
Yamamoto et al.
(2006, 0966711
Species: Mouse
Gender: Male
Strain: BALB/c
Age: 7 wk
Weight: 23 g
CB14: Printex 90 (Degussa) Route: IT Instillation
CB95:Flammruss101
(Degussa)
LTA: Lipoteichoic acid
14CLCB14+LTA
95CL: CB95 + LTA
Dose/Concentration: CB14: 0, 25,125, 625
pg/mouse
CB95: 0, 25,125, 625 pg/mouse
LTA: 10or50|jg/mouse
14CL: 125 pg CB14 + 10 or 50 pg LTA
CB14 measured Components: C 95CL: 125 pg CB95 + 10 or 50 pg LTA
96.79%, HR 0.19%, NO.13%, S
0.11%, Ash 0.05%, 0 2.74% Time to Analysis: Single, 4 and 24 h
CB95 measured Components: C
97.98%, HR 0.15%, N 0.28%, S
0.46%, Ash 0%, 01.14%
Particle Size: CB14:14 nm;
CB95: 90 nm
BALF Cells: CB95 induced dose-dependent increases
of PMN. CB14 induced an increase in PMNs but the
increases were not dose-dependent. LTA massively
increased PMN. LTA induced dose-dependent
increases in total cells, especially at high dose at 24 h.
LTA had massive synergistic effect with CB14 and
CB95 for total cells and PMNs. Total cell count and
PMN levels were highest in 14CL with levels at 24 h
higher than at 4 h. Macrophage data were inconsistent.
Cytokines: CB95 induced dose-dependent increases
in IL-6, TNF-a, CCL2 and CCL3. CB14 induced dose-
dependent increase in CCL2 and CCL3. Exposure
induced increases of IL-6 at the high dose only. Slight
effect on TNF-a was observed. LTA induced dose-
dependent increases of IL-6, TNF-a and CCL3.14CL
massively induced IL-6 and CCL2. No combination of
CB and LTA affected TNF-a or CCL3.
mRNA Expression: LTA, 14CL and 95CL increased
TLR2 mRNA expression with 95CL and 14CL inducing
higher increases than LTA. No effect on TLR4 mRNA
expression was observed.
December 2009
D-113
-------
Reference
Pollutant
Exposure
Effects
Reference:
Yanagisawa et al.
(2003, 0874871
Species: Mouse
Gender: Male
Strain: ICR
Age: 6 wk
Weight: 29-33 g
DEP:
(4JB1 light duty 4cyc 2, 74 liter
Isuzu engine)
LPS
DEP-OC: organic compounds
DL: DEP + IPS
DDL: DEP-OC + IPS
Particle Size: 04 |jm
Route: IT Instillation
Dose/Concentration: DEP/DEP-OC: 125
pg/mouse
IPS: 75 pg/mouse
Time to Analysis: Single, 24 h
BALF Cells: DEP and DEP-OC increased neutrophils
but the increases were not statistically significant. IPS
increased neutrophils significantly. DL and DDL
massively increased neutrophils at greater levels than
IPS alone. Macrophages were unaffected.
Cytokines: IPS increased IL-lp, MIP-1a, MCP-1 and
KC. DEP and DEP-OC had no effect. DL induced
further increases. DDL decreased cytokines compared
to LPS alone. DEP-OC increased IL-lp and MIP-1a
mRNA expression slightly. DEP had no effect. LPS
significantly increased IL-1|3and MIP-1a mRNA
expression. DL increased expressions while DOL did
not.
Pulmonary Edema: LPS, DEP and DEP-OC increased
edema. DL further increased this effect. DOL had no
effect compared to LPS alone.
Histology: DL elevated neutrophil inflammation
interstitial edema and alveolar hemorrhages. DOL
induced neutrophilic inflammation without the alveolar
hemorrhages.
mRNA Expression of TLRs: DEP-OC, DL, DOL and
LPS increased TLR2. DEP had no effect. All particles
increased TLR4 mRNA expression.
Reference:
Yokohira et al.
(2007, 0979761
Species: Rat
Gender: Male
Strain: F344/DuCrj
Age: 10 wk
DQ-12: Quartz dust (Douche
Montan)
HT: Hydrotalcite (Kyoward 500,
PL-1686, KYOWA)
POP: Potassium Octatitanate
fiber (TISMO, Otsuka)
PdO: Palladium Oxide
CB: Carbon Black (Mitsubishi
Kasei)
Particle Size: DQ12 <7 pm
HT7.8 ± 1.5pm
POP: <50 urn length; <2 pm
width
PdO: 0.54 ±1.11 pm
CB: 28 nm
Route: IT Instillation
Dose/Concentration: 4 mg/rat in 0.2 ml saline
Time to Analysis: Single, 1 and 28 days
Lung Weight/Body Weight Ratio: DQ-12, HT and
POP induced increases after 1 day. After 28 days, all
samples induced increases in lung weight.
BALF Cells: Neutrophils increased significantly in
walls and alveolar spaces in all groups on 1 day except
at HT. At 28 days, this increase was maintained only in
walls with severe and moderate elevations, except for
DQ-12.
Histopathology: DQ-12 caused pulmonary edema
both at 1 and 28 days. PdO and CB induced edema at
28 days. Fibrosis was observed after 28 days with the
most significant increase, in decreasing order, induced
by DQ-12, PdO, POP, HT, CB, and the control.
Histiocyte infiltration was observed after 1 day for DQ-
12, POP and PdO. At 28 days, infiltration was observed
for DQ-12, HT, POP and PdO. Restructuring of alveolar
walls and microgranulation was observed for all 5
particles but only at 28 days with DQ 12, PdO, HT,
POP, CB and control.
Immunohistochemistry: BrdU: At 1 day all 5 particles
elevated in both area and number. Activity declined
after 28 days but was still higher than the control.
iNOS: At 1 day DQ-12, POP and PdO induced
increases. At 28 days, DQ-12 and HT induced in-
creases.
MMP-3: DQ-12 induced increases at both 1 and 28
days and PdO at 28 days.
Toxicity scoring: The levels of toxicity are, in
decreasing order, as follows: DQ-12, HT/PdO/POF, and
CB.
December 2009
D-114
-------
Reference
Pollutant
Exposure
Effects
Reference: Zhao et DEP: SRM 2975
al. (2006, 1009961 DEPE: SRM 1975
Species: Rat
Strain: SD
Age: NR
Weight: 200 g
Particle Size: NR
Route: IT Instillation
Dose/Concentration: 35 mg/kg
Time to Analysis: Single, 1 day
AG group coexposed 30 pre and 3, 6, 9 h post
DEP/DEPE
iNOS Expression in AMs: Both DEP and DEPE
increased 12 and 6 fold respectively. NO and
peroxynitrite levels increased accordingly. AG had no
effect on iNOS expression but AG attenuated NO for
both DEP and DEPE but peroxynitrite only for DEPE.
DEP induced much higher levels of oxidants than
DEPE. Unlike DEPE, DEP was unaffected byAG
Role of iNOS in Lung Injury: DEP and DEPE induced
inflammation (PMN), cellular toxicity (LDH) and lung
injury (protein). AG significantly attenuated the DEPE
response but no effect was observed on the DEP
responses.
Cytokines: IL-12 levels were induced by both DEPE
and DEP, with DEPE inducing higher increases than
DEP, and both were significantly attenuated byAG
DEP and DEPE induced similar increases in IL-10
levels. AG increased DEP effect 3 fold and attenuated
DEPE to control.
CYP Enzymes: DEP and DEPE induced increases in
CYP1A1 level and activity. AG attenuated CYP1A1
activity for both DEP and DEPE. CYP2B1 level and
activity were slightly decreased by DEP and DEPE. AG
had no effect.
Cytosol Phase II Enzymes: DEPE had no effect; AG
treatment increased catalase activity. DEP reduced
catalase and GST activities. AG had no effect. Neither
DEP, DEPE nor AG affected OR quinone reductase.
Reference: Zhou et UFe: Ultrafme Fe particles
al. (2003, 0879401 „ ^. , „. ,„
Particle Size: 72 nm
Species: Rat
Gender: Male
Strain: SD
Age: 10-12 wk
Route: Whole-body Inhalation
Dose/Concentration: 57 or 90 pg/m3
BALF Cells: No significant changes observed in total
cell number, cell viability or cell differentials.
Cytokines: Only at the high dose was an increase in
Time to Analysis: 6 h/days for 3 days, parameters |L-1|3 observed. No effect on TNF-a or NF-KB-DNA
measured within 2 h post-exposure. bindjng actjvity was observed.
BAL Inflammatory/Injury Markers: At the high dose,
total protein increased. No significant changes were
observed in LDH.
Intracellular Ferritin: The high dose induced
increases. No significant differences were observed
between the low dose and control.
Oxidative stress: Antioxidant level by FRAP value
decreased at the high dose. GST (glutathione-S-
tranferase) activity increased at the high dose. No
effect on intracellular GSH and GSSG (glutathione
disulfide) was observed.
December 2009
D-115
-------
Table D-4. Effects related to immunity and allergy.
Study
Pollutant
Exposure
Effects
Reference:
Apicella et al.
(2006, 0965861
Species: Mouse
Strain: BALB/c
Cell Line: 112D5
hybridoma
Primary
Macrophages:
Peritoneal
Poly OVA (Ovalbumin on polystyrene
beads) Soluble OVA
Particle Size: NR
Route: Cell Culture
Dose/Concentration: PolyOVA and Soluble OVA:
0.2,1.0or5.0|jg/mL
Time to Analysis: 48 h
IL-6: Stimulation with PolyOVA higher than
stimulation with soluble OVA
TNF-a: Stimulation with PolyOVA higher than
stimulation with soluble OVA.
IL-10: No modifications in levels after PolyOVA
or soluble OVA stimulation.
Viability of Peritoneal Macrophages:
Stimulation with PolyOVA led to 33% decrease
in viability. Stimulation with soluble OVA led to
24% in viability.
Effects of PolyOVA Stimulated Macrophages:
Culture supernatants from PolyOVA stimulated
macrophages had a percentage increase of
asymmetric IgG; however, the addition of rmlL-6
at identical concentrations did not induce a
significant increase. It also decreased the
proliferation of 112D5 hybridoma.
Reference:
Arantes-Costa et
al. (2008, 1871371
Species: Mouse
Gender: Male
Strain: BALB/c
Age: 6 wk
Weight: NR
ROFA (solid waste incinerator powered
by combustible oil; Sao Paulo, Brazil)
Particle Size: NR
Route: Intranasal Instillation
Dose/Concentration: 60 pg ROFA in 50 pL saline
Time to Analysis: OVA sensitized days 1 and 14.
OVA-challenged days 22, 24, 26, and 28. ROFA
exposed 1-3 h after OVA challenge or saline.
Pulmonary responsiveness measured day 30 then
sacrificed. Lungs removed, fixed for 48 h.
ROFA increased pulmonary responsiveness and
decreased ciliated cells in nonsensitized mice,
which were both further amplified in the
presence of OVA. ROFA did not affect
eosinophils, macrophages, chronic
inflammation, or neutral or acidic mucus.
Reference: Archer PM = SRM 1648 (NIST)
et al. (2004,
0880971
Species: Mouse
Strain: BALB/c
D011.10+/+
transgenic - ova
specific receptor
for OVA peptide
323-339
Age: 4 wk
Ti02
Particle Size:
SRM1648:avg1.4|jm
Ti02: avg 0.3 pg (sic)
Route: Intranasal instillation
Dose/Concentration: 500 pg/30 pi sterile saline,
initial 0-750 pg range finding
Time to Analysis: Ova challenge at 68 h, Meth-
acholine aerosolization/AR at 72 h
Airway responsiveness (WBP): AR induced by
Ova/Mch challenge was significantly and dose-
dependently increased at doses of
SRM1648 >500 pg . Ti02/0va exposure was not
significantly different from saline. PM associated
endotoxin did not contribute to enhanced AR.
Lung inflammation/pathology: No increases
in BAL macrophages or eosinophils and no
histological alterations after PM exposure. Both
Ti02 and PM increased pulmonary neutrophils,
indicating particles alone were responsible for
this increase and that the inflammatory
response could occur independently of AR.
Reference: Barrett HWS (black/white oak)
et al. (2006,
1556771
Species: Mouse
Gender: Male
Strain: BALB/c
Age:8-10wk
CO
Total Vapor Hydrocarbon (TVH)
Particle Size: 0.25 ±3.3, 0.35 + 2.5,
0.35 ± 2.0, 0.36 ± 2.1 fjm (MMAD+GSD)
Route: Whole-body Inhalation
Dose/Concentration: HWS: 30, 100 300,
1000|jg/m3
00:0.7,1.6,4.0, 13 ppm
TVH: 0.3, 0.6, 1.3,3.1 ppm
Time to Analysis: Pretreatment: ip 10 pg OVA and
2 mg aluminum hydroxide post-OVA. OVA aerosol
challenge on day 14, followed by 3 days of HWS.
Pre-OVA received aerosol OVA challenge on day
14, then 3 days of HWS on days 26-28 and an
immediate (second) OVA challenge HWS 6 h/day
for 3 days. Sacrificed 18 h post-exposure.
Allergic Inflammation: A statistically significant
increase in eosinophils was observed at
300 pg/m3 HWS following OVA challenge as
compared to OVA alone. No changes in
macrophages, neurophils and lymphocytes were
observed. Post-OVA HWS did not significantly
alter BAL cytokine or serum antibody levels, but
linear trend analyses indicated decreases in IL-
2, IL-4, and IFN-v in the absence of OVA, as
well as a statistically significant upward trend in
OVA-specific IgE when HWS exposure followed
OVA challenge. HWS exposure pre-OVA (prior
to second OVA challenge) resulted in a
decrease in IL-13 (statistically significant at the
high dose but no evidence of an exposure-
dependent response), an increase in OVA lgG1
(trend significant) and no change in IL-2, IL-4,
IL-5, IFN-v, OVA IgE, total IgE or OVA lgG2a.
December 2009
D-116
-------
Study
Pollutant
Exposure
Effects
Reference:
Burchiel et al.
(2005, 088090)
Species: Mouse
Gender: Female
Strain: A/J
Age: 12-14 wk
HWS (black/white oak)
HWS particle Mass
BC
OC
CO
Total Vapor Hydrocarbon
29 other minor components PAH and
metals
Particle Size: 0.3 + 3. 0.4 +2.0.4 + 2.
Route: Inhalation
Dose/Concentration: HWS: 30, 100, 300,
1000|jg/m3
BC:3, 12, 25, 43pg/m3
CC40, 107, 281,908|jg/m3
CO: 1,2, 4, 13ppm
TVH:ND, 1, 1,3ppm
Time to Analysis: 6 h/day for 6 mo.
Proliferative Responses: HWS increased
splenic T cell proliferation at 100 pg/m3 with a
dose dependent decrease at 300 and
1000 pg/m exposures (p<0.05) HWS exposure
did not affect T (CDS), helper! cell (Th, CD4),
cytotoxic T cell (CTL, CDS), macrophage (Mac-
1), natural killer cell (NK, CD1 6) cell markers or
B cell proliferative response to IPS.
0.4 + 2|jm(MMAD + GSD)
Reference:
Burchiel et al.
(2004, 0555571
Species: Mouse
Strain: AJ
Age: 10-12 wk
DE generated alternatively from two
2000 Cummins ISB Turbo Diesel 5.9 L
engines using no 2 (chevron) oil and
15w/40 oil (RotellaT, Shell) run ac-
cording to USEPA Dynamometer
Schedule for Heavy-Duty Diesel Engines
18 PAHs quantified at exposure levels
(text mentions 65)
Particle Size: NR
Route: Inhalation
Dose/Concentration: 30,100, 300,1000 mg/m3
diesel PM
Time to Analysis: 6 h/day, 7 days/wk for 6 mo.
Proliferative Responses: DE depressed
splenic T cell proliferation at all exposure levels
but was not dose-dependent and most
pronounced at the 30 pg/m3 level. (p<0.05 at all
levels) Splenic B cell proliferation was increased
at the 30 pg/m3 level, but not at the other
exposure levels. Little, if any, PAH was found in
DE, and the majority of PAH tested in vitro
enhanced T cell proliferation (below), so PAH is
likely not responsible for the
immunosuppressive effect of DE on murine
spleen cell responses.
Reference: Chan
et al. (2006,
0974681
Species:
Mouse
Strain: D011.10,
BALB/c, Nrf2J-
Cell Types:
Primary bone
marrow dendritic
cells and dendritic
cell line (BC1),T
cells (BMDC)
DEP: DE particles
DEP methanol extract:
Particle Size: NR
Route: Cell Culture
Dose/Concentration: DEP: 10 pg/mL
LPS: 5 ng/mL
Time to Analysis: 24 h
Dendritic Cell Maturation: Organic DEP
chemicals interfered in the expression of several
DC maturation markers. Both DEP and DEP
extracts were found to inhibit CD86 expression
and IL-12 production in LPS-exposed DCs, and
intact particles were not as effective as DEP
extract. DEP extract treatment of BC1 cells
reduced their ability to stimulate co-cultured
antigen-specific T cells, leading to decreased
IFN- y and increased IL-10 without affecting IL-4
or IL-13. DEP extract also induced oxidative
stress and interfered with DC activation by
several other Toll-like receptor agonists as well
as the NF-kB cascade. Inhibition of IL-12
production by DEP extract was shown to be
mediated by pro-oxidative chemicals that
engage the Nrf2 pathway. Taken together the
inhibition of both IL-12 and IFN-y indicates a
suppression of the Th1 pathway and provides a
novel explanation for the adjuvant effect of
DEPs on allergic inflammation.
December 2009
D-117
-------
Study
Pollutant
Exposure
Effects
Reference:
Ciencewicki et al.
(2007, 0965571
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 10-12 wk
Weight: 17-20 g
DE: generated from a 30-kW(40 hp), 4-
cylinder Deutz BF4M1008 diesel engine
Influenza A/Bangkok/1/79 (H3N2
serotype) from Dr. Melinda Beck of the
University of North Carolina, Chapel Hill
02, CO, N02, NO, S02
02: 20.9-20.5% (Lo, Hi)
CO: 0.9-5.4 ppm
N02: 0.25-1.13 ppm
NO: 2.5-10.8 ppm
S02: 0.06-0.32 ppm
H3N2: NR
Particle Size: NR
Route: Inhalation; Oropharyngeal aspiration (virus)
Dose/Concentration: DE: 529 or 2070 pg/m3
Time to Analysis: 4 h/day for 5 days. Virus
exposure immediately after last DE exposure.
Analyzed 18 h post infection.
DE exposure on susceptibility to Influenza
Infection: Mice exposed to 0.5 mg/m3 had
significantly greater levels of HA mRNA
compared to air-exposed mice. HA levels not
significantly altered in mice exposed to 2.0
mg/m3.
DE Exposure on the Influenza-induced
Inflammatory Response: f IL-6 mRNA levels
were significantly greater when exposed to 0.5
mg/m3 of DE prior to infection compared to air
exposure. Significantly increased amount of IL-6
protein observed in exposed mice. Exposure to
DE in absence of influenza infection had no
significant effect on IL-6 mRNA or protein levels.
DE Exposure on Pulmonary Injury: Infection
with the influenza virus increases levels of PMN
in BAL fluid. Exposure to either dose of DE prior
to infection showed no significant effect on PMN
levels Exposure to DE alone had no effect on
PMNs in BAL fluid. Neither exposure to DE nor
infection with influenza significantly increased
BAL fluid protein levels when compared to non-
infected, air-exposed.
Other Markers of Injury, NAG and MIA were
not statistically affected by DE or influenza
exposure.
DE Exposure on the Influenza Induced
Interferon Response: No significant change in
TFN-a mRNA levels at either dose of DE,
although mice exposed to 0.5 mg/m3 of DE prior
to infection had significantly greater levels of
IFN-B MRNA compared to air controls. No effect
on any of the IFNs observed in uninfected mice
exposed to DE.
DE Exposure on Surfactant Protein
Expression: Influenza virus infection alone
significantly increased expression of SP-A in air-
exposed. Exposure to 0.5 mg/m3 of DE prior to
infection had significant decreases in levels of
SP-A mRNA in the lunas, this effect was not
observed in 2.0 mg/m DE exposed. Decrease
seen in expression of SP-A protein in lungs of
mice exposed to 0.5 mg/m3 DE prior to infection.
Levels of SP-D mRNA and protein were
significantly decreased in lungs of mice exposed
to 0.5 mg/m3 of DE prior to infection compared
with mice exposed to air or 2.0 mg/m3 DE prior
to infection. Exposure to 0.5 mg/m of DE prior
to infection with influenza decreased levels of
SP-D, especially in airways. Mice exposed to
2.0 mg/m DE prior to infection showed no
significant difference.
Reference: Day et GEE (General Motors 1996 model 4.3-L
al. (2008,1902041 V6 engine; regular unleaded fuel) (CO,
NO, N02, S02, NH3)
Species: Mouse
Gender: Male
Strain: BALB/c
Age:8-10wk
Weight: NR
Particle Size: NR
Route: Whole-body Inhalation
Dose/Concentration: Low(L)- 6.6 + 3.7 PM/m3,
MediumfMl- 30.3 ± 11.8 PM/m3, High(H)- 59.1 ±
28.3 PM/m , High-Filtered(HF)
Time to Analysis: Pre-OVA protocol: OVA or saline
sensitized 7 days. OVA challenge day 14. GEE or
air exposed 6 h/day on days 26-28. Immediately
after exposure on day 28 challenged with OVA.
Tested for MCh-induced changes 24 h post-
exposure then sacrificed. Post-OVA protocol: OVA
or saline sensitized 7 days. OVA challenge day 14.
GEE or air exposed days 15-17. Tested for MCh-
induced changes 24 h post-exposure then
sacrificed.
Pre-OVA: In nonsensitized mice, neutrophils
and IgE decreased in the H group. IL-2
increased in the HF group and was dose-
dependent. Eosinophils dose-dependently
decreased. OVA-specific IgE increased in the H
group, and OVA-specific lgG2a dose-
dependently increased. In OVA-sensitized mice,
OVA-specific lgG1 increased in the M group.
Airway hyperresponsiveness was lower in the M
and HF groups.
Post-OVA: In nonsensitized mice, neutrophils
dose-dependently decreased, IL-4 decreased in
the M group, IL-5 decreased in the HF group,
and IFN-y decreased at all exposures. In OVA-
sensitized mice, IL-13 dose-dependently
decreased.
December 2009
D-118
-------
Study
Pollutant
Exposure
Effects
Reference: de
Haaretal. (2005,
0978721
Species: Mouse
Gender: Female
Strain:
BALB/cANNCrl
Age: 6-8 wk
Weight: NR
CBP: Carbon black particles in
phosphate buffered saline, 1:10 & 1:
100 dilutions (Brunschwich Chemicals,
Amsterdam, The Netherlands)
OVA: Ovalbumin
Particle Size: CBP: 30-50 nm
Route: Intranasal Droplet
Dose/Concentration: CBP+ OVA 200, 20, 2 pg
(3.3, 0.33, 0.033 mg/ml)
OVA only: 20 pg (0.5 mg/ml)
Time to Analysis: Droplet applied on days 0,1, 2.
Sacrificed on day 4 or challenged with OVA droplet
on days 25, 26, & 27. Sacrificed on day 28
Acute Airway Damage and Inflammation:
Only day 4 had LDH increased in the 200 pg
CBP+OVA group. The 200 pg CBP+OVA group
induced significantly higher numbers of BAL
cells compared to OVA control. Total protein and
TNF-a levels were increased only in 200 pg
CBP+OVA group. RAS, parameter for
phagocytosis, 200 pg and 20 pg CBP+OVA had
higher levels than OVA controls.
Adjuvant Activity on PBLN: Total lymphocytes
in PBLN significantly increased 4-5 fold in the
200 pg CBP+OVA exposed. 20 pg and 2 pg
exposures did not increase the number of PBLN
cells compared to OVA control. All CBP+OVA
concentrations induced higher levels of IL-4, IL-
5, IL-10, and IL-13, with 200 pg concentration
having 10-200 times higher levels. IFN-y
cytokine was increased in the 200 pg dose.
IgE Production: In CBP+OVA, IgEwere
significantly increased.
PBLN and Lung Lymphocytes after OVA
Challenge: PBLN cell numbers increased in
OVA and CBP+OVA sensitized mice. CD4 and
CDS populations increased in both groups.
PBLN levels in CBP+OVA and challenged with
PBS were higher than mice treated with OVA
and challenged with PBS, both groups cytokine
production was low, only IL-5 levels were
significant in the CBP+OVA/PBS group. Higher
lung lymphocyte numbers were caused by
higher numbers of CD4 and CD19. Production
of IL-5 and IL-10 was four to five times higher
than in OVA treated mice.
OVA Challenge Induces Asthma like Airway
Inflammation in CBP+OVA Sensitized Mice:
Total number of cells in BAL increased 10 fold in
CBP+OVA mice challenged with OVA.
Eosinophils exhibited highest increase in CBP+
OVA/OVA group. Perivascular and peribronchial
infiltrates and goblet cell hyperplasia in
CBP+OVA/OVA was confirmed by histological
examination. Antigen specific inflammation
induced in CBP+OVA mice.
December 2009
D-119
-------
Study
Pollutant
Exposure
Effects
Reference: de
Haar (2006,
1447461
Species: Mouse
Gender: Female
Strain:
BALB/cANNCr
Age: 6-8 wk
Weight: NR
CBP: fine (F) and ultrafine (UF) carbon
black particles (Ken Donaldson Group)
Ti02: fine and ultrafine
OVA: Ovalbumin
Particle Size: F CBP: 260.0 nm
UF CBP: 14.0 nm
FTi02: 250.0 nm
UFTi02:29.0nm
Route: Intranasal Droplet
Dose/Concentration: CBP: 200 pg (3.3 mg/mL)
Ti02: 200 pg (3.3 mg/mL)
OVA: 10 pg
CBP+OVA:200+10|jg
Time to Analysis: Days 0,1,2: Exposed to OVA or
CBP+OVA. Sacrificed on day 8 & analyzed after 2
h, or continued to second group.
Second group: days 25, 26, 27 given OVA
challenge day 28: sacrificed , analyzed 24 h post
sacrifice
Ultrafine Particles Induce Lung
Inflammation: UF Ti02 and CBP induced a
local inflammatory response in the airways and
showed higher levels of LDH and total protein
as compared to mice exposed to the F particles.
Cytokine levels were much higher in groups
exposed to ultrafine particles. Histologic
analysis of the airways showed that exposure to
ultrafine Ti02 or CBP leads to peribronchial and
perivascular inflammatory infiltrates (mostly
neutrophils). Exposure to OVA alone, or
combined with fine Ti02 and fine CBP had no
effects on lung histology.
Ultrafine Stimulate Local Immune
Responses: Ti02 and CBP particles stimulated
the local immune response against co
administered OVA antigen. Fine Ti02 particles
induced a low but significant increase in PBLN
cell number. Both types of ultrafine particles
elicited higher levels of Th-2 associated
cytokines, with UF CBP stimulating a greater
response. IFN-y production was low, but
significantly higher than OVA exposures.
Ultrafine T\0> Increase OVA-specific IgE and
lgG1 Levels: Levels of OVA specific IgE were
significantly increased in animals exposed to the
UF Ti02+ OVA compared to F Ti02 or OVA-only
Average IgE level in mice exposed to ultrafine
CBP+OVA was not a significant increase. OVA-
specific lgG2a not detected in any groups.
Ultrafine Particles Stimulate Allergic Airway
Sensitization Against OVA: At day 28, the
PBLN cell numbers were significantly higher in
both ultrafine and combination with OVA.
Production of OVA specific IL-4, IL-5, IL-10and
IL-13 by PBLN cells was significantly increased
in both ultrafine Ti02 and CBP. IFN-y levels were
significantly increased in ultrafine CBP+OVA
treated animals. F Ti02 had low, but significant,
increases in IL-4 and IFN-y compared to OVA
only. Allergic airway inflammation and Influxes of
eosinophils, neutrophils and lymphocytes were
only found in both groups exposed to ultrafine
particles.
Reference: de
Haar (2008,
1871281
Species: Mouse
Gender: Female
Strain: BALB/c,
CD80/CD86-
deficient, D011.10
Age: 6-8 wk
Weight: NR
Ultrafine Carbon Black (UFCB)
(Brunschwich Chemicals; Amsterdam,
The Netherlands)
Particle Size: Diameter: 30-50 nm
Route: Intranasal Exposure
Dose/Concentration: 20 |jg/mL
Time to Analysis: Exposed days 1, 2, 3. OVA
challenge days 25, 26, 27. Spleens and lymph
nodes from D011.10 mice pooled and CD4+ T-cells
isolated. Solution injected into tail veins of BALB/c
mice day 0. CTLA4-lg ip injected days 0, 2. PBLN
cell suspensions plated, restimulated with OVA 4
day.
UFCB+OVA induced proliferation of CD4+ T-
cells, increased cytokine production.
UFCB+OVA did not induce any effects in
CD80/CD86-deficient mice. UFCB-induced
airway inflammation is dose-dependent.
Reference: de Ultrafine Carbon Black (UFCB)
Haar et al. (2008, (Brunschwich Chemicals; Amsterdam,
1871281 The Netherlands)
Species: Mouse Particle Size: Diameter: 30-50 nm
Cell Line: Myeloid
dendritic cells
(mDCs)
Route: Cell Culture
Dose/Concentration: 25 |jg/mL
Time to Analysis: 18 h
UFCB+OVA increased mDCs in the
peribronchial lymph nodes, and their
expressions of CD80, CD86, and MHC-11.
December 2009
D-120
-------
Study
Pollutant
Exposure
Effects
Reference: Dong
et al. (2005,
0880791
Species: Rat
Gender: Male
Strain: Brown-
Norway
(BN/CrlBR)
Age: NR
Weight: 200-225 g
DEP: SRM 2975 (NIST, Gaithersburg,
MD)
OVA: Ovalbumin
Particle Size: 0.5 pm(MMAD)
Route: Inhalation
Dose/Concentration: DEP: 20.6 + 2.7 mg/m3
OVA 40.5 +6.3 mg/m3
Time to Analysis: 4 h/dayfor 5 days + OVA 30
min/day1 xwk on days 8,158,29. Sacrificed on
days 9 or 30.
Lung Inflammation/Injury: Both the BAL
proteins and inflammatory cell counts for DEP
exposure alone were not different from those of
the air exposed control, suggesting that DEP
exposure did not cause lung injury at 9 or 30
days post-exposure. OVA exposure caused
significant increases in neutrophils,
lymphocytes, eosinophils, albumin and LDH
activity in the lung after two exposures. DEP did
show a strong effect on OVA-induced
inflammatory responses.
Alveolar Macrophage (AM) function: OVA
exposure resulted in an increase in NO levels in
the acellular BAL fluid and AM conditioned
media. This increase was significantly
attenuated in rats exposed to DEP. DEP
exposure had no significant effect on the
production of IL-10 or IL-12 by AM recovered
from rats 9 and 30 days post exposure. In
contrast, OVA sensitization elevated both IL-10
and IL-12 secretion by AM at both time points.
Lymphocyte population and cytokine
production: DEP exposure was found to
increase the numbers of total lymphocytes, T
cells and their CD4+ and CD8+ subsets in
LDLN. OVA exposure also significantly
increased these cell counts on days 9 and 30.
DEP+OVA exposure showed a significant
reduction in total lymphocytes, T cells, CD4+
and CD8+ subsets on day 30. Levels of IL-4 and
IFN-y in lymphocyte conditioned media were
below detection limit of the ELISA kits.
Intracellular GSH levels in AM and
Lymphocytes: DEP exposure alone slightly
decrease GSH levels in AM, but markedly
reduced GSH concentration in lymphocytes on
days 9 and 30. OVA exposure significantly
decreased intracellular GSH in both cell types.
Combined exposure showed AM and
lymphocytes to have depleted intracellular GSH.
OVA specific IgE and IgG levels in serum: In
all samples collected on day 9, both serum IgG
and IgE levels were under the detection limits.
On day 30, no measureable IgE levels were
found. The OVA exposure, however, resulted in
elevated IgE levels, and was enhanced in rats
preexposed to DEP. IgE and IgG levels for
DEP+OVA was twO times higher than OVA alone
indicating that DEP has an adjuvant effect on
the production of IgG and IgE.
Effects of DEP and OVA on Lung iNOS
expression: AM from various exposure groups
did not stain for iNOS. 1 rat at day 9 from the
combined DEP+OVA group showed a slightly
positive iNOS staining. On day 30, 2 of 5 rats
from combined exposure group and 1 from the
OVA group showed a positive airway staining.
December 2009
D-121
-------
Study
Pollutant
Exposure
Effects
Reference: Dong DEP: SRM 2975 Diesel Exhaust
et al. (2005,
0880831
Species: Rat
Gender: Male
Strain: Brown-
Norway
(BN/CrlBR)
Age: NR
Weight: 200-225 g
Particles (NIST)
OVA: Ovalbumin
Particle Size: 0.5 pm(MMAD)
Route: Nose-only Inhalation
Dose/Concentration: DEP 22.7 + 2.5 mg/m3
OVA 42.3 +5.7 mg/m3
Time to Analysis: Day 1, 8,15: OVA exposure 30
min/day
Days 24-28: DEP exposure 4 h/day
Day 29: OVA challenge
Day 30: Whole-body plethysmography
Day 31: Sacrifice
Effect of DEP on OVA Induced Allergic
Responses: DEP exposure had a synergistic
effect with OVA on inducing airway hyper-
responsiveness (AHR) in rats. DEP alone had
no effect on IgG production. Levels of OVA-
specific IgG and IgE increased in OVA+DEP
exposure. This indicates that DEP pre-exposure
augments the immune response of rats to OVA
in the production of allergen specific IgG and
Effect of DEP on OVA Induced Cell
Differentiation: Neither DEP, OVA nor the
combination induced elevated levels of LDH
activity or albumin content, indicating that the
exposure protocols did not cause significant
lung injury. DEP alone induced moderate but
significant increase of neutrophil numbers. OVA
exposure induced a greater infiltration of
neutrophils than DEP, and infiltration of
eosinophilsand lymphocytes. OVA-induced
eosinophil count markedly increased with DEP
exposure. Total lymphocytes, T cells, and their
CD4+ and CD8+ subsets in LDLN from rats
sensitized and challenged by OVA were
significantly higher than those of air-exposed
non sensitized rats. DEP+OVA exposure
resulted in substantial increase in T cells
compared to OVA alone.
Effect of DEP on OVA-induced Oxidant
Generation and GSH Depletion: Exposure to
DEP or OVA alone had no effect on ROS
production by AM. Substantial elevation seen in
ROS for the DEP+OVA exposed group. Both
OVA and DEP exposures resulted in an
increased presence of NO in the acellular BAL
fluid and in AM conditioned media; OVA+DEP
exposure further increased these levels. The
ATI I cells from OVA exposed rats exhibited a
higher percentage of cells that produce NO and
superoxide than air exposed, non sensitized
rats. DEP and OVA exposure resulted in a signi-
ficant increase in the percentage of cells that
produce NO and superoxide over the control.
iNOS Expression: Immunohistological analysis
in lung tissues showed no AM staining in any
group. Airway epithelium was found to be
positive in all 5 rats from the DEP+OVA group
and 3 of 5 rats from single exposure of DEP or
OVA and 2 of 5 in air only exposed rats. iNOS
expression was significantly higher in ATII cells
isolated from rats exposed to combined DEP
and OVA .
GSH levels in AM and lymphocytes: Levels
were slightly lowered by DEP or OVA exposure,
though not statistically significant. DEP+OVA
showed a significant reduction in GSH levels.
December 2009
D-122
-------
Study
Pollutant
Exposure
Effects
Reference: Drela ASM: Air suspended PM from Upper
etal. (2006, - • - • -
0963521
Species: Mouse
Strain: BALB/c
Age: 6 wk
Weight: NR
Silesia (Poland)
IpgofASM:
Pb(1.136ng)
Cu (0.004 pg)
Co (0.072 ng)
Mn (0.406 ng)
Fe(0.016|jg)
Route: Intraperitoneal Injection
Dose/Concentration: 170 mg/kg
Time to Analysis: Single, 72 h
Cr(0.418ng)
Ni (0.238 ng)
Particle Size: 0.3-1 Opm
CD28 Expression on Thymocytes at Different
Stages of Development: ASM exposure
accelerated thymocyte maturation but did not
alter the expression of CD28 on peripheral CD4
and CDS T cells isolated from lymph nodes. A
slight but not statistically significant decrease in
the expression of CD28 on spleen T cells from
ASM animals was observed.
Distribution of CD28(low) and CD28(high):
Acute exposure to ASM resulted in the increase
of CD28(low) and decrease of CD28 (high)
thymocyte percentages in the total thymocyte
population. The percentages of CD28 low and
high thymocytes did not differ between intact
and PBS controls. Acute ASM exposure resulted
in the increase of the percentage of CD28(low)
and the decrease of CD28(high) thymocytes in
the CDS low subset. The percentage of CD28
low and high positive thymocytes did not differ in
CDS high thymocyte subset.
Natural Regulatory CD4+ CD26+ T Cells in
the Thymus: The development of thymic
natural regulatory cells was unaffected by ASM.
Proliferation of Splenocytes and Lymph
Node Lymphocytes: Decreased proliferative
responses were evident in splenocytes from
ASM-exposed animals when cells were
stimulated with low but not high levels of anti-
CD3 mAb. In contrast, lymph node lymphocytes
from ASM treated mice had increased
proliferative responses independent of anti-CDS
concentration. Both CD4+ and CD8+ T cells
from ASM treated mice proliferated more
vigorously than from controls. Almost all CD8+ T
cells from ASM mice were induced to proliferate.
Reference: Dybing
et al. (2004,
0975451
Species: Mouse
Gender: NR
Strain: BALB/cA
Age: NR
Weight: NR
UP: Urban ambient particles collected in
5 different sites (Amsterdam, Lodz, Oslo,
Rome, Dutch seaside) during four-wk
periods in spring, summer, winter
seasons from March 2001 to March
2004.
DEP as reference std: SRM 1650 (NIST)
OVA: Ovalbumin (Sigma Chemical, St.
Louis, MO)
Particle Size: UP: PM,n and PM, 5
Route: Injection in hind foot pad
Dose/Concentration: UP: 100- 200 pg
DEP: 50 pg
OVA: 50 pg
Time to Analysis:
DayO: 1 exposure to OVA alone, OVAw/particles,
particles alone.
Day 6: Lymph nodes harvested
Day 21:1 OVA w/o particles exposure
Day 26: Antibody assay
Allergy Screening: All samples were
immunostimulatory in the popliteal lymph node
assay; activity was weak in the absence of OVA
but statistically significant when injected with
OVA, indicating an adjuvant effect. Particle
adjuvancy was further demonstrated via
significant enhancement of OVA-specific
antibody responses. All ambient particle
fractions from all seasons increased IgGl
Except for a few coarse samples, all fractions
significantly increased IgE. All fine fractions and
some coarse fractions significantly increased
lgG2a, indicating that most particles could exert
both Th1 and Th2 adjuvancy. In general, fine
particles demonstrated stronger adjuvant activity
than coarse in a pair-wise comparison of coarse
and fine particles from the same location.
Reference:
Dybing, et al.
(2004, 0975451
Species Rat
Cell Lines: Type 2
cells, AM
Dose/Concentration: 0-50 |jg/ml
Time to Analysis: 20 h
UP: Urban ambient particles collected in Route: Cell Culture
5 different sites (Amsterdam, Lodz, Oslo,
Rome, Dutch seaside) during four-wk
periods in spring, summer, winter
seasons from March 2001 to March
2004.
DEP: SRM 2975 (NIST)
OVA: Ovalbumin (Sigma Chemical, St.
Louis, MO)
Particle Size: PM10 and PM2 5
Inflammation: The coarse fractions were more
potent than the fine fractions. Among the
samples, the overall effects of the coarse
fractions on the cells were dependent on the site
of collection. High MIP-2 levels were found
using particles from some spring collections.
Coarse particles collected in summer
demonstrated the highest potency, and samples
collected during winter proved to be less potent
but seasonal variation was not obvious for all
sites. Only minor responses were observed
using fine fractions from urban sites.
December 2009
D-123
-------
Study
Pollutant
Exposure
Effects
Reference: Farraj
et al. (2006,
1417301
Species: Mouse
Gender: Male
Strain: BALB/c
Age: 6 wk
Weight: NR
DEP: SRM 2975 NIST
OVA: Ovalbumin
Anti-p75: Rabbit anti-mouse p75
neurotrophin receptor polyclonal
antibody (Chemicon, Temecula, CA)
Anti-trkA: anti-mouse trkA NGF receptor
antibody (Santa Cruz, Santa Cruz, CA)
Particle Size: DEP: 1.47 pm(MMAD),
2.75 GSD
Route: Nose-only Inhalation
Dose/Concentration: DEP: 1.78 to 2.18 mg/m3
Anti-p75: 50 \i\
Anti-trkA: 50 pi
OVA injection: 20 pg
MCH:0,16, 32, 64 mg/ml
Time to Analysis: On day 0: ip injection of 20 pg
OVA
Day 14: intranasal instillation of 50 pi anti-p75 or
anti-trkA, 1 h after 1st exposure challenged with
OVA aerosol for 1 h followed by a h exposure to
DEP
24 h after DEP exposure: MCH challenge
Airways Responsiveness: No significant
differences in avg baseline Penh values of any
treatment groups.
Vehicle sensitized mice: exposure to DEP, anti-
p75 or anti-trkA had no effect on MCH-induced
Penh values.
OVA-sensitized DEP-exposed: seen increase of
Penh values. Administration of anti-p75 or anti-
trkA to OVA sensitized mice reversed DEP
induced Penh increases.
Lung Function in Ventilated Mice: Compared
to vehicle sensitized mice, central airway
resistance (Rn) increased 62% in OVA
sensitized mice was not a significant increase.
OVA-sensitized DEP-exposed mice, anti-p75
decreased central airway resistance (Rn) and
anti-trkA did not significantly alter Rn. though Rn
response for anti-p75 was significantly less than
anti-trkA response, Constant phase model
parameter of tissue elastance not significantly
affected by any treatments or by increasing
MCH dose, indicating development of significant
regional ventilation inhomogeneity during
bronchoconstriction.
Airway Pathology: OVA-sensitized mice had
small increases in intraepithelial mucus
compared to vehicle-sensitized mice. DEP
exposure did not enhance severity of OVA-
induced airway pathology. Anti-p75 or anti-trkA
administration did not influence airway
morphology.
BAL Cells: Vehicle-sensitized DEP-exposed
mice had significantly enhanced macrophage
numbers by 92% compared to air-exposed,
vehicle-sensitized mice. Anti-p75 or Anti-trkA
administration significantly suppressed DEP-
induced macrophage increase to levels similar
to air-exposed, vehicle-sensitized group. DEP
co exposure significantly decreased number of
macrophages in OVA-sensitized mice to control
levels. Anti-trkA or anti-p75 had no effect in
OVA-sensitized, DEP-exposed. Eosinophil
number greater in OVA-sensitized DEP-exposed
mice than in vehicle-sensitized air-exposed
mice. No significant effects of DEP exposure on
neutrophils from vehicle- or OVA-sensitized
Cytokines: IL4: OVA-sensitized DEP-exposed
had five-fold increase over vehicle-sensitized,
air-exposed mice and anti-trkA or anti-p75
significantly inhibited the DEP-induced increase.
IL5, IL13: OVA-sensitized DEP-exposed had no
significant change. Anti-p75 or anti-trkA
administration had no significant effect.
Serum IgE: OVA sensitized mice had a 10 fold
increase in IgE levels for air and DEP exposed
mice. Anti-p75, anti-trkA treatment did not cause
significant effects on IgE levels.
December 2009
D-124
-------
Study
Pollutant
Exposure
Effects
Reference: Farraj DEP: SRM 2975 collected from diesel-
etal. (2006, " "'
Species: Mouse
Gender: Male
Strains: C57/BI6
Age: 6 wk
powered industrial forklift filter (NIST)
OVA: Ovalbumin
Anti-p75: Rabbit anti-mouse p75
neurotrophin receptor polyclonal
antibody
Route: Nose-only Inhalation
Dose/Concentration: DEP: 0.87 mg/m3
MCH:0,16, 32, 64 mg/ml
OVA: 20 pg ip
Anti-p75: 50 \i\
Particle Size: 1.47 (MMAD), 2.75 (GSD) T|me to Ana|ysis: Day Q. OVA in ge| vehic|e] ip
Day 14: anti-p75 exposure, intranasal instillation
1 h post anti-p75 exposure, OVA aerosol challenge
forl h
1 h post OVA challenge: DEP exposure for 5 h
48 h post DEP exposure: MCH challenge
Airway Responsiveness: No significant
differences in average Penh values among any
vehicle control groups. No significant differences
in treatment groups in OVA-sensitized mice at
baseline 0,16, or 32 mg/mL of MCH. At 64
mg/mL MCH, OVA-sensitized, DEP-exposed
mice had a 22% increase in Penh compared to
vehicle mice, and a 68% increase compared to
vehicle-sensitized, air-exposed mice. Instillation
of anti-p75 inhibited the DEP induced increased
Penh.
BALF Cells: DEP exposure in vehicle-
sensitized mice significantly increased
macrophages by 161 % compared to air-
exposed, vehicle-sensitized mice, while OVA-
sensitized mice had 69% increase. Anti-p75
administration significantly suppressed DEP-
induced macrophage increase in vehicle-
sensitized mice. No significant effects of DEP
exposure or anti-p75 treatment in OVA-allergic
OVA-sensitized air-exposed mice had a several
hundred fold increase in the number of
eosinophils. No significant effects of DEP
exposure or anti-p75 treatment on eosinophils
from OVA-sensitized mice. OVA-exposure or
DEP-exposure had no significant effects on
neutrophil or lymphocyte number.
Cytokines: No significant effects of DEP alone
orwithOVAonlL-4, IL-5, or IL-13.
Serum IgE: OVA sensitization in the presence
or absence of DEP or anti-p75 caused at least a
3 fold increase in IgE levels. No significant
effects of DEP or anti-p75 treatment on IgE
levels.
Reference:
Finkelman et al.
(2004, 0965721
Species: Mouse
Gender: Female
Strains: BALB/c,
C57BL/6
Age: 2-4 mo
DEP: 4JB1 type; Isuzu Automobile,
Tokyo, Japan
Particle Size: 2pm (MMAD)
Route: Groupl: 1 ip injection of 2 mg of DEP.
Group 2: daily ip injections of 2 mg of DEP
Dose/Concentration: 2 mg
Time to Analysis: 2-96 h
Serum Cytokines: Mice in group 1
demonstrated an increase in serum IL-6
production but no increase in IL-4 or IL-2
production. IFN-y levels were decreased in
group 2. TNF production was not affected.
Spleen Cytokines: When injected before IPS,
DEP had little effect on the LPS-induced TNF-a
and IL-6 response, but resulted in a minor
suppression of INF-y and IL-10. DEP LPS-
induced increase in INF-y mRNA responses in
spleen cells. DEP caused a dose related
suppression of LPS stimulated INF-y. DEP had
little or no effect on the percentage of NK or
NKT cells in the spleen and inhibited LPS-
induced IFN-y production by NK and NKT. DEP
failed to inhibit the IFN-y response by anti-CD3
mAb-activated NKT cells. Oxidant activity was
not responsible for DEP inhibition of LPS-
induced IFN-y production.
Reference:
Fujimaki and
Kurokawa (2004,
0965751
Species: Mouse
Gender: Male
Strains: BALB/c
Age: 4 wk
Cell Types:
Cervical lymph-
node (CLN) cells
DE + particles: Comparison of exposure
to DE including particles and exposure to
particle-filtered DE
DE: 12.09 +0.15 NOX, 1.99 + 0.02 N02,
10.02+ 0.12 NO, 0.18+ 0.002 S02 and
1769.2 +13.2 C02 (all in ppm).
DE gas: 11.93 + 0.13 NOX, 2.93 + 0.06
N02, 8.91 + 0.09 NO, 0.11 + 0.003 S02
and 1838.8 + 15.3 C02 (all in ppm)
Particle Size: 0.4 pm (MMAD)
Route: Whole-body Inhalation
Dose/Concentration: Exposure to: 0,1.0 mg/m3 or
1.0 mg/m3 DE gas only (0.04 mg/m3 PM)
Time to Analysis: Exposure for 12 h daily for 5 wk.
Days 14 and 35 challenge with sugi basic protein
(SBP), a cedar pollen allergen, intranasally.
Evaluation is 24 and 48 h after final SBP injection.
CLN Response: Exposure to DE or DE gas did
not affect B1 lymphocyte subpopulations of
CLN. Culture supernatants of CLN cells from DE
exposed/SBP immunized mice showed
significant increase in MCP-1 at 24 and 48 h.
Exposure to DE or DE gas significantly
increased the amount of TARC and MIP-1a in
CLN cells from SBP-immunized mice at 48 h.
December 2009
D-125
-------
Study
Pollutant
Exposure
Effects
Reference:
Fujimaki et al.
(2005, 1564561
Species: Mouse
Gender: Male
Strains: C57BL/6
Age: 4 wk
DE generated by 4 cyl 2.741 Isuzu diesel
DE gas = DE filtered to remove particles
Composition of Diesel Exhaust: DE
DEP: 1.01 mg/m3
1796ppmC02
12.09ppmNOx
0.18ppmS02
Composition of filtered DE Gas: DEP:
0.04 mg/m3
1839ppmC02
11.93ppm NOX
0.11 ppmS02
Sugi Basic Protein (SBP)- allergen
Particle Size: 0.4 pm (average
diameter)
Route: Whole-body Inhalation
Dose/Concentration: 1.0 mg DEP/m3 or 1.0 mg
DEP/m3 DE gas
Time to Analysis: 12 h daily, 5wk. All mice were
injected IP with 100 pg SBP before exposure to gas
or DE and again received 50 pg SBP intranasally
on days 14 and 35. Evaluation is 1 day after final
SBP-immunization (mice are euthanized and CLN
and blood samples are collected)
CLN: Exposure to DE and gas led to a decrease
in total number of CLN cells and percentage of
CD4+ and TCR-B levels. Cell proliferation
response to SBP was higher in gas-exposed
mice than in the control group. The production
of MCP-1 increased in CLN cells when
stimulated with SBP (in vitro) but the difference
was not significant at 24 and 48 h. SBP-
stimulated cells in gas-exposed mice showed
greatly enhanced MIP-1a production at 24 and
48 h. Exposure to gas increased the amount of
TARC in the culture supernatants of CLN cells.
Plasma: Exposure to DE or gas significantly
decreased the anti-SBP lgG1 antibody liters and
increased the anti-SBP lgG2a antibody liters in
mouse plasma.
Reference: DEP: generated by a 2369-cc diesel
Fujimoto et al. engine operated at 1050 rpm and 80%
(2005, 0965561 load with commercial light oil
Species: Mouse Particle Size: 0.4 pm (MMAD)
Gender: Female
1st day of
pregnancy)
Strains: Sic: IRC
Route: Whole-body Inhalation
Dose/Concentration: 0.3,1.0 and 3.0 mg DEP/m3
(Groups 1,2,3)
Time to Analysis: Exposure began at 2 days
postcoitum and was continued until 13 days
postcoitum. Exposure time was 12 h daily for 7
days/wk. Pregnant females were sacrificed 14 days
postcoitum.
mRNA Expression in Placentas: In groups
exposed to DE, the expression of CYP1A1
mRNA decreased to undetectable levels during
placental absorption and INF-y was increased.
Levels of CYP1A1 mRNA in normal placentas
from DE-exposed mice were unchanged. mRNA
levels of inflammatory cytokines IL-2, IL-5, IL-
12a, IL-12B and GM-CSF increased in
placentas of mice exposed to DE.
Reference: Gao et ROFA: collected near a power plant in
al. (2004, 0879501 FL burning low sulfur # 6 oil.
Species: Human
Cell Line: Lung
fibroblasts infected
with Mycoplasma
fermentans
(PM from Dusseldorf, volcanic ash for
Mt. St. Helens, PM from Utah used to
compare against ROFA in one
experiment)
NiS04, CuS04, VOS04, Na3V04
Particle Size: NR
Route: Cell Culture; seeded into 6-well plates (3-
4.5x105 cells/3 mL/well) or 24-well plates (0.6-
1xl05 cells/1.0 mL/well)
Dose/Concentration: PM: 3,10, 20, 40, 50 pg/ml
Metallic salts: 2, 20, 200 pM
Time to Analysis: 24, 48h
Cytokines: ROFA exposure in combination with
Mycoplasma fermentans infection synergistically
amplifies the induction of IL-6 production in
human lung fibroblasts (HLF). PM from the other
sources has little synergistic effect on IL-6
release. Exposing HLF cellsto,M. fermentans
derived macrophage activating lipopeptide-2
(MALP-2) and ROFA has the same synergistic
effect as M. fermentans infection and ROFA.
MALP-2 and ROFA extract have a similar
synergistic effect that requires more time to
appear. ROFA contains high levels of V, Ni, Fe
and Cu. Exposure of HLF to NiS04 alone and
NiS04 with MALP-2 produced 10 and 50 fold
increases, respectively, in IL-6 production.
Exposure of HLF to CuS04, VOS04 and
Na3V04, with and without the presence of
MALP-2, did not produce as dramatic results as
seen with Ni. The action of NiS04 and MALP-2
on IL-6 production was found to be dose
dependent.
December 2009
D-126
-------
Study
Pollutant
Exposure
Effects
Reference: Gavett PM2 5 from the German cities of Hettstedt Route: Oropharyngeal Aspirations
etal. (2003, orZerbst
0531531
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 7wk
Dose/Concentration: 50-100 |jg
PM Composition: samples from Hettstedt
have several-fold higher levels of Zn, Time to Analysis: Single, 18 h.
Mg, Pb, Cu and Cd than samples from
Zerbst.
Particle Size: PM2
Sensitization Model: Mice were exposed to 50 pg
PM 2 h before being sensitized with 10 pg OVA,
repeated two days later. On day 14 all mice were
challenged with 20 pg OVA.
Parameters measured on days 2 and 7 after final
exposure to OVA.
Challenge Model: Mice were sensitized IP with 20
fjg OVA or adjuvant only. 14 days later mice were
exposed to 100 pg PM25 followed 2 h later by 20 pg
OVA. Parameters measured on days 2 and 7 after
final exposure to OVA.
BAL Analysis: Hettstedt PM significantly
increased BAL protein and NAG levels. Zerbst
PM did not. Mice exposed to Zerbst had lower
levels of LDH than control groups. Hettstedt
exposed mice had increased levels of IL-1B, IL-
6 and MIP-2 in comparison to control and to
mice exposed to Zerbst PM. PM2 5 at a dose of
100 pg was not found to be toxic, therefore used
for subsequent studies.
Airway Responsiveness (PenH): In allergic
mice tested immediately after exposure,
Hettstedt PM increased PenH 190% compared
to baseline, Zerbst increased PenH by 120%
and the Control increased by 44%.:.: Two days
after OVA challenge, no differences in non-
allergic mice from either group. In allergic mice,
Hettstedt PM still caused a significant response
to Mch responsiveness, Zerbst none. No effects
on day seven.
IgE Levels: Serum collected on day 2 showed
antigen-specific IgE was increased by Hettstedt
PM25 in both the sensitization and challenge
phases when compared to the control and
exposure to Zerbst. Day 7 serum indicated no
effect.
BALF Cells: In non-allergic mice both Hettstedt
and Zerbst PM increased neutrophil numbers
(3-fold; not statistically significant) and in allergic
mice, only Hettstedt PM significantly increased
neutrophil count. Eosinophil numbers were
increased only in allergic mice exposed to
Hettstedt PM. Lymphocyte numbers were not
different among groups.
BAL Injury Markers: At 2 days after both
Hettstedt and Zerbst PM administered in allergic
mice caused significant increases in protein,
LDH and NAG compared to the non-allergic
groups. Both PMs caused an increase in LDH in
allergic mice compared to the allergic control,
but only Hettstedt caused an increase NAG in
allergic mice compared to control. At 7 days no
effect.
BAL Cytokines: Allergic mice had increased
levels of IL-4, IL-5 and IL-13 compared to non-
allergic mice (at 2 days after). IL-5 was
significantly increased by exposure to either PM
in allergic mice compared to non-allergic mice.
Exposure to either PM caused an increase in
TNF-a and IFN-y (by 6-8 fold) in allergic mice,
there was also an increase in these
inflammatory cytokines in the non-allergic group
but was not statistically significant. No
significant effects were observed in animals that
underwent the sensitization protocol alone for
any measurement or endpoint.
Reference: Gowdy
et al. (2008,
0972261
Species: Mouse
Gender: Female
Strain: BALB/c
Age:-12-14 wk
Weight: 17-20 g
DEP (30kW (40hp) 4-cylinder Deutz
BF4M1008 diesel engine, steady state,
20% full load) (Low dose: 21% 02, 0.4wt
ratio OC/EC; High dose: 20.7% 02,
0.4wt ratio OC/EC) (CO, NOX, S02)
Particle Size: Diameter: -240 nm
Route: Inhalation
Dose/Concentration: Low- 514 + 3 pg/m3, High-
2026 + 38 pg/m3
Time to Analysis: 4 h/day, 1 or 5 days
(consecutive). Necropsied immediately or 18 h
postexposure.
BAL Analysis: Neutrophils and lung injury
dose-dependently increased. ICAM-1 increased
immediately after both exposures and after 18h
postexposure in the low dose.
Cytokines: After 1 day exposure, IFN-y and
TNF-a increased immediately at both doses and
the high dose, respectively. Immediately after 5
days exposure TNF-a and IFN-y increased at
both concentrations and IL-6 increased at the
low dose. At 18 h postexposure IL-6 and IFN-y
increased at both doses, TNF-a and IL-13
increased at the low dose, and MIP-2 dose-
dependently increased.
CCSP, Surfactants: CCSP decreased. SP-A
and SP-D decreases were only significant after
5 days exposure, 18 h post-exposure.
December 2009
D-127
-------
Study
Pollutant
Exposure
Effects
Reference:
Hamada et al.
(2007, 0912351
Species: Mouse
Gender: Female
(Pregnant close to
partruition)
Strain: BALB/c
ROFA (obtained from a precipitator until
of a local power plant)
Composition of ROFA (in pg/mL): 341.2
Ni, 323.4V, 232.2 Zn, 18.3 Co, 15.8 Mn,
8.4 Ca, 6.7 Cu, 6.1 Sr, 5.0 mg, 0.9 Sb,
and 0.6 Cd.
Particle Size: NR
Route: Nebulized ROFA leachate
Dose/Concentration: 50 mg/mL dilution
Time to Analysis: Pregnant mice exposed to
nebulized ROFA leachate for 30 min/day at days
14,16 and 18 of pregnancy.
Newborns received a single injection (ip) of OVA (5
pg)+ alum (1 mg) at day 0 followed by exposure to:
1. aerosolized OVA days 12,13 and 14 (2-wkold
protocol)
OR
2. aerosolized OVA days 32, 33 and 34 (5 wk old
protocol)
Analysis 48 h after final allergen exposure
Susceptibility to Asthma: Exposure of mother
to PBS aerosols during pregnancy did not result
in prominent asthma features in young. The
offspring of the ROFA mothers revealed
increasing AHR and elevated numbers of
eosinophils in the BAL fluid. Similar results were
seen in both the 2-wk and 5-wk old groups.
IgE Levels: Histopathology revealed prominent
inflammation in the lungs of the ROFA neonates
and increased allergen-specific IgE and lgG1
levels in the 5-wk group.
Maternal Influence: Breast milk was not shown
to be responsible for the increased susceptibility
to allergy seen in offspring.
IL-4 and IFN-y: IL-4 and IFN-y levels in
maternal mice showed no difference between
PBS exposed or ROFA exposed mice. Cultured
spleen cells from mice born of ROFA-exposed
mothers showed either increased or similar
levels of IL-4 and decreased production of IFN-y
causing an increase in the ratio of IL-4/IFN-y
indicating greater susceptibility to develop Th2-
allergic response.
Eosinophils: Exposure of mothers to Ni levels
similar to those found in ROFA had no
appreciable effect on BAL eosinophil.
Reference: Hao et DEP (4-cylinder diesel engine under a
al. (2003, 0965651 10-torque load)
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 6-7 wk
Particle Size: NR
Route: Nebulization
Dose/Concentration: 2 mg DEP m3
Time to Analysis: Mild Sensitization- Mice receive
IP OVA alum and are challenge with aerosolized
OVA with and without DEPs. Mice sacrificed d19.
Postchallenge Model- DEPs are delivered to mice
sensitized by IP OVA and alum. Mice sacrificed
d23.
Transgenic Mice: Mice exposed to nebulized saline
or DEPs for 1 h daily for 3 days. Mice sacrificed
day5.
Mild Sensitization: Exposure of previously OVA
sensitized mice to aerosolized DEP and OVA did
not affect OVA-specific IgE production, BAL
eosinophilia or methacholine-induced AHR.
Aerosolized particles induced inflammation and
increased MBP deposition and MBP positive
eosinophils in the mucosa.
IL-6 Transgenic: Exposure to aerosolized DEP
did not change BAL cytokine levels, but did
increase AHR and BAL cell count.
Classic Sensitization, Post-Challenge: Did
not lead to a discernable increase in OVA-
induced AHR. DEP treatment was associated
with increased airway inflammation and mucin
production in larger and intermediary airways.
Reference:
Harkema et al.
(2004, 0568421
Species: Rat
Gender: Male
Strain: F344, BN
Age: 10-12 wk
Weight: NR
CAPs (Detroit; July-Sept. 2000; Harvard
Ambient Fine Particle Concentrator)
Particle Size: 2.5 pm (diameter)
Route: Inhalation; IT Instillation.
Dose/Concentration: 4 day concentration: 676 +
288 pg/m3, 5 day concentration: 313+119 pg/m3,
July concentration: 16-185 pg/m3, September
concentration: 81-755 pg/m3; IT Instillation- 200 pL
(soluble and insoluble)
Time to Analysis: 10 h/day 1, 4, 5 day
(consecutive); F344 rats sensitized to endotoxin,
BN rats to OVA. Both groups killed 24 h post-
exposure.
The retention of PM in the airways was
enhanced by allergic serialization. Recovery of
anthropogenic trace elements was greatest for
CAPs-exposed rats. Temporal increases in
these elements were associated with eosinophil
influx, BAL protein content and increased airway
mucosubstances. A mild pulmonary neutrophilic
inflammation was observed in rats instilled with
the insoluble fraction but instillation of total,
soluble or insoluble PM25 in allergic rats did not
result in differential effects.
December 2009
D-128
-------
Study
Pollutant
Exposure
Effects
Reference: Harrod DEE: Diesel Engine Emissions
et al. (2003,
0970461
Species: Mouse
Gender: NR
Strains: C57BL/6
Age:8-10wk
generated from a 5.9-liter turbo diesel
engine fueled by Number 2 fuel.
DEE Composition:
NOX: 2.0-43.3 ppm
CO: 0.94-29.0 ppm
S02: 8.3-364.9 ppb
Particle Size: 0.1-0.2 pm(MMAD)
Route: Whole-body Inhalation
RSV: IT administration
Dose/Concentration: DEE: 38.8 pg/m3 (low level)
or 10027 pg/m3 (high level)
RSV:100|jl
Time to Analysis: 6 h/day 7 days
After the final 6 h exposure period mice were
infected with RSV.
Parameters measured 4 days post infection
Viral Gene Expression: For air+RSV, RSV-F
gene expression was not apparent but RSV-G
gene expression was detectable at very low
levels. In DEE+RSV (for high and low levels),
RSV-F and -G were markedly elevated. B-Actin
mRNA levels not changed in DEE-exposed
compared to air-treated. DEE+RSV for high and
low levels show 10- to 20- fold induction of RSV-
G mRNA levels as compared to air+RSV.
BALF Cells: Uninfected low-level DEE did not
induce statistically significant increase in cell
numbers as compared to air+RSV. High level
DEE+RSV caused increase as compared to
air+RSV. Uninfected high-level DEE had
increase as compared to uninfected air group.
For all groups, alveolar macrophages were
predominant cell type and no substantial
changes in infiltrating cell populations by
exposure to DEE were noted.
Lung Inflammation & Airway Epithelial
Morphology: Lung sections from air- or DEE-
exposed, uninfected did not exhibit any
observable change. Low level DEE + RSV had
increased inflammatory cell infiltration in
peribronchial regions and loss of normal
cuboidal appearance of Clara cells as compared
to air+RSV. High level DEE+RSV had more
apparent lung-inflammation, especially
surrounding bronchi and bronchioles, and
increased appearance of pseudo-stratified,
columnar epithelial cell morphology and
apparent airway epithelial cell sloughing as
compared to low level DEE+RSV, indicating
dose-related increase in lung histopathology to
RSV infection by prior DEE exposure.
Cytokines: TNF-a and IFN-y were significantly
increased in RSV-infected mice exposed to low
or high level DEE and not increased in RSV-
infected mice exposed to air. TNF-a levels
elevated to similar levels for low and high level
DEE+RSV. IFN-y exhibited more dose-related
increase with higher levels in high level
DEE+RSV versus low level DEE+RSV.
Mucous Cell Metaplasia: DEE exposure in
uninfected was not altered. Mucous metaplasia
was increased in epithelium of RSV-infected
mice when exposed to DEE in a dose-
dependent manner. Following high level
DEE+RSV, mucous staining of airway epithelial
cells in more distal airways was occasionally
observed.
CCSP Production in Airway Epithelium: DEE
alone did not have an effect CCSP-producing
cells, or Clara cells, decreased in Low DEE +
RSV and further decreased in high level
DEE+RSV in large and terminal airways.
Surfactant Protein B: proSP-B staining post
RSV alone shows now discernible decrease
when compared to uninfected. Staining levels in
alveolar lung regions decreased when exposed
to low level DEE+RSV, and further decreased in
high level DEE+ RSV. Staining in airway
epithelium following high level DEE+RSV
diminished when compared to RSV alone or low
level DEE+RSV.
SP-A: In alveolar type II cells and airway
epithelial cells for untreated and air +RSV, no
discernible changes in levels. Prior exposure to
low or high level DEE decreased SP-A staining
in alveolar type II cells and airways epithelial
cells during RSV infection.
December 2009
D-129
-------
Study
Pollutant
Exposure
Effects
Reference: Harrod DEE (2 2000 model 5.9-1 Cummins ISB Route: Inhalation
et al. (2005,
0881441
Species: Mouse
Gender: Male
Strain: C57B1/6
Age: 10-12 wk
Weight: NR
turbo diesel engines, No. 2 certification
diesel fuel)
Particle Size: NR
Dose/Concentration: Low- 30 pg/m3 PM, Mid-
Low- 100 pg/m3 PM, Mid-High- 300 pg/m3 PM,
High-1000|jg/m3PM
Time to Analysis: 6 h/d, 7 days/wk, 1 wk or 6 mo.
1 wk exposure repeated on separate occasion.
Immediately after exposure, mice anesthetized, IT
instilled with Pseudomonas aeruginosa.
Bacterial Clearance: Lung bacterial clearance
was decreased at all levels after 1wk exposure
and was concentration-dependent 18h
postinfection. Bacterial clearance was not
affected at 6m and bacterial counts were higher.
Inflammation, Particle Deposition: Lung
inflammation and histopathology were increased
in all exposure groups postinfection. All
exposure groups possessed particle-laden
macrophages. Higher doses had a
concentration-dependent increase.
Ciliated, Clara Cells, TTF-1: Generally, ciliated
cells decreased with exposure dose, were more
discernible in inflamed airways, and higher
doses caused effects in small distal airways.
Clara cells decreased equally at all exposures
and were most notable in the distal airway
epithelium. TTF-1 decreased postinfection.
Reference:
Heidenfelder et al.
(2009, 1900261
Species: Rat
Gender: Male
Strain: Brown-
Norway
Age: 10-12 wk
Weight: NR
CAPs (Grand Rapids, Ml; July)
Particle Size: Diameter: 0.1-2.5 pm
Route: Whole-body Inhalation
Dose/Concentration: CAPs: 493 + 391 pg/m3; OC:
244 ± 144 pg/m3, EC: 10 ± 4 pg/m3, Sulfate: 79 ±
131 pg/m , Nitrate: 39 + 67 pg/m , Ammonium: 39 +
59 pg/m3; Urban dust (Fe, Al, Ca, Si): 18 + 6 pg/m3
Time to Analysis: Sensitized to OVA 3 day.
Challenged with OVA or saline 2wk later for 3 day.
Exposed to CAPs 8h/d, 13d. OVA or saline
challenge 9 day after first challenge. Sacrificed 24 h
after last CAPs exposure.
CAPs enhanced the effects of OVA by causing
differential expression in genes primarily
involved in inflammation and airway remodeling.
CAPs exposure alone had no effect on gene
expression. CAPs+OVA also increased IgE,
mucin glycoprotein, and BALF total protein, and
caused a more severe bronchopneumonia,
increased mucus cell metaplasia/hyperplasia
and mucosubstances.
Reference:
Hiramatsu et al.
(2003, 1558461
Species: Mouse
Gender: Female
Strains: BALB/c
and C57BL/6
Age: 8 wk
Weight: 17-22 g
DE -DE (generated by diesel engine and Route: Inhalation
diluted with filtered clean air)
Particle Size: NR
,
Dose/Concentration: Low -0.1 mg/m
High -3 mg/m3
Time to Analysis: 7 h/day, 5 days/wk, 1 or 3 mo
Lung Histopathology: DEP-laden
macrophages accumulated in the alveoli and
peribronchial tissues in a dose- and duration-
dependent manner in both strains. Lymphocytes
and neutrophils increased in both strains, but
were greatest in the BALB/c mice.
BALF and Mac-1 Positive Cells: BALT
formation in DEP-laden AMs was seen at the
high dose group and was greater in the BALB/c
mice. Mac-1 positive cells, a marker for
phagocytic activation of the AMs, was observed
in the high dose groups of both strains at 1 and
3 mo, and in the low dose group at 1 mo. in
BALB/c mice.
Cytokine and iNOS mRNA expression:!
month of exposure increased TNF-a, IL-12p40,
IL-4 and IL-10 mRNA in a dose-dependent
manner. IL-1B and iNOS decreased in a dose-
dependent manner. IFN-y mRNA expression
increased in BALB/c mice and decreased in
C57BL/6 mice. Similar results were seen at 3
mo, except IL-4 and IFN-y mRNA expression
decreased in the BALB/c mice. In C57BL/6
mice, IL-4 and IL-10 mRNA increased at the low
dose but decreased at the high dose. NF-KB
activation occurred after 1 wkand 1 month DE
exposure and was more prevalent in BALB/c
mice.
December 2009
D-130
-------
Study
Pollutant
Exposure
Effects
Reference:
Hiramatsu, (2005,
0882851
Species: Mouse
Gender: Female
Strains: BALB/c
Age: 8 wk
Weight: 17-22 g
DE (generated by diesel engine and
diluted with filtered clean air.)
Mycobacterial Infection -M.tuberculosis
(ATCC35812) Kurono strain
Particle Size: NR
Route: Inhalation
Dose/Concentration: Low-0.1 mg/m3
High-3mg/m3
Mycobaterial infection: 5 ml (nebulized) of a 106
colony-forming units (CPU) suspension
Time to Analysis: 7 h/day 5 day/wk, 1, 2 or 6 mo.
Subset infected on last day of DE exposure. CPU
evaluation 7 wk postinfection.
Histopathological Observations: DEP-laden
AMs and DEPs in the alveoli and peribronchial
tissues increased in a time-dependent manner.
DE-exposed mice had a greater number of
mycobacterial lesions, which were
disseminated. Lesions in the control mice had
clear borders and consisted of epithelial cells
and lymphocytes. Tubercle bacilli and DEPs
coexisted in AMs. BALT was seen around DEPs
in the 2 and 6-month exposure groups.
Inflammation cells increased in a time-
dependent manner with respect to DE exposure.
Granulomatous Lesions in Lungs: 6-month
DE-exposed mice had a significantly higher
amount of gross lesions than the 6-month
control mice.
Mycobacterial Burden: CPU in lungs were
increased in DE-exposed animals but only the 6
month exposure resulted in statistically
significant increases (a ~4-fold increase over
control). CPU in spleen were not significantly
altered by DE exposure.
Cytokines and iNOS mRNA Expression:
Infected DE-exposed mice had time-dependent
increases of TNF-a, IL-1B, IL-12p40, IFN-yand
iNOS mRNAs compared to the infected control
mice. IL-12 mRNA expression decreased in
infected 6-month DE-exposed mice.
Reference:
Ichinose, T. et al.
(2003, 0415251
Species: Mouse
Gender: NR
Strains:
BALB/cAnN, ICR,
C3H/HeN
Age: 6 wk
Weight: NR
DE: DE generated by 3059cc 4-cylinder
diesel engine
Der f: Crude extract of D. farinae
Particle Size: 0.4 pm(MMAD)
Route: Inhalation
Dose/Concentration: 1. Air
2. DE only: 3.0 mg/m3
3. Air + Der f: 1 mg Der f
4. DE 3.0 mg/m3 + 1 mgDerf
Time to Analysis: DE: 12 h/day, 7 days/wk, 8wk
Der f: 2 wk intervals, 6 wk
Analyzed 3 days after last instillation
Light Microscopic Observations: DE
exposure caused the proliferation of nonciliated
cells and epithelial cell hypertrophy. Soot-
containing macrophages were found in the
alveolar tissue spaces. Accumulated
lymphocytes were present in the peribronchiolar
lymphoid tissue. Inflammatory cells and soot-
containing macrophages were found in the
submucosal layer and the vessel interstitium of
mice treated with DE+Der f in all strains.
DE+Der f treated C3H/He mice had
desquamated goblet cells.
Eosinophil Infiltration: DE treated C3H/He
mice had a slight eosinophil infiltration in the
submucosal layer. DE+Der f treated mice in all
strains had a slight to moderate eosinophil
infiltration.
Lymphocyte Accumulation: Lymphocytes
significantly increased in all strains under the
DE treatment as compared to the air+saline
treatment, and further increased under the
DE+Der f treatment.
Goblet Cell Proliferation: Little proliferation
was seen in all strains under the DE treatment.
DE+Der [caused a significant increase in
proliferation compared to air+Derf in ICR mice,
but a significant decrease in C3H/He mice.
Local Cytokine and Chemokine Expression
in Lung Tissue Supernatant: DE+saline
significantly increased MIP-1a in all strains.
MCP-1 also increased but not significantly.
DE+Der f increased IL-5, RANTES, eotaxin,
MCP-1 and MIP-1ain all strains as compared
with air+saline and air+Der f. IL-5 decreased in
C3H/He mice treated with DE+Der f compared
to air+Der f. IL-3 decreased in ICR and C3H/He
mice compared to air+saline.
Derf-specific Immunoglobulin Production in
Plasma: Increased production of lgG1 was
statistically significant in ICR and C3H/He mice
treated with DE+Der f as compared to air+Der f.
IgE was low in all strains.
December 2009
D-131
-------
Study
Pollutant
Exposure
Effects
Reference:
Ichinose et al.
(2004, 1803671
Species: Mouse
Gender: NR
Strains: BALB/c,
ICR and C3H/He
Age: 5 wk
Weight: NR
DEP: 2740cc 4-cylinder engine
D. farinae: crude extract
Particle Size: 0.4 pm(MMAD)
Route: IT Instillation
Dose/Concentration: 1. D. farinae: 1 pm in PBS
2. D. farinae + DEP: 1 pg in PBS + 50 pg mg DEP
Time to Analysis: 4 times at 2 wk intervals. Mice
examined 3 wk after last instillation
Histological Changes: Mice in all three strains
treated with DEP+D. farinae had a significant
recruitment of eosinophils, more proliferation of
goblet cells, and more eotaxin positive
macrophages in the alveoli than mice treated
with D. farinae alone.
Local Cytokine Expression in Lung Tissue
Supernatant: DEP+D. farinae induced
significant elevation of IL-5 in ICR and C3H/He
mice as compared to D. farinae alone.
Production levels of IL-4 and RANTES did not
correlate with the manifestations of allergic
airway inflammation induced by the D. farinae
treatment with or without DEP.
Cytokine Expression in Plasma: IL-5 in
C3H/He mice treated with DEP+D. farinae was
significantly higher than D. farinae alone.
RANTES was unaffected by the DEP treatment
in all strains.
D. farinae-specific Immunoglobulin
Production in Plasma: The adjuvant effect of
DEP on lgG1 production was observed in all
three strains, with C3H/H3 being statistically
significant. The production levels of lgG1
correlated with the manifestations of
eosinophilic airway inflammation by both
treatments. No adjuvant effect on IgE production
was observed.
Reference: Inoue PM-OC: Urban PM, collected for 1
et al. (2007,
0967241
Species: Mouse
Gender: Male
Strain: ICR
Age: 6 wk
Weight: 29-33 g
month during early summer, 2001 in
Urawa city Saitama, Japan
IPS
Particle Size: <2.0|jm
Route: IT Instillation
Dose/Concentration: Vehicle group: PBS
PM-OC group: 4 mg/kg of PM-OC
IPS group: 2.5 mg/kg of IPS
PM-OC+LPS group: combined administration of
PM-OC +LPS
Time to Analysis: Single, 24 h
Effects of PM-OC on LPS Related Lung
Inflammation: PM-OC alone did not
significantly increase the infiltration of
neutrophils, but LPS challenge showed a
marked increase in the number of neutrophils
compared with vehicle. Administration of LPS
combined with PM-OC significantly increased
the infiltration of neutrophils compared with LPS
administration alone.
Effects of PM-OC on Histological Changes in
the Lung: Combined treatment with PM-OC
and LPS resulted in enhanced neutrophilic
inflammation.
Effects of PM-OC on Pulmonary Edema
Related to LPS: LPS group compared with
vehicle group had a significant increase in lung
water. The combined administration of PM-OC
and LPS resulted in further increase in the lung
water compared with LPS administration alone,
however it was not statistically significant.
Effects of PM-OC on Protein Expression IL-
1B,MIP-1a,MCP-1andKC:The
concentrations of these molecules were below
the detection limits in the PM-OC group. LPS
treatment significantly increased the protein
levels of these molecules compared with the
vehicle treatment. In the PM-OC + LPS group all
concentrations, particularly KC, were smaller
than in the LPS group.
December 2009
D-132
-------
Study
Pollutant
Exposure
Effects
Reference: Inoue
et al. (2006,
0909511
Species: Mouse
Gender: Male
Strain: ICR
Age: 6 wk
Weight: 29-33 g
Carbon black (14 nm PrinteX 90; PrinteX Route: IT Instillation
25; Degussa, Dusseldorf, Germany)
Particle Size: 14 nm - 300 rrn
56 nm - 45 m2/g
Dose/Concentration: Vehicle group: PBS at pH7.4
IPS group: 2.5 mg/kg of IPS in vehicle
Nanoparticle groups: 4 mg/kg carbon black
nanoparticles (14 nm or 56 nm) in vehicle
IPS + nanoparticle group: combined administration
of carbon black and IPS in vehicle
Time to Analysis: Single, 24 h
Effects of Nanoparticles: Nanoparticles alone
increased number of total cells and neutrophils,
but not statistically significant. IPS exposure
significantly increased numbers for both groups.
Nanoparticles and/or IPS enhance pulmonary
edema.
Histology: Treatment with LPS+14 nm
nanoparticles markedly enhanced neutrophil
sequestration into the lung parenchyma
compared to IPS alone. LPS+56 nm
nanoparticles did not.
Cytokines: IL-1B level significantly greater for
both LPS+ nanoparticles groups. TNF-a was not
significantly altered among the experimental
groups.
Chemokines: Challenge with 14 nm
nanoparticles alone elevated the levels of all
chemokines without significance except for KC.
IPS alone and with both nanoparticle groups
caused significant increases in all chemokines..
Formations of 8-OHdG in Lung: IPS plus
nanoparticles resulted in intensive expression 8-
OHdG, strongest in LPS+14 nm nanoparticle
Plasma Coagulatory Changes: PT - no
change for any group. APTT - some change with
IPS and IPS + nanoparticle groups, fibrinogen
level significantly elevated after IPS and for
LPS+14 nm nanoparticle. APC decrease with
LPS (significant) and LPS + nanoparticle
groups. vWF increase with LPS (significant) and
LPS+14 nm (significant).
Reference: Inoue
et al. (2004,
0879841
Species: Mouse
Gender: Male
Strain: ICR
Age: 6 wk
Weight: 29-33 g
DEPs [4JB-1 type light-duty, four-
cylinder, 2.74 liter Isuzu diesel engine
(Isuzu Automobile Co., Tokyo Japan)]
Washed DEP and DEP-OC - extracted
with dichloromethane
Particle Size: NR
Route: IT instillation
Dose/Concentration: Vehicle group: PBS;
Washed DEP group: 4mg/kg of DEP; DEP-OC
group: 4mg/kg of DEP-OQLPS group: 2.5mg/kg of
LPS; Washed DEP+LPS group: combined
administration of washed DEP +LPS; DEP-OC+
LPS group: combined administration of DEP-OC +
LPS
Time to Analysis: 4 h
COX-1 mRNA: Slightly elevated in both washed
DEP and DEP-OC groups, but slightly
decreased in other groups compared to vehicle
group.
COX-2 mRNA: Slightly increased with DEP-OC,
increased with LPS, washed DEP + LPS and
DEP-OC + LPS groups compared to vehicle.
COX-2 in the DEP-OC + LPS decreased when
compared to the LPS only group.
Pulmonary Edema: Washed DEP + LPS group
showed a synergistic enhancement of
pulmonary edema and local expression of
proinflammatory chemokines (MCP-1, MIP-1a,
KC, IL-1B).
Reference: Inoue
et al. (2006,
0967201
Species: Mouse
Gender: Male
Strain: ICR
Age: 6-7 wk
Weight: 29-33 g
Carbon black (PrinteX 90; PrinteX 25;
Degussa, Dusseldorf, Germany)
Particle Size: 14 nm - 300 m2/g
56 nm - 45 m /g
Route: IT instillation
Dose/Concentration: Vehicle group: PBS
Ovalbumin (OVA) group: 1mg OVA; Nanoparticle
groups: 50 mg carbon black nanoparticles (14 nm
or 56 nm);;OVA+ nanoparticle group: combined
administration of nanoparticles and OVA
Time to Analysis: Vehicle group - weekly for 6wk
OVA group - biweekly for 6 wk
Nanoparticle groups - weekly for 6 wk
OVA+Nanoparticle group (same protocol as OVA
and Nanoparticle) studied 24 h after last
administration
Nanoparticles: Exposure to carbon
nanoparticles resulted in the lung expression of
TARC, GM-CSFand MIP-1a. The levels were
higher in the 14 nm group compared to the
56 nm group.
OVA: In the presence of OVA, nanoparticles
enhanced levels of TARC, GM-CSF, MIP-1a, IL-
2 and IL-10, with the effects seen more
prominently in the 14 nm particles + OVA group.
December 2009
D-133
-------
Study
Pollutant
Exposure
Effects
Reference: Inoue
et al. (2005,
0886251
Species: Mouse
Gender: Male
Strain: ICR
Age: 6-7wk
Weight: 29-33 g
Carbon black (PrinteX 90; PrinteX 25;
Degussa, Dusseldorf, Germany)
Particle Size: 14 nm - 300 m2/g
56 nm - 45 m2/g
Route: IT Instillation
Dose/Concentration: Vehicle group: PBS;
Ovalbumin (OVA) group: 1mg OVA; Nanoparticle
groups: 50mg carbon black nanoparticles (14nm or
56 nm); OVA+ nanoparticle group: combined
administration of nanoparticles and OVA
Time to Analysis: Vehicle group - weekly for 6 wk
OVA group - biweekly for 6 wk
Nanoparticle groups - weekly for 6 wk
OVA+Nanoparticle group: same protocol as OVA
and Nanoparticle studied 24 h after last
administration
Nanoparticles + OVA: Nanoparticles given with
OVA enhanced airway inflammation,
characterized by increased eosinophils,
neutrophils, mononuclear cells and goblet cells.
In addition, nanoparticles + OVA significantly
increased local expression of IL-4, IL-5, eotaxin,
IL-13, RANTES, MCP-1 and IL-6. The formation
of 8-OHdG was enhanced by nanoparticles +
OVA.
14 nm Nanoparticles: All these effects were
more prominent when 14 nm nanoparticles were
used. The 14 nm nanoparticle + OVA group
significantly raised levels of total IgE and
antigen specific production of lgG1 and IgE.
Reference: Inoue
et al. (2006,
1901421
Species: Mouse
Gender: Male
Strain: ICR
Age: 6 wk
Weight: 29-33 g
Whole DE (generated by 4-cylinder,
3.0591, Isuzu diesel engine, Isuzu
automobile, Tokyo, Japan)
IPS
Particle Size: 110 nm (peak particle
size)
Route: Whole-body Inhalation
Dose/Concentration: 0.3 mgsoot/m3
1.0mgsoot/m3
3.0 mg soot/m
LPS:125mg/kg
Time to Analysis: IPS prior to
12 h exposure to exhaust
BAL fluid, total cells, neutrophils, protein and
gene levels (MCP-1 and KC) decreased
compared to control with IPS, but were smaller
with IPS + DE. Results are suggestive that
short-term exposure to DE does not exacerbate
LPS-related lung inflammation.
Reference: Inoue
et al. (2007,
0967021
Species: Mouse
Gender: Male
Strain: ICR
Age: 6 wk
Weight: 29-33 g
Cell Type
Splenocytes
DEPs [4JB-1 type light-duty, four-
cylinder, 2.74 liter Isuzu diesel engine
(Isuzu Automobile Co., Tokyo Japan)]
IPS
Particle Size: PM2 5
Route: Cell Culture (Splenocytes resuspended to
cell density of 1 xl06/mL and 1000 ml applied into
each of 12-well plate)
Dose/Concentration: DEP: 100 mg/mL; IPS: 1
mg/mL; LPS(1 mg/mL) + DEP (1,10 or 100 mg/mL)
Time to Analysis: 72 h
Cell viability: No effect.
Mononuclear cell response: Incubation with
DEP alone inhibited basal cytokine production.
LPS significantly increased protein levels of IFN-
Y, IL-2, and IL-10 compared to control. DEP
suppressed the LPS-enhanced protein levels in
a dose-dependent manner and moderately
elevated the IL-13 level.
Reference: Inoue
et al. (2007,
1988851
Species: Mouse
Gender: Male
Strain: ICR
Age: 6-7 wk
Weight: 20-30 g
Carbon nanoparticles (PrinteX 90,
PrinteX 25; Dusseldorf, Germany)
OVA
Particle Size: CB14 = 14 nm, CB56
= 56nm
Route: IT Instillation
Dose/Concentration: 50 pg and/or 1 pg OVA in
PBS
Time to Analysis: 1 */wk for 6 wk; sacrifice 24 h
after last exposure
Lung Responsiveness: Respiratory system
resistance, Newtonian resistance and tissue
dampening were significantly higher in the
nanoparticle + OVA groups. Elastance and
tissue elastance were higher in these groups but
not significantly so. Compliance was
significantly lower in the nanoparticle + OVA
groups compared to the control.
Lung mRNA Level for MucSac: Levels were
significantly higher in nanoparticle + OVA groups
compared to the control.
Reference: Inoue
et. al. (2007,
0966921
Species: Mouse
Gender: Male
Strain: ICR
Age: 6-7 wk
Weight: 29-34 g
DEP-OC collected from 4JB1 type, light
duty, 4 cylinder, 2.74 liter Isuzu diesel
engine, Isuzu Automobile Company,
Tokyo, Japan)
OVA
Particle Size: 0.4 |jm
Route: IT Instillation
Dose/Concentration: 50 pg and/or 1 pg OVA in
PBS
Time to Analysis: DEP or DEP-OC w/ or w/o OVA
initially; OVA or vehicle every 2 wk for 6 wk; DEP
components or vehicle 1 x/wk for 6 wk; sacrifice
24 h after last instillation
Total respiratory system resistance, elastance,
Newtonian resistance, tissue damping, tissue
elastance displayed general positive trends and
were significantly higher in OVA and OVA +
DEP-OC groups. Compliance displayed a
general negative trend and was significantly
lower in the washed DEP + OVA group.
December 2009
D-134
-------
Study
Pollutant
Exposure
Effects
Reference: Ito et
al. (2006, 0883911
Species: Rat
Cell Line: L2 cells
of alveolar
epithelial cell type
II origin
DEP - generated from 2982-cc common
rail direct injection diesel engine with
oxidation catalyst and exhaust gas
recirculation system.
Particle Size: PM2 5
Route: Cell Culture
Dose/Concentration: 1*106
1,10or30mg/mL
Time to Analysis: 3 h
ICAM-1 and LDL Receptor mRNA: Up-
regulation in a dose-dependent manner.
Statistically significant at 30 mg/mL compared to
control.
HO-1 and PAF Receptor mRNA: Up-regulation
in dose-dependent manner and statistically
significant at all doses compared to control.
Correlation Between HO-1 and ICAM-1, LDL,
and PAF: Significant correlation between HO-1
and each of these.
Reference: Jang
et al. (2005,
1553131
Species: Mouse
Gender: Female
Strains: BALB/c
Age: 5-6 wk
DEP -generated from 4JB1 type, light
duty, four-cylinder diesel engine (Isuzu
Automobile, Co, Tokyo, Japan)
03 - (generated with Sander Model 50
ozonizers, Sander, Eltze Germany)
OVA
Particle Size: NR
Route: Whole-body Inhalation
Dose/Concentration: DEP: 2,000 pg/pL (sic)
03: 2 ppm (avg 1.98 + 0.08 ppm)
OVA sensitization: 10 mg
Time to Analysis: OVA sensitization,
DEP, 03 and OVA Challenge on d21- 23
Exposed to 03 for 3 h and DEP for 1 h
AH and BAL measured 1 day after last challenge
Airway Responsiveness: OVA + 03 + DEP
exposure group had significantly higher
methacholine-induce Penh than sham group or
OVA group.
Total cells, proportion of eosinophils and
neutrophils: The OVA + 03 + DEP group was
significantly higher than OVA group and OVA+
03 group.
\L-4: OVA + 03, OVA + DEP and OVA + 03 +
DEP IL-4 level increased compare to OVA
group.
IFN-y: Levels significantly decreased in OVA +
DEP and OVA + 03 + DEP compared to OVA +
03.
Reference:
Jaspers et al.
(2005, 0881151
Species: Human
Cell Lines: A549
cells, primary
human bronchial
and nasal epithelial
cells
DEas: aqueous-trapped solution of DE
(emissions from Caterpillar diesel
engine, model 3304)
Influenza: A/Bangkok/1/79 (H3N2
serotype)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: Influenza: 3x105 cells
infected with 320 hemagglutination units (HAU)
DEas: For A549 cells: 6.25,12.5, 25 pg/cm2. For
bronchial and nasal cells: 22 or 44 pg/cm2.
Time to Analysis: 2 h incubation with DEas then
virus added.
HA RNA levels analyzed at 0,15, 30, 60 or 120 min
post infection.
IFN and MxA responses: analyzed 24 h post
infection.
Fluorescence: some cells treated with GSH-ET 30
min before DEas exposure. Measured 2 h post-
influenza infection.
A649 Cells Increased Susceptibility: DEas
enhances HA RNA levels in A549 cells in a
dose-dependent manner. 25 pg/cm2 significantly
enhanced levels in A549 cells compared to the
influenza-infected controls. Viral protein levels
were increased in A549 cells. Exposure to DEas
increased the number of influenza-infected
epithelial cells in A549 cells.
Human Nasal and Bronchial Cells
Susceptibility: Exposure to DEas increased HA
RNA levels in the nasal and bronchial cells.
Statistically significant at 22 pg/cm2 for nasal
cells and approaching significance at 44 pg/cm
for bronchial cells. Exposure of both types to 44
pg/cm2 enhanced viral protein levels.
Influenza Induced IFN Response in A649:
Exposure to DEas does not suppress but
enhances IFN-B mRNA levels. Treatment
enhanced influenza-induced nuclear levels of
both phospho-STAT-1 and ISFGSg. ISRE-
promoter activity was enhanced, but not
significantly. Treatment enhanced myxovirus
resistance protein (MxA) mRNA levels. This data
suggest that DEas exposure enhances influenza
virus replication without suppressing production
of IFN-B or IFN-B-inducible genes.
Influenza Induced IFN Response in Human
Nasal and Bronchial Cells: Exposure to DEas
increased IFN-B and MxA levels.
Oxidative Stress in A649: DEas exposure
dose-dependently increases oxidative stress in
A549 cells within 2-h post-exposure. Add the
antioxidant GSH-ET and it reverses the effect.
Pretreatment with GSH-ET A549 cells reversed
the effects of DEas on the number of influenza-
infected cells, and reduced HA RNA levels.
Oxidative Stress in Human Bronchial Cells:
The results were the same asA549 cells
pretreated with GSH-ET. Or Pretreatment with
GSH-ET also reversed effects of DEas on HA
RNA levels.
December 2009
D-135
-------
Study
Pollutant
Exposure
Effects
Reference: Kaan
and Hegele (2003,
0957531
Species: Guinea
pig
PMio - EHC-93 obtained (Environmental
Health Canada, Ottawa, ON, Canada)
RSV - Human RSV (long strain/lot!8D)
(American Tissue Culture Collection,
Bethesda, MD)
Gender: Female Particle Size: PMio (0.35 pm MMAD)
Strain: Cam
Hartley
Age: 22-29 days
Weight: 250-300 g
Cell Types: AM
Route: Cell Culture
Dose/Concentration: PM10: 500 pi/well (100 pg/ml
MEM)
RSV exposure:: 1 ml/well (6x106 pfu/ml MEM)
Groups: PM10+RSV
RSV+PM,o
RSV only
PM10only
negative control
Time to Analysis: PM10 - 60 min; RSV - 90 min
Parameters measured 24 h post treatment
Interaction on Phagocytic Ability of AM: Not
affected by sequential exposure to RSV and
PMi0. More than 95% of AM exposed to PM,0
engulfed PM. AM exposed to PMio showed
significant increase in mean side scatter in
comparison to negative control and RSV-
infected AM. No significant difference between
AM exposed only to PM10 and AM exposed to
both agents. No significant side mean side
scatter difference between AM exposed to PM
only and to both agents.
Interaction on RSV Immunopositivity: PIvl-;
exposure inhibits. All RSV-treated groups
showed significantly greater proportion of RSV-
immunopositive cells compared with negative
control. PM10+RSV showed significantly smaller
proportion of RSV-immunopositive cells
compared with RSV group. RSV+PMio group
similar to RSV group. Proportion of RSV-
immunopositive AM was influenced by the
sequence of exposure to RSV and PMio.
Interaction on RSV Replication: PM exposure
suppressed RSV replication. AM exposed to
both agents produced 3 to 9 fold less RSV
progeny compared with RSV alone group.
Quantity of RSV progency was not significantly
affected by the sequence of exposure RSV and
PMio. Negative control and PMio only did not
propagate progeny.
Interaction of RSV Yield: RSV alone group
produced the highest RSV yield, those exposed
to both agents, independent of sequence,
showed a 5-fold decrease.
Cytokine production: RSV infection stimulated
all three cytokines measure (IL-6, IL-8 and TNF-
a) compared to negative control. IL-6: PM10
significantly reduced RSV-induced IL-6
production. IL-6 was affected by the sequence
of exposure to PM,o and RSV (PMio+RSV vs.
RSV+ PM10). IL-8: PM10 significantly decreases
RSV-induced IL-8 production and baseline. No
affect on sequence of exposure. TNF-a:
production was increased when exposed to
RSV, PM10 or a combination of both agents. No
differences among treatments.
Reference:
Kleinman et al.
(2005, 0878801
Species: Mouse
Gender: Male
Strains: BALB/c
Age: 8-19 wk
Weight: NR
CAPS: fine (F) and ultrafme (UF) using
VACES system; performed a 2 sites in
Los Angeles, CA, one 50-m downwind
and another 150-m downwind from a
complex of three roadways, State Road
CA60, Interstate 10, and Interstate 5
F CAPS in 2001 and 2002, UF CAPS in
2002 only
OVA: Ovalbumin
Particle Size: UF: dp < 150 nm
F: dps2.5pm
Route: Whole-body Inhalation
OVA sensitization: nasal instillation
OVA challenge: inhalation
Dose/Concentration: UF at 50 m: 433 pg/m3
-UFat150m:283|jg/m3
F at 50 m or 150 m: average 400 pg /m3
OVA sensitization: 50 pg/5 pi
OVA challenge: 30 mg/m3
Time to Analysis: CAPS: 4 h/day, 5 days/wk for 2
wk
Sensitization: On morning of each exposure
1st Challenge: week after 10 days of treatment
2nd Challenge: one week following 1st challenge
Sacrificed: 24 h after 2nd challenge
There were significantly higher concentrations of
IL-5, IgE, lgG1 and eosinophils in mice exposed
to either CAPS compared to air. Mice exposed
to CAPS at 50-m downwind showed higher
levels of IL-5, lgG1, and eosinophils than those
exposed to CAPS 150-m downwind.
December 2009
D-136
-------
Study
Pollutant
Exposure
Effects
Reference:
Kleinman et al.
(2007, 0970821
Species: Mouse
Gender: NR
Strains: BALB/c
Age: 6-8 wk
CAPS - concentrated fine (F) and
ultrafine (UF) using VACES system -
performed a 2 sites in Los Angeles, CA,
on 50-m downwind and another 150-m
downwind from State Road CA60 and
Interstate 5. Fall 2001-summer 2004
OVA
Particle Size: F: PM25; UF: PM0.15
Route: Whole-body Chamber
Dose/Concentration: 50 m - F: 394 + 94 pg/m3
50 m - UF: 297 ± 189 pg/m3
150 m - F: 387 ± 68 pg/m3
150m-UF:213±95|jg/m3
OVA - 50 mg in 5 ml saline
Time to Analysis: 3, 4 h/day, 5 days/wk, 2wk
OVA the morning of each exposure
60m Site: higher levels and statistically
significant concentration curves of IL-5 and lgG1
in F-CAP mice at the 50 m site.
160m Site: in no cases were responses greater
than the 50m or control groups.
F vs. UF: The study was not able to differentiate
between the effects of F PM and UF PM
exposures.
Reference: Klein-
Patel et al. (2006,
0970921
Species: Cattle
and Human
Cell Types
Bovine tracheal
epithelial cells
(BTE)andA549
ROFA
V205, VOS04, Si02 Ti02, Fe2(S04)3,
NiS04, IPS
Particle Size: 1.95 pm(MMAD)
Route: Cell Culture
Dose/Concentration: ROFA: 0, 2.5, 5,10,15, 20
pg/cm2
IPS: 100 ng/mL
V205: 0, 0.15, 0.3, 0.61, 1.25, 2.5, 5, 10, 20 pg/cm2
NiS04 Fe2(S04)3, Ti02, Si02: 0, 1.23, 2.5, 5, 10, 20
pg/cm
VOS04: 0, 0.145, 0.29, 0.58, 1.16, 2.32 pg/cm2
Time to Analysis: IPS: 0, 6, or 18 h
ROFA: 0, 2,4,6 h
V205:0, 0.25, 0.5, 1,2, 4,6, 8h
NiS04, Fe2(S04)3, Ti02, Si02, VOS04: 6 h
ROFA in BTE: ROFA and ROFA leachate
inhibition of LPS-induced TAP gene expression
increases with exposure time and dose. Washed
particles of ROFA at doses 2.5 to 10 mg/cm
significantly increased inducible TAP
expression.
Soluble Metals in BTE: V205 inhibition of IPS
and IL-1|3 induced TAP gene expression
increases with exposure time and dose. NiS04
exhibits non-significant dose dependent
suppression of inducible TAP gene expression.
Fe2(S04)3, Ti02 and Si02 were found to have no
effect.
A549: Results with ROFA and V205 in BTE were
replicated using the A549 cell line and IL-1|3to
induce hBD2 gene expression.
Cellular Viability: Was not significantly affected
in ROFA doses below 20 pg/cm2 and
V205/VOS04 doses below 2.5 pg/cm2.
Reference: Koike
and Kobayashi
(2005, 0883031
Species: Rat
Gender: Male
Strains: Wstar
Kyoto
Age:8-10wk
Weight: 280-350 g
Cell Types: AM,
PBM (peripheral
blood monocytes),
T-cells (antigen
sensitized)
Whole DEP: Diesel Exhaust Particles
collected in the dilution tunnel of a diesel
inhalation facility. (Ratio of organic
extract to residual particles in the whole
DEP was 3:1.)
Organic extract of DEP
Residual particles of DEP
OVA: Ovalbumin
Particle Size: NR
Route: Cell Culture (1 xio6 cells/ml)
Dose/Concentration: Whole DEP: 10, 30, 100
pg/mL
Organic extract of DEP: 7.5, 22.5, 75 pg/mL
Residual particles: 2.5, 7.5, 25 pg/mL
Time to Analysis: 24 h post exposure
la Antigen and Costimulatory Molecules:
Most control AM did not express these
molecules. Whole DEP did not cause any
increase in expression level. 20% of control
PBM expressed la and 10% B7; expression of
these molecules was significantly increased by
whole DEP. Organic extract significantly
increased the expression of la and B7
molecules on PBM similar to whole DEP.
Residuals caused no effect. Organic extract-
induced expression of la antigen in PBM was
reduced by treatment with MAC.
AP Activity: After exposure to organic extract, T
cell proliferation was significantly increased by
the addition of control PBM in a cell number-
dependent manner. AP activity of PBM was
increased over control by exposure to 3 pg/mL
organic extract, although higher concentrations
suppressed the activity of PBM.
Cytokine Production: Organic extract
treatment of PBM decrease IFN-y production
from T-cells stimulated by PBM. No significant
effect on IL-4 observed.
HO-1 Protein Level: Levels in PBM were
significantly increased by exposure to whole
DEP or organic extract. Levels induced by
organic extract was diminished by MAC
treatment.
December 2009
D-137
-------
Study
Pollutant
Exposure
Effects
Reference: Last et PM - aerosol of soot and iron oxide
al. (2004, 0973341 OVA
Species: Mouse
Gender: NR
Strains: BALB/c
Age: 6 wk
Weight: 16-20 g
Particle Size: PMni-PM,
Route: Inhalation
OVA - Intraperitoneal Injections; Aerosol Exposure
Dose/Concentration: PM - 235-256 pg/m3
OVA-10|jg/0.1 ml injection
OVA aerosol -10 ml of 10 mg/mL (1%) solution
Time to Analysis: PM: 4 h/day, 3 days/wk; OVA: 2
ip injections days 1 and 15. Aerosol on day 28 after
first ip; 60 min 3x/wk
2 Wk PM Exposure/4 Wk OVA Aerosol
Treatment: The OVA alone group had
significantly more airway collagen than the PM
alone group. Histology showed significantly
more collagen in the treatment than the air
alone group. There was a significantly greater
amount of goblet cells than the OVA alone
group.
4 Wk OVA Aerosol/ 2 Wk PM Treatment: The
OVA treatment had significantly more goblet
cells than the PM alone group.
6 Wk Concurrent PM and OVA Treatment:
Significantly more cells were observed in the
OVA alone group over the treatment. The
treatment had significantly more lymphocytes
and significantly less macrophages than groups
exposed to PM before or after OVA. Histology
showed significantly more collagen in the
treatment than the air or PM alone groups. The
treatment had significantly more goblet cells
than the OVA alone group.
Reference: Li et DEP (2369-cc diesel engine
al. (2007, 0931561 manufactured by Isuzu Motor, operated
at 1050 rpm, 80% load, commercial light
Species: Mouse
Gender: Female
Strain: BALB/c,
C57BL/6
Age: 9 wk
Weight: NR
oil)
Particle Size: NR
Route: Inhalation
Dose/Concentration: DEP: 103.1 + 9.2 pg/m3,
CO: 3.5 + 0.1 ppm, N02: 2.2 + 0.3 ppm, S02:
O.01 ppm
Time to Analysis: Protocol 1: Exposed 7h/day,
5days/wk. Sacrificed at day 0, week 1, 4, 8.
Protocol 2: DE alone or DE+NAC 7h/d, 1-5 days.
Airway Hyperresponsiveness: Penh values
increased in BALB/c mice compared to the
control at day 0, but no significant changes
occurred after this time. Penh values increased
in C57BL/6 mice at 1wk compared to the control
but returned to control levels at 8 wk.
BALF: Compared to the other strain, the total
number of cells and macrophages increased
significantly at 1wk in C57BL/6 mice and at 8wk
in BALB/c mice. Neutrophils, lymphocytes,
MCP-1, IL-12, IL-10, IL-4, IL-13 increased
significantly for both strains. No eosinophils
were found. IL-1 B and IFN-y increased
significantly in BALB/c mice compared to
C57BL/6 mice.
HO-1 mRNAand Protein: HO-1 mRNA was
more marked in BALB/c mice at 1wk and
C57BL/6 mice at 4 and 8 wk. HO-1 protein
percentage changes from the control were
greater in BALB/c mice at 1wk and C57BL/C
mice at 8 wk.
MAC: NAC inhibited the increased Penh values,
total number of cells and macrophages in
C57BL/6 mice at 1 wk and neutrophils and
lymphocytes in both strains.
Reference: Li et
al. (2009, 1904571
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 6-8 wk
Weight: NR
CAPs (downtown Los Angeles, CA from
major freeway, traffic mainly passenger
cars and diesel trucks; Jan. 2007 or
Sept. 2006)
Ultrafine carbon black (UFCB; used as
control)
Particle Size: Fine- <2.5 pm (diameter),
UF-O.15|jm (diameter)
Route: Intranasal Instillation
Dose/Concentration: 0.5 pg PM in 50 pL
suspension
Time to Analysis: Day 1 exposed to PM or saline.
Day 2 exposed to PM+OVA or OVA or saline alone.
Repeated on days 4, 7, 9. Different experiment:
NAC ip injected 4 h pre-instillation on days 1, 2, 4,
7, 9. All animals rested and OVA aerosol
challenged 30 min on days 21, 22. Sacrificed day
23.
UFP alone had no effect on the lung. UFP+OVA
significantly increased eosinophils, and OVA-
specific lgG1 and IgE. The induction of
eosinophils and lgG1 were inhibited by NAC.
Generally, UFP+OVA mice had greater signs of
inflammation than the other groups as
determined by pulmonary histopathology and
airway morphometry. UFP had a greater PAH
content than fine particles. UFP significantly
increased IL-5, IL-13, TNF-a, IL-6, KC, MCP-1,
and MIP-1a.
Reference: Li et
al. (2009, 1904571
Species: Mouse
Cell Line: RAW
264.7
CAPs (downtown Los Angeles, CA from
major freeway, traffic mainly passenger
cars and diesel trucks; Jan. 2007 or
Sept. 2006)
Ultrafine carbon black (UFCB)
Particle Size: Fine- <2.5 pm (diameter),
UF-0.15pm (diameter)
Route: Cell Culture
Dose/Concentration: 1, 5, 8.3, 10 pg/mL
Time to Analysis: NR
UFP induced greater HO-1 expression than fine
particles. The higher PAH content of UFP
correlated with HO-1 expression.
December 2009
D-138
-------
Study
Pollutant
Exposure
Effects
Reference: Liu et
al. (2008, 1567091
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 11wk
Weight: NR
DEP (5500-watt single-cylinder diesel
engine generator (Yanmar, Model YDG
5500E), 406 cc displacement air-cooled
engine, Number 2 Diesel Certification
Fuel, 40 weight motor oil)
Particle Size:
^m
Route: Intranasal Exposure
Dose/Concentration: Average particle
concentration: 1.28 mg/m3
Time to Analysis: Four groups: saline+air control,
saline+DEP, A. fumigatus+air, A.fumigatus+DEP. A.
fumigatus exposure every 4 days for 6 doses. DEP
exposure 5 h/dayfor 3 wk concurrent with A.
fumigatus exposure.
A.fumigatus+DEP increased IgE, the mean BAL
eosinophil percentage, goblet cell hyperplasia,
and eosinophilic and mononuclear cell
inflammatory infiltrate around the airways and
blood vessels compared to the A. fumigatus or
DEP treatments. A.fumigatus+DEP also caused
methylation at the IFN-y promoter sites CpG-53,
CpG-45, and CpG-205.
Reference: Liu et
al. (2007, 0930931
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 11wk
DEP: 5500-watt single-cylinder diesel
engine.
Particle Size: NR
Route: Inhalation
Dose/Concentration: Average particle
concentration 1.28 mg/m3.
Time to Analysis: 1. Aerosol vehicle (saline) + air
2. Aerosol vehicle (saline) + DEP
3. A. fumigatus + air
4. A. fumigatus + DEP
A. fumigatus: 62.5 pg aerosolized protein extract in
50 pL PBS; 6 total doses, every 4 d.
DEP exposure 5 h/day 3wk concurrent with A.
fumigatus.
IgE Production: IgE production increased with
the A.fumigatus treatment and increased further
with the A.fumigatus and DEP treatment.
Histopathology: A. fumigatus with DEP caused
an increase in goblet cell hyperplasia and
eosinophil and mononuclear cell infiltrate around
the airways and blood vessels as compared to
the control and DEP treatments.
Gene Methylation: Greater methylation at the
CpG-53 site of the IFN-y promoter occurred
under the A. fumigatus + DEP treatment
compared to the A. fumigatus or DEP
treatments. The DEP treatment did not induce
methylation. Methylation correlated with
increased IgE and hypomethylation with
decreased IgE. Hypomethylation occurred in the
IL-4 promoter under the A. fumigatus + DEP
treatment.
Reference:
Lundborg et al.
(2007, 0960401
Species: Rat
Gender: Male
Strains: SD
Age: NR
Weight: 300-400 g
Cell Line: AM
Carbon-Black Particles (93% C)
DEPs (97% C) - toluene-extracted
10-fold Cr, Mn, N; 50-100 fold Al, Cd,
Cu, Fe, Mg, Pb, Zn more in DEP
aggregates
Particle Size: Carbon aggregates: 0.17
+ 0.08 pm (mean diameter)
Diesel Particles: 0.69 + 0.46 pm (mean
diameter)
Primary particles: 0.044 + 0.01 pm
(mean diameter)
Route: Cell Culture (0.5x106 AM/well)
Dose/Concentration: 20 |jg/mL
surface area: 159 + 4m2/g
Time to Analysis: 6 different experiments. AM pre-
exposed to carbon or washed DEP. Loaded with
particles. Incubated with S. pneumoniae, ATCC
strain or clinical isolates.
Effect of Time on Survival of S. Pneumoniae
when Incubated with Carbon Loaded AM:
Loading AM with carbon significantly increased
the bacterial survival. Bacteria opsonization
decreased bacterial survival.
Effect of Carbon Load in AM on Survival of
S. Pneumoniae: Bacterial survival increased in
a dose-dependent manner as the carbon
particle load of AM increased.
Survival of S. Pneumoniae after Incubation
with Carbon or Washed Diesel Loaded AM:
Bacterial survival increased in carbon loaded
AM compared to the control. No difference
existed with the washed diesel particles.
Survival of the ATCC Strain and Clinical
Isolates of S. Pneumoniae when Incubated
with Carbon Loaded AM or Control AM:
Carbon significantly increased the CFU of
opsonized and unopsonized bacteria for the
ATCC strain and clinical isolates.
Ability of carbon or washed diesel loaded
AM, incubated with the ATCC strain of S.
pneumoniae, to induce LPO of lung
surfactant: A 97% increase in the surfactant
LPO occurred after incubation with washed
diesel loaded AM compared to control AM. The
effect of washed diesel particles was
significantly greater than that of carbon
particles.
LPO by carbon loaded AM incubated with the
ATCC strain or clinical isolates in the
presence of absence of surfactant: LPO
induced by AM increased when incubated with
carbon loaded AM compared to control AM.
December 2009
D-139
-------
Study
Pollutant
Exposure
Effects
Reference:
Matsumoto et al.
(2006, 0980171
Species: Mouse
Gender: Female
Strains: BALB/c
Age: 6 wk
Weight: 15-20 g
DE (collected from a 2369 cm diesel
engine operated at 1050 rpm and 80%
load with commercial light oil; engine
exhaust passed through a particulate air
filter and charcoal filer) Diluted DE
introduced into the exposure chamber.
Composition of the DE: 3.5 + 0.1 ppm
CO, 2.2 + 0.3 ppm N02, <0.01 ppm S02
and 103.1 ± 9.2 pg/m3 DEP
Particle Size: NR
Route: Whole-body Inhalation
Dose/Concentration: 100 pg/m3 DE
Time to Analysis: Mice were initially sensitized w/
OVA (20ug absorbed to 2 mg alum diluted with 0.5
ml saline) via ip injection on day 0, 6 and 7. Two
wks later the mice were challenged with OVA
(0.1 mg in 0.1 ml saline) intranasallyonday21.
DEfor1dor1,2, 3, 4 or 8 wk (at 7 h/day for 5
days/wk).
Airway Hyperresponsiveness: Exposure to
DE significantly increased airway reactivity to
methacholine after 1 wk in both 24 and 48
mg/mL Mch and after 4 wk in the 48 mg/mL DE
exposure caused an increase in airway
sensitivity after 1 wk of exposure, 4 wk and 8 wk
of exposure did not result in a significant
increase.
BAL Cells: The total cell count was increased
after 1 wk of DE exposure. This increase was
mostly due to an increase in eosinophils. After 1
wkthe total cell count dropped drastically even
after continuous exposure to DE. DE did not
effect the number of CDS, CD4, CDS or NK1
cells at any point in time.
Cytokine/Chemokine mRNA Levels: DE
exposure on day 1 caused an increase in mRNA
levels of IL-4, IL-5 and IL-13 when compared to
the control mice but longer periods of DE
exposure failed to cause an increase. Protein
levels of IL-4 were significantly elevated at
compared to control at day 1, but did not persist
with time. mRNA levels of MDC were increased
at 1 wk of exposure (compared to control) but
also decreased at time periods after. mRNA
levels of RANTES were increased at 2 and 3 wk
after exposure and remained elevated at 4 wk
but not significantly. The level of RANTES
protein increased as the weeks went along, but
increased significantly only at 8 wk.
Histopathology: OVA sensitization caused an
increase peribronchial and perivascular
infiltration of inflammatory cells which peaked at
1 wk after exposure and decreased afterward.
DE exposure did not cause/show any additional
signs of inflammation.
Reference:
Morishita et al.
(2004, 0879791
Species: Rat
Gender: Male
Strain: Brown-
Norway
Age: 10-12 wk
CAPs (generated from ambient air in an
urban Detroit community).
Particle Size: 0.1-2.5pm
Route: Whole-body Inhalation
Dose/Concentration: July 676 pg/m3' September
313|jg/m3
Time to Analysis: First rats were sensitized (days
1-3) and challenged (days 14-16) with saline
(control) or OVA by intranasal instillation (5% in
saline, 150 pL/nasal passage).
4 days after the last intranasal challenge, rats
began exposure in the chambers. Exposures were
10 h long. The July exposure was for 4 consecutive
days. The September exposure was for 5
consecutive days.
Recovery of Trace Elements in Animal Lung
Tissues: July Exposure-Anthropogenic trace
elements were below limit of detection in
pulmonary tissue of animals exposed to July
CAPs. September Exposure- Several elements
were recovered from pulmonary tissue during
the Sept. exposure. La concentrations were
increased in both control/CAPs exposure and in
the OVA/CAPs exposure groups. V
concentration was increased in OVA/CAPs
exposed animals but not in rats exposed to just
CAPs. S content was only significant in animals
exposed to OVA/CAPs compared to the non-
exposed control.
Particle Characterization: July PM had an
average mass concentration twice as high as
the September mass concentration. S
concentration was four-folds higher in July PM.
In the September PM- the concentration of La
was 12.5 fold higher than in July PM, V was 2.7
fold higher than in July PM and Mnwas 1.5 fold
higher than in July PM.
BALF Analysis: Eosinophil concentration was
not significantly different when comparing rats
exposed to CAPs only in either July or
September (this was explained by the elapsed
time between exposure and BALF collection).
However OVA and CAP exposure in the
September group led to elevated eosinophil
levels. Similarly, the protein content was only
significantly increased in the September
OVA/CAP exposed rats, compared to the control
group.
December 2009
D-140
-------
Study
Pollutant
Exposure
Effects
Reference:
Nygaard et al.
(2005, 0886551
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 6-7 wk
Coarse and fine ambient air particles
collected in Rome (spring), Oslo (1-
summer, fine only, 2- following spring,
fine and coarse), Lodz (summer) and
Amsterdam (spring). These represent
areas with high population and
dominance of traffic.
DEP (Standard reference material
1650a)
Particle Size: Fine PM 0.1-2.5 pm;
Coarse PM 2.5-10 pm
Route: Subcutaneous Injection into mouse
footpads.
Dose/Concentration: 100 |jg of particle
Time to Analysis: Animals were in eight groups: 1.
Control- Hank's Balanced Salt Solution
2. OVA- 50 ug 3. OVA (50 pg)+ Amsterdam Coarse
PM (100 pg) 4. OVA (50 pg)+ Amsterdam Fine PM
(100 pg) 5. OVA (50 pg)+ Lodz Coarse PM (100
pg) 6. OVA (50 pg)+ Lodz Fine PM (100 pg) 7. 5.
OVA (50 pgj+ Oslo Coarse PM (100 pg) 8. OVA (50
pg)+ Oslo Fine PM (100 pg)
Analysis 5 days after injection.
Cell Numbers and Cell Phenotypes in the
Lymph Node: The overall number of B
lymphocytes, lymph node cells, PLN cells, and
the expression of MHC class II, CD86 and CD23
on B lymphocytes were increased by
coexposure of OVA+ the particles compared to
the OVA or particle groups alone. The OVA +
particle groups displayed a significant decrease
in T lymphocytes. Particles only significantly
increased the number of lymph node cells and
MHC Class II expression. There were no
differences observed between coarse and fine
PM fractions.
Cytokine Production by Lymph Node (ex
vivo culture of popliteal lymph node cells):
The OVA + particle (DEP and Oslol only)
significantly increased IL-4 and IL-10 levels. No
change was observed in IFN-y. The particle
groups only increased IL-4 and IL-10. All coarse
and fine particle fractions co-exposed with OVA
significantly increased IL-4 and IL-10 compared
to OVA alone. There was no significant
difference between coarse and fine particles.
IFN-y levels were not significantly affected by
most of the groups, but the fine fractions of PM
consistently produced higher levels of IFN-y.
Lymph Node Histology: OVA+ particle groups
resulted in significantly enlarged lymph nodes
and the formation of germinal centers.
Reference:
Nygaard et al.
(2005,,
Polysterene Particles (PSP)
Particle Size: 0.1 pm (diameter)
Species: Mouse
Gender: Female
Strains: BALB/c
Age: 6-8 wk
Route: Subcutaneous Injection into footpads.
Dose/Concentration: 40 pg PSP (5.94x1010
particles) per injection suspended in HBSS. One
injection per footpad
Time to Analysis:! HBSS
2. OVA (10 pg per injection)
3. PSP (40 pg per injection)
4. OVA (10 pg per injection) + PSP (40 pg per
injection).
Antibody experiments: reinjected with 10 pg OVA
on day 21. Killed on day 26.
Popliteal lymph node cell experiments-- animals
injected. Killed 1 to 21 days post-injection.
OVA-specific IgE, lgG1 and lgG2a
Antibodies: Analysis at day 26 indicated IgE,
lgG1 levels were significantly higher in mice
exposed to OVA+PSP compared to mice
injected with HBSS, OVA or PSP. No significant
difference was observed for lgG2a levels.
Number of Particle Containing Cells: There
was no significant difference between PSP
alone and OVA+PSP. Throughout days 0 -21 the
number of particle-containing cells in the PSP or
OVA+ PSP groups were significantly greater
than the HBSS group.
Total Cell Numbers, B and T Lymphocytes
and MHC class II Expression: The total cell
number and B lymphocytes significantly
increased by coexposure to OVA+ PSP when
compared to the other groups. Both OVA and
OVA+PSP increased T lymphocytes on Days 1,
3 and 5. MCH class II expression was
significantly higher in the OVA+PSP group on
days 5, 7 and 21 than other groups.
Cell Types and Surface Markers: The number
of CD40+ B Lymphocytes showed a slight but
significant decrease with OVA+PSP and OVA
compared to HBSS and PSP. CD86+, CD23+
and CD69+ B lymphocytes were significantly
higher in OVA+PSP group than other groups.
PSP alone did not affect CD86+ or CD23+
levels.
Cytokine Production: IL-4 and IL-10 were
significantly higher in the OVA+PSP group when
compared to the other groups. OVA alone
caused a slight increase compared to PSP. PSP
did not alter IL-4 or IL-10 levels.
December 2009
D-141
-------
Study
Pollutant
Exposure
Effects
Reference:
Nygaard et al.
(2004, 0585581
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 6-7 wk
Weight: NR
CB (carbon black/DEP)
Polystrene Particles (PSP)
Particle Size: PSP diameter: 0.0588,
0.202, 1.053, 4.64 or 11.14pm
Route: Single subcutaneous injection into footpad
Dose/Concentration: 10 pg OVA + 40 pg (low
dose) or 200 pg (high dose) of particles
Time to Analysis: 5 days after OVA injection
OVA Specific IgE and Ig2a: OVA with CB, DEP
or PSP of diameters 0.0588 and 0.202 pm
increased IgE compared to OVA alone, as well
as the 1.053, 4.64 and 11.14 pm PSP. OVA with
0.0588 pm PSP or CB significantly increased
lgG2a compared to OVA alone.
Primary Cellular Response: All OVA and PSP
groups (except the low dose of 11.14 pm PSP)
had more total lymph node cell numbers than
the OVA alone group. The low and high does
groups of 0.202 pm PSP had the greatest
amount of cell proliferation and lymphoblasts.
The OVA and 0.202 PSP treatment produced
the greatest amounts of B lymphocytes, IL-4, IL-
10 and IFN-Y. IL-2 in the PLN cells was
significantly lower in both dosage groups of OVA
and 0.202 PSP than the OVA control.
Particle Mass, Size, Number and Surface
Area: Total particle surface area explained 64%
of the variance in the IgE levels. 60-80%
variance of the PLN cellular parameters (except
CD23) were explained by total particle surface
area, number and diameter.
Reference: Reed
et al. (2008,
Species: Mouse
Gender: Male,
Female (only
BALB/c)
Strain: C57BL/6,
A/J, BALB/c
Age: NR
Weight: NR
GEE (two 1996 General Motors 4.3-L V-
6 engines; regular, unleaded, non-
oxygenated, non-reformulated gasoline
blended to US average consumption for
summer 2001 and winter 2001-2002-
Chevron-Phillips)
Particle Size: 150 nm(MMAD)
Route: Whole-body Inhalation
Dose/Concentration: PM: Low- 6.6 ± 3.7 pg/m3,
Medium- 30.3 ± 11.8 pg/m3, High- 59.1 ± 28.3
pg/m3
Time to Analysis: A/J- 6 h/day, 7days/wk, 3 days-6
mo. C57BL/6- 1wk exposure. Instillation of P.
aeruginosa. Killed 18 h postinstillation. BALB/c-
Pregnant females exposed GD1 and throughout
gestation. Offspring exposures continued until 4wk-
old. Half of offspring sensitized to OVA. Tested for
airway reactivity by methacholine challenge 48 h
post-instillation and euthanized.
The kidney weight of female A/J mice decreased
at 6 mo. and was strongly related to PM by the
removal of emission PM. PM-containing
macrophages increased by 6 mo.
Hypomethylation occurred in females at 1 wk.
The clearance of P. aeruginosa was unaffected
by exposure. Serum total IgE significantly and
dose-dependently increased. OVA-specific IgE
and lgG1 gave slight exposure-related evidence
but were not significant.
Reference:
Roberts et al.
(2007, 0976231
Species: Rat
Gender: Male
Strain: SD
Age: 10 wk
Weight: 250-300 g
R-Total = ROFA (Residual oily fish ash) Route: IT Instillation
R-Soluble = Soluble fraction of ROFA
R-Chelex = R-Soluble+Chelex (insoluble
resin)
Particle Size: 2.2 pm (mean diameter)
Dose/Concentration: 10 mg/kg (2.5-3 mg)
Time to Analysis: Pre-exposure to ROFA samples
on Day 0. Inoculation with 5*10 L. Monocytogenes
or saline on day 3. Sacrifice on days 6, 8,10.
Uninfected Groups: Compared to the controls,
the R-total and R-soluble groups had increased
LDH, PMNs, lymphocytes and AMs. The R-total
group had a slight, but significant increase in IL-
6 and the R-soluble group had a decrease in IL-
2.
Infected Groups: The R-soluble group had
increased levels of LDH (which also increased
for the R-total group), albumin, BALF cells, NK
cells, PMA-stimulated and zyomason-stimulated
CL compared to all other groups at various time
points. NOX was significantly elevated in the R-
soluble group at early time points, but in later
time points R-soluble and R-total AMs produced
less NOX than the controls. IL-10 and IL-6
increased in the R-soluble group, while IL-12,
IL-4 and IL-2 decreased. IL-12 also decreased
in the R-total group.
Reference:
Saxena et al.
(2003, 0543951
Species: Mouse
Gender: Female
Strain: C57B1/6J
Age: 18-30 wk
Weight: NR
DEPs (standard)
Particle Size: NR
Route: Intrapulmonary Instillation
Dose/Concentration: 100 pg/mouse
Time to Analysis: Pre-exposure to 2.5x104
bacillus Calmette-Guerin bacteria (BC G) with or
without coadministration of DEP. Sacrifice 5 wk
later.
The BC G + DEP group had four times the BC G
lung load than BC G alone. The load was
significantly greater in other organs in the BC G
+ DEP group. Interstitial lymphocytes, T, B and
NK cells were increased in the BC G + DEP
group over the DEP-alone group. DEP caused
no release of NO by AMs, but inhibited the
release of NO in response to IFN-y. Except for
CDS cells, no increase in IFN-y was seen in the
BC G + DEP group.
December 2009
D-142
-------
Study
Pollutant
Exposure
Effects
Reference:
Schneider etal.
(2005, 0883681
Species: Mouse
Strain: BALB/c
Cell Line: RAW
264.7
SRM 1648 (greater than 63% inOC; 4- Route: Cell Culture (625,000 cells/cm' in 96 well
7% OC; Si, S, Fe, Al, K greater than 1 % plate)
by weight; Mg, Pb, Na, Zn, Cl, Ti, Cu, As „-,„.,,-„„.,„ Jm?
Cr, Ba, Br, Mn less than 1%) Dose/Concentration: 0, 7.8,15.6, 31.2, and 62.5
Ti62
Particle Size: Ti02 = 0.3 pm average,
1.0pm max
SRM 1648 = 0.4 pm (mean diameter)
Time to Analysis: 1, 3, 6, and 12 h
No significant toxicity was exhibited by SRM
1648. The rate of dye oxidation was significantly
higher in SRM 1648-exposed cells. SRM 1648
significantly increased reduced glutathione
compared to the control at the 12-h time point.
SRM 1648 increased GSH and concurrently
caused significant PGE2 production compared
to the no ester control at the 6-h and 12-h time
points.
Reference:
Schober et al.
(2006, 0973211
Species: Human
Gender: Male and
Female
Age: 21-39 yr
treatment group;
23-32 yr control
group
Tissue Type:
Whole blood
samples
PM - organic extracts of airborne sample Route: Cell Culture
AERexld - urban aerosol 1 day sample
(total air volume-1270m3)
AERexSd - urban aerosol 5 day sample
(total air volume-6230m3)
rBetv 1 (birch pollen allergen 1a,
Biomay, Vienna, Austria)
Particle Size: NR
Dose/Concentration: 100 pi heparinized whole
blood
Time to Analysis: Blood stimulated with PBS/IL-3
foMOmin. Incubated with rBet v 1 alone or with
AERex for 20 min. Ice bath 5 min. Incubated with
antibody reagent 20 min.
Nine organic compound classes were identified
in AERexld and AERexSd, with AERexld
having 20 times more PAHs. Basophil activation
increased in all treatment groups up to 90%,
with AERexld being the most pronounced. 5-50
fold lower concentrations of AERexld were
needed to achieve the maximal effect on
basophil activation. AERex-induced
enhancement of CD63 upregulation of rBet v 1
in sensitized basophils occurred in a dose-
dependent manner. The AERex-alone treatment
did not affect CD63 expression.
Reference: Shwe
et al. (2005,
1115531
Species: Mouse
Gender: Male
Strains: BALB/c
Age: 8 wk
Weight: NR
CB = carbon black particles (Degussa,
Germany)
CB14:
C: 96.79%
1-1:0.19%
N:0.13%
3:0.11%
Ash: 0.05%
Others including 0: 2.74%
CB95:
C: 97.98%
H:0.15%
N: 0.28%
S: 0.46%
Others including 0:1.14%
Particle Size: CB14 = 14 nm (primary
particle size); CB95 = 95 nm (primary
particle size)
Route: IT instillation
Dose/Concentration: 25,125, or 625 pg in 1 ml
saline solution
Time to Analysis: 1/wk for 4 wk; 4or24h after
last instillation
BALF Cells: In CB14, the total number of BAL
cells increased significantly and dose-
dependently. In CB95, only the 625pg dose
showed a significant increase.
Cytokine and Chemokine: For CB14 and
CB95,125 or 625 pg showed a significant IL-1B
increase in a dose-dependent manner. For
CB14, only the 625 pg dose showed a
significant IL-6 increase. No difference was
observed in the CB95 group. For CB14, only
larger doses showed a significant TNF-a
increase. For CB95, no significant differences
were observed.
In BAL Fluid: CCL-2 production was
significantly increased for the 625pg dose in
both the CB14 and CB95 groups. CCL-3
production was significantly increased for the
larger doses in both the CB14 and CB95
groups.
Splenic Lymphocytes: No significant
differences were detected among the CB14
dosages, except for CD8+. No significant
differences were observed among the various
groups for CB95.
Deposition in Lymph Nodes: For all dosages,
greater deposition of CB14 than CB95 was
observed.
Chemokine mRNA Expression in Lungs and
Lymph Nodes: At 125 pg, significant increases
of CLL-3 mRNA expression was observed for
CB14; for CB95, no differences were detected.
December 2009
D-143
-------
Study
Pollutant
Exposure
Effects
Reference: Sigaud
et al. (2007,
0961001
Species: Mouse
Gender: Male
Strains: BALB/c
Age:8-10wk
Weight: NR
CAPs: Concentrated Ambient Particles
(Collected from ambient Boston air on
Teflon filters.)
Ti02
IFN-Y
S. pneumoniae (ATCC 6303, American
Type Culture Collection, Manassas, VA)
Particle Size: CAPs: <2.5 pm
Route: IFN-Y priming: aerosol
Particle exposure and infection: Intranasal
Instillation
Dose/Concentration: CAPS or Ti02: 50 pg/50 pL
PBS
S. pneumoniae: 105CFU/25 \A saline
Time to Analysis: Primed for 15 min
One time particle exposure 3 h post priming with
lung RNA analyzed 3, 6, 24 h after exposure
Sacrificed 24 h after exposure or one time infection
Inflammation: Saline-primed and unprimed
mice exposed to CAPs produced a significant
increase in PMNs in the lung (100% more than
mice exposed to Ti02.) Groups primed with IFN-
ythen exposed to CAPs produced a strong
inflammatory response, a 2.5 increase in PMNs
when compared to the increase caused by
PBS+ CAPs exposure.
Cytokine Levels: IFN-y primed and CAPs
exposed groups
Inflammation* S. Pneumo Infection: Saline-
primed and unprimed mice exposed to CAPs
produced a significant increase in PMNs in the
lung (100% more than mice exposed to Ti02.)
Groups primed with IFN-y then exposed to
CAPs produced a strong inflammatory
response, a 2.5 increase in PMNs when
compared to the increase caused by
PBS+CAPs exposure.
Cytokine Levels: IFN-y primed and CAPs
exposed groups showed a 1.5-fold increase
over the control.
PMNs: Treatment with CAPs enhanced
inflammation, causing a 2-fold increase in PMN
numbers as compared to the infected control.
IFN-y+CAPs+S. pneumo produced a 3.5 fold
increase compared to the infected control and a
1.6-fold increase compared to
PBS+CAPs+S.pneumo. Despite increased
numbers of PMNs in the IFN-y+CAPs groups,
the lungs were unable to clear the S. pneumo
infection.
Bacterial Load: Control groups showed
efficient clearance of bacteria after infection.
Unprimed, CAPs-treated, infected groups did
not show a decrease in bacterial numbers. IFN-
y+CAPs showed a 2.5-fold increase in bacterial
numbers.
Histopathology: Indicated moderate
pneumonia in PBS+CAPs and severe
pneumonia in IFN-y+CAPs. The other groups
did not indicate areas of pneumonia.
Bacterial Uptake AM and PMN Cells: In all
the treated groups, the bacterial content in AMs
showed a decrease, with a more marked
decrease in the IFN-y+CAPs group, but these
decreases were not statistically significant.
Groups exposed to CAPs showed a statistically
significant decrease in bacterial uptake by
PMNs.
ROS Levels in AM and PMN Cells:
Intracellular ROS significantly increased in AM
cells in the IFN-y+CAPs group, approximately
50% greater than controls. In PMNs, iROS
increased 100% in the IFN-y+CAPs groups as
compared to the controls.
December 2009
D-144
-------
Study
Pollutant
Exposure
Effects
Reference:
Steerenberg et al.
(2004, 0874741
Species: Rat
Gender: Male
Strain: Wistar
Kyoto
Age: 6-8 wk
DEP:SRM1650a (NIST, Gaithersburg,
MD
EHC-93: ambient PM (Ottawa, Canada)
03 (positive control)
L. mono: Listeria monocytogenes (strain
L242/73type4B)
Particle Size: DEP, EHC-93: NR
Route: DEP/EHC-93: intranasal droplet; 03:
Whole-body inhalation
Dose/Concentration: DEP/EHC-93: 50 pg (1.0
mg/ml)
03: 2mg/m3
L. mono: 0.2 or 0.3 ml (5x106 PFU/ml) 1 have
emailed author regarding correct dose
Time to Analysis: DEP/EHC-93:1/day for 7 days
(-7 days to -1 days)
03 24 h/day for 7 days (-7 days to -1 daysR
All rats infected on day 0. Sacrificed on days 3, 4,
or 5.
Body weight: Growth declined for 03 exposed
group while DEP or EHC-93 groups grew
progressively. Exposure to L. mono caused all
groups to decline in weight.
Bacterial Count in the Lung: The number of
bacteria in the lung of those rats exposed to 03
was significantly greater than those exposed to
saline. No differences in bacteria number were
found for rats exposed to saline, EHC-93 or
DEP at any time.
Bacterial Count in the Spleen: The 03
exposed group exhibited statistically significant
increases in bacteria numbers when compared
to the saline-treated group. No differences in
bacteria number were found for rats exposed to
saline, EHC-93 or DEP at any time. Exposure to
03 decreases the defense of the respiratory
tract against L. mono infection; however, DEP
and EHC-93 did not appear to affect the host
defense system in regards to clearing/fighting L.
mono.
Reference:
Steerenberg et al.
(2005, 0886491
Species: Mouse
Gender: Male
Strain:
BALB/cByJ.ico
Age: 6-8 wk
PM: collected from Rome, Oslo, Lodz,
Amsterdam and De Zilk during the
spring, summer and winter.
Rome, Oslo, Lodz and Amsterdam
represent areas with high population and
dominance of traffic. DeZilk, selected as
a negative control site, has low traffic
emissions and natural allergens.
EHC-93: used as a positive control
OVA: Ovalbumin
Particle Size: Coarse PM: 2.5 -10.0 pm
(MMAD
(MMAD
; Fine PM: 0.1 - 2.5pm
EHC-93: NR
; Ultrafine: <0.1 pm(MMAD);
Route: Intranasal Exposure
OVA challenge: aerosol
Dose/Concentration: PM: 450 pg PM (at 0, 3, or 9
mg/ml)
OVA sensitization: 50 pg (0.4 mg/ml).
OVA challenge: 20 pg (0.4 mg/ml)
EHC-93 was administered at
0 - 900 pg to evaluate any dose-response
relationship.
Time to Analysis: Sensitization and PM exposure
on days 0,14
Challenged on days 35, 38, 41 for 20 min/day
Sacrificed on day 42
Effects of Coarse and Fine Particles:
Immunoglobulins: 6/13 of the coarse and 9/13
fine PM samples induced an increase in IgE and
IgGlwhen compared to the control. lgG2a levels
were increased in 3/13 of the coarse and 5/13 of
the fine PM. Particles from De Zilk induced all
three immunoglobulins, except the fine PM did
not induce lgG2a. De Zilk was intended as a
negative control (see Table 3). Analysis among
the sites comparing the subclasses of
antibodies indicated a rank as follows: Lodz
>Rome 2 Oslo.
Histopathology: 9/13 of the coarse PM
samples and 5/13 of the fine PM samples
induced an inflammatory response.
BALF Cells: Lodz (spring/ summer) coarse and
fine PM induced a significant increase in
eosinophils, neutrophils and monocytes. The
coarse and fine PM from Rome (spring) induced
an increase in neutrophils and the coarse PM in-
duced an increase in eosinophils. Also both
Lodz and Rome from the coarse PM from the
spring induced an increase in macrophages.
Other PM samples did not have an effect on
BAL cell counts.
Cytokine Production: None of the samples
produced a significant effect on IL-4 levels. IFN-
Y levels were significantly decreased in mice
exposed to the fine PM fraction (in 8/13 of the
samples) when compared to control. Coarse
particle exposure did not appear to affect IFN-y
levels. TNF-a levels were significantly increased
(in 2 of the 13 samples) when exposed to
coarse PM; fine PM showed similar responses
compared to the OVA only group. IL-5 was
significantly increased in 4/13 of the coarse and
fine PM samples.
Analysis of PM Components: Samples from
Lodz, Oslo and Rome (all spring) were
evaluated and the water-soluble coarse PM
fraction showed increased immunoglobulin and
pathological responses and the water-insoluble
fine PM fraction from Lodz (Spring) showed
increased reactivity Leukocytes and cytokines
showed no major differences.
December 2009
D-145
-------
Study
Pollutant
Exposure
Effects
Reference: EHC-93
Steerenberg et al.
(2004,0879811 OVA
Species: Mouse Particle Size: NR
Gender: Male
Strain:
BALB/cByJ.ico
Age: 6-8 wk
Treatment:
1.C.D2-VH6:
NrampIS and
NrampIR deficient
2. B6.129P2:
Nos2tmLau: iNOS
deficient
3. BALB/CIL4
(tm2Nnt): deficient
in IL-4
4. BALB/c (wild
type) pretreated
with N-
Acetylcysteine
(MAC)
Route: Sensitization, Challenge: Intranasal
MAC: IP injection
Dose/Concentration: OVA: 200 pg (0.4 mg/ml)
EHC-93:150 pg (3 mg/ml)
MAC: 320 mg/kg
Time to Analysis: OVA-only or OVA+EHC-93
sensitization on days 0 and 14.
Some mice received MAC before intranasal
exposure on days 0 and 14
OVA challenge on days 35, 38 and 41
Sacrificed on day 42
Natural-Resistance-Associated Macrophage
Protein 1 (Nrampl): When exposed to only
OVA, NrampIS evoked less of an antibody
responses (IgE, lgG1 and lgG2a) compared to
Nrampl RHowever when coexposed to OVA and
EHC-93, the level of increased production of
antibodies was similar in both groups. After
coexposure, the wild-type showed increased
histopathological lesions, whereas the
macrophage-stimulation-deficient types showed
only a slight increase (not significant). IL-4, IFN-
Y, TNF-a and IL-5 levels were similar in wild-
type and the Nrampl strains.
Pretreatment with MAC: lgG2a concentration
was increased further in the group pretreated
with NAC. The wild-type mice and the NAC
pretreated mice showed similar
histopathological lesion patterns. IL-4 levels
were similar in wild-type and the NAC pretreated
mice. (IFN-y, TNF-a and IL-5 levels not
reported)
Inducible Nitric Oxide Synthase (iNOS): The
wild-type and the iNOS-deficient mice had
similar levels of increased IgE antibody
production. The lgG1 and lgG2a antibody
response was twice as great in the iNOS-
deficient mice compared to the wild type. The
wild-type and the iNOS-deficient mice showed
similar histopathological lesions. No differences
in BAL cells or cytokines were observed
between the wild-type and iNOS-deficient mice.
IL-4: The IL-4-deficient mice did not produce an
increase in IgE or lgG1 antibodies, as was seen
in the wild-type mice. The lgG2a antibody
response in the IL-4-deficient mice was similar
to the wild type response resulting in adjuvant
activity for the lgG2a antibodies. Overall the
histological response of the wild-type mice was
greater compared to the IL-4 deficient mice.
There was no real difference between the two
strains observed in the BAL cells, except IL-5
was significantly lower in the IL-4-deficient mice.
December 2009
D-146
-------
Study
Pollutant
Exposure
Effects
Reference:
Stevens et al.
(2008, 1553631
Species: Mouse
Strain: BALB/c
Age: 10-12 wk
Weight: 17-20 g
DE: generated using a 30 kW4-cylinder
Deutz BF4M1008 diesel engine
connected to a 22.3 kWSaylor Bell air
compressor. The engine was operated
on Diesel fue| purchased from a service
station in Research Triangle Park, NC.
The engjne wgs operated gt g steady.
s'a'e> approx. 20% of engine's full load.
Hi9n composition:
02: 4.3 ± 0.07 ppm
NO: 9.2 ± 0.30 ppm
N02: 1.1 ±0.05 ppm
S02: 0.2 ±0.10 ppm
Low composition:
02, NO, N02, S02 below detection limits
Particle Size: NR
Route: Whole-body Inhalation
OVA immunization and challenge: intranasal
Dose/Concentration: High = 2000 pg/m3
Low = 500 pg/m
Time to Analysis: DE exposure for 4 h/day on
days 0-4.
OVA immunization 40 min after DE exposure on
days 0-2
Challenged on days 18 and 28.
Sacrificed 4 h after last exposure of day 4 for gene
set analysis or 18, 48, or 96 h after the last
challenge
IgE Antibody Production: In the absence OVA,
IgE antibodies were not detected. 18, 48 and 96
h following OVA, mice exposed to low and high
doses of DE had an increase in antibodies over
time. Mice exposed to high dose had an
increase (non-significant) to the OVA exposed
control at the 48 h time mark
BAL Cells: Cell counts at 18 and 96 h after OVA
treatment did not differ among treatment groups.
At 48 h the number of eosinophils, neutrophils
and lymphocytes were significantly increased in
mice exposed to both high and low
concentrations of DE. With DE exposure alone,
only neutrophils were statistically increased in
the high DE concentration. This indicates the
combination exposure of DE and an antigen is
essential to promote the development of allergic
lung disease.
BAL Cytokines: IL-6 production showed a
dose-dependent and time-dependent increase,
but was significantly increase in the high dose
group at 96 h. The high dose group saw a non
significant increase in IL-10 levels over time.
The greatest increase in IL-10 for the low dose
group occurred18 h after OVA stimulation.
Pulmonary Inflammation and Lung Injury: No
differences among the groups were observed
for macrophage, lymphocyte, neutrophil and
eosinophil counts. Protein and LDH levels were
not found to be increased in the BALF of any
group.
Gene Analysis: Pair wise comparisons
revealed significant gene set difference between
the high DE and control groups. Comparison of
the high DE/OVA versus air/OVA showed
significant changes in 23 gene sets, including
genes involved in oxidative stress responses.
The high DE/saline versus the air/saline showed
significantly altered pathways. Altered pathways
include those for cell adhesion, cell cycle
control, apoptosis, growth and differentiation,
and cytokine signaling. The results show that
relatively short exposures to DE cause mild
increases in immunologic sensitization to
allergen.
December 2009
D-147
-------
Study
Pollutant
Exposure
Effects
Reference:
Takizawa et al.
(2003, 1570391
Species: Human
Cell Lines: Normal
Small Airway
Epithelial Cells and
Bronchial Epithelial
Cells (BET-1A)
Suspended DEP: collected using a
2,300-cc Isuzu diesel engine using
standard diesel fuel at 1,050 rpm under
a load of 6 torque.
DE exposure in vitro (air exposure):
collected using a 2,300-ml Isuzu diesel
engine at 1,050 rpm.
Composition:
Fine particles: 1 mg/m3
CO: 10.6 ppm
N02: 7.3 ppm
S02: 3.3 ppm
Particle Size: NR
Route: Cell Culture
Dose/Concentration: Suspended DEP: varying
doses from 0-50 pg/ml
IL-13: varying doses from 0-25 ng/ml
DE exposure in vitro (air exposure): 100 pg/m3
Time to Analysis: Cells were exposed to varying
concentrations of suspended DEP for up to 24 h.
NF-KB: analyzed at 6 h after suspended DEP
exposure
Air exposure at 0, 2, 4, 8 or 14 h
Preliminary experiments indicated that DEP at
0.1- 50 |jg/mL had no significant cytotoxicity to
BET-1A cells and human bronchial epithelial
cells (as analyzed by LDH levels).
Eotaxin Production: (Eotaxin is a cc
chemokine that plays a role in eosinophil
accumulation in a variety of allergic disorders)
Epithelial and BET-1 A cells treated with
suspended DEP or IL- showed a dose-
dependent stimulatory effect on eotaxin release
or production. Simultaneous exposure to 25
ng/mL IL-13 and DEP depicted an additive effect
for both cell types.
Eotaxin mRNA: At 25 pg/mL, suspended DEP
showed a time-dependent effect on eotaxin
mRNA levels up to 12 h in both cell types.
Extracted RNAfrom human bronchial epithelial
cells exposed to varying doses of DEP showed
a dose-dependent effect for both cell types (up
to 25 pg/mL DEP) on eotaxin mRNA levels after
12 h of exposure. IL-13 also induced a dose-
dependent increase on eotaxin mRNA levels in
cells in both cell types. Combination of IL-13
and DEP showed an additive effect on mRNA
levels in BET-1 A cells. DE exposure in vitro also
showed a time-dependent stimulatory effect on
eotaxin production in BET-1 A cells.
NF-KB / STAT6 Activation: (it has been
suggested that NF-KB plays a role in the trans-
criptional regulation of eotaxin gene expression)
Cells exposed to 1-25 pg/mL DEP for 6 h
increased NF-KB. BET-1 A cells treated with
suspended DEP failed to activate STAT6.
Effect of MAC and PDTC on Eotaxin mRNA
Levels: (NAC and PDTC are antioxidant
reagents with inhibitory effects on NF-KB
activation) in BET-1A, both NAC and PDTC
showed a dose-dependent inhibitory effect on
DEP-induced eotaxin production. Both reagents
also blocked DEP-induced eotaxin mRNA levels
in BET-1 A cells. NAC and PDTC did not
suppress eotaxin production or eotaxin mRNA
levels in IL-13 stimulated BET-1A cells. In
addition pre-treatment with NAC attenuated NF-
KB activation induced by DEP but had no effect
on STAT6 induction by IL-13.
These findings suggest that DEP stimulated
eotaxin gene expression via NF-KB dependent,
but STAT6-independent, pathways.
December 2009
D-148
-------
Study
Pollutant
Exposure
Effects
Reference:
Tesfaigzi et al.
(2005,1561161
Species: Rat
Gender: Male
Strain: Brown-
Norway
Age: 6 wk
PM: Wood smoke generated from a
conventional wood stove that has a
0.5m3 firebox and a sliding gate air
intake damper. The stove was operated
over a 3-phase burn cycle that spanned
6 h. Fire was started (initiated exposure)
with imprinted / unbleached newspaper
and a mix of black and white oak.
Wood smoke components: organic
material, small amounts of EC and
metals and associated analytes.
Particle Size: 0.36 pm(MMAD)
Route: Whole-body Inhalation
Dose/Concentration: PM: 1000 pg/m3
Time to Analysis: Exposed to wood smoke or
filtered air 6 h/day for 70 consecutive days
OVA IP injection immunization on days 2, 9
OVA aerosol exposure 2 h/day on days 67-70
following daily exposure to wood smoke or filtered
air
Sacrificed day 70
Body Weight and Respiratory Function: No
difference in clinical signs or body weight was
observed when comparing the two rat groups.
The wood smoke exposed group had a 45%
lower dynamic lung compliance when compared
to those exposed to the filtered air group before
the methacholine challenge. Challenging the
rats with methacholine caused a decrease in
dynamic lung compliance in both groups, but the
decrease was greater in the air-exposed group.
At the highest dose of methacholine, the
dynamic lung compliance in controls was similar
to the baseline value of the smoke-exposed
group. No significant differences in total
pulmonary resistance were observed. Wood
smoke exposed rats had a 10% increase in
functional residual capacity than the air-exposed
group.
BAL Cells and Cytokines: There was no
difference in lymphocyte, eosinophil or
neutrophils in the BALF of either group. There
was an increase, though not statistically
significant, in macrophages the wood smoke
exposed group when compared to the filtered air
group. In the BALF, IFN-y and IL-1IS levels were
significantly decreased, IL-4 and GRO-a levels
were increased in rats exposed to wood smoke
compared to filtered air. Serum IgE levels
experienced a reduction trend in the wood
smoke group, but it did not reach significance.
Both groups showed mild signs of inflammation.
The average eosinophils present in stained
tissue was 21 % higher in the wood smoke ex-
posed group compared to the air exposed.
Reference: Tomita
et al. (2006,
0978271
Species: Mouse
Gender: Female
Strain: C57BL/6J;
AHR-deficient;
mEH-deficient;
ARNT floxed (loxP
sequences
inserted in Arnt
gene);
Tcell-specific
ARNT-deficient
Age: 7 wk
Weight: 20 g
DEP: two independent preparations
fractionated into 13 different fractions
based on acidic and basic functionality
(one from light-duty, 4-cylinder diesel
engine using standard diesel fueled and
other generated from A4JB-type, Isuzu
automobile, Japan)
Individual PAH tested (Osaka, Japan):
BbF = benzo[b]fluoranthene
BeP = Benzo e]pyrene
IDP=lndeno1,2,3-cd]pyrene
BpPe = Benzo[ghi]perylene
BaP = Benzo[a]pyrene
BkF = Benzo[k]fluoranthene
Per = Perylene
DBA= Dibenzo[a,h]anthracene
Particle Size: NR
Route: Intraperitoneal Injection
Dose/Concentration: DEP, fractionated DEP or
PAH compounds: 0.5 pg -10 mg/kg bw in 50 pi of
olive oil
Time to Analysis: Single, sacrificed 3 days post-
exposure.
Effect on Thymus: DEP treatment (10 mg/kg of
body weight) caused severe atrophy of the
thymus while the spleen and lymph nodes
appeared normal. Three days following DEP
treatment showed a marked reduction in thymus
size. The total number of thymocytes was
reduced by more than 70% mostly due to a
massive reduction in DP cells (CD4+CD8+).
DEP induced no significant alterations in the cell
numbers of CD4/CD8 ratios in the spleen and
lymph nodes.
DEP Extracts: Only the WAC (carbonic acid
fraction) and BE (weak basic fraction) did not
produce a significant reduction in thymocyte
numbers in vivo. Among the active fractions, 7
produced a marked selective loss of immature
DP thymocytes, similar to the crude extract of
DEP.
PAH Effects: Thymic involution was severely
induced by the N and various other fractions. 7
out of the 8 PAH compounds were significantly
effective in decreasing the number of
thymocytes upon in vivo exposure. Only BpPe
did not have an effect.
AHR/ARNT and mEH Deficient Mice (BaP and
DEP only): In the absence of AHR, BaP
treatment did not result in a loss of thymocytes.
Like DEP, BaP produced severe thymic
involution in mEH-deficient mice. DEP-mediated
thymic involution was significantly enhanced in
mEH-deficient mice.
December 2009
D-149
-------
Study
Pollutant
Exposure
Effects
Reference:
Verstraelen et al.
(2005, 0968721
Species: Human
Tissue/Cell
Types: Monocyte-
derived dendritic
cells (Mo-DC)
Cord blood
samples of seven
women were
collected from
umbilical vessels
of placentas of
normal, full-term
infants.
DEP- SRM 2975
Particle Size: NR
Route: Cell Culture
Dose/Concentration: DEP in varying
concentrations: 0.2, 2, 20, 200, 2000 ng/mL
LPS100ng/mL
Time to Analysis: 24 h
Biological Markers: Exposure to DEP alone did
not alter expression levels of HLA-DR, CD86 or
CD83.
Treatment with IPS alone caused a non-
significant increase in all three markers when
compared to the control.
Treatment with DEP+LPS caused a significant
increase in the expression of CD83 and a non-
significant increased expression of HLA-DR and
CD86. DEP+ IPS induced a bell-shape dose-
response curve on the expression of all three
markers, with a dose of 20 ng/mL DEP + 100
ng/mL IPS causing the largest increase in
upregulation.
When only the results of the LPS-responsive
donors (5 out of 7 blood cord samples) were
included, the effects described above become
more pronounced.
Reference:
Walczak-
Drzewiecka et al.
(2003, 1888031
Species: Mouse
Cell Line:
C1.MC/C57.1
(C57) Mast Cells
Metal and Transition Metal Ions: Sr *,
Ni2*, Cd2*, Al3*, Pb2*
Particle Size: NR
Route: Cell culture,
Dose/Concentration: 0.1-5 pmol
Time to Analysis: 10min-4h
B-Hex Mediator Release in Mast Cells:
Incubation with SrCI2, NiS04, CdCI2 or AICI3
resulted in a 2-5% release of B-hexoaminidase
in mast cells. Incubation with a mixture of all
these compounds induced a greater (11%)
release in B-hexoaminidase, indicating there
might be a additive effect.
Cell Viability: Incubation of cells at
concentrations and incubation time employed
did not result in decrease in cell viability.
Antigen-Mediated Mediator Release in Mast
Cells: Al3* and Ni2* enhanced antigen-mediated
release. 10-7 M AICI3 released 23% of B-
hexoaminidase compared to antigen alone,
which induced 11 % release of B-
hexoaminidase. Cd *, Sr * and Pb * enhanced
antigen-mediated release to a lesser extent.
Ni2*, Al3*, Sr2* and Cd2* depicted a dose-
dependent relationship with antigen-mediated B-
hexoaminidase release.
Antigen-Induced Protein Phosphorylation:
Addition of the antigen induced the anticipated
phosphorylation of multiple proteins in C57 mast
cells The presence of Ni * and Pb * mediated an
increase in phosphorylation of several of the
proteins and Al3* mediated a decrease in
phosphorylation of multiple proteins (specifically
the 56 and 37 kD bands).
Antigen-Mediated Cytokine Secretion (IL-4):
At certain concentrations all tested metal and
transition metal ions were able to induce IL-4
secretion or enhance antigen-induced IL-4
secretion in mast cells, but no dose-dependent
relationship was established.
Reference: Wan
and Yu (2006,
1571041
Species: Human
Cell Lines:
Human, Bcell
lymphocytes
PMBC
(>98.5% B cells-
CD19+CD20+; <1
% T cells (CD3+))
Human lymphocyte
cell lines -- DG75
NQ01 wild type
DEP from 4 cyl Isuzu diesel methanol
extracts
Particle Size: NR
Route: Cell Culture, PMBC = 1 xio6 cell
DG 75 = 3x106 cells
lgEPMBC1xl06/mL
B-cells 0.5x106/mL
Dose/Concentration: 2.5, 5,10, 20 pg DEPX/
plate (20 pg/mL)
IgE DEPX 100 ng/mL
sulfurophane at 0 - 30 pmol
Time to Analysis: 6 h mRNA; 16 h protein assay.
IgE 14 days.
Induction of NQ01 by DEPX: In PBMCs and
DG75DEPX dose-dependently induced NQ01
mRNA expression NQ01 ARE was increased
NAC inhibited NQ01 gene expression dose
dependently. p38 MAPKand P13K inhibition
partially blocked NQ01 mRNA and ARE
induction by DEPX.
Induction of phase II enzymes: DEPX induced
IgE potentiation was reduced dose dependently
by induced phase II enzymes.
December 2009
D-150
-------
Study
Pollutant
Exposure
Effects
Reference: DEP (light-duty, four-cylinder engine-
Whitekus et al. 4JB1 type, Isuzu Automobile, Japan;
(2002,1571421 standard diesel fuel) (extracts)
Species: Mouse Particle Size: NR
Cell Line: RAW
264.7
Route: Cell Culture
Dose/Concentration: 50 |jg/mL
Time to Analysis: Exposed to antioxidants 5 h.
HO-1 western blot, determination of cellular
GSH:GSSG ratios, carbonyl protein content, lipid
hydroperoxides performed.
DEP significantly reduced the GSH:GSSG ratio.
This effect was prevented by adding thiol
antioxidants MAC or BUG. DEP increased lipid
peroxide levels, but the addition of all
antioxidants decreased these levels. DEP
increased carbonyl groups. MAC, BUG, and
luteolin reduced HO-1 expression.
Reference:
Whitekus et al.
(2002, 1571421
Species: Mouse
Gender: Female
Strain: BALB/c
Age: 6-8 wk
Weight: NR
DEP (light-duty, four-cylinder engine-
4JB1 type, Isuzu Automobile, Japan;
standard diesel fuel) (extracts)
Particle Size: 0.5-4 pm
Route: Inhalation
Dose/Concentration: 200, 600, 2000 pg/m3
Time to Analysis: Exposed 1 h/day, 10 days.
Animals receiving OVA had 20 min OVA exposure
after DEP exposure.
DEP+OVA dose-dependently increased IgE and
lgG1, being more effective than the OVA-alone
treatment. This effect was significantly
suppressed by thiol antioxidants NAG or BUG.
DEP+OVA increased carbonyl protein and lipid
peroxide over OVA. NAG or BUG suppressed
lipid peroxide and protein oxidation. No general
markers for inflammation were observed.
Reference: Witten
et al. (2005,
0874851
Species: Rat
Gender: Female
Strain: F344
Age: 8 wk
Weight: ~175 g
DEP (heavy-duty Cummins N14
research engine operated at 75%
throttle)
Particle Size: 7.234-294.27 nm
Route: Nose-only Inhalation
Dose/Concentration: Low- 35.3 ± 4.9 pg/m3, High-
632.9 ± 47.61 pg/m3
Time to Analysis: Exposed 4 h/day, 5 days/wk,
3 wk. Pretreated with saline or capsaicin.
There were no differences for substance P. The
low-exposure group had significantly less NK1.
DEP reduced NEP activity. Plasma extraversion
dose-dependently increased and was greatest
in capsaicin animals. Respiratory permeability
dose-dependently increased. IL-1|3was
significantly higher for the low-exposure group.
IL-12 was significantly lower in the capsaicin
high-exposure group. TNF-a increased in the
high-exposure group and capsaicin low-
exposure group. High exposure induced
particle-laden AMs in the lungs, perivascular
cuffing consisting of mononuclear cells, alveolar
edema and increased mast cell number.
Neutrophil and eosinophil influx was not seen.
Reference: Wong
et al. (2003,
0977071
Species: Rat
Gender: Female
Strain: F344/NH
Age: ~4 wk
Weight: ~175 g
DEP (Cummins N14 research engine at
75% throttle) (EC- 34.93-601.67 pg/m3,
OC-1.90-11.25 pg/m3, Sulfates 0.94-
17.96 pg/m3, Na- 4.07-4.78 ng/m3, Mg-
0.60-0.86 ng/m3, Ca- 5.05-10.66 na/m3,
Fe- 3.17-6.44, Cr- 0.68-1.31 ng/m3; Mn-
0.11-0.22 ng/m3, Pb-0.97-1.24 ng/m3)
Particle Size: 7.5-294.3 nm
Route: Nose-only Inhalation
Dose/Concentration: Low- 35.3 + 4.9 pg/m3, High-
669.3 ± 47.6 pg/m3
Time to Analysis: Exposed 4 h/day, 5 days/wk,
3 wk. Pretreated with saline or capsaicin.
DEP dose-dependently increased plasma
extraversion, which was further increased by
capsaicin. In the high-exposure group, particle-
laden AMs (which were reduced by capsaicin),
inflammatory cell margination, perivascular
cuffing with subsequent mononuclear cell
migration and dispersal, increased mast cells,
and decreased substance P were all seen. NK-
1R was downregulated in the low-exposure
group and upregulated in the capsaicin-
pretreated high-exposure group. NEP
decreased significantly for both groups.
December 2009
D-151
-------
Study
Pollutant
Exposure
Effects
Reference:
Yanagisawa et al.
(2006, 0964581
Species: Mouse
Gender: Male
Strain: ICR
Age: 5 wk
Weight: 25-28 g
Washed DEP (carbonaceous core),
DEP-OC(extracted organic chemicals)
and Whole DEP
Particles collected from: 4JB1-Type,
four-cylinder, 2.74 L, Isuzu diesel
engine, while operated on standard
diesel fuel at 200 g under a load of 10
torques.
Particle Size: 0.4 pm(MMAD)
Route: IT Instillation
Dose/Concentration: 50 pg/0.1L
1. Control-0.1mL PBS
2. DEP-OC- 50 pg
3. Washed DEP- 50 ug
4. Whole DEP- 50 ug DEP-OC + 50 ug Washed
DEP5. OVA-1 pg =
6. DEP-OC-1 pg + OVA
7. Washed DEP- 50 pg + OVA 8. Whole DEP- 50
pg DEP-OC + 50 pg Washed DEP + OVA
Time to Analysis: All groups received OVA or PBS
every 2 wk for 6 wk and the PM component or PBS
once a week for 6 wk.
BALF Cells: DEP-OC + OVA caused a
significant increase in PMN infiltration in the
BALF compared to the control. Exposure to
Whole DEP+ OVA caused PMN count to rise
further. OVA alone DEP-OC +OVA, Washed
DEP + OVA and Whole DEP + OVA all caused a
significant increase in macrophages com pared
to the control.
Lung Histology: Exposure to OVA, Washed
DEP, DEP-OC and Whole DEP caused a slight
increase in PMNs, mononuclear cells and goblet
cell proliferation. Treatment with all three DEP
groups + OVA caused a significant increase in
mononuclear cells, PMNs and goblet cell
proliferation. Whole DEP + OVA had the
greatest impact.
Th1 and Th2 Cytokine Expression: Washed
DEP+OVA caused a significant increase in IFN-
Y compared to control, whereas Whole
DEP+OVA caused a significant decrease
compared to control. No significant differences
in IL-2 and IL-4 levels were seen among groups.
DEP-OC+ OVA and Whole DEP+ OVA caused
significant increases in IL-5 compared to control
and compared to OVA Whole DEP+OVA caused
significant increase in IL-13 compared to control
Eotaxin and MIP-1o Expression: OVA
increased eotaxin levels and DEP-OC+OVA
caused a more significant increase in eotaxin.
Whole DEP alone caused a significant increase
in MIP-1a and Whole DEP+OVA caused an
even greater increase in MIP-1a.
lgG1 Levels: Exposure to DEP-OC+OVA
caused an increase in lgG1 and exposure to
Whole DEP+OVA caused greater elevation in
lgG1 levels.
Reference: Yang
et al. (2003,
0878861
Species: Mouse
Gender: Female
Strain: B6C3F1
Age: 6-8 wk
DEP-SRM 1650
Particle Size: 0.5 pm(MMAD)
Route: IT Aspiration
Dose/Concentration: 1, 5, or 15 mg /kg
Time to Analysis: 3 times in 2 wk or 6 times in
4wk.
Toxicity of DEP Exposure: DEP did not have a
significant effect on body, liver or spleen weight.
The highest dose of DEP caused an increase in
lung weight and lung weight relative to body
weight. None of the hematological parameters
were significantly different in the mice exposed
for 2 wk; in the 4 wk group there was a
significant decrease in platelet counts in mice
exposed to 15mg/kg.
Exposure on Spleen IgM AFC: DEP exposure
for 2 wk induced a dose-dependent decrease in
spleen AFC in response to sRBC immunization.
Mice exposed to 15 pg/kg depicted a 35%
reduction in total spleen activity. In the group
exposed to DEP for 4 wk, the decrease in AFC
was not significantly different than the control.
DEP Exposure on Spleen Cell
Number/Lymphocyte Counts: Exposure for 2
or 4 wk did not affect total number of nucleated
splenocytes. DEP caused a 30% reduction in
total T cells. The number of B cells were not
significantly affected.
DEP Exposure on Spleen T-Cell Function:
(evaluated in 2 wk exposure group only) DEP
induced a dose-dependent decrease in spleen
cell proliferation to ConA. DEP did not affect
spleen cell proliferation in response to anti-CD3
mAb. Production of IL-2 in response to ConA
was reduced in a dose-dependent manner by
DEP exposure. IFN-y production was decreased
by exposure to DEP. IL-4 production was not
measured.
December 2009
D-152
-------
Study
Pollutant
Exposure
Effects
Reference: Yin et DEP = SRM 2975 (NIST)
al. (2005, 0881331 Listeria
Species: Rat
Gender: Male
Strain: Brown-
Norway
(BN/CrlBR)
Age: NR
Weight: 200-250 g
Particle Size: NR
Route: Nose-only inhalation (DEP), IT instillation
(Listeria)
Dose/Concentration: 100,000 CPU (Listeria); 21.2
± 2.3 mg/m3 (DEP)
Time to Analysis: DEP exposure for 4 h/day for 5
days; infection with Listeria 7 days post-exposure;
sacrifice 3 and 7 days postinfection
Lung Deposit: Estimated mean lung deposit of
DEP = 406 + 29 pg/rat DEP prolonged growth of
bacteria in lung
Alveolar Macrophage (AM) Response: DEP
significantly inhibited Listeria-induced IL-1|3
secretion at day 7 and TNF-a and IL-12 at both
day 3 and day 7 IL-10 production was enhanced
at day 7.
T-Lymphocyte Response: DEP significantly
reduced the development of T cells in response
to Listeria infection. These lymphocytes
displayed increased production of IL-6 at day 7,
but significantly diminished levels of IL-10, IL-2
and IFN-y.
Reference: Yin et DEP = SRM 2975 (NIST)
al. (2004, 0976851 Listeria
Species: Rat
Gender: Male
Strain: Brown-
Norway
(BN/CrlBR)
Age: NR
Weight: 200-250 g
Particle Size: NR
Route: Inhalation (DEP), IT instillation (Listeria)
Dose/Concentration: 20.62 + 1.31 mg/m3 (DEP).
100,000 CPU Listeria
Time to Analysis: DEP exposure for 4 h/day for 5
days; inoculation with bacteria 2 h postexposure;
sacrifice 3, 7,10 days postinfection
Lung Deposit: Estimated mean lung deposit of
DEP = 389 ± 25 pg/rat
Pulmonary Responses and Bacterial
Clearance: DEP significantly augmented
Listeria-induced PMN infiltration, lung CPU and
recoverable AM at all times post-infection. LDH
activity was increased 3 days post-infection.
Bacterial count in DEP exposed rats remained
significantly higher through day 7.
Cytokine Production by AM: DEP exposure
significantly lowered Listeria-induced production
of IL-1P, TNF-a and IL-12. Production of IL-10
was strongly augmented.
T-lymphocyte Responses: DEP moderately
but not significantly lowered the total number of
lymphocytes, CD4+ cells and lymphocyte IL-10
production. Listeria-induced T-celi development
was strongly inhibited, as were the development
of CD8+ cells, IL-12 production and IFN-Y
secretion. DEP and Listeria exposure showed
and increased production of IL-6 at day 3 and
day 7 post-exposure.
Reference: Yin et DEP = SRM 2975
al. (2007,1989801 eDEP = organic DEP extract
wDEP = washed DEP
CB = carbon black
Species: Rat
Gender: Male
Strain: Brown-
Norway
(BN/CrlBR)
Age: NR
Weight: 225-250 g
Cell Line: AM
Particle Size: DEP: median diameter-
19.4 pm, surface area- 91 m2/g; CB: 0.1-
0.6 pm
Route: IT Instillation of Listeria; Cell Culture
(2.5x105 cells/well)
Dose/Concentration: DEP: 10, 50,100 pg/mL;
CB: 50 |jg/mL
Time to Analysis: Sacrifice 7 days postinfection or
no infection. Cell culture: 1, 4,16, 24 h.
AM Phagocytosis: None of the DEP or CB
treatments were cytotoxic or affected the
number of adherent cells. 10-100|jg/mL DEP
significantly decreased AM phagocytosis in a
concentration- and time-dependent manner, with
increased concentration and time decreasing
activity.
Bacterial Activity: The inhibition of AM
bactericidal activity by DEP was time- and
concentration-dependent. eDEP and wDEP
inhibited the AM bactericidal activity but were
less effective than DEP. The CB treatment was
not significant.
Cytokine Secretion by AM: DEP and eDEP
concentration-dependently decreased TNF-a,
IL-1B and IL-12, but increased IL-10. wDEP and
CB did not show a significant effect.
Cytokine Secretion by Lymphocytes: DEP
and eDEP concentration-dependently
decreased IL-2 and IFN-y. wDEP and CB had
little effect, except high concentrations of wDEP
decreased IFN-y.
December 2009
D-153
-------
Study
Pollutant
Exposure
Effects
Reference: Yin et
al. (2004, 0879831
Species: Rat
Gender: Male
Strain: Brown-
Norway
(BN/CrlBR)
Age: NR
Weight: 225-250 g
Cell Line: AM
DEP = SRM2975
eDEP = organic DEP extract
wDEP = washed DEP
CB = carbon black
Particle Size: DEP- NR, CB- 0.1-0.6 pm
Route: IT Instillation of Listeria; Cell Culture
Dose/Concentration: 50 pg/mL (DEP or CB)
DEP-lnduced ROS Production: ROSwas
induced by DEP or eDEP and inhibited by eDEP
with ANF or NAC. eDEP induction of ROS was
Time to Analysis: Killed 7 days postinstillation. AM Dependent. wDEP or CB did not induce
isolated then incubated. DEP treatments for up to
24 n DEP-lnduced HO-1 Expression: DEP- or
eDEP-induced HO-1 expression was inhibited
by ANF, NAC or SB203580. wDEP or CB did not
induce ROS. DEP or eDEP exposure resulted in
a 2.5- to 3-fold induction of HO-1 expression in
uninfectedAM.
eDEP-Modulated Cytokine Production: eDEP
exposure resulted in a time-dependent increase
in LPS-stimulated IL-10 or TNF-a production,
and both were inhibited by ANF or NAC. wDEP
did not affect either. SOD pretreatment
attenuated eDEP-upregulated HO-1 expression,
inhibited IL-10, and reversed eDEP inhibition of
IL-12. Znpp decreased IL-10.
Reference: Yin et DEP = SRM 1650a
al. (2003, 0961271
L. monocytogenes
Species: Rat
Gender: Male
Strain: Brown-
Norway
(BN/CrlBR)
Age: NR
Weight: 200-250 g
Particle Size: NR
Route: Nose-only Inhalation (DEP); IT Instillation
(Listeria)
Dose/Concentration: 50 or 100 mg/m3 (DEP);
100,000 bacteria per 500 pL sterile saline (Listeria)
Time to Analysis: DEP exposure for 4 h. Bacterial
inoculation. Sacrificed 3, 7 days post-exposure.
Lymphocyte Population: DEP-alone exposure
increased total lymphocytes, T cells and T-cell
subsets. Elevated cell counts in the combined
exposure were DEP dose-dependent, with the
100 mg/m treatment having significant
increases in the cell number and CD8+/CD4+
ratio.
IL-2: DEP exposure in noninfected rats at both
doses increased IL-2 in the 24 h culture and
decreased IL-2 in the 48 h culture. The increase
in IL-2 at 3 days postinfection was not
significant. DEP exposure increased IL-2Ra in
response to ConA stimulation. DEP-treated
infected rats had increases in ConA-inducible
CD4+/IL-2RQ+ and CD8+/ IL-2Ra+.
IL-6: IL-6 production was dose-dependent in
DEP-treated uninfected rats and infected rats.
The combined exposure produced less IL-6 than
the DEP-alone or Listeria-alone treatments.
IFN-y: DEP decreased IFN-y at 3 days post-
exposure, but increased at 7 days post-
exposure in a dose-dependent manner.
Uninfected DEP-treated rats did not
substantially respond to HKLM. HKLM-induced
IFN-y production is strongly inhibited at all
tested DEP doses.
Reference:
Zelikoffetal.
(2003, 0390091
Species: Rat
Gender: Male
Strain: F344
Age: 7-8 mo
Weight: NR
CAPS (concentrated ambient PM25 from
New York City)
S.pneumoniae
Particle Size: 0.4 pm(MMAD)
Route: Nose-only Inhalation (CAPS); IT Instillation
(S.pneumoniae)
Dose/Concentration: CAPS: Study 1- Mean- 345
pg/m ; 60-600 pg/m . .Study 2- Mean-107 pg/m ;
65-150|jg/m3
(S.pneumoniae 2-4*107)
Time to Analysis: Study 1: Uninfected rats
exposed to air or CAPS for 3 h. Sacrificed 3, 24, or
72 h post-exposure or IT instilled 4, 24, 72,120 h
and sacrificed 4, 24, 72 h postinfection
Study 2: Infection with bacteria. Exposed 48 h later
to CAPS or filtered air for 5 h. Sacrifice 9,18, 24,
72,120 h post-exposure.
Study 1: CAPS did not effect cell numbers,
viability, profiles, lavageable LDH activity, total
protein, or total circulating WBC counts.
Exposure to CAPs prior to infection significantly
increased PMN and decreased lymphocytes.
WBC levels returned to control levels by 4 h
postinfection. CAPS had no effect on circulating
monocyte values. CAPS significantly increased
bacterial burdens at 24 h, but thereafter the
burden decreased to below control levels.
Study 2: In CAPS exposed rats, PMN
decreased, Pam increased, and the cytokines
TNF-a, IL-1|3and IL-6 decreased. Lymphocytes
and monocytes were unaffected. Bacterial
burdens in CAP-exposed rats were about 10%
greater than air controls at 9 h and >300%
greater at 18 h. CAPS significantly increased
the percent of affected lung area and severity of
infection.
December 2009
D-154
-------
Study
Pollutant
Exposure
Effects
Reference:
Zelikoffetal.
(2002, 0377971
Species: Rat
Gender: Male
Strain: Fischer
344
Age: 7-9 mo.
Weight: NR
Ambient NYC PM
Single transition metals of Fe, Mn,
Streptococcus pneumoniae
Particle Size: NYC PM: PM2 5
Fe2*'Mn2*, Ni2*: 0.4 pm (MMAD)
Route: Nose-only Inhalation, IT instillation (S.
pneumoniae)
Dose/Concentration: Single metals/NYC PM: 65-
90 pg/m3; 15-20x106 (S.pneumoniae)
Time to Analysis: Infection/no infection followed
by 5 h exposure to NYC PM or single transition
metal. Sacrifice 4, 5, 9,18, 24, and 120 h after
exposure.
CAPs exposure to infected rats significantly
increased pulmonary bacterial burdens of S.
pneumo in a time-dependent manner. At 9 h,
18 h, 24 h, and 5 days after CAPs exposure,
bacterial burdens were 10%, 300%, 70% and
30% above controls. Uninfected rats exposed to
the single transition metals showed significant
alterations in PMNs and lymphocytes values at
1 h post-exposure.
Exposure to Fe of uninfected rats significantly
increased superoxide anion production by
pulmonary macrophages. Uninfected rats
exposed to inhaled Fe significantly reduced B-
lymphocyte proliferation at 48 h, but did not
affect T-lymphocyte production. Inhaled Ni, for
the uninfected, significantly decreased T-
lymphocyte production at 18 h, and did not
affect B-lymphocyte production. Inhalation of Fe
by infected rats facilitated an increase in
bacterial numbers while Ni inhibited bacterial
clearance. Inhaled Fe by infected also
significantly decreased PMNs and lymphocyte
numbers by 35% and increased pulmonary
macrophage numbers by 29% when compared
to the air exposed group. Results demonstrated
that inhalation of Fe altered innate and adaptive
immunity in uninfected hosts, and both Fe and
Ni reduced pulmonary bacterial clearance in
previously infected rats.
Reference: Zhong CAPs: Concentrated Air Particles
et al. (2006, (Boston, MA)
0932641
Urban air particles (UAP)SRM1649
Species: Mouse (Washington, DC)
Gender: Male
Strain: BALB/c
Age: 6-8 wk
Weight: NR
Cell Line:
J774A.1, IFN-y-
primedAMs,
unprimedAMs
Ti02
Carbon Black (CB) (Sigma, St. Louis,
MO)
Streptococcus pneumoniae: strain
ATCC6303
Particle Size: UAP = NR;
Ti02/CB = NR;CAPs:
-------
Reference
Reference:
Campbell et al.
(2005, 0872171
Species: Mouse
Strain: BALB/c
Age: 7 wk
Reference: Che
et al. (2007,
0964601
Species: Rat
Strain: SD
Gender: Male
and Female
Age: 9 wk
Weight: 190-
220 g
Reference:
Kleinman et al.
(2008, 1900741
Species: Mouse
Gender: Male
Strain: ApoE"'"
Age: 6 wk
Weight: NR
Reference: Liu
et al. (2005,
0886501
Species: Rat
Strain: Wistar
Gender: Male
Age: 8 wk
Reference:
Sirivelu et al.
(2006, 1111511
Species: Rat
Gender: Male
Strain: Brown
Norway
Age: 12-13 wk
Pollutant
CAPs from Los Angeles, lacking
reactive organic and H20 soluble
gases, 03, NOX, SOX
Particle Size: F+UF: <2.5 pm;
UF:<0.18|jm
Gasoline exhaust (collected from
1996 Guangzhou passenger car
with Dongfeng Gasoline Series 155
kw engine and no exhaust catalytic
converter fuelled with 90-octane Pb-
free gasoline from China
Petroleum).
Particle Size: NR
CAPs (Los Angeles, CA) (OC, EC =
-50%; sulfate, nitrate -11%)
Particle Size: NR
CAPs from Taiyuan, China
Particle Size: <2.5|jm
CAPs from Grand Rapids, Ml
Particle Size: <2.5|jm
Exposure
Route: Whole-body Inhalation
Dose/Concentration: 20-fold
concentration of near highway ambient air,
avg UF concentration: 282.5 pg/m3, avg F
concentration: 441. 7 pg/m3
Time to Analysis: 4 h/day, 5 days/wk for 2
wk
Route: IT Instillation
Dose/Concentration: 5.6, 16.7, or 50.0
L/kg, final volume 0.3 mL/rat
Time to Analysis: 1/wk for 4 wk; 24 h post
-instillation.
Route: Whole-body Inhalation
Dose/Concentration: High dose: Mass
concentration- 114.2 pg/m3, Low dose:
Mass concentration: 30.4 pg/m3
Time to Analysis: 5 h/day, 3days/wk, 6
wk; 24 h postexposure.
Route: IT Instillation
Dose/Concentration: 0, 1.5, 7.5, or 37. 5
mg/kg, final volume 0.2 mL/rat
Time to Analysis: 24 h
Route: Whole-body Inhalation
Dose/Concentration: 500 pg/m3
Time to Analysis: 8h; assayed at 24-h PE
Results
Mice were challenged with OVA prior to exposure and 1 and 2
wk following exposure, and then brains were assayed. F+UF
and UF exposure increased NF-KB DNA binding in brain. TNF-
a increased with F+UF. IL-1a increased with UFand F+UF.
This suggests a possible link between PM exposure and
neurodegenerative disease processes.
A dose-dependent increase was observed in brain DNA
damage starting at 5.6 L/kg. Increase in lipid peroxidation and
carbonyl protein was also observed at 50 L/kg. Decrease in
brain SOD occurred at all exposures. GPx activity was
unchanged with exposure. This suggests an association
between gasoline exhaust and oxidative damage to the brain.
Activated AP-1 dose-dependently increased. Activated NF-KB
significantly increased with the high CAPs dose. GFAP (which
represented activated astrocytes) and activated JNK
significantly increased with the low CAPs dose.
In the brain, SOD and CAT activity were significantly
decreased at the 2 highest doses; GSH levels were
significantly decreased at the highest dose. This suggests an
association between PM exposure and oxidative damage
mediated by prooxidant/antioxidant imbalance or high levels of
free radicals.
PVN: CAPs alone or with OVA increased NE.
MPA: CAPs increased Da when treated with OVA while no
changes in NE, 5-HIAA and DOPAC were observed.
Arcuate nucleus: OVA sensitization increased NE levels.
OB: CAPs and OVA increased NE levels, but no changes in
Da, DOPAC, or 5-HIAA were observed.
Other areas: No differences in other areas of hypothalamus,
substantia nigra, or cortex were observed. CAPs alone or with
OVA increased serum corticosterone. These results suggest
that CAPs can cause region-specific modulation of
neurotransmitters in brain and that the stress axis may be
activated causing aggravation of allergic airway disease.
Reference:
Veronesi et al.
(2005, 0874811
Species: Mouse
Strain: ApoE"'" or
C57BI/6
Age: Young
adults
CAPs from Tuxedo, NY
Particle Size: <2.5|jm
Route: Whole-body Inhalation
Dose/Concentration: Average daily
concentration 113|jg/m3
Time to Analysis: 6 h/day, 5days/wk for 4
CAPs-exposed ApoE" mice had an 29% reduction in TH-
stained neurons and a 8% increase in GFAP staining
compared to air-exposed ApoE"'". No differences were see in
C57 mice. The results suggest that ApoE" mice, characterized
by increased brain oxidative stress, are susceptible to PM-
induced neurodegeneration.
December 2009
D-156
-------
Reference
Reference: Win-
Shwe et al.
(2008, 1901461
Species: Mouse
Gender: Male
Strain: BALB/c
Age: 7 wk
Weight: NR
Reference:
Zanchi et al.
(2008, 1571731
Species: Rat
Gender: Male
Strain: Wistar
Age: 45 days
Pollutant
DEP (Nanoparticle-rich - NPDE; 81-
diesel engine, steady-state
condition, 5 h/d, 2000rpm, 0 Nm)
(CO, C02, NO, N02, S02)
Particle Size: 26.21 + 1. 50 nm
(diameter)
ROFAfrom Universidade de Sao
Paulo, Brazil
Particle Size: 1.2 + 2.24pm
(MMAD)
Exposure
Route: Whole-body Inhalation
Dose/Concentration: 148.86 + 8.44 pg/m3
Time to Analysis: 5 h/day, 5 days/wk,
4 wk. Some mice ip injected with
lipoteichoic acid (LTA) 1*/wk, 4wk. Morris
water maze behavioral test: 3 days
acquisition, 2 day probe trial.
Route: Intranasal Instillation
Dose/Concentration: 20 pg/10 pi saline
Time to Analysis: 30 days
Results
Mice in the LTA+NPDE group had significantly longer mean
escape latencies, indicating impaired acquisition of spatial
learning. NPDE directly increased NR1 and TNF-a.
LTA+NPDE increased NR2A, NR2B, and IL-lp, however LTA
was primarily responsible for the increases.
Exposed rats had increased lipid peroxidation in striatum and
cerebellum. This could be reversed with N-acetylcysteine
treatment. ROFA treatment altered motor activity shown by
decreased general exploration and peripheral walking, and
was not prevented by NAC. Results suggests that chronic
ROFA induces behavioral changes and brain oxidative stress.
Table D-6. Reproductive and developmental effects.
Reference
Pollutant
Exposure
Effects
Reference: Fedulov DEP
Route: Intranasal Instillation
DEP increased BAL PMN counts in normal and pregnant mice.
' ' — •-- IL-6
Species: Mouse
Gender: Female
(pregnant),
Offspring: NR
Strain: BALB/c
Age: NR
Weight: NR
Reference: Fujimoto
et al. (2005, 0965561
Species: Mouse
Strain: Sic: ICR
Gender: Females
(pregnant mice and
fetuses)
Age: NR (pregnant
females), 14 days of
gestation (fetuses)
Reference:
Hougaard et al.
(2008, 1565701
Species: Mouse
Strain: C57BI/6
Gender: Pregnant
females, male and
female offspring
Age: 12, 16 wk
(female offspring),
13, 17 wk (male
offspring)
Carbon black (CB)
Ti02
Particle Size: NR
DE: generated by 2369
ccdiesel engine at 1050
rpm at 80% load with
commercial light oil
Particle Size: 0.4 |jm
(MMAD)
DEP(SRM2975)
Particle Size: 90 m2/g
(SA)
Dose/Concentration: DEP, Ti02: 50 pg in 50 pL,
50 pg/mouse; CB: 250 pg in 50 pL
Time to Analysis: Particle samples baked 3 h.
Protocol 1a: Pregnant mice treated with DEP or
Ti02. Analyzed 1 9 or 48 h later. Protocol 1 b:
Pregnant mice DEP, Ti02 or CB treated day 14 of
pregnancy. 4 day-old offspring i.p. injected with
OVA+alum. 12-14 days-old exposed aerosolized
OVA.
Route: Inhalation
Dose/Concentration: 0.3, 1.0 or 3.0 mg DEP/m3
Time to Analysis: 1 2 h/day, 7 days/wk from 2 day
post coitum to 13 dpc. Sacrificed 14 dpc. mRNA
expression examined in female fetuses.
Route: Whole-body Inhalation
Dose/Concentration: 20 mg DEP/m3
Time to Analysis: Exposed 1 h/dayfrom gestation
days 7-19. Mice separated for behavioral testing
on PND 22 (day of delivery is PND 0). Behavioral
testing at 12, 16 wk for female offspring and 13,
1 7 wk for male offspring.
and KC compared to nonpregnant controls. Offspring of DEP,
CB or Ti02 exposed mice had increased AHR and airway
inflammation. Ti02 exclusively altered the expression of 80
genes in pregnant mice.
Significant increase in absorbed placentas were observed in
the 0.3 and 3.0 concentration. A decrease in absorbed
placentas was observed for the 1 .0 concentration. Increased
inflammatory cytokine mRNA in placentas from exposed
offspring were observed. An increased number of absorbed
placentas in DE-exposed offspring were seen.
Body weight of exposed unchanged at birth. Body weight
decreased at weaning.
Unchanged dams & pups at weaning. At 2 mo, exposed female
pups required less time to locate platform in spatial Reversal
task of Morris Water maze.
December 2009
D-157
-------
Reference
Pollutant
Exposure
Effects
Reference:
Hougaard et al.
(2008, 1565701
DEP (SRM 2975)
Particle Size: 240 nm
(MMAD); surface area 90
Species: Mouse mg /g, density 2.1 g/cm'
Gender: Female
(pregnant),
Offspring- male,
female
Strain: C57BL/6
Age: NR
Weight: NR
Route: Inhalation
Dose/Concentration: 19.1 + 1.13 mg DEP/m3
Time to Analysis: Pregnant dams exposed GD 7-
19,1 h/day. GD 20 named PND 0 for pups.
Weights recorded, 1 pup from each group
sacrificed PND 2. Weights recorded PND 9. PND
22 1 male and female removed from each group
for behavioral testing. Dams and remaining
offspring sacrificed PND 23 or 24.
DEP females gained more weight during gestation. Generally,
DEP pups weighed less. No significant DNA damage was
measured, but DEP caused slightly higher IL-6, MCP-1, and
MIP-2. Plasma thyroxin levels as well as learning and memory
were similar amongst the groups.
Reference: Huang
et al. (2008, 1565741
Species: Rat
Gender: Male
(adults), male and
female (fetuses)
Strain: Wstar
ME: Motorcycle Exhaust
(generated from 1992
Yamaha cabin motorcycle
with two-stroke 50 cc
engine).
Particle Size: NR
Route: Nose-only Inhalation
Dose/Concentration: 1: 10and 1: 50 dilutions
Time to Analysis: 2 h/day (1 h in morning and 1
h in afternoon), for 5 consecutive days/wk, for 4
wk (1 :50, 1 :10 dilutions) and 2 wk (1 :10 dilution).
Male mated with untreated females. Pregnant
females sacrificed on 20 days of gestation. Male
and female fetuses observed.
After exposure, decreased body weight and testicular
spermatid number were observed. 1: 10 ME exposure for 4 wk
(no recovery) decreased testicular weight and increased the
inflammatory cytokine mRNA. Glutathione system and lipid
peroxidation were not affected.
Age: 8 wk (male
adults), 20 days of
gestation (fetuses)
Reference:
Lichtenfels et al.
(2007, 0970411
Ambient air in Sao Paulo,
Brazil
Particle Size: NR
Route: Ambient Air Exposure
Dose/Concentration: NA
Time to Analysis: Males housed in
Decreased testicular, epididymal sperm counts, decreased
number of germ cells, and decreased elongated spermatids
were observed. Decreased SSR, and a sex ratio shift (fewer
open top males) also occurred after exposure.
Gender: Male and
Female
Strain: Swiss
Age: NR
chambers for 24 h/day, everyday for 4 mo,
beginning 10 days after birth. Males mated to non-
exposed females immediately following exposure.
Males sacrificed immediately following mating.
Pregnant females remain in chamber and
sacrificed on 19 days of pregnancy.
Reference: Mauad
et al. (2008, 1567431
Species: Mouse
Gender: Male,
Female
Strain: BALB/c
Age: 10 days
Weight: Parental:
21.4 + 4.0-26.3 +
2.8 g; 15 day-old
offspring: 7.8+ 1.1 -
9.0+ 1.0 g; 90 days-
old offspring: 20.3 +
2.3-27.4+1.8 g
PM (busy traffic street
Sao Paulo, Brazil; Aug.
2005-April 2006) (N02,
S02, CO)
Partjc|e size: 2.5, 1 0 pm
(diameter)
Route: Ambient Air Exposure Mild foci of macrophage accumulations containing black dots
of carbon pigment occurred in the alveolar areas on 90 day-old
Dose/Concentration:JM25: filtered chamber- 2.9 mice Surface-to-volume ratio decreased from 15 to 90 days of
age anc|was hjgner jnmjce exp0sec| to air pollution. PM
exposure reduced inspiratory and expiratory volumes at higher
levels of transpulmonary pressure.
+ 3.0 pg/m3, nonfiltered chamber- 16.9 + 8.3
pg/m3; Outdoor concentration: PM10-36.3+ 15.8
pg/m3, CO- 1.7 + 0.7 ppm, NO- 89.4 + 31.9 pg/m3,
S02- 8. 1+4.8 pg/m3
Time to Analysis: Nonfiltered exposure 24 h/day
for 4 mo. Mated at 120 days exposure. After birth,
30 females and offspring transferred to filtered or
nonfiltered chamber. Killed 15 or 90 days of age.
Reference:
Mohallem et al.
(2005, 0886571
Species: Mouse
Strain: BALB/c
Gender: Female
Age: 10 wk, 10 days
Filtered or ambient air in
downtown Sao Paulo
situated at crossroads
with high traffic density
(predominant source of
Partic|e size: NR
Route: Whole-body Inhalation
Dose/Concentration: PM10: 35.5 + 12.8 pg/m3;
CO: 2.2 + 1.0 ppm; N02:107.8 + 42.3 pg/m3; S02:
11.2+ 5.3 pg/m3
Time to Analysis: Exposed for 24 h/7days/wk for
4 mo. Newborns mated after reaching
reproductive age of 12 wk. All pregnant females
sacrificed between 19th and 20th day of
pregnancy.
No effects in adult exposed animals. Increased implantation
failure of neonatal exposed-dams.
Sex ratio, # of pregnancies, resorbtions, fetal deaths, and fetal
placenta Weights unchanged after neonatal ambient air
exposure.
December 2009
D-158
-------
Reference
Reference: Mori et
al. (2007, 0965641
Species: Mouse
Strain: C57/BL
Gender: Male
Age: 6 wk
Reference: Ono et
al. (Ono et al, 2007,
1560071
Species: Mouse
Strain: ICR
Gender: Pregnant
females, male
offspring
Age: NR (pregnant
females), 12wk
(offspring)
Reference: Ono et
al. (Ono et al, 2007,
1560071
Species: Mouse
Strain: ICR
Gender: Male
offspring, Pregnant
females
Age: 12 wk (male
offspring)
Reference:
Pinkerton et al.
(2004, 0874651
Species: Rat
Gender: Female
(pregnant),
Offspring- NR
Strain: SD
Age: 10 days (pups),
Pregnant females-
10- 14 days of
gestation
Pollutant
DEP: generated by 4-
cylinder diesel engine
Particle Size: NR
DE: generated from 4-cyl
diesel Isuzu engine at
1500rpm using standard
diesel fuel.
Particle Size: NR
DE: generated from 4JB-
2type, light duty 3060 cc
4-cyl Isuzu diesel engine
under 1 500 rpm
Particle Size: NR
PM (Fe and soot from
combustion of acetylene
and ethylene in a laminar
diffusion flame system)
Particle Size: Median
diameter: 72-74 nm; size
range: 10-50 nm
Exposure
Route: Dorsal Subcutaneous Instillation
Dose/Concentration: 0.2 ml (of 1. 1 mg/ml or 0.37
mg/ml)
Time to Analysis: 2*/wk for 10wk; 1 wk post last
instillation.
Route: Inhalation
Dose/Concentration: NR
Time to Analysis: Exposed from 2 day post
coitum to 16 dpc. Parameters for male offspring
measured on days 8, 16, 21, 35, 84 and sacrificed
at 84 days.
Route: Whole-body Inhalation
Dose/Concentration: 1.0 mg DEP/m3
Time to Analysis: Pregnant females exposed
from 2 day postcoitum- 16 dpc. Without
undergoing further exposure, male offspring
sacrificed at 12wk.
Route: Inhalation
Dose/Concentration: Mean mass concentration:
243 + 34 pg/m3; Average Fe concentration: 96
fjg/m
Time to Analysis: Exposed 10 days postnatal
age, 6 h/day, 3 days (consecutive).
Bromodeoxyuridine injected 2 h before necropsy. .
Effects
cDNA library screen after sub-cutaneous injection identified
activated clones related to prostanoids and arachadonic acid
(Platg2c2c, Acsl6) and sperm production (Stk35). However, the
route of exposure was unconventional.
PND 8 and 16 male reproductive accessory gland weight
decreased. PND 21 decreased serum testosterone (T); PND
84 increased serum T. FSHr, sTAR mRNA decreased PND 35
and 84. Relative testis and epididymal weight unchanged.
Sertoli cell degeneration observed.
Dose-dependent increase in seminiferous tubule degeneration
and decreased DSP. After 1 mo recovery, DSP recovered at
the lowest dose.
A significant reduction of cell proliferation occurred only within
the proximal alveolar region of exposed animals compared to
controls. There were no significant differences between the
groups for alveolar formation and separation within the
proximal alveolar region
Weight: NR
Reference: Silva et
al. (Silvaetal.,
2008,166981)
Species: Mouse
Strain: Swiss
Gender: Females
(pregnant mice)
Age: 1st, 2nd, 3rd
wk of pregnancy
(females), GD19
fetuses)
Ambient air: Sao Paulo,
Brazil
Particle Size: NR
Route: Ambient Air Exposure
Dose/Concentration: NR
Time to Analysis: 1stwk, 2ndwk, Srdwk
or combo of exposure during pregnancy.
Decreased fetal weight with exposure in 1st wk of pregnancy.
Decreased placental weight with exposure in any of the 3 wk of
pregnancy.
December 2009
D-159
-------
Reference
Reference: Somers
et al. (2002, 0781001
Species: Mouse
Strain: Swiss
Webster
Sender: Male and
Female
Age: 6-8 wk (adult
male and females), 5
days (pups)
Reference: Somers
et al. (2004, 0780981
Species: Mouse
Gender: NR
Strain: Sentinal Lab
Age: NR
Weight: NR
Reference:
Sugamata et al.
(2006, 1570251
Species: Mouse
Strain: ICR
Gender: Pregnant
-emales, male and
emale offspring
\ge: 11 wk
offspring), NR
pregnant females)
Reference: Tozuka
et al. (2004, 0908641
Species: Rat
Strain: F344
Sender: Pregnant
emales, male and
emale fetuses
Age: Gestation day
20 (fetuses), NR
(pregnant females)
Reference: Tsukue
et al. (2004, 0966431
Species: Mouse
Strain: Sic: ICR
Sender: Pregnant
emales, female
fetuses
Age: Gestation day
14 (fetuses)
Pollutant
Ambient air: 2 sites in
Canada (polluted
industrial area 1km
downwind from two
integrated steel mills &
rural location 30 km
gywgw\
Particle Size: NR
PM (rural or urban-
industrial)
Particle Size: >0.1 pm
DE
Particle Size: NR
DE: generated by diesel
engine (309 cc Model
NFAD-50)
Particle Size: NR
DE: generated by 2369
cc Isuzu diesel engine
operating at 1 050 rpm
with 80% load and using
commercial light oil.
Particle Size: NR
Exposure
Route: Ambient Air Exposure
Dose/Concentration: NR
Time to Analysis: Exposed 24 h/day, 7 days/wk
for 10 wk from September 10, 1999- November
21, 1999. Exposed to clean air for 6 wk post-
treatment. Paired with mice within exposure
group. 5d old pups measured.
Route: Ambient Air Exposure
Dose/Concentration: Mean TSP: Rural- 16.2 +
8.3-31.7+13.2 pg/m3, Urban-Industrial- 38.9 +
10.5 -115.3 + 25.3 pg/m3
Time to Analysis: Exposed 1 0 wk. Bred 9 wk
postexposure.
Route: Inhalation
Dose/Concentration: 0.3 mg DEP/m3
Time to Analysis: Pregnant females exposed
from 2 day post coitum to 16 dpc. Offspring
sacrificed 11 wk after birth.
Route: Whole-body Inhalation
Dose/Concentration: 1.73mg/m3
Time to Analysis: Exposed 6 h/day from GD 7-20
with no exposure on Saturdays or Sundays (4
non-exposure days total). Fetuses and maternal
blood collected on GD20. PAHs: Exposed 6 h/day
from GD 7-14 with no exposure on Saturdays or
Sundays. Breast milkcollected PND14.
Route: Whole-body Inhalation
Dose/Concentration: 0.1 mg DEP/m3 (at 1:8
dilution with clean air)
Time to Analysis: Exposed for 8h/day from 2 day
postcoitum to 13 dpc (with no exposure on days 4,
5, 11, 12). Sacrificed 14 dpc. Only female fetuses
studied.
Effects
ESTR germ line mutations following exposure.
Heritable mutation rate increased 1 .5 to 2 fold in urban vs.
rural site. Increased frequency is paternal line dependent.
The offspring of urban-industrial mice inherited paternal ESTR
mutations 1.9-2.1 times more than rural or HEPA-filtered
offspring. Maternal ESTR mutations were not significant.
Exposed pups had increased caspase 3 positive cells and
decreased purkinjie cell number (cerebellum), similar to human
Autism brain phenotype.
Gestational and lactational exposure to DE'sAnd PAHs. 7 milk
PAHs increased in DE-exposed dams. DE exposure can lead
to PAH pup exposure through breast milk.
SF-1 & MISmRNAdid not change. Other steroidogenic genes
were also unchanged. BMP-15 and oocyte differentiation
mRNA decreased.
December 2009
D-160
-------
Reference
Reference: Tsukue
et al. (2002, 0305931
Species: Mouse
Strain: C57/BL
Gender: Females,
male and female
offspring
Age: 6 wk, 70 days
post natal (offspring)
Reference: Ueng et
al. (2004, 0961991
Species: Mouse
Gender: Female
Strain: Wistar
Age: 21 days
Cell Line: MCF-7
Reference: Veras et
al. (2008, 1904931
Species: Mouse
Gender: Male,
Female
Strain: BALB/c
Age: 20 days,
newborns
Weight: NR
Reference: Veras et
al. (2009, 1904961
Species: Mouse
Gender: Male,
Female
Strain: BALB/c
Age: 20 days
Weight: NR
Reference:
Watanabe (2005,
0879851
Species: Rat
Gender: Female
(pregnant),
Offspring- male
Strain: F344/DuCrj
Age: 7 days of
gestation - parturition
(females), 96 days
(offspring)
Weight: 240-262 g
(offspring)
Pollutant
DE: generated by light-
duty, 4-cyl Isuzu diesel
engine at 1500 rpm.
Particle Size: NR
ME: generated from a
Yamaha Cabin
motorcycle 2-stroke 50-
cc engine and variable
venture carburetor
Pflrtirlp ^170* MR
ralUUIC wltd INf\
PM (downtown Sao
Paulo, Brazil near
crossroads with high
traffic density, 67% PM2 5
comprises air pollution)
Particle Size: 2.5 |jm
(diameter)
PM (Sao Paulo, Brazil;
near crossroads with high
traffic density) (Al, Ca,
Cu, Fe, K, Na, Ni, P, Pb,
S, Si, Ti, V, Zn, C)
Particle Size: 2.5 |jm
(diameter)
DE (309cc engine, Model
NFAD50, Yanmar Diesel
Co., Osaka, Japan,
1800rpm, 45% load) (PM,
N02)
Particle Size: 90% <0.5
pm
Exposure
Route: Whole-body Inhalation
Dose/Concentration: 0.3, 1.0 or 3.0 mg DEP/m3
Time to Analysis: Exposed 12 h/day, 7 days/wk
for 4 mo. Some females sacrificed immediately
following exposure. Remainder mated with
unexposed males. Parameters measured in
offspring at postnatal day 70.
Route: Intraperitoneal Instillation. Cell Culture.
Dose/Concentration: IP: 1, 10, 50 pg/ml
Cell Culture: 0.01, 0.1, 1, 10, 50, 100 pg/ml
Time to Analysis: IP: 1/day for 3 days and
sacrificed on 24 day. Cell Culture: 3, 24, 30, 48 h
and 2 days.
Route: Open-Top Chamber
Dose/Concentration: PM25- 27.5 pg/m3; N02-
101 pg/m ; CO- 1.81 pg/m ; S02- 7.66 ppm
Time to Analysis: 20 days-old mice maintained in
filtered or nonfiltered chamber until 60 days-old.
Offspring maintained in respective chambers until
21 days-old. Offspring mate at 60 days-old.
Females euthanized 18th GD.
Route: Open-Top Chamber
Dose/Concentration: Mean: Non-filtered- 27.5
pg/m , Filtered- 6.5 pg/m
Time to Analysis: 20 days-old mice maintained in
filtered or non-filtered chamber. Allowed to mate at
60 days. 2 generation model.
Route: Inhalation
Dose/Concentration: High dose total group: PM-
1.71 pg/m3, N02- 0.79 ppm; Low dose total group:
PM- 0.17 pg/m3, N02- 0.10 ppm
Time to Analysis: Pregnant rats exposed
gestational day 7 to delivery 6 h/day. 5 groups:
high dose total DE, high dose PM, N02 filtered,
low dose total DE, low dose PM, N02 filtered,
clean air control. Offspring sacrificed day 96 after
birth.
Effects
DE-exposed females had decreased uterine weight at 4 mo.
Offspring had decreased body weight at 6 and 8 wk of age.
Decreased rate of good nesting construction (3 mg/m3).
AGO decreased In males (30 and 70 days old).
Organ weight decreased in females and female crown to rump
length decreased.
10 mg/kg +E2 induced anti-estrogenic uterine effects and
antiestrogenic with in vitro (MCF-7 cells) E2 screen.
Fetal weight and maternal blood space volume and surfaces
declined in the groups exposed to nonfiltered air. Fetal
capillary surfaces were greater in nonfiltered air groups. There
was a significant gestational effect on maternal :fetal surface
ratios with values declining significantly in groups exposed
during pregnancy to nonfiltered air. The total oxygen diffusive
conductance of the intervascular barrier increased significantly
during pregestational exposure to nonfiltered air. Mass-specific
conductance increased during pregestational and gestational
periods of exposure to nonfiltered air.
Ambient air pollution extended the estrus cycle, which reduced
the number of cycles. Antral follicles decreased. Mating time
increased and fertility and pregnancy indices decreased. The
mean post-implantation loss rate increased, which was
influenced by both pre- and post-gestational exposure. Fetal
weight decreased and was also influenced by pre- and post-
gestational exposure, which exhibited a significant interaction.
All groups had significantly less daily sperm production than
the control. PM and N02 in DE decreased spermatogenia but
was not significant, however the high dose PM filtered group
achieved significance. Pachytene cells, spermatids, and Sertoli
cells were lower in all groups compared to the control.
December 2009
D-161
-------
Reference
Pollutant
Exposure
Effects
Reference: Yauk et
al. (2008, 1571641
Species: Mouse
Strain: C57BL/6x
CBA F1 hybrid
Gender: Male
Age: 7-9 wk
HEPA-Filtered air (PM Route: Ambient Air Exposure
removed) and ambient air
at 2 sites Dose/Concentration: NR
-2 km from two integrated Time to Analysis: Parameters measured 3,10
steel mills wk, or 10 + 6 wk recovery following exposure.
-1 km from major
highway on Hamilton
Harbor
Components:
Metals 3.6 + 0.7 pg/m3
TSP9.4+17|jg/m3
Particle Size: NR
10+6 wk exposure induced increased ESTR mutations in
sperm DMA of exposed v filtered. No testicular DMA adducts
seen in exposed males. At 3 wk DMA increased adducts seen
in lungs of exposed males, not in filtered males. Mutations
were PM dependent, and gas-phase independent.
Reference: Yokota
et al. (2009, 1905181
Species: Mouse
Gender: Female
(pregnant), Male
(offspring)
Strain: ICR
Age: NR
Weight: NR
DE (2369-cc diesel
engine, Isuzu Motors,
Ltd..Tokyo, Japan; 1050
rpm, 80% load,
commercial light oil)
Particle Size: NR
Route: Inhalation. Pre-natal Exposure
Dose/Concentration: DE: 1.0 mg/m3; CO: 2.67
ppm, N02: 0.23 ppm, S02: <0.01 ppm
Time to Analysis: Pregnant mice exposed 8 h for
5 days from GD 2-17. Mothers and pups kept in
clean room. Pups weaned on PND 21 then
transported to Tokyo University of Science. 2wk
acclimation. Exposed 12 h light/dark cycle. Activity
monitor with infrared ray sensor measured
spontaneous motor activity (SMA), 10 min
intervals 2 days. After behavioral test, mice
decapitated.
Prenatal DE exposure decreased SMA in the male offspring.
DE decreased locomotor activity during the light phase.
Dopamine levels in the striatum and nucleus accumbens did
not change, but HVA concentrations decreased in DE-exposed
mice.
Reference: Yoshida DEjgenerated from a 4-
et al. (2006,1561701 cyl, 2300 cc diesel Isuzu
engine at 1050 rpm and
Species: Mouse 80o/0 |oad).
Route: Whole-body Inhalation
Dose/Concentration: 0.1 mg DEP/m3
Time to Analysis: Exposure on 2-13 days of
Responses to exposure showed strain-related variations with
ICR as the most sensitive followed by C57 and ddY as the
least sensitive. MIS mRNAexpression, a factor in male
gonadal differentiation, was significantly decreased in the ICR
Strain: ICR,
C57BI/6J or DDY
Gender: Pregnant
Females, Male
fetuses
Age: 14 days of
gestation (fetuses),
2-13 days of
gestation (pregnant
females)
Reference: Yoshida
et al. (2006, 0970151
Species: Mouse
Strain: ICR
Gender: Pregnant
females and male
offspring
Age: 2-1 6 days
postcoitum (pregnant
females), 28 days
(male offspring)
Particle Size: NR
DE: generated by 4Jb1-
type, light duty 4-cylinder
Isuzu diesel engine using
standard diesel fuel at
1500 rpm.
Particle Size: NR
gestation. Parameters measured on 14 days of
gestation.
Route: Whole-body Inhalation
Dose/Concentration: 0.3, 1.0 or 3.0 mg DEP/m3
Time to Analysis: Pregnant females exposed 12
h/days, 7 days/wk from 2-16 days postcoitum.
Offspring sacrificed on postnatal day 28.
ana uo; strains. Amor/ar- 1 expression was signmcanny
decreased in the ICR strain only.
NOAEL 0.3 mg DEP/m3.
DE exposure induced increased reproductive gland weight
(two higher doses) in male mice. mRNA decreases in
aromatase and 3 p-hD (3.0 mg DEP/m3).
No change in sex ratio. Two higher doses induced significant
increased reproductive organ weights.
Male pup weight Increased at PND 28. Increased serum T was
observed in pups exposed to LOrng DEP/m3.
Serum T positively correlated with DSP, testis weight, steroid
enzyme mRNA.
Reference: Yoshida DE
et al. 2004 (2004,
0977601
Species: Mouse
Gender: Female
(pregnant),
Offspring- male
Strain: ICR
Age: 4, 6 wk
Weight: NR
Particle Size: NR
Route: Inhalation
Dose/Concentration: 6wk-old males, embryos:
0.3,1.0, 3.0 mg DEP/m , Pregnant mice: 0.1, 3.0
mg DEP/m3
Time to Analysis: 6 wk-old males: Exposed 12
h/day, 6 mo. 1 mo clean air exposure. Pregnant
mice: Exposed 2-13 p.c. 8 h/day. Male embryos:
Exposed 2-16 p.c. Examined at 4 wk-old.
6wk-old Males: In the seminiferous tubules, DE dose-
dependently caused degenerative and necrotic changes,
desquamation of the seminiferous epithelium, and loss of
spermatozoa. Spermatogenesis was still inhibited after a 1 m
clean air exposure.
Pregnant Mice: Ad4BP/5F-1 and MIS mRNA significantly and
dose-dependently decreased in male fetuses exposed to DE.
4wk-old Male Newborns: Tissue weight of the testis and
accessory reproductive glands were significantly greater in DE-
exposed mice. Blood testosterone concentration was 8X
higher than the control at 1.0 mg DEP/m3. No significant
differences occurred for testosterone synthetase mRNA.
December 2009
D-162
-------
Table D-7. Mutagenic/genotoxic effects in bacterial cultures.
Reference
Reference:
Binkova et al.
(2007, 1562731
Species:
Salmonella (+S9
(rat liver))
Cell Line: Calf
thymus DMA
Reference: Brits
et al. (2004,
0873971
Species: S.
typhimuriam
Strain: TA98 ±
S9 (Ames);
TA104recN2-4
andTA104pr1
(Vitotox)
Cell Line:
Human whole
blood (Comet,
MN assays)
Reference:
Brown et al.
(2005, 0959191
Species: S.
typhimuriam
Strain:TA98
Cell Line: Rat
hepatomaH4IIE
Reference:
Bungeretal.
(2006, 1563031
Species:
Salmonella
typhimuriam
Strain: TA98, TA
100
Reference:
Bungeretal.
(2007, 1563051
Species:
Salmonella
typhimuriam
Strain: TA98, TA
100
Pollutant
PM (Prague, Kosice, Sofia,
Czech Republic; summer,
winter) (organic extracts)
Particle Size: Diameter: <10
pm
PM (Flanders, Belgium;
urban, rural, industrial sites)
(organic extracts)
Particle Size: 10pm
(diameter)
PM (New Zealand, summer,
winter) (extracts)
Particle Size: 10pm
(diameter)
DEP (diesel fuel (DF), low-
sulfur diesel fuel (LSDF),
rapeseed oil methyl ester
(RME), and soybean oil
methyl ester (SME)) (SOF-
soluble organic fractions)
Particle Size: Total particulate
matter (no OCC) (gh-1): Mean
DF- 4.0 ±0.2; 2.8 ±0.5; 1.8 +
0.0; 3.4 ±0.2; 1.2 ±0.1
Diesel engine emissions
(DEE)-rapeseed oil (RSO)
and rapeseed methyl ester
(RME, biodiesel), natural gas
derived synthetic fuel (GTL),
and diesel fuel (DF)
(SOF- soluble organic
fractions)
Particle Size: NR
Exposure
Route: Cell Culture
Dose/Concentration: 100 pg EOM/mL
Time to Analysis: PM collected 24 h daily 3
mo, extracted. 24 h incubation BaP, c-PAH,
EOM, with or without S9. 32P-Postlabeling
4h. Autoradiography 1-24h.
Route: Cell Culture
Dose/Concentration: 2.5, 5, 10, 20m3 air
equivalents/ml
Time to Analysis: Air samples extracted.
Ames assay 48 h. Vivotox test. Comet assay
24 h. MN assay.
Route: Cell Culture
Dose/Concentration: 9.7-20.8 pg/m3
(summer), 21.8-61 pg/m (winter)
Time to Analysis: Air samples collected 15
days, extracted. Ames test: Bacteria growth
12 h, incubated 24 h. Hepatoma bioassay:
24 h incubation 2x. EROD assay.
Route: Cell Culture
Dose/Concentration: Log 2 dilutions of
extracts: 1.0, 0.5, 0.25, 0.125
Time to Analysis: SOF extracted 12 h.
Plates incubated 48 h.
Route: Cell Culture
Dose/Concentration: Log 2 dilutions of
extracts: 1.0, 0.5, 0.25, 0.125
Time to Analysis: SOF extracted 12 h.
Plates incubated 48 h.
Effects
DNA adducts in EOM treatments were greater with S9 than
without. Positive correlations were found between the amount of
DNA adducts and the PAH content (notably BaP) in the EOM
treatment.
Ames: S9 induced mutagenicity of all extracts from all areas in a
dose-dependent manner. Without S9, only extracts from the urban
and industrial areas were mutagenic at the highest dose.
Vitotox: Extracts were toxic at the highest dose.
Comet: Significant DNA damage in the extracts was seen and
enhanced by S9.
MN: A dose-response relationship was seen in the urban extracts
for increased micronucleated binuclear cells.
Generally, the mutagenic rate was positively correlated to PM10, as
well as PAH and BaP. PMi0 levels were higher and more
mutagenic in winter than summer.
No OCC: Without oxidation catalytic converter (OCC), DF extract
produced the highest number of revertant colonies at all load
modes in both TA98 and TA100 ± S9. RME, SME, and LSDF
extracts caused lower or no mutagenic effects, seen especially at
partial load modes and idle motion.
OCC: Wth OCC, all extracts reduced the number of revertant
colonies in TA98 and TA100 ±S9 at partial load modes B, C, and
D. At load mode A (rated power), there was an increase of the
number of revertant colonies in all assays -S9, significant for
extracts from RME (TA98, TA100) and SME (TA98). S9 lowered
frequency of mutations. At load mode E (idling), number of
revertant colonies of DF extracts increased ±S9.
Compared to DF, RSO significantly increased mutagenic effects of
particle extracts (i.e., revertants) by 9.7-59 in TA98 and by 5.4-
22.3 in TA100. (mRSO, RSO with lowered viscosity and fuel
preheating in tank, produced highest number of revertant colonies
in both strains ±S9.) RSO fuels condensates had 13.5 times
stronger mutagenicity than DF. RME extracts had moderate but
significantly higher mutagenic response in TA98 +S9 and TA100
-S9. Effects of GTL did not differ significantly from DF.
December 2009
D-163
-------
Reference
Pollutant
Exposure
Effects
Reference: de
Koketal. (2005,
0886561
Species: S.
typhimurium
Strain: TA98
(with and without
rat liver S9)
Cell Line:
Salmon testis
DMA
TSP (Total suspended
particulate, Maastricht, The
Netherlands; PMi0 and PM25
from 6 urban locations with
different traffic intensities.)
(organic extracts)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: Mutagenicity assay:
2.5, 9, or 18m3 sampled air in 100 pL DMSO;
DMA adduct assay: 5 pL DMSO containing
PMio or TSP from equivalent 50m3 sampled
air. PM25 concentration equivalent to 35m3
sampled air.
Time to Analysis: Mutagenicity assay: Cells
incubated 1 h with extracts. DMA adduct
assay: DMA incubated 4 h with extracts.
Overall, the direct mutagenicity and DMA reactivity of PM25
extracts were higher compared to PM10 and TSP. S9 generally
reduced mutagenic activity in TA98 but increased reactivity to
Salmon testis DMA. Total PAH and total carcinogenic PAH levels
correlated with the mutagenicity of TSP and the S9-mediated
mutagenicity of PM2 5. Neither transition metal composition nor
radical generating capacity of PM correlated with mutagenic
potential. Total PAH and carcinogenic PAH levels from PMio and
PM25 correlated with direct and S9-mediated DNAadducts; for
TSP these levels correlated with direct DMA reactivity only.
Reference:
DeMarini et al
(2004, 0663291
Species:
Salmonella
Strain: TA98,
TA98NR.TA98/1,
8-DNP6,
YG1021,
YG1024, TA100
A-DEP and forklift DEP (SRM
2975)
DEP (EOM)
Particle Size: 0.4 pm (mean
diameter)
Route: Cell Culture
Dose/Concentration: 0, 0.25, 0.5,1.0, 2.0
EOM pg/plate
Time to Analysis: DEPs sonicated 20min.
Centrifuged 10 min. Organic material
extracted and concentrated. Ames assay.
Incubated 3 days.
A-DEPs were more mutagenic in both TA98 and TA100 than SRM
2975. There was 22x more PAH-related and 8-45x more
nitroarene-related activity.
Reference: El PM (Jeddah, Saudi Arabia; 11
Assouli et al. sites, urban, winter) (organic
(2007,1869141 extracts)
Species: S.
typhimuriam
Strain: TA98
(±S9)
Particle Size: 10pm
(diameter)
Route: Cell Culture
Dose/Concentration: 2.5, 50, 100 pg/plate;
EOM range: 6-40 pg/m3
Time to Analysis: 24 h air samples,
PAHs varied from 0.83 to 0.18 ng/m . Only 2 locations of heavy
petrol driven cars showed strong genotoxic responses. A
correlation existed between DMA damage and the amount of
pollutants and PAHs. Toxicity and mutagenicity occurred only in
the presence of S9. Only 3 of the 11 sites exhibited moderate
. Comet mutagenic activities.
assay. 48 h incubation. Ames assay.
Reference: Endo PM (Tokyo, Japan; winter) Route: Cell Culture
Mutagenicity tests showed dose-response relationships that were
et al. (2003,
0972601
Species: S.
typhimuriam
Strain: YG1024
(±S9)
Reference:
Erdingeretal.
(2005, 1564231
Species: S.
typhimurium
Strain: TA98.
TA100, TA98NR
Reference: Iba et
al. (2006,
1565821
Species: S.
typhimuriam
Strain: TA98,
TA100(±S9(rat
liver))
Reference: Liu et
al. (2005,
0970191
Species: S.
typhimurium
Strain: YG1024,
YG1029
Cell Line:
Chinese hamster
lung V79 cells
(organic extracts)
Particle Size: Diameter:
>12.1 -0.06>|jm; Bimodal
mass concentration: 1-2 pm
PM (Baden-Wurttemberg,
Germany; urban, 8 locations,
glass fiber filters) (organic
extracts)
Particle Size: NR
PM (wood smoke (WS) (New
Jersey) and cigarette smoke
(CS) (Tobacco Research and
Health Institute, University of
Kentucky) (organic extracts)
Particle Size: 10 pi aliquots of
organic extracts
DEP extract (DP), gasoline
engine exhaust particulate
extract (GP), diesel exhaust
SVOC extract (DSVOC),
gasoline engine SVOC extract
(GSVOC), NISTSRM1650a
Particle Size: Gasoline PM:
0.554 mg extract (mgPM)-1;
Diesel PM: 0.363 mg extract
(mg PM)-1
Dose/Concentration: 2.5, 5, 10 pi; 0.30 -
22.76 pg/m3
Time to Analysis: Air samples collected,
extracted. 90 min pre-incubation. 48 h
incubation.
Route: Cell Culture
Dose/Concentration: 0.25, 2.5, 5, 12.5, 25
m3/plate
Time to Analysis: Standard Ames test
protocol followed.
Route: Cell Culture
Dose/Concentration: 62.5, 12.5 pg TPM
equivalent/plate
Time to Analysis: Incubation, shaking 25
min. Agar added. 48 h incubation. Rat lung
explants incubated 18 h. 12 h incubation with
treatments.
Route: Cell Culture
Dose/Concentration: 1.48, 4.44, 13.3, 40,
120, 360, 1080 pg/plate
Time to Analysis: 30 min preincubation. 48
h (YG1029). 66 h (YG1024). Overnight
preincubation 20 h.
higher without S9 and increased with decreasing size.
Extracts were mutagenic in all strains evaluated. No significant
difference in response with or without metabolic activation. Activity
in TA98NR suggests that the mutagenicity correlates with
concentrations of air pollutants such as NOX.
WS and CS were equally mutagenetic to TA98, but CS was 3-fold
more mutagenetic to TA100 than WS. CS induced CYP1A1 in the
explants, but WS did not.
Mutations: All samples induced mutations in both strains. The
increase was highly significant and dose-dependent. Response
with S9 was generally greater than without S9. PM extract was
more mutagenic than SVOC extract.
DP, GP, and GSVOC: Dose-response was seen for DNA damage
and micronuclei induction. GP, GSVOC and SRM 1650a were
stronger inducers of micronuclei than DP.
December 2009
D-164
-------
Reference
Pollutant
Exposure
Effects
Reference:
Matsumoto et al.
(2007, 1870201
Species: S.
typhimuriam
Strain: TA98,
TA100(±S9)
ARM (airborne participate
matter)
APE (airborne participate
extracts)
(Hokkaido, Japan; residential)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: Crude APE: 979mg/m!
air (CALUX BaP Equivalent (BaPEq)), 21
mg/m3 air (CALUX TCDD Equivalent
(TCDDEq)); Cleaned APE: 7.87 mg/m3 air
(CALUX BaPEq), 0.614 mg/m3 air (CALUX
TCDDEq)
Time to Analysis: Air samples collected,
extracted. Preincubation with S. typhimuriam.
3, 24 h exposure in CALUX assay. RNA
extracted from mice 6 days after last
application.
Most of the CALUX BaPEq for crude APE was derived from PAH-
like compounds, as suggested by the CALUX BaPEq of cleaned
APE accounting for 0.80% of CALUX BaPEq for crude APE.
CALUX TCDDEq showed TCDD and similar compounds to have a
low contribution. The TA100 strain was more mutagenic to APE,
with and without S9. S9 increased mutagenicity in both strains.
Reference:
Pastorkova et al.
(2004, 0874311
Species: S.
typhimuriam
Strain: TA98,
YG1041 (+S9)
Reference:
Rivedal et al.
(2003, 0976841
Species: S.
typhimurim
Strain: TA1 00,
TA98,TA100NR,
TA98NR,
TA98/1.8-DNP6
Reference:
Seagrave et al.
(2003, 0549791
Species:
Salmonella
Strain: TA98,
TA100
PM (EOM) (Plzeh, Prague,
Usti, Zd'ar - Czech Republic)
Particle Size: 10pm
(diameter)
DEP(SRM1650)(organic
extracts) (fractionated into
PAH, nitro-PAH, dinitro-PAH,
aliphatics, polar fraction)
Particle Size: NR
Compressed natural gas
(CNG) emissions (heavy-duty
vehicles): High emitter (HE),
Normal emitter (NE), New
technology (NT)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: TA98 (4 doses): 20-
200 fjg/plate, YG1041 (4 doses): 4-20
pg/piate
Time to Analysis: Collected 24 h every 18th
day, Oct-Mar, 1999-2003. Extracted. Ames
assay. 70 h incubation.
Route: Cell Culture
Dose/Concentration: Ames: 300, 600
DEP/plate; Gap junction: 100, 200 pg/mL
DEP
Time to Analysis: Extracted 16 h.
Fractionated. Ames assay. Gap junction
intracellular communication: exposed 1-6 h.
Western blot.
Route: Cell Culture
Dose/Concentration: PM (mg/mi)- NE- 7.0,
NT- 5.0, HE- 406; Recovered PM (mg/mi)-
NE-1.26, NT- 0.71, HE- 57.1; Recovered
SVOC- NE- 58, NT- 26.4, HE- 227.5
Time to Analysis: Samples collected in
filters 7x/day over several days. Recovered
Significant dose-response effects in mutagenic potency of EOM
occurred. Prague, one of the most polluted cities, had the highest
mutagenicity values. Increasing time-trends were observed in the
TA98 + S9 mutagenicity and PAH concentrations.
TA100 was the most mutagenic without S9 activation. GJIC was
dose- and time-dependently inhibited. The polar fraction was the
most potent inhibitor. Nitro-PAH and dinitro-PAH were the most
responsive fractions in the Ames assay.
All three CNG emissions were mutagenic in both strains.
Mutagenicity was reduced by S9 in TA100 but not in TA98. Activity
ranking in both strains was HE>NE>NT
PM, recovered SVOC extracts combined.
Ames assay.
Reference: PM (airborne, 4 sites: an oven
Sharma et al. hall and receiving hall in a
(2007, 1569751 waste incineration plant;
heavy-traffic street;
Species: S background; Mar-June 2005)
typhimurium
Particle Size: 2.5 urn
Strain: TA98, Miampfprt
YG1041, YG5161 (d'ameter)
Cell Line:
Human A549 lung
epithelial cells
Reference: Song PM (soluble organic fraction
et al. (2007, (SOF) extracts from diesel
1553061 engines using fuels blended
with ethanol by volume: EO -
Species: S. base diese| fue|; E5 . 5%- E10
typhimurium . 10%; E15 - 15%; E20 - 20%)
Strain: TA98, Particle Size: Density
TA10° (g/cm3): EO- 0.8379; E5-
Cell Line- Rat 0.8349; E1 0-0.8324; E1 5-
fibroc?esL929 0.8301; E20- 0.8279
cells
Route: Cell Culture
Dose/Concentration: 0.25 mg/ml
Time to Analysis: Samples taken over 7-1 6
days. A549 cells incubated 24 h. Comet and
microsuspension assays performed.
Route: Cell Culture
Dose/Concentration: Ames Assay: 0.025,
0.05, 0.1 mg/plate; Comet Assay: 0.125,
0.25,0.5, 1.0mg/mL
Time to Analysis: Samples extracted 24 h.
Ames and comet assays performed
DMA damage: Samples from all four sites induced DNA damage
in the comet assay with the street samples more damaging than
the oven hall sample.
Mutations: Microsuspension assay was used to assess
mutagenic activity. No mutagenic activity was observed for any of
the non-polar fractions from any sample sites. The moderately
polar fractions were all mutagenic, except for the oven hall
sample, only when S9 was added. Comparatively, the polar and
crude fractions were mutagenic without metabolic activation,
suggesting a direct mutagenic effect.
All PM extracts induced higher mutational response in TA98 (3- to
5-fold increase over spontaneous) than in TA100 (2-to 3-fold
increase). The highest brake specific revertants (BSR) ±S9 in both
strains occurred with E20 and lowest BSR was in E5 (except in
TA98 -S9). EO and E20 caused more significant DNA damage
(similar in effect) than the other extracts. Damage was dose-
dependent but variable with increasing ethanol volume.
December 2009
D-165
-------
Reference
Pollutant
Exposure
Effects
Reference:
Zhang et al.
(2007, 1571861
Species: S.
typhimurium
Strain: TA98,
TA100
Cell Line: A549
Gasoline engine exhaust
(GEE)
Methanol engine exhaust
(MEE)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: MTT Assay- 0.05-0.8
GEE or MEE L/ml; MN Assay- 0.025, 0.05,
0.1, 0.2 GEE or MEE L/ml; Comet Assay-
0.025, 0.05, 0.1, 0.2, 0.4 GEE or MEE L/ml;
Ames Assay- GEE: 0.625, 1.25, 2.5, 5.0, 10,
20 L/plate; MEE: 0.3125, 0.625, 1.25, 2.5,
5.0,10, 20 L/plate
Time to Analysis: Organic extracts from
GEE and MEE. MTT assay- 24 h incubation,
followed by 2 or 24 h incubation, followed by
4 h incubation. MN assay- 24 h incubation.
Comet assay. Ames assay- 72 h incubation.
Mutagenicity: GEE was mutagenic in TA98 but not TA100, -S9 at
10 and 20 L/plate and +S9 at >1.25 L/plate. Mutagenicity was
higher with S9 than without at 0.625-10 L/plate and a dose-
response was reported. MEE had no effect in either strain.
MN: GEE significantly and dose-dependently induced MN. MEE
had no significant effect at any dose.
DMA damage: GEE significantly induced DNA damage at all
doses compared to controls. MEE had no effect at any dose.
Reference: Zhao
et al. (2004,
1009721
Species: Rat
Gender: Male
Strain: SD
DEP (SRM 2975)
DEPE(SRM1975)
Carbon black (CB) (Elftex-12
furnace black, Cabot, Boston,
MA)
Particle Size: NR
Route: IT Instilled. Cell Culture.
Dose/Concentration: DEP orCB: 35mg/kg;
S9: 25, 50,100, 200 pg/plate; Cytosolic
protein: 20, 40, 80,160 pg/plate; Microsomal
protein: 5,10, 20, 40|jg/plate
Time to Analysis: Rats instilled. Sacrificed
1, 3, 7 days post-exposure. S9, cytosolic,
DEP and CB-exposed lung S9 time-dependently decreased 2-
aminoanthracene (2-AA) mutagenicity. Metyrapone and a-
napthoflavone inhibited the S9-activation of 2-AA in DEP and CB
exposed rats. Lung S9 increased the mutagenicity of DEPE but
not of DEP or CB. Liver S9 reduced DEPE dose-dependently.
CYP2B1 and CYP1A1 activated DEPE, with CYP2B1 being more
effective.
Age: NR
Weight: -200 g
Cell Line: S.
typhimurium
YG1024(±S9)
Reference: Zhao
et al. (2006,
1009961
Species: S.
typhimuriam
Strain: YGL024
(±S9)
DEP (SRM 2975)
DEPE (SRM 1975)
Aminoguanidine (AG)
Particle Size: NR
homogenates. Ames assay: 72 h incubation.
Route: Cell Culture
Dose/Concentration: NR
Time to Analysis: Lung S9 obtained from
rats used in in vivo experiment. Ames test.
Modified microsuspension assay. All assays
in duplicate plates. Repeated 3x.
AG significantly lowered 2-aminoanthracene mutagenic activity of
DEP or DEPE-exposed lung samples, with DEP being lowered the
most.
Table D-8. Mutagenicity and genotoxicity data summary: In vitro and in vivo.
Reference
Particle
Exposure
Effects
Reference: Abou
Chakra et al. (2007,
0988191
Species: Human
Gender: Male,
Female
Age: 6-1 Syr and
Adults
Participant
Characteristics:
Non-smokers
Cell Line: HeLa S3
cells
PM (3 French metropolitan cities: Urban Route: Cell Culture
PM25 and PM10 from "Residential
Sector," "Proximity Sector," "Industrial
Sector")
Dose/Concentration: 200 pL organic
extract; 20 pL aphidicoline
(organic extracts)
Particle Size: 2.5,10 pm (diameter)
Time to Analysis: 24 h
Seasonal variation was observed with genotoxic
effects being greater in winter. PM25 was more
active than PMio extracts. Samples from the
"Proximity Sector" (downtown area with heavy traffic)
exhibited the strongest genotoxic responses.
Reference: Arrieta et PM (El Paso, Texas; Juarez,
al. (2003, 0962101
Species: Rat
Cell Line: Hepatoma
(H4IIE)
Species: Mouse
Cell Line: Hepatoma
H1l1.1c2
Chihuahua, Mexico; Sunland Park, New
Mexico) (organic extracts)
Particle Size: 10 pm (diameter)
Route: Cell Culture
Dose/Concentration: EROD test: 0.03,
0.17, 0.34, 0.50, 0.68, 4.96, 9.93 extract
equivalents (m air); Luciferase: 0.17, 0.51,
1.26, 5.01 extract equivalents (m3 air)
Time to Analysis: 24 h
EROD activity declined at higher extract amounts,
but luciferase activity was not inhibited. Cytotoxicity
occurred only at extract equivalents to 0.47 m3 air.
PAH concentration increased with PM mass.
December 2009
D-166
-------
Reference
Particle
Exposure
Effects
Reference: Bao et al.
(2007, 0972581
Cell Line: Human-
hamster hybrid (AL)
DEP (organic extracts) (SRM 2975)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 10, 20, 50,100 pg/mL
Time to Analysis: Phagocytosis inhibitors:
Exposed 24 h with or without cytochalasin B
or ammonium chloride. Cytotoxicity: 24, 48 h
incubation. Mutations: Exposed 24 h. 5-7
days culture. Incubated additional 7-8 days.
The nucleus of DEP-treated cells was condensed
and shrunken compared to controls. DEPs
accumulated in cells, disrupting the mitochondria!
cristae, and were lodged in large cytoplasmic
vacuoles. DEP produced minimal toxicity CD59
locus mutations dose-dependently increased but
decreased when simultaneously treated with
cytochalasin B or ammonium chloride.
Reference: Carvalho- PM (Sao Paulo, Brazil; spring, bus strike Route: Cell Culture
Oliveria et al. (2005,
0778981
Species: T. pallida; A.
cepa
and non-strike days) (organic extracts)
Particle Size: 2.5 pm (diameter)
Dose/Concentration: Strike day: 47.32
pg/m ; Non-strike day: 43.01 pg/m
Time to Analysis: 8 h. 24 h recovery. A.
cepa roots induced 5 days. Exposed 30 h.
Fixed 24 h.
Element concentrations, sulfur and BTEX decreased
on the strike day. Micronuclei decreased in T. pallida
during the strike. Toxicity measured in A. cepa was
not significant, but higher on strike days.
Reference: Dybdahl
et al. (2004, 0890131
Species: Human
Cell Line: A549
DEP (SRM 1650)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 10, 50,100, 500 pg
DEP/mL
Time to Analysis: 2, 5, 24 h incubation.
DEP induced dose-dependent increases of IL-1a, IL-
6, IL-8, TNF-a. The cytokines increased 4-18-fold at
the highest dose. Cell viability did not decrease.
Comet tail length increased at 100 and 500 pg/mL
for 2, 5, 24 h.
Reference: Gabelova
et al. (2007, 1564581
Species: Human
Cell Line: Hepatoma
HepG2
PM (PRG-SM, PRG-LB, Kosice, Sofia; Route: Cell Culture
winter, summer) (organic extracts)
Dose/Concentration: 5-150 pg/mL
Particle Size: 10pm (diameter) *...„„„„„,
Time to Analysis: 2, 24, 48 h
Cell viability significantly decreased in the 24, 48 h
exposure groups compared to the 2 h exposure
group. DNA migration significantly dose-dependently
increased at most concentrations. In general,
oxidative DNA damage did not significantly increase.
Reference: Gabelova
et al. (2007, 1564571
Species: Human
Cell Line: Hepatoma
Hep G2 cell line
PM10 (Prague (Czech Republic), Ko'sice Route: Cell Culture
(Slovak Republic) and Sofia (Bulgaria);
urban, winter, summer) Dose/Concentration: 5 -150 pg/ml
(organic extracts)
Particle Size: 10 pm (diameter)
Time to Analysis: 24 h DNA adduct
formation. 2 h Comet assay. Oxidative DNA
damage measured by Fpg-sensitive sites.
Total DNA adducts ranged from -60 to 200 adducts
per 108 nucleotides. Extracts also produced
approximately the same levels of strand breaks.
Results suggested that the genotoxic potential of
ambient air was at least 6-fold greater in the winter
compared to summer. No substantial difference was
reported for oxidative DNA damage induced by
summer vs. winter samples.
Reference: Gong et
al. (2007, 0911551
Species: Human
Cell Line:
Microvascular
endothelial (HMEC)
DEP (aggregates, exhaust 4JB1-type Route: Cell Culture
LD.274 1,4-cylinder Isuzu diesel engine,
10 torque load, cyclone impactor,
dilution tunnel constant volume sampler)
Dose/Concentration: 5,15, 25 pg/mL
Particle Size: <1 pm (diameter)
Time to Analysis: Cells treated with DEP,
ox-PAPC (oxidized 1-palmitoyl-2-
arachidonyl-sn-glycero-3-
phosphorylchlorine), DEP+ox-PAPC
HO-1 expression was dose-dependent and greatest
with the DEP+ox-PAPC treatment. DEP significantly
dose-dependently upregulated or downregulated a
number of genes and was shown to have a
synergistic effect with co-treatment of ox-PAPC. The
most varying genes were significantly enriched for
EpRE, inflammatory response, UPR, immune
response, cell adhesion, lipid metabolism, apoptosis
and protein folding genes.
Reference:
Greenwell et al.
(2003, 0974781
Species: Rat
Cell Line: Epithelial
fluid; icosahedral
bacteriophage
cpX174-RFDNA
PM (South Wales, UK) (urban,
industrial)
Particle Size: Coarse diameter: 10-2.5
pm, Fine diameter: 2.5-0.1 pm
Route: Cell Culture
Dose/Concentration: Urban mean: 18.7 +
4.7 mg/day; Industrial mean: 22.6 + 2.5
mg/day
Time to Analysis: 24 h air samples 4-11
days. Substrates vortexed 1 h, suspended
4 h, centrifuged 1 h. Oxidation assay.
Industrial PM was more bioreactive than urban PM.
Coarse fractions had greater oxidative potential and
bioreactivity than fine fractions.
Reference: Gu et al.
(2005, 1959231
Species: Hamster
Strain: Chinese
Cell Line: Lung
fibroblast (V79)
DPM (1980 model General Motors 5.7-L Route: Cell Culture
V-8 enaine)
Dose/Concentration: 25, 50,100,150
Particle Size: NR pg/mL; 10 pg DPM in 10 pg in DPPC/mL; 10
pg DPM in 10 pg DMSO/mL
Time to Analysis: Chromosomal aberration:
24 h incubation. Treated 24 h. Incubated
again 24 h. MN assay: 24 h treatment. Gene
mutation: 24 h treatment. Cells replated. 7
days expression times. Staining at 8,10
days.
DPM significantly and dose-dependently increased
aberrant cells at 25-100 pg/mL DPM increased MN
formation dose-dependently. Mutant frequencies
were not significant and showed no dose-dependent
trends. DPM was toxic to cells at the highest
concentration.
December 2009
D-167
-------
Reference
Particle
Exposure
Effects
Reference: Gualtieri TD (Tire debris, generated by rotating Route: Cell Culture
et al. (2005, 0978411 new vehicle wheel against a steel brush,
significant component of PM10) (organic Dose/Concentration: 50, 60, 75 pg/mL
Species: Human extracts!
exlracls) Time to Analysis: Particles extracted 6 h.
Cell Line: A549 Particle Si?e-1 n.an i im MiamptPrt Cells subcultured every 3-4 days. After 24 h,
TD treatments 24, 48, 72 h.
Particle Size: 10-80 pm (diameter)
A time- and dose-dependent inhibitory effect on the
reduction of MTT was seen. Mortality increased
dose-dependently and was significantly greater than
the controls. DMA strand breaks increased
significantly in a dose-dependent manner. A
significant cell cycle block in the G1 phase with a
consequent decrease in the cell number in the S and
G2/M phases was seen. Exposed cells had a
modified morphology.
Reference:
Gutierrez-Castillo et
al. (2006, 0890301
Species: Human
Cell Line: A549
PM25 and PM10 (4 monitoring stations in Route: Cell Culture
Mexico City: (1) downtown high auto
traffic, (2) two industrial areas with high
levels of auto traffic and low vegetation,
(3) medium-traffic residential area)
(winter, spring , 4 sampling days in each
period)
(aqueous and organic extracts)
Particle Size: 2.5 or 10 pm (diameter)
Dose/Concentration: 0.05, 0.07, 0.1 m/r
equivalents PM25; 0.82,1.25,1.63m3/ml
equivalents PMi0
Time to Analysis: 48 h
Higher amounts of water-soluble metals were found
in samples collected during winter. V\Mer-soluble
extracts increased DMA damage 1.7-fold over the
background. Similar results were observed with
organic extracts. In general, PM25 extracts had
greater genotoxic potential than PMi0 extracts, and
water soluble fractions form both particle sizes were
more genotoxic than the corresponding organic
extracts.
Reference: Izawa H
et al. (2007, 1903871
Cell Line: NA
DEPE (4JB-1 Isuzu 4-cylinder direct-
injection 2740ccdiesel engine; 1500
rpm, 10 kg/m load)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: DEP: Ah-1
experiment-111, 55.5, 27.8,13.9, 6.9, 3.5,
1.7 pg/mL; Foods, polyphenols experiment-
27.8 pg/mL
Time to Analysis: DEPE incubated 2 h for
dioxin toxicity measurement. Absorbance at
405 nm measured. Food, polyphenol
inhibitory effects: food extract or polyphenol
solution added to cytosol solution, shaken 5
min. DEPE added, shaken 5 min. 2 h
incubation. Absorbance at 405 nm
measured.
The dioxin toxicity equivalent was 6,479 ± 58 ng
DEQ/g of DEP. The absorbance showed a sigmoid
curve and dose-dependently increased from 6.9 to
27.8 pg DEP/mL The Ginkgo biloba extract
significantly inhibited AhR activation significantly
more than the other foods, and was followed by
green tea, onions, and garlic. Quercetin and
myricetin dose-dependently inhibited AhR activation.
Ginkgolides A and B had weak inhibitory effects and
resveratol was the weakest.
Reference: Jacobsen DEP (SRM 1650b)
et al. (2008, 1565971
Carbon black (CB) (Printex 90)
Species: Mouse Partic|esize: DEP: 18.30 nm; CB: 14
Cell Line: FE1- nm; Agglomerates in suspensions: DEP
MutaTM lung Peaks- 249 nm, CB Peaks- 476 nm
epithelial cells
Route: Cell Culture
Dose/Concentration: 37.5, 75 pg/mL
Time to Analysis: 8 repeated 72 h
incubations.
Mutagenicity: The 75 pg/mL dose was significantly
increased compared to the 37.5 jjg/mL dose. Linear
regression showed a significant increasing trend by
increasing exposure. There was no change in the
total cell numbers.
ROS: ROS production increased in DEP-treated
cells after 3 h of exposure. CB-treated cells showed
a dose-dependent increase.
Reference: Karlsson PM (urban dust particles, SRM 1649) Route: Cell Culture
et al. (2004, 1989761
Species: Human
Cell Line:
Fibroblasts; calf
thymusDNAwith
human liver
microsomes or rat
liver S9
(extracted with DCM, acetone, DMSO,
water) (Fe 3% w/w, Ti 0.32% w/w, V
0.04% w/w, Mn 0.03% w/w, Cu 0.025%
w/w)
Particle Size: <10 pm (mean diameter)
Dose/Concentration: 0.1,1.0,10,100
pg/cm2
Time to Analysis: Fibroblasts exposed
24 h. Comet assay. Calf thymus incubated 2
h with microsomes or S9. 32P-labelled.
DMA damage increased dose-dependently, and a
significant amount of DMA-damaged cells had
particle interactions. DMA damage induced by the
insoluble particle core significantly increased after
each extraction. Native particles were more
genotoxic than those extracted with DMSO, DCM
and water, but not with acetone or hexane. DMSO
extracts had the most adduct-forming PACs, and
water extracts had the most oxidizing substances.
Reference: Karlsson PM (subway station, urban street)
et al. (2005, 0863921 Subway particles: 02, Fe (Fe from
Fe304) Street particles: Fe from Fe203
Species: Human
Particle Size: 10 urn (diameter)
Cell Line: A549
Route: Cell Culture
Dose/Concentration: Comet: 5,10, 20, 40
pg/cm ; 8-oxodG: 10 pg/cm
Time to Analysis: 4h.
Both PM types induced concentration-dependent
DMA damage, but subway particles were more
potent. Subway particles caused more 8-oxodG
formation and oxidation of dG, the latter of which
was inhibited by deferoxaminemesylate. Oxidation
from subway particles was due to nonsoluble, redox
active substances, and soluble substances from
street particles.
Reference: Karlsson
et al. (2006, 1566251
Species: Human
Cell Line: A549;
monocytes from
heparinized whole
blood
PM (wood- old, modern boiler; pellets- Route: Cell Culture
pellets burner, electrical ignition; tire-
road simulator studded, friction tires;
Street- busy street, Stockholm; Subway-
platform near street)
Dose/Concentration: 40 pg/cm2
Particle Size: 2.5,10 pm (diameter)
Time to Analysis: Cells grown 24 h. Comet
assay. Monocytes incubated 10 days.
Macrophages incubated 18 h.
All particles tested caused DMA damage, but there
was no significant difference between the size
fractions. Subway particles were the most genotoxic.
The urban street particles were the most potent
inducers of the cytokines. On the Teflon filters, PMi0
was somewhat more potent than PM2 5.
December 2009
D-168
-------
Reference
Particle
Exposure
Effects
Reference: Kubatova
et al. (2004, 0879861
Species: Monkey
Cell Line: African
green kidney COS-1
(CV-1 cells with
origin-defective SV40
mutants) (+S9)
PM (DE from diesel bus, wood smoke
(WS) from chimney, hardwood smoke)
(organic extracts)
Particle Size: NR
Route: Cell Culture WS had significantly increased cytotoxicity in
fractions of 25-250°C, and DE in nonpolar fractions
of 250 and 300°C and polar fractions of 50°C. The
cytotoxicity of DE PM nonpolar fractions
corresponded to increased concentrations of PAHs.
is- 24 h cvtotoxicitv 2 h SOS WS was not 9enotoxic and DE was Qenotoxic in
.\s. M n cytotoxicity i n bus> midpo|arjty fractjons (50-250°C). Genotoxic
response was not increased after S9 activation.
Dose/Concentration: 25, 50,100, 200
pg/mL; 50mg of each material used for all
experiments
Reference: Landvik
et al. (2007, 0967221
Species: Mouse
Cell Line: Hepatoma
Nepal dc7 cells
DEP extracts (DEPE in the paper)
Particle Size: NR
Route: Cell Culture
Dose/Concentration: 10, 20, 30, 50, 70
pg/mL
Time to Analysis: 24 h
50 and 70 pg/mL DEPE did not induce DMA
fragmentation but did cleave caspase 3 to a minor
extent.
Reference: Mehta et
al. (2008, 1904401
Species: Human
Cell Line: A549
PM(SRM1949a)
Particle Size: Ł 0.18 pm (diameter)
Route: Cell Culture
Dose/Concentration: 0, 50,100, 200, 400
pg/mL
Time to Analysis: Cell culture and cell
viability assay: PM treatment 24 h. 10 days
incubation. Host cell reactivation assay:
pGL3-luciferase plasmid UV irradiated 20
min. PM treatment 24 h. 16 h transfection.
24 h PM incubation. DMA repair synthesis
assay: PM treatment 24 h. Proteinase K
treatment 30 min. supf mutagenesis assay:
PM treatment 24 h. PM culture 60 h. DMA
extracted. Overnight incubation of
transformed bacteria.
PM reduced colony-forming ability and repair
synthesis capacity was proportional to the PM
concentration. PM dose-dependently decreased
HCR capacity and decreased more than TSP. PM
induced cyclobutane dimmers and pyrimidine<6-
4>pyrimidones mutations in UV-irradiated supf.
Reference: Meng
and Zhang (2007,
1989631
Species: Rat
Gender: Male
Strain: Wistar Kyoto
Age: NR
Weight: Mean: 230g;
Range: 200-250g
Dell Line: AMs from
reated rats
Reference: Motta et
al. (2004, 1989531
Species: Hamster
Strain: Chinese
Cell Line: Epithelial
iver, ovary
Reference: Oh and
Chung (2006,
0882961
Cell Line: A549
Comet CHO-K1
CBMN , H4IIE
EROD-microbiassay)
PM (Baotou, Wuwei, China) (normal
weather, dust storms, Mar 1-31)
(organic extracts, water soluble
fractions)
Particle Size: 2.5 |jm
PM (Catania, Sicily; spring) (organic
extracts)
Particle Size: NR
Crude extract (CE) DEP and fractions of
CE of DEP (organic extracts: F1 -
organic bases, F2 - organic acids, F3 -
aliphatic, F4 - aromatic, F5 - slightly
polar, F6 -moderately polar, F7 - high
poisrj
Particle Size: Diameter: <2.5 pm,
87.71%, 2.5-10 fjm, 3.87%, >10 pm,
n A^n/
Route: Cell Culture
Dose/Concentration: AM: 0, 33.3, 100, 300
|jg/mL; Water-soluble: 0, 75, 150, 300
pg/mL; Organic extracts: 0, 25, 50, 100
pg/mL; Mass concentration normal day:
68.49 + 28.83 pg/m3; Mass concentration
dust storm day: 221.83 + 69.89 pg/m
Time to Analysis: 24 h; cultures 4 h.
Route: Cell Culture
Dose/Concentration: 0.60, 1.21, 2.42, 4.85,
9.70, 19.40 pg/mL; 0.78, 1.56, 2.12, 6.25,
12.50, 25.00 pg/mL
Time to Analysis: 24 h
Route: Cell Culture
Dose/Concentration: 100 |jg/mL
Time to Analysis: DEP generated,
extracted. Comet assay- 24 h incubation,
CE, DEP exposed 24 h. MN assay- cultured
24 h, 4 h treatment, growth medium
incubation 20 h. EROD-microbioassay- 48 h.
OC, NH4*, N03" were higher in normal weather
PM25. S042", Ca2* were higher in dust storm PM25.
Fe, Al, Ca, Mg were 5x higher in dust storm PM2s.
Cell viability reduced in a concentration-dependent
manner, with normal weather being slightly more
cytotoxic. DNA damage was dose-dependently
induced, with normal weather and organic extracts
showing the greatest damage.
The treatment was only slightly cytotoxic at the
highest dose. DNA damage and aberrant cells
generally increased with dose. No effect was seen in
the Chinese hamster ovary cells without metabolic
activation.
DNA damage: CE significantly increased the
amount of DNA damage in A549 cells with and
without SKF-525A, a CYP450 inhibitor, and in CHO-
K1 cells. It significantly increased MN formation ±S9
com pared to controls.
Organic Extracts: Organic base (F1) and neutral
(F3-F7) fractions of CE of DEP significantly induced
DNA damage without SKF-525A compared to
controls. Adding SKF-525Acompletely inhibited
damage caused by F3, F4, F6 and F7 but kept the
effect of F1 similar to that without SKF and only
partially inhibited that of F5. F2 did not induce DNA
damage with or without SKF. All fractions except F6
induced MN formation ±S9.
December 2009
D-169
-------
Reference
Particle
Exposure
Effects
Reference: Poma et
al. (2006, 0969031
PM (L'Aquila, Italy; urban); air samples
collected weekly basis Jan-Mar 2004.
Species: Mouse Carbon black (CB)
Cell Line: RAW 264.7 Particle Size: 2.1-0.43 pm (diameter)
Route: Cell Culture
Dose/Concentration: 1,3,10 pg/cm2
Time to Analysis: Cells cultured 48 h.
Treatment 48 h. MN assay: 44 h incubation,
28 h incubation.
PM and CB dose-dependently reduced cell
proliferation and induced micronuclei. PM and CB
also reduced cellular metabolism of the
macrophages and induced significant amounts of
apoptosis. PM produced more micronuclei than
equally-weighted CB.
Reference: Roubicek PM (Mexico City from an industrial area
et al. (2007,1569291 with high-traffic and a medium-traffic
residential area)
Species: Human
Cell Line: A549 (aqUe°US °r °rganiC eXtraCtS)
Particle Size: 10 pm (diameter)
Route: Cell Culture
Dose/Concentration: 1.25,1.63, 2.5 mVml
equivalents of PMio
Time to Analysis: Cells treated 24 h
followed by 48 h incubation with
cytochalasin B. Micronuclei frequency
determined.
Water and organic extracts induced a significant
dose-dependent increase in the micronuclei
frequency. After doses of PM from different regions
were normalized for mass differences, the genotoxic
potency was higher for samples from the industrial
area.
Reference: Salonen
et al. (2004, 1870531
PM (Vallila, Finland; busy traffic site;
spring, winter)
Species: Mouse Particle Size: <10 pm (diameter)
Cell Line: RAW264.7
Route: Cell Culture
Dose/Concentration: 15, 50,150, 500,
1000|jg/mLofRPMI
Time to Analysis: 24 h
PAHs decreased from winter to spring. TNF-a dose-
dependently increased and was higher in spring
samples. IL-6 generally increased in spring but not in
winter. NO dose-dependently increased and was
higher in winter. Cell viability generally decreased
but there were no consistent potency differences
between the samples. Generally, proinflammatory
activity, cytotoxicity and IL-6 were associated with
the insoluble PM fractions. Polymyxin B inhibited IL-
6 and TNF-a. 'OH and 8-hydroxy-2'-deoxyguanosine
dose-dependently increased and were higher in the
spring and winter, respectively.
Reference: Seaton et PM (3 busy London underground (LU)
al. (2005,1989041 stations and cabs) (LU dust in PM2 5
samples: iron oxide 64-71%, chromium
0.1-0.2%, manganese 0.5-1%, copper
<0.1-0.9%; respirable dust samples: 1-
2%)
Particle Size: Diameter: <2.5 pm, 10
pm, Median diameter: 0.4 pm
Species: Human
Cell Line: A549
Route: Cell Culture
Dose/Concentration: Assays: 1,10, 50,
100|jg/mL
Time to Analysis: 8, 24 h.
PM10 caused less LDH release, IL-8 stimulation and
free radical activity than LU dust particles that
contained PM25. Chelation had little effect on PMio
soluble components.
Reference: PMio (Prague, Czech Republic; Ko'sice;
Sevastyanova et al. Slovak Republic; Sofia, Bulgaria)
(2007,1569691 (urban, summer, winter)
Species: Human (organic extracts)
Cell Line: HepG2 cell Particle Size: 10 pm (diameter)
line, embryonic lung
diploid fibroblasts
(HEL), or acute
monocytic leukemia
cells (THP-1)
Route: Cell Culture
Dose/Concentration: 10-100 pg/ml
Time to Analysis: 24 h
DNA adducts were observed in all cell types
evaluated. Highest adduct levels were observed in
HepG2 cells, followed by HEL and THP-1 cells. A
correlation between DNA adduct levels and
carcinogenic PAH content was observed in HepG2
cells at 50 pg/ml.
Reference: Shi et al.
(2003, 0882481
Species: Human
Cell Line: A549
PM (Dusseldorf, Germany, July-Dec.)
Weekly samplings July-Dec 1999.
Particle Size: Fine diameter: <2.5 pm;
Coarse diameter: 10-2.5 pm
Route: Cell Culture
Dose/Concentration: Fine: 0.57-2.49 mg;
Coarse: 0.66-1.89 mg; Concentration: 0.57
mg/mL
Time to Analysis: NR
Coarse and fine particles generated 'OH, but coarse
particles had significantly higher 'OH formation as
well as 8-hydroxy-2'-deoxyguanosine formation. 8-
hydroxy-2'-deoxyguanosine and 'OH had a
significant correlation.
Reference: Skarek et
al. (2007, 0968141
Species: Rat
Cell Line: Modified
hepatomaH4IIE.Iuc;
SOS: E. coli PQ37
(±S9)
PM (urban: Usti and Laben, Karvina;
background: Cervenohorske sedlo,
Kosetice - Czech Republic; July)
(organic extracts, TSP); GP (gas
phase). 24 h samples July 2002
Particle Size: <2.5 pm (diameter)
Route: Cell Culture
Dose/Concentration: SOS: 8, 4, 2,1 m3/ml;
Dioxin:TSP+GP-8,1.33, 0.22, 0.04m3/ml,
PM25+GP:4, 0.66, 0.11, 0.02m3 ml-1
Time to Analysis:. SOS chromotest: 22 h
incubation. Dioxin toxicity test: 24 h
exposure.
The urban areas had a much greater level of
carcinogenic PAHs and overall number of PAHs than
the background areas. Significant genotoxic activity
was only detected at TSP+GP without S9 from urban
areas. PM2 5+GP had lower dioxin activity at the
urban areas, but similar levels of toxicity were seen
for both treatments in the background areas.
Reference: Song et
al. (2007, 1553061
Species: S.
typhimurium
Strain: TA98, TA100
Cell Line: Rat
fibrocytes L-929 cells
PM (soluble organic fraction (SOF)
extracts from diesel engines using fuels
blended with ethanol by volume: EO -
base diesel fuel; E5 - 5%; E10 -10%;
E15-15%;E20-20%)
Particle Size: Density (g/cm3): EO-
0.8379; E5-0.8349; E10-0.8324; E15-
0.8301 ;E20-0.8279
Route: Cell Culture
Dose/Concentration: Ames Assay: 0.025,
0.05, 0.1 mg/plate; Comet Assay: 0.125,
0.25,0.5,1.0 mg/mL
Time to Analysis: 24 h
All PM extracts induced higher mutational response
in TA98 (3- to 5-fold increase over spontaneous)
than in TA100 (2-to 3-fold increase). The highest
brake specific revertants (BSR) ±S9 in both strains
occurred with E20 and lowest BSR was in E5
(except in TA98 -S9). EO and E20 caused more
significant DNA damage (similar in effect) than the
other extracts. Damage was dose-dependent but
variable with increasing ethanol volume.
December 2009
D-170
-------
Reference
Particle
Exposure
Effects
Reference: Ueng et MEP (Yamaha cabin motorcycle 2-strok Route: Cell Culture
al. (2005, 0970541 50-cc engine)
Species: Human Particle Size: NR
Cell Line: Lung
epithelium CL5
(cancerous), BEAS-
2B, WI-38 normal
lung fibroblast
Dose/Concentration: 1,10,100, 200 pg/mL
Time to Analysis: microarray analysis. RT-
PCR: 2 h.ELISA: 12 h incubation.
Centrifuged 24 h post-treatment. Bioactivity:
12 h incubation. Centrifuged 24 h post-
treatment. Medium replaced 48 h post-
incubation. Fibroblasts determined 96 h
post-incubation. Time response studies: 3-
48 h treatment. Concentration response
studies: 6 h treatment.
Drug Metabolism Array Study: MEP increased
CYP1A1,CYP3A7andUGT2B.
Cytokine Array Study: MEP increased fibroblast
growth factor (FGF)-6, FGF-9, IL-1a, IL-22 and
vascular endothelial growth factor (VEGF)-D mRNA.
Oncogene, Tumor Suppressor, Estrogen
Signaling Pathway: MEP increased fra-1, c-src,
SHC, p21, COX7RP, and decreased p53 and Rb
expression.
RT-PCR: MEP increased CYP1A1, CYP1B1, IL-6,
IL-11, IL-1a, FGF-6, FGF-9, VEGF-D, fra-1 and p21.
Concentration and Time Responses:
Concentration and time-dependent increases
occurred for FGF-9, IL-1a, IL-6, IL-11, but decreased
time-dependently after 6 h exposure.
BEAS-2B Cells: MEP had concentration-dependent
increases on CYP1A1 and CYP1B1 but did not
affect anything else.
Peroxide, MEP+NAC, WI-38 Cells: MEP increased
peroxide production. The MEP+NAC treatment
reduced MEP-elevated levels of IL-1a, IL-6, FGF-9,
VEGF-D to control levels. Fibroblasts increased in
WI-38 cells.
Reference:
Umbuzeiro et al.
(2008, 1904911
Species: Salmonella
typhimurium
Strain: TA98,
YG1041 (+/-S9)
Reference:
Upadhyay et al.
(2003, 0973701
Species: Human
Cell Line A549
Reference:
Valavanidis et al.
(2005, 0964321
Cell Line: NR
Reference: Xu and
Zhang (2004,
0972311
Species: Human
Cell Line: A549
PM (urban; Sao Paulo, Brazil- Cerqueira
Cesar street station, Ibirapuera park
station) (winter- June 17, 18; average
temperature: 16°C) (EOM)
Particle Size: NR
PM (Dusseldorf, Germany) (Particles
contain carbon (19.70%),
hydrogen(1.4%),nitrogen (<05%),
oxygen(14.12%), sulfur (2.09%), ash
(63.24%)) (lonizable metals
concentrations (ppm): Co(103),
Cu(48),Cr(104),Fe(14,521 , Mn(21.3),
Ni(1,519),Ti(131),V(2,767
Particle Size: NR
PM (TSP: high volume pumps, Athens;
DEP: 2.0L engine GM Astra; GEP: 1.6L
passenger vehicle Ford; Wood smoke
soot: domestic fireplace exhaust
chimney; PMi0: high volume sampling
system, Athens; PW2S\ high volume
cascade impactor (Anderson) system
Particle Size: >10.2-<0.41 pm
(diameter)
PM (Taiyuan, Beijing; Nov-Feb)
(Taiyuan: coal-fume pollution; Beijing:
coal-fume and vehicle exhaust)
Particle Size: 2.5 pm (diameter)
Route: Cell Culture
Dose/Concentration: Cerqueira Cesar:
UPM- 156 pg/m3, EOM- 57.7 mg/total UPM;
Ibirapuera Park: UPM- 32 pg/mr EOM- 41.7
mg/total UPM; Salmonella assay- 0.5, 1, 5,
10,50, 100UPMequiv/plate(|jg)
Time to Analysis: Organic extraction 20 h.
PAH fractionation.
Route: Cell Culture
Dose/Concentration: 1, 5, 25, 100 pg/cm2;
10,25,50, 100|jg/cm2
Time to Analysis: 1, 4, 8, 12, 24 h.
Route: Incubation
Dose/Concentration: 20, 40 mg/5mL
Time to Analysis: PM incubated with H202
and 2'-deoxyguanosine (dG). Stored 3-7
days at -20°C.
Route: Cell Culture
Dose/Concentration: 5, 50, 200 pg/mL
Time to Analysis: 12-24h
The TSP and EOM were similar for both sites. The
PAH fraction had very low mutagenicity for the
Cerqueira Cesar sample in the YG1041 strain and
no mutagenicity for the Ibirapuera sample. Nitro-PAH
and oxy-PAH had similar mutagenetic activities from
both samples. S9 decreased mutagenicity in nitro-
PAH but was increased in oxy-PAH. DNA adduct
levels were dose-dependent and not different
between the two sites.
PM induced dose- and time-dependent reductions in
ds-DNA due to the formation of DNA-SB. The
soluble component caused higher DNA damage.
Apoptosis and DNA fragmentation increased dose-
dependently AYm decreased dose-dependently in
control cells, but not in cells with Bcl-xl
overexpression. PM caused activation of caspase 9.
Pretreatment with iron chelators or a free radical
scavenger reduced PM-induced DNA-SB formation,
DNA fragmentation, caspase 9 activation, and
weakened AYm reductions.
PM generated 'OH by a Fenton reaction, which is
increased by the addition of EDTA but inhibited by
deferoxamine. PM dose-dependently induced dG
hydroxylation and 8-hydroxy-2'-deoxyguanosine
formation. Transition metals Ni, V, Co, Crthat are
capable of redox cycling electron producing ROS
were found in the PM samples.
Taiyuan had a significantly higher daily PM25
average than Beijing. It was shown that the smaller
the particulate diameter, the higher the concentration
of BaP and Pb. A dose- and time-response
relationship was seen in DNA fragmentation.
December 2009
D-171
-------
Annex D References
Aam BB; Fonnum F. (2007). ROS scavenging effects of organic extract of diesel exhaust particles on human neutrophil
granulocytes and rat alveolar macrophages. Toxicology, 230: 207-18. 155123
Abou Chakra OR; Joyeux M; Nerriere E; Strub MP; Zmirou-Navier D. (2007). Genotoxicity of organic extracts of urban
airborne particulate matter: an assessment within a personal exposure study. Chemosphere, 66: 1375-81. 098819
Adamson IYR; Vincent R; Bakowska J. (2003). Differential production of metalloproteinases after instilling various urban
air particle samples to rat lung. Exp Lung Res, 29: 375-388. 087943
AgopyanN; Head J; Yu S; Simon SA. (2004). TRPV1 receptors mediate particulate matter-induced apoptosis. Am J
Physiol Lung Cell Mol Physiol, 286: 563-572. 156198
Agopyan N; Li L; Yu S; Simon SA. (2003). Negatively charged 2- and 10-"mu"m particles activate vanilloid receptors,
increase cAMP, and induce cytokine release. Toxicol Appl Pharmacol, 186: 63-76. 056065
Ahn E-K; Yoon H-K; Jee Bo K; Ko H-J; Lee K-H; Kim Hyung J; Lim Y (2008). COX-2 expression and inflammatory
effects by diesel exhaust particles in vitro and in vivo. Toxicol Lett, 176: 178-187. 156199
Ahsan MK; Nakamura H; Tanito M; Yamada K; Utsumi H; Yodoi J. (2005). Thioredoxin-1 suppresses lung injury and
apoptosis induced by diesel exhaust particles (DEP) by scavenging reactive oxygen species and by inhibiting DEP-
induced downregulation of Akt. Free Radic Biol Med, 39: 1549-1559. 156200
Alfaro-Moreno E; Martinez L; Garcia-Cuellar C; Bonner JC; Murray JC; Rosas I; Resales SP; Osornio-Vargas AR. (2002).
Biologic effects induced in vitro by PM10 from three different zones of Mexico City. Environ Health Perspect, 110:
715-720. 156204
Amakawa K; Terashima T; Matsuzaki T; Matsumaru A; Sagai M; Yamaguchi K. (2003). Suppressive effects of diesel
exhaust particles on cytokine release from human and murine alveolar macrophages. Exp Lung Res, 29: 149-164.
156211
Amara N; Bachoual R; Desmard M; Golda S; Guichard C; Lanone S; Aubier M; Ogier-Denis E; Boczkowski J. (2007).
Diesel exhaust particles induce matrix metalloprotease-1 in human lung epithelial cells via a NADP(H)
oxidase/NOX4 redox-dependent mechanism. Am J Physiol Lung Cell Mol Physiol, 293: LI70-181. 156212
Andre E; Stoeger T; Takenaka S; Bahnweg M; Ritter B; Karg E; Lentner B; Reinhard C; Schulz H; Wjst M. (2006).
Inhalation of ultrafine carbon particles triggers biphasic pro-inflammatory response in the mouse lung. Eur Respir J,
28: 275-285. 091376
Anselme F; Loriot S; Henry J-P; Dionnet F; Napoleoni J-G; Thnillez C; Morin J-P (2007). Inhalation of diluted diesel
engine emission impacts heart rate variability and arrhythmia occurrence in a rat model of chronic ischemic heart
failure. Arch Toxicol, 81: 299-307. 097084
Anseth JW; Goffin AJ; Fuller GG; Ohio AJ; Kao PN; Upadhyay D. (2005). Lung surfactant gelation induced by epithelial
cells exposed to air pollution or oxidative stress. Am J Respir Cell Mol Biol, 33: 161-168. 088646
Antonini JM; Taylor MD; Leonard SS; Lawryk NJ; Shi X; Clarke RW; Roberts JR. (2004). Metal composition and
solubility determine lung toxicity induced by residual oil fly ash collected from different sites within a power plant.
Mol Cell Biochem, 255: 257-265. 097199
Apicella C; Custidiano A; Miranda S; Novoa L; Dokmetjian J; Gentile T. (2006). Differential macrophage modulation of
asymmetric IgG antibody synthesis by soluble or particulate stimuli. Immunol Lett, 103: 177-185. 096586
Arantes-Costa F; Lopes F; Toledo A; Magliarelli-Filho P; Moriya H; Carvalho-Oliveira R; Mauad T; Saldiva P; Martins M.
(2008). Effects of residual oil fly ash (ROFA) in mice with chronic allergic pulmonary inflammation. Toxicol
Pathol, 36: 680-686. 187137
Archer AJ; Cramton JL; Pfau JC; Colasurdo G; Holian A. (2004). Airway responsiveness after acute exposure to urban
particulate matter 1648 in a DO1110 murine model. Am J Physiol, 286: L337-L343. 088097
Arimoto T; Takano H; Inoue K; Yanagisawa R; Yoshino S; Yamaki K; Yoshikawa T. (2007). Pulmonary exposure to diesel
exhaust particle components enhances circulatory chemokines during lung inflammation. Int J Immunopathol
Pharmacol, 20: 197-201.097973
December 2009 D-172
-------
Arrieta DE; Ontiveros CC; Li W-W; Garcia JH; Denison MS; McDonald JD; Burchiel SW; Washburn BS. (2003). Aryl
hydrocarbon receptor-mediated activity of particulate organic matter from the Paso del Norte airshed along the US-
Mexico border. Environ Health Perspect, 111: 1299-1305. 096210
Auger F; Gendron MC; Chamot C; Marano F; Dazy AC. (2006). Responses of well-differentiated nasal epithelial cells
exposed to particles: role of the epithelium in airway inflammation. Toxicol Appl Pharmacol, 215: 285-294. 156235
Bachoual R; Boczkowski J; Goven D; Amara N; Tabet L; On D; Lecon-Malas V; Aubier M; Lanone S. (2007). Biological
effects of particles from the Paris subway system. Chem Res Toxicol, 20: 1426-1433. 155667
Bagate K; Meiring JJ; Cassee FR; Borm PJA. (2004). The effect of particulate matter on resistance and conductance
vessels in the rat. Inhal Toxicol, 16: 431-436. 055638
Bagate K; Meiring JJ; Gerlofs-Nijland ME; Cassee FR; Borm PJA. (2006). Signal transduction pathways involved in
particulate matter induced relaxation in rat aorta—spontaneous hypertensive versus Wistar Kyoto rats. Toxicol In
Vitro, 20: 52-62. 097608
Bagate K; Meiring JJ; Gerlofs-Nijland ME; Cassee FR; Wiegand H; Osornio-Vargas A; Borm PJA. (2006). Ambient
particulate matter affects cardiac recovery in a Langendorff ischemia model. Inhal Toxicol, 18: 633-643. 096157
Bao L; Chen S; Wu L; Hei Tom K; Wu Y; Yu Z; Xu A. (2007). Mutagenicity of diesel exhaust particles mediated by cell-
particle interaction in mammalian cells. Toxicol Sci, 229: 91-100. 097258
Barrett EG; Henson RD; Seilkop SK; McDonald JD; Reed MD. (2006). Effects of hardwood smoke exposure on allergic
airway inflammation in mice. Inhal Toxicol, 18: 33-43. 155677
Bartoli CR; Wellenius GA; Coull BA; Akiyama I; Diaz EA; Lawrence J; Okabe K; Verrier RL; Godleski JJ. (2009).
Concentrated ambient particles alter myocardial blood flow during acute ischemia in conscious canines. Environ
Health Perspect, 117: 333-337. 179904
Bartoli CR; Wellenius GA; Diaz EA; Lawrence J; Coull BA; Akiyama I; Lee LM; Okabe K; Verrier RL; Godleski JJ.
(2009). Mechanisms of Inhaled Fine Particulate Air Pollution-Induced Arterial Blood Pressure Changes. Environ
Health Perspect, 117: 361-366. 156256
Batalha JR; Saldiva P H; Clarke RW; Coull BA; Stearns RC; Lawrence J; Murthy GG; Koutrakis P; Godleski JJ. (2002).
Concentrated ambient air particles induce vasoconstriction of small pulmonary arteries in rats. Environ Health
Perspect, 110: 1191-1197. 088109
Baulig A; Blanche! S; Rumelhard M; Lacroix G; Marano F; Baeza-Squiban A. (2007). Fine urban atmospheric particulate
matter modulates inflammatory gene and protein expression in human bronchial epithelial cells. Front Biosci, 12:
771-82. 151733
Bayram H; Ito K; Issa R; Ito M; Sukkar M; Chung KF. (2006). Regulation of human lung epithelial cell numbers by diesel
exhaust particles. Eur Respir J, 27: 705-713. 088439
Becher R; Bucht A; Ovrevik J; Hongslo JK; Dahlman HJ; Samuelsen JT; Schwarze PE. (2007). Involvement of NADPH
oxidase and iNOS in rodent pulmonary cytokine responses to urban air and mineral particles. Inhal Toxicol, 19:
645-655. 097125
Beck-Speier I; Dayal N; Karg E; Maier KL; Schumann G; Schulz H; Semmler M; Takenaka S; Stettmaier K; Bors W; Ghio
A; Samet JM; Heyder J. (2005). Oxidative stress and lipid mediators induced in alveolar macrophages by ultrafine
particles. Free Radic Biol Med, 38: 1080-1092. 156262
Becker S; Dailey LA; Soukup JM; Grambow SC; Devlin RB; Huang YC. (2005). Seasonal variations in air pollution
particle-induced inflammatory mediator release and oxidative stress. Environ Health Perspect, 113: 1032-1038.
088592
Becker S; Mundandhara S; Devlin RB; Madden M. (2005). Regulation of cytokine production in human alveolar
macrophages and airway epithelial cells in response to ambient air pollution particles: further mechanistic studies.
Toxicol Appl Pharmacol, 207: 269-275. 088590
Bhattacharyya SN; Dubick MA; Yantis LD; Enriquez JI; Buchanan KG; Batra SK; Smiley RA. (2004). In vivo effect of
wood smoke on the expression of two mucin genes in rat airways. Inflammation, 28: 67-76. 088095
Binkova B; Chvatalova I; Lnenickova Z; Milcova A; Tulupova E; Farmer PB; Sram RJ. (2007). PAH-DNA adducts in
environmentally exposed population in relation to metabolic and DNA repair gene polymorphisms. Mutat Res Fund
Mol Mech Mutagen, 620: 49-61. 156273
December 2009 D-173
-------
Bitterle E; Karg E; Schroeppel A; Kreyling WG; Tippe A; Perron GA; Schmid O; Heyder J; Maier KL; Hofer T. (2006).
Dose-controlled exposure of A549 epithelial cells at the air-liquid interface to airborne ultrafine carbonaceous
particles. Chemosphere, 65: 1784-1790. 156276
Blanche! S; Ramgolam K; Baulig A; Marano F; Baeza-Squiban A. (2004). Fine particulate matter induces amphiregulin
secretion by bronchial epithelial cells. Am J Respir Cell Mol Biol, 30: 421-427. 087982
Bonvallot V; Baeza-Squiban A; Baulig A; Brulant S; Boland S; MuzeauF; Barouki R; Marano F. (2001). Organic
compounds from diesel exhaust particles elicit a proinflammatory response in human airway epithelial cells and
induce cytochrome p450 1A1 expression. Am J Respir Cell Mol Biol, 25: 515-521. 156283
Brits E; Schoeters G; Verschaeve L. (2004). Genotoxicity of PM10 and extracted organics collected in an industrial, urban,
and rural area in Flanders, Belgium. Environ Res, 96: 109-118. 087397
Brown DM; Hutchison L; Donaldson K; Stone V. (2007). The effects of PM10 particles and oxidative stress on
macrophages and lung epithelial cells: modulating effects of calcium-signaling antagonists. Am J Physiol Lung Cell
Mol Physiol, 292: 1444-1451. 156300
Brown LE; Trought KR; Bailey CI; demons JH. (2005). 2,3,7,8-TCDD equivalence and mutagenic activity associated
with PM10 from three urban locations inNew Zealand. Sci Total Environ, 349: 161-174. 095919
Bunger J; Krahl J; Munack A; Ruschel Y; Schroder O; Emmert B; Westphal G; Muller M; Hallier E; Bruning T. (2007).
Strong mutagenic effects of diesel engine emissions using vegetable oil as fuel. Arch Toxicol, 81: 599-603. 156305
Bunger J; Krahl J; Weigel A; Schroder O; Bruning T; Muller M; Hallier E; Westphal G. (2006). Influence of fuel properties,
nitrogen oxides, and exhaust treatment by an oxidation catalytic converter on the mutagenicity of diesel engine
emissions. Arch Toxicol, 80: 540-546. 156303
Burchiel SW; Lauer FT; Dunaway SL; Zawadzki J; McDonald JD; Reed MD. (2005). Hardwood smoke alters murine
splenic T cell responses to mitogens following a 6-month whole body inhalation exposure. Toxicol Appl Pharmacol,
202: 229-236. 088090
Burchiel SW; Lauer FT; McDonald JD; Reed MD. (2004). Systemic immunotoxicity in AJ mice following 6-month whole
body inhalation exposure to diesel exhaust. Toxicol Appl Pharmacol, 196: 337-345. 055557
Calcabrini A; Meschini S; Marra M; Falzano L; Colone M; De Berardis B; Paoletti L; Arancia G; Fiorentini C. (2004). Fine
environmental particulate engenders alterations in human lung epithelial A549 cells. Environ Res, 95: 82-91.
096865
Calderon-Garciduenas L; Maronpot RR; Torres-Jardon R; Henriquez-Roldan C; Schoonhoven R; Acuna-Ayala H;
Villarreal-Calderon A; Nakamura J; Fernando R; Reed W; Azzarelli B; Swenberg JA. (2003). DNA damage in nasal
and brain tissues of canines exposed to air pollutants is associated with evidence of chronic brain inflammation and
neurodegeneration. Toxicol Pathol, 31: 524-538. 156316
Campbell A; Oldham M; Becaria A; Bondy SC; Meacher D; Sioutas C; Misra C; Mendez LB; Kleinman M. (2005).
Particulate matter in polluted air may increase biomarkers of inflammation in mouse brain. Neurotoxicology, 26:
133-140.087217
Campen MJ; Babu NS; Helms GA; Pett S; Wernly J; Mehran R; McDonald JD. (2005). Nonparticulate components of
diesel exhaust promote constriction in coronary arteries from ApoE -/- mice. Toxicol Sci, 88: 95-102. 083977
Campen MJ; McDonald JD; Gigliotti AP; Seilkop SK; Reed MD; Benson JM. (2003). Cardiovascular effects of inhaled
diesel exhaust in spontaneously hypertensive rats. Cardiovasc Toxicol, 3: 353-361. 055626
Campen MJ; McDonald JD; Reed MD; Seagrave J. (2006). Fresh gasoline emissions, not paved road dust, alter cardiac
repolarization in ApoE-/- mice. Cardiovasc Toxicol, 6: 199-210. 096879
Cao D; Tal TL; Graves LM; Gilmour I; Linak W; Reed W; Bromberg PA; Samet JM. (2007). Diesel exhaust particulate-
induced activation of Stat3 requires activities of EGFR and Src in airway epithelial cells. Am J Physiol Lung Cell
Mol Physiol, 292: L422-L429. 156322
Cao Q; Zhang S; Dong C; Song W. (2007). Pulmonary responses to fine particles: Differences between the spontaneously
hypertensive rats and wistar kyoto rats. Toxicol Lett, 171: 126-137. 097491
December 2009 D-174
-------
Carter JM; Corson N; Driscoll KE; Elder A; Finkelstein JN; Harkema JN; Gelein R; Wade-Mercer P; Nguyen K;
Oberdorster G. (2006). A comparative dose-related response of several key pro- and antiinflammatory mediators in
the lungs of rats, mice, and hamsters after subchronic inhalation of carbon black. J Occup Environ Med, 48:1265-
1278. 095936
Carvalho-Oliveira R; Pozo RMK; Lobo DJA; Lichtenfels AJFC; Martins-Junior HA; Bustilho JOWV; Saiki M; Sato IM;
Saldiva PFfN. (2005). Diesel emissions significantly influence composition and mutagenicity of ambient particles: a
case study in Sao Paulo, Brazil. Environ Res, 98: 1-7. 077898
Cassee FR; Boere AJF; Fokkens PHB; Leseman DLAC; Sioutas C; Kooter IM; Dormans JAMA. (2005). Inhalation of
concentrated particulate matter produces pulmonary inflammation and systemic biological effects in compromised
rats. JToxicol Environ Health A, 68: 773-796. 087962
Chan RC-F; Wang M; Li N; Yanagawa Y; Onoe K; Lee JJ; Nel AE. (2006). Pro-oxidative diesel exhaust particle chemicals
inhibit LPS-induced dendritic cell responses involved in T-helper differentiation. J Allergy Clin Immunol, 118: 455-
465. 097468
Chang C-C; Chen S-H; Ho S-H; Yang C-Y; Wang H-D; Tsai M-L. (2007). Proteomic analysis of proteins from
bronchoalveolar lavage fluid reveals the action mechanism of ultrafine carbon black-induced lung injury in mice.
Proteomics, 7: 4388-4397. 097475
Chang C-C; Chiu H-F; Wu Y-S; Li Y-C; Tsai M-L; Shen C-K; Yang C-Y. (2005). The induction of vascular endothelial
growth factor by ultrafine carbon black contributes to the increase of alveolar-capillary permeability. Environ
Health Perspect, 113: 454-460. 097776
Chang C-C; Hwang J-S; Chan C-C; Cheng T-J. (2007). Interaction effects of ultrafine carbon black with iron and nickel on
heart rate variability in spontaneously hypertensive rats. Environ Health Perspect, 115: 1012-1017. 155720
Chang C-C; Hwang J-S; Chan C-C; Wang P-Y; Hu T-H; Cheng T-J. (2004). Effects of concentrated ambient particles on
heart rate, blood pressure, and cardiac contractility in spontaneously hypertensive rats. Inhal Toxicol, 16: 421-429.
055637
Chauhan V; Breznan D; Goegan P; Nadeau D; Karthikeyan S; Brook JR; Vincent R. (2004). Effects of ambient air particles
on nitric oxide production in macrophage cell lines. Cell Biol Toxicol, 20: 221-239. 096682
Chauhan V; Breznan D; Thomson E; Karthikeyan S; Vincent R. (2005). Effects of ambient air particles on the endothelin
system in human pulmonary epithelial cells (A549). Cell Biol Toxicol, 21: 191-205. 155722
Che W; Zhang Z; Zhang H; Wu M; Liang Y; Liu F; Shu Y; Li N. (2007). Compositions and oxidative damage of
condensate, particulate and semivolatile organic compounds from gasoline exhausts. Environ Toxicol Pharmacol,
24: 11-18.096460
Chen LC; Hwang JS. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice IV
Characterization of acute and chronic effects of ambient air fine particulate matter exposures on heart-rate
variability. Inhal Toxicol, 17: 209-216. 087218
Chen LC; Nadziejko C. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice V CAPs
exacerbate aortic plaque development in hyperlipidemic mice. Inhal Toxicol, 17: 217-224. 087219
Cheng MD; Malone B; Storey JM. (2003). Monitoring cellular responses of engine-emitted particles by using a direct air-
cell interface deposition technique. Chemosphere, 53: 237-243. 156337
Chin BY; Trush MA; Choi AM; Risby TH. (2003). Transcriptional regulation of the HO-1 gene in cultured macrophages
exposed to model airborne particulate matter. Am J Physiol Lung Cell Mol Physiol, 284: L473-480. 156340
Cho H-Y; Jedlicka AE; Clarke R; Kleeberger SR. (2005). Role of Toll-like receptor-4 in genetic susceptibility to lung
injury induced by residual oil fly ash. Physiol Genomics, 22: 108-117. 156344
Churg A; Brauer M; del Carmen Avila-Casado M; Fortoul TI; Wright JL. (2003). Chronic exposure to high levels of
particulate air pollution and small airway remodeling. Environ Health Perspect, 111: 714-718. 087899
Churg A; Xie C; Wang X; Vincent R; Wang RD. (2005). Air pollution particles activate NF-kappaB on contact with airway
epithelial cell surfaces. Toxicol Appl Pharmacol, 208: 37-45. 088281
Ciencewicki J; Gowdy K; Krantz QT; Linak WP; Brighton L; Gilmour MI; Jaspers I. (2007). Diesel exhaust enhanced
susceptibility to influenza infection is associated with decreased surfactant protein expression. Inhal Toxicol, 19:
1121-1133.096557
December 2009 D-175
-------
Corey LM; Baker C; Luchtel Daniel L. (2006). Heart-rate variability in the apolipoprotein E knockout transgenic mouse
following exposure to Seattle particulate matter. J Toxicol Environ Health A, 69: 953-965. 156366
Costa DL; Lehmann JR; Winsett D; Richards J; Ledbetter AD; Dreher KL. (2006). Comparative pulmonary toxicological
assessment of oil combustion particles following inhalation or instillation exposure. Toxicol Sci, 91: 237-246.
088438
Courtois A; Andujar P; Ladeiro Y; Baudrimont I; Delannoy E; Leblais V; Begueret H; Galland MAB; Brochard P; Marano
F. (2008). Impairment of NO-Dependent Relaxation in Intralobar Pulmonary Arteries: Comparison of Urban
Particulate Matter and Manufactured Nanoparticles. Environ Health Perspect, 116: 1294-1299. 156369
Cozzi E; Hazarika S; Stallings HW; Cascio WE; Devlin RB; Lust RM; Wingard CJ; Van Scott MR. (2006). Ultrafine
particulate matter exposure augments ischemia-reperfusion injury in mice. Am J Physiol, 291: H894-H903. 091380
Dagher Z; Garcon G; Billet S; Verdin A; Ledoux F; Courcot D; Aboukais A; Shirali P. (2007). Role of nuclear factor-kappa
B activation in the adverse effects induced by air pollution particulate matter (PM25) in human epithelial lung cells
(LI32) in culture. J Appl Toxicol, 27: 284-290. 097566
Dai J; Xie C; Vincent R; Churg A. (2003). Air pollution particles produce airway wall remodeling in rat tracheal explants.
Am J Respir Cell Mol Biol, 29: 352-358. 087944
Day KC; Reed MD; McDonald JD; Keilkop SK; Barrett EG. (2008). Effects of gasoline engine emissions on preexisting
allergic airway responses in mice. Inhal Toxicol, 20: 1145-1155. 190204
DeMarini DM; Brooks LR; Warren SH; Kobayashi T; Gilmour MI; Singh P. (2004). Bioassay-directed fractionation and
Salmonella mutagenicity of automobile and forklift diesel exhaust particles. Environ Health Perspect, 112: 814-819.
066329
de Haar C; Hassing I; Bol M; Bleumink R; Pieters R. (2005). Ultrafine carbon black particles cause early airway
inflammation and have adjuvant activity in a mouse allergic airway disease model. Toxicol Sci, 87: 409-418.
097872
de Haar C; Hassing I; Bol M; Bleumink R; Pieters R. (2006). Ultrafine but not fine particulate matter causes airway
inflammation and allergic airway sensitization to co-administered antigen in mice. Clin Exp Allergy, 36: 1469-
1479. 144746
de Haar C; Kool M; Hassing I; Bol M; Lambrecht BN; Pieters R. (2008). Lung dendritic cells are stimulated by ultrafine
particles and play a key role in particle adjuvant activity. J Allergy Clin Immunol, 121:1246-1254. 187128
De Kok TM; Hogervorst JG; Briede JJ; Van Herwijnen MH; Maas LM; Moonen EJ; Driece HA; Kleinjans JC. (2005).
Genotoxicity and physicochemical characteristics of traffic-related ambient particulate matter. Environ Mol
Mutagen, 46: 71-80. 088656
Dick CAJ; Brown DM; Donaldson K; Stone V (2003). The role of free radicals in the toxic and inflammatory effects of
four different ultrafine particle types. Inhal Toxicol, 15: 39-52. 036605
Doherty SP; Prophete C; Maciejczyk P; Salnikow K; Gould T; Larson T; Koenig J; Jaques P; Sioutas C; Zelikoff JT;
Lippmann M; Cohen MD. (2007). Detection of changes in alveolar macrophage iron status induced by select
PM25-associated components using iron-response protein binding activity. Inhal Toxicol, 19: 553-562. 096532
Dong CC; Yin XJ; Ma JYC; Millecchia L; Barger MW; Roberts JR; Zhang X-D; Antonini JM; Ma JKH. (2005). Exposure
of Brown Norway Rats to diesel exhaust particles Prior to ovalbumin (OVA) sensitization elicits IgE adjuvant
activity but attenuates OVA-induced airway inflammation. Toxicol Sci, 88: 150-160. 088079
Dong CC; Yin XJ; Ma JYC; Millecchia L; Wu Z-X; Barger MW; Roberts JR; Antonini JM; Dey RD; Ma JKH. (2005).
Effect of diesel exhaust particles on allergic reactions and airway responsiveness in ovalbumin-sensitized Brown
Norway Rats. Toxicol Sci, 88: 202-212. 088083
Doornaert B; Leblond V; Galiacy S; Gras G; Planus E; Laurent V; Isabey D; Lafuma C. (2003). Negative impact of DEP
exposure on human airway epithelial cell adhesion, stiffness, and repair. Am J Physiol Lung Cell Mol Physiol, 284:
L119-L132. 156410
Dostert C; Petrilli V; Van Bruggen R; Steele C; Mossman BT; Tschopp J. (2008). Innate Immune Activation Through Nalp3
Inflammasome Sensing of Asbestos and Silica. Science, 320: 674-677. 155753
Doyle M; Sexton KG; Jeffries H; Bridge K; Jaspers I. (2004). Effects of 1,3-butadiene, isoprene, and their photochemical
degradation products on human lung cells. Environ Health Perspect, 112: 1488-1495. 088404
December 2009 D-176
-------
Drela N; Zesko I; Jakubowska M; Biernacka M. (2006). CD28 in thymocyte development and peripheral T cell activation
in mice exposed to suspended particulate matter. Toxicol Appl Pharmacol, 215: 179-188. 096352
Duvall RM; Norris GA; Dailey LA; Burke JM; McGee JK; Gilmour MI; Gordon T; Devlin RB. (2008). Source
apportionment of particulate matter in the US and associations with lung inflammatory markers. Inhal Toxicol, 20:
671-683. 097969
Dvonch JT; Brook RD; Keeler GJ; Rajagopalan S; D'Alecy LG; Marsik FJ; Morishita M; Yip FY; Brook JR; Timm EJ;
Wagner JG; Harkema JR. (2004). Effects of concentrated fine ambient particles on rat plasma levels of asymmetric
dimethylarginine. Inhal Toxicol, 16: 473-480. 055741
Dybdahl M; Risom L; Bornholdt J; Autrup H; Loft S; Wallin H. (2004). Inflammatory and genotoxic effects of diesel
particles in vitro and in vivo. DNARepair (Amst), 562: 119-131. 089013
Dybing E; Lovdal T; Hetland RB; Lovik M; Schwarze PE. (2004). Respiratory allergy adjuvant and inflammatory effects
of urban ambient particles. Toxicol Sci, 198: 307-314. 097545
ElAssouli S; AlQahtani M; Milaat W. (2007). Genotoxicity of air borne particulates assessed by comet and the samonella
mutagenicity test in Jeddah, Saudi Arabia. Int J Environ Res Publ Health, 4: 216-223. 186914
Elder A; Gelein R; Finkelstein J; Phipps R; Frampton M; Utell M; Kittelson DB; Watts WF; Hopke P; Jeong CH; Kim E;
Liu W; Zhao W; Zhuo L; Vincent R; Kumarathasan P; Oberdorster G (2004). On-road exposure to highway
aerosols 2 Exposures of aged, compromised rats. Inhal Toxicol, 16 Suppl 1: 41-53. 087354
Elder A; Gelein R; Finkelstein JN; Driscoll KE; Harkema J; Oberdorster G. (2005). Effects of subchronically inhaled
carbon black in three species I Retention kinetics, lung inflammation, and histopathology. Toxicol Sci, 88: 614-629.
088194
Elder ACP; Gelein R; Azadniv M; Frampton M; Finkelstein J; Oberdorster G. (2004). Systemic effects of inhaled ultrafine
particles in two compromised, aged rat strains. Inhal Toxicol, 16: 461-471. 055642
Endo O; Sugita K; Goto S; Amagai T; Matsushita H. (2003). Mutagenicity of size-fractioned airborne particles collected
with Andersen low pressure impactor. Eisei Kagaku, 49: 11-11. 097260
Erdinger L; Durr M; Hopker KA. (2005). Correlations between mutagenic activity of organic extracts of airborne
particulate matter, NOx and sulphur dioxide in southern Germany: results of a two-year study. Environ Sci Pollut
Res Int, 12: 10-20. 156423
Evans S-A; Al-Mosawi A; Adams RA; Berube KA. (2006). Inflammation, edema, and peripheral blood changes in lung-
compromised rats after instillation with combustion-derived and manufactured nanoparticles. Exp Lung Res, 32:
363-378. 097066
Farraj AK; Haykal-Coates N; Ledbetter AD; Evansky PA; Gavett SH. (2006). Inhibition of pan neurotrophin receptor p75
attenuates diesel particulate-induced enhancement of allergic airway responses in C57/B16J mice. Inhal Toxicol,
18:483-491.088469
Farraj AK; Haykal-Coates N; Ledbetter AD; Evansky PA; Gavett SH. (2006). Neurotrophin mediation of allergic airways
responses to inhaled diesel particles in mice. Toxicol Sci, 94: 183-192. 141730
Fedulov AV; Leme A; Yang Z; Dahl M; Lim R; Mariani TJ; Kobzik L. (2008). Pulmonary exposure to particles during
pregnancy causes increased neonatal asthma susceptibility. Am J Respir Cell Mol Biol, 38: 57-67. 097482
Finkelman FD; Yang M; Orekhova T; Clyne E; Bernstein J; Whitekus M; Diaz-Sanchez D; Morris SC. (2004). Diesel
exhaust particles suppress in vivo IFN-gamma production by inhibiting cytokine effects onNK and NKT cells. J
Immunol, 172: 3808-3813. 096572
Finnerty K; Choi J-E; Lau A; Davis-Gorman G; Diven C; Seaver N; Linak William P; Witten M; McDonagh Paul F.
(2007). Instillation of coarse ash particulate matter and lipopolysaccharide produces a systemic inflammatory
response in mice. J Toxicol Environ Health A, 70: 1957-1966. 156434
Floyd HS; Chen LC; Vallanat B; Dreher K. (2009). Fine ambient air particulate matter exposure induces molecular
alterations associated with vascular disease progression within plaques of atherosclerotic susceptible mice. Inhal
Toxicol, 21: 394-403. 190350
Folkmann JK; Risom L; Hansen CS; Loft S; Moller P. (2007). Oxidatively damaged DNA and inflammation in the liver of
dyslipidemic ApoE-/- mice exposed to diesel exhaust particles. Toxicology, 237:134-44. 097344
December 2009 D-177
-------
Fritsch S; Diabate S; Krug HE (2006). Incinerator fly ash provokes alteration of redox equilibrium and liberation of
arachidonic acid in vitro. Biol Chem, 387: 1421-1428. 156452
Fujii T; Hayashi S; Hogg JC; Mukae H; Suwa T; Goto Y; Vincent R; Van Eeden SF. (2002). Interaction of alveolar
macrophages and airway epithelial cells following exposure to particulate matter produces mediators that stimulate
the bone marrow. Am J Respir Cell Mol Biol, 27: 34-41. 036478
Fujii T; Hayashi S; Hogg JC; Vincent R; Van Eeden SF. (2001). Particulate matter induces cytokine expression in human
bronchial epithelial cells. Am J Respir Cell Mol Biol, 25: 265-271. 156455
Fujimaki H; Kurokawa Y. (2004). Diesel exhaust-associated gas components enhance chemokine production by cervical
lymph-node cells from mice immunized with sugi basic proteins. , 16: 61-65. 096575
Fujimaki H; Kurokawa Y; Yamamoto S; Satoh M. (2006). Distinct requirements for interleukin-6 in airway inflammation
induced by diesel exhaust in mice. Immunopharmacol Immunotoxicol, 28: 703-714. 096601
Fujimaki H; Yamamoto S; Kurokawa Y. (2005). Effect of diesel exhaust on immune responses in C57BL/6 mice
intranasally immunized with pollen antigen. JUOEH, 27: 11-24. 156456
Fujimoto A; Tsukue N; Watanabe M; Sugawara I; Yanagisawa R; Takano H; Yoshida S; Takeda K. (2005). Diesel exhaust
affects immunological action in the placentas of mice. Environ Toxicol, 20: 431-440. 096556
Furuyama A; Hirano S; Koike E; Kobayashi T. (2006). Induction of oxidative stress and inhibition of plasminogen
activator inhibitor-1 production in endothelial cells following exposure to organic extracts of diesel exhaust
particles and urban fine particles. Arch Toxicol, 80: 154-162. 097056
Gabelova A; Valovicova Z; Bacova G; Labaj J; Binkova B; Topinka J; Sevastyanova O; Sram RJ; Kalina I; Habalova V;
Popov TA; Panev T; Farmer PB. (2007). Sensitivity of different endpoints for in vitro measurement of genotoxicity
of extractable organic matter associated with ambient airborne particles (PM10). Mutat Res Fund Mol Mech
Mutagen, 620: 103-113. 156458
Gabelova A; Valovicova Z; Labaj J; Bacova G; Binkova B; Farmer Peter B. (2007). Assessment of oxidative DNA damage
formation by organic complex mixtures from airborne particles PM(10). Mutat Res, 620: 135-144. 156457
Gao F; Barchowsky A; Nemec AA; Fabisiak JP. (2004). Microbial stimulation by Mycoplasma fermentans synergistically
amplifies IL-6 release by human lung fibroblasts in response to residual oil fly ash (ROFA) and nickel. Toxicol Sci,
81:467-479.087950
Garcon G; Dagher Z; Zerimech F; Ledoux F; Courcot D; Aboukais A; Puskaric E; Shirali P. (2006). Dunkerque City air
pollution particulate matter-induced cytotoxicity, oxidative stress and inflammation in human epithelial lung cells
(LI 32) in culture. Toxicol In Vitro, 20: 519-528. 096633
Gavett SH; Haykal-Coates N; Copeland L B; Heinrich J; Gilmour MI. (2003). Metal composition of ambient PM2.5
influences severity of allergic airways disease in mice. Environ Health Perspect, 111: 1471-1477. 053153
Geng H; Meng Z; Zhang Q. (2005). Effects of blowing sand fine particles on plasma membrane permeability and fluidity,
and intracellular calcium levels of rat alveolar macrophages. Toxicol Lett, 157: 129-137. 096689
Geng H; Meng Z; Zhang Q. (2006). In vitro responses of rat alveolar macrophages to particle suspensions and water-
soluble components of dust storm PM(2.5). Toxicol In Vitro, 20: 575-584. 097026
Gerlofs-Nijland ME; Rummelhard M; Boere AJF; Leseman DLAC; Duffin R; Schins RPF; Borm PJA; Sillanpaa M;
Salonen RO; Cassee FR. (2009). Particle induced toxicity in relation to transition metal and poly cyclic aromatic
hydrocarbon contents . Environ Sci Technol, 43: 4729-4736. 190353
Gerlofs-Nijland ME; Boere AJ; Leseman DL; Dormans JA; Sandstrom T; Salonen RO; van Bree L; Cassee FR. (2005).
Effects of particulate matter on the pulmonary and vascular system: time course in spontaneously hypertensive rats.
Part Fibre Toxicol, 2:2. 088652
Gerlofs-Nijland ME; Dormans JA; Bloemen HJ; Leseman DL; John A; Boere F; Kelly FJ; Mudway IS; Jimenez AA;
Donaldson K; Guastadisegni C; Janssen NA; Brunekreef B; Sandstrom T; van Bree L; Cassee FR. (2007). Toxicity
of coarse and fine particulate matter from sites with contrasting traffic profiles. Inhal Toxicol, 19: 1055-1069.
097840
Ghelfi E; Rhoden CR; Wellenius GA; Lawrence J; Gonzalez-Flecha B. (2008). Cardiac oxidative stress and
electrophysiological changes in rats exposed to concentrated air particles are mediated by TRP-dependent
pulmonary reflexes. Toxicol Sci, 102: 328-336. 156468
December 2009 D-178
-------
Ohio AJ; Cohen MD. (2005). Disruption of iron homeostasis as a mechanism of biologic effect by ambient air pollution
particles. Inhal Toxicol, 17: 709-716. 088272
Ohio AJ; Piantadosi CA; Wang X; Dailey LA; Stonehuerner JD; Madden MC; Yang F; Dolan KG; Garrick MD; Garrick
LM. (2005). Divalent metal transporter-1 decreases metal-related injury in the lung. Am J Physiol, 289: L460-L467.
088275
Gilmour MI; McGee J; Duvall Rachelle M; Dailey L; Daniels M; Boykin E; Cho S-H; Doerfler D; Gordon T; Devlin
Robert B. (2007). Comparative toxicity of size-fractionated airborne particulate matter obtained from different
cities in the United States. Inhal Toxicol, 19 Suppl 1: 7-16. 096433
Gilmour MI; O'Connor S; Dick CAJ; Miller CA; Linak WP. (2004). Differential pulmonary inflammation and in vitro
cytotoxicity of size-fractionated fly ash particles from pulverized coal combustion. J Air Waste Manag Assoc, 54:
286-295. 057420
Gilmour PS; Morrison ER; Vickers MA; Ford I; Ludlam CA; Greaves M; Donaldson K; MacNee W. (2005). The
procoagulant potential of environmental particles (PM10). Occup Environ Med, 62: 164-171. 087410
Gilmour PS; Rahman I; Donaldson K; MacNee W. (2003). Histone acetylation regulates epithelial IL-8 release mediated by
oxidative stress from environmental particles. Am J Physiol Lung Cell Mol Physiol, 284: L533-L540. 096959
Gilmour PS; Schladweiler MC; Nyska A; McGee JK; Thomas R; Jaskot RH; Schmid J; Kodavanti UP. (2006). Systemic
imbalance of essential metals and cardiac gene expression in rats following acute pulmonary zinc exposure. J
Toxicol Environ Health A, 69: 2011-2032. 156472
Gilmour PS; Schladweiler MC; Richards JH; Ledbetter AD; Kodavanti UP. (2004). Hypertensive rats are susceptible to
TLR4-mediated signaling following exposure to combustion source particulate matter. Inhal Toxicol, 16 Suppl 1: 5-
18.087948
Gilmour PS; Ziesenis A; Morrison ER; Vickers MA; Drost EM; Ford I; Karg E; Mossa C; Schroeppel A; Perron GA;
Heyder J; Greaves M; MacNee W; Donaldson K. (2004). Pulmonary and systemic effects of short-term inhalation
exposure to ultrafine carbon black particles. Toxicol Appl Pharmacol, 195: 35-44. 054175
Godleski JJ; Clarke RW; Coull BA; Saldiva PHN; Jiang NF; Lawrence J; Koutrakis P. (2002). Composition of inhaled
urban air particles determines acute pulmonary responses. Ann Occup Hyg, 46: 419-424. 156478
Gong KW; Zhao W; Li N; Barajas B; Kleinman M; Sioutas C; Horvath S; Lusis AJ; Nel A; Araujo JA. (2007). Air pollutant
chemicals and oxidized lipids exhibit genome wide synergistic effects on endothelial cells. Genome Biol, 8: R149.
091155
Goto Y; Hogg JC; Shih CH; Ishii H; Vincent R; van Eeden SF. (2004). Exposure to ambient particles accelerates monocyte
release from bone marrow in atherosclerotic rabbits. Am J Physiol, 287: L79-L85. 088100
Gottipolu RR; Wallenborn JG; Karoly ED; Schladweiler MC; Ledbetter AD; Krantz T; Linak WP; Nyska A; Johnson JA;
Thomas R; Richards JE; Jaskot RH; Kodavanti UP. (2009). One-month diesel exhaust inhalation produces
hypertensive gene expression pattern in healthy rats. Environ Health Perspect, 117: 38-46. 190360
Gowdy K; Krantz QT; Daniels M; Linak WP; Jaspers I; Gilmour MI. (2008). Modulation of pulmonary inflammatory
responses and antimicrobial defenses in mice exposed to diesel exhaust. Toxicol Appl Pharmacol, 229: 310-319.
097226
Graff DW; Schmitt MT; Dailey LA; Duvall RM; Karoly ED; Devlin RB. (2007). Assessing the role of particulate matter
size and composition on gene expression in pulmonary cells. Inhal Toxicol, 19 Suppl 1: 23-28. 156488
Greenwell LL; Moreno T; Richards RJ. (2003). Pulmonary antioxidants exert differential protective effects against urban
and industrial particulate matter. J Biosci, 28: 101-107. 097478
Gu Z-W; Keane MJ; Ong T; Wallac WE. (2005). Diesel exhaust particulate matter dispersed in a phospholipid surfactant
induces chromosomal aberrations and micronuclei but not 6-thioguanine-resistant gene mutation in V79 cells. J
Toxicol Environ Health A, 68: 431-444. 195923
Gualtieri M; Rigamonti L; Galeotti V; Camatini M. (2005). Toxicity of tire debris extracts on human lung cell line A549.
Toxicol In Vitro, 19: 1001-1008. 097841
Gunnison A; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice VI Gene
expression in heart and lung tissue. Inhal Toxicol, 17: 225-233. 087956
December 2009 D-179
-------
Gurgueira SA; Lawrence J; Coull B; Murthy GGK; Gonzalez-Flecha B. (2002). Rapid increases in the steady-state
concentration of reactive oxygen species in the lungs and heart after particulate air pollution inhalation. Environ
Health Perspect, 110: 749-755. 036535
Gutierrez-Castillo ME; Roubicek DA; Cebrian-Garcia ME; De Vizcaya-Ruiz A; Sordo-Cedeno M; Ostrosky-Wegman P.
(2006). Effect of chemical composition on the induction of DNA damage by urban airborne particulate matter.
Environ Mol Mutagen, 47: 199-211. 089030
Hamada K; Suzaki Y; Leme A; Ito T; Miyamoto K; Kobzik L; Kimura H. (2007). Exposure of pregnant mice to an air
pollutant aerosol increases asthma susceptibility in offspring. J Toxicol Environ Health A, 70: 688-695. 091235
Hamoir J; Nemmar A; Halloy D; Wirth D; Vincke G; Vanderplasschen A; Nemery B; Gustin P. (2003). Effect of
polystyrene particles on lung microvascular permeability in isolated perfused rabbit lungs: role of size and surface
properties. Toxicol Appl Pharmacol, 190: 278-285. 096664
Hansen C; Neller A; Williams G; Simpson R. (2007). Low levels of ambient air pollution during pregnancy and fetal
growth among term neonates in Brisbane, Australia. EnvironRes, 103: 383-389. 090703
Hao M; Cornier S; Wang M; Lee James J; Nel A. (2003). Diesel exhaust particles exert acute effects on airway
inflammation and function in murine allergen provocation models. J Allergy Clin Immunol, 112: 905-914. 096565
Happo MS; Salonen RO; Halinen AI; Jalava PI; Pennanen AS; Kosma VM; Sillanpaa M; Hillamo R; Brunekreef B;
Katsouyanni K; Sunyer J; Hirvonen MR. (2007). Dose and time dependency of inflammatory responses in the
mouse lung to urban air coarse, fine, and ultrafine particles from six European cities. Inhal Toxicol, 19: 227-246.
096630
Harder V; Gilmour P; Lentner B; Karg E; Takenaka S; Ziesenis A; Stampfl A; Kodavanti U; Heyder J; Schulz H. (2005).
Cardiovascular responses in unrestrained WKY rats to inhaled ultrafine carbon particles. Inhal Toxicol, 17: 29-42.
087371
Harkema JR; Keeler G; Wagner J; Morishita M; Timm E; Hotchkiss J; Marsik F; Dvonch T; Kaminski N; Barr E. (2004).
Effects of concentrated ambient particles on normal and hypersecretory airways in rats. Health Effects Institute.
Boston, MA. 056842
Harrod KS; Jaramillo RJ; Berger JA; Gigliotti AP; Seilkop SK; Reed MD. (2005). Inhaled diesel engine emissions reduce
bacterial clearance and exacerbate lung disease to Pseudomonas aeruginosa infection in vivo. Toxicol Sci, 83: 155-
165. 088144
Harrod KS; Jaramillo RJ; Rosenberger CL; Wang S-Z; Berger JA; McDonald JD; Reed MD. (2003). Increased
susceptibility to RSV infection by exposure to inhaled diesel engine emissions. Am J Respir Cell Mol Biol, 28: 451-
463. 097046
Heidenfelder BL; Reif DM; Harkema JR; Cohen Hubal EA; Hudgens EE; Bramble LA; Wagner JG; Morishita M; Keeler
GJ; Edwards SW; Gallagher JE. (2009). Comparative microarray analysis and pulmonary changes in brown
Norway rats exposed to ovalbumin and concentrated air particulates. Toxicol Sci, 108: 207-221. 190026
Hetland RB; Cassee FR; Lag M; Refsnes M; Dybing E; Schwarze PE. (2005). Cytokine release from alveolar macrophages
exposed to ambient particulate matter: heterogeneity in relation to size, city and season. Part Fibre Toxicol, 2: 1-15.
087887
Hetland RB; Cassee FR; Refsnes M; Schwarze PE; Lag M; Boere AJF; Dybing E. (2004). Release of inflammatory
cytokines, cell toxicity and apoptosis in epithelial lung cells after exposure to ambient air particles of different size
fractions. Toxicol In Vitro, 18: 203-212. 097535
Hiramatsu K; Azuma A; Kudoh S; Desaki M; Takizawa H; Sugawara I. (2003). Inhalation of diesel exhaust for three
months affects major cytokine expression and induces bronchus-associated lymphoid tissue formation in murine
lungs. Exp Lung Res, 29: 607-622. 155846
Hiramatsu K; Saito Y; Sakakibara K; Azuma A; Kudoh S; Takizawa H; Sugawara I. (2005). The effects of inhalation of
diesel exhaust on murine mycobacterial infection. Exp Lung Res, 31: 405-415. 088285
Hirano S; Furuyama A; Koike E; Kobayashi T. (2003). Oxidative-stress potency of organic extracts of diesel exhaust and
urban fine particles in rat heart microvessel endothelial cells. Toxicology, 187: 161-170. 097345
Holder AL; Lucas D; Goth-Goldstein R; Koshland CP (2008). Cellular Response to Diesel Exhaust Particles Strongly
Depends on the Exposure Method. Toxicol Sci, 103: 108-115. 093322
December 2009 D-180
-------
Hollingsworth JW; Cook DN; Brass DM; Walker JKL; Morgan DL; Foster WM; Schwartz DA. (2004). The role of Toll-
like receptor 4 in environmental airway injury in mice. Am J Respir Crit Care Med, 170: 126-132. 097816
Hougaard KS; Jensen KA; Nordly P; Taxvig C; Vogel U; Saber AT; Wallin H. (2008). Effects of prenatal exposure to diesel
exhaust particles on postnatal development, behavior, genotoxicity and inflammation in mice. Part Fibre Toxicol, 5:
3. 156570
Huang JY; Liao JW; Liu YC; Lu SY; Chou CP; Chan WH; Chen SU; Ueng TH. (2008). Motorcycle exhaust induces
reproductive toxicity and testicular interleukin-6 in male rats. Toxicol Sci, 103: 137-148. 156574
Huang SL; Hsu MK; Chan CC. (2003). Effects of submicrometer particle compositions on cytokine production and lipid
peroxidation of human bronchial epithelial cells. Environ Health Perspect, 111: 478-482. 087376
Hutchison GR; Brown DM; Hibbs LR; Heal MR; Donaldson K; Maynard RL; Monaghan M; Nicholl A; Stone V. (2005).
The effect of refurbishing a UK steel plant on PM10 metal composition and ability to induce inflammation. Respir
Res, 6: 43. 097750
Hwang B-F; Lee Y-L; Lin Y-C; Jaakkola JJK; Guo YL. (2005). Traffic related air pollution as a determinant of asthma
among Taiwanese school children. Thorax, 60: 467-473. 089454
Iba MM; Fung J; Chung L; Zhao J; Winnik B; Buckley BT; Chen LC; Zelikoff JT; Kou YR. (2006). Differential
inducibility of rat pulmonary CYP1 Al by cigarette smoke and wood smoke. Mutat Res Fund Mol Mech Mutagen,
606: 1-11. 156582
Ichinose T; Takano H; Sadakane K; Yanagisawa R; Kawazato H; Sagai M; Shibamoto T. (2003). Differences in airway-
inflammation development by house dust mite and diesel exhaust inhalation among mouse strains. Toxicol Appl
Pharmacol, 187: 29-37. 041525
Ichinose T; Takano H; Sadakane K; Yanagisawa R; Yoshikawa T; Sagai M; Shibamoto T. (2004). Mouse Strain Differences
in Eosinophilic Airway Inflammation Caused by Intratracheal Instillation of Mite Allergen and Diesel Exhaust
Particles. J Appl Toxicol, 24: 69-76. 180367
Imrich A; Ning Y; Lawrence J; Coull B; Gitin E; Knutson M; Kobzik L. (2007). Alveolar macrophage cytokine response to
air pollution particles: oxidant mechanisms. Toxicol Appl Pharmacol, 218: 256-264. 155859
Inoue K-I; Takano H; Yanagisawa R; Hirano S; Ichinose T; Shimada A; Yoshikawa T. (2006). The role of toll-like receptor
4 in airway inflammation induced by diesel exhaust particles. Arch Toxicol, 80: 275-279. 097815
Inoue K-i; Takano H; Yanagisawa R; Hirano S; Kobayashi T; Ichinose T; Yoshikawa T. (2007). Effects of organic
chemicals derived from ambient particulate matter on lung inflammation related to lipopolysaccharide. Arch
Toxicol, 80: 833-838. 096724
Inoue K-I; Takano H; Yanagisawa R; Ichinose T; Sakurai M; Yoshikawa T. (2006). Effects of nano particles on cytokine
expression in murine lung in the absence or presence of allergen. Arch Toxicol, 80: 614-619. 096720
Inoue K-i; Takano H; Yanagisawa R; Ichinose T; Shimada A; Yoshikawa T. (2005). Pulmonary exposure to diesel exhaust
particles induces airway inflammation and cytokine expression in NC/Nga mice. Arch Toxicol, 79: 595-599.
097481
Inoue K-I; Takano H; Yanagisawa R; Sakurai M; Abe S; Yoshino S; Yamaki K; Yoshikawa T. (2007). Effects of
components derived from diesel exhaust particles on lung physiology related to antigen. Immunopharmacol
Immunotoxicol, 29: 403-412. 096692
Inoue K-i; Takano H; Yanagisawa R; Sakurai M; Ueki N; Yoshikawa T. (2007). Effects of diesel exhaust particles on
cytokine production by splenocytes stimulated with lipopolysaccharide. J Appl Toxicol, 27: 95-100. 096702
Inoue K; Takano H; Yanagisawa R; Hirano S; Sakurai M; Shimada A; Yoshikawa T. (2006). Effects of airway exposure to
nanoparticles on lung inflammation induced by bacterial endotoxin in mice. Environ Health Perspect, 114:1325-
1330.090951
Inoue K; Takano H; Yanagisawa R; Ichinose T; Sadakane K; Yoshino S; Yamaki K; Uchiyama K; Yoshikawa T. (2004).
Components of diesel exhaust particles differentially affect lung expression of cyclooxygenase-2 related to bacterial
endotoxin. J Appl Toxicol, 24: 415-418. 087984
Inoue K; Takano H; Yanagisawa R; Sakurai M; Abe S; Yoshino S; Yamaki K; Yoshikawa T. (2007). Effects of nanoparticles
on lung physiology in the presence or absence of antigen. Int J Immunopathol Pharmacol, 20: 737-744. 198885
December 2009 D-181
-------
Inoue K; Takano H; Yanagisawa R; Sakurai M; Ichinose T; Sadakane K; Yoshikawa T. (2005). Effects of nano particles on
antigen-related airway inflammation in mice. Respir Res, 6: 106. 088625
Inoue K; Takano H; Yanagisawa R; Sakurai M; Ueki N; Yoshikawa T. (2006). Effects of diesel exhaust on lung
inflammation related to bacterial endotoxin in mice. Basic Clin Pharmacol Toxicol, 99: 346-352. 190142
Ishihara Y; Kagawa J. (2003). Chronic diesel exhaust exposures of rats demonstrate concentration and time-dependent
effects on pulmonary inflammation. Inhal Toxicol, 15: 473-492. 096404
Ishii H; Fujii T; Hogg JC; Hayashi S; Mukae H; Vincent R; van Eeden SF. (2004). Contribution ofIL-1 beta and TNF-
alpha to the initiation of the peripheral lung response to atmospheric particulates (PM10). Am J Physiol, 287: L176-
L183. 088103
Ishii H; Hayashi S; Hogg JC; Fujii T; Goto Y; Sakamoto N; Mukae H; Vincent R; van Eeden SF. (2005). Alveolar
macrophage-epithelial cell interaction following exposure to atmospheric particles induces the release of mediators
involved in monocyte mobilization and recruitment. Respir Res, 6: 87. 096138
Ito K; Christensen WF; Eatough DJ; Henry RC; Kim E; Laden F; Lall R; Larson TV; Neas L; Hopke PK; Thurston GD.
(2006). PM source apportionment and health effects: 2 An investigation of intermethod variability in associations
between source-apportioned fine particle mass and daily mortality in Washington, DC. J Expo Sci Environ
Epidemiol, 16: 300-310. 088391
Ito T; Suzuki T; Tamura K; Nezu T; Honda K; Kobayashi T. (2008). Examination of mRNA expression in rat hearts and
lungs for analysis of effects of exposure to concentrated ambient particles on cardiovascular function. Toxicol Sci,
243:271-283.096823
Izawa H; Watanabe G; Taya K; Sagai M. (2007). Inhibitory effects of foods and polyphenols on activation of aryl
hydrocarbon receptor induced by diesel exhaust particles. Environ Sci J Integr Environ Res, 14: 149-156. 190387
JacobsenNR; Mrller P; Cohn CA; Loft S; Vogel U; Wallin H. (2008). Diesel exhaust particles are mutagenic in FE1-
Muta™ Mouse lung epithelial cells. Mutat Res Fund Mol Mech Mutagen, 641: 54-57. 156597
Jalava P; Salonen RO; Halinen AI; Sillanpaa M; Sandell E; Hirvonen MR. (2005). Effects of sample preparation on
chemistry, cytotoxicity, and inflammatory responses induced by air particulate matter. Inhal Toxicol, 17: 107-117.
088648
Jalava PI; Salonen RO; Halinen AI; Penttinen P; Pennanen AS; Sillanpaa M; Sandell E; Hillamo R; Hirvonen M-R. (2006).
In vitro inflammatory and cytotoxic effects of size-segregated particulate samples collected during long-range
transport of wildfire smoke to Helsinki. Toxicol Appl Pharmacol, 215: 341-353. 155872
Jalava PI; Salonen RO; Pennanen AS; Sillanpaa M; Halinen AI; Happo MS; Hillamo R; Brunekreef B; Katsouyanni K;
Sunyer J; Hirvonen M-R. (2007). Heterogeneities in inflammatory and cytotoxic responses of RAW 264.7
macrophage cell line to urban air coarse, fine, and ultrafine particles from six European sampling campaigns. Inhal
Toxicol, 19: 213-225. 096950
Jang A-S; Choi Inseon S; Takizawa H; Rhim T; Lee J-H; Park S-W; Park C-S. (2005). Additive effect of diesel exhaust
particulates and ozone on airway hyperresponsiveness and inflammation in a mouse model of asthma. J Korean
Med Sci, 20: 759-763. 155313
Jaspers I; Ciencewicki JM; Zhang W; Brighton LE; Carson JL; Beck MA; Madden MC. (2005). Diesel exhaust enhances
influenza virus infections in respiratory epithelial cells. Toxicol Sci, 85: 990-1002. 088115
Jimenez LA; Drost EM; Gilmour PS; Rahman I; Antonicelli F; Ritchie H; MacNee W; Donaldson K. (2002). PM10-
exposed macrophages stimulate a proinflammatory response in lung epithelial cells via TNF-alpha. Am J Physiol
Lung Cell Mol Physiol, 282: L237-L248. 156610
Jones HA; Hamacher K; Clark JC; Schofield JB; Krausz T; Haslett C; Boobis AR. (2005). Positron emission tomography
in the quantification of cellular and biochemical responses to intrapulmonary particulates. Toxicol Appl Pharmacol,
207: 230-236. 198883
Jung H; Guo B; Anastasio C; Kennedy IM. (2006). Quantitative measurements of the generation of hydroxyl radicals by
soot particles in a surrogate lung fluid. Atmos Environ, 40: 1043-1052. 132421
Kaan PM; Hegele RG (2003). Interaction between respiratory syncytial virus and particulate matter in guinea pig alveolar
macrophages. Am J Respir Cell Mol Biol, 28: 697-704. 095753
December 2009 D-182
-------
Kafoury RM; Madden MC. (2005). Diesel exhaust particles induce the over expression of tumor necrosis factor-alpha
(TNF-alpha) gene in alveolar macrophages and failed to induce apoptosis through activation of nuclear factor-
kappaB (NF-kappaB). Int J Environ Health Res, 2: 107-113. 156617
Karlsson HL; Ljungman AG; Lindbom J; Moller L. (2006). Comparison of genotoxic and inflammatory effects of particles
generated by wood combustion, a road simulator and collected from street and subway. Toxicol Lett, 165: 203-211.
156625
Karlsson HL; Nilsson L; Moller L. (2005). Subway particles are more genotoxic than street particles and induce oxidative
stress in cultured human lung cells. Chem Res Toxicol, 18: 19-23. 086392
Karlsson HL; Nygren J; Moller L. (2004). Genotoxicity of airborne particulate matter: the role of cell-particle interaction
and of substances with adduct-forming and oxidizing capacity. Mutat Res, 565: 1-10. 198976
Kato A; Kagawa J. (2003). Morphological effects in rat lungs exposed to urban roadside air. Inhal Toxicol, 15: 799-818.
089563
Kato T; Yashiro T; Murata Y; Herbert DC; Oshikawa K; Bando M; Ohno S; Sugiyama Y (2003). Evidence that exogenous
substances can be phagocytized by alveolar epithelial cells and transported into blood capillaries. Cell Tissue Res,
311:47-51. 198882
Katterman ME; Birchard S; Seraphin S; Riley Mark R. (2007). Cellular evaluation of the toxicity of combustion derived
particulate matter: influence of particle grinding and washing on cellular response. Chemosphere, 66: 567-573.
096358
Kendall M; Guntern J; Lockyer NP; Jones FH; Hutton BM; Lippmann M; Tetley TD. (2004). Urban PM2.5 surface
chemistry and interactions with bronchoalveolar lavage fluid. Inhal Toxicol, 16 Suppl 1: 115-129. 156634
Khandoga A; Stampfl A; Takenaka S; Schulz H; Radykewicz R; Kreyling W; Krombach F. (2004). Ultrafine particles exert
prothrombotic but not inflammatory effects on the hepatic microcirculation in healthy mice in vivo. Circulation,
109: 1320-1325.087928
Kim YM; Reed W; Wu W; Bromberg PA; Graves LM; Samet JM. (2005). Zn2+-induced IL-8 expression involves AP-1,
JNK, and ERK activities in human airway epithelial cells. Am J Physiol, 290: L1028-L1035. 088454
Klein-Patel ME; Diamond G; Boniotto M; Saad S; Ryan LK. (2006). Inhibition of beta-defensin gene expression in airway
epithelial cells by low doses of residual oil fly ash is mediated by vanadium. Toxicol Sci, 92: 115-125. 097092
Kleinman M; Phalen R. (2006). Toxicological interactions in the respiratory system after inhalation of ozone and sulfuric
acid aerosol mixtures. Inhal Toxicol, 18: 295-303. 088596
Kleinman M; Sioutas C; Stram D; Froines J; Cho A; Chakrabarti B; Hamade A; Meacher D; Oldham M. (2005). Inhalation
of concentrated ambient particulate matter near a heavily trafficked road stimulates antigen-induced airway
responses in mice. J Air Waste Manag Assoc, 55: 1277-1288. 087880
Kleinman MT Sioutas C Chang MC Boere AJ Cassee FR. (2003). Ambient fine and coarse particle suppression of alveolar
macrophage functions. Toxicol Lett, 137: 151-158. 087938
Kleinman MT; Araujo JA; Nel A; Sioutas C; Campbell A; Cong PQ; Li H; Bondy SC. (2008). Inhaled ultrafine particulate
matter affects CNS inflammatory processes and may act via MAP kinase signaling pathways. Toxicol Lett, 178:
127-130. 190074
Kleinman MT; Hyde DM; Bufalino C; Basbaum C; Bhalla DK; Mautz WJ. (2003). Toxicity of chemical components of
fine particles inhaled by aged rats: effects of concentration. J Air Waste Manag Assoc, 53: 1080-1087. 053535
Kleinman MT; Sioutas C; Froines JR; Fanning E; Hamade A; Mendez L; Meacher D; Oldham M. (2007). Inhalation of
concentrated ambient particulate matter near a heavily trafficked road stimulates antigen-induced airway responses
in mice. Inhal Toxicol, 19 Suppl 1: 117-126. 097082
Knuckles T; Lund A; Lucas S; Campen M. (2008). Diesel exhaust exposure enhances venoconstriction via uncoupling of
eNOS. Toxicol Appl Pharmacol, 230: 346-351. 191987
Knuckles TL; Dreher KL. (2007). Fine oil combustion particle bioavailable constituents induce molecular profiles of
oxidative stress, altered function, and cellular injury in cardiomyocytes. J Toxicol Environ Health A, 70: 1824-
1837. 156652
December 2009 D-183
-------
Kocbach A; Namork E; Schwarze P. (2008). Pro-inflammatory potential of wood smoke and traffic-derived particles in a
monocytic cell line. Toxicology, 247: 123-132. 198874
Kodavanti UP; Schladweiler MC; Gilmour PS; Wallenborn JG; Mandavilli BS; Ledbetter AD; Christian! DC; Runge MS;
Karoly ED; Costa DL; Peddada S; Jaskot R; Richards JH; Thomas R; Madamanchi NR; Nyska A. (2008). The role
of participate matter-associated zinc in cardiac injury in rats. Environ Health Perspect, 116: 13-20. 155907
Kodavanti UP; Schladweiler MC; Ledbetter AD; McGee JK; Walsh L; Gilmour PS; Highfill JW; Davies D; Pinkerton KE;
Richards JH; Crissman K; Andrews D; Costa DL. (2005). Consistent pulmonary and systemic responses from
inhalation of fine concentrated ambient particles: roles of rat strains used and physicochemical properties. Environ
Health Perspect, 113: 1561-1568.087946
Koike E; Kobayashi T. (2005). Organic extract of diesel exhaust particles stimulates expression of la and costimulatory
molecules associated with antigen presentation in rat peripheral blood monocytes but not in alveolar macrophages.
Toxicol Appl Pharmacol, 209: 277-285. 088303
Kooter IM; Boere AJ; Fokkens PH; Leseman DL; Dormans JA; Cassee FR. (2006). Response of spontaneously
hypertensive rats to inhalation of fine and ultrafine particles from traffic: experimental controlled study. Part Fibre
Toxicol, 15:3-7. 09754Z
Kristovich R; Knight DA; Long JF; Williams MV; Dutta PK; Waldman WJ. (2004). Macrophage-mediated endothelial
inflammatory responses to airborne particulates: impact of particulate physicochemical properties. Chem Res
Toxicol, 17: 1303-1312. 087963
Kubatova A; Dronen LC; Picklo MJ Sr; Hawthorne SB. (2006). Midpolarity and nonpolar wood smoke particulate matter
fractions deplete glutathione in RAW 264.7 macrophages. Chem Res Toxicol, 19: 255-261. 198835
Kubatova A; Steckler TS; Gallagher JR; Hawthorne SB; Picklo MJSr. (2004). Toxicity of wide-range polarity fractions
from wood smoke and diesel exhaust particulate obtained using hot pressurized water. Environ Toxicol Chem, 23:
2243-2250. 087986
Kumar VS; Mani U; Prasad AK; Lai K; Gowri V; Gupta A. (2004). Effect of fly ash inhalation on biochemical and
histomorphological changes in rat lungs. Indian J Exp Biol, 42: 964-968. 096655
Kyoso M; Narisawa M; Ito E; Ishijima M; Yana K; Kato A; Ito T; Ishihara Y. (2005). Influence of Exposure to Diesel
Emissions in Rats and Distribution Profile for R-R Interval. Proc Inst Mech Eng H, 5: 5544-5547. 186998
Landvik NE; Gorria M; Arlt VM; Asare N; Solhaug A; Lagadic-Gossmann D; Holme JA. (2007). Effects of nitrated-
polycyclic aromatic hydrocarbons and diesel exhaust particle extracts on cell signalling related to apoptosis:
possible implications for their mutagenic and carcinogenic effects. Toxicology, 231: 159-174. 096722
Last JA; Ward R; Temple L; Pinkerton KE; Kenyon NJ. (2004). Ovalbumin-induced airway inflammation and fibrosis in
mice also exposed to ultrafine particles. Inhal Toxicol, 16: 93-102. 097334
Lee CC; Cheng YW; Kang JJ. (2005). Motorcycle exhaust particles induce IL-8 production through NF-kappaB activation
in human airway epithelial cells. J Toxicol Environ Health A, 68: 1537-1555. 156682
Lee CC; Kang JJ. (2002). Extract of motorcycle exhaust particles induced macrophages apoptosis by calcium-dependent
manner. Chem Res Toxicol, 15: 1534-1542. 198864
Lei Y-C; Chan C-C; Wang P-Y; Lee C-T; Cheng T-J. (2004). Effects of Asian dust event particles on inflammation markers
in peripheral blood and bronchoalveolar lavage in pulmonary hypertensive rats. Environ Res, 95: 71-76. 087884
Lei YC; Chen MC; Chan CC; Wang PY; Lee CT; Cheng TJ. (2004). Effects of concentrated ambient particles on airway
responsiveness and pulmonary inflammation in pulmonary hypertensive rats. Inhal Toxicol, 16: 785-792. 087999
Lei YC; Hwang JS; Chan CC; Lee CT; Cheng TJ. (2005). Enhanced oxidative stress and endothelial dysfunction in
streptozotocin-diabetic rats exposed to fine particles. Environ Res, 99: 335-343. 088660
Lemos M; Mohallen S; Macchione M; Dolhnikoff M; Assunaao J; Godleski J; Saldiva P. (2006). Chronic exposure to
urban air Pollution induces structural alterations in murine pulmonary and coronary arteries. Inhal Toxicol, 18: 247-
253. 088594
Li J; Li Q; Xu J; Li J; Cai X; Liu R; Li Y; Ma J; Li W. (2007). Comparative study on the acute pulmonary toxicity induced
by 3 and 20 nm TiO2 primary particles in mice. Environ Toxicol Pharmacol, 24: 239-244. 093156
December 2009 D-184
-------
Li N; Kim S; Wang M; Froines JR; Sioutas. (2002). Use of a stratified oxidative stress model to study the biological effects
of ambient concentrated and diesel exhaust particulate matter. Inhal Toxicol, 14: 459-486. 042080
Li N; Wang M; Bramble LA; Schmitz DA; Schauer JJ; Sioutas C; Harkema JR; Nel AE. (2009). The adjuvant effect of
ambient particulate matter is closely reflected by the particulate oxidant potential. Environ Health Perspect, 117:
1116-1123. 190457
Li N; Wang M; Oberley TD; Sempf JM; Ne. (2002). Comparison of the pro-oxidative and proinflammatory effects of
organic diesel exhaust particle chemicals in bronchial epithelial cells and macrophages. JImmunol, 169: 4531-
4541.087451
Li Y-J; Kawada T; Matsumoto A; Azuma A; Kudoh S; Takizawa H; Sugawara I. (2007). Airway inflammatory responses to
oxidative stress induced by low-dose diesel exhaust particle exposure differ between mouse strains. Exp Lung Res,
33: 227-244. 155929
Li Z; Carter JD; Dailey LA; Huang YC. (2005). Pollutant particles produce vasoconstriction and enhance MAPK signaling
via angiotensin type I receptor. Environ Health Perspect, 11: 1009-1014. 088647
Li Z; Hyseni X; Carter JD; Soukup JM; Dailey LA; Huang Y-CT. (2006). Pollutant particles enhanced H2O2 production
from NAD(P)H oxidase and mitochondria in human pulmonary artery endothelial cells. Am J Physiol Lung Cell
Mol Physiol, 291: C357-C365. 156693
Lichtenfels AJFC; Gomes JB; Fieri PC; Miraglia SGEK; Hallak J; Saldiva PHN. (2007). Increased levels of air pollution
and a decrease in the human and mouse male-to-female ration in Sao Paulo, Brazil. Fertil Steril, 87: 230-232.
097041
Lijinsky W; Reuber MD. (1987). Chronic carcinogenesis studies of acrolein and related compounds. Toxicol Ind Health, 3:
337-345. 007583
Lindbom J; Gustafsson M; Blomqvist G; Dahl A; Gudmundsson A; Swietlicki E; Ljungman AG (2007). Wear particles
generated from studded tires and pavement induces inflammatory reactions in mouse macrophage cells. Chem Res
Toxicol, 20: 937-946. 155934
Lippmann M; Hwang J; Maclejczyk P; Chen L. (2005). PM source apportionment for short-term cardiac function changes
inApoE-/- mice. Environ Health Perspect, 113: 1575-1579. 087453
Lippmann M; Ito K; Hwang JS; Maciejczyk P; Chen LC. (2006). Cardiovascular effects of nickel in ambient air. Environ
Health Perspect, 114: 1662-1669. 091165
Liu J; Ballaney M; Al-Alem U; Quan C; Jin X; Perera F; Chen LC; Miller RL. (2008). Combined Inhaled Diesel Exhaust
Particles and Allergen Exposure Alter Methylation of T Helper Genes and IgE Production In Vivo. Toxicol Sci, 102:
76-81. 156709
Liu L-JS; Curjuric I; Keidel D; Heldstab J; Kunzli N; Bayer-Oglesby L; Ackermann-Liebrich U; Schindler C; SAPALDIA
team. (2007). Characterization of source-specific air pollution exposure for a large population-based Swiss cohort
(SAPALDIA). Environ Health Perspect, 115: 1638-1645. 093093
Liu P-L; Chen Y-L; Chen Y-H; Lin S-J; Kou Y-R. (2005). Wood smoke extract induces oxidative stress-mediated caspase-
independent apoptosis in human lung endothelial cells: role of AIF and EndoG Am J Physiol, 289: L739-L749.
088304
Liu X; Meng Z. (2005). Effects of airborne fine particulate matter on antioxidant capacity and lipid peroxidation in
multiple organs of rats. Inhal Toxicol, 17: 467-473. 088650
Liu Y-Q; Keane M; Ensell M; Miller W; Kashon M; Ong T-m; Mauderly J; Lawson D; Gautam M; Zielinska B; Whitney
K; Eberhardt J; Wallace W. (2005). In vitro genotoxicity of exhaust emissions of diesel and gasoline engine vehicles
operated on a unified driving cycle. J Environ Monit, 7: 60-66. 097019
Long JF; Waldman WJ; Kristovich R; Williams M; Knight D; Dutta PK. (2005). Comparison of ultrastructural cytotoxic
effects of carbon and carbon/iron particulates on human monocyte-derived macrophages. Environ Health Perspect,
113: 170-174. 087454
Lopes FD; Pinto TS; Arantes-Costa FM; Moriya HT; Biselli PJ; Ferraz LF; Lichtenfels AJ; Saldiva PH; Mauad T; Martins
MA. (2009). Exposure to ambient levels of particles emitted by traffic worsens emphysema in mice. Environ Res,
109:544-551. 190430
December 2009 D-185
-------
Lund AK; Knuckles TL; Obot Akata C; ShohetR; McDonald JD; Gigliotti A; Seagrave JC; CampenMJ. (2007). Gasoline
exhaust emissions induce vascular remodeling pathways involved in atherosclerosis. Toxicol Sci, 95: 485-94.
125741
Lundborg M; Bouhafs R; Gerde P; Ewing P; Camner P; Dahlen SE; Jarstrand C. (2007). Aggregates of ultrafine particles
modulate lipid peroxidation and bacterial killing by alveolar macrophages. Environ Res, 104: 250-257. 096040
Ma C; Wang J; Luo J. (2004). Activation of nuclear factor kappa B by diesel exhaust particles in mouse epidermal cells
through phosphatidylinositol 3-kinase/Akt signaling pathway. Biochem Pharmacol, 67: 1975-1983. 088417
Maciejczyk P; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice: VIII
source-related daily variations in in vitro responses to CAPs. Inhal Toxicol, 17: 243-253. 087456
Madden MC; Dailey LA; Stonehuerner JG; Harris BD. (2003). Responses of cultured human airway epithelial cells treated
with diesel exhaust extracts will vary with the engine load. J Toxicol Environ Health A, 66: 2281-2297. 198877
Mangum JB; Bermudez E; Sar M; Everitt JI. (2004). Osteopontin expression in particle-induced lung disease. Exp Lung
Res, 30: 585-598. 097326
Martin S; Dawidowski L; Mandalunis P; Cereceda-Balic F; Tasat DR. (2007). Characterization and biological effect of
Buenos Aires urban air particles on mice lungs. Environ Res, 105: 340-349. 096366
Mason CE. (2002). Gasoline ETBE vapor condensate rat micronucleus test (satellite procedure). Princeton Research Center
. East Millstone, NJ. 087645
Matsumoto A; Hiramatsu K; Li Y; Azuma A; Kudoh S; Takizawa H; Sugawara I. (2006). Repeated exposure to low-dose
diesel exhaust after allergen challenge exaggerates asthmatic responses in mice. Clin Immunol, 121: 227-235.
098017
Matsumoto S; Ishii H; Tanabe T; Kawai J. (2007). Chemical state analysis of fine particles using XAFS. Tetsu-To-Hagane,
93: 132-137. 187020
Matsuo M; Shimada T; Uenishi R; Sasaki N; Sagai M. (2003). Diesel exhaust particle-induced cell death of cultured
normal human bronchial epithelial cells. Biol Pharm Bull, 26: 438-447. 198879
Matsuzaki T; Amakawa K; Yamaguchi K; Ishizaka A; Terashima T; Matsumaru A; Morishita T. (2006). Effects of diesel
exhaust particles on human neutrophil activation. Exp Lung Res, 32: 427-439. 199517
Mauad T; Rivero DH; de Oliveira RC; Lichtenfels AJ; Guimaraes ET; de Andre PA; Kasahara DI; Bueno HM; Saldiva PH.
(2008). Chronic exposure to ambient levels of urban particles affects mouse lung development. Am J Respir Crit
Care Med, 178: 721-728. 156743
McDonald JD; Harrod KS; Seagrave J; Seikop SK; Mauderly JL. (2004). Effects of low sulfur fuel composition and a
catalyzed particle trap on the composition and toxicity of diesel emissions. Environ Health Perspect, 112: 1307-
1312.087459
McQueen DS; Donaldson K; Bond SM; McNeilly JD; Newman S; Barton NJ; Duffin R. (2007). Bilateral vagotomy or
atropine pre-treatment reduces experimental diesel-soot induced lung inflammation. Toxicol Appl Pharmacol, 219:
62-71.096266
Medeiros N Jr; Rivero DH; Kasahara DI; Saiki M; Godleski JJ; Koutrakis P; Capelozzi VL; Saldiva PH; Antonangelo L.
(2004). Acute pulmonary and hematological effects of two types of particle surrogates are influenced by their
elemental composition. Environ Res, 95: 62-70. 096012
Mehta M; Chen LC; Gordon T; Rom W; Tang MS. (2008). Particulate matter inhibits DNA repair and enhances
mutagenesis. MutatRes Genet Toxicol Environ Mutagen, 657: 116-121. 190440
Meng Z; Zhang Q. (2007). Damage effects of dust storm PM2.5 on DNA in alveolar macrophages and lung cells of rats.
Food Chem Toxicol, 45: 1368-1374. 198963
Mohallem SV; de Araujo Lobo DJ; Pesquero CR; Assuncao JV; de Andre PA; Saldiva PH; Dolhnikoff M. (2005).
Decreased fertility in mice exposed to environmental air pollution in the city of Sao Paulo. Environ Res, 98: 196-
202. 088657
Molinelli AR; Santacana GE; Madden MC; Jimenez BD. (2006). Toxicity and metal content of organic solvent extracts
from airborne particulate matter in Puerto Rico. EnvironRes, 102: 314-325. 198949
December 2009 D-186
-------
Moller W; Hofer T; Ziesenis A; Karg E; Heyder J. (2002). Ultrafine particles cause cytoskeletal dysfunctions in
macrophages. Toxicol Appl Pharmacol, 182: 197-207. 036589
Montiel-Davalos A; Alfaro-Moreno E; Lopez-Marure R. (2007). PM2.5 andPMlO induce the expression of adhesion
molecules and the adhesion of monocytic cells to human umbilical vein endothelial cells. Inhal Toxicol, 19 Suppl 1:
91-98. 156778
Mori T; Watanuki T; Kashiwagura T. (2007). Diesel exhaust particles disturb gene expression in mouse testis. Environ
Toxicol, 22: 58-63. 096564
Morishita M; Keeler G; Wagner J; Marsik F; Timm E; Dvonch J; Harkema J. (2004). Pulmonary retention of particulate
matter is associated with airway inflammation in allergic rats exposed to air pollution in urban Detroit. Inhal
Toxicol, 16: 663-674. 087979
Motta S; Federico C; Saccone S; Librando V; Mosesso P. (2004). Cytogenetic evaluation of extractable agents from
airborne particulate matter generated in the city of Catania (Italy). Mutat Res, 561: 45-52. 198953
Moyer CF; Kodavanti UP; Haseman JK; Costa DL; Nyska A. (2002). Systemic vascular disease in male B6C3F1 mice
exposed to particulate matter by inhalation: studies conducted by the national toxicology program. Toxicol Pathol,
30: 427-434. 052222
Mutlu GM; Green D; Bellmeyer A; Baker CM; Burgess Z; Rajamannan N; Christman JW; Foiles N; Kamp DW; Ghio AJ;
Chandel NS; Dean DA; Sznajder JI; Budinger GR. (2007). Ambient particulate matter accelerates coagulation via
an IL-6-dependent pathway. J Clin Invest, 117: 2952-2961. 121441
Mutlu GM; Snyder C; Bellmeyer A; Wang H; Hawkins K; Soberanes S; Welch LC; Ghio AJ; Chandel NS; Kamp D;
Sznajder Jacob I; Budinger GRS. (2006). Airborne particulate matter inhibits alveolar fluid reabsorption in mice via
oxidant generation. Am J Respir Cell Mol Biol, 34: 670-676. 155994
Nadziejko C; Fang K; Nadziejko E; Narciso SP; Zhong M; Chen LC. (2002). Immediate effects of particulate air pollutants
on heart rate and respiratory rate in hypertensive rats. Cardiovasc Toxicol, 2: 245-252. 087460
Nadziejko C; Fang K; Narciso S; Zhong M; Su WC; Gordon T; Nadas A; Chen LC. (2004). Effect of particulate and
gaseous pollutants on spontaneous arrhythmias in aged rats. Inhal Toxicol, 16: 373-380. 055632
Nam HY; Choi BH; Lee JY; Lee SG; Kim YH; Lee KH; Yoon HK; Song JS; Kim HJ; Lim Y (2004). The role of nitric
oxide in the particulate matter (PM2.5)-induced NFkappaB activation in lung epithelial cells. Toxicol Lett, 148: 95-
102. 198887
Nemmar A; Al-Maskari S; Ali Badreldin H; Al-Amri Issa S. (2007). Cardiovascular and lung inflammatory effects induced
by systemically administered diesel exhaust particles in rats. Am J Physiol Lung Cell Mol Physiol, 292: L664-
L670. 156800
Nemmar A; Hoet PH; Dinsdale D; Vermylen J; Hoylaerts MF; Nemery B. (2003). Diesel exhaust particles in lung acutely
enhance experimental peripheral thrombosis. Circulation, 107: 1202-1208. 096567
Nemmar A; Hoet PH; Vermylen J; Nemery B; Hoylaerts MF. (2004). Pharmacological stabilization of mast cells abrogates
late thrombotic events induced by diesel exhaust particles in hamsters. Circulation, 110: 1670-1677. 087959
Nemmar A; Hoylaerts MF; Hoet PHM; Vermylen J; Nemery B. (2003). Size effect of intratracheally instilled particles on
pulmonary inflammation and vascular thrombosis. Toxicol Appl Pharmacol, 186: 38-45. 087931
Nemmar A; Inuwa IM. (2008). Diesel exhaust particles in blood trigger systemic and pulmonary morphological alterations.
Toxicol Lett, 176: 20-30. 096566
Nemmar A; Nemery B; Hoet PHM; Vermylen J; Hoylaerts MF. (2003). Pulmonary inflammation and thrombogenicity
caused by diesel particles in hamsters: role of histamine. Am J Respir Crit Care Med, 168: 1366-1372. 097487
Niwa Y; Hiura Y; Murayama T; Yokode M; Iwai N. (2007). Nano-sized carbon black exposure exacerbates atherosclerosis
in LDL-receptor knockout mice. Circ J, 71: 1157-1161. 091309
Niwa Y; Hiura Y; Sawamura H; Iwai N. (2008). Inhalation exposure to carbon black induces inflammatory response in rats.
Circ J, 72: 144-149. 156812
Nozaki JI; Yamamoto R; Ma L; Shima M. (2007). Trial to evaluate effects of ambient particulate matter on health: A
preliminary study using two-dimensional gel electrophoresis. Environ Health Prev Med, 12: 138-142. 097862
December 2009 D-187
-------
Nurkiewicz TR; Porter DW; Barger M; Castranova V; Boegehold MA. (2004). Particulate matter exposure impairs
systemic microvascular endothelium-dependent dilation. Environ Health Perspect, 112: 1299-1306. 087968
Nurkiewicz TR; Porter DW; Barger M; Millecchia L; Rao KMK; Marvar PJ; Hubbs AF; Castranova V; Boegehold MA.
(2006). Systemic microvascular dysfunction and inflammation after pulmonary particulate matter exposure.
Environ Health Perspect, 114: 412-419. 088611
Nurkiewicz TR; Porter DW; Hubbs AF; Cumpston JL; Chen BT; Frazer DG; Castranova V. (2008). Nanoparticle inhalation
augments particle-dependent systemic microvascular dysfunction. Part Fibre Toxicol, 5:1. 156816
Nurkiewicz TR; Porter DW; Hubbs AF; Stone S; Chen BT; Frazer DG; Boegehold MA; Castranova V. (2009). Pulmonary
nanoparticle exposure disrupts systemic microvascular nitric oxide signaling. Toxicol Sci, 110: 191-203. 191961
Nygaard UC; Alberg T; Bleumink R; Aase A; Dybing E; Pieters R; Lovik M. (2005). Ambient air particles from four
European cities increase the primary cellular response to allergen in the draining lymph node. Toxicology, 207: 241-
254. 088655
Nygaard UC; Ormstad H; Aase A; Lovik M. (2005). The IgE adjuvant effect of particles: characterisation of the primary
cellular response in the draining lymphnode. Toxicology, 206: 181-193. 087980
Nygaard UC; Samuelsen M; Aase A; Lovik M. (2004). The capacity of particles to increase allergic sensitization is
predicted by particle number and surface area, not by particle mass. Toxicol Sci, 82: 515-524. 058558
Obot CJ; Morandi MT; Beebe TP; Hamilton RF; Holian A. (2002). Surface components of airborne particulate matter
induce macrophage apoptosis through scavenger receptors. Toxicol Appl Pharmacol, 184: 98-106. 042370
Obot CJ; Morandi MT; Hamilton RF; Holian A. (2004). A comparison of murine and human alveolar macrophage
responses to urban particulate matter. Inhal Toxicol, 16: 69-76. 095938
Oh S-M; Chung K-H. (2006). Identification of mammalian cell genotoxins in respirable diesel exhaust particles by
bioassay-directed chemical analysis. Toxicol Lett, 161: 226-235. 088296
Okayama Y; Kuwahara M; Suzuki AK; Tsubone H. (2006). Role of reactive oxygen species on diesel exhaust particle-
induced cytotoxicity in rat cardiac myocytes. J Toxicol Environ Health A, 69: 1699-1710. 156824
Okeson CD; Riley MR; Fernandez A; Wendt JOL. (2003). Impact of the composition of combustion generated fine
particles on epithelial cell toxicity: influences of metals on metabolism. Chemosphere, 51: 1121-1128. 042292
Okeson CD; Riley MR; Riley-Saxton E. (2004). In vitro alveolar cytotoxicity of soluble components of airborne particulate
matter: effects of serum on toxicity of transition metals. Toxicol In Vitro, 18: 673-680. 087961
Ono N; Oshio S; Niwata Y; Yoshida S; Tsukue N; Sugawara I; Takano H; Takeda K. (2007). Prenatal exposure to diesel
exhaust impairs mouse spermatogenesis. Inhal Toxicol, 19: 275-281. 156007
Osornio-Vargas AR; Bonner JC; Alfaro-Moreno E; Martinez L; Garcia-Cuellar C; Resales SP; Miranda J; Rosas I. (2003).
Proinflammatory and cytotoxic effects of Mexico City air pollution particulate matter in vitro are dependent on
particle size and composition. Environ Health Perspect, 111:1289-1293. 052417
Pastorkova A; Cerna M; Smid J; Vrbikova V (2004). Mutagenicity of airborne particulate matter PM10. Cent Eur J Public
Health, 12: S72-S75. 087431
Perm A; Murphy G; Barker S; Henk W; Perm L. (2005). Combustion-derived ultrafine particles transport organic toxicants
to target respiratory cells. Environ Health Perspect, 113: 956-963. 088257
Pereira CEL; Heck TG; Saldiva PHN; Rhoden CR. (2007). Ambient particulate air pollution from vehicles promotes lipid
peroxidation and inflammatory responses in rat lung. Braz J Med Biol Res, 40: 1353-1359. 156019
Pinkerton KE; Zhou Y; Teague SV; Peake JL; Walther RC; Kennedy IM; Leppert VJ; Aust AE. (2004). Reduced lung cell
proliferation following short-term exposure to ultrafine soot and iron particles in neonatal rats: key to impaired lung
growth?. Inhal Toxicol, 1: 73-81. 087465
Pires-Neto RC; Lichtenfels AJ; Scares SR; Macchione M; Saldiva PHN; Dolhnikoff M. (2006). Effects of Sao Paulo air
pollution on the upper airways of mice. Environ Res, 101: 356-361. 096734
Poma A; Limongi T; Pisani C; Granato V; Picozzi P. (2006). Genotoxicity induced by fine urban air particulate matter in
the macrophages cell line RAW 264.7. Toxicol In Vitro, 20: 1023-1029. 096903
December 2009 D-188
-------
Pourazar J; Mudway IS; Samet JM; Helleday R; Blomberg A; Wilson SJ; Frew AJ; Kelly FJ; Sandstrom T. (2005). Diesel
exhaust activates redox-sensitive transcription factors and kinases in human airways. Am J Physiol, 289: L724-
L730. 088305
Pozzi R; De Berardis B; Paoletti L; Guastadisegni C. (2005). Winter urban air particles from Rome (Italy): effects on the
monocytic-macrophagic RAW 2647 cell line. Environ Res, 99: 344-354. 088610
Pradhan A; Waseem M; Dogra S; Khanna AK; Kaw JL. (2005). Alterations in bronchoalveolar lavage constituents,
oxidant/antioxidant status, and lung histology following intratracheal instillation of respirable suspended particulate
matter. J Environ Pathol Toxicol Oncol, 24: 19-32. 096128
Proctor SD; Dreher KL; Kelly SE; Russell JC. (2006). Hypersensitivity of prediabetic JCR: LA-cp rats to fine airborne
combustion particle-induced direct and noradrenergic-mediated vascular contraction. Toxicol Sci, 90: 385-391.
088480
Prophete C; Maciejczyk P; Salnikow K; Gould T; Larson T; Koenig J; Jaques P; Sioutas C; Lippmann M; Cohen M.
(2006). Effects of select PM-associated metals on alveolar macrophage phosphorylated ERK1 and -2 and iNOS
expression during ongoing alteration in iron homeostasis. J Toxicol Environ Health A, 69: 935-951. 156888
Radomski A; Jurasz P; Alonso-Escalano D; Drews M; Morandi M; Malinski T; Radomski MW (2005). Nanoparticle-
induced platelet aggregation and vascular thrombosis. Br J Pharmacol, 146: 882-893. 091377
Ramage L; Guy K. (2004). Expression of C-reactive protein and heat-shock protein-70 in the lung epithelial cell line A549,
in response to PM10 exposure. Inhal Toxicol, 16: 447-452. 055640
Ramos C; Cisneros J; Gonzalez-Avila G; Becerril C; Ruiz V; Montano M. (2009). Increase of matrix metalloproteinases in
woodsmoke-induced lung emphysema in guinea pigs. Inhal Toxicol, 21: 119-132. 190116
Rao KM; Ma JY; Meighan T; Barger MW; Pack D; Vallyathan V. (2005). Time course of gene expression of inflammatory
mediators in rat lung after diesel exhaust particle exposure. Environ Health Perspect, 113: 612-617. 095756
Reed MD; Barrett EG; Campen MJ; Divine KK; Gigliotti AP; McDonald JD; Seagrave JC; Mauderly JL; Seilkop SK;
Swenberg JA. (2008). Health effects of subchronic inhalation exposure to gasoline engine exhaust. Inhal Toxicol,
20: 1125-1143. 156903
Reed MD; Campen MJ; Gigliotti AP; Harrod KS; McDonald JD; Seagrave JC; Mauderly JL; Seilkop SK. (2006). Health
effects of subchronic exposure to environmental levels of hardwood smoke. Inhal Toxicol, 18: 523-539. 156043
Reed MD; Gigliotti AP; McDonald JD; Seagrave JC; Seilkop SK; Mauderly JL. (2004). Health effects of subchronic
exposure to environmental levels of diesel exhaust. Inhal Toxicol, 16: 177-193. 055625
Reibman J; Hsu Y; Chen LC; Bleck B; Gordon T. (2003). Airway epithelial cells release MIP-3alpha/CCL20 in response to
cytokines and ambient particulate matter. Am J Respir Cell Mol Biol, 28: 648-654. 156905
Rengasamy A; Barger MW; Kane E; Ma JKH; Castranova V; Ma JYC. (2003). Diesel exhaust particle-induced alterations
of pulmonary phase I and phase II enzymes of rats. J Toxicol Environ Health A, 66: 153-167. 156907
Renwick LC; Brown D; Clouter A; Donaldson K. (2004). Increased inflammation and altered macrophage chemotactic
responses caused by two ultrafine particle types. Occup Environ Med, 61: 442-447. 056067
Rhoden CR; Ghelfi E; Gonzalez-Flecha B. (2008). Pulmonary inflammation by ambient air particles is mediated by
superoxide anion. Inhal Toxicol, 20: 11-15. 190475
Rhoden CR; Lawrence J; Godleski JJ; Gonzalez-Flecha B. (2004). N-acetylcysteine prevents lung inflammation after
short-term inhalation exposure to concentrated ambient particles. Toxicol Sci, 79: 296-303. 087969
Rhoden CR; Wellenius GA; Ghelfi E; Lawrence J; Gonzalez-Flecha B. (2005). PM-induced cardiac oxidative stress and
dysfunction are mediated by autonomic stimulation. Biochim Biophys Acta, 1725: 305-313. 087878
Riley MR; Boesewetter DE; Kim AM; Sirvent FP (2003). Effects of metals Cu, Fe, Ni, V, and Zn on rat lung epithelial
cells. Toxicology, 190: 171-184. 053237
Riley MR; Boesewetter DE; Turner RA; Kim AM; Collier JM; Hamilton A. (2005). Comparison of the sensitivity of three
lung derived cell lines to metals from combustion derived particulate matter. Toxicol In Vitro, 19: 411-419. 096452
Ritz SA; Wan J; Diaz-Sanchez D. (2007). Sulforaphane-stimulated phase II enzyme induction inhibits cytokine production
by airway epithelial cells stimulated with diesel extract. Am J Physiol Lung Cell Mol Physiol, 292: L33-39. 198901
December 2009 D-189
-------
Rivedal E; Myhre O; Sanner T; Eide I. (2003). Supplemental role of the Ames mutation assay and gap junction intercellular
communication in studies of possible carcinogenic compounds from diesel exhaust particles. Arch Toxicol, 77: 533-
542. 097684
Rivero DH; Scares SR; Lorenzi-Filho G; Saiki M; Godleski JJ; Antonangelo L; Dolhnikoff M; Saldiva PH. (2005). Acute
cardiopulmonary alterations induced by fine particulate matter of Sao Paulo, Brazil. Toxicol Sci, 85: 898-905.
088653
Roberts E; Charboneau L; Espina V; Liotta L; Petricoin E; Dreher K. (2004). Application of laser capture microdissection
and protein microarray technologies in the molecular analysis of airway injury following pollution particle
exposure. J Toxicol Environ Health A, 67: 851-861. 198903
Roberts JR; Young S-H; Castranova V; Antonini JM. (2007). Soluble metals in residual oil fly ash alter innate and adaptive
pulmonary immune responses to bacterial infection in rats. Toxicol Appl Pharmacol, 221: 306-319. 097623
Rodriguez Ferreira Rivero DH; Sassaki C; Lorenzi-Filho G; Nascimento Saldiva PH. (2005). PM2.5 induces acute
electrocardiographic alterations in healthy rats. Environ Res, 99: 262-266. 088659
Rosas Perez I; Serrano J; Alfaro-Moreno E; Baumgardner D; Garcia-Cuellar C; Martin Del Campo JM; Raga GB;
Castillejos M; Colin RD; Osornio Vargas AR. (2007). Relations between PM10 composition and cell toxicity: a
multivariate and graphical approach. Chemosphere, 67: 1218-28. 097967
Roubicek DA; Gutierrez-Castillo ME; Sordo M; Cebrian-Garcia ME; Ostrosky-Wegman P. (2007). Micronuclei induced by
airborne particulate matter from Mexico City. Mutat Res Fund Mol Mech Mutagen, 631: 9-15. 156929
Saber AT; Bornholdt J; Dybdahl M; Sharma AK; Loft S; Vogel U; Wallin H. (2005). Tumor necrosis factor is not required
for particle-induced genotoxicity and pulmonary inflammation. Arch Toxicol, 79: 177-182. 097865
Sakamoto N; Hayashi S; Gosselink J; Ishii H; Ishimatsu Y; Mukae H; Hogg JC; van Eeden SF. (2007). Calcium dependent
and independent cytokine synthesis by air pollution particle-exposed human bronchial epithelial cells. Toxicol Appl
Pharmacol, 225: 134-141.096282
Salnikow K; Li X; Lippmann M. (2004). Effect of nickel and iron co-exposure on human lung cells. Toxicol Appl
Pharmacol, 196: 258-265. 087469
Salonen R; Halinen A; Pennanen A; Hirvonen M; Sillanpaa M; Hillamo R; Shi R; Borm P; Sandell E; Koskentalo T;
Aarnio P. (2004). Chemical and in vitro toxicologic characterization of wintertime and springtime urban-air
particles with an aerodynamic diameter below 10 um in Helsinki. Scand J Work Environ Health, 30: 80-90. 187053
Samet JM; Dewar BJ; Wu W; Graves LM. (2003). Mechanisms of Zn2+-induced signal initiation through the epidermal
growth factor receptor. Toxicol Appl Pharmacol, 191: 86-93. 113782
Santini MT; Rainaldi G; Ferrante A; Romano R; Clemente S; Motta A; De Berardis B; Balduzzi M; Paoletti L; Indovina
PL. (2004). Environmental fine particulate matter (PM25) activates the RAW 2647 macrophage cell line even at
very low concentrations as revealed by 1H NMR. Chem Res Toxicol, 17: 63-74. 087879
Saxena QB; Saxena RK; Siegel PD; Lewis DM. (2003). Identification of organic fractions of diesel exhaust particulate
(DEP) which inhibit nitric oxide (NO) production from a murine macrophage cell line. Toxicol Lett, 143: 317-322.
096986
Saxena RK; Saxena QB; Weissman DN; Simpson JP; Bledsoe TA; Lewis DM. (2003). Effect of diesel exhaust particulate
on bacillus Calmette-Guerin lung infection in mice and attendant changes in lung interstitial lymphoid
subpopulations and IFN gamma response. Toxicol Sci, 73: 66-71. 054395
Schins RPF; Lightbody JH; Borm PJA; Shi T; Donaldson K; Stone V. (2004). Inflammatory effects of coarse and fine
particulate matter in relation to chemical and biological constituents. Toxicol Appl Pharmacol, 195: 1-11. 054173
Schneider J; Hock N; Weimer S; Borrmann S; Kirchner U; Vogt R; Scheer V. (2005). Nucleation particles in diesel exhaust:
composition inferred from in situ mas spectrometric analysis. Environ Sci Technol, 39: 6153-6161. 088368
Schober W; Belloni B; Lubitz S; Eberlein-Konig B; Bohn P; Saritas Y; Lintelmann J; Matuschek G; Behrendt H; Buters J.
(2006). Organic extracts of urban aerosol (< or =PM25) enhance rBet v 1 -induced upregulation of CD63 in
basophils from birch pollen-allergic individuals. Toxicol Sci, 90: 377-384. 097321
Seagrave J; Campen M; McDonald J; Mauderly J; Rohr A. (2008). Oxidative stress, inflammation, and pulmonary function
assessment in rats exposed to laboratory-generated pollutant mixtures. J Toxicol Environ Health A, 71: 1352-1362.
191990
December 2009 D-190
-------
Seagrave J; Dunaway S; McDonald JD; Mauderly JL; Hay den P; Stidley C. (2007). Responses of differentiated primary
human lung epithelial cells to exposure to diesel exhaust at an air-liquid interface. Exp Lung Res, 33: 27-51.
097549
Seagrave J; Knall C; McDonald JD; Mauderly JL. (2004). Diesel particulate material binds and concentrates a
proinflammatory cytokine that causes neutrophil migration. Inhal Toxicol, 1: 93-98. 087470
Seagrave J; Mauderly JL; Seilkop SK. (2003). In vitro relative toxicity screening of combined particulate and semivolatile
organic fractions of gasoline and diesel engine emissions. J Toxicol Environ Health A, 66: 1113-1132. 054979
Seagrave J; McDonald JD; Reed MD; Seilkop SK; Mauderly JL. (2005). Responses to subchronic inhalation of low
concentrations of diesel exhaust and hardwood smoke measured in rat bronchoalveolar lavage fluid. Inhal Toxicol,
17: 657-670. 088000
Seagrave JC; McDonald JD; Bedrick E; Edgerton ES; Gigliotti AP; Jansen JJ; Ke L; Naeher LP; Seilkop SK; Zheng M;
Mauderley JL. (2006). Lung toxicity of ambient particulate matter from southeastern US sites with different
contributing sources: relationships between composition and effects. Environ Health Perspect, 114: 1387-93.
091291
Seaton A; Cherrie J; Dennekamp M; Donaldson K; Hurley JF; Tran CL. (2005). The London Underground: Dust and
hazards to health. Occup Environ Med, 62: 355-362. 198904
Sevastyanova O; Binkova B; Topinka J; Sram RJ; Kalina I; Popov T; Novakova Z; Farmer PB. (2007). In vitro
genotoxicity of PAH mixtures and organic extract from urban air particles part II: human cell lines. Mutat Res Fund
Mol Mech Mutagen, 620: 123-134. 156969
Sharma AK; Jensen KA; Rank J; White PA; Lundstedt S; Gagne R; Jacobsen NR; Kristiansen J; Vogel U; Wallin H. (2007).
Genotoxicity, inflammation and physico-chemical properties of fine particle samples from an incineration energy
plant and urban air. Mutat Res Fund Mol Mech Mutagen, 633: 95-111. 156975
Shi T; Knaapen AM; Begerow J; Birmili W; Borm PJ; Schins RP (2003). Temporal variation of hydroxyl radical
generation and 8-hydroxy-2'-deoxyguanosine formation by coarse and fine particulate matter. Occup Environ Med,
60: 315-321.088248
Shwe T-T-W; Yamamoto S; Kakeyama M; Kobayashi T; Fujimaki H. (2005). Effect of intratracheal instillation of ultrafine
carbon black on proinflammatory cytokine and chemokine release and mRNA expression in lung and lymph nodes
of mice. Toxicol Appl Pharmacol, 209: 51-61. 111553
Sigaud S; Goldsmith Carroll-Ann W; ZhouH; YangZ; FedulovA; ImrichA; KobzikL. (2007). Air pollution particles
diminish bacterial clearance in the primed lungs of mice. Toxicol Appl Pharmacol, 223: 1-9. 096100
Silva PJ; Erupe ME; Price D; Elias J; Malloy QG; Li Q; Warren B; Cocker DR 3rd. (2008). Trimethylamine as precursor to
secondary organic aerosol formation via nitrate radical reaction in the atmosphere. Environ Sci Technol, 42: 4689-
4696.156981
Simkhovich BZ; Marjoram P; Kleinman MT; Kloner RA. (2007). Direct and acute cardiotoxicity of ultrafine particles in
young adult and old rat hearts. Basic Res Cardiol, 102: 467-475. 096594
Singal M; Finkelstein JN. (2005). Use of indicator cell lines for determining inflammatory gene changes and screening the
inflammatory potential of particulate and non-particulate stimuli. Inhal Toxicol, 17: 415-425. 198905
Singh P; DeMarini DM; Dick CA; Tabor DG; Ryan JV; Linak WP; Kobayashi T; Gilmour MI. (2004). Sample
characterization of automobile and forklift diesel exhaust particles and comparative pulmonary toxicity in mice.
Environ Health Perspect, 112: 820-825. 087472
Sirivelu MP; MohanKumar SMJ; Wagner JG; Harkema JR; MohanKumar PS. (2006). Activation of the stress axis and
neurochemical alterations in specific brain areas by concentrated ambient particle exposure with concomitant
allergic airway disease. Environ Health Perspect, 114: 870-874. 111151
Skarek M; Janosek J; Cupr P; Kohoutek J; Novotna-Rychetska A; Holoubek I. (2007). Evaluation of genotoxic and non-
genotoxic effects of organic air pollution using in vitro bioassays. Environ Int, 33: 859-66. 096814
Smith KR; Kim S; Recendez JJ; Teague SV; Menache MG; Grubbs DE; Sioutas C; Pinkerton KE. (2003). Airborne
particles of the California Central Valley alter the lungs of healthy adult rats. Environ Health Perspect, 111: 902-
908. 042107
December 2009 D-191
-------
Smith KR; Veranth JM; Kodavanti UP; Aust AE; Pinkerton KE. (2006). Acute pulmonary and systemic effects of inhaled
coal fly ash in rats: comparison to ambient environmental particles. Toxicol Sci, 93: 390-399. 110864
Somers CM; McCarry BE; Malek F; Quinn JS. (2004). Reduction of particulate air pollution lowers the risk of heritable
mutations in mice. Science, 304: 1008-1010. 078098
Somers CM; Yauk CL; White PA; Parfett CLJ; Quinn JS. (2002). Air pollution induces heritable DNA mutations. PNAS,
99: 15904-15907.078100
Song CL; Zhou YC; Huang RJ; Wang YQ; Huang QF; Lu G; Liu KM. (2007). Influence of ethanol-diesel blended fuels on
diesel exhaust emissions and mutagenic and genotoxic activities of particulate extracts. J Hazard Mater, 149: 355-
363. 155306
Song H-M; Jang A-S; Ahn M-H; Takizawa H; Lee S-H; Kwon J-H; Lee Y-M; Rhim T; Park C-S. (2008). Yml and Ym2
expression in a mouse model exposed to diesel exhaust particles. Environ Toxicol, 23: 110-116. 156093
Steerenberg P; Verlaan A; De Klerk A; Boere A; Loveren H; Cassee F. (2004). Sensitivity to ozone, diesel exhaust particles,
and standardized ambient particulate matter in rats with a listeria monocytogenes-induced respiratory infection.
Inhal Toxicol, 16: 311-317. 087474
Steerenberg P; Withagen C; Dalen W; Dormans J; Loveren H. (2004). Adjuvant Activity of Ambient Particulate Matter in
Macrophage Activity-Suppressed, N-Acetylcysteine-Treated, iNOS- and IL-4-Deficient Mice. Inhal Toxicol, 16:
835-843. 087981
Steerenberg PA; Van Amelsvoort L; Lovik M; Hetland RB; Alberg T; Halatek T; Bloemen HJT; Rydzynski K; Swaen G;
Schwarze P; Dybing E; Cassee FR. (2006). Relation between sources of particulate air pollution and biological
effect parameters in samples from four European cities: an exploratory study. Inhal Toxicol, 18: 333-346. 088249
Steerenberg PA; Withagen CE; van Dalen WJ; Dormans JA; Heisterkamp SH; van Loveren H; Cassee FR. (2005). Dose
dependency of adjuvant activity of particulate matter from five European sites in three seasons in an ovalbumin-
mouse model. Inhal Toxicol, 17: 133-145. 088649
Stevens JP; Zahardis J; MacPherson M; Mossman BT; Petrucci GA. (2008). A new method for quantifiable and controlled
dosage of particulate matter for in vitro studies: the Electrostatic Particulate Dosage and Exposure System
(EPDExS). Toxicol In Vitro, 22(7): 1768-1774. 155363
Stinn W; Teredesai A; Anskeit E; Rustemeier K; Schepers G; Schnell P; Haussmann H-J; Carchman RA; Coggins CR;
Reininghaus W. (2005). Chronic nose-only inhalation study in rats, comparing room-aged sidestream cigarette
smoke and diesel engine exhaust. Inhal Toxicol, 17: 549-576. 088307
Sugamata M; Ihara T; Sugamata M; Takeda K. (2006). Maternal Exposure to Diesel Exhaust Leads to Pathological
Similarity to Autism inNewborns. Eisei Kagaku, 52: 486-488. 157025
Sun Q; Wang A; Jin X; Natanzon A; Duquaine D; Brook RD; Aguinaldo JG; Fayad ZA; Fuster V; Lippmann M; Chen LC;
Rajagopalan S. (2005). Long-term air pollution exposure and acceleration of atherosclerosis and vascular
inflammation in an animal model. JAMA, 294: 3003-3010. 087952
Sun Q; Yue P; Deiuliis JA; Lumeng CN; Kampfrath T; Mikolaj MB; Cai Y; Ostrowski MC; Lu B; Parthasarathy S; Brook
RD; Moffatt-Bruce SD; Chen LC; Rajagopalan S. (2009). Ambient air pollution exaggerates adipose inflammation
and insulin resistance in a mouse model of diet-induced obesity. Circulation, 119: 538-546. 190487
Sun Q; Yue P; Kirk RI; Wang A; Moatti D; Jin X; Lu B; Schecter AD; Lippmann M; Gordon T; Chen LC; Rajagopalan S.
(2008). Ambient air particulate matter exposure and tissue factor expression in atherosclerosis. Inhal Toxicol, 20:
127-137. 157033
Sun Q; Yue P; Ying Z; Cardounel AJ; Brook RD; Devlin R; Hwang JS; Zweier JL; Chen LC; Rajagopalan S. (2008). Air
Pollution Exposure Potentiates Hypertension Through Reactive Oxygen Species-Mediated Activation of
Rho/ROCK. Arterioscler Thromb Vase Biol, 28: 1760-1766. 157032
Sureshkumar V; Paul B; Uthirappan M; Pandey R; Sahu AP; Lai K; Prasad AK; Srivastava S; Saxena A; Mathur N; Gupta
YK. (2005). Proinflammatory and anti-inflammatory cytokine balance in gasoline exhaust induced pulmonary
injury in mice. Inhal Toxicol, 17: 161-168. 088306
December 2009 D-192
-------
Takizawa H; Abe S; Okazaki H; Kohyama T; Sugawara I; Saito Y; Ohtoshi T; Kawasaki S; Desaki M; Nakahara K;
Yamamoto K; Matsushima K; Tanaka M; Sagai M; Kudoh S. (2003). Diesel exhaust particles upregulate eotaxin
gene expression in human bronchial epithelial cells via nuclear factor-kappa B-dependent pathway. Am J Physiol
Lung Cell Mol Physiol, 284: L1055-L1062. 157039
Tal TL; Graves LM; Silbajoris R; Bromberg PA; Wu W; Samet JM. (2006). Inhibition of protein tyrosine phosphatase
activity mediates epidermal growth factor receptor signaling in human airway epithelial cells exposed to Zn2+.
Toxicol Appl Pharmacol, 214: 16-23. 108588
Tamagawa E; Bai N; Morimoto K; Gray C; Mui T; Yatera K; Zhang X; Xing L; Li Y; Laher I. (2008). Particulate matter
exposure induces persistent lung inflammation and endothelial dysfunction. Am J Physiol Lung Cell Mol Physiol,
295:L79-L85. 191988
Tamaoki J; Isono K; Takeyama K; Tagaya E; Nakata J; Nagai A. (2004). Ultrafine carbon black particles stimulate
proliferation of human airway epithelium via EGF receptor-mediated signaling pathway. Am J Physiol Lung Cell
Mol Physiol, 287: L1127-L1133. 157040
Tankersley CG; Bierman A; Rabold R. (2007). Variation in heart rate regulation and the effects of particle exposure in
inbred mice. Inhal Toxicol, 19: 621-629. 097910
Tankersley CG; Campen M; Bierman A; Flanders SE; Broman KW; Rabold R. (2004). Particle effects on heart-rate
regulation in senescent mice. Inhal Toxicol, 16: 381-390. 094378
Tankersley CG; Champion HC; Takimoto E; Gabrielson K; Bedja D; Misra V; El-Haddad H; Rabold R; Mitzner W. (2008).
Exposure to inhaled particulate matter impairs cardiac function in senescent mice. Am J Physiol Regul Integr Comp
Physiol, 295: R252-R263. 157043
Tao F; Kobzik L. (2002). Lung macrophage-epithelial cell interactions amplify particle-mediated cytokine release. Am J
Respir Cell Mol Biol, 26: 499-505. 157044
Tesfaigzi Y; McDonald JD; Reed MD; Singh SP; De Sanctis GT; Eynott PR; Hahn FF; Campen MJ; Mauderly JL. (2005).
Low-level subchronic exposure to wood smoke exacerbates inflammatory responses in allergic rats. Toxicol Sci,
88: 505-513. 156116
Tesfaigzi Y; Singh SP; Foster JE; Kubatko J; Barr EB; Fine PM; McDonald JD; Hahn FF; Mauderly JL. (2002). Health
effects of subchronic exposure to low levels of wood smoke in rats. Toxicol Sci, 65: 115-125. 025575
Thomson E; Kumarathasan P; Goegan P; Aubin RA; Vincent R. (2005). Differential regulation of the lung endothelin
system by urban particulate matter and ozone. Toxicol Sci, 88: 103-113. 087554
Thomson E; Kumarathasan P; Vincent R. (2006). Pulmonary expression of preproET-1 and preproET-3 mRNAs is altered
reciprocally in rats after inhalation of air pollutants. Exp Biol Med, 231: 979-984. 097483
Tomita S; Maekawa S-I; Rahman M; Saito F; Kizu R; Tohi K; Ueno T; Nakase H; Gonzalez FJ; Hayakawa K; Korenaga T;
Takahama Y. (2006). Thymic involution produced by diesel exhaust particles and their constituents in mice. Toxicol
Environ Chem, 88: 113-124. 097827
Tong Y; Zhang G; Li Y; Tan M; Wang W; Chen J; Hwu Y; Hsu P-C; Je JH; Margaritondo G; Song W; Jiang R; Jiang Z.
(2006). Synchrotron microradiography study on acute lung injury of mouse caused by PM(25) aerosols. Eur J
Radiol, 58: 266-272. 097699
Totlandsdal AI; Refsnes M; Skomedal T; Osnes JB; Schwarze PE; Lag M. (2008). Particle-induced cytokine responses in
cardiac cell cultures—the effect of particles versus soluble mediators released by particle-exposed lung cells.
Toxicol Sci, 106: 233-241. 157056
Tozuka Y; Watanabe N; Ohsawa M; Toriba A; Kizu R; Hayakawa K. (2004). Transfer of poly cyclic aromatic hydrocarbons
to fetuses and breast milk of rats exposed to diesel exhaust. Eisei Kagaku, 50: 497-502. 090864
Tsukue N; Tsubone H; Suzuki AK. (2002). Diesel exhaust affects the abnormal delivery in pregnant mice and the growth of
their young. Inhal Toxicol, 14: 635-651. 030593
Tsukue N; Yoshida S; Sugawara I; Taked K. (2004). Effect of diesel exhaust on development of fetal reproductive function
in ICR female mice. Eisei Kagaku, 50: 174-80. 096643
Tzeng H-P; Yang R-S; Ueng T-H; Lin-Shiau S-Y; Liu S-H. (2003). Motorcycle exhaust particulates enhance
vasoconstriction in organ culture of rat aortas and involve reactive oxygen species. Toxicol Sci, 75: 66-73. 097247
December 2009 D-193
-------
Tzeng H-P; Yang Rong S; Ueng T-H; Liu S-H. (2007). Upregulation of cyclooxygenase-2 by motorcycle exhaust
particulate-induced reactive oxygen species enhances rat vascular smooth muscle cell proliferation. Chem Res
Toxicol, 20: 1170-1176. 097883
U.S. EPA. (1976). Air quality data - 1970 annual statistics. U.S. Environmental Protection Agency. Research Triangle Park,
NC. EPA-450/2-76-019. 015607
Ueng T-H; Hung C-C; Kuo M-L; Chan P-K; Hu S-H; Yang P-C; Chang LW. (2005). Induction of fibroblast growth factor-9
and interleukin-1 alpha gene expression by motorcycle exhaust particulate extracts and benzo(a)pyrene in human
lung adenocarcinoma cells. Toxicol Sci, 87: 483-496. 097054
Ueng T-H; Wang H-W; Huang Y-P; Hung C-C. (2004). Antiestrogenic effects of motorcycle exhaust particulate in MCF-7
human breast cancer cells and immature female rats. Arch Environ Contam Toxicol, 46: 454-462. 096199
Umbuzeiro GA; Franco A; Martins MH; Kummrow F; Carvalho L; Schmeiser HH; Leykauf J; Stiborova M; Claxton LD.
(2008). Mutagenicity and DNA adduct formation of PAH, nitro-PAH, and oxy-PAH fractions of atmospheric
particulate matter from Sao Paulo, Brazil. Mutat Res Genet Toxicol Environ Mutagen, 652: 72-80. 190491
Upadhyay D; Panduri V; Ghio A; Kamp DW. (2003). Particulate matter induces alveolar epithelial cell DNA damage and
apoptosis: role of free radicals and the mitochondria. Am J Respir Cell Mol Biol, 29: 180-187. 097370
Upadhyay S; Stoeger T; Harder V; Thomas RF; Schladweiler MC; Semmler-Behnke M; Takenaka S; Karg E; Reitmeir P;
Bader M; Stampfl A; Kodovanti U; Schulz H. (2008). Exposure to ultrafine carbon particles at levels below
detectablepulmonary inflammation affects cardiovascular performance inspontaneously hypertensive rats. Part
Fibre Toxicol, 5:19. 159345
Valavanidis A; Vlahoyianni T; Fiotakis K. (2005). Comparative study of the formation of oxidative damage marker 8-
hydroxy-2'-deoxyguanosine (8-OHdG) adduct from the nucleoside 2'-deoxyguanosine by transition metals and
suspensions of particulate matter in relation to metal content and redox reactivity. Free Radic Res, 39: 1071-1081.
096432
Veranth J; Kaser E; Veranth M; Koch M; Yost G (2007). Cytokine responses of human lung cells (BEAS-2B) treated with
micron-sized and nanoparticles of metal oxides compared to soil dusts. Part Fibre Toxicol, 4: 2. 090346
Veranth JM; Moss TA; Chow JC; Labban R; Nichols WK; Walton JC; Walton JG; Yost GS. (2006). Correlation of in vitro
cytokine responses with the chemical composition of soil-derived particulate matter. Environ Health Perspect, 114:
341-349. Q87479
Veranth JM; Reilly CA; Veranth MM; Moss TA; Langelier CR; Lanza DL; Yost GS. (2004). Inflammatory cytokines and
cell death in BEAS-2B lung cells treated with soil dust, lipopolysaccharide, and surface-modified particles. Toxicol
Sci, 82: 88-96. 087480
Veras MM; Damaceno-Rodrigues NR; Caldini EG; Maciel Ribeiro AA; Mayhew TM; Saldiva PH; Dolhnikoff M. (2008).
Particulate urban air pollution affects the functional morphology of mouse placenta. Biol Reprod, 79: 578-584.
190493
Veras MM; Damaceno-Rodrigues NR; Guimaraes Silva RM; Scoriza JN; Saldiva PH; Caldini EG; Dolhnikoff M. (2009).
Chronic exposure to fine particulate matter emitted by traffic affects reproductive and fetal outcomes in mice.
Environ Res, 109: 536-543. 190496
Veronesi B; De Haar C; Lee L; Oortgiesen M. (2002). The surface charge of visible particulate matter predicts biological
activation in human bronchial epithelial cells. Toxicol Appl Pharmacol, 178: 144-154. 024599
Veronesi B; Makwana O; Pooler M; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles
VII Degeneration of dopaminergic neurons inApo E-/- mice. Inhal Toxicol, 17: 235-241. 087481
Verstraelen S; Van Den Heuvel R; Nelissen I; Witters H; Verheyen G; Schoeters G. (2005). Flow cytometric
characterisation of antigen presenting dendritic cells after in vitro exposure to diesel exhaust particles. Toxicol In
Vitro, 19:903-907. 096872
Vogel CFA; Sciullo E; Wong P; Kuzmicky P; Kado N; Matsumura F. (2005). Induction of proinflammatory cytokines and
C-reactive protein in human macrophage cell line U937 exposed to air pollution particulates. Environ Health
Perspect, 113: 1536-1541.087891
Walczak-Drzewiecka A; Wyczolkowska J; Dastych J. (2003). Environmentally relevant metal and transition metal ions
enhance Fc epsilon Rl-mediated mast cell activation. Environ Health Perspect, 111: 708-713. 188803
December 2009 D-194
-------
Wallenborn JG; Evansky P; Shannahan JH; Vallanat B; Ledbetter AD; Schladweiler MC; Richards JH; Gottipolu RR;
Nyska A; Kodavanti UP. (2008). Subchronic inhalation of zinc sulfate induces cardiac changes in healthy rats.
Toxicol Appl Pharmacol, 232: 69-77. 191171
Wallenborn JG; McGee John K; Schladweiler Mette C; Ledbetter Allen D; Kodavanti Urmila P. (2007). Systemic
translocation of particulate matter-associated metals following a single intratracheal instillation in rats. J Toxicol
Sci, 98: 231-239. 156144
Wan ECH; Yu JZ. (2006). Determination of sugar compounds in atmospheric aerosols by liquid chromatography combined
with positive electrospray ionization mass spectrometry. J Chromatogr, 1107: 175-181. 157104
Wang Y-Z; Ingram JL; Walters DM; Rice AB; Santos JH; Van Houten B; Bonner JC. (2003). Vanadium-induced STAT-1
activation in lung myofibroblasts requires H2O2 and P38 MAP kinase. Free Radic Biol Med, 35: 845-855. 157106
Watanabe N. (2005). Decreased number of sperms and Sertoli cells in mature rats exposed to diesel exhaust as fetuses.
Toxicol Lett, 155: 51-58. 087985
WegesserTC; Last JA. (2008). Lung response to coarse PM: bioassay in mice. Toxicol Appl Pharmacol, 230: 159-166.
190506
Wellenius GA; Batalha JRF; Diaz EA; Lawrence J; Coull BA; Katz T; Verrier RL; Godleski JJ. (2004). Cardiac effects of
carbon monoxide and ambient particles in a rat model of myocardial infarction. Toxicol Sci, 80: 367-376. 087874
Wellenius GA; Coull BA; Batalha JRF; Diaz EA; Lawrence J; Godleski JJ. (2006). Effects of ambient particles and carbon
monoxide on supraventricular arrhythmias in a rat model of myocardial infarction. Inhal Toxicol, 18: 1077-1082.
156152
Wellenius GA; Coull BA; Godleski JJ; Koutrakis P; Okabe K; Savage ST. (2003). Inhalation of concentrated ambient air
particles exacerbates myocardial ischemia in conscious dogs. Environ Health Perspect, 111: 402-408. 055691
Whitekus MJ; Li N; Zhang M; Wang M; Horwitz MA; Nelson SK; Horwitz LD; Brechun N; Diaz-Sanchez D; Nel AE.
(2002). Thiol antioxidants inhibit the adjuvant effects of aerosolized diesel exhaust particles in a murine model for
ovalbumin sensitization. J Immigr Minor Health, 168: 2560-2567. 157142
Wichers LB; Nolan JP; Winsett DW; Ledbetter AD; Kodavanti UP; Schladweiler MC J; Costa DL; Watkinson WP. (2004).
Effects of instilled combustion-derived particles in spontaneously hypertensive rats Part II: pulmonary responses.
Inhal Toxicol, 16: 407-419. 055636
Wichers LB; Rowan WH, 3rd; Nolan JP; Ledbetter AD; McGee JK; Costa DL; Watkinson WP. (2006). Particle deposition
in spontaneously hypertensive rats exposed via whole-body inhalation: measured and estimated dose. Toxicol Sci,
93: 400-410. 103806
Wilson MR; Foucaud L; Barlow PG; Hutchison GR; Sales J; Simpson RJ; Stone V (2007). Nanoparticle interactions with
zinc and iron: Implications for toxicology and inflammation. Toxicol Appl Pharmacol, 225: 80-89. 097268
Win-Shwe TT; Yamamoto S; Ahmed S; Kakeyama M; Kobayashi T; Fujimaki H. (2006). Brain cytokine and chemokine
mRNA expression in mice induced by intranasal instillation with ultrafine carbon black. Toxicol Lett, 63: 153-160.
088415
Win-Shwe TT; Yamamoto S; Fujitani Y; Hirano S; Fujimaki H. (2008). Spatial learning and memory function-related gene
expression in the hippocampus of mouse exposed to nanoparticle-rich diesel exhaust. Neurotoxicology, 29: 940-
947. 190146
Witten ML; Wong SS ; Sun NN; Keith I; Kweon C; Foster DE; Schauer JJ; Sherrill DL. (2005). Neurogenic responses in
rat lungs after nose-only exposure to diesel exhaust. Health Effects Institute. Boston, MA. 087485
Wold LE; Simkhovich BZ; Kleinman MT; Nordlie MA; Dow JS; Sioutas C; Kloner RA. (2006). In vivo and in vitro
models to test the hypothesis of particle-induced effects on cardiac function and arrhythmias. Cardiovasc Toxicol,
6: 69-78. 097028
Wong SS; Sun NN; Keith I; Kweon C-B; Foster DE; Schauer James J; Witten ML. (2003). Tachykinin substance P
signaling involved in diesel exhaust-induced bronchopulmonary neurogenic inflammation in rats. Arch Toxicol, 77:
638-650. 097707
Wottrich R; Diabate S; Krug HF. (2004). Biological effects of ultrafine model particles in human macrophages and
epithelial cells in mono- and co-culture. Int J Hyg Environ Health, 207: 353-361. 094518
December 2009 D-195
-------
Wu W; Samet JM; Silbajoris R; Dailey LA; Sheppard D; Bromberg PA; Graves LM. (2004). Heparin-binding epidermal
growth factor cleavage mediates zinc-induced epidermal growth factor receptor phosphorylation. Am J Respir Cell
Mol Biol, 30: 540-547. 096949
Wu W; Silbajoris RA; Whang YE; Graves LM; Bromberg PA; Samet JM. (2005). p38 and EGF receptor kinase-mediated
activation of the phosphatidylinositol 3-kinase/Akt pathway is required for Zn2+-induced cyclooxygenase-2
expression. Am J Physiol Lung Cell Mol Physiol, 289: L883-L889. 097350
Wu W; Wang X; Zhang W; Reed W; Samet JM; Whang YE; Ohio AJ. (2003). Zinc-induced PTEN protein degradation
through the proteasome pathway in human airway epithelial cells. J Biol Chem, 278: 28258-28263. 199749
Wu YH; Vincent JH. (2007). A modified Marple-type cascade impactor for assessing aerosol particle size distributions in
workplaces. J Occup Environ Hyg, 4: 798-807. 098412
Xu D-Q; Zhang W-L. (2004). Monitoring of pollution of air fine particles (PM25) and study on their genetic toxicity.
Biomed Environ Sci, 17: 452-458. 097231
Yacobi NR; Phuleria HC; Demaio L; Liang CH; Peng CA; Sioutas C; Borok Z; Kim KJ; Crandall ED. (2007). Nanoparticle
effects on rat alveolar epithelial cell monolayer barrier properties. Toxicol In Vitro, 21:1373-1381. 156166
Yamamoto S; Tin Tin Win S; Ahmed S; Kobayashi T; Fujimaki H. (2006). Effect of ultrafine carbon black particles on
lipoteichoic acid-induced early pulmonary inflammation inBALB/c mice. Toxicol Appl Pharmacol, 213: 256-266.
096671
Yanagisawa R; Takano H; Inoue K; Ichinose T; Sadakane K; Yoshino S; Yamaki K; Kumagai Y; Uchiyama K; Yoshikawa
T; Morita M. (2003). Enhancement of acute lung injury related to bacterial endotoxin by components of diesel
exhaust particles. Thorax, 58: 605-612. 087487
Yanagisawa R; Takano H; Inoue KI; Ichinose T; Sadakane K; Yoshino S; Yamaki K; Yoshikawa T; Hayakawa K. (2006).
Components of diesel exhaust particles differentially affect Thl/Th2 response in a murine model of allergic airway
inflammation. Clin Exp Allergy, 36: 386-395. 096458
Yang C-Y; Tseng Y-T; Chang C-C. (2003). Effects of air pollution on birthweight among children born between 1995 and
1997 inKaohsiung, Taiwan. J Toxicol Environ Health A, 66: 807-816. 087886
Yatera K; Hsieh J; Hogg James C; Tranfield E; Suzuki H; Shih C-H; Behzad Ali R; Vincent R; van Eeden Stephan F.
(2008). Particulate matter air pollution exposure promotes recruitment of monocytes into atherosclerotic plaques.
Am J Physiol Heart Circ Physiol, 294: H944-H953. 157162
Yauk C; Polyzos A; Rowan-Carroll A; Somers CM; Godschalk RW; Van Schooten FJ; Berndt ML; Pogribny IP; Koturbash
I; Williams A; Douglas GR; Kovalchuk O. (2008). Germ-line mutations, DNA damage, and global
hypermethylation in mice exposed to particulate air pollution in an urban/industrial location. PNAS, 105: 605-610.
157164
Yin H; Too HP; Chow GM. (2005). The effects of particle size and surface coating on the cytotoxicity of nickel ferrite.
Biomaterials, 26: 5818-5826. 088133
Yin X-J; Schafer R; Ma JYC; Antonini JM; Roberts JR; Weissman DN; Siegel PD; Ma JKH. (2003). Alteration of
pulmonary immunity to Listeria monocytogenes by diesel exhaust particles (DEPs) II Effects of DEPs on T-cell-
mediated immune responses in rats. Environ Health Perspect, 111: 524-530. 096127
Yin XJ; Dong CC; Ma JY; Roberts JR; Antonini JM; Ma JK. (2007). Suppression of phagocytic and bactericidal functions
of rat alveolar macrophages by the organic component of diesel exhaust particles. J Toxicol Environ Health A, 70:
820-828. 198980
Yin XJ; Dong CC; Ma JYC; Antonini JM; Roberts JR; Stanley CF; Schafer R; Ma JKH. (2004). Suppression of cell-
mediated immune responses to listeria infection by repeated exposure to diesel exhaust particles in brown Norway
rats. Toxicol Sci, 77: 263-271.097685
Yin XJ; Ma JYC; Antonini JM; Castranova V; Ma JKH. (2004). Roles of reactive oxygen species and heme oxygenase-1 in
modulation of alveolar macrophage-mediated pulmonary immune responses to listeria monocytogenes by diesel
exhaust particles. Toxicol Sci, 82: 143-153. 087983
Ying Z; Kampfrath T; Thurston G; Farrar B; Lippmann M; Wang A; Sun Q; Chen LC; Rajagopalan S. (2009). Ambient
particulates alter vascular function through induction of reactive oxygen and nitrogen species. Toxicol Sci, 1: 1-36.
190111
December 2009 D-196
-------
Yokohira M; Takeuchi H; Yamakawa K; Saoo K; Matsuda Y; Zeng Y; Hosokawa K; Imaida K. (2007). Bioassay by
intratracheal instillation for detection of lung toxicity due to fine particles in F344 male rats. Exp Toxicol Pathol,
58:211-221.097976
Yokota S; Furuya M; Seki T; Marumo H; Ohara N; Kato A. (2004). Delayed exacerbation of acute myocardial
ischemia/reperfusion-induced arrhythmia by tracheal instillation of diesel exhaust particles. Inhal Toxicol, 16: 319-
331.096516
Yokota S; Mizuo K; Moriya N; Oshio S; Sugawara I; Takeda K. (2009). Effect of prenatal exposure to diesel exhaust on
dopaminergic system in mice. Neurosci Lett, 449: 38-41. 190518
Yokota S; Seki T; Furuya M; Ohara N. (2005). Acute functional enhancement of circulatory neutrophils after intratracheal
instillation with diesel exhaust particles in rats. Inhal Toxicol, 17: 671-679. 096003
Yokota S; Seki T; Naito Y; Tachibana S; Hirabayashi N; Nakasaka T; Ohara N; Kobayashi H. (2008). Tracheal instillation
of diesel exhaust particles component causes blood and pulmonary neutrophilia and enhances myocardial oxidative
stress in mice. J Toxicol Sci, 33: 609-620. 190109
Yoshida S; Ono N; Tsukue N; Oshio S; Umeda T; Takano H; Takeda K. (2006). In utero exposure to diesel exhaust
increased accessory reproductive gland weight and serum testosterone concentration in male mice. Environ Sci J
Integr Environ Res, 13: 139-147. 097015
Yoshida S; Takedab K. (2004). The effects of diesel exhaust onmurine male reproductive function. Eisei Kagaku, 50: 210-
214. 097760
Yoshida S; Yoshida M; Sugawara I; Takeda K. (2006). Mice strain differences in effects of fetal exposure to diesel exhaust
gas on male gonadal differentiation. Environ Sci J Integr Environ Res, 13: 117-123. 156170
Yun Y-P; Joo J-D; Lee J-Y; Nam H-Y; Kim Y-H; Lee K-H; Lim C-S; Kim H-J; Lim Y-G; Lim Y. (2005). Induction of
nuclear factor-"kappa"B activation through TAK1 and NIK by diesel exhaust particles in L2 cell lines. Toxicol Lett,
155: 337-342. 088302
Zanchi AC; Venturini CD; Saiki M; Nascimento Saldiva PH; Tannhauser Barros HM; Rhoden CR. (2008). Chronic nasal
instillation of residual-oil fly ash (ROFA) induces brain lipid peroxidation and behavioral changes in rats. Inhal
Toxicol, 20: 795-800. 157173
Zelikoff JT; Chen LC; Cohen MD; Fang K; Gordon T; Li Y; Nadziejko C; Schlesinger RB. (2003). Effects of inhaled
ambient particulate matter on pulmonary antimicrobial immune defense. Inhal Toxicol, 15: 131-150. 039009
Zelikoff JT; Schermerhorn KR; Fang K; Cohen MD; Schlesinger RB. (2002). A role for associated transition metals in the
immunotoxicity of inhaled ambient particulate matter. Environ Health Perspect, 5: 871-875. 037797
Zhang J; Ghio AJ; Chang W; Kamdar O; Rosen GD; Upadhyay D. (2007). Bim mediates mitochondria-regulated
particulate matter-induced apoptosis in alveolar epithelial cells. FEES Lett, 581: 4148-4152. 156179
Zhang Q; Kleeberger SR; Reddy SP (2004). DEP-induced fra-1 expression correlates with a distinct activation of AP-1-
dependent gene transcription in the lung. Am J Physiol Lung Cell Mol Physiol, 286: L427-L436. 157183
Zhang Z; Che W; Liang Y; Wu M; Li N; Shu Y; Liu F; Wu D. (2007). Comparison of cytotoxicity and genotoxicity induced
by the extracts of methanol and gasoline engine exhausts. Toxicol In Vitro, 21: 1058-1065. 157186
Zhao H; Barger MW; Ma JKH; Castranova V; Ma JYC. (2006). Cooperation of the inducible nitric oxide synthase and
cytochrome P450 1 Al in mediating lung inflammation and mutagenicity induced by diesel exhaust particles.
Environ Health Perspect, 114: 1253-1258. 100996
Zhao L; Wang X; He Q; Wang H; Sheng G; Chan LY; Fu J; Blake DR. (2004). Exposure to hazardous volatile organic
compounds, PM sub(10) and CO while walking along streets in urban Guangzhou, China. Atmos Environ, 38:
6177-6184. 100972
Zhong W; Levin L; Reponen T; Hershey GK; Adhikari A; Shukla R; LeMasters G. (2006). Analysis of short-term
influences of ambient aeroallergens on pediatric asthma hospital visits. Sci Total Environ, 370: 330-336. 093264
Zhou Y-M; Zhong C-Y; Kennedy IM; Pinkerton KE. (2003). Pulmonary responses of acute exposure to ultrafine iron
particles in healthy adult rats. Environ Toxicol, 18: 227-235. 087940
December 2009 D-197
-------
Annex F. Source Apportionment Studies
Table F-1. Epidemiologic studies of ambient PM sources, factors, or constituents
Reference: Andersen et Subjects: NR N: NR
al. (2007.093201)
Exposure: NR
Location: 1 monitor in
Copenhagen, Denmark/
6yr, but apportionment
done for 1 .5 yr only
(2002-2003)
Particle Size: PM10
Number of
Constituents
considered for
grouping: 31
Grouping
method :PCA +
PMF/CMB hybrid
(COPREM)
# of groups: 12,
but only 6 used in
relating to health
effects, and CO,
N02
Groups/Factors/ Sources:
Road, vehicle, salt,
biomass, oil, coal, rock,
lime, NaN03, NH4N03,
(NH4)2S03, (NH4S04)
PM variables used:
Mass contribution of
sources
Results: Single pollutant models: Biomass, secondary compounds, oil, and crustal significantly associated with CVD HA
(4-day ma). Biomass and secondary components significantly associated with respiratory HA (5-day ma). No significant effects
for asthma HA in children (6-day ma).
Two pollutant models: Crustal effect for CVD admissions remained robust. Biomass effect for respiratory admissions was
highest. Effect of vehicle source remained robust for asthma admissions in children in presence of other PM 10 sources.
Reference: Bell et al. (Bell
etal., 2009, 191007)
Location: PM25:
2000-2005 (6 yr)/1 06 US
counties/EPA composition
data
Subjects: NR N: NR
Exposure: NR
Number of
Constituents
considered for
grouping: 16
elements + N03,
S04, EC, OC
Grouping
method: NR
# of groups: NR
Groups/Factors/ Sources:
NR
PM variables used:
Every component (16
elements + N03,
S04,EC, OC)
PM10:1987-2000/100
counties/EPA composition
data
Particle Size: PM10, PM2.5
Results: Mortality: Ni significantly increased PM10 mortality risks. However, effect of Ni was not significant when New York City
was removed, in a sensitivity analysis conducted by selectively removing cities from the overall estimate.
Hospital Admissions: CVD and respiratory HAs higher in counties with higher EC, Ni, and V PM25. In CVD association
between PM25, RR and V robust to inclusion of EC or V, and V robust to inclusion of EC.
Reference: Cakmak et al.
(2009,191995)
Location: 1 monitor in
Santiago, Chile
Particle Size: PM25
Subjects: NR N: NR
Exposure:
1998-2009
(8.3 yr)
Number of
Constituents
considered for
grouping: 16
elements + CO,
N02, S02, EC,
OC
Grouping
method: PCA
# of groups: 4
Groups/Factors/ Sources:
Vehicle (CO, N02, EC, OC),
Soil (Al, Ca, Fe, Si),
Combustion (Cr, Cu, Fe,
Mn.Zn), Factor 4 (Br.CI,
Pb)
PM variables used:
individual components,
then groupings
Results: Individual components: EC, OConly stat. sign, risk estimates for total, cardiac, and respiratory mortality for 1-day lag
after adjustment for other elements.
Groupings: Lag 1. Vehicle factor: Increased total mortality, cardiac mortality, and respiratory mortality. Soil factor: increased
cardiac mortality and respiratory mortality (but smaller than vehicle factor RRs). Combustion factor: greatest RR for respiratory
mortality, but significant for total and cardiac mortality. Factor 4: increased total, cardiac, and respiratory mortality. Point
estimates for Factor 1 significantly different from Factors 3 and 4. Elderly had higher risk estimates for combustion and soil
sources. No significant effect modification by gender or season.
Reference: Franklin et al.
(2008, 155779)
Location : STN/25
communities/2000-2005
(6yr)
Particle Size: PM25
Subjects: NR N: NR Number of
Constituents
Exposure: NR considered for
grouping: 15
elements + EC,
OC, N03
Grouping Groups/Factors/Sources: PM variables used:
method: NR NR Every component
# of groups: NR
Results: The PM25-mortality association was significantly modified by Al, As, Sulfate, Ni, and Si. When including a combination
of species proportions and using backwards elimination Al, sulfate, and Ni remained significant. Aland Ni explained most of the
residual heterogeneity.
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009
F-1
-------
Annex E. Epidemiologic Studies
E.1. Short-Term Exposure and Cardiovascular Outcomes
E.1.1.Cardiovascular Morbidity Studies
Table E-1 Short-term exposure - cardiovascular morbidity outcomes:
Study
Design & Methods
Concentrations
Effect Estimates (95% Cl)
Reference: Baccarelli et al. (2007,
0913101
Period of Study: Jan 1995-Aug 2005
Location: Lombardia region, Italy
Outcome: Fasting and postmethionine-
load total homocysteine (tHcy)
Age Groups: 11-84yr
Study Design: Cross-sectional / Panel
N: 1,213 participants
Statistical Analyses: Generalized
additive models
Covariates: Age, sex, BMI, smoking,
alcohol, hormone use, temperature, day
of the yr, and long-term trends
Season: Adjusted for long-term trends
to account for season
Dose-response Investigated? No
Statistical Package: R v2 2 1
Lags Considered: 1-day, 7-day ma.
Pollutant: PMio (some TSP measures
used to predict PM10)
Averaging Time: 24 h
Mean (SD): NR
Percentiles:
25th: 20.1
50th: 34.1
75th: 52.6
Max: 390.0
Monitoring Stations: 53
Copollutant: CO, N02, S02, 03
PM Increment: IQR
Percent Change: [Lower Cl, Upper
Cl]: Homocysteine, fasting: 0.4 (-2.4,
3.3)
Homocysteine, postmethionine-load:
1.1 (-1.5,3.7)
Percent Change: per 26.7m3
increase in 7-day ma of PMio
Homocysteine, fasting: 1.0 (-1.9, 3.9)
Homocysteine, postmethionine-load:
2.0 (-0.6, 4.7)
Percent Change: on fasting
homocysteine per IQR increase in
24-h PMio levels
Among smokers: 6.2 (0.0,12.7)
Among non-smokers: -1.6 (-5.5, 2.5)
Percent Change: on postmethionine-
load homocysteine per IQR increase
in 24-h PMio levels: Among smokers:
6.0(0.5,11.8)
Among non-smokers: -0.1 (-3.6, 3.5)
December 2009
E-1
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Baccarelli et al. (2007,
0907331
Period of Study: Jan 1995-Aug 2005
Location: Lombardia region, Italy
Outcome: Prothrombin time (PT)
Activated partial thromboplastin time
(APTT)
Fibrinogen
Functional antithrombin
Functional protein C
Protein C, antigen
Functional protein S
Free protein S
Age Groups: 11-84yr
Study Design: Cross-sectional / Panel
N: 1,218 participants
Statistical Analyses: Generalized
additive models
Covariates: Age, sex, BMI, smoking,
alcohol, hormone use, temperature, day
of the yr, and long-term trends
Season: Adjusted for long-term trends
to account for season
Dose-response Investigated? No
Statistical Package: R software v2.2.1
Pollutant: PMi0 (some TSP measures
used to predict PMio)
Averaging Time: Hourly concentrations
used to calculate lags of same day, 7-
day, 30-day, and h 0-6
Mean (SD): NR
Percentiles:
Sep-Nov:
5th: 33.1
50th: 51.2
75th: 76.5
Max: 148.9
Dec-Feb:
25th: 47.9
50th: 68.5
75th: 95.3
Max: 238.3
Mar-May:
25th: 30.0
50th: 64.1
75th: 64.8
Max: 158.5
Jun-Aug:
25th: 28.0
50th: 44.3
75th: 61.3
Max: 94.7
Monitoring Stations: 53 sites
Copollutant: CO, N02, S02, 03
PM Increment: SD
Effect Estimate [Lower Cl, Upper Cl]:
Estimated changes in endpoint
PT (international normalized ratio):
At time of blood sample: -0.06 (-0.12,
0.00)
Avg levels 7 days prior: -0.03 (-0.10,
0.04)
Avg levels 30 days prior: -0.08 (-0.14, -
0.01)
(Hourly ma presented in Fig 2)
APTT (ratio to reference plasma):
At time of blood sample: 0.02 (-0.04,
0.08)
Avg levels 7 days prior: 0.00 (-0.07,
0.06)
Avg levels 30 days prior: 0.01 (-0.06,
0.08)
Fibrinogen:
At time of blood sample: 0.01 (-0.05,
0.07)
Avg levels 7 days prior: -0.03 (-0.09,
0.04)
Avg levels 30 days prior: -0.02 (-0.09,
0.05)
Functional antithrombin:
At time of blood sample: -0.02 (-0.09,
0.04)
Avg levels 7 days prior: -0.06 (-0.13,
0.01)
Avg levels 30 days prior: -0.06 (-0.13,
0.02)
Functional protein C:
At time of blood sample: 0.00 (-0.06,
6.1)
Avg levels 7 days prior: -0.06 (-0.12,
0.01)
Avg levels 30 days prior: -0.06 (-0.14,
0.01)
Protein C, antigen:
At time of blood sample: 0.00 (-0.06,
6.0)
Avg levels 7 days prior: -0.04 (-0.10,
0.03)
Avg levels 30 days prior: -0.06 (-0.14,
0.01)
Functional protein S:
At time of blood sample: 0.04 (-0.03,
0.10)
Avg levels 7 days prior: -0.03 (-0.11,
0.06)
Avg levels 30 days prior: -0.14 (-0.23,
-0.05)
Free protein S:
At time of blood sample: 0.05 (-0.01,
0.10)
Avg levels 7 days prior: 0.01 (-0.05,
0.07)
Avg levels 30 days prior: -0.01 (-0.08,
0.06)
December 2009
E-2
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Barclay et al.
(2009, 1799351
Period of Study: Jan 2003-May 2005
Location: Aberdeen, Scotland
Outcome: Haematological outcomes,
Heart Rhythm outcomes, & Heart Rate
Variability outcomes
Age Groups: 70.4 (8.9)
Study Design: Panel
N: 132 patients w/ chronic heart failure
Statistical Analyses: Linear & Mixed
Effects Regression Model
Covariates: Age, temperature,
humidity, pressure
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: Lags 0-2 day
Pollutant: PM,0
Averaging Time: daily
Mean (SD): 20.25
Min:7.375
Max: 68.3
Monitoring Stations: 1
Copollutant: PM25, PNC, N02
Co-pollutant Correlation:
N02 city: 0.294
NO city: 0.112
N02 personal: 0.055
PNC DEOM: 0.241
PM25 total: 0.476*
PM25 traffic: 0.882*
PNC total: 0.125
PNC traffic: 0.190
'Correlations based on 3-day avg
concentrations
PM Increment: NR
Beta (Lower Cl, Upper Cl):
Haemoglobin: 0.136 (-0.274, 0.546)
Mean corpuscular haemoglobin: 0.030
(-0.232, 0.291)
Platelets: 0.096 (-0.923,1.115)
Haematocrit:0.131 (-0.289,0.551)
White blood cells: 0.034 (-1.175,1.244)
C reactive protein: -4.872 (-12.094,
2.351)
IL-6: 2.207 (-4.995, 9.410)
von Willebrand factor: 0.660 (-2.651,
3.970)
E-selectin:-0.536 (-2.528, 1.457)
Fibrinogen:-0.432 (-2.470,1.607)
Factor VII: 0.990 (-1.265, 3.245)
day-dimer:-1.225 (-4.505, 2.055)
All arrhythmias: -3.447 (-11.521, 4.627)
Ventricular ectopic beats: -2.110 (-
12.135,7.915)
Ventricular couplets: -1.561 (-10.811,
7.689)
Ventricular runs: -0.709 (-6.677, 5.259)
Supraventricular ectopic beats: 0.033
(-9.242, 9.308)
Supraventricular couplets: 0.006
(-8.618, 8.629)
Supraventricular runs: 3.710 (-2.847,
10.266)
Avg HR: 0.321 (-0.197, 0.838)
24 hSDNN: 1.040 (-0.415, 2.494)
24 hSDANN: 1.195 (-0.473, 2.863)
24 hRMSSD: 0.321 (-0.197, 0.838)
24 hPNN: 2.837 (-3.791, 9.465)
24 h LF power: 0583 (-3.622, 4.787)
24 hLF normalized:-3.137 (-5.540,
-0.733)*
24 h HF power: 0.872 (-4.649, 6.392)
24 h HF normalized: -2.223 (-4.952,
0.505)
24 h LF/HF ratio: -0.296 (-3.832, 3.240)
*p < 0.05
Notes: LF= low frequency
HF= high frequency
Reference: Briet et al. (2007, 0930491
Period of Study: NR
Location: Paris, France
Outcome: Endothelial Function
Age Groups: 20-40 yr
Study Design: Panel
N: 40 white male nonsmokers
Statistical Analyses: Multiple Robust
Regrssion
Covariates: R53R/R53H genotype,
diet, subject factor, visit, temperature
Dose-response Investigated? No
Statistical Package: NCSS
Lags Considered: 0-5 day
Pollutant: PM,0
Averaging Time: 24 h
5 day Mean (SD): 43 (10)
Monitoring Stations: NR
Co-pollutant: PM25, S02, NO, NO;
Co-pollutant Correlation: N/A
PM Increment: 1 SD
Beta (Lower Cl, Upper Cl), P, R2:
Flow-mediated brachial artery dilation:
0.07 (-0.62, 0.76), NS, 0.03
Reactive hyperemia:
CQ 15.91 (7.74, 24.0), O.001, 0.16
December 2009
E-3
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Choi et al. (2007, 0931961
Period of Study: 2001-2003
Location: Incheon, South Korea
Outcome: Blood pressure Pollutant: PMi0
Study Design: Cross-sectional Averaging Time: Measured hourly and
calculated 24-h means
N: 10459 subjects with a hospital health
examination Percentiles:
Warm season: Median: 36.7
Statistical Analyses: Linear regression Co|d season: Median: 45 7
Covariates: Season: Effect
modification by season
Monitoring Stations: 9 stations
Copollutant: N02, S02
PM Increment: 10 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
Estimate (p-value) for the relationship
between systolic blood pressure (SBP)
and diastolic blood pressure (DBP) and
an increase in PMi0 on lag day 1
SBP: Warm season: 0.0798 (p < 0.001)
DBP: Warm season: 0.0240 (p < 0.001)
Note: No evidence of associations
between PMi0 and BP during the cold
season
Reference: Chuang et al. (2007,
0910631
Period of Study:
Between Apr-Jun 2004 or 2005
Location: Taipei, Taiwan
Outcome: High-sensitivity C-reactive
protein (hs-CRP)
Fibrinogen, plasminogen activator
fibrinogen inhibitor-1 (PAI-1), tissue-
type plasminogen activator (tPA), 8-
hydroxy-2'-deoxyguanosine(8-OHdG),
and log-transformed HRV indices
(SDNN = standard deviation of NN
intervals, r-MSSD = square root of the
mean of the sum of the squares of
differences between adjacent NN
intervals, LF = low frequency [0.04-
0.15Hz], and HF= high frequency
[0.15-0.40HZ])
Age Groups: 18-25yr
Study Design: Panel (cross-sectional)
N: 76 students
Statistical Analyses: Linear mixed-
effects models
Covariates: Age, sex, BMI, weekday,
temperature of previous day, relative
humidity
Season: Only 1 season of data
collection
Dose-response Investigated? No
Statistical Package: NR
Pollutant: PM,0
Averaging Time: Hourly data used to
calculate avg over 1- to 3-day periods
Mean (SD): 1-day avg: 49.2 (18.0)
2-day avg: 55.3 (18.6)
3-day avg: 54.9 (18.2)
Range (Min, Max):
1-day avg: 29.5, 83.4
2-day avg: 25.5, 85.1
3-day avg: 22.2, 87.2
Monitoring Stations: 2 sites (each
pollutant measured at one site only)
Copollutant: PM25, Sulfate, Nitrate,
OC, EC, N02, CO, S02, 03
PM Increment: IQR (1-day avg: 32.7
2-day avg: 34.5
3-day avg: 26.0)
Effect Estimate [Lower Cl, Upper Cl]:
% change in health endpoint per
increase in IQR of PM10 (1-3 day
averaging period
single pollutant models)
hs-CRP: 1-day: 135.8 (1.8, 269.7)
2-day: 108.2 (-10.9, 227.3)
3-day: 109.6 (2.5, 216.7)
8-OHdG:1-day:-9.2(-21.5, 3.2)
2-day:-6.1 (-17.0, 4.8)
3-day:-5.6 (-13.8, 2.6)
PAI-1:1-day: 30.0 (12.4, 47.7)
2-day: 19.1 (3.6, 34.7
3-day: 21.2 (9.7, 32.8
tPA:1-day: 16.0 (-4.1,36.2)
2-day: 10.4 (-6.3, 27.2)
3-day: 8.8 (-2.8, 20.5)
Fibrinogen: 1-day: 5.3 (1.5,15.2)
2-day: 1.5
3-day: 3.3
-4.4, 7.5)
-1.1,7.7)
Heart Rate Variability
SDNN: 1-day:-4.9 (-7.8,-2.1)
2-day:-4.0 (-6.6,-1.4)
3-day:-4.1 (-6.1,-2.2)
r-MSSD: 1-day:-4.8 (-12.3, 2.7)
2-day: -2.2 (-9.0, 4.7)
3-day: -4.0 (-9.0, 0.9)
LF: 1-day:-6.1 (-10.1,-2.1)
2-day:-3.0 (-7.2,1.2)
3-day:-4.3 (-7.0,-1.6)
HF: 1-day:-5.5 (-13.0, 2.1)
2-day:-2.7 (-9.5, 4.1)
3-day: -2.0 (-7.2, 3.2)
December 2009
E-4
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ebelt et al. (2005, 0569071
Period of Study: Summer of 1998
Location: Vancouver, Canada
Outcome: CVD
Age Groups: Range from 54-86 yr
mean age= 74 yr
Study Design: Extended analysis of a
repeated-measures panel study
N: 16 persons with COPD
Statistical Analyses:
Earlier analysis expanded by
developing mixed-effect regression
models and by evaluating additional
exposure indicators
Dose-response Investigated? No
Statistical Package: SASV8
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD):
Ambient PM10:17 + 6
Exposure to ambient PMi0:10.3 ± 4.6
Range (Min, Max):
Ambient PM10.2.5:7-36
Exposure to ambient PM10-2s:
1.5-23.8
Monitoring Stations: 5
Copollutant (correlation):
Ambient concentrations and exposure
to ambient PM were highly correlated
for each respective metric: r > 0.71
PM10.2.5:r >0.72
PM25:r >0.92
Note: Total personal fine particle
exposure (T) were dominated by
exposures to non ambient particles
which were not correlated with ambient
fine particle exposure (A) or ambient
concentrations (C). Results for each of
these metrics are listed.
Effect estimates and 96% Cl for IQR
range increases in exposure
Increment: C10: IQR = 7 pg/m3
SBP (mm Hg): -2.2 (-4.78-0.38)
DBP (mm Hg):-0.78 (-2.65-1.09)
Ln-SVE(bph): 0.16 (-0.07-0.40)
HR(bpm): 1.02 (-0.79-2.82)
SDNN(ms):-2.14 (-6.94-2.65)
R-MSSD(ms):-2.24 (-4.27-0.21)
Increment: A10: IQR = 6.5|jg/m3
SBP (mm Hg): -2.81 (-5.67-0.05)
DBP (mm Hg):-0.59 (-2.79-1.62)
Ln-SVE (bph): 0.27 (0.03-0.52)
HR(bpm): 0.86 (-1.61-3.33)
SDNN(ms):-3.91 (-9.73-1.91)
R-MSSD(ms):-0.81 (-4.94-3.31)
Reference: Folino et al. (2009, 1919021
Period of Study: Jun 2006-May 2007
Location: Padua, Italy
Outcome: HRV & Inflammatory
Markers
Age Groups: 45-65 yr
Study Design: Panel
N: 39 patients w/ myocardial infarction
Statistical Analyses: Linear
Regression Model, ANOVA
Covariates: Temperature, relative
humidity, atmospheric pressure, beta-
blocker, aspirin, or nitrate consumption,
smoking habit
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD):
Summer: 46.4 (16.1)
Winter: 73.0 (30.9)
Spring: 38.3 (15. 4)
Monitoring Stations: NR
Copollutant: PM25, PM025
Co-pollutant Correlation: NR
PM Increment: 1 pg/m3
Beta (SE), p-value:
SDNN: 0.115 (0.093), 0.218
SDANN' 0138 (0103) 0182
RMSSD: 0.049 (0.034), 0.146
pH: 0.002 (0.001), 0.033
LTB4: 0.427 (0.0279), 0.126
eNO: 0.000 (0.002), 0.851
PTX3: -0.003 (0.001), 0.033
C-reactive protein: -0.006 (0.004),
0.161
CC16: -0.002 (0.002), 0.280
IL-8: 0.000 (0.003), 0.895
Dose-response Investigated? No
Statistical Package: Stata
Lags Considered: NR
Reference: Forbes et al. (2009,
1903511
Period of Study: 1994,1998, 2003
Location: England
Outcome: Inflammation markers
Age Groups: 16+yr
Study Design: Cross-sectional
N: 25,000 white adults w/fibrinogen
measurements & 17,000 white adults w/
C-reactive protein measurements
Statistical Analyses: Multilevel Linear
Regression Models
Covariates: Age, sex, BMI, social
class, region, cigarette smoking
Dose-response Investigated? No
Statistical Package: Stata
Lags Considered: NR
Pollutant: PM10
Averaging Time: Yearly
1994
Median: 19.5
Range: 12.5-36.1
IQR: 3.7
1998
Median: 17.9
Range: 12.6-27.0
IQR: 2.7
2003
Median: 16.2
Range: 11.0-22.7
IQR: 2.6
Monitoring Stations: NR
Copollutant: N02, S02, 03
Co-pollutant Correlation: N/A
PM Increment: 1 pg/m
Percent Change (Lower Cl, Upper
Cl):
Fibrinogen
1994 Crude:-0.068 (-0.367, 0.231)
1994 Adjusted: 0.080 (-0.164, 0.326)
1998 Crude:-0.592 (-0.902,-0.280)
1998 Adjusted: -0.388 (-0.727, -0.047)
2003 Crude:-0.339 (-0.696, 0.019)
2003 Adjusted: -0.069 (-0.458, 0.322)
Combined:-0.077 (-0.254, 0.100)
C-reactive protein
1998 Crude:-0.914 (-2.206, 0.395)
1998 Adjusted:-0.266 (-1.782, 1.274)
2003 Crude: 0.286 (-1.327, 1.925)
2003 Adjusted: 0.661(-1.068, 2.421)
Combined: 0.140 (-1.003,1.296)
December 2009
E-5
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Kaufman (1987, 1909601
Period of Study: Nov 2004-2005
Location: Isfahan, Iran
Outcome: Inflammation
Age Groups: 10-18 yr
Study Design: Panel
N: 374 children
Statistical Analyses: Linear
Regression, Logistic Regression
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 122.08 (33.63)
Oth: 11.00
25th: 86.50
50th: 153.0
Tfith- •iQ'i nn
/ win • i CM . uu
PM Increment: NR
Beta (SE):
CRP: 1.5(0.2)
Ox-LDL:1.4(0.1)
MDA:1.3(0.1)
CDE:1.1 (0.1)
HOMA-IR:1.1(0.3)
Covariates: Age, gender, BMI, waist
circumference, healthy eating index,
physical activity level
Dose-response Investigated? No
Statistical Package: SPSS
Lags Considered: 0- to 7-day avg
Monitoring Stations: 3
Copollutant: 03, S02, N02, CO
Co-pollutant Correlation: NR
Reference: Liao et al. (2004, 0565901
Period of Study: 1996-1998
Location: ARIC study cohort
(Washington County, MD
Forsyth County, NC
and selected suburbs of Minneapolis,
Outcome: 5-min HR, HRV indices (HF,
LF, SDNN)
Study Design: Cross-sectional
Statistical Analyses: Linear regression
The 4th quarter of the ARIC cohort was
sampled exclusively from black
residents of Jackson, MS.
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 24.3 (11.5)
Copollutant:
03
CO
S02
NO,
PM Increment: SD
Effect Estimate [Lower Cl, Upper Cl]:
Estimate (SE)
HF: -0.06 ms2 (0.018)
SDNN:-1.03ms (0.31)
H: 0.32 beats/min (0.158)
Reference: Liao et al. (2005, 0886771
Period of Study: 1987-1989 baseline
health exam
Location: 3 centers in the U.S.
(Forsyth County, NC
suburbs of Minneapolis, MN
black residents of Jackson, MS)
Outcome: Fibrinogen, factor VIII co-
agulant activity (VIII-C), von Willebrand
factor (vWF), white blood cell count
(WBC), and serum albumin
Age Groups: 45-64 yr
Study Design: Cross-sectional
N: 10,208 participants (7705 for PM)
Statistical Analyses: Multiple linear
regression
Covariates: Age, sex, ethnicity-center,
education, smoking, drinking status,
BMI, history of chronic respiratory
disease, humidity, season, cloud cover,
and temperature
Dose-response Investigated?
Yes, examined higher-ordered terms for
each pollutant
Statistical Package: SASv8.2
Pollutant: PM10
Averaging Time: 24-h avg (1, 2, and 3
days prior to the exam)
Mean (SD): 29.9 (29.9)
Mean (SD) within Quartiles:
Q1-3: 24.0 (6.96)
04:47.3(10.11)
Copollutant:
CO, S02, N02, 03
PM Increment: 1 SD (12.8 pg/m3)
Effect Estimate: Adjusted regression
coefficient (SE): Fibrinogen (mg/dl):
0.163(0.755)
Factor VIII-C (%): Non-linear
association: |3 (PM,o) = -5.30, p < 0.01
P (PM,o)2 = 0.80, p < 0.05
vWF(%): Diabetics: 3.93 (1.80)
Nondiabetics: -0.54 (0.58)
Albumin (g/dl): CVD: -0.006 (0.003)
Non-CVD: 0.029 (0.017)
p < 0.05
December 2009
E-6
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Liao et al. (2007, 1802721 Outcome: Ectopy
Period of Study: 1999-2004 Age Groups: Women 50-79 yr
Location: 24 U.S. states Study Design: Panel
N: 57,422
Statistical Analyses: Logistic
regression & random effects modeling
Covariates: Age, race, center,
education, history of CVD/chronic lung
disease, rel. humidity, temperature,
smoking
Pollutant: PM,0
Averaging Time: Daily
Mean (SD)*:
All: 27.5 (12.1)
No Ectopy: 27.5 (12.1)
Any Ectopy: 27.5 (11. 9)
6th, 96th percentile*:
All: 12.2, 48.9
No Ectopy: 12.3, 48.8
Any Ectopy: 11.8,49.3
Monitoring Stations: NR|
PM Increment: 10 pg/m3
Percent Change (Lower Cl, Upper
Cl):
All Ventricular Ectopy
Lag 0:1. 01
Lag 1:1. 02
Lag 2: 0.99
Current Smc
Lag 0:1. 21
Lag 1:1. 32
Lag 2: 1.22
0.95, 1.07)
0.96, 1.09)
0.93, 1.06)
)ker Ventricular Ectopy
0.96, 1.53)
1.07, 1.65)
0.95, 1.56)
Dose-response Investigated? No
Statistical Package: SAS, Stata
Lags Considered: Lags 0-365 day
I Monitors used in model for spatial
interpolation of daily PM values.
Copollutant: PM25
Co-pollutant Correlation: NR
*Lag1
Nonsmoker Ventricular Ectopy
Lag 0:1 (0.93,1.06)
Lag 1:1.01
Lag 2: 0.98
0.94, 1.07)
0.92, 1.05)
All Supraventricular Ectopy
Lag 0:1 (0.95,1.06)
Lag 1:1 (0.95,1.05)
Lag 2: 0.99 (0.94, 1.04)
All Ventricular or Supraventricular
Ectopy
Lag 0:1 (0.95,1.04)
Lag 1:1 (0.96,1.04)
Lag 2: 0.98 (0.94, 1.02)
Reference: Liu et al. (2007,1567051
Period of Study:
May 2005-Jul 2005
Location: Windsor, Ontario, Canada
Outcome: Heart rate, blood pressure,
brachial arterial diameter, flow-mediated
vasodilatation (FMD), plasma cytokines,
and thiobarbituric acid reactive
substances (TEARS)
Age Groups: 18-65yr
Study Design: Panel
N: 24 nonsmoking subjects with type I
or II diabetes over a 7 week period (2-
14 visits for subjects)
170 total vascular measurements and
134 total blood samples collected
Statistical Analyses: Mixed effects
regression models
Covariates: (Time-dependent
covariates) Daily temperature, relative
humidity, blood glucose level, also
checked for confounding by ambient air
pollutant concentrations (controlled for
ambient PM25)
Season: No adjustment since testing
was completed within a 7-wk period
during early summer
Dose-response Investigated? No
Statistical Package: S-Plus
Pollutant: PM10 (personal)
Averaging Time: Real-time monitor
measured exposure during 24-h period
prior to clinic measures
Median (6th-96th percentile):
0-24 h: 25.5 (9.8-133.0)
0-6 h: 15.3 (5.3-83.2)
7-12 h: 17.0 (7.1-186.3)
13-18 h: 28.5 (11.4-167.0)
19-24 h: 30.5 (10.1-148.2)
Monitoring Stations:
Personal monitoring
Copollutant (correlation):
Ambient PM25 (r = 0.34)
PM Increment: 10 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
**p < 0.05
*p < 0.10. Regression coefficients (SE)
End-diastolic basal diameter (urn): All
subjects (n=24): -2.52 (3.27)
subjects not taking vasoactive meds
(n=17): -3.93 (3.66)
subjects w/BMI < 29kg/m2 (n=14): 8.85
(5.85)
End-systolic basal diameter (urn): All
subjects (n=24): -9.02 (3.58)**
subjects not taking vasoactive meds
(n=17): -10.59 (4.36)**
subjects w/BMI <29kg/m2 (n=14): 3.85
(5.49)
End-diastolic FMD (%): All subjects
(n=24): 0.20 (0.08)**
subjects not taking vasoactive meds
(n=17): 0.23 (0.09)**
subjects w/BMI <29kg/m2 (n=14): 0.12
(0.05)**
End-systolic FMD (%): All subjects
(n=24): 0.38 (0.18)**
subjects not taking vasoactive meds
(n=17):0.51 (0.22)**
subjects w/BMI <29kg/m2 (n=14): 0.18
(0.10)*
Flow (cm/s): All subjects (n=24): -0.16
(0.19)
subjects not taking vasoactive meds
(n=17): -0.48 (0.21)**
subjects w/BMI < 29kg/m2 (n=14): -0.39
(0.23)*
Heart rate (bpm): All subjects (n=24):
0.01 (0.11)
subjects not taking vasoactive meds
(n=17): -0.06 (0.12)
subjects w/BMI <29kg/m2 (n=14): 0.15
(0.12)
Diastolic blood pressure (mm Hg): All
subjects (n=24): 0.19 (0.16)
subjects not taking vasoactive meds
December 2009
E-7
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
(n=17): 0.40 (0.18)**
subjects w/BMI < 29kg/m2 (n=14): 0.27
(0.21)
Systolic blood pressure (mm Hg): All
subjects (n=24): 0.17 (0.19)
subjects not taking vasoactive meds
(n=17): 0.43 (0.24)*
subjects w/ BMI < 29kg/m2 (n=14): 0.38
(0.24)
CRP (ug/mL): All subjects (n=24): 0.11
(0.07)
subjects not taking vasoactive meds
(n=17): 0.10 (0.09)
subjects w/ BMI < 29kg/m2 (n=14): 0.02
(0.03)
ET-1 (pg/mL): All subjects (n=24): 0.00
(0.00)
subjects not taking vasoactive meds
(n=17): 0.00 (0.00)
subjects w/BMI < 29kg/m2 (n=14): 0.00
(0.01)
IL-6 (pg/mL): All subjects (n=24): 0.00
(0.05)
subjects not taking vasoactive meds
(n=17):0.01 (0.05)
subjects w/BMI < 29kg/m2 (n=14): -0.00
(0.03)
TNF-a (pg/mL): All subjects (n=24):
0.03 (0.05)
subjects not taking vasoactive meds
(n=17): 0.02 (0.05)
subjects w/ BMI < 29kg/m2 (n=14): 0.03
TEARS (pmol/mL) All subjects (n=24):
16.12(4.00)**
subjects not taking vasoactive meds
(n=17): 8.10 (9.18)
subjects w/ BMI < 29kg/m2 (n=14): -
0.28 (6.60)
regression coefficients (SE) among
subjects not taking vasoactive
medications, with lag time
End-diastolic basal diameter (urn): 0-
6 h: 29.91 (10.64)**
7-12 h: 0.72 (3.95)
13-18 h:-3.62 (2.80)
19-24 h:-0.57 (1.7)
End-systolic basal diameter (urn): 0-6
h: 28.88 (11.22)**
7-12 h:-0.78 (4.58)
13-18 h:-7.70 (3.30)**
19-24 h: -2.87 (2.05)
End-diastolic FMD(%): 0-6 h:-0.12
(0.10)
7-12 h: 0.04 (0.05)
13-18 h: 0.11 (0.03)**
19-24 h: 0.12 (0.04)**
End-systolic FMD(%): 0-6 h: 0.36
(0.08)**
7-12 h: 0.48 (0.32)
13-18 h: 0.19 (0.06)**
19-24 h: 0.34 (0.13)**
Flow (cm/s): 0-6 h:-0.34 (0.22)
7-12 h:-0.26 (0.27
13-18 h:-0.27 (0.15)*
19-24 h:-0.30 (0.11)**
Heart rate (bpm): 0-6 h: 0.31 (0.13)**
7-12 h: 0.26 (0.12)**
December 2009 E-8
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
13-18 h: 0.01 (0.09)
19-24 h: -0.08 (0.05)
Diastolic blood pressure (mm Hg): 0-
6 h:-0.29 (0.12)**
7-12 h: 0.24 (0.12)**
13-18 h: 0.46 (0.17)**
19-24 h: 0.18 (0.14)
Systolic blood pressure (mm Hg): 0-
6 h:-0.65 (0.18)**
7-12 h: 0.17 (0.19)
13-18 h: 0.86 (0.24)**
19-24 h: 0.11 (0.10)
CRP (ug/mL): 0-6 h: 0.15 (0.13)
7-12 h: 0.15 (0.13)
13-18 h: 0.03 (0.06)
19-24 h: 0.04 (0.03)
ET-1 (pg/mL): 0-6 h: 0.02 (0.00)**: 7-12
h: -0.00 (0.00)
13-18 h:-0.00 (0.00)
19-24 h: 0.00 (0.00)
IL-6 (pg/mL): 0-6 h: 0.03 (0.06)
7-12 h: 0.00 (0.06)
13-18 h: 0.02 0.03
19-24 h: 0.00 0.02
TNF-a (pg/mL): 0-6 h: 0.01 (0.07)
7-12 h: 0.09 (0.04)**
13-18 h: 0.01 (0.04)
19-24 h: -0.00 (0.03)
TEARS (pmol/mL): 0-6 h: -4.44 (6.72)
7-12 h: 11.94 (5.08)**
13-18 h: 5.06 (4.03)
19-24 h: 1.06 (4.64)
Note: Adding ambient PM2s data as a
covariate in the model yielded similar
regression coefficients for personal
PM10
Reference: Lipsett et al. (2006,
0887531
Period of Study: Feb-May 2000
Location: Coachella Valley, CA
Outcome: HRV parameters: SDNN,
SDANN, r-MSSD, LF, HF, total power,
triangular index (TRII).
Study Design: Panel study
N: 19 non-smoking adults with coronary
artery disease
Statistical Analysis: Mixed linear re-
gression models with random effects
parameters
Pollutant: PM10
Averaging Time: 2 h
Mean (range):
Indio: 23.2 (6.3-90.4)
Palm Springs: 14 (4.7-52)
Monitoring Stations: 2
Copollutant: 0;
PM Increment: SE*1000
Effect Estimate (change in HRV per
unit increase in PM concentration):
SDNN: -0.71 msec (SE = 0.268)
Notes: Weekly ambulatory 24 h ECG
recordings (once per week for up to 12
wk), using Holter monitors, were made.
Subjects' residences were within 5
miles of 1 of 2 PM monitoring sites.
Regressed HRV parameters against 18:
00-20: 00 mean particulate pollution.
December 2009
E-9
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ljungman et al. (2008,
1802661
Period of Study: Aug 2001-Dec 2006
Location: Gothenburg & Stockholm,
Sweden
Outcome: Ventricular Arrhythmia
Age Groups: 28-85 yr
Study Design: Case-crossover
N: 88 patients w/ implantable
cardioverter defibrillators
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature, humidity,
pressure, ischemic heart disease,
ejection fraction, heart disease,
diabetes, use of beta-blockers, age,
BMI, location at time of arrhythmia,
distance from air pollution monitor
Dose-response Investigated? No
Statistical Package: Stata, S-plus
Lags Considered: Lags 2-24 h
Pollutant: PM,0
Averaging Time: Hourly
Gothenburg, Stockholm
Median:
2h: 18.95, 14.62
24 h: 19.92, 15.23
Min:
2h: 0.00, 0.33
24 h: 2.13, 3.96
Max:
2h: 203.75, 159.79
24 h: 78.01, 90.50
IQR:
2h: 14.16, 11.59
24 h: 11.49, 9.59
Monitoring Stations: 2
Copollutant: PM25, N02
Co-pollutant Correlation
2 h N02: 0.36
24 h N02: 0.29
PM Increment: Interquartile Range
Odds Ratio (Lower Cl, Upper Cl):
2 h: 1.31 (1.00, 1.72)
24 h: 1.24 (0.87, 1.76)
Notes: OR of ventricular arrhythmia for
an IQR increase of air pollutants in
different subgroups (Fig 2)
Reference: Ljungman et al. (2009,
1919831
Period of Study: May 2003-Jul 2004
Location:
Athens, Greece
Helsinki, Finland
Ausburg, Germany
Barcelona, Spain
Rome, Italy
Stokholm, Sweeden
Outcome: lnterleukin-6 Response
Age Groups: 35-80 yr
Study Design: Panel
N: 955 male myocardial infarction
survivors
Statistical Analyses: Additive Mixed
Models
Covariates: Age, sex, BMI, city,
HDL/total cholesterol, smoking,
alcohol intake, HbA1c, NT-proBNP,
history of Ml, heart failure, or
diabetes, phlegm
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 1 day
Pollutant: PM,0
Averaging Time: 24 h
Mean: 31.6
25th: 21.1
76th: 38.4
Monitoring Stations: NR
Copollutant: CO, N02, PNC, PM25
Co-pollutant Correlation
PM25: 0.81
PM Increment: Interquartile Range
(17.4 pg/m3)
Change of IL-6 (Lower Cl, Upper Cl),
p-value:
0.0 (-1.3, 1.3), 1.0
Reference: Mar et al. (2005, 0875661
Period of Study: 1999-2001
Location: Seattle, WA
Outcome: Change in arterial 02 satura-
tion, heart rate, and blood pressure
(SBP and DBP)
Age Groups: >75 yr
Study Design: Panel study
N: 88 elderly subjects
Statistical Analysis: GEE
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD):
Indoor: 12.6 (7.8)
Outdoor: 14.5 (7.0)
PM Increment: 10 pg/m
Unit change in measure(96% Cl):
Among all subjects:
Each increase in outdoor same day
PM10 was associated with: SBP: -0.10
mmHg(95%CI:-1.37,1.18)
DBP: -0.03 mmHg (95% Cl: -0.79, 0.73)
HR: -0.48 beats/min (95% Cl: -1.03,
0.06)
Each increase in indoor same day
PM25was associated with: SBP: 0.92
mmHg (95% Cl:-0.95, 2.78)
DBP: 0.63 mmHg (95% Cl:-0.29,1.56)
HR: 0.02 beats/min (95% Cl: -0.54,
0.58)
Notes: Results by health status
presented in Fig 1. Used 2 sessions
that each were 10 consecutive days of
measurement. Used personal, indoor,
and outdoor measures of PM2 5
December 2009
E-10
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Metzger et al. (2007,
0928561
Period of Study: Jan 1993-Dec 2002
Location: Atlanta, GA
Outcome: Days with any event
recorded by the [CD, days with [CD
shocks/defibrillation and days with
either cardiac pacing or defibrillation
Study Design: Repeated measures
N: 884 subjects
Statistical Analysis: Logistic
regression with GEE to account for
residual autocorrelation within subjects
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 28.0 (12.2)
Median: 26.4
Copollutant:
03, N02, CO, S02. Aug1998-Dec2002:
Oxygenated hydrocarbons
PM Increment: OR (96% Cl):
Outcome = Any event recorded by ICD
OR = 1.00 (95% 0:0.97,1.03)
Reference: Min et al. (2008,1919011
Period of Study: Dec 2003-Jan 2004
Location: Taein Isalnd, South Korea
Outcome: Heart Rate Variability
Age Mean (SD): 44.3 (21.9)
Study Design: Panel
N: 1.349 participants
Statistical Analyses: Linear
Regression
Covariates: Age, sex, BMI, smoking
Dose-response Investigated? No
Statistical Package: SAS, R
Lags Considered: 0-72 h
Pollutant: PM,0
Averaging Time: 1 h
Mean (SD): 33.244 (19.017)
Percentiles:
25th: 18.000
50th: 26.000
75th: 41.000
Range: 187.000.16.000
Monitoring Stations: 1
Copollutant: N02, S02
PM Increment: 1 SD (19 pg/rri)
Percent Change: [Lower Cl, Upper
Cl]:
SDNN
6-h avg: -4.34 (-7.99, -0.55 **
9-havg:-5.48 (-9.61,-1.17 **hA
12-h avg:-6.23 (-10.47,-1.79)***
24-h-avg: -4.73 (-9.73, 0.56)-
48-h avg:-1.25 (-5.59, 3.29)
72-h avg: -0.85 (-5.35, 3.86)
LF
6-h avg:-10.32
9-havg:-13.79
-18.05,-1.86)*1
-22.26, -4.39)*1
12-h avg:-14.48 (-23.18, -4.80)*1
24-h-avg:-13.15 (-23.36,-1.57)*'
48-h avg:-0.10 (-9.99,10.87)
72-h avg:-7.61 (-17.04, 2.88)
HF
6-h avg:-1.07 (-10.43, 9.28)
9-havg:-3.28 (-13.72, 8.43)
12-h avg:-4.06 (-14.77, 8.00)
24-h-avg:-1.22 (-13.96,13.41)
48-h avg: -3.55
72-h avg: -3.88
-14.01,8.18
-14.64, 8.23
Notes: Percent change in HRV for air
pollution children, adults, and the
elderly (Fig 2)
Percent change in HRV for PMi0
exposure in all ages (Fig 3)
December 2009
E-11
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Peters et al. (2009,
1919921
Period of Study: May 2003-Jul 2004
Location:
Helsinki, Finland
Ausburg, Germany
Barcelona, Spain
Rome, Italy
Stokholm, Sweeden
Outcome: Plasma Fibrinogen
Age Groups: 37-81
Study Design: Panel
N: 854 adults
Statistical Analyses: Additive Mixed
Models
Covariates: Age, sex, BMI, city,
HDL/total cholesterol, smoking,
HbA1c, NT-proBNP, history of
arrhythmia, asthma, arthrosis, stroke,
bronchitis, season, apparent
temperature, relative humidity,
weekday, hour of visit
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0- to 5-day avg
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 30.3
Min:0
Max: 194
Monitoring Stations: NR
Copollutant: PM25, PM10-25
Co-pollutant Correlation: NR
PM Increment: 13.5 pg/m
Change (Lower Cl, Upper Cl):
Genotype 1 1
rs2070006:1.22 (0.47, 1.96)
rs2070011:1.16 (0.41, 1.90)
rs1800790: 0.27 (-0.36, 0.91)
rs2227399: 0.27 (-0.36, 0.91)
rs6056: 0.19 (-0.45, 0.83
rs4220: 0.19 (-0.45, 0.83
Haplotype in cluster 2: 0.09 (-0.53,
0.76)
rs1800791: 0.18 (0.21, 1.40)
Genotype 1 2
rs2070006:0.5(-0.19, 2.15)
rs2070011: 0.42 (-0.28, 1.13)
rs1800790:1.28 (0.54, 2.01)
rs2227399:1.28 (0.55, 2.02)
rs6056:1.26 (0.49, 2.04
rs4220:1.27 (0.49, 2.04
Haplotype in cluster 2:1.17(0.35,1.S
rs1800791: 0.40 (-0.48, 1.28)
Genotype 2 2
rs2070006:0.11
rs2070011:0.08
-1.94,2.15)
-2.08, 2.24)
rs1800790: 2.15 (0.71, 3.60)
52227399:2.18(0.73,3.63)
S6056: 2.24 (0.72, 3.77)
S4220: 2.25 (0.73, 3.78)
Haplotype in cluster 2: 2.16 (0.61, 3.71)
rs1800791:-0.13 (-1.84, 1.58)
Reference: Rosenlund et al. (2007,
1146791
Period of Study: 1985-1996
Location: Stockholm County
Outcome: Myocardial Infarction
Age Groups: 15-79yr
Study Design: Case-control
N: 24,387 first event of myocardial
infarction cases and 276,926 population
based controls
Statistical Analyses: Logistic
Regression
Covariates: Age, sex, calendar yr, SES
Dose-response Investigated? No
Statistical Package: Stata
Lags Considered: 5 yr
Pollutant: PM,0
Averaging Time: 5 yr
Median: 2.4
6th-96th: 0.3-6.2
Median: 2.2
5th-95th: 0.3-6.0
Monitoring Stations: NR
Copollutant: N02, CO
Co-pollutant Correlation: HNR
PM Increment:
5th-95th percentile (5|jg/m3)
Odds Ratio (Lower Cl, Upper Cl):
All Subjects
Controls: 1.0
All Cases: 1.04 (1.00,1.09)
Nonfatal Cases: 0.98 (0.963,1.03)
Fatal Cases: 1.16 (1.09,1.24)
In-hospital death: 1.05 (0.95,1.17)
Out-of-hospital death: 1.23(1.14,1.33)
Subjects who did not move b/t
population censuses
Controls: 1.0
All Cases: 1.11 (1.02,1.21)
Nonfatal Cases: 1.05 (0.96,1.15)
Fatal Cases: 1.56 (1.28,1.91)
In-hospital death: 1.58 (1.13, 2.19)
Out-of-hospital death: 1.56(1.22,1.98)
December 2009
E-12
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ruckerl et al. (2007,
1569311
Outcome: lnterleukin-6 (IL-6),
fibrinogen, C-reactive protein (CRP)
Period of Study: May 2003-Jul 2004 Age Groups: 35-80 yr
Location: Athens, Augsburg,
Barcelona, Helsinki, Rome, and
Stockholm
Study Design: Repeated measures/
longitudinal
N: 1003 Ml survivors
Statistical Analyses: Mixed-effect
models
Covariates: City-specific confounders
(age, sex, BMI) long-term time trend
and apparent temperature RH, time of
day, day of week included if adjustment
improved model fit
Season: Long-term time trend
Dose-response Investigated? Used p-
splines to allow for nonparametric
exposure-response functions
Statistical Package: SASv9.1
Pollutant: PM,0
Averaging Time: Hourly and 24 h (lag
0-4, mean of lags 0-4, mean of lags 0-1,
mean of lags 2-3, means of lags 0-3)
Mean (SD): Presented by city only
Percentiles: NR
Range (Min, Max): NR
Monitoring Stations: Central
monitoring sites in each city
Copollutant:
S02
03
NO
NO,
PM Increment: IQR
Effect Estimate [Lower Cl, Upper Cl]:
% change in mean blood markers per
increase in IQR increase of air pollutant.
IL-6: Lag (IQR): % change in GM
(95%CI)
Lag 0(17.4):-0.34 (-1.66, 0.99)
Lag1
Lag 2
17.4):-0.69
17.4):-1.59
-1.95,0.58
-3.99, 0.88
5-day avg (13.5):-0.87 (-2.28, 0.55)
Fibrinogen: Lag (IQR): % change in
AM (95%CI)
Lag 0(17.4): 0.06 (-0.43, 0.55)
Lag 1(17.4): 0.14 (-0.35, 0.63)
Lag 2 (17.4): 0.24 (-0.24, 0.72)
5-day avg (13.5): 0.60 (0.10,1.09)
CRP: Lag (IQR): % change in GM
(95%CI)
Lag 0(17.4):-0.71 (-2.75, 1.37)
Lag1
Lag 2
17.4):-0.63
17.4):-1.42
-2.61, 1.39
-4.23, 1.47
5-day avg (13.5):-1.35 (-3.45, 0.79)
December 2009
E-13
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ruckerl et al. (2006,
0887541
Period of Study: Oct 2000-Apr 2001
Location: Erfurt, Germany
Outcome: C-reactive protein (CRP)
serum amyloid A (SAA)
E-selectin
vWF
intercellular adhesion molecule-1
(ICAM-1)
fibrinogen
Factor VII
prothrombin fragment 1+2
D-dimer
Age Groups: 50+ yr
Study Design: Panel (12 repeated
measures at 2-wk intervals)
N: 57 male subjects with coronary
disease
Statistical Analyses: Fixed effects
linear and logistic regression models
Covariates: Models adjusted for differ-
ent factors based on health endpoint
CRP: RH, temperature, trend, ID
ICAM-1: temperature, trend, ID
vWF: air pressure, RH, temperature,
trend, ID
FVII: air pressure, RH, temperature,
trend, ID, weekday
Season: Time trend as covariate
Dose-response Investigated? Sensi-
tivity analyses examined nonlinear
exposure-response functions
Statistical Package: SAS v8.2 and
S-Plusv6.0
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 20.0 (13.0)
Percentiles:
25th: 10.8
50th: 15.6
75th: 26.0
Range (Min, Max): 5.4, 74.5
Monitoring Stations: 1 site
Copollutant:
UFPs
AP
PM25
PM10
OC
EC
N02
CO
PM Increment: IQR (15.2
5-day avg: 12.8)
Effect Estimate [Lower Cl, Upper Cl]:
Effects of air pollution on blood markers
presented as OR (95%CI) for an
increase in the blood marker above the
90th percentile per increase in IQR air
pollutant.
CRP: Time before draw: 0-23 h: 1.2
(0.8,1.9)
24-47 h: 2.0 (1.1, 3.6
48-71 h: 2.2 (1.2, 3.8
5-day mean: 2.0 (1.2, 3.7)
ICAM-1: Time before draw: 0-23 h: 1.3
(0.9, 1.8)
24-47 h: 3.1 (2.0, 4.8)
48-71 h: 3.4 (2.2, 5.2)
5-day mean: 3.4 (2.2, 5.3)
Effects of air pollution on blood markers
presented as % change from the
mean/GM in the blood marker per
increase in IQR air pollutant.
vWF: Time before draw: 0-23 h: 4.0
(-0.6, 8.5)
24-47 h: 6.0 (0.6, 11.5)
48-71 h:1.1 (-4.9,7.0)
5-day mean: 6.1 (-0.6,12.8)
FVII: Time before draw: 0-23 h: -6.6
(-10.4--2.5)
24-47 h:-8.4 (-12.3--4.3)
48-71 h: -5.9 (-9.6, -2.0)
5-day mean:-8.0 (-12.4,-3.4)
Note: Summary of results presented in
figures. SAA results indicate increases
in association with PM (not as strong
and consistent as with CRP)
No association observed between E-
selectin and PM
An increase in prothrombin fragment
1+2 was consistently observed,
particularly with lag 4
Fibrinogen results revealed few
significant associations, potentially due
to chance
D-dimer results revealed null
associations in linear and logistic
analyses
December 2009
E-14
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ruckerl et al. (2007,
0913791
Period of Study: Oct 2000-Apr 2001
Location: Erfurt, Germany
Outcome: Soluble CD40 ligand
(SCD40L), platelets, leukocytes,
erythrocytes, hemoglobin
Age Groups: 50+ yr
Study Design: Panel (12 repeated
measures at 2-wk intervals)
N: 57 male subjects with coronary
disease
Statistical Analyses: Fixed effects
linear regression models
Covariates: Long-term time trend,
weekday of the visit, temperature, RH,
barometric pressure
Season: Time trend as covariate
Dose-response Investigated? No
Statistical Package: SAS v8.2 and
S-Plusv6.0
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 20.0 (13.0)
Percentiles:
25th: 10.8
50th: 15.6
75th: 26.0
Range (Min, Max): 5.4, 74.5
Monitoring Stations: 1 site
Copollutant:
UFPs
AP
PM25
PM10
NO
PM Increment: IQR (15.2
5-day avg: 12.8)
Effect Estimate [Lower Cl, Upper Cl]:
Effects of air pollution on blood markers
presented as % change from the
mean/GM in the blood marker per
increase in IQR air pollutant.
SCD40L, % change GM (pg/mL): lagO:
1.6 (-3.5, 7.0)
lag 1:1.1 (-5.4,7.9)
lag 2: -3.5 (-8.9, 2.2)
lag 3:-1.4 (-6.0, 3.4)
5-day mean:-1.2 (-7.8, 5.8)
Platelets, % change mean (103/ul):
lag 0:-0.4 (-1.9,1.0)
lag 1:0.4 (-1.4, 2.3)
lag 2: 0.5 (-1.4, 2.3)
lag 3:-0.1 (-1.6,1.4)
5-day mean: 0.0 (2.1, 0.0)
Leukocytes, % change in mean
(103/ul): lagO:-1.1 (-2.8,0.7)
lag 1:-0.5 (-2.6,1.5)
lag 2: 0.1 (-2.1,2.4)
lag 3:-0.7 (-2.6,1.2)
5-day mean:-1.1 (-3.6,1.4)
Erythrocytes, % change mean
(106/ul): lagO: 0.0 (-0.4, 0.5)
lag 1:-0.4 (-1.0, 0.1)
Reference: Steinvil et al. (2008,
1888931
Period of Study: 2003-2006
Location: Tel-Aviv, Israel
Outcome: Inflammation
Age Groups:
Mean(SD):46(12)yr
Study Design: Panel
N1 ?fi^Q
• \j\j\jy
Statistical Analyses: Linear
Regression
Covariates: Age, waist circumference,
BMI, HDL, OLDL, triglycerides, diastolic
& systolic BP, alcohol consumption,
sports intensity, medications, smoking
status, family history of CHD,
temperature, humidity, precipitation,
season, & yr
Dose-response Investigated? No
Statistical Package: SPSS
Lags Considered: 0-7 days
Pollutant: PM10
Averaging Time: 24 h
Mean (SD):
64 (100.8)
25th: 33.1
60th: 43.0
75th: 60.7
Monitoring Stations: NR
Copollutant: S02, N02, 03, CO
Co-pollutant Correlation:
S02: 0.043
N02: 0.082
03: -0.113
CO: 0.075
lag 2: -0.7 -1.2, -0.2)
lag 3: -0.4 -0.8, 0.0)
5-day mean: -0.6 (-1.2, -0.1)
Hemoglobin, % change mean (g/dl):
lag 0: -0.1 -0.7,0.6)
lag 1: -0.4 -1.2,0.3)
lag 2: -0.7 (-1.3, 0.0)
lag 3: -0.3 (-0.9, 0.2)
5-day mean: -0.7 (-1.5, 0.1)
PM Increment: Interquartile Range
(27.6 pg/m3)
hs-CRP Relative % Change (Lower
Cl, Upper Cl):
Men:
Lag 0: -1 (-2, 1)
Lag1:0(-1, 1);Lag2:-1 (-2, 1)
Lag 3: -1 (-2, 0)
Lag 4:0 (-1,1)
Lag 5:0 (-1,2)
Lag 6: 1 (0, 2)
Lag 7: 1(0,1)
0-7 avg: -2 (-5, 1)
V\fomen:
Lag 0: 0 (-2, 2)
Lag 1:0(-1, 2)
Lag 2: 1 (0, 2)
Lag3:0(-1, 1)
Laq 4' 0 -1 2
i_uy -T. w \ i , f-j
Lag 5:0 (-1,2)
Lag 6: -1 (-3, 1)
Lag 7: 0 (-2, 1)
0-7 avg: 1 (-2, 4)
Fibrinogen Absolute % Change
(Lower Cl, Upper Cl):
Men:
LagO: 0.7(0.0,1. 5); Lag1:0.4(-0.2, 0.9);
Lag2:-0.1(-0.9, 0.6)
Lag3:-0.1(-0.7, 0.6); Lag4: 0.0(-0.7,
0.7);Lag5:0.1(-0.7, 1.0)
Laq6:0.6(-0.1, 1.3); Laq7: 0.4(0.0, 0.8);
December 2009
E-15
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
0-7avg:-0.4(-1.9,1.0)
V\fomen:
LagO: 0.3(-0.6,1.2); Lag1: -0.1 (-0.8,
0.7); Lag2: -0.3(-0.9, 0.3)
Lag3:-0.1(-0.7, 0.5); Lag4: 0.2(-0.4,
0.9); Lag5: 0.2(-0.7,1.2)
Lag6:-0.3(-1.4, 0.8); Lag7:0.7(-0.1,
1.5);0-7avg:0.0(-1.5,1.5)
WBC Absolute Change (Lower Cl,
Upper Cl):
Men:
LagO: 2 (-22, 27)
Lag1:3(-14,19)
Lag2:1 (-22, 24)
Lag3:-7(-28,14)
Lag4: -22 (-44, -1)
Lag5: -20 (-46, 7
Lag6:-5(-27,16
Lag7:-4(-16, 9)
0-7avg:-11(-58, 36)
V\fomen:
Lag 0: 20 (-6, 46)
Reference: Su et al. (2006,1570221
Period of Study: Feb-Apr 2002
Location: Taipei, Taiwan
Outcome: Total cholesterol, HDL,
tryglycerides, LDL, hs-CRP, IL-6, TNF-
a, tPA, PAI-1, and fibrinogen
Age Groups: 40-75 yr
Study Design: Panel study
N: 49 subjects (31 males and 18
females) with coronary heart disease or
multiple risk factors for CHD
Statistical Analysis: Linear mixed
effects regression
Pollutant: PM,0
Averaging Time: 1 h
(High pollution day = PM10 from 08:
00-18: 00 >100)
Copollutant: 0;
PM Increment: High vs.. Low pollution
days
Effect Estimate [Lower Cl, Upper Cl]:
CHD patients (n = 23): P-value for
paired t-test comparing health endpoint
means on high and low pollution days
hs-CRP: p = 0.568
IL-6: p = 0.856
TNF-a: p = 0.246
PAI-1: p = 0.008
tPA: p = 0.322
Fibrinogen: p = 0.189
P-value for health endpoint in mixed-
effects models
PAI-1: p = 0.010
tPA: p = 0.329
Fibrinogen: p = 0.747
Patients with multiple CHD risk
factors (n = 26): P-value for paired t-
test comparing health endpoint means
on high and low pollution days
hs-CRP: p = 0.475
IL-6: p = 0.561
TNF-a: p = 0.572
PAI-1: p = 0.098
tPA: p = 0.260
Fibrinogen: p = 0.087
P-value for health endpoint in mixed-
effects models
PAI-1: p = 0.891
tPA: p = 0.789
Fibrinogen: p = 0.923
Notes: Subjects had paired fasting
blood samples taken during high and
low air pollution days.
December 2009
E-16
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Vedal et al., (2004, 0556301 Outcome: Implantable cardioverter
„..,„.. „„„-,„„„, defibrillator (ICD) discharge
Period of Study: 1997-2000
Location: Vancouver, British Columbia
Age Groups: All
Study Design: Time series
(Retrospective, longitudinal panel study)
N: 50 ICD patients with 1+ discharges
(40,328 person-days and 257
arrhythmia event days)
Statistical Analyses: Multiple logistic
regression with GEE
Covariates: Temperature, relative
humidity, barometric pressure, rainfall,
wind direction and speed
Season: Summer (May-Sep) and winter
(Oct-Apr)
Dose-response Investigated: No
Statistical Package: NR
Lags Considered: -3 day
Pollutant: PM,0
Averaging Time: 24 h
Mean (min-max): 12.9 (3.8-49.3)
SD = 5.6
Monitoring Stations: 8
Copollutant (correlation):
03:r = 0.11
S02:r = 0.70
N02:r = 0.49
CO: r = 0.43
Other variables:
Temp: r = 0.43
Humidity: r = -0.35
Baro Pressures = 0.26
Rain: r = -0.63
Wind: r = -0.53
PM Increment: 5.6 pg/m (SD)
Percent Change [Cl]: Values NR
Notes: The author states that significant
negative associations were found for
ICD discharge with same-day lag, and
also for 3-day lag with more arrhythmia-
prone patients. All other non-significant
percent change estimates are shown in
Fig 3 and 4.
Reference: Vedal et al. (2004, 0556301 Outcome: ICD discharges
PeriodofStudy:1997-2000
Location: Vancouver, British Columbia,
Canada
N: 150 patients w/ICD, 4 yr
Statistical Analysis: Logistic
regression, GEE
Covariates: Temporal trends,
temperature, relative humidity, wind
speed, rain
Season: Summer, winter
Dose-response Investigated? No
Lags Considered: 0, 1, 2, and 3 days
Pollutant: PM,0
Mean: 12.9 (SD = 5.6)
Copollutant): 03, S02, N02, CO
Increment: 1 SD
Effect Estimates, e.g., % change in the
rate of arrhythmia, were presented in
Fig 3. No association with PMiq was
observed while S02 was associated
with an increase in the rate of
arrhythmia among 16 patients with at
least 2 discharges per yr.
December 2009
E-17
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Whitsel et al. (2009,
1919801
Period of Study: 1993-2004
Location: U.S.
Outcome: Heart Rate Variability
Age Groups: 50-79 yr
Study Design: Panel
N: 4,295 women
Statistical Analyses: Random Effects
Model
Covariates: Temperature, humidity
Dose-response Investigated? No
Statistical Package: SUDAAN
Lags Considered: 0
Pollutant: PM,0
Averaging Time: 24 h
Amsterdam
Mean: 20.0
Min:3.8
26th: 10.4
50th: 16.9
76th: 23.9
Max: 82.2
Erfurt
Mean: 23.1
Min:4.5
25th: 10.5
60th: 16.3
76th: 27.4
Max: 118.1
Helsinki
Mean: 12.7
Min:3.1
26th: 8.1
60th: 10.6
76th: 16.0
Max: 39.8
Monitoring Stations: 3
Copollutant: NR
Co-pollutant Correlation: N/A
PM Increment: 10 pg/m
Beta (Lower Cl, Upper Cl):
Supine Position, Amsterdam
Lag 0: -0.06 (-0.95, 0.84)
Lag 1:0.18
Lag 2: 0.93
-0.74, 1.10)
0.01, 1.85)
5-day avg: 0.49 (-0.74,1.72)
Supine Position, Erfurt
Lag 0:-0.36 (-0.83, 0.11)
Lag 1:-0.40 (-0.91, 0.11)
Lag 2:-0.68 (-1.20,-0.17)
5-day avg:-0.68 (-1.44, 0.09)
Supine Position, Helsinki
Lag 0:-0.44 (-2.27, 1.40)
Lag 1:-0.17 (-1.69,1.3.5)
Lag 2:-1.14 (-2.51, 0.23)
5-day avg:-0.59 (-3.08,1.90)
Supine Position, Pooled
Lag 0:-0.30 (-0.71, 0.11)
Lag 1:-0.25 (-0.68, 0.18)
Lag 2:-0.26 (-1.22, 0.70)*
5-day avg:-0.36 (-0.99, 0.27)
Standing Position, Amsterdam
Lag 0:-0.44 (-1.6, 0.72)
Lag 1:-0.61 (-1.8,0.59)
Lag 2: 0.32 (-0.88, 1.51)
5-day avg:-0.55 (-2.15,1.04)
Standing Position, Erfurt
Lag 0:-0.59 (-1.24, 0.06)
Lag 1:-0.70 (-1.42, 0.03)
Lag 2:-0.65 (-1.37, 0.07)
5-day avg:-0.68 (-1.74, 0.39)
Standing Position, Helsinki
Lag 0:1.17 (-1.46, 3.80)
Lag 1:0.01 (-2.17,2.19)
Lag 2:-0.63 (-2.60, 1.34)
5-day avg:-1.96 (-5.51,1.60)
Standing Position, Pooled
Lag 0:-0.48 (-1.03, 0.07)
Lag 1:-0.62 (-1.21,-0.03)
Lag 2:-0.41 (-1.00, 0.17)
5-day avg:-0.72 (-1.57, 0.14)
*p<0.1
Period of Study:
12-wk period b/t Sep 2003-Jul 2004
Location: Chapel Hill, NC
Reference: Yeatts et al. (2007, 0912661 Outcome: Heart Rate Variability
Age Groups: 21-50 yr
Study Design: Panel
N: 12 asthmatics
Statistical Analyses: Linear Mixed
Model
Covariates: Temperature, humidity,
pressure
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 1 day
Pollutant: PM10
Averaging Time: 24 h
Mean (SD): 17.5 (7.8)
Min:1.4
Max: 45.6
Monitoring Stations: 1
Copollutant: PM2.5, PM10.2.5
Co-pollutant Correlation
PM25 = 0.90*
PMio-2.5 = 0.73*
*p < 0.01
PM Increment: 1 pg/m
Beta, SE, p-value (Lower Cl, Upper
Cl): NR
1AII units expressed in ug/m3 unless otherwise specified.
December 2009
E-18
-------
Table E-2. Short-term exposure - cardiovascular morbidity studies: PMio25.
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Chuang et al. (2007,
0910631
Period of Study: Nov 2002-Mar 2003
Location: Taipei, Taiwan
Outcome: Heart Rate Variability
Age Groups: 52-76 yr
Study Design: Panel
N: 10 CHD & 16 Hypertensive Patients
Statistical Analyses: Linear Mixed
Effects Model
Covariates: Age, sex, BMI, time of day,
temperature, humidity, pressure, HRV
Dose-response Investigated? No
Statistical Package: S-PLUS
Lags Considered: 1- to 4-h ma
Pollutant: PM10.2.5
Averaging Time: 1 h
among CHD, among hypertensive
Mean (SD): 16.4 (10.7), 14.0(11.1)
IQR: 14.8, 11.9
Min: 0.7, 0.3
Max: 59.6, 66.5
Monitoring Stations: 1 personal
monitor each
Copollutant: PM,
, PM0
Co-pollutant Correlation:
NR
PM Increment: Interquartile range
Percent Change (Lower Cl, Upper
Cl):
Cardiac Patients- SDNN
1h moving:-1.73 (-3.53, 0.08)
2h moving:-1.97 (-4.43, 0.49)
3h moving:-1.70 (-4.39, 0.89)
4h moving:-1.75 (-5.42,1.92)
Cardiac Patients- r-MSSD
1h moving:-4.39 (-9.54, 0.03)
2h moving: -4.36 (-8.99, 0.27)
3h moving:-4.20 (-9.02, 0.61)
4h moving: -2.70 (-9.24, 3.84)
Cardiac Patients- LF
1h moving:-1.85 (-4.33, 0.62)
2h moving: -3.87 (-8.22, 0.47
3h moving: -2.98 (-6.65, 0.69)
4h moving:-3.11 (-8.22,1.99)
Cardiac Patients- HF
1h moving:-4.46 (-9.23, 0.32)
2h moving: -4.41 (-9.55, 0.72)
3h moving:-3.80 (-9.12,1.53)
4h moving:-3.39 (-10.62, 3.84)
Cardiac Patients- LF: HF ratio
1h moving: 8.45 (-3.48, 20.38)
2h moving: 1.66 (-15.22,18.55)
3h moving: 11.69 (-7.27, 30.64)
4h moving: 8.18 (-17.22, 33.57)
Hypertensive Patients- SDNN
1h moving:-2.64 (-3.93, 0.55
2h moving: -3.51 (-7.87, 0.85)
3h moving: -2.74 (-6.22, 0.74)
4h moving:-2.49 (-6.13,1.15)
Hypertensive Patients- r-MSSD
1h moving:-2.53 (-5.10, 0.04)
2h moving:-5.42 (-10.92, 0.09)
3h moving:-3.15 (-6.32, 0.03)
4h moving: -4.23 (-8.88, 0.42)
Hypertensive Patients- LF
1h moving:-4.38 (-8.78, 0.03)
2h moving:-5.23 (-10.95, 0.05)
3h moving:-3.34 (-1.72, 0.04)
4h moving:-2.96 (-6.63, 0.71)
Hypertensive Patients- HF
1h moving:-4.92 (-9.94, 0.10)
2h moving:-6.07 (-12.28, 0.13)
3h moving:-1.94 (-5.44,1.55)
4h moving:-2.78 (-6.78,1.21)
Hypertensive Patients- LF: HF ratio
1h moving: 5.94 (-3.27,15.15)
2h moving: 10.70 (-2.19, 23.59)
3h moving:-1.51 (-17.02,14.00)
4h moving: 3.41 (-16.91, 23.74)
*p < 0.05
December 2009
E-19
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ebelt et al. (2005, 0569071
Period of Study: Summer of 1998
Location: Vancouver, Canada
Outcome: CVD
Age Groups: range from 54-86 yr
mean age= 74 yr
Study Design: extended analysis of a
repeated-measures panel study
N: 16 persons with COPD
Statistical Analyses:
Earlier analysis expanded by
developing mixed-effect regression
models and by evaluating additional
exposure indicators
Dose-response Investigated? No
Statistical Package: SASV8
Pollutant: PM10.2.5
Averaging Time: 24 h
Mean (SD):
Ambient PM10.2.5: 5.6 (3.0)
Exposure to ambient PM10.25: 2.4 (1.7)
Range (Min, Max):
Ambient PMio-25: (-1.2-11.9)
Exposure to ambient PM10.25: (-0.4-7.2)
Monitoring Stations: 5
Copollutant (correlation):
Ambient concentrations and exposure
to ambient PM were highly correlated
for each respective metric: r > 0.71
Note: Total personal fine particle
exposure (T) were dominated by
exposures to non ambient particles
which were not correlated with ambient
fine particle exposure (A) or ambient
concentrations (C). Results for each of
these metrics are listed.
PM Increment:
Increment: C10-2.5: IQR = 4.5 pg/m3
SBP (mm Hg):-2.12 (-5.07-0.82)
DBP (mm Hg): -0.92 (-3.37-0.36)
Ln-SVE (bph): 0.06 (-0.24-0.36)
HR(bpm): 1.09 (-0.69-2.86)
SDNN(ms): 2.64 (-2.85-8.13)
R-MSSD (ms): -0.33 (-4.49-3.82)
Increment: A10-2.5: IQR = 2.4 pg/m3
SBP (mm Hg):-2.55 (-6.15-1.05)
DBP (mm Hg):-0.75 (-3.50-2.01)
Ln-SVE (bph): 0.26 (-0.07-0.58)
HR(bpm): 1.04 (-0.95-3.03)
SDNN (ms): 0.68 (-3.07-4.42)
R-MSSD (ms): 1.10 (-3.08-5.28)
Reference: Lipsett et al. (2006,
0887531
Period of Study: Feb-May 2000
Location: Coachella Valley, CA
Outcome: HRV parameters, specifically Pollutant: PM10.25
SDNN, SDANN, r-MSSD, LF, HF, total
power, triangular index (TRII). Averaging Time: 2 h
Study Design: Panel study Monitoring Stations: 2
N: 19 non-smoking adults with coronary Copollutant: 03
artery disease
Statistical Analysis: Mixed linear
regression models with random effects
parameters
PM Increment: SE*1000
Effect Estimate (change in HRV per
unit increase in PM concentration):
SDNN: -0.72 msec (SE = 0.296)
Notes: PM10.2.5 calculated by
subtracting PM25 concentration from
PM10 concentration. Weekly ambulatory
24-h ECG recordings (once per wk for
up to 12 wk), using Holter monitors,
were made. Subjects' residences were
within 5 mi of 1 of 2 PM monitoring
sites. Regressed HRV parameters
against 18: 00-20: 00 mean particulate
pollution
Reference: Metzger et al. (2007,
0928561
Period of Study: Aug 1998-Dec 2002
Location: Atlanta, GA
Outcome: Days with any event
recorded by the ICD, days with ICD
shocks/defibrillation and days with
either cardiac pacing or defoliation
Study Design: Repeated measures
N: 884 subjects between 1993 and
2002
Statistical Analysis: Logistic
regression with GEE to account for
residual autocorrelation within subjects
Pollutant: PMio.25 (n/cm3)
Averaging Time: 24 h
Mean (SD): 9.6 (5.4)
Median: 8.7
Copollutant:
03, N02, CO, S02, oxygenated
hydrocarbons
PM Increment: OR (96% Cl):
OR = 1.03 (95% Cl: 1.00, 1.07)
Reference: Pekkanen et al. (2002,
0350501
Period of Study: Winter 1998-1999
Location: Helsinki, Finland
Outcome: ST Segment Depression
(>0.1mV)
Study Design: Panel of ULTRA Study
participants
N: 45 subjects, 342 biweekly
submaximal exercise tests, 72 exercise
induced ST Segment Depressions
Statistical Analysis: Logistic
regression / GAM
Pollutant: PM10.25 (n/cm3)
Averaging Time: 24 h
Median: 4.8
IQR: 5.5
Monitoring Stations: 1
Copollutant: N02, CO, PM25, PM,,
ACP, ultrafme
PM Increment: IQR
Effect Estimate(s): PMio.25: OR = 1.99
(0.70, 5.67), lag 2
Notes: The effect was strongest for
ACP and PM25, which in 2 pollutant
models appeared independent.
Increases in N02 and CO were also
associated with increased risk of ST
segment depression, but not with
coarse particles.
Reference: Timonen et al. (2006,
0887471
Period of Study: 1998-1999
Location Amsterdam, Netherlands
Erfurt, Germany
Helsinki, Finland
Outcome: HRV measurements: [LF,
HF, LFHFR, NN interval, SDNN, r-
MSSD]
Study Design: Panel study
N: 131 elderly subjects with stable
coronary heart disease
Statistical Analysis: Linear mixed
models
Pollutant: PMio.25
Means:
Amsterdam: 15.3
Erfurt: 3.7
Helsinki: 6.7
Copollutant: N02, CO
PM Increment: 10 pg/m
Effect Estimate: SDNN
0.69ms (95% Cl:-1.24, 2.63)
HF:2.9%(95%CI:-7.3, 13.1)
LFHFR:-3.3 (95% Cl:-12.7, 6.1)
Notes: Followed for 6 mo with biweekly
clinic visits
2-day lag. ULTRA Study
December 2009
E-20
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Period of Study:
12-wk period b/t Sep 2003-Jul 2004
Location: Chapel Hill, NC
Reference: Yeatts et al. (2007, 0912661 Outcome: Heart Rate Variability
Age Groups: 21-50 yr
Study Design: Panel
N: 12 asthmatics
Statistical Analyses: Linear Mixed
Model
Covariates: Temperature, humidity,
pressure
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 1 day
Pollutant: PM10.2.5
Averaging Time: 24 h
Mean (SD): 5.3 (2.8)
Min:0
Max: 14.6
Monitoring Stations: 1
Copollutant: PM25, PM10
Co-pollutant Correlation:
PM25 = 0.46*
PM10 = NR
*p < 0.01
PM Increment: 1 pg/m .
Beta, SE (Lower Cl, Upper Cl), p-
value
HRV
Max Heart Rate:-1.95, 0.88 (-3.67, -
0.23), 0.03
ASDNN5: -0.77, 0.37 (-1.580, -0.04),
0.05
SDANN5:-3.76, 1.53 (-6.76, -0.76),
0.02
SDNN24HR(mesc): -3.36, 1.38 (-6.06, -
0.65), 0.02
rMSSD:-0.75, 0.53 (-1.79, 0.28), 0.16
pNN50_24hr: -0.50, 0.27 (-1.03, 0.03),
0.07
pNN50_7min: -1.88, 0.55 (-2.95, -0.81),
0.07
Low-frequency power: -0.19, 0.42 (-
1.01, 0.63), 0.65
Percent low frequency: 0.57,1.08 (-
1.55, 2.69), 0.60
High-frequency power: -0.46, 0.17 (-
0.79,-0.14), 0.01
Percent high frequency: -2.14, 0.94 (-
3.98, -0.30), 0.03
Blood Lipids
Triglycerides: 4.78, 2.02(0.81,8.74),
0.02
VLDL: 1.15, 0.44 (0.29, 2.02), 0.01
Total cholesterol: 0.78, 0.54 (-0.28,
1.84), 0.15
Hematologic Factors
Circulating eosinophils: 0.16, 0.06
(0.04, 0.28), 0.01
Platelets:-1.71,1.11 (-3.89, 0.47), 0.13
Circulating Proteins
Plasminogen: -0.01, 0.01 (-0.02, 0.00),
0.08
Fibrenogen: -0.04, 0.02 (-0.08, 0.00),
0.07
Von Willibrand factor: -1.23, 0.66 (-2.53,
0.06), 0.07
Factor VII:-0.90, 0.85 (-2.58, 0.77),
0.29
1AII units expressed in ug/m3 unless otherwise specified.
Table E-3. Short-term exposure - cardiovascular morbidity studies: PIVhs (including PM
components/sources).
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Adar et al. (2007, 0014581
Period of Study: Mar-Jun 2002
Location: St. Louis, Missouri
Outcome: Heart rate variability: heart
rate, standard deviation of all normal-to-
normal intervals (SDNN), square root of
the mean squared difference between
adjacent normal-to-normal intervals
(rMSSD), percentage of adjacent
normal-to-normal intervals that differed
by more than 50 ms (pNNSO), high
frequency power (HF in the range of
0.15-0.4Hz), low frequency power (LF, in
the range of 0.04-0.15Hz), and the ratio
ofLF/HF
Age Groups: > 60 yr
Study Design: Panel (4 planned
repeated measures surrounding bus
Pollutant: PM25 (fjg/nf)
Averaging Time: Measurements
collected over 48 h period surrounding
the bus trip (during which health
endpoints were measured) used to
calculate 5-, 30-, 60-min, 4-h, 24-h ma
Median (IQR):
All: 7.7 (6.8)
Facility: 6.8 (5.1)
Bus: 17.2 (10.3)
Activity: 8.2 (16.1)
Lunch: 11.2 (5.9)
Monitoring Stations: 2 portable carts
Copollutant:
PM Increment: IQR
Effect Estimate [Lower Cl, Upper Cl]:
% change (95%CI) in HRV per IQR in
the 24-h ma of the microenvironmental
pollutant (IQr = 4.5 pg/m )
Single-pollutant models:
SDNN: -5.5 (-6.3, -4.8)
rMSSD:-9.1 (-9.8,-8.4)
pNN50+1:-12.2 (-13.3, -11.1). LF: -
10.8 (-12.3,-9.3)
HF:-15.1 (-16.7,-13.7)
LF/HF:5.1(3.9, 6.4)
H: 1.0 (0.9, 1.2)
Two-pollutant models (with particle
December 2009
E-21
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
trips with a total of 158 person-trips, 35
participating in all 4 trips)
N: 44 participants
Statistical Analyses: Generalized
additive models
Covariates: Subject, weekday, time,
apparent temperature, trip type, activity,
medications, and autoregressive terms
Season: Limited data collection period
Dose-response Investigated? No
Statistical Package: SASv8.02, R
V2.0.1
PM25
BC
Fine particle counts
coarse particle counts
Correlation notes: 24-h mean PM25,
BC, and fine particle count concentra-
tions ranged from 0.80-0.98
r = 0.76-0.97 when limited to time spent
on the bus
r = 0.55-0.86 when comparing bus
concentrations to 24-h ma
r = -0.003-0.51 when comparing 5-min
avg and 24-h ma
Poor correlations found between coarse
particle count concentrations and all fine
particulate measures during all times
periods
number count coarse):
SDNN: -5.7 (-6.5, -4.9)
rMSSD:-9.4(-10.1,-8.6)
pNN50+1:-13.1(-14.3,-11.9).
LF:-10.7(-12.4,-9.1)
HF: -14.9(-16.5, -13.3); LF/HF: 4.9 (3.6,
6.2)'H: 0.9 (0.7,1.1)
Independent short- and medium-term
associations with HRV across all time
periods
% change per IQR (95%CI)
IQR 5-min means = 6.8 pg/m3 and 23:
55-hmeans = 4.2|jg/m3
SDNN: 5-min mean:-0.5 (-0.8,-0.1)
23: 55-h mean: -4.6 (-5.3, -4.0)
rMSSD: 5-min mean: -0.9 (-1.3, -0.5)
23: 55-h mean:-7.5 (-8.1 to-6.8)
pNNSO + 1
5-min mean:-1.1 (-1.7 to-0.5)
23: 55-h mean:-9.9 (-10.9 to-8.9). LF
5-min mean: 0.4 (-0.5,1.2)
23: 55-h mean:-10.0 (-11.4 to-8.6)
HF
5-min mean:-1.5 (-2.3 to-0.6)
23: 55-h mean:-12.9 (-14.2 to-11.5)
LF/HF
5-min mean: 1.9(1.3, 2.4)
23: 55-h mean: 3.2 (2.1, 4.3)
H: 5-min mean: 0.1 (0.1,0.2)
23: 55-h mean: 0.8 (0.7, 0.9)
Independent associations of short-term
avg (5-min means) of PM with HRV by
bus and nonbus periods
IQR for bus = 10|jg/m3)and
nonbus = 5.6 pg/m )
% change (95%CI)
p-value of interaction
SDNN
Bus:-5.0 (-6.3 to-3.7)
Nonbus:-0.5 (-0.9 to-0.2)
p-value for interaction: O.0001. rMSSD
Bus:-4.8 (-6.2 to-3.5)
Nonbus: -0.7 (-1.1 to -0.4. p-value for
interaction: <0.0001
pNNSO + 1
Bus:-6.3 (-8.4 to-4.2)
Nonbus:-0.8 (-1.4 to-0.3)
p-value for interaction: <0.0001
LF: Bus: -7.0 (-9.8 to -4.1) Nonbus: 0.6
(-0.1, 1.4)
p-value for interaction: O.0001. HF:
Bus: -10.7 (-13.5to -7.9)' Nonbus: -0.7
(-1.5, 0.04) p-value for interaction:
O.0001. LF/HF: Bus: 3.9 (1.7, 6.0)
Nonbus: 1.4 (0.8,1.9)
p-value for interaction: 0.39. H: Bus: 0.7
(0.5,1.0)
Nonbus:-0.01 (-0.08,0.1)
p-value for interaction: <0.0001
Note: Exposure to health associations
by all lag periods presented in Fig 2
(magnitude of associations increased
with averaging period, with the largest
associations consistently found for 24-h
December 2009
E-22
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Adar et al. (2007, 0014581
Period of Study: Mar-Jun 2002
Location: St. Louis, Missouri
Outcome: Heart rate variability: heart
rate, standard deviation of all normal-to-
normal intervals (SDNN), square root of
the mean squared difference between
adjacent normal-to-normal intervals
(rMSSD), percentage of adjacent
normal-to-normal intervals that differed
by more than 50 ms (pNNSO), high
frequency power (HF
intherangeofO.15-0.4Hz), low
frequency power (LF, in the range of
0.04-0.15Hz), and the ratio of LF/HF
Age Groups: 2 60 yr
Study Design: Panel (4 planned
repeated measures with a total of 158
person-trips
35 participating in all 4 trips)
N: 44 participants
Statistical Analyses: Generalized
additive models
Covariates: Subject, weekday, time,
apparent temperature, trip type, activity,
medications, and autoregressive terms
Season: Limited data collection period
Dose-response Investigated? No
Statistical Package: SASv8.02, R
V2.0.1
Pollutant: BC (ng/rri)
Averaging Time: Measurements
collected over 48 h period surrounding
the bus trip (during which health
endpoints were measured) used to
calculate 5-, 30-, 60-min, 4-h, 24-h ma
Median (IQR): All: 330 (337)
Facility: 285 (270)
Bus: 2911(2464)
Activity: 482 (1168)
Lunch: 434 (276)
Monitoring Stations: 2 portable carts
Copollutant: PM25
BC
Fine particle counts
Coarse particle counts
Correlation notes: 24-h mean PM25,
BC, and fine particle count
concentrations ranged from 0.80 to 0.98
r = 0.76 to 0.97 when limited to time
spent on the bus
r = 0.55 to 0.86 when comparing bus
concentrations to 24-h ma
r = -0.003 to 0.51 when comparing 5-min
avg and 24-h ma
Poor correlations found between coarse
particle count concentrations and all fine
particulate measures during all times
periods
PM Increment: IQR
Effect Estimate [Lower Cl, Upper Cl]:
% change (95%CI) in HRV per IQR in
the 24-h ma of the microenvironmental
pollutant (IQr = 459 ng/m3)
Single-pollutant models
SDNN:-5.3 (-6.5 to-4.1)
rMSSD:-10.7 (-11.9 to-9.5)
pNN50+1:-13.2 (-15.0 to-11.4)
LF:-11.3(-13.7to-8.8)
HF: -18.8 (-21.1 to-16.5)
LF/HF: 9.3 (7.2, 11.4)
H: 1.0 (0.8, 1.3)
Independent short- and medium-term
associations with HRV across all time
periods
% change per IQR (95%CI)
IQR 5-min means = 337 ng/m3 and 23:
55-h means = 490 ng/m3)
SDNN: 5-min mean: -0.3 (-O.Sto -0.1)
23: 55-h mean:-4.7 (-5.9 to-3.5)
rMSSD: 5-min mean: -0.3 (-O.Sto -0.1)
23: 55-h mean:-9.3 (-10.5 to-8.1)
pNN50+ 1: 5-min mean:-0.3 (-0.6 to -
0.1)
23: 55-h mean:-10.5 (-12.3 to-8.7)
LF: 5-min mean: -0.5 (-0.9 to -0.1)
23: 55-h mean:-9.8 (-12.4 to-7.2)
HF: 5-min mean: -0.9 (-1.2 to -0.5)
23: 55-h mean:-15.4 (-17.8 to-12.9)
LF/HF: 5-min mean: 0.3 (0.1, 0.6)
23: 55-h mean: 6.5 (4.5, 8.6)
H: 5-min mean: 0.1 (0.1, 0.2)
23: 55-h mean: 0.4 (0.2, 0.7)
Independent associations of short-term
avg (5-min means) of PM with HRV by
bus and nonbus periods
IQR for bus = 2.6 pg/m3) and
nonbus = 0.27 pg/m )
% change (95%CI)
p-value of interaction
SDNN: Bus: -4.6 (-6.1 to -3.0)' Nonbus: -
0.1 (-0.3,0.1)
p-value for interaction: <0.0001
rMSSD: Bus: -2.6 (-4.2 to -0.9): Nonbus:
-0.3 (-0.5 to-0.1)
p-value for interaction: 0.64
pNN50+1:Bus:-2.0(-4.5, 0.5):
Nonbus:-0.5 (-0.8 to-0.1)
p-value for interaction: 0.34
LF: Bus: -6.0 (-9.3 to -2.5): Nonbus: -0.2
(-0.7, 0.3)
p-value for interaction: 0.028
HF: Bus: -5.8 (-9.1 to-2.3)
Nonbus:-0.9 (-1.4 to-0.4)
p-value for interaction: 0.50
LF/HF: Bus:-0.8 (-3.1, 1.7)
Nonbus: 0.8 (0.5,1.1)
p-value for interaction: <0.0001
H: Bus:-0.5 (-0.8 to-0.2)
Nonbus: 0.3 (0.26, 0.34)
p-value for interaction: <0.0001
Note: Exposure to health associations
by all lag periods presented in Fig 2
(magnitude of associations increased
with averaging period, with the largest
associations consistently found for 24-h
ma)
Reference: Auchincloss et al.
1562341
Outcome: Blood pressure: Systolic Pollutant: PM25
(SBP), diastolic (DBP), mean arterial
PM Increment: 10 pg/m (approx.
equivalent to difference between 90th
December 2009
E-23
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Period of Study: Jul 2000-Aug 2002 Pulse Pressure
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Note: Supplementary material available
on-line shows results for DBP and MAP,
among others
Reference: Baccarelli et al. (2009,
1881831
Period of Study: Nov 2000-Jun 2005
Location: Boston, Mass
Outcome: Heart rate variability
Age Groups: Elderly
Study Design: Panel
N:549 men
Statistical Analyses: Mixed-effects
model
Covariates: Age, past/current CHD,
BMI, mean arterial pressure, fasting
blood glucose, smoking, alcohol
consumption, use of beta-blockers, CA
channel blockers, angiotensin-
converting enzyme inhibitors, room
temperature, season, apparent
temperature
Season: No
Dose-response Investigated? No
Statistical Package: SAS
Pollutant: PM25
Averaging Time: 48-h ma
Geometric Mean (96%CI):
All Visits: 10.5 (10.0,10.9)
Visits w/Genotype Data: 10.4(9.9,11.0)
Visits w/o Genotype Data: 10.5 (9.8,
11.4)
Monitoring Stations: 1
Copollutant: NR
Correlation: N/A
PM Increment: 10 pg/m
Percent Change [Lower Cl, Upper Cl],
P:
All Subjects w/ Genotype Data
SDNN:-6.0 (-13.5, 2.0), 0.14
HF:-17.1 (-32.3, 1.6), 0.07
LF:-8.2 (-22.1, 8.2), 0.31
All Subjects
SDNN:-7.1 (-13.2, -0.6), 0.03
HF:-18.7 (-31.1,-4.0), 0.01
LF:-11.8 (-23.2, -1.3), 0.08
December 2009
E-25
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Barclay et al. (2007,
1922291
Period of Study: Jan 2003-May 2005
Location: Aberdeen, Scotland
Outcome: Haematological outcomes,
Heart Rhythm outcomes, & Heart Rate
Variability outcomes
Age Groups: 70.4 (8.9)
Study Design: Panel
N: 132 patients w/ chronic heart failure
Statistical Analyses: Linear & Mixed
Effects Regression Model
Covariates: Age, temperature, humidity,
pressure
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: Lags 0-2 day
Pollutant: PM25
Averaging Time: Daily
Mean: 7.454
Min: 1.092
Max: 21.97
Monitoring Stations: 0
Copollutant: PM10, PNC, N02
Co-pollutant Correlation
N02 city: 0.164
NO city: 0.048
PM,o city: 0.476*
N02 personal: 0.169
PNC DEOM: 0.115
PM25 traffic: 0.522*
PNC total: 0.367*
PNC traffic: 0.234
'correlations based on 3-day avg
concentrations
Notes: PM25 values model predicted
PM Increment: NR
Beta (Lower Cl, Upper Cl):
Haemoglobin: -0.509 (-1.560, 0.542)
Mean corpuscular haemoglobin: 0.188 (-
0.481,0.857)
Platelets: 3.022 (0.403, 5.642)
Haematocrit:-0.813 (-1.892, 0.267)
White blood cells: -1.652 (-4.727,1.424)
C reactive protein: 4.924 (-13.022,
22.869)
IL-6:-5.980 (-23.649, 11.690)
von Willebrand factor: 1.363 (-6.561,
9.287)
E-selectin: 2.136 (-2.946, 7.217)
Fibrinogen: -5.579 (-10.403, -0.755)*
Factor VII: 3.747 (-1.959, 9.452)
day-dimer: 5.211 (-2.974,13.397)
All arrhythmias: -7.082 (-28.789,14.626)
Ventricular ectopic beats: -12.203 (-
39.021, 14.615)
Ventricular couplets: -1.255 (-25.678,
23.168)
Ventricular runs: -2.548 (-17.448,
12.351)
Supraventricular ectopic beats: 4.898 (-
19.772, 29.568)
Supraventricular couplets: 6.138 (-
16.242, 28.518)
Supraventricular runs: -0.545 (-17.577,
16.487)
Avg HR: 0.617 (-0.782, 2.016)
24 hSDNN: 3.645 (-0.227, 7.517)
24 h SDANN: 4.437 (0.030, 8.844)*
24 hRMSSD: 0.617 (-0.782, 2.016)
24 h PNN 50%: 11.247 (-6.228, 28.722)
24 hLF power: 4.439 (-6.823, 15.701)
24 hLF normalized:-5.659 (-11.815,
0.497)
24 h HF power: 3.800 (-10.863, 18.464)
24 h HF normalized: -6.597 (-13.724,
0.531)
24 hLF/HF ratio: 1.033 (-8.355, 10.414)
*p < 0.05
Notes: Estimates also available for
PM25 traffic
LF= low frequency
HF= high frequency
Reference: Briet et al. (2007, 0930491
Period of Study: NR
Location: Paris, France
Outcome: Endothelial Function
Age Groups: 20-40 yr
Study Design: Panel
N: 40 white male nonsmokers
Statistical Analyses: Multiple Robust
Regrssion
Covariates: R53R/R53H genotype, diet,
subject factor, visit, temperature
Dose-response Investigated? No
Statistical Package: NCSS
Lags Considered: 0-5 days
Pollutant: PM25
Averaging Time: 24 h
6 day Mean (SD): 28 (6)
Monitoring Stations: NR
Co-pollutant: PM10, S02, NO, N02, CO
Co-pollutant Correlation: N/A
PM Increment: 1 SD
Beta (Lower Cl, Upper Cl), P, R2:
Flow-mediated brachial artery dilation:
-0.32 (-1.10, 0.46), NS, 0.04
Reactive hyperemia:
15.68(7.11,23.30), O.0001, 0.24
Changes in Endothelial function b/t
visits:
1.98(0.67, 3.259), 0.004, 0.44
December 2009
E-26
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Cardenas et al. (2008,
1919001
Period of Study: NR
Location: Mexico City, Mexico
Outcome: Heart Rate Variability
Age Groups: 20-40 yr
Study Design: Panel
N: 54 subjects
Statistical Analyses: Linear GEE
Pollutant: PM25
Averaging Time: NR
26th, 60th, 76th percentile:
Indoor: 14.8, 28.3, 47.9
Outdoor: 6.4, 10.8, 16.8
PM Increment: NR
Mean Difference (Lower Cl, Upper Cl),
lag:
Ln low frequency
Indoors: -0.028 (-0.0423, -0.0138)
Outdoors: -0.194 (-0.4509, 0.0627)
models
Covariates: Localization, supine
position, gender, age, humidity, heart
rate, orthostatic position, head-up tilt test
result
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: NR
Co-pollutant: NR
Co-pollutant Correlation: N/A
Ln high frequency
Indoors:-0.019 (-0.0338,-0.0044)
Outdoors: -0.298 (-0.5553, -0.0401)
Ln LF/HF ratio
Indoors:-0.017 (-0.0330,-0.0007)
Outdoors: -0.278 (-0.5540, 0.0030)
Reference: Cavallari et al. (2007,
1574251
Period of Study: 1999-2006
Location: Massachusetts
Outcome: Heart Rate Variability
Age Groups: 22-63
Study Design: Panel
N: 36 males
Statistical Analyses: Mixed Effects
Regression Model
Covariates: Age, smoking, heart rate at
work
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: Lags 0-14 h
Pollutant: PM25
Averaging Time: Hourly
Mean (SD): 1.12 (0.76)
Min:0.12
Max: 3.99
Monitoring Stations: NR
Copollutant: NR
Co-pollutant Correlation: N/A
PM 1 mg Increment: m
Beta (Lower Cl, Upper Cl):
Model 1
Lag1 h:-1.44 (-7.75, 4.87)
Lag 2 h:-5.33 (-10.97, 0.31)*
Lag 3 h:-6.86 (-11.91,-1.81)}
Lag 4 h:-2.17 (-9.33, 4.99)
Lag 5 h:-4.73 (-11.99, 2.53)
Lag 6 h: -3.52 (-9.89, 2.84)
Lag 7 h:-1.59 (-7.53, 4.35)
Lag 8 h: -0.72 (-7.63, 6.20)
Lag 9 h:-5.55 (-10.65,-0.45)}
Lag 10 h:-3.66 (-8.85, 1.53)
Lag 11 h:-8.60 (-17.45, 0.24)*
Lag 12 h:-5.98 (-14.67, 2.70)
Lag 13 h:-8.27 (-17.00, 0.46)*
Lag 14 h:-4.19 (-12.71, 4.33)
Model 2
Lag1 h: 4.10 (-0.39, 8.60)*
Lag 2 h:-3.21, (-8.78, 2.37)
Lag 3 h:-6.45 (-11.59, -1.31)}:
Lag 4 h: -0.01 (-6.96, 6.94)
Lag 5 h: -2.03 (-8.27, 4.22)
Lag 6 h:-1.99 (-8.46, 4.48)
Lag 7 h: -0.34 (-6.22, 5.54)
Lag 8 h: 0.72 (-6.35, 7.78)
Lag 9 h:-5.26 (-10.62, 0.11)*
Lag 10 h:-3.68 (-9.17,1.80)
Lag 11 h:-9.41 (-18.60,-0.23)}
Lag 12 h:-6.45 (-15.62, 2.72)
Lag 13 h:-7.33
Lag 14 h:-4.75
-16.55, 1.8
-13.81, 4.32
*p<0.05, }p<0.10
Notes: Model 1 adjusted for smoking
status and age only. Model 2 adjusted
for smoking status, age, and heart rate
during work.
December 2009
E-27
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Chahine et al. (2007,
1563271
Period of Study: Jan 2000-Jun 2005
Location: Boston, MA
Outcome: Heart Rate Variability
Age Groups: Mean 72.8(6.6) yr
Study Design: Panel
N: 539 white males
Statistical Analyses: Mixed Effects
Model
Covariates: Age, BMI, mean arterial
pressure, fasting blood glucose,
smoking, alcohol consumption, use of
beta-blockers, calcium channel blockers,
ACE inhibitors, room temperature,
season, outdoor temperature
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0- to 2-day ma
Pollutant: PM25
Averaging Time: 1 h
Mean (SD): 11.7 (7.8)
Monitoring Stations: 1
Copollutant: PM10
Co-pollutant Correlation: N/A
PM Increment: 10 pg/m
Percent Change (Lower Cl, Upper Cl),
p-value:
loglOSDNN
Total:-6.8 (-12.9, -0.2), 0.0436
GSTM1 wildtype: -2.0 (-11.3, 8.3),
0.6908
GSTM1 null:-10.5 (-18.2, -2.2), 0.0150
HMOX-1 <25 repeats: 7.4 (-8.7, 26.2),
0.3891
HMOX-1 >25 repeats: -8.5 (-14.8, -1.8),
0.0137
loglOHF
Total:-17.3 (-30.0, -2.3), 0.0263
GSTM1 wildtype: -4.0 (-24.8, 22.6),
0.7442
GSTM1 null:-24.2 (-39.2, -5.5), 0.0139
HMOX-1 <25 repeats: 8.9 (-27.1, 62.8),
0.6759
HMOX-1 >25 repeats: -20.1 (-32.9, -5.0),
0.0115
loglOLF
Total:-11.2 (-22.8, 2.2), 0.0986
GSTM1 wildtype:-0.6 (-19.0, 22.0),
0.9545
GSTM1 null: -17.0 (-31.0, -0.2), 0.0478
HMOX-1 <25 repeats: 14.0 (-18.6, 59.5),
0.4465
HMOX-1 >25 repeats: -14.0 (-25.7, -0.5),
0.0430
Reference: Chen and Schwartz (2008,
1901061
Period of Study: 1989-1991
Location: U.S.
Outcome: White Blood Cell count
Age Groups: 20-89 yr
Study Design: Panel
N: 2,978 participants
Statistical Analyses: Mixed Effects
Models
Covariates: Age, sex, race, SES,
smoking, alcohol consumption, MS
abnormalities, indoor air pollutants,
exercise
Dose-response Investigated? No
Statistical Package: Stata
Lags Considered: NR
Pollutant: PM25
Averaging Time: 24 h
Mean (SD): 36.8 (13.0) Median(range)
for
01:23.1(14.6-27.8)
PM Increment: Quartile, lyravg (36.8
pg/m3)
Avg WBC count(SE) by PM quartile:
01:6760
79
02:31.2
03: 38.8
27.9-34.3
34.3-43.3
04: 53.7 (43.3-78.5)
Monitoring Stations: NR
Copollutant: NR
Co-pollutant Correlation: N/A
02:6942
03:
04:7109(61)
Beta(Lower Cl, Upper Cl), p-value:
Crude: 239 (58, 420), 0.01
Model 1:145 (10, 281), 0.035
Model 2:141 (6, 277), 0.041
Model 3:138 (2, 273), 0.046
Model 1: Age, sex, race, SES, smoking,
alcohol consumption, MS abnormalities.
Model 2: Model 1 plus indoor air
pollutants, exercise. Model 3: Clean
areas (01) vs.. other more polluted
areas
December 2009
E-28
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Chuang et al. (Chuang et
al, 2007, 0910631
Period of Study: Between Apr-Jun
2004 or 2005
Location: Taipei, Taiwan
Outcome: High-sensitivity C-reactive
protein (hs-CRP)
Fibrinogen, plasminogen activator
fibrinogen inhibitor-1 (PAI-1), tissue-type
plasminogen activator (tPA), 8-hydroxy-
2'-deoxyguanosine (8-OHdG), and log-
transformed HRV indices
(SDNN = standard deviation of NN
intervals, r-MSSD = square root of the
mean of the sum of the squares of
differences between adjacent NN
intervals, LF = low frequency [0.04-
0.15Hz], and HF = high frequency[0.15-
0.40HZ])
Pollutant: PMi0, nitrate, sulfate
Averaging Time: Hourly data used to
calculate avg over 1- to 3-day periods
Mean (SD):
1-day avg: 31.8 (10.6)
2-day avg: 36.4 (12.6)
3-day avg: 36.5 (12.6)
Range (Min, Max):
1-day avg: 16.2, 50.1
2-day avg: 15.0,53.4
3-day avg: 12.7,59.5
Monitoring Stations: 2 sites (each
pollutant measured at 1 site only)
PMis Increment: IQR
(1-day avg: 20.4
2-day avg: 25.2
3-day avg: 20.0)
Effect Estimate [Lower Cl, Upper Cl]:
% change in health endpoint per
increase in IQR of PM25 (1-3 day
averaging period
single pollutant models)
hs-CRP:
1-day: 90.2
2-day: 99.1
-10.2,190.1
-26.1,224.3
3-day: 100.4 (-2.9, 203.7)
8-OHdG:
«yc wivupa. iu-^u yi
Study Design: Panel (cross-sectional) f^°gUtant: ™10
N: 76 students Nitrate
OC
Statistical Analyses: Linear mixed- EC
effects models N02
CO
Covariates: Age, sex, BMI, weekday, S02
temperature of previous day, relative QS
humidity
Season: Only 1 season of data
collection
Dose-response Investigated? No
Statistical Package: NR
1-day: -5.0 -14.3,4.4)
2-day: -5.5 -15.6,4.6)
3-day: -5.6 (-13.8, 2.6)
PAI-1:
1-day: 20.4 (17.3, 33.5)
2-day: 16.2 1.9,30.5)
3-day: 20.0 18.5,31.5)
tPA:
1-day: 12.0 (-2.4, 26.3)
2-day: 12.0 -2.9, 26.9);
3-day: 12.0 -2.7, 26.6)
Fibrinogen:
1-day: 2.6 (-2.7, 7.8)
2-day: 1.5 (-4. 1,7.1)
3-day: 3.6 (-0.8, 8.1)
Heart Rate Variability
SDNN:
1-day: -4.0 (-6.1 to -1.9)
2-day: -2.5 -4.6 to -0.4)
3-day: -3.0 -5.0 to -1.1)
r-MSSD:
1-day: -3.0 (-8.7, 2.7)
2-day: -2.0 (-8.4, 4.4);
3-day: -3.6 (-8.8, 1.6)
LF:
1 -day: -3.1 (-6.1 to -0.1)
2-day: -3.2 (-4.6, 0.1);
3-day: -3.4 (-6.1 to -0.6)
HF:
1-day: -3.7 -9.4,2.1
2-day: -2.1 -8.4,4.3
;
3-day: -4.0 (-9.3, 1.2)
December 2009
E-29
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Chuang et al. (2007,
0910631
Period of Study: Between Apr-Jun
2004 or 2005
Location: Taipei, Taiwan
Outcome: High-sensitivity C-reactive
protein (hs-CRP)
Fibrinogen, plasminogen activator
fibrinogen inhibitor-1 (PAI-1), tissue-type
plasminogen activator (tPA), 8-hydroxy-
2'-deoxyguanosine (8-OHdG), and log-
transformed HRV indices
(SDNN = standard deviation of NN
intervals, r-MSSD = square root of the
mean of the sum of the squares of
differences between adjacent NN
intervals, LF = low frequency [0.04-
0.15Hz], and HF = high frequency[0.15-
0.40Hzj)
Age Groups: 18-25 yr
Study Design: Panel (cross-sectional)
N: 76 students
Statistical Analyses: Linear mixed-
effects models
Covariates: Age, sex, BMI, weekday,
temperature of previous day, relative
humidity
Season: Only 1 season of data
collection
Dose-response Investigated? No
Statistical Package: NR
Pollutant: Nitrate
Averaging Time: Hourly data used to
calculate avg over 1-3 day periods
Mean (SD): 1-day avg: 4.5 (2.7)
2-day avg: 4.7 (2.4
3-day avg: 4.4 (2.2
Range (Min, Max): 1-day avg: 0.7,10.6
2-day avg: 0.7, 8.9
3-day avg: 0.8, 7.5
Monitoring Stations: 2 sites (each
pollutant measured at 1 site only)
Copollutant: PM10
Sulfate
PM25
OC
EC
N02
CO
S02
03
Nitrate Increment: IQR (1-day avg: 2.5
2-day avg: 4.0
3-day avg: 3.4)
Effect Estimate [Lower Cl, Upper Cl]:
% change in health endpoint per
increase in IQR of nitrate (1-3 day
averaging period
single pollutant models)
hs-CRP: 1-day:-2.1 (-21.9,17.8)
2-day:-11.6 (-58.6, 35.5)
3-day:-18.7 (-69.9, 32.5)
8-OHdG: 1-day: 9.0 (4.0,14.1)
2-day: 15.1 (5.9, 24.3)
3-day: 15.0 (4.9, 25.0)
PAI-1:1-day: 4.0 (-2.5,10.4)
2-day: 11.6 (0.1, 23.1)
3-day: 16.9 (4.3, 29.4)
tPA: 1-day: 2.0 (-6.2,10.3)
2-day: 12.9 (-1.6, 27.5)
3-day: 10.0 (-5.8, 25.8)
Fibrinogen: 1-day: 1.6 (-1.3, 4.5)
2-day: 1.3 (-3.9, 6.5)
3-day: 1.0 (-4.6, 6.6)
Heart Rate Variability
SDNN: 1-day: -1.5 (-2.6 to -0.3)
2-day:-2.6 (-4.7 to-0.5)
3-day:-3.0 (-5.3 to-0.7)
r-MSSD: 1-day:-5.5 (-8.7 to-2.2)
-14.0to-0.2
2-day:-7.1
3-day:-8.1
LF: 1-day:-1.0(-1.6to-0.5)
2-day:-2.0 (-5.6,1.6)
3-day:-2.0 (-5.2,1.2)
HF: 1-day: -2.0 (-5.3,14[potential typo,
possibly 1.4])
2-day:-4.9 (-10.9, 0.9)
3-day:-6.9 (-13.4 to-0.3)
December 2009
E-30
-------
Reference
Reference: Chuang et al. (2007,
0910631
Period of Study: Between Apr-Jun
2004 or 2005
Location: Taipei, Taiwan
Design & Methods Concentrations1 Effect Estimates (95% Cl)
Outcome: High-sensitivity C-reactive Pollutant: Sulfate
Sulfate Increment: IQR
protein (hs-CRP) . (1-day avg: 3.9
Averaging Time: Hourly data used to 2-dav aver 4 3
Fibrinogen, plasminogen activator calculate avg over
fibrinogen inhibitor-1 (PAI-1), tissue-type
plasminogen activator (tPA), 8-hydroxy- Mean (SD): 1-daY
2'-deoxyguanosine (8-OHdG), and log- 2-daY av9: 41 (37
transformed HRV indices 3-daY av9: i9 (3^5
(SDNN = standard deviation of NN R (Mi .. ,
intervals, r-MSSD = square root of the ™"8' „ '"' ""'
1- to 3-day periods 3-day avg: 3.8)
avg: 4. 1 (3.6) Effect Estimate [Lower Cl, Upper Cl]:
% change in health endpoint per
increase in IQR of sulfate (1-3 day
1 H -CIA mo averaging period
-, ~ ' ' ' ' single pollutant models)
mean of the sum of the squares of ^ "^n? 11 *
differences between adjacent NN 3-day avg. 0.4, 11.5 hs.CRP:
intervals, LF = low frequency [0.04- Monitoring Stations: 2 sites (each ^ay: °°'° ?,8k i« >\\
0.15HZJ, and HF = high frequency[0.15- po||utant measured at 1 site only) ^ %} %*• 159.4)
u.tunzj)
Copollutant: PIvl-;
Age Groups: 18-25 yr pM25
Mitrato
Study Design: Panel (cross-sectional) QCrale
N: 76 students EC
N02
Statistical Analyses: Linear mixed- CO
effects models S02
03
Covariates: Age, sex, BMI, weekday,
temperature of previous day, relative
humidity
Season: Only 1 season of data
collection
Dose-response Investigated? No
Statistical Package: NR
8-OHdG:
1-day: 1.0 (0.3, 1.3)
2-day: -0.4 (-5.4, 4.7)
3-day: -0.3 (-4.3, 3.7)
PAI-1:
1-day: 12.0 (5.4, 18.7)
2-day: 13.3 (6.6, 19.9)
3-day: 11. 2 (5.7, 16.6)
tPA:
1-day: 2.0 (-4.6, 8.7)
2-day: 3.8 (-2.8, 10.3)
3-day: 3.0 (-2.3, 8.2)
Fibrinogen:
1-day: 2.9 (0.2, 5.5)
2-day: 2.8 (0.1, 5.5)
3-day: 2.2 (0.4, 4.7)
Heart Rate Variability
SDNN:
1-day: -3.1 (-4.1 to -2.1)
2-day: -4.1 -5.2 to -3.1)
3-day: -2.0 -2.9 to -1.2)
r-MSSD:
1-day: -5.0 (-8.0 to -2.0)
2-day: -6.0 (-8.9 to -2.9)
3-day: -5.7 (-8.2 to -3.2)
LF:
1-day: -3.4 (-4.9 to -1.8)
2-day: -3.0 (-4.5 to -1.5)
3-day: -3.0 (-4.3 to -1.7)
HF:
1-day: -3.5 -6.5 to -0.4)
2-day: -3.9 -7.0 to -0.8)
3-day: -3.0 (-5.5 to -0.5)
December 2009
E-31
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Chuang et al. (2007,
0986291
Period of Study: NR
Location: Boston, MA
Outcome: ST Segment Depression
Age Groups: 43-75 yr
Study Design: Panel
N: 48 coronary artery disease patients
Statistical Analyses: Linear & Mixed
Logistic Regression models
Covariates: Participant, day of week,
order of visit, visit date, hour of day,
hourly temperature
Dose-response Investigated? No
Statistical Package: R
Lags Considered: Lags 1-72 h
Pollutant: PM25
Averaging Time: Hourly
26th, 60th, 76th percentile:
12-havg: 6.18, 8.91,13.18
24-havg: 6.38,9.20, 13.31
Max:
12-havg: 37.13
24-h avg: 40.38
Monitoring Stations: 1
Co-pollutant: BC, CO, 03, N02, S02
Co-pollutant Correlation
BC: 0.56
03: 0.20
N02: 0.38
S02: 0.25
PM Increment: Interquartile Increase
Change (Lower Cl, Upper Cl):
12-hmean
PM25:-0.022 (-0.032,-0.012)
PM25+ N02: -0.023 (-0.034, -0.012)
PM25+S02:-0.009 (-0.02, 0.001)
PM25+BC:-0.011 (-0.023,0.001)
24-h mean
PM25:-0.026 (-0.037,-0.015)
PM25+ N02: -0.017 (-0.029, 0.004)
PM2 5+S02:-0.014 (-0.025,-0.002)
PM25+BC:-0.012 (-0.026, 0.003)
Relative Risk (Lower Cl, Upper Cl):
12-h mean
PM25:1.02 (0.86, 1.21)
PM25+N02: 0.99 (0.82, 1.21)
PM25+S02: 0.87 (0.71, 1.05)
PM25+BC: 0.92 (0.74, 1.14)
24-h mean
PM25:1.22 (0.99, 1.50)
PM25+N02:1.00 (0.80, 1.25)
PM25+S02:1.04(0.83, 1.30)
PM25+BC: 0.87 (0.65, 1.17)
Mean (Lower Cl, Upper Cl):
12-h mean
Myocardial Infarction: -0.042 (-0.057, -
0.026)
No Myocardial Infarction: -0.012 (-0.023,
0.00)
p-for interaction: 0.002
Visit!:-0.102 (-0.12,-0.085)
Visits 2-4: 0.006 (-0.005, 0.017)
p-for interaction: O.001
Diabetic:-0.097 (-0.119,-0.074)
Non-diabetic: -0.009 (-0.019, 0.002)
p-for interaction: O.001
Diurnal daytime pattern: -0.032 (-0.043, -
0.021)
Diurnal nighttime pattern: -0.006 (-0.018,
0.006)
p-for interaction: O.001
24-h mean
Myocardial Infarction: -0.027 (-0.043, -
0.012)
No Myocardial Infarction: -0.025 (-0.038,
0.011)
p-for interaction: 0.787
Visit!:-0.127 (-0.148,-0.105)
Visits 2-4: 0.001 (-0.011,0.013)
p-for interaction: O.001
Diabetic:-0.118 (-0.144,-0.091)
Non-diabetic:-0.13 (-0.024,-0.002)
p-for interaction: O.001
Diurnal daytime pattern: -0.031 (-0.043, -
0.020)
Diurnal nighttime pattern: -0.018 (-0.030,
-0.005)
p-for interaction: 0.233
Notes: The effects of PM on half-h St
segment levels (Fig 1)
December 2009
E-32
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Dales et al. (2007,1557431
Period of Study: NR
Location: Ottawa, Canada
Outcome: Vascular Reactivity
Age Groups: 18-50yr
Study Design: Panel
N: 39 volunteers
Statistical Analyses: Mixed Effects
Model
Covariates: Temperature, humidity,
wind speed, time of day testing was
done, site
Dose-response Investigated? No
Statistical Package: S-PLUS
Lags Considered: NR
Pollutant: PM25
Averaging Time: 2 h
Mean (SD):
Downtown: 40 (20)
Tunney's Pasture: 10 (10)
p-value 0.000
Monitoring Stations: NR
Copollutant: PM1.0
Co-pollutant Correlation
N/A
PM Increment: Interquartile Range
(27.02 pg/m3)
Beta (SE), p-value:
Flow mediated vasodilation (%): -0.016
(0.0072) p=0.03
Heart Rate (beats/min): 0.081 (0.135)
p=0.55
Diastolic blood pressure (mmHg): 0.088
(0.088) p=0.32
Systolic blood pressure (mmHg): -0.108
(0.006) p=0.48
Reference: de Hartog et al. (2009,
1919041
Period of Study: 1998-1999
Location:
Amsterdam, The Netherlands
Erfurt, Germany
and Helsinki, Finland
Outcome: Heart Rate Variability
Age Groups: 50+
Study Design: Panel
N: 122 coronary heart disease patients
Statistical Analyses: Linear Regression
Covariates: Time trend, temperature,
humidity, pressure
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: Lags 0-3 days
Pollutant: PM25
Averaging Time: Daily
p26, p60, p76. p96:
Amsterdam: 10.4,16.7, 23.9, 47.0
Erfurt: 10.8, 16.3, 26.7, 62.3
Helsinki: 8.3,10.6,15.9,25.8
Monitoring Stations: NR
Copollutant: PM <0.1, PMO.1-1.0, N02,
S02
Co-pollutant Correlation
NR
Note: Correlations are provided for
source-specific PM25 & elements
PM Increment: 1 pg/m
Beta (Lower Cl, Upper Cl):
SDNN
Local traffic:-0.12 (-0.36, 0.12)
Long-range transport: -0.04 (-0.14, 0.06)
Oil combustion:-0.29 (-1.04, 0.45)
Industry: 0.03 (-0.12, 0.19)
Crustal:0.11 (-0.35,0.56)
Salt:-0.19 (-1.92, 1.55)
HF
Local traffic: 0.43 (-0.91,1.79)
Long-range transport: 0.19 (-0.38, 0.77)
Oil combustion: 1.05 (-2.70, 4.94)
Industry: 0.62 (-0.34,1.59)
Crustal: 1.57 (-1.28, 4.50)
Salt:-1.43 (-9.86, 7.78)
SDNN
ABS:-0.52 (-1.39, 0.31)
S:-0.51 (-1.36, 0.33)
V:-0.66 (-1.73, 0.41)
Zn: 0.12 (-0.55, 0.79)
Ca: 0.27 (-0.58, 1.11)
Cl: 0.14 (-0.39, 0.67)
Fe: 0.15 (-1.00,1.30)
Cu: -0.08 (-0.74, 0.57)
SDNN
ABS: 2.91 (-2.54, 8.67)
S:0.25 (-4.42, 5.14)
V: 0.73 (-4.74, 6.53)
Zn: 3.85 (-0.26, 8.13)
Ca: 3.39 (-1.80, 8.86)
CM.13 (-1.48, 3.81)
Fe: 6.69 (0.11, 13.69)
Cu: 3.00 (-0.85, 7.00)
Notes: Estimates provided are for all
subjects at lag 1, estimates are also
available at lags 0, 2, and 3, as well as
for subjects w/o beta-blockers at lags
0-3.
December 2009
E-33
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: DeMeo et al. (2004,
0873461
Period of Study: Jul-Aug 1999
Location: Boston, MA
Outcome: Oxygen saturation
Age Groups: 60.4-89.2 yr
Study Design: Cross-sectional study
N: 28 adult participants
Statistical Analyses: GLM, Natural
Spline Smoothing, Regression Analysis,
Random-effects model
Covariates: Mean temperature, Dew
point temperature, Barometric pressure,
Medication use
Season: Summer
Dose-response Investigated? No
Statistical Package:
S-PLUS, SAS
Lags Considered: Hourly lags between
2 and 7 h
Pollutant: PM25
Averaging Time: 6 h, 12 h, 24 h, 48 h
PM Increment: IQR (13.42 pg/rri)
increase
6 h: 13.42 pg/m3
12 h: 10.81 pg/m3
24h:10.26|jg/m3
48:10.57 pg/m3
Overall: 0.172% (-0.313, 0.031)
decrease
6 h: -0.769% (-1.21 to -0.327) decrease
B-blocker users: -0.062% (-0.248, 0.123)
Rest: 6 h:-0.173 (-0.345 to-0.001)
12 h:-0.160 (-0.308 to-0.012)
24 h:-0.169 (-0.316 to-0.022)
48 h:-0.153 (-0.304, 0.002)
Exercise: 6 h:-0.005 (-0.215, 0.205)
12 h:-0.014 (-0.196, 0.168)
24 h: 0.001 (-0.180,0.182)
48 h:-0.011 (-0.196, 0.174)
Post exercise Rest: 6 h: -0.173 (-0.332
to-0.014)
12 h:-0.128 (-0.266, 0.010)
4 h:-0.113 (-0.250, 0.023)
48 h:-0.157 (-295 to-0.019)
Paced breathing: 6 h: -0.142 (-0.292,
0.007)
12 h:-0.139 (-0.269 to-0.010)
24 h:-0.121 (-0.248,0.007)
48 h:-0.082 (0.211, 0.047)
Summary over protocol
6 h:-0.131 (-0.247 to-0.015)
12 h:-0.120 (-0.221, 0.020)
24 h:-0.112 (-0.212 to-0.013)
Notes: Fig of the variation in oxygen
saturation during the first rest period vs..
individual hourly lag measurements for
PM25
Reference: Diez-Roux et al. (2006,
1564001
Period of Study: Baseline data
collected Jun 2000-Aug 2002
Location:
USA
6 field centers:
Baltimore, MD
Chicago, IL
Forsyth Co, NC
Los Angeles, CA
New York, NY
St. Paul, MN
Outcome: C-reactive protein (CRP)
assessed continuously and as a
dichotomous variable (cutpoint, 3 mg/L)
interleukin-6 (IL-6)
Age Groups: 45-84 yr
Study Design: Cross-sectional
N: 5634 persons
Statistical Analyses: Linear regression
& logistic regression
Covariates: Age, sex, race/ethnicity,
general health status, BMI, diabetes,
cigarette status, secondhand smoke,
physical activity, arthritis flare in last 2
wk, medications, infections in last 2 wk
(also ran models including site,
copollutants, and weather)
Season: Examined seasonal patterns in
the residuals of fully adjusted models
stratified by season
Dose-response Investigated? No
Statistical Package: NR
Pollutant: PM25
Averaging Time: Prior day, prior 2 days,
prior wk, prior 30 days, and prior 60
days
Mean (SD): Presented in Fig 1 by site
Percentiles: Presented in Fig 1 by site
Range: NR
Monitoring Stations: NR
Long-term exposure to PM estimated
based on residential history reported
retrospectively
All addresses geocoded
Ambient AP obtained from U.S. EPA
Copollutant:
S02
N02
CO
03
PM Increment: 10 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
Adjusted (all personal-level covariates)
relative difference in CRP (mg/L) per
10 pg/m3 increase in PM25
Prior day: 0.99 (0.96,1.01)
Prior2 days: 0.99 (0.96,1.01)
Prior 7 days: 1.00 (0.96,1.04)
Prior 30 days: 1.03
Prior 60 days: 1.04
0.98,1.10)
0.97,1.11)
Odds Ratios of CRP of > 3 mg/L per
10 pg/m3 increase in PM25 (adjusted for
all personal-level covariates)
Prior day: 0.98 (0.92,1.04)
Prior 2 days: 0.99
Prior 7 days: 1.05
0.93,1.06
0.96, 1.15
Prior 30 days: 1.12 (0.98,1.29)
Prior 60 days: 1.12 (0.96,1.32)
December 2009
E-34
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Dubowsky et al. (2006,
0887501
Period of Study: Mar-Jun 2002
Location: St. Louis, Missouri
Outcome: White blood cells (WBC),
C-reactive protein (CRP), interleukin-6
(IL-6)
Age Groups: > 60 yr
Study Design: Panel (4 planned
repeated measures
n = 35 participated in 4 trips)
N: 44 participants
Statistical Analyses: Linear mixed
models
Covariates: Sex, obesity, diabetes,
smoking history, time-varying
parameters (apparent temperature, h,
day, trip, residence, mold, pollen, illness,
and juice intake), medication and vitamin Copollutant:
consumption (day of blood draw)
Pollutant: PM25 (ambient)
Averaging Time: Hourly data used to
calculate avg concentrations over 1-7
days preceding the blood draw (ambient
PM25)
Microenvironmental PM25 measures
were avgd over the 1-2 days preceding
the blood draw
Mean (SD) (1-day): 16 (6.0)
Percentiles (1-day): 0:6.5
25th: 12
75th: 22
100th: 28
Monitoring Stations: 1 ambient monitor
Season: Limited data collection period
Dose-response Investigated? No
Statistical Package: SAS v8.02
PM25 (ambient)
BC (ambient)
PM25(microenvironment)
CO
N02
S02
03
PM Increment: 6.1 pg/m (5-day mean)
Effect Estimate [Lower Cl, Upper Cl]:
Note: Most results presented in figures.
Selected result in abstract text: %
change in WBC per increase in IQR
(5.4 fjg/m3) of PM2 5 avgd over the
previous week: 5.5(0.1,11)
Associations (% changes and 95%CI)
between 5-day mean ambient
concentrations and markers of
inflammation per increase (IQR) in
pollutant.
CRP: All participants: 14 (-5.4, 37)
Among those with all 3 conditions
diabetes, obesity, and hypertension): 81
21,172)
Among those with at least 2 of the
conditions: 11 (-7.3, 33)
IL-6: All participants:-2.1 (-13,11)
Among those with all 3 conditions
(diabetes, obesity, and hypertension): 23
(-5.3, 59)
Among those with at least 2 of the
conditions:-3.1 (-14,9.7)
WBC (X109/L): All participants: 3.4 (-1.8,
8.9)
Among those with all 3 conditions
(diabetes, obesity, and hypertension):
0.4 (-8.8,11)
Among those with at least 2 of the
conditions: 3.6 (-1.7, 9.1)
Reference: Dubowsky et al. (2006,
0887501
Period of Study: Mar-Jun 2002
Location: St. Louis, Missouri
Outcome: White blood cells (WBC),
C-reactive protein (CRP), interleukin-6
(IL-6)
Age Groups: > 60 yr
Study Design: Panel (4 planned
repeated measures
n = 35 participated in 4 trips)
N: 44 participants
Statistical Analyses: Linear mixed
models
Covariates: Sex, obesity, diabetes,
smoking history, time-varying
parameters (apparent temperature, h,
day, trip, residence, mold, pollen, illness,
and juice intake), medication and vitamin Copollutant:
consumption (day of blood draw)
Pollutant: BC (ng/m3) (ambient)
Averaging Time: Hourly data used to
calculate avg concentrations over 1-7
days preceding the blood draw (ambient
PM)
microenvironmental PM25 measures
were avgd over the 1-2 days preceding
the blood draw
Mean (SD) (1-day): 900 (280)
Percentiles (1-day): 0:290
25th: 730
75th: 1,100
100th: 1,400
Monitoring Stations: 1 ambient monitor
Season: Limited data collection period
Dose-response Investigated? No
Statistical Package: SAS v8.02
PM25 (ambient)
BC (ambient)
PM25(microenvironment)
CO
N02
S02
03
PM Increment: 230 ng/m (5-day mean)
Effect Estimate [Lower Cl, Upper Cl]:
Note: Most results presented in figures.
Associations (% changes and 96%CI)
between 6-day mean ambient
concentrations and markers of
inflammation per increase (IQR) in
pollutant.
CRP: All participants: 13 (-0.34, 28)
Among those with all 3 conditions
(diabetes, obesity, and hypertension): 49
(16, 90)
Among those with at least 2 of the
conditions: 9.0 (-3.8, 24)
IL-6: All participants: -0.8 (-8.9, 8.0)
Among those with all 3 conditions
diabetes, obesity, and hypertension): 15
-2.2, 35)
Among those with at least 2 of the
conditions:-2.7 (-11, 6.2)
WBC (X109/L): All participants: 1.3 (-2.1,
4.8)
Among those with all 3 conditions
(diabetes, obesity, and hypertension):
0.05 (-5.9, 6.3)
Among those with at least 2 of the
conditions: 1.5 (-2.0, 5.1)
December 2009
E-35
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ebelt et al. (2005, 0569071
Period of Study: Summer of 1998
Location: Vancouver, Canada
Outcome: CVD
Age Groups: Range from 54-86 yr
mean age= 74 yr
Study Design: Extended analysis of a
repeated-measures panel study
N: 16 persons with COPD
Statistical Analyses: Earlier analysis
expanded by developing mixed-effect
regression models and by evaluating
additional exposure indicators
Dose-response Investigated? No
Statistical Package: SASV8
Pollutant: PM25
Averaging Time: 24 h
Mean (SD):
Ambient PM25:11.4 ±4.6
Exposure to ambient PM2.$. 7.9 ±3.7
Range (Min, Max):
Ambient PM25: 4.2-28.7
Exposure to ambient PM25: 0.9-21.3
Monitoring Stations: 5
Copollutant (correlation):
Ambient concentrations and exposure to
ambient PM were highly correlated for
each respective metric: r 2 0.71
PM Increment:
Increment: C2.5: IQR = 5.8
SBP (mm Hg):-1.70 (-3.48-0.08)
DBP (mm Hg): -0.58 (-2.02-0.85)
Ln-SVE (bph): 0.20 (0.00-0.40)
HR (bpm): 0.93 (-0.90-2.75)
SDNN (ms): -4.37 (-9.40-0.65)
R-MSSD(ms):-2.79 (-6.16-0.57)
Increment: NS_C2.5: IQR = 4.2
SBP (mm Hg):-1.52 (-2.94--0.09)
DBP (mm Hg):-0.77 (-1.87-0.32)
Ln-SVE (bph): 0.19 (-0.01-0.38)
HR (bpm): 1.03 (-0.43-2.48)
SDNN (ms):-3.83 (-7.77-0.11)
R-MSSD(ms):-2.90 (-5.55--0.25)
Increment: S_C2.5: IQR =1.5
SBP (mm Hg):-1.10 (-3.48-1.28)
DBP (mm Hg): 0.76 (-1.15-2.68)
Ln-SVE (bph): 0.09 (-0.05-0.23)
HR (bpm):-0.42 (-2.28-1.44)
SDNN (ms):-3.14 (-9.73-3.45)
R-MSSD(ms): 0.24 (-5.14-5.63)
Increment: A2.5: IQR = 4.4
SBP (mm Hg):-1.90 (-3.66--0.14)
DBP (mm Hg):-0.33 (-1.72-1.06)
Ln-SVE (bph): 0.20 (0.02-0.37)
HR (bpm): 0.57 (-1.34-2.47)
SDNN (ms): -3.91 (-8.79-0.97)
R-MSSD(ms):-1.05 (-4.79-2.17)
Increment: NS_A2.5: IQR = 3.4
SBP (mm Hg):-1.70 (-3.27--0.14)
DBP (mm Hg):-0.51 (-1.71-0.70)
Ln-SVE (bph): 0.20 (0.02-0.37)
HR (bpm): 0.69 (-0.96-2.35)
SDNN (ms):-4.18 (-8.51-0.15)
R-MSSD(ms):-1.40 (-4.40-1.60)
Increment: S_T2.5: IQR = 0.9
SBP (mm Hg):-1.55 (-3.35-0.26)
DBP (mm Hg): 0.49 (-0.91-1.90)
Ln-SVE (bph): 0.08 (-0.14-0.19)
HR (bpm):-0.24 (-1.75-1.26)
SDNN (ms): -0.68 (-4.74-3.38)
R-MSSD (ms): 0.91 (-3.51-5.33)
Increment: T2.5: IQR =10.1
SBP (mm Hg):-1.26 (-2.60-0.08)
DBP (mm Hg): 0.34 (-1.26-1.94)
Ln-SVE (bph): 0.01 (-0.10-0.11)
HR (bpm):-0.23 (-1.09-0.63)
SDNN (ms):-2.11 (-4.90-0.68)
R-MSSD (ms):-0.83 (-3.60-1.94)
Increment: N2.5: IQR = 8.9
SBP (mm Hg):-0.81 (-2.15-0.53)
DBP (mm Hg): 0.40 (-1.19-1.98)
Ln-SVE (bph):-0.04 (-0.18-0.10)
HR (bpm):-0.35 (-0.85-0.14)
SDNN (ms):-1.10 (-3.10-0.90)
R-MSSD (ms):-0.54 (-2.54-1.46)
Note: Total personal fine particle
exposure (T) were dominated by
exposures to non ambient particles
which were not correlated with ambient
fine particle exposure (A) or ambient
concentrations (C). Results for each of
these metrics are listed.
December 2009
E-36
-------
Reference Design & Methods
Reference: Fan et al. (2008, 191979) Outcome: Cardiopulmonary Health
„ • ., «~ ., ,-,„ ™nr (FEV, FVC, PEF, SDNN, HR)
Period of Study: Feb-May 2005
Age Groups: 61.2 (13.7)
Location: Paterson, New Jersey
Study Design: Panel
N:11
Statistical Analyses: Mixed Effects
models, Linear Regression models
Covariates: Temperature, humidity
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered :0
Concentrations1
Pollutant: PM,5
Averaging Time: Daily
Mean (SD): |
APM25avg
Morning: 35.2 (25.9)
Afternoon: 24.1 (22.1)
APM25peak
Morning: 71. 3 (56.1)
Afternoon: 64.3 (43.5)
Range:
APM2.5 avg
Morning: 1.1 -87
Afternoon: 1.2-98
APM25 peak
Morning: 4.0 -278
Afternoon: 3.0 -150
Monitoring Stations: NR
Copollutant: NR
Co-pollutant Correlation: N/A
Effect Estimates (95% Cl)
PM Increment: 10 pg/m3
Beta (SE), p-value:
ASDNN
Morning, APM25avg
15mirv-145(69) 006
2h: -18.9 (4.2), 0.0002
4h: -2.5 (8.6), 0.78
Morning, APM25 peak
15min: -9.2 (112), 0.43
2h: -5.1 (13.8), 0.72
4h: -7.4 (12.0), 0.55
Afternoon, APM25 avg
15min: -2.4 (7.6), 0.77
2h: -20.2 (10.8), 0.10
4h: -0.7 (11. 2), 0.95
Afternoon, APM25 peak
15min: 0.6 (8.9), 0.95
2h: 19.2 (14.6), 0.23
4h: -6.8 (14.1), 0.64
AHR
Morning, APM25avg
15min: 1.2 (3.1), 0.71
2h: -5.5 (2.9 , 0.08
4h: -3.1 (4.6, 0.51
Morning, APM25peak
15min: 0.8 (4.4), 0.86
2h: -7.2 (4.2, 0.11
4h: -7.1 (6.3, 0.28
Afternoon, APM25 avg
15min:-2.0 (4.0), 0.62
2h: 0.9 (5.4), 0.87
4h: 8.2 (5.2), 0.14
Afternoon, APM25 peak
15min:-5.6 (5.3), 0.31
2h: 3.1 (8.1), 0.71
4h:11.1 (8.1), 0.20
AFEV,
Morning, APM25 avg: 0.02 (0.04), 0.68
Morning, APM25 peak: -0.13 (0.08), 0.16
A FVC
Morning, APM25 avg: -0.10 (0.09), 0.31
Morning, APM25 peak: -0.12 (0.17), 0.51
A PEF
Morning, APM25 avg: -0.54 (0.62), 0.42
Morning, APM25 peak: -1.46 (1.12), 0.24
Notes: Estimates relative to increases in
the avg and peak PM25 concentrations
Reference: Folino et al. (2009,1919021
Period of Study: Jun 2006-May 2007
Location: Padua, Italy
Outcome: HRV & Inflammatory Markers
Age Groups: 45-65 yr
Study Design: Panel
N: 39 patients w/ myocardial infarction
Statistical Analyses: Linear Regression
Model, ANOVA
Covariates: Temperature, relative
humidity, atmospheric pressure, beta-
blocker, aspirin, or nitrate consumption,
smoking habit
Dose-response Investigated? No
Statistical Package: Stata
Lags Considered: NR
Pollutant: PM25
Averaging Time: 24 h
Mean (SD):
Summer: 33.9 (12.7)
Winter: 62.1 (27.9)
Spring: 30.8 (14.0)
Monitoring Stations: NR
Copollutant: PM10, PM0.25
Co-pollutant Correlation: NR
PM Increment: 1 pg/m
Beta (SE), p-value:
SDNN: 0.109 (0.115), 0.345
SDANN: 0.127 (0.126), 0.314
RMSSD: 0.045 (0.040), 0.256
pH: 0.002 (0.001), 0.041
LTB4: 0.590 (0.324), 0.069
eNO: -0.002 (0.003), 0.503
PTX3: -0.004 (0.002), 0.013
C-reactive protein: -0.008 (0.005), 0.115
CC16:-0.002 (0.002), 0.410
IL-8: 0.000 (0.003), 0.989
December 2009
E-37
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Folino et al. (2009,1919021
Period of Study: Jun 2006-May 2007
Location: Padua, Italy
Outcome: HRV & Inflammatory Markers
Age Groups: 45-65 yr
Study Design: Panel
N: 39 patients w/ myocardial infarction
Statistical Analyses: Linear Regression
Model, ANOVA
Covariates: Temperature, relative
humidity, atmospheric pressure, beta-
blocker, aspirin, or nitrate consumption,
smoking habit
Dose-response Investigated? No
Statistical Package: Stata
Lags Considered: NR
Pollutant: PM025
Averaging Time: 24 h
Mean (SD):
Summer: 17.6(7.5)
Winter: 30.5 (17.4)
Spring: 18.8 (10.8)
Monitoring Stations: NR
Copollutant: PM10, PM25
Co-pollutant Correlation: NR
PM Increment: 1 pg/m
Beta (SE), p-value:
SDNN: 0.214 (0.204), 0.295
SDANN: 0.214 (0.214), 0.316
RMSSD: 0.081 (0.077), 0.291
pH: 0.005 (0.002), 0.004
LTB4: 0.835 (0.533), 0.117
eNO:-0.006 (0.005), 0.182
PTX3: -0.006 (0.003), 0.071
C-reactive protein: -0.011 (0.007), 0.104
CC16: 0.001 (0.004), 0.890
IL-8: -0.004 (0.006), 0.527
Reference: Goldberg et al. (2008,
1803801
Period of Study: Jul 2002-Oct 2003
Location: Montreal, Canada
Outcome: Oxygen saturation & pulse
rate
Age Groups: 50-85 yr
Study Design: Panel
N:31
Statistical Analyses: Mixed Random
Effects Model
Covariates: Body temperature,
consumption of salt, intake of fluids,
being ill the day before, ambient
temperature, relative humidity,
barometric pressure
Dose-response Investigated? No
Statistical Package: Splus
Lags Considered: lags 1 day; 0- to
2-day avg
Pollutant: PM25
Averaging Time: Daily
IQR: 7.3
Monitoring Stations: 8
Co-pollutant: CO, N02, S02, 03
Co-pollutant Correlation
CO: 0.72
N02: 0.62
PM Increment: Interquartile Range
(7.3pg/m3)
Mean Difference (Lower Cl, Upper Cl),
lag:
Oxygen Saturation
Unadjusted:
-0.087 (-0.143, -0.031), lag 0
Unadjusted:
-0.058 (-0.114, -0.002), lag 1
Unadjusted:
-0.083 (-0.155, -0.010), lag 0-2-day avg
Adjusted: -0.056 (-0.117, 0.005), lag 0
Adjusted: -0.019 (-0.079, 0.041), lag 1
Adjusted: -0.039 (-0.118, 0.039), lag 0-
2-day avg
Pulse Rate
Unadjusted: 0.226 (-0.037, 0.489), lag 0
Unadjusted: 0.288 (0.022, 0.554), lag 1
Unadjusted: 0.420 (0.067, 0.772), lag 0-
2-day avg
Adjusted: 0.158 (-0.136, 0.451), lag 0
Adjusted: 0.246
Adjusted: 0.353
2-day avg
-0.040, 0.531), lag 1
-0.034, 0.740), lag 0-
December 2009
E-38
-------
Reference
Reference: Goldberg et al. (2008,
1803801
Period of Study: Jul 2002-Oct 2003
Location: Montreal, Canada
Design & Methods
Outcome: Shortness of Breath &
General health
Age Groups: 50-85 yr
Study Design: Panel
N:31
Statistical Analyses: Mixed Random
Fffprtc Mnrlpl
CIICULo IVIUUCI
Covariates: Body temperature,
consumption of salt, intake of fluids,
being ill the day before, ambient
temperature, relative humidity,
barometric pressure
Dose-response Investigated? No
Statistical Package: Splus
Lags Considered: lags 0-4 days; 0- to
2-day avg
Concentrations1
Pollutant: PM25
Averaging Time: Daily
Mean: 9.5
Median: 7.0
Mirr D R
IVIIIK U.u
Max: 50.2
IQR: 7.3
Monitoring Stations: 8
Co-pollutant: CO, N02, S02, 03
Co-pollutant Correlation
CO: 0.66
N02: 0.54
03: 0.32
S02: 0.50
Effect Estimates (95% Cl)
PM Increment: Interquartile Range
(7.3 pg/m3)
Mean Difference (Lower Cl, Upper Cl),
lag:
IMV).
General Health
Unadjusted: -0.317 (-0.699, 0.064), lag 0
Unadjusted: -0.284
Unadjusted: -0.048
-0.670, 0.103), lag 1
-0.427, 0.332), lag 2
Unadjusted: -0.241 (-0.620, 0.139), lag 3
Unadjusted: -0.010 (-0.390, 0.370), lag 4
Unadjusted: -0.482 (-1.053, 0.090), lag
0-2-day avg
Adjusted: -0.125 (-0.545, 0.295), lag 0
Adjusted: -0.167 (-0.568, 0.234 , lag 1
Adjusted: -0.081 (-0.464, 0.302 , lag 2
Adjusted: -0.222 (-0.602, 0.157), lag 3
Adjusted: 0.016 (-0.364, 0.396), lag 4
Adjusted: -0.281 (-0.886, 0.325), lag
0-2-day avg
Shortness of breath at night
Unadjusted: -0.421
Unadjusted: -0.278
-0.847, 0.006), lag 0
-0.711,0.155), lag 1
Unadjusted: -0.100 (-0.526, 0.327), lag 2
Unadjusted: -0.220
Unadjusted: -0.206
-0.645, 0.206), lag 3
-0.632, 0.220), lag 4
A *-jr\ n n/?o\ i
Unadjusted: -0.555 (-1.172, 0.063), lag
0-2-day avg
Adjusted: -0.171 (-0.639, 0.297), lag 0
Adjusted: -0.130 (-0.579, 0.319), lag 1
Adjusted:-0.127 (-0.553, 0.299
Adjusted:-0.192 (-0.616, 0.231
Jag 2
, Iag3
Adjusted: -0.171 (-0.594, 0.253), lag 4
Adjusted: -0.301 (-0.952, 0.350), lag
0-2-day avg
Reference: Ibald-Mulli et al. (2004,
0874151
Period of Study: Winter 1998-1999
Location:
Helsinki, Finland
Erfurt, Germany
Amsterdam, the Netherlands
Reference: Langrish et al. (2009,
1919081
Period of Study: Aug 2008
Location: Beijing, China
Outcome: Blood Pressure & Heart Rate
Age Groups: 40-84
Study Design: Panel
N: 131 adults w/CHD
Statistical Analyses: Linear Regression
Covariates: Trend, day of week,
temperature, barometric pressure,
relative humidity, medication use
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-2, 5-day avg
Outcome: Cardiovascular Effects
Age Groups: Median 28 yr
Study Design: Panel
N: 15
Statistical Analyses: NR
Covariates: NR
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: NR
Pollutant: PM25
Averaging Time: 24 h
Mean (SD):
Downtown: 40 (20)
Tunney's Pasture: 10 (10)
p-value 0.000
Monitoring Stations: NR
Copollutant: PM10
Co-pollutant Correlation: N/A
Pollutant: PM25
Averaging Time: NR
Mean:
W/o mask: 86
W/ mask: 140
Monitoring Stations: NR
Co-pollutant: CO, S02, N02
Co-pollutant Correlation: N/A
PM Increment: Interquartile Range
(27.02 pg/m3)
Beta (SE), p-value:
Flow mediated vasodilation (%):
-0.016 (0.0072) p=0.03
Heart Rate (beats/min):
0.081 (0.135) p=0.55
Diastolic blood pressure (mmHg):
0.088 (0.088) p=0.32
Systolic blood pressure (mmHg):
-0.108 (0.006) p=0.48
PM Increment: NR
Mean (Lower Cl, Upper Cl):
W/o Mask (Day)
SBP:100(104, 116)
DBP' 73 (69 76)
MAP: 85 (81, 88)
Heart Rate: 79 (74, 84)
Avg NN interval: 829 (789, 869)
pNNSO: 15.9 (10.7, 21.0)
RMSSD:35.1 (29.2,41.0)
SDNN:61.2(54.9, 67.5)
Triangular index: 12.9(11.9, 13.9)
LF power: 81 6 (628, 1004)
HF power: 460 (325, 595)
LFn: 62.8 (56.7, 68.9)
HFn: 29.2 (25.5, 32.8)
HF/LF ratio: 0.738 (0.507, 0.970)
W/ Mask (Day)
SBP: 109 (104, 114)
DBP: 73 (70-76)
MAP: 85 (81, 89)
December 2009
E-39
-------
Reference Design & Methods Concentrations1 Effect Estimates (95% Cl)
Heart Rate: 78 (73, 82)
Avg NN interval: 850 (805, 896)
pNNSO: 17.9 (14.2, 21.6)
RMSSD:37.1 (32.2,42.0)
SDNN: 65.5 (59.0, 72.2)*
Triangular index: 13.8 (13.0,14.5)
LF power: 919 (717,1122)*
HF power: 485 (400, 569)
LFn: 64.5 (60.6, 68.4)
HFn: 30.0 (27.0, 33.1)
HF/LF ratio: 0.680 (0.519, 0.842)
W/o Mask (During Walk)
SBP: 121 (115, 127)
DBP: 81 (75-87)
MAP: 94 (89, 99)
Heart Rate: 88 (82, 94)
Avg NN interval: 594 (562, 627)
pNNSO: 3.3 (0.8, 5.7)
RMSSD: 17.2 (13.4, 21.0)
SDNN: 45.8 (36.8, 54.8)
Triangular index: 10.7(9.1,12.4)
LF power: 313 (170, 455)
HF power: 76.5 (33.6, 120.0)
LFn: 68.2 (60.9, 75.5)
HFn: 16.1 (11.9, 20.3)
HF/LF ratio: 0.259 (0.173, 0.344)
VW Mask (During Walk)
SBP: 114 (108, 120)
DBP: 79 (74, 83)
MAP: 90 (86, 94)
Heart Rate: 91 (85, 97)
Avg NN interval: 613 (571, 655)
pNN50:2.1 (-0.1,-4.4)
RMSSD: 20.0 (15.5, 24.6)
SDNN: 54.8 (42.5, 67.0)
Triangular index: 11.4 (9.4,13.3)
W Mask (During Walk)
LF power: 414 (233, 595)
HF power: 116.8 (52.6,181.0)
LFn: 67.9 (61.9, 73.9)
HFn: 16.0 (12.5, 19.4)
HF/LF ratio: 0.247 (0.180, 0.314)
Mean (SD):
W/o Mask (After Walk)
Headache: 2.53 (5.55)
Dizziness: 1.07 (2.22)
Tiredness: 8.47 (12.14)
Sickness: 1.07 (2.22)
Cough: 1.80 (4.80)
Difficulty Breathing: 0.67 (0.90)
Eye irritation: 1.40(3.60)
Throat irritation: 1.47 (4.07)
Nose irritation: 1.53(3.78)
Unpleasant Smell: 0.93 (1.22)
Bad taste: 0.73 (0.96)
Difficulty walking: 12.53 (13.24)
Perception of Pollution: 19.80 (18.37)
W/Mask (After Walk)
Headache: 0.73 (1.03)
Dizziness: 0.80 (1.57
Tiredness: 7.40 (9.37)
Sickness: 0.87 (1.51)
Cough: 1.00 (1.73)
Difficulty Breathing: 3.80 (8.10)
Eye irritation: 1.67 (3.27)
Throat irritation: 1.07 (2.63)
Nose irritation: 1.07 (1.91)
Unpleasant Smell: 0.60 (0.91)
Bad taste: 0.60 (1.18)
Difficulty walking: 15.13 (11.51)
Perception of Pollution: 11.60 (10.44)
*p < 0.05
Notes: Estimates also available for 24 h,
night, before walk, and 24 h after walk.
December 2009 E-40
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Lanki et al. (2006, 0884121
Period of Study: Fall 1998-spring 1999
Location: Helsinki, Finland
Outcome: ST segment depressions (2
endpoints: >0.1mV regardless of the
direction of the ST slope and >0.1mV
with horizontal or downward slope
[stricter criteria])
Age Groups: Mean = 68.2 (6.5) yr
Study Design: Panel
N: 45 elderly nonsmoking persons with
stable coronary heart disease
342 total exercise tests for analyses
Statistical Analyses: Generalized
additive models with penalized splines
(logistic regression)
principal components analysis and linear
regression of 13 measured elements
used to apportion PM25 mass between
different sources
Covariates: Subject, linear terms for
time trend, temperature, relative
humidity, penalized spline for change in
heart rate during the exercise test
Season: NR
Dose-response Investigated? No
Statistical Package: S-plus 2000 and R
Pollutant: PM25 (Analyses conducted
for source specific PM25)
Averaging Time: Daily filter samples
Mean:
Crustal: 0.6
Long-range transported: 6.4
Oil combustion: 1.6
Salt: 0.9
Local traffic: 2.9
Total: 12.8
Percentiles: Crustal
25: 0.0
50: 0.4
75:1.1;
Max: 5.3
Long-range transported
25:2.2
50: 5.5
75: 9.8;
Max: 26.5
Oil combustion
25: 0.6
50:1.3
75:2.3;
Max: 12.2
Salt
25: 0.3
50: 0.8
75:1.2;
Max: 5.9
Local traffic
25:1.7
50: 2.5
75: 3.4;
Max: 12.0
Total
25: 8.3
50:10.6
75:15.9;
Max: 39.8
Monitoring Stations: 1 monitor
Copollutant (correlation):
Correlations with PM25:
Crustal: r =-0.01
Long-range transported: r = 0.82
Oil combustion: r = 0.35
Salt: r = 0.19
Local traffic: r = 0.26
PM Increment: 1 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
Adjusted ORs between daily source-
specific PM25 concentrations and
ST-segment depressions. ST-segment
depression defined as >0.1 mV (n = 62)
Crustal
Lag 0: 0.80
Lag 1:0.66
0.47, 1.36)
0.40, 1.10)
Lag 2:1.18 (0.68, 2.06)
Lag 3:1.87 (0.85, 4.09)
Long-range transport
Lag 0:0.94 (0.84, 1.05)
Lag 1:1.00 (0.92,1.08)
Lag 2:1.11 (1.02,1.20)
Lag 3:1.06 (0.95,1.18)
Oil combustion
Lag 0:0.87 (0.57, 1.32)
Lag 1:1.04 (0.75,1.45)
Lag 2:1.10
Lag 3:1.12
0.83, 1.46)
0.79, 1.58)
Salt
LagO: 1.03 (0.57,1.85)
Lag1: 0.72 (0.37,1.40)
Lag2: 0.66 (0.31,1.40)
Lag3:1.55 (0.83, 2.89)
Local traffic
Lag 0:0.91 (0.69,1.21)
Lag 1:1.22
Lag 2:1.53
0.88, 1.69)
1.19,1.97)
Lag 3: 0.98 (0.78, 1.23)
ST-segment depression defined as >0.1
mV with horizontal or downward slope
(n = 46)
Crustal
LagO: 0.76 (0.42, 1.35)
Lag1:0.41 (0.22,0.79)
Lag2:1.17 (0.65, 2.09)
Lag3:1.60 (0.72, 3.59)
Long-range transport
Lag 0: 0.98 (0.86,1.10)
Lag 1:1.03 (0.95,1.12)
Lag 2: 1.11 (1.02, 1.21)
Lag 3:1.02 (0.95,1.10)
Oil combustion
Lag 0:0.95 (0.61,1.49)
Lag 1:1.13
Lag 2:1.33
0.76, 1.68)
0.98, 1.80)
Lag 3:1.29 (0.90, 1.86)
Salt
Lag 0:1.15
Lag 1:0.90
0.56, 2.38)
0.44, 1.81)
Lag 2:1.39 (0.63, 3.08)
Lag 3:1.93 (1.00, 3.72)
Local traffic
Lag 0:0.89 (0.64, 1.23)
Lag 1:1.21 (0.86,1.71)
Lag 2:1.37 (1.03,1.83)
Lag 3:1.03 (0.80,1.32)
Adjusted ORs for the association of
indicator elements of PM25 sources and
ST-segment depressions in
multipollutant models (models include all
5 indicator elements). ST-segment
depression defined as >0.1 mV (n = 62)
Si (Crustal)
LagO: 0.73 (0.39, 1.38)
December 2009
E-41
-------
Reference Design & Methods Concentrations1 Effect Estimates (95% Cl)
Lag 1:0.48 (0.25, 0.93)
Lag2: 0.78 (0.35, 1.71)
Lag3:1.95 (0.69, 5.48)
S (Long-range transport)
LagO: 0.70 (0.25, 1.95)
Lag 1:0.58 (0.23, 1.47)
Lag2:1.08 (0.44, 2.63)
Lag3:1.60 (0.73, 3.48)
Ni (Oil combustion)
LagO: 0.78 (0.30, 2.04)
Lag 1:1.20 (0.58, 2.46)
Lag2:1.15(0.61,2.18)
Lag3:1.02 (0.41, 2.54)
Cl (Salt)
LagO: 1.03 (0.79,1.34)
Lag 1:0.88 (0.56, 1.38)
Lag2:1.02 (0.62, 1.69)
Lag3:1.27 (0.85,1.91)
ABS (Local traffic)
LagO: 0.92 (0.36, 2.37)
Lag 1:1.83 (0.73, 4.59)
Lag2: 4.46 (1.69, 11.79)
Lag3: 0.92 (0.40, 2.12)
ST-segment depression defined as >0.1
mV with horizontal or downward slope
(n = 46)
Si (Crustal)
LagO: 0.67 (0.33, 1.36)
Lag1: 0.34 (0.15, 0.81)
Lag2: 0.81 (0.33, 2.00)
Lag3:1.90 (0.64, 5.65)
S (Long-range transport)
LagO: 0.84 (0.29, 2.47)
Lag 1:0.89 (0.34, 2.32)
Lag2:1.36 (0.54, 3.45)
Lag3:1.12 (0.53, 2.40)
Ni (Oil combustion)
LagO: 1.10(0.36, 3.37)
Lag1:1.16 (0.45, 2.96)
Lag2:1.64 (0.84, 3.20)
Lag3:1.63 (0.64, 4.14)
Cl (Salt)
LagO: 1.13 (0.80,1.62)
Lag 1:0.99 (0.58, 1.68)
Lag2:1.55 (0.87, 2.76)
Lag3:1.45 (0.94, 2.25)
ABS (Local traffic)
LagO: 0.74 (0.25, 2.23)
Lag1:1.76 (0.62, 5.00)
Lag2: 4.86 (1.55, 15.26)
Lag3: 0.97 (0.39, 2.41)
December 2009 E-42
-------
Reference
Reference: Lanki et al. (2008, 191984)
Period of Study: Jan 1999-Apr 1999
Location: Helsinki, Finland
Design & Methods
Outcome: ST Segment Depressions
>0.1 mV
Age Groups: 50+
Study Design: Panel
N: 41 elderly people w/CHD
Statistical Analyses: Logistic
Regression Model
Covariates: Long-term time trend,
temperature, humidity, change in heart
rate following exercise test
Dose-response Investigated? No
Statistical Package: R
Lags Considered: lags 0-24 h
Concentrations1
Pollutant: PM,5
Averaging Time: Hourly
26th, 60th, 76th, Max:
Personal PM25
1h:6.9, 11.2, 15.8, 41.5
4h: 5.9, 10.0, 14.6, 41.3
8h:5.0, 7.9, 13.0, 34.9
12h: 5.2, 7.8, 12.1,28.8
22h:6.6, 9.3, 13.0,30.2
Outdoor PM2 5
1h:8.9, 12.9, 17.8,42.9
4h:8.8, 12.5, 17.6, 40.8
8h:8.3, 12.1, 17.2,39.2
12h: 8.3, 11.9, 17.0, 37.0
24 h: 9.0, 12.5, 17.7, 30.5
Monitoring Stations: 1
Co-pollutant: PM<0.1
Effect Estimates (95% Cl)
PM Increment: 10 pg/m3
Odds Ratio (Lower Cl, Upper Cl):
Personal PM2 5
1-havg: 3.26 (1.07, 9.99)*
4-h avg: 2.42 0.75, 7.83)
8-havg:1.57 0.49,5.09)
12-h avg: 1.96 (0.44, 8.64)
22-h avg: 2.06 (0.30, 14.10)
Outdoor PM2 5
1-havg: 1.77 (0.87, 3.58)
4-h avg: 2.47 (1.05, 5.85)*
8-h avg: 1.83 (0.80, 4.20)
12-h avg: 1.90 (0.77, 4.65)
24-h avg: 1.60 (0.59, 4.39)
*p < 0.05
Co-pollutant Correlation
Personals Outdoor PM2 5
1 h&1 h:0.70
4hS4h:0.54
8hS8h:0.60
12hS12h:0.50
22hS24h:0.80
Notes: 1-22 h pollutant averaging times.
Correlations also available for personal-
personal and outdoor-outdoor.
Reference: Liao et al. (2007, 1802721 Outcome: Ectopy Pollutant: PM25
Period of Study: 1999-2004 Age Groups: women 50-79 yr Averaging Time: Daily
Location: 24 U.S. States Study Design: Panel Mean (SD)*:
„„,„ All: 13.8 (79)
N: 57>422 No Ectopy: 13.8 (7.9)
Statistical Analyses: logistic regression AnV EctoPy; 13'8 <7'6'
& random effects modeling 6th, 95th percentile*:
AH1 5 29 1
Covariates: Age, race, center, M V t ^ ?Q ?
education, history of CVD/chronic lung A° SfL i ns OR *
disease, rel. humidity, temperature, Any Ectopy. 5.06, 28.5
smokin9 Monitoring Stations: NRJ
Dose-response Investigated? No Copollutant: PM10
Statistical Package: SAS, State Co-pollutant Correlation: NR
Lags Considered: lags 0-365 days *|_ag 1
^Monitors used in model for spatial
interpolation of daily PM values.
PM Increment: 10 pg/m3
Percent Change (Lower Cl, Upper Cl):
All Ventricular Ectopy
Lag 0:1. 01 (0.91, 1.1 3)
Lag 1:1.07 (0.96, 1.20)
Lag 2: 1.09 (0.98, 1.21)
Current Smoker Ventricular Ectopy
Lag 0:1. 52 (1.04, 2.24)
Lag 1:2 (1.32, 3.03)
Lag 2: 1.59 (0.99, 2.55)
Nonsmoker Ventricular Ectopy
Lag 0: 0.99
Lag 1:1. 05
0.89, 1.11)
0.94, 1.17)
Lag 2: 1.08 (0.97, 1.21)
All Supraventricular Ectopy
Lag 0:1. 04 (0.96, 1.13)
Lag 1:1. 01
Lag 2: 0.96
0.93, 1.10)
0.87, 1.05)
All Ventricular or Supraventricular
Ectopy
Lag 0:1. 03
Lag 1:1. 04
0.96, 1.11)
0.97,1.11)
Lag 2:1 (0.94, 1.07)
Reference: Lipsett et al. (2006, 0887531 Outcome: HRV parameters, specifically Pollutant: PM25
PM Increment: SE*1 00
SDNN, SDANN, r-MSSD, LF, HF, total
Period of Study: Feb-May 2000 power] triangular index (TRII).
Location: Coachella Valley, CA
Study Design: Panel study
N: 19 non-smoking adults with coronary
artery disease
Statistical Analysis: Mixed linear
regression models with random effects
parameters
Averaging Time: 2 h
Mean (range)
Indio: 23.2 (6.3-90.4)
Palm Springs: 14 (4.7-52)
Monitoring Stations: 2
Copollutant: 0;
Effect Estimate (change in HRV per
unit increase in PM concentration):
SDNN: -0.37 msec (SE= 1.01)
Notes: Weekly ambulatory 24 h ECG
recordings (once per week for up to 12
wk), using Holter monitors, were made.
Subjects' residences were within 5 mi of
1 of 2 PM monitoring sites. Decreased
HRV was associated with PM2 5, but
these effects were not statistically
significant. Regressed HRV parameters
against 18: 00-20: 00 mean particulate
pollution.
December 2009
E-43
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ljungman et al. (2008,
1802661
Period of Study: Aug 2001-Dec 2006
Location: Stockholm, Sweden
Outcome: Ventricular Arrhythmia
Age Groups: 28-85 yr
Study Design: Case-crossover
N: 88 patients w/ implantable
cardioverter defibrillators
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature, humidity,
pressure, ischemic heart disease,
ejection fraction, heart disease,
diabetes, use of beta-blockers, age,
BMI, location at time of arrhythmia,
distance from air pollution monitor
Dose-response Investigated? No
Statistical Package: Stata, S-plus
Lags Considered: lags 2-24 h
Pollutant: PM25
Averaging Time: Hourly
Median:
2 h: 9.17
24 h: 9.49
Min:
2 h: 0.15
24 h: 2.97
Max:
2 h: 99.25
24 h: 47.07
IQR:
2 h: 6.69
24 h: 5.27
Monitoring Stations: 1
Copollutant: PM10, N02
Co-pollutant Correlation: NR
PM Increment: Interquartile Range
Odds Ratio (Lower Cl, Upper Cl):
2 h: 1.23 (0.84, 1.80)
24 h: 1.28 (0.90, 1.84)
Notes: OR of ventricular arrhythmia for
an IQR increase of air pollutants in
different subgroups (Fig 2)
Reference: Ljungman et al. (2009,
191983)
Period of Study: May 2003-Jul 2004
Location: Athens, Greece
Helsinki, Finland
Ausburg, Germany
Barcelona, Spain
Rome, Italy
Stokholm, Sweeden
Outcome: lnterleukin-6 Response
Age Groups: 35-80 yr
Study Design: Panel
N: 955 male myocardial infarction
survivors
Statistical Analyses: Additive Mixed
Models
Covariates: Age, sex, BMI, city,
HDL/total cholesterol, smoking, alcohol
intake, HbA1c, NT-proBNP, history of Ml,
heart failure, or diabetes, phlegm
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 1 day
Pollutant: PM25
Averaging Time: 24 h
Mean: 17.7
26th: 10.9
75th: 21.9
Monitoring Stations: NR
Copollutant: CO, N02, PNC, PM25
Co-pollutant Correlation:
PM,0:0.81
PM Increment: Interquartile Range
Change of IL-6 (Lower Cl, Upper Cl),
p-value:
0.6 (-0.8, 2.0), 0.40
Reference: Luttman-Gibson et al.
(2006, 0897941
Period of Study: Jun-Dec 2000
Location: Steubenville, OH
Outcome: Heart rate variability
Age Groups:
Study Design: Panel study
N: 32 participants
Statistical Analysis: Linear mixed
models
Pollutant: PM25
Averaging Time:
1 h
24 h
Mean (IQR)
PM25: 20.0 (15.2)
Sulfate:6.9(5.1)
EC: 1.1 (0.6)
Copollutant: N02, S02, 03
PM Increment: IQR
Percent change (96% Cl): Each
13.4 pg/m3 increase in 24 h mean PM25
concentration was associated with:
SDNN: -4.0% (95% Cl: -7.0% to -0.9%)
r-MSSD:-6.5%(95%CI:-12.1%to
-0.6%)
HF: -11.4% (95% Cl: -21.5% to -0.1%)
Each 5.1 pg/m3 increase in sulfates on
the previous day was associated with:
SDNN: -3.3% (95% Cl: -6.0% to -0.5%)
r-MSSD: -5.6% (95% Cl: -10.7%, 0.2%)
HF: -10.3% (95% Cl: -19.5% to -0.1%)
Notes: The authors conclude that
increases in both traffic related particles
and sulfates may adversely effect
autonomic function.
December 2009
E-44
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Mar et al. (2005, 0875661
Period of Study: 1999-2001
Location: Seattle, WA
Outcome: Change in arterial 02
saturation, heart rate, and blood
pressure (SBP and DBP)
Age Groups: >75 yr
Study Design: Panel study
N: 88 elderly subjects
Statistical Analysis: GEE
Pollutant: PM25
Averaging Time: 24 h
Mean (SD):
Personal: 9.3(8.4)
Indoor: 7.4 (4.8)
Outdoor: 9.0 (4.6)
PM Increment: 10 pg/m
Unit change in measure (96% Cl):
Among all subjects: Each increase in
outdoor same day PM2 5 was associated
with: SBP: -0.81 mmHg (95% Cl: -2.34,
0.73)
DBP: -0.46 mmHg (95% Cl: -1.49 to
0.57)
H: -0.75 beats/min (95% Cl: -1.42 to
-0.07)
Each increase in indoor same day PM25
was associated with: SBP: 0.92 mmHg
(95% Cl:-2.04 to 3.87)
DBP: 0.38 mmHg (95% Cl:-1.43 to
2.20)
H: 0.22 beats/min (95% Cl:-0.71 to
1.16)
Each increase in personal same day
PM25 was associated with: SBP: 0.37
mmHg (95% Cl:-0.93 to 1.67)
DBP: -0.20 mmHg (95% Cl: -0.85 to
0.46)
H: 0.44 beats/min (95% Cl: 0.04to 0.84)
Notes: Results by health status
presented in Fig 1
Used 2 sessions that each were 10
consecutive days of measurements
Used personal, indoor, and outdoor
measures of PM25
Reference: Metzger et al. (2007,
0928561
Period of Study: Aug 1998-Dec 2002
Location: Atlanta, GA
Outcome: Days with any event
recorded by the ICD, days with ICD
shocks/defibrillation and days with either
cardiac pacing or defibrillation
Study Design: Repeated measures
N: 884 subjects between 1993 and 2002
Statistical Analysis: Logistic regression
with GEE to account for residual
autocorrelation within subjects
Pollutant: PM25
Averaging Time: 24 h
Mean (SD):
PM25: 17.8 (8.6)
PM25sulfates:5.0(3.4)
PM25EC:1.7(1.2)
PM25OC:4.4(2.4)
PM2 5 water-soluble metals: 0.029
(0.024)
Percentiles:
PM25: Median: 16.2
PM25sulfates: Median: 4.1
PM25 EC: Median: 1.4
PM25 OC: Median: 3.9
PM2 5 water-soluble metals:
Median: 0.022
PM Increment: OR (96% Cl):
Outcome = Any event recorded by ICD
PM25
OR = 1.00
(95% Cl: 0.95, 1.04)
PM25EC
OR = 1.01
(95% Cl: 0.98, 1.05)
PM25OC
OR = 1.01
(95% Cl: 0.98, 1.03)
PM25Sulfates
OR = 0.99
(95% Cl: 0.93, 1.06)
PM25V\feter soluble metals
OR = 0.95
(95% Cl: 0.90, 1.00
Copollutant:
03
N02
CO
S02
Oxygenated hydrocarbons
Reference: O'Neill et al. (2007, 0913621
Period of Study: May 1998-Dec 2002
Location: Boston, MA
Outcome: Soluble intercellular adhesion Pollutant: PM
molecule 1 (ICAM-1)
Vascular cell adhesion molecule 1
(VCAM-1)
von Willebrand factor (vWF)
Age Groups: Mean (SD): 56.6 (10.6)
Averaging Time: 24 h (lagged ma of
days 0 to 1, 2, 3, 4, and 5)
Mean (SD): 11.4 (5.9)
Descriptive statistics represent entire
study period
PM Increment: IQR (specific to lag
period)
Effect Estimate [Lower Cl, Upper Cl]:
% change per IQR of PM2 5
ICAM-1-All subjects
Lag 0: 2.87 (-4.63, 10.95)
2 dma: 2.25 (-5.15,10.22)
Study Design: Cross-sectional
N: 92 participants (type 2 diabetic
patients)
Percentiles: IQR range: 7.6
Range (Min, Max): 0.07, 33.7)
3 dma: 1.48
4 dma: 1.80
5 dma: 1.51
6 dma: 2. 12
-5.63, 9.11)
-4.98, 9.07)
-5.30, 8.80)
-4.23, 8.89)
December 2009
E-45
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Statistical Analyses: linear regression Monitoring Stations: 1 site
Covariates: Apparent temperature,
season, age, race, sex, glycosylated
hemoglobin, cholesterol, smoking
history, BMI
Dose-response Investigated? No
Statistical Package: NR
Copollutant:
PM25
BC.
S042"
Subjects not known to be taking
statins
Lag 0: 5.47 (-3.74, 15.57)
2 dma: 5.70 (-3.70,16.01)
3dma:4.57
4 dma: 4.57
-4.31, 14.27
-4.27, 14.23
5 dma: 3.80 (-4.84, 13.22)
6 dma: 3.79 (-4.49, 12.80)
Subjects who report smoking in the
past (but not within 6 mo)
Lag 0: 0.9 (-9.56, 12.66)
2 dma: 0.40
3 dma: 1.34
-12.08, 14.65)
-9.23, 13.14)
4 dma: 2.29 (-6.84, 12.30)
5 dma: 1.09 -8.30,11.44)
6 dma: 3.08 -6.30,13.40);
Subjects who did not report smoking
in the past
Lag 0: 0.46 (-8.23, 9.97)
2 dma: 1.37 (-7.96,11.65)
3 dma:-0.96 (-10.01, 9.00)
4 dma:-1.34 (-10.35, 8.58)
5 dma: -0.87
6 dma:-1.78
-10.17, 9.40)
-10.64, 7.94)
VCAM-1- All subjects
Lag 0: 6.88 (-2.88, 17.62)
2 dma: 8.18
3 dma: 6.92
-1.43, 18.72
-1.66,16.25
4 dma: 6.46 (-1.16,14.66)
5 dma: 8.57 (0.05,17.80)
6 dma: 11.76 (3.48, 20.70)
Subjects not known to be taking
statins
Lag 0:10.26 (-0.64, 22.35)
2 dma: 15.02
3 dma: 14.59
3.76, 27.49
3.94, 26.34
4 dma: 15.15 (4.54, 26.84)
5 dma: 16.16 5.77,27.58
6 dma: 17.66 7.77,28.45
Subjects who report smoking in the
past (but not within 6 mo)
Lag 0:13.2 (-1.30, 29.72)
2 dma: 18.4 (0.69, 39.33)
3 dma: 15.7 (1.19, 32.30)
4 dma: 13.1
5 dma: 13.2
0.88, 26.78)
0.49, 27.58)
6 dma: 16.2 (3.76, 30.10)
Subjects who did not report smoking
in the past
Lag 0:-3.12 (-12.41, 7.17)
2 dma:-0.34 (-10.57,11.05)
3 dma:-1.09 (-11.15,10.12)
4 dma:-0.81 (-10.91,10.43)
5 dma: 2.07 (-8.59, 13.96)
6 dma: 4.89 (-5.56, 16.50)
vWF-All subjects
Lag 0:15.16 (-9.79, 47.01)
2 dma: 12.57 (-9.19, 39.55)
3 dma: 25.14 (-9.87, 73.74)
4 dma: 23.42
5 dma: 17.92
-9.47, 68.25)
-10.22, 54.87)
6 dma: 20.48 (-8.82, 59.22)
Subjects not known to be taking
statins
Lag 0:7.40 (-19.82, 43.88)
2 dma: 7.10 (-19.09, 41.76)
3 dma: 10.78 (-17.92, 49.52)
4 dma: 11.61 (-16.64,49.42)
5 dma: 9.15 (-20.32, 49.53)
6 dma: 7.91 (-20.70,46.85)
Subjects who report smoking in the
December 2009
E-46
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
past (but not within 6 mo)
Lag 0:19.23 (-24.29, 87.77)
2dma:19.92
3dma:29.54
-29.65,104.41)
-17.24, 102.76)
4 dma: 41.98 (-6.95, 116.63)
5 dma: 44.05 (-1.23,110.07)
6 dma: 50.39 (9.35, 106.82)
Subjects who did not report smoking
in the past
Lag 0:-14.21 (-53.20,57.24)
2 dma:-20.66 (-63.14, 70.77
3 dma:-28.89 (-68.43, 60.19
4 dma:-23.51 (-55.11, 30.34)
5 dma:-29.18 (-60.08, 25.66)
6 dma:-30.68 (-55.95, 9.08)
Reference: O'Neill et al. (2007, 0913621
Period of Study: May 1998-Dec 2002
Location: Boston, MA
Outcome: Soluble intercellular adhesion Pollutant: BC
molecule 1 (ICAM-1)
Vascular cell adhesion molecule 1
(VCAM-1)
von Willebrand factor (vWF)
Age Groups: Mean (SD): 56.6 (10.6)
Study Design: Cross-sectional
N: 92 participants (type 2 diabetic
patients)
Statistical Analyses: Linear regression
Covariates: Apparent temperature,
season, age, race, sex, glycosylated
hemoglobin, cholesterol, smoking
history, BMI
Dose-response Investigated? No
Statistical Package: NR
Averaging Time: 24 h (lagged ma of
days 0 to 1,2, 3, 4, and 5)
Mean (SD): 1.1 (0.8)
descriptive statistics represent entire
study period
Percentiles: IQR range: 0.8
Range (Min, Max): 0.2, 5.8
Monitoring Stations: 1 site
Copollutant:
PM25
BC.
S042"
PM Increment: IQR (specific to lag
period)
Effect Estimate [Lower Cl, Upper Cl]:
% change per IQR of BC
ICAM-1--All subjects
Lag 0: 5.09 (-2.37, 13.11)
2 dma: 3.97 (-10.24, 20.42)
3 dma: 5.10 (-10.17, 22.96)
4 dma: 8.38 (-6.46, 25.56)
5 dma: 10.09 -7.36,30.83
6 dma: 10.58 -5.34,29.18
Subjects not known to be taking
statins
Lag 0: 5.77 (-3.92, 16.44)
2 dma: 2.39 (-7.65, 13.52)
3 dma: 0.84 (-8.16,10.73)
4 dma: 1.67
5 dma: 1.55
-6.71, 10.80
-6.46, 10.24
6 dma: 2.20 (-6.47,11.6
Subjects who report smoking in the
past (but not within 6 mo)
Lag 0:5.84 (0.87, 11.05)
2 dma: 5.08 (-2.34, 13.07)
3 dma: 4.44 (-2.70,12.11)
4 dma: 5.02
5 dma: 5.89
-1.78, 12.29
-2.14, 14.58
6 dma: 6.73 (-1.54,15.70)
Subjects who did not report smoking
in the past
Lag 0:6.04 (0.87, 11.48)
2 dma: 6.54 (-1.64, 15.39)
3 dma: 5.86 (-1.90,14.22)
4 dma: 6.11 (-1.18,13.94)
5 dma: 6.89 (-1.42, 15.89)
6 dma: 7.86 (-1.35, 17.94)
VCAM-1-All subjects
Lag 0:9.26 (2.98, 15.91)
2 dma: 10.18 (1.93,19.10)
3 dma: 15.45 (2.70, 29.78)
4 dma: 17.97
5 dma: 23.83
3.63, 34.30
8.41,41.44
6 dma: 27.51 (11.96,45.21)
Subjects not known to be taking
statins
Lag 0:9.19 (3.23, 15.49)
2 dma: 14.64 (5.02, 25.14)
3 dma: 14.39
4 dma: 14.19
5.30, 24.28
5.71,23.36
5 dma: 19.11 (9.44, 29.65)
6 dma: 22.60 (11.79, 34.45)
Subjects who report smoking in the
past (but not within 6 mo)
Lag 0:12.4 (2.77, 22.92)
2 dma: 28.5 (8.38, 52.24)
3 dma: 25.14 (3.50, 51.30)
4 dma: 23.1 (2.70,47.58)
December 2009
E-47
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
5 dma: 32.0 (7.29, 62.30)
6 dma: 31.8 (9.74, 58.26)
Subjects who did not report smoking
in the past
Lag 0:5.15 (-5.63, 17.17)
2 dma: 2.09 (-9.07,14.61)
3 dma: 3.90 (-6.38,15.31)
4 dma: 4.92
5 dma: 7.89
-4.63, 15.43
-1.31, 17.95
6 dma: 10.97 (0.98, 21.96)
vWF-All subjects
Lag 0:7.96 (-4.34, 21.84)
2 dma: 14.87 (-2.85, 35.82)
3 dma: 15.34 (-3.22, 37.45)
4 dma: 15.47 (-7.60, 44.31)
5 dma: 19.50 -8.89,56.74
6 dma: 20.53 -9.80,61.05
Subjects not known to be taking
statins
Lag 0:3.23 (-8.91, 17.00)
2 dma: 9.82 (-8.39, 31.66)
3 dma: 17.79 (-16.03, 65.21)
4 dma: 13.14
5 dma: 16.14
-18.71,57.47)
-20.43, 69.52)
6 dma: 13.25 (-22.09, 64.62)
Subjects who report smoking in the
past (but not within 6 mo)
Lag 0:7.63 (-17.01, 39.58)
2 dma: 37.64 (-7.18, 104.10)
3 dma: 75.41 (6.16, 189.85)
4 dma: 72.05
5 dma: 73.14
-3.34, 206.22)
6.94, 180.32)
6 dma: 71.23 (14.00, 157.19)
Subjects who did not report smoking
in the past
Lag 0:10.22 (-23.14, 58.04)
2 dma: 17.07 (-18.86, 68.91)
3 dma: 6.56 (-42.75, 98.36)
4 dma:-9.20 (-65.79, 140.99)
5 dma:-23.86 (-71.05, 100.29)
6 dma:-48.69 (-77.75, 18.29)
Reference: O'Neill et al. (2005, 0884231
Period of Study:
Baseline period: May 1998-Jan 2000
Time trial: 2000-2002
Location: Boston, MA
Outcome: Changes in vascular
reactivity, specifically percent change in
brachial artery diameter (flow-mediated
and nitroglycerin-mediated)
Pollutant: PM25
Mean (SD): 11.5 (6.4)
Range: 1.1-40.0
N: 270 patients with diabetes or at risk of ./,.„:,....:„„ «»«„..• 1
diabetes, who participated in non-air Monitoring Stations. 1
pollution related studies at the Joselyn
Diabetes Center in Boston
Statistical Analysis: Linear regression
Copollutant:
Sulfates
BC
Ultrafine particle counts
PM Increment: IQR (value not given)
Percent change (96% Cl): PM2 5 6-day
ma
Nitroglycerin-mediated reactivity: -7.6%
(95% Cl: 12.8% to-2.1%)
Notes: PM25 was positively associated
with nitroglycerin-mediated reactivity
an association was also reported with
ultrafine particles. Effect estimates were
larger in type II than type I diabetes. BC
and sulfate increases were associated
with decreased flow-mediated reactivity
among those with diabetes. Although the
largest associations were with the 6-day
ma, similar patterns and quantitatively
similar results appear in the other lags.
December 2009
E-48
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: O'Neill et al. (2007, 0913621
Period of Study: May 1998-Dec 2002
Location: Boston, MA
Outcome: soluble intercellular adhesion Pollutant: S04'
molecule 1 (ICAM-1)
vascular cell adhesion molecule 1
(VCAM-1)
von Willebrand factor (vWF)
Mean Age: 56.6 (10.6)
Study Design: Cross-sectional
N: 92 participants (type 2 diabetic
patients)
Statistical Analyses: Linear regression
Covariates: Apparent temperature,
season, age, race, sex, glycosylated
hemoglobin, cholesterol, smoking
history, BMI
Dose-response Investigated? No
Statistical Package: NR
Averaging Time: 24 h (lagged ma of
days 0 to 1, 2, 3, 4, and 5)
Mean (SD): 3.0 (2.0)
descriptive statistics represent entire
study period
Percentiles: IQR range: 2.2
Range (Min, Max): 0.5, 9.6)
Monitoring Stations: 1 site
Copollutant: PM25, BC, S042"
PM Increment: IQR (specific to lag
period)
Effect Estimate [Lower Cl, Upper Cl]:
% change per IQR of PM2 5
ICAM-1 All subjects
Lag 0: 5.30 (-2.60, 13.83)
2 dma: 4.02 (-3.26,11.85)
3 dma: 4.03 (-5.34, 14.34)
4 dma:-0.79 (-7.30, 6.18)
5 dma: 1.06 (-7.10, 9.93)
6 dma: 3.15 (-5.66, 12.78)
Subjects not known to be taking
statins
Lag 0:10.14 (0.44, 20.77)
2 dma: 9.39 (-1.28, 21.20)
3 dma: 10.93 (-2.23, 25.85)
4 dma:-0.24 (-9.66,10.16)
5 dma: 4.03 -8.66, 18.47
6 dma: 5.66 -7.52,20.72
Subjects who report smoking in the
past (but not within 6 mo)
Lag 0: -4.00 (-24.79, 22.52)
2 dma: -4.82
3 dma:-7.19
-18.01, 10.48)
-23.66, 12.83)
4 dma:-9.8 (-27.96, 12.97)
5 dma:-10.4 (-29.92, 14.44)
6 dma:-6.8 (-25.72, 17.03)
Subjects who did not report smoking
in the past
Lag 0: 6.67 (-4.34, 18.94)
2 dma: 5.65 (-4.67,17.10)
3 dma: 10.21 (-5.83, 28.99)
4 dma: 0.80
5 dma: 2.80
-9.94, 12.83)
-10.85, 18.54)
6 dma: 5.15 (-7.78, 19.89)
VCAM-1 All subjects
Lag 0:-0.04 (-3.75, 3.80)
2 dma: 0.94 (-4.79, 7.01)
3 dma:-0.87 (-3.50,1.82)
4 dma: 0.13 (-2.02, 2.34)
5 dma:-0.47 -2.67,1.78
6 dma:-0.46 -1.99,1.09
Subjects not known to be taking
statins
Lag 0:-1.34 (-11.23, 9.66)
2 dma:-0.19 (-11.13,12.09)
3 dma:-2.84 (-13.90, 9.64)
4 dma: 4.28 (-6.18,15.90)
5 dma:-0.26 (-13.44, 14.93)
6 dma:-3.44 (-16.51,11.67)
Subjects who report smoking in the
past (but not within 6 mo)
Lag 0: 0.07 (-23.40, 30.73)
2 dma:-5.62 (-20.77, 12.43)
3 dma:-26.92 (-33.31 to-19.91)
4 dma: -3.06
5 dma: -6.42
-28.01,30.56)
-30.75, 26.47)
6 dma: -6.46 (-28.55, 22.47)
Subjects who did not report smoking
in the past
Lag 0:-3.28 (-12.66, 7.12)
2 dma:-3.17 (-11.75, 6.23)
3 dma: -9.67
4 dma:-5.51
-22.07, 4.70)
-14.28, 4.15)
5 dma:-12.17 (-22.05 to-1.05)
6dma: -11.77 (-20.95to-1.52)
vWF (sulfate measures not available)
December 2009
E-49
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Park et al. (2008,1568451 Outcome: Total homocysteine (tHcy) Pollutant: PM2
Period of Study: Jan 1995-Jun 2005
Location: Greater Boston area, MA
Mean Age: 73.6 +6.9 yr
Study Design: Cross-sectional and
longitudinal analyses performed
N:960 men
Averaging Time: 24 h (ma up to 7 days
prior to blood collection)
Mean (SD): 12.0 (6.6)
Median: 10.6
Statistical Analyses: Generalized Range (Min, Max): 2.0, 62.0
additive models (also hierarchical mixed-
effects regression models to assess
repeated measures of tHcy)
Monitoring Stations: 1 site
Copollutant:
Covariates: Model 1: season, age, long- PM25_
termtrend, apparent temperature BC(r-0.51)
Model 2: further adjustment for BMI,
systolic blood pressure, smoking status,
pack yr of cigarettes, alcohol consump-
tion
Model 3: further adjustment for serum
creatinine, plasma folate, vitamin B6,
and vitamin B12
Dose-response Investigated? Modeled
continuous covariates as penalized
splines to determine if association with
tHcy was linear
Statistical Package: R software
UL/ (r ~ u.oi j
S04 (r = 0.85)
PM Increment: IQR
Effect Estimate [Lower Cl, Upper Cl]:
Estimated % change in tHcy per IQR
increase in pollutant.
Lag model
Concurrent day. IQR: 7.66
Model 1:1.32 (-0.83, 3.52)
Model 2:1.55 (-0.77, 3.91)
Model 3:1.57 (-0.38, 3.56)
1-day previous. IQR: 6.91
Modell:-1.43 (-3.51, 0.69
Model 2:-1.41 (-3.53, 0.76
Model 3:-1.28 (-3.12, 0.60)
2-day ma. IQR: 6.47
Model 1:0.04 (-2.13, 2.26)
Model 2:-0.07 (-2.26, 2.17)
Model 3: 0.25 (-1.69, 2.22)
3-day ma. IQR: 5.83
Model!:-0.64 (-2.92, 1.69)
Model 2:-0.74 (-3.04, 1.61
Model 3:-0.59 (-2.63, 1.49
4-day ma. IQR: 5.21
Modell:-0.63 (-2.94, 1.72)
Model 2:-0.86 (-3.19, 1.52
Model 3:-0.73 (-2.78, 1.37
5-day ma. IQR: 4.68
Model!:-0.51 (-2.79, 1.83)
Model 2:-0.82 (-3.13, 1.54)
Model 3:-0.84 (-2.85, 1.22)
6-day ma. IQR: 4.50
Model!:-0.91 (-3.32, 1.56)
Model 2:-1.32 (-3.76, 1.17)
Model 3:-1.44 (-3.58, 0.74)
7-day ma. IQR: 4.20
Modell:-0.84 (-3.27, 1.64
Model 2:-1.19 (-3.64, 1.33
Model 3:-1.69 (-3.84, 0.51)
Stratified analyses: No significant
difference in effect of PM25 among those
with high and low levels of vitamins
December 2009
E-50
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Park et al. (2008,1568451 Outcome: Total homocysteine (tHcy) Pollutant: BC
Period of Study: Jan 1995-Jun 2005
Location: Greater Boston area, MA
Mean Age: 73.6 +6.9 yr
Study Design: cross-sectional and
longitudinal analyses performed
N:960 men
Averaging Time: 24 h (ma up to 7 days
prior to blood collection)
Mean (SD): 0.99 (0.56)
Median: 0.87
Statistical Analyses: Generalized Range (Min, Max): 0.07, 3.7
additive models (also hierarchical mixed-
effects regression models to assess
repeated measures of tHcy)
BC
OC(r = 0.0.51)
S04 (r = 0.50)
Monitoring Stations: 1 site
Co pollutant
Covariates: Model 1: season, age, long- ,,nr...,..tjnn\.
term trend, apparent temperature PM2 5 (r = 0 51)
Model 2: further adjustment for BMI,
systolic blood pressure, smoking status,
pack yr of cigarettes, alcohol
consumption
Model 3: further adjustment for serum
creatinine, plasma folate, vitamin B6,
and vitamin B12
Dose-response Investigated? Modeled
continuous covariates as penalized
splines to determine if association with
tHcy was linear
Statistical Package: R software
PM Increment: IQR
Effect Estimate [Lower Cl, Upper Cl]:
Estimated % change in tHcy per IQR
increase in pollutant.
Lag model Concurrent day. IQR: 0.66
Model 1:2.64 (-0.12, 5.48)
Model 2: 2.62
Model 3: 3.13
-0.17, 5.48)
0.76, 5.55)
1-day previous. IQR: 0.66
Model 1:1.46 (-0.98, 3.96)
Model 2:1.32 (-1.14, 3.85)
Model 3: 0.95 (-1.12, 3.05)
2-day ma. IQR: 0.60
Model 1:2.75
Model 2: 2.63
-0.18, 5.76
-0.33, 5.67
Model 3: 2.59 (0.10,5.14)
3-day ma. IQR: 0.57
Model 1:2.95
Model 2: 2.97
-0.44, 6.46
-0.46, 6.51
Model 3: 3.12 (0.21, 6.11)
4-day ma. IQR: 0.52
Model 1:3.94 (0.24, 7.78)
Model 2: 3.76
Model 3: 3.00
0.02, 7.64)
-0.13, 6.22)
5-day ma. IQR: 0.49
Model 1:3.26 (-0.60, 7.27)
Model 2: 2.64 -1.23,6.67
Model 3: 2.38 -0.89, 5.77
6-day ma IQR: 0.44
Model 1:1.63 (-1.99, 5.38)
Model 2:1.03 (-2.62, 4.80)
Model 3: 0.93 (-2.15, 4.11)
7-day ma. IQR: 0.44
Model 1:1.38 (-2.45, 5.36)
Model 2: 0.69 (-3.16, 4.70)
Model 3: 0.45 (-2.81, 3.83)
% change in tHcy per IQR increase in
BC, 24-h avg
Among those with low folate: 5.31 (2.26,
8.42)
Among those with low B12: 5.06 (2.03,
8.17)
nearly null associations among those
with high levels
December 2009
E-51
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Park et al. (2008,1568451 Outcome: Total homocysteine (tHcy) Pollutant: OC
Period of Study: Jan 1995-Jun 2005
Location: Greater Boston area, MA
Mean Age: 73.6 +6.9 yr
Study Design: Cross-sectional and
longitudinal analyses performed
N:960 men
Averaging Time: 24 h (ma up to 7 days
prior to blood collection)
Mean (SD): 3.5 (1.8)
Median: 3.1
Statistical Analyses: Generalized Range (Min, Max): 0.29,11!
additive models (also hierarchical mixed-
effects regression models to assess Monitoring Stations: 1 site
repeated measures of tHcy)
Copollutant (correlation):
Covariates: Model 1: season, age, long- PM2= (r - 0.51)
term trend, apparent temperature BC (r - °'51)
Model 2: further adjustment for BMI, sys- S042" (r = 0.41)
tolic blood pressure, smoking status,
pack yr of cigarettes, alcohol consump-
tion
Model 3: further adjustment for serum
creatinine, plasma folate, vitamin B6,
and vitamin B12
Dose-response Investigated? Modeled
continuous covariates as penalized
splines to determine if association with
tHcy was linear
Statistical Package: R software
PM Increment: IQR
Effect Estimate [Lower Cl, Upper Cl]:
Estimated % change in tHcy per IQR
increase in pollutant.
Lag model
Concurrent day. IQR: NA
Model 1:NA
Model 2: NA
Model 3: NA
1-day previous. IQR: 2.00
Model 1:2.12
Model 2:1.69
-0.98, 5.31
-1.51,5.00
Model 3:1.87 (-0.81, 4.62)
2-day ma. IQR: 1.93
Model!:-0.39 (-3.67, 3.01
Model 2: -0.88 (-4.26, 2.61
Model 3:1.05 (-1.86, 4.06)'
3-day ma. IQR: 1.68
Model 1:0.53 (-2.66, 3.83)
Model 2: 0.14 -3.15,3.54
Model 3:1.32 -1.44,4.16
4-day ma. IQR: 1.64
Model 1:1.57 (-1.89, 5.15)
Model 2:1.42 -2.14,5.12
Model 3:1.89 -1.15,5.03
5-day ma, IQR: 1.60
Model 1:2.27 (-1.49, 6.16)
Model 2: 2.11 (-1.77,6.15)
Model 3: 2.12 (-1.29, 5.65)
6-day ma. IQR: 1.43
Model 1:2.83 (-0.74, 6.52)
Model 2: 2.78 (-0.90, 6.60)
Model 3: 2.53 (-0.59, 5.74)
7-day ma. IQR: 1.23
Model 1:2.75
Model 2: 2.55
-0.41,6.02
-0.71,5.92
Model 3: 2.55 (-0.21, 5.39)
% change in tHcy per IQR increase in
OC, 7-day avg.
Among those with low B12: 5.23 (1.59,
9.01)
Nearly null associations among those
with high levels
December 2009
E-52
-------
Reference
Reference: Park et al. (2005, 0573311
Period of Study: Nov 2000-Oct 2003
Location: Greater Boston area, MA
Design & Methods
Outcome: Change in HRV (SDNN, HF,
LF, LFHFR)
Mean age: 72.7 yr
Study Design: Cross-sectional
N: 497 adult males living in the Greater
Boston, MA area
Concentrations1
Pollutant: PM25
Averaging Time:
4h
24 h
48 h
Mean (SD): 11. 4 (8.0)
Range: 6.45-62.9
Effect Estimates (95% Cl)
PM Increment: 8 pg/m3
Percent change (96% Cl): 48h mean
PM25: 20.8% decrease in HF (95% Cl:
4.6%, 34.2%)
18.6% increase in LFHFR (4.1%,
35.2%).
Notes: Subjects were monitored during
Copollutant:
03, Particle number count, BC, N02,
SO,, CO
a 4-min rest period between 8 a.m. and
1 p.m. Modifying effects of hypertension,
IHD, diabetes, and use of cardiac/anti-
hypertensive medications also
examined. Linear regression analyses.
This subject group is from the VA
Normative Aging Study. The 4-h
averaging period was most strongly
associated with HRV indices. The PM
effect was robust in models including 03.
The HRV change per IQR increase in
PM25 were larger in subjects with
hypertension (n = 335) IHD (n = 142),
and diabetes (n = 72). In addition, those
who did not use calcium-channel
blockers had a greater decline in LF
associated with each IQR increase in
PM25 than did those who did use
calcium channel blockers. IQR increases
in 48h mean BC concentration were also
associated with adverse changes in
HRV, suggesting traffic pollution may be
particularly toxic.
Reference: Park et al. (2006, 0912451
Period of Study: Nov 2000-Dec 2004
Location: Greater Boston area, MA
Outcome: Change in HF
Study Design: Cross-sectional
N: Statistical Analysis: Linear
regression models
Pollutant: PM25
Averaging Time: 48 h
Mean (SD):
PM25: 11.7 (7.8)
Sulfates: 3.3 (3.3)
BC: 0.92 (0.46)
Copollutant: 0;
PM Increment: 10 pg/m3
Percent change (96% Cl): Wild-type
HFE genotype: 31.7% (95% Cl: 10.3,
48.1)
Among those with either of the 2 HFE
variants, there was no association
between 48h PM25 and HF (shown in a
graph, ~10% non-significant increase).
Notes: Normative Aging Study.
Examining association between PM and
HF among those with and without the
wild-type HFE genotype.
Reference: Pekkanen et al. (2002,
0350501
Period of Study: Winter 1998-1999
Location: Helsinki, Finland
Outcome: ST-Segment Depression
(>0.1mV)
Study Design: Panel of ULTRA Study
participants
N: 45 Subjects, n = 342 biweekly
submaximal exercise tests, 72 exercise
induced ST Segment Depressions
Statistical Analysis: Logistic regression
/GAM
Pollutant: PM25
Averaging Time: 24 h
Median: 10.6
IQR: 7.9
Pollutant: PM1
Median: 7.0
IQR: 5.6
Pollutant: ACP (100 to 1000nm) (n/cm3)
Median: 1200
IQR: 760
Copollutant: N02, CO, PM10.25, ultrafine
PM Increment: IQR
Effect Estimate(s): ACP: OR = 3.29
(1.57, 6.92), lag 2
PM,: OR = 4.56 (1.73, 12.03), lag 2
PM2.5: OR = 2.84 (1.42, 5.66), lag 2
Notes: The effect was strongest for ACP
and PM25, which in 2 pollutant models
appeared independent. Increases in N02
and CO were also associated with
increased risk of ST-segment
depression, but not with coarse particles.
December 2009
E-53
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Park et al. (2008,1568451
Period of Study: Jan 1995-Jun 2005
Location: Greater Boston area, MA
Outcome: Total homocysteine (tHcy) Pollutant: S04'
Mean Age: 73.6 +6.9 yr
Study Design: Cross-sectional and
longitudinal analyses performed
N: 960 men
Averaging Time: 24 h (ma up to 7 days
prior to blood collection)
Mean (SD): 3.2 (3.0)
Median: 2.4
Statistical Analyses: Generalized Range (Min, Max): 0.39, 29.0
additive models (also hierarchical mixed-
effects regression models to assess
repeated measures of tHcy)
Covariates: Model 1 : season, age, long-
term trend, apparent temperature
Model 2: further adjustment for BMI,
systolic blood pressure, smoking status,
pack yr of cigarettes, alcohol consump-
tion
Model 3: further adjustment for serum
creatinine, plasma folate, vitamin B6,
and vitamin B12
Dose-response Investigated? Modeled
continuous covariates as penalized
splines to determine if association with
tHcy was linear
Statistical Package: R software
Monitoring Stations: 1 site
Copollutant (correlation):
(r - 0.85)
BC(r-O.SO)
UL/ (f ~ U.41 J
S04
PM Increment: IQR
Effect Estimate [Lower Cl, Upper Cl]:
Estimated % change in tHcy per IQR
increase in pollutant.
Lag model
Concurrent day: IQR: NA
Model 1:NA
Model 2: NA
Model 3: NA
1-day previous: IQR: 2.61
Model 1:0.91
Model 2: 0.99
-0.77, 2.62
-0.94, 2.95
Model 3: 0.91 (-0.72, 2.57)
2-day ma: IQR: 2.10
Model!:-0.25 (-2.07, 1.60
Model 2:-0.29 (-2.35, 1.82
Model 3: 0.05 (-1.74, 1.86)'
3-day ma: IQR: 1.73
Model!:-0.15 (-1.97, 1.69)
Model 2:-0.17 (-2.23, 1.93
Model 3: -0.01 (-1.78,1."
4-day ma: IQR: 1.64
Model!:-0.69 (-2.74, 1.41)
Model 2:-0.60 (-2.95, 1.81
Model 3:-0.58 (-2.63, 1.51
5-day ma: IQR: 1.60
Modell:-1.14 (-3.53, 1.30)
Model 2:-0.90 (-3.64, 1.92)
Model 3:-1.09 (-3.48, 1.36)
6-day ma;' IQR: 1.40
Model 1:0.00 (-2.39, 2.44)
Model 2: 0.36 (-2.36, 3.16)
Model 3: 0.41 (-2.01,2.89)
7-day ma
IQR: 1.30
Modell:-0.16 (-2.51, 2.24)
Model 2: 0.30 (-2.37, 3.04)
Model 3: 0.07 (-2.25, 2.43)
Stratified analyses: No significant
difference in effect of SO?' among those
with high and low levels of vitamins
Reference: Peters et al. (2005, 0957471
Also Peters et al. (2005,1568591
Period of Study: Feb 1999-Jul 2001
Location: Augsburg, Germany
Outcome: Myocardial infarction
Study Design: Case-crossover
N: 691 myocardial infarction patients
Pollutant: PM25
Averaging Time:
1 h: Median = 14.5
IQR: 9.1
Statistical Analysis: Conditional logistic ^"h:,M,edlan = 149
regression u
Dose-response Investigated? No
Copollutant: N02, S02, CO
Effect Estimate: 2-h lag: OR = 0.93
95% 0:0,83,1.04
24-h mean, 2-day lag: OR = 1.18
95% Cl: 1.03, 1.34
Notes: Examined triggering for Ml at
various lags before Ml onset (up to 6 h
before Ml, up to 5 days before Ml). PM2 5
levels 2 days before Ml onset were
associated with increased risk of Ml, but
not on the concurrent day, or lags 1, 3,
4, or 5. These findings are consistent
with the prior Boston Ml study for a 1- to
2-day lagged effect of PM2 5.
December 2009
E-54
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Pope et al. (2004, 0552381
Period of Study: Winter 1999-2000 (in
Wasatch Front, UT). Summer 2000 (in
Hawthorne, UT).
Winter 2000-2001 (in Bountiful, UT and
Lindon, UT)
Location: Utah: Wasatch Front,
Hawthorne, Bountiful, and Lindon
Outcome: Change in autonomic Pollutant: PM25 (TEOM)
function (measured by changes in HRV),
C-reactive protein (CRP), blood cell Averaging Time: 24 h
counts, platelets, and blood viscosity
associated with short-term changes in
PM25
Mean (SD): 18.9 (13.4)
Copollutant: None
Age Groups: Elderly (specific age range
not given)
Study Design: Panel study
N: 88 elderly subjects
Statistical Analysis: Linear regression
Season: Wnter, summer
Dose-response Investigated? No
PM Increment: 100 pg/m
Effect Estimate: Each 100 pg/m3
increase associated with: -35 (SE = 8)
msec decline in SDNN
0.81 (SE 0.17) mg/dL increase in CRP
0.31 (SE 9.34) k/pL increase in platelets
0.07 (SE 0.21) cP increase in blood
viscosity
Notes: The study observed small but
statistically significant adverse
associations between daily mean PM25
and HRV and C-reactive protein (CRP).
The authors point out, however, that
most of the variability in the temporal
deviation of these physiological
endpoints was not explained by PM2 5.
These observations therefore suggest
that PM25 may be 1 of multiple factors
that influence HRV and CRP.
Reference: Pope et al. (2006, 0912461 Outcome: Acute ischemic heart disease Pollutant: PM25 (FRM)
Period of Study: 1994-2004
Location: Wasatch Front, Utah
Study Design: Case-crossover study
(time-stratified control selection)
N: Statistical Analysis: Conditional
logistic regression
Averaging Time: 24 h
Mean (SD): Site 1:10.1
Site 2:10.8
Site 3:11.3
Monitoring Stations: 3
Copollutant: PM10 (FRM) measured at
4 monitoring sites
PM Increment: 10 pg/m
Effect Estimate: For same-day increase
inPM25:OR = 1.045
95% Cl: 1.011, 1.080
Notes: Case-crossover study (time-
stratified control selection) triggering of
acute ischemic heart disease by ambient
PM25 concentrations on the same and
previous 3 days. PM25 measured at 3
sites and estimated for missing days.
Effect estimates were larger for those
with angiographically demonstrated
coronary artery disease.
Reference: Pope et al. (2004, 0552381
Period of Study: 1999-2001
Location: Wasatch Front, Utah
Outcome: Heart rate variability (HRV)
C-reactive protein (CRP)
Blood cell counts, whole blood viscosity
Age Groups: 54-89 yr
Study Design: Panel study
N: 88 participants
Statistical Analyses: Linear regression
Covariates: Subject-specific fixed
effects
Interactive spline smooths for temp, RH
(partial control for H)
Season: Temperature as covariate
Dose-response Investigated?
Yes, also assessed PM by including
cubic smoothing splines with 3 df
Statistical Package: SAS
Pollutant: PM25
Averaging Time: 24 h
Mean (SD): 23.7 (20.2)
Range (Min, Max): 1.7, 74.0
Monitoring Stations: NR
Copollutant: None
PM Increment: 100 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
Regression coefficients (SE) for
associations with concurrent day
pollutant: Mean H:-4.49 (1.73)
SDNN: -34.94 (8.32)
SDANN:-18.98 (8.67)
r-MSSD: -42.25 (10.90)
CRP: 0.81 (0.18)
Whole blood viscosity: 0.07 (0.21)
WBC: -0.07 (0.38)
Granulocytes: 0.02 (0.37)
Lymphocytes:-0.07 (0.14)
Monocytes:0.12(0.04)
Basophils:-0.01 (0.01)
Eosinophils: -0.01 (0.02)
RBC: 0.03 (0.06)
Platelets: 0.31 (9.34)
Reference: Rich et al. (2005, 0796201
Period of Study: Jul1995-Jul 2002
Location: Eastern Massachusetts, USA
Outcome: Confirmed ventricular
arrhythmias
Study Design: Case-crossover (time-
stratified control selection)
N: 203 patients with implantable
cardioverter defibrillators
Statistical Analysis: Conditional logistic
regression
Pollutant: PM25 (TEOM)
Averaging Time: 1-h avg
24-h avg
Median (IQR):
1-h avg: Median = 9.2 pg/m3
24-h avg: Median = 9.8 pg/m3
IQr = 7.8
Copollutant: 03, BC, CO, N02, S02
PM Increment: 7.8 pg/m
Effect Estimate: For mean PM25 in the
24 h before ventricular arrhythmia: OR
= 1.19
95% 0:1.02,1.38
Notes: 794 ventricular arrhythmias
among 84 subjects.
Lag h: 0-2, 0-6, 0-23, 0-47
December 2009
E-55
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Rich et al. (2006, 0884271
Period of Study: Jul 1995-Jul 2002
Location: Eastern Massachusetts, USA
Outcome: Confirmed episodes of
paroxysmal atrial fibrillation
Study Design: Case-crossover (time-
stratified control selection)
N: 203 patients with implantable
cardioverter defibrillators
Statistical Analysis: Conditional logistic
regression
Pollutant: PM25 (TEOM)
Averaging Time: 1-h avg
24-h avg
Median (IQR):
1-h avg: Median = 9.2 pg/m3
24-h avg: Median = 9.8 pg/m3
IQr = 7.8
Copollutant: 03, BC, CO, N02, S02
PM Increment: 9.4 pg/m
Effect Estimate: 0-h lag: OR 1.41 (0.82,
2.42)
Notes: 91 paroxysmal atrial fibrillation
(PAF) episodes among 29 subjects.
Lag h: 0, 0-23
Positive, but not significant increases in
the relative odds of PAF associated with
PM25 concentrations in the same h and
24-h before PAF episode onset. Authors
note reduced statistical power for PM25
analyses due to missing data.
Reference: Rich et al. (2006, 0884271
Period of Study: Jul 1995-Jul 2002
Location: Eastern Massachusetts, USA
Outcome: Confirmed episodes of
paroxysmal atrial fibrillation
Study Design: Case-crossover (time-
stratified control selection)
N: 203 patients with implantable
cardioverter defibrillators
Statistical Analysis: Conditional logistic
regression
Pollutant: BC
Averaging Time: 1-h avg, 24-h avg
Median (IQR): IQR: 0.91 pg/m3
Copollutant: 03, PM25, CO, N02, S02
PM Increment: 0.91 pg/m3 (IQR)
Effect Estimate: 0- to 23-h lag period:
OR 1.46 (95% 0:0.67,3.17)
Notes: 91 paroxysmal atrial fibrillation
(PAF) episodes among 29 subjects.
Lag h: 0, 0-23
Positive, but not significant increases in
the relative odds of PAF associated with
BC concentrations in the same h and 24
h before PAF episode onset. Authors
note reduced statistical power for BC
analyses due to missing data.
Reference: Rich et al. (2006, 0898141
Period of Study: May 2001-Dec 2002
Location: St. Louis, MO metropolitan
area
Outcome: Confirmed ventricular
arrhythmia
Study Design: Case-crossover design
(time-stratified control selection)
Dose-response Investigated? No
Pollutant: PM25 (CAMM)
Averaging Time: 24 h
Median (IQR): 16.2 |jg/m3(IQr = 9.7)
Copollutant: N02, S02, CO, 03, EC, OC
PM Increment: 9.7 pg/rri (IQR)
Effect Estimate: OR (PM25) = 0.95
(95% Cl: 0.72, 1.27)
OR (S02) = OR = 1.24 (95% Cl: 1.07,
1.44)
Notes: 139 confirmed ventricular
arrhythmia episodes among 56 subjects.
Lags:0-2h, 0-6h, 0-11h, 0-23h, 0-47h
Authors did not find increased relative
odds of VA associated with each IQR
increase in 24-h mean PM25, but did find
non-significantly increased relative odds
of VA associated with 24-h EC. Shorter
and longer lag times' relative odds
estimates provided no evidence of
immediate ventricular arrhythmic effects
of air pollution.
Reference: Rich et al. (2004, 0556311
Period of Study: Feb-Dec 2000
Location: Vancouver, British Columbia,
Canada
Outcome: ICD discharges (as a proxy
forVTA/F)
Age Groups: 15-85yr
Study Design: Case-crossover design
(ambidirectional control selection + 7
days)
N: 34 patients with implantable
cardioverter defibrillators
Statistical Analysis: Conditional logistic
regression
Dose-response Investigated? No
Pollutant: PM25 (Partisol)
Averaging Time: 1 h
Mean (SD), IQR:
Mean:: 8.2 pg/m3(SD= 10.7)
IQr = 5.2
Copollutant: 03, EC, OC, S042", CO,
N02, S02, PM,o
PM10: Mean:: 13.3 pg/m3
(SD = 4.9)
IQr = 7.4
PM Increment: Effect Estimate: Odds
ratios were less than 1.0 at all lags (0,1,
2, 3)forPM25.
No consistent association between any
of the air pollutants and implantable
cardioverter defibrillators discharges.
Notes: Same study as Vedal et al.
(2004, 0556301, except Rich (2004)
used data from a shorter time period so
as to estimate relative odds of ICD
discharge associated with acute
increases in more pollutants than Vedal
(2004, 0556301.
December 2009
E-56
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Rich et al. (2008,1569101
Period of Study: NR
Location: New Jersey
Outcome: Pulmonary Artery and Right
Ventricular Pressures
Age Groups: 25-68
Study Design: Panel
N: 11 subjects
Statistical Analyses: Repeated
Measures
Covariates: Long-term trends, calendar
month, weekday, apparent temperature
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered :0-6d
Pollutant: PM25
Averaging Time: 24 h
Mean (SD): NR
Monitoring Stations: NR
Copollutant: NR
Co-pollutant Correlation: N/A
PM Increment: 11.62|jg/m
Change (Lower Cl, Upper Cl), p-value:
ePAD:0.19(0.05, 0.33), 0.01
RV diastolic pressure: 0.23 (0.11, 0.34),
<0.001
RV systolic pressure: 0.12 (-0.07, 0.31),
0.23
MPAP: 0.12 (-0.05, 0.28), 0.16/
Reference: Riediker et al. (2004,
0912611
Period of Study: Fall 2001
Location: Wake County, North Carolina
Outcome: Heart rate variability
(measured 10 h after shift): mean cycle
length of normal R-R intervals (MCL),
the standard deviation of normal R-R
intervals (SDNN), and percentage of
normal R-R interval differences greater
than 50 msec (PNN50), low frequency
(0.04-0.15Hz), high frequency (0.15-
0.40Hz), the ratio of low to high
frequency.
Blood analysis (measured 16 h after
shift): Uric acid, blood urea nitrogen,
gamma glutamyl transpeptidase, white
blood cell count, red blood cell count,
hematocrit, hemoglobin, mean red blood
cell volume (MCV), neutronphils (count
and %), lymphocytes (count and %), C-
reactive protein, plasminogen,
plasminogen activator inhibitor type 1,
von Willebrand factor (vWF), endo-
vthzelin-1, protein C, and interleukin-6
Age Groups: 23-30 yr
Study Design: Panel
N: 9 healthy male troopers, repeated
measures (36 person-days)
Statistical Analyses: Mixed effects
regression models (principal factor
analysis for classification of exposure)
Covariates: Potential confounders:
temperature, relative humidity, number
of law-enforcement activities during the
shift and the avg speed during the shift
Controlling had no effect on effect esti-
mates for "crustal" and "speed-change"
factors
However, confounder inclusion in the
"speed change" and blood urea nitrogen
and vWF reduced the effect estimate
and the Cl included zero
Season: Only 1 season included
Dose-response Investigated? No
Statistical Package: S-Plus 6.1
Pollutant: In-vehicle PM25 components
identified with factor analysis (crustal
material, wear of steel automotive
components, gasoline combustion,
speed-changing traffic with engine
emissions and brake wear
Averaging Time: Exposure assessed
during 3 p.m. to 12a.m. work shifts
Mean: PM25mass= 23.0 pg/m3
Monitoring Stations: Per vehicle
Copollutant (correlation): Correlation
to PM25Mass
Benzene: r = 0.50
Aldehydes: r = 0.34
CO: r = 0.52
Aluminum: r = 0.58
Silicon: r = 0.66
Sulfur: r = 0.58
Calcium: r = 0.37
Titanium: r = 0.41
Chromium: r = 0.51
Iron: r = 0.71
Copper: r = 0.16
Selenium: r = 0.38
Tungsten: r = 0.37
PM2.Lightscatter:r = 0.71
PM Increment: 1 SD change in source
factor
Effect Estimate: % change in the health
outcome per 1 SD change in the "speed
change" factor
MCL: 7%
HRV: 16%
supraventricular ectopic beats: 39%
% Neutrophils: 7%
% lymphocytes:-10%
red blood cell volume MCV: 1%
vWF: 9%
blood urea nitrogen: 7%
protein C:-11%
% change in the health outcome per 1
SD change in the "crustal" factor
MCL: 3% serum uric acid
concentrations: 5%
Note: Results (including CIs) are
reported in figures 2 & 3.
December 2009
E-57
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Riojas-Rodriguez et al.
(2006, 1569131
Outcome: Heart rate variability (5-
minute periods)
Period of Study: Dec 2001 -Apr 2002 Study Design: Panel study
Location: Mexico City metropolitan area N: 30 patients from the outpatient clinic
of the National Institute of Cardiology of
Mexico, where each subject had existing
ischemic heart disease.
Statistical Analysis: Mixed models
Pollutant: PM25 (nephelometry)
Averaging Time: 5 min
Mean (SD), Range:
46.8 pg/rri (SD= 1.82)
Range: 0-483 pg/m3
Copollutant: CO
PM Increment: 10 pg/m
Effect Estimate: Each 20 pg/m3
increase in 5 min PM25 was associated
with a: -0.008 decrease in the
ln(HF)(95%CI: -0.015, 0.0004
Notes: Population of subjects with
known ischemic heart disease (25 men
and 5 women who had at least 1 prior Ml
[not in last 6 mo])
Each 10 pg/m3 increase in 5-min mean
PM25 was associated with non-
significantly decreased HF, and with
similar, but smaller changes in LF and
VLF
Reference: Romieu et al. (2005,
0862971
Period of Study: 2000-2001
Location: Mexico City, Mexico
Outcome: Heart rate variability (HF, LF, Pollutant: PM25
VLF, PNN50, SDNN, r-MSSD)
Averaging Time: 24 h
Age Groups: >60 yr of age
a K * a Copollutant: 03, N02, S02, PM1(
Study Design: Double blind randomized
controlled trial
N: 50 elderly residents of a Mexico City
nursing home
PM Increment: 8 pg/m
Effect Estimate: In the group receiving
the fish oil supplement, each 8 pg/m3
change in 24-h mean total exposure
PM25 was associated with a: a) 54%
reduction (95% Cl: -72% to -24%) in HF
(log transformed) in the pre-
supplementation phase
b) 7% reduction (95% Cl: -20%, 7%) in
the supplementation phase.
Changes in other HRV parameters were
also smaller in the supplementation
phase. In the group receiving soy oil
supplementation, the % reduction in HF
was also smaller in the supplementation
phase, but the differences were smaller
and not statistically significant.
Notes: Study of the effect of omega-3-
fatty acid supplementation (2 g/day of
fish oil vs.. 2 g/day of soy oil) to mitigate
the effect of ambient PM25 on HRV.
Subjects had no cardiac arrhythmias,
cardiac pacemakers, allergies to omega-
3 fatty acids or fish, treatment with oral
anticoagulants, or history of bleeding
diathesis. PM25 was measured and
estimated indoors, outdoors, and with
regards to total exposure (the same as
Holguin et al. (2003)).
December 2009
E-58
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Romieu et al. (2008,
1569221
Period of Study: Sep 2001-Apr 2002
Location: Mexico City, Mexico
Outcome: Copper/zinc superoxide
dismutase activity (Cu/Zn SOD)
Lipoperoxidation (LPO)
Reduced glutathione (GSH)
Age Groups: 60-96 yr
Study Design: Intervention (randomly
assigned fish oil or soy oil)
N: 52 participants
Statistical Analyses: Linear mixed
models
Covariates: Time
Dose-response Investigated?
Assessed possible nonlinearity using
generalized additive mixed models with
p-splines
Statistical Package: STATAv8.2 and
SASv9.1
Pollutant: PM25 (indoor)
Averaging Time: 24 h (same day)
Mean (SD): 38.7 (14.7)
Percentiles:
25th: 30.62
50th: 35.11
75th: 41.10
Range (Min, Max): 14.8, 70.9
Monitoring Stations: Indoor measured
inside nursing home
Copollutant: 0;
PM Increment: 10 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
Regression coefficient (SE
p-value):
Cu/Zn SOD: -0.05 (0.02
0.001)
LPO (square root transformed): 0.08
(0.09
0.381)
GSH (log-transformed
quadratic term for PM): -0.05 (0.01
0.002)
Regression coefficient (SE
p-value) by supplementation groups
(same transformations as above): Cu/Zn
SOD
Soy Oil: -0.06 (0.02, O.001)
Fish Oil:* 0.04 (0.02, 0.009)
LPO
Soy Oil:-0.02 (0.14, 0.904)
Fish Oil: *0.16 (0.07, 0.024)
GSH
Soy Oil:-0.03 (0.04, 0.406)
Fish Oil:-0.09 (0.04, 0.017)
'Quadratic term for PM
Reference: Ruckerl et al. (2007,
1569311
Outcome: lnterleukin-6 (IL-6),
fibrinogen, C-reactive protein (CRP)
Period of Study: May 2003-Jul 2004 Age Groups: 35-80 yr
Location:
Athens, Augsburg, Barcelona, Helsinki,
Rome, and Stockholm
Study Design: Repeated measures /
longitudinal
N: 1003 Ml survivors
Statistical Analyses: Mixed-effect
models
Covariates: City-specific confounders
(age, sex, BMI)
Long-term time trend and apparent
temperature
RH, time of day, day of week included if
adjustment improved model fit
Season: Long-term time trend
Dose-response Investigated? Used p-
splines to allow for nonparametric
exposure-response functions
Statistical Package: SASv91
Pollutant: PM25
Averaging Time: Hourly and 24-h (lag
0-4, mean of lags 0-4, mean of lags 0-1,
mean of lags 2-3, means of lags 0-3)
Mean (SD): Presented by city only
Monitoring Stations: Central
monitoring sites in each city
Copollutant:
S02
03
NO
N02
PM Increment: IQR
Effect Estimate [Lower Cl, Upper Cl]:
% change in mean blood markers per
increase in IQR of air pollutant.
IL-6
Lag (IQR): % change in GM (95%CI)
Lag 0(11.0): 0.46 (-0.89, 1.83)
Lag1 (11.0):-0.39
Lag 2 (11.0):-0.23
-1.69,0.93)
-1.53, 1.07)
5-day avg (8.6): 0.05 (-1.37,1.50)
Fibrinogen
Lag (IQR): % change in AM (95%CI)
Lag 0(11.0): 0.05 (-0.48, 0.58
Lag 1(11.0): 0.17 (-0.35, 0.69
Lag 2 (11.0): 0.20 (-0.32, 0.71)
5-day avg (8.6): 0.38 (-0.21, 0.96)
CRP
Lag (IQR): % change in GM (95%CI)
Lag 0(11.0): 0.11 (-1.95, 2.21)
Lag1 (11.0):-0.06 (-1.98,1.90)
Lag 2 (11.0): 0.11 (-1.80, 2.06)
5-day avg (8.6):-0.13 (-2.15,1.92)
December 2009
E-59
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ruckerl et al. (2006,
0887541
Period of Study: Oct 2000-Apr 2001
Location: Erfurt, Germany
Outcome: C-reactive protein (CRP)
serum amyloid A (SAA)
E-selectin
von Willebrand Factor (vWF)
intercellular adhesion molecule-1 (ICAM-
1)
fibrinogen
Factor VII
prothrombin fragment 1+2
D-dimer
Age Groups: 50+
Study Design: Panel (12 repeated
measures at 2-wk intervals)
N: 57 male subjects with coronary
disease
Statistical Analyses: Fixed effects
linear and logistic regression models
Covariates: Models adjusted for
different factors based on health
endpoint
CRP: RH, temperature, trend, ID
ICAM-1: temperature, trend, ID
vWF: air pressure, RH, temperature,
trend, ID
FVII: air pressure, RH, temperature,
trend, ID, weekday
Season: Time trend as covariate
Dose-response Investigated?
Sensitivity analyses examined nonlinear
exposure-response functions
Statistical Package: SAS v8.2 and S-
piusve.o
Pollutant: PM25
Averaging Time: 24 h
Mean (SD): 20.0 (15.0)
Percentiles:
25th: 9.7
50th: 14.9
75th: 26.1
Range (Min, Max): 2.6, 83.7
Monitoring Stations: 1 site
Copollutant:
UFPs
AP
PM25
PM10
OC
EC
N02
CO
PM Increment: IQR (16.4
5-day avg: 12.2)
Effect Estimate [Lower Cl, Upper Cl]:
Effects of air pollution on blood markers
presented as OR (95%CI) for an
increase in the blood marker above the
90th percentile per increase in IQR air
pollutant.
CRP
Time before draw:0to 23 h: 1.1 (0.7,
1.8)
24-47 h: 1.5 (0.9, 2.5)
48-71 h: 1.2 (0.8, 1.9)
5-day mean: 1.4 (0.9, 2.3)
ICAM-1
Time before draw: 0-23 h: 0.7 (0.4, 0.9)
24-47 h: 1.3 (0.8, 1.8
48-71 h: 1.8 (1.2, 2.7
5-day mean: 1.1 (0.8,1.5)
Effects of air pollution on blood markers
presented as % change from the
mean/GM in the blood marker per
increase in IQR air pollutant.
vWF
Time before draw: 0-23 h: 3.9 (-0.3, 8.1)
24-47 h: 3.1 (-1.6, 7.8)
48-71 h: 3.6 (-1.1, 8.3)
5-day mean: 5.6 (0.5,10.8)
FVII
Time before draw: 0-23 h: -2.5 (-6.2 to
1.4)
24-47 h:-2.8 (-6.1 to 0.6)
48-71 h:-2.3 (-5.0 to 0.6)
5-day mean:-3.5 (-6.4 to-0.4)
Note: Summary of results presented in
figures. SAA results indicate increase in
association with PM (not as strong and
consistent as with CRP)
No association observed between E-
selectin and PM
An increase in prothrombin fragment
1+2 was consistently observed,
particularly with lag 4
Fibrinogen results revealed few
significant associations, potentially due
to chance
D-dimer results revealed null
associations in linear and logistic
analyses
December 2009
E-60
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ruckerl et al. (2006,
0887541
Period of Study: Oct 2000-Apr 2001
Location: Erfurt, Germany
Outcome: C-reactive protein (CRP)
serum amyloid A (SAA)
E-selectin
von Willebrand Factor (vWF)
intercellular adhesion molecule-1
(ICAM-1)
fibrinogen
Factor VII
prothrombin fragment 1+2
D-dimer
Age Groups: 50+ yr
Study Design: Panel (12 repeated
measures at 2-wk intervals)
N: 57 male subjects with coronary
disease
Statistical Analyses: Fixed effects
linear and logistic regression models
Covariates: Models adjusted for
different factors based on health
endpoint
CRP: RH, temperature, trend, ID
ICAM-1: temperature, trend, ID
vWF: air pressure, RH, temperature,
trend, ID
FVII: air pressure, RH, temperature,
trend, ID, weekday
Season: Time trend as covariate
Dose-response Investigated?
Sensitivity analyses examined nonlinear
exposure-response functions
Statistical Package: SAS v8.2 and S-
piusve.o
Pollutant: EC
Averaging Time: 24 h
Mean (SD): 2.6 (2.4)
Percentiles:
25th: 1.0
50th: 1.8
75th: 3.2
Range (Min, Max): 0.2,12.4
Monitoring Stations: 1 site
Copollutant:
UFPs
AP
PM25
PM10
OC
EC
N02
CO
PM Increment: IQR (2.3
5-day avg: 1.8)
Effect Estimate [Lower Cl, Upper Cl]:
Effects of air pollution on blood markers
presented as OR (95%CI) for an
increase in the blood marker above the
90th percentile per increase in IQR air
pollutant.
CRP
Time before draw: 0-23 h: 1.2 (0.7, 2.0)
24-47 h: 1.3 (0.7, 2.4)
48-71 h: 1.6 (0.9, 2.7)
5-day mean: 1.2 (0.7, 2.1)
ICAM-1
Time before draw: 0-23 h: 1.0 (0.7,1.6)
24-47 h: 2.6 (1.7, 3.8)
48-71 h: 4.0 (2.5, 6.1)
5-day mean: 2.2 (1.4, 3.3)
Effects of air pollution on blood markers
presented as % change from the
mean/GM in the blood marker per
increase in IQR air pollutant.
vWF
Time before draw: 0-23 h: 5.0 (0.0,10.1)
24-47 h: 7.6 (1.4, 13.7)
48-71 h:1.1 (-5.2,7.4)
5-day mean: 5.7 (-0.5,12.0)
FVII
Time before draw: 0-23 h: -5.7 (-10.5 to -
0.7)
24-47 h:-6.9 (-11.2 to-2.3)
48-71 h: -4.2 (-8.4, 0.2)
5-day mean:-6.0(-10.5to-1.2)
Note: Summary of results presented in
figures. SAA results indicate increase in
association with PM (not as strong and
consistent as with CRP)
No association observed between
E-selectin and PM
An increase in prothrombin fragment
1+2 was consistently observed,
particularly with lag 4
Fibrinogen results revealed few
significant associations, potentially due
to chance
D-dimer results revealed null
associations in linear and logistic
analyses
December 2009
E-61
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ruckerl et al. (2006,
0887541
Period of Study:
Oct 2000-Apr 2001
Location: Erfurt, Germany
Outcome (ICD9 and ICD10): C-reactive Pollutant: OC
protein (CRP) . . „
Serum amyloid A (SAA) Averaging Time: 24 h
v'onw'febrand Factor (vWF) Mean (SD): 1.5(0.6)
intercellular adhesion molecule-1 (ICAM- percentiles:
1) 25th: 1.1
Fibrinogen 50th: 1.4
FactorVII 75th: 1.8
Prothrombin fragment 1+2
D-dimer Range (Min, Max): 0.3, 3.4
Age Groups: 50+ yr
Study Design: Panel (12 repeated
measures at 2-wk intervals)
N: 57 male subjects with coronary
disease
Statistical Analyses: Fixed effects
linear and logistic regression models
Covariates: Models adjusted for
different factors based on health
endpoint
CRP: RH, temperature, trend, ID
ICAM-1: temperature, trend, ID
vWF: air pressure, RH, temperature,
trend, ID
FVII: air pressure, RH, temperature,
trend, ID, weekday
Season: Time trend as covariate
Dose-response Investigated?
Sensitivity analyses examined nonlinear
exposure-response functions
Statistical Package: SAS v8.2 and S-
piusve.o
Monitoring Stations: 1 site
Copollutant:
UFPs
AP
PM25
PM10
OC
EC
N02
CO
PM Increment: IQR (0.7
5-day avg: 0.5)
Effect Estimate [Lower Cl, Upper Cl]:
Effects of air pollution on blood markers
presented as OR (95%CI) for an
increase in the blood marker above the
90th percentile per increase in IQR air
pollutant.
CRP
Time before draw: 0-23 h: 1.2 (0.7,1.9)
24-47 h: 1.3 (0.8, 2.1)
48-71 h: 1.4 (0.8, 2.4)
5-day mean: 1.2(0.7,1.8)
ICAM-1
Time before draw: 0-23 h: 0.9 (0.6,1.3)
24-47 h: 2.0 (1.3, 3.2)
48-71 h: 3.0 (1.8, 4.8)
5-day mean: 1.3(0.8,2.0)
Effects of air pollution on blood markers
presented as % change from the
mean/GM in the blood marker per
increase in IQR air pollutant.
vWF
Time before draw: 0-23 h: 5.5 (0.2,10.8)
24-47 h: 8.0 (2.1, 13.9)
48-71 h: 3.5 (-2.6, 9.6)
5-day mean: 7.4 (2.0,12.8)
FVII
Time before draw: 0-23 h: -6.1 (-10.6 to -
1.4)
24-47 h:-7.2 (-11.4 to-2.8)
48-71 h: -3.8 (-8.2, 0.9)
5-day mean:-5.6 (-9.8 to-1.1)
Note: Summary of results presented in
figures. SAA results indicate increase in
association with PM (not as strong and
consistent as with CRP)
No association observed between E-
selectin and PM
An increase in prothrombin fragment
1+2 was consistently observed,
particularly with lag 4
Fibrinogen results revealed few
significant associations, potentially due
to chance
D-dimer results revealed null
associations in linear and logistic
analyses
December 2009
E-62
-------
Reference
Reference: Ruckerl et al. (2007,
0913791
Period of Study: Oct 2000-Apr 2001
Location: Erfurt, Germany
Design & Methods
Outcome: Soluble CD40 ligand
(sCD40L), platelets, leukocytes,
erythrocytes, hemoglobin
Age Groups: 50+ yr
Study Design: Panel (12 repeated
measures at 2-wk intervals)
N: 57 male subjects with coronary
disease
Statistical Analyses: Fixed effects
linear regression models
Covariates: Long-term time trend,
weekday of the visit, temperature, RH,
barometric pressure
Season: Time trend as covariate
Dose-response Investigated? No
Concentrations1
Pollutant: PM25
Averaging Time: 24 h
Mean (SD): 20.0 (15.0)
Percentiles:
25th: 9.7
50th: 14.9
75th: 26.1
Range (Min, Max): 2.6, 83.7
Monitoring Stations: 1 site
Co pollutants:
UFPs
AP
PM
r IVI2.5
PM,o
MO
INW
Effect Estimates (95% Cl)
PM Increment: IQR (16.4
5-day avg: 12.2)
Effect Estimate [Lower Cl, Upper Cl]:
Effects of air pollution on blood markers
presented as % change from the
mean/GM in the blood marker per
increase in IQR air pollutant.
SCD40L, % change GM (pg/mL)
lagO' 1 5 (-4 0 73)
Lag 1:0. 2 (-5. 4, 6.2)
Lag2: -2.6 (-8.0, 3.1)
Lag3: 0.5 (-3.9, 5.0)
5-day mean: 0.2 (-5.4, 6.2)
Platelets, % change mean (103/ul)
LagO: -0.6 (-1.9, 0.7)
Lag1: 0.1 (-1.3, 1.5)
Lag2:0.5 -0.9, 1.9
Lag3:0.2 -1.1,1.5
(T j— i, «.._«. n A i A r\ A o\
Statistical Package: SAS v8.2 and S-
piusve.o
5-day mean:-0.4 (-1.9,1.2)
Leukocytes, % change in mean
(103/ul)
LagO:-1.6 (-3.2, 0.0)
Lag 1:-0.4 (-2.2, 1.4)
Lag2:-0.2 -2.1, 1.7
Lag3: -0.8 (-2.4, 0.7
5-day mean:-1.6 (-3.5, 0.3)
Erythrocytes, % change mean (106/ul)
LagO:-0.1 (-0.5,0.3)
Lag 1:-0.3 (-0.7, 0.2
Lag2: -0.4 (-0.8, 0.0
Lag3:-0.2 (-0.5, 0.1)
5-day mean: -0.4 (-0.8, 0.0)
Hemoglobin, % change mean (g/dl)
LagO: 0.0 (-0.6, 0.5)
Lag 1:-0.2 (-0.8, 0.3)
Lag2:-0.5 -1.1, 0.0
Lag3: -0.2 (-0.7, 0.2
5-day mean:-0.5 (-1.0, 0.1)
Reference: Sarnat et al. (2006, 0904891 Outcome: Supraventricular ectopy
(SVE) or ventricular ectopy (VE)
Period of Study: Summer and fall 2000
Location: Steubenville, OH
N: 32 nonsmoking older adults
Statistical Analysis: Logistic mixed
effects regression
Season: Summer and fall
Dose-response Investigated? No
Pollutant: PM25
Averaging Time: 5 days
Median (IQR): PM25: Median: 19.0
|jg/m3
IQr=10.0
Sulfate: Median: 6.1. IQR: 4.2
EC: Median: 0.9. IQR: 0.5
Copollutants: 03, N02, S02
PM Increment: IQR
Effect Estimate: PM25: SVE: OR = 1.42
(95% 01:0.99, 2.04)
VE: OR =1.02 (95% Cl: 0.63-1.65)
Sulfate: SVE: OR =1.70 (95% 01:1.12,
2.57)
VE: OR =1.08 (95% 01:0.65, 1.80)
EC: SVE: OR = 1.15 (95% 01:0.73,
1.81)
VE: OR =1.00 (95% 01:0.57, 1.75)
Notes: Longitudinal study of 32
nonsmoking older adults who had ECG
measurements made every week for 24
wk. PM measured within 1 mile of
subjects' residences, and central site
pollutant measurements were also
made.
December 2009
E-63
-------
Reference
Reference: Schneider et al. (2008,
1919851
Period of Study: Nov 2004-Dec 2005
Location: Chapel Hill, NC
Design & Methods
Outcome: Endothelial Function
Parameters
Age Groups: 48-80 yr
Study Design: Panel
N: 22 diabetics
Statistical Analyses: Mixed Models
Covariates: Season, day of the week,
temperature, relative humidity,
barometric pressure
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-4 days; 5-day ma
Concentrations1
Pollutant: PM25
Averaging Time: Daily
Mean (SD): 13.6 (7.0)
Min:2.0
Max: 38.9
Monitoring Stations: 2
Copollutant: NR
Effect Estimates (95% Cl)
PM Increment: 10 pg/m3
Percent Change: (Lower Cl, Upper
Cl), lag:
cyr> AI
rlvlU. I
-1 7.3 (-34.6, 0.0, lag 0
-4.4 (-24.6, 15.8, lag 1
-18.6 (-44.8, 7.6), lag 2
1.6 (-23.6, 26.9), lag 3
18.4 (-3.5, 40.3), lag 4
-19.4 (-62.6, 23.8), 5-day ma
NTGMD:
2.5 (-9.0, 13.9), lag 0
-13.6 (-24.5, -2.6), lag 1*
-10.2 (-23.5, 3.0), lag 2
-8.0 (-22.4, 6.4), lag 3
3.6 (-7.9, 15.0), lag 4
-19.4 (-44.3, 5.5), 5-day ma
LAEI:
0.4 (-4.2, 5.0), lag 0
-0.3 (-6.0, 5.4), lag 1
2.5 (-4.3, 9.4), lag 2
-7.3 (-13.5, -1.1), Iag3*
-2.3 (-8.0, 3.3), lag 4
-4.6 (-15.3, 6.1), 5-day ma
SAEI:
-3.0 (-13.0, 7.0), lag 0
-17.0 (-27.5, -6.4), Iag1**
-9.7 (-23.5, 4.2), lag 2
-15.1 (-29.3,-0.9)*, lag 3
-2.1 (-14.0, 9.7), lag 4
-25.4 (-45.4, -5.3), 5-day ma*
SVR:
-1.6 (-3.7, 0.4), lagO
1.6 -0.9, 4.1), lag 1
3.5 0.5, 6.5), lag 2
2.4 (-0.5, 5.3), lag 3
" " 0.7, 5.6), lag 4*
4.5
-0.3, 9.2), 5-day ma
*p<0.05, **p<0.01
Notes: Percent change (95% Cl) per
10 pg/m3 PM25 by GSTM1 genotype
(Fig 3)
Reference: Schwartz et al. (2005,
0743171
Period of Study: 12 wk during the
summer of 1999
Location: Boston, MA
Outcome: Heart rate variability (HRV),
(SDNN,
r-MSSD, PNN50, LFHFR)
Age Groups: 61-89 yr
Study Design: Panel study
N: 28 elderly subjects
Statistical Analysis: Mixed models. To
examine heterogeneity of effects,
hierarchical modeling was used.
Season: Summer
Dose-response Investigated? No
Pollutant: PM25
Averaging Time: 1 h, 24 h
Median: 24-h: 10 pg/m3
Monitoring Stations: 1
Copollutant: BC, 03, CO, S02, N02
PM Increment: IQR (not given)
Effect Estimate: 24 h: 2.6 ms decrease
in SDNN (95% Cl: 0.8 to-6.0)
10.1 ms decrease in r-MSSD (95% Cl: -
2.8to-16.9).
1 h: 3.4 ms decrease in SDNN (95% Cl:
0.6to-7.3)
7.4 ms decrease in r-MSSD (95% Cl:
1.6to-15.5).
Notes: Various log-transformed HRV
parameters were measured for 30
minutes once a week. The random
effects model indicated that the negative
effect of BC on HRV was not restricted
to a few subjects.
Same study population as Gold et al.
(2005). Boston Elders Study
For each pollutant/averaging time,
similarly sized changes were observed
for PNN50 (%) and LFHFR.
December 2009
E-64
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Schwartz et al. (2005,
0743171
Period of Study: 12 wk during the
summer of 1999
Location: Boston, MA
Outcome: Heart rate variability (HRV),
(SDNN, r-MSSD, PNN50, LFHFR)
Age Groups: 61 -89 yr
Study Design: Panel study
N: 28 elderly subjects
Statistical Analysis: Mixed models. To
examine heterogeneity of effects,
hierarchical modeling was used.
Season: Summer
Dose-response Investigated? No
Pollutant: BC
Averaging Time: 24 h
Median: 1.0|jg/m3
Monitoring Stations: 1
Copollutant: PM25, 03, CO, S02, N02
PM Increment: IQR
Effect Estimate: 5.1 ms decrease in
SDNN (-1.5 to-8.6)
10.1 ms decrease in r-MSSD (-2.4 to -
17.2).
Notes: Various log-transformed HRV
parameters were measured for 30
minutes once a week. The random
effects model indicated that the negative
effect of BC on HRV was not restricted
to a few subjects. Same study
population as Gold et al. (2005). Boston
Elders Study. Subjects with a prior Ml
experienced greater declines in BC
associated HRV. For each
pollutant/averaging time, similarly sized
changes were observed for PNN50 (%)
and LFHFR.
Reference: Schwartz et al. (2005,
0743171
Period of Study: 2000
Location: Boston, Massachusetts
Outcome: HF (high frequency
component of heart rate variability)
Study Design: Cross-sectional
N: 497 subjects
Statistical Analysis: Linear regression,
controlling for covariates
Pollutant: PM25
Averaging Time: 48 h
Mean(SD):11.4|jg/m3(8.0)
Copollutant: None
PM Increment: 10 pg/m
Effect Estimate: 34% decrease in HF
(95% Cl: -9% to -52%) in subjects
without the GSTM1 allele. In subjects
with the allele, no effect was noted.
Similar findings for obese subjects and
those with high neutrophil counts.
Notes: Study population: Normative
Aging Study.
Effects of PM2 5 appear to be mediated
by ROS.
Reference: Sorensen et al. (2005,
0894281
Period of Study: Nov 1999-Aug 2000
Location: Copenhagen, Denmark
Outcome: 7-Hydro-8-Oxo-2'-
Deoxyguanosine (8-oxodG) (measured
in lymphocytes and urine)
Age Groups: 20-33 yr
Study Design: Panel (repeated
measures)
N: 49 students living and studying in
central Copenhagen
50 students examined each season (66
subjects total
32 participated in each season
total of 98 measurements)
Statistical Analyses: Mixed models
repeated measures
Covariates: PM25, season, subject
(random factor)
Dose-response Investigated? No
Statistical Package: SAS v8e
Pollutant: PM25
Averaging Time: 48 h
Mean (SD): Fall: 20.7
Summer: 12.6
Percentiles: IQR Fall: 13.1-27.7
IQR summer: 9.4-24.3
Range (Min, Max): NR
Monitoring Stations: NA (personal
assessment)
Copollutant (correlation):
Spearman correlations with PM25 mass:
chromium (r = 0.22)
copper (r = 0.33)
iron (r = 0.29)
vanadium (p>0.5)
nickel (p>0.5)
platinum (p>0.5)
PM Increment: see below
Effect Estimate [Lower Cl, Upper Cl]:
Association between 8-oxodG in
lymphocytes and personal exposure to
transition metals in PM2 5.
% increase in 8-oxodG per increase in
metal concentration indicated
Vanadium: 1.9% per 1 pg/L (0.6, 3.3)
Chromium: 2.2% per 1 pg/L (0.8, 3.5)
Platinum: 6.1% perl ng/L(-0.6,13.2)
Nickel: 0.8% per 10 pg/L (-2.1, 3.7)
Copper: -0.8% per 10 pg/L (-2.7,1.0)
Iron: 0.6% per 10 pg/L (-1.4, 2.6)
Note: PM25 mass was independently
associated with 8-oxodG in 5 of 6
transition metal models (p < 0.02 in
models with vanadium, chromium,
nickel, copper, and iron
p = 0.07 in platinum model). No
transition metals were associated with 8-
oxodG measured in urine
December 2009
E-65
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Sorensen et al. (2003,
0427001
Period of Study: Nov 1999-Aug 2000
Location: Copenhagen, Denmark
Outcome: RBC count, hemoglobin,
platelet count, fibrinogen, PLAAS (2-
aminoadipic semialdehyde in plasma
proteins), HBGGS (v-glutamyl
semialdehyde in hemoglobin), HBAAS
(2-aminoadipic semialdehyde in
hemoglobin), MDA (malondialdehyde)
Age Groups: 20-33 yr
Study Design: Panel (repeated
measures)
N: 50 students living and studying in
central Copenhagen
50 students examined each season (68
subjects total
31 participated in each season
total of 195 measurements)
Statistical Analyses: Mixed model
repeated-measures analysis
Covariates: Season, avg outdoor
temperature, and sex
Season: Repeated measures 4 times
(once per season)
Dose-response Investigated? No
Statistical Package: SAS v8e
Pollutant: PM25 (personal)
Averaging Time: 48 h
Median: 16.1 pg/m3
Percentiles: Q25-Q75:10.0-24.5
Copollutant:
Urban background PM25
Personal PM25
PM Increment: 1 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
Relationship between exposure and
biomarkers
Estimate (p-value): Platelet count (x
106/g protein): 0.0008 (0.37)
Fibrinogen (nmol/g protein): 0.0006
(0.69)
PLAAS (pmol/mg protein): 0.0016
(0.061)
HBGGS (pmol/mg protein): 0.0001
(0.94)
HBAAS (pmol/mg protein): 0.0006 (0.64)
Increase (96%CI) in biomarkers per 10
ug/m3 increase in PM25
RBC
Men: 0% (-1.6,1.6)
Vtomen: 2.3% (0.5, 4.1)
Hemoglobin
Men: 0.0% (-1.7,1.5)
Vtomen: 2.6% (0.8, 4.5)
Reference: Sorensen et al. (2003,
0427001
Period of Study: Nov 1999-Aug 2000
Location: Copenhagen, Denmark
Outcome: RBC count, hemoglobin,
platelet count, fibrinogen, PLAAS (2-
aminoadipic semialdehyde in plasma
proteins), HBGGS (Y-glutamyl semi-
aldehyde in hemoglobin), HBAAS (2-
aminoadipic semialdehyde in
hemoglobin), MDA (malondialdehyde)
Age Groups: 20-33 yr
Study Design: Panel (repeated
measures)
N: 50 students living and studying in
central Copenhagen
50 students examined each season (68
subjects total
31 participated in each season
total of 195 measurements)
Statistical Analyses: Mixed model
repeated-measures analysis
Covariates: Season, avg outdoor
temperature, and sex
Season: Repeated measures 4 times
(once per season)
Dose-response Investigated? No
Statistical Package: SAS v8e
Pollutant: Personal exposure to black
carbon (10-6/m)
Averaging Time: 48 h
Median: 8.1
Percentiles: Q25-Q75: 5.0-13.2
Copollutant:
Urban background PM25
Personal PM25
PM Increment: 10-6/m
Effect Estimate [Lower Cl, Upper Cl]:
Relationship between exposure and
biomarkers
Estimate (p-value): RBC count (x 109/g
protein): 0.0003 (0.75)
Hemoglobin (pmol/g protein): 0.0004
(0.65)
Platelet count (x 106/g protein): 0.0009
(0.51)
Fibrinogen (nmol/g protein): -0.0027
(0.29)
PLAAS (pmol/mg protein): 0.0041
(0.0009)
HBGGS (pmol/mg protein): 0.0024
(0.25)
HBAAS (pmol/mg protein): 0.0022 (0.20)
MDA (pmol/mg protein): 0.0018 (0.30)
December 2009
E-66
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Sorensen et al. (2003,
0427001
Period of Study: Nov 1999-Aug 2000
Location: Copenhagen, Denmark
Outcome: RBC count, hemoglobin,
platelet count, fibrinogen, PLAAS (2-
aminoadipic semialdehyde in plasma
proteins), HBGGS (v-glutamyl
semialdehyde in hemoglobin), HBAAS
(2-aminoadipic semialdehyde in
hemoglobin), MDA (malondialdehyde)
Age Groups: 20-33 yr
Study Design: Panel (repeated
measures)
N: 50 students living and studying in
central Copenhagen
50 students examined each season (68
subjects total
31 participated in each season
total of 195 measurements)
Statistical Analyses: Mixed model
repeated-measures analysis
Covariates: Season, avg outdoor
temperature, and sex
Season: Repeated measures 4 times
(once per season)
Dose-response Investigated? No
Statistical Package: SAS v8e
Pollutant: PM25 (urban background
concentration)
Averaging Time: 48 h
Median: 9.2 pg/m3
Percentiles: Q25-Q75: 5.3-14.8
Copollutant:
Urban background PM25
Personal carbon black
PM Increment: 1 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
Relationship between exposure and
biomarkers
Estimate (p-value): RBC count (x 109/g
protein): 0.0008 (0.36)
Hemoglobin (pmol/g protein): 0.0005
(0.53)
Platelet count (x 106/g protein): -0.0008
(0.49)
Fibrinogen (nmol/g protein): 0.0004
(0.84)
PLAAS (pmol/mg protein): 0.0004 (0.76)
HBGGS (pmol/mg protein): -0.0020
(0.39)
HBAAS (pmol/mg protein): -0.0021
(0.29)
MDA (pmol/mg protein): 0.0012 (0.52)
Reference: Sullivan et al. (2007,
1000831
Period of Study: Feb 2000-Mar 2002
Location: Seattle, Washington, USA
Outcome: Blood CRP, fibrinogen, D-
dimer
Age Groups: >55 yr of age
Study Design: Panel study
N: 47 elderly subjects
Pollutant: PM25
Averaging Time: 24 h
Median (IQR): 7.7 pg/m3 (6.4)
Monitoring Stations: 1
Copollutant: Indoor PM25
PM Increment: 10 pg/m
Effect Estimate: Among those with
CVD, PM251 day earlier: CRP: 1.25
(95% 0:0.97,1.58)
Fibrinogen: 1.01 (95% Cl: 0.97,1.05)
D-dimer: 1.04 (95% Cl: 0.93,1.15)
With COPD: CRP: 0.69 (95% Cl: 0.34,
1.42)
Fibrinogen: 1.05 (95% Cl: 0.97,1.13)
D-dimer: 1.10 (95% Cl: 0.95,1.28)
Healthy: CRP: 1.01 (95% Cl: 0.85,1.19)
Fibrinogen: 0.88 (95% Cl: 0.81, 0.95)
D-dimer: 1.10 (95% Cl: 0.75,1.58)
Notes: Out of 47 subjects, n = 23 with
CVD and n = 24 (n = 16 COPD and 8
healthy) without CVD. Blood markers
were measured on 2-3 morning over a
5-10 day period, and outdoor PM25 was
measured at a central monitoring site.
These findings are not consistent with
and effect of fine PM on markers of
inflammation and thrombosis in the
elderly.
December 2009
E-67
-------
Reference Design & Methods Concentrations1 Effect Estimates (95% Cl)
Reference: Sullivan et al. (2005, Outcome: Heart rate variability (H, LF, Pollutant: PM25 PM Increment: 10 pg/m3
1094181 HF, r-MSSD, SDNN) . . „ „ t ™ .-.-.„,
Averaging Time: 1 h Effect Estimate: 1 h:
Period of Study: Feb 2000-Mar 2002 Study Design: Panel study
y y a > Median (IQR): 10.7 (7.6) With CVD: HF: (3% increase, 95% Cl:
Location: Seattle, Washington, USA N: 34 elderly subjects with (n = 21) and -19, 32)
without (n = 13) CVD Copollutant: CO, N02
1 ' Wthout CVD: H F(5% decrease, 95% Cl:
Statistical Analysis: Linear mixed -34, 36)
effects regression .
Similarly, no association was found for
4-h or 24-h mean PM25 concentrations.
Notes: 285 daily 20 min HRV measures
were made in the homes of study
subjects over a 10-day period.
December 2009 E-68
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Sullivan et al. (2005,
1094181
Period of Study: Feb 2000-Mar 2002
Location: Seattle area, WA
Outcome (ICD9 and ICD10): High-
sensitivity C-reactive protein (hs-CRP)
fibrinogen
D-dimer
Endothelin-1 (ET-1)
lnterleukin-6 (IL-6
lnterleukin-6 receptor (IL-6r)
Tumor necrosis factor-a (TNF-8- a)
Tumor necrosis factor-receptors (p55,
p75)
Monocyte chemoattractant protein-1
(MCP-1)
Age Groups: > 55 yr
Study Design: Panel (repeated
measures)
N: 47 participants with (23) and without
(10COPD and 8 healthy) CVD
Statistical Analyses: Mixed models
Covariates: Age, gender, medication
use, meteorological variables
(temperature and RH)
Dose-response Investigated? No
Statistical Package: SAS v8.02
Pollutant: PM25
Averaging Time: 24 h
(0-day and 1-day lags)
Mean (SD): NR
Percentiles: For all subject-days:
25th: 5.2
50th: 7.7
75th: 11.5
90th: 19.9
Range (Min, Max): 1.3, 33.9
Monitoring Stations: NA, measured at
participant's residence
Copollutant: None
PM Increment: 10 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
Multiplicative change in mean outcome
associated with 10 pg/m3 increase in PM
Among those with different disease
status.
CRP Fold-rise (96%CI)
CV
0-day lag: 1.21 (0.86,1.70)
CV
1-day lag: 1.25 (0.97,1.58);
COPD
0-day lag: 0.93 (0.48,1.80)
COPD
1-day lag: 0.69 (0.33,1.46)
Healthy
0-day lag: 0.98 (0.88,1.08)
Healthy
1-day lag: 1.01 (0.841.21)
Fibrinogen Fold-rise (96%CI)
CV
0-day lag: 1.02 (0.98,1.06)
CV
1-day lag: 1.0 (0.97,1.03);
COPD
0-day lag: 1.0 (0.91,1.09)
COPD
1-day lag: 1.08 (0.99,1.17)
Healthy
0-day lag: 0.94 (0.87,1.01)
Healthy
1-day lag: 0.99 (0.88,1.17)
D-dimer Fold-rise (96%CI)
CV
0-day lag: 1.02 (0.88,1.17)
CV
1-day lag: 1.03 (0.93,1.15);
COPD
0-day lag: 1.04 (0.93,1.16)
COPD
1-day lag: 1.09 (0.94,1.27)
Healthy
0-day lag: 0.95 (0.79,1.14)
Healthy
1-day lag: 0.97 (0.71,1.31)
Among those with cardiovascular
disease
MCP-1 Fold-rise (96%CI)
0-day lag: 1.3 (1.1,1.7)
1-day lag: 1.0 (0.9,1.3)
ET-1 Fold-rise (96%CI)
0-day lag: 1.1 (0.8,1.2)
1-day lag: 1.1 (0.9,1.2)
Note: TNF-a and IL-6 measures were
below the limit of detection of assays
Reference: Timonen et al. (2006,
0887471
Period of Study: 1998-1999
Location Amsterdam, Netherlands
Erfurt, Germany
Helsinki, Finland
Outcome: Heart variability (HRV)
measurements: [LF, HF, LFHFR, Nl\
interval, SDNN, r-MSSD]
Study Design: Panel study
N:131 elderly subjects with stable
coronary heart disease
Statistical Analysis: Linear mixed
models
Pollutant: PM25
Means:
Amsterdam: 20.0
Erfurt: 23.3
Helsinki: 12.7
Copollutant: N02, CO
PM Increment: 10 pg/m
Effect Estimate: SDNN
-0.33ms (95% Cl:-1.05, 0.38)
HF: -0.3% (95% Cl: -10.6, 5.4)
LFHFR:-1.4 (95% Cl:-5.9, 8.7)
Notes: Followed for 6 mo with biweekly
clinic visits
2-day lag. ULTRA Study
December 2009
E-69
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Vallejo et al. (2006,1570811 Outcome: Heart rate variability
„ • ., «~ ., , , .vw, measures (SDNN, pNNSO)
Period of Study: Apr-Aug 2002
Location: Mexico City metropolitan area
Age Groups: Mean age 27 yr
Study Design: Panel study
N: 40 young healthy participants (non-
smokers, no meds or history of CVD,
respiratory, neurological, or endocrine
disease)
Statistical Analysis: Linear mixed
effects models
Pollutant: PM25
(pDR nephelometric method-DataRAM)
Copollutant: None
PM Increment: 30 pg/m
Effect Estimate: pNNSO:
0 h lag:-0.01% (95% Cl:-0.03, 0.01)
1 h:-0.01% (95% Cl:-0.04, 0.02)
2 h: -0.05% (95% Cl: -0.09, 0.00)
3 h:-0.07% (95% Cl:-0.13 to-0.02)
4 h:-0.08% (95% Cl:-0.14 to-0.01)
5 h:-0.06% (95% Cl:-0.13, 0.02)
6 h:-0.05% (95% Cl:-0.13, 0.04)
SDNN:
Oh: 0.00% (95%Cl:0.00, 0.01)
1 h: 0.00%
2 h: 0.00%
95% Cl:-0.01, 0.01
95% Cl:-0.02, 0.01
3 h:-0.01% (95% Cl:-0.02, 0.00)
4 h:-0.01% (95% Cl:-0.02, 0.01)
5 h:-0.01% (95% Cl:-0.02, 0.01)
6 h: 0.00% (95% Cl:-0.02, 0.02)
Notes: Subjects underwent 13 h of ECG
monitoring and personal PM25
measurement. HRV measures were
regressed against different lags of PM2 5
concentration.
Reference: Van Hee et al. (2009,
1921101
Period of Study: Jul 2000-Aug 2002
Location: Baltimore, Maryland
Chicago, Illinois
Winston-Salem, North Carolina
St. Paul
Minnesota
New York, New York
Los Angeles, California
Outcome: Left Ventricular Mass Index
and Ejection Fraction
Age Groups: 45-84 yr
Study Design: Cross-sectional
N: 3,827 participants
Statistical Analyses: Linear Regression
Models
Covariates: Age, race, income, sex,
education, medication use, LDL, HDL,
physical activity, alcohol consumption,
smoking, diabetes, systolic BP, diastolic
BP
Dose-response Investigated? No
Statistical Package: Stata
Lags Considered: NR
Pollutant: PM25
Averaging Time: NR
Mean (SD): Fig only
Monitoring Stations: N/A
Interpolation used
Copollutant: NR
Co-pollutant Correlation: N/A
PM Increment: 10 pg/m
Difference (Lower Cl, Upper Cl), p-
value:
Left Ventricular Mass Index
Unadjusted: -6.0 (-7.8, -4.2), O.0001
All covariates except center, BP: -6.1 (-
7.8, -4.4), O.0001
All covariates except BP: 3.7 (-6.0,
13.4), 0.46
Full model: 4.6 (-4.7,13.9), 0.33
Full model plus center/race interaction:
3.8 (-6.1,13.7), 0.45
Left Ventricular Ejection Fraction
Unadjusted: 3.0 (2.2, 3.8), <0.0001
All covariates except center, BP: 1.4
(0.5, 2.2), 0.001
All covariates except BP: -1.1 (-5.8, 3.7),
0.66
Full model:-1.3 (-6.0, 3.5), 0.60
Full model plus center/race interaction: -
3.0 (-8.0, 2.0), 0.24
Reference: Wellenius et al. (2007,
0928301
Outcome: Circulating levels of B-type
natriuretic peptide (BNP
Period of Study: Feb 2002-Mar 2003 Measured in whole blood at 0, 6,12 wk)
Location: Boston, Massachusetts, USA Study Design: Panel study
N: 28 subjects (each with chronic stable
HF and impaired systolic function)
Statistical Analysis: Linear mixed
effects models
Pollutant: PM25
Copollutant: N02, S02, 03, CO, BC
PM Increment: 10 pg/m
Effect Estimate: Same day PM25: 0.8%
increase in BNP (95% Cl: -16.4, 21.5)
Notes: The study found no association
between any pollutant and measures of
BNP at any lag. Further, the within
subject coefficient of variation was large
suggesting the magnitude of effected air
pollutant health effects are small in
relation to within subject variability in
BNP.
December 2009
E-70
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Wellenius et al. (2007,
0928301
Period of Study: Feb 2002-Mar 2003
Location: Boston, Massachusetts
Outcome (ICD9 and ICD10): B-type
natriuretic peptide (BMP) (natural-log
transformed)
Age Groups: 33-88 yr
Study Design: Panel (blood collected at
0, 6, and12wk)
N: 28 patients with chronic stable heart
failure and impaired systolic function
Statistical Analyses: Linear mixed-
effects models
Covariates: Temperature, dew point,
mean dew point over the past 3 days,
calendar month of blood draw, mea-
surement occasion, treatment
assignment, measurement occasion by
treatment assignment interaction
Season: Adjusted for calendar month
Dose-response Investigated? No
Statistical Package: SASv91
Pollutant: PM25
Averaging Time: Daily (assessed lags
of 0-3 days)
Mean (SD): 10.9 (8.4)
Percentiles: 50th: 8.0 pg/m3
Range (Min, Max): 07-50.9 pg/m3
Monitoring Stations: 1 monitor
Copollutant (correlation):
CO (r = 0.35)
N02(r = 0.31)
S02(r = 0.18)
03(r = 0.35)
BC(r = 0.68)
PM Increment: IQr = 8.1 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
% change in BMP per IQR increase in
PM25
LagO: 1.5 (-18.7,19.2)
Lag1:2.1 (-20.0,30.3)
Lag2:1.3 (12.3,17.1)
Lag3: 5.6 (-16.8, 34.0)
Note: No significant associations
observed between any pollutant and
BMP levels at any lags (presented in
Fig 2)
Reference: Wheeler et al. (2006,
0884531
Period of Study: Fall 1999 and spring
2000
Location: Atlanta, GA
Outcome: Heart rate variability
Age Groups: 49-76 yr
N: 18 subjects with COPD and 12
subjects with a recent Ml
Statistical Analysis: Linear-mixed
effect model
Season: Fall and spring
Pollutant: PM25
Averaging Time:
1 h
4h
24 h
Mean: 24-h: 17.8 pg/m3
Copollutant: 03, CO, S02, N02
PM Increment: 11.65 pg/m3 (IQR) in 4 h
PM25
Effect Estimate: Among COPD
patients: 8.3% increase in SDNN (95%
0:1.7,15.3)
Among Ml patients: 2.9% decrease in
SDNN (95% Cl:-7.8, 2.3)
Results for 1-h and 24-h averaging times
were similar.
Notes: Data was collected on 7 days in
the fall of 1999 or spring of 2000.
Effects were modified by medication
use, baseline pulmonary function, and
health status.
December 2009
E-71
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Yeatts et al. (2007, 0912661
Period of Study: 12-wk period b/t
Sep 2003-Jul 2004
Location: Chapel Hill, NC
Outcome: Heart Rate Variability
Age Groups: 21-50 yr
Study Design: Panel
N: 12 asthmatics
Statistical Analyses: Linear Mixed
Model
Covariates: Temperature, humidity,
pressure
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 1 day
Pollutant: PM25
Averaging Time: 24 h
Mean (SD): 12.5 (6.0)
Min:0.6
Max: 37.1
Monitoring Stations: 1
Copollutant: PM10.2.5, PM10
Co-pollutant Correlation:
PMio-25 = 0.46*
PM10 = NR
*p < 0.01
PM Increment: 1 pg/m
Beta, SE (Lower Cl, Upper Cl), p-
value:
HRV
Max Heart Rate: 0.40, 0.43 (-0.45,1.24),
0.36
ASDNN5: -0.07, 0.15 (-0.37, 0.22), 0.63
SDANN5:1.66, 0.65 (0.39, 2.93), 0.02
SDNN24HR(mesc): 1.16, 0.58(0.02,
2.29), 0.06
rMSSD:0.53, 0.20(0.14, 0.91), 0.01
pNN50_24hr:-0.06, 0.11 (-0.27, 0.15),
0.58
pNN50_7min: 0.47, 0.42 (-0.35, 1.29),
0.27
Low-frequency power: -0.23, 0.14 (-0.51,
0.05), 0.11
Percent low frequency: -0.78, 0.41 (-
1.59, 0.03), 0.07
High-frequency power: 0.14, 0.07 (-0.01,
0.28), 0.07
Percent high frequency: 0.64, 0.36 (-
0.07, 1.34), 0.09
Blood Lipids
Triglycerides: -0.63, 0.84 (-2.29,1.02),
0.46
VLDL:-0.17, 0.22 (-0.61, 0.26), 0.44
Total cholesterol: -0.06, 0.22 (-0.49,
0.36), 0.77
Hematologic Factor
Circulating eosinophils: -0.02, 0.00 (-
0.02, -0.02), 0.27
Platelets:-0.01, 0.45 (-0.88, 0.86), 0.98
Circulating Proteins
Plasminogen: 0.00, 0.00 (-0.01, 0.00),
0.82
Fibrenogen: 0.00, 0.01 (-0.01,0.02),
0.59
Von Willibrand factor: -0.31, 0.29 (-0.87,
0.25), 0.28
Factor VII:-0.65, 0.33 (-1.29, -0.01),
0.05
December 2009
E-72
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Yue et al. (2007, 0979681
Period of Study: Oct 2000-Apr 2001
Location: Erfurt, Germany
Outcome: QT interval and T-wave
amplitude for ECG recordings, and vWF,
CRP from blood samples
Study Design: Panel study
N: 56 patients (male CAD patients with
12 clinical visits)
Statistical Analysis: Linear and logistic
regression models
Dose-response Investigated? No
Pollutant: PM25, PNC (n/crri)
Averaging Time: Mean:
Mass concentrations of PNC
(0.1-2.84 n/cm3)
Monitoring Stations: 1
Copollutant: None
PM Increment:. IQR
Effect Estimate: Each IQR increase in
0-23 h mean traffic particle concentration
was associated with: QT interval: 0.6%
(95% Cl:-0.3, 1.4)
T wave amplitude: -1.6% (95% Cl: -3.3,
0.1)
vWF: 3.2% (95% Cl: -0.5, 7.0)
CRP: (OR = 1.5
95% Cl 1.0-2.3)
Each IQR increase in 0-23 h mean
combustion-generated particle
concentration was associated with: QT
interval: 0.1%(-0.3, 0.6)
T wave amplitude: -0.2% (-1.2, 0.7)
vWF: 2.8% (0.8, 4.8)
CRP (OR = 1.0
0.8, 1.2)
Notes: Five sources of particles were
identified (airborne soil, local traffic-
related ultrafine particles, combustion-
generated aerosols, diesel traffic-related
particles, and secondary aerosols).
Reference: Yue et al. (2007, 0979681
Period of Study:
Oct 12, 2000-Apr 27, 2001
Location: Erfurt, Germany
Outcome: QT interval, T wave
amplitude, von Willebrand factor (vWF),
C-reactive protein (CRP above 90th
percentile compared to below)
Age Groups: >50 yr
Study Design: Panel (12 visits
625 observations for repolarization
parameters and 578 observations for
inflammatory markers)
N: 57 male coronary artery disease
patients
Statistical Analyses: Linear and logistic
fixed-effects regression models
(generalized additive models)
Covariates: Trend, weekday, and
meteorological variables (temperature,
relative humidity, barometric pressure)
Dose-response Investigated? No
Statistical Package: SASv9.1 and S-
piusve.o
Pollutant: Five particle source factors PM Increment: IQR
(airborne soil, local traffic-related
ultrafine particles, combustion-generated
aerosols, diesel traffic-related particles,
and secondary aerosols); see below for
size fractions (factor scores)
Averaging Time: Used daily factor
scores in analyses
Mean (SD):
Factor 1: particles from airborne soil
(1.0-2.8 fjm): 2390 (1696)
Factor 2: ultrafine particles from local
traffic (0.01-0.1 pm):9931 (5858)
Factor 3: secondary aerosols from local
fuel combustion (0.1-0.5 pm): 3770
(6129)
Factor 4: particles from traffic (0.01-
0.5 pm): 6865 (5689)
Factor 5: secondary aerosols from
multiple sources (0.2-1.0 pm): 4732
(3890)
Median:
Factor 1:2053
Factor 2: 8531
Factor 3:1348
Factor 4: 5045
Factor 5: 3752
IQR (6-day avg):
Factor 1:1110
Factor 2: 5749
Factor 3: 4124
Factor 4: 5000
Factor 5: 3393
Range (Min, Max):
Factor 1:284,12960
Factor 2: 866, 26632
Factor 3:139, 39097
Factor 4: 283, 27605
Factor 5: 67, 20129
Monitoring Stations: 1 monitor
Effect Estimate [Lower Cl, Upper Cl]:
QT interval, % change (96%CI)
Factor 1:
0-5 h:-0.1 (-0.6,0.6)
6-11 h: -0.5 (-1.1, 0.2)
12-17 h: 0.1 (-0.4,0.4)
18-23 h:-0.2 (-0.7, 0.2)
0-23 h: -0.2 (-0.9, 0.4)
1 day:-0.1 (-0.7,0.6)
2 day: -0.3 (-0.9, 0.4)
3 day:-0.7 (-1.4, 0.1
4 day: -0.2 (-0.9, 0.5
0-4 day avg:-0.7 (-1.8, 0.3)
Factor 2:
0-5 h: 0.2 (-0.4, 0.8)
6-11h:0.8(-0.0, 1.7)
12-17 h: 0.6 (-0.2, 1.4)
18-23 h: 0.5 (-0.4, 1.4)
0-23 h: 0.9 (-0.1, 2.0)
1 day: 1.5 (0.3, 2.7)
2 day:-0.4 (-1.7,1.0)
3 day: 0.5 (-0.9,1.9)
4 day: 0.1 (-1.2,1.4)
0-4 day avg: 1.6 (-0.1, 3.3)
Factors:
0-5 h: 0.1 (-0.3,0.5)
6-11 h: 0.2 (-0.3, 0.6)
12-17 h: 0.2 (-0.3, 0.6)
18-23 h: 0.1 (-0.3,0.4)
0-23 h: 0.1 (-0.3, 0.6)
1 day: 0.1 (-0.3, 0.4)
2 day:-0.1 (-0.4,0.3
3 day:-0.2 (-0.5, 0.2
4 day:-0.1 (-0.5, 0.2)
0-4 day avg:-0.1 (-0.7,0.6)
Factor 4:
0-5 h: 0.2 (-0.4, 0.8)
6-11 h: 0.8 (0.0, 1.6)
12-17 h: 0.5 (-0.2, 1.3)
18-23 h: 0.5 (-0.2, 1.2)
0-23 h: 0.6 (-0.3, 1.4)
1 day:-0.4 (-1.5, 0.7)
2 day:-0.9 (-2.0, 0.1)
3 day:-0.5 (-1.4, 0.5)
December 2009
E-73
-------
Reference Design & Methods Concentrations1 Effect Estimates (95% Cl)
Copollutant:NA4 day: -0.5 (-130.2)
f 0-4 day avg: -0.3 (-1.7,1.1)
Factor 6:
nO-5h:1.0(-0.1,2.1)
6-11 h: 0.9 (-0.2, 2.0)
12-17 h: 0.3 (-0.7, 1.4)
18-23 h:-0.1 (-1.2, 1.0)
0-23h: 0.7 (-0.6, 1.9)
1 day: 0.1 (-1.1,1.3)
2 day:-0.2 (-1.5,1.1)
3 day:-0.6 (-1.9, 0.8
4 day:-0.9 (-2.0, 0.2
0-4 day avg:-0.4 (-1.9,1.2)
Twave amplitude, % change (96%CI)
Factor!:
0-5 h:-0.3 (-1.5, 0.9)
6-11 h:-0.6 (-1.9, 0.7)
12-17 h: 0.1 (-0.8,0.9)
18-23 h:-0.6 (-1.5, 0.4)
0-23 h:-0.5 (-1.8, 0.9)
1 day: 0.4 (-0.9,1.7)
2 day: 1.2
3 day: 0.2
-0.3, 2.7)
-1.2, 1.7)
4 day:-0.2 (-1.3,1.0)
0-4 day avg: 0.8 (-1.1, 2.6)
Factor 2:
0-5 h:-1.7 (-3.0 to-0.4)
6-11 h:-2.6 (-4.5 to-0.6)
12-17 h:-1.0 (-2.6, 0.7)
18-23 h:-1.1 (-2.8,0.7)
0-23 h:-3.1 (-5.3 to-0.9)
1 day: -0.3 (-2.9, 2.2)
2 day:-1.2 (-4.1,1.7)
3 day:-0.5 (-3.2, 2.1
4 day:-3.4 (-9.9, 3.1
0-4 day avg:-1.5 (-4.4,1.5)
Factor 31
0-5 h:-0.3 (-1.1, 0.6)
6-11 h:-0.1 (-0.9,0.9)
12-17 h: 0.1 (-0.9, 1.0)
18-23 h:-0.4 (-1.2, 0.4)
0-23 h:-0.2 (-1.2, 0.7)
1 day: 0.1 (-0.7,0.8)
2 day:-0.1 (-0.7, 0.7)
3 day: 0.4 (-0.3,1.1)
4 day: 0.1 (-0.7,0.7)
0-4 day avg: 0.3 (-0.9,1.5)
Factor 4:
0-5 h:-1.5 (-2.8 to-0.2)
6-11 h:-1.3 (-3.0, 0.3)
12-17 h:-1.1 (-2.7,0.4)
18-23 h: -0.9 (-2.4, 0.6)
0-23 h:-1.6 (-3.3, 0.1)
1 day:-1.2 (-3.3, 0.9)
2 day:-1.0 (-3.2,1.2)
3 day: 0.2 (-1.5,1.9)
4 day: 0.5 (-1.0, 2.0)
0-4 day avg:-1.7 (-4.1,0.7)
Factor 6:
0-5 h:-1.6 (-3.6, 0.4)
6-11 h:-0.1 (-2.1,2.0)
12-17 h:-0.2 (-2.2, 1.8)
18-23 h:-1.8 (-3.8, 0.2)
0-23 h:-1.2 (-3.4, 1.0)
1 day:-1.8 (-4.2, 0.6)
2 day:-0.7 (-3.5, 2.1)
3 day: 0.8
4 day: 0.5
-1.5, 3.2)
-1.5, 2.5)
0-4 day avg:-1.4 (-4.0,1.2)
vWF, % change (95%CI)Factor 1:
0-5 h: 1.1 (-1.5,3.6)
6-11 h: 1.6 (-1.2, 4.5)
12-17 h: 0.4 (-1.4, 2.1)
18-23 h: 1.4 (-0.6, 3.5)
0-23 h: 1.6 (-1.3, 4.4)
1 day:-1.0 (-3.9,1.9)
2 day:-1.8 (-4.8,1.2)
December 2009 E-74
-------
Reference Design & Methods Concentrations1 Effect Estimates (95% Cl)
3 day:-2.5 (-5.8, 0.9)
4 day: 0.5 (-2.9, 3.9)
0-4 day avg:-2.5 (-7.1,2.2)
Factor 2:
0-5 h: 0.4 (-2.4, 3.2)
6-11 h:-0.4 (-4.3, 3.4)
12-17 h: 2.1 (-1.4, 5.7)
18-23 h: 2.3 (-1.4, 5.9)
0-23 h: 1.9 (-2.8, 6.6)
1 day: 2.8 (-2.8, 8.3)
2 day: 5.1 (-0.8,11.1)
3 day: 11.4 (5.3,17.6)
4 day: 6.6 (0.0,13.1)
0-4 day avg: 11.4 (3.7,19.1)
Factor 31
0-5 h: 1.8 (0.1, 3.6)
6-11 h: 1.7 (-0.3, 3.7)
12-17 h: 2.2 (0.3, 4.2)
18-23 h: 2.8 (1.1, 4.5)
0-23 h: 2.8 (0.8, 4.8)
1 day: 2.7 (1.0, 4.4)
2 day: 3.4 (1.8, 5.0)
3 day: 2.3
4 day: 1.4
0.8, 3.8)
-0.2, 2.9)
0-4 day avg: 4.8 (2.0, 7.6)
Factor 4:
0-5h:1.5(-1.4, 4.3)
6-11h:2.0(-1.7, 5.6)
12-17h:2.6
18-23h: 3.5
-0.8, 5.9)
0.4, 6.6)
0-23h: 3.2 (-0.5, 7.0)
1 day: 5.4
2 day: 4.5
0.6, 10.2)
-0.6, 9.5)
3 day: 3. 8 (-0.6, 8.1)
4 day: 3.0 (-0.6, 6.6)
0-4d avg: 11.3 (5.0, 17.6)
Factors:
0-5 h: 1.9 (-2.8, 6.6)
6-11 h: 3.2 (-1.6, 8.0)
12-17 h: 2.4 (-2.3, 7.1)
18-23 h: 1.6 -3.1,6.2
0-23 h: 2.9 (-2.5, 8.2)
1 day: -2.2 (-7.6, 3.2)
2 day: -1.3 (-7.4, 4.9)
3 day: 1.1 (-4.8, 7.1)
4 day: 1.3 (-4. 2, 6.7)
0-4 day avg: 3.3 (-4.1, 10.6)
CRP, Odds Ratio (96%CI)
Factor 1
0-5 h: 0.9 (0.7, 1.1)
6-11 h: 1.4 (1.1, 1.8)
12-17 h: 1.2 (1.0, 1.4
18-23 h: 1.0 (0.8, 1.3
0-23 h: 1.1 (0.9, 1.5)
1 day: 1.4
2 day: 1.3
1.1, 1.8)
1.0, 1.7)
3 day: 1.0 (0.7, 1.4)
4 day: 1.1 (0.9, 1.5)
0-4 day avg: 1.6 (1.1, 2.2)
Factor 2
0-5h: 0.8 (0.6, 1.0)
6-11h:1.0(0.7, 1.4)
12-17h:1.1(0.8, 1.5)
18-23h:1.0 0.8, 1.4)
0-23h: 0.9 (0.6, 1.4)
1 day: 0.9 (0.6, 1.5)
2 day: 2.1
3 day: 1.9
1.3,3.3)
1.0,3.6)
4 day: 1.4 (0.8, 2.3)
0-4d avg: 1.4 (0.8, 2.6)
Factor 3
0-5 h: 1.0 (0.8, 1.1)
6-11 h: 0.9 (0.8, 1.1)
12-17 h: 1.0 (0.9, 1.2)
18-23 h: 1.0 (0.8, 1.2)
0-23 h: 1.0 (0.8, 1.2)
1 day: 1.1 (1.0,1.3)
December 2009 E-75
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
2 day: 1.0(0.9,1.2)
3 day: 1.2
1.1, 1.4)
1.0, 1.3)
4 day: 1.1
0-4dYavg:1.2(1.0,1.5)
Factor 4
0-5 h: 0.8 (0.6, 1.1)
6-11 h: 0.8 (0.6, 1.1)
12-17 h: 1.3 (1.0, 1.8
18-23 h: 1.1 (0.8, 1.5
0-23 h: 1.0 (0.7, 1.4)
1 day: 1.5
2 day: 2.0
1.0,2.3)
1.3,3.2)
3 day: 1.5 (0.9, 2.3)
4 day: 1.3 (0.9,1.8)
0-4 day avg: 1.7 (1.0, 2.9)
Factor 6
0-5 h: 0.7 (0.5, 1.1)
6-11 h: 1.4 (0.9, 2.1)
12-17 h: 1.9 (1.3, 2.8)
18-23 h: 1.4 (1.0, 2.0)
0-23 h: 1.4 (0.9, 2.2)
1 day: 1.6 (1.0, 2.6)
2 day: 1.6
3 day: 2.3
0.9, 2.6)
1.3,4.1)
4 day: 1.6 (0.9, 2.8)
0-4 day avg: 2.1 (1.2,3.8)
Reference: Zanobetti et al. (2004,
0874891
Period of Study: 1999-2001
Location: Boston, Massachusetts, USA
Outcome: Blood pressure (systolic
blood pressure, diastolic blood pressure,
mean arterial blood pressure)
Age Groups: Elderly
Study Design: Panel study
N: 62 elderly subjects with n = 631
repeated visits for cardiac rehabilitation
Statistical Analysis: Linear mixed
effects models
Pollutant: PM25
Averaging Time: 24 h
Median (10th-90th percentile)
Median: 8.8 pg/m3
10th-90th: 13.4
Monitoring Stations: 1
Copollutant: S02, 03, CO, N02, BC
120-havg
Median: 0.651
10th-90th: 0.376
PM Increment: .10.4 pg/m for 5-day
mean, 13.9pg/m for2-daymean
Effect Estimate: Each 10.4 pg/m3
increase in 5-day mean PM25
concentration was associated with:
Systolic BP: 2.8mmHg (95% Cl: 0.1, 5.5)
Diastolic BP: 2.7mmHg (95% Cl: 1.2,
4.3)
Mean arterial BP: 2.7mmHg (95% Cl:
1.0,4.5)
Each 13.9 pg/m3 increase in 2-day mean
PM25, during exercise in person with
HJObprn
Diastolic: 7.0mmHg (95% Cl: 2.3,12.1)
Mean arterial BP: 4.7mmHg (95% Cl:
0.5,9.1)
December 2009
E-76
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Zeka et al. (2006,1571771
Period of Study: Nov 2000-Dec 2004
Location: Greater Boston area
(Massachusetts)
Outcome: White blood cells (WBC), C-
reactive protein (CRP), sediment rate,
fibrinogen
Age Groups: Mean age (SD) = 73.0
(6.7)
Study Design: Cross-sectional
N: 710 subjects
Statistical Analyses: Linear regression
Covariates: Age, BMI, season (also
assessed potential for confounding by
temperature, RH, barometric pressure,
hypertensive or cardiac medications,
hypertension, smoking, alcohol, and
fasting glucose levels)
Dose-response Investigated? No
Pollutant: BC
Averaging Time: Hourly (PN, BC,
PM25) and 24-h (S042~) measurements
used to create 48-h, 1-wk, and 4-wk ma
Mean (SD): 0.77 (0.63)
Percentiles: 50th: 0.61
75th: 1.00
90th: 1.51
Monitoring Stations: 2 sites
Units: ng/m3
Copollutant (correlation):
PM25(r = 0.52)
BC
PN(r = 0.30)
S042"(r = 0.30)
PM Increment: 1 SD increase
Effect Estimate [Lower Cl, Upper Cl]:
% increase (95%CI) in biomarker per 1
SD increase in pollutant.
Fibrinogen
48 h: 0.84 (-0.63, 2.31)
1wk: 0.60 (-0.95, 2.15)
4 wk: 1.78 (0.19, 3.36)
CRP
48 h: 4.51 (-2.03, 11.06)
1wk:1.07
4wk:5.41
-5.55, 7.68)
-1.00, 11.81)
Sediment rate
48 h:-4.56 (-25.55, 16.43)
1 wk: 1.98 (-18.15, 22.11)
4 wk: 21.65 (1.48, 41.82)
WBC count
48 h:-0.63 (-2.45, 1.19)
1 wk:-0.13 (-1.87,1.60)
4 wk:-0.55 (-2.36, 1.26)
Note: No statistically significant
difference was reported for any category
of effect modifiers (age, obesity,
medications, homozygousforthe
deletion of GSTM1-null, hypertension)
However, results suggested almost all
the effect of BC on sediment rate was
among the younger group (<78 yr)
There was a 4-fold difference for the
association between BC and CRP in the
presence of obesity
Also evidence for effect modification by
obesity of the association between BC
and sediment rate
There was a suggestive greater effect of
BC on CRP among GSTM1-null subjects
(9.73% [1.48, 17.98]) vs.. GSTM1-
present subjects (-2.97% [-14.05, 8.10]
for concentrations 4-wk prior)
A stronger effect of BC on sediment rate
was seen among non-users of statins
36.01% [13.88, 58.13]) vs.. users
•12.29% [39.13, 14.55])
Reference: Zeka et al. (2006,1571771
Period of Study: Nov 2000-Dec 2004
Location: Greater Boston area
(Massachusetts)
Outcome: White blood cells (WBC), C-
reactive protein (CRP), sediment rate,
fibrinogen
Age Groups: Mean age (SD) = 73.0
(6.7)
Study Design: Cross-sectional
N: 710 subjects
Statistical Analyses: Linear regression
Covariates: Age, BMI, season (also
assessed potential for confounding by
temperature, RH, barometric pressure,
hypertensive or cardiac medications,
hypertension, smoking, alcohol, and
fasting glucose levels)
Dose-response Investigated? No
Pollutant: S04
Averaging Time: Hourly (PN, BC,
PM25) and 24-h (S042~) measurements
used to create 48-h, 1-wk, and 4-wk ma
Mean (SD): 2.29 (1.62)
Percentiles:
50th: 1.84
75th: 2.81
90th: 4.10
Monitoring Stations: 2 sites
Copollutant (correlation):
PM25(r = 0.50)
BC(r = 0.30)
PN(r = -0.15)
S042"
PM Increment: 1 SD increase
Effect Estimate [Lower Cl, Upper Cl]:
% increase (95%CI) in biomarker per 1
SD increase in pollutant.
Fibrinogen:
48 h: 0.60 (-1.23, 2.42)
1wk:0.03 -1.93,1.99
4wk:1.12 -0.52,2.77
CRP:
48 h: 1.57 (-7.13, 10.27)
1 wk: 0.21 (-8.27, 8.69)
4 wk: 5.29 (-1.91, 12.49)
Sediment rate:
48 h: 4.05 (-23.26, 31.36)
1 wk: -5.87 (-32.39, 20.64)
4 wk:-1.60 (-25.24, 22.04)
WBC count:
48 h:-0.12 (-2.35, 2.11)
1 wk:-0.48 (-2.87, 1.90)
4 wk: 0.75 (-1.30, 2.80)
Note: No statistically significant
difference was reported for any category
of effect modifiers (age, obesity,
medications, homozygousforthe
deletion of GSTM1-null, hypertension)
December 2009
E-77
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Zeka et al. (2006,1571771
Period of Study: Nov 2000-Dec 2004
Location: Greater Boston area
(Massachusetts)
Outcome (ICD9 and ICD10): White
blood cells (WBC), C-reactive protein
(CRP), sediment rate, fibrinogen
Age Groups: Mean age (SD) = 73.0
(6.7)
Study Design: Cross-sectional
N: 710 subjects
Statistical Analyses: Linear regression
Covariates: Age, BMI, season (also
assessed potential for confounding by
temperature, RH, barometric pressure,
hypertensive or cardiac medications,
hypertension, smoking, alcohol, and
fasting glucose levels)
Dose-response Investigated? No
Statistical Package: NR
Pollutant: PM25
Averaging Time: Hourly (PN, BC,
PM25) and 24-h (S042~) measurements
used to create 48-h, 1-wk, and 4-wk ma
Mean (SD): 11.16 (7.95)
Percentiles:
50th: 9.39
75th: 14.57
90th: 21.48
Monitoring Stations: 2 sites
Copollutant (correlation):
PM25
BC(r = 0.52)
PN(r = -0.02)
S042"(r = 0.50)
PM Increment: 1 SD increase
Effect Estimate [Lower Cl, Upper Cl]:
% increase (95%CI) in biomarker per 1
SD increase in pollutant.
Fibrinogen: 48 h:-0.18 (-1.93,1.57)
1wk:-1.39 (-3.46, 0.67)
4wk: 1.14 (-0.60, 2.88)
CRP: 48 h: -4.88 (-13.29, 3.53)
1 wk: -1.37 (-10.44, 7.71)
4 wk: 4.36 (-3.25, 11.96)
Sediment rate: 48 h: -16.91 (-43.66,
9.84)
1 wk: -18.89 (-47.48, 9.70)
4 wk: 24.93 (0.68, 49.18)
WBC count: 48 h: -3.18 (-5.39 to -0.97)
1 wk: -0.51 (-3.02, 2.00)
4 wk:-0.03 (-2.17, 2.10)
Note: No statistically significant
difference was reported for any category
of effect modifiers (age, obesity,
medications, homozygousforthe
deletion of GSTM1-null, hypertension)
Reference: Zhang et al. (2009,1919701
Period of Study: 1999-2003
Location: U.S.
Outcome: Myocardial Ischemia
Age Groups: 52-90
Study Design: Panel
N: 55,529
Statistical Analyses: Logistic & Linear
Regression
Covariates: Age, race/ethnicity,
education, exam site, BMI, current
smoking status, history of CHD,
diabetes, hypertension, SBP, chronic
lung disease, or hypercholesterolemia,
day of week, time of day, temperature,
dew point, pressure, season
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-5-day
Pollutant: PM25
Averaging Time: NR
Mean (SD):
Lag 0:14.1 (8)
Lag 1:13.8 (8)
Lag 2:13.8 (8)
Lag 3:13.8 (8)
Lag 4:13.9 (8)
Lag 5:14.1 (8)
Lag 0-2:13.9 (7)
Monitoring Stations: NRJ
Co-pollutant: NR
I Monitors used in model for spatial
interpolation of daily PM values.
PM Increment: 10 pg/m
Odds Ratio (Lower Cl, Upper Cl), l
Minnesota Codes*
MC4:1.04
MC4:1.04
0.97, 1.10), lag 0-2
0),
0.98, 1.11), lag 3-5
MC5: 1.05 (1.00, 1.09), lag 0-2
MC5: 1.04 (1.00, 1.08), lag 3-5
MC 4 or 5: 1.04 (1.00, 1.09), lag 0-2
MC 4 or 5: 1.03 (0.99, 1.07), lag 3-5
Change (Lower Cl, Upper Cl), lag:
ST-segment amplitude
Lead I: -0.07 (-0.36, 0.21), lag 0-2
Lead I: 0.1 8 (-0.10, 0.46), lag 3-5
Lead II: -0.12 (-0.47, 0.23), lag 0-2
Lead II: 0.16 (-0.18, 0.50), lag 3-5
Lead aVL: -0.01 (-0.25, 0.23), lag 0-2
Lead aVL: 0.11 (-0.12, 0.34), lag 3-5
Lead V1: -0.02 (-0.39, 0.35), lag 0-2
Lead V1: -0.22 (-0.58, 0.14), lag 3-5
Lead V2: 0.07 (-0.57, 0.70), lag 0-2
Lead V2: -0.01 (-0.61, 0.62), lag 3-5
Lead V3: -0.11 (-0.68, 0.47), lag 0-2
Lead V3: -0.02 (-0.58, 0.54), lag 3-5
Lead V4: -0.0.3 (-0.51, 0.45), lag 0-2
Lead V4: 0.24 (-0.23, 0.71), lag 3-5
Lead V5: -0.01 (-0.41, 0.39), lag 0-2
Lead V5: 0.35 (-0.04, 0.74), lag 3-5
Lead V6: 0.02 (-0.30, 0.33), lag 0-2
Lead V6: 0.35 (0.04, 0.65), lag 3-5
T-wave amplitude
Lead I:-1.60
Lead I: -0.31
-3.07,-0.13), lag 0-2
-1.73,1.11), lag 3-5
Lead 11:-0.54 (-1.99, 0.92), lag 0-2
Lead II: 0.71 (-0.70, 2.13), lag 3-5
Lead aVL:-1.21 (-2.50, 0.10), lag 0-2
Lead aVL:-0.55 (-1.18, 0.71), lag 3-5
Lead V1:1.45 (-0.16, 3.06, lag 0-2
Lead V1: 0.03 (-1.53,1.59, lag 3-5
Lead V2:-0.18 (-2.96, 2.60), lag 0-2
Lead V2: 0.57 (-2.12, 3.27), lag 3-5
Lead V3:-2.33 (-5.15, 0.49), lag 0-2
Lead V3:-0.13 (-2.87, 2.60), lag 3-5
Lead V4:-2.03 (-4.69, 0.63), lag 0-2
Lead V4: 0.64 (-1.94, 3.22), lag 3-5
Lead V5:-1.92 (-4.22, 0.38), lag 0-2
Lead V5: 0.55 (-1.69, 2.78), lag 3-5
Lead V6:-0.63 (-2.36,1.10), lag 0-2
Lead V6: 0.82 (-0.86, 2.49), lag 3-5
QRS/T angles and heart rate (change)
December 2009
E-78
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
QRS/T angle-spatial (°):0.19 (-0.21,
0.59), lag 0-2
QRS/T angle-spatial (°): -0.20 (-0.59,
0.19), lag 3-5
QRS/T angle-frontal (°): 0.13 (-0.24,
0.50), lag 0-2
QRS/T angle-frontal (°): 0.35 (-0.01,
0.71), lag 3-5
Heart Rate (beats/min): 0.16 (0.02,
0.30), lag 0-2
Heart Rate (beats/min): 0.04 (-0.10,
0.18), lag 3-5
*Any ST abnormality (MC 4.1 -4.4)
Any T abnormality (MC 5.1-5.4)
1AII units expressed in ug/m3 unless otherwise specified.
Table E-4. Short-term exposure-cardiovascular morbidity studies: Other size fractions.
Reference
Design & Methods
Concentrations
Effect Estimates (95% Cl)
Reference: Adar et al. (2007, 0014581
Period of Study: Mar-Jun 2002
Location: St. Louis, Missouri
Outcome: Heart rate variability: heart
rate, standard deviation of all normal-to-
normal intervals (SDNN), square root of
the mean squared difference between
adjacent normal-to-normal intervals
(rMSSD), percentage of adjacent
normal-to-normal intervals that differed
by more than 50 ms (pNNSO), high
frequency power (HF in the range of
0.15-0.4Hz), low frequency power (LF,
in the range of 0.04-0.15Hz), and the
ratio of LF/HF
Age Groups: > 60 yr
Study Design: Panel (4 planned
repeated measures with a total of 158
person-trips 35 participating in all 4
trips)
N: 44 participants
Statistical Analyses: Generalized
additive models
Covariates: Subject, weekday, time,
apparent temperature, trip type, activity,
medications, and autoregressive terms
Season: Limited data collection period
Dose-response Investigated? No
Statistical Package: SAS v8.02, R
V2.0.1
Pollutant: Particle count fine (PC fine)
(particles/cm3)
Averaging Time: Measurements
collected over 48-h period surrounding
the bus trip (during which health
endpoints were measured) used to
calculate 5-, 30-, 60-min, 4-h, 24-h ma
Median (IQR):
All: 42 (57)
Facility: 36 (45)
Bus: 105 (96)
Activity: 50 (133)
Lunch: f" ""
Monitoring Stations: 2 portable carts
Copollutant:
PM25
BC
Fine particle counts
Coarse particle counts
Correlation notes: 24-h mean PM25,
BC, and fine particle count
concentrations ranged from 0.80 to 0.98
r = 0.76 to 0.97 when limited to time
spent on the bus
r = 0.55 to 0.86 when comparing bus
concentrations to 24-h ma
r = -0.003 to 0.51 when comparing 5-
min avg and 24-h ma. Poor correlations
found between coarse particle count
concentrations and all fine particulate
measures during all times periods
PM Increment: IQR
Effect Estimate [Lower Cl, Upper Cl]:
% change (95%CI) in HRV per IQR in
the 24-h ma of the microenvironmental
pollutant (IQr = 39 pt/cm3)
Single-pollutant models
SDNN:-5.1 (-5.8 to-4.4)
rMSSD:-8.0 (-8.7 to-7.2)
pNN50+1:-10.2 (-11.3to-9.0)
LF:-9.9 (-11.4 to-8.4)
HF:-13.7 (-15.1 to-12.2)
LF/HF: 4.3 (3.1, 5.5)
H: 0.9 (0.8, 1.1)
Note: Exposure to health associations
by all lag periods presented in Fig 2
(magnitude of associations increased
with averaging period, with the largest
associations consistently found for 24-h
December 2009
E-79
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Adar et al. (Adar et al.,
2007, 0014581
Period of Study: Mar-Jun 2002
Location: St. Louis, Missouri
Outcome: Heart rate variability: heart
rate, standard deviation of all normal-to-
normal intervals (SDNN), square root of
the mean squared difference between
adjacent normal-to-normal intervals
(rMSSD), percentage of adjacent
normal-to-normal intervals that differed
by more than 50 ms (pNNSO), high
frequency power (HF in the range of
0.15-0.4Hz), low frequency power (LF,
in the range of 0.04-0.15Hz), and the
ratio of LF/HF
Age Groups: > 60 yr
Study Design: Panel (4 planned
repeated measures with a total of 158
person-trips
35 participating in all 4 trips)
N: 44 participants
Statistical Analyses: Generalized
additive models
Covariates: Subject, weekday, time,
apparent temperature, trip type, activity,
medications, and autoregressive terms
Season: Limited data collection period
Dose-response Investigated? No
Statistical Package: SAS v8.02, R
V2.0.1
Pollutant: Particle count coarse (PT PM Increment: IQR
coarse) (pt/cm ]
Averaging Time: Measurements
collected over 48-h period surrounding
the bus trip (during which health
endpoints were measured) used to
calculate 5-, 30-, 60-min, 4-h, and 24-h
ma
Median (IQR):
All: 0.02 (0.11)
Facility: 0.01 (0.04)
Bus: 0.16 (0.13)
Activity: 0.29 (0.26)
Lunch: 0.16 (0.36)
Monitoring Stations: 2 portable carts
Copollutant:
PM25
BC
Fine particle counts
Coarse particle counts
Correlation notes: 24-h mean PM25,
BC, and fine particle count
concentrations ranged from 0.80 to 0.98
r = 0.76 to 0.97 when limited to time
spent on the bus
r = 0.55 to 0.86 when comparing bus
concentrations to 24-h ma
r = -0.003 to 0.51 when comparing 5-
min avg and 24-h ma. Poor correlations
found between coarse particle count
concentrations and all fine particulate
measures during all times periods
Effect Estimate [Lower Cl, Upper Cl]:
% change (95%CI) in HRV per IQR in
the 24-h ma of the microenvironmental
pollutant (IQr = 0.066 pt/cm3)
Single-pollutant models
SDNN: 2.4 (1.3, 3.6)
rMSSD: 3.9 (2.6, 5.1)
pNN50+1:2.9 (1.0, 4.9)
LF: 6.4 (3.7, 9.1)
HF: 10.2 (7.4, 13.1)
LF/HF:-3.3 (-5.0 to-1.6)
H:-1.1 (-1.3(0-0.8)
Two-pollutant models (with PM25):
SDNN:-0.7 (-1.9, 0.6)
rMSSD:-1.3 (-2.6 to-0.05)
pNN50+1:-4.3(-6.3to-2.4)
LF: 0.2 (-2.5, 3.0)
HF: 1.3 (-1.5, 4.1)
LF/HF:-0.9 (-2.7, 1.0)
H:-0.6 (-0.9 to-0.4)
Note: Exposure to health associations
by all lag periods presented in Fig 2
(magnitude of associations increased
with averaging period, with the largest
associations consistently found for 24-h
ma)
Reference: Delfmo et al. (2008,
1563901
Period of Study: 2005-2006
Location: Los Angeles, California, air
basin
Outcome: C-reactive protein (CRP)
Fibrinogen, tumor necrosis factor-a
TNF-a) and its soluble receptor-ll
TNF-RII)
lnterleukin-6 (IL-6) and its soluble
receptor (IL-6sR)
Fibrin D-dimer
Soluble platelet selectin (sP-selectin)
Soluble vascular cell adhesion mole-
cule-1 (sVCAM-1)
Intracellular adhesion molecule-1
(slCAM-1) and myeloperoxidase (MPO)
Erythrocyte lysates for glutathione
peroxidase-1 (GPx-1)
Copper-zinc superoxide dismutase (cu,
Zn-SOD)
Age Groups: 2 65 yr
Study Design: Panel (biomarkers
measured weekly 12 times)
N: 29 participants (nonsmoking with
history of coronary artery disease)
Statistical Analyses: Mixed models
Covariates: temperature (infectious
illnesses were excluded by excluding
weeks with such observations)
Season: Collected 6 wk of data during
warm period and 6 wk of data during
Pollutant: PM (multiple size fractions
and components)
Averaging Time: 24-h avg preceding
the blood draw (lag 0) and cumulative
avg up to 5 days preceding the draw
Outdoor hourly PM: EC: Mean (SD):
1.61 (0.62)
Median: 1.56
IQR: 0.92
Min, Max: 0.24, 3.94
OC: Mean (SD): 5.94 (2.11)
Median: 5.58
IQR: 2.79
Min-Max: 2.51,13.60
BC: Mean (SD): 2.00 (0.77)
Median: 1.89
IQR: 0.96
Min-Max: 0.58, 5.11
OCpri: Mean (SD): 3.37 (1.21)
Median: 3.21
IQR: 1.63
Min-Max: 0.99, 7.11
Secondary OC: Mean (SD): 2.49 (1.50)
Median: 2.10
IQR: 1.86
Min-Max: 0,8.10
PN (p/cm3): Mean (SD): 16,043 (5886)
Median: 13,968
IQR: 7,386
Min-Max: 6837, 31263
Indoor hourly PM EC: Mean (SD):
1.31 (0.52)
Median: 1.30
IQR: 0.70
Min-Max: 0.19, 2.89
PM Increment: IQR
Effect Estimate [Lower Cl, Upper Cl]:
Note: Nearly all results presented in
figures
Results: The authors found significant
positive associations for CRP, IL-6,
sTNF-RII, and sP-selectin with outdoor
and/or indoor concentrations of quasi-
ultrafme PM Ł 0.25 pm in diameter, EC,
OCpri, BC, PN, CO, and nitrogen
dioxide from the current-day and
multiday avg. There were consistent
positive but largely nonsignificant
coefficients for TNF-a, sVCAM-1, and
slCAM-1, but not fibrinogen, IL-6sR, or
D-dimer. The authors found inverse
associations for erythrocyte Cu, Zn-
SOD with these pollutants and other PM
size fractions (0.25-2.5 and 2.5-10 pm).
Inverse associations of GPx-1 and MPO
with pollutants were largely
nonsignificant. Indoor associations were
often stronger for estimated indoor EC,
OCpri, and PN of outdoor origin than for
uncharacterized indoor measurements.
There was no evidence for positive
associations with SOA.
December 2009
E-80
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
cool period
Dose-response Investigated? No
Statistical Package: NR
EC of outdoor origin: Mean (SD): 1.11
(0.39)
Median: 1.06
IQR: 0.51
Min-Max: 0.41, 2.97
OC:Mean(SD):5.69(1.51)
Median: 5.60
IQR: 1.96
Min-Max: 2.34,10.79
OCpri of outdoor origin: Mean (SD):
2.18(0.82)
Median: 2.15
IQR: 1.07
Min-Max: 0.32, 5.21
Secondary OC of outdoor origin: Mean
(SD): 2.08 (1.26)
Median: 1.75
IQR: 1.45
Min-Max: 0, 6.87
PN (particles/cm3): Mean (SD): 14,494
(6770)
Median: 12,341
IQR: 7,337
Min-Max: 1016, 43027
PN of outdoor origin (p/cm3): Mean
(SD):10,108(3108)
Median: 9,580
IQR: 3,684
Min-Max: 1016,17700
Outdoor PM mass PM0.25: Mean
(SD): 9.47 (2.97)
Median: 9.4
IQR: 4.2
Min-Max: 3.31,18.75
PMO.25-2.5: Mean (SD): 13.53 (10.67)
Median: 11.7
IQR: 11.5
Min-Max: 1.29, 66.77
PM10.2.5: Mean (SD): 10.04 (4.07)
Median: 9.9
IQR: 5.9
Min-Max: 1.76, 22.38
Indoor PM mass PM0.25: Mean (SD):
10.45(6.77)
Median: 9.5
IQR: 4.5
Min-Max: 1.42, 69.86
PMO.25-2.5 (pg/m3): Mean (SD): 7.36
(4.57)
Median: 6.5
IQR: 5.7
Min-Max: 0.77, 30.86
PMio-25: Mean (SD): 4.12 (4.76)
Median: 2.8
IQR: 3.5
Min-Max: 0.12, 37.63
Copollutant: Outdoor hourly gases
(N02, CO, 03) and indoor hourly gases
(N02, CO)
Reference: Pekkanen et al. (2002,
0350501
Period of Study: Winter 1998-1999
Location: Helsinki, Finland
Outcome: ST Segment Depression
(>0.1mV)
Study Design: Panel of ULTRA Study
participants
N: 45 Subjects, n = 342 biweekly
submaximal exercise tests, 72 exercise
induced ST Segment Depressions
Statistical Analysis: Logistic
regression / GAM
Pollutant: UltrafmeNCO.01-0.1 pm
(n/cm3)
Averaging Time: 24 h
Median: 14,890
IQR: 9830
Monitoring Stations: 1
Copollutant: N02, CO, PM25, PMi0.25,
PMi.ACP
PM Increment: IQR
Effect Estimate(s): NCO.01-0.1: OR
= 3.14(1.56, 6.32), lag 2
Notes: The effect was strongest for
ACP and PM25, which in 2 pollutant
models appeared independent.
Increases in N02 and CO were also
associated with increased risk of ST
segment depression, but not with
coarse particles.
December 2009
E-81
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Peters et al. (2005,
0957471
Also Peters et al, 2005 (2005,1568591
Period of Study: Feb 1999-Jul 2001
Location: Augsburg, Germany
Outcome: Myocardial infarction
Study Design: Case-crossover
N: 691 myocardial infarction patients
Statistical Analysis: Conditional
logistic regression
Dose-response investigated
(yes/no)? No
Pollutant: Ultrafine (TNC) (n/cm ,
Averaging Time:
1 h: Median = 10,001
IQR: 7919
24 h: Median = 10,934
IQR: 6276
Copollutant: N02, S02, CO
PM Increment: Effect Estimate: 2-h
lag: OR = 0.95
95% Cl: 0.84, 1.06
24-h mean, 2-day lag: OR = 1.04
95% Cl: 0.90, 1.20
Notes: Examined triggering for Ml at
various lags before Ml onset (up to 6 h
before Ml, up to 5 days before Ml). No
statistically significant increases in
lagged ultrafine particle concentration
were found.
Reference: Ruckerl et al. (2006,
0887541
Period of Study: Oct 2000-Apr 2001
Location: Erfurt, Germany
Outcome (ICD9 and ICD10):
C-reactive protein (CRP)
Serum amyloid A (SAA)
E-selectin
von Willebrand Factor (vWF)
Intercellular adhesion molecule-1
(ICAM-1)
Fibrinogen
Factor VII
Prothrombin fragment 1+2
D-dimer
Age Groups: 50+ yr
Study Design: Panel (12 repeated
measures at 2-wk intervals)
N: 57 male subjects with coronary
disease
Statistical Analyses: Fixed effects
linear and logistic regression models
Covariates: Models adjusted for
different factors based on health
end point
CRP: RH, temperature, trend, ID
ICAM-1: temperature, trend, ID
vWF: air pressure, RH, temperature,
trend, ID
FVII: air pressure, RH, temperature,
trend, ID, weekday
Season: Time trend as covariate
Dose-response Investigated?
Sensitivity analyses examined nonlinear
exposure-response functions
Statistical Package: SAS v8.2 and S-
Dh ir« wC n
Pollutant: AP (n/cm3)
Averaging Time: 24 h
Mean (SD): 1593 (1034)
Percentiles:
9"v R91
ZJ. OZ I
50: 1238
75: 2120
Range (Min, Max): 328, 4908
Unit (i.e. pg/m3): n/cm3
Monitoring Stations: 1 site
Copollutant:
UFPs
AP
PM25
PM
r Ivlio
oc
EC
N02
CO
PM Increment: IQR (1299
5-day avg: 11 27)
Effect Estimate [Lower Cl, Upper Cl]:
Effects of air pollution on blood markers
presented as OR (95%CI) for an
increase in the blood marker above the
90th percentile per increase in IQR air
pollutant.
CRP
Time before draw:
0 to 23 h: 0.7 (0.5, 1.2)
24 to 47 h: 1.5 0.9,2.6
48 to 71 h: 3.2 1.7,6.0
5-day mean: 1.5(0.8, 3.0)
ICAM-1
Time before draw:
0 to 23 h: 0.6 (0.4, 0.9)
24 to 47 h' 1 8 12 28
48 to 71 h: 1.6 1.o|2.5
5-day mean: 0.9 (0.6, 1.5)
Effects of air pollution on blood markers
presented as % change from the
mean/GM in the blood marker per
increase in IQR air pollutant.
vWF
Time before draw:
0 to 23 h: 4.8 (0.2, 9.3)
24 to 47 h: 5.9 (0.4, 11.5)
48 to 71 h: 7.0 (0.7, 13.4)
5-day mean: 13.5(6.3,20.6)
FVII
Time before draw:
0 to 23 h: 0.0 (-2.9, 3.0)
24 to 47 h: -2.9 (-6.1, 0.4)
48 to 71 h: -3.6 (-6.8 to -0.3)
5-day mean: -4.1 (-7.9 to -0.3)
Note: Summary of results presented in
figures.
SAA results indicate increase in
association with PM (not as strong and
consistent as with CRP)
No association observed between E-
selectin and PM
An increase in prothrombin fragment
1+2 was consistently observed,
particularly with lag 4
Fibrinogen results revealed few
significant associations, potentially due
to chance
D-dimer results revealed null
associations in linear and logistic
analyses
December 2009
E-82
-------
Reference
Reference: Ruckerl et al. (2006,
0887541
Period of Study: Oct 2000-Apr 2001
Location: Erfurt, Germany
Design & Methods
Outcome: Soluble CD40 ligand
(SCD40L), platelets, leukocytes,
erythrocytes, hemoglobin
Age Groups: 50+ yr
Study Design: Panel (12 repeated
measures at 2-wk intervals)
N: 57 male subjects with coronary
disease
Statistical Analyses: Fixed effects
linear regression models
Covariates: Long-term time trend,
weekday of the visit, temperature, RH,
barometric pressure
Season: Time trend as covariate
Dose-response Investigated? No
Statistical Package: SAS v8.2 and S-
piusve.o
Concentrations1
Pollutant: AP (n/cm3)
Averaging Time: 24 h
Mean (SD): 1593 (1034)
Percentiles:
25th: 821
50th: 1238
75th: 2120
Range (Min, Max): 328, 4908
Monitoring Stations: 1 site
Copollutant:
UFPs
AP
PM25
PM10
NO
Effect Estimates (95% Cl)
PM Increment: IQR (1299
5-day avg: 11 27)
Effect Estimate [Lower Cl, Upper Cl]:
Effects of air pollution on blood markers
presented as % change from the
mean/GM in the blood marker per
increase in IQR air pollutant.
SCD40L, % change GM (pg/mL)
lagO: 6.9 (0.5, 13.8)
Iag1:-1.1(-8.0, 6.4)
Iag2: -4.9 (-11.9, 2.7)
Iag3: -3.8 (-10.3, 3.2)
5-day mean: -1.3 (-9.9, 8.1)
Platelets, % change mean (103/ul)
Iag0:-1.0 -2.5,0.5
Iag1:-0.4 -2.1, 1.6
Iag2:0.8(-1.0,2.4)
Iag3:0.0(-1.8, 1.7)
5-day mean: -0.9 (-3.0, 1.3)
Leukocytes, % change in mean
(103/ul)
Iag0:-1.9(-3.8to-0.1)
lag1:-0.6(-2.9, 1.6)
Iag2: -0.6 (-3.2, 2.0)
Iag3: -2.3 (-4.6, 0.1)
5-day mean: -2.7 (-5.5, 0.1)
Erythrocytes, % change mean
(106/ul)
Iag0:-0.1(-0.5, 0.3)
Iag1:-0.4 -0.9,0.2
Iag2: -0.4 -0.9, 0.2
Iag3: -0.4 (-0.6, 0.3)
5-day mean:-0.4 (-1.0, 0.2)
Hemoglobin, % change mean (g/dl)
lagO: -0.2 (-0.7, 0.4)
lag 1:-0.3
Iag2:-0.1
-1.0,0.4
-0.9, 0.7
Iag3:-0.1(-0.8, 0.6)
5-day mean: -0.2 (-1.1, 0.6)
Reference: Ruckerl et al. (2007,
1569311
Period of Study: May 2003-Jul 2004
Location: Athens, Augsburg,
Barcelona, Helsinki, Rome, and
Stockholm
Outcome: lnterleukin-6
(IL-6), fibrinogen, C-reactive protein
(CRP)
Age Groups: 35-80 yr
Study Design: Repeated measures /
longitudinal
N: 1003 Ml survivors
Statistical Analyses: Mixed-effect
models
Covariates: City-specific confounders
(age, sex, BMI)
Long-term time trend and apparent
temperature
RH, time of day, day of week included if
adjustment improved model fit
Season: Long-term time trend
Dose-response Investigated? Used p-
splines to allow for nonparametric
exposure-response functions
Statistical Package: SAS v9.1
Pollutant: UFP (n/cm3)
Averaging Time: Hourly and 24 h (lag
0-4, mean of lags 0-4, mean of lags 0-1,
mean of lags2-3, means of lags 0-3)
Mean (SD): Presented by city only
Percentiles: NR
Range (Min, Max): NR
Monitoring Stations: Central
monitoring sites in each city
Copollutant:
S02
03
NO
NO,
PM Increment: IQR
Effect Estimate [Lower Cl, Upper Cl]:
% change in mean blood markers per
increase in IQR of air pollutant.
IL-6
Lag (IQR): % change in GM (95%CI)
Lag 0(11852): 1.88 (-0.16, 3.97)
Lag 1(11852):-0.67 (-2.56, 1.25)
Lag2(11852):-2.12 (-4.03 to-0.17)
5-day avg (11003): -0.93 (-3.37,1.56)
Fibrinogen
Lag (IQR): % change in AM (95%CI)
Lag 0(11852): 0.40 (-0.40, 1.19)
Lag1 (11852): 0.11 (-0.69,0.91)
Lag 2 (11852): 0.09 (-0.71, 0.90)
5-day avg (11003): 0.50 (-2.20, 3.20)
CRP
Lag (IQR): % change in GM (95%CI)
Lag 0(11852): 1.33 (-3.05, 5.90)
Lag1
Lag 2
11852
11852
-1.52
-4.39, 1.45
-6.70, 3.71
5-day avg (11003): -0.08 (-3.78, 3.75)
December 2009
E-83
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Pekkanen et al. (2002,
0350501
Period of Study: Winter 1998-1999
Location: Helsinki, Finland
Outcome: ST Segment Depression
(>0.1mV)
Age Groups: Study Design: Panel of
ULTRA Study participants
N: 45 Subjects, n = 342 biweekly
submaximal exercise tests, 72 exercise
induced ST Segment Depressions
Statistical Analysis: Logistic
regression / GAM
Pollutant: Ultrafme NCO.01-0.1 pm PM Increment: IQR
(n/cm ]
Averaging Time: 24 h
Median: 14,890
IQR: 9830
Monitoring Stations: 1
Copollutant: N02, CO, PM25, PMio.25,
PM,,ACP
Effect Estimate(s): NCO.01-0.1: OR
= 3.14(1.56, 6.32), lag 2
Notes: The effect was strongest for
ACP and PM25, which in 2 pollutant
models appeared independent.
Increases in N02 and CO were also
associated with increased risk of ST
segment depression, but not with
coarse particles.
Reference: Peters et al. (2005,
0957471
Also Peters et al, 2005 (2005,1568591
Period of Study: Feb 1999-Jul 2001
Location: Augsburg, Germany
Outcome: Myocardial infarction
Study Design: Case-crossover
N: 691 myocardial infarction patients
Statistical Analysis: Conditional
logistic regression
Dose-response Investigated? No
Pollutant: Ultrafme (TNC) (n/cm3)
Averaging Time: 1 h: Median = 10,001
IQR: 7919
24-h: Median = 10,934
IQR: 6276
Copollutant: N02, S02, CO
PM Increment: Effect Estimate:
2-h lag: OR = 0.95
95% Cl: 0.84, 1.06
24-h mean, 2-day lag: OR = 1.04
95% Cl: 0.90, 1.20
Notes: Examined triggering for Ml at
various lags before Ml onset (up to 6 h
before Ml, up to 5 days before Ml). No
statistically significant increases in
lagged ultrafme particle concentration
were found.
Reference: Ruckerl et al. (2007,
0913791
Period of Study: Oct 2000-Apr 2001
Location: Erfurt, Germany
Outcome (ICD9 and ICD10): Soluble
CD40 ligand (sCD40L), platelets,
leukocytes, erythrocytes, hemoglobin
Age Groups: 50+ yr
Study Design: Panel (12 repeated
measures at 2-wk intervals)
N: 57 male subjects with coronary
disease
Statistical Analyses: Fixed effects
linear regression models
Covariates: Long-term time trend,
weekday of the visit, temperature, RH,
barometric pressure
Season: Time trend as covariate
Pollutant: UFP
Averaging Time: 24 h
Mean (SD): 12,602 (6455)
Percentiles:
25th: 7326
50th: 11,444
75th: 17,332
Range (Min, Max): 328, 4908
Monitoring Stations: 1 site
Copollutant:
AP
PM Increment: IQR (10,005
5-day avg: 6,821)
Effect Estimate [Lower Cl, Upper Cl]:
SCD40L, % change GM (pg/mL)
lag 0:7.1 (0.1,14.5)
lag 1:0.3 (-6.6, 8.6
lag 2: 0.6 (-5.9, 8.6
lag 3:-8.5 (-15.8,-0.5)
5-day mean: -0.7 (-7.6, 6.8)
Platelets, % change mean (103/ul)
lag 0:-1.8 (-3.4,-0.2)
lag 1:-1.1 (-2.9,0.6)
lag 2:1.0 (-2.9, 0.8)
lag 3: -2.4(4.5, -0.3)
5-day mean: -2.2 (-4.0, -0.3)
Leukocytes, [103/ul]
lag 0: -2.4 (-4.5, -0.2)
Dose-response Investigated? No
Statistical Package: SAS v8.2 and S-
piusve.o
nvi25
PM10
NO
lag 1: -2.1
lag 2: -0.2
lag 3: -1.5
5-day mea
-4.4, 0.2)
-2.4, 2.8)
-4.4, 1.4)
r. -1.6 (-4.1, 0.8)
All units expressed in pg/m unless otherwise specified.
December 2009
E-84
-------
E.1.2.Cardiovascular Emergency Department Visits and Hospital
Admissions
Table E-5. Short-term exposure-cardiovascular: ED/HA PM
10
Reference
Design & Methods
Concentrations
Effect Estimates (95% Cl)
Reference: Anderson et al. (2003,
0548201
Period of Study: 1992-1994
Location: London, U.K.
Outcome: Ischemic Heart Disease
Age Groups: 0-15, 15-64, 65-74, 75+
Study Design: Time series
N:NR
Statistical Analyses: NR
Covariates: NR
Dose-response Investigated? No
Statistical Package: NR
Pollutant: PM10
Averaging Time: 24 h
Mean (min-max): NR
Monitoring Stations: NR
Copollutant: NR
PM Increment: 10th-90th percentile
% Change in Daily IHD Admissions
byAge[CI]:0-15yr:NR
15-64 yr: 2.6 [0.3,5]
65-74 yr: 2.5 [0.1,4.9]
75+ yr: 2.2 [0.2,4.6]
Notes: RRs are presented in graph
form showing little change with
increasing age (PM increment of
10 |jg/m3). This article is primarily a
systematic literature review of other
studies.
Reference: Andersen et al. (2008,
1896511
Period of Study: May 2001-Dec 2004
Location: Copenhagen, Denmark
Pollutant: PM,,
Outcome (ICD-10): CVD, including
angina pectoris (I20), myocardial
infarction (121-22), other acute ischemic Averaging Time: 24 h
heart diseases (I24), chronic ischemic
heart disease (I25), pulmonary
embolism (I26), cardiac arrest (I46),
cardiac arrhythmias (148-48), and heart
failure (ISO).
..„„ ,.ra. ,,,,,>
Mean ' 24<14'
Median: 21
IQR: 16-28
Age Groups: >65 yr (CVD and RD),
5-18 yr (asthma)
Study Design: Time series
N:NR
Statistical Analyses: Poisson GAM
Covariates: Temperature, dew-point
temperature, long-term trend,
seasonally, influenza, day of the week,
public holidays.
Season: NR
Dose-response Investigated: No
Statistical Package: R (gam
procedure, mgcv package)
Lags Considered: Lag 0 -5 days, 4-
day pollutant avg (lag 0 -3) for CVD.
99th percentile): 72
Monitoring Stations: 1
Copollutant (correlation):
NCtot:r = 0.39
NC100:r = 0.28
NCa12:r = 0.02
NCa23:r = -0.12
NCa57:r = 0.45
NCa212:r = 0.63
PM25:r = 0.80
CO: r = 0.37
N02:r = 0.35
N0x:r = 0.32
N0xcurbside:r = 0.18
03:r = -0.21
Other variables:
Temperature: r = 0.12
Relative humidity: r = 0.05
PM Increment: 13 pg/m3 (IQR)
Relative risk (RR) Estimate [Cl]:
CVD hospital admissions
(4-day avg, lag 0 -3), age 65+:
One-pollutant model: 1.03 [1.01-1.05]
Adj for NCtot: 1.04 [1.02-1.06]
Adj for NCa212:1.05 [1.01-1.09]
RD hospital admissions
(5-day avg, lag 0 -4), age 65+:
One-pollutant model: 1.06 [1.02-1.09]
Adj for NCtot: 1.05 [1.01-1.10]
Adj for NCa212:1.04 [0.98-1.11]
Asthma hospital admissions
(6-day avg lag 0-5), age 5-18:
One-pollutant model: 1.02 [0.93-1.12]
Adj for NCtot: 1.01 [0.91-1.12]
Adj for NCa212: 0.94 [0.81-1.09]
Estimates for individual day lags
reported only in Fig form (see notes):
Notes: Fig 2: Relative risks and 95%
confidence intervals per IQR in single
day concentration (0- to 5-day lag).
Summary of Fig 2: CVD: Positive,
marginally or statistically significant
associations at Lag 0-Lag 2.
December 2009
E-85
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Anderson et al. (2007,
1562141
Period of Study: January
1999-December2004
Location: Copenhagen, Denmark
Outcome (ICD10): Hospital Admission,
CVD, including angina pectoris (I20),
myocardial infarction (121-22), other
acute ischemic heart diseases (I24),
chronic ischaemic heart disease (I25),
pulmonary embolism (I26), cardiac
arrest (I46), cardiac arrhythmias
(148-48), and heart failure (ISO).
Age Groups Analyzed: Age >65
Study Design: Time series
N: 2192 days, 9 Hospitals
Statistical Analyses: Principal
Component Analysis and Constrained
Physical Receptor Model (COPREM),
Poisson regression, GAM,
Covariates: Season, day of the wk,
public holidays, influenza epidemics
and meteorology
Season: All yr
Dose-response Investigated? No
Statistical package: R, gam/mgcv
package
Lags Considered: 0-6 days
Pollutant: Source specific PMi0
components
Averaging Time: 24 h
Mean (SD): Percentiles:
25th: 16
50th (Median): NR
75th: 30
Monitoring Stations: 1
Copollutant (correlation):
PM10:
Biomass: r = 0.53
Secondary: r = 0.73
Oil: r = 0.57
Crustal:r = 0.37
Sea salt: r = 0.04
Vehicle: r = 0.02
Notes: Correlations between source
specific PM10 components presented in
paper
PM Increment: IQR
RR Estimate
Respiratory disease (age >66)
Single pollutant model:
PM10:1.027 (1.013, 1.042), IQR=14
PM,o (other 5 sources): 1.045 (1.016,
1.074), IQR=13
Biomass: 1.040 (0.009, 1.072), IQR=5.4
Secondary: 1.050 (1.021,1.081),
IQR=6.1
011:1.035(1.006, 1.065), IQR=2.8
Crustal: 1.054 (1.028, 1.081), IQR=1.8
Sea salt: 0.98 (0.947, 1.017), IQR=2.2
Vehicle: 0.989 (0.949, 1.032), IQR=0.6
Notes: 2 pollutant model results for
PMio with source specific components
and gases also presented in
manuscript.
Reference: Baccarelli et al. (2007,
0913101
Period of Study: Jan 1995-Aug 2005
Location: Lombardia region, Italy
Outcome (ICD9 and ICD10): Fasting
and postmethionine-load total
homocysteine (tHcy)
Age Groups: 11-84yr
Study Design: Cross-sectional/Panel
N: 1,213 participants
Statistical Analyses: Generalized
additive models
Covariates: age, sex, BMI, smoking,
alcohol, hormone use, temperature, day
of the yr, and long-term trends
Season: Adjusted for long-term trends
to account for season
Dose-response Investigated? No
Statistical Package: R software v2.2.1
Pollutant: PM10 (some TSP measures
used to predict PMio)
Averaging Time: Hourly concentrations
used to calculate 24-h ma and 7-day
ma
Mean (SD): NR
Percentiles:
25th: 20.1
50th: 34.1
75th: 52.6
Range (Min, Max): Max: 390.0
Monitoring Stations: 53 sites
Copollutant:
CO
N02
S02
03
PM Increment: IQR
Effect Estimate [Lower Cl, Upper Cl]:
Estimates (%) per 32.5 pg/m3 increase
in 24-h maofPM10
Homocysteine, fasting: 0.4 (-2.4, 3.3)
Homocysteine, postmethionine-load:
(-1.5,3.7)
Estimates (%) per 25.7m3 increase in 7-
day maof PMio
Homocysteine, fasting: 1.0 (-1.9, 3.9)
Homocysteine, postmethionine-load:
2.0 (-0.6, 4.7)
Estimates of effect (%) on fasting
homocysteine per IQR increase in 24-h
PM,o levels
Among smokers: 6.2 (0.0,12.7)
Among non-smokers: -1.6 (-5.5, 2.5)
Estimates of effect (%) on
postmethionine-load homocysteine per
IQR increase in 24-h PMio levels
Among smokers: 6.0 (0.5,11.8)
Among non-smokers: -0.1 (-3.6, 3.5)
December 2009
E-86
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ballester et al. (2006,
0887461
Period of Study: 1995-1999
Location: 5 Spanish cities: Granada,
Huelva, Madrid, Seville, Zaragoza
Outcome (ICD-9): All cardiovascular
disease (390-459), including all heart
diseases (410-414, 427, 428)
Age Groups: All ages
Study Design: Time series
N:NR
Statistical Analyses: Poisson GAMs
Covariates: Dilytemp, barometric
pressure relative humidity
Daily influenza incidence, day of the
week, holidays, unusual events (ex.
medical strikes), seasonal variation,
trend
Dose-response Investigated: No
Statistical Package: S-Plus GAM
function
Lags Considered: lag 0-3 days, lag
0-1 avg
Pollutant: PM,0
Averaging Time: 24 h
Mean (10-90th percentile): overall
mean NR.
City specific means
Granada: 43.2 (24.8, 62.6)
Huelva: 38.6 (23.1, 57.3)
Madrid: 35.7 (21.4, 54.4)
Seville: 41.9 (27.3, 57.6)
Zaragoza: 32.8 (17.3, 50.3)
Monitoring Stations: At least three
stations/city (15+)
Copollutant (correlation): Summary of
the correlation coefficients between
each pair of pollutants within cities:
BS:r = 0.48
TSP: N/A
N02: from r = 0.13tor = 0.62
(median r = 0.40)
S02: from r = 0.20tor = 0.51
(median r = 0.46)
CO: from r = 0.34 to r = 0.45
(median r = 0.37)
03:from r = -0.07 to r = 0.16
(median r = 0.11)
PM Increment: 10 pg/m
Relative risk [Cl]: Relative risks are
expressed only in the form of figures
(see notes).
Percentage change in risk [Cl]: All
cardiovascular diseases (avg of lags 0 -
1): 0.91% [0.35, 1.47]
Heart disease (avg of lags 0 -1)
1.56% [0.82, 2.31]
Notes: Relative risks for the single
pollutant models are expressed in
Fig 2.
Fig 2: Time sequence of the combined
association between PM10 and hospital
admissions for all CVD (A) and heart
disease (B).
Summary of results: Significant, positive
association of PMi0 with both overall
CVD and heart disease hospitalizations
at Lag 0 and Lag 1.
Relative risks for 2 pollutant models
are expressed in Fig 3: Fig 3:
Combined estimates of the association
between hospital admissions for heart
diseases and air pollutants (avg of lags
0-1
Adjusted for CO, N02, 03, orS02)
Summary of results: Significant, positive
association remains after adjusting for
pollutants.
December 2009
E-87
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Bell et al. (2008, 0912681
Period of Study: 1995-2002
Location: Taipei, Taiwan
Outcome (ICD-9): Hospital admissions
for ischemic heart disease (410, 411,
414), cerebrovascular disease
(430-437).
Age Groups: All
Study Design: Time series
N: 6,909 hospital admissions for
ischaemic heart diseases, 11,466 for
cerebrovascular disease.
Statistical Analyses: Poisson
regression
Covariates: Day of the week, time,
apparent temperature, long-term trends,
seasonality
Season: All
Dose-response Investigated: No
Statistical Package: NR
Lags Considered: lags 0-3 days, avg
of lags 0-3
Pollutant: PM,0
Averaging Time: 24 h
Mean (range
IQR):
49.1 (127-215.5
27.6)
Monitoring Stations: Taipei area: 13
monitors
Taipei City: 5 monitors
Monitors with correlations of 0.75 + for
PMi0:12 monitors
Copollutant: NR
PM Increment: 28 pg/m (near IQR)
Percentage increase estimate [96%
Cl]: Ischemic heart disease: Taipei
area (13 monitors): LO: 1.91 (-1.25,
5.17)
L1:0.39 (-2.73, 3.61)
L2:1.80 (-1.33, 5.04)
L3:2.01 (-1.14,5.26)
LOS: 2.91 (-1.52, 7.55)
Taipei City (5 monitors): LO: 2.08 (-1.04,
5.30)
L1: 0.43 (-2.64, 3.60)
L2: 2.17 (-0.92, 5.36)
L3: 2.16 (-0.94, 5.36)
LOS: 3.40 (-1.19, 8.20)
Monitors with > = 0.75 between monitor
correlations (12 monitors): LO: 1.82
(-1.29,5.03)
L1: 0.35 (-2.72, 3.52)
L2:1.93 (-1.15, 5.10)
L3:1.93 (-1.16, 5.12)
LOS: 2.86 (-1.63, 7.54)
Cerebrovascular disease: Taipei area
(13 monitors): LO:-1.41 (-3.80,1.04)
L1:-1.95 (4.31, 0.48)
L2: 0.77 (-1.62, 3.23)
L3: 2.64 (0.21, 5.12)
LOS: 0.01 (-3.33, 3.47)
Taipei City (5 monitors): LO: -1.27
(-3.64, 1.16)
L1:-2.13 (-4.47, 0.27)
L2: 0.85 (-1.52, 3.28)
L3: 2.52 (0.13, 4.97)
LOS: -0.07 (-3.53, 3.51)
Monitors with > = 0.75 between monitor
correlations (12 monitors): LO: -1.34
(-3.70,1.07)
L1:-1.98 (-4.31, 0.40)
L2: 0.80 (-1.56, 3.22)
L3: 2.61 (0.22, 5.05)
L03: -0.02 (-3.40, 3.49)
Reference: Chan et al. (2007,1477871 Outcome: Cerebrovascular Emergency Pollutant: PM10
Period of Study: Apr 1997-Dec 2002 Adm lsslons
Age Groups: 50+ yr
Location: Boston, MA
Study Design: Tme series
Statistical Analyses: GAM Poisosn
Regression
Covariates: Yr, mo, day of wk,
temperature, dew point
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0-3 days
Averaging Time: 24 h
Mean (SD): 50.2 (22.1)
Min:16.0
Max: 325.4
IQR: 25.4
Monitoring Stations: 16
Copollutant: 03, CO, S02, N02, PM25
Co-pollutant Correlation:
03: 0.43
CO: 0.47
S02: 0.59
N02: 0.64
PM25: 0.61
PM Increment: Interquartile Range
(25.4 pg/m3)
Percent Change (Lower Cl, Upper Cl),
p-value:
Cerebrovascular Disease
Lag 0:1.001 (0.969, 1.033)
Lag 1:0.999 (0.9787, 1.020)
Lag 2:1.023 (0.989, 1.057)
Lag 3:1.030 (1.011,1.049)
Lag 3+ 03:1.018 (0.987, 1.049)
Lag 3 + 00:1.019(0.988, 1.050)
Lag 3+ 03+ 00:1.015(0.985, 1.045)
Stroke
Lag 0: 0.969 (0.897, 1.041)
Lag 1:0.992 (0.918, 1.066)
Lag 2:1.004 (0.993, 1.015)
Lag 3:1.009 (0.988, 1.030)
Ischaemic stroke
Lag 0: 0.984 (0.932, 1.036)
Lag 1:0.993 (0.939, 1.047)
Lag 2: 0.989 (0.927, 1.041)
Lag 3:1.042 (0.981, 1.103)
Haemorrhagic stroke
Lag 0: 0.966 (0.884, 1.048)
Lag 1:0.990 (0.908, 1.072)
Lag 2:1.002 (0.920, 1.084)
Lag 3: 0.974 (0.902, 1.046)
December 2009
E-88
-------
Reference
Reference: Chan et al. (2008, 0932971
Period of Study: 1995-2002
Location: Taipei Metropolitan area,
Taiwan
Reference: Chang et al. (2007,
1476211
Period of Study: 1997-2001
Location: Taipei, Taiwan
Reference: D'lppoliti et al. (2003,
0743111
Period of Study: Jan 1995-Jun 1997
Location: Rorne Italy
Design & Methods
Outcome (ICD-9): Emergency visits for
ischaemic heart diseases (410-411,
414), cerebrovascular diseases
(430-437), and COPD (493, 496)
Age Groups: All
Study Design: Time series
N:NR
Statistical Analyses: Poisson
regression models
Covariates: Yr, mo, day of wk,
temperature, dew point temperature,
PM2.5, N02
Season: All
Dose-response Investigated: No
Statistical Package: SAS version 8.0
Lags Considered: 0- to 7-day lags
Outcome: CVD HA
Age Groups: NR
Study Design: Case-crossover
Statistical Analyses: Conditional
Logistic Regression
Covariates: Temperature, humidity
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-2 days
Outcome: Myocardial Infarction HA
Age Groups: 18+yr
Study Design: Case-crossover
Statistical Analyses: Conditional
Logistic Regression
Covariates: Temperature, humidity
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0-4 days
Concentrations1
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): High dust events: Pre-dust
periods: 45.5 (17.6)
Asian dust events: 122.7 (24.4)
Low dust events: Pre-dust periods: 59.4
(31.0)
Asian dust events: 61.1 (17.8)
Monitoring Stations: 1
Copollutant: NR
Pollutant: PM10
Averaging Time: 24 h
Mean: 48.32
Min: 14.44
26th: 32.65
60th: 42.80
76th: 57. 16
Max: 234.91
Monitoring Stations: 6
Copollutant: 03, CO, S02, N02
Co-pollutant Correlation: NR
Pollutant: TSP
Averaging Time: 24 h
Mean (SD): 66.9 (19.7)
26th: 54.7
60th: 66.4
76th: 78.4
IQR: 23.7
Monitoring Stations: 3
Copollutant: CO, S02, N02
Co-pollutant Correlation:
CO: 0.35
S02: 0.29
N02: 0.38
Effect Estimates (95% Cl)
PM Increment: 25.4 pg/m3 (IQR)
OR [96% Cl]: In environmental
conditions without dust storms (results
only shown for best-fitting model)
Lag 3 days: 1.023 (1.003, 1.041)
PM Increment: Interquartile Range
(24.51 pg/m3)
Odds Ratio (Lower Cl, Upper Cl):
ion°p
ŁzU U
PM,0: 1.085 (1.061, 1.110)
PM10+S02: 1.131 (1.103, 1.161)
PM10+N02: 10.977 (0.950, 1.006)
PM,o+ CO: 1.025 (0.999, 1.052)
PM10+03: 1.064 (1.039, 1.090)
<20°C
PM,0: 1.142 (1.105, 1.180)
PM10+S02: 1.235 (1.184, 1.288)
PM,o+N02: 1.148 (1.103, 1.194)
PM10+ CO: 1.165 (1.121, 1.212)
PM10+03: 1.142 (1.105, 1.180)
PM Increment: Quartiles
Odds Ratio (Lower Cl, Upper Cl):
Lag 0-2-day avg
f-\i. A n lrc.f\
ui. i.u (ret)
011:1.048(0.957, 1.148)
Qlll: 1.105(1.007, 1.214)
QIV: 1.132 (1.023, 1.253)
Various Lags
Lag 0:1.023 (1.004, 1.042)
Lag 1:1. 01 5 (0.996, 1.034)
Lag 2: 1.017 (0.999, 1.035)
Lag 3: 0.989 (0.974, 1.003)
Lag 4: 1.001 (0.987, 1.016)
December 2009
E-89
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Fung et al, (2005, 0932621
Period of Study:
Nov1995-Dec2000
Location: London, Ontario
Outcome (ICD-9): Cardiovascular
diseases
(410-414, 427-428)
Age Groups: <65 yr, 65+ yr
Study Design: Time series
N: 12,947 CVD admissions
Statistical Analyses: GAM with locally
weighted regression smoothers
(LOESS)
Covariates: Maximum and minimum
temp, humidity, day of the week,
seasonal cycles, secular trends
Season: NR
Dose-response Investigated? No
Statistical Package: S-Plus
Lags Considered: Current to 3-day
Pollutant: PM,0
Averaging Time: 24 h
Mean (min-max): 38.0 (5-248)
SD = 23.5
Monitoring Stations: 4
Copollutant (correlation):
N02:r = 0.30
S02:r = 0.24
CO: r = 0.21
03:r = 0.53
COM: r = 0.29
PM Increment: 26 pg/m
% Change in Daily Admission [Cl]:
Age <65
Current day mean: 2.6 [-2.3,7.7]
2-day mean:-1.2 [-7.2,5.1]
3-day mean: -3 [-9.6,4]
Age 65+
Current day mean: 0.9 [-2.3,4.2]
2-day mean: -0.9 [-4.8,3.2]
3-day mean:-0.1 [-4.4,4.5]
Reference: Hanigan et al. (2008,
1565181
Period of Study: 1996-2005
(Apr-Nov of each yr)
Location: Darwin, Australia
Outcome: Daily emergency hospital
admissions for total cardiovascular
(ICD-9: 390-459
ICD-10: IOO-I99), ischemic heart
disease (ICD-9: 410-414
ICD-10:120-I25).
Age Groups: All
Study Design: Time series
N: 8,279 hospital admissions
Statistical Analyses: Poisson
generalized linear models
Covariates: Indigenous status, time in
days, temperature, relative humidity,
day of the week, influenza epidemics,
change between ICD editions, holidays,
yrly population
Season: Apr-Nov (corresponding to the
dry season)
Dose-response Investigated? No
Statistical Package: R version 2.3.1
Lags Considered: 0-3
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD
range): 21.2 (8.2
55.2)
Monitoring Stations: N/A (see notes)
Copollutant: NR
PM Increment: 10 pg/m
Percent change [96% Cl]: Overall
CVD: Lag 0 (indigenous): -3.78 [-13.4,
6.91]
Lag 0 (non-indigenous): -3.43 [-9.00,
2.49]
All unstratified associations either
negative or zero and not statistically
significant.
All other results of stratified analysis (by
indigenous status) reported in a Fig
(see notes).
Notes: Fig 3: Associations between
hospitalizations for non-indigenous and
indigenous people with estimated
ambient PM10. Summary: Confidence
intervals were wide, but indigenous
people generally had stronger
associations with PM10 than non-
indigenous people. Daily PMi0 exposure
levels were estimated for the population
of the city from visibility data using a
previously validated models.
December 2009
E-90
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Hanigan et al. (2008,
1565181
Period of Study: 1996-2005
(Apr-Nov of each yr)
Location: Darwin, Australia
Outcome: Cardiorespiratory Disease
HA (ICD 9: 390-519
ICD10:IOO-99SJOO-99)
Age Groups: NR
Study Design: Time series
N: 8279 events
Statistical Analyses: poisson
regression
Covariates: Indigenous status, time in
days, temperature, relative humidity,
day of the week, influenza epidemics,
change between ICD editions, holidays,
yearly population
Dose-response Investigated? No
Statistical Package: R
Lags Considered: lags 0-3
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 21.2 (8.2)
Range: 55.2
Monitoring Stations: 2 (monitored &
modeled)
Copollutant: NR
Co-pollutant Correlation: N/A
PM Increment: 10 pg/m
Percent Change (Lower Cl, Upper Cl),
lag:
Tot. Cardiovascular, Indigenous: -3.43 (-
9.00, 2.49), lag 0
Tot Cardiovascular, Non-Indigenous: -
3.78 (-13.4, 6.91), lag 0
*Fig 3. percent change in hospital
admissions per 10 pg/m3 increase in
PM10
Reference: Henrotin et al. (2007,
0932701
Period of Study: Mar 1994-Dec 2004
Location: Dijon, France
Outcome: Ischemic and hemorrhagic
strokes
Age Groups: All
Study Design: Bi-directional case-
crossover
N: 1487 (ischemic) and 220
(hemorrhagic) stroke patients
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature, relative
humidity, influenza epidemics, holidays
Season: NR
Dose-response Investigated? Yes
Statistical Package: STATA software v.
8.2
Lags Considered: 0-3 days
Pollutant: PM10
Averaging Time: 24 h
Mean (min-max):
21.1 (2-103)
SD=11.3
Monitoring Stations: 1
Copollutant: NR
PM Increment: 10 pg/m3
OR Estimate [Cl]: Ischemic stroke
Same-day lag: 1.009 [0.930,1.094]
1-day lag: 1.011 [0.998,1.094]
2-day lag: 0.960 [0.889,1.036]
3-day lag: 0.990 [0.919,1.066]
Hemorrhagic stroke
Same-day lag: 0.901 [0.730,1.111]
1-day lag: 1.014 [0.828,1.241]
2-day lag: 1.100 [0.903,1.339]
3-day lag: 0.991 [0.881,1.212]
Notes: Ischemic stroke ORswere also
categorized into male and female,
yielding similar results (none were
significant for any lag days).
Reference: Issever et al. (2005,
0977361
Period of Study:
Jan1997-Dec2001
Location: Istanbul, Turkey
Outcome: Acute coronary syndrome
(ACS)
Age Groups: All
Study Design: Time series
N: 2889 ACS admissions
Statistical Analyses: Multiple stepwise
regression, Pearson correlation
Covariates: Humidity, temperature,
pressure
Season: NR
Dose-response Investigated? No
Pollutant: PM10
Averaging Time: 24 h
Mean: NR
Monitoring Stations: 1
Copollutant (correlation):
ACS: r = 0.37 (p = 0.003)
ACS controlled for temp: r = 0.29
(p = 0.02)
PM Increment: NR
RR Estimate [Cl]: NR
Notes: This study focused more on the
seasonal change in acute coronary
syndrome admissions.
December 2009
E-91
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Jalaludin et al. (2006,
1894161
Period of Study: Jan 1997-Dec 2001
Location: Sydney, Australia
Outcome (ICD-9): Cardiovascular
disease (390-459), cardiac disease
(390-429), ischemic heart disease
(410-413) and cerebrovascular disease
or stroke (430-438)
Age Groups: 65+ yr
Study Design: Time series
N:NR
Statistical Analyses: GAM, GLM
Covariates: Temperature, humidity
Season: Warm (Nov-Apr) and cool
(May-Oct)
Dose-response Investigated? No
Statistical Package: S-Plus
Lags Considered: 0-3
Pollutant: PM,0
Averaging Time: 24 h
Mean (min-max): 16.8 (3.8-103.9)
SD = 7.2
Monitoring Stations: 14
Copollutant (correlation):
Warm
BSP:r = 0.82
PM25:r = 0.89
03:r = 0.59
N02:r = 0.44;
CO: r = 0.31
S02:r = 0.37
Cool
BSP:r = 0.75
PM25:r = 0.88
03:r = 0.22
N02:r = 0.67
CO: r = 0.48
S02:r = 0.46
Other variables:
Warm
Temp: r = 0.36
Pel humidity: r = -0.25
Cool
Temp: r = 0.13
Pel humidity: r = 0.05
PM Increment: 7.8 pg/rri (IQR)
Percent Change Estimate [Cl]:
All CVD
Same-day lag: 0.72 [-0.14,1.60]
Avg 0-1 day lag: 0.25
[-0.61,1.12]
Cool (same-day lag): 1.34 [0.08,2.61]
Warm (same-day lag): 0.33 [-0.83,1.50]
Cardiac disease
Same-day lag: 1.15 [0.14,2.18]
Avg 0-1 day lag: 0.97
[-0.07,2.02]
Cool (same-day lag): 1.35 [-0.16,2.89]
Warm (same-day lag): 1.12 [-0.23,2.48]
Ischemic heart disease
Same-day lag: 0.59 [-0.95,2.17]
Avg 0-1 day lag: 0.61
[-0.95,2.20]
Cool (same-day lag): 0.33 [-2.00,2.72]
Warm (same-day lag): 0.79 [-1.23,2.85]
Stroke
Same-day lag: -1.66 [-3.48,0.20]
Avg 0-1 day lag: -2.05
[-3.88.-0.20]
Cool (same-day lag): 0.46 [-2.17,3.17]
Warm (same-day lag): -3.49 [-5.97,-
0.95]
Notes: All other lag-day ORs were
provided, yet none were significant.
Percent change in ED attendance was
also reported graphically
(Fig 1-5).
Reference: Johnston et al. (2007,
1558821
Period of Study: 2000, 2004, 2005
(Apr-Nov of each yr)
Location: Darwin, Australia
Outcome (ICD-10): All cardiovascular
conditions (IOO-I99), including ischemic
heart disease (I20-I25).
Age Groups: All
Study Design: Case-crossover
N: 2466 emergency admissions
Statistical Analyses: Conditional
logistic regression
Covariates: Wsekly influenza rates,
temperature, humidity, days with rainfall
>5mm, public holidays, school holiday
periods (for respiratory conditions only)
Season: Apr-Nov (dry season)
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0-3
Pollutant: PM,0
Averaging Time: 24 h
Median: 17.4
IQR: 13.6-22.3
10-90th Percentile: 10.3-27.7
Range: 1.1-70.0
Monitoring Stations: 1
Copollutant: NR
PM Increment: 10 pg/m
OR Estimate [96% Cl]: All respiratory
conditions: Ischemic heart disease:
Lag 0: 0.82 [0.68-0.98]
Lag 0 (non-indigenous): 0.75
[0.61-0.93]
Lag 3 (indigenous): 1.71 [1.14-2.55]
Notes:
Fig 6: OR and 95% Cl for hospital
admissions for cardiovascular
conditions.
Summary: Negative associations in
overall study population and in non-
indigenous people. Positive
associations in Indigenous people at
Lag 1, Lag 2, and Lag 3.
Fig 6: OR and 95% Cl for hospital
admissions for ischaemic heart disease.
Summary: Negative associations in
overall study population and non-
indigenous people. Positive association
in indigenous people.
December 2009
E-92
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Koken et al. (2003,
0494661
Period of Study: Jul and Aug,
1993-1997
Location: Denver, Colorado
Outcome (ICD-9): Acute myocardial
infarction (410.00-410.92), pulmonary
heart disease (416.0-416.9), cardiac
dysrhythmias (427.0-427.9), congestive
heart failure (428.0)
Age Groups: 65+ yr
Study Design: Time series
N: 298 days
Statistical Analyses: GLM, GEE
Covariates: Maximum temp and dew
point temp
Season: NR
Dose-response Investigated: Yes
Statistical Package: SAS (PROC
GENMOD)
Lags Considered: 0-4 days
Pollutant: PM,0
Averaging Time: 24 h
Mean (min-max): 24.2 (7.0-51.6)
SD = 6.25
Monitoring Stations: 3
Copollutant (correlation):
N02:r = 0.56
S02:r = 0.36
03:r = 0.03
CO: r = 0.25
Other variables: Max temp: r = 0.38
Dew point temp: r = -0.24
PM Increment: 8.0 pg/m (IQR)
Percent Change Estimate [Cl]: No PM
data reported
Reference: Lanki et al, (2006, 0897881
Period of Study: 1992-2000
Location:
Augsburg, Barcelona, Helsinki, Rome,
and Stockholm
Outcome (ICD-9): Acute myocardial
infarction
(410
ICD-10:121,122)
Age Groups: 35+ yr, <75 yr, 75+ yr
Study Design: Time series
N: 26,854 hospitalizations
Statistical Analyses: GAM
Covariates: Temperature, barometric
pressure
Season: V\ferm (Apr-Sep) and cold
(Oct-Mar)
Dose-response Investigated: No
Statistical Package: R package mgcv
0.9-5
Lags Considered: 0-3 days
Pollutant: PM10
Averaging Time: 24 h
Median:
Augsburg: 43.5
Barcelona: 57.4
Helsinki: 21.0
Rome: 48.5
Stockholm: 12.5
Copollutant (correlation):
Augsburg
PNC: r = 0.53
CO: r = 0.56
N02:r = 0.64
03:r = 0.43
Barcelona: PNC: r = 0.38
CO: r = 0.44
N02:r = 0.48
03:r = 0.01
Helsinki: PNC: r = 0.45
CO: r = 0.21
N02:r = 0.40
03:r = 0.40
Rome: PNC: r = 0.32
CO: r = 0.41
N02:r = 0.29
03:r = 0.59
Stockholm: PNC: r = 0.06
CO: r = 0.41
N02:r = 0.29
03:r = 0.59
PM Increment: 10 pg/m
Pooled Rate Ratio [Cl]:
All 5 cities (35+ yr)
Same-day lag: 1.003 [0.995,1.011]
1-day lag: 1.001 [0.990,1.011]
2-day lag: 1.002 0.994,1.010'
3-day lag: 1.002 0.991,1.013
3 cities with hospital discharge register
(35+ yr)
Same-day lag: 1.003 [0.994,1.012]
1-day lag: 0.997 [0.988,1.006]
2-day lag: 1.003 ' ""
3-day lag: 1.003
0.995,1.012
0.986,1.020
V\ferm season (35+ yr)
Same-day lag: 1.006 [0.990,1.022]
1-day lag: 1.000 [0.985,1.016]
2-day lag: 1.005 [0.990,1.020]
3-day lag: 1.010 [0.995,1.025]
Cold season (35+ yr)
Same-day lag: 1.001 [0.991,1.012]
1-day lag: 0.998 ' ""
2-day lag: 1.001
0.987,1.009
0.991,1.012
3-day lag: 0.991 [0.981,1.002]
Age >75
Non-fatal
Same-day lag: 1.012 [0.995,1.029]
1-day lag: 1.000
0.983,1.017
0.982,1.017
2-day lag: 0.999
3-day lag: 1.001 [0.984,1.018]
Fatal
Same-day lag: 1.009 [0.985,1.034]
1-day lag: 0.998 [0.974,1.023]
2-day lag: 1.003
3-day lag: 1.018
0.978,1.028
0.975,1.063
Notes: Pooled rate ratios were also
provided for groups <75 yielding similar
results to the overall 3-city data.
December 2009
E-93
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Lee et al. (2003, 0955521
Period of Study:
Dec1997-Dec1999
Location: Seoul, Korea
Outcome (ICD-10): Angina pectoris
(I20), acute/subsequent myocardial
infarction (121-123), other acute
ischemic heart diseases (124)
Age Groups: All ages, 64+ yr
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 64.0 (31.8)
Monitoring Stations: 27
PM Increment: 40.4 pg/rri (IQR)
RR Estimate [Cl]: All yr
All ages: 0.99 [0.96,1.01]
64+yr: 1.05 [1.01,1.10]
Summer
Study Design: Time series
N: 822 days
Statistical Analyses: GAM with
LOESS, Pearson correlation
Covariates: Temperature, relative
humidity, day of the week
Season: Summer (Jun-Aug) and winter
Dose-response Investigated: Yes
Statistical Package: NR
Lags Considered: 0-6 days
Copollutant (correlation):
Allyr
S02:r = 0.59
N02:r = 0.74
03:r = 0.11
CO: r = 0.60
Temp: r = -0.07
Humidity: r = 0.02
Summer
S02:r = 0.61
N02:r = 0.73
03:r = 0.64
CO: r = 0.55
Temp: r = -0.01
Humidity: r = -0.11
All ages: 1.03 [0.97,1. 09]
64+ yr: 1.09 [1.00,1. 19]
Two-pollutant model
00(1 ppmlQI): 1.04 [0.98,1. 11]
03 (21.7 ppblQI): 1.07 [1.03,1. 11]
N02 (14.6 ppblQI): 1.09 [1.02,1. 16]
S02 (4.4 ppb): 0.98 [0.94,1. 03]
Reference: Lee et al. (2008,1920761
Period of Study: 1996-2005
Location: Taipei, Taiwan
Outcome: Congestive Heart Failure HA
(ICD9:428)
Age Groups: NR
Study Design: Case-crossover
N: 18593 events
Statistical Analyses: conditional
logistic regression
Covariates: Temperature, humidity
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: Lags 0-2
Pollutant: PM,0
Averaging Time: 24 h
Mean: 49.94
Min: 11.33
25th: 33.37
60th: 45.05
76th: 60.82
Max: 234.92
Monitoring Stations: 6
Copollutant: S02, CO, N02, 03
Co-pollutant Correlation
PM Increment: Interquartile Range
(27.45 pg/m3)
Odds Ratio (Lower Cl, Upper Cl):
VW Hypertension: 1.23 (1.15,1.32)
W/o Hypertension: 1.20 (1.15,1.25)
VW Diabetes: 1.20 (1.12,1.40)
W/o Diabetes: 1.21 (1.15,1.26)
W/Dysrhythmia: 1.17 (1.08,1.27)
W/o Dysrhythmia: 1.22 (1.17,1.27)
W/COPD:1.21 (1.07, 1.36)
W/o COPD: 1.21 (1.16, 1.25)
S02: 0.52
CO: 0.67
N02: 0.35
03: 0.39
Reference: Larrieu et al. (2007,
0930311
Period of Study: 1998-2003
1 oration1
LUUdllUII.
8 French urban area: Bordeaux, Le
Havre, Lille, Lyon, Marseille, Paris,
Rouen, and Toulouse
Outcome (ICD-10): Hospital
admissions for cardiovascular disease
(IOO-I99), cardiac disease (IOO-I52),
ischemic heart disease (I20-I25), and
stroke (cerebrovascular disease: 160-64
and transient ischemic attack:
G45-G46).
Age Groups: All, and 65 +
Study Design: Time series
N: Statistical Analyses: generalized
additive Poisson regression
Covariates: Temperature, holidays,
influenza epidemic periods, long-term
trend, season, day of the week,
Season: NR
Dose-response Investigated: No
Pollutant: PM10
Averaging Time: 24 h
Mean: Bordeaux: 21.0
Le Havre: 21. 7
Lille: 22.1
Lyon: 24.6
Marseille: 28.9
Paris: 23.1
Rouen: 21. 2
Toulouse: 21.8
Monitoring Stations: 32
Copollutant: NR
PM Increment: 10 pg/m3
ERR[96%CI]:
CVD: All ages: 0.7 [0.1, 1.2]
65+ yr: 1.1 [0.5, 1.7]
Cardiac diseases: All ages: 0.8 [0.2,
1.4]
65+ yr: 1.5 [0.7, 2.2]
Ischemic heart diseases: All ages: 1.9
[0.8, 3.0]
65+ yr: 2.9 [1.5, 4.3]
Strokes: All ages: 0.2 [-1.6, 1.9]
65+ yr: 0.8 [-0.9, 2.5]
Statistical Package: R 2.2.1
Lags Considered: 0 -to 1-day lag
(mean)
December 2009
E-94
-------
Reference
Reference: Le Tertre et al. (2002,
0237461
Period of Study: 1990-1997
Location:
Barcelona, Birmingham, London, Milan,
the Netherlands, Paris, Rome, and
Stockholm
Design & Methods
Outcome (ICD-9): Cardiac diseases
(390-429), ischemic heart disease (410-
413), and stroke (430-438)
Age Groups: <65 yr, 65+ yr
Study Design: Time series
N:NR
Statistical Analyses: GAM
Covariates: Long term trend, season,
days of the week, holidays, influenza
epidemics, temperature, and humidity
Season: NR
Dose-response Investigated: No
Statistical Package: S-Plus
Lags Considered: 0-3 days
Concentrations1
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD):
Barcelona: 55.7 (18. 4)
Birmingham: 24.8 (13.1)
London: 28.4 (12.3)
Milan: 51. 5 (22.7)
Netherlands: 39.5 (19.9)
Paris: 22.7 (10.8)
Rome: 52.5 (12.9)
Stockholm: 15.5 (7.2)
Monitoring Stations: 1-12
Copollutant: NR
Effect Estimates (95% Cl)
PM Increment: 10 pg/m3
Pooled Percent Increase [Cl]: Cardiac
(all ages)
Fixed: 0.5 [0.3,0.7]
Random: 0.5 [0.2,0.8]
Cardiac (over 65)
Fixed: 0.7 [0.4,1.0]
Random: 0.7 [0.4,1.0]
IHD (<65)
Fixed: 0.3 [-0.1, 0.6]
Random: 0.3 [-0.2,0.7]
IHD (over 65)
Fixed: 0.6 [0.3,0.8]; Random: 0.8
[0.3,1.2]
Stroke (over 65)
Fixed: 0.0 [-0.3,0.3]; Random: 0.0 [-
0.3,0.3]
Deaths: Cardiac: 0.5 [0.2,0.8]; Cardiac
(65+): 0.7 [0.4,1.0]
IHD (65+): 0.8 [0.3,1.2]
Notes: Estimated percentage increases
are also provided by city for cardiac
admissions and ischemic heart disease
in Fig 1-3.
Reference: Mann et al. (2002, 0367231 Outcome (ICD-9): Ischemic heart
disease (410-414), secondary
Period of Study: 1988-1995
Location: South Coast Air Basin,
California
congestive heart failure (sCHF) (428),
and secondary arrhythmia (sARR) (426,
427)
Age Groups: All, 40-59 yr, >60 yr
Study Design: Time series
N: 54,863 IHD admissions
Statistical Analyses: GAM
Covariates: Temperature, day of the
week, relative humidity
Season: NR
Dose-response Investigated: No
Statistical Package: S-Plus
Lags Considered: 0-5 days
Pollutant: PM,0
Averaging Time: 24 h
Mean (min-max): 43.7 (0.22-251)
SD = 27.7
Monitoring Stations: 20
Copollutant (correlation):
Region 1:
CO: r = 0.28
03:r = 0.20
N02:r = 0.36
Region 2:
CO: r = 0.15
03:r = 0.57
N02:r = 0.53
Region 3:
CO: r = 0.36
03:r = 0.30
N02:r = 0.46
Region 4:
CO: r = 0.27
03:r = 0.33
N02:r = 0.50
Region 5:
CO: r = 0.40
03:r = 0.43
N02:r = 0.53
Region 6:
CO: r = 0.33
03:r = 0.20
N02:r = 0.42
Region 7:
CO: r = 0.28
03:r = 0.48
N02:r = 0.60
PM Increment: 10 pg/m
Percent Change in IHD Admissions
[Cl]: Secondary ARR
Same-day lag: 0.59 [-0.71,1.91]
1-day lag: 0.46 [-0.86,1.80]
2-day lag:-0.04 [-1.37,1.31]
Secondary CHF
Same-day lag:-0.62 [-1.77,0.55]
1-day lag:-0.45 [-1.60,0.71]
2-day lag:-0.36 [-1.52,0.82]
No secondary diagnosis
Same-day lag:-0.25 [-1.23,0.75]
1-day lag: 0.04 [-0.97,1.06]
2-day lag: 0.18 [-0.82,1.20]
All IHD admissions: 0.19 [-0.576,0.955]
Ml admissions:-0.10 [-1.33,1.12]
Other acute IHD admissions: 0.36 [-
0.87,1.60]
December 2009
E-95
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Metzger et al. (2004,
0442221
Period of Study: Aug 1993-Aug 2000
Location: Atlanta Metropolitan area
(Georgia)
Outcome (ICD-9): Emergency visits for
ischemic heart disease (410-414),
cardiac dysrhythmias (427), cardiac
arrest (427.5), congestive heart failure
(428), peripheral vascular and
cerebrovascular disease (433-437, 440,
443.444, 451-453), atherosclerosis
(440), and stroke (436).
Age Groups: All
Study Design: Time series
N: 4,407,535 emergency department
visits
Statistical Analyses: Poisson
generalized linear modeling
Covariates: Day of the wk, hospital
entry and exit indicator variables,
federally observed holidays, temporal
trends, temperature, dew point
temperature
Season: All
Dose-response Investigated: No
Statistical Package: SAS
Lags Considered: 3-day ma, lags 0 -7
Pollutant: PM,0
Averaging Time: 24 h
Median (10% - 90% range): 26.3 (13.2,
44.7)
Monitoring Stations: NR
Copollutant (correlation):
03:r = 0.59
N02:r = 0.49
CO: r = 0.47
S02:r = 0.20
PM25:r = 0.84
UFP:"2r5=-0.13
PM2 5 water-sol
metals: r = 0.74
PM25sulfates:r = 0.74
PM25 acidity: r = 0.68
PM25OC:r = 0.69
PM25EC:r = 0.56
oxygenated hydrocarbon: r = 0.58
Other variables: Temperature: r = 0.58
Dew point: r = 0.44
PM Increment: 10 pg/m
(approximately 1 SD)
RR [96% Cl]: For 3-day ma: All CVD:
1.009 [0.998, 1.019]
Dysrhythmia: 1.008 [0.989,1.029]
Congestive heart failure: 0.992
[0.968-1.016]
Ischemic heart disease: 1.011
[0.992-1.030]
Peripheral vascular and
cerebrovascular disease: 1.020
[0.999-1.043]
Notes: Results for Lags 0-7 expressed
in figures
Fig 1: RR (95% Cl) for single-day lag
models for the association of ER visits
for CVD with daily ambient PMi0.
Summary: Statistically significant
association at Lag 0. Positive but not
statistically significant association at
Lag 1. Negative, statistically significant
association at Lag 7, and negative
associations at Lag 2 through Lag 6.
Reference: Middleton et al. (2008,
1567601
Period of Study: 1995-1998,
2000-2004
Location: Nicosia, Cyprus
Outcome: Hospital admissions for all
cardiovascular disease (ICD-10:
IOO-I52).
Age Groups: All, also stratified by age
(<15vs.. >15yr)
Study Design: Time series
Statistical Analyses: Generalized
additive Poisson models
Covariates: Seasonality, day of the
week, long- and short-term trend,
temperature, relative humidity
Dose-response Investigated: No
Statistical Package: STATASE 9.0, R
2.2.0
Lags Considered: Lag 0 -2 days
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD median 6% - 96% range):
Cold: 57.6 (52.5
50.8
20.0-103.0
5.0-1370.6)
Warm: 53.4 (50.5
30.7
32.0-77.6
18.4-933.5)
Monitoring Stations: 2
Copollutant: NR
PM Increment: 10 pg/m , and across
quartiles of increasing levels of PMi0
Percentage increase estimate [Cl]:
All age/sex groups (Lag 0): All
admissions: 0.85 (0.55,1.15)
Cardiovascular: 1.18 (-0.01, 2.37)
Nicosia residents (Lag 0):
Cardiovascular: 0.73 (-0.62, 2.09)
Males (Lag 0): All admissions: 0.96
(0.54, 1.39)
Cardiovascular: 1.27 (-0.15, 2.72)
Females (Lag 0): All admissions: 0.74
(0.31,1.18)
Cardiovascular: 0.99 (-1.11,3.14)
Aged <16 yr (Lag 0): All admissions:
0.47 (-0.13, 1.08)
Aged >16 yr (Lag 0): All admissions:
0.98(0.63,1.33)
December 2009
E-96
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Peel et al. (2007, 0904421
Period of Study: Jan 1993-Aug 2000
Location: Atlanta, GA
Outcome (ICD-9): Ischemic heart
disease (410-414), dysrhythmia (427),
congestive heart failure (428),
peripheral vascular and
cerebrovascular disease (433-437, 440,
443, 444, 451-453)
Age Groups: All
Study Design: Case-crossover
N: 4,407,535 ED visits
Statistical Analyses: Conditional
logistic regression
Covariates: Avg temp and dew point
temp
Season: NR
Dose-response Investigated: No
Statistical Package: SASv. 9.1
Lags Considered: 0-2 days
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): Daily levels: 27.9 (12.3)
Diff in case and control-day avg: 9.1
(7.5)
Monitoring Stations: 1
Copollutant: NR
PM Increment: 10 pg/m
OR Estimate [Cl]: All CVD: 1.010
[1.000,1.020]
IHD: 1.009 [0.991,1.027]
Dysrhythmia: 1.011 [0.991,1.031]
Peripheral/Cerebrovascular disease:
1.017 [0.996,1.039]
CHF: 1.001 [0.978,1.024]
With comorbid hypertension
IHD: 1.003 [0.973,1.034]
Dysrhythmia: 1.037 [0.988,1.089]
Peripheral/Cerebrovascular disease:
1.024 [0.990,1.060]
CHF: 1.041 [0.999,1.084]
No comorbid hypertension
IHD: 1.013 [0.991,1.036]
Dysrhythmia: 1.006 [0.985,1.028]
Peripheral/Cerebrovascular disease:
1.013 [0.987,1.040]
CHF: 0.982 [0.955,1.010]
With comorbid diabetes
IHD: 1.022 [0.979,1.067]
Dysrhythmia: 1.049 [0.968,1.137]
Peripheral/Cerebrovascular disease:
1.016 [0.965,1.069]
CHF: 1.029 [0.982,1.078]
No comorbid diabetes
IHD: 1.006 [0.987,1.026]
Dysrhythmia: 1.009 [0.989,1.029]
Peripheral/Cerebrovascular disease:
1.018 [0.995,1.042]
CHF: 0.992 [0.966,1.019]
Wth comorbid COPD
IHD: 0.981 [0.921,1.044]
Dysrhythmia: 0.984 [0.889,1.088]
Peripheral/Cerebrovascular disease:
1.086 [0.998,1.181]
CHF: 1.010 [0.954,1.069]
No comorbid COPD
IHD: 1.012 [0.993,1.031]
Dysrhythmia: 1.012 [0.992,1.032]
Peripheral/Cerebrovascular disease:
1.013 [0.991,1.035]
CHF: 0.999 [0.974,1.025]
Reference: Pope et al., (2006, 0912461
Period of Study: 1994-2004
Location: Wasatch Front area, Utah
Outcome: Myocardial infarction or
unstable angina (ICD codes not
reported)
Age Groups: All
Study Design: Case-crossover
N: 12,865 patients who underwent
coronary arteriography
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature and dew
point temperature
Season: NR
Dose-response Investigated: No
Statistical Package: NR
Lags Considered: 0- to 3-day lag, 2- to
4-day lagged ma
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD maximum):
Ogden: 28.5 (16.5
163)
SLC Hawthorne: 27.7 (17.4
162)
Provo/Orem, Lindom: 32.7 (21.1
240)
SLC AMC: 35.9 (20.4
161)
SLC North: 45.1(25.1
199)
Monitoring Stations: 5
Copollutant: NR
PM Increment: 10 pg/m
Percent increase in risk [96% Cl]:
Results summarized in Fig (see notes).
Notes: Fig 1: Percent increase in risk
(and 95% Cl) of acute coronary events
associated with 10 pg/m3 of PM10 for
different lag structures.
Summary of Fig 1: Positive, statistically
significant or marginally significant
associations between association seen
for Lag 0, Lag 1 and 2-, 3-, and 4-day
ma. Non-statistically significant
associations
December 2009
E-97
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Santos et al. (2008,
1920041
Period of Study: Jan 1998-Aug 1999
Location: Sao Paulo, Bazil
Outcome: Cardiac Arrhythmia ER Visits
(ICD 10:145-149)
Age Groups: 17+yr
Study Design: Time series
N:3251 ER visits
Statistical Analyses: Poisson
Covariates: Temperature, humidity,
seasonality
Dose-response Investigated? Yes
Statistical Package: S-Plus
Lags Considered: Lags 0-13
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 48.64 (20.34)
Min: 18.68
Max: 137.76
Monitoring Stations: 14
Copollutant: S02, CO, N02, 03
Co-pollutant Correlation:
S02: 0.675*
CO: 0.580*
N02: 0.781*
03: 0.438*
*p < 0.01
PM Increment: Interquartile Range
(22.2 pg/m3)
Percent Increase (Lower Cl, Upper
Cl):
PMio+N02,CO:-5.6 (-12.7, 2.1)
PM10+CO:-1.1(-7.0, 5.1)
PM10+N02:-2.4 (-9.4, 5.1)
Fig 1. PM10 effects, reported as percent
increase, on arrhythmia ER visits
caused by interquartile range increases,
lags 0-6.
Fig 2. Relative risks and 95% Cl for
arrhythmia ER visits according to the
division of air pollutant daily
concentrations in quintiles.
Reference: Tolbert et al. (2007,
0903161
Period of Study: 1993-2004
Location: Atlanta Metropolitan area,
Georgia
Outcome (ICD-9): Combined CVD
group, including: Ischemic heart
disease (410-414), cardiac
dysrhythmias (427), congestive heart
failure (428), and peripheral vascular
and cardiovascular disease (433-437,
440, 443-445, and 451-453).
Age Groups: All
Study Design: Time series
N: 10,234,490 ER visits (283,360 and
1,072,429 visits included in the CVD
and RD groups, respectively)
Statistical Analyses: Poisson
generalized linear models
Covariates: Long-term temporal trends,
season (for RD outcome), temperature,
dew point, days of week, federal
holidays, hospital entry and exit
Season: All
Dose-response Investigated: No
Statistical Package: SAS version 9.1
Lags Considered: 3-day ma (lag 0-2)
Pollutant: PM10
Averaging Time: 24 h
Mean (median
IQR, range, 10th-90th percentiles):
26.6 (24.8
17.5-33.8
0.5-98.4
12.3-42.8)
Monitoring Stations: NR
Copollutant (correlation):
03:r = 0.59
N02:r = 0.53
CO: r = 0.51
S02:r = 0.21
Coarse PM:r = 0.67
PM25:r = 0.84
PM25S04:r = 0.69
PM25EC:r = 0.61
PM25OC:r = 0.65
PM25TC:r = 0.67
P M2 5 water-sol metals: r = 0.73
OHC:r = 0.53
PM Increment: 16.30 pg/rri (IQR)
Risk ratio [96% Cl]: Single pollutant
models: CVD: 1.008 (0.997-1.020)
Reference: Tsai et al. (2003, 0801331
Period of Study: 1997-2000
Location: Kaohsiung, Taiwan
Outcome (ICD-9): Cerebrovascular
diseases (430-438), subarachnoid
hemorrhagic stroke (430), primary
intracerebral hemorrhage (431-432),
ischemic stroke (433-435), and others
(436-438)
Age Groups: All
Study Design: Case-crossover
N: 23,179 admissions
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature and humidity
Season: NR
Dose-response Investigated: No
Statistical Package: SAS
Lags Considered: Cumulative 0-2
days
Pollutant: PM,0
Averaging Time: 24 h
Mean (min-max): 78.82 (20.50-217.33)
Monitoring Stations: 6
Copollutant: NR
PM Increment: 66.33 pg/m3 (IQR)
OR Estimate [Cl]: Two-pollutant
model (all stroke admissions)
Primary intracerebral hemorrhage (PIH)
Adj for S02:1.55 [1.31,1.83]
Adj for N02:1.28 [1.01,1.61];
Adj for CO: 1.45 [1.20,1.74]
Adj for 03:1.56 [1.27,1.91]
Ischemic stroke (IS)
Adj for S02:1.46 [1.32,1.61
Adj for N02:1.16 [1.01,1.34
Adj for CO: 1.35 [1.21,1.51]
Adj for 03:1.51 [1.34,1.71]
Single-pollutant model
Temp >20°C
PIH: 1.54 [1.31,1.81]
IS: 1.46 [1.32,1.61]
Temp <20°C
PIH: 0.82 [0.48,1.40]
IS: 0.97 [0.65,1.44]
December 2009
E-98
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ulirsch et al. (2007,
0913321
Period of Study: Nov 1994-Mar 2000
Location:
Pocatello, Idaho and Chubbuck, Idaho
Outcome (ICD-9): CVD (390-429).
Age Groups: 65 +
Study Design: Time series
N: 39,347 admissions/visits
Statistical Analyses: Log-linear
generalized linear models
Covariates: Time, temperature, relative
humidity, influenza, day of the week
Season: All, and separate analyses
were performed for the all-age group for
cool months (Oct-Mar) vs.. warm
months (Apr-Sep).
Dose-response Investigated: No
Statistical Package: S-plus version 6.1
Lags Considered: 0- to 4-day lags,
and mean of days 0 -4
Pollutant: PM,0
Averaging Time: 24 h
Mean (range 10th - 90th percentiles):
24.2(3.0-183.0
10.5-40.7)
Monitoring Stations: 4
Copollutant (correlation):
N02:r = 0.47
Other variables: Correlation for PM10
between monitors: r = 0.42-0.87
PM Increment: 50 pg/m , and
24.3 pg/m3 (mean increase in PM10)
Mean percent of change (% change
in the mean number of daily
admissions and visits) [96% Cl]:
For 24.3 pg/m3 increase in PM10:
All-age RD/CVD: 3.7 [1.3, 6.3]
All-age CVD (Lag 0
All-age CVD (Lag 1
: -0.02 [-5.9, 6.3]
: 1.9 [-4.1,8.4]
All-age CVD (Lag 2):-3.1 [-9.1,3.4]
All-age CVD (Lag 3
: 0.5 [-5.6, 6.9]
:-1.7 [-4.3, 0.9]
All-age CVD (Lag 4 :
Lag 0-4 days:-0.5 [-8.0, 7.6]
For 60 ug/m3 increase in PM10
(single pollutant models, CIs not
given):
All-age respiratory disease: 8.4
All-age RD/CVD: 7.9
18-64 yrRD: 7.2
All-age CVD
All-age CVD
Lag3
Lag 4
:-3.6
All-age CVD (Lag 0-4):-1.1
Notes: Included urgent care visits as
well as emergency department visits
and hospital admissions.
Reference: Yang et al. (2007, 0928471 Outcome: Congestive Heart Failure HA
Period of Study: 1996-2005 ('CD 9' 428)
Location: Taipei, Taiwan
Age Groups: NR
Study Design: case-crossover
N: 24,240 events
Statistical Analyses: Poisson
Covariates: Temperature, humidity
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: lags 0-3
Pollutant: PM10
Averaging Time: 24 h
Mean: 49.47
Min: 14.42
26th: 33.08
60th: 44.71
76th: 60.10
Max: 234.91
Monitoring Stations: 6
Copollutant: NR
Co-pollutant Correlation: N/A
PM Increment: Interquartile Range
(27.02 pg/m3)
Odds Ratio (Lower Cl, Upper Cl):
Temp >20°C
PM,0:1.15 (1.10-1.21)*
PM10+S02:1.23 (1.17, 1.30)*
PM10+N02:1.03 (0.97, 1.10)
PMio+C02:1.09 (1.03, 1.15*
PM10+03:1.10 (1.04, 1.15)*
Temp <20°C
PM,0: 0.99 (0.93, 1.05)
PM10+S02: 0.96 (0.89, 1.03)
PM10+N02: 0.97 (0.90, 1.04)
PM,o+C02: 0.96 (0.90, 1.03)
PM10+03:1.00 (0.94, 1.05)
*p < 0.05
Reference: Yang et al. (2007, 0928471 Outcome: Congestive Heart Failure HA
Period of Study: 1996-2001
Location: Taipei, Taiwan
Age Groups: NR
Study Design: case-crossover
N:NR
Statistical Analyses: Poisson
Covariates: NR
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: Lags 0-3
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD):
Index days: 111.68 (38.32)
Comparison days: 55.43 (24.66)
Monitoring Stations: 7
Copollutant: NR
Co-pollutant Correlation: N/A
PM Increment: Index (>125 pg/m ) vs..
Comparison (<125 pg/m3)
Relative Risk (Lower Cl, Upper Cl),
0.915(0.805, 1.041), lag 0
1.114(0.993, 1.250), lag 1
0.983(0.873, 1.106), lag 2
0.974(0.870, 1.090), Iag3
December 2009
E-99
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Villeneuve et al. (2006,
0901911
Period of Study: Apr 1992-Mar 2002
Location: Edmonton, Canada
Outcome (ICD-9): Stroke (430-438),
including ischemic stroke (434-436),
hemorrhagic stroke (430,432), and
transient ischemic attacks (TIA) (435).
Age Groups: 65+ yr
Study Design: Case-crossover
N: 12,422 visits
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature and relative
humidity
Season: summer (Apr-Sep), winter
(Oct-Mar)
Dose-response Investigated: No
Statistical Package: SAS (PHREG)
Lags Considered: 0,1, and 3 days
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD):
All yr:
24.2(14.8)
Summer: 25.9 (16.4)
Winter: 22.6 (12.9)
Monitoring Stations: 3
Copollutant (correlation):
Allyr
S02:r = 0.19
N02:r = 0.34;
CO: r = 0.30
Os-mean: r = 0.07;
03-max:r = 0.22
PM25:r = 0.79
Summer
S02:r = 0.18
N02:r = 0.57;
CO: r = 0.38
Oj-mean: r = 0.20;
Os-max: r = 0.40
PM25:r = 0.85
Winter
S02:r = 0.27
N02:r = 0.48;
CO: r = 0.53
03-mean:r = -0.26;
03-max: r = -0.09
PM25:r = 0.70
PM Increment: pg/m (IQR)
Allyr: 16.0
Summer: 17.5
Wnter: 16.0
Adjusted OR Estimate [Cl]: Acute
ischemic stroke
Allyr
Same-day lag: 0.98 [0.94,1.03]
1-day lag: 1.00 [0.96,1.05]
3-day lag: 0.99 [.93,1.05]
summer
Same-day lag: 0.93 [0.87,1.00]
1-day lag: 1.01 [0.94,1.08
3-day lag: 0.96 [0.88,1.04
Wnter
Same-day lag: 1.04 [0.97,1.11]
1-day lag: 1.00 [0.94,1.06];
3-day lag: 1.05 [0.95,1.15]
Hemorrhagic stroke
Allyr
Same-day lag: 1.01 [0.90,1.12]
1-day lag: 1.03 [0.93,1.15
3-day lag: 1.13 [0.98,1.30
summer
Same-day lag: 1.02 [0.88,1.20]
1-day lag: 1.07 [0.91,1.26]
3-day lag: 1.20 [0.98,1.46]
Wnter
Same-day lag: 1.05 [0.90,1.22]
1-day lag: 1.04 [0.91,1.19]
3-day lag: 1.11 [0.90,1.37]
Transient cerebral ischemic attack
Allyr
Same-day lag: 0.96 [0.90,1.02]
1-day lag: 0.99 [0.94,1.05]
3-day lag: 0.94 [0.87,1.01]
summer
Same-day lag: 0.97 [0.89,1.09]
1-day lag: 0.99 [0.91,1.08]
3-day lag: 0.94 [0.84,1.04]
Wnter
Same-day lag: 0.95 [0.87,1.04]
1-day lag: 0.99 [0.92,1.07
3-day lag: 0.93 [0.83,1.05
Notes: Adjusted ORs are provided for
an IQR increase in the 3-day mean in
Fig 1-4 for single and two-pollutant
models.
December 2009
E-100
-------
Reference
Reference: von Klot et al. (2005,
0880701
Period of Study: 1992-2001
Location :
Augsburg, Germany
Barcelona, Spain
Helsinki, Finland
Rome, Italy
Stockholm, Sweden
Reference: Wellenius et al. (2005,
0874831
Period of Study: Jan 1987-Nov 1999
Location: Pittsburgh, Pennsylvania
Design & Methods
Outcome (ICD-9): Acute myocardial
infarction (410
ICD-10: 121-122), angina pectoris (411,
413
ICD-10: 120, 124), dysrhythmia (427
ICD-10: 146.0, 46.9, I47-I49, R00.1,
R00.8), heart failure (428
ICD-10: 150)
Age Groups: 35+ yr
Study Design: Cohort
N: 22,006 Ml survivors
Statistical Analyses: GAM, Spearman
correlation
Covariates: Temperature, dew point
temp, avg barometric pressure, relative
humidity
Season: NR
Dose-response Investigated: No
Statistical Package: R
Lags Considered: 0-3 days
Outcome (ICD-9): Congestive heart
failure (428.0-428.1)
Age Groups: 65+ yr
Study Design: Case-crossover
N: 55,019 patients
Statistical Analyses: Conditional
logistic regression, Pearson's pain/vise
correlation
Covariates: Temperature, barometric
pressure, dew point
Season: NR
Dose-response Investigated: No
Statistical Package: SAS
Lags Considered: 0-3 days
Concentrations1
Pollutant: PM,0
Averaging Time: 24 h
Mean (6th-96th percentile):
Augsburg: 44.7 (16.8-81. 4)
Barcelona: 52.2 (25.3-89.2)
Helsinki: 25.3 (9.5-57.6)
Rome: 51.1 (23.3-89.4)
Stockholm: 14.6 (6.4-30.0)
Monitoring Stations: NR
Copollutant (correlation):
Augsburg
PNC: r = 0.52
CO: r = 0.57;
N02:r = 0.64
03: r = -0.32
Barcelona
PNC: r = 0.29
CO: r = 0.39;
N02:r = 0.36
03:r = -0.14
Helsinki
PNC: r = 0.46
CO: r = 0.21;
N02:r = 0.40
03: r = 0.02
Rome
PNC: r = 0.33
CO: r = 0.31;
N02:r = 0.48
03:r = -0.22
Stockholm
PNC: r = 0.06
CO: r = 0.38;
N02:r = 0.29
03:r = 0.15
Pollutant: PM,0
Averaging Time: 24 h
Mean 5th-95th percentile):
31.06 8.89-70.49)
SD = 20.10
Monitoring Stations: 17
Copollutant (correlation):
CO: r = 0.57
N02:r = 0.64
03:r = 0.29
S02:r = 0.51
Effect Estimates (95% Cl)
PM Increment: 10 pg/m3
Pooled RR Estimate [Cl]:
All cardiac admissions: 1.021
[1.005,1.048]
Myocardial infarction: 1.026
[0.995,1.058]
Angina pectoris: 1.008 [0.986,1.032]
Notes: Rate ratios for 0-3 day lags are
provided in graphical form (Fig 1).
Same-day levels were significantly
associated with cardiac readmissions.
PM Increment: 24 pg/m3 (IQR)
Percent Increase [Cl]: Single-pollutant:
3.07 [1.59,4.57]
Adj. for CO: -1.10 [-3.02,0.86]
Adj. for N02: 0.52 [-1.46,2.53]
Adj. for 03: 2.80 [1.29,4.33]
Adj. for S02: 2. 18 [0.37,4.02]
Percent Increase (with 10 pg/m3
increment)
1.27 [0.66, 1.88]
December 2009
E-101
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Wellenius et al. (2005,
0886851
Period of Study:
Jan 1986-Nov 1999
Location:
Birmingham, Chicago, Cleveland,
Detroit, Minneapolis, New Haven,
Pittsburgh, Salt Lake City, Seattle
Outcome (ICD-NR): Ischemic stroke
and hemorrhagic stroke
Age Groups: 65+ yr
Study Design: Case-crossover (time-
stratified)
N: 115,503 hospital admissions
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature and humidity
Season: NR
Dose-response Investigated: No
Statistical Package: SAS (v.9) and R-
statistical package
Lags Considered: 0-2 days
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 32.69 (19.75)
Monitoring Stations: NR
(data obtained from the U.S. EPA)
Copollutant (correlation):
CO: r = 0.43
N02:r = 0.53
S02:r = 0.39
Other variables: Temp: r = 0.22
PM Increment: 22.96 pg/rri (IQR)
Percent Increase [Cl]: Ischemic (same-
day lag): 1.03 [0.04,2.04]
Hemorrhagic: -0.58 [-5.48,4.58]
Notes: Percent increase in rate for
ischemic and hemorrhagic stroke are
provided for each city in graphical form
(Fig A and B).
Reference: Wellenius et al.,(2006,
0887481
Period of Study:
Jan 1986-Nov 1999
Location:
Birmingham, Chicago, Cleveland,
Detroit, Minneapolis, New Haven,
Pittsburgh, Salt Lake City, Seattle
Outcome (ICD-9): Congestive heart
failure (428)
Age Groups: 65+ yr
Study Design: Case-crossover (time-
stratified)
N: 292,918 admissions
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature and
barometric pressure
Season: NR
Dose-response Investigated: No
Statistical Package: SAS (v.9) and R-
statistical package
Lags Considered: 0-3 days
Pollutant: PM,0
Averaging Time: 24 h
Median: Overall: 28.3
Birmingham: 33.0
Chicago: 31.5
Cleveland: 34.5
Detroit: 29.5
Minneapolis: 24.0
New Haven: 22.
Seattle: 25.8
Monitoring Stations: NR
(data obtained from the U.S. EPA)
Copollutant: NR
PM Increment: 10 pg/m
Percent Increase [Cl]: Same-day lag:
0.72 [0.35,1.10]
p-value = 0.0002
Notes: City-specific percent increases
are graphed in Fig 1 for same-day lag
showing a significant association in
Chicago, Detroit, Seattle, and the
summary values.
Percent increase in admission rate s
are provided for lag 0-3 days in Fig 2
where same-day lag showed a
significant association.
Reference: Yang et al. (2004, 0943761
Period of Study: 1997-2000
Location: Kaohsiung, Taiwan
Outcome (ICD-9): Cardiovascular
diseases (410-429)
Age Groups: All
Study Design: Case-crossover
N: 29,661 admissions
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature and humidity
Season: NR
Dose-response Investigated: No
Statistical Package: SAS
Lags Considered: Cumulative 0-2
days
Pollutant: PM,0
Averaging Time: 24 h
Median (min-max): 78.82 (20.50-
217.33)
Monitoring Stations: 6
Copollutant: NR
PM Increment: 66.33 pg/m3 (IQR)
OR Estimate [Cl]: Temp >25°C: 1.439
[1.316,1.573]
Temp <25°C: 1.568 [1.433,1.715]
Adj for S02
Temp >25°C: 1.460 [1.333,1.599]
Temp <25°C: 1.543 [1.404,1.696]
Adj for N02
Temp >25°C: 1.306 [1.154,1.478]
Temp <25°C: 0.912 [0.809,1.028]
Adj for CO
Temp >25°C: 1.260 [1.144,1.388]
Temp <25°C: 1.259 [1.128,1.406]
Adj for 03
Temp >25°C: 1.086 [0.967,1.220]
Temp <25°C: 1.703 [1.541,1.883]
December 2009
E-102
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Yang et al. (2008,1571601 Outcome (ICD-9): Congestive heart
Period of Study: 1996-2004
Age Groups: All
Location: Taipei, Taiwan
Study Design: Case-crossover
N: 24,240 CHF hospital admissions
Statistical Analyses: Conditional
logistic regression
Covariates: temperature, humidity
Season: All
Dose-response Investigated: No
Statistical Package: SAS
Lags Considered: Cumulative lag 0-2
days
Pollutant: PM,0
Averaging Time: 24 h
Mean (median, range, IQR):
49.47(4471,14.42-234.91,
33.08-44.71)
Monitoring Stations: 6
Copollutant: NR
PM Increment: 27.02 pg/rri (IQR)
OR [95% Cl]:
Single pollutant models: >20 °C: 1.15
[1.10-1.21]
<20°C: 0.99 [0.93-1.05]
Adjusted for S02:> 20 °C: 1.23
[1.17-1.30]
<20°C: 0.96 [0.89-1.03]
Adjusted for N02:> 20 °C: 1.03
[0.97-1.10]
<20°C: 0.97 [0.90-1.04]
Adjusted for CO: > 20 °C: 1.09
[1.03-1.15]
<20°C: 0.96 [0.90-1.03]
Adjusted for 03:> 20 °C: 1.10
[1.04-1.15]
<20°C: 1.00 [0.94-1.05]
Reference: Zanobetti and Schwartz
(2002, 0348211
Period of Study: 1988-1994
Location:
Cook county (Chicago), Illinois
Wayne county (Detroit), Michigan
Allegheny county (Pittsburgh),
Pennsylvania
and King county (Seattle), Washington
Outcome (ICD-9): Cardiovascular Pollutant: PM10
disease (390-429) with/without diabetes
(250) Averaging Time: 24 h
Age Groups: 65-74 and 75+ yr with Median (25-75th percentile):
diabetes, 65-74 and 75+ yr without Chicago: 33 (23-46)
dlabetes Detroit: 32 (2 1-49)
Study Design: Time series pittsburgh. 3Q (19.4y)
N: NR Seattle: 27 (18-39)
Statistical Analyses: GAM, meta- Monjtorjng statjons. NR (ob(ained
regression from USEPAAerometric Information
Covariates: Temperature, prior day's Retrieval System)
temperature, relative humidity, r«n«ii,,*an*. MP
barometric pressure, day of the week c°P°llutant- NR
Season: NR
PM Increment: 10 pg/m3
Perc
All 4
<75
75+
<75
75+
ent Change [Cl]:
cities
w/ diabetes): 1.6 [1.2,2.0]
w/ diabetes): 2.0 [1.6,2.4]
w/o diabetes): 0.9 [0.6,1.1]
w/o diabetes): 1.3 [1.0,1.5]
ChicPnn
<75
75+
<75
75+
Platr
uetr
<75
75+
<75
w/ diabetes): 1.9 [1.1,2.7]
w/ diabetes): 2.0 [1.1,3.0]
w/o diabetes): 0.7 [0.2,1.2]
w/o diabetes): 1.2 [0.8,1.7]
w/ diabetes): 1.3 [0.5,2.2]
w/ diabetes): 2.1 [1.0,3.1]
w/o diabetes): 1.2 [0.7,1.7]
75+ (w/o diabetes): 1.2 [0.7,1.6]
Dose-response Investigated: No
Pittsburgh
<75(w/ diabetes): 1.8 [0.9,2.7]
75+
<75
w/ diabetes): 0.9 [-0.2,2.0]
w/o diabetes): 0.6 [0.1, 1.2]
75+ (w/o diabetes): 1.6 [1.2,2.1]
Seattle
<75(w/ diabetes): 1.9 [0.1,3.7]
75+
<75
w/ diabetes): 2.7 [0.7,4.8]
w/o diabetes): 0.8 [0.0,1.6]
75+ (w/o diabetes): 0.9 [0.2,1.6]
Notes: Overall percent increases were
also provided for each city, yielding
similar results.
December 2009
E-103
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Zanobetti and Schwartz
(2005, 0880691
Period of Study: 1985-1999
Location:
21 U.S. cities (Birmingham, Alabama
Boulder, Colorado
Canton, Ohio
Chicago, Illinois
Cincinnati, Ohio
Cleveland, Ohio
Colorado Springs, Colorado
Detroit, Michigan
Honolulu, Hawaii
Houston, Texas
Minneapolis-St. Paul, Minnesota
Nashville, Tennessee
New Haven, Connecticut
Pittsburgh, Pennsylvania
Provo-Orem, Utah
Salt Lake City, Utah
Seattle, Washington
Steubenville, Ohio
Youngstown, Ohio)
Outcome (ICD-9): Myocardial infarction
(410)
Age Groups: >65 yr
Study Design: Case-crossover
N: 302,453 admissions
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature
Season: NR
Dose-response Investigated: Yes
Statistical Package: SAS (PROC
PHREG)
Lags Considered: 0-2 days
Pollutant: PM,0
Averaging Time: 24 h
Median: Ranged from 15.5-34.1Avg
across all cities = 27
PM Increment: 10 pg/m
Percent Increase [Cl]: Ml only: 0.65
[0.3,1]
Previous COPD admission: 1.3 [-
0128]
Monitoring Stations: 1+ (data obtained
from USEPAsAerometric Information Secondary pneumonia diagnosis: 1.4 [-
Retrieval System)
Copollutant: NR
0.8,3.6]
Notes: Fig 1 presents percent change
in Ml per lag day, showing same-day
lag to be significant. Fig 2 shows
percent change with/without other co-
morbidities.
1AII units expressed in ug/m3 unless otherwise specified.
Table E-6. Short-term exposure-cardiovascular-ED/HA - PMi0.2.6-
Reference
Reference: Halonen et al. (2009,
1803791
Period of Study: 1998-2004
Location: Helsinki, Finland
Design & Methods
Outcome: Cardiovascular
Hospitalizations & Mortality (ICD 10:
100-99)
Age Groups: 65+ yr
Study Design: Time series
N:NR
Statistical Analyses: Poisson, GAM
Covariates: Temperature, humidity,
influenza epidemics, high pollen
episodes, holidays
Dose-response Investigated? No
Statistical Package: R
Lags Considered: lags 0-3 days; 5-day
(0-4) mean
Concentrations1
Pollutant: PM10.2.5
Averaging Time: Daily
Mean (SD): NR
Min:0.0
26th percentile: 4.9
60th percentile: 7.5
76th percentile: 12.1
Max: 101. 4
Monitoring Stations: NR
Copollutant:
PMO.03, PM0.03-0.1,PM<0.1,
PM<0. 10.29, PM25, CO, N02
Co-pollutant Correlation
PMO.03: 0.14
PMO.03-0. 1:0.28
PM<0. 1:0.24
PM<0.10.29:0.20
PM25: 0.25
Effect Estimates (95% Cl)
PM Increment: Interquartile Range
Percent Change (Lower Cl, Upper
Cl):
All Cardiovascular Morality
Lag 0: -0.01 (-1.52, 1.53)
Lag 1: -0.26 (-1.69, 1.18)
Lag 2: -0.61 (-2.03, 0.83)
Lag 3: -0.57 (-1.98, 0.85)
5-day mean: -0.70 (-2.56, 1.20)
Coronary Heart Disease HA
Lag 0:1. 12 (-0.28, 2.55)
Lag 1: -0.38 (-1.68, 0.94)
Lag 2: 0.01 (-1.33, 1.37)
Lag 3: -0.53 (-1.82, 0.78)
5-day mean: 0.23 (-0.29, 0.75)
Stroke HA
Lag 0: -1.33 (-3.26, 0.63)
Lag 1: -1.90 (-3.82, 0.07) J
Lag 2: -1.09 (-3.04, 0.89)
Lag 3: -0.51 (-2.40, 1.43)
5-day mean: -2.21 (-4.75, 0.39)
Arrhythmia HA
Lag 0:0.57 (-1.33, 2.49)
Lag 1: -0.65 (-2.55, 1.29)
Lag 2: 0.02 (-1.93, 2.00)
Lag 3: -1.34 (-3.26, 0.62)
5-day mean: -1.11 (-3.68, 1.53)
*p < 0.05, Jp< 0.10
December 2009
E-104
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Host et al. (2008,
155852)(Host et al, 2008,1558521
Period of Study: 2000-2003
Location: Six French cities: Le Havre,
Lille, Marseille, Paris, Rouen, and
Toulouse
Outcome (ICD-10): Daily
hospitalizations for all cardiovascular
(IOO-I99), cardiac (IOO-I52), and
ischemic heart diseases (I20-I25).
Age Groups: For cardiovascular
diseases: All ages, and restricted to 2
65 yr
Study Design: Time series
N: NR (Total population of cities:
approximately 10 million)
Statistical Analyses: Poisson
regression
Covariates: Seasons, days of the
week, holidays, influenza epidemics,
pollen counts, temperature, and
temporal trends
Dose-response Investigated: No
Statistical Package: MGCV package in
R software (R 2.1.1)
Lags Considered: Avg of 0-1 days
Pollutant: PM10.2.5
Averaging Time: 24 h
Mean ug/m3 (6th -96th percentile):
Le Havre: 7.3 (2.5-14.0)
Lille: 7.9 (2.2-137)
Marseille: 11.0(4.5-21.0)
Paris: 8.3 (3.2-15.9)
Rouen: 7.0 (3.0-12.5)
Toulouse: 7.7 (3.0-15.0)
Monitoring Stations:
13 total:
1 in Toulouse
4 in Paris
2 each in other cities
Copollutant (correlation): PM25:
Overall: r>0.6
Ranged between r = 0.28 and r = 0.73
across the six cities.
PM Increment: 10 pg/m , and an
18.8 pg/m3 increase (corresponding to
an increase in pollutant levels between
the lowest of the 5th percentiles and the
highest of the 95th percentiles of the
cities' distributions)
ERR (excess relative risk) Estimate [Cl]:
For all cardiovascular diseases
(10 pg/m increase): All ages: 0.5% [-
1.2,2.3]
> 65 yr: 1.0% [-1.0, 3.0]
For all cardiovascular diseases
(18 pg/m3 increase): All ages: 1.0% [-
2.3, 4.3]
> 65 yr: 1.9% [-2.0, 5.9]
For cardiac diseases (10 pg/m3
increase): All ages: 0.1% [-1.9, 2.1]
> 65 yr: 1.6% [-0.8, 4.1]
For cardiac diseases (18.8 pg/m3
increase): All ages: 0.1% [-3.6, 4.0]
> 65 yr: 3.1% [-1.5, 7.9]
For ischemic heart diseases (10 pg/m3
increase): All ages: 2.8% [-0.8, 6.6]
> 65 yr: 6.4% [1.6,11.4]
For ischemic heart diseases (18 pg/m3
increase): All ages: 5.4% [-1.5,12.8]
> 65 yr: 12.4 [3.1, 22.6]
Reference: Metzger et al. (2004,
0442221
Period of Study: Aug 1998-Aug 2000
Location: Atlanta Metropolitan area
(Georgia)
Outcome (ICD-9): Emergency visits for
ischemic heart disease (410-414),
cardiac dysrhythmias (427), cardiac
arrest (427.5), congestive heart failure
(428), peripheral vascular and
cerebrovascular disease (433-437, 440,
443-444, 451-453), atherosclerosis
(440), and stroke (436).
Age Groups: All
Study Design: Time series
N: 4,407,535 emergency department
visits between 1993-2000 (data not
reported for 1998-2000)
Statistical Analyses: Poisson
generalized linear modeling
Covariates: Day of the wk, hospital
entry and exit indicator variables,
federally observed holidays, temporal
trends, temperature, dew point
temperature
Season: All
Dose-response Investigated: No
Statistical Package: SAS
Lags Considered: 3-day ma; lags 0 -7
Pollutant: PM,0 -2.5
Averaging Time: 24 h
Median ug/m3 (10% - 90% range):
9.1 (4.4, 16.2)
Monitoring Stations: 1
Copollutant (correlation):
PM10:r = 0.59
03:r = 0.35
N02:r = 0.46
CO: r = 0.32
S02:r = 0.21
PM25:r = 0.43
UFP:r = 0.13
PM25water
soluble metals: r = 0.47
PM25sulfates:r = 0.26
PM25 acidity: r = 0.23
PM25OC:r = 0.51
PM25EC:r = 0.48
PM25 oxygenated hydrocarbon: r = 0.31
Other variables: Temperature: r = 0.20
Dew point: r = 0.00
PM Increment: 5 pg/m (approximately
1SD)
RR [95% Cl]: For 3 day ma: All CVD:
1.012 [0.985, 1.040]
Dysrhythmia: 1.021 [0.974,1.070]
Congestive heart failure: 1.020
[0.964-1.079]
Ischemic heart disease: 0.994
[0.946-1.045]
Peripheral vascular and
cerebrovascular disease: 1.022
[0.972-1.074]]
Results for Lags 0-7 expressed in
figures (see notes).
Notes: Fig 1: RR (95% Cl) for single-
day lag models for the association of
ER visits for CVD with daily ambient
PM,0-2.5.
Summary of Fig 1 results: Positive
association at Lag 0.
December 2009
E-105
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Peng et al. (2008,1568501
Period of Study: Jan 1999-Dec 2005
Location: 108 U.S. counties in the
following states: Alabama, Arizona,
California, Colorado, Connecticut,
District of Columbia, Florida, Georgia,
Idaho, Illinois, Indiana, Kentucky,
Louisiana, Maine, Maryland,
Massachusetts, Michigan, Minnesota,
Missouri, Nevada, New Hampshire,
New Jersey, New Mexico, New York,
North Carolina, Ohio, Oklahoma,
Oregon, Pennsylvania, Rhode Island,
South Carolina, Tennessee, Texas,
Utah, Virginia, Washington, West
Virginia, Wisconsin
Outcome (ICD-9): Emergency
hospitalizations for: Cardiovascular
disease, including heart failure (428),
heart rhythm disturbances (426-427),
cerebrovascular events (430-438),
ischemic heart disease (410-414, 429),
and peripheral vascular disease
(440-448).
Age Groups: 65 + yr, 65-74, 75+
Study Design: Time series
N: approximately 12 million Medicare
enrollees (3.7 million CVD and 1.4
million RD admissions)
Statistical Analyses: Two-stage
Bayesian hierarchical models:
Overdispersed Poisson models for
county-specific data. Bayesian
hierarchical models to obtain national
avg estimate
Covariates: Day of the wk, age-specific
intercept, temperature, dew point
temperature, calendar time, indicator for
age of 75 yr or older. Some models
were adjusted for PM25.
Dose-response Investigated: No
Statistical Package: R version 2.6.2
Lags Considered: 0-2 days
Pollutant: PM10.2.5
Averaging Time: 24 h
Mean ug/m3 (IQR):
All counties assessed: 9.8 (6.9-15.0)
Counties in Eastern U.S.: 9.1 (6.6-13.1)
Counties in Western U.S.: 15.4
(10.3-21.8)
Monitoring Stations: At least 1 pair of
co-located monitors (physically located
in the same place) for PM10 and PM25
per county
Copollutant (correlation):
PM25:r = 0.12
PM10:r = 0.75
Other variables: Median within-county
correlations between monitors: r = 0.60
PM Increment: 10 pg/m
Percentage change [95% Cl]: CVD: Lag
0 (unadjusted for PM25): 0.36 [0.05,
0.68]
Lag 0 (adjusted for PM25): 0.25 [-0.11,
0.60]
Notes: Effect estimates for PMi0.25 (0-2
day lags) are showing in Fig 2-5.
Fig 2: Percentage change in emergency
hospital admissions for CVD per
10 pg/m3 increase in PM (single
pollutant model and model adjusted for
PM2 5 concentration)
Fig 4: Percentage change in emergency
hospital admissions rate for CVD and
RD per a 10 pg/m increase in PMi0.2.5
(0-2 day lags, Eastern vs.. Western
USA)
Fig 5: County-specific log relative risks
of emergency hospital admissions for
CVD per 10 pg/m increase in PMi0.25
at Lag 0 (unadjusted for PM25 and
plotted vs.. percentage of urbanicity)
No significant associations between
PMio.25and cause-specific
cardiovascular disease.
Reference: Tolbert et al. (2007,
0903161
Period of Study: Aug 1998-Dec 2004
Location: Atlanta Metropolitan area,
Georgia
Outcome (ICD-9): Combined CVD
group, including: Ischemic heart
disease (410-414), cardiac
dysrhythmias (427), congestive heart
failure (428), and peripheral vascular
and cardiovascular disease (433-437,
440, 443-445, and 451-453)
Age Groups: All
Study Design: Time series
N: NR for 1998-2004. For 1993-2004:
10,234,490 ER visits (283,360 visits).
Statistical Analyses: Poisson
generalized linear models
Covariates: Long-term temporal trends,
temperature, dew point, days of week,
federal holidays, hospital entry and exit
Season: All
Dose-response Investigated: No
Statistical Package: SAS version 9.1
Lags Considered: 3-day ma (lag 0-2)
Pollutant: PMi0.25
Averaging Time: 24 h
Mean (ug/m3) (median IQR, range,
10th-90th percentiles):
9.0(8.2
5.6-11.5
0.5-50.3
3.6-15.1)
Monitoring Stations: 1
Copollutant (correlation):
PMi0:r = 0.67
03:r = 0.36
N02:r = 0.48
CO:r = 0.38S02:r = 0.16
PM25:r = 0.47
PM25S04:r = 0.32
PM25EC:r = 0.49
PM25OC:r = 0.49
PM25TC:r = 0.51
PM2 5 water-sol metals: r = 0.50
OHC:r = 0.41
PM Increment: 5.89 pg/m3 (IQR)
Risk ratio [95% Cl]: CVD: 1.004
(0.990-1.019)
1AII units expressed in ug/m3 unless otherwise specified.
December 2009
E-106
-------
Table E-7. Short-term exposure - cardiovascular: ED/HA PM2.6 (including PM components/sources)
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Andersen et al. (2008,
1896511
Period of Study: May 2001-Dec 2004
Location: Copenhagen, Denmark
Outcome (ICD-10): CVD, including
angina pectoris (I20), myocardial
infarction (121-22), other acute ischemic
heart diseases (I24), chronic ischaemic
heart disease (I25), pulmonary
embolism (I26), cardiac arrest (I46),
cardiac arrhythmias (148-48), and heart
failure (ISO). RD, including chronic
bronchitis (J41-42), emphysema (J43),
other chronic obstructive pulmonary
disease (J44), asthma (J45), and status
asthmaticus (J46). Pediatric hospital
admissions for asthma (J45) and status
asthmaticus (J46).
Age Groups: > 65 yr (CVD and RD),
5-18 yr (asthma)
Study Design: Time series
N:NR
Statistical Analyses: Poisson GAM
Covariates: Temperature, dew-point
temperature, long-term trend,
seasonality, influenza, day of the week,
public holidays, school holidays (only
for 5-18 yr olds), pollen (only for
pediatric asthma outcome)
Season: NR
Dose-response Investigated: No
Statistical Package: R statistical
software (gam procedure, mgcv
package)
Lags Considered: Lag 0-5 days, 4-day
pollutant avg (lag 0-3) for CVD, 5-day
avg (lag 0-4) for RD, and a 6-day avg
(lag 0-5) for asthma.
Pollutant: PM25
Averaging Time: 24 h
Mean ugm3 (SD): 10(5)
Median: 9
IQR:7-12
99th percentile): 28
Monitoring Stations: 1
Copollutant (correlation):
NCtot:r = 0.40
NC100:r = 0.29
NCa12:r = 0.07
Nca23:r = -0.25
NCa57:r = 0.51
NCa212:r = 0.82
PM10:r = 0.80
CO: r = 0.46
N02:r = 0.42
N0x:r = 0.40
curbside:r = 0.28
03:r = -0.20
Other variables:
Temperature: r = -0.01
Relative humidity: r = 0.21
PM Increment: 5 pg/m (IQR)
Relative risk (RR) Estimate [Cl]: CVD
hospital admissions (4-day avg, lag 0 -
3), age 65+: One-pollutant model: 1.03
[1.01-1.06]
Adj for NCtot: 1.03 [1.01-1.06]
RD hospital admissions (5-day avg, lag
0-4), age 65+:
One-pollutant model: 1.00 [0.95-1.00]
Adj for NCtot: 1.00 [0.95-1.06]
Asthma hospital admissions (6-day avg
lag 0-5), age 5-18:
One-pollutant model: 1.15 [1.00-1.32]
Adj for NCtot: 1.13 [0.98-1.32]
Estimates for individual day lags
reported only in Fig form (see notes):
Notes: Fig 2: Relative risks and 95%
confidence intervals per IQR in single
day concentration (0-5 day lag).
Summary: CVD: Marginally significant
association at Lag 0. RD: No
statistically or marginally significant
associations. Positive associations at
Lag 4-5.Asthma: Wide confidence
intervals make interpretation difficult.
Positive associations at Lag 1, 2, 3.
December 2009
E-107
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ballester et al. (2006,
0887461
Period of Study: 1995-1999
Location: 6 Spanish cities: Barcelona,
Bilbao, Pamplona, Valencia, Vigo,
Zaragoza
Outcome (ICD-9): The number of daily
emergency admissions with primary
diagnosis for all cardiovascular disease
(390-459) and heart diseases (410-414,
427, 428)
Age Groups: All ages
Study Design: Time series
N:NR
Statistical Analyses: Poisson GAMs
Covariates: Daily temperature,
barometric pressure, and relative
humidity
Daily influenza incidence, day of the
week, holidays, unusual events (ex.
medical strikes), seasonal variation,
trend of the series
Season: NR
Dose-response Investigated: No
Statistical Package: S-Plus GAM
function
Lags Considered: 0-3 days, 0- to
1-dayavg
Pollutant: Black smoke (BS)
Averaging Time: 24 h
Mean ug/m3 (10-90th percentile):
Overall mean NR.
City specific means
Barcelona: 35.0 (19.4, 53.0)
Bilbao: 18.5 (8.8, 31.0)
Pamplona: 7.4 (2.3,13.0)
Valencia: 40.3 (20.3, 66.4)
Vigo: 79.4 (43.9, 122.3)
Zaragoza: 40.4 (23.8, 61.3)
Monitoring Stations: NR
(at least 3 stations per city)
Copollutant (correlation): Summary of
the correlation coefficients between
each pair of pollutants within cities:
PM,0:r = 0.48
TSP:fromr = 0.16tor = 0.69
(median r = 0.43)
N02: from r = 0.23tor = 0.69
(median r = 0.48)
S02: from r = 0.09tor = 0.59
(median r = 0.24)
CO: from r = 0.62 to r = 0.69
(median r = 0.69)
03: from r = -0.43 to r = -0.06
(median r =-0.16)
PM Increment: 10 pg/m
Relative risk [Cl]: Relative risks are
expressed only in the form of figures
(see notes).
Percentage change in risk [Cl]: All
cardiovascular diseases (avg of lags 0 -
1)0.24% [-0.18, 0.67]
Heart disease (avg of lags 0 -1) 0.71%
[0.13,1.29]
Notes: Relative risks for the single
pollutant models are expressed in Fig 2.
Fig 2: Time sequence of the combined
association between BS and hospital
admissions for all CVD (A) and heart
disease (B). Summary: Significant,
positive association of TSP with both
overall CVD and heart disease
hospitalizations at Lag 0.
Relative risks for 2 pollutant models are
expressed in Fig 3: Combined
estimates of the association between
hospital admissions for heart diseases
and air pollutants (avg of lags 0-1
Adjusted for CO, N02, 03, or S02).
Summary: Significant, positive
association remains after adjusting for
N02, 03, and S02. Association remains
positive but becomes marginally
significant after adjusting for CO.
Reference: Ballester et al. (2006,
0887461
Period of Study: 1993-1999
Location: 7 Spanish cities: Barcelona,
Bilbao, Cartagena, Castellon, Gijon,
Oviedo, Valencia
Outcome (ICD-9): The number of daily Pollutant: TSP
emergency admissions with primary
diagnosis for all cardiovascular disease Averaging Time: 24 h
(390-459) and heart diseases (410-414,
427, 428)
Mean ug/m3 (10-90th percentile):
Overall mean NR.
Age Groups: All ages
Study Design: Time series
N:NR
Statistical Analyses: Poisson GAMs
Covariates: Daily temperature,
barometric pressure, and relative
humidity
Daily influenza incidence, day of the
week, holidays, unusual events (ex.
medical strikes), seasonal variation,
trend of the series
Season: NR
Dose-response Investigated: No
Statistical Package: S-Plus GAM
function
Lags Considered: 0-3 days, 0- to
1-dayavg
City specific means
Barcelona: 51.8 (29.4, 78.8)
Bilbao: 58.3 (30.3, 92.3)
Cartagena: 54.9 (32.5, 79.9)
Castellon: 60.4 (32.0, 92.1)
Gijon: 77.4 (47.4, 118.3)
Oviedo: 76.0 (48.3,111.8)
Valencia: 61.0(44.1, 80.7)
Monitoring Stations: NR (at least
three stations per city)
Copollutant (correlation): Summary of
the correlation coefficients between
each pair of pollutants within cities:
BS: from r = 0.16tor = 0.69
(median r = 0.43)
PM,0:NA
N02: from r =-0.13tor = 0.65
(median r = 0.48)
S02: from r = 0.06tor = 0.69
(median r = 0.31)
CO: from r = 0.06 to r = 0.59
(median r = 0.47)
03: from r =-0.27 tor = 0.07
(median r =-0.03)
PM Increment: 10 pg/m
Relative risk [Cl]: Relative risks are
expressed only in the form of figures
(see notes).
Percentage change in risk [Cl]: All
cardiovascular diseases: 0.07% [-0.23,
0.36]
Heart disease 0.45% [0.04, 0.86]
Notes: Relative risks for the single
pollutant models are expressed in Fig 2.
Fig 2: Time sequence of the combined
association between TSP and hospital
admissions for all CVD (A) and heart
disease (B).
Summary of results: Positive, marginally
significant association of TSP with
overall CVD at Lag 0. Positive,
statistically significant relation between
TSP and heart disease hospitalizations
at Lag 0.
Relative risks for 2 pollutant models are
expressed in Fig 3:
Fig 3: Combined estimates of the
association between hospital
admissions for heart diseases and air
pollutants (avg of lags 0-1 adjusted for
CO, N02, 03, or S02).
Summary of results: Small positive
significant or marginally significant
associations between TSP and general
CVD and heart disease hospitalizations
remain constant after adjustment for
CO, N02, 03, or S02.
December 2009
E-108
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Bell et al. (2008, 0912681
Period of Study: 1995-2002
Location: Taipei, Taiwan
Outcome (ICD-9): Hospital admissions
for ischemic heart disease (410, 411,
414), cerebrovascular disease
(430-437), asthma (493), and
pneumonia (486).
Age Groups: All
Study Design: Time series
N: 6,909 hospital admissions for
ischaemic heart diseases, 11,466 for
cerebrovascular disease, 19,966 for
pneumonia, and 10,231 for asthma
Statistical Analyses: Poisson
regression
Covariates: Day of the week, time,
apparent temperature, long-term trends,
seasonality
Season: All
Dose-response Investigated: No
Statistical Package: NR
Lags Considered: lags 0-3 days, mean
of lags 0-3
Pollutant: PM25
Averaging Time: 24 h
Mean ug/m3 (range IQR):
31.6(0.50-355.020.2)
Monitoring Stations: 2
Copollutant (correlation): NR
PM Increment: 20 pg/m (near IQR)
Percentage increase estimate [95% Cl]:
Ischemic heart disease: LO: 3.48 (-0.39,
7.51)
L1: 3.55 (-0.30, 7.56)
L2: 3.32 (-0.50, 7.29)
L3: 2.80 (-1.04, 6.79)
LOS: 8.38 (2.28, 14.84)
Cerebrovascular disease: LO: -2.22 (-
50.2, 0.67)
L1:-1.30 (-4.08, 1.55)
L2: 0.24 (-2.49, 3.040
L3:1.21 (-1.41,3.90)
LOS:-1.45 (-5.58, 2.87)
Asthma: LO: 0.46 (-2.41, 3.42)
L1:-1.36 (-4.33, 1.71)
L2:-0.83 (-3.67, 2.10)
L3:-0.78 (-3.63, 2.16)
LOS:-1.75 (-6.21, 2.92)
Pneumonia: LO: 0.06 (-2.74, 2.94)
L1: 0.34 (-2.446, 3.20)
L2: -0.59 (-3.38, 2.29)
L3:-0.44 (-3.22, 2.41)
L03: -0.61 (-4.87, 3.85)
Reference: Bell et al. (2008, 0912681
Period of Study: 1999-2005
Location: 202 U.S. counties
Outcome (ICD-9): Heart failure (428),
heart rhythm disturbances (426-427),
cerebrovascular events (430-438),
ischemic heart disease (410-414, 429),
peripheral vascular disease (440-449),
COPD (490-492), respiratory tract
infections (464 - 466, 480 - 487)
Age Groups: 65+
Study Design: Time series
N:NR
Statistical Analyses: Two-stage
Bayesian hierarchical model to find
national avg
First stage: Poisson regression (county-
specific)
Covariates: Day of the week,
temperature, dew point temperature,
temporal trends, indicator for persons
75+ yr, population size
Season: All, Jun-Aug (Summer),
Sep-Nov (Fall), Dec-Feb (Winter),
Mar-May (Spring)
Dose-response Investigated: No
Statistical Package: NR
Lags Considered: 0- to 2-day lags
Pollutant: PM25
Averaging Time: 24 h
Mean (ug/m3): Descriptive information
presented in Fig S2 (boxplots):
IQR: 8.7 pg/m3
Monitoring Stations: NR
Copollutant (correlation): NR
PM Increment: 10 pg/m
Percent increase [96% PI]:
Cardiovascular admissions:
Lag 0 (all seasons): 0.80 [0.59-1.01]
Lag 0 (winter, national): 1.49 [1.09-1.8
Lag 0 (winter, northeast): 2.01
[1.39-2.63]
Lag 0 (winter, southeast): 1.06
[-0.07-2.21]
Lag 0 (winter, northwest): 0.85
[-4.11-6.07]
Lag 0 (winter, southwest): 0.76
[-0.25-1.79]
Lag 0 (spring, national): 0.91
[0.47-1.35]
Lag 0 (spring, northeast)
0.95 [0.32-1.58]
Lag 0 (spring, southeast): 0.75
[-0.26-1.78]
Lag 0 (spring, northwest): -0.07
[-12.40-13.98]
Lag 0 (spring, southwest): 1.78
[-0.87-4.51]
Lag 0 (summer, national): 0.18
[-0.23-0.58]
Lag 0 (summer, northeast): 0.55
[0.08-1.02]
Lag 0 (summer, southeast): -0.67
[-1.60-0.26]
Lag 0 (summer, northwest): -1.55
[-15.22-14.31]
Lag 0 (summer, southwest): -1.20
[-4.90-2.65]
LagO
LagO
fall, national): 0.68 [0.29-1.07]
fall, northeast): 1.03 [0.48-1.58]
Lag 0 (fall, southeast): 0.17 [-0.72-1.07]
Lag 0 (fall, northwest): -0.67
[-6.96-6.05]
Lag 0 (fall, southwest): 0.30 [-0.98-1.59]
Lag1
Lag1
all seasons): 0.07 [-0.12-0.26]
winter): 0.56 [0.16-0.96]
Lag1 (spring):-0.10 [-0.58-0.39]
Lag1
Lag1
summer):-0.16 [-0.54-0.22]
fall): 0.04 [-0.28-0.35]
Lag2 (all seasons): [0.06 [-0.12-0.23]
Lag 2 winter): 0.27 [-0.12-0.65]
Lag 2 (spring): 0.19 [-0.23-0.60]
December 2009
E-109
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Lag 2 (summer): -0.12 [-0.50-0.26]
Lag 2 (fall): 0.02 [-0.30-0.34]
Respiratory admissions: Lag 0 (all
seasons): 0.22 [-0.12-0.56]
LagO
LagO
winter, national): 1.05 [0.29-1.82]
winter, northeast): 1.76
[0.60-2.93]
Lag 0 (winter, southeast): 0.59
[-1.35-2.58]
Lag 0 (winter, northwest): -0.07
[-6.74-7.08]
Lag 0 (winter, southwest): 0.03
[-1.25-1.34]
Lag 0 (spring, national): 0.31
[-0.47-1.11]
Lag 0 (spring, northeast): 0.34
[-0.66-1.34]
Lag 0 (spring, southeast): -0.06
[-1.77-1.68]
Lag 0 (spring, northwest): -8.52
[-25.62-12.51]
Lag 0 (spring, southwest): 1.87
[-2.00-5.90]
Lag 0 (summer, national): -0.62
[-1.33-0.09]
Lag 0 (summer, northeast): -0.8
[-1.65-0.07]
Lag 0 (summer, southeast): -0.15
[-1.88-1.61]
Lag 0 (summer, northwest): 0.25
[-21.46-27.96]
Lag 0 (summer, southwest): 0.64
[-5.38-7.04]
Lag 0 (fall, national): 0.02 [-0.63-0.67]
LagO
LagO
fall, northeast): -0.01 [-0.87-0.85]
fall, southeast): -0.58
[-2.06-0.91]
LagO (fall, northwest):-1.38
[-11.84-10.32]
Lag 0 (fall, southwest): 1.77 [-0.73-4.33]
Lag1
Lag1
all seasons): 0.05 [-0.29-0.39]
winter): 0.50 [-0.27-1.27]
Lag1 (spring):-0.24 [-1.01-0.53]
Lag1
Lag1
summer): 0.28 [-0.39-0.95]
fall): 0.15 [-0.49-0.79
Lag 2 (all seasons): 0.41 [0.09-0.74]
Lag 2
Lag 2
winter, national): 0.72 [0.01-1.43]
winter, northeast): 0.79
[-0:21-1.80]
Lag 2 (winter, southeast): 0.4 [-1.45,
2.27]
Lag 2 (winter, northwest): -0.06
[-6.52-6.85]
Lag 2 (winter, southwest): 1.2
[-0.10-2.52]
Lag 2 (spring, national): 0.35
[-0.29-0.99]
Lag 2 (spring, northeast): 0.04
[-0.88-0.97]
Lag 2 (spring, southeast): 0.75
[-0.82-2.34]
Lag 2 (spring, northwest): 2.29
[-14.26-22.03]
Lag 2 (spring, southwest): 1.05
[-2.18-4.39]
Lag 2 (summer, national): 0.57
[-0.07-1.23]
Lag 2 (summer, northeast): 0.77
-0.01-1.56]
Lag 2 (summer, southeast): -0.52
[-2.07-1.06]
Lag 2 (summer, northwest): 0.74
[-18.73-24.86]
Lag 2 (summer, southwest): 2.41
[-2.61-7.69]
Lag 2 (fall, national): 0.39 [-0.22-1.01]
Lag 2 (fall, northeast): 0.12 [-0.82-1.07]
Lag 2 (fall, southeast): 0.14 [-1.29-1.59]
December 2009
E-110
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Lag 2 (fall, northwest): -0.74
[-10.08-9.58]
Lag 2 (fall, southwest): 0.97[-1.36-3.36]
Reference: Bell et al. (2009,1910071
Period of Study: 1999-2005
Location: 168 U.S. Counties
Outcome: CVD hospital admissions
Study Design: Retrospective Cohort
Covariates: Socio-economic
conditions, long term temperature
Statistical Analysis: Bayesian
hierarchical model
Age Groups: 2 65 yr
Pollutant: PM25
Averaging Time: 24 h
Mean (SD) Unit: NR
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: 20% of the population
acquiring air conditioning
Percent Change (96% Cl) in
community-specific PM health effect
estimates for CVD hospital
admissions
Any AC, including window units
Yearly health effect: -4.3 (-72.7 to 4.2)
Summer health effect: -148 (-327 to
31.1)
Winter health effect: -80.0 (-182 to 22.0)
Central AC
Yearly health effect: -42.5(-63.4-21.6)
Summer health effect: -79.5 (-143 to
15.7)
Winter health effect: -41.9 (-124 to 40.0)
Reference: Bell et al. (2009,1919971
Period of Study: 1999-2005
Location: U.S.
Outcome: Cardiovascular HA
Age Groups: 65+
Study Design: time series
N:NR
Statistical Analyses: Bayesian
Hierarchical Regression
Covariates: time trend, day of week,
seasonality, dew point, temperature
Statistical Package: NR
Lags Considered: 0-2
Pollutant: PM25
Averaging Time: Daily
Mean:
EC: 0.715
Ni: 0.002
V: 0.003
Min:
EC: 0.309
Ni: 0.003
V: 0.001
Max:
EC: 1.73
Ni: 0.021
V: 0.010
Interquartile Range:
EC: 0.245
Ni: 0.001
V: 0.001
Interquartile Range of Percents:
EC: 1.7
Ni:0.01
V: 0.01
Monitoring Stations: NR
Copollutant: Al, NH4+, As, Ca, Cl, Cu,
EC, CMC, Fe, Pb, Mg. Ni, N03-, K, Si,
Na+, S04=, Ti, V, Zn
Co-pollutant Correlation:
Ni, V: 0.48
V, EC: 0.33
Ni, EC: 0.30
Note: Pollutant concentrations available
for all fractions of PM2 5
PM Increment: Interquartile Range in
the fraction of PM2 5
Percent Increase in PM Health Effect
(Lower Cl, Upper Cl), lag
EC: 25.8 (4.4, 47.2), lag 0
EC+Ni: 14.0 (-7.6, 35.5), lag 0
EC+ V: 14.9 (-7.8, 37.6), lag 0
EC+ V, HS education: 15.0 (3.3, 26.8),
lagO
EC+ V, median income: 15.8 (4.1, 27.5),
lagO
EC+ V, racial composition: 14.2 (2.8,
25.6), lag 0
EC+ V, percent living in urban area:
14.7(3.1,26.3), lag 0
EC+ V, population: 13.6 (2.2, 25.0), lag
0
EC+Ni, V: 11.9(-10.4, 43.2), lagO
Ni: 19.0 (9.9, 28.2), lag 0
Ni +EC: 17.3 (7.7, 26.9), lag 0
Ni+ V: 15.5 (4.1,26.9), lag 0
Ni + EC,V:14.9(3.4, 26.4), lag 0
V: 27.5 (10.6, 44.4), lag 0
V+EC: 23.1 (4.9, 41.4), lag 0
V+Ni: 10.9 (-9.6, 31.5), lag 0
V+EC, Ni: 8.1 (-13.3, 29.5), lag 0
EC: 11.8 (-69.2, 92.8), lag 1
EC: 21.0 (-46.6, 88.6), lag 2
Ni: 20.6 (-15.5, 56.7), lag 1
Ni: -2.3 (-32.5, 27.9), lag 2
V: 34.0 (-31.2, 99.1), lag 1
V: 8.0 (-46.8, 62.7), lag 2
Percent HS education: -17.4 (-46.8,
11.9), lag 0
Median income: 21.3 (-20.0, 62.5), lag 0
Percent black: 26.9 (-15.8, 69.6), lag 0
Percent living in urban area: 34.4 (-
29.0, 97.8), lagO
Population: -4.3 (-13.3, 4.8), lag 0
Notes: Interquartile ranges in percent
HS education, median income, percent
black, percent living in urban area, and
population are 5.2 %, $9,223,17.3%,
11.0%, and 549,283 respectively.
December 2009
E-111
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Chan et al. (2007,1477871 Outcome: Cerebrovascular Emergency Pollutant: PM25
Admissions
Averaging Time: 24 h
Age Groups: 50+yr Mean (SD): 31.5 (16.0)
Study Design: Time series
Period of Study: Apr 1997-Dec 2002
Location: Boston, MA
Statistical Analyses: GAM Poisosn
Regression
Covariates: Yr, mo, day of wk,
temperature, dew point
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0-3 days
Min:15.6
Max: 200.6
IQR: 19.7
Monitoring Stations: 16
Copollutant: 03, CO, S02, N02, PM1(
Co-pollutant Correlation
03: 0.33
CO: 0.44
S02: 0.51
N02: 0.50
PM10:0.61
PM Increment: Interquartile Range
(19.7 |jg/m3)
Percent Change (Lower Cl, Upper Cl),
p-value:
Cerebrovascular Disease
Lag 0:1.006 (0.993, 1.019)
Lag 1:1.002 (0.990,1.014)
Lag 2:1.015 (0.978, 1.052)
Lag 3:1.021 (1.005, 1.037)
Lag 3+ 03:1.009 (0.987, 1.031)
Lag 3+ 00:1.014(0.993, 1.035)
Lag 3+ 03+ 00:1.009(0.987, 1.031)
Stroke
Lag 0:0.931 (0.831, 1.031)
Lag 1:0.936 (0.845, 1.027)
Lag 2: 0.931 (0.820, 1.042)
Lag 3: 0.991 (0.969, 1.013)
Ischaemic stroke
Lag 0:0.981 (0.907, 1.055)
Lag 1:0.994 (0.920, 1.078)
Lag 2: 0.960 (0.885, 1.035)
Lag 3:1.059 (0.984, 1.134)
Haemorrhagic stroke
Lag 0: 0.870 (0.740, 1.010)
Lag 1:0.882 (0.761, 1.003)
Lag 2: 0.909 (0.810, 1.008)
Lag 3: 0.921 (0.830, 1.012)
Reference: Chan et al. (2008, 0932971 Outcome (ICD-9): Emergency visits for Pollutant: PM25
ischaemic heart diseases (410-411,
Period of Study: 1995-2002
Location: Taipei Metropolitan area,
Taiwan
414), Cerebrovascular diseases (430-
437), and COPD (493, 496)
Age Groups: All
Study Design: Time series
N:NR
Statistical Analyses: Poisson
regression
Covariates: Yr, mo, day of wk,
temperature, dewpoint temperature,
PM,o, N02
Season: All
Dose-response Investigated: No
Statistical Package: SAS version 8.0
Lags Considered: 0- to 7-day lags
Averaging Time: 24 h
Mean ug/m3 (SD): NR
Monitoring Stations: 1
Copollutant (correlation): NR
PM Increment: 19.7 pg/rri (IQR)
OR [95% Cl]: In environmental
conditions without dust storms (results
only given for best-fitting model)
Lag 6 days: 1.024 (1.004,1.044)
Reference: Delfmo et al, (2008,
1563901
Period of Study: October
2001-2003-November 2003
Location: Southern California
Outcome: Cardiovascular hospital
admissions
Study Design: Time series
Statistical Analysis: Poisson
regression with GEE
Age Groups: All
Pollutant: PM25
Averaging Time: Hourly
Mean (SD) Unit by county:
Los Angeles
Before Fires: 27.2 (12.4) pg/m3
During Fires: 54.1 (21.0) pg/m3
After Fires: 15.9 (5.5) pg/m
Orange
Before Fires: 23.2 (9.6) pg/m3
During Fires: 64.3 (26.5) pg/m3
After Fires: 15.5(10.2) pg/m3
Riverside
Before Fires: 32.7 (14.7) pg/m3
During Fires: 42.1 (25.5) pg/m
After Fires: 16.9(10.2) pg/m3
San Bernadino
Before Fires: 35.7 (16.6) pg/m3
During Fires: 45.3 (28.7) pg/m3
After Fires: 18.5 (8.3) pg/m
San Diego
Before Fires: 18.5 (6.7) pg/m3
Increment: 10|jg/m
Relative Rate (Min Cl, Max Cl)
All Cardiovascular
All Periods: 0.996 (0.989-1.003)
Pre-Wildfire: 0.992 (0.976-1.009)
Wildfire: 1.008 (0.999-1.018), p = 0.104
Post-Wildfire: 0.991 (0.964-1.019),
p = 0.955
Ischaemic Heart Disease
All Periods: 0.991 (0.980-1.003)
Pre-Wildfire: 0.990 (0.963-1.017)
Wildfire: 0.117 (0.990-1.024), p = 0.313
Post-Wildfire: 0.989 (0.950-1.030),
p = 0.976
Congestive Heart Failure
All Periods: 0.989 (0.974-1.004)
Pre-Wildfire: 0.978 0.942-1.015)
Wldfire: 1.016 (0.933-1.039), p = 0.096
Post-Wildfire: 0.969 (0.914-1.027),
p = 0.791
December 2009
E-112
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
During Fires: 76.1 (66.6) pg/rrf
After Fires: 14.2 (7.2) pg/nr
Ventura
Before Fires: 18.4 (8.3) pg/m3
During Fires: 50.1 (50.5) pg/m3
After Fires: 12.9 (4.3) pg/nf
Copollutant (correlation): NR
Cardiac Dysrhythmia
All Periods: 0.980 (0.962-0.998)
Pre-Wildfire: 0.979 (0.935-1.025)
Wildfire: 0.989 (0.961-1.017), p = 0.721
Post-Wildfire: 0.976 (0.912-1.044),
p = 0.934
Cerebrovascular Disease and Stroke
All Periods: 1.019 (1.004-1.035)
Pre-Wildfire: 1.015 (0.980-1.052)
Wildfire: 1.016 (0.997-1.036), p = 0.971
Post-Wildfire: 1.044 (0.987-1.104),
p = 0.379
Relative Rate (Min Cl, Max Cl) in
relation to pre-wildfire period (1)
All Cardiovascular: Wldfire, unadjusted
for PM25: 0.958 (0.920-0.997)
Wldfire, adjusted for PM25: 0.947
(0.902-0.994)
Post-wildfire, unadjusted for PM25:
1.061 (1.006-1.119)
Post-wildfire, adjusted for PM25:1.053
(0.994-1.114)
Ischaemic Heart Disease: Wldfire,
unadjusted for PM25: 0.913 (0.852-
0.978)
Wldfire, adjusted for PM25: 0.905
(0.832-0.985)
Post-wildfire, unadjusted for PM25:
1.029(0.943-1.123)
Post-wildfire, adjusted for PM25:1.029
(0.936-1.131)
Congestive Heart Failure: Wldfire,
unadjusted for PM25: 0.981 (0.817-
0.972)
Wldfire, adjusted for PM25: 0.911
(0.819-1.014)
Post-wildfire, unadjusted for PM25:
1.113(0.997-1.242)
Post-wildfire, adjusted for PM25:1.105
(0.982-1.244)
Cardiac Dysrhythmia: Wldfire,
unadjusted for PM25: 0.968 (0.874-
1.072)
Wldfire, adjusted for PM25: 0.964
(0.851-1.093)
Post-wildfire, unadjusted for PM25:
1.089(0.949-1.251)
Post-wildfire, adjusted for PM25:1.057
(0.914-1.223)
Cerebrovascular Disease and Stroke:
Wldfire, unadjusted for PM25:1.066
(0.981-1.159)
Wldfire, adjusted for PM25:1.017
(0.922-1.123)
Post-wildfire, unadjusted for PM25:
1.013(0.907-1.132)
Post-wildfire, adjusted for PM25:1.013
(0.902-1.138)
Reference: Dominici et al. (2006,
Period of Study: 1999-2002
Location: 204 U.S. counties, located
in: Alabama, Alaska, Arizona, Arkansas,
California, Colorado, Connecticut,
Delaware, District of Columbia, Florida,
Georgia, Hawaii, Idaho, Illinois, Indiana,
Iowa, Kansas, Kentucky, Louisiana,
Maine, Maryland, Massachusetts,
Michigan, Minnesota, Mississippi,
Missouri, Nevada, New Hampshire,
New Jersey, New Mexico, New York,
North Carolina, Ohio, Oklahoma,
Oregon, Pennsylvania, Rhode Island,
South Carolina, Tennessee, Texas,
Outcome (ICD-9): Daily counts of
hospital admissions for primary
diagnosis of heart failure (428), heart
rhythm disturbances (426-427),
Cerebrovascular events (430-438),
ischemic heart disease (410-414, 429),
peripheral vascular disease (440-448),
chronic obstructive pulmonary disease
(490-492), and respiratory tract
infections (464-466, 480-487).
Age Groups: >65 yr
Study Design: Time series
N: 11.5 million Medicare enrollees
Statistical Analyses: Bayesian 2-stage
Pollutant: PM25
Averaging Time: 24 h
Mean (pg/m3) (IQR): 13.4 (11.3-15.2)
Monitoring Stations: NR
Copollutant (correlation): NR
Other variables: Median of pairwise
correlations among PM25 monitors
within the same county for 2000: r =
0.91 (IQR: 0.81-0.95)
PM Increment: 10 pg/m (Results in
figures; see notes)
Percent increase in risk [95% PI]:
Cerebrovascular disease (Lag 0):
Age 65+: 0.81 [0.30, 1.32]
Age 65-74: 0.91 [0.01, 1.82]
Age 75+: 0.80 [0.21, 1.38]
Peripheral vascular disease (Lag 0):
Age 65+: 0.86 [-0.06, 1.79]
Age 65-74:1.21 [-0.26,2.67]
Age 75+: 0.86 [-0.39, 2.11]
Ischemic heart disease (Lag 2):
Age 65+: 0.44 [0.02, 0.86]
Age 65-74: 0.37 [-0.22, 0.96]
Age 75+: 0.52 [-0.01, 1.04]
December 2009
E-113
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Utah, Virginia, Washington, West
Virginia, Wisconsin
hierarchical models.
First stage: Poisson regression (county-
specific)
Second stage: Bayesian hierarchical
models, to produce a national avg
estimate
Covariates: Day of the week,
seasonality, temperature, dew point
temperature, long-term trends
Season: NR
Dose-response Investigated: No
Statistical Package: R statistical
software version 2.2.0
Lags Considered: 0-2 days, avg of
days 0-2
Heart rhythm disturbances (Lag 0):
Age 65+: 0.57 [-0.01,1.15]
Age 65-74: 0.46 [-0.63, 1.54]
Age 75+: 0.72 [0.02, 1.42]
Heart failure (Lag 0):
Age 65+: 1.28 [0.78, 1.78]
Age 65-74:1.21 [0.35, 2.07]
Age 75+: 1.36 [0.78, 1.94]
COPD (Lag 0):
Age 65+: 0.91 [0.91, 1.64]
Age 65-74: 0.42 [-0.64, 1.48]
Age 75+: 1.47 [0.54, 2.40]
Respiratory tract infection:
Age 65+: 0.92 [0.41, 1.43]
Age 65-74: 0.93 [0.04, 1.82]
Age 75+: 0.92 [0.32, 1.53]
Annual reduction in admissions
attributable to a 10 pg/m reduction in
daily PM25 level (95% PI):
Cerebrovascular disease: Annual
number of admissions: 226,641
Annual reduction in admissions: 1836
[680, 2992]
Peripheral vascular disease: Annual
number of admissions: 70,061
Annual reduction in admissions: 602 [-
42,1254]
Ischemic heart disease: Annual number
of admissions: 346,082
Annual reduction in admissions: 1523
[69, 2976]
Heart rhythm disturbances: Annual
number of admissions: 169,627
Annual reduction in admissions: 967 [-
17,1951]
Heart failure: Annual number of
admissions: 246,598
Annual reduction in admissions: 3156
[1923, 4389]
COPD: Annual number of admissions:
108,812
Annual reduction in admissions: 990
[196,1785]
Respiratory tract infections: Annual
number of admissions: 226,620
Annual reduction in admissions: 2085
[929, 3241]
Notes: Fig 2: Point estimates and 95%
posterior intervals of the % change in
admissions rates per 10 pg/m3 (national
avg relative rates) for single lag (0,1,
and 2 days) and distributed lag models
for 0 to 2 days (total) for all outcomes.
Summary: Positive significant or
marginally significant associations
between PM2 5 and cerebrovascular
disease at Lag 0
Peripheral vascular disease at Lags 0
and 2
Ischemic heart disease at Lag 2
Heart rhythm disturbances at Lag 0
December 2009
E-114
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Heart failure at Lag 0, Lag 2, and Lags
0-2
COPD at Lag 0, Lag 1, and Lags 0-2
and respiratory tract infections at Lag 2
and Lags 0-2.
Fig 3: Point estimates and 95%
posterior intervals of the % change in
admission rates per 10 pg/m3 (regional
relative rates). Summary: For
cardiovascular diseases, all estimates
in the Midwestern, Northeastern, and
Southern regions were positive, while
estimates in the other regions (South,
V\fest, Central, Northwest) were close to
0. For respiratory disease, there were
larger effects in the Central,
Southeastern, Southern, and Wfestern
regions than in the other regions.
Fig 4: Point estimates and 95%
posterior intervals of the % change in
admission per 10 pg/m3 (Eastern vs..
Wfestern regions): Summary: All
estimates for cardiovascular outcomes
were positive in the U.S. Eastern region
but not in the U.S. Wfestern region. The
estimates for respiratory tract infections
were larger in the Wfestern region than
in the Eastern region. The estimates for
CCPDwere positive in the both regions.
Reference: Halonen et al. (2009,
1803791
Period of Study: 1998-2004
Location: Helsinki, Finland
Outcome: Cardiovascular
Hospitalizations & Mortality (ICD 10:
100-99)
Age Groups: 65+ yr
Study Design: Time series
N:NR
Statistical Analyses: Poisson, GAM
Covariates: Temperature, humidity,
influenza epidemics, high pollen
episodes, holidays
Dose-response Investigated? No
Statistical Package: R
Lags Considered: lags 0-3 days; 5-day
(0-4) mean
Pollutant: PM25
Averaging Time: Daily
Mean (SD): NR
Min:1.1
26th percentile: 5.5
60th percentile: 9.5
76th percentile: 11. 7
Max: 69.5
Monitoring Stations: NR
Copollutant:
PMO.03, PM0.03-0.1,PM<0.1,
PM<0. 10.29, PMio-25, CO, N02
Co-pollutant Correlation
PMO.03: 0.14
PMO.03-0.1: 0.48
PMO.1: 0.35
PMO. 10.29: 0.88
PMio-2.5: 0.25
PM Increment: Interquartile Range
Percent Change (Lower Cl, Upper
Cl):
All Cardiovascular Morality
Lag 0:0.73 -0.66, 2.13
Lag 1:0.74 -0.63,2.13
Lag 2: 0.74 (-0.62, 2.11)
Lag 3: 0.06 (-1.29, 1.43)
5-day mean: 0.87 (-0.94, 2.70)
Coronary Heart Disease HA
Lag 0: -0.1 7 (-1.5.0, 1.18)
Lag 1: -0.03 (-1.31, 1.26)
Lag 2: -0.63 (-1.87, 0.62)
Lag 3: 0.48 (-0.78, 1.76)
5-day mean: 0.80 (-0.94 2.58)
Stroke HA
Lag 0: -0.99 (-2.78, 0.84)
Lag 1:0.02 (-1.74, 1.82)
Lag 2: -1.38 (-3.13, 0.40)
Lag 3: -0.1 7 (-1.92, 1.61)
5-day mean: -0.78 (-3.10, 1.60)
Arrhythmia HA
Lag 0:0.82 -1.03,2.68
Lag 1:0. 18 -1.58, 1.97
Lag 2: -0.09 (-1.82, 1.67)
Lag 3: -0.48 (-2.22, 1.29)
5-day mean: 0.16 (-2.16, 2.54)
*p < 0.05,
0.10
December 2009
E-115
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Host et al. (2008,1558521
Period of Study: 2000-2003
Location: Six French cities: Le Havre,
Lille, Marseille, Paris, Rouen, and
Toulouse
Outcome (ICD-10): Daily
hospitalizations for all cardiovascular
(IOO-I99), cardiac (IOO-I52), and
ischemic heart diseases (I20-I25), all
respiratory diseases (JOO-J99),
respiratory infections (J10-J22).
Age Groups: For cardiovascular
diseases: All ages, and restricted to >
65 yr. For all respiratory diseases: 0-14
yr, 15-64 yr, and > 65 yr. For respiratory
infections: All ages
Study Design: Time series
N: NR (Total population of cities:
approximately 10 million)
Statistical Analyses: Poisson
regression
Covariates: Seasons, days of the wk,
holidays, influenza epidemics, pollen
counts, temperature, and temporal
trends
Season: NR
Dose-response Investigated: No
Statistical Package: MGCV package in
R software (R 2.1.1)
Lags Considered: Avg of 0-1 days
Pollutant: PM25
Averaging Time: 24 h
Mean (6th -96th percentile):
Le Havre: 13.8 (6.0-30.5)
Lille: 15.9 (6.9-26.3)
Marseille: 18.8 (8.0-33.0)
Paris: 14.7 (6.5-28.8)
Rouen: 14.4 (7.5-28.0)
Toulouse: 13.8 (6.0-25.0)
Monitoring Stations:
13 total: 1 in Toulouse
4 in Paris
2 each in other cities
Copollutant (correlation):
PM10.2.5. Overall: r>0.6
Ranged between r = 0.28 and
r = 0.73 across the six cities.
PM Increment: 10 pg/m increase, and
a 27 pg/m3 increase (corresponding to
the difference between the lowest of the
5th percentiles and the highest of the
95th percentiles of the cities'
distributions)
ERR (excess relative risk) Estimate [Cl]:
For all cardiovascular diseases (10
pg/m increase): All ages: 0.9% [0.1,
1.8]
> 65 yr: 1.9% [0.9, 3.0]
For all cardiovascular diseases (27
pg/m3 increase): All ages: 2.5% [0.2,
4.9]
> 65 yr: 5.3% [2.6, 8.2]
For ischemic heart diseases (27 pg/m3
increase): All ages: 5.2% [-0.6,11.3]
> 65 yr: 12.7% [6.3, 19.5]
For cardiac diseases (10 pg/m3
increase): All ages: 0.9% [-0.1, 2.0]
> 65 yr: 2.4% [1.2, 3.7]
For cardiac diseases (27 pg/m3
increase): All ages: 2.5% [-0.3, 5.4]
> 65 yr: 6.8% [3.3, 10.3]
For ischemic heart diseases (10 pg/m3
increase): All ages: 1.9 % [-0.2, 4.0]
> 65 yr: 4.5% [2.3, 6.8]
For all respiratory diseases (10 pg/m3
increase): 0-14 yr: 0.4% [-1.2, 2.0]
15-64 yr: 0.8% [-0.7, 2.3];
> 65 yr: 0.5% [-2.0, 3.0]
For all respiratory diseases (27 pg/m3
increase): 0-14 yr: 1.1% [-3.1, 5.5]
15-64 yr: 2.2% [-1.8, 6.4];
> 65 yr: 1.3% [-5.3, 8.2]
For respiratory infections
increase): All ages: 2.5%
10 pg/m3
0.1,4.8]
For respiratory infections (27 pg/m3
increase): All ages: 7.0% [0.7,13.6]
December 2009
E-116
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Jalaludin et al. (2006,
1894161
Period of Study: Jan 1997-Dec 2001
Location: Sydney, Australia
Outcome (ICD-9): Cardiovascular
disease (390-459), cardiac disease
(390-429), ischemic heart disease (410-
413) and cerebrovascular disease or
stroke (430-438)
Age Groups: 65+ yr
Study Design: Time series
N:NR
Statistical Analyses: GAM, GLM
Covariates: Temperature, humidity
Season: Warm (Nov-Apr) and cool
(May-Oct)
Dose-response Investigated: No
Statistical Package: S-Plus
Lags Considered: 0-3 days
Pollutant: PM25
Averaging Time: 24 h
Mean (min-max): 9.5 (2.4-82.1)
SD = 5.1
Monitoring Stations: 14
Copollutant (correlation):
Warm
BSP:r = 0.93
PM,o:r = 0.89
03:r = 0.57
N02:r = 0.45
CO: r = 0.35
S02:r = 0.27
Cool
BSP:r = 0.90
PMi0:r = 0.88
03:r = 0.05
N02:r = 0.68
CO: r = 0.60
S02:r = 0.46
Other variables:
Warm
Temp: r = 0.24
Pel humidity: r = -0.15
Cool
Temp: r = -0.04
Pel humidity: r = 0.20
PM Increment: 4.8 pg/rri (IQR)
Percent Change Estimate [Cl]: All CVD
Same-day lag: 1.26 [0.56,1.96]
Avg 0-1 day lag: 0.85 [0.18,1.52]
Cool (same-day lag): 2.23 [0.98,3.50]
Warm (same-day lag): 0.73 [-0.05,1.52]
Cardiac disease
Same-day lag: 1.55 [0.74,2.38]
Avg 0-1 day lag: 1.33 [0.54,2.13]
Cool (same-day lag): 2.37 [0.87,3.89]
Warm (same-day lag): 1.13 [0.22,2.04]
Ischemic heart disease
Same-day lag: 1.17 [-0.08,2.44]
Avg 0-1 day lag: 1.24 [0.04,2.45]
Cool (same-day lag): 0.57 [-1.74,2.94]
Warm (same-day lag): 1.31 [-0.04,2.68]
Stroke
Same-day lag: -0.89 [-2.41,0.65]
Avg 0-1 day lag:-1.08 [-2.54,0.41]
Cool (same-day lag): 1.45 [-1.17,4.15]
Warm (same-day lag): -2.19
[-4.00.-0.36]
Notes: All other lag-day ORs were
provided, yet none were significant.
Percent change in ED attendance was
also reported graphically (Fig 1-5).
Reference: Jalaludin et al. (2006,
1894161
Period of Study: Jan 1997-Dec 2001
Location: Sydney, Australia
Outcome (ICD-9): Cardiovascular
disease (390-459), cardiac disease
(390-429), ischemic heart disease (410-
413) and cerebrovascular disease or
stroke (430-438)
Age Groups: 65+ yr
Study Design: Time series
N:NR
Statistical Analyses: GAM, GLM
Covariates: Temperature, humidity
Season:
Warm (Nov-Apr) and cool (May-Oct)
Dose-response Investigated: No
Statistical Package: S-Plus
Lags Considered: 0-3 days
Pollutant: BS,P
Averaging Time: 24 h
Mean/104/m (min-max):
0.26 (0.04-3.37)
SD = 0.22
Monitoring Stations: 14
Copollutant (correlation):
Warm
PM25:r = 0.93
PM,0:r = 0.82
03:r = 0.48
N02:r = 0.35
CO: r = 0.33
S02:r = 0.21
Cool
PM25:r = 0.90
PM10:r = 0.75
03:r = -0.08
N02:r = 0.59
CO: r = 0.62
S02:r = 0.48
Other variables:
Warm
Temp: r = 0.23
Rel humidity: r = -0.04
Cool
Temp: r = -0.09
Rel humidity: r = 0.36
PM Increment: 0.18/104/m (IQR)
Percent Change Estimate [Cl]:
All CVD
Same-day lag: 1.05 [0.44,1.66]
Avg 0-1 day lag: 0.79 [0.20,1.38];
Cool (same-day lag): 2.38 [1.15,3.62]
Warm (same-day lag): 0.45 [-0.18,1.09]
Cardiac disease
Same-day lag: 1.34 [0.63,2.05]
Avg 0-1 day lag: 1.13 [0.44,1.82];
Cool (same-day lag): 2.50 [1.04,3.98]
Warm (same-day lag): 0.80 [0.07,1.54]
Ischemic heart disease
Same-day lag: 0.91 [-0.17,2.02]
Avg 0-1 day lag: 0.90 [-0.14,1.95];
Cool (same-day lag): 0.52 [-1.74,2.83]
Warm (same-day lag): 0.93 [-0.15,2.03]
Stroke
Same-day lag:-0.93 [-2.27,0.42]
Avg 0-1 day lag:-0.82 [-2.11,0.49];
Cool (same-day lag): 1.38 [-1.19,4.01];
Warm (same-day lag): -1.85
[-3.31,-0.36]
Notes: All other lag-day ORs were
provided, yet none were significant.
Percent change in ED attendance was
also reported graphically (Fig 1-5).
December 2009
E-117
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Lisabeth et al. (2008,
1559391
Period of Study: 2001-2005
Location: Nueces County, Texas
Outcome: Ischemic stroke and
transient ischemic attacks (ICD codes
not reported).
Age Groups: 45+ yr
Study Design: Time series
N: 3,508 stroke/TIAs (2,350 strokes,
and1,158TIAs)
Statistical Analyses: Poisson
regression
Covariates: Temperature, day of week,
temporal trends
Season: All, but looked at potential
effect modification by season (Summer:
Jun-Sep; Non-summer: Oct-May)
Dose-response Investigated: No
Statistical Package: S-plus7.Q
Lags Considered: Lags 0-5 days, and
avg lag effect (0-5 days)
Pollutant: PM25
Averaging Time: 24 h
Median ug/m3 (IQR): 7.0 (4.8-10.0)
Monitoring Stations: 6
Copollutant (correlation): NR
PM Increment: 5.1 pg/rri (IQR)
RR Estimate [Cl]: Lag 0:1.03 (0.99,
1.07)
Lag 1:1.03 (1.00-1.07)
All other lags and avg (lag 0-5) were not
statistically or marginally significant.
Adjusted for03: Lag 0:1.03 (0.99,1.07)
Lag 1:1.03 (0.99-1.06)
All other lags and avg (lag 0-5) were not
statistically or marginally significant.
Notes: Fig 3: % change in stroke/TIA
risk associated with an IQR increase in
PM25
Reference: Metzger et al. (2004,
0442221
Period of Study: Aug 1998-Aug 2000
Location: Atlanta Metropolitan area
(Georgia)
Outcome (ICD-9): Emergency visits for
ischemic heart disease (410-414),
cardiac dysrhythmias (427), cardiac
arrest (427.5), congestive heart failure
(428), peripheral vascular and
cerebrovascular disease (433-437, 440,
443.444, 451-453), atherosclerosis
(440), and stroke (436).
Age Groups: All
Study Design: Time series
N: 4,407,535 emergency department
visits for 1993-2000 (data not reported
for 1998-2000)
Statistical Analyses: Poisson
generalized linear modeling
Covariates: Day of the wk, hospital
entry and exit indicator variables,
federally observed holidays, temporal
trends, temperature, dew point
temperature
Season: All
Dose-response Investigated: No
Statistical Package: SAS
Lags Considered: 3-day ma, lags 0 -7
Pollutant: PM25
Averaging Time: 24 h
Median ug/m3 (10%-90% range):
PM25:17.8 (8.9, 32.3)
PM2 5 water soluble metals: 0.021
(0.006-0.061)
PM25 acidity: 4.5 (1.9-1.07)
PM25OC: 0.010
(-0.001-0.045)
PM25EC:4.1 (2.2-7.1)
Monitoring Stations: 1
Copollutant (correlation):
PM10:r = 0.84
03:r = 0.65
N02:r = 0.46
CO: r = 0.44
S02:r = 0.17
PMi0.25:r=.43
UFP:r = -0.16
PM2 5 water-sol metals: r = 0.70
PM25sulfates:r = 0.77
PM25 acidity: r = 0.58
PM25OC:r = 0.51
PM25EC:r = 0.48
oxygenated hydrocarbon: r = 31
Other variables:
Temperature: r = 0.20
Dew point: r = 0.00
PM Increment: Approximately 1 SD
increase: PM25:10|jg/m3
PM2 5 water-sol metals: 0.03 pg/m3
PM25sulfates:5|jg/m3
PM2 5 acidity: 0.02 pequ/m3
PM25OC:2|jg/m3
PM25EC:1|jg/m3
RR [95% Cl]: PM25 (3-day ma):
All CVD: 1.033 [1.010, 1.056]
Dysrhythmia: 1.015 [0.976,1.055]
Congestive heart failure:
1.055 [1.006-1.105]
Ischemic heart disease:
1.023 [0.983-1.064]
Peripheral vascular and
cerebrovascular disease:
1.050 [1.008-1.093]
PM;: water soluble metals (3-day
ma):
All CVD: 1.027(0.998, 1.056]
Dysrhythmia: 1.031 [0.982,1.082]
Congestive heart failure: 1.040
[0.981-1.103]
Ischemic heart disease: 1.000
[0.951-1.051]
Peripheral vascular and
cerebrovascular disease: 1.043
[0.991-1.098]
PM25sulfates (3-day ma):
All CVD: 1.003 [0.968, 1.039]
Dysrhythmia: 0.986 [0.926,1.048]
Congestive heart failure: 1.009
[0.938-1.085]
Ischemic heart disease: 0.997
[0.936-1.062]
Peripheral vascular and
cerebrovascular disease: 1.025
[0.964-1.090]
PM2 5 acidity (3-day ma):
December 2009
E-118
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
All CVD: 0.994 [0.966, 1.022]
Dysrhythmia: 0.991 [0.942,1.043]
Congestive heart failure: 0.989
[0.930-1.052]
Ischemic heart disease: 0.992
[0.944-1.043]
Peripheral vascular and
cerebrovascular disease: 1.004
[0.955-1.056]
PM;: OC (3-day ma)
All CVD: 1.026 [1.006, 1.046]
Dysrhythmia: 1.008 [0.975,1.044]
Congestive heart failure: 1.048
[1.007-1.091]
Ischemic heart disease: 1.028
[0.994-1.064]
Peripheral vascular and
cerebrovascular disease:
1.026 [0.990-1.062]
hydrocarbons simultaneously.
PM;: OC (3-day ma):
All CVD: 1.020 [1.005, 1.036]
Dysrhythmia: 1.011 [0.985,1.037]
Congestive heart failure: 1.035
[1.003-1.068]
Ischemic heart disease: 1.019
[0.992-1.046]
Peripheral vascular and
cerebrovascular disease: 1.021
[0.994-1.049]
Results for Lags 0-7 expressed in
figures (see notes).
Notes: Fig 1: RR (95% Cl) for single-
day lag models for the association of
ER visits for CVD with daily ambient
PM2 5 and associated components.
Summary of Fig 1 results: Statistically
significant positive associations at Lag
0 and Lag 1 for PM25, at Lag 0 for PM25
water soluble metals (inverse
association at Lag 7), at Lag 0, Lag 1,
and Lag 3 for organic and EC (inverse
association at Lag 7).
Fig2:RR(95%)ofmultipollutant
models for the association of ER visits
for CVD with daily ambient air quality
measurements.
Summary of Fig 2 results: Positive
association after adjustment for N02,
CO, and oxygenated hydrocarbons, but
attenuated when adjusted for total
carbon and null when adjusted for N02,
CO, total carbon, and oxygenated
December 2009 E-119
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Peng et al. (2008,1568501
Period of Study: Jan 1999-Dec 2005
Location: 108 U.S. counties in the
following states: Alabama, Arizona,
California, Colorado, Connecticut,
District of Columbia, Florida, Georgia,
Idaho, Illinois, Indiana, Kentucky,
Louisiana, Maine, Maryland,
Massachusetts, Michigan, Minnesota,
Missouri, Nevada, New Hampshire,
New Jersey, New Mexico, New York,
North Carolina, Ohio, Oklahoma,
Oregon, Pennsylvania, Rhode Island,
South Carolina, Tennessee, Texas,
Utah, Virginia, Washington, West
Virginia, Wisconsin
Outcome (ICD-9): Emergency
hospitalizations for: Cardiovascular
disease, including heart failure (428),
heart rhythm disturbances (426-427),
cerebrovascular events (430-438),
ischemic heart disease (410-414, 429),
and peripheral vascular disease
(440-448). Respiratory disease,
including COPD (490-492) and
respiratory tract infections (464-466,
480-487)
Age Groups: 65 + yr, 65-74, ,75 +
Study Design: Time series
N: ~12 million Medicare enrollees (3.7
million CVD and 1.4 million RD
admissions)
Statistical Analyses: Two-stage
Bayesian hierarchical models:
Overdispersed Poisson models for
county-specific data
Bayesian hierarchical models to obtain
national avg estimate
Covariates: Day of the week, age-
specific intercept, temperature, dew
point temperature, calendar time,
indicator for age of 75 yr or older. Some
models were adjusted for PM10.2.5.
Season: NR
Dose-response Investigated: No
Statistical Package: R version 2.6.2
Lags Considered: 0-2 days
Pollutant: PM25
Averaging Time: 24 h
Mean ug/m3 (IQR):
All counties assessed: 13.5 (11.1-15.8)
Counties in Eastern U.S.:
13.8(12.3-15.8)
Counties in Western U.S.:
11.1 (10.1-14.3)
Monitoring Stations: At least 1 pair of
co-located monitors (physically located
in the same place) for PM10 and PM25
per county
Other variables: Median within-county
correlations between monitors: r = 0.92
PM Increment: 10 pg/m
Percentage change [95% Cl]: CVD and
RD (unadjusted for PM10-25): Lag 0:
0.71 [0.45, 0.96]
Lag 2: 0.44 [0.06, 0.82]
Most values NR (see note)
Notes: Effect estimates for PM10.2.5 (0-2
day lags) are showing in Fig 2-5.
Fig 2: Percentage change in emergency
hospital admissions for CVD per 10
pg/m3 increase in PM25 (single pollutant
model and model adjusted for PMi0.25
concentration)
Fig 3: Percentage change in emergency
hospital admissions for RD per 10
pg/m increase in PM25 (single pollutant
model and model adjusted for PMi0.25
concentration)
No significant associations between
PM25 and cause-specific cardiovascular
disease.
Reference: Peters et al. (2005,
1568591
Period of Study: Feb 1999-Jul 31,
2001
Location: Germany: City of Augsburg,
County Augsburg, and County Aichach-
Friedlberg
Outcome: Transmural or
nontransmural acute Ml
Age Groups: NR
Study Design: Case-crossover and
time series
N: 851 Ml survivors
Statistical Analyses: Conditional
logistic regression for case-crossover
element. Poisson regression for time
series element.
Covariates: Case-crossover: Season,
temperature, day of the week, time
series: trend, season, influenza,
weather, and day of the week
Season: All
Dose-response Investigated: No
Statistical Package:
SAS, version 8.2
Poisson: R, version 1.7.1
Lags Considered: Lags 0-6 h, 0-5
days
Poisson: Single lagged days, 5-day,
15-day, 30-day, and 45-day ma
Pollutant: PM25
Averaging Time: 1 h and 24 h
Mean ug/m3 (range IQR/ median
IQR):
1-h avg: 16.3 (-6.9-355.2
10.7-19.8
14.5)
24-h avg: 16.3 (6.1-58.5
11.6-19.3
14.9)
Monitoring Stations: 1
Copollutant (correlation):
24-h avg:
TNC:r = 0.37
TSP:r = 0.89
PM10:r=0.92
CO: r = 0.57
N02:r = 0.67
NO: r = 0.59
S02:r = 0.58
03:r = -0.24
Ihravg:
TNC:r = 0.42
CO: r = 0.52
N02:r = 0.58
NO: r = 0.50
S02:r = 0.48
03:r = -0.35
Other variables:
24-h avg: Temperature: r = 0.05
1-h avg: Temperature: r = -0.01
PM Increment: 1-h avg: 9.1 pg/m
(IQR)
24-h avg: 7.7 pg/m3 (IQR)
OR [95% Cl]: Case-Crossover (control
selection method (unidirectional with
three control periods):
1-h avg: Lag 0:0.98 (0.88,1.10)
Lag 1:0.97
Lag 2: 0.93
0.87, 1.09)
0.83, 1.04)
Lag 3: 0.98 (0.88, 1.09)
Lag 4: 0.96
Lag 5: 0.94
0.86, 1.07)
0.84, 1.05)
Lag 6: 0.90 (0.80,1.01).
24-h avg: Lag 0:0.95 (0.83,1.080)
Lag 1:1.10 0.96,1.25)
Lag 2:1.18 1.03,1.34)
Lag 3:1.07 (0.94, 1.22)
Lag 4: 0.94 (0.83, 1.07)
Lag 5: 0.90 (0.79, 1.02)
Case-Crossover (control selection
method: bidirectional with 16 control
periods):
24-h avg: Lag 0:1.03 (0.94,1.12)
Lag 1:1.07
Lag 2:1.0'
0.98,1.16)
0.99, 1.17)
Lag 3:1.01 (0.92,1.10)
Lag 4: 0.96 (0.88, 1.04)
Lag 5: 0.93 (0.85, 1.02)
Lag 0-4 (IQR = 5.8): 1.03 (0.94, 1.14)
Unidirectional: Model 1 (unadjusted):
1.175(1.033,1.337)
Model 2 (adjusted for day of week using
indicator variables): 1.179(1.035,
December 2009
E-120
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
1343)
Model 3 (adjusted for temperature-
quadratic, linear air pressure): 1.170
(1.028, 1.333)
Model 4 (adjusted for temperature-
quadratic, linear air pressure, day of
week): 1.176 (1.031,1.341)
Model 5 (temperature-quadratic, air
pressure-quadratic, relative humidity-
quadratic, day of week using indicator
variables): 1.170 (1.026,1.336)
Model 6 (temperature-penalized spline,
4.4 df, linear air pressure, day of week
using indicator variables): 1.175 (1.030,
1.340
Model 7 (temperature-penalized spline,
4.4 df, linear air pressure, relative
humidity-penalized spline, 7.8 df, day of
week using indicator variables: 1.177
(1.030, 1.344)
Bidirectional (16 control periods): Model
1 (unadjusted): 1.077 (0.988,1.174)
Model 2 (adjusted for day of the week
using indicator variables): 1.078 (0.988,
1.175)
Model 3 (adjusted for temperature-
quadratic, linear air pressure): 1.060
(0.970,1.160)
Model 4 (adjusted for temperature-
quadratic, linear air pressure, day of the
week): 1.060 (0.969,1.160)
Model 5
(temperature-quadratic, air pressure-
quadratic, relative humidity-quadratic,
day of the week using indicator
variables): 1.065 (0.973,1.166)
Model 6 (temperature-penalized spline,
4.4 df, linear air pressure, day of the
week using indicator variables): 1.068
(0.976,1.168)
Model 7 (temperature-penalized spline,
4.4 df, linear air pressure, relative
humidity-penalized spline, 7.8 df, day of
the week using indicator variables:
1.077(0.983,1.179)
Bidirectional (4 control periods): Model
1 (unadjusted): NR
Model 2 (adjusted for day of the week
by design): 1.049 (0.964,1.141)
Model 3 (adjusted for temperature-
quadratic, linear air pressure): NR
Model 4 (adjusted for temperature-
quadratic, linear air pressure, day of the
week): 1.032 (0.944,1.128)
Model 5 (temperature-quadratic, air
pressure-quadratic, relative humidity-
quadratic, day of the week by design):
1.033(0.945, 1.130)
Model 6 (temperature-penalized spline,
4.4 df, linear air pressure, day of the
week by design): 1.036(0.947,1.132)
Model 7 (temperature-penalized spline,
4.4 df, linear air pressure, relative
December 2009 E-121
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
humidity-penalized spline, 7.8 df, day of
the week by design): 1.039(0.950,
1.136)
Stratified: Model 1 (unadjusted): NR
Model 2 (adjusted for day of week by
design): 1.059 (0.972,1.154)
Model 3 (adjusted for temperature-
quadratic, linear air pressure): NR
Model 4 (adjusted for temperature-
quadratic, linear air pressure, day of
week): 1.047 (0.957,1.145)
Model 5 (temperature-quadratic, air
pressure-quadratic, relative humidity-
quadratic, day of week by design):
1.045(0.954,1.144)
Model 6 (temperature-penalized spline,
4.4 df, linear air pressure, day of week
by design): 1.054 (0.964,1.153odel7
(temperature-penalized spline, 4.4 df,
linear air pressure, relative
humidity-penalized spline, 7.8 df, day of
week by design): 1.056(0.965,1.156)
RR (95% Cl): Time series (24 h avg):
Lag 0: 0.97
Lag 1:1.04
0.89, 1.07)
0.96, 1.13)
Lag 2:1.07 (0.98,1.15)
Lag 3:1.03
Lag 4: 0.98
0.95, 1.11)
0.90, 1.07)
Lag 5: 0.98 (0.90, 1.06)
Lag 0-4:1.03 (0.94,1.12)
Lag 0-14:1.03 (0.95, 1.13)
Lag 0-29:1.09 (1.01, 1.18)
Lag 0-44:1.08 (1.00, 1.17)
Time series (OR [96% Cl]): Model 1
(unadjusted): 1.059 (0.981,1.142)
Model 2 (adjusted for day of week using
indicator variables): 1.056 (0.979,
1.140)
Model 3 (adjusted for temperature-
quadratic, linear air pressure): 1.062
(0.982, 1.148)
Model 4 (adjusted for temperature-
quadratic, linear air pressure, day of
week): 1.059 (0.979,1.146)
Model 5 (temperature-quadratic, air
pressure-quadratic, relative humidity-
quadratic, day of week using indicator
variables): 1.063 (0.981,1.151)
Model 6 (temperature-penalized spline,
4.4 df, linear air pressure, day of week
using indicator variables): 1.065 (0.985,
1.153)
Model 7 (temperature-penalized spline,
4.4 df, linear air pressure, relative
humidity-penalized spline, 7.8 df, day of
week using indicator variables: 1.069
(0.988,1.157)
December 2009 E-122
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Pope et al.(2006, 0912461
Period of Study: 1994-2004
Location: Wasatch Front area, Utah
Outcome: Myocardial infarction or
unstable angina (ICD codes not
reported)
Age Groups: All, <65, 65+
Study Design: Case-crossover
N: 12,865 patients who underwent
coronary arteriography
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature and dew
point temperature
Season: NR
Dose-response Investigated: No
Statistical Package: NR
Lags Considered: 0- to 3-day lag, 2- to
4-day lagged ma
Pollutant: PM25
Averaging Time: 24 h
Mean (ug/m3) (SD maximum):
Ogden: 10.8 (10.6
108)
SLC Hawthorne: 11.3 (11.9
94)
Provo/Orem, Lindom: 10.1 (9.8
82)
Monitoring Stations: 3
Copollutant (correlation): NR
PM Increment: 10 pg/m
Percent increase in risk [95% Cl]:
Same-day increase in PM25 (Lag 0):
Index Ml and unstable angina: 4.81
[0.98-879]
Subsequent Ml: 3.23 [-3.87,10.85]
All acute coronary events: 4.46
[1.07-7.97]
All acute coronary events excluding
observations using imputed PM25 data:
4.24 [0.33-8.31]
Stable presentation: -2.57 [-5.39, 0.34]
Remaining results summarized in
figures (see notes).
Notes: Fig 1: Percent increase in risk
(and 95% Cl) of acute coronary events
associated with 10 pg/m3 of PM25 for
different lag structures.
Summary of Fig 1: Positive, statistically
significant association seen for Lag 0,
Lag 1 and 2, 3, and 4 day ma. Positive
but non-statistically significant
associations seen for Lags 2 and 3.
Fig 2: Percent increase in risk (and 95%
Cl) of acute coronary events associated
with 10 pg/m3 of PM25 stratified by
various characteristics.
Reference: Pope et al. (2008,1919691
Period of Study: 1994-2006
Location: Ogden, Salt Lake City, &
Provo/Orem, Utah
Outcome: Heart Failure
Hospitalizations
Age Groups: NR
Study Design: Case-crossover
N: 2,618
Statistical Analyses: Conditional
Logistic Regression
Covariates: Age, sex, length of stay,
temperature, pressure, clearing index,
day of the week, seasonality, and long-
term trends
Season: Adjusted for long-term trends
to account for season
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0- to 28-day ma.
Pollutant: PM25
Averaging Time: NR
Mean (SD):
Ogden: 10.6(9.9)
SLC, Hawthorne: 11.1 (11.2)
Provo/Orem, Lindon: 10.1 (9.3)
Max:
Ogden: 108
SLC, Hawthorne: 94
Provo/Orem, Lindon: 82
Monitoring Stations: NR
Copollutant: PM10
PM Increment: 10 pg/m
Percent Increase: (Lower Cl, Upper
Cl):
All HF Admissions
All: 13.1 (1.3,26.2)*
Men: 13.4 (-1.7, 30.7)|
Vtomen: 12.7 (-5.1,33.9)
Age <65yr: 3.5 (-13.5, 23.8)
Age >65yr: 19.6 (4.0, 37.5)*
Length of stay 0-2 days:
24.4 (-0.8, 56.0) J
Length of stay 3-7 days:
10.8 (-4.6, 28.7)
Length of stay 8+ days:
6.5 (-15.9, 34.8)
First HF Admissions: 2.1 (-11.3,17.5)
Subsequent HF Admits: 32.4 (10.7,
58.4) T
All HF Admissions
All: 32.4(10.7,58.4)1
Men: 29.2 (2.7, 62.6)*
V\fomen:41.5(5.4, 89.9)*
Age<65yr:-3.1 (-26.5,27.8)
Age >65yr: 64.1 (28.6,109)t
Length of stay 0-2 days:
68.9(12.5,154)*
Length of stay 3-7 days:
35.7 (5.9, 73.9)*
Length of stay 8+ days:
2.6 (-28.5, 47.1)
*p<0.05, tp<0.01,tp<0.10
December 2009
E-123
-------
Study
Reference: Sarnat et al. (2008,
0979721
Period of Study: Nov 1998-Dec 2002
Location: Atlanta (Georgia)
metropolitan area
Design & Methods
Outcome (ICD-9): Cardiovascular
disease ED visits: Ischemic heart
disease (41 0-41 4), cardiac
dysrhythmias (427), congestive heart
failure (428), and peripheral vascular
and cerebrovascular disease (433-437,
440, 443-444, 451-453)
Age Groups: All
Study Design: Time series
N: >4.5 million emergency department
wjcjtc
VIolLo
Statistical Analyses: Poisson
generalized linear models
Covariates: Day of the week, holidays,
hospital, long-term trends, temperature,
dew point temperature
Season: All, warm season (Apr 15-Oct
14), and cool season (Oct 15-Apr 14).
Dose-response Investigated: No
Statistical Package: NR
Lags Considered: 0-day lag
Concentrations1
Pollutant: PM25
Averaging Time: 24 h
Mean (ug/m3) (median 10th-90th
percentile):
Total PM25:
Cool season: 15.8 (14.3
7.5-25.5).
Warm season: 18.2(17.0
9.1-29.0)
PM25 EC:
Cool: 1.7(1.4
0.6-3.3).
Warm: 1.4 (1.3
0.6-2.5)
PM25Zn(ng/m3):
Cool: 15.7 (11. 7
A fi "30. OA
4.0-oU.z)
Warm: 10.9 (8.5
-3 -5 on <-)\
o.o-zU.z)
PM25K(ng/m3):
Cool: 63.0 (53.9
24.3-114.2)
Warm: 52.7 (43.3
23.2-93.5)
PM25Si (ng/m3):
Cool: 67.7 (54.1
24.3-123.5).
Warm: 110.9 (89.0
32.9-186.3)
PM25S042":
Cool: 3.4 (0.6
1.5-5.8).
Warm: 6.0 (5.2
2.3-10.8)
PM25NOr:
Cool: 1.4 (1.2
0.5-2.6).
Warm: 0.7 (2.9
0.3-1.2)
PM25Se(ng/m3):
Cool: 1.4(1.1
0.4-3.0).
Warm1 1 2 (09
0.4-2.7) '
PM25OC:
Cool: 4.6 (3.9
1.9-8.0)
Warm: 4.0 (3.7
2.1-6.4)
Monitoring Stations: 1
Copollutants: NR
Effect Estimates (95% Cl)
PM Increment: IQR (specific values not
given)
Risk ratio [95% Cl]: CVD (Lag 0): All
seasons: Total PM25: 1.022 [1.007,
1.038]
PM25 EC: 1.02 [1.013-1.037]
PM2.5 zinc: 1.013 [1.005-1.022]
PM2.5 potassium: 1.030 [1.018-1.042]
PM25 silicon: 1.008 [1.00-1. 01 6]
PM25 sulfate: 1.007 [0.994-1.019]
PM25 nitrate: 1.002 [0.990-1.014]
PM25 selenium: 1.002 [0.991-1. 012]
PM25OC: 1.024 [1.013-1.035]
Cool season: Total PM25: 1.028
[1.012-1.044]
PM25 EC: 1.029 [1.015-1.044]
PM2.5 Zinc: 1.012 [1.002-1.022]
PM25K: 1.037 [1.021-1.054]
PM2 5 Si: 1.022 [1.002-1. 043]
PM2.5 sulfate: 1.014 [0.991-1.037]
PM25 nitrate: 1.006 [0.993-1.019]
PM25Se: 1.012 [0.997-1. 027]
PM25OC: 1.027 [1.013-1.040]
Warm season: Total PM25: 1.006
[0.990-1.022]
PM2.5 EC: 1.021 [1.000-1.043]
PM2.5 Zinc: 1.017 [1.002-1.033]
PM2 5 K: 1.024 [1.007-1. 041]
PM2.5 Si: 1.005 [0.996-1.014]
PM25 sulfate: 1.001 [0.988-1.015]
PM2 5 nitrate: 1.000 [0.969-1. 033]
PM25Se: 0.996 [0.981-1. 011]
PM25OC: 1.027 [1.004-1.051]
December 2009
E-124
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Schreuder et al. (2006,
0979591
Period of Study: Sep 1995-May 2002
Location: Spokane, WA
Outcome: Cardiac HA
Age Groups: NR
Study Design: Time series
Statistical Analyses: GAM Poisosn
Regression
Covariates: Season, temperature,
relative humidity, day of week
Dose-response Investigated? No
Statistical Package: S-Plus
Lags Considered: 0-1 day
Pollutant: PM25 (ng/rri)
Averaging Time: 24 h
Arithmetic Mean: 10,580
Geometric Mean: 8,790
Min: 930
Max: 43,230
IQR:
Entire period: 7.7 pg/m3
Heating season: 10.1|jg/m
Non-heating season: 5.5|jg/m3
Monitoring Stations: NR
Copollutant: NR
Co-pollutant Correlation: NR
PM Increment: Interquartile Range
Relative Risk (Lower Cl, Upper Cl):
Entire Period, Lag 0:1.008 (0.985,
1.032)
Entire Period, Lag 1:1.000 (0.978,
1.023)
Heating Season, Lag 1:1.015 (0.968,
1.063)
Non-Heating Season, Lag 1: 0.995
(0.969, 1.021)
December 2009
E-125
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Sullivan et al. (2005,
1094181
Period of Study: 1988-1994
Location: King County, Washington
Outcome: Acute Ml
Age Groups: All, <50, 50-59, 70+
Study Design: Case-crossover
N: 5793 cases of acute Ml (5793 case
days and 20,134 referent exposure
days from these case individuals)
Statistical Analyses: Conditional
logistic regression
Covariates: Relative humidity,
temperature, season, day of week
Season: All, and also conducted
stratified analysis by season of event
(heating season: Nov-Feb
nonheating season: Mar-Oct)
Dose-response Investigated: No
Statistical Package: SAS version 8.0
and SPSS version 10
Lags Considered: Lag 1 and Lag 2 for
24-h avg
Pollutant: PM25
Averaging Time:
1 h, 2 h, 4 h, and 24 h
Summary of PM2 51 h before Ml onset:
Mean (ug/m3) (median IQR, 90th
percentile range):
12.8(8.6
5.3-15.9
27.3
2.0-147)
Monitoring Stations: 3
Copollutant (correlation):
1-havg:
PMi0:r = 0.78
CO: r = 0.47
SO,: r = 0.16
PM Increment: 10 pg/m
Odds ratio [96% Cl]:
1-h Averaging Time: 1.01 [0.98,1.05]
2-h Averaging Time: 1.01 [0.97,1.05]
4-h Averaging Time: 1.02 [0.98,1.04]
24-h Averaging Time: 1.02 [0.98,1.07]
Association between PM25 (24 h)
lagged 1 or 2 days non-significant (data
not shown)
Season (1-h avg): Heating: 1.01
[0.98-1.05]
Nonheating: 0.99 [0.91-1.09]
Age (1-h avg): <50yr: 1.04 [0.95,1.14]
50-60 yr: 0.99 [0.94, 1.05]
70+yr: 1.03 [0.98,1.08]
Age (24-h avg): <50yr: 1.07 [0.98,1.19]
50-69 yr: 0.99 [0.93, 1.06]
70+yr: 1.04 [0.99,1.11]
Sex (1-h avg): Men: 1.02 [0.98,1.06]
Women: 1.00 [0.95,1.06]
Sex (24-h avg): Men: 1.03 [0.99,1.08]
Women: 1.00 [0.94,1.07]
Race (1-h avg): White: 1.01 [0.97,1.04]
Nonwhite: 1.06 [0.97,1.17]
Race (24-h avg): White: 1.01 [0.97,
1.06]
Nonwhite: 1.10 [0.99,1.23]
Smoking status (1-h avg): Current: 0.99
[0.93,1.06]
Nonsmoker: 1.03 [0.97,1.08]
Smoking status (24-h avg): Current:
0.99 [0.95,1.14]
Nonsmoker: 1.03 [0.98,1.09]
Survivor of Ml* (1-h avg): Yes: 1.02
[0.98, 1.06]; No: 0.96 [0.86, 1.08]
Survivor of Ml * (24-h avg): Yes: 1.03
[0.98, 1.07]; No: 0.97 [0.85, 1.10]
Previous congestive heart failure (1 h
avg): Yes: 1.06 [0.97,1.16]; No: 1.00
[0.97, 1.04]
Previous congestive heart failure (24-h
avg): Yes: 1.08 [0.97,1.2]; No: 1.00
[0.97, 1.04]
Previous Ml (1-h avg): Yes: 1.03 [0.97,
1.1]; No: 1.01 [0.96,1.06]
Previous Ml (24-h avg): Yes: 1.04 [0.97,
1.17]; No: 1.02 [0.98, 1.08]
Hypertension (1-h avg): Yes: 1.02 [0.97,
1.07]; No: 1.01 [0.96, 1.06]
Hypertension (24-h avg): Yes: 1.02
[0.97, 1.07]; No: 1.02 [0.97, 1.08]
Diabetes mellitus (1-h avg): Yes: 1.06
[0.98, 1.14]; No: 1.01 [0.97, 1.05]
Diabetes mellitus (24-h avg): Yes: 1.04
[0.95, 1.14]; No: 1.01 [0.97, 1.06]
'Compares those who survive
hospitalization (yes) with those who
died in hospital from complications of
December 2009
E-126
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Symons et al. (2006,
0912581
Period of Study: Apr-Dec 2002
Location: Baltimore, Maryland
Outcome: Congestive heart failure
Age Groups: All
Study Design: Case-crossover
N: 125 patients
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature and humidity
Season: NR
Dose-response Investigated: Yes
Statistical Package: SAS and S-Plus
Lags Considered: 0-3 days (single and
cumulative)
Pollutant: PM25
Averaging Time: 8 h & 24 h
Mean (min-max):
8h
17.0(0.1-111.9)
SD=12.7
24 h
16.0(3.5-69.2)
SD=10.0
Monitoring Stations: 8
Copollutant (correlation): NR
PM Increment: 9.2 pg/m3 (IQR)
RR Estimate [Cl]:
8 h (participant's onset period)
Same-day lag: 0.87 [0.69,1.09]
1-day lag: 0.96 [0.78,1.18
2-day lag: 1.09 [0.91,1.30
3-day lag: 0.99 [0.79,1.23]
Cumulative 1-day lag: 0.89
Cumulative 2-day lag: 0.99
0.67,1.16]
0.74,1.33]
Cumulative 3-day lag: 0.98 [0.70,1.36]
24 h avg
Same-day lag: 0.81 [0.65,1.01]
1-day lag: 0.90 [0.74,1.11]
2-day lag: 0.85 [0.68,1.07]
3-day lag: 0.86 [0.70,1.05]
Cumulative 1-day lag: 0.82
0.64,1.04]
0.57,1.01]
Cumulative 2-day lag: 0.76
Cumulative 3-day lag: 0.70 [0.51,0.97]
Notes: |3 coefficients presented in Fig 5
Reference: Tolbert et al. (2007,
0903161
Period of Study: Aug 1998-Dec 2004
Location: Atlanta Metropolitan area,
Georgia
Outcome (ICD-9):
Combined CVD group, including:
Ischemic heart disease (410-414),
cardiac dysrhythmias (427), congestive
heart failure (428), and peripheral
vascular and cardiovascular disease
(433-437, 440, 443-445, and 451-453)
Age Groups: All
Study Design: Time series
N:NR for 1998-2004.
For 1993-2004:10,234,490 ER visits
(283,360 and 1,072,429 visits included
in the CVD and RD groups,
respectively)
Statistical Analyses: Poisson
generalized linear models
Covariates: Long-term temporal trends,
season (for RD outcome), temperature,
dew point, days of week, federal
holidays, hospital entry and exit
Season: All
Dose-response Investigated: No
Statistical Package: SAS version 9.1
Lags Considered: 3-day maflag 0 -2)
Pollutant: PM25
Averaging Time: 24 h
Mean (ug/m3) (median IQR, range,
10th -90th percentiles):
PM25:17.1 (15.6
11.0-21.9
0.8-65.8
7.9-28.8).
PM25sulfate:4.9(3.9
2.4-6.2
0.5-21.9
1.7-9.5).
PM25OC:4.4(3.8
2.7-5.3
0.4-25.9
2.1-7.2).
PM25EC:1.6(1.3
0.9-2.0
0.1-11.9
0.6-3.0).
PM25 water-soluble metals: 0.030
(0.023
0.014-0.039
0.003-0.202
0.009-0.059)
Monitoring Stations: 1
Copollutant (correlation):
Between PM25and:
PM,0:r = 0.84
03:r = 0.62
N02:r = 0.47
CO: r = 0.47
S02:r = 0.17
PM10.25:r = 0.47
PM25S04:r = 0.76
PM25EC:r = 0.65
PM25OC:r = 0.70
PM25TC:r = 0.71
P M2 5 water-sol metals: r = 0.69
OHC:r = 0.50
Between PM25 S04 and: PM,0: r = 0.69
03:r = 0.56
N02:r = 0.14
CO: r = 0.14
S02:r = 0.09
PMi0.25:r = 0.32
PM25:r = 0.76
PM25EC:r = 0.32
PM25OC:r = 0.33
PM Increment:
PM25:10.96 pg/m3 (IQR)
PM25sulfate:3.82|jg/ms(IQR)
PM25total carbon: 3.63 pg/m3 (IQR)
PM25OC:2.61 pg/m3(IQR)
PM25EC:1.15|jg/m3(IQR)
PM25 water-soluble metals: 0.03 pg/m3
(IQR)
Risk ratio [95% Cl] (single pollutant
models):
PM25:
CVD: 1.005 [0.993-1.017]
PM25sulfate:
CVD: 0.999 [0.987-1.011]
PM25total carbon:
CVD: 1.016 [1.005-1.026]
PM25OC:
CVD: 1.015 [1.005-1.026]
PM25EC:
CVD: 1.015 [1.005-1.025]
PM2 5 water-soluble metals:
CVD: 1.009 [0.997-1.021]
Notes: Results of selected multi-
pollutant models for cardiovascular
disease are presented in Fig 1.
Fig 1: PM25 total carbon adjusted for
CO, N02, or N02+C0
Summary of results: PM2 5 total carbon
continued to have a positive, statistically
significant association with CVD after
adjustment for N02 but not after
adjustment
December 2009
E-127
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
PM25TC:r = 0.34
PM2 5 water-sol metals: r = 0.65
OHC:r = 0.47
Between PM2 5 EC and:
PM10:r = 0.61
03:r = 0.40
N02:r = 0.64
CO: r = 0.66
S02:r = 0.22
PMi0.25:r = 0.49
PM25:r = 0.65PM25
S04:r = 0.32
PM25OC:r = 0.82
PM25TC:r = 0.91
PM2 5 water-sol metals: r = 0.52
OHC:r = 0.35
Between PM25 OC and:
PMi0:r = 0.65
03:r = 0.54
N02:r = 0.62
CO: r = 0.59
S02:r = 0.17
PM10.25:r = 0.49
PM25:r = 0.70
PM25S04:r = 0.33
PM25EC:r = 0.82
PM25TC:r = 0.98
P M2 5 water-sol metals: r = 0.49
OHC:r = 0.37
Between PM25 total carbon and:
PM10:r = 0.67
03:r = 0.52
N02:r = 0.65
CO: r = 0.63
S02:r = 0.19
PM10.25:r = 0.51
PM25:r = 0.71
PM25S04:r = 0.34
PM25EC:r = 0.91
PM25OC:r = 0.98
PM2 5 water-sol metals: r = 0.52
OHC:r = 0.38
Between PM25 water-soluble metals
and: PM,0:r = 0.73
03:r = 0.43
N02:r = 0.32
CO: r = 0.35
S02:r = 0.06
PMi0.25:r = 0.50
PM25:r = 0.69
PM25S04:r = 0.65
PM25EC:r = 0.52
PM25OC:r = 0.49
PM25TC:r = 0.52
December 2009 E-128
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Villeneuve et al. (2006,
0901911
Period of Study: Apr 1992-Mar 2002
Location: Edmonton, Canada
Outcome (ICD-9): Stroke (430-438),
including ischemic stroke (434-436),
hemorrhagic stroke (430,432), and
transient ischemic attacks (TIA) (435).
Age Groups: 65+ yr
Study Design: Case-crossover
N: 12,422 visits
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature and relative
humidity
Season: Summer (Apr-Sep), winter
(Oct-Mar)
Dose-response Investigated: No
Statistical Package: SAS (PHREG)
Lags Considered: 0,1, and 3 days
Pollutant: PM25
Averaging Time: 24 h
Mean ug/m3 (SD):
All yr: 8.5 (6.2)
Summer: 8.7 (7.1)
Winter: 8.3 (5.2)
Monitoring Stations: 3
Copollutant (correlation):
Allyr
S02:r = 0.22
N02:r = 0.41
CO: r = 0.43
03-mean:r = -0.07
Os-max: r = 0.07
PMi0:r = 0.79
Summer
S02:r = 0.20
N02:r = 0.52
CO: r = 0.42
Oj-mean: r = 0.11
03-max: r = 0.34
PM10:r = 0.85
Winter
S02:r = 0.28
N02:r = 0.57
CO: r = 0.71
Os-mean: r = -0.45
Os-max: r = -0.35
PM10:r = 0.70
PM Increment: pg/m (IQR)
All yr: 6.3
Summer: 6.5
Wnter: 6.0
Adjusted OR Estimate [Cl]:
Acute ischemic stroke
Allyr: Same-day lag: 1.00 [0.96,1.04]
1-day lag: 1.00 [0.96,1.05]
3-day lag: 1.01 [0.96,1.06]
Summer: Same-day lag: 0.96
[0.90,1.03]
1-day lag: 1.01 [0.94,1.07]
3-day lag: 0.98 [0.89 [1.07]
Wnter: Same-day lag: 1.04 [0.99,1.10]
1-day lag: 1.01 [0.96,1.07]
3-day lag: 1.05 [0.98,1.13]
Hemorrhagic stroke
All yr: Same-day lag: 0.99 [0.90,1.08]
1-day lag: 1.07 [0.98,1.16]
3-day lag: 1.05 [0.93,1.19]
Summer: Same-day lag: 0.99
[0.86,1.15]
1-day lag: 1.12 [0.97,1.30]
3-day lag: 1.08 [0.88,1.31]
Wnter: Same-day lag: 1.04 [0.92,1.18]
1-day lag: 1.08 [0.97,1.20]
3-day lag: 1.11 [0.94,1.31]
Transient cerebral ischemic attack
All yr: Same-day lag: 0.98 [0.93,1.03]
1-day lag: 0.99 [0.95,1.04]
3-day lag: 0.96 [0.90,1.03]
Summer: Same-day lag: 1.00
[0.92,1.08]
1-day lag: 1.03 [0.95,1.12]
3-day lag: 0.98 [0.88,1.09]
Wnter: Same-day lag: 0.97 [0.90,1.05]
1-day lag: 0.97 [0.91,1.04]
3-day lag: 0.94 [0.86,1.03]
Notes: Adjusted ORs are provided for
an IQR increase in the 3-day mean in
Fig 1-4 for single and two-pollutant
models.
Reference: Zanobetti and Schwartz
(2006, 0901951
Period of Study: 1995-1999
Location: Boston Metropolitan area
Outcome (ICD-9): Myocardial infarction
(410) or pneumonia (480-487)
Age Groups: 65+ yr
Study Design: Case-crossover
N: 15,578 patients admitted for Ml and
25,857 admitted for pneumonia
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature, day of the
week.
Season: All, and also tested for
interaction by warm (Apr-Sep) vs.. cold
season
Dose-response Investigated: No
Statistical Package: SAS version 8.2
(PROC PHREG)
Lags Considered: Lag 0, and mean of
lags 0 -1
Pollutant: PM25
Averaging Time: 24 h
Median (ug/m3) (IQR 5th-95th
percentile):
11.1 (7.23-16.14
3.87-26.31)
Monitoring Stations: 1
Copollutant (correlation):
BC:r = 0.66
N02:r = 0.55
CO: r = 0.52
03:r = 0.20
PM non-traffic: r = 0.74
PM Increment: Difference between the
90th and 10th percentile for PM2 5
Myocardial infarction cohort (Lag 0):
17.17 pg/m3
Myocardial infarction cohort (Lag 0-1):
16.32 pg/m3
Pneumonia cohort (Lag 0): 17.14 pg/m3
Pneumonia cohort (Lag 0): 16.32 pg/m3
Percentage (%) increase in risk [95%
Cl]:
Myocardial infarction cohort:
Lag 0:8.50 (1.89-14.43)
Lag 0-1: 8.65 (1.22-15.38)
Pneumonia cohort:
Lag 0:6.48 (1.13-11.43)
Lag 0-1: 5.56 (-0.45, 11.27)
Notes: Assessed for effect modification
by season. Results are reported in Fig
2. Summary of results: PM25 is
associated with pneumonia
hospitalization in the cold season but
not the hot season. PM2 5 is associated
with Ml hospitalization in the hot season
but not the cold season.
December 2009
E-129
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Zanobetti and Schwartz
(2006, 0901951
Period of Study: 1995-1999
Location: Boston Metropolitan area
Outcome (ICD-9): Myocardial infarction
(410) or pneumonia (480-487)
Age Groups: 65 + yr
Study Design: Case-crossover
N: 15,578 patients admitted for Ml and
25,857 admitted for pneumonia
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature, day of the
week.
Season: All, and also assessed for
interaction by hot (Apr-Sep) vs.. cold
season
Dose-response Investigated: No
Statistical Package: SAS Software
Release 8.2
Lags Considered: Lag 0, and mean of
lags 0 -1
Pollutant: BC
Averaging Time: 24 h
Median (ug/m3) (IQR 5th-95th
percentiles):
1.15(0.74-1.72
0.42-2.83)
Monitoring Stations: 1
Copollutant (correlation):
PM25:r = 0.66
N02:r = 0.70
CO: r = 0.82
03:r = -0.25
PM non-traffic: r = -0.01
PM Increment: Difference between the
90th and 10th percentile for BC
Myocardial infarction cohort (Lag 0):
2.01 pg/m3
Myocardial infarction cohort (Lag 0-1):
1.69|jg/m3
Pneumonia cohort (Lag 0): 2.05 pg/m3
Pneumonia cohort (Lag 0 -1):
1.69|jg/m3
Percentage (%) increase in risk [95%
Cl]:
Myocardial infarction cohort:
Lag 0:6.98 (-0.27-13.76)
Lag 0-1: 8.34 (0.21-15.82)
Pneumonia cohort:
|Lag 0:10.76 (4.54-15.89)
Lag 0-1:11.71 (4.79, 17.36)
Notes: Assessed for effect modification
by season. Results are reported in Fig
2. Summary of results: PM2.BC is
associated with pneumonia
hospitalization in the cold season but
not the hot season. BC had a stronger
positive association with Ml
hospitalization in the cold season, but
the confidence interval was wide.
All units expressed in pg/m unless otherwise specified.
Table E-8. Short-term exposure-cardiovascular-ED/HA-other size fractions.
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Andersen et al, (2008,
1896511
Period of Study: May 2001-Dec 2004
Location: Copenhagen, Denmark
Outcome (ICD-10): CVD, including
angina pectoris (I20), myocardial
infarction (121-22), other acute ischemic
heart diseases (I24), chronic ischaemic
heart disease (I25), pulmonary
embolism (I26), cardiac arrest (I46),
cardiac arrhythmias (148-48), and heart
failure (ISO).
RD, including chronic bronchitis
(J41-42), emphysema (J43), other
chronic obstructive pulmonary disease
(J44), asthma (J45), and status
asthmaticus (J46).
Pediatric hospital admissions for
asthma (J45) and status asthmaticus
(J46).
Age Groups: >65 yr (CVD and RD),
5-1 Syr (asthma)
Study Design: Time series
N:NR
Statistical Analyses: Poisson GAM
Covariates: Temperature, dew-point
temperature, long-term trend,
seasonality, influenza, day of the week,
public holidays, school holidays (only
for 5-18 yr olds), pollen (only for
pediatric asthma outcome)
Season: NR
Pollutant: Total number concentration
of ultrafme and accumulation mode
particles (NCtot) (particles/cm3)
Averaging Time: 24 h
NCtotal
Mean(SD): 8,116(3502)
Median: 7,358
IQR: 5,738-9,645
99th Percentile: 19,895
NCa12
Mean (SD): 493 (315)
Median: 463
IQR: 308-650
99th Percentile: 1,263
NCa23
Mean (SD): 2,253 (1,364)
Median: 2,057
IQR: 1,280-3,066
99th Percentile: 6,096
NCa67
Mean (SD): 5,104 (2,687)
Median: 4,562
IQR: 3,248-6,274
99th Percentile: 14,410
NCalOO
Mean (SD): 6,847 (2,846)
Median: 6,243
IQR: 4,959-8,218
99th Percentile: 16,189
NCa212
PM Increment: IQR increase in
pollutant level: Nctot: 3907 particles/cm
(IQR)
Ncai2:342 particles/cm3 (IQR)
Nca23:1786 particles/cm* (IQR)
Nca57:3026 particles/cm3 (IQR)
NCIOO: 3259 particles/cm (IQR)
Nca2i2: 495 particles/cm3 (IQR)
Relative risk (RR) Estimate [Cl]: CVD
hospital admissions (4-day avg, lag 0 -
3), age 65+
One-pollutant model (NCtot):
1.00 [0.99-1.02]
Adj for PM,0: 0.98 [0.96-1.01]
Adj for PM25: 0.99 [0.95-1.03]
Adj for CO: 0.99 [0.97-1.02]
Adj for N02:1.01 [0.98-1.03]
Adj for 03:1.01 [0.96-1.06]
One-pollutant model (NC100):
1.00 [0.98-1.02]
One pollutant model (Nca12):
0.99 [.97-1.01]
Adj for other size fractions:
0.99 [0.97-1.02]
One pollutant model (Nca23):
0.99 [0.96-1.01]
Adj for other size fractions:
0.99 [0.96-1.02]
One pollutant model (NCa57):
1.01 [0.98-1.02]
Adj for other size fractions:
0.99 [0.97-1.02]
One pollutant model (Nca212):
1.02 [1.00-1.04]
December 2009
E-130
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Dose-response Investigated: No
Statistical Package: R statistical
software (gam procedure, mgcv
package)
Lags Considered: Lag 0 -5 days, 4-
day pollutant avg (lag 0 -3) for CVD, 5-
day avg (lag 0-4) for RD, and a 6-day
avg (lag 0-5) for asthma.
Mean (SD): 392 (441)
Median: 246
IQR: 89-584
99th Percentile: 2,248
*NC, number concentration tot, total (all
particles 6-700 in diameter) a12, size
mode with mean diameter of 12 nm
a23, size mode with median diameter of
23 nm
a57, size mode with median diameter of
57 nm a212 size mode with median
diameter of 212 nm
NC100 = a12+a23+0.797*a57+0.084*a
212.
Monitoring Stations: 1
Copollutant (correlation):
Correlation of NCtot with:
PM,o:r = 0.39
PM25:r = 0.40
N02:r = 0.68
:r = 0.66
NCi00:r = 0.98
NCa12:r = 0.31
NCa23:r = 0.57
NCa57:r = 0.87
NCa212:r = 0.29
CO: r = 0.54
N0xcurbside:r = 0.36
03:r = -0.12
Other variables:
Temperature: r = -0.06
Relative humidity: r = -0.04
Adj for other size fractions:
1.02 [1.00-1.05]
Adj for PM10: 0.98 [0.95-1.01]
RD hospital admissions (5-day avg, lag
0 -4), age 65+: One-pollutant model:
1.04 [1.00-1.07]
Adj for PM,0:1.00 [0.96-1.05]
Adj for PM25: 0.97 [0.89-1.05]
Adj for CO: 1.03 [0.98-1.07]
Adj for N02:1.00 [0.95-1.05]
Adj for 03: 0.95 [0.87-1.04]
One pollutant model (NC100):
1.03 [0.99-1.07]
One pollutant model (Nca12):
1.01 [0.98-1.05]
Adj for other size fractions:
1.01 [0.97-1.05]
One pollutant model (Nca23):
0.99 [0.94-1.03]
Adj for other size fractions:
0.98 [0.94-1.03]
One pollutant model (Nca57):
1.04 [1.00-1.08]
Adj for other size fractions:
1.02 [0.97-1.06]
One pollutant model (Nca212):
1.04 [1.01-1.08]
Adj for other size fractions:
1.03 [0.99-1.07]
Adj for PM10:1.01 [0.96-1.07]
Asthma hospital admissions (6-day avg
lag 0-5), age 5-18: One-pollutant model:
1.07 [0.98-1.17]
Adj for PM10:1.03 [0.92-1.15]
Adj for PM25:1.04 [0.85-1.28]
Adj for CO: 1.09 [0.99-1.21]
Adj for N02:1.07 [0.96-1.19]
Adj for 03:1.08 [0.87-1.35]
One pollutant model (NC100):
1.06 [0.97-1.16]
One pollutant model (Nca212):
1.08 [0.99-1.18]
Adj for other size fractions:
1.07 [0.97-1.19]
One pollutant model (Nca23):
1.09 [0.98-1.21]
Adj for other size fractions:
1.08 [0.97-1.21]
One pollutant model (Nca57):
1.02 [0.94-1.12]
Adj for other size fractions:
0.93 [0.83-1.04]
One pollutant model (Nca212):
1.08 [1.00-1.17]
Adj for other size fractions: 1.12
[1.02-1.23]
Adj for PM10:1.10 [0.96-1.13]
Notes: Fig 2: Relative risks and 95%
confidence intervals per IQR in single
day concentration (0-5 day lag).
Summary of Fig 2: CVD: Positive,
marginally or statistically significant
associations at Lag 2 (Nctot, Nca57,
Nca212), Lag3 (Nca212), and Lag 1
(Nca212). RD: Positive, statistically or
marginally significant associations at
Lag 4 (Nctot, Nca57, NCa212) and Lag
5 (Nctot, Nca57, Nca212), and to a
lesser extent Lag 2 (Nctot, Nca212) and
Lag 3 (Nctot, Nca212). Asthma: Wide
confidence intervals make interpretation
difficult. Positive, significant association
forNca212atLagl
December 2009
E-131
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Lanki et al. (2006, 0897881
Period of Study: 1992-2000
Location: Augsburg, Barcelona,
Helsinki, Rome, and Stockholm
Outcome (ICD-9): Acute myocardial
infarction (410
ICD-10:I21, I22)
Age Groups: 35+ yr, <75 yr, 75+ yr
Study Design: Time series
N: 26,854 hospitalizations
Statistical Analyses: GAM
Covariates: Temperature, barometric
pressure
Season: V\ferm (Apr-Sep) and cold
(Oct-Mar)
Dose-response Investigated: No
Statistical Package: R package mgcv
0.9-5
Lags Considered: 0-3 days
Pollutant: UFP (PNC)
Averaging Time: 24 h
Median particles/cm3:
Augsburg: 12,400
Barcelona: 76,300
Helsinki: 13,600
Rome: 46,000
Stockholm: 11,800
Copollutant (correlation):
Augsburg
PM,o:r = 0.53
CO: r = 0.63
N02:r = 0.65
03:r = 0.26
Barcelona:
PM,o:r = 0.38
CO: r = 0.80
N02:r = 0.49
03:r = -0.35
Helsinki:
PM,0:r = 0.45
CO: r = 0.48
N02:r = 0.82
03:r = 0.01
Rome:
PM,0:r = 0.32
CO: r = 0.83
N02:r = 0.68
03:r = 0.03
Stockholm:
PMi0:r = 0.06
CO: r = 0.56
N02:r = 0.83
03:r = -0.01
PM Increment: 10,000 particles/cm
Pooled Rate Ratio [Cl]: All 5 cities
(35+ yr)
Same-day lag: 1.005 [0.996,1.015]
1-day lag: 0.997 [0.982,1.012
2-day lag: 0.999 [0.990,1.008
3-day lag: 0.998 [0.979,1.017]
3 cities with hospital discharge register
(35+ yr)
Same-day lag: 1.013 [1.000,1.026]
1-day lag: 0.995 [0.953,1.039
2-day lag: 1.001 [0.989,1.014
3-day lag: 1.009 [0.974,1.046]
V\ferm season (35+ yr)
Same-day lag: 1.009 [0.972,1.048]
1-day lag: 1.023
2-day lag: 1.050
0.988,1.060
1.016,1.085
3-day lag: 1.022 [0.987,1.C
Cold season (35+ yr)
Same-day lag: 1.014 [1.001,1.028]
1-day lag: 1.001 [0.956,1.048
2-day lag: 1.001 [0.989,1.014
3-day lag: 1.009 [0.971,1.049]
Age >75Non-fatal
Same-day lag: 1.032 [1.008,1.056]
1-day lag: 1.009 [0.985,1.032
2-day lag: 0.989 [0.966,1.013
3-day lag: 1.009 [0.969,1.051]
Fatal
Same-day lag: 1.016 [0.978,1.055]
1-day lag: 1.001 [0.966,1.038
2-day lag: 1.005 [0.969,1.041
3-day lag: 0.984 [0.948,1.021]
Notes: Rate ratios for PNC are given
for 0-5 lag days in graph form (Fig 1) for
each city. Pooled rate ratios were also
provided for groups <75 yielding similar
results to the overall 3-city data.
Reference: Metzger et al. (2004,
0442221
Period of Study: Aug 1998-Aug 2000
Location: Atlanta Metropolitan area
(Georgia)
Outcome (ICD-9): Emergency visits for
ischemic heart disease (410-414),
cardiac dysrhythmias (427), cardiac
arrest (427.5), congestive heart failure
(428), peripheral vascular and
cerebrovascular disease (433-437, 440,
443-444, 451-453), atherosclerosis
(440), and stroke (436).
Age Groups: All
Study Design: Time series
N: 4,407,535 emergency department
visits between 1993-2000 (data not
reported for 1998-2000)
Statistical Analyses: Poisson
generalized linear modeling
Covariates: Day of the week, hospital
entry and exit indicator variables,
federally observed holidays, temporal
trends, temperature, dew point
temperature
Season: All
Dose-response Investigated: No
Statistical Package: SAS
Lags Considered: 3-day ma, lags 0-7
Pollutant: UFP (10-100 nm particle
count) (no/cm3)
Averaging Time: 24 h
Median (10%-90% range): 25,900
(11,500-74,600)
Monitoring Stations: 1
Copollutant (correlation):
PM,0:r = -0.13
03:r = -0.13
N02:r = 0.26
CO: r = 0.10
S02:r = 0.24
PM25:r = -0.16
PM2 5 water soluble metals: r = -0.27
PM25sulfates:r = -0.31;
PM25 acidity: r = -0.39;
PM25OC:r = 0.08;
PM25EC:r = 0.08;
PM25 oxygenated hydrocarbon: r = 0.05
Other variables:
Temperature: r = -0.33
Dew point: r = -0.41
PM Increment: 30,000 no/cm
(approximately 1 SD)3
RR [95% Cl]: For 3 day ma: All CVD:
0.985 [0.965, 1.005]
Dysrhythmia: 0.972 [0.937,1.008]
Congestive heart failure: 0.983
[0.943-1.025]
Ischemic heart disease: 0.989
[0.953-1.026]
Peripheral vascular and
cerebrovascular disease: 0.998
[0.960-1.039]
Results for Lags 0-7 expressed in
figures (see notes).
Notes: Fig 1: RR (95% Cl) for single-
day lag models for the association of
ER visits for CVD with daily ambient
UFP.
Summary of Fig 1 results: Null or
negative associations.
December 2009
E-132
-------
Study
Reference: von Klot et al. (2005,
0880701
Period of Study: 1992-2001
Location:
Augsburg, Germany
Barcelona, Spain
Helsinki, Finland
Rome, Italy
Stockholm, Sweden
Design & Methods
Outcome (ICD-9): Acute myocardial
infarction (410
ICD-10: 121-122), angina pectoris (411,
413
ICD-1 0:120, 124), dysrhythmia (427
ICD-10: 146.0, 46.9, I47-I49, ROD. 1,
R00.8), heart failure (428
ICD-10: 150)
Age Groups: 35+ yr
Study Design: Cohort
N: 22,006 Ml survivors
Statistical Analyses: GAM, Spearman
correlation
Covariates: Temperature, dew point
temp, avg barometric pressure, relative
humidity
Season: NR
Dose-response Investigated: No
Statistical Package: R-software with
"mgcv" package
Lags Considered: 0-3 days
Concentrations'!
Pollutant: UFP (PNC)
Averaging Time: 24 h
Mean particle/cm3 (6th-96th
percent! le):
Augsburg:
Barcelona:
Helsinki:
Rome:
Stockholm:
Monitoring Stations: NR
Copollutant (correlation):
Augsburg
PM,0:r = 0.52
CO: r = 0.63
N02' r = 064
03: r =-0.32
Barcelona
PM • r - n 9Q
riviio. r - u.zy
CO: r = 0.71;
N02:r = 0.44
03: r =-0.55
Helsinki
PM10' r = 046
CO: r = 0.47;
N02:r = 0.83
03:r=-0.16
Rome
PM,0:r = 0.33
CO: r = 0.80;
N02:r = 0.71
03: r =-0.47
Stockholm
PMi0:r = 0.06
CO: r = 0.54;
N02:r = 0.80
03:r=-0.17
Effect Estimates (95% Cl)
PM Increment: 10,000 particles/cm3
Pooled RR Estimate [Cl]:
All cardiac admissions: 1.026
[1.005,1.048]
Myocardial infarction: 1.039
[0.998.1.082]
Angina pectoris: 1.020 [0.992,1.048]
All units expressed in pg/m unless otherwise specified.
December 2009
E-133
-------
E.2. Short-Term Exposure and Respiratory Outcomes
E.2.1. Respiratory Morbidity Studies
Table E-9. Short-term exposure-respiratory morbidity outcomes -PMio.
Study
Reference: Aekplakorn, et al. (2003,
0899081
Period of Study: 107 days,
from Oct1997-Jan 1998
Location: Mae Mo district, Lampang
Province, North Thailand
Design & Methods
Outcome: Upper respiratory symptoms,
lower respiratory symptoms, cough
Age Groups: 6-1 4 yr old
Study Design: Logistic regression
N: 98 asthmatic school children, 98
non-asthmatic school children
Statistical Analyses: GEE, stratified
analysis, PROCGENMOD
Covariates: Temperature and relative
humidity
Season: Winter
Dose-response Investigated? No
Statistical Package: SASv 8.1
Concentrations!
Pollutant: PM,0
Averaging Time: Daily
Mean (SD):
Sob Pad station: 31. 92
Sob Mo station: 33.64
Hua Fai station: 37.45
Range (Min, Max):
Sob Pad: 6.63, 153.25
Sob Mo: 4.23, 121.80
Hua Fai: 6.98, 113.30
Monitoring Stations: 3
Copollutant: PM25, S02
Effect Estimates (95% Cl)
PM Increment: 10 pg/m3
Odds Ratios [Lower Cl, Upper Cl]
Ian1
Asthmatics: URS: 1.03 (0.99, 1.07)
lagO
LRS: 1.04 (0.99, 1.09)
lagO
Cough: 1.04 (1.00, 1.07)
lagO
Non-Asthmatics: URS: 1.04 (0.99, 1.08)
lagO
LRS: 1.0 (0.93, 1.07)
lagO
Cough: 0.99 (0.94, 1.05)
lagO
PM10 + S02
Asthmatics: URS: 1.03 (0.99, 1.07)
lagO
LRS: 1.03 (0.98, 1.09)
lagO
Cough: 1.04 (1.00, 1.08)
lagO
Non-Asthmatics: URS: 1.04 (0.99, 1.08)
lagO
LRS: 1.0 (0.93, 1.07)
lagO
Cough: 0.99 (0.95, 1.05)
lagO
December 2009
E-134
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Andersen et al. (2008,
1896511
Period of Study:
Dec 1998-Dec 2004
Location: Copenhagen, Denmark
Outcome: Daily symptoms (prospective
daily recording of symptoms via diary)
Age Groups: 0-3 yr
Study Design: Panel study of children
with genetic susceptibility to asthma
(mothers had asthma)
N: 205 children (living within a 15km
radius of the central monitor during the
first Syr of life)
born between Aug 2, 1998 and Dec 12,
2001
Statistical Analyses: Logistic
regression model (GEE)
Covariates: Temperature, season,
gender, age, exposure to smoking, and
paternal history of asthma
Effect modification: gender, medication
use, and paternal history of asthma
Statistical Package: SASv91
Lag: 0,1, 2,3,4,2-4
Pollutant: PM,0
Mean: 25.1
SD: 16.7
Percentiles:
25th: 15.7
75th: 30.2
IQR: 14.5
Copollutant (correlation):
PM25(r = 0.79)
Number concentration of ultrafine
particles,
UFP(r = 0.37)
N02(r = 0.43)
N0x(r = 0.40)
CO (r = 0.45)
03 (r = -0.32)
Temp (r = 0.25)
PM Increment: IQR (14.5 pg/m3)
increase
Odds Ratios (95%CI) for incident
wheezing symptoms
Age 0-1
10:1.05(0.88,1.25)
11:1.00(0.82,1.22)
L2:1.01 (0.83, 1
13:1.20(0.98,1
L4: 1.23(1.02, 1
L2-4: 1.21 (0.99,
Age 1-2
LO' 1 00 (0 86 1
L1 : 1.02 (0.8?! 1
12:1.05(0.93,1
L3: 0.96 (0.84, 1
14:1.04(0.90,1
12-4:1.03(0.88,
Age 2-3
LO: 0.87 (0.72, 1
L1: 0.95 (0.78, 1
L2: 0.99 (0.82, 1
L3: 1.03(0.84,1
L4: 0.89 (0.74, 1
L2-4: 0.94 (0.74,
Age 0-3
LO: 0.97 (0.87, 1
L1: 0.99(0.89,1
L2: 1.01(0.92,1
L3: 1.03 (0.93, 1
L4: 1.04 (0.94, 1
L2-4: 1.04 (0.92,
23
46
48)
1.48)
15
19
19)
09
21
1.22)
06)
15
17
25)
09)
1.19)
08
10
12)
14
15
1.17)
Two pollutant models (lag 2-4)
1-pollutant model: 1.21 (0.99,1.48)
2-pollutant (adj for N02): 1.13 (0.88,
1.45)
2-pollutant (adj for): 1.16 (0.90,1.48)
2-pollutant (adj for CO): 1.23 (0.96,
1.57)
110 children living within 5km radius
from monitor (sensitivity analysis): Age
0-1:1.32(0.95,1.82)
Age 1-2:1.20 (0.87, 1.67)
Age 2-3: 0.78 (0.52, 1.16)
Age 0-3:1.11 (0.88,1.39)
December 2009
E-135
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Boezen et al. (2005,
0873961
Period of Study: Two consecutive
winters (winter 1993-winter 1995)
Location: Rural (Meppel, Nunspeet)
and urban (Amsterdam) areas in the
Netherlands
Outcome: FEVi, airway
hyperresponsiveness (AHR), serum
total IgE and daily data on lower
respiratory symptoms (LRS), upper
respiratory symptoms (URS), cough
and morning and evening peak
expiratory flow
Age Groups: 50-70 yr
Study Design: Case-control study
N: 327 patients
Statistical Analyses: Logistic
regression
Covariates: daily minimum
temperature, linear, quadratic and cubic
time trend, weekend/holidays, and
influenza incidence for the rural and
urban areas and two winters separately
Season: winter
Dose-response Investigated? No
Lags Considered: 0,1, 2, and 5-day
mean
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD):
Winter 93/94 Urban: 41.5
Winter 93/94 Rural: 44.1
Wnter 94/95 Urban: 31.1
Wnter 94/95 Rural: 26.6
Percentiles: SOth(Median):
Wnter 93/94 Urban: 34.6
Wnter 93/94 Rural: 30.4
Wnter 94/95 Urban: 28.9
Wnter 94/95 Rural: 23.7
Range (Min, Max):
93/94 Urban: (12.1-112.7)
93/94 Rural: (7.9-242.2)
94/95 Urban: (8.8-89.9)
94/95 Rural: (7.1-96.9)
Copollutant:
S02
N02
BS
PM Increment: 10 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
AHFWIgE-
Upper Respiratory Symptoms
Lag 0: OR = 0.99 (0.97-1.01)
Lag 1: OR = 1.01 (0.99-1.03)
Lag 2: OR = 1.00 (0.96-1.02)
5-day mean: OR = 1.00 (0.96-1.04)
Cough
Lag 0: OR = 1.00 (0.99-1.02)
Lag 1: OR = 0.99 (0.98-1.01)
Lag 2: OR = 1.00 (0.98-1.01)
5-day mean: OR = 0.98 (0.95-1.01)
>10% fall in morning peak expiratory
flow
Lag 1: OR = 1.01 (0.98-1.04)
Lag 2: OR = 0.97 (0.94-1.00)
5-day mean: OR = 0.97 (0.92-1.02)
AHR-/lgE+
Upper Respiratory Symptoms
Lag 0: OR = 1.01 (0.99-1.03)
Lag 1: OR = 1.02 (1.00-1.04)
Lag 2: OR = 1.01 (0.99-1.03)
5-day mean: OR = 1.08 (1.04-1.11)
Cough
Lag 0: OR = 1.01 (0.99-1.03)
Lag 1: OR = 0.99 (0.98-1.01)
Lag 2: OR = 1.00 (0.98-1.02)
5-day mean: OR = 1.01 (0.97-1.05)
>10% fall in morning peak expiratory
flow
Lag 1: OR = 0.99 (0.97-1.02)
Lag 2: OR = 0.99 (0.97-1.02)
5-day mean: OR = 0.97 (0.93-1.01)
AHR+/lgE-
Upper Respiratory Symptoms
Lag 0: OR = 0.99 (0.95-1.03)
Lag 1: OR = 1.01 (0.97-1.05)
Lag 2: OR = 0.99 (0.96-1.03)
5-day mean: OR = 0.98 (0.91-1.06)
Cough
Lag 0: OR = 1.00 (0.97-1.02)
Lag 1: OR = 1.01 (0.98-1.03)
Lag 2: OR = 0.99 (0.96-1.02)
5-day mean: OR = 1.02 (0.96-1.08)
>10% fall in morning peak expiratory
flow
Lag 1: OR = 0.99 (0.95-1.03)
Lag 2: OR = 0.99 (0.95-1.03)
5-day mean: OR = 0.99 (0.93-1.06)
AHR+/lgE+
Upper Respiratory Symptoms
Lag 0: OR = 1.01 (0.98-1.04)
Lag 1: OR = 1.03 (1.00-1.05)
Lag 2: OR = 1.02 (0.99-1.05)
5-day mean: OR = 1.06 (1.00-1.11)
Cough
Lag 0: OR = 1.03 (1.01-1.06)
Lag 1: OR = 1.00 (0.98-1.02)
Lag 2: OR = 0.99 (0.97-1.01)
5-day mean: OR = 0.99 (0.95-1.04)
Lag 2: OR = 0.99 (0.96-1.03)
5-day mean: OR = 0.99 (0.92-1.05)
>10% fall in morning peak expiratory
flow
Lag 1: OR = 1.04 (1.00-1.07)
Lag 2: OR = 1.03 (0.99-1.06)
5-day mean: OR = 1.05 (0.99-1.11)
Reference: Boezen et al. (1999,
0404101
Periods of Study:
3 Wnters (1992-1995)
Location: Urban and rural areas of the
Netherlands
Outcome: Respiratory symptoms
Lower respiratory symptoms (wheeze,
attacks of wheezing, shortness of
breath)
Upper respiratory symptoms (sore
throat, runny or blocked nose)
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD):
Wnter 1992-93
Urban: 54.8
Rural: 44.7
Wnter 1993-94
Increment: 100|jg/m
Odds Ratio (Lower Cl, Upper Cl) lag:
OR for respiratory symptoms and
exposure to PMio in children with BHR
and high serum total IgE
Lower Respiratory Symptoms
1.32(1.07,1.63)0
December 2009
E-136
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Bronchial hyperresponsiveness (BHR)
Study Design: Time-series
Statistical Analyses: Logistic
regression (PROC model)
Age Groups: 7-11
Urban: 41.5 3
Rural: 44.1
Winter 1994-95
Urban: 31.1
Rural: 26.6
Range (Min, Max):
Winter 1992-93
Urban: (4.7,145.6)
Rural: (4.8,103.8)
Wnter 1993-94
Urban: (12.1,112.7)
Rural: (7.9, 242.2)
Wnter 1994-95
Urban: (8.8, 89.9)
Rural: (7.1, 96.9)
Co pollutants:
BS
S02
N02
1.36(1.13, 1.64) 1
1.36(1.13, 1.65)2
2.39 (1.71, 3.35) 0-5 avg.
Upper Respiratory Symptoms
1.13(0.97,1.32)0
1.00(0.87,1.16)1
0.96(0.84,1.11)2
0.91 (0.70, 1.18) 0-5 avg
>10% morning peak expiratory flow
(PEF) decrease
1.10(0.92,1.33)0
1.08(0.90,1.28)1
1.03(0.87,1.23)2
1.10(0.83, 1.46) 0-5 avg
>10% evening peak expiratory flow
(PEF) increase
1.37(1.16, 1.63)0
1.09(0.92,1.29)1
1.16(0.98.1.36)2
1.35(1.04, 1.77) 0-5 avg.
OR for respiratory symptoms and
exposure to PMi0 in children without
BHR and low serum total IgE
Lower Respiratory Symptoms
1.08(0.75,1.57)0
1.04(0.70, 1.53)1
0.98(0.69,1.39)2
1.15 (0.61, 2.15) 0-5 avg.
Upper Respiratory Symptoms
1.12(0.99,1.28)0
1.01(0.89,1.15)1
1.01 (0.89, 1.15)2
0.93(0.67, 1.28) 0-5 avg
>10% morning PEF decrease
1.07(0.93, 1.23)0
0.86 (0.75, 0.99) 1
0.97(0.85,1.11)2
0.99(0.79, 1.23) 0-5 avg
>10% evening PEF decrease
1.13(0.98,1.30)0
1.05(0.91, 1.21)1
0.99(0.87,1.14)2
0.94(0.75, 1.17) 0-5 avg
OR for respiratory symptoms and
exposure to PM10 in children with BHR
and low serum total IgE
Lower Respiratory Symptoms
0.77(0.48,1.24)0
1.34(0.94,1.93)1
1.24(0.86, 1.81)2
1.92(0.84, 4.41) 0-5 avg
Upper Respiratory Symptoms
1.13(0.92, 1.40)0
0.98(0.79,1.22)1
0.97(0.79,1.20)2
0.83(0.54, 1.25) 0-5 avg
>10% morning PEF decrease
1.04(0.78,1.38)0
0.86(0.66, 1.12)1
0.91(0.71,1.17)2
0.78(0.51, 1.20) 0-5 avg
>10% evening PEF decrease
1.07(0.82,1.41)0
0.98(0.76,1.26)1
0.93(0.73, 1.19)2
0.83(0.55, 1.26) 0-5 avg
OR for respiratory symptoms and
exposure to PMi0 in children without
BHR and high serum total IgE
Lower Respiratory Symptoms
1.04(0.80,1.35)0
1.21(0.98,1.51)1
1.18(0.96,1.45)2
1.35(0.89, 2.04) 0-5 avg
Upper Respiratory Symptoms
1.01(0.85,1.20)0
0.95(0.81, 1.12)1
0.93(0.80,1.09)2
December 2009
E-137
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
0.93(0.69, 1.25) 0-5 avg
>10% morning PEF decrease
0.97(0.80,1.17)0
1.09(0.91,1.30)1
1.02(0.85, 1.21)2
0.95(0.71, 1.28) 0-5 avg
>10% evening PEF decrease
1.02(0.85, 1.22)0
1.06(0.90,1.25)1
1.08(0.93,1.27)2
1.04(0.80, 1.34) 0-5 avg.
Reference: Chattopadhyay et al. (2007, Outcome: pulmonary function tests
(respiratory impairments)
147471
Period of Study: NR
Location: Three different points in
Kolkata, India: North, South, and
Central
Age Groups: All ages
Study Design: Cross-sectional
N: 505 people studied for PFT
total population of Kolkata not given
Statistical Analyses: Frequencies
Covariates: Meteorologic data (i.e.
temperature, wind direction, wind
speed, and humidity)
Dose-response Investigated? No
Pollutant: PM,0
Averaging Time: 8 h
Mean (SD):
North Kolkata: 535.9
Central Kolkata: 1114.5
South Kolkata: 909.2
Monitoring Stations: 1
Copollutant:
PM<10-3.3
PM<3.3-0.4
PM Increment: NR
Respiratory impairments (SD):
North Kolkata
Male (n = 137)
Restrictive: 4 (2.92)
Obstructive: 5 (3.64)
Combined Res. And Obs.: 6 (4.37)
Total: 15 (10.95)
Female (n = 152)
Restrictive: 3 (1.97)
Obstructive: 5 (3.28)
Combined Res. And Obs.: N/A
Total: 8 (5.26)
Total (n = 289)
Restrictive: 7 (2.42)
Obstructive: 10 (3.46)
Combined Res. And Obs: 6 (2.07)
Total: 23 (7.96)
Central Kolkata
Male (n = 44)
Restrictive: 6 (13.63)
Obstructive: 1 (2.27)
Combined Res. And Obs.:1 (2.27)
Total: 8 (18.18)
Female (n = 50)
Restrictive: 3 (6.00)
Obstructive: 2 (4.00)
Combined Res. And Obs.: N/A
Total: 5 (10.00)
Total (n = 94)
Restrictive: 9 (9.57)
Obstructive: 3 (3.19)
Combined Res. And Obs.: 1 (1.06)
Total: 13 (13.82)
South Kolkata
Male (n = 52)
Restrictive:! (1.92)
Obstructive: 2 (3.84)
Combined Res. And Obs.: 3 (5.76)
Total: 6 (11.53)
Female (n = 70)
Restrictive: 2 (2.85)
Obstructive:! (1.42)
Combined Res. And Obs.: N/A
Total: 3 (4.28)
Total (n = 122)
Restrictive: 3 (2.45)
Obstructive: 3 (2.45)
Combined Res. And Obs.: 3 (2.45)
Total: 9 (7.37)
December 2009
E-138
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Dales et al. (2006, 0907441
Period of Study: Jan 1986-Dec 2000
Location: 11 Canadian Cities: Calgary,
Edmonton, Halifax, London, Hamilton,
Ottawa, St. John, Toronto, Vancouver,
Windsor, Winnipeg
Health Outcome: Respiratory Illness:
Asphyxia (799)
Respiratory failure (799.1)
Dyspnea and respiratory abnormalities
(786)
Respiratory distress syndrome (769)
Unspecified birth asphyxia in live-born
infant (768.9)
Other respiratory problems after birth
(770.8)
Pneumonia (486)
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups: 0-27 days
Pollutant: PM,0
Averaging Time: 24-h avg
Copollutants (correlation):
03:r =-0.29 to 0.41
N02:r =-0.26 to 0.69
S02:r =-0.09 to 0.61
CO: r = -0.13 to 0.71
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl) Lag:
In respiratory illness and exposure to
PMio in neonates
PM10 alone: 2.13 (-0.50, 4.76)
Multipollutant model
PM10:1.45 (-1.90, 4.80)
PM,o, 03: 2.67 (0.98, 4.39)
PM,o, N02: 2.48 (1.18, 3.80)
PM10, S02:1.41 (0.35,2.47)
PM,o, 00:1.30(0.13, 2.49)
Reference: de Hartog et al. (2003,
0010611
Period of Study: Wnter of 1998-1999
Amsterdam, from Nov 1998 to Jun 1999
Erfurt, from Oct 1998 to Apr 1999
Helsinki, from Nov 1998 to Apr 1999
Location:
Amsterdam, the Netherlands;
Erfurt, Germany; Helsinki, Finland
Outcome: Respiratory symptoms
Age Groups: > 50 yr
Study Design: Panel
N: 131 subjects with history of coronary
heart disease
Statistical Analyses: Logistic
regression
Covariates: Ambient temperature,
relative humidity, atmospheric pressure,
incidence of influenza-like illness
Season: Wnter
Dose-response Investigated? No
Statistical Package: S-PLUS 2000
Lags Considered: 0-, 1-, 2-, 3-, and
5-day avg
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD):
Amsterdam: 36.5
Erfurt: 27.1
Helsinki: 19.6
Range (Min, Max):
Amsterdam: (13.6-112.0)
Erfurt: (5.2-104.2)
Helsinki: (6.4-67.4)
Monitoring Stations: 1
Copollutant:
PM25
NCO.01-0.1
CO
N02
S02
There was a tendency toward positive
associations between avoidance of
activities and both particulate air
pollution (PMio) and gases, but none of
the associations were statistically
significant....In both incidence analyses
and prevalence analyses, odds ratios
for PMio were generally similar to the
corresponding odds ratios for PM25, but
were somewhat less significant.'
Reference: Delfmo et al. (1998,
0514061
Period of Study :Aug-Oct 1995
Location: Alpine, CA
Outcome: asthma symptom severity
Age Groups: 9-17
Study Design: Panel Study
N: 24 non-smoking pediatric asthmatics
Statistical Analyses: GEE
Covariates: Day of week, temperature,
humidity, wind speed
Statistical Package: SAS
Lags Considered: 0-5, 0, 0-4
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD):
31(8)
90th: 42
Range (Min, Max): 16, 54
Copollutant (correlation):
03(r = 0.32)
PM Increment: 42 pg/m3 (90th
percentile increase)
Asthma symptoms:
Everyone: 1.47 (0.90, 2.39) lag 0
Everyone: 1.73 (1.03, 2.89) lag 0-4
Less symptomatic: 2.47 (1.23-4.95)
lagO
Less symptomatic: 4.03 (1.22,13.33)
lag 0-4
More symptomatic: 1.50 (0.80, 2.80)
lagO
More symptomatic: 1.95 (1.12, 3.43)
lag 0-4
PMio + 03
Asthma symptoms: 1.31 (0.84, 2.06)
lagO
1.65(1.03, 2.66) lag 0-4
Less symptomatic: 2.08 (1.12-3.83)
lagO
Less symptomatic: 3.35 (1.06,10.51)
lag 0-4
More symptomatic: 1.40 (0.77, 2.53)
lagO
More symptomatic: 1.87(1.11,3.13)
lag 0-4
December 2009
E-139
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Delfino et al. (2002,
0937401
Period of Study: Mar-Apr 1996
Location: Alpine, California
(a semi-rural area)
Outcome: Asthma symptoms that
interfere with daily activities
Age Groups: 9-19 yr
Study Design: Daily panel study
N: 22 asthmatic children
Statistical Analyses: GEE
Covariates: Temperature, relative
humidity, day-of-weektrends, linear
time trend across the 61 days, and
upper or lower respiratory infection
Season: "Early spring season" of
Mar-Apr
Dose-response Investigated? Yes
Statistical Package: SAS, version 8
Lags Considered: 0-, 1-, 2-, 3-, 4-, 5-,
and 3-day ma
Pollutant: PM10
Averaging Time: 1 h max
Mean (SD): 38(15)
Percentiles: 90th: 63
Range (Min, Max): (12-69)
Averaging Time: 8 h max
Mean (SD): 28(12)
Percentiles: 90th: 46
Range (Min, Max): (8-57)
Averaging Time: 24 h
Mean (SD): 20(9)
Percentiles: 90th: 32
Range (Min, Max): (7-42)
Copollutant (correlation):
1 h maxPM10
8h maxPM10: r = 0.93
24hPM,o:r = 0.84
1 h max 03:r = 0.68
8h max03: r = 0.95
1 h max N02:r = 0.49
8hmaxN02:r = 0.55
8 h max PM10:1 h max PM10: r = 0.93
24hPM,0:r = 0.95
1 h max 03:r = 0.72
8h max03: r = 0.65
1 h maxN02: r = 0.48
8hmaxN02:r = 0.55
24hPM10:1 h max PM10:r = 0.84
8h maxPM10: r = 0.95
1 h max 03:r = 0.74
8h max03: r = 0.71
1 h max N02:r = 0.37
8hmaxN02:r = 0.44
PM Increment: 90th percentile
increase
Effect Estimate [Lower Cl, Upper Cl]:
ORs for risk of asthma symptoms in
those who report a respiratory infection
compared to those who do not have a
respiratory infection
1h max PM10 lag 0:4.88 (1.31-18.2)
8 h max PM,o lag 0:6.78 (1.38-33.3)
24 h mean PM10 lag 0: 4.68 (0.71-30.7)
3-day ma 1 h max PM10:11.1 (1.10-112)
3-dayma8hmaxPM10:10.1 (1.42-
72.0)
3-day ma 24 h PM,0: 2.67 (0.60-11.8)
Effect modification by anti-inflammatory
medication use on the relationship of
asthma symptoms in children
1 h max PM10 lag 0:1.41 (0.87-2.30)
On medication: 0.96 (0.25-3.69)
Not on medication: 1.92 (1.22-3.02)
8 h max PM10 lag 0:1.19 (0.74-1.94)
On medication: 0.75 (0.18-3.04)
Not on medication: 1.68 (0.91-3.09)
24 h mean PM,o lag 0:1.08 (0.73-1.61)
On medication: 0.80 (0.24-2.69)
Not on medication: 1.35 (0.82-2.22)
3-day ma 1 h max PM10:1.45 (0.76-
2.76)
On medication: 1.01 (0.14-7.02)
Not on medication: 1.92 (0.99-3.71)
3-dayma8hmaxPM10:1.32(0.76-
2.29)
On medication: 0.82 (0.17-3.94)
Not on medication: 1.89 (1.10-3.24)
3-day ma 24 h PM,0:1.22 (0.84-1.77)
On medication: 0.75 (0.26-2.14)
Not on medication: 1.75 (1.15-2.68)
Dose-response results are found in Fig
2 and not quantitatively reported
elsewhere.
Reference: Delfino et al. (2003,
0909411
Period of Study: Nov 1999-Jan 2000
Location:
Huntington Park, Los Angeles
Outcome: Asthma severity scale
Peak Expiratory Flow Rate (PEF)
Age Groups: Ages 10 to 16
Study Design: Longitudinal study panel
N: 22 children
Statistical Analyses: Regression
analysis (GEE, GLM)
multivariate regression models
Covariates: Day of the week, Maximum
Temperature, Respiratory Infections
Season: Winter
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0,1
Pollutant: PM,0
Mean (SD): 59.9 (24.7)
Range (Min, Max): 20-126
IQR: 37
90th: 86.0
Monitoring Stations: 1
Copollutant (correlation):
8-h max N02 = 0.38
8-hmax03 = -0.16
8-h max CO = 0.50
8-h max S02 = 0.73
PM Increment: IQR 37.0 pg/m
OR Estimate [Lower Cl, Upper Cl]
lag:
LagO
Symptom Scores >1:1.45(1.11,1.90)
Symptom Scores >2: NR
Lag1
Symptom Scores >1:1.07 (0.64,1.77)
Symptom Scores >2: NR
December 2009
E-140
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Delfino et al. (2004,
0568971
Period of Study :Sep-Oct 1999
Apr-Jun 2000
Location: Alpine, California
Outcome: FEVi
Age Groups: 9-19 yr old
Study Design: Panel study
N: 24 children
Statistical Analyses: GLM
Akaike's information criterion and
Bayesian information criterion
Covariates: Day of week, Personal
temperature and relative humidity, time
of FEVi maneuver (morning, afternoon,
or evening), Season (fall 1999 or spring
2000)
As-needed medication use
Presence or absence of upper or lower
respiratory infections
Season: Spring, Fall
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: Lag 0-4
Pollutant: PM,0
Averaging Time: 4 h, 8 h, 12 h, 24 h
Personal Monitor
1-h max personal PM last
24-h
Mean (SD): 151.0 (12.03)
90th: 292.4
Range (Min, Max): (9.1,996.8)
Mean personal PM last 24-h
Mean (SD): 37.9 (19.9)
90th: 65.1
Range (Min, Max):
(3.9, 113.8)
Central outdoor stationary-site PM
1-h Maximum TEOM
PM,o last 24-h
Mean (SD): 54.4 (13.8)
90th: 71.0
Range (Min, Max): (24.4, 95.4)
Mean TEOM PM,o last 24-h
Mean (SD): 29.7 (8.6)
90th: 40.9
Range (Min, Max): (12.9, 50.7)
24-h mean PM10
Mean (SD): 23.6 (9.1)
90th: 34.6
Range (Min, Max): (3.2, 48.0)
Copollutant (correlation): 8-h max
personal PM
8-h max 03 = 0.03
8-h Max N02 = 0.26
24-h Mean Personal
PM = 0.94
8-h Max TEOM PM,o = 0.38
24-h Mean TEOM PM10 = 0.40
24-h Central HI PM,o = 0.37
24-h Central HI PM25 = 0.38
24-h Outdoor HI PM,o = 0.32
24-h Outdoor HI PM2 5 = 0.39
24-h Indoor HI PM10 = 0.23
24-h Indoor HI PM25 = 0.37
24-h mean personal PM
8-h max 03 = 0.01
8-h Max N02 = 0.27
8-h Max Personal PM = 0.94
8-h Max TEOM PM10 = 0.36
24-h Mean TEOM PM10 = 0.39
24-h Central HI PM,o = 0.36
24-h Central HI PM25 = 0.43
24-h Outdoor HI PM,o = 0.34
24-h Outdoor HI PM2 5 = 0.44
24-h Indoor HI PM10 = 0.29
24-h Indoor HI PM25 = 0.46
24-h Mean TEOM PM,o
8-h max 03 = 0.41
8-h Max N02 = 0.58
8-h Max Personal PM = 0.40
24-h Mean Personal PM = 0.39
8-h Max TEOM PM10 = 0.92
24-h Central HI PM,o = 0.86
24-h Central HI PM25 = 0.78
24-h Outdoor HI PM,o = 0.79
24-h Outdoor HI PM2 5 = 0.78
24-h Indoor HI PM10 = 0.36
24-h Indoor HI PM25 = 0.59
Results presented graphically: Percent
predicted FEV, was inversely
associated with personal exposure to
fine particles.
- Inverse associations of FEV, with
stationary-site indoor, outdoor and
central-site gravimetric PM25 and PM10,
and with hourly TEOM PM,o
December 2009
E-141
-------
Study
Reference: Delfino et al. (2006,
0907451
Period of Study: Region 1 : Aug to Mid
Dec 2003. Region 2: Jul through Nov
2004
Design & Methods
Outcome: Fractional Concentration of
Nitric Oxide in exhaled air (FENO)
Age Groups: 9 through 18
Study Design: Longitudinal Panel
Study
Concentrations!
Pollutant: PM,0
Central Site
Averaging Time: 24 h
Riverside
Effect Estimates (95% Cl)
PM Increment: IQR increase
(Riverside: 28.41 pg/m3, Whittier
21.87|jg/m3)
Coefficient [Lower Cl, Upper Cl]
lag: Lag = 2-day ma
Location: Region 1: Riverside, CA.
Region 2: Whittier, CA
N: 45 children
Statistical Analyses: Linear mixed-
effects models
Two-stage hierarchical model
Empirical Variograms
Fourth-order polynomial distributed lag
mixed-effects model
Covariates: Personal temperature,
Personal Rel. Humid., 10-day exposure
run, Respiratory infections, Region of
study, Sex, Cumulative daily use of as-
needed B-agonist inhalers
Dose-response Investigated? No
Lags Considered: Lag 0, Lag 1, 2-day
ma
Mean (SD): 70.82 (29.36)
SOth(Median): 65.96
Range (Min, Max): (30.75,54.05) pg/m3
Whittier
Mean (SD): 35.73 (16.6)
SOth(Median): 34.65
Range (Min, Max):
(5.86, 105.46) pg/m3
Monitoring Stations: 48 personal
nephelometers, 2 central sites
Stratified by Medication Use
Not Taking Anti-lnflamm. Medication
Central 0.76 (-1.54, 3.07)
Taking Anti-lnflamm. Medication
Central 0.53 (-0.83,1.90)
Inhaled Corticosteroids
Central 1.28 (-0.01, 2.58)
Antileukotrienes +- inhaled
corticosteroids
Central-2.10 (-5.33,1.12)
Notes: Fig of Estimated lag effect of
hourly personal PM25 on FENO.
Fig of the Estimated lag effect of hourly
personal PM25 on FENO by use of
medications.
Fig of one- and two-pollutant models for
change in FENO using 2-day Ma
personal and central-site pollutant
measurements.
December 2009
E-142
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Desqueyroux et al. (2002,
0260521
Period of Study: Nov 1995-Nov 1996
Location: Paris, France
Outcome: Asthma attacks
Age Groups: Adults.
Study Design: Panel study
N: 60 moderate to severe adult
asthmatics
Statistical Analyses: Marginal logistic
regression
Covariates: FEV,, smoking, allergy,
oral steroid treatment, mean daily
temperature, relative humidity, pollen
counts, season, holiday period
Season: winter, summer
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 1,2, 3, 4, 5, 3-5
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD):
Summer: 23 (9)
Winter: 28 (14)
Range (Min, Max):
Summer: 6, 63
Winter: 9, 84
Monitoring Stations: 7
Copollutant: S02, N02, 03
0.93
1.11
1.17
1.16
1.21
0.80, 1.08
0.98, 1.26
1.03, 1.33
1.01, 1.34
1.01, 1.34'
PM Increment: 10 pg/m
OR Estimate [Lower Cl, Upper Cl]
g: 0.87 [0.71,1.06] lag 1
•" lag 2
Iag3
lag 4
Iag5
lag 3-5
Vs seasons alone:
Winter: 1.41 [1.16,1.71] lag 3-5
Summer: 1.03 [0.72,1.47] lag 3-5
Vs link to explanatory factors:
No link: [1.71 [1.20, 2.43] lag 3-5
Link: 1.27 [1.06,1.52] lag 3-5
Vs occurrence of infection:
Wthout infection:
1.52 [1.16, 2.00] lag 3-5
Wth infection: 1.30 [1.03,1.65] lag 3-5
Vs baseline pulmonary function:
FEV,>/ = 68% predicted:
1.38 [1.06, 1.79] lag 3-5
FEV <68% predicted:
1.45 [1.11, 1.90] lag 3-5
Vs smoking habits:
Nonsmokers: 1.53 [1.18,1.98] lag 3-5
Current & ex-smokers:
1.18 [0.90, 1.54] lag 3-5
Vs allergy:
Non-allergic: 1.29 [0.94,1.77] lag 3-5
Allergic: 1.49 [1.17,1.90] lag 3-5
Vs regular oral steroid treatment:
No: 1.41 [1.15,1.73] lag 3-5
Yes: 1.41 [0.88, 2.25] lag 3-5
Multipollutant model: PM,0 + N02:1.43
[1.16, 1.76] Lag 3-5
PMio + S02:1.51 [1.20, 1.90] Lag 3-5
PM10 + 03:1.09 [0.71, 1.67] Lag 3-5
Reference: Diette et al. (2007,1563991 Outcome: Asthma in the last 12 mo
(493 xl
Period of Study: Sep 2001 -Dec v '
2003
Location: East Baltimore, MD
Age Groups: 2 to 6 yr old
Study Design: Prospective cohort
N: 150 with asthma
150 without asthma
Statistical Analyses:
Student's two-tailed t-test
Kruskal-Wallistest
Pearson's chi square
Fisher's exact test
Covariates: Season of collection
Dose-response Investigated? No
Statistical Package: STATASE 8.0
Pollutant: PM,0
Averaging Time: 72 h
60th(Median): 43.7
IQR: (29-70)
Notes: "Pollutant concentrations in the
homes of asthmatic and control children
who lived in the same home for their
whole life were not different compared
with those who had moved at least
once."
December 2009
E-143
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Ebelt et al. (2005, 0569071
Period of Study: Summer of 1998
Location: Vancouver, Canada
Outcome: spirometry
Age Groups: Range from 54-86 yr
mean age = 74 yr
Study Design: Extended analysis of a
repeated-measures panel study
N: 16 persons with COPD
Statistical Analyses: Earlier analysis
expanded by developing mixed-effect
regression models and by evaluating
additional exposure indicators
Dose-response Investigated? No
Statistical Package: SASV8
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD):
Ambient PM10:17 (6)
Exposure to ambient PM10:10.3 (4.6)
Range (Min, Max):
Ambient PM10: (7-36)
Exposure to ambient PM10: (1.5-23.8)
Monitoring Stations: 5
Copollutant (correlation):
Ambient PMio-25:r = 0.69
Ambient PM25r = 0.78
Exposure to Ambient PM10: r = 0.71
PM Increment: Ambient PM10: 7 (IQR)
Exposure to ambient PM10: 6.5 (IQR)
Notes: Effect estimates are presented
in Fig 2 and Electronic Appendix Table 1
(only available with electronic version of
article) and not provided quantitatively
elsewhere.
Reference: Fischer et al. (2007,
1564351
Period of Study: 7 wk (dates not
specified)
Location: The Netherlands
Outcome: Respiratory Symptoms, Sore
throat, Runny nose, Cold, Sick at home
Study Design: Prospective cohort
N:68
Statistical Analyses: Linear regression
model (PROC mixed)
Age Groups: 10-11
Lag: 1-2
Statistical Package: SASv 6.11
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): 56 pg/m3
IQ (25th, 75th): (21,187) pg/m3
Co pollutants:
BS
N02
CO
NO
Increment: 10|jg/m
% Increase in eNO and PM10 and
change in spirometric lung function lag
eNO and PM10 only
6.5
7.8
0.9, 12.4) 1
-11.3,31.0)2
FVC mean (SEM)
0.4
0.6
0.5)1
1.6)2
FEV, mean (SEM)
-0.3 (0.5
-2.1 (1.9
PEF mean (SEM)
-2.8 (3.3) 1
7.1(12.0)2
MMEF mean (SEM)
-0.5(1.7
-2.5 (5.9
Reference: Forsberg et al. (1998,
0517141
Period of Study: Jan 1994-March
1994
Location: Urban and rural areas of
Umea, Sweden
Outcome: Respiratory Symptoms,
Shortness of breath
Wheeze, Asthma attacks, Recent
asthma, Dry cough, Doctor-diagnosed
asthma, Recently treated for asthma,
Early chest illness
Study Design: Cohort panel
Statistical Analyses: Logistic linear
regression
Age Groups: 6-12
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD):
Urban: 13.4|jg/m3
Rural: 11.5|jg/m3
Range (Min, Max):
Urban: (0, 40.5) pg/m3
Rural: (1.6, 29.0) pg/m3
Copollutants (correlation):
BS:r = 0.73
Increment: 10|jg/m
OR between prevalence of acute
respiratory symptoms and PMi0
exposure for urban and rural children
lag
Urban children:
Cough: 1.031 (0.957, 1.112)0
0.997
1.018
0.923, 1.077
0.940, 1.103
1.094(0.895, 1.338) 0-6 avg
Phlegm:
0.998(0.899,1.108)0
1.035(0.928,1.154)1
1.121
1.043
1.013, 1.240
0.822, 1.324
0-6 avg
Upper respiratory symptoms:
1.004
0.975
0.949, 1.063
0.922,1.031 1
0.951(0.895,1.010)2
0.849(0.687, 1.050) 0-6 avg
Lower respiratory symptoms:
0.984(0.872,1.110)0
0.919
0.894
0.812, 1.039
0.771, 1.036 2
0.800(0.617, 1.038) 0-6 avg
Rural children (control)
Cough:
0.997(0.900,1.105)0
1.003
0.997
0.906, 1.112
0.891,1.116
0.855(0.655, 1.115) 0-6 avg
Phlegm:
1.024(0.880,1.192)0
0.995(0.853,1.160)1
1.117(0.956, 1.305)2
1.041(0.742, 1.459) 0-6 avg
Upper respiratory symptoms:
1.093 0.989, 1.208 0
1.018(0.918,1.130)1
December 2009
E-144
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
1.075(0.962, 1.201)2
1.052(0.786, 1.407) 0-6 avg
Lower respiratory symptoms:
1.022(0.855,1.180)0
0.998
1.000
0.855, 1.164
0.830, 1.206
0.939(0.703, 1.253) 0-6 avg
OR between incidence of acute
respiratory symptoms and PM10
exposure in urban and rural children
lag
Urban Children:
Cough:
1.114(0.886,1.401)0
0.891 (0.703, 1.130)1
0.766
0.817
Phlegm:
0.577, 1.017
0.523, 1.276
0.954
1.056
0.664, 1.371
0.744, 1.501
0-6 avg
1.416(0.969,2.069)2
0.808(0.357, 1.827) 0-6 avg
Upper respiratory symptoms:
1.155(0.965, 1.383)0
0.788
0.629, 0.986
0.728, 1.077 2
0.770(0.549, 1.081) 0-6 avg
Lower respiratory symptoms:
1.060(0.828,1.356)0
0.763 (0.584, 0.996) 1
0.652
0.519
0.493, 0.863
0.306, 0.882
Rural Children:
Cough:
1.052(0.767,1.444)0
0.753(0.547, 1.038)1
0-6 avg
0.840
0.800
Phlegm:
0.571, 1.235
0.409, 1.565
1.051
1.010
0.731, 1.509
0.693, 1.472
0-6 avg
0.998(0.652, 1.528)2
0.797(0.344, 1.847) 0-6 avg
Upper respiratory symptoms:
1.044(0.813, 1.341)0
0.810
0.800
0.612, 1.072)1
0.611,1.048)2
0.714(0.417, 1.220) 0-6 avg
Lower respiratory symptoms:
1.079(0.756,1.539)0
0.888(0.615, 1.281)1
0.472, 1.083)2
0.395, 1.711) 0-6 avg
0.715
0.822 .
OR between prevalence of medication
use and PM10 exposure in urban and
rural children lag
Bronchodilator use - Urban children:
0.998
0.999
0.951, 1.048
0.952, 1.049
1.006(0.953, 1.062)2
0.919(0.775, 1.090) 0-6 avg
Rural children:
0.970(0.904,1.040)0
0.959
1.008
0.893, 1.030
0.927, 1.095
1.087(0.914, 1.292) 0-6 avg
OR between incidence of medication
use and PM10 exposure in urban and
rural children lag
Bronchodilator use - Urban children:
1.498(0.899,2.498)0
1.049(0.565, 1.947)1
0.674, 1.954)2
0.611, 5.227) 0-6 avg
1.148
1.787 .
Rural children:
1.275 0.702,2.315 0
0.924(0.437,1.956)1
December 2009
E-145
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
1.005(0.522, 1.936)2
1.823(0.534, 6.277) 0-6 avg
Reference: Goncalves et al. (2005,
0898841
Period of Study: Dec 1992-Mar 1993.
Dec 1992-Mar 1994
Outcome: Respiratory
morbidity/admissions
Age Groups: Children <13 yr
Study Design: Time series
Pollutant: PM,0
Averaging Time: 24 h
Copollutant: S02 , 03
PCA coefficients: PC1.PC2, PCS:
Summer 1992/1 993:
PMi0:0.69, 0.45, 0.13
Solar Radiation: -0.04, 0.94 to -0.12
Location: Sao Paulo
Statistical Analyses: Principal
component analysis
Covariates: Daily mean temperature,
daily mean water vapor density, solar
radiation
Season: Summer
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: Lag 3
Mean Temperature: 0.62, 0.44 to -0.47
Mean V\Mer Vapor Density:
0.73 to-0.46 to-0.26
S02: 0.78 to-0.03, 0.33
03:0.18, 0.63, 0.37
Respiratory Mortality:
0.05(0-0.02,0.81
Variations explained by Principal
Component:
PC1:0.29
PC2: 0.27
PCS: 0.17
Summer 1993/1994:
PM,0: 0.38, 0.80(0-0.23
Solar Radiation: 0.02, 0.09 to -0.97
Mean Temperature: 0.71, 0.40 to -0.37
Mean V\Mer Vapor Density:
0.88, 0.25, 0.09
S02: 0.01, 0.92, 0.00
03: 0.47 to-0.06 to-0.35
Respiratory Mortality: -0.73, 0.11, 0.08
Variations explained by Principal
Component:
PC1:0.31
PC2: 0.25
PCS: 0.18
Notes: Association between respiratory
morbidity and air pollution more likely
during summer with smaller contrasts in
synoptic weather condition (summer
1992/93) but respiratory morbidity more
related to weather variables during
summer with larger contrasts (summer
1993/94).
Reference: Gordian and Choudhury
(2003, 0548421
Period of Study: 1994-Dec 1996
Location: Anchorage, Alaska
Outcome: Asthma medication among
school children
Age Groups: Elementary school
children (kindergarten-6th grade)
Study Design: Time series
Statistical Analyses: Time series
regression model
Covariates: Day of the week, month,
time trend, temperature
Season: All seasons
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 1,2, 7,14,21,28
Pollutant: PM10
Averaging Time: 24 h
Mean (SD): 36.11 (30.46)
Range (Min, Max): 2.96, 210.0
Monitoring Stations: 1
Model regression slope coefficient for
PM,o (estimated SE) lag:
7.25 (2.88)
lag 21
RR: 1.075 (1.016, 1.138)
Notes: PMi0 coefficients for other lags
were also statistically significant but not
reported.
December 2009
E-146
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Pollutant: PMi0
Reference: Harre et al. (1997, 0957261 Outcome: Respiratory symptoms,
_..,-.. , „„, „ .,„„,, Cough, Wheeze, Chest tightness, .
Period of Study: Jun 994-Aug 1 994 shortness of breath, Change in sputum Averaging Time: 24-h avg
volume, Nose, throat, or eye irritation,
'
Location: Christchurch, New Zealand
Study Design: Prospective cohort
Statistical Analyses: Poisson, log
linear regression
Age Groups: >55
- „ . ._
Copollutants:
SO
2
|\|Q2
Increment: 35.04 pg/m
Relative Risk (Lower Cl, Upper Cl)
lag:
Chest symptoms: 1.38 (1.07,1.78)1
Wheeze: 0.97 (0.75,1.26)1
Nebulizer Use: 0.71 (0.42,1.18)1
Inhaler Use: 0.94 (0.78,1.13)1
Reference: Hastings and Jardine
(2002, 0303441
Period of Study: 1997-1998
Location: Bosnia (U.S. military camps)
Outcome: Weekly rates of upper
respiratory disease (URD), reported by
the medical treatment facility in each
military camp
Age Groups: U.S. soldiers
Study Design: Ecologic (at level of
military camp)
N: 5 camps
Statistical Analyses:
1. Pearson correlations between weekly
URD rates and weekly PMi0 (avg and
max)
2.Kruskal Wallace test to compare URD
rates in the 4 exposure quartiles
3. Mann Whitney test to compare
dichotomized exposure groups (above
and below 50th percentile)
Dose-response Investigated? Yes
Lags Considered: Weekly rates of
URD disease were related to avg
weekly PM levels in the same week
Pollutant: PM,0
Mean (SD):
PM10 avg: 75.5
PM10 max: 92.9
Percentiles:
PM10 max:
25th: 58.57
50th: 74.55
75th: 107.56
PM,o avg:
25th: 42.19
50th: 64.17
75th: 81.75
Range (Min, Max):
PM10 avg: 25.0, 338.7
PM,o max: 25.0, 338.7
Monitoring Stations: At least 1 in each
of the 5 camps
PM max Quartiles (combining all
camps):
Q1:<58.7|jg/m3
Q2:60.1to<75.54|jg/m3
Q3: 78.56 to <107.56 pg/m3
Q4: >107.56 pg/m3
For dichotomous analysis
cutoff = 74.55 pg/m3
PM avg Quartiles (combining all
camps):
Q1:<42.19|jg/m3
Q2: 42.19to64.17 pg/m3
03: 64.17 to 81.75 fjg/m3
Q4:>81.75|jg/m3
For dichotomous analysis
cutoff =64.17 fjg/m
Pearson correlation coefficients
between URD rate and PM category [p-
value]: PM,0 max: quartiles of PMURD
rates
All camps 0.203 [0.041]
Blue Factory camp 0.277 [0.095]
Comanche 0.165 [0.237]
Demi 0.639 [0.123]
McGovern 0.535 [0.177]
Tuzla Main 0.107 [0.327]
PMio max: dichotomous PMURD rates:
All camps 0.283 [0.007]
Blue Factory camp 0.038 [0.430]
Comanche 0.282 [0.107]
Demi 0.927 [0.012]
McGovern 0.853 [0.033]
Tuzla Main 0.155 [0.258]
PM,o avg: quartiles of PM*URD rates:
All camps 0.149 [0.101]
Blue Factory camp 0.301 [0.077]
Comanche 0.246 [0.141]
Demi 0.437 [0.231]
McGovern 0.853 [0.033]
Tuzla Main 0.182 [0.222]
PMio avg: dichotomous PMURD rates:
All camps 0.060 [0.305]
Blue Factory camp -0.075 [0.365]
Comanche 0.143 [0.268]
Demi N/A*
McGovern N/A*
Tuzla Main 0.123 [0.303]
Kruskal Wallace p-value comparing
URD rates across exposure quartiles:
PM10 max
All camps 0.047
Blue Factory camp 0.321
Comanche 0.556
Demi 0.165
McGovern 0.202
Tuzla Main 0.554
PMio avg
December 2009
E-147
-------
Study Design & Methods Concentrations! Effect Estimates (95% Cl)
All camps 0.672
Blue Factory camp 0.809
Comanche 0.658
Demi 0.564
McGovernO.157
Tuzla Main 0.891
Mann-Whitney p-value comparing URD
rates between upper and lower 50th
percentileofPM:
PM10 max
All camps 0.034
Blue Factory camp 0.173
Comanche 0.314
Demi 0.083
McGovern 0.401
Tuzla Main 0.481
PMio avg
All camps 0.824
Blue Factory camp 0.682
Comanche 0.508
Demi N/A*
McGovern N/A*
Tuzla Main 0.656
Notes: * There were no days that fell in
the upper 50 percentile for PM avg in
these camps
-Rates of URD by PM quartiles for each
camp presented in figures. Authors
state, "Generally the avg URD rate
increased with quartile of maximum
exposure...the trend was not as clear
for quartiles of PM10 avg exposure"
December 2009 E-148
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Hong et al. (2007, 0913471 Outcome: Peak expiratory flow rate
Period of Study: Mar 23-May 2004 ' '
Age Groups: 3rd to 6th grade (mean
Location: School on the Dukjeok Island age = 95 yr)
near Incheon City, Korea
Study Design: Panel study
N: 43 schoolchildren
Statistical Analyses: Mixed linear
regression
Covariates: Age, sex, height, weight,
asthma history, and passive smoking
exposure at home
Dose-response Investigated? No
Lags Considered: 0,1,2,3,4,5
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 35.30 (23.48)
50th (Median): 29.36
Range (Min, Max):
(12.24-124.87)
PM Component:
Fe: mean = 0.208 (0.203) pg/m3
Median = 0.112
Range (Min, Max): (0.061-0.806)
Mn: mean = 0.008 (0.005) pg/m3
Median = 0.007
Range (Min, Max): (0.000-0.019)
Pb: mean = 0.051 (0.031) pg/m3
Median = 0.051
Range (Min, Max): (0.011-0.155)
Zn: mean = 0.021 (0.021) pg/m3
Median = 0.013
Range (Min, Max): (0.006-0.112)
Al: mean = 0.085 (0.100) pg/m3
Median = 0.031
Range (Min, Max): (0.017-0.344)
Copollutant: PM25
Effect Estimate: Regression
coefficients of morning and daily mean
PEFR on PMio and metal components
using linear mixed-effects regression
Lag 1 (PM10)
Morning PEFR
Crude: IS = -0.00, p = 0.99
Adjusted: IS = -0.04, p = 0.37
Mean PEFR
Crude: IS = 0.00, p = 0.93
Adjusted: IS = -0.05, p = 0.12
Lag 1 (logFe)
Morning PEFR
Crude: IS =-1.26, p = 0.31
Adjusted: IS = -3.24, p = 0.13
Mean PEFR
Crude: IS =-1.20, p = 0.20
Adjusted: IS = -2.37, p = 0.15
Lag 1 (logMn)
Morning PEFR
Crude: IS = -4.40, p< 0.01
Adjusted: IS = -9.82, p< 0.01
Mean PEFR
Crude: IS = -4.05, p< 0.01
Adjusted: IS = -8.44, p< 0.01
Lag 1 (logPb)
Morning PEFR
Crude: IS = -6.79, p< 0.01
Adjusted: IS = -6.83, p<0.01
Mean PEFR
Crude: IS = -6.23, p< 0.01
Adjusted: IS = -6.37, p<0.01
Lag 1 (logZn)
Morning PEFR
Crude: IS = -0.55, p = 0.71
Adjusted: IS = -0.98, p = 0.59
Mean PEFR
Crude: IS =1.33, p = 0.24
Adjusted: IS =1.53, p = 0.28
Lag1 (logAI)
Morning PEFR
Crude: IS = -0.58, p = 0.57
Adjusted: IS = -2.22, p = 0.25
Mean PEFR
Crude: IS = -0.59, p = 0.45
Adjusted: IS = -1.48, p = 0.32
Regression coefficients of morning and
daily mean PEFR on metal components
ofPM10andGSTM1andGSTT1
genotype using linear mixed-effects
regression
Lag 1 (logPb)
Morning PEFR: IS =-7.26, p<0.01
Mean PEFR: IS =-6.43, p< 0.01
GSTM1
Morning PEFR: IS = 21.19, p = 0.23
Mean PEFR: IS = 20.09, p = 0.25
Lag 1 (logMn)
Morning PEFR: IS = -10.31, p<0.01
Mean PEFR: IS =-8.66, p< 0.01
GSTM1
Morning PEFR: IS = 21.02, p = 0.23
Mean PEFR: IS = 19.84, p = 0.25
Lag 1 (logPb)
Morning PEFR: IS =-7.26, p<0.01
Mean PEFR: IS =-6.43, p< 0.01
GSTT1
Morning PEFR: IS = 2.07, p = 0.90
Mean PEFR: IS =-2.39, p< 0.88
Lag 1 (logMn)
Morning PEFR: IS = -10.32, p < 0.01
Mean PEFR: IS =-8.67, p< 0.01
GSTT1
Morning PEFR: IS = 2.02, p = 0.90
Mean PEFR: IS = 2.33, p = 0.88
December 2009
E-149
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Hwang et al. (2006,
0889711
Period of Study: 2001
Location: Taiwan
Outcome: Allergic rhinitis
Study Design: Cross-sectional
Statistical Analyses: Two-stage
hierarchical models
Age Groups: 6-15 yr
Pollutant: PM,0
Averaging Time: 1-h avg
Mean (SD): 55.58 (16.57)
Range (Min, Max):
(29.36, 99.58)
Copollutants (correlation):
CO: r = 0.27
N0x:r = 0.34
03:r = 0.28
S02:r = 0.58
Increment: 10|jg/m
Odds Ratio (Lower Cl, Upper Cl)
lag:
PM10 alone: 1.00 (0.99,1.02)
, PM10: 0.99 (0.97, 1.00)
CO, PM10:1.00 (0.99, 1.01)
03,PM10:1.00 (0.99, 1.02)
Gender
Male: 1.02 (0.99,1.04)
Female: 0.99 (0.97,1.02)
Parental atopy*
Yes: 1.00 (0.98, 1.03)
No: 1.01 (0.99, 1.03)
Parental education
<6yr: 1.05 (0.96,1.14)
6-8 yr: 1.03 (0.98,1.07)
9-11 yr: 1.00 (0.98,1.03)
12+yr: 0.99 (0.97,1.02)
Environmental tobacco smoke
Yes: 1.01 (0.99,1.03)
No: 1.00 (0.98, 1.03)
Visible mold**
Yes: 1.02 (0.99, 1.06)
No: 1.00 (0.98, 1.02)
* Parental atopy was a measure of
genetic predisposition and was defined
as the father or the mother of the index
child ever having been diagnosed as
having asthma, allergic rhinitis, or atopic
eczema.
"Visible mold found in the home.
Reference: Jalaludin et al. (2004,
0565951
Outcome: Respiratory symptoms,
Wheeze, Dry cough, Wet cough
Period of Study: Feb 1994-Dec 1994 Study Design: Longitudinal study panel
Location: Western and southwestern
Sydney, Australia
Statistical Analyses: Logistic
regression model (GEE)
Age Groups: 9-11 yr
Pollutant: PM10
Averaging Time: 24-h avg
Mean (SD): 22.8 (13.8)
IQ Range (26th,76th): (12.00,122.8)
Copollutants (correlation):
03:r = 0.13
N02:r = 0.26
Increment: 10|jg/m
Odds Ratio (Lower Cl, Upper Cl)
Lag
Wheeze
1.01 (0.99, 1.03)0
1.01(0.97,1.04)1
0.99(0.96,1.03)2
1.02(0.98, 1.06) 0-2 avg
1.04(0.99, 1.10) 0-5 avg
Dry Cough
1.00(0.98, 1.03)0
1.00(0.97, 1.03)1
1.00(0.97,1.02)2
1.00(0.97, 1.03) 0-2 avg
1.03(0.98, 1.08) 0-5 avg
Wet Cough
1.01 (0.99, 1.04)0
0.99(0.97, 1.01)1
1.00(0.97,1.03)2
0.99(0.96, 1.02) 0-2 avg
0.99(0.94, 1.04) 0-5 avg
Inhaled B2-agonist Use
0.99(0.98, 1.01)0
1.00(0.98, 1.03)1
0.99(0.97,1.01)2
1.00(0.97, 1.02) 0-2 avg
1.02(0.98, 1.06) 0-5 avg
Inhaled Corticosteroid Use
1.00(0.99, 1.01)0
1.00(0.99, 1.02)1
1.00(0.99,1.02)2
1.00(0.98, 1.02) 0-2 avg
1.00(0.97, 1.02) 0-5 avg
Doctor Visit for Asthma
1.11 (1.04, 1.19)0
1.10(1.02, 1.19)1
1.15(1.06,1.24)2
1.11 (1.03,1.20) 0-2 avg
1.14(0.98, 1.31) 0-5 avg
OR for respiratory symptoms and
PM10 exposure by different groups
December 2009
E-150
-------
Study Design & Methods Concentrations! Effect Estimates (95% Cl)
All children
Wheeze: 1.01 (0.99,1.04)
Dry Cough: 1.00 (0.97,1.02)
Wet Cough: 1.01 (0.98,1.04)
Inhaled B2-agonist Use:
1.00(0.98,1.02)
Inhaled Corticosteroid Use:
0.99(0.98, 1.01)
Doctor Visit for asthma:
1.11(1.03,1.19)
Group 1*
Wheeze: 1.01 (0.98,1.04)
Dry Cough: 0.97 (0.94, 0.99)
Wet Cough: 1.00 (0.97,1.03)
Inhaled B2-agonist use:
1.00(0.98,1.02)
Inhaled Corticosteroid Use:
1.00(0.98,1.01)
Doctor Visit for asthma: 1.09
(0.98,1.21)
Group 2**
Wheeze: 1.01 (0.97,1.05)
Dry Cough: 1.02 (0.98,1.06)
Wet Cough: 1.01 (0.96,1.06)
Inhaled B2-agonist use:
0.99(0.94, 1.05)
Inhaled Corticosteroid Use:
0.99(0.97,1.01)
Doctor Visit for asthma:
1.12(1.02,1.23)
Group 3***
Wheeze: 1.08 (0.90,1.31)
Dry Cough: 1.01 (0.91,1.11)
Wet Cough: 1.02 (0.94,1.11)
Inhaled B2-agonist use:
0.98(0.84,1.11)
Inhaled Corticosteroid Use:
1.27(1.08, 1.49)
Doctor Visit for asthma: NR
'Group 1 consists of children with a
history of wheeze in the past 12 mo,
positive histamine challenge, and doctor
diagnosed asthma.
"Group 2 consists of children with a
history of wheeze in the past 12 mo and
doctor diagnosed asthma.
***Group 3 consists of children only with
a history y of wheeze in the past 12 mo.
December 2009 E-151
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Jansen, et al. (2005,
0822361
Period of Study: 1987-2000
Location: Seattle, WA
Outcome: FENO: fractional exhaled
nitrogen oxide, Spirometry, Blood
pressure, Sa02: oxygen saturation,
Pulse rate
Age Groups: 60-86 yr old
Study Design: Short-term cross-
sectional case series
N: 16 subjects diagnosed with COPD,
asthma, or both
Statistical Analyses: Linear mixed
effects model with random intercepts
Covariates: Age, relative humidity,
temperature, medication use
Season: winter 2002-2003
Dose-response Investigated? No
Statistical Package: STATA
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD):
Fixed-site Monitor: 18.0
All Subjects (N = 16)
Indoor, home: 11.93
Outdoor, home: 13.47
Personal: 23.34
Asthmatic Subjects (N = 7)
Indoor, home: 12.54
Outdoor, home: 11.86
Personal: 26.88
COPD Subjects (N = 9)
Indoor, home: 11.45
Outdoor, home: 14.76
Personal: 19.91
Range (Min, Max):
Fixed-site Monitor 2.5, 51
IQR:
All Subjects
Indoor, home: 6.93
Outdoor, home: 9.53
Personal: 20.72
Asthmatic Subjects
Indoor, home: 10.19
Outdoor, home: 8.77
Personal: 20.08
COPD Subjects
Indoor, home: 4.56
Outdoor, home: 6.14
Personal: 19.94
PM Increment: 10 |ig/m
Slope [95% Cl]: dependence of FENO
concentration [ppb] on PMi0
Asthmatic Subjects
Indoor, home: 3.81 [-0.86: 8.50]
Outdoor, home: 5.87 [2.87: 8.88]*
Personal: 0.66 [-0.56:1.88]
COPD Subjects
Indoor, home: 2.19 [-3.48: 7.87]
Outdoor, home: 4.45 [-1.11:10.01]
Personal: 0.17 [-1.61:1.96]
Results indicate that FENO may be a
more sensitive biomarker of PM
exposure than other traditional health
endpoints.
December 2009
E-152
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Johnston, et al. (2006,
0913861
Period of Study:
7mo(Apr7-Nov7, 2004)
Location: Darwin, Australia
Outcome: Asthma symptoms
Age Groups: All ages
Study Design: Time-series
N: 251 people (130 adults, 121 children
Statistical Analyses: Logistic
regression model
Covariates: Minimum air temperature,
doctor visits for influenza and the
prevalence of asthma symptoms and,
the fungal spore count and both onset
of asthma symptoms and
commencement of reliever medication
Season: "Dry season"-specific months
NR, note Southern Hemisphere
Dose-response Investigated? No
Statistical Package: STATA8
Lags Considered: 0-5 days
Pollutant: PM,0
Averaging Time: Daily
Mean (SD): 20 (6.4)
Range (Min, Max): 2.6-43.3
PM Component: Vegetation fire smoke
(95%) and motor vehicle emissions
(5%)
Monitoring Stations: 1
Correlation: PM25 r = 0.90
PM Increment: 10 pg/m
RR Estimate [Lower Cl, Upper Cl]
Symptoms attributable to asthma
Overall:!.010 (0.98,1.04)
Adults:!.027 (0.987,1.068)
Children:0.930 (0.966,1.060)
Using preventer: 1.022 (0.985,1.060)
Became symptomatic
Overall: 1.240 (1.106,1.39)
Adults: 1.277 (1.084,1.504)
Children: 1.247 (1.058,1.468)
Using preventer:1.317 (1.124,1.543)
Used Reliever
Overall: 1.010 (0.99, 1.04)
Adults: 1.026 (0.990, 1.063)
Children: 1.006 (0.960,1.055)
Using preventer: 1.035 (1.004,1.060)
Commenced Reliever
Overall: 1.132 (0.99, 1.29)
Adults: 1.199 (0.994, 1.446)
Children: 1.093 (0.906,1.319)
Using preventer: 1.194 (0.996,1.432)
Commenced Oral Steroids
Overall: 1.540 (1.01, 2.34)
Adults: 1.752 (1.008, 3.045)
Children: 1.292 (0.682, 2.448)
Using preventer: 1.430 (0.888, 2.304)
Asthma Attack
Overall: 1.030 (0.95, 1.12)
Adults: 1.08 (0.976, 1.202)
Children: 0.861 (0.710,1.044)
Using preventer:!.051 (0.939,1.175)
Exercise induced asthma
Overall: 0.980 (0.92, 1.05)
Adults: 0.988 (0.902, 1.081)
Children: 0.972 (0.844,1.119)
Using preventer: 1.026 (0.928,1.134)
Saw a health professional for asthma
Overall: 1.030 (0.85, 1.26)
Adults: 1.064 (0.794, 1.424)
Children: 0.998 (0.749,1.328)
Using preventer:0.924 (0.731,1.169)
Missed school or work due to asthma
Overall: 1.102 (0.941,1.290)
Adults: 1.135 (0.897, 1.435)
Children: 1.073 (0.862,1.333)
Using preventer: 1.025 (0.857,1.228)
Mean daily number of asthma
symptoms
Overall: 1.020 (1.001,1.031)
Adults: 1.027 (1.005,1.049)
Children: 1.016 (0.986,1.047)
Using preventer: 1.034 (1.011,1.058)
Mean Daily number of applications of
reliever
Overall: 1.020 (1.00,1.030)
Adults: 1.032 (1.008, 1.057)
Children: 1.002 (0.969,1.034)
Using preventer: 1.022 (1.001,1.043)
December 2009
E-153
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Just et al. (2002, 0354291
Period of Study:Apr 1996-Jun 1996
Location: Paris, France
Outcome: Incident and prevalent
episodes of asthma attacks, nocturnal
cough, wheeze, symptoms of irritation,
respiratory infections, supplementary
use of |32-agonists, Z-transformed peak
expiratory flow (PEF), daily PEF
variability
Age Groups: 7-15 yr old
Study Design: Cohort
N: 82 children
Statistical Analyses: Linear
regression, logistic regression, GEE
Covariates: Effects of time trend, day
of the week, weather, pollen levels
Season: Spring/summer
Lags Considered: 0, 0-2 mean, 0-4
mean
Pollutant: PM,0
Averaging Time: Daily
Mean (SD): 23.5 (8.4)
Range (Min, Max): 9.0, 44.0
Monitoring Stations: 5
Copollutant (correlation):
BS: 0.59
S02: 0.70
N02: 0.54
03: 0.21
Temp: 0.04
Humid: -0.41
PM Increment: 10 pg/m for binary
responses data (results that use odds
ratios [ORs])
Incident episodes of
1) Asthma
a lag 0:1.06 (0.61,1.83)
b 0-2 mean: 1.09 (0.48, 2.49)
c) 0-4 mean: 1.07 (0.44, 2.65)
2) Nocturnal cough
a lag 0:1.10 (0.88,1.37)
b 0-2 mean: 1.03 (0.77,1.37)
c) 0-4 mean: 1.11 (0.86,1.42)
3) Respiratory infections
a) lag 0:0.64 (0.35,1.15)
b) 0-2 mean: 0.74 (0.38,1.43)
c) 0-4 mean: 0.99 (0.58,1.68)
Prevalent episodes of
1) Asthma
a) lag 0:1.07 (0.72,1.59)
b) 0-2 mean: 1.18 (0.64, 2.17)
c) 0-4 mean: 1.16 (0.63, 2.13)
2) Nocturnal cough
a) lag 0:1.05 (0.83,1.34)
b) 0-2 mean: 1.10 (0.81,1.50)
c) 0-4 mean: 1.09 (0.79,1.52)
3) Respiratory infections
a) lag 0:1.17 (0.68, 2.03)
b) 0-2 mean: 1.31 (0.51,3.36)
c) 0-4 mean: 1.71 (0.71,4.12)
4) Eye irritation
a lag 0:1.18 (1.01,1.39)
b 0-2 mean: 1.28 (1.03,1.59)
c) 0-4 mean: 1.42 (1.12,1.80)
Analysis restricted to days with no
steroid use:
Incident episodes of
1) Eye irritation
a) lag 0:1.07 (0.66,1.71)
b) 0-2 mean: 0.83 (0.45,1.53)
c) 0-4 mean: 0.92 (0.46,1.83)
2) Throat irritation
a) lag 0:1.33 (0.66, 2.69)
b) 0-2 mean: 1.28 (0.58, 2.80)
c) 0-4 mean: 1.06 (0.38, 2.95)
3) Nose irritation
a lag 0: 0.74 (0.48,1.13)
b 0-2 mean: 0.76 (0.42,1.36)
c) 0-4 mean: 0.96 (0.53,1.73)
Prevalent episodes of
1) Eye irritation
a lag 0:1.20 (0.88,1.65)
b 0-2 mean: 1.71 (0.97,3.01)
c) 0-4 mean: 1.97 (1.03, 3.76)
2) Throat irritation
a lag 0:1.23 (0.83,1.82)
b 0-2 mean: 1.08 (0.68,1.73)
c) 0-4 mean: 0.91 (0.47,1.73)
3) Nose irritation
a) lag 0:1.20 (0.91,1.58)
b) 0-2 mean: 1.09 (0.78,1.52)
c) 0-4 mean: 1.09 (0.73,1.61)
Notes: The authors noted that incident
or prevalent wheeze was not correlated
with levels of any type of pollutant.
Also, they state no relationship was
observed between PEF variables and
levels of PM.
The authors also note that in a
multipollutant model assessing
independent effects of PM and 03 on
prevalent episodes of eye irritation
(mean 0-4), the PM parameter
decreased and was not significant
(p = 0.19).
December 2009
E-154
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Kulkarni et al. (2006,
0892571
Period of Study: Nov 2002-Dec
2003
Location: Leicester, United Kingdom
Outcome: Lung function by spirometry:
FVC, REV,, FEV,: FVC, FEF25.75
Age Groups: 8-15 yr
Study Design: Cross-sectional
N: 114 children, 64 provided sputum for
assessment of carbon content of
macrophages.
Statistical Analyses: Linear
regressions, Spearman rank
correlations. Mann-Whitney, Chi-square
and unpaired t tests were used to
compare results between asthmatic and
non asthmatic children
Covariates: BMI, sex, exercise, traffic
PM10
Dose-response Investigated? Yes
Statistical Package: SPSS
Pollutant: Primary PMi0 (pg/m )
concentration was modeled, and was
considered a covariate for carbon
content of macrophages. Carbon
content of alveolar macrophages was
the primary variable of interest.
Averaging Time: 1 yr
SOth(Median):
Children without asthma, 1.21
Children with asthma, 1.81
Range (Min, Max):
Children without asthma, 0.10, 2.17
Children with asthma, 0.17, 2.13
PM Component: Carbon content in
alveolar macrophages
Monitoring Stations: NR.
Copollutant (correlation):
Vs carbon content in macrophages
(increment, coefficient range])
-1.0|jg/m3, 0.1 [0.01-0.18]
PM Increment: 1.0|jg/m
% Change [Lower Cl, Upper Cl]:
Single pollutant model:
FEV,: -4.3 [-8.5, 0.2] p = 0.04
R2 = 0.06
Single pollutant model:
FVC: -1.2 [-5.6, 3.2] p = 0.59
R2 = 0.005
Single pollutant model:
FEF25.75: -8.6 [-17.3, 0.1] p = 0.05
R2 = 0.06
2 pollutant model with Macrophage
Carbon:
FEVi:PMio-2.9[-6.9, 1.2]
p = 0.17
FVC:PM100.1 [-4.4,4.6]
p = 0.96
FEF25.75:PMio-5.5 [-14.2, 3.1]
p = 0.21
Reference: Kuo, et al. (2002, 0363101
Period of Study: 1-yr period (yr not
specified)
Location: Central Taiwan
Outcome: Asthma (yes/no)
Age Groups: 13-16 yr
Study Design: Cohort
N: 12,926 total children
775 asthmatic children
8 junior high schools
Statistical Analyses: Pearson
correlation coefficients
Logistic regression
Covariates: Gender, age, residential
area, level of parental education,
number cigarettes smoked by family
members, incense burning in the home,
frequency of physical activities
Dose-response Investigated? No
Statistical Package: SAS 6.12
Lags Considered: Monthly avg at each
school
Pollutant: PM,0
Averaging Time: 1 h
Mean (SD):
School A: 59.7
School B: 65.3
School C: 84.3
School D: 59.2
School E: 75.3
School F: 60.2
School G: 54.1
School H: 69.0
Monitoring Stations: 8 (1 for each
school)
PM Increment: Dichotomized annual
avg:
<65.9 pg/m3
>65.9|jg/m3
OR Estimate [Lower Cl, Upper Cl]
lag:
Crude (outcome = asthma, yes/no)
<65.9 pg/m3:1 (ref)
> 65.9 pg/m3: 0.837 [NR]
Adjusted (outcome = asthma, yes/no)
<65.9 pg/m3:1 (ref)
> 65.9 pg/m3: 0.947 [0.640, 1.401]
Notes: Asthma prevalence was highest
in urban areas and lowest in rural areas
Pearson correlation between annual
PM levels at each school and asthma
prevalence at each school: 0.214
(p > 0.05)
Reference: Lagorio et al. (2006,
0898001
Period of Study:
May1999-Jun 1999
Jan 1999-Dec 1999
Location: Rome, Italy
Outcome: Lung function of subjects
(FVC and FEV,) with COPD, Asthma
Age Groups:
COPD: 50 to 80 yr
Asthma: 18to64yr
Study Design: Time series panel
N: COPDN = 11; Asthma N = 11
Statistical Analyses: Non-parametric
Spearman correlation
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD):
Overall: 42.8 (21. 8)
Spring: 36.9(10.8)
Winter: 49.0(28.1)
Range (Min, Max): (7.9, 123)
PM Component: NR
PM Increment: 1 pg/m3
They observed negative association
between ambient PM10 and respiratory
function (FVC and FEV,) in the COPD
panel. The effect on FVC was seen at
lag 24 h, 48 h, and 72 h. The effect on
FEV, was evident at lag 72 h. There
was no statistically significant effect of
PM,0 on FVC and FEV, in the asthmatic
and IHD panels.
GEE
Covariates: COPD and IHD: daily
mean temperature, season variable
(spring or winter), relative humidity, day
of week
Asthma: season variable, temperature,
humidity, and (3-2-agonist use
Season: Spring and winter
Dose-response Investigated? Yes
Statistical Package: STATA
Lags Considered: 1-3 days
Monitoring Stations: Two fixed sites:
(Villa Ada and Istituto superior di Sanita)
Copollutant (correlation):
N02r = 0.45
03r = -0.36
CO r = 0.55
S02r = 0.21
PM,0.25r = 0.61
PM25r = 0.93
P Coefficient (SE)
COPD
FVC(%) 24 h -0.66 (0.30)
48-h -0.75 (0.35)
72-h -0.94 (0.47)
FEV,(%) 24 h -0.37 (0.27)
48-h-0.58 (0.31)
72-h -0.87 (0.43)
Asthma
FVC(%)24h-0.12(0.24)
48-h -0.09 (0.29
72-h -0.08 (0.36
FEV,(%) 24 h -0.28 (0.28)
48-h-0.40 0.34
72-h -0.40 0.43
December 2009
E-155
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Lee, et al. (2007, 0930421
Period of Study: 2000-2001
Location: South-Western Seoul
Metropolitan area, Seoul, South Korea
Outcome: PEFR (peak expiratory flow
rate), lower respiratory symptoms (cold,
cough, wheeze)
Age Groups: 61-89 yr (77.8 mean age)
Study Design: Longitudinal panel
survey
N: 61 adults
Statistical Analyses: Logistic
regression model
Covariates: Temperature (Celsius),
relative humidity, age, season
Dose-response Investigated? No
Statistical Package: SAS 8.0
Lags Considered: 0-4 days
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 71.40 (30.69)
Percentiles: 25th: 43.47
SOth(Median): 74.92
75th: 87.54
Range (Min, Max):
26.23, 148.34
Monitoring Stations: 2
PM Increment: 10 pg/m
Effect Estimate [Lower Cl, Upper Cl]
lag:
PEFR (peak expiratory flow rate)
-0.39 (-0.63 to-0.14)
1day
relative odds of a lower respiratory
symptom (cold, cough, wheeze)
1.015(0.900,1.144)
1day
Reference: Lewis, et al. (2005,
081079)
Period of Study:
Winter 2001-spring 2002
Location: Detroit, Michigan, USA
Outcome: Poorer lung function
(increased diurnal variability and
decreased forced expiratory volume)
Age Groups: 7-11 yr
Study Design: longitudinal cohort study
N: 86 children
Statistical Analyses: descriptive
statistics and bivariate analyses of
exposures, multivariable regression
models that included interaction terms
between exposure measures and CS
use or, alternatively, presence of a URI,
multivariate analog of linear regression.
Covariates: sex, home location, annual
family income, presence of one or more
smokers in household, race, season
(entered as dummy variables), and
parameters to account for intervention
group effect.
Season: Winter 2001 (Feb 10-23),
spring 2001 (May 5-18), summer 2001
(Jul 14-27), fall 2001 (Sep 22-Oct 5),
winter 2002 (Jan 18-31), and spring
2002 (May 18-31).
Dose-response Investigated? No
Lags Considered: 1-2 days
3-5 days
Pollutant: PM,0
Averaging Time: 2 wk
Mean (SD): Eastside 23.0 (13.5)
Southwest 28.2 (16.1)
Range (Min, Max): 2.9, 70.9
PM Component: ("likely" in southwest
site) carbon and diesel emissions
Monitoring Stations: 2
Copollutant:
PM25 0.93
03 Daily mean 0.59
038-h peak 0.57
PM Increment: 19.1 pg/m
Lung function among children
reporting use of maintenance CSs
Diurnal variability FEVi
Lag 1:1.53 [-0.85, 3.90]
Lag 1:2.94 [-1.07, 6.96] PM10 + 03
Lag 2: 5.32 [0.32, 10.33]
Lag 2:13.73 [8.23, 19.23] PM10 + 03
Lag 3-5:1.46 [-2.21,5.13]
Lag 3-5: 3.30 [0.58, 6.02] PM,0 + 03
Lowest daily value FEV,
Lag 1:-0.28
Lag 1:-6.25
Lag 2:-2.21
Lag 2:-5.97
-2.34, 1.77]
-1.36] PM,o
-3.97 to-0.46]
-11.06 to-0.87] PM10 + 03
Lag 3-5: -2.58 [-7.65, 2.49]
Lag 3-5:1.98 [-0.38, 4.33] PM10 + 03
Lung function among children
reporting presence of URI on day of
lung function assessment
Diurnal variability FEVi
Lag 1:3.51
Lag 1:3.21
Lag 2:1.12
Lag 2: 5.40
-4.52,11.55]
-1.28,7.71] PM10 + 03
-4.62, 6.86]
-0.82, 11.62] PM10 + 03
Lag 3-5: 3.90 [0.34, 7.47]
Lag 3-5: 6.27 [0.07, 12.47] PM,0 + 03
Lowest daily value FEV,
Lag 1:-2.72 [-9.47, 4.03]
Lag 1:-13.11 [-21.59 to-4.62] PM,0 +
03
Lag 2: 0.24 [-5.10, 4.63]
Lag 2:-3.32 [-6.83, 0.18] PM,o + 03
Lag 3-5: -4.48 [-8.36, 0.60]
Lag 3-5: -3.17 [-5.82 to -0.51] PM10 + 03
December 2009
E-156
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Mar et al. (2004, 0573091
Period of Study: 1997-1999
Location: Spokane, Washington
Outcome: Respiratory symptoms
Age Groups: Adults: Ages 20-51 yr
Children: Ages 7-12 yr
Study Design: Time-series
N: 25 people
Statistical Analyses: Logistic
regression
Covariates: Temperature, relative
humidity, day of-the-wk
Statistical Package: STATA 6
Lags Considered: 0-2 days
Pollutant: PM,0
Mean (SD):
1997:24.5(18.5)
1998:20.6(12.3)
1999:16.8(8.0)
Monitoring Stations: 1 station
Copollutant (correlation):
PM10
PM1:r = 0.48
PM25:r = 0.61
PM10.25:r = 0.93
PM Increment: 10 pg/m
OR Estimate [Lower Cl, Upper Cl]
Adult Respiratory symptoms:
Wheeze: 1.01 [0.93,1.09] lag 0
0.98[0.91, 1.06] lag 1
0.99[0.92, 1.06] lag 2
Breath: 1.02[0.96,1.08] lag 0
1.01[0.97,1.06] lag 1
1.02(0.97, 1.06] lag 2
Cough: 0.96[0.88,1.05] lag 0
0.97[0.90, 1.04
0.98[0.92, 1.05
lag 2
Sputum: 1.01 [0.92,1.12] lag 0
0.99[0.91, 1.08] lag 1
1.00(0.93, 1.08] lag 2
Runny Nose: 0.98[0.93,1.04] lag 0
0.97[0.93, 1.02] lag 1;0.97[0.94, 1.01]
lag 2
Eye Irritation: 0.97[0.87,1.08] lag 0
0.97[0.88, 1.06
0.97[0.91, 1.04
lag 2
Lower Symptoms: 0.96[0.91, 1.02]
lagO
0.95[0.89, 1.00] lag 1
0.95[0.90, 1.00] lag 2
Any Symptoms: 0.97[0.93,1.02] lag 0
0.96[0.91, 1.00
0.95(0.91,0.99
1
lag 2
Children Respiratory symptoms:
Wheeze: 0.92(0.71,1.18] lag 0
0.89[0.64, 1.24] lag 1
0.95[0.69, 1.31] lag 2
Breath: 1.04(0.95,1.15] lag 0
1.04(0.95, 1.15
1.06(0.95, 1.19
lag 2
Cough: 1.09(1.02,1.16] lag 0
1.08(1.02,1.14] lag 1
1.10(1.02,1.18] lag 2
Sputum: 1.08(0.98,1.17] lag 0
1.07(0.98,1.17]lag1
1.07(0.98, 1.16] lag 2
Runny Nose: 1.08(1.00,1.16] lag 0
1.081.02, 1.15
1.081.02,1.14
lag 2
Eye Irritation: 1.06(0.74,1.51] lag 0
0.94(0.70, 1.26] lag 1
0.99[0.88, 1.12]lag2
Lower Symptoms: 1.07(1.00,1.14]
lagO
1.06(0.98, 1.15
1.07(0.95, 1.19
lag 2
Any Symptoms: 1.07(1.02,1.11] lag 0
1.09(1.03,1.15] lag 1
1.10(1.03,1.17] lag 2
December 2009
E-157
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Mar et al. (2005, 0875661
Period of Study: 1999-2001
Location: Seattle, Washington
Outcome: Pulmonary function (arterial Pollutant: PMi0
oxygen saturation) and cardiac function
(heart rate and blood pressure) Averaging Time: 24-h avg
Study Design: Time series
N:88
Statistical Analyses: Linear logistic
regression
Age Groups: >57
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl)
Lag
Indoor
Systolic: 0.92 (-0.95, 2.78) 0
Diastolic: 0.63 (-0.29, 1.56)0
Outdoor
Systolic:-0.10 (-1.37,1.18)0
Diastolic: -0.03 (-0.79, 0.73) 0
Nephelometer
Systolic: 0.35 (-0.91,1.61)0
Diastolic:-0.12 (-0.91, 0.67)0
% Increase between heart rate and
PM10 exposure for people >67
PM10
Indoor: 0.02 (-0.54, 0.58) 0
Outdoor:-0.48 (-1.03, 0.06)0
Nephelometer:-0.31 (-0.76,0.14)0
Reference: McCormack et al. (2009,
1998331
Period of Study: Sep 2001-Apr 2004
Location: East Baltimore, Maryland
Outcome: Asthma symptoms
Study Design: Panel
Statistical Analysis: Chi-square,
Student t-test, Negative binomial
regression models with GEE, Logistic
regression with GEE
Statistical Package: StataSE
Age Groups: Asthmatic children aged
2-6 yr
Pollutant: PMio.2.5, PM2.5
Averaging Time: 3 days
Mean (SD) Unit:
PMio-25:17.4 + 21.2 pg/m3
PM25: 40.3 + 35.4 pg/m3
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: 10|jg/m
Relative Risk (Min Cl, Max Cl) Lag
Bivariate Models, PM10.2.5
Cough, wheezing, chest tightness:
1.05(0.99-1.10), p = 0.08
Slow down: 1.08 (1.03-1.13). p< 0.01
Symptoms with running:
1.03(0.97-1.09). p = 0.39
Nocturnal symptoms:
1.06(1.01-1.11), p = 0.03
Limited speech:
1.11 (1.05-1.18), p< 0.01
Rescue medication use:
1.06(1.02-1.11), p< 0.01
Bivariate Models, PM25
Cough, wheezing, chest tightness:
1.01 (0.98-1.05), p = 0.41
Slowdown: 1.00 (0.97-1.04), p = 0.85
Symptoms with running:
1.04(1.01-1.07), p = 0.14
Nocturnal symptoms:
1.02(0.98-1.05), p = 0.37
Limited speech:
1.01 (0.95-1.07), p = 0.33
Rescue medication use:
1.03(1.00-1.60), p = 0.06
Multivariate Models, PM10-25
Cough, wheezing, chest tightness: 1
.06(1.01-1.12), p = 0.02
Slow down: 1.08 (1.02-1.14), p = 0.01
Symptoms with running:
1.00(0.94-1.08), p = 0.81
Nocturnal symptoms:
1.08(1.01-1.14), p = 0.02
Limited speech:
1.11 (1.03-1.19), p< 0.01
Rescue medication use:
1.06(1.01-1.10), p = 0.02
Multivariate Models, PM25
Cough, wheezing, chest tightness:
1.03(0.99-1.07), p = 0.18
Slow down:
1.04(1.00-1.09), p = 0.06
Symptoms with running:
1.07(1.02-1.11), p< 0.01
Nocturnal symptoms:
1.06(1.01-1.10), p = 0.01
Limited speech:
1.07(1.00-1.14), p = 0.04
Rescue medication use:
1.04(1.01-1.08), p = 0.04
December 2009
E-158
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Mortimer et al. (2008,
1872801
Period of Study: 1989-2000
Location: Joaquin Valley, California
Outcome: Respiratory Symptoms,
Decreased lung function
Study Design: Time series
Pollutant: PM,0
Averaging Time: 24-h avg
Copollutants (correlation):
Statistical Analyses:
Deletion/Substitution/Addition algorithm C0: r = °'05
(GEE) N02:r = 0.30
Logistic linear regression 0 • r = 0 39
Age Groups: 6-11
Increment: NR
p(SE):
FVC:
PM,o (age 0-3 yr): 0.0121 (0.0037)
FEV,:PM10 (age 0-3 yr): 0.0102
(0.0034)
PEF:
PM10 (Mother smoked during
pregnancy):
-0.0102(0.0039)
Reference: Mortimer et al. (2002,
0302811
Period of Study: Jun-Aug 1993
Location: Eight urban areas of the
U.S.: Bronx and East Harlem, NY
Baltimore, MD
Washington, DC
Detroit, Ml
Cleveland, OH
Chicago, IL
and St. Louis, MO.
Outcome: peak expiratory flow rate
(PEFR) and symptoms
Age Groups: 4-9 yr
Study Design: Cohort study
N: 846 children with a history of asthma
Statistical Analyses: Mixed linear
models and GEE
Covariates: Day of study, previous 12-
h mean temperature, urban area, diary
number, rain in the past 24 h
Season: Summer
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0,1,2,3,4,5,6,1-
5 avg, 1-4 avg, 0-4 avg, 0-3 avg
Pollutant: PM10
Averaging Time: 24 h
Mean (SD): 53
Monitoring Stations: NR
Copollutant (correlation):
8-havg03: r = 0.51
PM Increment: 20 pg/m3
Effect Estimate [Lower Cl, Upper Cl]:
(RR estimates are odds ratios for
incidence of morning asthma symptoms
using the avg of lag 1-2)
3 urban areas (DE, CL, CH)
Single pollutant: OR = 1.26 (1.00-1.59)
03+PM,o: OR = 1.25 (0.97-1.61)
03+S02 +N02+PM,o: OR = 1.14 (0.80-
Reference: Moshammer and
Neuberger (2003, 0419561
Period of Study: 2000-2001
Location: Linz, Austria
Outcome: Lung Function: FVC, FEV,,
MEF25, MEF50, MEF75, PEF, LQ Signal,
PAS Signal
Age Groups: Ages 7 to 10
Study Design: Case-crossover
N: 161 children
1898-2120 "half-h means"
Statistical Analyses: Correlations
Regression Analysis
Covariates: Morning, evening, night
Season: Spring, summer, winter, fall
Dose-response Investigated? No
Pollutant: PM10
Averaging Time: 8 h
Daily Means
Mean (SD): 23.13 (20.08)
Range (Min, Max): (NR, 190.79)
Monitoring Stations: 1
Copollutant (correlation):
LQ = 0.751
PAS = 0.406
Notes: "Acute effects of 'active particle
surface' as measured by diffusion
charging were found on pulmonary
function (FVC, FEVi,MEF50) of
elementary school children and on
asthma-like symptoms of children who
had been classified as sensitive."
Reference: Moshammer et al. (2006,
0907711
Period of Study: 2000-2001
Location: Linz, Austria
Outcome: Respiratory symptoms and
decreased lung function
Age Groups: Children ages 7-10
Study Design: Time-series
N: 163 children
Statistical Analyses: GEE model
Covariates: Sex, age, height, weight
Dose-response Investigated? NR
Statistical Package: NR
Lags Considered: 1
Pollutant: PM10
Averaging Time: 8 h
Mean (SD):
Maximum 24 h: 76.39
Annual avg: 19.06
Percentiles:
8-h mean 25th: 14.39
8-h mean SOth(Median): 24.85
8-h mean 75th: 38.82
Monitoring Stations: 1 station
Copollutant (correlation):
PM,:r = 0.91
PM25:r = 0.93
N02:r = 0.62
PM Increment: 10 pg/m3
% change in Lung Function per
Af\ nnltVkt
lu |jg/m3
FEV:0.11
FVC: 0.06
FEV05:-0.19
MEF75%: -0.30
MEF50%: -0.36
MEF25%: 0.41
PEF: 0.22
% change in Lung Function per IQR
FEV: -0.27
FVC: -0.07
FEV05:-0.47
MEF75%: -0.74
MEF50%: -0.86
MEF25%: 0.98
PEF: -0.54
December 2009
E-159
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Neuberger et al. (2004,
0932491
Period of Study: Sep 1999-Mar 2000
Location: Vienna, Austria
Outcome: Ratio measure: Time to peak
tidal expiratory flow divided by total
expiration time (i.e., tidal lung function,
a surrogate for bronchial obstruction)
Age Groups: 3.0-5.9 yr (preschool
children)
Study Design: Longitudinal prospective
cohort
N: 56 children
Statistical Analyses: Mixed models
linear regression, with autoregressive
correlation structure
Covariates: Age, sex, respiratory rate,
phase angle, temperature,
kindergarten, parental education,
observer (also in sensitivity analyses:
height, weight, cold/sneeze on same
day, heating with fossil fuels, hair
cotinine, number of tidal slopes used to
measure tidal lung function)
Dose-response Investigated? No
Statistical Package: SAS 8.0
Lags Considered: 0
Pollutant: PM,0
Averaging Time: 24 h
Copollutant (correlation):
PM2.5 (r = 0.94) in Vienna
PM Increment: Interquartile range (NR)
Change in mean associated with an
IQR increase in PM (p-value)
lag
-1.067(0.241)
lagO
Reference: Neuberger et al. (2004,
0932491
Period of Study: Oct. 2000-May 2001
Location: Linz, Austria
Outcome: Forced oscillatory resistance
(at zero Hz), FVC, FEV,, MEF25, MEF50,
MEF75, PEF
Age Groups: 7-10 yr
Study Design: Longitudinal prospective
cohort
N: 164 children
Statistical Analyses: Mixed models
linear regression with autoregressive
correlation structure
Covariates: sex, time and individual
Season: Oct-May
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0-7
Pollutant: PM,0
Averaging Time: 24 h
Monitoring Stations: 1
PM Increment: 1 pg/m
Notes: No significant associations
between PM10 and the metrics of lung
function were reported. The authors
state they only reported significant
associations, so results are assumed to
be null.
Reference: Odajima et al. (2008,
1920051
Period of Study: Apr 2003-Mar 2004
Location: Fukuoka, Japan
Outcome: PEF
Study Design: Panel/Field
Statistical Analysis: GEE
Statistical Package: SAS
Covariates: Age, sex, growth index,
temperature, N02, 03
Age Groups: Asthmatic children, 4-11
yrold
Pollutant: PM,0
Averaging Time: 3 h
Mean (SD) Unit:
Warmer months, 5-8 am
SPM:40.7|jg/m3
N02:15.2 ppb
03:17.7ppb
Warmer months, 7-1 Opm
SPM:41.5|jg/m3
N02: 20.0 ppb
03:28.1 ppb
Colder months, 5-8am
SPM:32.6|jg/m3
N02: 20.5 ppb
03:17.5 ppb
Colder months, 7-1 Opm
SPM:34.7|jg/m3
N02: 28.0 ppb
03:19.4 ppb
Range (Min, Max):
Increment: 10|jg/m
Relative Risk (Min Cl, Max Cl)
Lag
Apr-Sep, morning sample, multi-
pollutant::
SPM, 5am-8am: -0.6 (-1.228, 0.028)
SPM, 2am-5am: -0.78 (-1.399, -0.161)
SPM, 11pm-2am:-0.612 (-1.180,-
0.045)
SPM, 8pm-11am:-0.732 (-1.318,-
0.145)
03,5am-8am:-0.575 (-1.569, 0.419
03, 2am-5am: -0.052 (-0.997, 0.893
03,11pm-2am:-0.305 (-1.269, 0.658)
03,8pm-11am:-0.416 (-1.283, 0.451)
N02, 5am-8am:-0.3 (-2.246,1.645)
N02,2am-5am: 0.265 (-1.354, 1.885)
N02, 11pm-2am: -0.187 (-1.447, 1.073)
N02,8pm-11am: 0.432 (-0.689, 1.553)
Single-pollutant model:
SPM, 5am-8am:-0.67 (-1.236,-0.104)
SPM, 2am-5am:-0.761 (-1.328,-0.194)
SPM, 11pm-2am:-0.661 (-1.159,-
December 2009
E-160
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Warmer months, 5-8am
SPM:(11.0, 126.0)
N02: (1.3, 44.7)
03: (0.3, 52.3)
Warmer months, 7-1 Opm
SPM: (8.3, 191.3)
N02: (3.0, 51.3)
03: (1.3, 71.3)
Colder months, 5-8am
SPM: (9.0, 160.0)
N02: (1.3, 44.0)
03: (0.6, 48.7)
Colder months, 7-1 Opm
SPM: (10.3, 131.0)
N02: (3.6, 49.0)
03: (1.0, 60.0)
Copollutant (correlation):
Warmer months (24-h mean):
03:r = 0.32
N02:r = 0.30
Colder months (24-h mean):
03:r = -0.02
N02:r = 0.45
0.163)
SPM, 8pm-11am:-0.714 (-1.212, -
0.215)
Evening sample, multi-pollutant model
SPM, 7pm-10pm: -0.449 (-1.071, 0.174)
SPM, 4pm-7pm:-0.434 (-1.122, 0.254)
SPM, 1pm-4pm: -0.415 (-1.015, 0.184)
SPM, 10am-1pm:-0.522 (-1.199, 0.155)
03,7pm-10pm:-0.22 (-1.171, 0.731)
03, 4pm-7pm: -0.118 (-0.809, 0.574)
03, 1pm-4pm:-1.086 (-0.888, 0.516)
03,10am-1pm:-0.315 (-1.123, 0.493)
N02, 7pm-10pm: 0.296 (-0.806, 1.397)
N02,4pm-7pm: 0.220 (-0.818, 1.258)
N02, 1pm-4pm: 0.438 (-0.568, 1.444)
N02, 10am-1pm: 0.536 (-0.546, 1.617)
Single-pollutant model:
SPM, 7pm-10pm: -0.449 (-0.956, 0.058)
SPM, 4pm-7pm:-0.449 (-1.029, 0.131)
SPM, 1pm-4pm: -0.414 (-0.943, 0.115)
SPM, 10am-1pm: -0.486 (-1.051, 0.079)
Oct-Mar, morning sample, multi-
pollutant::
SPM, 5am-8am: 0.290 (-0.279, 0.859)
SPM, 2am-5am: 0.431 (-0.173,1.036)
SPM, 11pm-2am: 0.304
SPM, 8pm-11am: 0.010
-0.311,0.919)
-0.523, 0.543)
03, 5am-8am: -0.415 (-1.568, 0.738)
03,2am-5am:-0.046 (-1.245,1.153)
03,11pm-2am: 0.004 (-1.265,1.273)
03,8pm-11am:-0.470 (-2.017,1.077)
N02, 5am-8am: -0.319 (-2.269,1.631)
N02,2am-5am: 0.262 (-1.777, 2.300)
N02, 11pm-2am: 0.609 (-1.132, 2.350)
N02, 8pm-11am: 0.155 (-1.545, 1.856)
Single-pollutant model:
SPM, 5am-8am: 0.308 (-0.189, 0.805
SPM, 2am-5am: 0.485 (-0.026, 0.996
SPM, 11pm-2am: 0.486 (-0.049, 1.022)
SPM, 8pm-11am: 0.100 (-0.414, 0.613)
Evening Sample, Multi-pollutant Model
SPM, 7pm-10pm: 0.059 (-0.397, 0.515)
SPM, 4pm-7pm: 0.360 (-0.093, 0.812)
SPM, 1pm-4pm: 0.357 (-0.157, 0.871)
SPM, 10am-1pm: 0.169 (-0.394, 0.731)
03,7pm-10pm:-0.656 (-2.394, 1.083)
03,4pm-7pm: 0.046 (-1.140, 1.232)
03, 1pm-4pm: 0.164 (-1.038, 1.365)
03,10am-1pm: 0.665 (-0.613,1.942)
N02, 7pm-10pm: -0.415 (-2.444, 1.613)
N02, 4pm-7pm: -0.144
N02,1pm-4pm:-0.181
-1.490, 1.202)
-1.821, 1.459)
N02,10am-1pm: 0.194 (-1.503,1.890)
Single-pollutant model:
SPM, 7pm-10pm: 0.071 (-0.388, 0.529)
SPM, 4pm-7pm: 0.318 (-0.123, 0.758
SPM, 1pm-4pm: 0.317 (-0.171, 0.804
SPM, 10am-1pm: 0.112 (-0.412, 0.636)
December 2009
E-161
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Peacock et al. (2003,
0420261
Period of Study: Nov 1996-Feb 1997
Location: Southern England
Outcome: Reduced peak expiratory
flow rate (PEFR)
Age Groups: 7-13 yr
Study Design: Time-series
N:179
Statistical Analyses: GEE, multiple
regression
Covariates: Day of the week, 24-h
mean outside temperature.
Season: Winter
Dose-response Investigated? No
Statistical Package: STATA
Lags Considered: Same day, lag 1, lag
2, 5-day ma
Pollutant: PM,0
Averaging Time: Daily
Mean (SD):
Rural (nationally validated) 21.2 (11.3)
Rural (locally validated) 18.7 (11.3)
Urban 118.4 (9.8)
Urban 2 22.7 (10.6)
Percentiles:
10th
Rural (nationally validated) 11.0
Rural (locally validated) 9.0
Urban 1 10.5
Urban 2 12.5
90th
Rural (nationally validated) 33.0
Rural (locally validated) 32.5
Urban 1 32.0
Urban 2 36.0
Range (Min, Max):
Rural (nationally validated) 7.0, 82.0
Rural (locally validated) 6.6, 87.9
Urban 1 4.7, 62.8
Urban 2 6.7, 63.7
Monitoring Stations: 3
Co pollutants:
N02
03
S02
S042"
Increment: 10|jg/rrf
Odds Ratio (Lower Cl, Upper Cl)
Lag
Change in PEFR
Community
-0.04 (-0.11, 0.03)0
0.03 (-0.04, 0.05) 1
-0.01
-0.10
-0.07, 0.05) 2
-0.25, 0.05)
0-4 avg
Local
-0.01 (-0.06, 0.03) 0
0.04(0.01,0.08)1
0.01 (-0.04, 0.05) 2
0.04 (-0.05, 0.13)
0-4 avg
20% decrease in PEFR
All children
1.012(0.992,1.031)0
1.016(0.995, 1.036)1
1.013
1.037
1.000, 1.025
0.992, 1.084
0-4 avg
Wheezy Children Only
1.016(0.986,1.047)0
1.030
1.018
1.001, 1.060
0.995, 1.041
1.114(1.057,1.174)
0-4 avg
Reference: Peled, et al. (2005,
1560151
Period of Study: 5-6 wk between Mar-
Jun 1999 and Sep-Dec 1999.
Location: Ashdod, Ashkelon and
Sderot, Israel
Outcome: Reduced peak expiratory
flow (PEF)
Age Groups: 7-10 yr
Study Design: Nested cohort study
N:285
Statistical Analyses: Time series
analysis, generalized linear model,
GEE, one-way ANOVA
Covariates: seasonal changes,
meteorological conditions and personal
physiological, clinical and
socioeconomic measurements
Season: Spring, fall
Dose-response Investigated? No
Statistical Package: STATA
Pollutant: PM,0
Averaging Time: Daily
Mean:
Ashkelon: 67.1
Sderot: 52.9
Ashdod: 31.0
PM Component: Local industrial
emissions, desert dust, vehicle
emissions and emissions from two
electric power plants
Monitoring Stations: 6
Copollutant: PM25
PM Increment: 1 pg/m
P coefficient (SE) [96% Cl]
Sderot:
PM10 MAX:-0.34 (0.41) [-1.16, 0.46]
PM10 MAX x sin(u2 day): 0.84 (0.22)
[0.405, 1.28]
PM,o MAX x cos (u1 day):-1.61 (0.41)
[-2.43, 0.79]
PM,o MAXx sin (u1 day): 0.44 (0.120)
[-0.68-0.21]
In Sderot, an interaction between PM10
and the sequential day were
significantly associated with PEF.
Reference: Pitard, et al. (2004,
0874331
Period of Study: 732 days (Jul
1998-Jun2000)
Location: City of Rouen, France
Outcome: Respiratory drug sales
Age Groups: 0-14,15-64, 65-74, over
75 yr
Study Design: Ecological time-series
N: 106,592
Statistical Analyses: Generalized
additive model
Covariates: Days of the weeks, trend,
seasonal variations, influenza
epidemics, meteorological variables,
holidays
Dose-response Investigated? No
Statistical Package: S-plus
Lags Considered: 0 to 10 days
Pollutant: PM,0
Averaging Time: Daily
Mean (SD): 16.7 (13.3)
Percentiles:
25th: 8.00
SOth(Median): 13.0
75th: 20
Range (Min, Max): 2.00,126
Monitoring Stations: 2
Copollutant (correlation):
S02 (0.39)
N02(0.61)
PM Increment: 10 pg/m
Percent increase in sales of anti-
asthmatics and bronchodilators (Lower
Cl, Upper Cl)
6.2(2.4,10.1)
lag 10 days
Percent increase in sales of cough and
cold preparation for children under 15 yr
of age (Lower Cl, Upper Cl)
9.2(5.9,12.6)
10 days
December 2009
E-162
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Preutthipan et al. (2004,
0555981
Period of Study: 31 days (school
days) from Jan-Feb 1999
Location: Mae Pra Fatima School,
central Bangkok, Thailand
Outcome: Decreases in peak
expiratory flow rates (PEFR),
respiratory symptoms including wheeze,
shortness of breath, runny/stuffed nose,
sneezing, cough, phlegm, and sore
throat
Age Groups: Third to ninth grade
Study Design: Time- Series
N: 133 children (93 asthmatics, 40
nonasthmatics)
Statistical Analyses: For continuous
data, an unpaired t-test or Mann-
Whitney U test was used. For
categorical data, the chi-square test or
Fisher's exact test was used. One-way
analysis of covariance (ANCOVA) was
used to compare avg daily reported
respiratory symptoms, diurnal PEFR
variability, and the prevalence of PEFR
decrements between groups of days.
Covariates: Age, sex, weight, height,
parents smoking, person smoking in
home, daily number of household
cigarettes, air-conditioned bedroom,
fuel used for cooking (charcoal, gas),
distance from home to main road
Dose-response Investigated? No
Lags Considered: Up to 5 days
Pollutant: PM,0
Averaging Time: Daily
Mean (SD): 111.0 (39)
Range (Min, Max): 46, 201
Monitoring Stations: 1
Copollutant:
S02
CO
03
PM Increment: Authors classified
exposure according to High and Low
PM10 days:
High = >120|jg/m3
Low = <120|jg/m3
Daily reported respiratory symptoms
and diurnal PEFR variability as
classified by concurrent days with high
vs.. lowPM10
Mean % reporting (SEM)
Asthmatics: High PM10
Wheeze/shortness of breath =
21.3(1.4)
Runny/stuffed nose or sneezing =
42.3(1.8)
Cough = 59.9 (1.9)
Phlegm = 60.5 (2.3)
Sore throat = 23.7 (1.5)
Any respiratory symptoms = 72.2 (3.2)
Diurnal PEFR variability = 3.0 (0.4)
Asthmatics: Low PMi0
Wheeze/shortness of breath =
19.3(1.3)
Runny/stuffed nose or sneezing =
35.8(1.6)
Cough = 59.1 (1.6)
Phlegm = 58.6 (2.0)
Sore throat = 21.0(1.4)
Any respiratory symptoms = 63.8 (2.8)
Diurnal PEFR variability = 2.8 (0.3)
Nonasthmatics: High PMi0
Wheeze/shortness of breath =
11.7(1.4)
Runny/stuffed nose or sneezing =
40.9 (2.5)
Cough = 50.4 (2.6)
Phlegm = 50.2 (2.5)
Sore throat = 27.1 (1.7)
Any respiratory symptoms = 67.8 (3.7)
Diurnal PEFR variability = 2.4 (0.4)
Nonasthmatics: Low PM10
Wheeze/shortness of breath = 9.3 (1.2)
Runny/stuffed nose or sneezing =
33.1 (2.2)
Cough = 54.0 (2.2)
Phlegm = 49.9 (2.2)
Sore throat = 23.9 (1.5)
Any respiratory symptoms = 56.4 (3.2)
Diurnal PEFR variability = 2.1 (0.4)
Notes: None of the daily reported
respiratory symptoms had significant
direct correlations with daily PM10
levels, according to the authors.
December 2009
E-163
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Rabinovitch et al. (2004,
0967531
Periods of Study: Nov 1999-Mar 2000
Nov 2000-Mar 2001
Nov 2001-Mar 2002
Location: Denver, Colorado
Outcome: Respiratory symptoms,
Asthma symptoms (cough and
wheeze), Upper respiratory symptoms
Study Design: Time-series panel
Statistical Analyses: Logistic linear
regression
Age Groups: 6-12
Pollutants: PM10
Averaging Time: 24-h avg
Mean (SD): 28.1 (13.2)
Range (Min, Max):
(6.0, 102.0)
Copollutant:
CO
NO,
Increment: 1 pg/rrf
AM:-0.010 (0.008)
PM:-0.011 (0.010)
Odds Ratio (Lower Cl, Upper Cl)
Lag
1.016(0.911,1.133)
0-3 avg.
OR for respiratory symptoms and PMi0
exposure for children age 6-12
Asthma exacerbation:
1.00(0.75, 1.25) 0-3 avg
Medication: 0.85 (0.75, 0.95) 0-3 avg
Previous night's symptoms:
1.10(1.00,1.20) 0-3 avg
Current day's symptoms:
1.00(0.90,1.10) 0-3 avg
% Increase (Lower Cl, Upper Cl)
Lag
% Increase in FEV, or PEF and PM,0
exposure for children age 6-12
AM FEV,:-0.01 (-0.02, 0.01) 0-3 avg
PMFEV,:-0.02 (-0.03, 0.02) 0-3 avg
AM PEF: -0.025 (-0.035, 0.02) 0-3 avg
PM PEF: 0.00 (-0.03, 0.03) 0-3 avg.
Reference: Renzetti et al. (2009,
1998341
Period of Study: Jun 2006-Jul 2006
Location: Pescara and Ovindoli, Italy
Outcome: Airway inflammation and
function
Study Design: Panel
Covariates: NR
Statistical Analysis: Student T-test,
Pearson's correlation coefficients
Statistical Package: StatView
Age Groups: Children, mean age
9.9 yr
Pollutant: PM10
Averaging Time: Daily
Mean (SD) Unit:
Urban: 56.9 ±13.1 pg/m3
Rural: 13.8 + 5.6 pg/m3
Copollutant (correlation): NR
All results are presented in Fig format.
December 2009
E-164
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Rojas-Martinez et al. (2007, Outcome: Lung function: FEVi, FVC,
FEF25.75%
0910641
Period of Study: 1996-1999
Location: Mexico City, Mexico
Age Groups: Children 8 yr old at time
of cohort recruitment
Study Design: School-based "dynamic'
cohort study
N: 3170 children
14,545 observations
Statistical Analyses: Three-level
generalized linear mixed models with
unstructured variance-covariance matrix
Covariates: Age, body mass index,
height, height by age, weekday spent
outdoors, environmental tobacco
smoke, previous-day mean air pollutant
concentration, time since first test
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-1 days
Pollutant: PM,0
Averaging Time: 24 h, 6 mo
Mean (SD): 24-h averaging
Tlalnepantla: 66.7 (35.6)
Xalostoc: 96.7 (49.4)
Merced: 79.3 (40.8)
Pedregal:53.4(31.9)
Cerrode la Estrella: 69.6 (35.3)
6-mo averaging
Mean: 75.6
Percentiles: 6-mo averaging
25th: 55.8
SOth(Median): 67.5
75th: 92.2
Monitoring Stations: 5 sites for PIvl-
10 for other pollutants
Copollutant:
03
N02
PM Increment: IQR
PM10, 6-LC: 36.4
GIRLS
One-pollutant model
FVC:-39 [-47:-31]
FEV: -29 [-36: -21]
FEF25.75%:-17[-36:1]
FEV,/FVC: 0.12 [0.07: 0.17]
Two-pollutant model
PM10, 6-LC&03
FVC: -30 [-39: -22]
FEV:-24 [-31:-16]
FEF25.75%:-9[-26:9]
FEV,/FVC: 0.10 [0.06: 0.15]
PM10, 6-LCSN02
FVC:-21 [-30:-13]
FEV:-17 [-25:-8]
FEF25.75%: -23 [-43: -4]
FEV,/FVC: 0.07 [0.02: 0.13]
Multipollutant model
PM10, 6-LC, 03, & N02
FVC: -14 [-23: -5]
FEV:-11 [-20:-3]
FEF25-75%: -7 [-27: 12]
FEV,/FVC: 0.08 [0.03: 0.13]
BOYS
One-pollutant model
FVC:-33 [-41:-25]
FEV:-27 [-34:-19]
FEF25.75%:-18[-34:-2]
FEV,/FVC: 0.04 [-0.01: 0.09]
Two-pollutant model
PM10, 6-LC&03
FVC:-28 [-36:-19]
FEV:-22 [-30:-15]
FEF25.75%:-10[-27:7]
FEV,/FVC: 0.04 [-0.01: 0.09]
PM10, 6-LCSN02
FVC: -16 [-26: -7]
FEV:-19 [-27:-10]
FEF25.75%: -26 [-44: -9]
FEV,/FVC: 0.005 [-0.06: 0.05]
Multipollutant model
PM10, 6-LC, 03, & N02
FVC: -12 [-22: -3]
FEV:-15 [-23:-6]
FEF25.75%:-12[-30:6]
FEV,/FVC: -0.002 [-0.06: 0.05]
Long-term exposure to 03, PM10, and
N02 is associated with decrements in
FVC and FEVi growth in Mexico City
schoolchildren. In a multipollutant
model, PM,o (-12%), 03 (-9%), and N02
(-41%) each contribute independently
and statistically significantly to
diminished FVC growth. For FEV,,
however, the multipollutant model
indicates that only PM10 (-15%) and
N02 (-25%) each contribute
independently and statistically
significantly to diminished FEVi growth.
December 2009
E-165
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Location: Hamilton, Canada
Reference: Sahsuvaroglu et al. (2009, Outcome: Asthma symptoms
1909831
Study Design: Panel
Period of Study: 1994-1995
Covariates: Neighborhood income,
dwelling value, state of housing,
deprivation index, smoking
Statistical Analysis: Logistic
regressions
Statistical Package: SPSS
N:6388
Age Groups:
Children in grades 1 and 8
Pollutant: PM,0
Averaging Time: 3-yr avg
Avg:
All Subjects: 20.90 pg/m3
Boys: 20.88 pg/m3
Girls: 20.92 pg/m3
Range:
All Subjects: 26.98
Boys: 26.98
Girls: 20.10
Copollutant (correlation):
NOxTheissen: 0.083
SOJheissen: -0.021
03Theissen: -0.251
N02Kriged:0.126
N02LUR: 0.072
Increment: NR
Odds Ratio (96%CI) for copollutant
model PMIOSpline and N02LUR
All Girls: 1.063 (0.969-1.666)
Older Girls: 1.058 (0.918-1.219)
Odds Ratio (96%CI) for copollutant
model PMIOSpline and N02LUR,
S02Theissen and OSTheissen
All Girls: 1.045 (0.943-1.158)
Older Girls: 1.044 (0.891-1.225)
Regression coefficients (96%CI)
between non-allergic asthma and
PMIOSpline exposure
All Children: 1.043 (0.996-1.092)
Younger Children: 1.011 (0.929-1.100)
Older Children: 1.073 (1.013-1.136)
All Girls: 1.069 (0.999-1.144)
All Boys: 1.024 (0.962-1.091)
Younger Girls: 1.065 (0.943-1.203)
Younger Boys: 0.962 (0.853-1.085)
Older Girls: 1.072 (0.984-1.169)
Older Boys: 1.075 (0.995-1.160)
Reference: Sanchez-Carrillo et al.
(2003, 0984281
Period of Study: 1996-1997
Location: metropolitan Mexico City,
Mexico
Outcome: Upper respiratory symptom
indicator (wet cough, sore throat,
hoarseness, nose dryness, and head
cold); Lower respiratory symptom
indicator (dry cough, lack of air, and
chest sounds); and Ocular symptom
indicator (eye irritation, eye itch, eye
burning, teary eyes, red eyes, and eye
infection)
Age Groups: All ages
Study Design: Cohort
N: 151,418 interviews
Statistical Analyses: Logistic
regression models
Covariates: Sex, age, education,
cigarette smoking, season, emergency
episode mass media report,
temperature, and relative humidity
Dose-response Investigated? Yes
Statistical Package: NR
Lags Considered: 1
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD):
Northeast: 132 (52)
Northwest: 87 (46)
Central: 85 (37)
Southeast: 79 (35)
Southwest: 55 (28)
Range (Min, Max):
Northeast: (34-269)
Northwest: (10-275)
Central: (9-319)
Southeast: (14-225)
Southwest: (12-264)
Monitoring Stations: Up to 32
Copollutant (correlation):
03:r = 0.067
03 8: 00-18: 00 h:r = 0.075
S02:r = 0.265
N02:r = 0.265
Effect Estimate [Lower Cl, Upper Cl]:
PMio quartiles:
10.04-52.62 (ref) 52.63-73.58
Upper respiratory indicator:
1.02(0.99-1.06)
Lower respiratory indicator:
1.04(0.99-1.09)
Ocular indicator:
0.99(0.95-1.03)73.59-101.91
Upper respiratory indicator:
1.07(1.03-1.10)
Lower respiratory indicator:
1.09(1.04-1.14)
Ocular indicator:
0.89(0.86-0.92)101.92-318.80
Upper respiratory indicator:
0.93 (0.90-0.97)
Lower respiratory indicator:
1.03(0.98-1.08)
Ocular indicator: 0.84 (0.81-0.87)
Northeast - 2nd quartile
Upper respiratory indicator:
0.354(0.112-1.222)
Lower respiratory indicator:
0.215(0.040-1.160)
Ocular indicator: 1.080 (0.915-1.274)
3rd quartile
Upper respiratory indicator:
0.118(0.039-0.356)
Lower respiratory indicator:
0.126(0.023-0.690)
Ocular indicator: 1.228 (0.720-2.095)
4th quartile
Upper respiratory indicator:
0.095 (0.034-0.267)
Lower respiratory indicator:
0.119(0.026-0.549)
Ocular indicator: 0.878 (0.619-1.246)
Northwest - 2nd quartile
Upper respiratory indicator:
0.990(0.898-1.090)
Lower respiratory indicator:
1.246(1.087-1.429)
Ocular indicator: 1.218 (0.808-1.834)
3rd quartile
Upper respiratory indicator:
1.133(0.974-1.317)
Lower respiratory indicator:
1.202(1.044-1.385)
Ocular indicator:
0.345(0.125-0.951)
4th quartile
Upper respiratory indicator:
December 2009
E-166
-------
Study Design & Methods Concentrations! Effect Estimates (95% Cl)
1.019(0.904-1.149)
Lower respiratory indicator:
1.344(1.137-1.589)
Ocular indicator: 1.949 (1.416-2.683)
Central - 2nd quartile
Upper respiratory indicator: 1.088
(1.002-1.183)
Lower respiratory indicator:
1.046(0.930-1.176)
Ocular indicator: 1.220 (1.115-1.335)
3rd quartile
Upper respiratory indicator:
1.054(0.977-1.137)
Lower respiratory indicator:
1.055(0.948-1.175)
Ocular indicator: 1.049 (0.965-1.142)
4th quartile
Upper respiratory indicator:
0.899 (0.826-0.979)
Lower respiratory indicator:
0.952(0.845-1.073)
Ocular indicator: 0.875 (0.796-0.963)
Southeast - 2nd quartile
Upper respiratory indicator:
0.778(0.575-1.052)
Lower respiratory indicator:
1.047(0.916-1.196)
Ocular indicator: 0.460 (0.299-0.708)
3rd quartile
Upper respiratory indicator:
1.297(1.127-1.491)
Lower respiratory indicator:
1.391 (1.131-1.711)
Ocular indicator: 0.474 (0.314-0.715)
4th quartile
Upper respiratory indicator:
0.893 (0.812-0.983)
Lower respiratory indicator:
0.937(0.818-1.073)
Ocular indicator: 0.314 (0.182-0.542)
Southwest - 2nd quartile
Upper respiratory indicator:
0.987(0.913-1.066)
Lower respiratory indicator:
2.181 (1.177-4.040)
Ocular indicator: 1.026 (0.928-1.135)
3rd quartile
Upper respiratory indicator:
0.673(0.673-1.886)
Lower respiratory indicator:
0.899(0.790-1.024)
Ocular indicator: 1.017 (0.862-1.200)
4th quartile
Upper respiratory indicator:
0.524(0.524-1.787)
Lower respiratory indicator:
4.346 (0.917-20.606)
Ocular indicator: 0.187 (0.090-0.387)
December 2009 E-167
-------
Study
Reference: Schildcrout et al. (2006,
0898121
Period of Study: Nov 1993-Sep 1995
Location:
Albuquerque, New Mexico
Baltimore, Maryland
Boston, Massachusetts
Denver, Colorado
San Diego, California
Seattle, Washington
St. Louis, Missouri
Toronto, Ontario, Canada
Design & Methods
Outcome: Asthma Symptoms, Rescue
Inhaler Uses
Age Groups: 5-12 yr
Study Design: Meta-analysis of CAMP
N: 990 children
Statistical Analyses: "Working
independence covariance structure"
Logistic Regression
Poisson Regression
"GEE Procedure"
Covariates: Season, age, race-
ethnicity, annual family income, day of
the week
Dose-response Investigated?
Statistical Package: SAS 8.2
R
Lags Considered: 0 day lag, 1 day lag,
2 day lag, 3-day moving sum
Concentrations!
Pollutant: PM,0
Averaging Time: 24-h avg
Seattle: Daily
Albuquerque: Daily
Baltimore: 50% of study days measured
Boston: 23% of study days measured
Denver: 37% of study days measured
San Diego: 24% of study days
measured
St. Louis: 19% of study days measured
Toronto: 47% of study days measured
Percentiles:
10th: 6.8-14.0
25th: 12.0-22.4
SOth(Median): 17.7-32.4
75th: 26.2-42.7
90th: 32.5-53.9
Monitoring Stations: 1-12
Copollutant (correlation):
N02r = 0.26-0.64
S02r = 0.31-0.65
03r = 0.03-0.73
CO r = 0.24-0.88
Effect Estimates (95% Cl)
PM Increment: 25 pg/m3
One-pollutant model
Asthma Symptoms:
1.02 0.94, 1.11 0
1.01 0.97, 1.06 1
1.02 0.98, 1.07 2
1.01 0.98, 1.05
3-day moving sum
Rescue Inhaler Uses:
[0 97 1 05 0
[0.97, 1.05 1
1.00 [0.97, 1.03] 2
1.01 [0.98, 1.03]
3-day moving sum
Two-pollutant model
Asthma Symptoms:
CO-PM,o
1.08 1.01,1.15 0
1.06 0.99, 1.14 1
1.08 1.02, 1.14 2
1.05 1.01, 1.08
3-day moving sum
N02~ PM10
1 06 099 1 13 0
1.04 0.97 1.11 1
1.08 1.02, 1.15 2
1.04 1.00, 1.07;
3-day moving sum
S02-PM10
1.05 0.98, 1.13 0
1.04 0.96, 1.14 1
1.05 0.98, 1.12 2
1.04 0.99, 1.08;
3-day moving sum
Rescue Inhaler Uses:
CO-PM,o
1.06 0.99, 1.13 0
1.05 0.99, 1.11 1;
1.05 1.01, 1.09 2
1.03 1.00, 1.07;
3-day moving sum
N02" PM10
1.03 0.97, 1.08 0
1.03 0.98, 1.08 1
1.04 1.00, 1.09 2
1.02 1.00, 1.05
3-day moving sum
S02-PM10
1.01 0.95, 1.07 0
1.02 0.97, 1.07 1
1.03 0.98, 1.09 2
1.02 0.98, 1.05;
3-day moving sum
All units expressed in pg/m unless otherwise specified.
December 2009
E-168
-------
Table E-10. Short-term exposure - respiratory morbidity outcomes - PMio25.
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Aekplakorn et al.
(2003, 089908)
Period of Study: 107 days,
Oct1997-Jan1998
Location: Mae Mo district,
Lampang Province, north
Thailand
Outcome: Upper respiratory
symptoms, lower respiratory
symptoms, cough
Age Groups: 6-1 4 yr
Study Design: Logistic regression
N: 98 asthmatic school children
Statistical Analyses: Generalized
Pollutant: PM10.2.5
Averaging Time: Daily
Mean (SD): NR
Range (Min, Max): NR
Monitoring Stations: 3
Copollutant: PM10, S02
PM Increment: 10fjg/m°
Odds Ratios [Lower Cl, Upper Cl] lag:
Asthmatics:
URS: 1.04 (0.93, 1.17) lag 0
LRS: 1.09 (0.95, 1.26) lag 0
Cough: 1.08 (0.96, 1.21) lag 0
Non-Asthmatics:
i IDC' ^ CK /n on -i ^\ i-^n n
Estimating Equations, stratified
analysis, PROCGENMOD
Covariates: Temperature and relative
humidity
Season: Winter
Dose-response Investigated? No
Statistical Package: SASv 8.1
URS: 1.05 (0.99,1.19) lag 0
LRS: 0.90 (0.72,1.11) lag 0
Cough: 0.95 (0.81,1.11) lag 0
December 2009
E-169
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Bourotte et al.
(2007,1500401
Period of Study:
May 2002-Jul 2002
Location: Sao Paulo, Brazil
Outcome: Peak expiratory flow (PEF) Pollutant: PM:
Age Groups: Avg age 39.8 ± 12.3 yr
Study Design: Cross-sectional
N: 33 patients
Statistical Analyses: Linear mixed-
effects model
Covariates: Gender, Age, BMI, Air
Pollutants, Ambient temperature,
Relative Humidity
Season: Winter
Dose-response Investigated? No
Statistical Package: S-plus
Lags Considered: 2-day lag, 3-day
lag
Averaging Time: 24 h
Mean(SD):21.7(12.9)|jg/m3
Range (Min, Max): (4.13, 62.0)
Components:
Na*
K*
Mg2*
Ca2*
Finf
Cl-
N03:
S042"
Monitoring Stations: 1
PM Increment: NR
Effect [Lower Cl, Upper Cl] lag:
Morning PEF
Na* concurrent day = -0.454 (-1.605, 0.697)
Na* 2-day lag = -0.907 (-2.288, 0.474)
Na* 3-day lag = -1.361 (-2.972, 0.251)
K* concurrent day = 1.685 (-0.492, 3.862)
K* 2-day lag = 1.838 (-1.272, 4.984)
K* 3-day lag = 2.604 (-0.812, 6.025)
Mg2* concurrent day = 2.265* (-0.427, 4.956)
Mg2* 2-day lag = 1.271 (-1.869,4.410)
Mg2* 3-day lag = 0.939 (-2.425, 4.303)
Ca2* concurrent day = 5.491* (2.558, 8.424)
Ca2* 2-day lag = 6.358* (2.251,10.465)
Ca2* 3-day lag = 6.069 (1.962,10.176)
Finf concurrent day = 1.572 (-0.792, 3.935)
Finf 2-day lag = 1.630 (-1.679, 4.939)
Finf 3-day lag = 2.736* (-1.754, 7.226)
Cl" concurrent day = -0.951 (-2.238, 0.336)
Cl" 2-day lag = -1.871 (-3.242 to -0.4997)
Cf 3-day lag = -2.286* (-3.934 to -0.638)\
N03" concurrent day = 4.195* (-0.063, 8.452)
N03" 2-day lag = 6.292* (2.034, 10.55)
N03" 3-day lag = 7.341* (3.083,11.60)
S042" concurrent day = 3.528 (-0.053, 7.110)
S042" 2-day lag = 4.411* (0.829, 7.991)|
S042" 3-day lag = 6.175* (2.593, 9.756)
Evening PEF
Na* concurrent day = -0.680 (-1.831, 0.471)
Na* 2-day lag = -1.90 (-3.316 to -0.494)
Na* 3-day lag = -2.336* (-3.878 to -0.794)
K* concurrent day = 0.613 (-1.564, 2.790)
K* 2-day lag = 0.613 (-2.497, 3.723)
K* 3-day lag = 0.000 (-3.421, 3.421)|
Mg2+ concurrent day = -0.718 (-3.522, 2.085)
Mg2+2-day lag = -1.933 (-5.073,1.206)
Mg2+ 3-day lag = -3.591 (-7.056 to -0.126)
Ca2+ concurrent day = 2.312* (-1.208, 5.832)
Ca2+2-day lag = 2.023 (-2.084, 6.130)
Ca2+ 3-day lag = 0.578 (-3.530, 4.685)
Fin, concurrent day = -1.281 (-3.644,1.083)
Fin, 2-day lag = -2.503 (-5.930, 0.924)
Fin, 3-day lag = -4.540 (-9.149, 0.068)
Cf concurrent day = -0.317 (-1.604, 0.970)
Cf 2-day lag = -1.268 (-2.556, 0.019)
Cl" 3-day lag = -1.902 (-3.589 to -0.216)
N03"concurrent day = 3.146 (-1.112, 7.404)
N03" 2-day lag = 3.146 (-1.112, 7.404)
N03" 3-day lag = 1.049 (-3.209, 5.306)
S042" concurrent day = 1.764 (-1.817, 5.346)
S042" 2-day lag = 2.646 (-0.935, 6.228)
S042" 3-day lag = 1.764 (-1.817, 5.346)
Reference: Ebelt et al. (2005,
0569071
Period of Study: Summer of
1998
Location: Vancouver,
Canada
Outcome: Spirometry
Age Groups: Range from 54-86 yr
Mean age= 74 yr
Study Design: Extended analysis of a
repeated-measures panel study
N: 16 persons with COPD
Statistical Analyses: Earlier analysis
expanded by developing mixed-effect
regression models and by evaluating
additional exposure indicators
Dose-response Investigated? No
Statistical Package: SASV8
Pollutant: PM10.25
Averaging Time: 24 h
Mean (SD):
Ambient PMi0.25: 5.6 (3.0)
Exposure to ambient PMio-2 5:
2.4(1.7)
Range (Min, Max):
Ambient PMio.25: (-1.2-11.9)
Exposure to ambient
PMio.25: (-0.4-7.2)
Monitoring Stations: 5
Copollutant (correlation):
Ambient PM,0: r= 0.69
Ambient PM25:r= 0.15
Nonsulfate Ambient
PM25:r=0.14
Exposure to Ambient
PM10.25: r= 0.73
PM Increment: Ambient PM10.25: 4.5 (IQR)
Exposure to ambient PM10.2 5: 2.4 (IQR)
Notes: Effect estimates are presented in Fig 2 and
Electronic Appendix Table 1 (only available with electronic
version of article) and not provided quantitatively
elsewhere.
December 2009
E-170
-------
Study
Reference: Lagorio et
al.(2006, 089800)
Period of Study: May
1999-June 1999 and Nov
1999-Dec1999
Location: Rome, Italy
Reference: Laurent et al.
(2008, 1566721
Period of Study: Dec
2003-Dec 2004
Location: Strasbourg, France
Reference: Tang et al. (2007,
0912691
Period of Study:
Dec 2003-Feb 2005
Location: Sin-Chung City,
Taipei County, Taiwan
Design & Methods
Outcome: Lung function of subjects
(FVC and FEV,) with COPD, Asthma
Age Groups: COPD 50-80 yr
Asthma 1 8-64 yr
Study Design: Time series
N: COPD N = 11; Asthma N = 11
Statistical Analyses: Non-parametric
Spearman correlation
GEE
Covariates:
COPD: daily mean temperature,
season variable (spring or winter),
relative humidity, day of week
Asthma: season variable, temperature,
humidity, and D-2-agonist use
Season: Spring and winter
Dose-response Investigated? Yes
Statistical Package: STATA
Lags Considered: 1-3 days
Outcome: Sales of short acting |3-
agonists
Study Design: Case-crossover
Covariates: NR
Statistical Analysis: Conditional
logistic regression
Age Groups: 0-39 yr
Outcome: Peak expiratory flow rate
(PEFR) of asthmatic children
Age Groups: 6-1 2 yr
Study Design: Panel study
N: 30 children
Statistical Analyses:
Linear mixed-effect models were used
to estimate the effect of PM exposure
on PEFR
Covariates: Gender, age, BMI, history
of respiratory or atopic disease in
family, SHS, acute asthmatic
exacerbation in past 12 mo, ambient
temperature and relative humidity,
presence of indoor pollutants, and
presence of outdoor pollutants,
Dose-response Investigated? yes
Statistical Package: S-Plus2000
Lags Considered: 0-2
Concentrations1
PM Size: PM10.2.5
Averaging Time: 24 h
Mean (SD):
Overall: 15.6 (7.2)
Spring: 18.7 (7.4)
Winter: 12.3 (5.4)
Range (Min, Max): (3.4, 39.6)
PM Component:
Cd: 0.46+0.40 ng/m3
Cr:1.9±1.7ng/rr?
Fe: 283+167 ng/m3
Ni: 4.8+6.5 ng/m
Pb: 30.6+19.0 ng/m3
Pt: 5.0+8.6 pg/rr?
V: 1.8+1. 4 ng/m3
Zn: 45.8+33.1 ng/m3
Monitoring Stations: Two fixed
sites: (Villa Ada and Istituto
superior di Sanita)
Copollutant (correlation):
N02r = 0.51
03r = 0.31
CO r = -0.09
S02r = -0.16
PM10r = 0.61
PM25r = 0.34
Pollutant: PM,0
Averaging Time: NR
Mean (SD) Unit: 20.8 (10.2)
pg/m3
Range (Min, Max): NR
Copollutant (correlation): N02,
03, correlations NR
Pollutant: PMi0.25
Averaging Time: 1 h
Mean (SD):
Personal: 17.8 (19.6)
Ambient: 17.0 (10.6)
Range (Min, Max):
Personal: 0.3-195.7
Ambient: 0.1-80.2
Monitoring Stations: 1
Effect Estimates (95% Cl)
PM Increment: 1 pg/m3
They observed no statistically significant effect of PM10.2.5
on FVC and FEV, on any of the panels (COPD, Asthma).
P Coefficient (SE)
COPD
FVC(%)
24 h -1.32 (1.06)1
48-h -1.46 (1.31)
72-h -1.38 (1.53)
FEVi(%)
24 h -0.59 (0.95)
48-h -1.01 (1.19)
72-h -0.90 (1.42)
Asthma
FVC(%)
24 h -0.1 7 (0.75)
48-h -0.36 (0.91)
72-h -0.24 (1.07)
FEV,(%)
24 h -0.67 (0.89)
48-h -1.19 (1.07)
72-h -0.51 (1.26)
Increment: 10 pg/m3
Percent Increase in Short Acting p-agonists sold
Per increment increase in ambient PM10 at lags 4-7, a
7.5% increase (95% Cl: 4-11.2%) was seen in SABA
sales.
All other results were given in Fig 1 and 2
PM Increment:! 5.9 pg/m3
RR Estimate [Lower Cl, Upper Cl]
lag:
Change in morning PEFR:
-20.55 (-45.83, 4.73) lag 0
-39.05 (-104.16, 26.06) lag 1
-39.56 (-79.56, 0.44) lag 2
-37.15 (-105.01, 30.7) 2-day mean
-35.47 (-27.32, 56.38) 3-day mean
Change in evening PEFR:
-1.68 (-19. 13, 15.78) lag 0
1.59 (-14.32, 17.5) lag 1
0.86 (-30.84, 32.57) lag 2
5.97 (-15.57, 27.5) 2-day mean
29.75 (-1.69, 61. 18) 3-day mean
December 2009
E-171
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Trenga et al.,
(2006, 1552091
Period of Study: 1999-2002
Location: Seattle, WA
Outcome: Lung function: FEVi, PEF,
MMEF (maximal midexpiratory flow
assessed only for children)
Age Groups: Adults (56-89-yr-old)
healthy & with COPD
Asthmatic children 6-13-yr-old
Study Design: Adult and pediatric
panel study over 3 yr with 1 monitoring
period ("session") per yr
N: 57 adults (33 healthy, 24 with
COPD) = 692 subject-days = 207
study-days
17 asthmatic children = 319 subject-
days = 98 study-days
Statistical Analyses: Mixed effects,
longitudinal regression models, with
the effects of pollutant decomposed
into each subject's a) overall mean
b) Difference between their session-
specific mean and overall mean
c) Difference between their daily
values and session-specific mean
Covariates: Gender, age, ventral site
temperature and relative humidity, CO,
N02
Season: NR
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-1 days
Pollutant: PMi0.2.5 (coarse)
Averaging Time: 24 h
Percentiles:
Subject-specific exposure
PM,o-PM2.5
Outdoor
25th: 3.3
50th (Median): 4.7
75th: 6.9
Adults
Outdoor
25th: 3.3
50th (Median): 5.0
75th: 7.1
Range (Min, Max):
Subject-specific exposure
Children
Outdoor (0.0, 25.3)
Adults
Outdoor (0.0, 25.7)
Monitoring Stations: 2
Also subject-specific local
outdoors (i.e., at each home),
indoor, and personal
Copollutant (correlation):
CO
N02
PM2.5
PM Increment: 10|jg/m
Adult
Outdoor Home PM10-PM25
FEV,
Overall: Lag 0-27.9 [-87.5: 31.8]
Lag 147.1 [-5.1:99.4]
No-COPD: Lag 0 -49.2 [-22.3: 23.9]
Lag 174.3 [6.8:141.8]
COPD: Lag 07.3 [-84.7: 99.4]
Lag1 11.5 [-65.4: 88.3]
PEF
Overall: Lag 05.3 [-5.1:15.7]
Lag 1-2.5 [-11.6: 6.5]
No-COPD: Lag 05.1 [-7.7:17.8]
Lag 1-5.8 [-17.5: 5.9]
COPD: Lag 05.7 [-10.3: 21.6]
Lag 11.7 [-11.5:14.9]
Pediatric
FEV,
Outdoor Home PMi0-PM25
Overall
Lag 0-7.43 [-69.41: 54.55]
Lag 1-25.61 [-88.16: 36.94]
NoAnti-inflam. Medication
Lag 0-63.87 [-199.58: 71.84]
Lag 1 -96.48 [-232.48: 39.52]
Anti-inflam. Medication
Lag 06.57 [-96.90:110.04]
Lag 1-8.63 [-217.39: 200.14]
PEF
Outdoor Home PM10-PM25
Overall
Lag 04.53 [-6.60:15.67]
Lag 1-3.35 [-14.31: 7.62]
NoAnti-inflam. Medication
Lag 02.05 [-22.36: 26.45]
Lag 1-6.56 [-30.90:17.78]
Anti-inflam. Medication
Lag 05.15 [-7.90:18.19]
Lag 1-2.58 [-15.35:10.19]
MMEF
Outdoor Home PM10-PM25
Overall
Lag 0-0.01 [-7.29: 7.28]
Lag 1-2.07 [-9.25: 5.12]
NoAnti-inflam. Medication
Lag 0-7.14 [-23.16: 8.87]
Lag 1-14.39 [-30.11:1.32]
Anti-inflam. Medication
Lag 01.76 [-6.78:10.30]
Lag 1 0.89 [-7.56: 9.33]
All units expressed in pg/m unless otherwise specified.
December 2009
E-172
-------
Table E-11. Short-term exposure - respiratory morbidity outcomes - PM2.e (including
components/sources).
Study
Reference: Adamkiewicz et al. (2004,
0879251
Period of Study: Aug-Dec 2000
Location: Steubenville, Ohio
Reference: Adar et al. (2007, 0986351
Period of Study: Mar-Jun 2002
Location: St. Louis, MO
Reference: Aekplakorn et al. (2003,
0899081
Period of Study: 107 days, from Oct
1997-Jan 1998
Location: Mae Mo district, Lampang
Province, north Thailand
Design & Methods
Outcome: FENO
Age Groups: Ranged 53.5-90.6 yr
Study Design: Prospective cohort
N: Total of 294 breaths from 29 subjects
Statistical Analyses: Fixed effect
models, ANOVA, GLM procedure
Covariates: Subject, week of study,
day of the week, h of the day, ambient
barometric pressure, temperature, and
relative humidity
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: Hourly lags, 0-48 h
Outcome: FENO
Age Groups: 60+
Study Design: Panel Study
N: 44 non-smoking seniors
Statistical Analyses: Mixed models
containing random subject effects
Covariates: Day of week, trip type,
FENO collection device, current illness,
use of vitamins, antihistamines, statins,
steroids, and asthma medications,
temperature, pollen, mold, NO
concentration in testing room
Statistical Package: SAS
Lags Considered :0
Outcome: Upper respiratory
symptoms, lower respiratory symptoms,
cough
Age Groups: 6-1 4 yr old
Study Design: Logistic regression
N: 98 asthmatic school children
Statistical Analyses: Generalized
Estimating Equations, stratified
analysis, PROCGENMOD
Covariates: Temperature and relative
humidity
Season: Winter
Dose-response Investigated? No
Statistical Package: SAS v 8.1
Concentrations1
Pollutant: PM25
Averaging Time: 1 h
Mean (SD): 19.5
Percentiles:
O^th' 7 K
zotn. /.D
75th: 25.5
Range (Min, Max): NR, 105.8
Monitoring Stations: 1
Averaging Time: 24 h
Mean (SD): 19.7
Percentiles:
25th: 9.7
75th: 27.4
Range (Min, Max): NR, 57.8
Monitoring Stations: 1
Copollutant (correlation):
Ambient NO
Indoor NO
N02
03
S02
Pollutant: PM25
Averaging Time: 24 h
Mean (SD):
Pretrip: 14.8
Post-trip: 16.5
Percentiles:
25th (pretrip): 11.2
75th (pretrip): 20.1
25th (post-trip: 11. 7
75th (post-trip: 21. 6
Monitoring Stations: 1
Copollutant (correlation): BC
CO
N02
S02
03
Pollutant: PM25
Averaging Time: Daily
Mean (SD):
Sob Pad station: 24.77
Sob Mo station: 24.89
Hua Fai station: 26.27
Range (Min, Max):
Sob Pad: 4.52 24.77
Sob Mo: 3.13, 24.89
Hua Fai: 3.67, 26.27
Monitoring Stations: 3
Copollutant:
PM,o
S02
Effect Estimates (95% Cl)
PM Increment: 17.9 pg/m3
Effect Estimate [Lower Cl, Upper Cl]:
1-h Single pollutant models:
0.36(0.58-2.14)
PM Increment: 17.7
Effect Estimate [Lower Cl, Upper Cl]:
24-h ma: 1.45 (0.33-2.57)
Multipollutant models for PM25, ambient
NO and room NO and estimated
change in FENO (ppb) for an IQR in
pollutant measure
Model! 1.95(0.47-3.43)
Model 2 1.38 (0.26-2.51)
Model 4 1.97 (0.48-3. 46)
Notes: Association of FENO with PM25
at different lags presented in Fig 1 are
not presented quantitatively elsewhere.
PM Increment: 9.8 pg/m3
Effect Estimate [Lower Cl, Upper Cl]:
Pre-trip% change: 21. 9 (6.7, 39.4)
Post-trip % change: -4.7 (-17.1, 9.6)
PM Increment: 10 pg/m3
Odds Ratios [Lower Cl, Upper Cl]
lag:
Asthmatics:
URS: 1.04 (0.99, 1.09) lag 0
LRS: 1.05 (0.98, 1.2)lagO
Cough: 1.05 (0.99, 1.10) lag 0
Non-Asthmatics1
URS: 1.03 (0.96', 1.09) lag 0
LRS: 1. 02 (0.93, 1.10) lag 0
Cough: 1.00 (0.93, 1.07) lag 0
PM,0 + S02
Asthmatics:
URS: 1.04 (0.99, 1.10)lagO
LRS: 1.05 (0.98, 1.10)lagO
Cough: 1.05 (0.99, 1.11) lag 0
Non-Asthmatics:
URS: 1.03 (0.97, 1.09) lag 0
LRS: 1.02 (0.93, 1.11) lag 0
Cough: 1.00 (0.93, 1.07) lag 0
December 2009
E-173
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Allen et al. (2008,1562081
Period of Study: 1999-2002 (additional
PM composition data collected Dec
2000 and May 2001)
Location: Seattle, USA
Outcome: Daily changes in exhaled
nitric oxide (FENO) and 4 lung function
measures, midexpiratory flow (MEF),
peak expiratory flow (PEF), forced
expiratory volume in 1 second (FEVi),
and forced vital capacity (FVC)
Age Groups: 6-1 Syr
Study Design: Panel study
N: 19 children with asthma
Statistical Analyses: Linear mixed
effects model with random intercept to
test for within participant associations
Covariates: Temperature, relative
humidity, BMI, age, and, in the case of
FENO, ambient NO measured at a
centrally located monitoring site
Models also included a term for within-
participant, within-session effects, and a
term for participant between-session
effects
Effect modification: Decided a priori to
include interaction term for PM25
exposure and inhaled corticosteroids
Pollutant: PM25
Mean (SD): 11.23 (6.48)
Range (Min, Max):
276-40.38
25th: 6.38
75th: 14.73
Copollutant (correlation):
Ambient LAC* r=0.83
Ambient LG**r=0.84
Personal PM25: r=0.34
Personal LAC: r=0.54
Ambient-generated PM25: r=0.87
Nonambient-generated PM25: r=-0.06
* LAC Light-absorbing carbon
** LG: Leroglucosan (a marker of wood
smoke)
Health effect estimates presented in
graphic form (Fig 1). Summary from text
is as follows:
Personal LAC, personal PM25, and
ambient-generated PM25were
associated with (p < 0.05) and ambient
PM2 5 was marginally associated
(p=0.09) with increased FENO. Neither
of the ambient combustion markers
(LAC, LG) nor nonambient-generated
PM2 5 was associated with FENO
changes.
All of the ambient concentrations were
associated with decrements in PEF and
MEF while ambient-generated PM25
was marginally associated (p < 0.10).
Only ambient LG was associated with a
decrease in FEVi and there were no
associations between exposure metrics
and FVC.
Reference: Barraza-Villarreal et
al.(2008, 1562541
Period of Study: Jun 2003-Jun 2005
Location: Mexico City
Outcome: Respiratory Symptoms,
Coughing, Wheezing, Airway
inflammation, Asthma
Study Design: Prospective cohort
Statistical Analyses: Bivarate analysis
Age Groups: 6-14
Pollutant: PM25
Averaging Time: Maximum 8-h avg
Mean (SD) unit:
28.9 (2.8)
Range (Min, Max):
(4.2, 102.8)
Copollutants (correlation):
03
N02
Increment: 17.5|jg/m
% Increase (Lower Cl, Upper Cl)
Asthmatic children
Inflammatory Marker:
FENO: 1.08 (1.01, 1.16)0
IL-8:1.08 (0.98, 1.19)0
ph_EBC: -0.03 (-0.09, 0.03) 0
Lung Function:
FEV,:-16.0(-31. Oto-0.13) 0-4 avg
FVC:-23.0 (-42.0 to-5.21) 0-4 avg
FEV25-75:-11.0 (-42.0, 20.3) 0-4 avg
Nonasthmatic children
Inflammatory Marker:
FENO: 0.89 (0.78, 1.01)0
IL-8:1.16 (1.00, 1.36)0;
ph_EBC:-0.05 (-0.14, 0.04)0
Lung Function:
FEV,:-21.0 (-42.3, 0.38) 0-4 avg
FVC:-29.0 (-52.8 to-4.35) 0-4 avg
FEV25.75: -20.0 (-69.0, 29.0) 0-4 avg
All children age 6-14
Respiratory Symptom:
Cough: 1.11 (1.06,1.17)
Wheezing: 1.06 (0.99,1.13)
December 2009
E-174
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Bennett et al. (2007,
1562681
Period of Study: 1992-2005
Location: Melbourne, Australia
Outcome: Adverse respiratory
symptoms (wheeze, shortness of breath
on waking, cough in the morning,
phlegm in the morning, cough with
phlegm in the morning, asthma attack)
Age Groups: All ages with a mean of
37.2 yr
Study Design: Cohort study
N: 1446 persons
Statistical Analyses: Logistic
regression models
Covariates: Age, gender, current
smoking status, medication use (IS2-
agonist and inhaled steroid), atopy
Dose-response Investigated? No
Statistical Package: STATA statistical
software, version 9 (Statcorp, 2005)
Pollutant: PM25
Averaging Time: 24 h
Mean (SD): 6.8
Range (Min, Max): (1.8-73.3)
Monitoring Stations: 1
PM Increment: 1 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
Within-person (longitudinal effects)
Wheeze: OR=1.08 (0.79-1.48)
SOB on waking: OR=1.34 (0.84-2.16)
Cough in the morning:
OR=0.74 (0.47-1.15)
Phlegm in the morning:
OR=1.55 (0.95-2.53)
Cough w/phlegm morning:
OR=1.28 (0.70-2.33)
Asthma attack: OR=0.91 (0.55-1.49)
Between-person (cross-sectional)
effects
Wheeze: OR=1.32 (0.82-2.10)
SOB on waking: OR=1.29 (0.46-3.60)
Cough in the morning:
OR=0.21 (0.07-0.62)
Phlegm in the morning:
OR=0.49(0.16-1.44)
Cough w/phlegm morning:
OR=0.28 (0.08-0.97)
Asthma attack: OR=0.52 (0.17-1.59)
December 2009
E-175
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Bourotte et al. (2007,
1500401
Period of Study: May 2002-Jul 2002
Location: Sao Paulo, Brazil
Outcome: Peak expiratory flow (PEF)
Age Groups: Avg age 39.8 ± 12.3 yr
Study Design: Cross-sectional
N: 33 patients
Statistical Analyses:
Linear mixed-effects model
Covariates: Gender, Age, BMI, Air
Pollutants, Ambient temperature,
Relative Humidity
Season: Winter
Dose-response Investigated? No
Statistical Package: S-plus
Lags Considered: 2-day lag, 3-day lag
Pollutant: PM25 (Fine)
Averaging Time: 24 h
Mean(SD):11.9(5.12)
Range (Min, Max):
(2.82, 26.6)
Components:
K*
Mg2*
Ca2*
Fin,
Cl"
N03:
S042"
Monitoring Stations: 1
PM Increment: NR
Effect [Lower Cl, Upper Cl] lag:
Morning PEF
Na* concurrent day =
-0.409 (-2.485, 1.667)
Na* 2-day lag = -0.818 (-4.139, 2.503)
Na* 3-day lag = -0.205 (-4.356, 3.974)
K* concurrent day =
-0.211 (-2.778, 2.357)
K* 2-day lag = -0.843 (-4.695, 3.008)
K* 3-day lag = 0.843 (-4.292, 5.978)
Mg2* concurrent day =
-1.750 (-5.302, 1.802)
Mg2* 2-day lag = -5.016 (-10.79, 0.762
Mg2* 3-day lag = -3.850 (-10.15, 2.449
Ca2* concurrent day =
3.192*(-0.599, 6.943)
Ca2* 2-day lag = 5.880 (1.105,10.65)
Ca2* 3-day lag = 7.560* (2.103,13.02)
Fjnf concurrent day =
2.218*(-0.033,4.470)
Finf 2-day lag = 3.697* (1.446, 5.949)
Fin, 3-day lag =4.067* (1.065, 7.069)
Cl" concurrent day =
-1.010 (-3.469, 1.450)
Cl" 2-day lag = -1.615
Cl" 3-day lag = -1.615
-5.714, 2.483
-6.534, 3.303
N03 concurrent day =
3.144(0.409,5.878)
N03" 2-day lag = 3.593 (0.858, 6.328)
N03" 3-day lag = 4.491 (1.756, 7.226)
S042 concurrent day =
2.210 (-0.032, 4.272)
S042" 2-day lag = 3.180 (1.028, 5.332)
S042" 3-day lag = 3.180 (1.028, 5.332)
Evening PEF
Na* concurrent day =
-1.636 (-3.712, 0.440)
Na* 2-day lag = -0.205 (-3.256, 3.117)
Na* 3-day lag = -1.023 (-5.174, 3.129)
K* concurrent day =
-1.897 (-4.465, 0.670)
K* 2-day lag = -1.686 (-5.966, 2.592)
K* 3-day lag = -1.054 (-6.189, 4.081)
Mg2* concurrent day =
-2.753 (-6.400, 0.894)
Mg2* 2-day lag = -2.567 (-8.534, 3.401)
Mg2* 3-day lag = -4.876 (-11.36,1.612)
Ca2* concurrent day =
2.184 (-1.567, 5.935)
Ca2*2-day lag = 5.040 (0.265, 9.815)
Ca2* 3-day lag = 5.040 (-0.417,10.50)
FM f concurrent day =
1.479 (-0.773, 3.730)
Fin, 2-day lag = 1.819 (-0.403, 4.100)
Fn, 3-day lag = 2.958 (-0.044, 5.960)
Cl~ concurrent day =
-0.404 (-2.863, 2.055)
Cl" 2-day lag = 0.000 (-4.099, 4.099)
Cl"3_-day lag =0.202 (-4.716, 5.120)
N03 concurrent day =
1.796 (-0.939, 4.531)
N03" 2-day lag = 2.695 (-0.040, 5.430)
N03" 3-day lag = 3.144 (0.409, 5.878)
S042 concurrent day =
2.120 (-0.032, 4.272)
S042" 2-day lag = 2.120 (-0.032, 4.272)
S042" 3-day lag =2.120 (-0.032, 4.272)
December 2009
E-176
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: de Hartog et al. (2003,
0010611
Period of Study: Winter of 1998-1999
(in Amsterdam, from Nov 1998-Jun
1999; in Erfurt, from pet 1998-Apr
1999; and in Helsinki, from Nov
1998-Apr 1999.)
Location:
Amsterdam, the Netherlands
Erfurt, Germany
and Helsinki, Finland
Outcome: Respiratory symptoms
Age Groups: 2 50 yr
Study Design: Cohort
N: 131 subjects with history of coronary
heart disease
Statistical Analyses: Logistic
regression
Covariates: Ambient temperature,
relative humidity, atmospheric pressure,
incidence of influenza-like illness
Season: Winter
Dose-response Investigated? No
Statistical Package: S-PLUS 2000
Lags Considered: 0-, 1-, 2-, 3-, and
5-day avg
Pollutant: PM25
Averaging Time: 24 h
Mean (SD):
Amsterdam, the Netherlands: 20.0
Erfurt, Germany: 23.4
Helsinki, Finland: 12.8
Range (Min, Max):
Amsterdam, the Netherlands: (3.8-82.2)
Erfurt, Germany: (4.5-118.1)
Helsinki, Finland: (3.1-39.8)
Unit (i.e. pg/m3): pg/m3
Monitoring Stations: 1
Co pollutant:
PM10
NCO.01-0.1
CO
N02
SO,
PM Increment: 10 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
Association of air pollution and
incidence of symptoms in three panels
of elderly subjects
LagO
Chest pain w/ physical exertion: 1.04
(0.96-1.13)
Shortness of breath: 1.04 (0.96-1.12)
Awakened, breathing problems: NA
Avoidance of activities: 1.04 (0.96-1.14)
Phlegm: 1.03 (0.93-1.13)
Lagt
Chest pain w/ physical exertion: 1.01
(0.93-1.09)
Shortness of breath: 1.06 (0.99-1.14)
Awakened, breathing problems: 1.09
(1.00-1.20)
Avoidance of activities: 1.03 (0.95-1.12)
Phlegm: 1.10 (1.01-1.19)
Lag 2
Chest pain w/ physical exertion: 0.98
(0.90-1.05)
Shortness of breath: 1.05 (0.98-1.12)
Awakened, breathing problems: 1.04
(0.95-1.14)
Avoidance of activities: 1.05 (0.97-1.14)
Phlegm: 1.08 (1.00-1.18)
Lag 3
Chest pain w/ physical exertion: 1.00
(0.93-1.08)
Shortness of breath: 1.08 (1.01-1.15)
Awakened, breathing problems: 0.99
(0.91-1.08)
Avoidance of activities: 1.06 (0.98-1.14)
Phlegm: 1.10 (1.01-1.19)
6-day
Chest pain w/ physical exertion: 1.02
(0.91-1.13)
Shortness of breath: 1.12 (1.02-1.24)
Awakened, breathing problems: 1.03
(0.90-1.18)
Avoidance of activities: OR= 1.09 (0.97-
1.22)
Phlegm: OR= 1.16 (1.03-1.32)
December 2009
E-177
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Delfino et al. (2004,
0568971
Period of Study :Sep-Oct 1999
Apr-Jun 2000
Location: Alpine, California
Outcome: FEVi
Age Groups: 9-19 yr old
Study Design: Panel study
N: 24 children
Statistical Analyses: GLM
Akaike's information criterion and
Bayesian information criterion
Covariates: Day of wk, personal
temperature and relative humidity, time
of FEVi maneuver (morning, afternoon,
or evening), Season (fall 1999 or spring
2000), As-needed medication use,
Presence or absence of upper or lower
respiratory infections
Season: Spring, fall
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-4
Pollutant: PM25
Averaging Time: 24-h avg 1-h max
personal PMIast24h
Mean (SD): 151.0 (12.03) 90th: 292.4
Range (Min, Max): (9.1,996.8)
Mean personal PMIast24h
Mean (SD): 37.9 (19.9)
90th: 65.1
Range (Min, Max): 3.9,113.8
Home stationary-site PM
24-h Mean indoor PM25
Mean (SD): 12.1 (5.4)
90th: 20.2
Range (Min, Max): 2.8, 35.3
24-h Mean outdoor PM25
Mean (SD): 11.0(5.4)
90th: 18.4
Range (Min, Max): 1.8, 31.0
Central outdoor stationary-site PM
24-h Mean PM25
Mean (SD): 10.3 (5.6)
90th: 18.4
Range (Min, Max): 1.7, 29.1
Copollutant (correlation):
24-h Central HI PM2 5
8-h max 03 = 0.24
8-h Max N02 = 0.73
8-h Max Personal PM = 0.38
24-h Mean Personal PM = 0.43
8-h Max TEOM PM10 = 0.71
24-h Mean TEOM PM,0 = 0.78
24-h Central HI PM10 = 0.90
24-h Outdoor HI PM2 5 = 0.89
24-h Outdoor HI PM,o = 0.72
24-h Indoor HI PM10 = 0.40
24-h Indoor HI PM2 5 = 0.73
Results presented graphically;
% predicted FEV, was inversely
associated with personal exposure to
fine particles.
Inverse associations of FEV, with
stationary-site indoor, outdoor and
central-site gravimetric PM25 and PM10,
and with hourly TEOM PM,o
December 2009
E-178
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Delfmo et al. (2006,
0907451
Period of Study: Region 1: Aug-Mid
Dec 2003. Region 2: Jul-Nov 2004
Location: Region 1: Riverside, CA.
Region 2: Whittier, CA
Outcome: Fractional Concentration of
Nitric Oxide in exhaled air (FENO)
Age Groups: 9 through 18
Study Design: Longitudinal Panel
Study
N: 45 children
Riverside children
32 Whittier children
Statistical Analyses: Linear mixed-
effects models
Two-stage hierarchical model
Empirical Variograms
Fourth-order polynomial distributed lag
mixed-effects model
Covariates: Personal temperature,
Personal Rel. Humid., 10-day exposure
run, Respiratory infections, Region of
study, Sex, Cumulative daily use of as-
needed B-agonist inhalers
Dose-response Investigated? No
Lags Considered: 0,1, 2, MA day
Pollutant: PM25
Personal Exposure
Averaging Time: 24 h
Riverside
Mean (SD): 32.78 (21.84)
SOth(Median): 28.14
Range (Min, Max): 7.27, 98.43
Whittier
Mean (SD): 36.2 (25.46) SOth(Median):
29.07
Range (Min, Max): 7.55,197.05
Personal Exposure
Averaging Time: 1 h
Riverside
Mean (SD): 97.94 (70.29)
SOth(Median): 83.7
Range (Min, Max): 14.9, 431.8
Whittier
Mean (SD): 93.63 (75.19)
SOth(Median): 71.95
Range (Min, Max): 5.8, 572.9
Personal Exposure
Averaging Time: 8 h
Riverside
Mean (SD): 47.21 (30.9) SOth(Median):
38.5
Range (Min, Max): 8.9,132.1
Whittier
Mean (SD): 51.75 (36.88)
SOth(Median): 40.15
Range (Min, Max): 8.7, 254.1
Central Site
Averaging Time: 24 h
Riverside
Mean (SD): 36.63 (23.46)
SOth(Median): 29.26
Range (Min, Max): (9.52, 87.22)
Whittier
Mean (SD): 18 (12.14) SOth(Median):
16.3
Range (Min, Max): 2.7, 77.09
Monitoring Stations: 48 personal
nephelometers
2 central sites
Copollutant (correlation):
Personal
24-h personal PM251.00
24-h personal EC 0.18
24-h personal OC 0.15
24-h personal N020.33
24-h central PM25 0.64
24-h central EC 0.12
24-h central OC 0.21
24-h central N02 0.22
Central
24-h personal PM250.64
24-h personal EC 0.00
24-h personal OC-0.11
24-h personal N020.12
24-h central PM251.00
24-h central EC 0.55
24-h central OC 0.66
24-h central N02 0.25
PM Increment: IQR increase
(Riverside: 28.41 pg/m3, Whittier 21.87
pg/m3)
Coefficient [Lower Cl, Upper Cl]
Mixed-model estimates of the
association between personal and
central-site air pollutant exposure and
FENO
LagO
Personal 0.42 (-0.15, 0.99)
Central 0.03 (-0.68, 0.74)
Lag1
Personal 0.51 (-0.10,1.12)
Central 0.44 (-0.28,1.16)
2-day ma
Personal 1.01 (0.14,1.88)
Central 0.52 (-0.43,1.47)
Stratified by Medication Use
Lag = 2-day ma
Not Taking Anti-lnflamm. Medication
Personal 1.11 (-1.39, 3.60)
Central 0.44 (-1.65, 2.53)
Taking Anti-lnflamm. Medication
Personal 1.01 (0.19,1.84)
Central 0.55 (-0.47,1.57)
Inhaled Corticosteroids
Personal 1.58 (0.72, 2.43)
Central 1.16 (0.11,2.20)
Antileukotrienes +- inhaled
corticosteroids
Personal -0.89 (-2.73, 0.95)
Central-0.75 (-2.83,1.32)
Notes:
Fig of Estimated lag effect of hourly
personal PM25 on FENO.
Fig of the Estimated lag effect of hourly
personal PM25 on FENO by use of
medications.
Fig of one- and two-pollutant models for
change in FENO using 2-day Ma
personal and central-site pollutant
measurements.
December 2009
E-179
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Delfmo et al. (2006,
0907451
Period of Study: Region 1: Aug-Mid
Dec 2003. Region 2: Jul-Nov 2004
Location: Region 1: Riverside, CA.
Region 2: Whittier, CA
Outcome: Fractional Concentration of
Nitric Oxide in exhaled air (FENO)
Age Groups: 9 through 18
Study Design: Longitudinal Panel
Study
N: 45 children
Statistical Analyses: Linear mixed-
effects models
Two-stage hierarchical model
Empirical Variograms
Fourth-order polynomial distributed lag
mixed-effects model
Covariates: Personal temperature,
personal rel. humid., 10-day exposure
run, respiratory infections, region of
study, sex, cumulative daily use of as-
needed B-agonist inhalers
Dose-response Investigated? No
Lags Considered: Lag 0, Lag 1, 2-day
ma
Pollutant: PM25
PM Component: EC
Personal Exposure
Averaging Time: 24 h
Riverside
Mean (SD): 0.42 (0.69) SOth(Median):
0.34 fjg/m
Range (Min, Max): 0.01, 6.94
Whittier
Mean (SD): 0.78 (1.42)
SOth(Median): 0.47
Range (Min, Max): 0,17.2
Central Site
Averaging Time: 24 h
Riverside
Mean (SD): 1.61 (0.78) SOth(Median):
1.35
Range (Min, Max): 0.52, 3.64
Whittier
Mean (SD): 0.71 (0.43) SOth(Median):
0.63
Range (Min, Max): 0.14, 2.95
Monitoring Stations: 48 personal
nephelometers,
2 central sites
Copollutant (correlation):
Personal
24-h personal PM2 5 0.18
24-h personal EC 1.00
24-h personal OC 0.41
24-h personal N02 0.0.21
24-h central PM25 0.00
24-h central EC 0.04
24-h central OC-0.01
24-h central N02 0.23
Central
24-h personal PM2 5 0.12
24-h personal EC 0.04
24-h personal OC 0.03
24-h personal N02 0.19
24-h central PM25 0.55
24-h central EC 1.00
24-h central OC 0.87
24-h central N02 0.70
PM Increment: IQR increase
(Riverside: 28.41 pg/m3, Whittier 21.87
pg/m3)
Coefficient [Lower Cl, Upper Cl] lag:
Mixed-model estimates of the
association between personal and
central-site air pollutant exposure and
FENO
LagO
Personal 0.29 (0.10,0.48)
Central 0.10 (-0.65, 0.85)
Lag1
Personal-0.01 (-0.23,0.21)
Central 0.99 (0.27,1.71)
2-day ma
Personal 0.72 (0.32,1.12)
Central 1.38 (0.15, 2.61)
Stratified by Medication Use
Lag = 2-day ma
Not Taking Anti-lnflamm. Medication
Personal 0.84 (0.08,1.60)
Central 1.02 (-2.55, 4.60)
Taking Anti-lnflamm. Medication
Personal 0.71 (0.28,1.15)
Central 1.42 (0.25, 2.60)
Inhaled Corticosteroids
Personal 0.67 (0.28,1.07)
Central 1.28 (0.07, 2.49)
Antileukotrienes +- inhaled
corticosteroids
Personal 0.03 (-3.29, 3.35)
Central 1.15 (-1.58, 3.88)
Notes:
Fig of Estimated lag effect of hourly
personal PM25 on FENO.
Fig of the estimated lag effect of hourly
personal PM25 on FENO by use of
medications.
Fig of one- and two-pollutant models for
change in FENO using 2-day Ma
personal and central-site pollutant
measurements.
December 2009
E-180
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Delfino et al. (2006,
0907451
Period of Study: Region 1: Aug-Mid
Dec 2003. Region 2: Jul through Nov
2004
Location: Region 1: Riverside, CA.
Region 2: Whittier, CA
Outcome: Fractional Concentration of
Nitric Oxide in exhaled air (FENO)
Age Groups: 9 through 18
Study Design: Longitudinal Panel
Study
N: 45 children
Statistical Analyses: Linear mixed-
effects models
Two-stage hierarchical model
Empirical Variograms
Fourth-order polynomial distributed lag
mixed-effects model
Covariates: Personal temperature,
personal rel. humid., 10-day exposure
run, respiratory infections, region of
study, sex, cumulative daily use of as-
needed B-agonist inhalers
Dose-response Investigated? No
Lags Considered: Lag 0, Lag 1, 2-day
ma
Pollutant: PM25
PM Component: OC
Personal Exposure
Averaging Time: 24 h
Riverside
Mean (SD): 5.63 (2.59) SOth(Median):
4.98
Range (Min, Max): 1.94,12.38
Whittier
Mean (SD): 6.81 (3.45) SOth(Median):
6.43
Range (Min, Max): 2.18, 31.5
Central Site
Averaging Time: 24 h
Riverside
Mean (SD): 6.88 (1.86)
Percentiles: 50th
Median: 6.07
Range (Min, Max): 4.11,11.62
Whittier
Mean (SD): 3.93 (1.49) SOth(Median):
3.76
Range (Min, Max): 1.64, 8.82
Monitoring Stations: 48 personal
nephelometers,
2 central sites
Copollutant (correlation):
Personal
24-h personal PM250.15
24-h personal EC 0.41
24-h personal OC 1.00
24-h personal N020.20
24-h central PM2 5-0.11
24-h central EC 0.03
24-h central OC-0.02
24-h central N02 0.21
Central
24-h personal PM250.21
24-h personal EC-0.01
24-h personal OC-0.02
24-h personal N020.17
24-h central PM25 0.66
24-h central EC 0.87
24-h central OC 1.00
24-h central N02 0.62
PM Increment: IQR increase
(Riverside: 28.41 pg/m3, Whittier 21.87
pg/m3)
Mixed-model estimates of the
association between personal and
central-site air pollutant exposure and
FENO
LagO
Personal 0.51 (-0.28,1.30)
Central 0.93 (-0.20, 2.06)
Lag1
Personal 0.13 (-0.77,1.03)
CentralO.51 (-0.64,1.66)
2-day ma
Personal 0.94 (-0.47, 2.35)
Central 1.6 (-0.17, 3.37)
Stratified by Medication Use
Lag = 2-day ma.
Not Taking Anti-lnflamm. Medication
Personal 0.88 (-1.62, 3.38)
Central 0.36 (-4.07, 4.79)
Taking Anti-lnflamm. Medication
Personal 0.87 (-0.79, 2.53)
Central 2.05 (0.24, 3.86)
Inhaled Corticosteroids
Personal 2.47 (0.30, 4.64)
Central 1.96 (0.14, 3.78)
Antileukotrienes +- inhaled
corticosteroids
Personal 0.52 (-1.99, 3.02)
Central 1.29 (-2.58, 5.15)
Notes:
Fig of Estimated lag effect of hourly
personal PM25 on FENO.
Fig of the Estimated lag effect of hourly
personal PM25 on FENO by use of
medications.
Fig of one- and two-pollutant models for
change in FENO using 2-day Ma
personal and central-site pollutant
measurements
December 2009
E-181
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Dubowsky et al. (2006,
0887501
Period of Study: Mar 2002-Jun
2002
Location: St. Louis, Missouri
Outcome: Chronic inflammation,
Diabetes, Obesity, Hypertension,
Cardiac Risk
Study Design:
Prospective Cohort
Statistical Analyses:
Poisson, LOESS
Age Groups:
>60
Pollutant: PM25
Averaging Time: 24-h avg
Mean (SD) unit: 16 (6.0)
Range (Min, Max): 6.5, 28
Co pollutants:
BC
CO
N02
S02
03
Increment: 5.4 pg/m
% Increase (Lower Cl, Upper Cl)
Lag
% increase in inflammatory response
and exposure to PM25 in people > 60
Inflammatory Marker:
IL-6:-8(-16, 8)
1:-6(-10, 5
2:-5 (-11, 6)
3: -3 (-9, 6)
4:-4 (-12, 10)
5:-5 (-13, 8)
6:-6 (-14, 9)
7
CRP:-2(-22, 15)
1:3 (-8, 17)
2: 4 (-9, 20)
3: 9 (-4, 27)
4:11 (-5, 35)
5: 8 (-9, 29)
6: 5 (-12, 26)
7
WBC: 0 (-2, 4)
1:1 (-1,2)
2: 2 (-1,3)
3:1 (-2, 5)
4: 3 (-1,10)
5:5(0,12)
6:8(0, 14)
% Increase in inflammatory responses
and exposure to ambient PM25
concentrations in people > 60
Inflammatory Marker:
CRP
All conditions*: 14 (-5.4, 37)
0-5 avg
3 conditions met*: 81 (21,172)
0-5 avg
2 conditions met*: 11 (-7.3,33)
0-5 avg
IL-6
All conditions*:-2.1 (-13,11)
0-5 avg
3 conditions met*: 23 (-5.3, 59)
0-5 avg
2 conditions met*:-3.1 (-14,9.7)
0-5 avg
WBC
All conditions*: 3.4 (-1.8, 8.9)
0-5 avg
3 conditions met*: 0.4 (-8.8,11)
0-5 avg
2 conditions met*: 3.6 (-1.7, 9.1)
0-5 avg
*AII conditions met means model is
adjusted for sex, obesity, diabetes,
smoking history, ambient and
microenvironmental apparent
temperature, mold, pollen, trip, h, and
vitamins.
Three conditions met means model is
adjusted for three of the variables.
Two conditions met means model is
adjusted for 2 of the variables.
December 2009
E-182
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ebelt et al. (2005, 0569071
Period of Study: Summer of 1998
Location: Vancouver, Canada
Outcome: spirometry,
Age Groups: range from 54-86 yr
Mean age= 74 yr
Study Design: extended analysis of a
repeated-measures panel study
N: 16 persons with COPD
Statistical Analyses: Earlier analysis
expanded by developing mixed-effect
regression models and by evaluating
additional exposure indicators
Dose-response Investigated? No
Statistical Package: SAS V8
Pollutant: PM25
Averaging Time: 24 h
Mean (SD):
Ambient PM25:11.4 (4.6)
Exposure to ambient PM25: 7.9 (3.7)
Nonsulfate ambient PM25: 9.3 (3.7)
Exposure to nonsulfate ambient PM25:
6.5 (3.0)
Total exposure to PM25:18.5 (14.9)
Exposure to nonambient PM25:10.6
(14.5)
Range (Min, Max):
Ambient PM25: (4.2-28.7)
Exposure to ambient PM25: (0.9-21.3)
Nonsulfate ambient PM25: (3.3-23.3)
Exposure to nonsulfate ambient PM25:
(0.7-16.9)
Total exposure to PM25: (2.2-90.9)
Exposure to nonambient PM25: (-2.6-
85.0)
Monitoring Stations: 5
Copollutant (correlation):
Ambient PM,0: r= 0.78
Ambient PM10.25: r=0.15
Ambient Sulfate-0.82
Nonsufate Ambient PM25: r= 0.98
PM Increment: Ambient PM25: 5.8
(IQR)
Exposure to ambient PM25: 4.4 (IQR)
Nonsulfate ambient PM25: 4.2 (IQR)
Exposure to nonsulfate ambient PM25:
3.4 (IQR)
Total exposure to PM25:10.1 (IQR)
Exposure to nonambient PM25: 8.9
(IQR)
Notes: Effect estimates are presented
in Fig 2 and Electronic Appendix Table
1 (only available with electronic version
of article) and not provided
quantitatively elsewhere.
Reference: Ebelt et al. (2005, 0569071
Period of Study: Summer of 1998
Location: Vancouver, Canada
Outcome: spirometry
Age Groups: Range from 54-86 yr
Mean age= 74 yr
Study Design: extended analysis of a
repeated-measures panel study
N: 16 persons with COPD
Statistical Analyses: Earlier analysis
expanded by developing mixed-effect
regression models and by evaluating
additional exposure indicators
Dose-response Investigated? No
Statistical Package: SAS V8
Pollutant: Sulfate (S04)
Averaging Time: 24 h
Mean (SD):
Ambient Sulfate: 2.0 (1.1)
Exposure to Ambient Sulfate: 0.2 (4.7)
Range (Min, Max):
Ambient Sulfate: (0.4-5.4)
Exposure to ambient Sulfate: (0.2-4.7)
Monitoring Stations: 5
Copollutant (correlation):
Ambient PM25:r= 0.82
Nonsulfate Ambient PM25: r= 0.74
Exposure to Ambient Sulfate: r= 0.82
PM Increment: Ambient Sulfate: 1.5
(IQR)
Exposure to Ambient Sulfate: 0.9 (IQR)
Notes: Effect estimates are presented
in Fig 2 and Electronic Appendix Table
1 (only available with electronic version
of article) and not provided
quantitatively elsewhere.
Reference: Ferdinands et al. (2008,
1564331
Period of Study: Aug 2004
Location: Atlanta, Georgia
Outcome: Respiratory Symptoms,
airway inflammation
Study Design: Prospective cohort
Statistical Analyses: Pearson
Correlation Analysis
Age Groups: 14-18
Pollutant: PM25
Averaging Time: 24-h avg
Mean (SD) unit: 27.2 (11.9)
Range (Min, Max): 21.7, 34.7
Copollutants (correlation):
03: r= 0.8-0.9
The study presents results qualitatively
not quantitatively.
Reference: Gent et al. (2003, 0528851
Period of Study: Apr-Sep 2001
Location: Connecticut
Springfield, MA
Outcome: Respiratory symptoms
including: Wheeze, persistent cough,
chest tightness, shortness of breath
Age Groups: Infants
Study Design: 1-yr prospective cohort
study
N: 1002 infants
171 60 observations
Statistical Analyses: Logistic
regression analysis
GEEs
Tests for linear trend
Test for goodness of fit
Hosmer-Lemeshow statistic for
regression
Pollutant: PM25
Averaging Time: 24 h
Mean (SD): 13. 1(7.9)
Percentiles: 20th: 6.9
40th: 9.0
SOth(Median): 10.3
60th: 12.1
80th: 19.0
Range (Min, Max): 3.7, 44.2
Monitoring Stations: 4 sites
Copollutant (correlation):
Temperature: 0.58
PM Increment: 12 pg/m3 same day
19 pg/m3 previous day
Model 6 (same day)
Wheeze <6. 9 =1.00
6.9-8.9 = 0.95(0.83,1.10)
9.0-12.0=1.04(0.89, 1.20)
12.1-18.9=1.05(0.92, 1.20)
> 19.0 = 0.93 (0.78, 1.11)
Persistent Cough <6.9=1.00
6.9-8.9 = 0.95(0.87, 1.04)
9.0-12.0 = 0.96(0.87,1.06)
12.1-18.9=1.00(0.91, 1.09)
> 19.0 = 0.95 (0.83, 1.09)
Chest Tightness <6. 9 =1.00
6.9-8.9=1.01 (0.86, 1.19)
9.0-12.0=1.06(0.89, 1.26)
12.1-18.9=1.24(1.06,1.45)
> 19.0 =1.05 (0.84, 1.33)
Shortness of Breath <6.9 = 1 .00
6.9-8.9=1.01(0.87,1.17)
9.0-12.0=1.03(0.87, 1.22)
December 2009
E-183
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Covariates: Temperature
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 1-day lag
12.1-18.9=1.07(0.91, 1.25)
> 19.0 =1.03 (0.83, 1.28)
Bronchodilator<6.9 = 1.00
6.9-8.9=1.04(0.99, 1.09)
9.0-12.0=1.02(0.96, 1.08)
12.1-18.9=1.04(0.99,1.09)
> 19.0 =1.02 (0.97, 1.08)
Model 6 (previous day)
Wheeze <6.9 =1.00
6.9-8.9=1.06(0.95, 1.20)
9.0-12.0=1.09(0.94, 1.28)
12.1-18.9=1.03(0.89,1.19)
> 19.0 =1.14 (0.97, 1.34)
Persistent Cough <6.9=1.00
6.9-8.9=1.04(0.94,1.14)
9.0-12.0=1.05(0.94, 1.17)
12.1-18.9=1.03(0.94, 1.14)
> 19.0= 1.12 (1.02 1.24)
Chest Tightness <6.9 =1.00
6.9-8.9=1.03(0.87, 1.23)
9.0-12.0=1.04(0.85,1.27)
12.1-18.9=1.00(0.84, 1.19)
> 19.0 =1.21 (1.00, 1.46)
Shortness of Breath <6.9 = 1.00
6.9-8.9=1.00(0.84, 1.19)
9.0-12.0=1.09(0.90, 1.31)
12.1-18.9=1.09(0.90,1.31)
> 19.0 =1.26 (1.02, 1.54)
Bronchodilator <6.9 = 1.00
6.9-8.9 = 0.98(0.94,1.03)
9.0-12.0 = 0.99(0.95, 1.03)
12.1-18.9 = 0.97(0.94, 1.01)
> 19.0 = 0.99 (0.95, 1.04)
PM2.5 + 03:
Medication Users: Same-day
Wheeze <6.9 =1.00
6.9-8.9 = 0.89(0.75, 1.29)
9.0-12.0=1.02(0.87, 1.19)
12.1-18.9 = 0.94(0.77,1.15)
> 19.0 = 0.83 (0.65, 1.06)
Persistent Cough <6.9 = 1.00
6.9-8.9 = 0.95(0.84,1.06)
9.0-12.0 = 0.97(0.86, 1.10)
12.1-18.9 = 0.94(0.77, 1.15)
> 19.0 = 0.83 (0.65, 1.06)
Chest Tightness <6.9 =1.00
6.9-8.9 = 0.90(0.74, 1.09)
9.0-12.0 = 0.97(0.79,1.18)
12.1-18.9 = 0.97(0.76,1.25)
> 19.0 = 0.76 (0.54, 1.05)
Shortness of Breath <6.9 = 1.00
6.9-8.9 = 0.95(0.80, 1.12)
9.0-12.0=1.00(0.82, 1.21)
12.1-18.9 = 0.90(0.73,1.12)
> 19.0 = 0.87 (0.65, 1.17)
Bronchodilator <6.9 = 1.00
6.9-8.9=1.03(0.98,1.08)
9.0-12.0=1.01 (0.96, 1.07)
12.1-18.9=1.02(0.95, 1.08)
> 19.0 = 0.99 (0.91, 1.07)
Previous Day
Wheeze <6.9 =1.00
6.9-8.9=1.03(0.89,1.18)
9.0-12.0=1.05(0.88, 1.24)
12.1-18.9 = 0.98(0.82, 1.17)
> 19.0 =1.05 (0.85, 1.29)
Persistent Cough <6.9 = 1.00
6.9-8.9 = 0.99(0.89, 1.11)
9.0-12.0 = 0.98(0.86,1.10)
12.1-18.9 = 0.95(0.83,1.10)
> 19.0 =1.00 (0.88, 1.15)
Chest Tightness <6.9 =1.00
6.9-8.9 = 0.89(0.72, 1.10)
9.0-12.0 = 0.90(0.70, 1.16)
12.1-18.9 = 0.81(0.63,1.03)
> 19.0 = 0.91 (0.71, 1.17)
Shortness of Breath <6.9 = 1.00
6.9-8.9 = 0.96(0.78,1.18)
December 2009
E-184
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
9.0-12.0=1.00(0.81, 1.25)
12.1-18.9 = 0.96(074, 1.24)
> 19.0 =1.20 (0.94, 1.52)
Bronchodilator <6.9 = 1.00
6.9-8.9 = 0.99(0.94, 1.04)
9.0-12.0 = 0.97(0.93,1.02)
12.1-18.9 = 0.96(0.91,1.02)
> 19.0 = 0.97 (0.89, 1.04)
PM25 + 03:
Non-users: Same-day
Wheeze <6.9 =1.00
6.9-8.9 = 0.92(0.72,1.17)
9.0-12.0=1.08(0.85, 1.36)
12.1-18.9 = 0.94(0.73, 1.22)
> 19.0 =1.15 (0.75, 1.75)
Persistent Cough <6.9 = 1.00
6.9-8.9 = 0.96(0.83, 1.12)
9.0-12.0=1.02(0.89,1.18)
12.1-18.9 = 0.93(0.78,1.12)
> 19.0 =1.07 (0.85, 1.34)
Chest Tightness <6.9 =1.00
6.9-8.9 = 0.84(0.54, 1.31)
9.0-12.0=1.09(0.74, 1.61)
12.1-18.9 = 0.78(0.47,1.30)
> 19.0 = 0.71 (0.36, 1.39)
Shortness of Breath <6.9 = 1.00
6.9-8.9 = 0.61(0.39,0.95)
9.0-12.0=1.13(0.85, 1.50)
12.1-18.9 = 0.72(0.42, 1.23)
> 19.0 =1.17 (0.72, 1.90)
Bronchodilator Use: <6.9 = 1.00
6.9-8.9 = 0.95(0.78, 1.15)
9.0-12.0 = 0.95(0.78,1.16)
12.1-18.9 = 0.85(0.69,1.06)
> 19.0 = 0.99 (0.76, 1.30)
Previous-day
Wheeze <6.9= 1.00
6.9-8.9=1.01 (0.78, 1.31)
9.0-12.0=1.15(0.88,1.51)
12.1-18.9=1.08(0.78,1.51)
a 19.0=1.18(0.71,1.97)
Persistent Cough <6.9= 1.00
6.9-8.9=1.07(0.94, 1.22)
9.0-12.0=1.13(0.97, 1.32)
12.1-18.9=1.03(0.87,1.22)
> 19.0 =1.14 (0.88, 1.46)
Chest Tightness <6.9 =1.00
6.9-8.9=1.44(0.90,2.30)
9.0-12.0=1.50(0.97,2.33)
12.1-18.9=1.56(0.91,2.66)
> 19.0 =1.76 (0.83, 3.73)
Shortness of Breath <6.9 = 1.00
6.9-8.9 = 0.99(0.75, 1.30)
9.0-12.0=1.30(0.88,1.91)
12.1-18.9 = 0.84(0.57,1.24)
> 19.0 =1.48 (0.94, 2.34)
Bronchodilator Use <6.9 = 1.00
6.9-8.9=1.05(0.85, 1.34)
9.0-12.0=1.28(1.01, 1.62)
12.1-18.9=1.05(0.80,1.37)
a 19.0=1.19(0.83,1.71)
Notes: Line graphs of daily levels of 03
and PM25 and daily temperature with
daily prevalence of respiratory
symptoms for users of asthma
maintenance medication
Reference: Gent et al, (2009,1803991 Outcome: Increased asthma symptoms Pollutant: PM25 and components
and medication use
Averaging Time: Daily
Study Design: Panel
Period of Study: 2000-2003
Location: New Haven County CT
Covariates: Season, day of the week,
date
Statistical Analysis: Logistic
regression
Statistical Package: SAS
Mean: (estimated sources, pg/m ]
Motor Vehicle: 6.6
Road Dust: 2.3
Sulfur: 5.5
Odds Ratio and p-value for
sources and components of PIVh.s.
Lags are 0,1 or 2 days, and the mean
of days 0-2 (L02).
Source: Motor Vehicle
EC, Increment = 1000ng/m3
Wheeze
L0:1.04, p = 0.04
L1:1.01, p = 0.70
December 2009
E-185
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
L2:1.00, p = 0.99
L02:1.07, p = 0.06
Persistent Cough
L0:1.01,p = 0.42
L1:1.01, p = 0.38
L2:0.99, p = 0.44
L02:1.03, p = 0.23
Shortness of Breath
LO: 1.06, p = 0.001
L1:1.01,p = 0.65
L2:1.01,p = 0.63
L02:1.12, p = 0.01
Chest Tightness
L0:1.03, p = 0.20
L1:1.02, p = 0.24
L2:1.01,p = 0.59
L02:1.10, p = 0.04
Inhaler Use
LO: 1.01, p = 0.15
L1:1.00, p = 0.72
L2:1.00, p = 0.75
L02:1.02, p = 0.40
Zn, Increment = 10ng/m3
Wheeze
L0:1.00, p = 0.69
L1: 0.99, p = 0.54
L2:1.00, p = 0.89
L02:1.00, p = 0.98
Persistent Cough
L0:1.00, p = 0.60
L1:1.00, p = 0.77
L2: 0.99, p = 0.24
L02:1.00, p = 0.94
Shortness of Breath
LO: 1.02, p = 0.001
L1:1.00, p = 0.57
L2:1.01,p = 0.49
L02:1.04, p = 0.06
Chest Tightness
L0:1.00, p = 0.72
L1:1.00, p = 0.96
L2:1.01,p = 0.38
L02:1.03, p = 0.13
Inhaler Use
L0:1.00, p = 0.41
L1:1.00, p = 0.44
L2:1.00, p = 0.52
L02:1.01,p = 0.53
Pb, Increment = 5 ng/m3
Wheeze
L0:1.02, p = 0.31
L1:1.00, p = 0.91
L2:1.01,p = 0.62
L02:1.07, p = 0.13
Persistent Cough
L0:1.02, p = 0.25
L1:1.00, p = 0.88
L2:1.00, p = 0.87
L02:1.05, p = 0.12
Shortness of Breath
L0:1.03, p = 0.11
L1: 0.98, p = 0.51
L2:1.03, p = 0.05
L02:1.12, p = 0.01
Chest Tightness
L0:1.02, p = 0.31
L1: 0.99, p = 0.79
L2:1.03, p = 0.13
L02:1.10, p = 0.02
Inhaler Use
L0:1.01,p = 0.06
L1: 0.98, p = 0.11
L2:1.02, p = 0.04
L02:1.04, p = 0.10
Cu, Increment = 5 ng/m
Wheeze
Age Groups: Children aged 4-12
Biomass Burning: 0.9
Oil: 0.8
Sea Salt: 0.5
Range (Min, Max): NR
Copollutant (correlation): NR
December 2009
E-186
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
LO: 1.01, p = 0.59
L1:0.99, p = 0.55
L2:0.99, p = 0.82
L02:1.02, p = 0.67
Persistent Cough
L0:1.02, p = 0.13
L1:1.02, p = 0.21
L2:0.98, p = 0.26
L02:1.05, p = 0.04
Shortness of Breath
L0:1.06, p = 0.01
L1:1.01,p = 0.74
L2:0.96, p = 0.10
L02:1.06, p = 0.21
Chest Tightness
L0:10.3, p = 0.23
L1:1.02, p = 0.42
L2:0.97, p = 0.17
L02:1.04, p = 0.39
Inhaler Use
L0:1.01,p = 0.22
L1:0.99, p = 0.37
L2:1.00, p = 0.70
L02:1.01,p = 0.46
Se, Increment = 1 ng/m3
Wheeze
L0:1.00, p = 0.97
L1:0.99, p = 0.52
L2:1.00, p = 0.91
L02:1.02, p = 0.71
Persistent Cough
L0:1.00, p = 0.84
L1:0.99, p = 0.32
L2:1.00, p = 0.93
L02:0.98, p = 0.43
Shortness of Breath
L0:1.02, p = 0.40
L1:0.97, p = 0.10
L2:1.01,p = 0.55
L02:1.02, p = 0.67
Chest Tightness
L0:1.00, p = 0.79
L1:0.97, p = 0.13
L2:1.01,p = 0.72
L02:0.98, p = 0.61
Inhaler Use
L0:0.99, p = 0.20
L1:1.01, p = 0.02
L2:0.99, p = 0.32
L02:0.99, p = 0.75
Source: Road Dust
Si, Increment =100 ng/m3
Wheeze
L0:1.03, p = 0.03
L1:1.00, p = 0.99
L2:1.02, p = 0.26
L02:1.07, p = 0.04
Persistent Cough
L0:1.02, p = 0.01
L1:1.00, p = 0.78
L2:1.01,p = 0.60
L02:1.05, p = 0.02
Shortness of Breatl04, p = 0.01h
L0:1.04, p = 0.01
L1:1.01,p = 0.60
L2:1.01,p = 0.63
L02:1.08, p = 0.02
Chest Tightness
L0:1.02, p = 0.20
L1:1.02, p = 0.17
L2:1.00, p = 0.88
L02:1.06, p = 0.10
Inhaler Use
LO: 1.02, p = 0.004
L1:0.99, p = 0.18
L2:1.01,p = 0.45
December 2009 E-187
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
L02:1.03, p = 0.09
Fe, Increment = 100 ng/m3
Wheeze
L0:1.04, p = 0.02
L1:1.00, p = 0.80
L2:1.00, p = 0.87
L02:1.07, p = 0.05
Persistent Cough
L0:1.02, p = 0.06
L1:1.01, p = 0.52
L2:0.99, p = 0.52
L02:1.04, p = 0.04
Shortness of Breath
LO: 1.06, p = 0.002
L1:1.01,p = 0.65
L2:0.98, p = 0.27
L02:1.08, p = 0.04
Chest Tightness
L0:1.01,p = 0.47
L1:1.02, p = 0.22
L2:0.98, p = 0.35
L02:1.05, p = 0.21
Inhaler Use
LO: 1.02, p = 0.004
L1:0.99, p = 0.44
L2:1.00, p = 0.91
L02:1.03, p = 0.08
Al, Increment = 50 ng/m3
Wheeze
L0:1.02, p = 0.17
L1:1.01,p = 0.73
L2:1.02, p = 0.30
L02:1.07, p = 0.03
Persistent Cough
LO: 1.03, p = 0.001
L1:1.00, p = 0.96
L2:1.00, p = 0.68
L02:1.06, p = 0.01
Shortness of Breath
LO: 1.05, p = 0.002
L1:1.01,p = 0.63
L2:1.01,p = 0.59
L02:1.09, p = 0.004
Chest Tightness
L0:1.02, p = 0.21
L1:1.02, p = 0.18
L2:1.00, p = 0.94
L02:1.07, p = 0.04
Inhaler Use
L0:1.02, p = 0.02
L1: 0.99, p = 0.27
L2:1.01,p = 0.50
L02:1.02, p = 0.11
Ca, Increment = 50 ng/m
Wheeze
L0:1.07, p = 0.02
L1:1.00, p = 0.97
L2:1.01,p = 0.74
L02:1.14, p = 0.04
Persistent Cough
L0:1.05, p = 0.01
L1: 0.99, p = 0.64
L2:1.00, p = 0.90
L02:1.09, p = 0.03
Shortness of Breath
LO: 1.10, p = 0.002
L1:1.02, p = 0.66
L2:1.00, p = 0.89
L02:1.18, p = 0.01
Chest Tightness
L0:1.04, p = 0.26
L1:1.03, p = 0.43
L2:1.00, p = 0.93
L02:1.14, p = 0.07
Inhaler Use
L0:1.04, p = 0.01
December 2009 E-188
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
L1:0.97, p = 0.06
L2:1.01,p = 0.44
L02:1.04, p = 0.17
Ba, Increment = 10ng/m3
Wheeze
L0:0.99, p = 0.57
L1:1.00, p = 0.92
L2:0.99, p = 0.48
L02:0.99, p = 0.81
Persistent Cough
L0:1.00, p = 0.83
L1:1.01, p = 0.38
L2:0.99, p = 0.32
L02:1.00, p = 0.81
Shortness of Breath
L0:1.04, p = 0.02
L1:1.00, p = 0.96
L2:0.96, p = 0.05
L02:1.03, p = 0.38
Chest Tightness
L0:1.01,p = 0.63
L1:1.00, p = 0.88
L2:0.98, p = 0.30
L02:1.02, p = 0.51
Inhaler Use
L0:1.01,p = 0.08
L1:0.99, p = 0.19
L2:1.00, p = 0.92
L02:1.01,p = 0.36
Ti, lncrement = 5ng/m3
Wheeze
L0:1.00, p = 0.59
L1:0.99, p = 0.49
L2:1.01,p = 0.34
L02:1.01,p = 0.56
Persistent Cough
L0:1.00, p = 0.57
L1:1.00, p = 0.55
L2:1.00, p = 0.30
L02:1.01,p = 0.29
Shortness of Breath
L0:1.01,p = 0.01
L1:1.00, p = 0.56
L2:1.00, p = 0.60
L02:1.03, p = 0.05
Chest Tightness
L0:1.00, p = 0.34
L1:1.00, p = 0.55
L2:0.99, p = 0.49
L02:1.01,p = 0.52
Inhaler Use
L0:1.00, p = 0.72
L1:1.00, p = 0.30
L2:1.00, p = 0.67
L02:1.00, p = 0.66
Source: Sulfur
S, Increments000ng/m3
Wheeze
L0:0.98, p = 0.43
L1:0.99, p = 0.62
L2:1.02, p = 0.29
L02:1.00, p = 0.99
Persistent Cough
L0:1.00, p = 0.84
L1:1.00, p = 0.69
L2:1.02, p = 0.21
L02:1.02, p = 0.27
Shortness of Breath
L0:1.01,p = 0.63
L1:0.99, p = 0.71
L2:1.01,p = 0.55
L02:1.01,p = 0.79
Chest Tightness
L0:0.99, p = 0.80
L1:1.01,p = 0.62
December 2009 E-189
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
L2:1.01, p = 0.81
L02:1.02, p = 0.68
Inhaler Use
L0:0.99, p = 0.13
L1:1.00, p = 0.81
L2:1.02, p = 0.04
L02:1.00, p = 0.81
P, lncrement = 50ng/m3
Wheeze
L0:0.98, p = 0.39
L1:0.98, p = 0.48
L2:1.02, p = 0.38
L02:0.99, p = 0.89
Persistent Cough
L0:1.00, p = 0.75
L1:0.99, p = 0.69
L2:1.01,p = 0.38
L02:1.03, p = 0.30
Shortness of Breath
L0:1.01,p = 0.61
L1:0.99, p = 0.71
L2:1.01,p = 0.67
L02:1.01,p = 0.78
Chest Tightness
L0:1.00, p = 0.88
L1:1.01,p = 0.72
L2:1.00, p = 0.87
L02:1.02, p = 0.67
Inhaler Use
L0:0.98, p = 0.15
L1:1.00, p = 0.83
L2:1.01,p = 0.11
L02:1.00, p = 0.99
Source: Biomass Burning
K, Increment = 50 ng/m3
Wheeze
L0:0.98, p = 0.06
L1:0.99, p = 0.43
L2:1.00, p = 0.85
L02:0.96, p = 0.04
Persistent Cough
L0:1.00, p = 0.64
L1:1.00, p = 0.83
L2:1.00, p = 0.46
L02:1.00, p = 0.86
Shortness of Breath
L0:1.01,p = 0.01
L1:0.98, p = 0.09
L2:1.00, p = 0.38
L02:1.00, p = 0.79
Chest Tightness
L0:1.00, p = 0.02
L1:0.99, p = 0.24
L2:0.98, p = 0.07
L02:0.99, p = 0.67
Inhaler Use
L0:1.00, p = 0.68
L1:0.99, p = 0.05
L2:1.00, p = 0.59
L02:0.99, p = 0.28
Source: Oil
V, lncrement= 10 ng/m
Wheeze
L0:0.99, p = 0.73
L1:0.96, p = 0.03
L2:0.99, p = 0.56
L02:0.93, p = 0.04
Persistent Cough
L0:1.01,p = 0.56
L1:0.99, p = 0.24
L2:0.98, p = 0.01
L02:0.96, p = 0.05
Shortness of Breath
L0:1.01,p = 0.46
L1:0.98, p = 0.24
December 2009 E-190
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
L2:1.00, p = 0.83
L02:0.98, p = 0.58
Chest Tightness
L0:0.99, p = 0.71
L1:0.98, p = 0.32
L2:0.98, p = 0.23
L02:0.94, p = 0.12
Inhaler Use
L0:0.98, p = 0.12
L1:1.00, p = 0.68
L2:0.99, p = 0.22
L02:0.96, p = 0.03
Ni, Increment = 5 ng/m3
Wheeze
L0:1.01,p = 0.59
L1:0.97, p = 0.09
L2:1.00, p = 0.76
L02:0.99, p = 0.72
Persistent Cough
LO: 1.01, p = 0.21
L1:0.99, p = 0.57
L2:0.99, p = 0.23
L02:1.00, p = 0.99
Shortness of Breath
L0:1.04, p = 0.05
L1:0.98, p = 0.36
L2:1.00, p = 0.81
L02:1.04, p = 0.32
Chest Tightness
L0:1.01,p = 0.58
L1:1.00, p = 0.89
L2:0.98, p = 0.27
L02:1.01,p = 0.84
Inhaler Use
L0:1.01,p = 0.48
L1:1.00, p = 0.83
L2:0.99, p = 0.51
L02:1.01,p = 0.48
Source: Sea Salt
Na, Increment = 100 ng/m3
Wheeze
L0:0.98, p = 0.23
L1:1.00, p = 0.80
L2:1.00, p = 0.88
L02:0.97, p = 0.29
Persistent Cough
L0:1.00, p = 0.58
L1:0.99, p = 0.19
L2:1.00, p = 0.61
L02:0.98, p = 0.21
Shortness of Breath
L0:1.00, p = 0.94
L1:0.99, p = 0.46
L2:1.01,p = 0.63
L02:0.99, p = 0.74
Chest Tightness
L0:0.99, p = 0.43
L1:0.99, p = 0.75
L2:1.00, p = 0.88
L02:0.98, p = 0.61
Inhaler Use
L0:0.99, p = 0.35
L1:1.00, p = 0.61
L2:1.00, p = 0.85
L02:0.99, p = 0.37
Cl, Increment = 10 ng/m3
Wheeze
L0:1.00, p = 0.89
L1:1.00, p = 0.88
L2:1.00, p = 0.38
L02:1.00, p = 0.81
Persistent Cough
L0:1.00, p = 0.31
L1:1.00, p = 0.31
L2:1.00, p = 0.51
December 2009 E-191
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
L02:1.00, p = 0.06
Shortness of Breath
L0:1.00, p = 0.89
L1:1.00, p = 0.94
L2:1.00, p = 0.70
L02:1.00, p = 0.80
Chest Tightness
L0:1.00, p = 0.24
L1:1.00, p = 0.28
L2:1.00, p = 0.52
L02:1.00, p = 0.65
Inhaler Use
L0:1.00, p = 0.69
L1:1.00, p = 0.52
L2:1.00, p = 0.51
L02:1.00, p = 0.83
Odds Ratio (96%CI) from repeated
measures logistic regression models
of respiratory symptoms and daily
source concentrations of PMu.
Lag 0 Model
Wheeze, p = 0.23
Motor Vehicle: 1.05 (0.99-1.10)
Road Dust: 1.10 (1.01-1.19)
Sulfur: 0.97 (0.94-1.00)
Biomass Burning: 0.80 (0.66-0.98)
011:1.02(0.86-1.20)
Sea Salt: 0.96 (0.86-1.07)
Persistent Cough, p < 0.001
Motor Vehicle: 1.02 (0.99-1.04)
Road Dust: 1.06 (1.01-1.11)
Sulfur: 1.00 (0.98-1.01)
Biomass Burning: 0.97 (0.92-1.03)
011:1.02(0.95-1.10)
Sea Salt: 0.99 (0.92-1.07)
Shortness of Breath, p< 0.001
Motor Vehicle: 1.06 (1.01-1.11)
Road Dust: 1.12 (1.02-1.22)
Sulfur: 0.98 (0.94-1.02)
Biomass Burning: 1.05 (0.95-1.17)
011:1.07(0.92-1.26)
Sea Salt: 1.01 (0.92-1.12)
Chest Tightness, p < 0.001
Motor Vehicle: 1.02 (0.97-1.08)
Road Dust: 1.04 (0.95-1.15)
Sulfur: 0.99 (0.94-1.03)
Biomass Burning: 1.06 (0.95-1.18)
011:0.99(0.82-1.18)
Sea Salt: 0.95 (0.84-1.08)
Inhaler Use, p < 0.001
Motor Vehicle: 1.02 (1.00-1.05)
Road Dust: 1.06 (1.02-1.11)
Sulfur: 0.98 (0.97-1.00)
Biomass Burning: 1.00 (0.96-1.03)
011:0.98(0.91-1.05)
Sea Salt: 0.99 (0.94-1.04)
Lag 02 Model
Wheeze, p = 0.86
Motor Vehicle: 1.10(1.01-1.19)
Road Dust: 1.26 (1.05-1.51)
Sulfur: 0.98 (0.92-1.04)
Biomass Burning: 0.64 (0.46-0.88)
011:0.80(0.56-1.08)
Sea Salt: 0.91 (0.82-1.16)
Persistent Cough, p < 0.001
Motor Vehicle: 1.03 (0.98-1.09)
Road Dust: 1.16 (1.02-1.32)
Sulfur: 1.01 (0.98-1.05)
Biomass Burning: 0.93 (0.81-1.06)
011:0.84(0.71-1.00)
Sea Salt: 0.88 (0.77-1.01)
December 2009 E-192
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
Shortness of Breath, p = 0.006
Motor Vehicle: 1.12 (1.01-1.24)
Road Dust: 1.28 (1.05-1.55)
Sulfur: 0.97 (0.90-1.04)
Biomass Burning: 0.78 (0.52-1.18)
011:0.94(0.69-1.29)
Sea Salt: 1.01 (0.79-1.29)
Chest Tightness, p = 0.39
Motor Vehicle: 1.08 (0.98-1.20)
Road Dust: 1.20 (0.97-1.49)
Sulfur: 1.00 (0.92-1.08)
Biomass Burning: 0.87 (0.62-1.22)
011:0.80(0.58-1.10)
Sea Salt: 0.95 (0.71-1.27)
Inhaler Use, p < 0.001
Motor Vehicle: 1.03 (0.98-1.08)
Road Dust: 1.09 (1.00-1.19)
Sulfur: 1.00 (0.97-1.03)
Biomass Burning: 0.95 (0.87-1.04)
011:0.92(0.81-1.05)
Sea Salt: 0.97 (0.88-1.07)
Odds Ratio (96%CI) from repeated
measures logistic regression models
of respiratory symptoms and daily
source concentrations of PMu when
copollutants are included.
Wheeze
Motor Vehicle
N02:1.03 (0.98-1.08)
00:1.05(0.99-1.11)
S02:1.04 (0.99-1.09)
03:1.06 (0.97-1.16)
Road Dust
N02:1.11 (1.02-1.20)
00:1.10(1.01-1.19)
S02:1.10 (1.01-1.19)
03:1.11 (1.01-1.23)
Sulfur
N02: 0.96 (0.92-0.99)
00:0.97(0.94-1.01)
S02: 0.97 (0.93-1.00)
03: 0.95 (0.91-1.00)
Biomass Burning
N02: 0.79 (0.65-0.98)
CO: 0.80 (0.66-0.98)
S02: 0.79 (0.64-0.98)
03: 0.74 (0.57-0.97)
Oil
N02:1.02 (0.87-1.21)
00:1.02(0.86-1.20)
S02:1.01 (0.86-1.19)
03: 0.92 (0.62-1.39)
Sea Salt
N02: 0.96 (0.85-1.07)
00:0.96(0.86-1.08)
S02: 0.95 (0.85-1.07)
03:1.01 (0.72-1.40)
Inhaler Use
Motor Vehicle
N02:1.02 (0.99-1.04)
00:1.02(0.99-1.05)
S02:1.02 (0.99-1.04)
03:1.02 (0.98-1.07)
Road Dust
N02:1.06 (1.02-1.10)
00:1.06(1.02-1.11)
S02:1.06 (1.02-1.11)
03:1.06 (1.00-1.13)
Sulfur
N02: 0.98 (0.96-1.00)
00:0.98(0.96-1.00)
S02: 0.98 (0.96-1.00)
December 2009 E-193
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
03: 0.97 (0.95-1.00)
Biomass Burning
N02:1.00 (0.96-1.03)
00:0.99(0.96-1.03)
S02: 0.99 (0.96-1.03)
03: 0.99 (0.95-1.03)
Oil
N02: 0.98 (0.91-1.05)
00:0.97(0.91-1.04)
S02: 0.97 (0.91-1.04)
03:1.03 (0.88-1.22)
Sea Salt
N02: 0.99 (0.94-1.04)
00:0.99(0.94-1.04)
S02: 0.99 (0.94-1.04)
03:1.01 (0.88-1.15)
Reference: Girardot et al. (2006,
0882711
Period of Study:
Aug 2002-Oct 2002
Jun 2003-Aug 2003
Outcome: Pulmonary
function/spirometry-FVC, FEV,, PEF,
FVC/FEV,, FEF25-75
Age Groups: 18-82yr
Study Design: Cohort
Location: Charlies Bunion Trail (portion N: 354 hikers
of Appalachia Trail)
Statistical Analyses: Multiple linear
regression
Covariates: Age, h hiked, mean
temperature, sex, smoking status,
history of asthma or wheeze symptoms,
carriage of backpack, whether reaching
summit or not
Season: Fall 2002, Summer 2003
Dose-response Investigated? No
Statistical Package: SAS
Pollutant: PM25
Averaging Time: 24 h
Mean:
Trail: 13.9 ±8.2
Estimated personal: 15.0 ± 7.4
Range (Min, Max):
Trail: 1.6, 38.4
Estimated personal:
0.21,41.9
Copollutant (correlation): 0; (i-u.67.
for estimated personal exposure)
PM Increment: 1 pg/m
% Change ± Cl
p value
Univariate:
FVC: 0.023 ± 0.035
0.51
FEV,: 0.015 ±0.029
0.607
PEF: 0.185 ±0.091
0.043
FVC/FEV,: 0.003 ±0.023
0.905
FEF25-75%: 0.052 ± 0.093
0.578
Adjusted: FVC: 0.007+/0.040
0.966
FEV,: 0.003 ±0.033
0.937
PEF: 0.258 ±0.103
0.013
FVC/FEV,:-0.011 ±0.027
0.676
FEF25-75%:-0.041 ±0.109
0.707
Spirometry result for each quintile ± Cl
Quintile 1 (6.0 pg/m3): FVC (L):
Prehike: 4.32 ±0.13
Posthike:4.33±0.12
FEV,(L): Prehike: 3.39 ±0.10
Posthike:3.40±0.10
FEV,/FVC (%): Prehike: 78.66 ± 0.86
Posthike: 78.63 ± 0.81
FEF25-75% (L/sec): Prehike: 3.27 ±
0.14
Posthike: 3.26 ±0.14
PEF (L/sec): Prehike: 7.91+/0.22
Posthike: 7.58 ±0.22
Quintile 2 (10.4 pg/m3): FVC (L):
Prehike: 4.30 ±0.11
Posthike: 4.30 ±0.11
FEV, (L): Prehike: 3.42 ± 0.09
Posthike: 3.43 ±0.09
FEV,/FVC (%): Prehike: 79.37 ± 0.71
Posthike: 79.55 ± 0.69
FEF25-75% (L/sec): Prehike: 3.39 ±
0.14
Posthike: 3.38 ±0.14
PEF (L/sec): Prehike: 8.37+/0.23
Posthike: 8.26 ±0.25
Quintile 3 (14.8 pg/m3): FVC (L):
Prehike: 4.34 ±0.12
Posthike: 4.33 ±0.12
FEV, (L): Prehike: 3.42 ±0.10
Posthike: 3.40 ±0.09
FEV,/FVC (%): Prehike: 79.20 ± 0.81
Posthike: 78.83 ±0.80
FEF25-75% (L/sec): Prehike: 3.19 ±
0.13
Posthike: 3.21 ±0.13
PEF (L/sec): Prehike: 8.12+/0.25
Posthike: 7.89 ±0.25
Quintile 4 (17.9 |jg/m3):
December 2009
E-194
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
FVC(L):Prehike: 4.23 ±0.11
Posthike: 4.23 ±0.11
FEV, (L): Prehike: 3.36 ±0.10
Posthike: 3.36 ±0.10
FEV,/FVC (%): Prehike: 79.18 ± 0.81
Posthike: 79.26 ±0.79
FEF25-75% (L/sec): Prehike: 3.34 ±
0.15
Posthike: 3.30 ±0.15
PEF (L/sec): Prehike: 7.75+/0.25
Posthike: 7.73 ±0.26
Quintile6(26.6|jg/m3):FVC(L):
Prehike: 4.15 ±0.11
Posthike: 4.18 ±0.12
FEV, (L): Prehike: 3.31 ± 0.09
Posthike: 3.33 ±0.10
FEV,/FVC (%): Prehike: 79.73 ± 0.66
Posthike: 79.55 ±0.64
FEF25-75% (L/sec): Prehike: 3.22 ±
0.14
Posthike: 3.24 ±0.14
PEF (L/sec): Prehike: 7.72+/0.22
Posthike: 7.77 ± 0.23
Overall (16.0 pg/m3): FVC (L): Prehike:
4.27 ± 0.05
Posthike: 4.27 ± 0.05
FEV,(L): Prehike: 3.38 ±0.04
Posthike: 3.38 ±0.04
FEV,/FVC (%): Prehike: 79.2 ± 0.34
Posthike: 79.2 ± 0.33
FEF25-75% (L/sec): Prehike: 3.28 ±
0.06
Posthike: 3.28 ±0.06
PEF (L/sec): Prehike: 7.97+/0.11
Posthike: 7.97 ±0.11
Reference: Hertz-Picciotta et al. (2007,
1359171
Period of Study: 1994-2003
Location: Teplice and Prachatice,
Czech Republic
Outcome: Lower respiratory
illness-croup (JOS, J04), acute
bronchitis (J20), acute bronchiolitis
(J21)
Age Groups: Neonates followed for
2-4. Syr
Study Design: Cohort
N: 1133 children
Statistical Analyses: Generalized
linear longitudinal models
Covariates: District, mother's age,
mother's education, mother or adult
smoke, child's sex, season, day of the
week, fuel for heating and/or cooking,
breastfeeding category, number of other
children, temperature
Season: Winter, spring, summer and
Pollutant: PM25
Averaging Time: 24 h
Mean (SD):
PAH: 22.3 (SD-16 for 3-day avg and 11
for 45-day avg)
PM Increment: 25 pg/m
RR Estimate [Lower Cl, Upper Cl]
lag:
Birth-23 mo:
1.30 [1.08, 1.58] lag 1-30
2-4.5 yr:
1.23 [0.94, 1.62] lag 1-30
RR Estimate for categories of
exposure [Lower Cl, Upper Cl] lag:
Crude RR:
Birth-23 mo:
> 50 pg/m3: 2.26 [1.81, 2.82] lag 1-30
25-50 pg/m3:1.48 [1.32,1.65] lag 1-30
< 25 pg/m3:
Referent
2-4.5 yr:
> 50 pg/m3: 3.66 [2.07, 6.48] lag 1-30
25-50 pg/m3:1.60 [1.41,1.82] lag 1-30
< 25 pg/m3:
Referent
Dose-response Investigated? No
Statistical Package: SUDAAN version
Lags Considered: 1-3,1-7,1-14,1-30,
1-45
December 2009
E-195
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Hertz- Picciotta et al.
(2007, 1359171
Period of Study: 1994-2003
Location: Teplice and Prachatice,
Czech Republic
Outcome: Lower respiratory
illness-croup (JOS, J04), acute
bronchitis (J20), acute bronchiolitis
(J21)
Age Groups: Neonates followed for
2-4. Syr
Study Design: Cohort
N: 1133 children
Statistical Analyses: Generalized
linear longitudinal models
Covariates: District, mother's age,
mother's education, mother or adult
smoke, child's sex, season, day of the
week, fuel for heating and/or cooking,
breastfeeding category, number of other
children, temperature
Dose-response Investigated? No
Statistical Package: SUDAAN version
Pollutant: PM25
Averaging Time: 24 h
Mean (SD):
PAH:
52.5 ng/m3 (SD-57 ng/m3 for 3-day avg
and 46 ng/m3 for 45-day avg)
PAH Increment: 100 ng/m
RR Estimate [Lower Cl, Upper Cl]
lag:
Birth-23 mos:
1.29 [1.07, 1.54] lag 1-30
2-4.5 yr:
1.56 [1.22, 2.00] lag 1-30
RR Estimate for categories of exposure
[Lower Cl, Upper Cl] lag:
Crude RR:
Birth-23 mos:
> 100 ng/m3: 2.52 [2.22, 2.87] lag 1-30
40-100 ng/m3:1.87 [1.65, 2.13] lag 1-30
< 40 ng/m3: Reference
2-4.5 yr:
> 100 ng/m3: 2.26 [1.93, 2.65] lag 1-30
40-100 ng/m3:1.40 [1.20,1.64] lag 1-30
< 40 ng/m3: Reference
Lags Considered: 1-3,1-7,1-14,1-30,
1-45
Reference: Hogervorst, et al. (2006,
A ccccn\
1bbbb9)
Period of Study: 2002
Location: Maastricht, the Netherlands
(six schools selected)
Outcome: Decreased lung function
Age Groups: 8-1 Syr
Study Design: Multivariate linear
regression (enter method) analysis
N: 342 children
Statistical Analyses: ANOVA, chi
square
Covariates: Independent variables:
Pollutant: PM25
Averaging Time: Daily
Mean (SD): 19.0 (3.2)
Monitoring Stations: 6
Co pollutant:
PM10
Total Suspended Particles (TSP)
PM Increment: 10 pg/m3
RR Estimate [Lower Cl, Upper Cl]
lag:
FEV: 3.62 [0.50,7.63] lag NR
FVC: 1.80 [-2.10, 5.80] lag NR
FEF: 5.93 [-2.34, 14.89] lag NR
Age, height, gender, smoking at home
by parents, pets, use of ventilation
hoods during cooking, presence of
unvented geysers, tapestry in the
home, indoor/outdoor time, education
level of parents.
Dependent variables: lung function
indices
Dose-response Investigated? No
Reference: Holguin et al, (2007,
0990001
Period of Study:
Location: Ciudad Juarez, Mexico
Outcome: FeNO, FEV,
Study Design: Panel
Covariates: sex, age, body mass
index, day of week, season, yr of
maternal and paternal education,
passive smoking
Statistical Analysis: linear and
nonlinear mixed effects models
Age Groups: 6-12 yr
Pollutant: PM25
Averaging Time: 48 h
Mean (SD) Unit: 17.5 (8.9) pg/m3
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: NR
Relative Risk (Min Cl, Max Cl)
Lag
Results not given in table form, but
abstract states that no significant
associations with PM25 were observed.
December 2009
E-196
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Hong et al. (2007, 0913471 Outcome: Peak expiratory flow rate
Period of Study: Mar 23-May 2004 ' '
Age Groups: 3rd-6th grade (mean
Location: School on the Dukjeok Island age=9.6 yr)
near Incheon City, Korea
Study Design: Panel study
N: 43 schoolchildren
Statistical Analyses: Mixed linear
regression
Covariates: age, sex, height, weight,
asthma history, and passive smoking
exposure at home
Dose-response Investigated? No
Lags Considered :0,1,2,3,4,5
Pollutant: PM25
Averaging Time: 24 h
Mean (SD): 20.27 (8.23)
SOth(Median): 22.07
Range (Min, Max): 5.94-36.28
Co pollutant:
PM10
Components of PM10
(Fe, Mn, Pb.Zn.AI)
Effect Estimate:
Regression coefficients of morning and
daily mean PEFR on PM25
Lag 1 (PM25)
Morning PEFR
Crude: B=-0.14, p=0.12
Adjusted: IS= -0.54, p,0.01
Mean PEFR
Crude: IS= -0.15, p=0.02
Adjusted: IS= -0.54, p,0.01
Regression coefficients of morning and
daily mean PEFR on PM25 and GSTM1
and GSTT1 genotype using linear
mixed-effects regression
Lag 1 (PM25)
Morning PEFR: IS= -0.57, p < 0.01
Mean PEFR: IS= -0.56, p < 0.01
GSTM1
Morning PEFR: IS= 20.04, p=0.25
Mean PEFR: IS= 18.75, p=0.28
GSTT1
Morning PEFR: IS= 2.31, p=0.89
Mean PEFR: S=1.75, p=0.91
Reference: Jansen, et al. (2005,
0822361
Period of Study: 1987-2000
Location: Seattle, WA
Outcome: FENO: fractional exhaled
nitrogen oxide, Spirometry, Blood
pressure, Sa02: oxygen saturation,
Pulse rate
Age Groups: 60-86-yr-old
Study Design: Short-term cross-
sectional case series
N: 16 subjects diagnosed with COPD,
asthma, or both
Statistical Analyses: Linear mixed
effects model with random intercepts
Covariates: Age, relative humidity,
temperature, medication use
Season: Winter 2002-2003
Dose-response Investigated? No
Statistical Package: STATA
Pollutant: PM25
Averaging Time: 24 h
Mean (SD):
Fixed-Site Monitor: 14.0
All Subjects (N=16)
Indoor, home: 7.29
Outdoor, home: 10.47
Asthmatic Subjects (N=7)
Indoor, home: 7.25
Outdoor, home: 8.99
COPD Subjects (N=9)
Indoor, home: 7.33
Outdoor, home: 11.66
Range (Min, Max):
Fixed-Site Monitor: 1.3, 44
IQR
All Subjects
Indoor, home: 4.05
Outdoor, home: 8.87
Asthmatic Subjects
Indoor, home: 5.72
Outdoor, home: 7.55
COPD Subjects
Indoor, home: (3.18
Outdoor, home: 6.71
PM Increment: PM25:10 |ig/m
Slope [95% Cl]: dependence of FENO
concentration [ppb] on PM25
Asthmatic Subjects
Indoor, home: 3.69 [-0.74: 8.12]
Outdoor, home: 4.23 [1.33: 7.13]*
Copd Subjects
Indoor, home: -0.35 [-7.45: 6.75]
Outdoor, home: 3.83 [-1.84: 9.49]
Results indicate that FENO may be a
more sensitive biomarker of PM
exposure than other traditional health
endpoints.
December 2009
E-197
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Johnston, et al. (2006,
0913861
Period of Study: 7 mo
(Apr-Nov 2004)
Location: Darwin, Australia
Outcome: Asthma symptoms
Age Groups: All Ages
Study Design: Time-series
N: 251 people
(130 adults, 121 children
Statistical Analyses: Logistic
regression model
Covariates: Minimum air temperature,
doctor visits for influenza and the
prevalence of asthma symptoms and,
the fungal spore count and both onset
of asthma symptoms and
commencement of reliever medication
Season: "Dry season"- note Southern
Hemisphere
Dose-response Investigated? No
Statistical Package: STATA8
Lags Considered: 0-5 days
Pollutant: PM25
Averaging Time: Daily
Mean (SD): 11.1(5.4)
Range (Min, Max): 2.2, 36.5
PM Component: Vegetation fire smoke
(95%) and motor vehicle emissions
(5%)
Monitoring Stations: 1
PM Increment: 5 pg/m
RR Estimate [Lower Cl, Upper Cl]
lag:
Symptoms attributable to asthma
Overall: 1.000 (0.98,1.01)
Adults: 1.000 (0.976,1.026)
Children: 1.008 (0.980,1.037)
Using preventer: 1.013 (0.990,1.037)
Became symptomatic
Overall: 1.150 (1.07,1.23)
Adults: 1.165 (1.058,1.284)
Children: 1.148 (1.042,1.264)
Using preventer: 1.181 (1.076,1.296)
Used Reliever
Overall: 1.000 (0.98,1.02)
Adults: 1.007 (0.980, 1.035)
Children: 1.002 (0.972,1.034)
Using preventer: 1.020 (1.000,1.042)
Commenced Reliever
Overall: 1.120 (1.03,1.210)
Adults: 1.141 (1.021, 1.275)
Children: 1.112 (0.994,1.243)
Using preventer: 1.129 (1.013,1.257)
Commenced Oral Steroids
Overall: 1.310 (103,1.66)
Adults: 1.601 (1.192,2.150)
Children: 0.995 (0.625,1.459)
Using preventer: 1.350 (1.040,1.752)
Asthma Attack
Overall: 0.980 (0.94,1.04)
Adults: 1.026 (0.962, 1.095)
Children: 0.832 (0.731, 0.946)
Using preventer: 1.002 (0.934,1.075)
Exercise induced asthma
Overall: 0.990 (0.95,1.03)
Adults: 0.998 (0.943, 1.056)
Children: 0.982 (0.899,1.071)
Using preventer: 1.002 (0.942,1.067)
Saw a health professional for asthma
Overall: 1.030 (0.91,1.16)
Adults: 1.079 (0.899, 1.296)
Children: 1.003 (0.841,1.195)
Using preventer: 0.980 (0.847,1.133)
Missed school or work due to
asthma
Overall: 1.025 (0.9284,1.131)
Adults: 1.077 (0.923, 1.247)
Children: 1.000 (0.873,1.458)
Using preventer: 1.005 (0.897,1.124)
Mean daily number of asthma
symptoms
Overall: 1.003 (0.99,1.01)
Adults: 0.998 (0.984, 1.012)
Children: 1.004 (0.985,1.023)
Using preventer: 1.013 (0.999,1.028)
Mean Daily number of applications of
reliever
Overall: 1.002 (0.993,1.010)
Adults: 1.001 (0.986, 1.016)
Children: 1.000 (0.980,1.021)
Using preventer: 1.005 (0.994,1.017)
December 2009
E-198
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Koenig et al. (2003,
1566531
Period of Study: Winter 2000-2001,
Spring 2001
Location: Seattle, WA
Outcome: Exhaled NO (eNO)
Age Groups: 6-1 Syr old
Study Design: Cohort
N: 19 children
Statistical Analyses: Linear mixed-
effects regression
Covariates: Medication use, ambient
NO reading for specific individual on
specific day of session, mean ambient
NO for subject during session, mean
ambient NO for subject during all
sessions
Season: Winter, Spring
Dose-response Investigated? No
Statistical Package: STATA
Pollutant: PM25
Averaging Time: 10 consecutive days
Mean (SD): Outdoor: 13.3(1.4)
Indoor: 11.1 (4.9)
Personal: 13.4 (3.2)
Central-site: 10.1 (5.7)
Range (Min, Max): Outdoor: Max: 40.4
Indoor: Max: 36.3
Personal: Max: 49.4
Central-site: NR
Monitoring Stations: Outdoor: NR
Indoor: NR
Personal: NR
Central-site: 3
Copollutant (correlation): Outdoor
PM-central-site NO: 0.50
For NO values < 100 ppb, outdoor PM-
central-site NO: 0.04
PM Increment: 10 pg/m
Results presented as change in eNO
(95% Cl)
Among ICS* nonuser
Personal monitor 4.48 (1.02, 7.93)
Outdoor monitor 4.28 (1.38, 7.17)
Indoor monitor 4.21 (1.02,7.41)
Central site 3.82 (1.22, 6.43)
Among ICS* user
Personal monitor-0.09 (-2.39, 2.21)
Outdoor monitor 0.74 (-2.28, 3.76)
Indoor monitor-1.11 (-5.08,2.87)
Central site 1.28 (-1.23, 3.79)
* ICS: Inhaled corticosteroid
Reference: Koenig et al. (2003,
1566531
Period of Study: Wnter 2000-2001,
Spring 2001
Location: Seattle, WA
Outcome: Increased exhaled nitric
oxide (eNO)
Age Groups: 6-13 yr of age
Study Design: Combined recursive
and predictive model
N: 19 children with asthma
Statistical Analyses: Linear mixed
effects model
Covariates: Residence type, air
cleaner, avg outdoor temperature, avg
daily rainfall
Season: Wnter Spring
Dose-response Investigated? No
Statistical Package:
STATA 7.0 for health analyses, SAS 8.0
Pollutant: PM25
Averaging Time: Daily
Mean: Home indoor 9.5
Home outdoor 11.1
Recursive model Bag: 7.0
Recursive model Eig:2.1
Predictive model Bag: 6.0
Predictive model Eig: 4.0
Combined model Bag: 6.4
Combined model Eig: 3.2
25th: Home indoor 5. 7
Home outdoor 6.3
Recursive model Bag: 4.2
Recursive model Eig: 0.0
Predictive model Bag: 3.4
Predictive model Eig: 0.9
Combined model Bag: 3.7
Combined model Eig: 0.5
60th(Median): Home indoor 7.6
Home outdoor 9.5
Recursive model Bag: 5.9
Recursive model Eig: 1.2
Predictive model Bag: 5.0
Predictive model Eig: 2.2
Combined model Bag: 5.5
Combined model Eig: 1.7
76th: Home indoor 10.8
Home outdoor 14.6
Recursive model Bag: 9.2
Recursive model Eig: 2.3
Predictive model Bag: 7.5
Predictive model Eig: 4.9
Combined model Bag: 7.8
Combined model Eig: 4.2
PM Increment: 10-pg/m3
RR Estimate [Lower Cl, Upper Cl]
Ian1
lag.
Eag= ambient-generated personal
exposure
Eig= indoor-generated personal
exposure
eNO= exhaled nitric oxide
Recursive model with 8 children, Eag
was marginally associated with
increases in eNO [5.6 ppb [-0.6,11.9].
Eig was not associated with eNO (-0. 19
ppb).
For those combined estimates, only
Eag was significantly associated with
an increase in eNO:
Eag: 5.0 ppb [0.3, 9.7]
Eig: 3.3 ppb [1.1, 7.7]
Notes: Effects were seen only in
children who were not using
corticosteroid therapy
Range (Min, Max): Home indoor 2.3,
36.3
Home outdoor 2.8, 40.4
Recursive Eag: 1.8,22.6
Recursive Eig: 0.0,17.2
Predictive Eag: 1.3,22.6
Predictive Eig: 0.0,33.0
Combined Eag: 1.3,22.6
Combined Eig: 0.0,33.0
Monitoring Stations: 19 personal
environmental monitors
December 2009
E-199
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Kongtip et al. (2006,
0969201
Period of Study: Sep-Oct 2004
Location: Dindang district, Bangkok
metropolitan, Thailand
Outcome: respiratory and other
Outcomes reported
Age Groups: Age range 15-55 yr
Study Design: Panel study
N: 77 street vendors
Statistical Analyses: Binary logistic
regression
Covariates: Gender, age, type of fuel
used, working duration (months)
Dose-response Investigated? No
Pollutant: PM25
Averaging Time: 24 h
Mean (SD): 70.94
Percentiles: SOth(Median): 72.05
Range (Min, Max): 23.20-120.00
Monitoring Stations: 1
Copollutant (correlation):
S02
N02
03
VOCs
CO
PM Increment: 1 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
Model 1
Headache: 1.011 (0.999-1.022)
Nose congestion: 1.006 (0.997-1.015)
Sore throat: 1.000 (0.991-1.008)
Cold: 1.006 (0.995-1.017)
Cough: 0.989 (0.980-0.998)
Phlegm: 0.998 (0.992-1.003)
Chest tightness: 0.995 (0.955-1.036)
Fever: 1.008 (0.993-1.024)
Eye irritation: 1.022 (1.011-1.033)
Dizziness: 1.027 (1.013-1.041)
Weakness: 0.996 (0.983-1.008)
Upper respiratory symptom: 1.001
(0.994-1.008)
Lower respiratory symptom: 0.997
(0.992-1.002)
Model 2
Headache: 1.004 (0.996-1.013)
Nose congestion: 1.003 (0.996-1.010)
Sore throat: 0.995 (0.989-1.001)
Cold: 0.996 (0.988-1.004)
Cough: 0.990 (0.983-0.996)
Phlegm: 0.995 (0.991-0.999)
Chest tightness: 0.997 (0.970-1.025)
Fever: 1.010 (0.998-1.022)
Eye irritation: 1.019 (1.010-1.028)
Dizziness: 1.020 (1.009-1.032)
Weakness: 1.003 (0.994-1.012)
Upper respiratory symptom: 0.995
(0.990-1.000)
Lower respiratory symptom: 0.995
(0.991-0.999)
Reference: Lagorio et al. (2006,
0898001
Period of Study:
May-Jun1999 and Nov-Dec 1999
Location: Rome, Italy
Outcome: Lung function (FVC and
FEV,) of subjects with COPD, Asthma
Age Groups: COPD 50-80 yr
Asthma 18-64 yr
Study Design: Time series
N:COPD= 11
Asthma = 11
Statistical Analyses: Non-parametric
Spearman correlation
GEE
Covariates: COPD and IHD: daily
mean temperature, season variable
(spring or winter), relative humidity, day
of week
Asthma: season variable, temperature,
humidity, and (3-2-agonist use
Season: Spring and Winter
Dose-response Investigated? Yes
Statistical Package: STATA
Lags Considered: 1-3 days
Pollutant: PM25
Averaging Time: 24 h
Mean (SD):
Overall: 27.2 (19.4)
Spring: 18.2 (5.0)
Winter: 36.7 (24.1)
Range (Min, Max): 4.5, 100
PM Component:
Cd: 0.46+0.40 ng/m3
Cr: 1.9+1.7 ng/m
Fe: 283+167 ng/m3
Ni: 4.8+6.5 ng/m3
Pb: 30.6+19.0 ng/m3
Pt: 5.0+8.6 pg/nr
V: 1.8+1. 4 ng/m3
Zn: 45.8+33.1 ng/m3
Monitoring Stations: 2 fixed sites:
(Villa Ada and Istituto superior di Sanita)
Copollutant (correlation):
N02r = 0.43
03r = -0.51
CO r = 0.67
S02r = 0.34
PM10.25r = 0.34
PMi0r = 0.93
PM Increment: 1 pg/m3
They observed negative association
between ambient PM25 and respiratory
function (FVC and FEV,) in the COPD
panel. The effect on FVC was seen at
lag 24 h, 48 h, and 72 h. The effect on
FEVi was evident at lag 72 h. There
was no statistically significant effect of
PM2 5 on FVC and FEV, in the
asthmatic and IHD panels.
P Coefficient (SE)
COPD
FVC(%)
24 h -0.80 (0.36)
48-h -0.89 (0.41)
72-h -1.10 (0.55)
FEV,(%)
24 h -0.47 (0.33)
48-h -0.69 (0.37)
72-h -1.06 (0.50)
Asthma
FVC(%)
24 h -0.1 4 (0.29)
48-h -0.07 (0.33)
72-h -0.06 (0.39)
FEV,(%)
24 h -0.30 (0.34)
48-h -0.36 (0.39)
72-h -0.40 (0.46)
December 2009
E-200
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Lee et al. (2007, 0930421
Period of Study: 2000-2001
Location: South-Western Seoul
Metropolitan area, Seoul, South Korea
Outcome: PEFR (peak expiratory flow
rate), lower respiratory symptoms (cold,
cough, wheeze)
Age Groups: 61-89 yr of age (77.8
mean age)
Study Design: longitudinal panel
survey
N: 61 adults
Statistical Analyses: SAS MIXED,
logistic regression model
Covariates: Temperature (Celsius),
relative humidity, age,
Dose-response Investigated? No
Statistical Package: SAS 8.0
Lags Considered: 0-4 days
Pollutant: PM25
Averaging Time: 24 h
Mean (SD): 51.15 (19.94)
Percentiles:
25th: 33.00
SOth(Median): 53.20
75th: 87.54
Range (Min, Max):
17.94, 92.71
Monitoring Stations: 2
PM Increment: 10 pg/m
Effect Estimate [Lower Cl, Upper Cl]
lag:
PEFR (peak expiratory flow rate)
-0.54 (-0.89,-0.19)
1day
relative odds of a lower respiratory
symptom (cold, cough, wheeze)
0.976(0.849,1.121)
1day
Reference: Lewis et al. (2005, 0810791
Period of Study: Wnter 2001-Spring
2002
Location: Detroit, Michigan, USA
Outcome: Poorer lung function
(increased diurnal variability and
decreased forced expiratory volume)
Age Groups: 7-11 yrold
Study Design: Longitudinal cohort
study
N: 86 children
Statistical Analyses: Descriptive
statistics and bivariate analyses of
exposures, multivariable regression
multivariate analog of linear regression.
Covariates: Sex, home location,
annual family income, presence of one
or more smokers in household, race,
season (entered as dummy variables),
and parameters to account for
intervention group effect.
Season: Wnter 2001 (Feb 10-23),
Spring 2001 (May 5-1 8), Summer 2001
(Jul 14-27), Fall 2001 (Sep22-0ct 5),
Wnter 2002 (Jan 18-31), and Spring
2002 (May 18-31)].
Dose-response Investigated? No
Lags Considered: 1-2 days, 3-5 days
Pollutant: PM25
Averaging Time: 2 wk
Mean (SD):
Eastside
15.7(10.6)
Southwest
17.5(12.2)
Range (Min, Max): 1.0, 56.1
Monitoring Stations: 2
Copollutant (correlation):
PM100.93
03 Daily mean 0.57
03 8-h peak 0.53
PM Increment: 12.5 pg/m3
RR Estimate [Lower Cl, Upper Cl]
lag:
Lung function among children reporting
use of maintenance CSs
Diurnal variability FEV1
Lag 1:1. 61 -0.5,3.72]
Lag 1:0.99 -5.64, 7.62] PM2
5 + 03
Lag 2: 2.96 -1.74,7.66]
Lag 2: 4.62 -4.31, 13.54] PM25 + 03
Lag 3-5: 1.37 [-1.49,4.22]
Lag 3-5: 2.70 [1.0, 4.40] PM2
Lowest daily value FEV1
Lag 1: -2.23 [-6.99,2.53]
5 + 03
Lag 1:3.36 [-3.92, 10.63] PM2 5 + 03
Lag 2: -0.21 [-4.09,3.68]
Lag 2: 0.88 [-8.69, 10.46] PM25 + 03
Lag 3-5: -0.76 [-5.00, 3.49]
Lag 3-5: -2.78 [-4.87 to -0.70] PM25 +
03
Lung function among children reporting
presence of URI on day of lung function
assessment
Diurnal variability FEV-
I an 1 • 4 OR -1 7R Q Q41
Lag 1:3^99 -276| 10J4] PM25 + 03
Lag 2: 7.62 -0.49, 15.73]
Lag 2: 4. 10 -1.41, 9.60] PM2
Lag 3-5: 1.47 [-7.73, 10.67]
5 + 03
Lag 3-5: 3.81 [-1.83, 9.45] PM25 + 03
Lowest daily value FEV1
Lag 1: -1.21 5.62,3.21]
Lag 1: -0.74 -4.14, 2.65] PM25 + 03
Lag 2: -0.10 4.36,4.16]
Lag 2: -1.67 -5.09, 1.75] PM25 + 03
Lag 3-5: -2.88 -5.46 to -0.30
Lag 3-5: -2.78 -4.79 to -0.77
03
PM2.5 +
December 2009
E-201
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Liu et al. (2009,1920031
Period of Study: 4 wk in 2005
Location: Windsor, Ontario, Canada
Outcome: Decreased lung function
Study Design: Panel
Statistical Analysis: mixed-effects
regression models
Statistical Package: S-PLUS
Age Groups: Asthmatic children,
9-14 yr
Pollutant: PM25
Averaging Time: 1, 2 & 3 days
Mean (SD) Unit (1d): 6.5 pg/m3
Range (Min, Max): 2.0-19.0
Copollutant (correlation):
S02: 0.56
N02: 0.71
03: -0.41
Increment: 5.4 pg/m
Percent Change (Min Cl, Max Cl)
Lag
FEV,
Same Day:-0.5 (-1.3-0.3)
Lag 1 Day:-0.5 (-1.1-0.5)
2-Day Avg:-0.6 (-1.5-0.4)
3-Day Avg:-1.1 (-3.1-0.9)
FEF25%-75%
SameDay:-1.9(-3.5--0.3)
Lag1 Day:-1.2 (-2.8-0.3)
2-Day Avg:-2.0 (-3.8--0.2)
3-Day Avg: -3.3 (-7.2-0.8)
FeNO
Same Day: 5.3 (-3.6-15)
Lag1 Day: 1.7 (-6.3-15)
2-Day Avg: 4.3 (-5.4-15.1)
3-Day Avg:-17.3 (-33.5-2.9)
TEARS
Same Day: 16.9 (2.2-33.6)
Lag 1 Day: 14.6 (0.8-30.4)
2-Day Avg: 22.0 (4.8-42.1)
3-Day Avg: 69.1 (20.1-138.2)
8-lsoprostane
Same Day: 5.1 (-3.6-14.5)
Lag1 Day:-3.8 (-12.1-5.3)
2-Day Avg: 0.1 (-9.8-11.1)
3-Day Avg: 5.8 (-15.8-33.0)/
December 2009
E-202
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Mar et al. (2004, 0573091
Period of Study: 1997-1999
Location: Spokane, Washington
Outcome: Respiratory Symptoms
Age Groups: Adults: Ages 20-51 yr
Children: Ages 7-12 yr
N: 25 people
Statistical Analyses: Logistic
regression
Covariates: Temperature, relative
humidity, day of-the-wk
Statistical Package: STATA 6
Lags Considered: 0-2 days
Pollutant: PM25
Mean (SD):
1997:11.0(5.9)
1998:10.3(5.4)
1999:8.1(3.8)
Unit (i.e. pg/m3):
Monitoring Stations: 1 station
Copollutant (correlation):
PM25
PM,r = 0.92
PM10r = 0.61
PM Increment: 10 pg/m
OR Estimate [Lower Cl, Upper Cl]
lag:
Adult Respiratory symptoms:
Wheeze:
1.04[0.86, 1.26] lag 0
1.00(0.83,1.19] lag 1
0.99(0.84, 1.17] lag 2
Breath:
0.97[0.87, 1.08] lag 0
0.980.87, 1.10] lag 1
0.950.80, 1.13] lag 2
Cough:
0.86[0.62, 1.21] lag 0
0.87[0.63, 1.20] lag 1
0.89[0.66, 1.20] lag 2
Sputum:
0.94[0.63, 1.41] lag 0
0.90(0.62, 1.31]lag1
0.92(0.66, 1.27] lag 2
Runny Nose:
0.98[0.83, 1.15] lag 0
0.950.82, 1.10] lag 1
0.930.80, 1.08] lag 2
Eye Irritation:
0.91
0.70,1.20] lag 0
0.890.70, 1.13] lag 1
0.86(0.68, 1.08] lag 2
Lower Symptoms:
0.91[0.73, 1.13] lag 0
0.89(0.72, 1.10] lag 1
0.89(0.72, 1.10] lag 2
Any Symptoms:
0.92[0.80, 1.07] lag 0
0.890.76, 1.04] lag 1
0.890.75,1.05]
lag 2
Children Respiratory symptoms:
Wheeze:
0.55[0.26, 1.19] lag 0
0.530.18, 1.58] lag 1
0.550.19, 1.64] lag 2
Breath:
1.130.86, 1.48] lag 0
1.120.86, 1.44] lag 1
1.10(0.82, 1.48] lag 2
Cough:
1.17[0.98, 1.40] lag 0
1.21(1.00,1.47] lag 1
1.18(0.99, 1.42] lag 2
Sputum:
1.06[0.92, 1.22] lag 0
1.100.91,1.34] lag 1
1.090.92, 1.30] lag 2
Runny Nose:
1.09[0.85, 1.39] lag 0
1.12(0.89,1.41]lag1
1.16(0.94, 1.42] lag 2
Eye Irritation:
0.93(0.53,1.64] lag 0
0.75(0.45, 1.27] lag 1
0.77(0.65, 0.91] lag 2
Lower Symptoms:
1.18(1.00,1.38] lag 0
1.21
1.17
Any
1.17
1.22
1.23
1.00, 1.46
0.96, 1.43
Symptom
1.03, 1.34
1.04, 1.43
1.07, 1.42
Iag1
lag 2
s:
lagO
Iag1
lag 2
December 2009
E-203
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Mar et al. (2005, 0875661
Period of Study: 1999-2001
Location: Seattle, Washington
Outcome: Pulmonary function (arterial Pollutant: PM2 5
oxygen saturation) and cardiac function
(heart rate and blood pressure) Averaging Time: 24-h avg
Study Design: Time series
Statistical Analyses: Linear logistic
regression
Age Groups: > 57
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl)
Lag
Personal:
Systolic: 0.37 (-0.93, 1.67)0
Diastolic: -0.20 (-0.85, 0.46) 0
Indoor:
Systolic: 0.92 (-2.04, 3.87) 0
Diastolic: 0.38 (-1.43, 2.20)0
Outdoor:
Systolic: -0.81 (-2.34, 0.73) 0
Diastolic:-0.46 (-1.49, 0.57)
0
% Increase between heart rate and
PM2 5 exposure for people > 57
PM25:
Personal: 0.44 (0.04, 0.84) 0
Indoor: 0.22 (-0.71,1.16)0
Outdoor:-0.75 (-1.42 to-0.07)0
Reference: Mar et al. (2005, 0887591
Period of Study: 1999-2002
Location: Seattle, Washington
Outcome: Respiratory Symptoms
Age Groups: 6-13 yr
Study Design: Time-Series
N: 19 children
Statistical Analyses: Polynomial
distributed lag model, Poisson
regression
Covariates: Age, ambient NO levels,
temperature, relative humidity,
modification of use of inhaled
corticosteroids
Season: Winter, Spring
Dose-response Investigated? No
Statistical Package: STATA
Lags Considered: 0-8 h
Pollutant: PM25
Averaging Time: 24 h
Mean (SD):
Results presented in Fig 1.
Monitoring Stations: 3 Stations
PM Increment: 10 pg/m
Change in FE(NO) (exhaled NO
concentration) with air pollution
[Lower Cl, Upper Cl] lag:
Medication use:
No meds: 6.99[3.43,10.55] lag 1-h
Meds:-0.18[-3.33, 2.97] lag 1-h
No meds: 6.30[2.64, 9.97] lag 4-h
Meds: -0.77[-4.58, 3.04] lag 4-h
No meds: 0.46[-1.18,2.11] lag 8-h
Meds: 0.40[-1.94, 2.74] lag 8-h
Reference: McCreanor et al. (2007,
0928411
Period of Study: 2003-2005
Location: London, England
Outcome: Decreased Lung Function
Age Groups: Adults
Study Design: Crossover study
N: 60 adults
Statistical Analyses: Linear regression
Covariates: Temperature, relative
humidity, age, sex, bod-mass index,
and race or ethnic group
Pollutant: PM25
Averaging Time: 1 h
Mean (SD): NR
SOth(Median): Oxford St: 28.3
Hyde Park: 11.9
Range (Min, Max):
Oxford St: (13.9, 76.1)
Hyde Park: (3, 55.9)
% changes in FEV and FVC are
presented in Fig 1-3. Results are not
presented quantitatively in text or
tables. The authors did not find any
significant differences in respiratory
symptoms between the two locations.
Also, there were no significant
differences in sputum eosinophil counts
or eosinophil cationic protein levels.
Reference: Moshammer and
Neuberger (2003, 0419561
Period of Study: 2000-2001
Location: Linz, Austria
Outcome: Lung Function: FVC, FEVi,
MEF25, MEF50, MEF75, PEF, LQ Signal,
PAS Signal
Age Groups: Ages 7-10
Study Design: Case-crossover
N: 161 children
1898-2120 "half-h means"
Statistical Analyses: Correlations
Regression Analysis
Covariates: Morning, evening, night
Season: Spring, Summer, Winter, Fall
Dose-response Investigated? No
Pollutant: PM25
Averaging Time: 8 h means & daily
means
Mean (SD): 14.61 (10.83)
Range (Min, Max):
(NR, 119.92)
Monitoring Stations: 1
Copollutant (correlation):
LQ = 0.751
PAS = 0.354
Notes: "Acute effects of 'active particle
surface' as measured by diffusion
charging were found on pulmonary
function (FVC, FEVi,MEF50) of
elementary school children and on
asthma-like symptoms of children who
had been classified as sensitive."
December 2009
E-204
-------
Study
Reference: Moshammer et al. (2006,
0907711
Period of Study: 2000-2001
Location: Linz, Austria
Reference: Murata et al. (2007,
A nr\A rr\\
Design & Methods
Outcome: Respiratory symptoms and
decreased lung function
Age Groups: Children ages 7-10
Study Design: Time-series
N: 163 children
Statistical Analyses: Generalized
estimating equations model
Covariates: Sex, age, height, weight
Dose-response Investigated? NR
Statistical Package: NR
Lags Considered: 1
Outcome: Exhaled nitric oxide levels,
Concentrations1
Pollutant: PM25
Averaging Time: 8 h
Mean (SD):
Maximum 24 h: 76.39
Annual avg: 19.06
Percentiles: 8-h mean 25th: 8.64
8-h mean SOth(Median): 15.70
8-h mean 75th: 25.82
Monitoring Stations: 1 station
Copollutant (correlation):
PM, r = 0.95
PMio r= 0.93
N02r = 0.54
Pollutant: PM25
Effect Estimates (95% Cl)
PM Increment: 10 pg/m3
% change in Lung Function per 10
ug/m3
PP\/. n 0*3
rtv. U.ZO
FVC: 0.08
FEV05:0.33
MEF75%: -0.49
MEF50%: -0.58
MEF25%: -0.83
PEF: 0.41
% change in Lung Function per IQR
FEV: -0.59
FVC: -0.2
FEV0.5: 0.85
MEFy5%' -1 25
MEF50%:-148
MEF25%:-2.14
PEF: -1.06
Multiple pollutant model
FEV: 0.10
FVC: 0.21
FEV05:0.06
MEF75%:-0.15
MEF50%: 0.04
MEF25%: -0.21
PEF: -0.1 8
% change in Lung Function per IQR
FEV: 0.27
FVC: 0.54
FEV05:0.15
MEF75%: -0.39
MEF50%:0.11
MEF25%: 0.54
PEF: 0.015: -0.47
PM Increment: IQR 110 pg/m3
Period of Study: Nov 2004
Location: Tokyo, Japan
(eNO), a marker of airway inflammation
Age Groups: 5-1 Oyr
Study Design: Cohort/Panel study
N: 19 schoolchildren*
Statistical Analyses: Linear regression
Covariates: None
Season: Nov (fall)
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: Lag h 1-24, 8-h ma,
7-h ma, 6-h ma, 24-h ma
Averaging Time: Hourly, 24 h
Mean (SD):
39.0 (16.9) pg/m3 (daily mean)
Range (Min, Max):
10,120 (range of hourly values)
Monitoring Stations: 1, on the street
where the children lived
Mean [Lower Cl, Upper Cl] lag:
0.145 [0.62, 0.228] ppb eNO
8-h ma
Notes: Associations for lag h 1-24
presented in figures. Authors state
"Individual hourly lag models showed a
consistent association between the
eNO value and PM2 5 for exposure in
the previous 24 h"
"The trend on the graphs strongly
suggest that fluctuations in eNO were
affected by changes in air pollutants
over at least the previous 8-h period"
PM25, black carbon, and NOX were all
highly correlated (shown in figures), so
effects are difficult to separate
Pollutant concentrations peaked in the
morning and evening h during traffic
peaks
December 2009
E-205
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Neuberger et al. (2004,
0932491
Period of Study: Jun 1999-Jun 2000
Location: Austria (Vienna and a rural
area near Linz)
Outcome: Questionnaire derived
asthma score, and a 1-5 point
respiratory health rating by parent
Age Groups: 7-1 Oyr
Study Design: Cross-sectional survey
N: about 2000 children
Statistical Analyses: mixed models
linear regression-used factor analysis to
develop the "asthma score"
Covariates: Pre-existing respiratory
conditions, temperature, rainy days, #
smokers in household, heavy traffic on
residential street, gas stove or heating,
molds, sex, age of child, allergies of
child, asthma in other family members
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 4-wk avg (preceding
interview)
Pollutant: PM25
Averaging Time: 24 h
Copollutant (correlation):
PMio(r=0.94) in Vienna
PM Increment: 10 pg/m
Change in mean associated unit
increase in PM
(p-value)
Respiratory Health score
Vienna: 0.016 (p>0.05)
lag 4 week avg
Rural area: 0.022 (p < 0.05)
lag 4 week avg
Asthma score
Vienna: 0.006 (p>0.05)
lag 4 week avg
Rural area: 0.004 (p>0.05)
lag 4 week avg
Reference: Neuberger et al. (2004,
0932491
Period of Study: Sep 1999-Mar 2000
Location: Vienna, Austria
Outcome: Ratio measure: Time to peak
tidal expiratory flow divided by total
expiration time (i.e., tidal lung function,
a surrogate for bronchial obstruction)
Age Groups: 3.0-5.9 yr (preschool
children)
Study Design:
Longitudinal prospective cohort
N: 56 children
Statistical Analyses: mixed models
linear regression, with autoregressive
correlation structure
Covariates: Age, sex, respiratory rate,
phase angle, temperature,
kindergarten, parental education,
observer (also in sensitivity analyses:
height, weight, cold/sneeze on same
day, heating with fossil fuels, hair
cotinine, number of tidal slopes used to
measure tidal lung function)
Dose-response Investigated? No
Statistical Package: SAS 8.0
Lags Considered: Lag 0
Pollutant: PM25
Averaging Time: 24 h
PM Component: Total carbon
EC
OC
Copollutant (correlation):
PM10(r=0.94) in Vienna
PM Increment: Interquartile range (NR)
Change in mean associated with an
IQR increase in PM (p-value)
PM2.5 mass: -0.987 (0.091)
lagO
Total carbon:-0.815 (0.041)
lagO
EC:-0.657 (0.126)
lagO
OC: -0.942 (0.025)
lagO
Reference: Neuberger et al. (2004,
0932491
Period of Study: Oct. 2000-May 2001
Location: Linz, Austria
Outcome: Forced oscillatory resistance
(at zero Hz), FVC, FEV,, MEF25, MEF50,
MEF75, PEF
Age Groups: 7-1 Oyr
Study Design: Longitudinal
prospective cohort
N: 164 children
Statistical Analyses: Mixed models
linear regression with autoregressive
correlation structure
Covariates: Sex, time and individual
Season: Oct-May
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: Lag 0-7
Pollutant: PM25
Averaging Time: 24 h
Monitoring Stations: 1
PM Increment: 1 pg/m
Notes: Authors report increased
oscillatory resistance significantly
associated with PM25 (lag 0)
December 2009
E-206
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: O'Connor et al. (2008,
1568181
Period of Study: Aug 1998-Jul 2001
Location: Boston, the Bronx, Chicago,
Dallas, New York, Seattle, Tucson
Outcome: Pulmonary function and
respiratory symptoms
Age Groups: 5-12 yr
Study Design: Inner-City Asthma Study
(ICAS)-Panel/cohort study
N: 861 children
Statistical Analyses: Mixed effects
models
Lags Considered: Lag 0-6, 0-4
Pollutant: PM25
Averaging Time: 24 h
Mean (SD): 14
Range (Min, Max):
5-35 (estimated from Fig)
Copollutant (correlation):
N02 (F0.59)
S02 (F0.37)
CO (F0.44)
03 (F-0.02)
PM Increment: 13.2 pg/rri 90th-10th
percentile
Change in pulmonary function
FEV,: -1.47 (-2.00 to -0.94) lag 0-4
PEFR:-1.10 (-1.65 to-0.56) lag 0-4
PM25+03+N02
FEVi:-0.73(-1.33to-0.12)lagO-4
PEFR: -0.25 (-0.88, 0.38) lag 0-4
Risk of Respiratory Symptoms
lag
Wheeze: 0.98 (0.88,1.09) lag 0-4
Nighttime asthma: 1.11 (0.94,1.30)
lag 0-4
Slow play: 1.01 (0.89,1.15) lag 0-4
Missed school: 1.33 (1.06,1.66) lag 0-4
PM25+03+N02
Wheeze: 0.92 (0.81,1.05) lag 0-4
Nighttime asthma: 1.03 (0.86,1.23)
lag 0-4
Slow play: 0.92 (0.79,1.06) lag 0-4
Missed school: 1.13 (0.87,1.45) lag 0-4
Reference: Peacock et al. (2003,
0420261
Period of Study: Nov 1996-Feb 1997
Location: northern Kent, UK
Outcome: Reduced peak expiratory
flow rate (PEFR)
Age Groups: 7-13 yr of age
Study Design: Time Series
N:179
Statistical Analyses: generalized
estimating equations
Covariates: Day of the week, 24-h
mean outside temperature.
Season: Wnter
Dose-response Investigated? No
Statistical Package: STATA
Lags Considered: Same day, lag 1,
lag 2, 5-day ma
Pollutant: Sulfate (SOi)
Averaging Time: Daily avg
Mean (SD): Urban 2
24 h avg: 1.3 (1.1)
Percentiles:
10th: Urban 2 0.5
90th: Urban 2 2.4
Range (Min, Max):
Urban 2 0.3, 6.7
Monitoring Stations: 3
Sulfate (S042~)
Increment: 1.3 pg/m3
Odds ratio [Lower Cl, Upper Cl]
lag:
1.090 [0.898, 1.322]
5 days
Reference: Peled, et al. (2005,
1560151
Period of Study: 5-6 wk between
Mar-Jun 1999 and Sep-Dec 1999.
Location: Ashdod, Ashkelon and
Sderot, Israel
Outcome: Reduced peak expiratory
flow (PEF)
Age Groups: 7-1 Oyr
Study Design: Nested cohort study
N:285
Statistical Analyses: Time series
analysis
Generalized linear model, generalized
estimating equations, one-way ANOVA,
generalized linear model
Covariates: Seasonal changes,
meteorological conditions and personal
physiological, clinical and
socioeconomic measurements
Season: Spring, Fall
Dose-response Investigated? No
Statistical Package: STATA
Pollutant: PM25
Averaging Time: Daily
Mean:
Ashkelon: 24.0
Sderot: 29.2
Ashdod: 23.9
PM Component: Local industrial
emissions, desert dust, vehicle
emissions and emissions from two
electric power plants
Monitoring Stations: 6
Copollutant: PM,0
PM Increment: 1 pg/m
P coefficient (SE) [95% Cl]
Ashkelon:
PM25 MAX:-0.144 (0.12) [-0.38-0.09]
Ashdod:
PM25 MAX:-2.74 (0.61) [-3.95-1.53]
PM25 MAX TMAX: 0.11 (0.02) [0.06-
0.16]
In Ashdod, PM25 and an interaction
between PM2 5 and temperature were
significantly associated.
December 2009
E-207
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Penttinen et al. (2006,
0879881
Period of Study: Nov 1996-Apr 1997
Location: Helsinki, Finland
Outcome: Decreased lung function and Pollutant: PM25
respiratory symptoms
PM Component: Soil, heavy fuel oil,
Age Groups: Adults, mean age 53 yr sea salt
Study Design: Time Series
N: 78 people
Statistical Analyses: Generalized least
squares autoregressive model
Covariates: Temperature, relative
humidity, day of study, day of study
squared, binary dummy variable for
weekends
Season: Winter, Spring
Dose-response Investigated? NR
Statistical Package: SAS version 6
Lags Considered: 0-3
Averaging Time: 24 h
Percentiles: 26th:
Long range transport: 2.44
Local combustion: 1.75
Soil: 0.14
Heavy fuel oil:-0.13
Sea Salt: 0.22
Unidentifiable:-1.41
All sources: 6.47
60th(Median):
Long range transport: 4.15
Local combustion: 2.41
Soil: 0.64
Heavy fuel oil: 0.10
Sea Salt: 0.27
Unidentifiable: 0.02
All sources: 8.37
76th:
Long range transport: 7.33
Local combustion: 3.05
Soil: 1.46
Heavy fuel oil: 0.52
Sea Salt: 0.42
Unidentifiable: 0.74
All sources: 11.15
Range (Min, Max):
Long range transport: (-0.89, 28.31)
Local combustion: (0.83, 6.51)
Soil: (-1.13, 6.43)
Heavy fuel oil: (-0.67, 4.74)
Sea Salt: (0.09, 0.98)
Unidentifiable: (-4.40, 4.77)
All sources: (4.11, 33.53)
Monitoring Stations: 1 site
PM Increment: 1.3|jg/m
PMu, long range:
PEF Morning:
0.37[-0.59, 1.34] lag 0
-1.04[-1.88 to-0.19] lag 1
-0.82[-1.81,0.16] lag 2
0.22[-0.64,1.08] lag 3
-0.24[-1.12,0.64] 5 day mean.
PEF Afternoon:
0.20[-0.67, 1.06] lag 0
-0.20[-1.24, 0.83] lag 1
-0.30[-1.14,0.53] lag 2
0.45
0.03
-0.57, 1.47] lag 3
-0.79, 0.85] 5 day mean.
PEF Evening:
-0.33[-1.30, 0.64] lag 0
-0.29[-1.13, 0.55] lag 1
-0.41-1.46, 0.64 lag 2
0.39[-0.47,1.24] lag 3
0.07[-0.81,0.95]5daymean
PMu, local combustion:
PEF Morning:
-0.73[-1.69, 0.23
-0.461-1.24, 0.32
lagO
Iag1
-0.43[-1.49, 0.63] lag 2
0.34[-0.47, 1.15] lag 3
-0.25[-1.03,0.53] 5 day mean.
PEF Afternoon:
-0.21 [-1.07, 0.65] lag 0
-0.81[-1.77, 0.16] lag 1
-0.83[-1.74, 0.09] lag 2
0.20[-0.80,1.20] lag 3
-0.87[-1.63 to -0.12] 5 day mean.
PEF Evening:
-0.51[-1.48, 0.45] lag 0
-1.16[-1.93 to-0.39 lag 1
0.23[-1.35, 0.90] lag 2
0.56[-0.21,1.32] lag 3
-1.14[-1.95 to-0.33]5day mean
PMis, soil:
PEF Morning:
0.81(0.05,1.57] lag 0
0.03 [-0.65, 0.71] lag 1
0.50[-0.34, 1.35] lag 2
-0.07[-0.74, 0.61] lag 3
0.39[-0.46,1.23] 5 day mean.
PEF Afternoon:
1.05[0.38, 1.72] lag 0
0.40-0.38, 1.19] lag 1
0.66 [0.03, 1.30] lag 2
-0.36[-1.12,0.41] lag 3
0.55 [-0.21,1.32] 5 day mean.
PEF Evening:
1.080.33, 1.84] lag 0
1.000.31,1.68] lag 1
0.33[-0.56,1.22] lag 2
-0.84 [-1.53 to-0.15]lag3
0.90[0.08,1.73] 5 day mean
PM2.s,oil:
PEF Morning:
-0.22[-1.00,0.56] lag 0
-0.20[-1.24, 0.84] lag 1
0.66[-0.68, 2.00] lag 2
0.57 [-0.18, 1.32] lag 3
0.10[-0.61,0.81]5daymean.
PEF Afternoon:
-0.04[-0.75, 0.67] lag 0
December 2009
E-208
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
0.29[-0.98, 1.55] lag 1
0.08 [-1.13, 1.28] lag 2
0.62[-0.31,1.54] lag 3
0.07 [-0.64, 0.78] 5 day mean.
PEF Evening:
0.57[-0.23, 1.37] lag 0
0.12-0.92,1.15 lag 1
-0.97[-2.39, 0.45] lag 2
0.40[-0.31,1.12] lag 3
0.43[-0.33,1.19] 5 day mean
PM2.s, salt:
PEF Morning:
0.76[-0.13, 1.65] lag 0
0.43 [-0.30, 1.16] lag 1
0.13
0.38
-0.75,1.02] lag 2
-0.47, 1.23] lag 3
0.95[-0.18, 2.09] 5 day mean.
PEF Afternoon:
0.62-0.18,1.41 lagO
0.80-0.08, 1.69 Iag1
0.14-0.62,0.90 lag 2
0.16-0.83,1.15 Iag3
0.88 [-0.18,1.94] 5 day mean.
PEF Evening:
1.090.19, 1.98] lag 0
0.63-0.10, 1.35] lag 1
0.32[-0.62,1.26] lag 2
-0.31[-1.16,0.54] lag 3
0.88[-0.27, 2.02] 5 day mean
PMu, unidentified:
PEF Morning:
0.38[-0.67,1.43] lag 0
0.09 " " '
0.22
0.78 [-0.10, 1.66] lag 3
0.78[-0.14,1.69] 5 day mean.
-0.83,1.00] lag 1
-0.82, 1.26] lag 2
PEF Afternoon:
0.02
0.65
0.17
0.69
-0.92, 0.96
-0.48, 1.77
-0.71, 1.05
-0.36, 1.75
lagO
Iag1
lag 2
Iag3
0.17 [-0.72,1.06] 5 day mean.
PEF Evening:
-0.11[-1.17, 0.95] lag 0
0.19[-0.72,1.10] lag 1
-0.25, 1.96] lag 2
-0.70,1.01] lag 3
0.86
0.15
-0.19[-1.15, 0.77] 5'day mean
PMu, local combustion:
PEF morning:
Cu:-0.25 [-1.25, 0.75]
Zn:-0.45[-1.19,0.29]
Mn: 0.13[-0.83, 1.08]
Fe: 0.08[-0.70, 0.85],
PEF afternoon:
Cu:-0.37[-1.29, 0.55]
Zn:-0.19[-0.87, 0.50]
Mn:-0.48[-1.37,0.42]
Fe: 0.29[-0.45, 1.04],
PEF evening:
Cu:-0.48[-1.47, 0.52]
Zn:-0.17[-0.92, 0.57]
Mn: 0.51[-0.44, 1.47]
Fe: 0.34[-0.46, 1.14]
PMu, long range:
PEF morning:
S: 0.11 [-0.886, 1.07]
K:-0.10[-1.00, 0.80]
Pb:-0.62[-1.37, 0.13]
December 2009
E-209
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
Br:-0.40 [-1.40, 0.60].
PEF afternoon:
S:-0.05[-0.92,0.81]
K: 0.26[-0.56, 1.07]
Pb:-0.12[-0.84, 0.60]
Br: 0.15[-0.81,1.12].
PEF evening:
S: 0.08[-0.86, 1.02
K: 0.18[-0.70,1.07;
Pb: -0.20[-0.97, 0.58]
Br: 0.35[-0.71, 1.40]
PMis, soil:
PEF morning:
Si: 0.27[-0.43, 0.97]
Al: 0.17 [-0.72, 1.05]
Ca:0.13[-1.08, 1.35].
PEF afternoon:
Si: 0.39[-0.24,1.01]
Al: 0.49-0.29, 1.27
Ca: 0.15[-0.92, 1.22]
PEF evening:
Si: 0.60-0.06, 1.26
Al: 0.76[-0.08,1.60
Ca: 0.90[-0.22, 2.03]
Plfe, Oil combustion:
PEF morning:
V:-0.01 [-0.87, 0.86]
Ni:-0.09[-1.08,0.90].
PEF afternoon:
V:-0.48[-1.32,0.35
Ni: 0.26[-0.72,1.23.
PEF evening:
V: 0.02[-00.88, 0.92]
Ni: 0.50[-0.55,1.55]
Plfe, Sea salt:
PEF morning:
Na: 0.92[-0.34, 2.17]
Cl: 0.93[0.08, 1.79]
PEF afternoon:
Na: 0.96[-0.24, 2.16]
Cl: 0.57[-0.22, 1.36]
PEF evening
Na: 0.87[-0.40, 2.15]
Cl:0.65[-0.19, 1.49]
December 2009 E-210
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Pino et al. (2004, 0502201
Period of Study: Apr 1995-Oct 1996
Location: Santiago, Chile
Outcome: Respiratory Symptoms,
Wheezing bronchitis
Study Design: Time-series
Statistical Analyses: Bayesian
hierarchical analysis, cubic spline
Age Groups: 4 mo-2 yr old
Pollutant: PM25
Averaging Time: 24-h avg
Mean (SD) unit: 52.0 (31.6)
Range (5th, 95th): 17.0,114.0
Copollutants (correlation):
S02: r= 0.73
NO,: F 0.85
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl) lag:
% increase in wheezing bronchitis and
PM2 5 exposure for infants 4 mo-2 yr old
4.75(1.25,8.25)1
3.85
2.25
1.75 (-2.20, 5.75)4
0.45, 7.75) 2
-1.00,6.00)3
4.00
5.00
0.25, 8.00
1.00,8.50
7.00(3.50,11.00)7
8.10(4.00, 11.25)8
9.00
8.75
6.00, 12.00)9
5.75, 12.00) 10
1.50 (-3.50, 4.75)11
0.25
0.00
-3.75, 4.25) 12
-4.00, 4.00) 13
1.00 (-3.50, 4.50)14
1.50 (-3.50, 4.50)15
OR for wheezing bronchitis and PM2 5
exposure in infants 4 mo to 2 yr old
according to family history of asthma
Yes to family history of asthma
1.09(1.00,1.19)1
1.10(1.02, 1.20)2
1.11 (1.02, 1.22)3
No to family history of asthma
1.04
1.02
1.00, 1.08
0.98, 1.06
1.01(0.96,1.05)3
Reference: Rabinovitch et al., (2006,
0880311
Period of Study: 2001-2003 (two
winters 2001-2002 and 2002-2003)
Location: Denver, CO
Outcome: Bronchodilator doser
activations (daily) and urinary
leukotriene E4 (daily)
Age Groups: Children 6-1 Syr old
Study Design: School-based cohort
study
N: 73 children
Statistical Analyses: Doser activation:
Poisson regression with GEE with AR1
working covariance
Urinary leukotriene E4: linear mixed
model with spatial exponential
covariance
Covariates: Temperature, pressure,
humidity, time trend, Friday indicator,
upper respiratory infection (URI), height
(leukotriene E4 only).
Season: Winter
Dose-response Investigated? NR
Statistical Package: SAS
Lags Considered: 0-2 days
Pollutant: PM25
Averaging Time:
Morning (midnight to 11: 00AM) mean
Morning (midnight to 11: 00AM)
maximum
24-h mean
Mean (SD):
24-h mean, TEOM
Year 1,N: 55 days: 6.5 (3.2)
Year 2, N: 128 days: 8.2 (3.7)
24-h mean, FRM
Year 1,N: 55 days: 11.8 (7.2)
Year 2, N: 122 days: 11.2 (5.5)
Morning mean, TEOM
Year 1,N: 71 days: 7.4 (4.7)
Year 2, N: 127 days: 9.1 (5.0)
Morning maximum, TEOM
Year 1,N: 71 days: 15.5 (9.5)
Year 2, N: 127 days: 18.4 (9.6)
Percentiles:
24-h mean, TEOM
Yearl
25th: 4.4
SOth(Median): 6.2
75th: 7.9
Year 2
25th: 55
SOth(Median): 7.3
75th: 9.9
24-h mean, FRM
Yearl
25th: 7.8
50th(Median):10.1
75th: 14.1
Year 2
25th: 7.5
SOth(Median): 9.3
75th: 13.3
Morning mean, TEOM
Yearl
25th: 4.0
PM Increment: IQR (over current and
previous day)
Doser Activation
Morning avg PM2s TEOM
Yearl:
Pet Increase: 3.0 [-0.5: 6.6] p = 0.10
Year 2:
Pet Increase: 2.7 [1.1: 4.4] p = 0.006
Aggregated yr: 2.2 [0.7: 3.6] p = 0.005
Morning max PM2s TEOM
Yearl
Pet Increase: 4.0 [0.5: 7.6] p = 0.02
Year 2
Pet Increase: 2.3 [0.7: 4.0] p = 0.009
Aggregated yr 2.6 [0.9: 4.2] p= 0.002
24-h PM2S
TEOM
Lag 0:0.4 [-0.7:1.6] p-value = 0.45
Lag 1:0.9 [-0.7: 2.4] p-value = 0.27
Lag2:-0.4[-1.7:0.9]p-value = 0.59
Lag 0-2 Avg: 0.6 [-1.0: 2.2]
p-value = 0.43
FRM
Lag 0:0.2 [-1.2:1.6] p-value = 0.81
Lag 1:0.9 [-0.9: 2.6] p-value = 0.31
Lag 2:-0.2 [-2.2:1.8] p-value = 0.88
Lag 0-2 Avg: 1.2 [-0.6: 2.9]
p-value = 0.20
Morning avg PMu TEOM
URI not adjusted
Mild/Moderate Asthmatics:
1.5 [-0.5: 3.4] p = 0.14
Severe Asthmatics: 3.7 [1.6: 5.8]
p- = 0.0006
Difference between severity groups,
p = 0.12
Aggregated severity group:
2.2 [0.7: 3.6] p= 0.005
URI adjusted
Mild/Moderate Asthmatics:
1.0 [-1.9: 3.9]p= 0.50
Severe Asthmatics: 6.0 [1.8:10.1]
p = 0.006
Difference between severity groups,
December 2009
E-211
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
SOth(Median): 5.9
75th: 9.6
Year 2 25th: 5.2
50th (Median): 8.5
75th: 11.6
Morning maximum, TEOM
Year 125th: 8
50th (Median): 13
75th: 20
Year 2 25th: 11
50th (Median): 16
75th: 23
Range (Min, Max):
24-h mean, TEOM
YeaM
Year 2
2.1,23.7
1.7, 20.5
24-h mean, FRM
Year!
Year 2
4.3, 53.5
3.4, 26.3
Morning mean, TEOM
Year!
Year 2
1.4, 22.7
1.6, 30.2
Morning maximum, TEOM
Year!
Year 2
4,42
4,46
Monitoring Stations: 2 (1 TEOM and 1
Federal Reference Monitor [FRM])
p = 0.08
Aggregated severity groups:
2.7[-0.1:5.4]p=0.06
Morning maximum PIfcTEOM
URI not adjusted
Mild/Moderate Asthmatics:
1.9 [-0.2: 4.1] p= 0.07
Severe Asthmatics: 3.9 [1.1: 6.8]
p = 0.006
Difference between severity groups,
p = 0.29
Aggregated severity groups: 2.6 [0.9:
4.2] p= 0.002
URI adjusted
Mild/Moderate Asthmatics:
1.6 [-2.2: 5.4] p = 0.41
Severe Asthmatics: 8.1 [2.9:13.4]
p = 0.003
Difference between severity groups,
p = 0.03
Aggregated severity groups:
3.8 [0.2: 7.4] p = 0.04
Leukotriene E4
24-h PM2.j
TEOM
Lag 0:3.3 [-0.7: 7.2] p = 0.09
Lag 1: -1.6[-5.7: 2.5] p = 0.40
Lag 2:1.1 [-2.8: 5.1] p= 0.64
Lag 0-2 Avg: 2.3 [-4.0: 8.6] p = 0.45
Lag 0: 2.7 [1.1: 6.5] p = 0.12
Lag 1:-0.8
Lag 2: -0.8
-4.9: 3.3] p = 0.65
-4.9: 3.3] p = 0.71
Lag 0-2 Avg: 2.6 [-2.3: 7.5] p = 0.27
Leukotriene E4
Morning avg PMu TEOM
Height 25percentile: 8.9 [3.0:14.7] p=
0.004
Height SOpercentile: 5.9 [1.4:10.4] p =
0.01
Height 75percentile: 1.9 [-3.4: 7.3] p =
0.47
Model w/o Height x Pollutant: 5.6 [1.0:
10.2] p = 0.02
Morning maximum PIfcTEOM
Height 25percentile: 8.3 [3.4:13.2] p =
0.001
Height SOpercentile: 6.1 [2.1:10.2] p=
0.004
Height 75 percentile: 3.2 [-2.0: 8.4] p=
0.23
Model w/o Height x
Pollutant: 6.2 [1.9:10.5] p = 0.006
Reference: Rabinovitch et al. (2004,
0967531
Periods of Study:
Nov1999-Mar2000
Nov2000-Mar2001
Nov 2001-Mar 2002
Location: Denver, Colorado
Outcome: Respiratory symptoms,
Asthma symptoms (cough and
wheeze), Upper respiratory symptoms
Study Design: Time-series
Statistical Analyses: Logistic linear
regression, PROC Mixed, PROC
Genmod
Age Groups: 6-12
Pollutant: PM25
Averaging Time: 24-h avg
Mean (SD): 10.8 (7.1)
Range (Min, Max): (1.8, 53.5)
Copollutant (correlation):
CO
N02
S02
PM Increment: 1 pg/m
p(SE)
AM:-0.003 (0.009)
PM: 0.004 (0.011)
Odds Ratio (Lower Cl, Upper Cl)
Lag
0.971 (0.843, 1.118)
0-3 avg.
December 2009
E-212
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ranzi et al. (2004, 0895001
Period of Study: Feb-May 1999
Location: Emilia-Romagmna, Italy
(urban-industrial and rural area)
Outcome: respiratory symptoms, PEF
measurements, drug consumption and
daily activity
Age Groups: Children, mean age=(7.2-
7.9 yr)
Study Design: Panel study
N: 120 children
Statistical Analyses: Ecological
analysis and Panel analysis
Pollutant: PM2 5
Averaging Time: 24 h
Mean (SD):
Urban= 53.07
Rural=29.11
Monitoring Stations: 3
Copollutant (correlation):
TSP: r=0.613
PM Increment: 10 pg/m3
Effect Estimate:
Urban-industrial panel
Cough and Phlegm: RR=1.0044
(1.0011-1.0077)
gender, medicinal use, symptomatic
status of previous day
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0,1, 2, 3, 0-3 ma
Daily air pollution concentrations: r=
0.658
Reference: Rodriguez et al. (2007,
0928421
Period of Study: 1996-2003
Location: Perth, Australia
Outcome: Body temperature, cough,
runny/ blocked nose, wheeze/ rattle
chest (daily)
Age Groups: Children 0-5 yr old
Pollutant: PM25
Averaging Time: 1 h and 24 h
Mean (SD): 1-h averaging, 20.767
PM Increment: NR
[Lower Cl, Upper Cl]
lag: NR
LAG :0 day
Study Design: hospital-based cohort
study
N: 198-263 children
Statistical Analyses: Logistic
regression with GEE and AR (order not
specified) working covariance
Covariates: temperature, humidity
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-5 days
24-h averaging, 8.534
Range (Min, Max): 1-h averaging
(0.012:93.433)
24-h averaging
(0.004: 39.404)
Monitoring Stations: 10 total, usually
3-5 sites for each pollutant
Copollutant (correlation):
03
NO*
CO
PM2S, 1-h
Body temperature: 1.004 [0.998:1.011]
Cough: 1.006 [1.000:1.012]
Wheeze/rattle chest:
1.004 [0.998:1.010]
Runny/blocked nose:
0.997 [0.983:1.010]
PM2.j, 24-h
Body temperature: 1.005 [0.986:1.024]
Cough: 1.019 [0.999:1.040]
Wheeze/rattle chest:
0.990 [0.969:1.012]
Runny/blocked nose:
0.968 [0.926:1.013]
LAG: 5 days
PM2.j, 1-h
Body temperature: 1.005 [0.999:1.040]
Cough: 1.003 [0.995:1.010]
Wheeze/rattle chest: 1.005 [0.998:
10.12]
Runny/blocked nose: 1.015 [1.000:
1.030]
PM2.s, 24-h
Body temperature: 1.020 [0.998:1.011]
Cough: 1.006 [0.984:1.011]
Wheeze/rattle chest:
1.018 [0.997:1.040]
Runny/blocked nose:
1.039 [0.990:1.089]
LAG: 0-5 days
PM2.s, 1-h
Body temperature: 1.000 [0.998:1.002]
Cough: 1.001 [0.999:1.003]
Wheeze/rattle chest:
1.002 [1.000:1.004]
Runny/blocked nose:
1.01 [0.997:1.006]
1.02
PM2.j, 24-h
Body temperature: 1.000 [0.994:1.005]
Cough: 1.004 [0.997:1.011]
Wheeze/rattle chest:
1.001 [0.995:1.007]
Runny/blocked nose:
0.998 [0.985:1.011]
December 2009
E-213
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Sakai et al. (2004, 0874351
Period of Study:
Nov1999-Mar2001
Location: Diesel-powered ship from
Tokyo, Japan to Showa Station on Ongul
Island, Antarctica for 366 days (from Feb
1, 2000) and then heading back to
Japan on Feb1, 2001
Outcome: circulating leukocyte counts
and serum inflammatory cytokine levels
Age Groups: 24-57 yr, mean=36.1 + 4.7
yr
Study Design: Cohort
N: 39 members of 41st Japanese
Antarctic Research Expedition
(JARE-41)
Statistical Analyses: ANOVA
Covariates: Smoking history,
occupational pollutant exposure
Dose-response Investigated? No
Statistical Package: SPSS 11.5J
Pollutant: PM5.0.2.o
Averaging Time: 24 h
Unit (i.e. ug/m3): particles/L
PM Component: organic and inorganic
substances, including microorganisms
Copollutant (correlation):
PM2.0-0.3
PMlO-5.0
Effect Estimate:
Multiple regression analysis between
inhaled factors in Antarctica
Total leukocyte
Cigarette smoking= 0.211, p< 0.001
Support staff= 0.139, p=0.024
Total PM= 0.168, p=0.004
Segmented PMN
Cigarette smoking= 0.015, p=0.805
Support staff= 0.097, p=0.119
Total PM= 0.272, p < 0.001
Band-formed PMN
Cigarette smoking= 0.035, p=0.543
Support staff= 0.010, p=0.864
Total PM= 0.470, p < 0.001
Monocyte
Cigarette smoking= 0.081, p=0.187
Support staff=-0.019, p=0.759
Total PM= 0.328, p < 0.001
G-CSF
Cigarette smoking= 0.131, p < 0.038
Support staff= 0.176, p=0.005
Total P M=0.078, p=0.186
IL-6
Cigarette smoking= 0.182, p=0.004
Support staff= 0.076, p=0.228
Total PM= 0.158, p=0.008
Reference: Sakai et al. (2004, 0874351
Period of Study: Nov 1999-Mar 28,
2001
Location: Diesel-powered ship from
Tokyo, Japan to Showa Station on Ongul
Island, Antarctica for 366 days (from Feb
1, 2000) and then heading back to
Japan on Feb1, 2001
Outcome: circulating leukocyte counts
and serum inflammatory cytokine levels
Age Groups: 24-57 yr, mean=36.1 + 4.7
yr
Study Design: cohort
N: 39 members of 41st Japanese
Antarctic Research Expedition (JARE-
41)
Statistical Analyses: ANOVA
Covariates: Smoking history,
occupational pollutant exposure
Dose-response Investigated? No
Statistical Package: SPSS 11.5J
Pollutant: PMi0.5.o
Averaging Time: 24 h
Unit (i.e. ug/m3): particles/L
Monitoring Stations: NR
Copollutant (correlation):
PM2.0-0.3
PMlO-5.0
Effect Estimate:
Multiple regression analysis between
inhaled factors in Antarctica
Total leukocyte
Cigarette smoking= 0.211, p < 0.001
Support staff= 0.139, p=0.024
Total PM= 0.168, p=0.004
Segmented PMN
Cigarette smoking= 0.015, p=0.805
Support staff= 0.097, p=0.119
Total PM= 0.272, p < 0.001
Band-formed PMN
Cigarette smoking= 0.035, p=0.543
Support staff= 0.010, p=0.864
Total PM= 0.470, p < 0.001
Monocyte
Cigarette smoking= 0.081, p=0.187
Support staff= -0.019, p=0.759
Total PM= 0.328, p < 0.001
G-CSF
Cigarette smoking= 0.131, p < 0.038
Support staff= 0.176, p=0.005
Total P M=0.078, p=0.186
IL-6
Cigarette smoking= 0.182, p=0.004
Support staff= 0.076, p=0.228
Total PM= 0.158, p=0.008
December 2009
E-214
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Silkoff et al. (2005, 0874711 Outcome: Lung function: FEV,, PEF
Period of Study: Winter 1999-2000,
Winter 2000-2001
Location: Denver, CO
Age Groups: Adults (>40 yr-old) with
COPD, as well as >10 pack-yr tobacco
use, FEV, < 70%, FEV,/FVC < 60%,
and no other lung disease
Study Design: COPD patient panel
study (2 independent panels
One for each winter)
N: 34 subjects (16 1st winter, 18
second winter)
Statistical Analyses: Mixed effects
models with first-order, autoregressive,
ma variance-covariance
Binary outcomes (rescue medication
use, total symptom score) assessed
using Poisson regression with GEE and
first-order, auto-regressive variance-
covariance
Covariates: Temperature, relative
humidity, barometric pressure analysis
run separately for each winter
Season: Wnter
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-2 days
Pollutant: PM25
Averaging Time: 24 h
Mean (SD):
Wnter 1999-2000: 9.0 (5.2)
Wnter 2000-2001:14.3 (9.6)
Percentiles:
Wnter 1999-2000
25th 5.4
SOth(Median): 7.7
75th: 11.3
Wnter 2000-2001
25th 7.6
SOth(Median): 11.7
75th: 17.2
Range (Min, Max):
Wnter 1999-2000
(1.8,36.6)
Wnter 2000-2001
(3.4, 59.6)
Monitoring Stations: multiple sites
Copollutant (correlation):
CO
N02
PM,0
PM Increment: SD
Wnter 1999-2000: 5.2
Wnter 2000-2001: 9.6
Model results reported graphically only.
No quantitative results reported.
Direction of slope (±) and statistical
significance (SIG: yes; NS: no) inferred
from graphs.
Among subjects with severe COPD
observed in Wnter 1999-2000,
statistically significant, but marginal,
improvements in PEF associated with
morning lag 0 PM25.
There were no statistically significant
associations between rescue
medication use and symptom score
with PM.
Reference: Sivacoumar et al. (2006,
1111151
Period of Study: Apr 1998-May 1998;
Sep 1998-Oct 1998
Location: Pammal, India
Outcome: Respiratory symptoms,
Decreased pulmonary function
Study Design: Case-control
Statistical Analyses: Poisson
Age Groups: >18
Pollutant: PM25
Averaging Time: 24-h avg
The study does not present quantitative
results of association.
Reference: Slaughter et al. (2003,
0862941
Period of Study: 1994
Location: Seattle, WA
Outcome: Asthma attacks, asthma
severity, medication use
Age Groups: 5.1-13.1 yrold
Study Design: Cross-sectional study
N: 133 children
Statistical Analyses: Ordinal Logistic
Regression
Poisson Modeling
Covariates: Temperature, Day of the
Week, Seasonality
Dose-response Investigated? No
Statistical Package: STATA
Lags Considered: 1-, 2-, 3-day lag
Pollutant: PM25
Averaging Time:
Daily Avg
25th: 5.0
50th(Median):7.33
75th: 11.3
Monitoring Stations: 3
Copollutant (correlation):
PM10 = 0.75
CO = 0.82
PM Increment: 10 pg/m increase
RR Estimate [Lower Cl, Upper Cl]
lag:
Inhaler use:
1-day lag: 1.04 (0.98,1.10)
OR Estimate [Lower Cl, Upper Cl]
lag:
Asthma Attack:
1-day lag: 1.20 (1.05,1.37)
Previous day: 1.13 (1.03,1.23)
Medication Use
Nontransition model:
Previous Day: 1.08 (1.01,1.15)
Notes: Figures of estimated odds ratios
for having a more serious asthma
attack for short-term, within-subject
increases in PM25, PM10, and CO.
Transition models additionally control
for the previous day's severity.
Figures of estimated relative risks for
having inhaler use for short-term,
within-subject increases in PM25, PM10,
and CO. Transition models additionally
control for the previous day's severity.
December 2009
E-215
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Strand et al. (2006,
0892031
Period of Study: 2002-2004
Location: Denver, Colorado, United
States
Outcome: Reduced forced expiratory
volume (FEVi)
Age Groups: 6-12 yr old
Study Design: Mixed model analysis
(using the default restricted maximum
likelihood (REML) estimators)
N: 50 children
Statistical Analyses: least squares
regression, SAS "Output Delivery
System" (ODS)
Season: Fall and Winter
Dose-response Investigated? Yes
Statistical Package: SAS
Pollutant: PM25
Averaging Time: Daily
Mean (SD):
Outdoor: 12.699 (6.426)
Indoor: 8.148 (4.348)
Sulfate/PM25/outdoor: 0.079 (0.067)
Sulfate/PM25/indoor: 0.074 (0.060)
Range (Min, Max):
Mean Personal: (0, 3.035)
Outdoor: (0, 6.303)
Indoor: (0, 2.759)
PM Component: EC, sulfate, nitrate and
ETS.
Monitoring Stations: 2 fixed monitors
and up to 10 personal monitors on a
given day.
Copollutant (correlation): Sulfate
(0.63)
PM Increment: 10 pg/m
Effects Estimate:
Using the estimated slope for the
validation study model [Lower Cl, Upper
Cl] lag:
2.2 percent decrease in FEVi per 10
pg/m3 increase in ambient PM25 [0.0,
4.3 decrease] 1 day
Reference: Tang et al. (2007, 0912691
Period of Study: Dec 2003-Feb 2005
Location: Sin-Chung City, Taipei
County, Taiwan
Outcome: Peak expiratory flow rate
(PEFR) of asthmatic children
Age Groups: 6-12 yr
Study Design: Panel study
N: 30 children
Statistical Analyses: Linear mixed-
effect models were used to estimate the
effect of PM exposure on PEFR
Covariates: Gender, age, BMI, history
of respiratory or atopic disease in
family, SHS, acute asthmatic
exacerbation in past 12 mo, ambient
temp and relative humidity, presence of
indoor pollutants, and presence of
outdoor pollutants,
Dose-response Investigated? yes
Statistical Package: S-Plus 2000
Lags Considered: 0-2
Pollutant: PM25
Averaging Time: 1 h
Mean (SD):
Personal: 27.8 (25.3)
Range (Min, Max):
Personal: 1.4-263.4
Monitoring Stations: 1
PM Increment: 24.5 pg/m
RR Estimate [Lower Cl, Upper Cl]
lag:
Change in morning PEFR:
-6.00 (-29.85, 17.85
-12.52 (-77.93, 52.9
lagO
Iag1
-24.87 (-71.49, 21.74) lag 2
-45.67 (-117.09, 25.74) 2-day mean
-5.69 (-105.96, 94.59) 3-day mean
Change in evening PEFR:
0.50 (-18.82, 19.82) lag 0
16.66 (-7.59, 40.9) lag 1
11.60 (-11.1,34.31) lag 2
39.97 (7.1,72.85) 2-day mean
-3.32 (-66.14, 59.5) 3-day mean
Reference: Timonen et al. (2004,
0879151
Period of Study: Oct 1998-Apr 1999
Location Amsterdam, Netherlands
Erfurt, Germany
Helsinki, Finland
Outcome: Urinary concentration of Pollutant: PM25
Clara cell protein CC1 6 of subjects with
coronary heart disease Averaging Time: 24 h
Mean (SD)1
Age Groups: 50+ Amsterdam: 20.03pg/m3
Study Design: Longitudinal cohort ur7lrt:l23;,l'i9/m/ 3
study (panel) Helslnkl: 127 «/m
N: 37 (Amsterdam) Range (Min, Max):
Amsterdam: 3.8-82.2
47 (Erfurt) Erfurt: 4.5-118.1
Helsinki: 3. 1-39.8
PM Increment: 10 pg/m3
RR Estimate [Lower Cl, Upper Cl]
lag:
Pooled estimate;
2.8 (-1.1-6.7) lag 0
2.9 (-0.6-6.5) lag 1
5.0 (-2.4-12.4) lag 2
1.6 (-4.7-7.9) lag 3
9.7 (-6.0-25.4) 5-day mean
CC16 was not associated to PM25
47 (Helsinki)
Statistical Analyses: The response of
interest was log transformed, creatinine
adjusted CC16. Mixed-effect model was
used to investigate the association
between CC16 and air pollutants.
Covariates: Subjects, long term time
trend, temperature (lags 0-3), relative
humidity (lags 0-3), barometric pressure
(lags 0-3), and weekday of visit.
Dose-response Investigated? yes
Statistical Package: S-Plus and SAS
Lags Considered: 0-3
Monitoring Stations: 3
Copollutant (correlation):
Spearman Correlation:
NC 0.01 -0.1: Amsterdam -0.15
Erfurt 0.62
Helsinki 0.14
NCO.1-1.0: Amsterdam 0.80
Erfurt 0.84
Helsinki 0.80
N02: Amsterdam 0.49
Erfurt 0.82
Helsinki 0.35
CO: Amsterdam 0.58
Erfurt 0.77
Helsinki 0.40
in the pooled analysis but CC16 was
significantly associated to PM25
in Helsinki:
23.3 (6.3-40.3) lag 0
6.4 (-8.2-21.1) lag 1
20.2 (6.9-33.5) lag 2
17.6 (4.3-30.9) lag 3
38.8 (15.8-61.8) 5-day mean
December 2009
E-216
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Trenga et al. (2006,
1552091
Period of Study: 1999-2002
Location: Seattle, WA
Outcome: Lung function: FEVi, PEF,
MMEF (maximal midexpiratory flow
assessed only for children)
Age Groups: Adults (56-89-yr-old)
healthy & with COPD
Asthmatic children 6-13-yr-old
Study Design: Adult and pediatric
panel study over 3 yr with 1 monitoring
period ("session") per yr
N: 57 adults (33 healthy, 24 with COPD)
= 692 subject-days = 207 study-days
17 asthmatic children = 319 subject-
days = 98 study-days
Statistical Analyses: Mixed effects,
longitudinal regression models, with the
effects of pollutant decomposed into
each subject's
a) overall mean
b) Difference between their session-
specific mean and overall mean
c) Difference between their daily values
and session-specific mean
Covariates: Gender, age, ventral site
temperature and relative humidity, CO,
N02
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-1 days
Pollutant: PM25
Averaging Time: 24 h
Percentiles:
Children
Personal
25th: 8.1
SOth(Median): 11.3
75th: 16.3
Indoor
25th: 5.7
SOth(Median): 7.5
75th: 10.2
Local outdoor
25th: 6.4
SOth(Median): 9.6
75th: 14.
Adults
Personal
25th: 5.9
SOth(Median): 8.5
75th: 12.4
Indoor
25th: 5.1
SOth(Median): 7.6
75th: 10.8
Local outdoor
25th: 6
SOth(Median): 8.6
75th: 13.1
Range (Min, Max):
Children, Personal 1.0, 49.4
Indoor (2.2, 36.3)
Local outdoor (2.8, 40.4)
Adults, Personal 1.3, 66.6
lndoor(1.6, 65.3)
Local outdoor (0.0, 41.5)
Monitoring Stations: 2
also subject-specific local outdoors (i.e.
at each home), indoor, and personal
Copollutant (correlation):
CO
N02
PM25
PMio-2.s (coarse)
PM Increment: 10 pg/m
ADULT
Personal PM2 5-FEV,
Overall: Lag 0-6.0 [-29.1:17.2]
Lag1 12.0 [-12.9: 36.9]
No-COPD: Lag 0-4.6 [-31.0: 21.9]
Lag 1 19.3 [-8.2: 46.7]
COPD: Lag 0-10.2 [-55.8: 35.4]
Lag 1-19.0 [-74.1: 36.2]
PEF: Lag 01.5 [-2.2: 5.2]
Lag1 2.1 [-1.9:6.1]
No-COPD: Lag 03.4 [-0.9: 7.6]
Lag1 1.9 [-2.5: 6.3]
COPD: Lag 0-4.3 [-11.5: 3.0]
Lag 12.6 [-6.3:11.5]
Indoor PM2 5-FEV,
Overall: Lag 0-12.8 [-44.5:19.0]
Lag1 19.4 [-11.3: 50.1]
No-COPD: Lag 0-15.8 [-50.0:18.4]
Lag1 28.4[-4.6:61.3]
COPD: Lag 02.6 [-71.7: 76.8]
Lag 1-29.7 [-102.9: 43.5]
PEF
Overall: Lag 0-0.5 [-5.6: 4.6]
Lag 1 2.3 [-3.3: 7.8]
No-COPD: Lag 00.1 [-5.4:5.6]
Lag 1 2.5 [-3.5: 8.4]
COPD: Lag 0-3.2 [-15.1: 8.7]
Lag1 1.1 [-12.0:14.3]
Outdoor Home PM2 5-FEV,
Overall: Lag 0-1.4 [-35.6: 32.7]
Lag 1 -2.4 [-37.6: 32.7], No-COPD: Lag
01.5 [-36.1:39.2]
Lag 1 10.7 [-26.9: 48.4]
COPD: Lag 0 -8.9 [-62.2: 44.4]
Lag 1-45.2 [-102.6:12.1]
PEF
Overall: Lag 02.3 [-3.3: 7.9]
Lag 1 0.4 [-5.6: 6.4]
No-COPD: Lag 04.0 [-2.2:10.1]
Lag 1 2.0 [-4.4: 8.4]
COPD: Lag 0-1.8 [-10.6: 6.9]
Lag 1-4.8 [-14.6: 4.9]
Central Sites PM25 -FEV,
Overall: Lag 0-35.5 [-70.0:-1.0]
Lag 1 -40.4 [-71.1: -9.6], No-COPD: Lag
0-32.6 [-69.5: 4.3]
Lag 1 -29.0 [-62.5: 4.5]
COPD: Lag 0 -43.6 [-95.0: 7.8]
Lag1 -70.8 [-118.4: 23.1]
PEF
Overall: Lag 01.5 [-4.2: 7.1]
Lag 1 -2.3 [-7.4: 2.9]
No-COPD: Lag 02.5 [-3.5: 8.6]
Lag 1-0.5 [-6.1:5.0]
COPD: Lag 0-1.5 [-9.9: 6.9]
Lag 1-7.1 [-15.0: 0.9]
PEDIATRIC FEV,
Personal PM25
Overall:
Lag 0-13.08 [-38.26: 12.10]
Lag1 -16.12 [-42.61: 10.37],
No anti-inflammatory medication:
Lag 0-41.73 [-94.31:10.84]
Lag! -30.99 [-82.17: 20.19],
Anti-inflammatory medication:
Lag 0 -4.61 [-34.49: 25.28]
Lag1 -10.87 [-45.01: 23.27]
Indoor PM2 5
Overall:
Lag 0 -45.90
Lag 1 -64.78
-89.92: 1.88]
-111.27:18.2!
No anti-inflammatory medication:
Lag 0 -75.92
Lag 1 -65.08
-145.16:6.6'
-136.98: 6.82],
Anti-inflammatory medication:
Lag 0-28.50 [-94.72: 37.71]
December 2009
E-217
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
Lag1 -64.60-147.23:18.04]
Outdoor Home P f.l; :
Overall:
Lag 0-13.11 [-57.41:31.19]
Lag 1 -9.37 [-54.73: 36.00],
No anti-inflammatory medication:
Lag 0-24.42 [-81.22: 32.38]
Lag 1 16.52 [-45.76: 78.80],
Anti-inflammatory medication:
Lag 0-3.59 [-75.88: 68.70]
Lag 1-26.76 [-89.53: 36.01]
Central Sites PM25.
Overall:
Lag 0-12.32 [-53.21: 28.56]
Lag 1 5.75 [-33.27: 44.76],
No anti-inflammatory medication:
Lag 0 -33.59 [-89.99: 22.82]
Lag 131.30 [-29.91: 92.51]
Anti-inflammatory medication:
Lag 0-2.13 [-71.99: 67.73]
Lag 1 -3.53 [-67.32: 60.27]
PEF:
Personal PM25
Overall:
Lag 0 0.31 [-4.02: 4.64]
Lag1 -2.19 [-6.49: 2.12]
No anti-inflammatory medication:
Lag 0 0.22 [-8.85: 9.29]
Lag! -10.48 [-18.68: 2.28]
Anti-inflammatory medication:
Lag 0 0.34 [-4.67: 5.35]
Lag1 0.74 [-4.21: 5.69]
Indoor PM2 5
Overall:
Lag 0-8.68 [-16.64:-0.72
Lag 1-9.22 [-17.51:-0.93;
No anti-inflammatory medication:
Lag 0-13.34
Lag1 -17.13
-25.90: -0.79]
-29.86: 4.41],
Anti-inflammatory medication:
Lag 0-5.98 [-15.85: 3.89]
Lag 1-4.19 [-14.59: 6.20]
Outdoor Home P f.l :
Overall:
Lag 0-6.27 [-14.07:1.53]
Lag1 -5.64 [-13.73: 2.44],
No anti-inflammatory medication:
Lag 0-7.52 [-17.56: 2.51]
Lag 1-6.92 [-18.03: 4.19],
Anti-inflammatory medication:
Lag 0-5.22 [-14.77: 4.34]
Lag 1-4.78 [-14.42: 4.86]
Central Sites Pf.l; :
Overall:
Lag 0-5.62 [-12.86:1.62]
Lag 1 -2.45 [-9.34: 4.43],
No anti-inflammatory medication:
Lag 0-6.32 [-16.31: 3.68]
Lag! -0.83[-11.60: 9.95]
Anti-inflammatory medication:
Lag 0-5.29 [-13.42: 2.85]
Lag1 -3.04-10.76:4.67
MMEF
Personal PM25
Overall:
Lag 0-0.99 [-3.96:1.98]
Lag1 -1.08 [-4.05:1.88],
No anti-inflammatory medication:
Lag 0 -3.32 [-9.52: 2.88]
Lag 1 -2.49 [-8.23: 3.25],
Anti-inflammatory medication:
Lag 0-0.31 [-3.77: 3.16]
Lag 1 -0.59 [-4.06: 2.89]
Indoor PM2 5
Overall: Lag 0-3.29 [-8.52:1.94]
Lagl -11.08 [-16.26: 5.90].
Anti-inflammatory medication:
Lag 0-12.65 [-20.74:-4.56]'Lag 1-
December2009 E-218
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Covariates: Gender, age, BMI, history
of respiratory or atopic disease in family,
SHS, acute asthmatic exacerbation in
past 12 mo, ambient temp and relative
humidity, presence of indoor pollutants,
and presence of outdoor pollutants,
Dose-response Investigated? Yes
Statistical Package: S-Plus 2000
Lags Considered: 0-2
13.84 [-21.82: 5.85]
Anti-inflammatory medication:
Lag 02.14 [-4.17: 8.45]
Lag 1-9.33 [-15.89:-2.78]
Outdoor Home P f.l; :
Overall:
Lag 0-4.13 [-9.28:1.01]
Lag 1 -0.73 [-6.02: 4.56]
No anti-inflammatory medication:
Lag 0-8.23 [-14.77:1.69]
Lag! -1.19[-8.45: 6.07]
Anti-inflammatory medication:
Lag 0 -0.68 [-6.87: 5.50]
Lag 1 -0.42 [-6.72: 5.87]
Central Sites PM25.
Overall:
Lag 0-2.10 [-6.99: 2.79
Lag 1-0.12 [-4.67: 4.42
No anti-inflammatory medication:
Lag 0-8.21 [-14.79:1.62]
Lag 1 -0.22 [-7.34: 6.90]
Anti-inflammatory medication:
Lag 00.82 [-4.48: 6.12],
Lag 1-0.09 [-5.19: 5.01]
Reference: Tang et al. (2007, 0912691
Period of Study: Dec 2003-Feb 2005
Location: Sin-Chung City, Taipei
County, Taiwan
Outcome: Peak expiratory flow rate
(PEFR) of asthmatic children
Age Groups: 6-12 yr
Study Design: Panel study
N: 30 children
Statistical Analyses: Linear mixed-
effect models were used to estimate the
effect of PM exposure on PEFR
Pollutant: PM25.i
Averaging Time: 1 h
Mean (SD):
Personal: 6.2 (4.8)
Range (Min, Max):
Personal: 0.3-86.8
Monitoring Stations: 1
No quantitative effects reported.
Reference: Tang et al. (2007, 0912691
Period of Study: Dec 2003-Feb 2005
Location: Sin-Chung City, Taipei
County, Taiwan
Outcome: Peak expiratory flow rate
(PEFR) of asthmatic children
Age Groups: 6-12 yr
Study Design: Panel study
N: 30 children
Statistical Analyses: Linear mixed-
effect models were used to estimate the
effect of PM exposure on PEFR
Covariates: Gender, age, BMI, history
of respiratory or atopic disease in family,
SHS, acute asthmatic exacerbation in
past 12 mo, ambient temp and relative
humidity, presence of indoor pollutants,
and presence of outdoor pollutants,
Dose-response Investigated? yes
Statistical Package: S-Plus 2000
Lags Considered: 0-2
Pollutant: PM1
Averaging Time: 1 h
Mean (SD):
Personal: 34.0 (28.9)
Ambient: 31.4 (18.8)
Range (Min, Max):
Personal: 1.8-284.6
Ambient: 0.1-128.4
Monitoring Stations: 1
PM Increment: 27.6 pg/m
RR Estimate [Lower Cl, Upper Cl]
lag:
Change in morning PEFR:
-6.44 (-30.18, 17.29) lag 0
-12.26 (-77.6, 53.09) lag 1
-4.38 (-54.79, 46.03) lag 2
-44.06 (-113.79, 25.67) 2-day mean
-6.01 (-101.48, 89.46) 3-day mean
Change in evening PEFR:
1.17 (-17.79, 20.13) lag 0
-4.98 (-27.77, 17.81) lag 1
11.30 (-11.55, 34.16) lag 2
41.74(11.36, 72.13) 2-day mean
28.21 (-19.08, 75.5) 3-day mean
December 2009
E-219
-------
Study
Reference: Timonen et al. (2004,
0879151
Period of Study: Oct 1998-Apr 1999
Location:
Amsterdam, The Netherlands
Erfurt, Germany
Helsinki, Finland
Reference: Timonen et al. (2004,
0879151
Period of Study: Oct 1998-Apr 1999
Location :
Amsterdam, The Netherlands
Erfurt, Germany
Helsinki, Finland
Design & Methods
Outcome: Urinary concentration of
Clara cell protein CC16 of subjects with
coronary heart disease
Age Groups: 50+
Study Design: Longitudinal cohort
study (panel)
N:
N=37 (Amsterdam)
N=47 (Erfurt)
N=47 (Helsinki)
Statistical Analyses: The response of
interest was log transformed, creatinine
adjusted CC16. Mixed-effect model was
used to investigate the association
between CC16 and air pollutants.
Covariates: Subjects, long term time
trend, temperature (lags 0-3), relative
humidity (lags 0-3), barometric pressure
(lags 0-3), and weekday of visit.
Dose-response Investigated? yes
Statistical Package:
S-Plus and SAS
Lags Considered: 0-3
Outcome: Urinary concentration of
Clara cell protein CC16 of subjects with
coronary heart disease
Age Groups: 50+
Study Design: Longitudinal cohort
study (panel)
N:
N=37 Amsterdam)
N=47 Erfurt)
N=47 (Helsinki)
Statistical Analyses: The response of
interest was log transformed, creatinine
adjusted CC16. Mixed-effect model was
used to investigate the association
between CC16 and air pollutants.
Covariates: Subjects, long term time
trend, temperature (lags 0-3), relative
humidity (lags 0-3), barometric pressure
(lags 0-3), and weekday of visit.
Dose-response Investigated? Yes
Statistical Package: S-Plus and SAS
Lags Considered: 0-3
Concentrations1
Pollutant: NC 0.01 -0.1
Averaging Time: 24 h
Mean (SD):
Amsterdam: 17338 /cm3
Erfurt: 21124 /cm3
Helsinki: 17041 /cm3
Range (Min, Max):
Amsterdam: 5699-37 195
Erfurt: 3867-96678
Helsinki: 2305-50306
Unit (i.e. pg/m3): 1/cm3
Monitoring Stations: 3
PM25:
Amsterdam -0.1 5
Erfurt 0.62
Helsinki 0.1 4
N02:
Amsterdam 0.49
Erfurt 0.82
Helsinki 0 72
CO:
Amsterdam 0.22
Erfurt 0.72
Helsinki 0.35
Pollutant: NC1 0-0.1
Averaging Time: 24 h
Mean (SD):
Amsterdam: 2131 /cm3
Erfurt: 1829 /cm3
Helsinki: 1390 /cm3
Range (Min, Max):
Amsterdam: 413-6413
Erfurt: 303-6848
Helsinki: 344-3782
Unit (i.e. pg/m3): 1/cm3
Monitoring Stations: 3
Copollutant (correlation):
Spearman Correlation:
NC 0.1-0.01:
Amsterdam 0.1 6
Erfurt 0.67
Helsinki 0.53
PM25:
Amsterdam 0.80
Erfurt 0.84
Helsinki 0 80
N02:
Amsterdam 0.67
Erfurt 0.82
Helsinki 0.72
CO:
Amsterdam 0.60
Erfurt 0.78
Helsinki 0.51
Effect Estimates (95% Cl)
PM Increment: 10,000 /cm3
RR Estimate [Lower Cl, Upper Cl]
lag:
Pooled estimate;
1.7 (-4.4-7.8) lag 0
-1.8 (-8.3-4.6) lag 1
1.5 (-5.6-8.6) lag 2
2.3 (-4.8-9.3) lag 3
1.8 (-9.4-13.0) 5-day mean
There was no association between NC
0.01-0.1 and CC16 in the pooled
analysis.
PM Increment: 1000 /cm3
RR Estimate [Lower Cl, Upper Cl] lag:
Pooled estimate;
4.3 (-1.4-10.0) lag 0
5.1 (-0.6-10.7) lag 1
4.5 (-0.5-9.6) lag 2
1.6 (-3.5-6.7) lag 3
13.1 (-4.3-30.5) 5-day mean
CC16 was not associated to NC 0.1-1.0
in the pooled analysis but CC16 was
significantly associated to NC 0.1-1.0 in
Helsinki:
15.5 (0.001 -30.9) lag 0
10.8 (-4.2-25.8) lag 1
10. 5 9-4. 1-25.1) lag 2
17.4 (3.4-31. 4) lag 3
43.2 (17.4-69.0) 5-day mean
December 2009
E-220
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: von Klot et al. (2002,
0347061
Period of Study: Sep 1996-Mar 1997
(winter)
Location: Erfurt, Germany
Outcome: Asthma symptoms
(wheezing, shortness of breath at rest,
waking up with breathing problems, or
coughing without having a cold) and
Asthma medication (inhaled short-acting
IS2- agonists, inhaled long-acting IJ2-
agonists, inhaled corticosteroids,
cromolyn sodium, theophylline, oral
corticosteroids, and N-acetylcysteine)
Age Groups: Adults, mean=59.0 yr and
range =37-77 yr
Study Design: Panel study
N: 53 adult asthmatics
Statistical Analyses: Logistic
regression models
Covariates: Seasonal variation in
medication use or symptom prevalence,
meteorological factors (relative humidity,
temperature), weekend, Christmas
holidays
Season: Winter
Dose-response Investigated? No
Statistical Package: NR
Lags Considered:0,1,2,3,4,5,6,7,
8, 9,10, ma calculated from same day
and preceding days
Pollutant: MCO.5-0.1
Averaging Time: 10-min intervals
Mean (SD): 24.8
Percentiles:
25th: 11.4
SOth(Median): 19.6
75th: 33.1
Range (Min, Max): (2.4-108.3)
Copollutant (correlation):
PM10.25:r=0.51
NCO. 1-0.01 :r= 0.45
NC0.5-0.1:r=0.95
NC2.5-0.5: r= 0.92
MC2.5-0.01:r=1.00
PMi0:r=0.91
N02: r= 0.69
CO: r= 0.66
SO,: r= 0.60
NC Increment: 1 IQR
Effect Estimate [Lower Cl, Upper Cl]:
Association between the prevalence of
inhaled B2- agonist use and MCO. 1-0.5
Same day, IQR=21,
OR= 0.98 (0.92-1.04)
5-day mean, IQR= 21
OR= 1.11 (1.02-1.20)
14-day mean IQR= 17,
OR=1.01 (0.93-1.10)
Association between the prevalence of
inhaled corticosteroid use and
MCO. 1-0.5
Same day, IQR= 2,
OR= 1.09 (1.02-1.17)
5-day mean IQR=21,
OR= 1.28 (1.18-1.39)
14-day mean, IQR=17,
OR= 1.49 (1.38-1.61)
Association between the prevalence of
wheezing and MCO.1-0.5
Same day, IQR=21,
OR= 1.01 (0.94-1.08)
5-day mean, IQR=21,
OR= 1.08 (0.99-1.17)
14-day mean, IQR=17,
OR= 1.05 (0.96-1.15)
Reference: von Klot et al. (2002,
0347061
Period of Study: Sep 1996-Mar 1997
(winter)
Location: Erfurt, Germany
Outcome: Asthma symptoms
(wheezing, shortness of breath at rest,
waking up with breathing problems, or
coughing without having a cold) and
Asthma medication (inhaled short-acting
B2- agonists, inhaled long-acting B2-
agonists, inhaled corticosteroids,
cromolyn sodium, theophylline, oral
corticosteroids, and N-acetylcysteine)
Age Groups: Adults, mean=59.0 yr and
range =37-77 yr
Study Design: Panel study
N: 53 adult asthmatics
Statistical Analyses: Logistic
regression models
Covariates: Seasonal variation in
medication use or symptom prevalence,
meteorological factors (relative humidity,
temperature), weekend, Christmas
holidays
Season: Winter
Dose-response Investigated? No
Statistical Package: NR
Lags Considered:0,1,2,3,4,5,6,7,
8, 9,10, ma calculated from same day
and preceding days
Pollutant: MC2.5-0.01
Averaging Time: 10-min intervals
Mean (SD): 30.3
Percentiles:
25th: 13.5
SOth(Median): 24.6
75th: 41.3
Range (Min, Max): (3.6-133.8)
Copollutant (correlation):
PMio-25: r= 0.52
NC0.5-o.i: r=0.45
NC25.o5: r= 0.94
MCo.5-o.i:r=1.00
NC0.i-0.oi:r=0.45
PM10:r=0.94
N02: r= 0.68
CO: r= 0.65
S02: r= 0.62
NC Increment: 1 IQR
Effect Estimate [Lower Cl, Upper Cl]:
Association between the prevalence of
inhaled B2- agonist use and MC0.01-
2.5
Same day, IQR= 28, OR= 0.96 (0.90-
1.04)
5-day mean, IQR= 26 , OR= 1.10 (1.01-
1.20)
14-day mean, IQR=20, OR= 1.03
(0.95-1.12)
December 2009
E-221
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: von Klot et al. (2002,
0347061
Period of Study: Sep 1996-Mar 1997
(winter)
Location: Erfurt, Germany
Outcome: Asthma symptoms
(wheezing, shortness of breath at rest,
waking up with breathing problems, or
coughing without having a cold) and
Asthma medication (inhaled short-acting
IS2- agonists, inhaled long-acting IJ2-
agonists, inhaled corticosteroids,
cromolyn sodium, theophylline, oral
corticosteroids, and N-acetylcysteine)
Age Groups: Adults, mean=59.0 yr and
range =37-77 yr
Study Design: Panel study
N: 53 adult asthmatics
Statistical Analyses: Logistic
regression models
Covariates: Seasonal variation in
medication use or symptom prevalence,
meteorological factors (relative humidity,
temperature), weekend, Christmas
holidays
Season: Winter
Dose-response Investigated? No
Statistical Package: NR
Lags Considered:0,1,2,3,4,5,6,7,
8, 9,10, ma calculated from same day
and preceding days
Pollutant: NCO. 1-0.01
Averaging Time: 10-min intervals
Mean (SD): 17,300/cm3
Percentiles:
25th: 9286
SOth(Median): 16940
75th: 24484
Range (Min, Max): (3272-46195)
Unit (i.e. pg/m3): 1/cm3
Copollutant (correlation):
PM10.2.5: r= 0.41
NC0.5-o.i: r=0.55
NC25.o5: r= 0.34
MC0.5.0,:r=0.45
MC2.5-o.oi:r=0.45
PMi0:r=0.51
N02: r= 0.66
CO: r= 0.66
SO,: r= 0.36
NC Increment: 1 IQR
Effect Estimate [Lower Cl, Upper Cl]:
Association between the prevalence of
inhaled B2- agonist use and NCO.01-0.1
Same day, IQR= 15000,
OR= 0.97 (0.90-1.04)
5-day mean, IQR= 10000,
OR= 1.11 (1.01-1.21)
14-day mean, IQR=7700,
OR= 1.08 (0.96-1.21)
Association between two pollutants,
jointly in one model, and the
Outcomes
Inhaled short-acting B2- agonist use
NCO. 1-0.01 OR= 1.07 (0.97-1.18)
MC0.5-0.1:OR= 1.07 (0.98-1.18)
Inhaled corticosteroid use
NCO. 1-0.01 OR= 1.01 (0.87-1.18)
MC0.5-0.1:OR= 1.53 (1.39-1.69)
Wheezing
NCO. 1-0.01 OR= 1.12 (1.01-1.24)
MC0.5-0.1:OR= 1.02 (0.92-1.12)
Association between the prevalence of
inhaled corticosteroid use and NCO.01-
0.1
Same day, IQR= 15000,
OR= 1.07 (1.00-1.15)
5-day mean, IQR= 10000,
OR= 1.22 (1.12-1.33)
14-day mean, IQR=7700,
OR= 1.45 (1.29-1.63)
Association between the prevalence of
wheezing and NCO. 1-0.01
Same day, IQR= 15000,
OR= 0.94 (0.86-1.01)
5-day mean, IQR= 10000,
OR= 1.13 (1.03-1.24)
14-day mean, IQR=7700,
OR= 1.27 (1.13-1.43)
Association between the prevalence of
respiratory symptoms and NCO.1-0.01
Attack of shortness of breath and
wheezing
Same day, IQR= 15000,
OR= 1.01 (0.91-1.12)
5-day mean, IQR= 10000,
OR= 1.08 (0.96-1.21)
14-day mean, IQR=7700,
OR= 1.26 (1.08-1.48)
Walking up with breathing problems
Same day, IQR= 15000,
OR= 1.04 (0.96-1.13)
5-day mean, IQR= 10000,
OR= 1.09 (0.99-1.19)
14-day mean, IQR=7700,
OR= 1.26 (1.13-1.41)
Shortness of breath
Same day, IQR= 15000,
OR= 0.98 (0.90-1.06)
5-day mean, IQR= 10000,
OR= 1.09 (0.99-1.19)
14-day mean, IQR=7700,
OR= 1.24 (1.11-1.40)
Phlegm
Same day, IQR= 15000,
OR=1.01 (0.94-1.09)
December 2009
E-222
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
5-day mean, IQR= 10000,
OR=1.11 (1.02-1.21)
14-day mean, IQR=7700,
OR= 1.11(0.99-1.25)
Cough
Same day, IQR= 15000,
OR= 1.07 (0.98-1.16)
5-day mean, IQR= 10000,
OR= 1.17 (1.07-1.28)
14-day mean, IQR=7700,
OR= 1.20 (1.06-1.35)
Reference: von Klot et al. (2002,
0347061
Period of Study: Sep 1996-Mar 1997
(winter)
Location: Erfurt, Germany
Outcome: Asthma symptoms
(wheezing, shortness of breath at rest,
waking up with breathing problems, or
coughing without having a cold) and
Asthma medication (inhaled short-acting
IS2- agonists, inhaled long-acting IJ2-
agonists, inhaled corticosteroids,
cromolyn sodium, theophylline, oral
corticosteroids, and N-acetylcysteine)
Age Groups: Adults, mean=59.0 yr and
range =37-77 yr
Study Design: Panel study
N: 53 adult asthmatics
Statistical Analyses: Logistic
regression models
Covariates: Seasonal variation in
medication use or symptom prevalence,
meteorological factors (relative humidity,
temperature), weekend, Christmas
holidays
Season: Winter
Dose-response Investigated? No
Statistical Package: NR
Lags Considered:0,1,2,3,4,5,6,7,
8, 9,10, ma calculated from same day
and preceding days
Pollutant: NCO.5-0.1
Averaging Time: 10-min intervals
Mean (SD): 2005/cm3
Percentiles:
25th: 958
SOth(Median): 1610
75th: 2767
Range (Min, Max): (291 -6700)
Unit (i.e. pg/m3): 1/cm3
Copollutant (correlation):
PM10.25: r= 0.50
NC0.i-0.oi:r=0.55
NC2.5-o.5: r= 0.76
MC0 5.0 ,:r= 0.95
MC25-ooi:r=0.93
PM10:r=0.85
N02: r= 0.75
CO: r= 0.79
S02: r= 0.51
NC Increment: 1 IQR
Effect Estimate [Lower Cl, Upper Cl]:
Association between the prevalence of
inhaled IS2- agonist use and NCO.5-0.1
Same day, IQR=1800,
OR= 0.99 (0.92-1.05)
5-day mean, IQR= 1500,
OR= 1.10 (1.03-1.19)
14-day mean, IQR=1450,
OR= 0.95 (0.86-1.05)
Association between the prevalence of
inhaled corticosteroid use and NCO.5-
0.1
Same day, IQR=1800,
OR= 1.06 (0.99-1.14)
5-day mean, IQR= 1500,
OR= 1.23 (1.14-1.32)
14-day mean, IQR=1450,
OR=1.51 (1.37-1.67)
Association between the prevalence of
wheezing and NCO.5-0.1
Same day, IQR=1800,
OR= 1.00 (0.93-1.07)
5-day mean, IQR= 1500,
OR= 1.08 (1.00-1.17)
14-day mean, IQR=1450,
OR=1.11 (1.00-1.24)
December 2009
E-223
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: von Klot et al. (2002,
0347061
Period of Study: Sep 1996-Mar 1997
(winter)
Location: Erfurt, Germany
Outcome: Asthma symptoms
(wheezing, shortness of breath at rest,
waking up with breathing problems, or
coughing without having a cold) and
Asthma medication (inhaled short-acting
IS2- agonists, inhaled long-acting IJ2-
agonists, inhaled corticosteroids,
cromolyn sodium, theophylline, oral
corticosteroids, and N-acetylcysteine)
Age Groups: Adults, mean=59.0 yr and
range =37-77 yr
Study Design: Panel study
N: 53 adult asthmatics
Statistical Analyses: Logistic
regression models
Covariates: Seasonal variation in
medication use or symptom prevalence,
meteorological factors (relative humidity,
temperature), weekend, Christmas
holidays
Season: Winter
Dose-response Investigated? No
Statistical Package: NR
Lags Considered:0,1,2,3,4,5,6,7,
8, 9,10, ma calculated from same day
and preceding days
Pollutant: NC2.5-0.5
Averaging Time: 10-min intervals
Mean (SD): 21.4/cm3
Percentiles:
25th: 5.6
SOth(Median): 13.0
75th: 31.6
Range (Min, Max): (0.9-127.6)
Unit (i.e. pg/m3): 1/cm3
Copollutant (correlation):
PM10.25: r= 0.48
NC0.i-0.oi:r=0.34
NC0 5-0 a= 0.76
MC0 5.0 ,:r= 0.92
MC2.5-o.oi:r=0.94
PM10:r=0.88
N02: r= 0.54
CO: r= 0.46
SO,: r= 0.66
NC Increment: 1 IQR
Effect Estimate [Lower Cl, Upper Cl]:
Association between the prevalence of
inhaled B2- agonist use and NC2.5-0.5
Same day, IQR= 26, OR= 0.99 (0.93-
1.05)
5-day mean, IQR= 22, OR= 1.09 (1.01-
1.17)
14-day mean, IQR= 17, OR= 1.08
(1.02-1.15)
Association between the prevalence of
inhaled corticosteroid use and NC2.5-
0.5
Same day, IQR= 26, OR= 1.13 (1.06-
1.21)
5-day mean, IQR= 22, OR= 1.28 (1.19-
1.37)
14-day mean, IQR= 17, OR= 1.44
(1.36-1.53)
Association between the prevalence of
wheezing and NC2.5-0.5
Same day, IQR= 26, OR= 1.03 (0.95-
1.10)
5-day mean, IQR= 22, OR= 1.05 (0.97-
1.13)
14-day mean, IQR= 17, OR= 1.03
(0.96-1.10)
Reference: von Klot et al. (2002,
0347061
Period of Study: Sep 1996-Mar 1997
(winter)
Location: Erfurt, Germany
Outcome: Asthma symptoms
(wheezing, shortness of breath at rest,
waking up with breathing problems, or
coughing without having a cold) and
Asthma medication (inhaled short-acting
B2- agonists, inhaled long-acting B2-
agonists, inhaled corticosteroids,
cromolyn sodium, theophylline, oral
corticosteroids, and N-acetylcysteine)
Age Groups: Adults, mean=59.0 yr and
range =37-77 yr
Study Design: Panel study
N: 53 adult asthmatics
Statistical Analyses: Logistic
regression models
Covariates: Seasonal variation in
medication use or symptom prevalence,
meteorological factors (relative humidity,
temperature), weekend, Christmas
holidays
Season: Winter
Dose-response Investigated? No
Statistical Package: NR
Lags Considered:0,1,2,3,4,5,6,7,
8, 9,10, ma calculated from same day
and preceding days
Pollutant: PMi0.2.5
Averaging Time: 24 h
Mean (SD): 10.3
Percentiles:
25th: 2.9
SOth(Median): 6.9
75th: 14.6
Range (Min, Max): (-8.7-64.3)
Copollutant (correlation):
NC0.i-0.oi:r=0.41
NC0.5-o.i: r=0.50
NC25.05: r= 0.48
MCo.5-o.i:r=0.51
MC2.5.o.oi:r=0.52
PM10:r=0.67
N02: r= 0.45
CO: r= 0.42
S02: r= 0.28
PM Increment: 1 IQR
Effect Estimate [Lower Cl, Upper Cl]:
Association between the prevalence of
inhaled B2- agonist use and PM10.25
Same day, IQR= 12, OR= 1.01 (0.95-
1.06)
5-day mean, IQR= 11, OR= 1.01 (0.94-
1.09)
14-day mean, IQR= 6.7, OR= 0.92
(0.86-1.00)
Association between the prevalence of
inhaled corticosteroid use and PMi0.25
Same day, IQR= 12, OR= 1.03 (0.98-
1.08)
5-day mean, IQR= 11, OR= 1.12 (1.04-
1.20)
14-day mean, IQR= 6.7, OR= 1.27
(1.18-1.37)
Association between the prevalence of
wheezing and PMi0.2.5
Same day, IQR= 12, OR= 0.97 (0.91-
1.02)
5-day mean, IQR= 11, OR= 1.06 (0.98-
1.15)
14-day mean, IQR= 6.7, OR= 1.05
(0.96-1.15)
Reference: Ward et al. (2002, 0258391
Period of Study: 1997 (two 8-wk
periods)
Location: Birmingham and Sandwell,
UK
Outcome: Change in PEF (peak
expiratory flow), self reported
respiratory symptoms (same day
cough, illness, short of breath, waking
up at night with cough or wheeze,
wheeze)
Age Groups: 9 yr olds
Study Design:
Pollutant: PM25
Averaging Time: 24 h
Mean (SD):
Wnter:12.7|jg/m3
Summer: 12.3 pg/m3
Range (Min, Max):
PM Increment:
Wnter: 12.3 pg/m3
Summer: 6.3 pg/m3
Mean (PEF l/min) [Lower Cl, Upper Cl]
lag:
Winter morning:
0.80 [-1.97, 3.67] lag 0
0.62 [-2.22, 3.54] lag 1
December 2009
E-224
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Time-series Panel study
N: 162 children from 5 schools
Statistical Analyses: Linear regression
(PEF),
Logistic regression (respiratory
symptoms)
Covariates: Trend, temperature,
schooldays (yes/no)
Season: Winter (Jan 13-Mar 10)
Summer (May 19-Jul 14)
Dose-response Investigated? No
Statistical Package: Nr
Lags Considered: Lag 0, lag 1, lag 2,
lag 3, 7-day ma
Winter: 4, 37
Summer: 5, 28
PM Component:
Total mass
Monitoring Stations:
5 stations near the 5 schools
Copollutant (correlation):
Wnter:
PM10(r=0.93)
N02 (F0.88)
03 (p-0.83)
Summer:
HN03(r=0.81)
-0.86
-2.47
-4.07
-4.32, 2.47] lag 2
-5.30, 0.36] lag 3
-10.60, 2.42] 7-day mean
Winter afternoon:
0.95 [-2.22, 4.23] lag 0
-0.99 [-4.69, 2.72] lag 1
-1.60 [-5.18, 2.01] lag 2
-3.45 [-6.53 to-0.25] lag 3
1.00 [-11.47,13.56] 7-day mean
Summer morning:
-1.49 [-3.65, 0.67] lag 0
0.21
2.50
3.41
3.90
-2.12, 2.55] lag 1
0.28, 4.72] lag 2
1.40, 5.44] lag 3
-2.53,10.33] 7-day mean
Summer afternoon:
-0.49 [-2.43, 1.45
-0.78 [-2.72, 1.16
lagO
Iag1
0.57 [-1.35, 2.49] lag 2
0.16 [-1.85, 2.17] lag 3
-0.08 [-5.43, 5.27] 7-day mean
Winter morning in atopy/recent
wheezing subgroup:
-0.072
-0.271
-0.527, 0.383
lagO
Iag1
-0.701,0.159
0.127[-0.354, 0.608]"lag'2
0.055 [-0.391, 0.501] lag 3
Winter morning in no atopy or recent
wheezing subgroup:
0.126 [-0.413, 0.666] lag0
0.193 [-0.340, 0.728] lag 1
-0.170
-0.314
-0.788, 0.447
-0.846,0.216
lag 2
Iag3
Winter morning in subgroup with
parental atopy/recent wheezing:
0.187 [-0.008, 0.382] lag0
-0.006 [-0.207, 0.195] lag 1
-0.011 [-0.226, 0.204] lag 2
-0.037 [-0.228, 0.154] lag3
Winter morning in subgroup without
parental atopy/recent wheezing:
0.026 [-0.341 , 0.395] lag 0
0.068 [-0.307 , 0.444] lag 1
-0.099 -0.535 , 0.335] lag 2
-0.252 -0.61 5, 0.1 10] lag 3
RR Estimate [Lower Cl, Upper Cl]
lag:
Cough:
Winter: 0.98 [0.80, 1.1 8] lag 0
0.95
1.02
1.01
1.31
0.77, 1.17
0.83, 1.24
0.83, 1.23
0.82, 2.09
Summer: 1.13
1.04
0.94
0.89
0.81
0.94, 1.13
0.87, 1.02
0.82, 0.96
0.62, 1.06
Iag1
lag 2
Iag3
7-day mean
1.04, 1.22] lag 0
Iag1
lag 2
lagS
7 day mean
Illness:
Winter: 1.17 [1.05, 1.32] lag 0
1.07
1.16
1.01
1.57
0.95, 1.23
1.01, 1.35
0.90, 1.16
1.15,2.13
Iag1
lag 2
lagS
7-day mean
Summer: 1.02 [0.91, 1.13] lag 0
1.00
0.96
0.97
0.68
0.89, 1.13
0.85, 1.07
Iag1
lag 2
0.86, 1.09] lag 3
0.41, 1.1 3] 7-day mean
December 2009
E-225
-------
Study
Reference: Ward et al. (2002, 0258391
Period of Study: 1997 (two 8-wk
periods)
Location: Birmingham and Sandwell,
UK
Design & Methods
Outcome: Change in PEF (peak
expiratory flow), self reported
respiratory symptoms (same day
cough, illness, short of breath, waking
up at night with cough or wheeze,
wheeze)
Age Groups: 9 yr olds
Study Design:
Time-series panel study
N: 162 children from 5 schools
Statistical Analyses: Linear regression
(PEF),
Logistic regression (respiratory
symptoms)
Covariates: Trend, temperature,
schooldays (yes/no)
Season: Winter (Jan 13-Mar 10)
Summer (May 19- Jul 14)
Concentrations1
Pollutant: Sulfate
Averaging Time: 24 h
Mean (SD):
Wnter: 2.4 pg/m3
Summer: 3.8 pg/m3
Range (Min, Max):
Wnter: 0.8, 14.9
Summer: 1.1, 7.8
PM Component:
S04
Monitoring Stations: 2 stations
Effect Estimates (95% Cl)
Shortness of breath:
Winter: 1.07 [0.94, 1.24] lag 0
0.98 0.84, 1.13] lag 1
0.96 0.82, 1.13]lag2
0.91 0.79, 1.07] lag 3
0.82 0.58, 1.1 8] 7-day mean
Summer: 1.04 [0.90, 1.20] lag 0
1.08 0.93, 1.25] lag 1
0.97 0.84, 1.13] lag 2
0.93 0.81, 1.08] lag 3
1.16 0.76, 1.77] 7-day mean
Wake at night with cough/wheeze:
Winter: 1.10 [0.96, 1.26] lag 0
1.05 0.90, 1.22 Iag1
0.98 0.83, 1.13; lag 2
0.94 0.81, 1.09; lag 3
0.93 0.66, 1.32 7-day mean
Summer: 0.93 0.78, 1.10]lagO
0.81 0.67, 0.98 lag 1
0.91 0.77,1.09 lag 2
0.97 0.83, 1.13 Iag3
1.04 0.57, 1.90 7-day mean
Wheeze:
Winter: 0.98 [0.83, 1.1 6] lag 0
0.90 0.75, 1.05] lag 1
1.00 0.83, 1.20] lag 2
1.13 0.95, 1.35] lag 3
1.02 0.68, 1.57]; 7-day mean
Summer: 1.02 [0.88, 1.19]lagO
0.98 0.84, 1.16] lag 1
0.87 0.74, 1.02] lag 2
0.85 0.72, 0.99] lag 3
0.96 0.51, 1.81] 7-day mean
PM Increment:
Wnter: 4.8 pg/m3
Summer: 3.1 pg/m3
Mean (PEF l/min) [Lower Cl, Upper Cl]
lag
Winter morning:
-1.75 -4.00, 0.50] lag 0
-0.91 -3.44, 1.62] lag 1
-0.62 -3.16, 1.91] lag 2
-1.82 -4.27, 0.64] lag 3
-3.22 -8.03, 1.58] 7-day mean
Winter afternoon:
0.99 [-1.58, 3.55] lag 0
0.79 [-2.42, 4.00] lag 1
-1.89 [-4.99, 1.21] lag 2
-1.73 [-4.69, 1.23] lag 3
-1.96 [-13.35, 9.42] 7-day mean
Summer morning:
-0.72 [-3.27, 1.82] lag 0
-1. 69 [-4.28, 0.90] lag 1
1.35 -1.27, 3.97] lag 2
3.38 1.03, 5.72] lag 3
Statistical Package: Nr
Lags Considered: Lag 0, lag 1, lag 2,
lag 3, 7-day ma
2.98 [-4.17,10.13] 7-day mean
Summer afternoon:
-0.32 [-2.81, 2.17] lag 0
0.84 [-1.63, 3.30] lag 1
-0.08 [-2.61, 2.44] lag 2
-0.25 [-2.69, 2.19]lag3
-2.20 [-9.51, 5.12] 7-day mean
Winter morning in atopy/recent
wheezing subgroup:
0.200 [-0.755, 1.156] lag 0
-0.219
-0.431
-1.318,0.881
-1.526,0.664
Iag1
;lag2
1.200 [0.095, 2.305] lag 3
December 2009
E-226
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Winter morning in no atopy or recent
wheezing subgroup:
-0.613 [-1.714, 0.488] lag 0
-0.174 [-1.423, 1.075] lag 1
0.006 [-1.243, 1.253] lag 2
-1.080 [-2.308, 0.148] lag 3
Winter morning in subgroup with
parental atopy/recent wheezing:
0.457 [0.003, 0.910] lag0
0.078 [-0.503, 0.660] lag 1
-0.102[-0.656, 0.452] lag 2
0.002 [-0.609, 0.613] lag 3
Winter morning in subgroup without
parental atopy/recent wheezing:
-0.622 -1.379, 0.1 36] lag 0
-0.272 -1.147, 0.602] lag 1
-0.138 -1.005, 0.728] lag 2
-0.496 -1.359, 0.367] lag 3
RR Estimate [Lower Cl, Upper Cl]
lag:
Cough:
Winter: 1.01 [0.84, 1.20] lag 0
1.02
0.99
0.86
0.78
0.85, 1.24
0.82, 1.20
0.71, 1.05
0.53, 1.14
Summer: 1.08
1.03
0.97
0.90
0.73
0.93, 1.15
0.88, 1.07
0.82, 0.99
0.54, 0.97
Iag1
lag 2
Iag3
7-day mean
0.98, 1.20] lag 0
Iag1
lag 2
Iag3
7 day mean
Illness:
Winter: 1.06 [0.96, 1.1 7] lag 0
1.15
1.14
1.04
1.30
1.03, 1.28
1.00, 1.28
0.92, 1.18
1.00, 1.66
Iag1
lag 2
lagS
7-day mean
Summer: 0.98 [0.86 1.11]lagO
0.97
1.01
0.95
0.72
0.84, 1.12
0.88, 1.16
0.84, 1.09
0.46, 1.12
Iag1
lag 2
lagS
7-day mean
Shortness of breath:
Winter: 0.96 [0.85, 1.07] lag 0
0.98
0.94
0.93
0.80
0.86, 1.12
0.82, 1.07
0.81, 1.08
0.59, 1.07
Iag1
Iag2
lagS
7-day mean
Summer: 0.95 [0.80, 1.14]lagO
1. 07 [0.89, 1.28] lag 1
1.04
0.94
0.87, 1.24
0.80, 1.12
lag 2
lagS
|0.58[0.33, 1.04] 7-day mean
Wake at night with cough/wheeze:
Winter: 0.97 [0.87, 1.08] lag 0
1.01
1.00
0.93
0.79
0.89, 1.15
0.88, 1.14
0.82, 1.07
0.59, 1.05
Summer: 0.95
0.81
0.93
0.87
0.77
0.67, 0.99
0.76, 1.13
0.72, 1.05
0.41, 1.48
Iag1
;lag2
;lag3
7-day mean
0.78, 1.16] lag 0
Iag1
lag 2
lagS
7-day mean
Wheeze:
Winter: 1.00 [0.87, 1.1 5] lag 0
0.96 [0.82, 1.13]lag1
0.88
1.12
0.75, 1.04
0.95, 1.32
lag 2
lagS
December 2009
E-227
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
0.83 [0.58,1.20]; 7-day mean
Summer: 0.97 [0.80,1.17]lagO
.09 [0.89, 1.32] lag 1
1.00 [0.82, 1.22] lag 2
0.81
1.30
0.69, 0.97
0.68, 2.50
Iag3
7-day mean
Reference: Ward et al. (2002, 0258391
Period of Study: 1997 (two 8-week
periods)
Location: Birmingham and Sandwell,
UK
Outcome: Change in PEF (peak
expiratory flow), self reported
respiratory symptoms (same day
cough, illness, short of breath, waking
up at night with cough or wheeze,
wheeze)
Age Groups: 9 yr olds
Study Design: Time-series panel study
N: 162 children from 5 schools
Statistical Analyses: Linear regression
(PEF),
Logistic regression (respiratory
symptoms)
Covariates: Trend, temperature,
schooldays (yes/no)
Season: Winter (Jan 13-Mar 10)
Summer (May 19-Jul 14)
Dose-response Investigated? No
Statistical Package: Nr
Lags Considered: Lag 0, lag 1, lag 2,
lag 3, 7-day ma
Pollutant: N03
Averaging Time: 24 h
Mean (SD):
Wnter: 3.6 pg/m3
Summer: 3.5 pg/m3
Range (Min, Max):
Wnter: 0.1, 29.9
Summer: 0.7,13.2
Monitoring Stations: 2 stations
PM Increment: Wnter: 6.7 pg/m3
Summer: 3.7 pg/m3
Mean (PEF l/min) [Lower Cl, Upper Cl]
lag:
Winter morning:
-2.08 [-4.02 to-0.15] lagO
-0.64 [-2.87, 1.59] lag 1
0.71 [-1.69, 3.11] lag 2
-1.38 [-3.61, 0.84] lag 3
-0.92 [-5.32, 3.47] 7-day mean
Winter afternoon:
0.24 [-1.89, 2.38] lagO
-0.72 [-3.87, 2.43] lag 1
-1.37 [-5.11, 2.38] lag 2
-2.54 [-5.74, 0.66] lag 3
0.21 [-7.67, 8.11] 7-day mean
Summer morning:
-0.80 [-2.74, 1.15] lag 0
0.68
1.42
2.54
1.74
-1.31, 2.67]Iag1
-0.73, 3.58] Iag2
0.48, 4.59] Iag3
-2.66, 6.13] 7-day mean
Summer afternoon:
-0.72 [-2.47, 1.03
-0.59 [-2.36, 1.18
lagO
Iag1
-0.33 [-2.11, 1.45] lag 2
0.66 [-1.26, 2.58] lag 3
0.47 [-3.36, 4.29] 7-day mean
Winter morning in atopy/recent
wheezing subgroup:
-0.036 [-0.627 , 0.555] lag 0
0.142 [-0.573, 0.857] lag 1
0.000 [-0.760, 0.759] lag 2
0.689 [-0.061, 1.439] lag 3
Winter morning in no atopy or recent
wheezing subgroup:
-0.434 [-1.116, 0.248] lag 0
-0.201 [-1.002, 0.600] lag 1
0.154 [-0.703, 1.010] lag 2
-0.605 [-1.422, 0.210] lag 3
Winter morning in subgroup with
parental atopy/recent wheezing:
0.228
0.476
0.196
0.083
-0.054, 0.511] lag 0
0.060, 0.892] lag 1
-0.202, 0.594] lag 2
-0.321, 0.487] lag 3
Winter morning in subgroup without
parental atopy/recent wheezing:
-0.482 [-0.952,-0.012] lag 0
-0.276 [-0.846, 0.294] lag 1
0.078 [-0.520, 0.675] lag 2
-0.298 [-0.864, 0.268] lag 3
RR Estimate [Lower Cl, Upper Cl]
Cough: Winter:
0.92
0.91
0.99
0.87
0.71
0.80, 1.07
0.77, 1.07
0.83, 1.17
0.73, 1.03
0.52, 0.97
lagO
Iag1
lag 2
Iag3
7-day mean
Summer: 1.05 [0.97, 1.13]lagO
December 2009
E-228
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
1.01 [0.93,1.10] lag 1
0.95 [0.88, 1.03] lag 2
0.89 [0.83, 0.96] lag 3
0.81 [0.68, 0.97] 7 day mean
Illness: Winter: 1.05 [0.97,1.14] lag 0
1.11
1.13
1.13
1.13
1.01,1.22
1.01, 1.26
1.04, 1.26
0.92, 1.38
Iag1
lag 2
Iag3
7-day mean
Summer: 0.97 [0.87, 1.09] lag 0
0.98
0.95
0.94
0.74
0.87, 1.10
0.85, 1.06
0.85, 1.05
0.54, 1.03
Iag1
lag 2
Iag3
7-day mean
Shortness of breath: Winter:
0.99
1.01
0.93
0.98
0.85
0.90, 1.10
0.90, 1.13
0.82, 1.05
0.86, 1.13
0.67, 1.08
Summer: 1.04
1.12
1.04
0.90
1.06
0.98, 1.28
0.90, 1.20
0.79, 1.03
0.78, 1.43
lagO
Iag1
Iag2
lags
7-day mean
0.90, 1.18] lag 0
Iag1
lag 2
lagS
7-day mean
Wake at night with cough/wheeze:
Winter:
0.98
1.05
0.99
0.99
0.84
0.89, 1.08
0.94, 1.16
0.88, 1.12
0.87, 1.12
0.67, 1.05
lagO
Iag1
;lag2
;lag3
7-day mean
Summer: 0.94 [0.80, 1.09] lag 0
0.86
0.94
0.92
0.95
0.72, 1.01
0.79, 1.11
0.79, 1.07
0.62, 1.47
Iag1
lag 2
lagS
7-day mean
Wheeze: Winter: 0.98 [0.87,1.10] lag 0
1.00
0.89
1.11
0.80
0.87, 1.14
0.77, 1.03
0.95, 1.30
0.61, 1.07
Summer: 1.01
0.96
0.95
0.87
1.04
0.83, 1.11
0.82, 1.10
0.75, 1.01
0.67, 1.60
Iag1
lag 2
lagS
7-day mean
0.87, 1.17] lag 0
Iag1
lag 2
lagS
7-day mean
All units expressed in pg/m unless otherwise specified.
December 2009
E-229
-------
E.2.2. Respiratory Emergency Department Visits and Hospital Admissions
Table E-12. Short-term
Reference
Reference: Andersen et al. (2008,
1896511
1st page: 458
Period of Study: May 2001- Dec 2004
Location: Copenhagen, Denmark
Reference: Cheng et al. (2007,
0930341
Period of Study: 1996-2004
Location: Kaohsiung, Taiwan
exposure-respiratory-ED/HA-PMi .
Design & Methods
Hospital Admissions/ED visits
Outcome (ICD-10):
RD, including chronic bronchitis
(J41-42), emphysema (J43), other
chronic obstructive pulmonary disease
(J44), asthma (J45), and status
asthmaticus (J46).
Pediatric hospital admissions for
asthma (J45) and status asthmaticus
/ \AR\
(JtO).
Age Groups Analyzed: >65 yr (RD
combined), 5-1 Syr (asthma)
Study Design: Time series
N: NR
Statistical Analyses: Poisson GAM
Covariates: temperature, dew-point
temperature, long-term trend,
seasonality, influenza, day of the week,
public holidays, school holidays (only
for 5-18 yr olds), pollen (only for
pediatric asthma outcome)
Season: NR
Dose-response Investigated: No
Statistical package: R statistical
software (gam procedure, mgcv
package)
Lags Considered: Lag 0 -5 days,
5-day avg (lag 0-4) for RD, and a 6-day
avg (lag 0-5) for asthma.
Outcome (ICD-9: 480-486):
Pneumonia
Age Groups: NR
Study Design: Case-crossover
N: 82,587 pneumonia hospital
admissions
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature and humidity
on the same day
Season: NR
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: Cumulative lag
period up to 2 previous days
Concentrations1
Pollutant: PM,0 (pg/m3)
Averaging Time: 24 h
Mean(SD): 24(14)
Median: 21
IQR: 16-29
99th percentile: 72
Monitoring Stations: 1
Copollutant (correlation):
NCtot:r = 0.39
NC100:r = 0.28
NCa12:r = 0.02
Nca23:r = -0.12
NCa57:r = 0.45
Nca212:r = 0.63
PM25:r = 0.80
CO: r = 0.37
N02: r = 0.35
:r = 0.32
curbside:r = 0.18
03:r = -0.21
Other variables:
Temperature: r = 0.12
Relative humidity: r = 0.05
Pollutant: PM10
Averaging Time: 24 h
Mean (min-max):
77.01 (16.7-232)
Percentiles: 25%: 42.12
50%: 75.27
75%: 104.65
Monitoring Stations: 6
Copollutant: NR
Effect Estimates (95% Cl)
PM Increment: 13 pg/m3 3 (IQR)
Relative risk (RR) Estimate [Cl]:
RD hospital admissions (6 day avg,
lag 0 -4), age 66+: One-pollutant
model: 1.06 [1.02-1. 09]
Adj for NCtot: 1.05 [1.01-1. 10]
Adj for NCa212: 1.04 [0.98-1. 11]
Asthma hospital admissions (6-day
avg lag 0-6), age 6 - 18: One-pollutant
model: 1.02 [0.93-1. 12]
Adj for NCtot: 1.01 [0.91-1.12]
Adj for NCa212: 0.94 [0.81-1.09]
Estimates for individual day lags
reported only in Fig form (see notes):
Notes: Fig 2: Relative risks and 95%
confidence intervals per IQR in single
day concentration (0-5 day lag).
Summary of Fig 2: RD: Positive,
statistically or marginally significant
associations at Lag 2-5. Asthma: Wide
confidence intervals make interpretation
difficult. Positive associations at Lag 1,
2, 3, and 5.
PM Increment: 62.53 pg/m3 (IQR)
OR Estimate [Cl]: Single Pollutant
Model: Temp>25(1C: 1.21 [1.15,1.28]
Temp <25°C: 1.57 [1.50,1. 65]
Two-Pollutant Model: Temp>25°C
Adj. for S02: 1.21 [1.14,1.28]
Adj. for N02: 1.15 [1.07,1. 24]
Adj. for CO: 1.10 [1.03, 1.1 7]
Adj. for 03: 0.96 [0.89,1. 03]
Temp < 25°C
Adj. for S02: 1.56 [1.48,1.65]
Adj. for N02: 1.09 [1.02,1. 16]
Adj. for CO: 1.30 [1.22, 1.39]
Adj. for 03: 1.56 [1.48,1. 65]
December 2009
E-230
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Chimonas and Gessner
(2007, 0932611
Period of Study: Jan 1999-Jun 2003
Location: Anchorage, Alaska
Outcome (ICD-9): Asthma (493.0-
493.9); Lower respiratory illness-LRI
(466.1,466.0,480-487,490,510-511);
Inhaled quick-relief medication; Steroid
medication
Age Groups: <20 yr old
Study Design: Time series
N: 42,667 admissions
Statistical Analyses: GEE for
multivariable modeling
Covariates: Season, serial correlation,
yr, weekend, temperature, precipitation,
and wind speed
Season: NR
Dose-response Investigated? No
Statistical Package: SPSS (dataset),
SAS (analysis)
Lags Considered: 1 day and 1 week
Pollutant: PM,0
Averaging Time: 24 h and 1 wk
Mean (min-max):
Daily: 27.6 (2-421)
Weekly: 25.3 (5.0-116.0)
Monitoring Stations: NR
Copollutant: Daily PM25
p = 0.25 (p< 0.01)
Weekly PM25
p = 0.08 (p = 0.21)
PM Increment: 10 pg/m
RR Estimate [Cl]:
Same Day
Outpatient Asthma: 1.006 [1.001,1.013]
Outpatient LRI: 1.001 [0.987,1.015]
Inpatient Asthma: 1.003 [0.922,1.091]
Inpatient LRI: 1.015 [0.978,1.053]
Inhaled Steroid Prescriptions:
1.006 [0.996,1.011]
Quick-relief Medication:
1.018 [1.006,1.030]
Weekly (median increase)
Outpatient Asthma: 1.021 [1.004,1.038]
Outpatient LRI: 1.013 [0.978,1.049]
Inpatient Asthma: 1.023 [0.948,1.104]
Inpatient LRI: 1.025 [0.981,1.072]
Inhaled Steroid Prescriptions:
0.989 [0.969,1.010]
Quick-relief Medication:
1.057 [1.037,1.077]
Reference: Chiu et al. (2008,1919891
Period of Study: 1996-2001
Location: Taipei, Taiwan
Outcome: Hospital admissions for
COPD
Study Design: Time-series
Covariates: Temperature, humidity,
PM10and03
Statistical Analysis: Poisson
regression
Statistical Package: SAS
Age Groups: All
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD) Unit:
Index Days: 111.68 + 38.32 pg/m3
Comparison Days: 55.43 + 24.66 pg/m3
Range (Min, Max): NR
Copollutant (correlation): NR
All results refer to "dust storm days" and
can be found in Table 3
Reference: Chiu et al. (2009,1902491
Period of Study: 1996-2004
Location: Taipei, Taiwan
Outcome: Hospital admissions for
pneumonia (ICD-9 480-486)
Study Design: Time-series
Covariates: Weather variables, day of
the week, seasonality, long-term time
trends
Statistical Analysis: Conditional
logistic regression
Statistical Package: SAS
Age Groups: All
Pollutant: PM,0
Averaging Time: 24 h
Mean Unit: 49.47 pg/m3
Range (Min, Max): 14.42, 234.91
Copollutant (correlation):
S02: 0.50
N02: 0.58
CO: 0.34
03: 0.31
Increment: IQR
Odds Ratio (96% Cl)
Temperatures 23° C:
Temperature < 23° C:
Adjusted for S02
Temperatures 23° C:
Temperature < 23° C:
Adjusted for N02
Temperatures 23° C:
Temperature < 23° C:
Adjusted for CO
Temperatures 23° C:
Temperature < 23° C:
Adjusted for 03
Temperatures 23° C:
Temperature < 23° C:
1.11 (1.08-1.14)
1.09(1.07-1.11)
1.10(1.08-1.13)
1.19(1.17-1.22)
0.90 (0.88-0.93)
1.09(1.07-1.12)
1.03(1.00-1.05)
1.07(1.05-1.10)
1.05(1.03-1.08)
1.09(1.07-1.11)
December 2009
E-231
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Erbas et al. (2005, 0738491 Hospital Admissions
Period of Study: Jan 2000-Dec 2001
Location: Melbourne, Australia
Outcome (ICD-10): Asthma (J45, J46)
Age Groups: 1-1 Syr
Study Design: Time series
N: 8955 asthma cases
Statistical Analyses: GAM, GEE (if
autocorrelation was present in
residuals)
Covariates: Temp and humidity
Season: NR
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0, 1, 2 days
Pollutant: PM,0
Averaging Time: 1 h
Mean (SD):
Western: 2.99 (2.11)
10th percentile: 13.67
90th percentile: 48.00
Inner Melbourne: 4.54 (2.65)
10th percentile: 15.63
90th percentile: 59.73
South/Southeastern: 1.13 (1.18)
10th percentile: 12.00
90th percentile: 36.05
Eastern: 3.61 (2.39)
10th percentile: 16.00
90th percentile: 51.05
Combined: 30.07 (10.55-112.33)
SD= 15.27
10th percentile: 16.00
90th percentile: 50.51
Monitoring Stations: Data obtained
from an air quality simulation model
(TAPM) by CSIRO Atmospheric
Research
Copollutant: NR
PM Increment: Increase from 10th to
90th percentile
RR Estimate [Cl]:
Same day lag
Western: NR
Inner Melbourne: 1.17 [1.05,1.31]
South/Southeastern: 1.14 [0.95,1.33]
Eastern: 1.09 [1.01,1.18]
Notes: All other lags NR
Reference: Farhat et al. (2005,
0894611
Period of Study: Aug 1996-Aug 1997
Location: Sao Paulo, Brazil
Hospital Admissions and Emergency
Room Visits
Outcome (ICD-9): Lower respiratory
tract diseases (466, 480-519) including
pneumonia or bronchopneumonia (480-
486), asthma (493), bronchiolitis (466)
Age Groups: <13yr
Study Design: Time series
N: 43,635
Statistical Analyses: GAM, Poisson
regression, Pearson correlation
Covariates: Time, temperature,
humidity, weekday
Season: NR
Dose-response Investigated? No
Statistical Package: S-Plus
Lags Considered: 0-7 days
Pollutant: PM,0
Averaging Time: 24 h
Mean (min-max):
62.6(25.5-186.3)
SD = 26.6
IQr = 30
N = 396
Monitoring Stations: 13
Copollutant (correlation): S02:
r = 0.69
N02:r = 0.83
03:r = 0.35
CO: r = 0.72
(all p < 0.05)
Additional correlations:
Rel humidity: r = -0.55
Mintemp: r = -0.44
(both p< 0.05)
PM Increment: 30 pg/m (IQR)
RR Estimate [Cl]:
Lower respiratory tract disease
5-day ma
Copollutant model:
N02: 2.1 [-7.1,11.3]
S02:16.5 [10.5,22.6]
03: 10.1 [5.0,15.2]
CO: 14.1 [8.1,20.2]
Multipollutant model: 5.2 [-4.6,15.1]
Pneumonia or bronchopneumonia
6-day ma
Copollutant model:
N02:14.8 [-3.8,33.4]
S02:14.8 [-0.3,30.0]; 03:16.2 [1.0,31.3]
CO: 17.6 [0.4,34.8]
Multipollutant model: 5.23 [-16.2,26.6]
Asthma or bronchiolitis
2-day ma
Copollutant model:
N02:-11.04 [-50.0,28.0]
S02:15.8 [-7.8,39.3]
03: 11.7 [-10.4, 33.9]
CO: 12.4 [-14.8,39.7]
Multipollutant model: -15.5 [-61.2,30.2]
December 2009
E-232
-------
Reference
Reference: Fung et al. (2006, 0897891
Period of Study: June 1995-Mar 99
Location: Vancouver, Canada
Reference: Fung al. (2005, 0932621
Period of Study: Nov 1995-Dec 2000
Location: London, Ontario
Design & Methods
Hospital Admission/ED
Outcome: Respiratory diseases (460-
519)
Age Groups: Age >65
Study Design: Time series
N: 26,275 individuals admitted
Statistical Analyses: Poisson
regression (spline 12 knots), case-
crossover (controls +17 days from case
date), Dewanji and Moolgavkar (DM)
method
Covariates: Long-term trends, day-of-
the-week effect, weather
Season: All yr
Dose-response Investigated? No
Statistical Package: SPIus, R
Lags Considered: 0-7 days
Hospital Admissions
Outcome (ICD-9): Asthma (493) and all
other respiratory diseases (460-519)
Age Groups: <65 yr
65+ yr
Study Design: Time series
N: 5574 respiratory admissions
Statistical Analyses: GAM with locally
weighted regression smoothers
(LOESS)
Covariates: Maximum and minimum
temp, humidity, day of the week,
seasonal cycles, secular trends
Qeacniv MP
Concentrations1
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): 13.31(6.13) pg/m3
Range (Min, Max): (3.77, 52.17)
Monitoring Stations: NR
Copollutant (correlation): PIvl-;
PM25 r = 0.80
P M r - n 11
r lvlio-2.5 1 - -U. II
CO r = 0.46
Coh r = 0.61
03 r = -0.08
N02r = 0.54
S02r = 0.61
Pollutant: PM,0
Averaging Time: 24 h
Mean (min-max):
38.0 (5-248)
SD = 23.5
Monitoring Stations: 4
Copollutant (correlation):
N02:r = 0.30
S02: r = 0.24
CO: r = 0.21
0 • r = o 53
COH: r = 0.29
Effect Estimates (95% Cl)
PM Increment: : 7.9 pg/m3
Rr Estimate (65+ Yr)
Dm Method:
1.014(0.998,1.029]
I -an 0
Lag u
1.016[0.998,1.034]
3-day avg
0.988(0.970, 1.006]
5-day avg
0.983[0.963, 1.004]
7-day avg Time Series:
1.016(0.999, 1.033]
LagO
1.015(0.996, 1.035]
3-day avg
1.009(0.987, 1.032]
5-day avg
1.009[0.983, 1.036]
7-day avg
Case-Crossover:
1.017(0.998, 1.036]
LagO
1.015(0.993, 1.037]
3-day avg 1.008(0.984, 1.033]
5-day avg
1.003(0.976, 1.031]
7-day avg
PM Increment: 26 pg/m3
% Change in Daily Admission [Cl]:
Age <65
Current day mean: -0.9 [-6.8,5.4]
2-day mean: -1.3 [-8.5,6.6]
3-day mean: 1.9 [-6.5,11]
Age 65+
Current day mean: 3.3 [-1.7,8.6]
2-day mean: 5 [-1.5,11.9]
3-day mean: 1.2 [-6.1,9.1]
Dose-response Investigated? No
Statistical Package: S-Plus
Lags Considered: Current to 3-day
December 2009
E-233
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Galan et al. (2003, 0874081
Period of Study: 1995-1998
Location: Madrid, Spain
Hospital Admissions
Outcome (ICD): Asthma (493)
Age Groups: All ages
Study Design: Time series
N: 555,1 53 at-risk
Pollutant: PM,0
Averaging Time: 24 h
Mean (min-max):
32.1 (11.2-108.6)
SD=12.1
Monitoring Stations: 13
PM Increment: 10 pg/m3
RR Estimate [Cl]:
Single-pollutant
Current-day lag: 1.011 (0.980-1.042)
1-day lag: 1.006 (0.976-1.037)
Statistical Analyses: GAM,
autoregressive Poisson regression
Covariates: Temperature, relative
humidity, pollen, yr, day of the week,
public holiday
Season: NR
Dose-response Investigated? No
Statistical Package: S-Plus
Lags Considered: 0,1, 2, 3, and 4 day
Copollutant (correlation):
S02:r = 0.581
N02:r = 0.717
03:r =-0.188
Other variables:
O.europaea: r = -0.066
Plantagosp.: r = -0.202
Poaceae: r = -0.132
Urticaceae: r = -0.104
Temp: r = -0.122
Humidity: r = 0.119
2-day lag: 1.008 (0.978-1.038)
3-day lag: 1.039 (1.010-1.068)
4-day lag: 1.027 (0.999-1.056)
Adjustment for pollen (PM10 3-day lag)
O.europaea: 1.041 (1.011-1.071)
Plantagosp.: 1.046 (1.017-1.076)
Poaceae: 1.043 (1.015-1.073)
Urticaceae: 1.038 (1.009-1.068)
All four: 1.045 (1.016-1.074)
Reference: Hajat et al. (2002, 0303581
Period of Study: Jan 1992-Dec 1994
Location: London, England
Family Practice consultations
Outcome: Upper Resp Disease
(excluding allergic rhinitis) (460-3),
(465), (470-5), (478)
Age Groups: 0-14,
15-64, >65yr
Study Design: Time series
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 28.5 (13.7) pg/m3
Percentiles: 10th: 15.8
90th: 46.5
Monitoring Stations: 1
N: 268,718-295,740 registered patients Copollutant: NR
Statistical Analyses: Poisson
regression, GAM, LOESS smoothers,
default convergence criteria
Covariates: Long term trends, pollen
counts, flu, meteorological variables
Season: All yr
Dose-response Investigated? No
Statistical Package: SPLUS
Lags Considered: 2-3
PM Increment: All Year: 18
Warm Season: 15
Cold Season: 20
% Change, Single Pollutant Models:
All Year: Ages
0-14:2.0[-0.2, 4.2] Lag 3
Ages 15-64: 5.7[2.9, 8.6]
Lag 2
Ages>65:10.2[5.3,15.3] Lag 2
Warm Season: Ages 0-14:1.1[-2.4, 4.8]
Lag3
Ages 15-64: 6.0[2.7, 9.4]
Lag 2
Ages >65: 0.1 [-7.7, 8.5] Lag 2
Cold Season: Ages 0-14: 2.7[-0.1, 5.5]
Lag3
Ages 15-64: 3.6[1.0, 6.4]
Lag 2
Ages>65:18.9[11.7, 26.7] Lag 2
% Change, 2 Pollutant Models:
0-14 Yr
PM10w/N02:3.8[1.6, 6.1]
PMi0w/03:1.8[-0.4, 3.9]
PM10w/S02:2.0[-0.6, 4.6]
15-65Yr
PMi0w/N02:2.8[0.7, 4.9]
PMi0w/03:4.8[2.6, 7.0]
PMi0w/S02:4.8[2.2, 7.5]
>65Yr
PM10w/N02:4.6[0.5, 8.8]
PM10w/03:10.7[5.7, 16.0]
PM10w/S02:10.6[4.5, 17.1]
December 2009
E-234
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Hanigan et al. (2008,
1565181
Period of Study: 1996-2005 (Apr-Nov
of each yr)
Location: Darwin, Australia
Outcome: Cardiorespiratory Disease
HA (ICD 9: 390-519
ICD10:IOO-99SJOO-99)
Age Groups: NR
Study Design: Time series
N: 8279 events
Statistical Analyses: Poisson
regression
Covariates: Indigenous status,
Dose-response Investigated? No
Statistical Package: R
Lags Considered: Lags 0-3
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 21.2 (8.2)
Range: 55.2
Monitoring Stations: 2 (monitored &
modeled)
Copollutant: NR
Co-pollutant Correlation
N/A
PM Increment: 10 pg/m
Percent Change (Lower Cl, Upper Cl),
lag:
Total Respiratory: 4.81 (-1.04,11.01),
lagO
Total Resp., Indigenous: 9.40 (1.04,
18.46), lag 0
Total Resp., Non-Indigenous: 3.14 (-
2.99, 9.66), lag
Resp. Infection, Indigenous: 15.02
(3.73, 27.54), lag 3
Resp. Infection, Non-Indigenous: 0.67 (-
7.55, 9.61), Iag3
Asthma Indigenous: 16.27 (3.55,
40.17), lag 1
Asthma Non-Indigenous: 8.54 (-5.60,
24.80), lag 1
*Fig 3. percent change in hospital
admissions per 10 pg/m3 increase in
PM10
Reference: Hanigan et al. (2008,
1565181
Period of Study: 1996-2005 (Apr-Nov
of each yr)
Location: Darwin, Australia
Hospital Admissions/ED visits
Outcome (ICD-9 or ICD-10):
Daily emergency hospital admissions
for total respiratory (ICD-9: 460-519
ICD-10: JOO-J99), asthma (ICD-9: 493
ICD-10:J45-J47), COPD (ICD-9:
490-492, 494-496
ICD-10: J40-J44, J47, J67), and
respiratory infections (ICD-9: 461-466,
480-487, 514
ICD-10: JOO-J22).
Age Groups Analyzed: All
Study Design: Time series
N: 8,279 hospital admissions
Statistical Analyses: Poisson
generalized linear models
Covariates: Indigenous status, time in
days, temperature, relative humidity,
day of the week, influenza epidemics,
change between ICD editions, holidays,
yrly population
Season: Apr-Nov (corresponding to the
dry season)
Dose-response Investigated? No
Statistical package: R version 2.3.1
Lags Considered: Lag 0 -3
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD range): 21.2 (8.2-55.2)
Monitoring Stations: N/A (see notes)
Copollutant (correlation): NR
PM Increment: 10 pg/m
Percent change [96% Cl]:
Overall respiratory disease:
Lag 0:4.81 [-1.04,11.01]
Lag 0 (indigenous people):
9.40 [1.04, 18.46]
Lag 0 (non-indigenous people):
3.14[-2.99,9.66]
In unstratified analyses, the subgroups
of respiratory infections, asthma, and
COPD all had positive associations with
PM10 Lag 0.
Asthma:
Lag 1 (indigenous people):
16.27 [-3.55, 40.17]
Lag 1 (non-indigenous people):
8.54
[-5.60, 24.80]
Respiratory infections:
Lag 3 (indigenous people):
15.02 [3.73, 27.54]
Lag 3 (non-indigenous people):
0.67 [-7.55, 9.61]
Notes:
Fig 3: Associations between
hospitalizations for non-indigenous and
indigenous people with estimated
ambient PMi0.
Summary of Fig 3: Confidence
intervals were wide, but indigenous
people generally had stronger
associations with PMi0 than non-
indigenous people. Daily PMi0 exposure
levels were estimated for the population
of the city from visibility data using a
previously validated models.
December 2009
E-235
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Hapcioglu et al. (2006,
0932631
Period of Study: Jan 1997-Dec 2001
Location: Istanbul, Turkey
Hospital Admissions
Outcome (ICD-9): COPD (ICD: NR)
Age Groups: NR
Study Design: Time series
N: 1586 patients
Statistical Analyses: Multiple stepwise
regression, Pearson correlation
Covariates: Humidity, temperature, and
pressure
Season: Summer, fall, winter, spring
Dose-response Investigated? No
Statistical Package: SPSS
Lags Considered: NR
Pollutant: PM,0
Averaging Time: 1 mo
Mean (SD): NR
Monitoring Stations: 1
Copollutant: NR
Correlation with COPD:
r = 0.28
p = 0.03
Adjfortemp:r = 0.16
p = 0.23
PM Increment: NR
Notes: RRs only provided for season,
notPM
Reference: Hwang and Chan (2002,
0232221
Period of Study: 1998
Location: Taiwan
Clinic visits
Outcome: LRI
466, 480-486 (acute bronchitis, acute
bronchiolitis, pneumonia)
Age Groups: 0-14 yr, 15-64, 65+ yr
Study Design: Cluster analysis of small
study areas
N: 50 communities
Statistical Analyses: GLM to model
temporal patterns, hierarchical model to
obtain estimates across 50 communities
Covariates: Day of week, temperature,
dew point, summer/Winter
Season: All
Dose-response Investigated? Yes
Statistical Package: NR
Lags Considered: 0-2
Pollutant: PM10
Averaging Time: 24 h
Mean (SD): 58.9 pg/m3 (14.0)
Range (Min, Max): 33.3, 83.1 pg/m3
PM Component:
Monitoring Stations: 59
Notes: Number Of stations estimated
from fig.
Copollutant: NR
PM Increment:
10% Increase In PM,0 (5.9 pg/m
Percent Change:
0-14
0.5%(-0.1,0.8]LagO
[-0.3, 0.3] Lag1
0.3 [0.0, 0.6] Lag2
15-64
0.6
0.2
0.2, 0.9] LagO
-0.1,0.5] Lag 1
0.3 [0.0, 0.6] Lag2
65+
0.8 [0.4,1.1] LagO
0.3 [-0.1, 0.6] Lag 1
0.5[0.1,0.8]Lag2
All Ages
0.5 [0.2, N0.8] LagO
[-0.3, 0.3] Lag1
0.3 [0.0, 0.6] Lag2
Reference: Jaffe et al. (2003, 0419571
Period of Study: July 1991 -June
1996
Location: Cincinnati, Cleveland,
Columbus, Ohio
ED visits
Outcome (ICD10): Asthma (493)
Age Groups: Age 5-34 yr
Study Design: Time-series
N: 4,416 recipients
Statistical Analyses: Poisson
regression, GAM
Covariates: City, day of week, wk, yr,
minimum temperature, dispersion
parameter
Season: Jun-Aug only
Dose-response Investigated? Yes
Statistical Package: NR
Lags Considered: 0-3 days
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD):
Cincinnati: 43.0(16.4)
Cleveland: 60.8(28.4)
Columbus: 37.4(16.3)
Range (Min, Max):
Cincinnati: (16,90)
Cleveland: (12,183)
Columbus: (7,87)
Monitoring Stations: 3
Copollutant (correlation):
Cincinnati:
PM10
03r = 0.42
N02r = 0.36
S02r = 0.31
Cleveland:
PM10
03r = 0.42
N02r = 0.34
S02r = 0.29
Columbus:
PM10
03r = 0.51
N02r=Na
S02r = 0.42
PM Increment: 50 pg/m
% Change
Asthma
Cincinnati:-22%[-49,-19] Lag 3
Cleveland: 12%[0,27] Lag 2
Columbus: 32%[-6,-85] Lag 3
Ar Estimate [Lower Ci, Upper Ci]
Lag:
Asthma
Cincinnati: PMi0: Nr
Cleveland: PM,0:1.32
Columbus: PM,0: 3.62
Notes: Dose response was investigated
by assessing the relationship between
odds of ed visit by quintile of PMi0.
Results are displayed in Fig. "no
consistent effects for all three cities
were observed for PMi0." Rate ratios
were also reported for each city.
December 2009
E-236
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Jalaludin et al. (2004,
0565951
Period of Study: Feb-Dec 1994
Location: Sydney, Australia
Doctor Visits
Outcome (ICD- NR): Respiratory
symptoms (wheeze, dry cough, and wet
cough), asthma medication use, and
doctor visits for asthma
Age Groups: Primary school children
Study Design: Longitudinal cohort
study
N: 125 children
Statistical Analyses: GEE logistic
regression models
Covariates: Temperature, humidity,
daily pollen count, daily alternaria count,
number of h spend outdoors, season
Season: Fall (Feb-Apr), winter (May-
Aug), spring/summer (Sep-Dec)
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-2 days
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 22.8 (13.8)
Monitoring Stations: 4
Copollutant (correlation):
03:r = 0.13
N02:r = 0.26
Other variables:
Temp: r = 0.04
Humidity: r = -0.29
Total pollen: r = 0.04
Alternaria: r = 0.04
PM Increment: IQR (pg/m )
Same day: 12.0
1-day lag: 12.02
2-day lag: 12.25
2-day avg: 11.15
5-day avg: 10.23
OR Estimate [Cl]:
Doctor Visits for Asthma
Same day: 1.11 [1.04,1.19]
1-day lag: 1.10 [1.02,1.19]
2-day lag: 1.15 [1.06,1.24]
2-day avg: 1.11 [1.03,1.20]
5-day avg: 1.14 [0.98,1.31]
Prevalence of Doctor Visits for
Asthma:
Quartile 1:0.50 (mean PM = 12.4)
Quartile 2: 0.38 (mean PM = 17.2)
Quartile 3: 0.65 (mean PM = 23.0)
Quartile 4: 0.63 (mean PM = 38.3)
Notes: ORs and prevalence are also
provided for wheeze, dry cough, wet
cough, inhaled |32-agonist use, and
inhaled corticosteroid use. None were
statistically significant.
December 2009
E-237
-------
Reference
Design & Methods
Reference: Johnston et al. (2007, Hospital Admissions/ED visits
Outcome (ICD-10):
Period of Study: 2000, 2004, 2005
Apr-Nov of each yr) Al1 respiratory conditions (JOO-J99),
Location: Darwin, Australia
including asthma (J45-46), COPD
(J40-J44
(JOO-J22
, and respiratory infections
Age Groups Analyzed: All
Study Design: Case-crossover
N: 2466 emergency admissions
Concentrations1
Pollutant: PM,0
Averaging Time: 24 h
Median: 17.4
IQR: 13.6-22.3
10-90th Percentile: 10.3-27.7
Range: 1.1-70.0
Monitoring Stations: 1
Copollutant (correlation): NR
Effect Estimates (95% Cl)
PM Increment: 10 pg/m3
OR Estimate [96% Cl]: All respiratory
conditions: Lag 0:1. 08 [0.98-1. 18]
Lag 0 (indigenous): 1.17 [0.98-1. 40]
COPD: Lag 0:1. 21 [1.0-1.47]
Lag 0 (indigenous): 1.98 [1.10-3.59]
Asthma: Lag 0:1. 14 [0.90-1. 44]
Asthma* COPD: Lag 0:1. 19
[1.03-1.38]
Statistical Analyses: Conditional
logistic regression
Covariates: Weekly influenza rates,
temperature, humidity, days with rainfall
>5mm, public holidays, school holiday
periods (for respiratory conditions only)
Season: Apr-Nov (dry season)
Dose-response Investigated? No
Statistical package: NR
Lags Considered: 0-3 days
Notes: Fig 1: Adjusted OR and 95% Cl
for hospital admissions for all
respiratory conditions per 10 pg/m rise
in PM10 for the same day and lags up to
3 days, overall and stratified by
indigenous status.
Summary of Fig 1 results: Marginally
significant positive association at Lag 0
in overall study population. Larger
marginally significant positive
association among indigenous people.
Fig 2: OR and 95% Cl for hospital
admissions for COPD. Summary of Fig
2 results: Marginally significant positive
associations at Lag 0 and Lag 1 in
overall study population and among
non-indigenous people. Large,
statistically significant positive
association at Lag 0 for indigenous
people, with smaller, non-significant
positive associations at Lag 1 and Lag2.
Fig 3: OR and 95% Cl for hospital
admissions for asthma.
Summary of Fig 3 results: Positive,
non-significant (sometime marginally
significant) associations at Lag 0, Lag 2,
and Lag 3 for overall population and
indigenous status strata.
Fig 4: OR and 95% Cl for hospital
admissions for respiratory infections.
Summary of Fig 4 results: Negative
associations at Lag 2 and Lag 3 in all
population strata.
December 2009
E-238
-------
Reference
Reference: Kim et al. (2007, 0928371
Period of Study: 2002
Location: Seoul, Korea
Reference: Ko et al. (2007, 0916391
Period of Study: Jan 2000-Dec 2004
Location: Hong Kong, China
Design & Methods
Ed Visits
Outcome (ICD10): Asthma (J45), (J46)
Age Groups: All Ages
Study Design: Cass-Crossover
N: 92, 535 Visits
Statistical Analyses: Conditional
Logistic Regression, Relative Effect
Modification (Rem)
Covariates: Time Trend, Season, Daily
Mean Temperature, Relative Humidity,
Air Pressure. Sep As Modifier Of Air
Pollution Asthma Visit Association.
Season: All Year
Dose-response Investigated? No
Statistical Package: Nr
Lags Considered: 0-2 days
Ed Visits
Outcome (ICD-9): COPD: chronic
bronchitis (491), emphysema (492),
chronic airway obstruction (496)
Age Groups: All Ages
Study Design: Time Series
N: 15 hospitals, 119,225 admissions
Statistical Analyses: Poisson
regression, gam with stringent
convergence criteria, aphea2 protocol.
Covariates: Time trend, season,
temperature, humidity, other cyclical
factors, day, day of wk, holidays
Season: All yr, interactions with season
tested
Dose-response Investigated? No
Statistical Package: Splus 4.0
Lags Considered: 0-5 days
Concentrations1
Pollutant: PM,0
Averaging Time: 8 h
Mean (SD): Daily Concentration: 67.6
(39.0) pg/m3
Relevant Exposure Term (Difference
Between Concentration On Event Day
And Mean Of Concentrations On
Control Days): 26.0 (19.7)
Percentiles: SOth(Median): Daily
Concentration: 61.9
Relevant Exposure Term: 21.6
Range (Min, Max): Daily
Concentration: (4.9, 302.0)
Relevant Exposure Term: (0.0, 143.1)
Monitoring Stations: 3
Copollutant: Nr
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 50.1 (23.9) pg/m3
Percentiles: 25th: 31. 9
SOth(Median): 44.5
75th: 64.1
Range (Min, Max): (13.6, 172.2)
Monitoring Stations: 14 Stations
Copollutant (correlation):
PM10:
S02r = 0.436
N02r = 0.229
03r = 0.421
PM25 r = 0.952
Effect Estimates (95% Cl)
PM Increment: 47.4 pg/m3
Rr Estimate For Asthma (Stratified
By Sep):
Individual Level Sep:
Quintile1-1.06[1.02 1.09]
Quintile2-1.07[1.04 1.10]
Quintile3-1.06[1.03 1.10]
Quintile 4-1 .03[0.99 1.07]
Quintile5-1.10[1.05 1.14]
Regional Level Sep:
Quintile 1-1 .04[0.99 1.10]
Quintile2-1.03[1.00 1.07]
Quintile3-1.05[1.03 1.08]
Quintile4-1.06[1.02 1.10]
Quintile5-1.09[1.06, 1.13]
Total-1.06[1.04, 1.08], 3D Ma
Notes: Relative Effect Modification
(Rem) Estimates Presented In Paper.
PM Increment: 10 pg/m3
Rr Estimate
COPD:
1.0031.000,1.005 LagO
1.005 1.002, 1.007 Lag 1
1.0101.007, 1.012 Lag 2
1.011 1.008, 1.013] Lag 3
1.0081.006, 1.011] Lag 4
1.0071.004, 1.009 Lag5
1.0051.002, 1.008 Lag 0-1
1.011 1.008, 1.014] Lag 0-2
1.0161.013,1.019 Lag 0-3
1.0201.017, 1.024 Lag 0-4
1.0241.021, 1.028 Lag 0-5
December 2009
E-239
-------
Reference
Reference: Ko et al. (2007, 0916391
Period of Study:
Jan 2000-Dec 2004
Location: Hong Kong, China
Design & Methods
Hospital Admission
Outcome (ICD-9): Asthma (493)
Age Groups: All, 0-14, 15-56,65+
Study Design: Time series
N: 69,716 admissions, 15 hospitals
Statistical Analyses: Poisson
regression, with GAM with stringent
convergence criteria.
Covariates: Time trend, season,
temperature, humidity, other cyclical
factors
Season: All yr, evaluated effect of
season in analysis
Dose-response Investigated? No
Concentrations1
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 52.5(27. 1) pg/m3
Percentiles: 26th: 30.9
60th(Median):
47 1
HI . i
75th: 68.8
Range (Min, Max): (13.4, 198.9)
Monitoring Stations: 14 stations
Copollutant (correlation):
PM10:
S02 r = 0.436
N02 r = 0.761
03r = 0.600
PM25 r = 0.956
Effect Estimates (95% Cl)
PM Increment: 10.0 pg/m3
RR Estimate:
Asthma (Single-pollutant model):
1.0061.003, 1.010 lagO
1.0051.002, 1.009 Iag1
1.0051.002,1.009 lag 2
1.0081.005, 1.012 Iag3
1.0061.002, 1.009 lag 4
1.0060.999,1.006 lag 5
1.008 1.004, 1.012 ; lag 0-1
1.0121.008, 1.016 lag 0-2
1.0151.011, 1.019] lag 0-3
1.0181.013, 1.022 lag 0-4
1.0191.015, 1.024 lag 0-5
Asthma by age group
0-14: 1.023(1.015, 1.031] lag 0-5
14-65:1.014(1.006, 1.022] lag 0-5
*,ŁŁ• -i msM nno 1 moi i^n n_/i
Statistical Package: SPLUS 4.0
Lags Considered: 0-5 days
Asthma-Effect of season: 1.148(1.051,
1.245] lag 0-5
Reference: Kuo et al. (2002, 0363101
Period of Study: 1 yr
Location: central Taiwan
Hospital Admissions
Outcome (ICD-NR): Asthma
Age Groups: 13-1 6 yr
Study Design: Cohort
N: 12,926
Statistical Analyses: Multiple logistic
regression, Pearson correlation
Pollutant: PM,0
Averaging Time: 1 h
Mean (min-max): NR
Range: (54.1-84.3)
Monitoring Stations: 8
Copollutant: Values NR
Notes: Author states that a positive
PM Increment: NR
OR Estimate:
PM10 <65.9 |jg/m3-referent
PM10 >65.9 pg/m3
Crude OR: 0.837
Adj OR: 0.947
95% Cl: (0.640, 1.401)
_l
Covariates: Sex, age, residential area
level of parents' education, number of
cigarettes smoked by smokers in the
family, incense burning, frequency of
physical activity
Season: NR
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: NR
correlation was found between N02 and
PM10
Reference: Langley-Turnbaugh et al.
(2005, 0932691
Period of Study: 2000-2001
Location: Portland, Bridgeton, and
Presque Isle, Maine
Hospital Admissions
Outcome (ICD-9): Asthma (493xx)
Age Groups: 0-18 yr, 19+ yr
Study Design: Time series
N:NR
Statistical Analyses: NR
Covariates: NR
Season: Winter, spring, summer, fall
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: NR
Notes: Hospital admissions were used
to determine seasonality of asthma
admissions so that PM components
from those time periods could be
analyzed
Pollutant: PM10
Averaging Time: NR
Mean (min-max): NR
Monitoring Stations: NR
Copollutant: NR
PM Increment: NR
RR Estimate [Cl]: NR
Notes: Portland filters contained more
PM in the winter (Jan) and Bridgeton
filters contained more PM in the spring
(May)
study analyzed metal components of
PM,o(Mn, Cu, Pb.As, V, Ni.AI)
Clinical data shows a strong peak in fall
and weaker peaks in Jan and May for
asthma admissions
December 2009
E-240
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Lee et al. (2002, 0348261
Period of Study: Dec 1997-Dec 1999
Location: Seoul, Korea
Hospital Admissions
Outcome (ICD10): Asthma, J45, J46,
Age Groups: Children <15 yr
Study Design: Time-Series
N: 822 days, 6,436 admissions
Statistical Analyses: Poisson
regression, GAM, LOESS smoothers.
Covariates: Days of the week,
temperature, humidity
Season: All
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0-5, 0-1 ma for 1-2,
2-3, and 3-4 days
Pollutant: PM,0
Averaging Time: 24 h
Mean(SD):64.0(31.8)|jg/m3
Percentiles: 26th: 40.5 pg/m3
50th(Median):59.1|jg/m3
76th: 80.9 pg/m3
Range (Min, Max): NR
Monitoring Stations: 27
Notes: Copollutant (correlation):
PMio-S02: 0.585
PM10-N02: 0.738
PM10-03: 0.106
PM10-CO: 0.598
PM Increment: IQR: 40.4 pg/m
RR Estimate:
Single Pollutant:
1.07(1.04,1.11) lag 1
Two pollutant models:
+S02:1.05 (1.01,1.09) lag 1
+N02:1.03 (0.99,1.07) lag 1
+03:1.06 (1.03,1.10)lag1
+00:1.04(1.00,1.08) lag 1
Three pollutant models:
+03 +CO: 1.02 (0.98,1.06), lag 1
Four pollutant models:
+03 + CO+S02:1.02(0.98, 1.06), lag 1
Five pollutant model:
1.016(0.975, 1.059) lag 1
Notes: Investigated the association
between outdoor air pollution and
asthma attacks in children <15 yr.
Reference: Lee et al. (2006, 0901761
Period of Study: Jan 1997-Dec 2002
Location: Hong Kong, China
Hospital Admission
Outcome: Asthma (493)
Age Groups: <18yr
Study Design: Time series
N: 26,663 asthma admissions for
asthma and 5821 admissions for
influenza
Statistical Analyses: Poisson
regression, GAM
Covariates: Temperature, atmospheric
pressure, relative humidity
Season: All
Dose-response Investigated? No
Statistical Package: SAS 8.02
Lags Considered: 0-5
Notes: Controls were admissions for
influenza ICD9 487
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 56.1 (24.2)
Percentiles: 26th: 37.3
50th(Median):51.1
75th: 70.7
Monitoring Stations: 10
Notes: Copollutant (correlation):
PM10-PM25: 0.90
PM,o-S02: 0.39
PM,o-N02: 0.80
PM10-03: 0.60
PM Increment: IQr = 33.4
Percent Increase:
Single pollutant model:
4.97 [2.96, 7.03], lag 0
5.71 [3.78, 7.68], lag 1
6.40 [4.51, 8.32], lag 2
7.25 [5.38, 9.16], lag 3
7.45 [5.58, 9.35], lag 4
5.96 [4.11, 7.85], Iag5
Multipollutant model (S02, CO, N02, 03)
3.67 [1.52,5.86] Iag4
Reference: Lin et al. (2005, 0878281
Period of Study: 1998-2001
Location: Toronto, North York, East
York, Etobicoke, Scarborough, and York
(Canada)
Hospital Admissions
Outcome (ICD-9): Respiratory
infections including laryngitis, tracheitis,
bronchitis, bronchiolitis, pneumonia,
and influenza (464, 466, 480-487)
Age Groups: 0-14 yr
Study Design: Bidirectional case-
crossover
N: 6782 respiratory infection
hospital izations
Statistical Analyses: Conditional
logistic regression (Cox proportional
hazards model)
Covariates: Daily mean temp and dew
point temp
Season: NR
Dose-response Investigated? No
Statistical Package: SAS 8.2 PHREG
procedure
Lags Considered: 1-7 day avg
Pollutant: PM10
Averaging Time: 24 h
Mean (min-max):
20.41 (4.00-73.00)
SD= 10.14
Monitoring Stations: 4
Copollutant (correlation):
PM25:r = 0.87
PMi0.25:r = 0.76
CO: r = 0.10
S02:r = 0.48
N02:r = 0.54
03:r = 0.54
PM Increment: 12.5 pg/m
OR Estimate [Cl]:
Adjusted for weather
4-day avg: 1.22 [1.10,1.34]
6-day avg: 1.25 [1.11,1.40]
Adj for weather and other gaseous
pollutants
4-day avg: 1.14 [0.99,1.32]
6-day avg: 1.20 [1.01,1.42]
Notes: OR's were also categorized into
"Boys" and "Girls," yielding similar
results
December 2009
E-241
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Lin et al, (2008,1268121
Period of Study: 1991-2001
Location: New York State, U.S.
Outcome: Respiratory hospital
admissions (ICD-9 466, 490-493, 496
Study Design: Time-series
Covariates: Demographic
characteristics, PM10, meteorological
conditions, day of the week,
seasonality, long term trends and
different lag periods
Statistical Analysis: GAM and case-
crossover design at the regional level
and Bayesian hierarchical model at the
state level
Age Groups: Children 0-17 yr
Pollutant: 03 (PMi0 is secondary)
Averaging Time: 24 h
Mean (SD) Unit: 19.56 (10.92) pg/m3
Range (Min, Max): 1.0, 90.00
Copollutant (correlation):
Given in Fig 3
All PMio results are given in Fig 3
Reference: Lin et al. (2002, 0260671
Period of Study: Jan 1981-Dec 1993
Location: Toronto
Hospital Admissions
Outcome (ICD-9): Asthma (493)
Age Groups: 6-12 yr
Study Design: Uni- and bi-directional
case-crossover (UCC, BCC) and time-
series (TS)
N: 7,319 asthma admissions
Statistical Analyses: Conditional
logistic regression, GAM
Covariates: Maximum and minimum
temp, avg relative humidity
Season: Apr-Sep, Oct-Mar
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 1-7 day avg
Pollutant: PM10
Averaging Time: 6 days (predicted
daily values)
Mean (min-max):
30.16(3.03-116.20)
SD= 13.61
Monitoring Stations: 1
Copollutant (correlation):
PM25:r = 0.87
PM10.25:r = 0.83
CO: r = 0.38
S02:r = 0.44
N02:r = 0.52
03:r = 0.44
PM Increment: 14.8 pg/m
RR Estimate [CI]:
Adj for weather and gaseous pollutants
BCC 5-day avg: 0.99 [0.90,1.09]
BCC 6-day avg: 1.01 [0.90,1.12]
TS 5-day avg: 1.03 [0.95,1.11
TS 6-day avg: 1.02 [0.94,1.11
Boys-adj for weather
1.04,1.17
1.02,1.17
0.98,1.09
0.95,1.08:
TS 1-day avg:"1.03 [0:99,1.07] '
TS 2-day avg: 1.01 [0.96,1.05]
Girls-adj for weather
UCC 1-day avg: 1.10
UCC 2-day avg: 1.10
BCC 1-day avg: 1.04
BCC 2-day avg: 1.01
UCC 1-day avg: 1.07
UCC 2-day avg: 1.15
BCC 1-day avg: 0.99
BCC 2-day avg: 1.03
0.99,1.16
1.04,1.26
0.92,1.06'
0.95,1.12:
TS1-day avg: 0.99 [0.94,1.04]
TS 2-day avg: 1.02 [0.96,1.08]
Notes: The author also provides RR
using UCC, BCC, and TS analysis for
female and male groups for days 3-7,
yielding similar results
Reference: Linares et al. (2006,
0928461
Period of Study: Jan 1995-Dec 2000
Location: Madrid, Spain
Outcome: Respiratory system diseases
460-519, bronchitis 460-496,
pneumonia 480-487
Age Groups: <10yr
Study Design: Time series
N: -15,000 admissions, 2192 days
Statistical Analyses: Poisson
regression, dummy variables to adjust
for season and weather
Covariates: Temperature, difference in
barometric pressure, relative humidity,
pollen counts, influenza epidemics
Season: All
Dose-response Investigated? Yes
Statistical Package: S-Plus 2000
Lags Considered: 0-13
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 33.4 pg/m3, (13.7)
Range (Min, Max): 6,109 pg/m3
Monitoring Stations: 24
Notes: Copollutant (correlation):
PM,o-S02: 0.532
PMio-03: -0.289
PM10-: 0.721
PM10-N02: 0.711
PM Increment: 10 pg/m
RR Estimate
Bronchitis
1.09 [1.01,1.16] lag 2
AR% Estimate
Bronchitis
7.9 [Cl NR] Iag2
Notes: Only statistically significant
relative and attributable risks were
presented by the authors.
The authors conducted multivariate
modeling using a linear term to
represent PM10. They also report an
apparent estimated PMio effect
threshold of 60 pg/m3, based on
examination of a scatter plot of
respiratory emergency hospital
admissions and PMio levels.
December 2009
E-242
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Luginaah, et al. (2005,
0573271
Period of Study: Apr 1995-Dec 2000
Location: Windsor, Ontario, Canada
Hospital Admission/ED:
admission
Outcome: All respiratory: 460-519
Age Groups: All, 0-14,15-64, and >65
Study Design: Times-series, bi-
directional case-crossover
N: 4214 admissions
Statistical Analyses: Poisson
regression, GAM w/ stringent
convergence criteria or natural splines,
conditional logistic regression
Covariates: Age, sex
Maximum & minimum temperature,
change in barometric pressure from
previous day
Season: All
Dose-response Investigated? No
Statistical Package: S-Plus
Lags Considered: 1-3
Pollutant: PM,0
Averaging Time: 24-h max
Mean (SD): 50.6 ,(35.5)
Range (Min, Max): 9, 349
Monitoring Stations: 4
Notes: Copollutant (correlation):
PM10-N02: 0.33
PM10-S02: 0.22
PM,o-CO: 0.21
PM,o-03: 0.33
PM Increment: Interquartile range
(75th-25th) 31 pg/m3
RR Estimates (Time Series)
All Age Groups Females
0.996 [0.950, 1.044], lag 1
1.015 [0.963, 1.069], lag 2
1.022 [0.968, 1.078], lag 3
All Age Groups Males
1.008 [0.965, 1.054], lag 1
1.036 [0.986, 1.089], lag 2
1.027 [0.974, 1.083], lag 3
RR Estimates (Case Crossover)
All Age Groups Females
1.034 [0.974, 1.098], lag 1
1.045 [0.972, 1.124], lag 2
1.054 [0.970, 1.145], lag 3
All Age Groups Males
0.997 [0.942, 1.056], lag 1
1.022 [0.953, 1.097], lag 2
1.008 [0.930, 1.092], lag 3
Notes: Results, stratified by age group
available in manuscript.
Reference: Martins et al. (2002,
0350591
Period of Study: May 1996-Sep 1998
Location: Sao Paulo, Brazil
Hospital Admission/ED:
ER visits
Outcome (ICD10): Chronic lower
respiratory disease (CLRD) (40-47)
Includes chronic bronchitis,
emphysema, other COPDs, asthma,
bronchiectasia
Age Groups: >64 yr
Study Design: Time series
N: 712 for CLRD
1 hospital
Statistical Analyses: Poisson
regression GAM, LOESS smoothers, no
mention of stringent criteria
Covariates: Day of week, time
minimum temperature, relative humidity
Season: All
Statistical Package: S-Plus
Lags Considered: 2-7 3 day ma
Pollutant: PM10
Averaging Time: Daily
Mean (SD): 60.0 pg/m3, (26.3)
Range (Min, Max):
22.8. 186.5 pg/m3
PM Component: None
Monitoring Stations: 12
Notes: Copollutant (correlation):
PM,o-CO: 0.73
PM10-N02:0.83
PM10-S02: 0.72
PM10-03: 0.35
PM Increment: 1 pg/m
Regression Coefficients (SE):
0.0024(0.0023), 6 day ma
Notes: % Increase (SD) for ER visits
per 2435 pg/m3 (IQR) PM,o (lag 6 day
ma) presented graphically in text.
December 2009
E-243
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Masjedi et al. (2003,
0521001
Period of Study: Sep 1997-Feb 1998
Location: Tehran, Iran
Hospital Admissions
Outcome (ICD-9): Acute asthma and
COPD exacerbations (ICD: NR)
Age Groups: NR
Study Design: Time series
N: 355 patients
Statistical Analyses: Multiple stepwise
regression, autoregression method
(time series), Pearson correlation
Covariates: NR
Season: NR
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 3-, 7-, and 10-day
mean
Pollutant: PM,0
Averaging Time: 24 h
Mean (min-max):
108.41 (14.5-506.60)
SD = 59.55
Monitoring Stations: 3
Copollutant: NR
PM Increment: NR
Results:
Time-series analysis
Asthma: (3 = 0.002
p = 0.32
COPD: (3 = 0.004
p = 0.02
Total Acute Resp Conditions: (3 = 0.006
p = 0.27
Correlation of 3-day mean
Asthma: r =-0.21
(3 =-0.16
p = 0.08
Correlation of weekly mean
Asthma: r =-0.27
(3 = -0.008
p = 0.12
Correlation of 10-day mean
Asthma: r = -0.38
(3 = -0.066
p = 0.089
Reference: McGowan et al. (2002,
0303251
Period of Study: Jun 1988-Dec 1998
Location: Christchurch, New Zealand
Hospital Admissions
Outcome (ICD-9): Pneumonia (480-
487), acute respiratory infections (460-
466), chronic lung diseases (491-492,
494-496), asthma (493)
Age Groups: <15 yr, 15-64, 65+
Study Design: Time series
N: 20,938 admissions
Statistical Analyses: GAM with log
link, Linear Regression Model
Covariates: Wind speed, relative
humidity, temperature
Season: NR
Dose-response Investigated? No
Statistical Package: S-PLUS
Lags Considered: 0-6 days
Pollutant: PM10
Averaging Time: 24 h
Mean (min-max):
25.17(0-283)
SD = 25.49
Monitoring Stations: 1
Copollutant: NR
PM Increment: 14.8 fjg/nr (IQR)
% Increase [Cl]:
Respiratory Admissions (2-day lag)
0-14 yr: 3.62 [2.34,4.90]
15-64 yr: 3.39 [1.85,4.93]
65+yr: 2.86 [1.23,4.49]
All ages: 3.37 [2.34,4.40]
Overall
Acute respiratory infections: 4.53
[2.82,6.24]
Pneumonia/influenza: 5.32 [3.46,7.18]
Chronic lung diseases: 3.95 [2.15,5.75]
Asthma: 1.86 [0.48,.3.24]
Total Respiratory Admissions
Same day lag: 2.52 [1.49,3.55]
1-day lag: 2.56
2-day lag: 3.37
3-day lag: 3.09
4-day lag: 3.13
5-day lag: 3.21
6-day lag: 3.09
1.53,3.59'
2.34,4.40:
2.06,4.12:
2.10,4.16:
2.18,4.24:
2.06,4.12:
December 2009
E-244
-------
Reference
Reference: Medina-Ramon et al.
(2006, 0877211
Period of Study: 1986-99
Location: 36 U.S. Cities
Design & Methods
Outcome: 490-496, except 493
(COPD), 480-487 (Pneumonia)
Age Groups: 65 + (U.S. Medicare
beneficiaries)
Concentrations1
Pollutant: PM,0
Averaging Time: 24-h avg
Mean(SD):30.4|jg/m3(5.1)
Effect Estimates (95% Cl)
PM Increment: 10 pg/m3
% change [Lower Cl, Upper Cl]
lag:
Study Design: Case crossover
N: 578,006 COPD admissions
1,384,813 Pneumonia admissions
Statistical Analyses: Conditional
logistic regression, Meta-analysis using Copollutant: NR
REML random effects models
Covariates: Mean and variance of daily
summer apparent temperature index, %
65+ living in poverty, % households with
central air-conditioning mortality rate for
emphysema among 65+(surrogate for
smoking history), % PMi0 from traffic
Season: WarmjMay -Sepnd
Cold(Oct-Apr)
Dose-response Investigated? No
Statistical Package: SAS
STATA
Lags Considered: 0-1 days
Monitoring Stations: at least one per
city
Notes: PM10 measurements made
every 2, 3 or 6 days depending on the
city.
COPD warm season
0.81(0.22,1.41) at lag 0
1.47(0.93,2.01) at lag 1
COPD cold season
0.06(-0.40,0.51)atlagO
0.10(-0.30,0.49)atlag1
Pneumonia warm season
0.84 (0.50,1.19) at lag 0
0.79 (0.45,1.13) at lag 1
Pneumonia cold season
0.30 (0.07,0.53) at lag 0
0.14 (-0.17,0.45) at lag 1
Reference: Meng et al, (2007, 0932751
Period of Study: Nov 2000-Sep 2001
Location: Los Angeles and San Diego
counties, California
Outcome (ICD-NR): Poorly controlled
asthma defined as (1) daily or weekly
asthma symptoms or (2) at least 1 ED
visit or hospitalization due to asthma
over the past 12 mo
Age Groups: >18yr
Study Design: Time series
N: 1609 asthma patients
Statistical Analyses: Logistic
regression
Covariates: Age, sex, race/ethnicity,
poverty level, insurance status, smoking
behavior, employment, asthma
medication use, and county
Season: NR
Dose-response Investigated: No
Statistical Package: NR
Lags Considered: NR
Pollutant: PM10
Averaging Time: 24 h
Mean (26-76th percentile): NR
Monitoring Stations: NR
03:r = -0.72
N02:r = 0.83
CO: r = 0.42
Other variables:
Traffic: r = 0.14
PM Increment: 10 pg/m3
OR Estimate [Cl]:
All Adults: 1.08 [0.82,1.43]
18-64 yr: 1.14 [0.84,1.55]
65+: 0.84 [0.41,1.73]
Men: 0.72 [0.42,1.21]
Women: 1.38 [0.99,1.94]
Exposure above 44.01 pg/m3 (annual
concentration)
All Adults: 1.56 [0.96,2.52]
18-64 yr: 1.40 [0.81,2.41]
65+: 2.23 [0.60,8.27]
Men: 0.80 [0.27,2.41]
Women: 2.06 [1.17,3.61]
Notes: This study focused more on the
relation between poorly controlled
asthma and traffic density.
December 2009
E-245
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Middleton et al. (2008,
1567601
Period of Study: 1995-1998,
2000-2004
Location: Nicosia, Cyprus
Hospital Admissions/ED visits
Outcome (ICD-NR):
Hospital admissions for all respiratory
disease (ICD-10: JOO-J99).
Age Groups Analyzed: All, also
stratified by age (<15 vs.. >15 yr)
Study Design: Time series
N: Statistical Analyses: Generalized
additive Poisson models
Covariates: Seasonality, day of the
week, long- and short-term trend,
temperature, relative humidity
Season: NR
Dose-response Investigated: No
Statistical package: STATA SE 9.0,
and the MGCV package in the R
software (R 2.2.0)
Lags Considered: lag 0 -2 days
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD
median
5%-95%
range):
Cold:
57.6 (52.5
50.8
20.0-103.0
5.0-1370.6)
Warm:
53.4 (50.5
30.7
32.0-77.6
18.4-933.5)
Monitoring Stations: 2
Copollutant (correlation): NR
Other variables:
PM Increment: 10 pg/m , and across
quartiles of increasing levels of PMi0
Percentage increase estimate [Cl]:
All age/sex groups (Lag 0):
All admissions: 0.85 (0.55,1.15)
Respiratory (all): 0.10 (-0.91,1.11)
Respiratory (cold months):
-0.33 (-1.47, 0.82)
Respiratory (warm months):
1.42 (-0.42, 3.31)
CVD+RD: 0.56 (-0.21, 1.34)
Nicosia residents (Lag 0):
Respiratory (all): 0.25 (-0.84,1.36)
Respiratory (cold months):
-0.22 (-1.45, 1.02)
Respiratory (warm months):
1.80 (-0.22, 3.85)
CVD+RD: 0.38 (-0.47, 1.23)
Males (Lag 0):
All admissions: 0.96 (0.54,1.39)
Respiratory (all):-0.06 (-1.37,1.26)
Respiratory (cold months):
-0.16 (-1.76,1.46)
Respiratory (warm months):
1.10 (-1.47, 3.74)
CVD+RD: 0.63 (-0.34, 1.62)
Females (Lag 0):
All admissions: 0.74 (0.31,1.18)
Respiratory (all): 0.39 (-1.21, 2.02)
Respiratory (cold months):
-0.26 (-2.18,1.70)
Respiratory (warm months):
3.27 (-0.00, 6.65)
CVD+RD: 0.59 (-0.68, 1.87)
Aged <16yr (Lag 0):
All admissions: 0.47 (-0.13,1.08)
Respiratory (all):-0.35 (-1.77,1.08)
Respiratory (cold months):
-0.31 (-2.02, 1.42)
Respiratory (warm months):
-0.59 (-3.53, 2.45)
Aged >15yr (Lag 0):
All admissions: 0.98 (0.63,1.33)
Respiratory (all): 0.59 (-0.87, 2.07)
Respiratory (cold months):
0.02 (-1.76,1.83)
Respiratory (warm months):
3.89(1.05,6.80)
Reference: Moore et al. (2008, 1966851
Period of Study: 1983-2000
Location: California's South Coast Air
Basin
Outcome: Hospital admissions for
asthma (ICD-9 493)
Study Design: Time-series
Covariates: Income, demographic and
residential variables
Statistical Analysis: HRMSM
Age Groups: Children ages 0-19 yr
Pollutant: 03 (PM10 secondary)
Averaging Time: Quarterly
Mean (SD) Unit: NR
Range (Min, Max): NR
Copollutant (correlation):
1hr03:0.52
8hr03:0.46
24 h N02: 0.53
24 h CO: 0.36
24 h SO,: 0.13
Results given are for 03
December 2009
E-246
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Nascimento et al. (2006,
0932471
Period of Study: May 2000-Dec 2001
Location: Sao Jose dos Campos,
Brazil
Hospital Admissions
Outcome (ICD-10): Pneumonia (J12-
J18)
Age Groups: 0-1 Oyr
Study Design: Time series
N: 1265 admissions
Statistical Analyses: GAM, Poisson
regression
Covariates: Temperature, humidity
Season: NR
Dose-response Investigated? Yes
Statistical Package: S-Plus, SPSS
Lags Considered: 0-7 days
Pollutant: PM,0
Averaging Time: 24 h
Mean (min-max):
40.2(3.4-196.6)
SD = 26.9
Monitoring Stations: 2
Copollutant (correlation):
S02:r = 0.30
03:r = 0.09
Other variables:
Admissions: r = 0.21
Temp: r = -0.14
Notes: All p < 0.05
PM Increment: 24.7 pg/m
Regression coefficients (SE):
Same day:-0.00053 (0.00125)
1-day lag: 0.00029 (0.00057)
2-day lag: 0.00089 (0.00069)
3-day lag: 0.00122 (0.00053)*
4-day lag: 0.00126 (0.00055)*
5-day lag: 0.00098 (0.00071)
6-day lag: 0.00035 (0.00056)
7-day lag:-0.00067 (0.00123)
*p < 0.05
Notes: Percent increase over all lag
days is displayed in Fig 2
Reference: Neuberger et al. (2004,
0932491
Period of Study: 1999-2000 (1 yr
period)
Location: Vienna and Lower Austria
Hospital Admissions
Outcome (ICD-9): Bronchitis,
emphysema, asthma, bronchiectasis,
extrinsic allergic alveolitis, and chronic
airway obstruction (490-496)
Age Groups: 3.0-5.9 yr
7-1 Oyr
65+
Study Design: Time series
N: 366 days (admissions NR)
Statistical Analyses: GAM
Covariates: S02, NO, N02, 03,
temperature, humidity, and day of the
week
Season: NR
Dose-response Investigated? Yes
Statistical Package: S-Plus 2000
Lags Considered: 0-14 days
Pollutant: PM10
Averaging Time: 24 h
Maximum daily mean:
Vienna: 105
Rural area: NR
Monitoring Stations: NR
Copollutant: NR
PM Increment: 10 pg/m3
Log Relative Rate Estimate (p-value):
Vienna
Male: 2 day lag = 4.217 (0.030)
Association with tidal lung function:
P = -1.067 (p-value = 0.241)
Notes: Effect parameters with
significant coefficients for respiratory
health included: male sex, allergy,
asthma in family, and traffic for Vienna
and age, allergy, asthma in family, and
passive smoking for the rural area.
Effect parameters with significant
coefficients for log asthma score were
allergy, asthma in family, and rain for
Vienna and allergy, asthma in family,
and passive smoking for the rural area.
Reference: Oftedal et al. (2003,
0556231
Period of Study: 1995-2000
Location: Drammen, Norway
Hospital Admissions
Outcome: All Respiratory (460-517)
Age Groups: All
Study Design: Time-series
N: ~4,458 admissions
Statistical Analyses: Poisson
regression, GAM w/ stringent
convergence criteria
Covariates: Temperature, humidity,
influenza epidemics, summer and
Christmas vacation
Season: All
Dose-response Investigated? Yes
Statistical Package: S-Plus
Lags Considered: 2-3
Pollutant: PM10
Averaging Time: 24 h
Mean (SD): 16.8 pg/m3, (10.2) 1994-
1997
16.5 pg/m3, (10.3)1998-2000
16.6, pg/m3 (10.2) total period
PM Component: Benzene,
formaldehyde, toluene
Monitoring Stations: NR
Notes: Copollutant (correlation):
Correlation between pollutants ranged
from -0.47-0.78 with the exception of
the VOCs studied
Notes: Benzene, formaldehyde and
toluene also evaluated
PM Increment: IQr=11.04
RR Estimate
1.035 [0.990, 1.083] 1994-1997
0.992 [0.948, 1.037] 1998-2000
1.021 [0.990, 1.053] 1994-2000
2 Pollutant Model
PM10w/benzene: 1.01 (0.978,1.043)
December 2009
E-247
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Peel et al. (2005, 0563051
Period of Study: Jan 1993-Aug 2000
Location: Atlanta, Georgia
ED visits
Outcome: Asthma (493, 786.09)
COPD(491,492, 496)
URI (460-466, 477)
Pneumonia (480-486)
Age Groups: All ages. Secondary
analyses conducted by age group: 0-1,
2-18, >18
Study Design: Time series
N: 31 hospitals
Statistical Analyses: Poisson GEE for
URI, asthma and all RD
Poisson GLM for pneumonia and
COPD)
Covariates: Avg temperature and dew
point, pollen counts
Season: All (secondary analyses of
warm season)
Dose-response Investigated? Yes
Statistical Package: SAS 8.3, S-Plus
2000
Lags Considered: 0-7 day, 3 day ma,
0-13 days unconstrained distributed lag
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): 27.9 (12.3) pg/m3
Percentiles: 10th: 13.2
90th: 44.7
Monitoring Stations:
"Several"
Copollutant (correlation):
8h03:r = 0.59
1 hN02:r = 0.49
1 hCO:r = 0.47
1hS02:r = 0.20
24-h PM25: 0.84
24hPM10.25:r = 0.59
24hUF:r = -0.13
Components: r ranged from 0.42-0.74
PM Increment: PMi0:10 pg/m
RR Estimate [Lower Cl, Upper Cl]
All Respiratory Outcomes:
1.013(1.004-1.021), 3 day ma
URI:
1.014 (1.004-1.025), 3 day ma
1.073(1.048-1.099), 14-day dist. lag
Asthma:
1.009(0.996-1.022), 3 day ma
1.099(1.065-1.135), 14-day dist. lag:
Pediatric Asthma 2-18yrs):
1.016(0.998-1.034)
Pneumonia:
1.011 (0.996-1.027), 3 day ma
1.087(1.044-1.132), 14-day dist. lag
COPD:
1.018(0.994-1.043), 3 day ma
1.092(1.023-1.165), 14-day dist. lag
Notes: RRs obtained using AQS 1993-
2000, AQS 1998-2000 and ARIES data
compared. Infant (0-1 y) and pediatric
(2-18 y) asthma was associated more
strongly with PMi0, PM25 and OC than
adult asthma.
Reference: Ren et al.
(2006, 0928241
Period of Study: Jan 1996-Dec 2001
Location: Brisbane, Australia
Hospital Admissions
Outcome (ICD-9): Respiratory
diseases (460-519) excluding influenza
(487.0-487.8)
Age Groups: NR
Study Design: Time series
N:NR
Statistical Analyses: GAM
Pollutant: PM,0
Averaging Time: 24 h
Mean (min-max): 15.84 (2.5-60)
Monitoring Stations: 1
Copollutant: NR
PM Increment: NR
Coefficient Estimates:
Respiratory Hospital Admissions
Same day: -0.004296
1-day lag: -0.002474
2-day lag: -0.004229
*all statistically significant
Covariates: Day of week, relative
humidity, influenza outbreaks
Season: NR
Dose-response Investigated? Yes
Statistical Package: S-Plus
Lags Considered: 0, 1, and 2 days
Respiratory Emergency Visits
Same day: -0.000887
1-day lag: -0.004209
2-day lag: -0.003440
Notes: Relative risks were provided in
graphical form (Fig 3)
Reference: Sauerzapfetal. (2009,
1800821
Period of Study: Mar 2006-Mar 2007
Location: Norfolk, UK
Outcome: COPD
Study Design: Case-Crossover
Covariates: Environmental factors and
Influenza
Statistical Analysis: Logistic
regression
Statistical Package: SPSS 14
Age Groups: >18yr
N: 1050 adult COPD admissions
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD) Unit:
Control: 19.87 (8.51) pg/m3
Case: 20.47 (9.27) pg/m3
Range (Min, Max):
Control: 9.77-34.27
Case: 10.04-35.03
Copollutant (correlation): NR
Increment: 10|jg/m
Odds Ratio (96% Cl)
Lag 0-7, unadjusted: 1.079 (0.980-
1.188)
Lag 0-8, adjusted: 1.101 (0.988-1.226)
Lag 1-8, unadjusted: 1.056 (0.961-
1.161)
Lag 1-8, adjusted: 1.054 (0.949-1.170)
December 2009
E-248
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Sinclair and Tolsma (2004,
0886961
Period of Study: 25 Months
Location: Atlanta, Georgia
Outpatient Visits
Outcome: Asthma (493)
URI(460, 461,462, 463, 464, 465, 466,
477)
LRI (466.1,480, 481,482, 483, 484,
485, 486).
Age Groups: < = 18 yr, 18+ yr (asthma)
All ages (URI//LRI)
Study Design: Times series
N: 25 mo
260,000-275,000 health plan members
(Aug 1998-Aug 2000)
Statistical Analyses: Poisson GLM
Covariates: Season, Day of week,
Federal Holidays, Study Months
Season: NR
Dose-response Investigated?: No
Statistical Package: SAS
Lags Considered: Three 3-day ma (0-
2, 2-5, 6-8)
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): PM,0 mass-29.03 pg/m3
(11.61)
Monitoring Stations: 1
Notes: Copollutant: NR
PM Increment: 11.61 (1 SD)
RR Estimate [Lower Cl, Upper Cl]
lag:
Child Asthma: 1.049(3), lag 3-5 day
LRI: 1.074(3), 3-5 day lag
Notes: Numerical findings for significant
results only presented in manuscript.
Results for all lags presented
graphically for each outcome (asthma,
URI, and LRI).
Reference: Slaughter et al. (2005,
0738541
Period of Study: Jan 1995-Jun 2001
Location: Spokane, WA
Hospital Admissions and ED visits
Outcome: All respiratory (460-519)
Asthma (493)
COPD (491,492, 494,496)
Pneumonia (480-487)
Acute URI not including colds and
sinusitis (464, 466, 490)
Age Groups: All, 15+ yr for COPD
Study Design: Time series
N: 2373 visit records
Statistical Analyses: Poisson
regression, GLM with natural splines.
For comparison also used GAM with
smoothing splines and default
Pollutant: PM,0
Averaging Time: 24-h avg
Range (90% of concentrations): 7.9-
41.9|jg/m3
Monitoring Stations:
1
Notes: Copollutant (correlation):
PM10
PM1 r = 0.50
PM25r = 0.62
CO r = 0.32
Tern peraturer = 0.11
PM Increment: 25 pg/m
RR Estimate [Lower Cl, Upper Cl]
lag:
ER visits -PM10
All Respiratory
Lag 1:1.01
Lag 2: 1.01
0.99, 1.04]
0.98,
1.04]
1.03]
Lag 3: 1.02 [0.99, 1.04]
Acute Asthma
Lag 1:1.03 [0.98, 1.07]
Lag 2: 1.01 [0.96, 1.05]
Lag 3: 1.00 [0.95, 1.04]
COPD (adult)
Lag 1:1.00 [0.93, 1.07]
Lag 2: 0.99
Lag 3:1.02
0.92, 1.0
0.95, 1.0
Hospital Admissions - PIvl-;
All Respiratory
uui ivci yci i^c ui ILCI la.
Covariates: Season, temperature,
relative humidity, day of week
Season: All
Dose-response Investigated?: No
Statistical Package: SAS, SPLUS
Lags Considered: 1 -3 day
Lag 1:0.99
Lag 2: 0.99
0.95,
0.96,
Lag 3: 1.00 [0.97,
Asthma
Lag 1:1. 03
Lag 2: 1.01
Lag 3: 1.00
COPD (adu
Lag 1:0.98
Lag 2: 1.03
Lag 3: 1.02
0.95,
0.94,
0.92,
t)
0.90,
0.96,
0.94,
1.02
1.02
1.03]
1.12]
1.10]
1.09]
1.07]
1.11]
1.09]
December 2009
E-249
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Sun et al. (2006, 0907681
Period of Study: Jan 2004-Dec 2004
Location: Taichung, Taiwan (Central
Taiwan)
ED visits
Outcome: Asthma (493.xx)
Age Groups: <55, <16,16-55 yr
Study Design: Cross-sectional
N:NR
All diagnoses for all patients at 4
medical centers
Statistical Analyses: Pearson's
correlations, multiple correlation
coefficients from regression analyses.
Covariates: Only copollutants
considered
Dose-response Investigated? No
Statistical Package: SPSS
Lags Considered: None
Pollutant: PM,0 Children ED Visits
Averaging Time: Monthly avg for 2004 r = 0.626
P = 0.015
Adult ED Visits
r = 0.384
P = 0.109
Mean(SD):~60.3|jg/m3
(estimated from Fig)*
Range (Min, Max): (-35, 80)
Monitoring Stations: 11
Copollutant: NR
Reference: Szyszkowicz (2007,
Period of Study: Jan 1992-Mar 2002
Location: Edmonton, Canada
Outcome: ED visits for asthma (ICD-
(493)
Study Design: Time-series
Covariates: Temperature, relative
humidity
Statistical Analysis: Poisson
regression
Age Groups: All
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD) Unit: 22.6 (13.1) pg/m3
Median, IQR: 19.4, 15.0
Copollutant (correlation): NR
Increment: IQR
Percent Relative Risk (96% Cl)
"Only statistically significant results are
presented in the paper*
No lag, > 10 yr
Apr-Sep, All: 3.7 (1.5-6.0)
Apr-Sep, Female: 4.5 (1.8-7.3)
Apr-Sep, Male: 3.3 (0.1-6.7)
2 day lag, <10yr
Year round, All: 2.7 (0.1-5.4)
Apr-Sep, All: 6.3 (2.6-10.2)
Apr-Sep, Male: 7.4 (3.1-11.9)
2 day lag, > 10 yr
Apr-Sep, All: 2.4 (0.1-4.7)
Apr-Sep, Female: 3.9 (1.1-6.7)
Reference: Tecer et al, (2008, 1 80030) Outcome: ED visits for respiratory
Period of Study: Dec 2004-Oct 2005 pr°b'emS ^ 47M7* 493)
Study Design: Bidirectional Case-
Location: Zonguldak, Turkey crossover
Covariates: Daily meteorological
parameters
Statistical Analysis: Conditional
logistic regression
Statistical Package: Stata
Age Groups: 0-1 4 yr
Pollutant: PM,0
Averaging Time: NR
Mean, Unit: 53.3 pg/m3
Range (Min, Max): 12-237.5
Copollutant (correlation):
PM25/PM10
Mean: 0.56
Range: 0.17-0.88
Increment:
10|jg/m3
Odds Ratio (96% Cl)
Asthma
Lag 0:1. 14 (1.03-1. 26)
Lag 1:0.92
Lag 2: 0.92
Lag 3: 1.01
0.83-1.02)
0.81-1.03)
0.92-1.11)
Lag 4: 1.16 (1.06-1. 26)
Allergic Rhinitis with Asthma
Lag 0:1. 07
Lag 1:0.96
Lag 2: 0.93
Lag 3: 0.96
1.01-1.13)
0.91-1.02)
0.88-0.99)
0.90-1.02)
Lag 4: 1.08 (1.02-1. 14)
Allergic Rhinitis
Lag 0:1. 06 (0.99-1. 13)
Lag 1:1. 08
Lag 2: 0.92
1.01-1.16)
0.87-0.99)
Lag 3: 0.97 (0.92-1. 03)
Lag 4: 1.09 (1.03-1. 16)
Upper Respiratory Disease
Lag 0: 0.88
Lag 1:1. 17
Lag 2: 1.00
0.68-1.14)
0.91-1.51)
0.76-1.31)
Lag 3: 0.95 (0.76-1. 19)
Lag 4: 1.15 (0.97-1. 35)
Lower Respiratory Disease
Lag 0:1. 01 (0.86-1. 19)
Lag 1:1. 04
Lag 2: 1.04
0.88-1.23)
0.92-1.18)
Lag 3: 1.23 (1.07-1. 41)
Lag 4: 0.99 (0.90-1. 08)
December 2009
E-250
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Tolbert et al. (2007,
0903161
Period of Study: 1993-2004
Location: Atlanta Metropolitan area,
Georgia
Hospital Admissions/ED visits
Outcome (ICD-9):
Combined RD group, including:
Asthma (493, 786.07, 786.09), COPD
(491, 492, 496), URI (460-465, 460.0,
477), pneumonia (480-486), and
bronchiolitis (466.1,466.11, and
466.19))
Age Groups Analyzed: All
Study Design: Time series
N: 10,234,490 ER visits (283,360 and
1,072,429 visits included in the CVD
and RD groups, respectively)
Statistical Analyses: Poisson
generalized linear models
Covariates: Long-term temporal trends,
season (for RD outcome), temperature,
dew point, days of week, federal
holidays, hospital entry and exit
Season: All
Dose-response Investigated: No
Statistical package: SAS version 9.1
Lags Considered: 3-day ma(lag 0 -2)
Pollutant: PM,0
Averaging Time: 24 h
Mean (median
IQR, range, 10th-90th percentiles):
26.6 (24.8
17.5-33.8
0.5-98.4
12.3-42.8)
Monitoring Stations: NR
Copollutant (correlation):
03:r = 0.59
N02:r = 0.53
CO: r = 0.51
S02:r = 0.21
Coarse PM:r = 0.67
PM25:r = 0.84
PM25S04:r = 0.69
PM25EC:r = 0.61
PM25OC:r = 0.65
PM25TC:r = 0.67
PM25 water-sol metals: r = 0.73
OHC:r = 0.53
PM Increment: 16.30 pg/rri (IQR)
Risk ratio [96% Cl]:
Single pollutant models:
RD: 1.015 (1.006-1.024)
Notes: Results of selected multi-
pollutant models for respiratory disease
are presented in Fig 2.
Fig 2: PM10 adjusted for CO, 03, N02,
or N02/03 (nonwinter months only)
Summary of results: PMi0 remained
predictive of RD in non-winter months
after adjustment for pollutants.
Reference: Tsai et al. (2006, 0897681
Period of Study: 1996-2003
Location: Kaohsiung City, Taiwan
Outcome: Asthma (493)
Age Groups: All (universal health care
covers >96% of the population)
Study Design: Case crossover
N: 17,682 admissions
63 hospitals
Statistical Analyses: Conditional
Logistic Regression
Covariates: Temperature, humidity
Season: Warm and cool seasons
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-2 day cumulative
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): 76.62 pg/m3 (NR)
Percentiles: 25th: 41.73
SOth(Median): 74.40
75th: 104.01
Range (Min, Max): (16.70, 232.00)
Monitoring Stations: 6
Copollutant: NR
PM Increment: 62.28 pg/m
OR Estimate [Lower Cl, Upper Cl]
Single-pollutant model, 0-2 day
cumulative lag
>25oC: 1.302 [1.155, 1.467]
<25oC: 1.556 [1.398, 1.371]
Two-pollutant models, 0-2 day
cumulative lag
PM,oW/S02
>25oC: 1.305 [1.156, 1.473]
<25oC: 1.540 [1.374, 1.727]
PM10w/03
>25oC: 0.985 [0.842, 1.152]
<25oC: 1.581 [1.402, 1.783]
PM,0w/N02
>25oC: 1.237 [1.052, 1.455]
<25oC: 1.009 [0.875, 1.163]
PM,oW/CO
>25oC: 1.156 [1.012, 1.320]
<25oC: 1.300 [1.134, 1.490]
December 2009
E-251
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ulirsch et al. (2007,
0913321
Period of Study: Nov 1994-Mar 2000
Location: Pocatello, Idaho
Chubbuck, Idaho
Outcome: Respiratory Disease (460-
499, 509-519)
Reactive Airway Disease (786.09)
Age Groups: All age groups
Study Design: Time series
N: 39,347 visits (TS1)
29,513 visits (TS2)
Statistical Analyses: Poisson
regression, GLM. Sensitivity Analyses
Covariates: Time, Temperature,
Relative Humidity Influenza
Season: Warm/Cool
Dose-response Investigated? No
Statistical Package: S-Plus
Lags Considered: 0-4 day lags
Notes: Time series (TS) 1 includes HA,
ED and urgent care visits. TS 2 includes
family practice data available after 1997
Pollutant: PM,0
Averaging Time: NR
Mean (SD):TS1: 24.2 pg/m3(NR)
10th: 10.5
90th: 40.7
TS2: 23.2
10th: 10.0
90th: 37.4
Range (Min, Max):
TS1: (3.0, 183.0)
TS2: (3.0, 183.0)
Monitoring Stations: 4
Notes: Copollutant (correlation):
PM10 w/ N02: r = 0.47. PM10 with other
copollutants weakly correlated.
PM Increment: Single Pollutant
Models, TS1: 24.4 pg/m3
Single Pollutant Models: TS2:
23.2 pg/m3
Multipollutant Models: TS1/TS2:
50 pg/m3
Mean Percentage Change, lag 0
TS1: Single Pollutant
All-age (all yr): 4.0
1.4,6.7]
18-64: 3.4 [0.2, 6.7
0-17: 4.3 [-0.1, 8.9]
65+: 5.6 [-1.4, 13.1]
0-17/65+:5.5[1.4, 9.6]
All age (Cool season): 4.3 [1.3, 7.5]
All age (Warm season): 6.7 [-0.8,14.8]
TS2: Single Pollutant
All-age: 3.3 [0.3, 6.3]
18-64: 3.3 [-0.4, 7.0]
0-17: 5.0 [0.1, 10.1
65+: 6.9 [-0.4, 14.7
Multipollutant (PM10 + S02)
All-age (all yr): TS1 10.8
TS2 17.5
18-64: TS1 8.0
TS29.1
0-17: TS1 10.8
TS2 32.7
65+:TS18.7
TS231.3
0-17/65+:TS1 14.2
TS2 25.3
All age (Cool season) TS1 11.9
Multipollutant (PM10 + N02)
All-age (all yr)TS1:TS2 16.3
18-64: TS1 9.3
TS2 17.3
0-17: TS1 4.6
TS2 18.7
65+:TS1 12.4
TS2 32.7
0-17/65+:TS19.5
32.7
All age (Cool season): TS1 11.1
TS2 16.8
Notes: Results from multipollutant
model with PMi0, S02 and N02 also
available.
Reference: Vegni and Ros (2004,
0874481
Period of Study: Sep 2001-Sep 2002
Location: Milan area, Italy
Hospital Admissions
Outcome (ICD-9): Respiratory, non-
infectious admissions (ICD: NR)
Age Groups: NR
Study Design: Time series
N: 9881 admissions
Statistical Analyses: Poisson
regression
Covariates: Temperature, wind velocity,
relative humidity, week day, holidays
Season: Spring, summer, fall, winter
Dose-response Investigated? No
Statistical Package: STATAv. 5
Lags Considered: 0, 1, and 2 days
Pollutant: PM,0
Averaging Time: 24 h
Mean (6th-96th percentile):
Overall: 41.5 (13-98) SD = 28.2
Spring: 29.0 (10-51) SD=12.6
Summer: 24.8 (10-40) SD = 9.9
Fall: 51.8 (21-114) SD = 27.1
Winter: 64.1(20-135) SD = 35.7
Monitoring Stations: 1
Copollutant: NR
PM Increment: Increase from 5th-95th
percentile
Spring: 85 pg/m3
summer: 30 pg/m3
Fall: 93 pg/m3
Wnter: 115 pg/m3
RR Estimate [Cl]:
Overall: 1.10 [0.83,1.46]
Adjusted: 0.97 [0.67,1.41]
Notes: 1-day and 2-day lags show
similar results, with no association
between PM10 and daily hospital
admissions
December 2009
E-252
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Vigotti et al. (2007, 0907111 ED Visits
Period of Study: Jan 2000-Dec 2000
Location: Pisa, Italy
Outcome: Asthmatic attack (493), dry
cough (468), acute bronchitis (466)
Age Groups: <10 yr; 65+ yr
Study Design: Time series
N: 966 Emergency room visits
Statistical Analyses: Poisson
regression, GAM, LOESS smoothers,
stringent criteria
Covariates: Temperature, humidity,
relative humidity, day of study, rainfall,
influenza, day of-the-wk, holidays, time
trend
Season: All yr
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0-5 days
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 35.4 (15.8) pg/m3
Percentiles: 25th: NR
50th(Median):31.6
75th: NR
Range (Min, Max): (9.5,100.1)
Monitoring Stations:
2
Copollutant (correlation):
PM10:
N02r = 0.58
CO r = 0.70
PM Increment: 10 pg/m
RR Estimate [Lower Cl, Upper Cl]
lag:
<10y:10%[2.3,18.2]
Iag1
65+: 8.5% [1.5, 16.1]
lag 2
Reference: Xirasagar et al. (2006,
0932671
Period of Study: 1998-2001
Location: Taiwan
Hospital Admission/ED:
Outcome: Asthma or Asthmatic
Bronchitis (493)
Age Groups: <2 yr old, 2~5 yr old,
6-14 yr old
Study Design:
N: 27, 275 pediatric hospitalizations
Statistical Analyses: ARIMA Modeling
Spearman's Correlations
Covariates: Season, ambient temp.,
rel. humidity, atmospheric pressure,
rainfall, h of sunshine
Season: Spring: Feb-Apr
Summer: May-Jul
Fall:Aug-Oct
Winter: Nov-Jan
Dose-response Investigated? No
Statistical Package: EViews 4
Lags Considered: NR
Pollutant: PM,0
Averaging Time: Monthly means
Mean (SD): 24.4 pg/m3 (NR)
Percentiles: NR
Range (Min, Max): NR
PM Component: NR
Monitoring Stations: 44 air quality
monitoring banks. 23 weather
observatories
Notes: Copollutant (correlation):
<2 yr old: r = 0.315
2-5 yr old: r = 0.589
6-14 yr old: r = 0.493
PM Increment: NR
RR Estimate [Lower Cl, Upper Cl]
lag: NR
AR Estimate [Lower Cl, Upper Cl]
lag: NR
Notes: Plot of monthly asthma
admission rates per 100,000 population
by age group
Plot of mean monthly concentration
trends of criteria air pollutants
Mean monthly trends of climatic factors
Other Outcomes Assessed? NR
Other Exposures Assessed?
Seasonal ity
December 2009
E-253
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Wong et al., (2002,
0232321
Period of Study: 1995-1997 (Hong
Kong) and 1992-1994 (London)
Location: Hong Kong and London
Hospital Admissions
Outcome (ICD- NR): Asthma (493) for
ages 15-64 and respiratory disease
(460-519) for ages 65+
Age Groups: 15-64, 65+
Study Design: Time series
N:NR
Statistical Analyses: Poisson
regression, GAM
Covariates: Temperature, humidity, and
influenza
Season: Warm (Apr-Sep) and cool
(Oct-Mar)
Dose-response Investigated? Yes
Statistical Package: S-Plus
Lags Considered: 0-3 days
Pollutant: PM,0
Averaging Time: 24 h
Mean (min-max): Hong Kong: 51.i
(14.1-163.8)30 = 25.0
London: 28.5 (6.8-99.8)
SD=13.7
Monitoring Stations: NR
Copollutant (correlation):
Hong Kong
N02:r = 0.82
S02:r = 0.30
03:r = 0.54
London
N02:r = 0.68
S02:r = 0.64
03:r = 0.17
Other variables: Hong Kong
Temp: r = -0.42
Humidity: r = -0.53
London
Temp: r = 0.02
Humidity: r = -0.05
PM Increment: 10 pg/m
ER Estimate [Cl]:
Single-pollutant excess risk (mean lag
0-1 day)
Asthma-Hong Kong:-1.1 [-2.4,0.1]
Asthma-London: 1.4 [-0.1,3.0]
Respiratory Disease-Hong Kong:
1.0 [0.5,1.5]
Respiratory Disease-London:
0.4 [-0.3,1.2]
Warm season
Asthma-Hong Kong: -1.0 [-2.8, 0.8]
Asthma-London: 0.6 [-1.9,3.1]
Respiratory Disease-Hong Kong:
0.8 [0.1,1.4]
Respiratory Disease-London:
1.8 [0.5,3.1]
Cool season
Asthma-Hong Kong: -1.2 [-2.8,0.4]
Asthma-London: 1.6 [-0.3,3.6]
Respiratory Disease-Hong Kong:
1.2 [0.6,1.9]
Respiratory Disease-London:
-0.5 [-1.5,0.5]
Notes: RRs are shown graphically in
Fig 1 and 2. Exposure response curves
are provided in Fig 5 of the article
Reference: Wong et al. (2006, 0932661
Period of Study: 2000-2002
Location: Hong Kong (8 districts)
General Practitioner Visits
Outcome (ICPC-2): Respiratory
diseases/symptoms: upper respiratory
tract infections (URTI), lower respiratory
infections, influenza, asthma, COPD,
allergic rhinitis, cough, and other
respiratory diseases
Age Groups: All ages
Study Design: Time series
N: 269,579 visits
Pollutant: PM,0
Averaging Time: 24 h
Mean (min-max): Ranged from 43.4-
56.9 (dependent on location)
Monitoring Stations: 1 per district
Copollutant (correlation):
PM25:r = 0.94
03:r = 0.40
S02:r = 0.28
PM Increment: 10 pg/m3
RR Estimate [C I]:
Overall URTI
1.020 [1.016,1.025]
Overall Non-UTRI
1.025 [1.018,1.032]
Notes: RRs are also reported for each
individual general practitioner yielding
similar results
Statistical Analyses: GAM, Poisson
regression
Covariates: Season, day of the week,
climate
Season: NR
Dose-response Investigated? No
Statistical Package: S-Plus
Lags Considered: 0-3 days
December 2009
E-254
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Yang et al. (2007, 0928471 Hospital Admission/ED:
Period of Study: 1996-2003
Location: Taipei, Taiwan
Outcome: Asthma (493)
Age Groups: All ages
Study Design: Case-crossover
N: 25,602 asthma hospital admissions
Statistical Analyses: NR
Covariates: Temperature, humidity, day
of-the-wk, seasonality, long term trends
Season: All yr
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-2
Pollutant: PM,0
Averaging Time: NR
Mean (SD): 48.99 pg/m3
Percentiles: 26th: 32.64
60th(Median):44.13
76th: 59.05
Range (Min, Max): (14.44, 234.91)
PM Component: NR
Monitoring Stations:
6 Stations
Notes: Copollutant: NR
PM Increment: 26.41 pg/m
OR Estimate [Lower Cl, Upper Cl]
lag:
Asthma
Single-Pollutant Model: Temperature
>25°C:1.046[0.971, 1.128]
Temperature <25° C:
1.048(1.011,1.251]
Two-Pollutant Model: Adjusted for S02:
>25° 01.006(0.920, 1.099]
<25° 01.088(1.040, 1.138]
Adjusted for N02:
>25° 00.800(0.717, 0.892]
<25° 00.982(0.937, 1.029]
Adjusted for CO:
>25° 00.920(0.844, 1.002]
<25° 01.029(0.984, 1.076]
Adjusted for 03:
>25° 01.038(0.950, 1.134]
<25° 01.042(1.004, 1.081]
AR Estimate [Lower Cl, Upper Cl]
lag: NR
Notes: Other Outcomes Assessed?
NR
Other Exposures Assessed? S02,
N02, CO, 03
Reference: Yang et al. (2007, 0928471 Hospital Admission
Period of Study: 1996-2003
Location: Taipei, Taiwan
Outcome: COPD (490-192), (494),
(496)
Age Groups: All ages
Study Design: Case-crossover
N: 46,491 COPD admissions, 47
hospitals
Statistical Analyses: Conditional
logistic regression
Covariates: Wfeather, day of-the-wk,
seasonality, long term trends
Season: Warm/Cool
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-2 cumulative
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 48.99 pg/m3
26th: 32.64
60th(Median):44.13
76th: 59.05
Range (Min, Max):
(14.44, 48.99)
Monitoring Stations:
6 Stations
Notes: Copollutant: NR
PM Increment: 26.41 pg/m
OR Estimate [Lower Cl, Upper Cl]
Single-Pollutant Model (0-2 day cum
lag):
Temperature >20° 0:1.133(1.098,
1.168]
Temperature <20° C: 1.035(0.994,
1.077]
Two-Pollutant Model:
PM,0w/S02:
>20° 0-1.180(1.139, 1.223]
<20° 0-1.004(0.954, 1.057]
PM,0w/N02:
>20° 0-1.013(0.973, 1.055]
<20° 0-1.074(1.022, 1.129]
PM,oW/CO:
>20° 0-1.061(1.023, 1.100]
<20° 0-1.067(1.016, 1.120]
PM10w/03:
>20° 0-1.097(1.062, 1.133]
<20° 0-1.036(0.996, 1.079]
December 2009
E-255
-------
Reference
Reference: Yang et al. (2004, 0874881
Period of Study: Jun 1995-Mar 1999
Location: Vancouver area, British
Columbia
Design & Methods
Hospital Admissions
Outcome (ICD-9): Respiratory
diseases (460-519), pneumonia only
(480-486), asthma only (493)
Age Groups: 0-3 yr
Studv Desian: Case control
Concentrations1
Pollutant: PM,0
Averaging Time: 24 h
Mean (min-max):
13.3(3.8-52.2)
SD = 6.1
Effect Estimates (95% Cl)
PM Increment: 7.9 pg/m3 (IQR)
OR Estimate [Cl]:
Values NR
Notes: Author states that ORs for PMi0
increased with lag time up to 3 days for
both sinale and multiole-Dollutant
bidirectional case-crossover (BCC), and
time series (TS)
N: 1610 cases
Statistical Analyses: Chi-square test,
Logistic regression, GAM (time-series),
GLM with parametric natural cubic
splines
Covariates: Gender, socioeconomic
status, weekday, season, study yr,
influenza epidemic month
Season: Spring, summer, fall, winter
Dose-response Investigated? No
Statistical Package: SAS (Case
control and BCC), S-Plus (TS)
Lags Considered: 0-7 days
Monitoring Stations: NR (data
obtained from Greater Vancouver
Regional District Air Quality Dept)
Copollutant (correlation):
PM25:r = 0.83
PM10.2.5: r = 0.83
CO: r = 0.46
03:r = -0.08
N02:r = 0.54
SO,: r = 0.61
models.
All units expressed in pg/m unless otherwise specified.
Table E-13. Short-term exposure-respiratory-ED/HA-PMi 25.
Reference
Design & Methods
Concentrations'!
Effect Estimates (95% Cl)
Reference: Chen et al.
(2005, 0875551
Period of Study:
Jun1995-Mar1999
Location: Vancouver area,
BC
Hospital Admissions
Outcome (ICD-9): Acute respiratory
infections (460-466), upper
respiratory tract infections (470-478),
pneumonia and influenza (480-487),
COPD and allied conditions (490-
496), other respiratory diseases (500-
519)
Age Groups: >65 yr
Study Design: Time series
N: 12,869
Statistical Analyses: GLM
Covariates: Temp and relative
humidity
Season: NR
Dose-response Investigated? No
Statistical Package: S-Plus
Lags Considered: 1,2,3,4,5,6,
and 7-day avg
Pollutant: PMi0.2.5 (fjg/m)
Averaging Time: 24 h
Mean (min-max):
5.6(0.1-24.6)
SD = 3.6
Monitoring Stations: 13
Copollutant (correlation):
PM25:r = 0.38
PM,o:r = 0.83
COH:r = 0.63
CO: r = 0.53
03:r = -0.13
N02:r = 0.54
S02:r = 0.57
Other variables:
Mean temp: r = 0.13
Rel humidity: r = -0.27
PM Increment: 4.2 pg/m
RR Estimate [Cl]:
Adj for weather conditions
Overall admission
1-day avg: 1.03 [1.00,1.06]
2-day avg: 1.05 [1.02,1.08]
3-day avg: 1.06 [1.02,1.09]
Adj for weather conditions and copollutants
Overall admission
1-day avg: 1.02 [0.98,1.06]
2-day avg: 1.05 [1.01,1.10]
3-day avg: 1.06 [1.02,1.11]
Notes: RR's were also provided for lags 4-7 in Table 3,
yielding similar results
December 2009
E-256
-------
Reference
Reference: Fung et al. (2006,
0897891
Period of Study:
Jun1995-Mar1999
Location: Vancouver,
Canada
Reference: Halonen et al.
(2009, 1803791
Period of Study: 1998-2004
Location: Helsinki, Finland
Design & Methods
Hospital Admission/ED: Hospital
Admission
Outcome: Respiratory diseases
(460-519)
Age Groups: Age >65
Study Design: Time series
N: 26,275 individuals admitted
Statistical Analyses: Poisson
regression (spline 12 knots), case-
crossover (controls +17 days from
case date), Dewanji and Moolgavkar
(DM) method
Covariates: Long-term trends, day-
of-the-week effect, weather
Season: All yr
Dose-response Investigated? No
Statistical Package: SPIus, R
Lags Considered: 0-7 days
Outcome: Hospital Admissions
Age Groups: 65+ yr
Study Design: Time series
N:NR
Statistical Analyses: Poisson, GAM
Covariates: Temperature, humidity,
influenza epidemics, high pollen
episodes, holidays
Dose-response Investigated? No
Statistical Package: R
Lags Considered: Lags 0-3 & 5-day
(0-4) mean
Concentrations!
Pollutant: PM10.2.5 (|jg/m3)
Averaging Time: 24-h avg
Mean (SD)
5.6(3.88) pg/m3
Range (Min, Max):
(-2.9, 27.07)
Monitoring Stations:
NR
Notes: Copollutant
(correlation):
PM,0-2.5
PM,0r = 0.83
PM25 r = 0.34
CO r = 0.51
CoH r = 0.61
03r = -0.11
N02r = 0.52
S02r = 0.57
Pollutant: PM10.25
Averaging Time: Daily
Mean (SD): NR
Min: 0.0
26th percentile: 4.9
60th percentile: 7.5
76th percentile: 12.1
Max: 101. 4
Monitoring Stations: NR
Copollutant:
PMO.03, PM0.03-0.1,PM<0.1,
PM<0.10.29, PM25, CO, N02
Co-pollutant Correlation
pM
-------
Reference
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Host et al. (2007,
1558511
Period of Study: 2000-2003
Location: Six French cities:
Le Havre, Lille, Marseille,
Paris, Rouen, and Toulouse
Outcome (ICD-10): Daily
hospitalizations for all respiratory
diseases (JOO-J99), respiratory
infections (J10-J22).
Age Groups: For all respiratory
diseases: 0-14 yr, 15-64 yr, and 2
65 yr
For respiratory infections: All ages
Study Design: Time series
N: NR (Total population of cities:
approximately 10 million)
Statistical Analyses: Poisson
regression
Covariates: Seasons, days of the
week, holidays, influenza epidemics,
pollen counts, temperature, and
temporal trends
Pollutant: PM10.2.5
Averaging Time: 24 h
Mean ug/m3 (6th -96th
percentile):
Le Havre: 7.3 (2.5-14.0)
Lille: 7.9 (2.2-13.7)
Marseille: 11.0 (4.5-21.0)
Paris: 8.3 (3.2-15.9)
Rouen: 7.0 (3.0-12.5)
Toulouse: 7.7 (3.0-15.0)
Monitoring Stations:
13 total: 1 in Toulouse
4 in Paris
2 each in other cities
PM Increment: 10 pg/m3 , and an 18.8 pg/m3 increase
(corresponding to an increase in pollutant levels between
the lowest of the 5th percentiles and the highest of the 95th
percentiles of the cities' distributions)
ERR (excess relative risk) Estimate [Cl]: For all respiratory
diseases (10 pg/m3 increase): 0-14 yr: 6.2% [0.4, 12.3]
15-64yr:2.6%
[-0.5, 5.8]
> 65 yr: 1.9% [-1.9, 5.9]
For all respiratory diseases (18.8 pg/m3 increase): 0-14 yr:
12.0 [0.8, 24.3]
15-64 yr: 5.0 [-0.9, 11.1]
> 65 yr: 3.7 [-3.6, 11.4]
For respiratory infections (10 pg/m3): All ages: 4.4% [0.9,
Q ni
O.UJ
Dose-response Investigated: No
Statistical Package: MGCV package
in R software (R 2.1.1)
Lags Considered: Avg of 0-1 days
Copollutant (correlation):
PM25: Overall: r>0.6
Ranged between r = 0.28 and
r = 0.73 across the six cities.
For respiratory infections (18 pg/m ): All ages:
15.5]
.7,
Reference: Lin et al. (2005,
0878281
Period of Study: 1998-2001
Location: Toronto, North
York, East York, Etobicoke,
Scarborough, and York
(Canada)
Hospital Admissions
Outcome (ICD-9): Respiratory
infections including laryngitis,
tracheitis, bronchitis, bronchiolitis,
pneumonia, and influenza (464, 466,
480-487)
Age Groups: 0-14 yr
Study Design: Bidirectional case-
crossover
N: 6782 respiratory infection
hospitalizations
Statistical Analyses: Conditional
logistic regression (Cox proportional
hazards model)
Covariates: Daily mean temp and
dew point temp
Season: NR
Dose-response Investigated? No
Statistical Package: SAS 8.2
PHREG procedure
Lags Considered: 1- to 7-day avg
Pollutant: PM10.25 (pg/m3)
Averaging Time: 24 h
Mean (min-max):
10.86 (0-45.00)
SD = 5.37
Monitoring Stations: 4
Copollutant (correlation):
PM25:r = 0.33
PM10:r = 0.76
CO: r = 0.06
S02:r = 0.29
N02:r = 0.40
03:r = 0.30
PM Increment: 6.5 pg/m3
OR Estimate [Cl]:
Adjusted for weather
4-day avg: 1.16 [1.07,1.26]
6-day avg: 1.21 [1.10,1.32]
Adj for weather and other gaseous pollutants
4-day avg: 1.13 [1.03,1.23]
6-day avg: 1.17 [1.06,1.29]
Notes: OR's were also categorized into "Boys" and "Girls,"
yielding similar results
December 2009
E-258
-------
Reference
Reference: Lin et al. (2002,
0260671
Period of Study:
Jan 1981-Dec 1993
Location: Toronto
Reference: Peel et al. (2005,
ncc*3nc\
UbbJUb)
Period of Study:
Jan 1993-Aug 2000
Location: Atlanta, Georgia
Design & Methods
Hospital Admissions
Outcome (ICD-9): Asthma (493)
Age Groups: 6-12 yr
Study Design: Uni- and bi-directional
case-crossover (UCC, BCC) and
time-series (TS)
N: 7,319 asthma admissions
Statistical Analyses: Conditional
logistic regression, GAM
Covariates: Maximum and minimum
temp, avg relative humidity
Season: Apr-Sep, Oct-Mar
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 1- to 7-day avg
ED visits
Outcome: Asthma (493, 786.09)
COPD(491,492, 496)
URI (460-466, 477)
Pneumonia (480-486)
Age Groups: All ages. Secondary
analyses conducted by age group: 0-
1,2-18, >18
Study Design: Time series
N: 31 hospitals
Statistical Analyses: Poisson GEE
for URI, asthma and all RD
Poisson GLM for pneumonia and
COPD)
Covariates: Avg temperature and
dew point, pollen counts
Season: All (secondary analyses of
warm season)
Dose-response Investigated? Yes
Statistical Package: SAS 8.3
S-Plus 2000
Lags Considered: 0-7 days, 3-day
ma, 0-13 days unconstrained
distributed lag
Concentrations!
Pollutant: PM10.2.5 (|jg/m3)
Averaging Time: 6 days
(predicted daily values)
Mean (min-max):
12.17(0-68.00)
SD = 7.55
Monitoring Stations: 1
Copollutant (correlation):
PM25; r = 0.44
PM10:r = 0.83
CO: r = 0.17
S02: r = 0.28
N02:r = 0.38
03:r = 0.56
Pollutant: PMi0.25 (pg/rn3)
Averaging Time: 24 h avg
Mean (SD): 9.7 (4.7)
Percentiles: 10th: 4.4
90th: 16.2
Monitoring Stations:
"Several"
Copollutant (correlation):
24hPM,0:r = 0.59
8h03:r = 0.35
1hN02:r = 0.46
1 hCO:r = 0.32
1 hS02:r = 0.21
24hPM25:r = 0.43
Components: r ranged from
0.23-0.51
Effect Estimates (95% Cl)
PM Increment: 8.4 pg/m3
RR Estimate [Cl]:
Adj for weather and gaseous pollutants
BCC 5-day avg: 1.14 [1.01, 1.28]
BCC 6-day avg: 1.17 [1.03,1. 33]
TS 5-day avg: 1.14 [1.05,1.23]
TS 6-day avg: 1.15 [1.06,1.25]
Boys-adj for weather
UCC 1-day avg: 1.08 1.01,1.16
UCC 2-day avg: 1.08 0.99,1. 17:
BCC 1-day avg: 1.06 1.00,1.14
BCC 2-day avg: 1.06 0.98,1. 14:
TS 1-day avg: 1.08 [1.03,1. 12]
TS 2-day avg: 1.07 [1.01, 1.13]
Girls-adj for weather
UCC 1-day avg: 1.07 0.97,1.18
UCC 2-day avg: 1.1 6 1.03,1.3f
BCC 1-day avg: 0.98 0.90,1.07'
BCC 2-day avg: 1.05 0.94,1.16
TS 1-day avg: 1.00 [0.94,1.06
TS 2-day avg: 1.05 [0.98,1. 13
Notes: The author also provides RR using UCC, BCC, and
TS analysis for female and male groups for day 3-7, yielding
similar results
PM Increment: 5
RR Estimate [Lower Cl, Upper Cl]
All Respiratory Outcomes: 1.003 [0.982, 1.025]
URI: 1.013 [0.987, 1.039]
Asthma: 0.998 [0.987 1.039]
Pneumonia: 0.975 [0.940, 1.011]
COPD: 0.948 [0.897, 1.003]
December 2009
E-259
-------
Reference
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Peng et al. (2008,
1568501
Period of Study: Jan
1999-Dec2005
Location: 108 U.S. counties
in the following states:
Alabama, Arizona, California,
Colorado, Connecticut,
District of Columbia, Florida,
Georgia, Idaho, Illinois,
Indiana, Kentucky, Louisiana,
Maine, Maryland,
Massachusetts, Michigan,
Minnesota, Missouri, Nevada,
New Hampshire, New Jersey,
New Mexico, New York, North
Carolina, Ohio, Oklahoma,
Oregon, Pennsylvania, Rhode
Island, South Carolina,
Tennessee, Texas, Utah,
Virginia, Washington, West
Virginia, Wisconsin
Outcome (ICD-9): Emergency
hospitalizations for respiratory
disease, including COPD (490-492)
and respiratory tract infections
(464-466, 480 - 487)
Age Groups: 65 + yr, 65-74, ,75 +
Study Design: Time series
N: Approximately 12 million Medicare
enroilees (1.4 million RD admissions)
Statistical Analyses: Two-stage
Bayesian hierarchical models: Over
dispersed Poisson models for county-
specific data. Bayesian hierarchical
models to obtain national avg
estimate
Covariates: Day of the week, age-
specific intercept, temperature, dew
point temperature, calendar time,
indicator for age of 75 yr or older.
Some models were adjusted for
PM25.
Dose-response Investigated: No
Statistical Package: R version 2.6.2
Lags Considered: 0-2 days
Pollutant: PM10.2.5
Averaging Time: 24 h
Mean (IQR): All counties
assessed: 9.8 (6.9-15.0)
Counties in Eastern U.S.: 9.1
(6.6-13.1)
Counties in Western U.S.: 15.4
(10.3-21.8)
Monitoring Stations: At least 1
pair of co-located monitors
(physically located in the same
place) for PMi0 and PM25 per
county
Copollutant (correlation):
PM25:r = 0.12
PM10:r = 0.75
Other variables: Median within-
county correlations between
monitors: r = 0.60
PM Increment: 10|jg/m
Percentage change [95% Cl]: Respiratory disease (RD):
Lag 0 (unadjusted for PM25): 0.33 [-0.21, 0.86]
Lag 0 (adjusted for PM25): 0.26 [-0.32, 0.84]
Most values NR (see note)
Notes: Fig 3: Percentage change in emergency hospital
admissions for RD per 10 pg/m increase in PM (single
pollutant model and model adjusted for PM25 concentration)
Fig 4: Percentage change in emergency hospital
admissions rate for CVD and RD per a 10 pg/m3 increase in
PM10.25 (0-2 day lags, Eastern vs.. Western USA)
Reference: Slaughter et al.
(2005, 0738541
Period of Study:
Jan 1995-Jun 2001
Location: Spokane, WA
Hospital Admissions and ED visits
Outcome: All respiratory (460-519)
Asthma (493)
COPD (491,492, 494,496)
Pneumonia (480-487)
Acute URI not including colds and
sinusitis (464, 466, 490)
Age Groups: All, 15+ yr for COPD
Study Design: Time series
N: 2373 visit records
Statistical Analyses: Poisson
regression, GLM with natural splines.
For comparison also used GAM with
smoothing splines and default
convergence criteria.
Covariates: Season, temperature,
relative humidity, day of week
Season: All
Dose-response Investigated?: No
Statistical Package: SAS, SPLUS
Lags Considered: 1 -3 days
Pollutant: PMi0.25 (fjg/m )
Averaging Time: 24 h avg
Range (90% of
Concentrations): Reported for
PM25and PM10only
Monitoring Stations: 1
Copollutant (correlation):
PM,0-2.5
PM1r = 0.19
PM25r = 0.31
PM10r = 0.94
CO r = 0.32
Temperature r = 0.11
PM Increment: 25 pg/m
RR Estimate [Lower Cl, Upper Cl] I
ER visits:
PM,0-2.5
All Respiratory
Lag 1:1.01 [0.98,1.04]
Lag 2:1.01 [0.98,1.04]
Lag 3:1.02 [0.99,1.05]
Acute Asthma
Lag 1:1.03 [0.98,1.08]
Lag 2:1.01 [0.96,1.07]
Lag 3: 0.99 [0.94, 1.05]
COPD (adult)
Lag 1:1.01 [0.93,1.09]
Lag 2: 0.98 [0.90, 1.06]
Lag 3:1.02 [0.95,1.10]
December 2009
E-260
-------
Reference
Reference: Tecer et al.
(2008, 1800301
Period of Study:
Dec 2004-Oct 2005
Location: Zonguldak, Turkey
Reference: Yang etal.,
(2004, 087488)
Period of Study:
Jun1995-MaM999
Location: Vancouver area,
British Columbia
Design & Methods
Outcome: ED visits for respiratory
problems (ICD-9 470-478, 493)
Study Design: Bidirectional Case-
crossover
Covariates: Daily meteorological
parameters
Statistical Analysis: Conditional
logistic regression
Statistical Package: Stata
Age Groups: 0-14 yr
Hospital Admissions
Outcome (ICD-9): Respiratory
diseases (460-51 9), pneumonia only
(480-486), asthma only (493)
Age Groups: 0-3 yr
Study Design: Case control,
bidirectional case-crossover (BCC),
and time series (TS)
N: 1610 cases
Statistical Analyses: Chi-square
test, Logistic regression, GAM (time-
series), GLM with parametric natural
cubic splines
Covariates: Gender, socioeconomic
status, weekday, season, study yr,
influenza epidemic month
Season: Spring, summer, fall, winter
Dose-response Investigated? No
Statistical Package: SAS (Case
control and BCC), S-Plus (TS)
Lags Considered: 0-7 days
Concentrations!
Pollutant: PM10.2.5
Averaging Time: NR
Mean, Unit: 24.3 pg/m3
Range (Min, Max): 4, 195.8
Copollutant (correlation):
PM2.5/PM10.2,
Mean: 1.49
Range: 0.21, 7.53
Pollutant: PM10.2.5 (|jg/m3)
Averaging Time: 24 h
Mean (min-max):
5.6 (0-24.6)
SD = 3.6
Monitoring Stations: NR (data
obtained from Greater
Vancouver Regional District Air
Quality Dept)
Copollutant (correlation):
PM25: r = 0.39
PM10:r = 0.83
CO1 r = 0 33
03:r = -0.16
N02:r = 0.37
S02:r = 0.54
Effect Estimates (95%
Increment: 10 pg/m3
Odds Ratio (96% Cl)
Asthma
Lag 0:1. 18 (1.01-1.39)
Lag 1:0.92 0.78-1.08
Lag 2: 0.98 0.84-1.15
Lag 3: 1.11 (0.97-1. 27)
Lag 4: 1.17 (1.05-1.31)
Allergic Rhinitis with Asthma
Lag 0:0.96 (0.88-1. 04)
Lag 1:1. 08 0.99-1.18
Lag 2: 0.93 0.86-1.02
Lag 3: 0.94 (0.86-1. 03)
Lag 4: 1.10 (1.03-1. 18)
Allergic Rhinitis
Lag 0:1. 06 (0.95-1. 19)
Lag 1:1. 17 1.04-1.31
Lag 2: 0.92 0.84-1.02
Lag 3: 0.99 (0.91-1.08)
Lag 4: 1.15 (1.06-1. 25)
Upper Respiratory Disease
Lag 0:0.80 (0.54-1. 19)
Lag 1:1. 22 0.92-1.61
Lag 2: 0.97 0.70-1.33
Lag 3: 0.94 (0.66-1. 33)
Lag 4: 1.08 (0.88-1. 32)
Lower Respiratory Disease
Lag 0:0.90 (0.71-1. 16)
Lag 1:1. 20 0.97-1.50
Lag 2: 1.00 0.84-1.19
Lag 3: 1.26 (1.08-1. 47)
Lag 4: 1.02 (0.93-1. 13)
PM Increment: 4.2 pg/m3 (IQR)
OR Estimate [Cl]:
3-day lag
1.12 [0.98,1.28]
Adj for gaseous pollutants: 1.22 [1.02,1.45
Cl)
]
Notes: Author states that ORs for PM10.2.5 increased with lag
time up to 3 days for both single and multiple-pollutant
models. More adjusted ORs and RRs are provided in Fig 1.
All units expressed in pg/m unless otherwise specified.
December 2009
E-261
-------
Table E-14. Short-term exposure-respiratory-ED/HA-PWhs (including PM components/sources).
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Andersen et al. (2008,
1896511
Period of Study: May 2001-Dec 2004
Location: Copenhagen, Denmark
Outcome (ICD-10): RD, including
chronic bronchitis (J41-42),
emphysema (J43), other chronic
obstructive pulmonary disease (J44),
asthma (J45), and status asthmaticus
(J46). Pediatric hospital admissions for
asthma (J45) and status asthmaticus
(J46).
Age Groups: > 5-18 yr (asthma)
Study Design: Time series
N:NR
Statistical Analyses: Poisson GAM
Covariates: Temperature, dew-point
temperature, long-term trend,
seasonality, influenza, day of the week,
public holidays, school holidays (only
for 5-18 yr olds), pollen (only for
pediatric asthma outcome)
Season: NR
Dose-response Investigated: No
Statistical Package: R statistical
software (gam procedure, mgcv
package)
Lags Considered: Lag 0-5 days, 4-day
pollutant avg (lag 0-3) for CVD, 5-day
avg (lag 0-4) for RD, and a 6-day avg
(lag 0-5) for asthma.
Pollutant: PM25
Averaging Time: 24 h
Mean |jg/m3(SD): 10(5)
Median: 9
IQR:7-12
99th percentile: 28
Monitoring Stations: 1
Copollutant (correlation):
NCtot:r = 0.40
NC100:r = 0.29
NCa12:r = 0.07
Nca23:r = -0.25
NCa57:r = 0.51
NCa212:r = 0.82
PM10:r = 0.80
CO: r = 0.46
N02:r = 0.42
:r = 0.40
NCycurbside: r = 0.28
03:r = -0.20
Other variables:
Temperature: r = -0.01
Relative humidity: r = 0.21
PM Increment: 5 pg/m (IQR)
Relative risk (RR) Estimate [Cl]: RD
hospital admissions (5-day avg, lag
0-4), age 65+:
One-pollutant model: 1.00 [0.95-1.00]
Adj for NCtot: 1.00 [0.95-1.06]
Asthma hospital admissions (6-day avg
lag 0-5), age 5-18:
One-pollutant model: 1.15 [1.00-1.32]
Adj for NCtot: 1.13 [0.98-1.32]
Estimates for individual day lags
reported only in Fig form (see notes):
Notes: RD: No statistically or marginally
significant associations. Positive
associatons at Lag 4-5.Asthma: Wide
confidence intervals make interpretation
dificult. Positive associations at Lag 1,
2,3.
Reference: Babin et al. (2007, 0932501
Period of Study: Oct 2001-Sep 2004
Location: Washington, DC
ED Visit/Admissions
Outcome: Asthma-493
Age Groups: 1-17 yr,1-4, 5-12,13-17
Study Design: Time-series
N:NR
Statistical Analyses: Poisson
regression, spline w/12 knots to adjust
for long term trend
Covariates: Temperature, mold, pollen,
seasonal trends,
Season: All
Dose-response lnvestigated?No
Statistical Package: STATA
Lags Considered: 0-4
Pollutant: PM25
Averaging Time: 24 h
Mean: "low, never reached code red"
Percentiles: NR
Range (Min, Max): NR
Monitoring Stations: 3
Copollutant (correlation): NR
PM Increment: 1 pg/m
%Change ED Visits
Ages 5-12:
-0.2 (-0.6,0.2), lag 0
% Change ED Admissions:
Ages 5-12:
-0.4 (-1.6,0.8), lag 0
Ages 1-17:
0.2 (-0.6,1.1), lag 0
AR Estimate [Lower Cl, Upper Cl] lag:
NR
Notes: No significant interactions
between PM and 03 or other covariates
were observed.
December 2009
E-262
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Barnett et al. (2005,
0873941
Period of Study: 1998-2001
Location: 5 Australian cities (Brisbane,
Canberra, Melbourne, Perth, and
Sydney) and 2 New Zealand cities
(Auckland, Christchurch)
Outcome (ICD: NR): All respiratory
admissions (including asthma,
pneumonia, and acute bronchitis)
Age Groups: Children aged <1 yr, 1-
4yr, and 5-14 yr
Study Design: Matched case-
crossover
N: ~2.4 million children <15 yr old
Statistical Analyses: Random effects
meta-analysis
Covariates: Temperature, current
minus previous day's temperature,
relative humidity, pressure, extremes
of hot and cold, day of the week,
public holiday, and day after public
holiday
Season: Warm (Nov-Apr) and Cool
(May-Oct)
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: NR
Pollutant: PM25
Averaging Time: 24 h
Mean (min-max):
Auckland (A): 11.0(2.1-37.6)
Brisbane (B): 9.7 (3.2-122.8)
Canberra (Ca): NR
Christchurch (Ch): NR
Melbourne (M): 8.9 (2.8-43.3)
Perth (P): 8.1 (1.7-29.3)
Sydney (S): 9.4 (2.4-82.1)
Monitoring Stations: 1-3 per city
Copollutant: NR
PM Increment: 3.8 pg/m (IQR)
Percent Increase Estimate [Cl]:
Pneumonia S Acute Bronchitis:
Single Pollutant Model
<1yr(B,M,P,S): 1.7 [0.0,3.4
1-4 yr(B,M,P,S): 2.4 [0.1,4.7
Matched Multipollutant Model
1-4 yr with 1-hS02(B,S): 1.9 [-1.7,5.6]
1-4 yr with temp (B,M,P,S): 2.3 [-0.4,5.1]
Respiratory Admissions:
Single Pollutant Model
<1yr(B,M,P,S): 2.4 [1.0,3.8]
1-4 yr(B,M,P,S): 1.7 [0.7,2.7]
Matched Pollutant Model
<1yr with 1-hS02(B,S): 3.1 [0.5,5.7]
<1 yr with temp (B,M,P,S): 1.8 [0.2,3.4]
1-4 yrwith PM,0 (B,M,P,S): 2.9 [0.2,5.6]
1-4 yr with 1-hS02(B,S): 1.3 [-1.8,4.4]
1-4 yr with 1-h N02 (B,M,P,S):
-1.5 [-3.2,0.2]
1-4 yrwith temp (B,M,P,S,):
1.5 [-0.2,3.1]
Reference: Bell et al. (2008, 0912681
Period of Study: 1995-2002
Location: Taipei, Taiwan
Outcome (ICD-9): Hospital admissions
for asthma (493), and pneumonia (486).
Age Groups: All
Study Design: Time series
N: 19,966 hospital admissions for
pneumonia, and 10,231 for asthma
Statistical Analyses: Poisson
regression
Covariates: Day of the week, time,
apparent temperature, long-term trends,
seasonality
Season: All
Dose-response Investigated: No
Statistical Package: NR
Lags Considered: lags 0-3 days, mean
of lags 0-3
Pollutant: PM25
Averaging Time: 24 h
Mean (range
IQR): 31.6 (0.50-355.0
20.2)
Monitoring Stations: 2
Copollutant (correlation): NR
PM Increment: 20 pg/m (near IQR)
Percentage increase estimate [95% Cl]:
Asthma: L0:0.46 (-2.41, 3.42)
L1:-1.36 (-4.33, 1.71)
L2:-0.83 (-3.67, 2.10)
L3:-0.78 (-3.63, 2.16)
LOS:-1.75 (-6.21, 2.92)
Pneumonia: LO: 0.06 (-2.74, 2.94)
L1:0.34 (-2.446, 3.20)
L2: -0.59 (-3.38, 2.29)
L3:-0.44 (-3.22, 2.41)
LOS: -0.61 (-4.87, 3.85)
Reference: Bell et al. (2008, 0912681
Period of Study: 1999-2005
Location: 202 U.S. counties
Outcome (ICD-9): COPD (490-492),
respiratory tract infections (464-466,
480-487)
Age Groups: 65+
Study Design: Time series
N:NR
Statistical Analyses: Two-stage
Bayesian hierarchical model to find
national avg
First stage: Poisson regression (county-
specific)
Covariates: Day of the week,
temperature, dew point temperature,
temporal trends, indicator for persons
75+ yr, population size
Season: All, Jun-Aug (Summer),
Sep-Nov (Fall), Dec-Feb (Winter),
Mar-May (Spring)
Dose-response Investigated: No
Pollutant: PM25
Averaging Time: 24 h
Mean (ug/m3):
Descriptive information presented in Fig
S2 (boxplots):
IQR: 8.7 pg/m3
Monitoring Stations: NR
Copollutant (correlation): NR
PM Increment: 10 pg/m
Percent increase [95% PI]:
Respiratory admissions:
Lag 0 (all seasons): 0.22 [-0.12-0.56]
Lag 0 (winter, national): 1.05 [0.29-1.82]
Lag 0 (winter, northeast):
1.76 [0.60-2.93]
Lag 0 (winter, southeast):
0.59 [-1.35-2.58]
Lag 0 (winter, northwest):
-0.07 [-6.74-7.08]
Lag 0 (winter, southwest):
0.03[-1.25-1.34]
Lag 0 (spring, national):
0.31 [-0.47-1.11]
Lag 0 (spring, northeast):
0.34 [-0.66-1.34]
Lag 0 (spring, southeast):
-0.06 [-1.77-1.68]
Lag 0 (spring, northwest):
-8.52 [-25.62-12.51]
Lag 0 (spring, southwest):
1.87 [-2.00-5.90]
Lag 0 (summer, national):
-0.62 [-1.33-0.09]
December 2009
E-263
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Statistical Package: NR
Lags Considered: 0-2 day lags
Lag 0 (summer, northeast):
-0.8 [-1.65-0.07]
Lag 0 (summer, southeast):
-0.15 [-1.88-1.61]
Lag 0 (summer, northwest):
0.25 [-21.46-27.96]
Lag 0 (summer, southwest):
0.64 [-5.38-7.04]
Lag 0 (fall, national):
0.02 [-0.63-0.67]
Lag 0 (fall, northeast):
-0.01 [-0.87-0.85]
Lag 0 (fall, southeast):
-0.58 [-2.06-0.91]
Lag 0 (fall, northwest):
-1.38[-11.84-10.32]
Lag 0 (fall, southwest):
1.77 [-0.73-4.33]
Lag 1 (all seasons): 0.05 [-0.29-0.39]
Lag1
Lag1
winter): 0.50 [-0.27-1.27]
spring):-0.24 [-1.01-0.53]
Lag 1 (summer): 0.28 [-0.39-0.95]
Lag1
Lag 2
fall): 0.15 [-0.49-0.79]
all seasons): 0.41 [0.09-0.74]
Lag 2 (winter, national): 0.72 [0.01-1.43]
Lag 2 (winter, northeast):
0.79 [-0.21-1.80]
Lag 2 (winter, southeast):
0.4 [-1.45, 2.27]
Lag 2 (winter, northwest):
-0.06 [-6.52-6.85]
Lag 2 (winter, southwest):
1.2 [-0.10-2.52]
Lag 2 (spring, national):
0.35 [-0.29-0.99]
Lag 2 (spring, northeast):
0.04 [-0.88-0.97]
Lag 2 (spring, southeast):
0.75 [-0.82-2.34]
Lag 2 (spring, northwest):
2.29 [-14.26-22.03]
Lag 2 (spring, southwest):
1.05 [-2.18-4.39]
Lag 2 (summer, national): 0.57 [-
0.07-1.23]
Lag 2 (summer, northeast):
0.77[-0.01-1.56]
|Lag 2 (summer, southeast):
-0.52 [-2.07-1.06]
Lag 2 (summer, northwest):
0.74 [-18.73-24.86]
Lag 2 (summer, southwest):
2.41 [-2.61-7.69]
Lag 2 (fall, national):
0.39 [-0.22-1.01]
Lag 2 (fall, northeast):
0.12 [-0.82-1.07]
Lag 2 (fall, southeast):
0.14[-1.29-1.59]
Lag 2 (fall, northwest):
-0.74 [-10.08-9.58]
Lag 2 (fall, southwest):
0.97[-1.36-3.36]
Reference: Bell et al. (2009,1910071
Period of Study: 1999-2005
Location: 168 U.S. Counties
Outcome: Respiratory hospital
admissions
Study Design: Retrospective Cohort
Covariates: Socio-economic
conditions, long term temperature
Statistical Analysis: Bayesian
hierarchical model
Age Groups: >65
Pollutant: PM25
Averaging Time: 24 h
Mean (SD) Unit: NR
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: 20% of the population
acquiring air conditioning
Percent Change (96% Cl) in
community-specific PM health effect
estimates for respiratory hospital
admissions
Any AC, including window units
Yearly health effect: 44.5 (-87.5-176)
Summer health effect: -74.8 (-417-267)
Winter health effect: -32.5 (-245-180)
Central AC
Yearly health effect: 27.6 (-46.7-102)
Summer health effect: -38.6 (-160-82.6)
Winter health effect: 43.8 (-125-213)
December 2009
E-264
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Bell et al. (2009,1910071
Period of Study: 1999-2005
Location: 168 U.S. Counties
Outcome: Respiratory HA
Age Groups: 65+
Study Design: Time series
N:NR
Statistical Analyses: Bayesian
Hierarchical Regression
Covariates: Time trend, day of week,
seasonality, dew point, temperature
Statistical Package: NR
Lags Considered: 0-2
Pollutant: PM25
Averaging Time: Daily
Mean:
EC: 0.715
Ni: 0.002
V: 0.003
Min:
EC: 0.309
Ni: 0.003
V: 0.001
Max:
EC: 1.73
Ni: 0.021
V: 0.010
Interquartile Range:
EC: 0.245
Ni: 0.001
V: 0.001
Interquartile Range of Percents:
EC: 1.7
Ni:0.01
V: 0.01
Monitoring Stations: NR
Copollutant: Al, NH4+, As, Ca, Cl, Cu,
EC, CMC, Fe, Pb, Mg. Ni, N0r, K, Si,
Na+, S04=, Ti, V, Zn
Co-pollutant Correlation:
Ni, V: 0.48
V, EC: 0.33
Ni, EC: 0.30
Note: Pollutant concentrations available
for all fractions of PM2 5
PM Increment: Interquartile Range in
the fraction of PM2 5
Percent Increase (Lower Cl, Upper
Cl):
EC: 511 (80.7, 941), lag 0
EC+Ni: 399 (-45.1, 843), lag 0
EC + V: 386 (-74.8, 846), lag 0
EC + Ni, V: 362 (-98.0, 823), lag 0
Ni: 223 (36.9, 410), lagO
Ni +EC: 176 (-18.7, 370), lag 0
Ni + V:151 (-78.4, 381), lag 0
Ni + EC, V: 136 (-94.9, 368), lag 0
V: 392 (46.3, 738), lag 0
V + EC: 279 (-93.2, 651), lagO
V+Ni: 230(-193.7, 653), lagO
V+EC, Ni: 140 (-300, 579), lagO
EC:-1.5 (80.7, 941), lag 1
EC: 17.5 (-22.3, 57.3), lag 2
Ni:-7.2 (-66.6. 52.1), lag 1
Ni: -4.9 (-22.3, 12.5), lag 2
V:-19.6 (-127, 88.3), lag 1
V: 10.5 (-21.5, 42.4), lag 2
HS education: -77.8 (-390, 234), lag 0
median income: 45.9 (-411, 503), lagO
Percent black: -53.1 (-557, 451), lag 0
Percent living in urban area: -41.9 (-
774.7, 691), lag 0
Population:-22.9 (-121, 75.3), lag 0
Notes: Interquartile ranges in percent
HS education, median income, percent
black, percent living in urban area, and
population are 5.2 %, $9,223,17.3%,
11.0%, and 549,283 respectively.
Reference: Chardon et al. (2007,
0913081
Period of Study: 2000-2003
Location: Greater Paris Area, France
Doctors house calls
Outcome (ICPC2): Asthma (R96),
Upper respiratory disease (URD R07,
R21.R29, R75, R76, R02), Lower
respiratory disease (LRD, R05, R78)
Age Groups: All
Study Design: Time series
N: 8027 for asthma
52928 for LRD
74845 for URD
Statistical Analyses: Quasi-Poisson,
GAM, parametric penalized spline
smoothers.
Covariates: Lagged and current
temperature, humidity, long term trends,
seasonality, pollen counts, influenza
epidemic, days of the week, holidays,
bank holidays
Season: All
Dose-response Investigated? No
Statistical Package: R
Lags Considered: 0-3 days
Pollutant: PM25
Averaging Time: Mean of the daily
means
Mean (SD): 14.7(7.34) pg/m3
Percentiles: 25th: 9.5
SOth(Median): 12.9
75th: 18.2
Range (Min, Max): (3, 69.6)
Monitoring Stations: 1-4
Copollutant:
PM,0:r = 0.95
N02:r = 0.68
PM Increment: 10 pg/m3
% Change, lag 0-3-day avg
URD
6.0(3.1,9.1)
LRD
5.8 (2.8, 8.9)
Asthma
4.4 (-1.3, 10.4)
December 2009
E-265
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Chen et al. (2005, 0875551
Period of Study: Jun 1995-Mar 1999
Location: Vancouver area, BC
Hospital Admissions
Outcome (ICD-9): Acute respiratory
infections (460-466), upper respiratory
tract infections (470-478), pneumonia
and influenza (480-487), COPD and
allied conditions (490-496), other
respiratory diseases (500-519)
Age Groups: >65 yr
Study Design: Time series
N: 12,869
Statistical Analyses: GLM
Covariates: Temp and relative humidity
Season: NR
Dose-response Investigated? No
Statistical Package: S-Plus
Lags Considered: 1-, 2-, 3-, 4-, 5-, 6-,
and 7-day avg
Pollutant: PM25
Averaging Time: 24 h
Mean (min-max):
7.7 (2.0-32.0)
SD = 3.7
Monitoring Stations: 13
Copollutant (correlation):
PM,o:r = 0.83
PMi0.25:r = 0.38
COM: r = 0.39
CO: r = 0.23
03:r = -0.01
N02:r = 0.36
S02:r = 0.42
Other variables:
Mean temp: r = 0.41
Pel humidity: r = -0.23
PM Increment: 4.0 pg/rri (IQR)
RR Estimate [Cl]:
Adj for weather conditions
Overall admission
1-day avg: 1.02 [0.99,1.05]
2-day avg: 1.02 [0.99,1.06]
3-day avg: 1.02 [0.98,1.05]
Adj for weather conditions and
co pollutants
Overall admission
1-day avg: 1.01 [0.98,1.06]
2-day avg: 1.01 [0.98,1.05]
3-day avg: 1.00 [0.96,1.04]
Notes: PR's were also provided for lags
4-7 in Table 3, yielding similar results
Reference: Chimonas and Gessner
(2007, 0932611
Period of Study: Jan 1999-Jun 2003
Location: Anchorage, Alaska
Outcome (ICD-9): Asthma (493.0-
493.9)
Lower respiratory illness-LRI (466.1,
466.0, 480-487, 490, 510-511)
Inhaled quick-relief medication
Steroid medication
Age Groups: <20 yr old
Study Design: Time series
N: 42,667 admissions
Statistical Analyses: GEE for
multivaliable modeling
Covariates: Season, serial correlation,
yr, weekend, temperature, precipitation,
and wind speed
Season: NR
Dose-response Investigated? No
Statistical Package: SPSS (dataset),
SAS (analysis)
Lags Considered: 1 day and 1 wk
Pollutant: PM25
Averaging Time: 24 h and 1 wk
Mean (min-max):
Daily: 6.1 (0.5-69.8)
Weekly: 5.8 (1.8-45.0)
Monitoring Stations: NR
Copollutant: N/A
PM Increment: 5 pg/m
RR Estimate [Cl]:
Same Day
Outpatient Asthma: 0.992 [0.964,1.024]
Outpatient LRI: 0.952 [0.907,1.001]
Inpatient Asthma: 0.936 [0.798,1.098]
Inpatient LRI: 0.919 [0.823,1.027]
Inhaled Steroid Prescriptions:
0.988 [0.902,1.083]
Quick-relief Medication:
0.962 [0.901,1.028]
Weekly (median increase)
Outpatient Asthma: 0.983 [0.935,1.038]
Outpatient LRI: 0.969 [0.874,1.075]
Inpatient Asthma: 0.754 [0.513.1.109]
Inpatient LRI: 0.943 [0.715,1.245]
Inhaled Steroid Prescriptions:
1.018 [0.883,1.175]
Quick-relief Medication:
0.978 [0.882,1.087]
Reference: Delfmo et al. (2009,
1919941
Period of Study: Oct 2003-Nov 2003
Location: Southern California
Outcome: Respiratory hospital
admissions
Study Design: Time series
Statistical Analysis: Poisson
regression with GEE
Age Groups: All
Pollutant: PM25
Averaging Time: Hourly
Mean (SD) Unit by county:
Los Angeles
Before Fires: 27.2 (12.4) pg/m3
During Fires: 54.1 (21.0)|jg/m3
After Fires: 15.9(5.5) pg/m
Orange
Before Fires: 23.2 (9.6) pg/m3
During Fires: 64.3 (26.5) pg/m3
After Fires: 15.5 (10.2) pg/m3
Riverside
Before Fires: 32.7 (14.7) pg/m3
During Fires: 42.1 (25.5) pg/m
After Fires: 16.9 (10.2) pg/m3
San Bernadino
Before Fires: 35.7 (16.6) pg/m3
During Fires: 45.3 (28.7) pg/m3
After Fires: 18.5(8.3) pg/m
San Diego
Before Fires: 18.5 (6.7) pg/m3
During Fires: 76.1 (66.6) pg/m3
After Fires: 14.2 (7.2) pg/nf
Ventura
Increment: 10|jg/m
Relative Rate (Min Cl, Max Cl)
All Respiratory, All Ages: All Periods:
1.009(0.999-1.018)
Pre-Wildfire: 1.022 (1.004-1.040)
Wildfire: 1.028 (1.014-1.041), p = 0.639
Post-Wildfire: 0.999 (0.968-1.031),
p = 0.198
All Respiratory, Ages 0-4: All Periods:
0.994(0.967-1.021)
Pre-Wildfire: 0.982 (0.921-1.046)
Wildfire: 1.045 (1.010-1.082), p = 0.103
Post-Wildfire: 0.894 (0.807-0.991),
p = 0.126
All Respiratory, Ages 5-19: All Periods:
1.014(0.983-1.046)
Pre-Wildfire: 1.026 (0.946-1.113)
Wldfire: 1.027 (0.984-1.076), p = 0.990
Post-Wildfire: 0.958 (0.852-1.077),
p = 0.354
All Respiratory, Ages 20-64: All Periods:
December 2009
E-266
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Before Fires: 18.4 (8.3) pg/nf
During Fires: 50.1 (50.5) pg/m3
After Fires: 12.9 (4.3) pg/rri
Copollutant (correlation): NR
1.015(1.002-1.029)
Pre-Wildfire: 1.036 (1.007-1.066)
Wildfire: 1.024(1.005-1.044), p = 0.534
Post-Wildfire: 1.007 (0.960-1.056),
p = 0.315
All Respiratory, Ages 65-99: All Periods:
1.009(0.996-1.022)
Pre-Wildfire: 1.022 (0.994-1.050)
Wildfire: 1.030(1.011-1.049), p = 0.649
Post-Wildfire: 1.024 (0.967-1.074),
p = 0.932
Asthma, All Ages, Male and Female: All
Periods: 1.022 (1.001-1.042)
Pre-Wildfire: 0.998 (0.949-1.050)
Wldfire: 1.048 (1.021-1.076), p = 0.097
Post-Wildfire: 0.986 (0.910-1.068),
p = 0.792
Asthma, All Ages, Male: All Periods:
1.010(0.980-1.040)
Pre-Wldfire: 1.021 (0.944-1.106)
Wldfire: 1.031 (0.990-1.073), p = 0.848
Post-Wildfire: 1.063 (0.948-1.192),
p = 0.553
Asthma, All Ages, Female: All Periods:
1.029(1.001-1.058)
Pre-Wldfire: 0.979 (0.913-1.050)
Wldfire: 1.059(1.022-1.097), p = 0.056
Post-Wldfire: 0.928 (0.829-1.037),
p = 0.412
Asthma, Ages 0-4, Males and Females:
All Periods: 0.996 (0.947-1.048)
Pre-Wldfire: 0.924 (0.824-1.035)
Wldfire: 1.083 (1.021-1.149), p = 0.017
Post-Wldfire: 0.924 (0.767-1.113),
p = 0.999
Asthma, Ages 0-4, Males: All Periods:
1.018(0.963-1.076)
Pre-Wldfire: 0.942 (0.815-1.089)
Wldfire: 1.086 (1.016-1.162), p = 0.101
Post-Wldfire: 1.057 (0.839-1.332),
p = 0.380
Asthma, Ages 0-4, Females: All
Periods: 0.937 (0.845-1.040)
Pre-Wldfire: 0.880 (0.706-1.099)
Wldfire: 1.073 (0.965-1.194), p = 0.116
Post-Wldfire: 0.699 (0.515-0.949),
p = 0.214
Asthma, Ages 5-19, Males and
Females: All Periods:
1.006(0.966-1.048)
Pre-Wldfire: 1.045 (0.936-1.167)
Wldfire: 0.999 (0.935-1.068), p = 0.492
Post-Wldfire: 0.918 (0.788-1.069),
p = 0.198
Asthma, Ages 5-19, Males: All Periods:
0.991 (0.935-1.051)
Pre-Wldfire: 1.034 (0.892-1.198)
Wldfire: 0.969 (0.883-1.064), p = 0.462
Post-Wldfire: 0.979 (0.806-1.189),
p = 0.671
Asthma, Ages 5-19, Females: All
Periods: 1.026 (0.964-1.092)
Pre-Wldfire: 1.065 (0.901-1.260)
Wldfire: 1.033 (0.943-1.132), p = 0.768
Post-Wldfire: 0.831 (0.640-1.079),
p = 0.136
Asthma, Ages 20-64, Males and
Females: All Periods: 1.043(1.012-
December 2009
E-267
-------
Reference Design & Methods Concentrations1 Effect Estimates (95% Cl)
1076)
Pre-Wildfire: 1.037 (0.957-1.123)
Wildfire: 1.041 (0.995-1.090), p = 0.931
Post-Wildfire: 1.000 (0.882-1.132),
p = 0.624
Asthma, Ages 20-64, Males: All Periods:
1.013(0.954-1.077)
Pre-Wildfire: 1.159 (0.996-1.349)
Wildfire: 0.939 (0.837-1.053), p = 0.026
Post-Wildfire: 1.275 (1.020-1.595),
p = 0.486
Asthma, Ages 20-64, Females: All
Periods: 1.052 (1.015-1.090)
Pre-Wldfire: 0.995 (0.904-1.096)
Wldfire: 1.064 (1.014-1.116), p = 0.247
Post-Wildfire: 0.908 (0.780-1.056),
p = 0.310
Asthma, Ages 65-99, Males and
Females: All Periods: 1.027(0.974-
1.082)
Pre-Wldfire: 0.951 (0.849-1.064)
Wldfire: 1.101 (1.030-1.178), p = 0.032
Post-Wildfire: 1.168 (0.967-1.412),
p = 0.072
Asthma, Ages 65-99, Males: All Periods:
1.046(0.957-1.142)
Pre-Wldfire: 0.948 (0.804-1.116)
Wldfire: 1.185 1.077-1.305), p = 0.029
Post-Wildfire: 0.902 (0.629-1.294),
p = 0.804
Asthma, Ages 65-99, Females: All
Periods: 1.018 (0.958-1.081)
Pre-Wldfire: 0.947 (0.813-1.102)
Wldfire: 1.065 (0.977-1.162), p = 0.195
Post-Wildfire: 1.263 (1.024-1.557),
p = 0.032
Acute Bronchitis and Bronchiolitis, All
Ages: All Periods: 1.044(0.990-1.102)
Pre-Wldfire: 1.001 (0.890-1.126)
Wldfire: 1.096 (1.018-1.179),
p = 0.223
Post-Wildfire: 1.031 (0.870-1.222),
p = 0.779
Acute Bronchitis and Bronchiolitis, Ages
0-4: All Periods: 1.017 (0.949-1.089)
Pre-Wldfire: 0.987 (0.847-1.149)
Wldfire: 1.092 (0.997-1.195), p = 0.276
Post-Wildfire: 0.910 (0.700-1.183),
p = 0.588
Acute Bronchitis and Bronchiolitis, Ages
5-19: No Convergence
Acute Bronchitis and Bronchiolitis, Ages
20-64: All Periods: 1.039 (0.912-1.183)
Pre-Wldfire: 1.001 (0.792-1.266)
Wldfire: 1.044 (0.872-1.252), p = 0.778
Post-Wildfire: 1.259 (0.921-1.722),
p = 0.275
Acute Bronchitis and Bronchiolitis, Ages
65-99: All Periods: 1.134 (1039-1.238)
Pre-Wldfire: 1.073 (0.764-1.505)
Wldfire: 1.143(1.032-1.265), p = 0.730
Post-Wildfire: 1.190 (0.865-1.638),
p = 0.652
COPD, Ages 20-99: All Periods: 1.018
(0.994-1.042)
Pre-Wldfire: 1.007 (0.958-1.058)
Wldfire: 1.038 (1.004-1.075), p = 0.320
Post-Wildfire: 1.024 (0.943-1.112),
December 2009 E-268
-------
Reference Design & Methods Concentrations1 Effect Estimates (95% Cl)
p = 0.728
COPD, Ages 20-64: All Periods: 1.022
(0.980-1.066)
Pre-Wildfire: 0.995 (0.916-1.081)
Wildfire: 1.068 (1.009-1.131), p = 0.161
Post-Wildfire: 1.015 (0.893-1.153),
p = 0.728
COPD, Ages 65-99: All Periods: 1.019
(0.992-1.048)
Pre-Wildfire: 1.014 (0.955-1.077)
Wildfire: 1.031 (0.990-1.074), p = 0.660
Post-Wildfire: 1.023 (0.928-1.128),
p = 0.878
Pneumonia, All Ages: All Periods: 1.009
(0.994-1.024)
Pre-Wildfire: 1.045 (0.931-1.180)
Wldfire: 1.028 (1.007-1.050), p = 0.420
Post-Wildfire: 0.980 (0.927-1.035),
p = 0.045
Pneumonia, Ages 0-4: All Periods:
0.995(0.944-1.049)
Pre-Wldfire: 1.048 (0.931-1.180)
Wldfire: 1.018 (0.948-1.092), p = 0.691
Post-Wildfire: 0.823 (0.649-1.044),
p = 0.089
Pneumonia, Ages 5-19: All Periods:
1.031 (0.966-1.098)
Pre-Wldfire: 1.017 (0.882-1.172)
Wldfire: 1.064 (0.990-1.142), p = 0.586
Post-Wildfire: 1.017 (0.767-1.349),
p = 0.998
Pneumonia, Ages 20-64: All Periods:
1.008(0.982-1.035)
Pre-Wldfire: 1.041 (0.982-1.104)
Wldfire: 1.032 (0.994-1.072), p = 0.823
Post-Wildfire: 1.013 (0.913-1.124),
p = 0.633
Pneumonia, Ages 65-99: All Periods:
1.011 (0.993-1.030)
Pre-Wldfire: 1.050 (1.006-1.097)
Wldfire: 1.029 (1.002-1.057), p = 0.445
Post-Wildfire: 0.985 (0.920-1.055),
p = 0.127
Relative Rate (Min Cl, Max Cl) in
relation to pre-wildfire period (1)
All Respiratory, All Ages: Wldfire,
unadjusted for PM25: 0.961 (0.916-
1.008)
Wldfire, adjusted for PM25: 0.903
(0.850-0.960)
Post-wildfire, unadjusted for PM25:
1.143(1.072-1.219)
Post-wildfire, adjusted for PM25:
1.173(1.097-1.253)
All Respiratory, Ages 0-4: Wldfire,
unadjusted for PM25:
0.865 (0.757-0.989)
Wldfire, adjusted for PM25:
0.842 (0.717-0.988)
Post-wildfire, unadjusted for PM25:
1.152(0.957-1.388)
Post-wildfire, adjusted for PM25:
1.162(0.954-1.415)
All Respiratory, Ages 5-19: Wldfire,
unadjusted for PM25:
1.098(0.901-1.324)
Wldfire, adjusted for PM25:
1.087(0.863-1.370)
December 2009 E-269
-------
Reference Design & Methods Concentrations1 Effect Estimates (95% Cl)
Post-wildfire, unadjusted for PM25:
1.373(1.089-1.732)
Post-wildfire, adjusted for PM25:
1.467(1.142-1.883)
All Respiratory, Ages 20-64: Wildfire,
unadjusted for PM25:
0.991 (0.922-1.066)
Wildfire, adjusted for PM25:
0.923(0.843-1.012)
Post-wildfire, unadjusted for PM25:
1.074(0.971-1.188)
Post-wildfire, adjusted for PM25:
1.104(0.992-1.228)
All Respiratory, Ages 65-99: Wldfire,
unadjusted for PM25:
0.932(0.867-1.003)
Wldfire, adjusted for PM25:
0.874 (0.795-0.959)
Post-wildfire, unadjusted for PM25:
1.147(1.045-1.259)
Post-wildfire, adjusted for PM25:
1.193(1.084-1.313)
Asthma, All Ages: Wildfire, unadjusted
for PM25:1.088 (0.965-1.227)
Wldfire, adjusted for PM25:
0.992(0.856-1.149)
Post-wildfire, unadjusted for PM25:
1.264(1.085-1.473)
Post-wildfire, adjusted for PM25:
1.336(1.134-1.573)
Asthma, Ages 0-4: Wldfire, unadjusted
for PM25: 0.806 (0.632-1.029)
Wldfire, adjusted for PM25:
1.282(0.958-1.716)
Post-wildfire, unadjusted for PM25:
1.092(1.759-1.572)
Post-wildfire, adjusted for PM25:
1.133(0.777-1.654)
Asthma, Ages 5-19: Wldfire,
unadjusted for PM25:
1.254(0.999-1.575)
Wldfire, adjusted for PM25:
1.282(0.958-1.716)
Post-wildfire, unadjusted for PM25:
1.564(1.160-2.109)
Post-wildfire, adjusted for PM25:
1.629(1.184-2.243)
Asthma, Ages 20-64: Wldfire,
unadjusted for PM25:
1.273(1.067-1.518)
Wldfire, adjusted for PM25:
1.221 (0.979-1.524)
Post-wildfire, unadjusted for PM25:
1.362(1.043-1.779)
Post-wildfire, adjusted for PM25:
1.486(1.111-1.987)
Asthma, Ages 65-99: Wldfire,
unadjusted for PM25:
0.869(0.657-1.151)
Wldfire, adjusted for PM25:
0.645 (0.450-0.925)
Post-wildfire, unadjusted for PM25:
0.924(0.606-1.408)
Post-wildfire, adjusted for PM25:
1.005(0.650-1.552)
Acute Bronchitis and Bronchiolitis, All
Ages: Wldfire, unadjusted for PM25:
1.143(0.878-1.490)
Wldfire, adjusted for PM25:
0.959(0.696-1.321)
December 2009 E-270
-------
Reference Design & Methods Concentrations1 Effect Estimates (95% Cl)
Post-wildfire, unadjusted for PM25:
1.482(1.042-2.109)
Post-wildfire, adjusted for PM25:
1.580(1.089-2.291)
Acute Bronchitis and Bronchiolitis, Ages
0-4: Wildfire, unadjusted for PM25:
1.128(0.819-1.555)
Wildfire, adjusted for PM25:
0.899(0.607-1.333)
Post-wildfire, unadjusted for PM25:
1.520(0.947-2.440)
Post-wildfire, adjusted for PM25:
1.547 (0.954-2.507)
Acute Bronchitis and Bronchiolitis,
Ages 5-19
No Correlation
Acute Bronchitis and Bronchiolitis,
Ages 20-64: Wldfire, unadjusted for
PM25:1.350 (0.688-2.648)
Wldfire, adjusted for PM25:
1.320(0.608-2.863)
Post-wildfire, unadjusted for PM25:
2.454(1.068-5.640)
Post-wildfire, adjusted for PM25:
2.515(1.055-5.998)
Acute Bronchitis and Bronchiolitis,
Ages 65-99: Wldfire, unadjusted for
PM25:1.166 (0.643-2.115)
Wldfire, adjusted for PM25:
0.934 (0.422-20.66)
Post-wildfire, unadjusted for PM25:
0.911 (0.428-1.942)
Post-wildfire, adjusted for PM25:
0.997 (0.439-2.262)
COPD, Ages 20-99: Wldfire,
unadjusted for PM25:
0.988(0.875-1.115)
Wldfire, adjusted for PM25:
0.913(0.779-1.069)
Post-wildfire, unadjusted for PM25:
1.043(0.885-1.228)
Post-wildfire, adjusted for PM25:
1.064(0.897-1.262)
COPD, Ages 20-64: Wldfire,
unadjusted for PM25:
0.967(0.779-1.201)
Wldfire, adjusted for PM25:
0.873(0.660-1.156)
Post-wildfire, unadjusted for PM25:
1.175(0.862-1.601)
Post-wildfire, adjusted for PM25:
1.311 (0.954-1.802)
COPD, Ages 65-99: Wldfire,
unadjusted for PM25:
1.002(0.869-1.156)
Wldfire, adjusted for PM25:
0.926(0.767-1.117)
Post-wildfire, unadjusted for PM25:
0.985(0.811-1.196)
Post-wildfire, adjusted for PM25:
0.981 (0.798-1.206)
Pneumonia, All Ages: Wldfire,
unadjusted for PM25:
0.943(0.868-1.025)
Wldfire, adjusted for PM25:
0.888 (0.799-0.986)
Post-wildfire, unadjusted for PM25:
1.294(1.158-1.446)
Post-wildfire, adjusted for PM25:
1.318(1.174-1.479)
December 2009 E-271
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Pneumonia, Ages 0-4: Wildfire,
unadjusted for PM25:
0.938(0.705-1.247)
Wildfire, adjusted for PM25:
0.951 (0.678-1.333)
Post-wildfire, unadjusted for PM25:
1.458(0.974-20182)
Post-wildfire, adjusted for PM25:
1.374(0.885-2.133)
Pneumonia, Ages 5-19: Wldfire,
unadjusted for PM25:
0.891 (0.604-1.312)
Wldfire, adjusted for PM25:
0.830(0.541-1.272)
Post-wildfire, unadjusted for PM25:
0.960(0.588-1.569)
Post-wildfire, adjusted for PM25:
0.969(0.578-1.624)
Pneumonia, Ages 20-64: Wldfire,
unadjusted for PM25:
0.927(0.795-1.081)
Wldfire, adjusted for PM25:
0.837(0.690-1.016)
Post-wildfire, unadjusted for PM25:
1.314(1.064-1.622)
Post-wildfire, adjusted for PM25:1.300
(1.047-1.615)
Pneumonia, Ages 65-99: Wldfire,
unadjusted for PM25:
0.959(0.861-1.068)
Wldfire, adjusted for PM25:
0.899(1.782-1.033)
Post-wildfire, unadjusted for PM25:
1.277(1.102-1.481)
Post-wildfire, adjusted for PM25:
1.331 (1.142-1.552)
Reference: Dominici et al. (2006,
Period of Study: 1999-2002
Location: 204 U.S. counties, located
in: Alabama, Alaska, Arizona, Arkansas,
California, Colorado, Connecticut,
Delaware, District of Columbia, Florida,
Georgia, Hawaii, Idaho, Illinois, Indiana,
Iowa, Kansas, Kentucky, Louisiana,
Maine, Maryland, Massachusetts,
Michigan, Minnesota, Mississippi,
Missouri, Nevada, New Hampshire,
New Jersey, New Mexico, New York,
North Carolina, Ohio, Oklahoma,
Oregon, Pennsylvania, Rhode Island,
South Carolina, Tennessee, Texas,
Utah, Virginia, Washington, West
Virginia, Wsconsin
Outcome (ICD-9: Daily counts of
hospital admissions for primary
diagnosis of chronic obstructive
pulmonary disease (490-492), and
respiratory tract infections (464-466,
480-487).
Age Groups: >65 yr
Study Design: Time series
N: 11.5 million Medicare enrollees
Statistical Analyses: Bayesian 2-stage
hierarchical models.
First stage: Poisson regression (county-
specific)
Second stage: Bayesian hierarchical
models, to produce a national avg
estimate
Covariates: Day of the week,
seasonality, temperature, dew point
temperature, long-term trends
Season: NR
Dose-response Investigated: No
Statistical Package: R statistical
software version 2.2.0
Lags Considered: 0-2 days, avg of
days 0-2
Pollutant: PM25
Averaging Time: 24 h
Mean ftjg/m3) (IQR): 13.4 (11.3-15.2)
Monitoring Stations: NR
Copollutant (correlation): NR
Other variables: Median of pain/vise
correlations among PM25 monitors
within the same county for 2000: r =
0.91 (IQR: 0.81-0.95)
PM Increment: 10 pg/m3 (Results in
figures see notes)
Percent increase in risk [95% PI]:
COPD (Lag 0): Age 65+: 0.91 [0.18,
1.64]
Age 65-74: 0.42 [-0.64, 1.48]
Age 75+: 1.47 [0.54, 2.40]
Respiratory tract infection: Age 65+:
0.92 [0.41,1.43]
Age 65-74: 0.93 [0.04, 1.82]
Age 75+: 0.92 [0.32, 1.53]
Annual reduction in admissions
attributable to a 10 pg/m3 reduction in
daily PM25 level (95% PI):
Cerebrovascular disease: Annual
number of admissions: 226,641
Annual reduction in admissions: 1836
[680, 2992]
COPD: Annual number of admissions:
108,812
Annual reduction in admissions: 990
[196,1785]
Respiratory tract infections: Annual
number of admissions: 226,620
Annual reduction in admissions: 2085
[929, 3241]
December 2009
E-272
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Dominici et al. (2006,
0883981
Period of Study: 1999-2002
Location: U.S. (mainland)
Outcome (ICD-9): Respiratory tract
infections (464-466, 480-487) and
Chronic Obstructive Pulmonary Disease
(490-492)
Age Groups: All >65 yr
65-74 yr
>75yr
Study Design: Time series
N: 11.5 million at-risk
Statistical Analyses: Bayesian 2-stage
hierarchical models (day-to-day
variation), Poisson regression (county-
specific RRs)
Covariates: Calendar time (seasonality
and yr), temperature, dew point
Season: NR
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0,1, 2 days
Pollutant: PM25
Averaging Time: Daily or every 3 days
(depending on county)
Mean: 13.4 (IQR: 11.3-15.2)
Monitoring Stations: NR (used data
from Air Quality System database)
Copollutant: NR
PM Increment: 10 pg/m
Percentage Change in Hospital
Admission Rates [PI]:
COPD-Same day
All >65: 0.91 [0.18,1.64]
65-74 yr: 0.42 [-0.64,1.48]
>75:1.47 [0.54,2.40]
Respiratory Tract lnfections-2-day lag
All >65: 0.92 [0.41,1.43]
65-74 yr: 0.93 [0.04,1.82]
>75: 0.92 [0.32,1.53]
Notes: Other lag data shown in Fig 2-4
Reference: Erbas et al. (2005, 0738491
Period of Study: Jul 1989-Dec 1992
Location: Melbourne, Australia
Outcome (ICD):
COPD (490-492, 494, 496)
Asthma (493)
Age Groups: NR
Study Design: Time series
N:NR
Statistical Analyses: GLM, GAM,
Parameter Driven Poisson Regression,
Transitional Regression, Seasonal-
Trend decomposition based on Loess
smoothing for seasonal adjustment
Covariates: Secular trends,
seasonality, relative humidity, dry bulb
temp, dew point temp
Season: NR
Dose-response Investigated? Yes
Statistical Package: S-Plus, SAS
Lags Considered: 0-5 days
Pollutant: PMO.1-1 (API)
Averaging Time: 24 h
Mean (min-max): NR
Monitoring Stations: 9
Copollutant (correlation): NR
PM Increment: Increase from the
10th-90thpercentile (value NR)
RR Estimate [Cl]:
COPD
GAM:
0.95 [0.91,1.00]
GLM, PDM, TRM:NR
Asthma
NR
Notes: This study was used to
demonstrate that conclusions are highly
dependent on the type of model used
December 2009
E-273
-------
Reference
Reference: Fung et al. (2006, 0897891
Period of Study: Jun 1995-Mar 1999
Location: Vancouver, Canada
Reference: Hinwood et al. (2006,
Design & Methods
Hospital Admission/ED:
Hospital Admission
Outcome: Respiratory diseases
(460-519)
Age Groups: Age >65
Study Design: Time series, case
crossover
N: 26,275 individuals admitted
Statistical Analyses: Poisson
regression (spline 12 knots), case-
crossover (controls +17 days from case
date), Dewanji and Moolgavkar (DM)
method
Covariates: Long-term trends, day-of-
the-week effect, weather
Season: All yr
Dose-response Investigated? No
Statistical Package: SPIus, R
Lags Considered: 0-7 days
Hospital Admission
Concentrations1
Pollutant: PM25
Averaging Time: 24-h avg
Mean (SD): 7.72(3.61)
Range (Mm, Max): (2, 32)
Monitoring Stations: NR
Copollutant (correlation):
PM25:
PM10r = 0.80
PMio-25 r = 0.34
rn r - n 9?
**s\J 1 ~ U.ZO
CoH r = 0.38
03r = -0.03
N02r = 0.36
S02r = 0.42
Pollutant: PM25
Effect Estimates (95% Cl)
PM Increment: : 4 pg/m3
RR Estimate (66+ yr)
DM method:
1.007(0.994, 1.020]
Current
1.007[0.990,1.023]3day
0.995[0.979,1.012]5day
0.995(0.971, 1.020] 7 day
Tirnp cpripc1
I II 1 1C oCI ICo.
1.003(0.989, 1.018]
Current
1.000(0.982, 1.018] 3 day
0.993(0.972, 1.014] 5 day
0.995(0.971, 1.020] 7 day
Case-crossover:
1.002(0.986, 1.019]
Current
1.001(0.981, 1.021] 3 day
0.988(0.966, 1.011] 5 day
0.984(0.959, 1.010] 7 day
Increment: 1 pg/m3
088976'
Period of Study: Jan 1992-Dec 1998
Location: Perth, Australia
Outcome (ICD-9): COPD (490-496.99,
except asthma), pneumonia /influenza
(480-489.99), asthma
Age Groups: All ages
Study Design: Time stratified case-
crossover
N:NR
Statistical Analyses: Conditional
logistic regression
Covariates: Time trend, season,
temperature, humidity, dayofwk,
holidays
Season: All yr
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 0-3 days
Averaging Time: 24-h avg
Mean (SD): 9.2 (4.3)
Percent! les:
10th: 5.0
90th: 14.5
Monitoring Stations: 13
Notes: Copollutant: NR
Notes: Odds ratio for PM25 and all
respiratory, COPD, pneumonia and
asthma. Authors found an elevation in
the odds ratio for lags 2 and 3
reaaching significance in all age groups
for lag 3. For each increase of 1 pg/m ,
the number of hospitalizations
increases 0.2% for respiratory disease,
0.5% for pneumonia and 0.3% for
asthna. PM25 concentrations were also
significantly associated with asthma for
those aged under 15 yrwith an
estimated 0.5% increase in
hospitalizations.
December 2009
E-274
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Hirshon et al. (2008,
1803751
Period of Study: Jun 2002-Nov 2002
Location: Baltimore, Maryland
Outcome: Hospital admissions for
asthma
Study Design: Time-series
Covariates: Spatial distance from
pollution monitor, demographic
variation, long term, seasonal and daily
trends, weather and other pollutants
Statistical Analysis: Overdispersed
Poisson regression
Age Groups: 0-17 yr
Pollutant: Plfezinc
Averaging Time: 24 h
Mean (SD) Unit: 22.42 (25.14) pg/m3
Range (Min, Max): NR
Copollutant (correlation):
Ni:0.41
Cr:0.17
Fe: 0.54
Sulfate: 0.01
CO: 0.40
PM25: 0.39
03: 0.01
N02: 0.66
EC: 0.48
Increment: NR
Relative Risk (96% Cl), Best fit Model
Medium = 8.63-20.76 ng/m3
High = >20.76 ng/m3
No Lag
Medium: 1.12 (0.98-1.28)
High: 1.09 (0.91-1.30)
1-day Lag
Medium: 1.23 (1.07-1.41)
High: 1.16 (0.97-1.39)
2-day Lag
Medium: 1.11 (0.94-1.30)
High: 1.15 (0.96-1.38)
Controlling for Time Trends
No Lag
Medium: 1.08 (0.95-1.23)
High: 0.98 (0.86-1.11)
1-day Lag
Medium: 1.13 (1.003-1.28)
High: 1.03 (0.91-1.16)
2-day Lag
Medium:! 13 ()
High: 0.98-1.31
Controlling for Time Trends and
Additional Copollutants
No Lag
Medium: 1.12 (0.98-1.29)
High: 1.09 (1.01-1.30)
1-day Lag
Medium: 1.20 (1.04-1.38)
High: 1.12 (0.93-1.35)
2-day Lag
Medium: 1.12 (0.95-1.32)
High: 1.19 (0.98-1.44)
Reference: Host et al. (2007,1558511
Period of Study: 2000-2003
Location: Six French cities: Le Havre,
Lille, Marseille, Paris, Rouen, and
Toulouse
Outcome (ICD-10): Daily
hospitalizations for all respiratory
diseases (JOO-J99), respiratory
infections (J10-J22).
Age Groups: For all respiratory
diseases: 0-14 yr, 15-64 yr, and > 65 yr.
For respiratory infections: All ages
Study Design: Time series
N: NR (Total population of cities:
approximately 10 million)
Statistical Analyses: Poisson
regression
Covariates: Seasons, days of the
week, holidays, influenza epidemics,
pollen counts, temperature, and
temporal trends
Season: NR
Dose-response Investigated: No
Statistical Package: MGCV package in
R software (R 2.1.1)
Lags Considered: Avg of 0-1 days
Pollutant: PM25
Averaging Time: 24 h
Mean (6th -96th percentile):
Le Havre: 13.8 (6.0-30.5)
Lille: 15.9 (6.9-26.3)
Marseille: 18.8 (8.0-33.0)
Paris: 14.7 (6.5-28.8)
Rouen: 14.4 (7.5-28.0)
Toulouse: 13.8 (6.0-25.0)
Monitoring Stations:
13 total: 1 in Toulouse
4 in Paris
2 each in other cities
Copollutant (correlation):
PMio.25: Overall: r> 0.6
Ranged between r = 0.28 and
r = 0.73 across the six cities.
PM Increment: 10 pg/m increase, and
a 27 pg/m3 increase (corresponding to
the difference between the lowest of the
5th percentiles and the highest of the
95th percentiles of the cities'
distributions)
ERR (excess relative risk) Estimate [Cl]:
For all respiratory diseases (27 pg/m
increase): 0-14 yr: 1.1% [-3.1, 5.5]
15-64 yr: 2.2% [-1.8, 6.4];
> 65 yr: 1.3% [-5.3, 8.2]
For respiratory infections (10 pg/m3
increase): All ages: 2.5% [0.1, 4.8]
For respiratory infections
increase): All ages: 7.0%
27 pg/m3
0.7,13.6]
December 2009
E-275
-------
Reference
Reference: Ko et al. (2007, 0916391
Period of Study: Jan 2000-Dec 2004
Location: Hong Kong, China
Reference: Ko et al. (2007, 0928441
Period of Study: Jan 2000-Dec 2005
Location: Hong Kong, China
Reference: Lee et al. (2006, 0901761
Period of Study: Jan 1997-Dec 2002
Location: Hong Kong, China
Design & Methods
ED Visits
Outcome (ICD-9): COPD: Chronic
bronchitis (491), Emphysema (492),
Chronic airway obstruction (496)
Age Groups: All ages
Study Design: Time series
N: 15 hospitals, 11 9,225 admissions
Statistical Analyses: Poisson
regression, GAM with stringent
convergence criteria, APHEA2 protocol.
Covariates: Time trend, season,
temperature, humidity, other cyclical
factors, day, dayofwk, holidays
Season: All yr, interactions with season
tested
Dose-response Investigated? No
Statistical Package: SPLUS 4.0
Lags Considered: 0-5 days
Hospital Admission
Outcome (ICD-9): Asthma (493)
Age Groups: All, 0-14, 15-56,65+
Study Design: Time series
N: 69,716 admissions, 15 hospitals
Statistical Analyses: Poisson
regression, with GAM with stringent
convergence criteria.
Covariates: Time trend, season,
temperature, humidity, other cyclical
factors
Season: All yr, evaluated effect of
season in analysis
Dose-response Investigated? No
Statistical Package: SPLUS 4.0
Lags Considered: 0-5 days
Hospital Admission
Outcome: Asthma (493)
Age Groups: <18yr
Study Design: Time series
N: 26,663 asthma admissions for
asthma and 5821 admissions for
influenza
Statistical Analyses: Poisson
regression, GAM
Covariates: Temperature, atmospheric
pressure, relative humidity
Season: All
Dose-response Investigated? No
Statistical Package: SAS 8.02
Lags Considered: 0-5
Notes: Controls were admissions for
influenza ICD9 487
Concentrations1
Pollutant: PM25
Averaging Time: 24 h
Mean (SD): 35.7 (20.6)
Percent! les:
25th: 19.4
50th(Median):31.7
75th: 46.7
Range (Min, Max): (6.0, 163.2)
Monitoring Stations: 14
Copollutant (correlation):
PM25:
PM,0r = 0.952
N02r = 0.441
03r = 0.394
S02 r = 0.282
Pollutant: PM25
Averaging Time: 24 h
Mean (SD): 36.4 (21.1)
Percent! les:
25th: 20.0
SOth(Median): 32.5
75th: 47.7
Range (Min, Max): (6, 163)
Monitoring Stations: 14
Copollutant (correlation):
PM25:
PM,o r = 0.956
N02 r = 0.774
03 r = 0.585
S02 r = 0.482
Pollutant: PM25
Averaging Time: 24 h
Mean (SD): 45.3 pg/m3, (16.2)
Percentiles: 25th: 33.4
SOth(Median): 43.0
75th: 54.0
Range (Min, Max): NR
Monitoring Stations: 10
Copollutant (correlation):
PM25-PM10: 0.89
PM25-S02: 0.48
PM25-N02: 0.74
PM25-03: 0.47
Effect Estimates (95% Cl)
PM Increment: PMi0
RR Estimate
COPD:
1.0020.998, 1.001 lagO
1.0030.999, 1.007 lag 1
1.0111.007, 1.014] lag 2
1.0131.010,1.017 Iag3
1.011 1.008, 1.015] lag 4
1.0091.006,1.013 Iag5
1.0040.999, 1.008 lag 0-1
1.0101.006, 1.015 lag 0-2
1.0181.013, 1.022 lag 0-3
1.0241.019, 1.029 lag 0-4
1.031 1.026, 1.036 lag 0-5
4-Pollutant model:
1.014(1.007, 1.022] lag 0-5
3-Pollutant model:
1.011(1.004, 1.017] lag 0-5
PM Increment: 10.0 pg/m3
RR Estimate
Asthma (Single-pollutant model):
1.0081.004,1.013 lagO
1.0041.000,1.009 Iag1
1.0041.000, 1.009 lag 2
1.0091.005, 1.014 Iag3
1.006 1.001, 1.011] lag 4
1.0020.998! 1.007 lag 5
1.0091.004 1.014 lag 0-1
1.0121.007,1.018 lag 0-2
1.0171.011, 1.022] lag 0-3
1.020 1.014, 1.026 lag 0-4
1.021 1.015,1.028 lag 0-5
Asthma in Age:
0-14: 1.024(1. 013, 1.034] lag 0-5
14-65:1.018(1.008, 1.029] lag 0-5
>65: 1.021(1. 012, 1.030] lag 0-4
Asthma-Cold Season:
1.139(1.043, 1.244] lag 0-5
PM Increment: IQr= 20.6 pg/m3
Percent increase:
Single pollutant model:
5.10(2.95, 7.30], lag 0
5.00(2.88, 7.16], lag 1
5.48 [2.75, 6.95], lag 2
4.83 [2.78, 6.93], lag 3
6.59 [4.51, 8.72], lag 4
5.24 [3.18, 7.34], lag 5
Multipollutant model (S02, N02, CO, 03)
3.24 [0.93, 5.60], lag 4
December 2009
E-276
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Letz and Quinn (2005,
0887521
Period of Study: Oct 2001-Aug 2002
Location: San Antonio,Texas
Emergency Dept Visits
Outcome (ICD-9): Asthma or reactive
airway disease (493.0-493.9), wheezing
(786.07), dyspnea (786.01-786.9),
shortness of breath (786.05), bronchitis
(490-496), or cough (786.2)
Age Groups: NR (basic air force
trainees)
Study Design: Historic (retrospective)
cohort
N: 149 ED visits
Statistical Analyses: Pearson
correlation
Covariates: NR
Season: NR
Dose-response Investigated? No
Statistical Package: SPSS
Lags Considered: NR
Pollutant: PM25
Averaging Time: 24-h AQI
AQI Range (min-max): (4-109)
Monitoring Stations: Data obtained
from the Texas Commission on
Environmental Quality
Copollutant (correlation): NR
N: 6782 respiratory infection
hospitalizations
Statistical Analyses: Conditional
logistic regression (Cox proportional
hazards model)
Covariates: Daily mean temp and dew
point temp
Season: NR
Dose-response Investigated? No
Statistical Package: SAS 8.2 PHREG
procedure
Lags Considered: 1- to 7-day avg
Copollutant (correlation):
PM10.25:r = 0.33
PM10:r = 0.87
CO: r = 0.10
S02:r = 0.47
N02:r = 0.48
03:r = 0.56
PM Increment: NR
Correlation with Outcomes:
Same-day
All visits: r = 0.082
Proven asthmatic events: r = -0.042
3-day
All visits: r = 0.097
Proven asthmatic events: r = 0.011
Reference: Lin et al. (2005, 0878281
Period of Study: 1998-2001
Location: Toronto, North York, East
York, Etobicoke, Scarborough, and York
(Canada)
Hospital Admissions
Outcome (ICD-9): Respiratory
infections including laryngitis, tracheitis,
bronchitis, bronchiolitis, pneumonia,
and influenza (464, 466, 480-487)
Age Groups: 0-14 yr
Study Design: Bidirectional case-
crossover
Pollutant: PM25
Averaging Time: 24 h
Mean (min-max):
9.59 (0.25-50.50)
SD = 7.06
Monitoring Stations: 4
PM Increment: 7.8 pg/m3
OR Estimate [Cl]:
Adjusted for weather
4-day avg: 1.11 [1.02,1.22]
6-day avg: 1.11 [1.00,1.24]
Adj for weather and other gaseous
4-day avg: 0.94 [0.81,1.08]
6-day avg: 0.90 [0.76,1.07]
Notes: OR's were also categorized into
"Boys" and "Girls," yielding similar
results
December 2009
E-277
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Lin et al. (2002, 0260671
Period of Study: Jan 1981-Dec 1993
Location: Toronto
Hospital Admissions
Outcome (ICD-9): Asthma (493)
Age Groups: 6-12 yr
Study Design: Uni- and bi-directional
case-crossover (UCC, BCC) and time-
series (TS)
N: 7,319 asthma admissions
Statistical Analyses: Conditional
logistic regression, GAM
Covariates: Maximum and minimum
temp, avg relative humidity
Season: Apr-Sep, Oct-Mar
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 1- to 7-day avg
Pollutant: PM25
Averaging Time: 6 days (predicted
daily values)
Mean (min-max):
17.99(1.22-89.59)
SD = 8.49
Monitoring Stations: 1
Copollutant (correlation):
PM,o:r = 0.87
PMi0.25:r = 0.44
CO: r = 0.45
S02:r = 0.46
N02:r = 0.50
03:r = 0.21
PM Increment: 9.3 pg/m
RR Estimate [Cl]:
Adj for weather and gaseous pollutants
BCC 5-day avg: 0.94 [0.85,1.03]
BCC 6-day avg: 0.92 [0.83,1.02]
TS 5-day avg: 0.96 [0.90,1.02]
TS 6-day avg: 0.94 [0.88,1.01]
Boys-adj for weather
'1.04,1.15
1.02,1.16
0.97,1.06'
0.93,1.05:
TS 1-day avg: 1.00 [0.97,1.04]
TS 2-day avg: 0.98 [0.94,1.02]
Girls-adj for weather
UCC 1-day avg: 1.09
UCC 2-day avg: 1.09
BCC 1-day avg: 1.01
BCC 2-day avg: 0.99
UCC 1-day avg: 1.06
UCC 2-day avg: 1.11
BCC 1-day avg: 0.99
BCC 2-day avg: 1.02
0.99,1.14
1.02,1.21'
0.93,1.06:
0.94,1.09:
o:95,1.04] '
0.95,1.06]
TS 1-day avg: 0.99
TS 2-day avg: 1.00
Notes: The author also provides RR
using UCC, BCC, and TS analysis for
female and male groups for days 3-7,
yielding similar results
Reference: Magas et al. (2007,
0907141
Period of Study: 2001-2003
Location: Oklahoma City Metro area,
Oklahoma and Cleveland counties
Hospital Admission/ED: Admissions
Outcome: Asthma 493.01-493.99
Age Groups: <1 Syr
Study Design: Time series
N: 1,270 admissions
Statistical Analyses: Negative
binomial regression
Covariates: Temperature, humidity,
pollen count, mold
Season: All
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 1
Pollutant: PM25
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Monitoring Stations: 10
Copollutant (correlation): NR
Notes: Coefficient for PM25 was not
significant and thus not reported.
Reference: Mohr et al. (2008,1802151
Period of Study: Jun 2001-May 2003
Location: St. Louis, MO
Outcome: Asthma ER Visits
Age Groups: 2-17 yr
Study Design: Time series
Statistical Analyses: GEE Poisson
models
Covariates: Season, weekend
exposure, allergens
Dose-response Investigated: No
Statistical Package: SAS
Lags Considered: 1 day
Pollutant: PM25 EC
Averaging Time: 24 h
StdDev:0.1
Monitoring Stations: 1
Copollutant: NOX, S02, 03
Co-pollutant Correlation
NOX: 0.68*
S02: 0.09
03: -0 06
*p<0.05
PM Increment: 0.1 pg/m
Relative Risk Effect (Lower Cl, Upper
Cl):
Weekend Exposure
Summer: 1.05 (1.00,1.11)
Fall: 0.99 (0.97, 1.01)
Winter: 0.96 (0.92,1.00)
Spring: 0.96 (0.92,1.00)
Weekday Exposure
Summer: 1.01 (0.98,1.03)
Fall: 1.00 (0.99,1.01)
Wnter: 0.99 (0.96,1.01)
Spring: 0.98 (0.96,1.01)
December 2009
E-278
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Neuberger et al. (2004,
0932491
Period of Study: 1999-2000 (1-yr
period)
Location: Vienna and Lower Austria
Hospital Admissions
Outcome (ICD-9): Bronchitis,
emphysema, asthma, bronchiectasis,
extrinsic allergic alveolitis, and chronic
airway obstruction (490-496)
Age Groups: 3.0-5.9 yr
7-10 yr
65+yr
Study Design: Time series
N: 366 days (admissions NR)
Statistical Analyses: GAM
Covariates: S02, NO, N02, 03,
temperature, humidity, and day of the
week
Season: NR
Dose-response Investigated? Yes
Statistical Package: S-Plus 2000
Lags Considered: 0-14 days
Pollutant: PM25
Averaging Time: 24 h
Maximum daily mean:
Vienna: 96.4
Rural area: 48.0
Monitoring Stations: NR
Copollutant (correlation): NR
PM Increment: 10 pg/m
Log Relative Rate Estimate (p-value):
Vienna
Male: 2-day lag = 5.467 (0.019)
Female: 3-day lag = 5.596 (0.009)
Rural
Male: 10-day lag = 9.893 (0.012)
Female: 11-day lag = 10.529 (0.011)
Association with tidal lung functioN:
P = -0.987 (p-value = 0.091)
Notes: Effect parameters with
significant coefficients for respiratory
health included: male sex, allergy,
asthma in family, and traffic for Vienna
and age, allergy, asthma in family,
passive smoking, and PM fraction for
the rural area. Effect parameters with
significant coefficients for log asthma
score were allergy, asthma in family,
and rain for Vienna and allergy, asthma
in family, and passive smoking for the
rural area. Cross-correlation coefficients
are provided in Fig 1.
Reference: Ostro et al. (2008, 0979711
Period of Study: 2000-2003
Location: Six California Counties
Outcome: Respiratory disease
(ICD-9 460-519)
Study Design: Time-Series
Statistical Analysis: Poisson
Regression
Statistical Package: R
Age Groups: Children <19 yr
Pollutant: PM25 and components
Averaging Time: 24 h
Mean (SD) Unit: 19.4 pg/m3
IQR: 14.6 pg/m3
Copollutants:
EC, OC, N02, S04, Cu, Fe, K, Si, Zn
Increment: NR
Relative Risk (Min Cl, Max Cl)
Full results are presented graphically in
figures 1 and 2.
Excess risks for all-yr respiratory
hospital admissions in children <19yrs,
3-day lag
PM25: 4.1% (1.8-6.4)
EC: 5.4% (0.8-10.3)
Fe: 4.7% (2.2-7.2)
OC: 3.4% (1.1-5.7)
Nitrates: 3.3% (1.1-5.5)
Sulfates: 3.0% (0.4-5.7)
Excess risks for cool season (Oct-Mar)
respiratory hospital admissions in
children <19yrs, 3 day lag
PM25: 5.1% (1.6-8.9)
EC: 6.8% (-0.2-14.2)
Fe: 4.8% (1.7-8.0)
K: 4.0% (0.3-7.7)
December 2009
E-279
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Slaughter et al. (2005,
0738541
Period of Study: Jan 1995-Jun 2001
Location: Spokane, WA
Hospital Admissions and ED visits
Outcome: All respiratory (460-519)
Asthma (493)
COPD (491,492, 494,496)
Pneumonia (480-487)
Acute URI not including colds and
sinusitis (464, 466, 490)
Age Groups: All, 15+ yr for COPD
Study Design: Time series
N: 2373 visit records
Statistical Analyses: Poisson
Pollutant: PM25
Averaging Time: 24-h avg
Range (90% of Concentrations):
4.2-20.2 pg/m3
Monitoring Stations:
One
Notes: Copollutant (correlation):
PM2.5
PM1 r = 0.95
PM Increment: 10 pg/m
RR Estimate [Lower Cl, Upper Cl]
lag:
ER visits:
PM25
All Respiratory
Lag 1:1.01 [0.98,1.04]
Lag 2:1.02 [0.99,1.04]
Lag 3:1.02 [0.99,1.05]
Acute Asthma
Lag 1:1.03 [0.98,1.09]
Lag 2:1.00 [0.95,1.05]
Lag 3:1.01 [0.96,1.06]
COPD (adult)
Reference: Tecer et al. (2008, 1800301
Period of Study: Dec 2004-Oct 2005
Location: Zonguldak, Turkey
Reference: Tolbert et al. (2007,
0903161
Period of Study: Aug 1998-Dec 2004
Location: Atlanta Metropolitan area,
Georgia
regression, GLM with natural splines.
For comparison also used GAM with
smoothing splines and default
convergence criteria.
Covariates: Season, temperature,
relative humidity, day of week
Season: All
Dose-response Investigated?: No
Statistical Package: SAS, SPLUS
Lags Considered: 1 -3 days
Outcome: ED visits for respiratory
problems (ICD-9 470-478, 493)
Study Design: Bidirectional Case-
crossover
Covariates: Daily meteorological
parameters
Statistical Analysis: Conditional
logistic regression
Statistical Package: Stata
Age Groups: 0-14 yr
Outcome (ICD-9):
Combined RD group, including:
Asthma (493, 786.07, 786.09), COPD
(491, 492, 496), URI (460-465, 460.0,
477), pneumonia (480-486), and
bronchiolitis (466.1, 466. 11, and
400. iyjj
Age Groups: All
PM10r = 0.62
PM10.25r = 0.31
CO r = 0.62
Temperature r = 0.21
Pollutant: PM25
Averaging Time: NR
Mean, Unit: 29.1 pg/m3
Range (Min, Max): 4.55, 95.65
Copollutant (correlation):
PM25/PM10
Mparv D ^fi
ivicai i. u.uu
Range: 0.17-0. 88
PM2,/PM10.2,
Mean: 1.49
Range: 0.21-7.53
Pollutant: PM25
Averaging Time: 24 h
Mean (median IQR, range, 10th-90th
percentiles):
PM25' 17 1 (156
11.0-21.9
0.8-65.8
7.9-28.8)
PM25sulfate:4.9(3.9
2.4-6.2
Lag 1:0. 96
Lag 2: 1.01
Lag 3: 1.00
Hospital Ac
0.89, 1.04]
0.93, 1.09]
0.93, 1.08]
missions:
PM25
All Respiratory
Lag 1:0.98 [0.94, 1.01]
Lag 2: 0.99 [0.96, 1.03]
Lag 3: 1.01 [0.98, 1.05]
Asthma
Lag 1:1. 01 [0.91, 1.11]
Lag 2: 1.03 [0.94, 1.13]
Lag 3: 1.02 [0.93, 1.13]
COPD (adult)
Lag 1:0.99 [0.91, 1.08]
Lag 2: 1.06 [0.98, 1.16]
Lag 3: 1.03 [0.94, 1.12]
Increment: 10 pg/m3
Odds Ratio (96% Cl)
Asthma
Lag 0:1. 15 0.99-1.34)
Lag 1:0.85
Lag 2: 0.87
Lag 3: 0.93
Lag 4: 1.25
Allergic Rhir
LagO: 1.21
Lag 1:0.84
0.70-1.03)
0.73-1.04)
0.79-1.10)
1.05-1.50)
itis with Asthma
1.10-1.33)
0.75-0.93)
Lag 2: 0.89 (0.81-0.98)
Lag 3: 0.99 0.90-1.09)
Lag 4: 1.06
0.95-1.19)
Allergic Rhinitis
Lag 0:1. 08
Lag 1:1. 03
Lag 2: 0.89
Lag 3: 0.98
Lag 4: 1.18
0.98-1.20)
0.93-1.13)
0.80-0.99)
0.89-1.09)
1.00-1.24)
Upper Respiratory Disease
Lag 0: 0.99
Lag 1:0.52
0.49-2.00)
0.22-1.20)
Lag 2: 1.29 (0.75-2.22)
Lag 3: 1.29
Lag 4: 1.47
0.69-2.43)
0.87-2.50)
Lower Respiratory Disease
Lag 0:1. 06
Lag 1:0.85
0.78-1.44)
0.59-1.22)
Lag 2: 1.08 (0.72-1.61)
Lag 3: 1.18
Lag 4: 0.72
0.92-1.52)
0.54-0.96)h
PM Increment:
PM25: 10.96 pg/m3 (IQR)
PM25sulfate:3.82pg/m3(IQR)
PM2.5 total carbon: 3.63 pg/m3 (IQR)
PM25OC:2.61 pg/m3 (IQR)
PM2 5 EC: 1.1 5 pg/m3 (IQR)
PM2 5 water-soluble metals: 0.03 pg/m3
December 2009
E-280
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Study Design: Time series
N:NR for 1998-2004.
For 1993-2004:10,234,490 ER visits
(283,360 and 1,072,429 visits included
in the CVD and RD groups,
respectively)
Statistical Analyses: Poisson
generalized linear models
Covariates: Long-term temporal
trends, season (for RD outcome),
temperature, dew point, days of week,
federal holidays, hospital entry and exit
Season: All
Dose-response Investigated: No
Statistical Package: SAS version 9.1
Lags Considered: 3-day ma(lag 0 -2)
0.5-21.9
1.7-9.5)
PM25OC:4.4(3.8
2.7-5.3
0.4-25.9
2.1-7.2)
PM25EC:1.6(1.3
0.9-2.0
0.1-11.9
0.6-3.0)
PM2 5 water-soluble metals: 0.030
(0.023
0.014-0.039
0.003-0.202
0.009-0.059)
Monitoring Stations: 1
Copollutant (correlation): Between
PM25and:
PM10:r = 0.84
03:r = 0.62
N02:r = 0.47
CO: r = 0.47
S02:r = 0.17
PMi0.25:r = 0.47;
PM25S04:r = 0.76;
PM25EC:r = 0.65;
PM25OC:r = 0.70;
PM25TC:r = 0.71;
PM2 5 water-sol metals:
r = 0.69
OHC:r = 0.50
Between PM25 S04 and: PMi0: r = 0.69
03:r = 0.56
N02:r = 0.14
CO: r = 0.14
S02:r = 0.09
PM10.25:r = 0.32;
PM25:r = 0.76;
PM25EC:r = 0.32;
PM25OC:r = 0.33;
PM25TC:r = 0.34;
PM2 5 water-sol metals:
r = 0.65
OHC:r = 0.47
Between PM25ECand: PM10: r = 0.61
03:r = 0.40
N02:r = 0.64
CO: r = 0.66
S02:r = 0.22
PMi0.25:r = 0.49
PM25:r = 0.65
PM25S04:r = 0.32
PM25OC:r = 0.82
PM25TC:r = 0.91
PM2 5 water soluble metals: r = 0.52
OHC:r = 0.35
Between PM25OCand: PM10: r = 0.65
03:r = 0.54
N02:r = 0.62
CO: r = 0.59
S02:r = 0.17
PMi0.25:r = 0.49
PM25:r = 0.70
PM25S04:r = 0.33
PM25EC:r = 0.82
PM25TC:r = 0.98
PM2 5 water-sol metals:
r = 0.49
OHC:r = 0.37
Between PM25 total carbon and: PM10: r
= 0.67
03:r = 0.52
N02:r = 0.65
CO: r = 0.63
S02:r = 0.19
PM10.25:r = 0.51
PM25:r = 0.71
(IQR)
Risk ratio [95% Cl] (single pollutant
models):
PM25:
RD: 1.005 [0.995-1.015]
PM2 5 sulfate:
RD: 1.007 [0.996-1.018]
PM25 total carbon:
RD: 1.001 [0.993-1.008]
PM25OC:
RD: 1.003 [0.995-1.011]
PM25 EC:
RD: 0.996 [0.989-1.004]
PM2 5 water-soluble metals:
RD: 1.005 [0.995-1.015]
December 2009
E-281
-------
Reference Design & Methods
Reference: Wong et al. (2006, 0932661 General Practitioner Visits
Period of Study: 2000-2002 Outcome (ICPC-2): Respiratory
diseases/symptoms: upper respiratory
Location: Hong Kong (8 districts) tract infections (URTI), lower respiratory
infections, influenza, asthma, COPD,
allergic rhinitis, cough, and other
respiratory diseases
Age Groups: All ages
Study Design: Time series
N: 269,579 visits
Concentrations1
PM25S04:r = 0.34
PM25EC:r = 0.91
PM25OC:r = 0.98
PM2 5 water-sol metals:
r = 0.52
OHC:r = 0.38
Between PM25 water-soluble metals
and: PM10: r = 073
03:r = 0.43
N02:r = 0.32
CO: r = 0.35
S02:r = 0.06
PM10.25:r = 0.50
PM25:r = 0.69
PM25S04:r = 0.65
PM25EC:r = 0.52
PM25OC:r = 0.49
PM25TC:r = 0.52
Pollutant: PM25
Averaging Time: 24 h
Mean (min-max):
35.7(9-120)
SD=16.7
Monitoring Stations: 1 per district
Copollutant (correlation):
PM10:r = 0.94
Effect Estimates (95% Cl)
PM Increment: 10 pg/m3
RR Estimate [Cl]:
Overall URTI
1.021 [1.010,1.032]
Notes: RRs are also reported for each
individual general practitioner yielding
similar results
Statistical Analyses: GAM, Poisson
regression
Covariates: Season, day of the week,
climate
Season: NR
Dose-response Investigated? No
Statistical Package: S-Plus
Lags Considered: 0-3 days
Reference: Yang Q et al. (2004,
0874881
Period of Study: Jun 1995-Mar 1999
Location: Vancouver area, British
Columbia
Hospital Admissions
Outcome (ICD-9): Respiratory
diseases (460-51 9), pneumonia only
(480-486), asthma only (493)
Age Groups: 0-3 yr
Study Design: Case control,
Pollutant: PM25
Averaging Time: 24 h
Mean (min-max):
7.7 (2.0-32.0)
SD = 3.7
PM Increment: 4.0 pg/m3 (IQR)
OR Estimate [Cl]:
Values NR
Notes: Author states that no significant
association was found between PM25
and respiratory disease hospitalizations.
bidirectional case-crossover (BCC), and
time series (TS)
N: 1610 cases
Statistical Analyses: Chi-square test,
Logistic regression, GAM (time-series),
GLM with parametric natural cubic
splines
Covariates: Gender, socioeconomic
status, weekday, season, study yr,
influenza epidemic month
Season: Spring, summer, fall, winter
Dose-response Investigated? No
Statistical Package: SAS (Case
control and BCC), S-Plus (TS)
Lags Considered: 0-7 days
Monitoring Stations: NR (data
obtained from Greater Vancouver
Regional District Air Quality Dept)
Copollutant (correlation):
PM,0:r = 0.83
PM10.25:r = 0.39
CO: r = 0.24
03:r = -0.03
N02:r = 0.37
S02:r = 0.43
December 2009
E-282
-------
Reference
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Zanobetti and Schwartz
(2006, 0901951
Period of Study: 1995-1999
Location: Boston, MA
Hospital Admission/ED:
Outcome: Pneumonia (480-487)
Age Groups: >65 y
Study Design: Case-crossover, time
stratified
N: 24,857 for Pneumonia
Statistical Analyses: Condition logistic
regression
Covariates: Season, long term trend,
day of-the-wk, mean temperature,
relative humidity, barometric pressure,
extinction coefficient
Pollutant: PM non-traffic
Averaging Time: 24 h
Percentiles (pneumonia cohort):
5th: -7.3
25th: -3.28 pg/m3
SOth(Median): -0.88
75th: 1.92
95th: 12.11
PM Component: BC
Monitoring Stations: 4-5 monitors
PM Increment: PM non-traffic lag 0:
13.44|jg/m3
PM non-traffic lag 0-1 avg: 10.28 pg/m3
% change in Pneumonia:
PM non-traffic-0.57 [-7.51, 6.36]
lagO
PM non-traffic-0.94 [-7.20, 5.32]
mean lag 1
Reference: Zhong et al. (2006,
0932641
Period of Study: Apr-Oct 2002
Location: Cincinnati, Ohio
Season: All yr
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-1
Notes: Also looked at Ml cohort
Hospital Admissions
Outcome (ICD-9): Asthma (493-
493.91)
Age Groups: 1-18 yr
Study Design: Time series
N: 1254 admissions
Statistical Analyses: Poisson multiple
regression, GAM
Covariates: Season, temperature,
humidity, 03, day of the week
i/Qpuiiurani iCQiieiauunj.
PM non-traffic:
PM25r = 0.74
CO r = -0.01
N02r = 0.14
03r = -0.47
BC r = -0.01
Pollutant: PM25
Averaging Time: 24 h
Mean (SD):
Apr: 12.4 (3.8)
May: 13.6 (5.8)
Jun: 21.6 (9.9)
Jul: 25.8 (11.9)
Aug: 20.3 (8.7)
Sep: 19.5 (11.1)
Oct: 12.8 (6.4)
Monitoring Stations: NR (data
obtained from the National Virtual Data
System)
PM Increment: NR
RR Estimate [Cl]:
NR
Notes: This study focused primarily on
aeroallergens and asthma visits
Season: NR
Dose-response Investigated? Yes
Statistical Package: NR
Lags Considered: 1-5 days
Copollutant (correlation): NR
Notes: Author states all pairwise
correlations were insignificant
Reference: Zanobetti and Schwartz
(2006, 0901951
Period of Study: 1995-1999
Location: Boston, MA
Outcome: Pneumonia (480-487)
Age Groups: >65 y
Study Design: Case-crossover, time
stratified
N: 24,857 for Pneumonia
Statistical Analyses: Condition logistic
regression
Covariates: Season, long term trend,
day of-the-wk, mean temperature,
relative humidity, barometric pressure,
extinction coefficient
Season: All yr
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-1
Notes: Also looked at Ml cohort
Pollutant: PM25
Averaging Time: 24 h
Percentiles (pneumonia cohort):
25th: 7.23 pg/m3
50th(Median):11.10
75th: 16.14
PM Component: Black Carbon (BC),
PM non-traffic
Monitoring Stations: 4-5 monitors
Copollutant (correlation):
PM25:
CO r = 0.52
N02r = 0.55
03r = 0.20
BC r = 0.66
PM non-traffic r = 0.74
PM Increment: PM25 lag 0:
17.17 pg/m3
PM25lagO-1 avg: 16.32 pg/m3
% change in Pneumonia:
6.48[1.13,11.43]
lagO
5.56[-0.45, 11.27]
mean lag 1
'All units expressed in ug/m3 unless otherwise specified.
December 2009
E-283
-------
Table E-15. Short-term exposure-respiratory-ED/HA-Other Size Fractions.
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Andersen et al. (2007,
0932011
Period of Study: 2001-2004
Location: Copenhagen, Denmark
Outcome (ICD10):
Respiratory disease (J41-46)
Asthma (J45, 46)
Age Groups: 5-18 and >65
Study Design: Time-series
N: 1327 days
-1.5 million people at-risk
Statistical Analyses: Poisson
regression, GAM.
Covariates: Influenza epidemics,
pollen, temperature, dew point, day-of-
week, holiday, season.
Season: All
Dose-response Investigated? No
Statistical Package: R with gam and
mgcv packages.
Lags Considered: 0-5
Pollutant: Number concentration (NC)
of ultrafme & accumulation mode
particles
Averaging Time: 24 h
Mean particles/cm3 (SD):
NCtot (total): 8116 (3502)
25th: 4959
50th: 6243
75th: 8218
99th: 16189
IQR: 3259
NC100(<100nm): 6847 (2864)
25th: 5738
50th (Median): 7358
75th: 9645
99th: 19895
IQR: 3907
Mean particles/cm3 for 4 size modes
(median diameter (nm) noted):
NCa12: 493(315)
NCa23: 2253 (1364)
NCa57: 5104 (2687)
NCa212:6847(2864)
Monitoring Stations: 3 (Background,
rural Background, urban Curbside,
urban)
Notes: NC exposure data available for
n = 578 days. Information on
distribution of 4 size modes provided in
the paper.
Copollutant (correlation):
NCtot and PM,0:r = 0.39
NCtot and PM25:r = 0.40
NCtot and N02:r = 0.68
PM10andPM25: r = 0.8
"Low or no" correlations between 4 size
modes
NCa212andPM25:r = 0.8
NCa212 and PM,0:r = 0.63
NCa57 and N02:r = 0.57
Notes: selected correlations reported in
text, all correlations in annex to the
manuscript
PM Increment: Based on the IQR,
specific to metric (see below).
RR Estimate:
Single pollutant results, Asthma,
(5-18 yr), lag 0-5:
PM25:1.15 [1,1.32], IQr = 5
NCtot: 1.07 [0.98,1.17], IQr = 3907
NC100:1.06 [0.97, 1.16], IQr = 3259
NCa12:1.08 [0.99,1.18], IQr = 342
NCa212:1.08[1,1.17], IQr = 495
NCa23:1.09 [0.98,1.21 , IQr=1786
NCa57:1.02 [0.94,1.12, IQr = 3026
2-pollutant results:
NCa212w/PM10:1.1 [0.96, 1.13],
IQr = 495
NCtot w/PM,0:1.03 [0.92, 1.15]
NCtotw/PM25:1.04 [0.85, 1.28]
All RD, (>65 yr), lag 0-4, single pollutant
results:
PM25:1 [0.95, 1.05]
NCtot: 1.04 [1,1.07] IQr = 3907
NC100:1.03 [0.99, 1.07], IQr = 3259
NC12:1.01 [0.98, 1.05], IQr = 342
NC212:1.04 [1.01, 1.08], IQr = 495
NCa23: 0.99 [0.94, 1.03], IQr= 1786
NCa57:1.04[1, 1.08], IQr = 3026
2-pollutant results:
NCa212w/PM10:1.01 [0.96, 1.07],
IQr = 495
NCtot w/PM25: 0.97 [0.89, 1.05]
NCtot w/PM10:1 [0.96, 1.05]
Notes: Multipollutant model results also
included for models with 4 size modes.
Reference: Agarwal et al. (2006,
0990861
Period of Study: 2000-2003
Location: Safdarjung area of Delhi
Outcome (ICD-NR): COPD, asthma,
emphysema
Age Groups: NR
Study Design: Time series
N:NR
Statistical Analyses: Kruskal-VVallis
one-way analysis, Chi-square,
Multivariate linear regression
Covariates: Temp (min & max), relative
humidity at 0830 and 1730 h, wind
speed
Season: I (Jan-Mar), II (Apr-Jun), III
(Jul-Sep), IV (dot-Dec)
Dose-response Investigated? Yes
Statistical Package: SPSS
Lags Considered: NR
Pollutant: SPM (Suspended PM)
Averaging Time: 8 h
Mean pg/m3 (SD):
Qtr I: 297.5 (34.6)
Qtr II: 398.0 (85.6)
Qtr III: 220.0 (78.0)
Qtr IV: 399.0 (54.6)
Monitoring Stations: 2
Copollutant (correlation):
RSPM:r = 0.771
Other variables:
RH0830:r =-0.482
RH1730:r =-0.531
COPD: r = 0.474
PM Increment: NR
RR Estimate [Cl]: NR
Notes: This study analyzed seasonal
variation of pollutants and health
outcomes and correlations among the
variables
December 2009
E-284
-------
Study
Reference: Agarwal et al. (2006,
0990861
Period of Study: 2000-2003
Location: Safdarjung area of Delhi
Reference: Arbex et al. (2007, 0916371
Period of Study: Mar 2003-Jul 2004
Location: Araraquara, Sao Paulo State,
Brazil
Design & Methods
Outcome (ICD-NR): COPD, asthma,
emphysema
Age Groups: NR
Study Design: Time series
N:NR
Statistical Analyses: Kruskal-Wallis
one-way analysis, Chi-square,
Multivariate linear regression
Covariates: Temp (min & max), relative
humidity at 0830 and 1730 h, wind
speed
Season: I (Jan-Mar), II (Apr-Jun), III
(Jul-Sep), IV (dot-Dec)
Dose-response Investigated? Yes
Statistical Package: SPSS
Lags Considered: NR
Hospital Admission
Outcome (ICD10): Asthma (J15, J45)
Age Groups: All
Study Design: Time-series
N: 493 days, 1 hospital, 640 admissions
Statistical Analyses: Generalized
linear Poisson regression model with
natural cubic spline, Mann-Whitney U
Test
Covariates: Temperature and humidity
Season: All
Concentrations1
Pollutant: RSPM (Respirable
Suspended PM <10|jm)
Averaging Time: 8 h
Mean pg/m3 (SD):
Qtr 1: 119.0 (19.8)
Qtr II: 132.0 (28.4)
Qtr III: 75.0(23.4)
Qtr IV: 168. 0(40. 6)
Monitoring Stations: 2
Copollutant (correlation): SPM:
r - n 771
I ~ U. 1 1 \
Other variables:
Temp (min): r = -0.420
COPD: r = 0.353
Pollutant: TSP
Averaging Time: 24 h
Mean (SD): 46.8 pg/m3 (24.4)
Range (Min, Max):
6.7-1 37.8 pg/m3
Monitoring Stations: 1
Notes: TSP used as a proxy for fine &
ultrafine particles since it is composed
of 85-95% PM2.5.
Copollutant (correlation): NR
Effect Estimates (95% Cl)
PM Increment: NR
RR Estimate [Cl]: NR
Notes: This study analyzed seasonal
variation of pollutants and health
outcomes and correlations among the
variables
PM Increment: 10 pg/m3
% Increase
6.96 [1.4-12.86] 2-day ma
9.090 3. 12-1 5.40] 3 day ma
10.28 4.05-16.90] 4-day ma
11.63 5. 46-19.318] 5 day ma
12.61 5.68-20.00] 6-day ma
12.56 5. 47-20. 13] 7-day ma
% Increase by TSP quintile:
9.25-28. 45 pg/m3:: 1.00
28.46-48.85 pg/m3: : 1.55 045-5.77]
48.86-69.06 pg/m3: : 2.46 1.08-5.60]
69.07-88.44 pg/m3: : 2.77 1.32-5.84]
88. 45-108.9 pg/m3:: 2.94 1.48-5.85]
Dose-response Investigated? Yes,
quintile analysis
Statistical Package: SPSS V11 &
Splus 4.5
Lags Considered: 0-9
Notes: No TSP threshold for asthma
admissions noted. Analysis of lag
structure indicated that the acute effect
of TSP on admissions started 1 day
after TSP concentration increase and
remained unchanged for next 4 days.
Notes: To evaluate the association
between TSP generated from burning
sugar cane and asthma hospital
admissions.
Reference: Bartzokas et al. (2004,
0932521
Period of Study: Jun 1992-May 2000
Location: Athens, Greece
Outcome: Respiratory and
cardiovascular diseases (combined)
Age Groups: NR
Study Design: Time series
N: 1554 patients
Statistical Analyses: Simple linear
regression and linear stepwise
regression, Pearson correlation
Covariates: Temperature, atmospheric
pressure, relative humidity, wind speed
Season: Warm (May-Sep) and cold
(Nov-Mar)
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: NR
Pollutant: PM4.5 (black smoke)
Averaging Time: 10-day ma
Mean|jg/m3(SD):NR
Monitoring Stations: 1
Copollutant (correlation): N
PM Increment: NR
Correlation with Number of
Admissions:
Entire yr
Original: r = 0.18
Smoothed: r = 0.31
Warm period
Original: r = 0.19
Smoothed: r = 0.30
Cold period
Original: r = 0.18
Smoothed: r = 0.34
*AII above values are statistically
significant
December 2009
E-285
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Erbas et al. (2005, 0738491
Period of Study: Jul 1989-Dec 1992
Location: Melbourne, Australia
Outcome (ICD):
COPD (490-492, 494, 496)
Asthma (493)
Age Groups: NR
Study Design: Time series
N:NR
Statistical Analyses: GLM, GAM,
Parameter Driven Poisson Regression,
Transitional Regression, Seasonal-
Trend decomposition based on Loess
smoothing for seasonal adjustment
Covariates: Secular trends,
seasonality, relative humidity, dry bulb
temp, dew point temp
Season: NR
Dose-response Investigated? Yes
Statistical Package: S-Plus, SAS
Lags Considered: 0-5 days
Pollutant: PM 0.1-1 (API)
Averaging Time: 24 h
Mean (min-max): NR
Monitoring Stations: 9
Copollutant (correlation): NR
PM Increment: Increase from the 10th-
90th percentile (value NR)
RR Estimate [Cl]:
COPD
GAM:
0.95 [0.91,1.00]
GLM, PDM,TRM:NR
Asthma
NR
Notes: This study was used to
demonstrate that conclusions are highly
dependent on the type of model used
Reference: Halonen et al. (2008,
1895071
Period of Study: 1998-2004
Location: Helsinki, Finland
Outcome: Respiratory Hospitalizations
& Mortality (ICD 10: JOO-99)
Age Groups: 65+ yr
Study Design: Time series
N:NR
Statistical Analyses: Poisson, GAM
Covariates: Temperature, humidity,
influenza epidemics, high pollen
episodes, holidays
Dose-response Investigated? No
Statistical Package: R
Lags Considered: Lags 0-3 & 5-day
(0-4) mean
Pollutant: PM25
Averaging Time: Daily
Mean (SD): NR
Min:1.1
26th percentile: 5.5
60th percentile: 9.5
76th percentile: 11.7
Max: 69.5
Monitoring Stations: NR
Copollutant: PM<0.03, PMO.03-0.1,
PM<0.1,
PM<0.10.29, PM10.25, CO, N02
Co-pollutant Correlation
PM<0.03:0.14
PMO.03-0.1: 0.48
PM<0.1:0.35
PM<0.10.29:0.88
PM10.2.5: 0.25
PM Increment: Interquartile
Percent Change (Lower Cl, Upper
Cl):
All Respiratory Mortality
Lag 0: 2.67 (-0.39, 5.82 J
Lag 1:1.59 (-1.43, 4.70
Lag 2: 0.03 (-2.99, 3.16)
Lag 3:-0.11 (-3.13,3.01)
5-day mean: 1.39 (-2.83, 5.81)
Pneumonia HA
Lag 0: 0.93 (-0.85, 2.75)
Lag 1:2.41 (0.64, 4.21)
Lag 2:1.48 (-0.27, 3.26)
Lag 3:1.91 (0.14, 3.70)
5-day mean: 3.10 (0.60, 5.65)
Asthma + COPD HA
Lag 0: 2.48 (0.60, 4.39)
Lag 1:2.62 (0.78, 4.49)
Lag2:1.22(-0.62, 3.10)
Lag 3: 0.59 (-1.28, 2.49)
5-day mean: 2.49 (-0.08, 5.12)
Other HA
Lag 0: 0.05 (-2.38, 2.54)
Lag 1:0.2 (-2.17, 2.62)
Lag 2: 2.03 (-0.29, 4.41)
Lag 3:1.72 (-0.63, 4.12)
5-day mean: 1.88 (-1.50, 5.36)
*p<0.05, Jp<0.10
December 2009
E-286
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Llorca et al. (2005, 0878251
Period of Study: Jan 1992-Dec 1995
Location: Torrelavega, Spain
Outcome (ICD-9): Respiratory (460-
519) and cardiac (390-459) admissions
(analyzed combined and individually)
Age Groups: NR
Study Design: Time series
N: 18,137 admissions
Statistical Analyses: Stepwise multiple
linear regression, Poisson regression,
Spearman correlation
Covariates: Influenza, day of week,
wind speed, northeast and southwest
winds, minimum and maximum
temperature
Season: NR
Dose-response Investigated? No
Statistical Package: STATA
Interceded, Release 6
Lags Considered: NR
Pollutant: TSP (total suspended
particles)
Averaging Time: 24 h
Mean pg/m3 (SD):
48.8 (23.7)
Monitoring Stations: 3
Copollutant (correlation):
S02:r =-0.400
SH2:r =-0.392
NO: r = -0.109
N02:r =-0.120
Other variables:
Rain: r = -0.339
Max temp: r = 0.071
Min temp: r = -0.003
Avg temp: r = 0.035
Wind speed: r = -0.357
PM Increment: NR
Rate Ratio Estimate [Cl]:
Cardiorespiratory Admissions
Single-pollutant model: 0.92 [0.86,0.98]
Five-pollutant model: 1.05 [0.97,1.14]
Respiratory Admissions
Single-pollutant model: 0.98 [0.89,1.08]
Five-pollutant model: 0.91 [0.80,1.02]
Reference: Michaud et al. (2004,
1885301
Period of Study: Jan 1997-May 2001
Location: Hilo, Hawaii
ED visits
Outcome: Asthma/COPD (490-496)
Respiratory Irritation (506-508)
Age Groups: All
Study Design: Time-series
N:1,561 ER visits
Statistical Analyses: Multiple linear
regression
Covariates: Hourly temperature,
minimum daily temperature, minimum
daily temperature, humidity, yr, month,
day of the week
Season: all
Dose-response Investigated? No
Statistical Package:
STATA 6.0
SAS
Lags Considered: Previous night,
1,2,3
Pollutant: PM1
Averaging Time: 24-h avg
Mean (SD): 1.91 (2.95) pg/m3
Range (Min, Max):
0.0, 56.6 pg/m3
Monitoring Stations: 2
Notes: Copollutant (correlation): NR
PM Increment: 10 pg/m
RR Estimate [Lower Cl, Upper Cl]
lag:
Asthma, COPD (499-496): Adjusted for
day, month & yr:
1.11(0.92, 1.34), 00: 00-6: 00AM
1.14(1.03,1.26), lag 1
1.06(0.83, 0.94), lag 2
0.91 (0.06, 1.05), Iag3
Asthma (493, 495): Adjusted for day,
month & yr:
1.03(0.90, 1.42), 00: 00-6: 00AM
1.02(0.94,1.21), lag 1
1.02(0.99, 1.23), lag 2
0.97(0.69, 1.15), Iag3
Bronchitis (490, 491): Adjusted for day,
month & yr:
1.02(0.82, 1.41), 00: 00-6: 00AM
1.07(1.18,1.49), lag 1
0.97(0.60, 1.34), lag 2
0.93(0.43, 1.18), Iag3
Notes: Crude and estimates adjusted
for month and yr only also presented.
Notes: Volcanic fog = vog
December 2009
E-287
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Migliaretti et al. (2005,
0886891
Period of Study: 1997-1999
Location: Turin, Italy
Cases: Asthma (493)
Controls: Admissions for non-
respiratory or cardiac conditions (460-
487, 490-493, 494-496, 500-519, 390-
405, 410-429)
Age Groups: 0-14,15-64, >64
Study Design: Case-control
N: Cases: 1,401
Controls: 201,071
Statistical Analyses: Logistic
regression
Covariates: Gender, age, daily mean
temperature, season, day of week,
holidays, education level
Season: All
Dose-response Investigated? No
Lag :0-to 2-day avg
Pollutant: TSP
Averaging Time: Means of daily total
levels at stations
Mean (SD): 105.3 pg/m3, (44.2)
Percentiles: 26th: NR
60th(Median): 96.0 pg/m3
75th NR
Monitoring Stations:
10
Notes: Copollutant (correlation):
All seasons: N03"TSP = 0.80
Winter: NOfTSP = 0.77
Summer: N03~TSP = 0.69
PM Increment: 10 pg/m increase
% Increase, lag 0-2-day avg
1 pollutant model:
<15:1.90[0.40, 3.40]
15-64: 2.30 [-0.01, 5.20]
>64: 2.30 [1.10, 3.60]
Total: 2.30[1.10, 3.60]
% Increase, lag 0-2-day avg
2 pollutant model:
<15:-0.12 [-0.03, 2.50]
15-64: 0.90 [-0.04, 5.61]
>64:1.2 [-0.01, 4.32]
Total: 0.91 [-0.02,3.11]
Reference: Migliaretti et al. (2004,
0874251
Period of Study: 1997-1999
Location: Turin, Italy
Outcome:
Cases: Asthma (493)
Controls: Non-respiratory or cardiac
admissions (460-487, 490-493, 494-
496, 500-519, 390-405, 410-429)
Age Groups: 0-15
Study Design: Case-control
N: Cases: 1,060
Controls: 25,523
Statistical Analyses: Logistic
regression pg/m increase
Covariates: Gender, age, daily mean
temperature, season, day of week,
holidays, solar radiation
Season: All
Lags Considered: 1- to 3-day avg
Pollutant: Total suspended particulate PM Increment: 10 pg/m
Averaging Time: Mean of admission
day and 3 preceding days
Mean (SD): 114.5 pg/m3, (42.8)
Percentiles:
25th: NR
SOth(Median): 109.9 pg/m3
75th: NR
Monitoring Stations: 10
Notes: Copollutant (correlation):
TSP-NO: 0.76
% Increase, lag 1-3-day avg
<4yr: 1.8% [0.00, 3.05]
4-15 yr: 3.0% [0.01, 5.08]
all: 1.8% [0.03, 3.02]
adjusted for all covariates
Notes: Multipollutant models also used
Reference: Neuberger et al. (2004,
0932491
Period of Study: 1999-2000 (1-yr
period)
Location: Vienna and Lower Austria
Outcome (ICD-9): Bronchitis,
emphysema, asthma, bronchiectasis,
extrinsic allergic alveolitis, and chronic
airway obstruction (490-496)
Age Groups: 3.0-5.9 yr
7-1 Oyr
65+yr
Study Design: Time series
N: 366 days (admissions NR)
Statistical Analyses: GAM
Covariates: S02, NO, N02, 03,
temperature, humidity, and day of the
week
Season: NR
Dose-response Investigated? Yes
Statistical Package: S-Plus 2000
Lags Considered: 0-14 days
Pollutant: PMi
Averaging Time: 24 h
Mean|jg/m3(SD):NR
Monitoring Stations: NR
Copollutant (correlation): NR
PM Increment: NR
Effect parameters (Vienna children):
Respiratory Health
Male sex = 0.098
Allergy = 0.238
Asthma in family = 0.190
Traffic = 0.112
Log Asthma Score
Allergy = 0.210
Asthma in family = 0.112
Rain = 0.257
"only significant coefficients are
presented
Association with tidal lung function:
P =-1.059 (p-value = 0.060)
Notes: No significant associations
between PM and respiratory mortality
were found for either sex. Data is also
provided for children in the rural area
where age, allergy, asthma in family,
passive smoking, and PM fraction had
significant coefficients.
December 2009
E-288
-------
Study
Reference: Peel et al. (2005, 0563051
Period of Study: Jan 1993-Aug 2000
Location: Atlanta, Georgia
Reference: Simpson et al. (2005,
0874381
Period of Study: 1996-1999
Location: Brisbane, Sydney,
Melbourne, and Perth, Australia
Design & Methods
Hospital Admission/ED:
ED visits
Outcome: Asthma 493, 786.09
COPD 491, 492, 496
URI 460-466, 477
Pneumonia 480-486
Age Groups: All ages. Secondary
analyses conducted by age group:
Infants 0-1 yr
Pediatric asthma 2-1 Syr
Adults >18yr
Study Design: Case-control
All respiratory disease vs. finger
wounds
N: 31 hospitals
ED visits NR
Statistical Analyses: Poisson
generalized linear models
General linear models
Covariates: Avg temperature and dew
point, pollen counts
Season: All
Dose-response Investigated? Yes
Statistical Package: SAS 8.3
S-Plus2000
Lags Considered: 0-7 days and
14-day distributed lag
Outcome: All Respiratory (460-519)
Asthma (493)
COPD (490-492)
Pneumonia, acute bronchitis (466,
480-486)
Age Groups: All ages, split into f15-64
and >64 yr
Study Design: Time-series
N: NR -64,000 admissions
Concentrations1
Pollutant: UF(1 0-1 OOnm)
Averaging Time: 24-h avg
Mean (SD): 3800 (40700)
Percentiles:
10th: 11500
90th: 74600
PM Component: Oxygenated
hydrocarbons (OH), sulfate, acidity, EC
(EC), OC (OC), water-soluble transition
metals
Monitoring Stations: "Several"
Copollutant (correlation):
PM,0:r = -0.13
03:r = -0.13
N02:r = 0.26
CO: r = 0.10
S02: r = 0.24
PM25:r = -0.16
PM10.25:r = 0.13
Pollutant: BSP (indicator of particles
<2 pm in diameter)
Averaging Time: 24-h avg
Mean (SD): Means only
Brisbane 0.3 10 -4 m-1
Sydney 0.3 10 -4 m-1
Effect Estimates (95% Cl)
Increment:
30,000 #/cm3
All Respiratory Disease
0.984 [0.968-1. 000]
URI
0.986 [0.966, 1.006]
Asthma
0.999 [0.977, 1.021]
Pneumonia
0.997 [0.953, 1.002]
COPD
0.982 [0.942, 1.022]
PM Increment: "per unit increase"
RR Estimate [Lower Cl, Upper Cl]
Ian-
lag.
Single pollutant model
Respiratory >64 yr
1.0401 1.0045, 1.0770] Iag1
1.0520 1.0164, 1.0889] Iag2;
1.0451 1.0093, 1.0821] Iag3
1.0552 1.0082, 1.1 045] lag 0-1 avg
Statistical Analyses: GAM w/ LOESS
smoothers
GLM w/ natural and penalized spline
smoothers
Covariates: Temperature, relative
humidity, rain, day of the week, public
and school holidays, influenza
epidemics, and controlled burn events
Season: All
Dose-response Investigated? Yes
Statistical Package: S-Plus
R Lags Considered: 1-3 days, 0- to
1-day avg
Melbourne 0.3 10-4 m-1
Perth 0.3 10-4 m-1
Range (Min, Max):
Brisbane0.0, 2.5 10-4m-1
Sydney 0.0, 1.6 10-4 m-1
Melbourne0.0, 2.2 10-4m-1
Perth 0.1,1.8 10-4 m-1
PM Component: Monitoring Stations:
"network of sites across each city"
Copollutant (correlation): NR
Asthma 15-64 yr
1.0641 [1.0006, 1.1315] Iag2
1.0893 [1.0240, 1.1587] Iag3
Asthma + COPD >64 yr
1.0713 [1.0179, 1.1276] lagS
1.0552 [1.0082,1.1045] lag 0-1 avg
Pneumonia & Acute Bronchitis >64 yr
1.0587 [1.0013, 1.1193] Iag1
1.0636 [1.0056, 1.1249] lag 2
1.0769 [1.0046,1.1544] lag 0-1 avg
Multipollutant model
Respiratory admissions >64 yr
No other pollutants:
1.0552 [1.0082,1.1045] lag 0-1 avg
Max 1 h N02
1.0028 [0.9513,1.0572] lag 0-1 avg
Max 1 h 03
1.0534 [1.0058-1.1033] lag 0-1 avg
December 2009
E-289
-------
Study
Reference: Sinclair and Tolsma (2004,
0886961
Period of Study: 25 mo
Location: Atlanta, Georgia
Design & Methods
Outpatient Visits
Outcome: Asthma (493)
URI (460, 461, 462, 463, 464, 465, 466,
477)
LRI (466. 1,480, 481,482, 483, 484,
485, 486).
Age Groups: < = 18 yr; 18+ yr
(asthma); All ages (URI//LRI)
Study Design: Times series
N:25mo
260,000-275,000 health plan members
(Aug1998-Aug2000)
Statistical Analyses: Poisson GLM
Covariates: Season, day of week,
federal holidays, study months
Concentrations1
Pollutant: PM2.5.io (|jg/m3)
Averaging Time: 24-h avg
Mean (SD): PM coarse mass
(2.5-10 |jm)-9.67 pg/m3 (4.74)
Monitoring Stations: 1
Copollutant (correlation): NR
Effect Estimates (95% Cl)
PM Increment: 4.74 (1 SD)
RR Estimate [Lower Cl, Upper Cl]
lag:
Child Asthma:
Coarse PM = 1.053(3)
3-5 day lag
URI:
Course PM = 1.021 (S)
3-5 day lag
LRI:
Coarse PM = 1.07(3)
3-5 day lag
Season: NR
Dose-response Investigated?: No
Statistical Package: SAS
Lags Considered: Three 3-day ma
(0-2, 2-5, 6-8)
Notes: Numerical findings for significant
results only presented in manuscript.
Results for all lags presented
graphically for each outcome (asthma,
URI, and LRI).
Reference: Sinclair and Tolsma (2004,
0886961
Period of Study: 25 mo
Location: Atlanta, Georgia
Outpatient Visits
Outcome: Asthma (493)
URI (460, 461, 462, 463, 464, 465, 466,
477)
LRI (466.1,480, 481,482, 483, 484,
485, 486).
Age Groups: < = 18 yr, 18+ yr (asthma)
All ages (URI//LRI)
Study Design: Times series
N:25mo
260,000-275,000 health plan members
(Aug1998-Aug2000)
Statistical Analyses: Poisson GLM
Covariates: Season, day of week,
federal holidays, study months
Season: NR
Dose-response Investigated?: No
Statistical Package: SAS
Lags Considered: Three 3-day ma
(0-2, 2-5, 6-8)
Pollutant: UF(PM,o-100nm)
Averaging Time: 24 h avg
Mean (SD): PM10-100 nm area
(|jm2/cm3)- 249.33 (244.09)
Monitoring Stations: 1
Copollutant (correlation): NR
PM Increment: NR
RR Estimate [Lower Cl, Upper Cl]
lag:
Adult Asthma:
UltrafinePM area = 1.223(3)
3-5 days lag
URI:
UltrafinePM: = 1.041 (S)
0-2 days lag
LRI:
UltrafinePM area = 1.099(3)
6-8 days lag
Notes: Numerical findings for significant
results only presented in manuscript.
Results for all lags presented
graphically for each outcome (asthma,
URI, and LRI).
December 2009
E-290
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Slaughter et al. (2005,
0738541
Period of Study: Jan 1995-Jun 2001
Location: Spokane, WA
Hospital Admissions and ED visits
Outcome: All respiratory (460-519)
Asthma (493)
COPD (491,492, 494,496)
Pneumonia (480-487)
Acute URI not including colds and
sinusitis (464, 466, 490)
Age Groups: All, 15+ yr for COPD
Study Design: Time series
N: 2373 visit records
Statistical Analyses: Poisson
regression, GLM with natural splines.
For comparison also used GAM with
smoothing splines and default
convergence criteria.
Covariates: Season, temperature,
relative humidity, day of week
Season: All
Dose-response Investigated?: No
Statistical Package: SAS, SPLUS
Lags Considered: 1 -3 days
Pollutant: PMi
Averaging Time: 24-h avg
Range (90% of concentrations):
3.3-17.6 pg/m3
Monitoring Stations: 1
Copollutant (correlation):
PM,
PM25r = 0.95
PM10r = 0.50
PMi0.25r = 0.19
CO r = 0.63
PM Increment: 10 pg/m
RR Estimate [Lower Cl, Upper Cl]
lag:
ED visits:
PM,
All Respiratory
Lag1:1.01 [0.98,1.04]
Lag 2:1.02 [0.99,1.06]
Lag 3:1.02 [0.99,1.06]
Acute Asthma
Lag 1:1.03 [0.97,1.09]
Lag 2: 0.99 [0.93, 1.05]
Lag 3:1.02 [0.96,1.08]
COPD (adult)
Lag 1:0.96 [0.87,1.05]
Lag 2:1.02 [0.93,1.12]
Lag 3: 0.99 [0.90, 1.09]
Reference: Slaughter et al. (2005,
0738541
Period of Study: Jan 1995-Jun 2001
Location: Spokane, WA
Hospital Admissions and ED visits
Outcome: All respiratory (460-519)
Asthma (493)
COPD (491,492, 494,496)
Pneumonia (480-487)
Acute URI not including colds and
sinusitis (464, 466, 490)
Age Groups: All, 15+ yr for COPD
Study Design: Time series
N: 2373 visit records
Statistical Analyses: Poisson
regression, GLM with natural splines.
For comparison also used GAM with
smoothing splines and default
convergence criteria.
Covariates: Season, temperature,
relative humidity, day of week
Season: All
Dose-response Investigated?: No
Statistical Package: SAS, SPLUS
Lags Considered: 1 -3 days
Pollutant: PM25
Averaging Time: 24-h avg
Range (90% of Concentrations):
4.2-20.2 pg/m3
Monitoring Stations: 1
Notes: Copollutant (correlation):
PM25
PM1 r = 0.95
PM10r = 0.62
PM10.2.5r = 0.31
CO r = 0.62
Temperature r = 0.21
PM Increment: 10 pg/m
RR Estimate [Lower Cl, Upper Cl]
lag:
ER visits:
PM25
All Respiratory
Lag 1:1.01 [0.98,1.04]
Lag 2:1.02 [0.99,1.04]
Lag 3:1.02 [0.99,1.05]
Acute Asthma
Lag 1:1.03 [0.98,1.09]
Lag 2:1.00 [0.95,1.05]
Lag 3:1.01 [0.96,1.06]
COPD (adult)
Lag 1:0.96 [0.89,1.04]
Lag 2:1.01 [0.93,1.09]
Lag 3:1.00 [0.93,1.08]
Hospital Admissions:
PM25
All Respiratory
Lag 1:0.98 [0.94,1.01]
Lag 2: 0.99 [0.96, 1.03]
Lag 3:1.01 [0.98,1.05]
Asthma
Lag 1:1.01 [0.91,1.11]
Lag 2:1.03 [0.94,1.13]
Lag 3:1.02 [0.93,1.13]
COPD (adult)
Lag 1:0.99 [0.91,1.08]
Lag 2:1.06 [0.98,1.16]
Lag 3:1.03 [0.94,1.12]
December 2009
E-291
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Zanobetti and Schwartz
(2006, 0901951
Period of Study: 1995-1999
Location: Boston, MA
Outcome: Pneumonia (480-487)
Age Groups: >65 y
Study Design: Case-crossover, time
stratified
N: 24,857 for Pneumonia
Statistical Analyses: Condition logistic
regression
Covariates: Season, long term trend,
day of-the-wk, mean temperature,
relative humidity, barometric pressure,
extinction coefficient
Season: All yr
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-1
Notes: Also looked at Ml cohort
Pollutant: PM25
Averaging Time: 24 h
Percentiles (pneumonia cohort):
25th: 7.23 pg/m3
50th(Median):11.10
75th: 16.14
PM Component: Black Carbon (BC),
PM non-traffic
Monitoring Stations: 4-5 monitors
Copollutant (correlation):
PM25:
CO r = 0.52
N02r = 0.55
03r = 0.20
BC r = 0.66
PM non-traffic r = 0.74
PM Increment: PM25 lag 0:
17.17pg/m3
PM25lagO-1 avg: 16.32 pg/m3
% change in Pneumonia:
6.48[1.13,11.43]
lagO
5.56[-0.45, 11.27]
mean lag 1
Reference: Zanobetti and Schwartz
(2006, 0901951
Period of Study: 1995-1999
Location: Boston, MA
Outcome: Pneumonia (480-487)
Age Groups: >65 y
Study Design: Case-crossover, time
stratified
N: 24,857 for Pneumonia
Statistical Analyses: Condition logistic
regression
Covariates: Season, long term trend,
day of-the-wk, mean temperature,
relative humidity, barometric pressure,
extinction coefficient
Season: All yr
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: 0-1
Notes: Also looked at Ml cohort
Pollutant: BC
Averaging Time: 24 h
Percentiles (pneumonia cohort):
5th: 0.42
25th: 0.74 pg/m3
50th(Median):1.15
75th: 1.72
95th: 2.83
PM Component: PM non-traffic
Monitoring Stations: 4-5 monitors
Copollutant (correlation):
BC:
PM25 r = 0.66
CO r = 0.82
N02 r = 0.70
03 r = -0.25
PM non-traffic r = -0.01
PM Increment: BC lag 0: 2.05 pg/m
BC lag 0-1 avg: 1.69|jg/m3
% change in Pneumonia:
BC-10.76[4.54, 15.89]
lagO
BC-11.71[4.79, 17.36]
mean lag 1
1AII units expressed in ug/m3 unless otherwise specified.
December 2009
E-292
-------
E.3. Short-Term Exposure and Mortality
Table E-16. Short-term exposure-mortality - PMio.
Study
Design & Methods
Concentrations
Effect Estimates (95% Cl)
Reference: Aga et al. (2003, 0548081
Period of Study: ~5 yr for most cities,
during the 1990s
Location: 28 European cities
(APHEA2)
Outcome: Nonaccidental Mortality
(<800)
Study Design: Time-series
Statistical Analyses: Poisson GAM,
LOESS
Age Groups: All ages
>65
Pollutant: PM10
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): (15, 66)
Copollutant: BS
Note: PM10 only measured in 21 cities.
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl) lag:
All ages
Fixed effects: 0.71% (0.60,0.83) 0-1
Random effects: 0.67% (0.47,0.87) 0-1
>65
Fixed effects: 0.79% (0.66,0.92) 0-1
Random effects: 0.74% (0.52,0.95) 0-1
Models with effect modifiers (>65)
24-h N02:
25th Percentile: 0.30% (0.07,0.53)
75th Percentile: 0.97% (0.82,1.11)
24-h temperature:
25th Percentile: 0.44% (0.25,0.64
75th Percentile: 0.91% (0.77,1.05
24-h relative humidity:
25th Percentile: 0.98% (0.82,1.14
75th Percentile: 0.52% (0.33,0.71
Age standardized annual mortality rate:
25th Percentile: 0.93% (0.77,1.09
75th Percentile: 0.61% (0.43,0.79
Proportion individuals >65
25th Percentile: 0.67% (0.50,0.83
75th Percentile: 0.85% (0.71,0.99
Northwest/Central East:
25th Percentile: 0.81% (0.63,0.98)
75th Percentile: 0.26% (-0.05,0.57)
Northwest/South:
25th Percentile: 0.81% (0.63,0.98
75th Percentile: 1.04% (0.81,1.27
Reference: Analitis et al. (2006,
0881771
Period of Study: NR
Location: 29 European cities
(APHEA2)
Outcome: Mortality: Cardiovascular
diseases (390-459)
Respiratory diseases (460-519)
Study Design: Time-series
Statistical Analyses: 2-stage
hierarchical modeling
Age Groups: All ages
Pollutant: PM10
Averaging Time: 24-h avg
Median (SD) unit: Range: 9-64 pg/m3
Range (Min, Max): NR
Copollutant: BS
Note: PM10 only measured in 21 cities.
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl) lag:
Cardiovascular: Fixed effects: (
(0.47, 0.80) 0-1
Random effects:
0.76% (0.47, 1.05)0-1
0.90% (0.57, 1.23)0-5
Respiratory: Fixed effects:
0.58% (0.21, 0.95) 0-1
Random effects:
0.71% (0.22,1.20)0-1
1.24% (0.49, 1.99)0-5
Reference: Ballester et al. (2002,
0303711
Period of Study: 1990-1996
Location: 13 Spanish cities
Outcome: Mortality:
Nonaccidental (<800)
Cardiovascular diseases (390-459)
Respiratory diseases (460-519)
Study Design: Ecological time series
Statistical Analyses: Poisson GAM,
LOESS
Age Groups: All ages
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): Huelva: 42.5 (15)
Madrid: 37.8 (17.7)
Sevilla:45.1(14)
Range (Min, Max): NR
Copollutant:
BS
TSP
S02
Note: PMio only measured in 3 cities.
Increment: 10|jg/m
Relative Risk (Lower Cl, Upper Cl)
lag:
Nonaccidental:
Random effects:
1.006(0.998,1.015)0-1
Fixed Effects: 1.005 (1.001,1.010) 0-1
PM,o+S02:1.013 (1.006, 1.020)0-1
Cardiovascular:
1.012(1.005, 1.018)0-1
PM,o+S02:
Random effects:
1.024(1.001, 1.048)0-1
Fixed effects: 1.021 (1.007,1.035)0-1
Respiratory:
1.013(1.001, 1.026)0-1
PM10+S02:1.003 (0.983, 1.023)0-1
December 2009
E-293
-------
Study
Reference: Bateson and Schwartz
(2004, 0862441
Period of Study: 1988-1991
Location: Cook County, Illinois
Reference: Bell et al. (2009, 1910071
Period of Study: 1987-2000
Location: 84 U.S. Counties
Reference: Bell et al. (2007, 0932561
Period of Study: 1999-2005
Location: U.S.
Design & Methods
Outcome: Mortality:
Heart Disease (390-429)
Respiratory (460-51 9)
Study Design: Bi-directional case-
crossover
Statistical Analyses: Conditional
logistic regression
Age Groups: 2 65
Study population:
65, 1 80 elderly residents with history of
hospitalization for heart or lung disease
Outcome: Mortality
Study Design: Time-series
Covariates: Socio-economic
conditions, long term temperature
Statistical Analysis: Bayesian
hierarchical model
Age Groups: All
Outcome: Mortality
Age Groups: 65+
Study Design: Time series
N:NR
Statistical Analyses: Bayesian
Hierarchical Regression
Covariates: Time trend, day of week,
seasonality, dew point, temperature
Statistical Package: NR
Lags Considered: 0-2
Concentrations1
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SE) unit: 37.6(1 5.5) pg/m3
Range (Mm, Max): (3.7, 128)
Copollutant: NR
Pollutant: PM10
Averaging Time: 24 h
Mean (SD) Unit: NR
Range (Min, Max): NR
Copollutant (correlation): NR
Pollutant: PM,0
Averaging Time: Daily
Mean: Ni: 0.002
Min: Ni: 0.003
Max: Ni: 0.021
Interquartile Range: Ni: 0.001
Interquartile Range of Percents:
Ni: 0.01
Monitoring Stations: NR
Copollutant: Al, NH4+, As, Ca, Cl, Cu,
EC, CMC, Fe, Pb, Mg. Ni, N03-, K, Si,
Na+, S04=, Ti, V, Zn
Co-pollutant Correlation
Ni V- 0 4R
INI, V . U.HO
Mi Fp1 n ?n
INI, C\s. U.OU
Note: Pollutant concentrations available
for all fractions of PM2 5
Effect Estimates (95% Cl)
Increment: 10|jg/m3
% Increase (Lower Cl, Upper Cl) lag:
All-cause: 1.1 4% (0.44, 1.85)0-1
Modification of Effect by Prior Diagnosis
Myocardial Infarction:
1.98% (-0.25, 4.26)0-1
Diabetes: 1.49% (-0.06, 3.07)0-1
Congestive heart failure:
1.28% (-0.06, 2.64)0-1
COPD: 0.58% (-0.82, 2.00) 0-1
Conduction Disorders:
0.64% (-0.61, 1.90)0-1
All other heart or lung diseases:
0.74% (-0.29, 1.79)0-1
All-cause
Men
65: 2.0% (0.3, 3.8) 0-1
75: 1.5% (-0.2, 3.1)0-1
85: 0.9% -0.7, 2.5 0-1
95:0.3% -1.3,1.9 0-1
All: 1.3% (0.4, 2.3)0-1
Women
65: 0.1% (-1.6, 1.9)0-1
75: 0.7% (-1.1, 2.4) 0-1
85:1.2% -0.5,3.0)0-1
95:1.8% 0.03,3.6)0-1
All: 1.0% (0.1, 1.9) 0-1
Total
65: 1.1% (-0.12, 2.3)0-1
75: 1.1% (-0.1, 2.3) 0-1
85:1.2% -0.0,2.4)0-1
95:1.2% 0.0,2.4)0-1
All: 1.1% (0.4, 1.9)0-1
Increment: 20% of the population
acquiring air conditioning
Percent Change (96% Cl) in
community-specific PM health effect
estimates for mortality
Any AC, including window units
Yearly health effect: -30.4 (-80.4-19.6)
Summer health effect: 29.9 (-84-144)
Winter health effect: -573 (-9100-7955)
Central AC
Yearly health effect: -39 (-81. 4-3.3)
Summer health effect: 20. (-60.3-64.3)
Winter health effect: -1777 (-5755-2201)
PM Increment: Interquartile Range in
the fraction of PM25
Percent Increase in PM10 Health
Effect (Lower Cl, Upper Cl)
Ni: 14.8 (-8.1, 37.7), lag 0
Ni: 14.7 (4.0, 25.3) lag 1
Ni: 14.7 (1.8, 27.5) lag 2
HS education: -31. 9 (-82.4, 18.6)
median income: -12.3 (-62.3, 37.7)
Percent black: 48.7 (-15.8, 113)
Percent living in urban area: -20.1 (-
102,61.7)
Population: 5.1 (-14.4,24.5)
Notes: Interquartile ranges in percent
HS education, median income, percent
black, percent living in urban area, and
population are 5.2 %, $9,223, 17.3%,
11.0%, and 549,283 respectively.
December 2009
E-294
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Bellini et al. (2007, 0977871 Outcome: Mortality
Period of Study: 1996-2002
Location: 15 Italian cities
All-cause (nonaccidental) (<800)
Cardiovascular (390-459)
Respiratory (460-519)
Study Design: Meta-analysis
Statistical Analyses: Poisson GLM
Age Groups: All ages
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant:
S02
CO
03
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl) lag:
All-cause:
0.31% (-0.19, 0.74)0-1
Wnter: 0.08% 0-1
Summer: 1.95% 0-1
PM,o+03: 0.30% 0-1
PM,o+N02: 0.08% 0-1
Respiratory:
0.54% (-0.91, 1.74) 0-1
Wnter: 0.27% 0-1
Summer: 3.61% 0-1
PM,o+03: 0.55% 0-1
PMio+N02:0.19% 0-1
Cardiovascular:
0.54% (0.02, 1.02) 0-1
Wnter: 0.20% 0-1
Summer: 2.79% 0-1
PM,o+03: 0.57% 0-1
PMio+N02: 0.39% 0-1
Reference: Burnett et al. (2004,
0862471
Period of Study: 1981-1999
Location: 12 Canadian cities
Outcome: Mortality:
Nonaccidental (<800)
Study Design: Time-series
Statistical Analyses: 1. Poisson,
natural splines
2. Random effects regression model
Age Groups: All ages
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD):
PM25:12.8
PMi0.2.5: 11.4
Range (Min, Max): NR
Copollutant (correlation):
N02
03
S02
CO
Increment: 10|jg/m3
% Increase (Lower Cl,
lag:
1981-1999
Upper Cl)
PM,0: 0.57% (0.05, 0.89) 1
PM10+N02: 0.07% (-0.44, 0.58) 1
Note: PM10 measurement calculated as
the sum of PM2 5 and PMi0.2 5
measurements.
Reference: Cakmak et al. (2007,
0911701
Period of Study: Jan 1997-Dec 2003
Location: Chile-7 cities
Outcome: Mortality:
Nonaccidental (<800)
Cardiovascular diseases (390-459)
Respiratory diseases (460-519)
Study Design: Time-series
Statistical Analyses: Poisson
Random effects regression model
Age Groups: All age
<64yr
65-74 yr
75-84 yr
>85yr
Pollutant: PM10
Averaging Time: 24-h avg
Mean (SD): 84.9
Range (Min, Max): NR
Copollutant (correlation):
03:r =-0.16-0.13
S02:r = 0.37-0.77
CO: r = 0.49-0.82
Note: Correlations are between
pollutants for seven monitoring stations.
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl) lag:
Nonaccidental:
0.97% (-1.09, 2.76 0
1.31% (-1.56, 3.68 0-5
PMio+03+S02+CO:
0.80% (-0.87, 2.28) 0
0.52% (-0.55, 1.51) 0
0.49% (-0.51, 1.43) 0-5
65-75:
1.07% (-1.23, 3.03) 0
1.31% (-1.57, 3.69) 0-5
75-84:
1.41% (-1.71, 3.94) 0
1.93% -2.57, 5.30 0-5
>85:
1.56% (-1.94, 4.34) 0
2.14% (-2.97, 5.85) 0-5
Apr-Sep:
1.03% (-1.17, 2.93) 0
1.37% (-1.64, 3.82) 0-5
Oct-Mar:
0.07% (-0.07, 0.21) 0
0.15% (-0.15, 0.44) 0-5
Cardiovascular:
1.14% (-1.31, 3.21) 0
1.49% (-1.82, 4.14) 0-5
Respiratory:
2.03% (-2.75, 5.56) 0
3.11% (-5.25, 8.25) 0-5
December 2009
E-295
-------
Study
Reference: Chen et al. (2008, 1901061
Period of Study: 2001 -2004
Location: Shanghai, China
Reference: Daniels et al. (2004,
087343)
Period of Study: 1987-1994
Location: 20 Largest U.S. cities
Design & Methods
Outcome
(ICD9: 2001; ICD10: 2002-2004):
Mortality:
Nonaccidental causes (ICD9 <800;
ICD10AOO-R99)
Cardiovascular (ICD9 390-459; I
CD10IOO-I99)
Respiratory (ICD9 460-51 9;
ICD10JOO-J98)
Study Design: Time-series
Statistical Analyses: Poisson GAM
Age Groups: All ages
Outcome:
Mortality:
Total (Nonaccidental) mortality
Cardiovascular-Respiratory (390-448)
(480-486, 487, 490-496, 507)
Other-cause mortality
Study Design: Time-series
Statistical Analyses: City-Specific
Estimates: Poisson GLM, natural cubic
splines
Combined Estimates: 2-stage Bayesian
hierarchical model
Age Groups: All ages
Concentrations1
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): 102.0
Range (Mm, Max): (14.0-566.8)
Copollutant (correlation):
S02r = 0.64
N02r = 0.71
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD):
Los Angeles: 46.0
New York: 28. 8
Chicago: 35.6
Dallas-Ft. Worth: 23.8
Houston: 30.0
San Diego: 33.6
Santa Ana-Anaheim: 37. 4
Phoenix: 39.7
Detroit: 40.9
Miami: 25.7
Philadelphia: 35.4
Minneapolis: 26.9
Seattle: 25.3
San Jose: 30.4
Cleveland: 45.1
San Bernardino: 37.0
Pittsburgh: 31. 6
Oakland: 26.3
Atlanta: 34.4
San Antonio: 23.8
Effect Estimates (95% Cl)
Increment: 10|jg/m3
% Increase (Lower Cl, Upper Cl) lag:
Nonaccidental
Single Pollutant: 0.26% (0.14, 0.37)
PM,o+S02: 0.08% (-0.07, 0.22)
PM10+N02: 0.01% (-0.14, 0.17)
PM10+S02+N02: 0.00% (-0.16, 0.16)
Cardiovascular mortality
Single Pollutant: 0.27% (0.10, 0.44)
PM,o+S02: 0.12% (-0.10, 0.34)
PM,o+N02: 0.01% (-0.22, 0.25)
PM10+S02+N02: 0.01% (-0.23, 0.25)
Respiratory mortality
Single Pollutant: 0.27% (-0.01, 0.56)
PM10+S02: -0.04% (-0.41, 0.33)
PM10+N02: -0.05% (-0.45, 0.34)
PMio+S02+N02: -0.10% (-0.50, 0.30)
Increment: 10|jg/m3
% Increase (Lower Cl, Upper Cl) lag:
Total (nonaccidental):
0.17% (0.03, 0.30)0
0.20% (0.07, 0.33) 1
0.28% (0.16, 0.41)0-1 avg
Cardiovascular-Respiratory:
0.17% (-0.01, 0.35)0
0.27% (0.09, 0.44) 1
0.30% (0.18, 0.51)0-1 avg
Other-cause:
0.17% (-0.03, 0.37)0
0.12% (-0.07, 0.31)1
0.20% (0.01, 0.38) 0-1 avg
Threshold Models: Total Mortality
Threshold = 15 pg/m3
0.30% (0.17, 0.42)0-1 avg
Threshold = 0 pg/m3
0.28% (0.16, 0.41)0-1 avg
Threshold = 20 pg/m3
0.30% (0.16, 0.43)0-1 avg
December 2009
E-296
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: De Leon et al. (2003,
0556881
Period of Study: Jan 1985-Dec 1994
Location: New York, New York
Outcome: Mortality:
Circulatory (390-459)
Cancer (140-239)
Study Design: Time-series
Statistical Analyses: Poisson GAM
Age Groups: All ages
<75yr
>75yr
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD):
33.27 pg/m3
IQR (25th, 75th):
(22.67, 40.83)
Copollutant (correlation):
03
CO
S02
NO,
Increment: 18.16 pg/m
Relative Risk (Lower Cl, Upper Cl)
lag:
All Ages
Cancer: 1.014 (1.000,1.029)0-1
-w/out respiratory:
1.011(0.996,1.026)0-1
-w/ respiratory:
1.051(0.998,1.107)0-1
Circulatory: 1.025 (1.014,1.035)0-1
-w/out respiratory:
1.022(1.012,1.033)0-1
-w/ respiratory:
1.054(1.022, 1.086)0-1
<75yr
Cancer: 1.003 (0.985,1.021)0-1
-w/out respiratory:
1.002(0.983,1.022)0-1
-w/ respiratory:
1.009(0.943, 1.078)0-1
Circulatory: 1.027 (1.012,1.043)0-1
-w/out respiratory:
1.027(1.011, 1.043)0-1
-w/ respiratory:
1.033(0.980,1.089)0-1
>75yr
Cancer: 1.033 (1.009,1.058)0-1
-w/out respiratory:
1.025(1.000, 1.050)0-1
-w/ respiratory:
1.129(1.041,1.225)0-1
-w/out pneumonia:
1.026(1.002,1.050)0-1
-w/ pneumonia:
1.183(1.058, 1.323)0-1
-w/out COPD: 1.032 (1.008, 1.057)0-1
-w/COPD: 1.008 (0.849, 1.197)0-1
Circulatory: 1.025 (1.012,1.038)0-1
-w/out respiratory:
1.022(1.008,1.035)0-1
-w/ respiratory:
1.066(1.027,1.106)0-1
-w/out pneumonia:
1.023(1.010, 1.036)0-1
-w/ pneumonia:
1.078(1.018,1.141)0-1
-w/out COPD:
1.025(1.012,1.038)0-1
-w/COPD:
1.058(0.991, 1.130)0-1
Reference: Dominici et al. (2003,
0428041
Period of Study: 1987-1994
Location: 88 U.S. cities
Outcome: Mortality:
All-cause (nonaccidental) (<800)
Cardiac (390-448)
Respiratory (490-496)
Influenza (487)
Pneumonia (480-486, 507)
Other causes
Study Design: Time-series
Statistical Analyses: 2-stage Bayesian
hierarchical model
Age Groups: <65 yr; 65-74 yr; 2 75 yr
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: 10 pg/m
% Increase (Lower Cl, Upper Cl) lag:
Cardio-respiratory
0.31% (0.15, 0.50)1
All-cause
0.22% (0.10, 0.38)1
Other causes
0.13% (-0.05, 0.29)1
Reference: Dominici et al. (2004,
0591581
Period of Study: 1987-1994
Location: 90 U.S. cities (NMMAPS)
Outcome: Mortality:
Total (nonaccidental)
Study Design: Time-series
Statistical Analyses: Poisson. GAM,
GLM
Age Groups: All ages
Pollutant: PM10
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Increment: 10 pg/m3
% Increase (Lower Cl, Upper Cl)
lag:
a = 3
0.2% (0.05, 0.35)
December 2009
E-297
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Dominici et al. (2004,
0969511
Period of Study: 1986-1993
Location: 10 U.S. cities
Outcome: Mortality:
Total (nonaccidental)
Study Design: Time-series
Statistical Analyses: 2-stage Bayesian
hierarchical model
Age Groups: All ages
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD):
Birmingham 34.8
Canton 28.4
Colorado Springs 27.5
Minneapolis/St. Paul 28.1
Seattle 32.2
Spokane 42.9
Chicago 36.3
Detroit 36.7
New Haven 28.6
Pittsburgh 36.0
New York: 28.8
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl) lag:
Combined analysis:
0.26% (-0.37, 0.65) 0-1
Separate analysis:
0.28% (-0.12, 0.63)0-1
Notes: A separate analysis assumes
the mortality data does not provide any
information on the log relative rates of
mortality.
Reference: Dominici et al. (2007,
0973611
Period of Study: PM10:1987-2000
PM25:1999-2000
Location: 100 U.S. counties
(NMMAPS)
Outcome: Mortality:
All-cause (nonaccidental)
Cardiorespiratory
Other-cause
Study Design: Time-series
Statistical Analyses: 2-stage Bayesian
hierarchical model
Age Groups: All ages
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl) lag:
PM10
All-cause:
East:
1987-1994: 0.29% (0.12, 0.46)1
1995-2000: 0.13% (-0.19, 0.44)1
1987-2000: 0.25% (0.11, 0.39)1
V\fest:
1987-1994: 0.12% (-0.07, 0.30)1
1995-2000: 0.18% (-0.07, 0.44)1
1987-2000: 0.12% (-0.02, 0.26)1
National:
1987-1994: 0.21% (0.10, 0.32)1
1995-2000: 0.18% (0.00, 0.35)1
1987-2000: 0.19% (0.10, 0.28)1
Cardiorespiratory:
East:
1987-1994: 0.39% (0.16, 0.63)1
1995-2000: 0.30% (-0.13, 0.73)1
1987-2000: 0.34% (0.15, 0.54)1
V\fest:
1987-1994: 0.17% (-0.07, 0.40)1
1995-2000: 0.13% (-0.23, 0.50
1987-2000: 0.14% (-0.05, 0.33
National:
1987-1994: 0.28% (0.14, 0.43)
1
1
1
1995-2000: 0.21% (-0.03, 0.44)1
1987-2000: 0.24% (0.13, 0.36)
Other-cause:
East:
1
1987-1994: 0.21% (-0.03, 0.44)1
1995-2000: 0.00% (-0.49, 0.50
1987-2000: 0.15% (-0.09, 0.39
V\fest:
1987-1994: 0.09% (-0.21, 0.38
1995-2000: 0.23% (-0.15, 0.62
1
1
1
1
1987-2000: 0.17% (-0.07, 0.41)1
National:
1987-1994: 0.15% (-0.02, 0.32)1
1995-2000: 0.17% (-0.07, 0.41)1
1987-2000: 0.15% (0.00, 0.29)1
Reference: Dominici et al. (2007, Outcome: Total mortality Pollutant: PM10
099135)
Study Design: Time-series Averaging Time: 24-h avg
Period of Study: 2000-2005
Statistical Analyses: 2-stage Bayesian Mean (SD): NR
Location: 72 U.S. counties hierarchical model
representing 69 communities Range (Mm, Max): NR
Age Groups: All ages ....... ...,„,„
Copollutant correlation : NR
The study does not provide results
quantitatively.
Note: The study investigated whether
county-specific short-term effects of
PM10 on mortality are modified by long-
term county-specific nickel or vanadium
PM2 .5 concentrations.
December 2009
E-298
-------
Study
Reference: Fischer et al. (2003,
0437391
Period of Study: 1986-1994
Location: The Netherlands
Reference: Fischer et al. (2004,
0556051
Period of Study: Jun 2003-Aug 2003
Location: The Netherlands
Reference: Forastiere et al. (2005,
0863231
Period of Study: 1998-2000
Location: Rome, Italy
Design & Methods
Outcome: Mortality:
Nonaccidental (<800)
Pneumonia (480-486)
COPD (490-496)
Cardiovascular (390-448)
Study Design: Time-series
Statistical Analyses: Poisson GAM,
LOESS
Age Groups:
<45yr
45-64 yr
65-74 yr
>75yr
Outcome: Total mortality
Study Design: NR
Statistical Analyses: NR
Age Groups: All ages
Outcome: Mortality:
Ischemic heart disease (410-414)
Study Design: Time-stratified case-
crossover
Statistical Analyses: Conditional
logistic regression
Age Groups: >35
Concentrations1
Pollutant: PM,0
Averaging Time: 24-h avg
Median (SD) unit: 34
Range (Min, Max): (10, 278)
Copollutant:
BS
03
N02
S02
CO
Pollutant: PM,0
Averaging Time: Wfeekly avg
Mean (SD):
2000: 31
2002: 33
2003: 35
IQR (26th, 76th): NR
Copollutant: 0;
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD):
CO -I /OO O\
IQR (25th, 75th):
(36.0, 65.7)
Copollutant (correlation):
PNC1 r = 038
CO: f = 0.34
N02:r = 0.45
S02:r = 0.23
03:r = 0.13
Effect Estimates (95% Cl)
Increment: 80 pg/m3
Relative Risk (Lower Cl, Upper Cl)
lag:
Cardiovascular
<45: 0.906 (0.728, 1.128)0-6
45-64:1.023(0.945,1.106)0-6
65-74:1.002(0.945,1.062)0-6
> 75: 1.016 (0.981, 1.052)0-6
COPD
<45: 1.153 (0.587, 2.268)0-6
45-64:1.139(0.841, 1.541)0-6
65-74:1.166(0.991,1.372)0-6
> 75: 1.066 (0.965, 1.178)0-6
Pneumonia
<45: 1.427 (0.806, 2.525)0-6
45-64:1.712(1.042,2.815)0-6
65-74:1.240 0.879, 1.748 0-6
> 75: 1.1 23 (1.01 1,1. 247) 0-6
The study does not present quantitative
results.
Notes: The study estimates the number
of deaths attributable to PM10 during the
summers of 2000, 2002, and 2003.
Increment: 29.7 pg/m3
% Increase (Lower Cl, Upper Cl)
lag:
4. 8% (0.1, 9. 8)0
4. 9% (0.0, 10.1)1
3. 8% (-1.0, 8.9)2
2.8% (-2.0, 7.7) 3
6.1% (0.6, 11.9)0-1
Reference: Forastiere et al. (2007,
0907201
Period of Study: 1998-2001
Location: Rome, Italy
Outcome: Mortality:
Natural (<800)
Malignant neoplasms (140-208)
Diabetes mellitus (250)
Hypertensive disease (401-405)
Previous acute myocardial infarction
(410, 412)
Other ischemic heart diseases (411,
413-414)
Conduction disorders (426)
Dysrhythmia (427)
Heart failure (428)
Cerebrovascular disease (430-438)
Peripherical artery disease (440-448)
COPD (490-496)
Study Design: Time-stratified case-
crossover
Statistical Analyses: Conditional
logistic regression
Age Groups: >35
Pollutant: PM10
Averaging Time: 24-h avg
Mean Range (SD) unit: 51.0
(21.0)fjg/rr?
IQR (25th, 75th):
(36.1,63.0)
Copollutant (correlation): NR
Increment: 10 pg/m3
% Increase (Lower Cl, Upper Cl)
lag:
Nonaccidental: 1.1% (0.7,1.6)0-1
Low income: 1.9% 0-1
LowSES:1.4% 0-1
High income: 0.0% 0-1
High SES: 0.1% 0-1
Low PM Area: 0.9% (-0.4, 2.1)0-1
High PM Area: 1.47% (0.4, 2.5)0-1
December 2009
E-299
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Forastiere et al. (2008,
1869371
Period of Study: 1997-2004
Location: 9 Italian cities
Outcome: Mortality:
Nonaccidental (<800)
Study Design: Time-stratified case-
crossover
Statistical Analyses: Conditional
logistic regression
Age Groups: >35
Pollutant: PM,0
Averaging Time: 24-h avg
Mean Range (SD) unit:
35.1-71.5
Range (5th, 95th):
Lowest 5th: 14.3
Highest 95th: 147.0
Copollutant (correlation): NR
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl) lag:
Total: 0.60% (0.31, 0.89) 0-1
Age
35-64: -0.20% (-0.77, 0.37) 0-1
65-74: 0.51% (0.05, 0.98)0-1
75-84: 0.59%(0.20, 0.97) 0-1
> 85: 0.97% (0.53, 1.42)0-1
> 65: 0.75% (0.42, 1.09)
Sex
Men: 0.72% (0.37, 1.07)0-1
Vtomen: 0.83% (0.33,1.33)0-1
Median income (by census block)
Low (<20th percentile):
0.80% (-0.02, 1.62)0-1
Mid-low (20th-50th percentile):
0.68% (0.25, 1.12)0-1
Mid-high (51st-80th percentile):
0.85% (0.40, 1.30)0-1
High (>80th percentile):
0.30% (-0.25, 0.86) 0-1
Location of death
Out-of-hospital: 0.71% (0.32,1.11) 0-1
Discharged 2-28 days before death:
1.34% (0.49, 2.20) 0-1
In-hospital: 0.65% (0.33, 0.97) 0-1
Nursing home: -0.04% (-1.02, 0.95) 0-1
Reference: Goldberg et al. (2003,
0352021
Period of Study: 1984-1993
Location: Montreal, Quebec, Canada
Reference: Goldberg et al. (2003,
0352021
Period of Study: 1984-1993
Location: Montreal, Quebec, Canada
Reference: Kan and Chen (2003,
0873721
Period of Study: Jan 2000-Dec 2001
Location: Shanghai, China
Outcome: Mortality: Congestive Heart
Failure (428)
Study Design: Time-series
Statistical Analyses: Poisson, natural
splines
Age Groups: 2 65
Outcome: Mortality:
Diabetes (250)
Study Design: Time-series
Statistical Analyses: Poisson, natural
spline
Age Groups: 2 65
Outcome: Mortality:
Nonaccidental (<800)
Cardiovascular (390-459)
COPD (490-496)
Study Design: Time-series
Statistical Analyses: Poisson GAM,
LOESS
Age Groups: All ages
<65yr
65-75 yr
>75yr
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD):PM10: 32.2 (17.6)
IQR(25th, 75th): PM,0: (19.7, 41.1)
Copollutant (correlation): PM25, TSP,
Sulfate, CoH, S02, N02, CO, 03
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD):
PM,0: 32.2 (17.6) pg/m3
IQR (25th, 75th):
PM10: (19.7, 41.1)
Copollutant (correlation): PM25,
Sulfate, CoH, S02, N02, CO, 03
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): 91. 14 (51. 85)
Range (Min, Max): (17.0, 385.0)
Copollutant (correlation):
S02:r = 0.71
N02: r = 0.73
This study does not present results
quantitatively for PM10
This study does not present results
quantitatively for PM10
Increment: 10|jg/m3
Relative Risk (Lower Cl, Upper Cl)
lag:
Nonaccidental:
All ages: 1.003 (1.001, 1.005)0
<65: 1.001 (0.997, 1.005)0
65-75:1.005(1.001, 1.008)0
>75: 1.003 (1.001, 1.006)0
Cardiovascular:
All ages: 1.003 (1.000, 1.006)0
<65: 1.002 (0.994, 1.010)0
65-75:1.003(0.998, 1.008)0
>75: 1.003 (1.000, 1.006)0
COPD:
All ages: 1.005 (0.999, 1.011)0
<65: 1.004 (0.981, 1.027)0
65-75:0.996(0.986, 1.007)0
>75: 1.006 (1.000, 1.012)0
Multipollutant models:
S02: 1.001 (0.998, 1.003)0
N02: 1.001 (0.998, 1.003)0
S02+N02: 1.000 (0.997, 1.003)0
December 2009
E-300
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Kan and Chen (2003,
0873721
Period of Study: Jan 2000-Dec 2001
Location: Shanghai, China
Outcome: Mortality:
Nonaccidental (<800)
Cardiovascular (390-459)
COPD (490-496)
Study Design: Case-crossover
Statistical Analyses: Conditional
logistic regression
Age Groups: All ages
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): 91.14 (51.85)
IQR (25th, 75th): (54,114)
Copollutant (correlation):
S02:r = 0.71
NO,: r = 0.73
Increment: 10|jg/m
Odds Ratio (Lower Cl, Upper Cl) lag:
Nonaccidental:
Bidirectional referent days:
7 days: 1.000 (0.9988,1.002) 0-1 ma
7 and 14 days:
1.002(1.000, 1.004) 0-1 ma
7, 14, and 21 days:
1.003 (1.001, 1.005) 0-1 ma
Unidirectional referent days:
7 days: 1.015 (1.012,1.018)0-1 ma
7 and 14 days: 1.017 (1.015,1.019)
0-1 ma
7, 14, and 21 days: 1.019(1.012,
1.021) 0-1 ma
Bidirectional referent days (7,14, and
21 days):
Cardiovascular:
1.004 (1.001, 1.007) 0-1 ma
COPD:
1.006(0.999, 1.013) 0-1 ma
Nonaccidental:
PMio+S02: 0.997 (0.994, 1.025) 0-1 ma
PM10+N02: 0.997 (0.994, 1.025) 0-1 ma
PM10+S02+N02: 0.995 (0.992, 1.025)
0-1 ma
Reference: Kan et al. (2005, 0875611
Period of Study: Apr 2003-May 2003
Location: Beijing, China
Outcome: Mortality:
Severe acute respiratory syndrome
(SARS)
Study Design: Time-series
Statistical Analyses: Poisson, GAM,
smoothing spline
Age Groups: All ages
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): 149.1 (8.1)
Range (Min, Max): (34, 246)
Copollutant:
S02
N02
Increment: 10|jg/m
Relative Risk (Lower Cl, Upper Cl)
lag:
0.99(0.96-1.03)0
1.00(0.97-1.04)1
1.02 0.98-1.06 2
1.04(0.99-1.09)3
1.06(1.00-1.11)4
1.06 1.00-1.12)5
1.05(0.98-1.12)6
Reference: Kan et al. (2007, 0912671
Period of Study: Mar 2004-Dec 2005
Location: Shanghai, China
Outcome (ICD10): Mortality:
Total (nonaccidental) (AOO-R99)
Cardiovascular (IOO-I99)
Respiratory (JOO-J98)
Study Design: Time-series
Statistical Analyses: Poisson GAM,
penalized splines
Age Groups: All ages
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): 107.9 (2.39) pg/m3
Range (Min, Max): (22.0, 403.0)
Copollutant (correlation):
PM10
PM25:r = 0.84
PMi0.25:r = 0.88
03:r = 0.21
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl) lag:
PM10
Total: 0.16% (0.02, 0.30)0-1
Cardiovascular: 0.31% (0.10, 0.53) 0-1
Respiratory: 0.33% (-0.08, 0.75) 0-1
December 2009
E-301
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Kan et al. (2008,1566211
Period of Study: Jan 2001-Dec 2004
Location: Shanghai, China
Outcome: Mortality: Total
(nonaccidental) (AOO-R99)
Cardiovascular (IOO-I99)
Respiratory (JOO-J98)
Study Design: Time-series
Statistical Analyses: Poisson GLM,
natural splines
Age Groups:
All ages;
0-4 yr
5-44 yr
45-64 yr
>65yr
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD):
Warm season: 87.4 (1.8)
Cool season: 116.7 (2.8)
Entire period: 102.0 (1.7)
Range (Min, Max): NR
Copollutant (correlation):
S02
N02
03
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl) lag:
Nonaccidental
Warm season: 0.21 (0.09, 0.3) 0-1
Cool season: 0.26 (0.22, 0.30) 0-1
Entire period: 0.25 (0.14, 0.37)0-1
Female: 0.33 (0.18, 0.48)0-1
Male: 0.17 (0.03, 0.32)0-1
5-44: 0.04 (-0.52, 0.59) 0-1
45-64: 0.17 (-0.11, 0.45) 0-1
> 65: 0.26 (0.15, 0.38)0-1
Cardiovascular
Warm season: 0.22 (-0.14, 0.58) 0-1
Cool season: 0.25 (0.05, 0.45) 0-1
Entire period: 0.27 (0.10, 0.44)0-1
Respiratory
Warm season: -0.28 (-0.93, 0.38) 0-1
Cool season: 0.58 (0.25, 0.92) 0-1
Entire period: 0.27 (-0.01, 0.56) 0-1
Stratified by Educational Attainment
Nonaccidental:
Low: 0.33 (0.19, 0.47)0-1
High: 0.18 (0.01, 0.36) 0-1
Cardiovascular:
Low: 0.30 (0.10, 0.51)0-1
High: 0.23 (-0.03, 0.50) 0-1
Respiratory:
Low: 0.36 (0.00, 0.72) 0-1
High: 0.02 (-0.43, 0.47) 0-1
Reference: Keatinge and Donaldson
(2006, 0875361
Period of Study: 1991-2002
Location: London, England
Outcome: Mortality: Total
(nonaccidental)
Study Design: Time-series
Statistical Analyses: Poisson GAM
Age Groups: > 65 yr
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant:
03
SO,
Increment: 10|jg/m
Mortality per 106 (Lower Cl, Upper
Cl) lag:
PM10+Temp:
2.1 (0.9, 3.3) 0-2 avg
PMio+Temp+Acclim:
1.6(0.4, 2.8) 0-2 avg
PM10+Temp+Acclim+Acclim x T:
1.5(0.3, 2.6) 0-2 avg
PM10+Temp+Acclim+Acclim x T+Sun:
1.4(0.2, 2.5) 0-2 avg
PMio+Temp+Acclim+Acclim x
T+Sun+Wnd: 0.8 (-0.4,1.9) 0-2 avg
PM10+Temp+Acclim+Acclim x
T+Sun+Wind+Abs. Humid.:
0.8 (-0.3,1.9) 0-2 avg
PMio+Temp+Acclim+Acclim x
T+Sun+Wind+Abs. Humid.+ Rain:
0.9 (-0.3, 2.0) 0-2 avg
PMio+Temp+Abs. Humid.:
1.9(0.7, 3.1) 0-2 avg
Reference: Kettunen et al. (2007,
0912421
Period of Study: 1998-2004
Location: Helsinki, Finland
Outcome (ICD10): Mortality:
Stroke (160-161,163-164)
Study Design: Time-series
Statistical Analyses: Poisson GAM,
penalized thin-plate splines
Age Groups: > 65
Pollutant: PM,0
Averaging Time: 24-h avg
Median (SD) unit:
Cold Season: 16.3
Warm Season: 16.5
Range (Min, Max):
Cold Season: (3.1,136.7)
Warm Season: (3.3, 67.4)
Copollutant:
PM25
PMlO-25
UFP
03
CO
NO,
Increment:
Cold Season: 13.8 pg/m3
Warm Season: 9.8 pg/m3
% Increase (Lower Cl, Upper Cl) lag:
Cold Season
-0.56%
-0.93%
-3.32, 2.29) 0
-3.55,1.75)1
-1.68% (-4.30, 1.00)2
-1.53% (-4.14, 1.14)3
Warm Season
10.89% (0.95, 21.81)0
8.56% (-0.88, 18.90) 1
2.06% (-6.76, 11.71)2
-2.89% (-11.32, 6.34)3
December 2009
E-302
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Kim et al. (2003,1558991
Period of Study: Jan 1995-Dec 1999
Location: Seoul, Korea
Outcome (ICD10): Mortality:
Nonaccidental (all except S01-S99,
T01-T98)
Cardiovascular (IOO-I52)
Respiratory (JOO-J98)
Cerebrovascular (I60-I69)
Study Design: Time-series
Statistical Analyses: Poisson GAM
Age Groups: All ages
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): 69.19 (10.36)
IQR (25th, 75th):
(44.82, 87.95)
Copollutant (correlation): NR
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl) lag:
All cause:
2.8% (1.8, 3.7)0
2.8%
1.9, 3.7) 1
0.5, 2.3) 2
3.7% (2.1, 5.4) distributed lag (6-day)
Respiratory:
8.3% (4.3, 12.5) 0
6.4%
6.5%
2.7, 10.2) 1
2.7,10.4)2
13.9% (6.8, 21.5) distributed lag (6-day)
Pneumonia:
11.6% (4.2,19.6)0
9.0%
7.7%
2.1, 16.3)1
0.8, 15.2) 2
17.1% (4.1, 31.7) distributed lag (6-day
COPD:
4.2% (-1.2, 10.0)0
3.5%
-1.5,8.9)1
-3.7, 6.8) 2
12.2% (2.5, 22.9) distributed lag (6-day
)
Cardiovascular:
2.0% (-0.9, 5.0) 0
3.3%
2.9%
0.6, 6.2) 1
0.1,5.8)2
4.4% (-0.6, 9.6) distributed lag (6-day)
Myocardial infarction:
2.6% (-2.3, 7.8) 0
5.8%
5.5%
1.0, 10.7) 1
0.7, 10.6) 2
4.9% (-3.4,13.9) distributed lag (6-day)
Cerebrovascular:
3.2% (0.8, 5.5) 0
3.1%
2.4%
0.9, 5.3) 1
0.1,4.6)2
6.3% (2.3,10.5) distributed lag (6-day)
Ischemic stroke:
-0.6% (-5.6, 4.7) 0
0.6% (-4.2, 5.7) 1
-0.1% (-4.9, 5.1)2
10.3% (1.0, 20.4) distributed lag (6-day)
Reference: Kim et al. (2004, 0874171
Period of Study: Jan 1997-Dec 2001
Location: Seoul, Korea
Outcome: Mortality: Nonaccidental
Study Design: Time-series
Statistical Analyses: Poisson GAM,
LOESS
Age Groups: All ages
Pollutant: PM10
Averaging Time: 24-h avg
Mean (SD): 68.23 (36.36) pg/m3
IQR (25th, 75th): (42.56, 84.67)
Copollutant (correlation): NR
Increment: 42.11 pg/m3
Relative Risk (Lower Cl, Upper Cl)
lag:
1.021 (1.009, 1.035)
December 2009
E-303
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Le Tertre et al. (2005,
0875601
Period of Study: NR
Location: 21 European cities
(APHEA-2)
Outcome: Mortality: Pollutant: PMi0
Nonaccidental (<800) Averaging Time: 24-h avg
Study Design: Time-series Mean (SD): NR
Statistical Analyses: Empirical Bayes Range (Min, Max): NR
Age Groups: All ages Copollutant: N0;
Increment: 1.0|jg/m
P coefficient (SE) lag:
Athens: 0.001311 (0.0003)
Barcelona: 0.000575 (0.0002)
Basel: 0.000462 (0.0005)
Birmingham: 0.000305 (0.0003)
Budapest: -0.000248 (0.0005)
Cracow: 0.000155 (0.0004)
Erfurt: -0.000465 (0.0004)
Geneva: -0.000059 (0.0005)
Helsinki: 0.000389 (0.0004)
London: 0.000591 (0.0002)
Lyon: 0.001554 (0.0005)
Madrid: 0.000372 (0.0003)
Milan: 0.000901 (0.0002)
Paris: 0.000411 (0.0003)
Prague: 0.000097 (0.0002)
Rome: (0.001333 (0.0003)
Stockholm: 0.000479 (0.0009)
Tel Aviv: 0.000522 (0.0003)
Teplice: 0.000876 (0.0004)
Torino: 0.000938 (0.0002)
Zurich: 0.000365 (0.0004)
Toulouse: NR (NR)
Overall: 0.00055 (0.000098)
Reference: Lee et al. (2007, 0930421
Period of Study: Jan 2000-Dec 2004
Location: Seoul, Korea
Outcome (ICD10): Mortality:
Nonaccidental (AOO-R99)
Study Design: Time-series
Statistical Analyses: Poisson GAM
Age Groups: All ages
Pollutant: PM10
Averaging Time: 24-h avg
Mean (SD):
w/Asian dust days: 70.00 (47.80)
w/o Asian dust days: 65.77 (33.60)
Asian dust days only: 188.49 (142.85)
Copollutant:
CO
N02
S02
Increment: 41.49 pg/m
% Increase (Lower Cl, Upper Cl) lag:
Model with Asian Dust Days
0.7% (0.2, 1.3)1-3
Model without Asian dust days
1.0% (0.2,1.8)1-3
Reference: Lee and Shaddick (2007,
1566851
Period of Study: Jan 1993-Dec 1997
Location: Cleveland, Ohio
Detroit, Michigan
Minneapolis, Minnesota
Pittsburgh, Pennsylvania
Outcome (ICD10): Mortality:
Nonaccidental
Study Design: Time-series
Statistical Analyses:
1. Bayesian, penalized spline
2. Likelihood, penalized spline
Age Groups: All ages
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Increment: 10|jg/m
Relative Risk (Lower Cl, Upper Cl)
lag:
Constant model
Cleveland: 1.0049
1
Detroit: 1.0046
1
Minneapolis: 1.0052
1
Pittsburgh: 1.0045
Reference: Martins et al. (2004,
0874571
Period of Study: Jan 1997-Dec 1999
Location: Sao Paulo, Brazil
Outcome (ICD10): Mortality:
Respiratory (JOO-J99)
Study Design: Time-series
Statistical Analyses: Poisson GLM,
natural cubic splines
Age Groups: > 60
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD):
Cerqueira Cesar: 42.5(22.9)
Santa Amaro: 49.6(32.1)
Central: 52.1(23.5)
Penha: 40.4(23.8)
Santana: 72.6(24.5)
Sao Miguel Paulista: 68.6(31.0)
Range (Min, Max): NR
The study does not present quantitative
results.
December 2009
E-304
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Nawrot et al. (2007,
0986191
Period of Study: Jan 1997-Dec 2003
Location: Flanders, Belgium
Outcome: Mortality:
Nonaccidental (<800)
Cardiovascular (390-459)
Respiratory (460-519)
Study Design: Time-series
Statistical Analyses: Main analysis:
Segmented regression models
Sensitivity analysis: Poisson GAM,
LOESS
Age Groups: All ages
Pollutant: PM,0
Averaging Time: 24-h avg
Median (SD) unit:
Winter: 43.3(0.88)
Spring: 39.5(0.88)
summer: 37.7(0.91)
Fall: 37.2(0.88)
Range (Min, Max): NR
Copollutant (correlation): NR
Increment:
Main analysis: NR
Sensitivity analysis: 10 pg/m3
% Increase (Lower Cl, Upper Cl) lag:
Highest season-specific PM10 quartile
vs. the lowest season-specific PM10
quartile
Summer: 7.8% (6.1,9.6)
Spring: 6.3% (4.7, 7.8)
Fall: 2.2% (0.58, 3.8)
Winter: 1.4% (0.06, 2.9)
Warm months (Jun, Jul, Aug):
7.9% (6.2, 9.6)
Cold months (Dec, Jan, Feb):
1.5% (0.22, 3.3)
Intermediate months (Mar, Apr, May,
Sep, Oct, Nov): 4.2% (2.9, 5.6)
Warmer Periods (Apr-Sep)
Nonaccidental: 1.5% (1.1, 2.0)0
Respiratory: 2.0% (0.6, 3.7) 0
Cardiovascular: 1.8% (1.1, 2.4)0
December 2009
E-305
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: O'Neill et al. (2004,
0874291
Period of Study: 1996-1998
1994-1995
Location: Mexico City, Mexico
Outcome: Mortality:
Nonaccidental
Study Design: Time-series
Statistical Analyses: Poisson, natural
cubic spline
Age Groups: All ages
Pollutant: PM,0
Averaging Time: 24-h avg
Range:
Hi-Vol: 46.3-164.0
TEOM: 48.2-107.5
Predicted: 30.2-162.4
Impactor: 58.4
Range (Min, Max):
Xalostoc
Hi-Vol: (40.0, 335.0)
TEOM: (16.5, 291.2)
Predicted: (60.6, 320.0)
Tlalnepantla
Hi-Vol: (25.0, 264.0)
TEOM: (10.4, 275.9)
Predicted: (17.7,175.0)
Merced
Hi-Vol: (17.0, 266.0)
TEOM: (9.4, 318.7)
Predicted: (12.3,160.8)
Cerro de la Estrella
Hi-Vol: (15.0, 292.0)
TEOM: (13.7, 268.3)
Predicted: (11.2,154.4)
Pedregal (1996-1998)
Hi-Vol: (5.0, 226.0)
TEOM: (7.8, 264.4)
Predicted: (-0.5, 86.3)
Pedregal (1994-1995)
Hi-Vol: (24.0, 114.0)
TEOM: (8.7, 152.5)
Impactor: (15.0,154.0)
Predicted: (3.9, 75.9)
Reference: O'Neill et al. (2005,
0980941
Period of Study: 1996-1998
1996-1999
Location: Mexico City and Monterrey,
Mexico
Outcome: Mortality: Nonaccidental
Cardiovascular (390-460)
Respiratory (460-520)
Other-causes
Study Design: Time-series
Statistical Analyses: Poisson, natural
cubic splines
Age Groups: All ages, 0-15, > 65
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD):
Mexico City: 75.8 (31.4)
Monterrey: 50.0 (23.5)
Range (Min, Max):
Mexico City: (18.0, 233.9)
Monterrey: (6.2, 230.8)
Copollutant: 0;
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl) lag:
TEOM
0.04% (-0.12, 0.20)0
-0.02%
-0.01%
-0.18,0.13)1
-0.27, 0.25) 2
-0.03% (-0.19, 0.13)3
-0.03% (-0.19, 0.13)4
-0.05% (-0.21, 0.11) 5
0.05% (-0.25, 0.35) 0-5
Predicted
-0.05% (-0.29, 0.19)0
0.09% (-0.16, 0.34)1
-0.12%
-0.02%
-0.43, 0.20) 2
-0.26,0.21)3
-0.14% (-0.37, 0.09)4
-0.05% (-0.28, 0.18)5
0.00% (-0.39, 0.38) 0-5
Sierra-Anderson High Volume Air
Sampler
0.02% (-0.29, 0.32) 0
0.13% (-0.27, 0.54)1
0.21% (-0.10, 0.52)2
0.53% (0.07, 0.99) 3
0.11% (-0.20, 0.41)4
0.38% (0.07, 0.70) 5
GAM: 2 LOESS terms, default
convergence
1.68% (0.45, 2.93)0
-0.36% (-1.56, 0.86)1
-0.21% (-1.40, 1.00)2
-0.18% (-1.40, 1.05)3
1.31% (0.08, 2.55)4
1.49% (0.25, 2.73)5
1.77% (-0.26, 3.83)0-5
Parametric: cubic splines
5df
1.45% (0.09, 2.83)0
-0.71% (-2.06, 0.67) 1
-0.59% (-1.95, 0.79)2
-0.70% (-2.09, 0.71)3
0.92% (-0.46, 2.32) 4
1.17% (-0.19, 2.55)5
1.17% (-1.54, 3.95)0-5
10 df
1.60% (0.20, 3.02)0
-0.80% (-2.18, 0.60)1
-0.73% (-2.11, 0.68) 2
-1.05% (-2.49, 0.40)3
0.64% (-0.79, 2.10
1.05% (-0.36, 2.48
0.51% (-2.60, 3.71
2df
1.79% (0.48, 3.11)
-0.09% (-1.38, 1.2:
0.10% (-1.18, 1.40
0.20% (-1.10, 1.52
1.60% (0.30, 2.91)
1.72% (0.43, 3.04)
1.90% (-0.36, 4.21
4
5
0-5
0
)1
2
3
4
5
0-5
The study focuses on the
temperature-mortality relationship and
only includes PM10 as a covariate in
models.
December 2009
E-306
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: O'Neill et al. (2008,
1923141
Period of Study: Jan 1998-Dec 2002
Location:
Mexico City, Mexico
Santiago, Chile
Sao Paulo, Brazil
Outcome:
Study Design: Time-series
Pollutant: PM,0
Averaging Time: 24 h
Covariates: Temperature, day of week, Mean (SD) ug/m3:
temporal trends, sex
Statistical Analysis: Poisson
regression
Statistical Package: S-Plus
Age Groups: Adults over 21 yr
Mexico City: 53.8 (24.9)
Sao Paulo: 48.9 (21.9)
Santiago: 78.7 (33.0)
Range (Min, Max):
Mexico City: 1.08-192.2
Sao Paulo: 12.0-171.3
Santiago: 8.0-218.6
Copollutant: NR
Increment: 10|jg/m
Percent increase (96% Cl) in all-
cause adult mortality (>22yrs) by
educational level and sex
Mexico City
All Adults, Concurrent Day
None: 0.76 (0.17-1.36)
Primary: 0.27 (-0.19-0.72)
Secondary: 0.19 (-0.19-0.57)
> 12 yr: 0.83 (0.03-1.63)
All Adults, Lag 1
None: 0.62 (0.02-1.22)
Primary: 0.62 (0.17-1.08)
Secondary: 0.29 (-0.09-0.90)
> 12 yr: 0.58 (-0.21-1.38)
All Adults, Distributed Lags 0-5
None: 0.91 (-0.07-1.89)
Primary: 0.48 (-0.27-1.24)
Secondary: 0.27 (-0.36-0.90)
> 12 yr: 0.75 (-0.49-2.02)
All Adults, df(yr)
None: 5.4
Primary: 6.0
Secondary: 6.0
>12yr:3.0
V\fomen, Concurrent Day
None: 0.65 (-0.08-1.38)
Primary: 0.48 (-0.13-1.09)
Secondary: 0.35 (-0.16-0.86)
> 12 yr: 1.64 (0.69-2.59)
Wfomen, Lag 1
None: 0.62 (-0.12-1.36)
Primary: 1.03 (0.42-1.64)
Secondary: 0.59 (0.08-1.11)
> 12 yr: 1.79 (0.84-2.75)
Wfomen, Distributed Lags 0-5
None: 0.46 (-0.74-1.68)
Primary: 1.39 (0.42-2.36)
Secondary: 0.51 (-0.30-1.33)
a 12 yr: 1.71 (0.61-2.83)
V\fomen, df (yr)
None: 5.4
Primary: 4.4
Secondary: 4.8
>12yr: 1.0
Men, Concurrent Day
None: 0.75 (-0.21-1.72)
Primary: 0.52 (-0.11-1.15)
Secondary: 0.56 (0.08-1.05)
> 12 yr: 1.20 (0.25-2.17)
Men, Lag 1
None: 0.45 (-0.51-1.42)
Primary: 0.70 (0.06-1.34)
Secondary: 0.47 (-0.02-0.95)
> 12 yr: 0.74 (-0.22-1.70)
Men, Distributed Lags 0-5
None: 1.24 (-0.25-2.75)
Primary: 0.65 (-0.39-1.69)
Secondary: 0.88 (0.11-1.66)
> 12 yr: 1.07 (-0.41-2.57)
Men, df (yr)
None: 3.8
Primary: 5.6
Secondary: 4.6
>12yr:3.8
Sao Paulo
All Adults, Concurrent Day
None: 0.77 (-0.28-1.82)
Primary: 1.27 (0.78-1.76)
Secondary: 0.93 (-0.07-1.94)
> 12 yr: 2.93 (2.00-2.88)
All Adults, Lag 1
None: 0.70 (-0.34-1.76)
Primary: 1.32 (0.83-1.82)
Secondary: 1.91 (0.58-2.60)
> 12 yr: 2.20 (1.27-3.15)
December 2009
E-307
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
All Adults, Distributed Lags 0-5
None: 0.76 (-0.91-2.46)
Primary: 1.34 (0.55-2.14)
Secondary: 1.91 (0.35-2.60)
> 12 yr: 2.20 (1.27-3.15)
All Adults, df(yr)
None: 4.0
Primary: 4.0
Secondary: 2.8
>12yr: 1.6
V\fomen, Concurrent Day
None: 1.93 (0.87-3.00)
Primary: 1.72 (1.04-2.41)
Secondary: 0.85 (-0.21-1.92)
> 12 yr: 1.84 (0.56-3.13)
Wfomen, Lag 1
None: 1.41 (0.34-2.48)
Primary: 1.64 (0.96-2.33)
Secondary: 1.43 (0.36-2.50)
> 12 yr: 2.27 (0.99-3.56)
Wfomen, Distributed Lags 0-5
None: 2.00 (0.40-3.63)
Primary: 2.05 (0.96-3.14)
Secondary: 1.61 (0.07-3.17)
> 12 yr: 3.35 (1.49-5.25)
V\fomen, df (yr)
None: 2.4
Primary: 3.6
Secondary: 1.4
>12yr:0.8
Men, Concurrent Day
None:-0.43 (-2.15-1.32)
Primary: 1.36 (0.71-2.02)
Secondary: 1.74 (0.77-2.72)
>12yr:2.81 (1.71-3.92)
Men, Lag 1
None:-0.44 (-2.17-1.33)
Primary: 1.44 (0.79-2.10)
Secondary: 1.52 (0.55-2.49)
> 12 yr: 1.48 (0.38-2.59)
Men, Distributed Lags 0-5
None: -0.30 (-3.09-2.56)
Primary: 1.67 (0.65-2.70)
Secondary: 1.06 (-0.34-2.49)
> 12 yr: 3.18 (1.60-4.79)
Men, df (yr)
None: 4.4
Primary: 3.2
Secondary: 0.8
a 12 yr: 1.2
Santiago
All Adults, Concurrent Day
None: 1.44 (0.53-2.36)
Primary: 0.06 (-0.21-0.34)
Secondary: 0.42 (0.06-0.78)
> 12 yr: 1.32 (0.60-2.05)
All Adults, Lag 1
None: 2.08 (1.16-30.1)
Primary: 0.53 (0.25-0.81)
Secondary: 0.55 (0.19-0.91)
a 12 yr: 1.31 (0.59-2.04)
All Adults, Distributed Lags 0-5
None: 3.18 (1.60-4.78)
Primary: 0.58 (0.10-1.06)
Secondary: 1.10 (0.48-1.73)
> 12 yr: 2.00 (0.93-3.07)
All Adults, df(yr)
None: 3.6
Primary: 5.6
Secondary: 4.0
>12yr: 1.6
Wfomen, Concurrent Day
None: 0.91 (-0.06-1.89)
Primary: 0.31 (-0.06-0.68)
Secondary: 0.84 (0.33-1.36)
> 12 yr: 0.60 (-0.32-1.52)
V\fomen, Lag 1
December 2009 E-308
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
None: 1.58(0.58-2.58)
Primary: 0.79 (0.42-1.17)
Secondary: 0.76 (0.25-1.28)
> 12 yr: 0.53 (-0.39-1.45)
V\fomen, Distributed Lags 0-5
None: 1.15 (-0.48-2.80)
Primary: 1.05 (0.41-1.69)
Secondary: 1.29 (0.40-2.19)
> 12 yr: 1.06 (-0.27-2.41)
Wfomen, df (yr)
None: 2.6
Primary: 4.8
Secondary: 4.4
a 12 yr: 1.0
Men, Concurrent Day
None: 0.05 (-1.02-1.12)
Primary:-0.11 (-0.5-0.28)
Secondary: 0.18 (-0.31-0.68)
> 12 yr: 1.52 (0.70-2.35)
Men, Lag 1
None: 0.61 (-0.44-1.68)
Primary: 0.23 (-0.16-0.62)
Secondary: 0.49 (0.00-0.98)
> 12 yr: 1.03 (0.21-1.86)
Men, Distributed Lags 0-5
None: 2.08 (0.28-3.91)
Primary: 0.16 (-0.50-0.82)
Secondary: 1.27 (0.43-2.12)
> 12 yr: 1.98 (0.76-3.20)
Men, df (yr)
None: 2.8
Primary: 4.8
Secondary: 4.4
>12yr: 1.6
Percent increase (96% Cl) in all-
cause adult mortality (>66yrs) by
educational level and sex
Mexico City
All Adults, Concurrent Day
None: 0.41 (-0.25-1.08)
Primary: 0.40 (-0.15-0.95)
Secondary: 0.50 (-0.01-1.01)
a 12 yr: 1.51 (0.39-2.63)
All Adults, Lag 1
None: 0.20 (-0.47-0.87)
Primary: 0.80 (0.24-1.36)
Secondary: 0.60 (0.09-1.12)
> 12 yr: 1.09 (-0.02-2.22)
All Adults, Distributed Lags 0-5
None: 0.27 (-0.83-1.38)
Primary: 0.99 (0.07-1.91)
Secondary: 0.30 (-0.56-1.16)
> 12 yr: 1.83 (0.09-3.59)
All Adults, df(yr)
None: 5.6
Primary: 5.4
Secondary: 6.0
>12yr:3.2
Wfomen, Concurrent Day
None: 0.49 (-0.30-1.29)
Primary: 0.39 (-0.33-1.11)
Secondary: 0.52 (-0.16-1.20)
> 12 yr: 1.29 (0.12-2.48)
V\fomen, Lag 1
None: 0.73 (-0.07-1.54)
Primary: 1.24 (0.52-1.97)
Secondary: 0.55 (-0.13-1.23)
> 12 yr: 1.50 (0.32-2.70)
V\fomen, Distributed Lags 0-5
None: 0.75 (-0.56-2.08)
Primary: 1.43 (0.29-2.59)
Secondary: 0.06 (-1.01-1.15)
> 12 yr: 1.48 (0.10-2.87)
Wfomen, df (yr)
None: 5.4
Primary: 4.2
Secondary: 4.8
December 2009 E-309
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
>12yr:0.6
Men, Concurrent Day
None: 0.90 (-0.23-2.04)
Primary: 0.37 (-0.40-1.16)
Secondary: 0.78 (0.07-1.49)
> 12 yr: 1.66 (0.30-3.04)
Men, Lag 1
None:-0.15 (-1.27-0.98)
Primary: 0.26 (-0.53-1.05)
Secondary: 0.93 (0.22-1.65)
> 12 yr: 0.95 (-0.41-2.32)
Men, Distributed Lags 0-5
None: 0.80 (-0.95-2.58)
Primary: 0.29 (-0.99-1.58)
Secondary: 1.06 (-0.08-2.21)
> 12 yr: 1.76 (-0.35-3.91)
Men, df (yr)
None: 3.8
Primary: 5.6
Secondary: 4.6
>12yr:3.8
Sao Paulo
All Adults, Concurrent Day
None: 0.60 (-0.48-1.70)
Primary: 0.59 (1.00-2.19)
Secondary: 1.21 (-0.01-2.44)
> 12 yr: 2.80 (1.67-3.94)
All Adults, Lag 1
None: 0.62 (-0.47-1.72)
Primary: 1.48 (0.89-2.07)
Secondary: 2.31 (1.08-3.55)
> 12 yr: 2.52 (1.40-3.66)
All Adults, Distributed Lags 0-5
None: 0.91 (-0.84-2.69)
Primary: 1.73 (0.79-2.67)
Secondary: 3.25 (1.39-5.16)
> 12 yr: 3.63 (2.01-5.29)
All Adults, df(yr)
None: 4.0
Primary: 3.8
Secondary: 2.6
>12yr: 1.6
V\fomen, Concurrent Day
None: 1.82 (0.71-2.94)
Primary: 1.84 (1.05-2.64)
Secondary: 0.62 (-0.55-1.81)
> 12 yr: 1.00 (-0.27-2.29)
Wfomen, Lag 1
None: 1.36 (0.25-2.49)
Primary: 1.76 (0.97-2.56)
Secondary: 1.57 (0.39-2.76)
> 12 yr: 1.39 (0.12-2.68)
Wfomen, Distributed Lags 0-5
None: 1.80 (0.12-3.51)
Primary: 1.97 (0.73-3.22)
Secondary: 1.89 (0.19-3.61)
> 12 yr: 2.53 (0.70-4.40)
V\fomen, df (yr)
None: 2.4
Primary: 3.4
Secondary: 1.2
>12yr:0.8
Men, Concurrent Day
None:-0.67 (-2.50-1.19)
Primary: 1.82 (1.00-2.65)
Secondary: 2.46 (1.31-3.63)
> 12 yr: 1.73 (0.47-3.00)
Men, Lag 1
None:-0.59 (-2.42-1.26)
Primary: 1.59 (0.78-2.41)
Secondary: 2.64 (1.49-3.80)
> 12 yr: 0.89 (-0.35-2.15)
Men, Distributed Lags 0-5
None: 1.50 (-1.52-4.60)
Primary: 2.46 (1.20-3.74)
Secondary: 2.24 (0.56-3.95)
> 12 yr: 1.45 (-0.34-3.29)
December 2009 E-310
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
Men, df (yr)
None: 4.6
Primary: 3.0
Secondary: 0.8
a 12 yr: 1.0
Santiago
All Adults, Concurrent Day
None: 1.49 (0.54-2.45)
Primary: 0.28 (-0.03-0.59)
Secondary: 0.58 (0.13-1.04)
> 12 yr: 2.32 (1.50-3.15)
All Adults, Lag 1
None: 2.20 (1.24-3.17)
Primary: 0.74 (0.43-1.05)
Secondary: 0.64 (0.20-1.11)
> 12 yr: 2.20 (1.36-3.04)
All Adults, Distributed Lags 0-5
None: 3.21 (1.54-4.90)
Primary: 0.92 (0.38-1.46)
Secondary: 1.46 (0.67-2.25)
> 12 yr: 4.02 (2.78-5.27)
All Adults, df(yr)
None: 3.8
Primary: 5.2
Secondary: 4.0
>12yr: 1.8
Wfomen, Concurrent Day
None: 1.39 (0.41-2.39)
Primary: 0.4 (0.01-0.8)
Secondary: 0.91 (0.29-1.53)
> 12 yr: 0.87 (-0.02-1.78)
Wfomen, Lag 1
None: 1.83 (0.83-2.85)
Primary: 0.98 (0.58-1.38)
Secondary: 0.73 (0.11-1.35)
> 12 yr: 0.76 (-0.15-1.68)
V\fomen, Distributed Lags 0-5
None: 2.47 (0.85-4.11)
Primary: 1.2 (0.52-1.88)
Secondary: 1.71 (0.65-2.78)
> 12 yr: 0.87 (-0.02-1.78)
Wfomen, df (yr)
None: 2.4
Primary: 4.8
Secondary: 4.4
>12yr:0.6
Men, Concurrent Day
None: 0.54 (-0.51-1.61)
Primary: 0.34 (-0.12-0.80)
Secondary: 0.25 (-0.40-0.91)
> 12 yr: 1.97 (1.09-2.86)
Men, Lag 1
None: 0.84 (-0.21-1.91)
Primary: 0.43 (-0.03-0.89)
Secondary: 0.61 (-0.04-1.26)
> 12 yr: 1.57 (0.67-2.46)
Men, Distributed Lags 0-5
None: 2.41 (0.64-4.22)
Primary: 0.80 (0.02-1.59)
Secondary: 1.58 (0.45-2.71)
> 12 yr: 2.99 (1.66-4.33)
Men, df (yr)
None: 2.0
Primary: 4.4
Secondary: 4.4
>12yr: 1.8
December 2009 E-311
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Peng et al. (2005, 0874631 Outcome: Mortality:
Period of Study: 1987-2000 Nonaccidental
Location: 100 U.S. cities (NMMAPS) Study Design: Time-series
Statistical Analyses: Bayesian
semiparametric hierarchical models
Age Groups: All ages
Pollutant: PM,0
Averaging Time: 24-h avg
Median (SD) unit: 27.1
Range (Mm, Max): (13.2, 48.7)
Copollutant (correlation): NR
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl) lag:
Winter:
-0.4% (-0.30, 0.21)0
0.15% (-0.08, 0.39 1
0.10% (-0.13, 0.33 2
Spring:
0.32% (0.08, 0.56) 0
0.14% (-0.14, 0.42)1
0.05% (-0.21, 0.32) 2
Summer:
0.13% (-0.11,0.37)0
0.36% (0.11, 0.61)1
-0.03% (-0.27, 0.21)2
Fall:
0.05% (-0.16, 0.25)0
0.14% (-0.06, 0.34 1
0.13% (-0.08, 0.35 2
All Seasons:
0.09% (-0.01, 0.19)0
0.19% (0.10, 0.28)1
0.08% (-0.03, 0.19)2
PM10 only (46 cities):
Winter: 0.15% (-0.16, 0.45)1
Spring: 0.13% (-0.21, 0.48)1
Summer: 0.30% (-0.10, 0.69)1
Fall: 0.07% (-0.23, 0.37) 1
PM10 +03 (46 cities):
Wnter: 0.18% (-0.16, 0.52)1
Spring: 0.10% (-0.30, 0.49)1
Summer: 0.33% (-0.14, 0.81)1
Fall: 0.08% (-0.25, 0.41)1
PM10 + 03 (45 cities):
Wnter: 0.13% (-0.24, 0.49)1
Spring: 0.1% 9(-0.18, 0.56)1
Summer: 0.28% (-0.13, 0.70)1
Fall:-0.01% (-0.34, 0.31)1
PM10 + N02 (46 cities):
Wnter: 0.21% (-0.18, 0.60)1
Spring: 0.19% (-0.17, 0.54)1
Summer: 0.34% (0.01, 0.68)1
Fall: 0.13% (-0.12, 0.39)1
Reference: Penttinen et al. (2004,
0874321
Period of Study: 1988-1996
Location: Helsinki, Finland
Reference: Qian et al. (2007, 0930541
Period of Study: 2001 -2004
Location: Wuhan, China
Outcome: Mortality:
Total (nonaccidental) (<800)
Cardiovascular (390-459)
Respiratory (460-51 9)
Study Design: Time-series
Statistical Analyses: Poisson GAM,
LOESS
Age Groups:
15-64yr
65-74 yr
>75yr
Outcome: Mortality:
Total (nonaccidental) (<800)
Cardiovascular (390-459)
Stroke (430-438)
Cardiac Diseases (390-398)
Respiratory (460-51 9)
Cardiopulmonary
Study Design: Time-series
Statistical Analyses: Poisson GAM,
natural splines
Age Groups: All ages
Pollutant: PM,0
Averaging Time: 24-h avg
Median (SD) unit: 21 |jg/m;
Range (Min, Max): (0.2, 213)
Copollutant (correlation):
rv r- nnQ
W3. I U.UiJ
N02:r = 0.50
CO: r = 0.45
en,' r- Dfil
OW2. I U.U I
TSP:r = 0.72
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): 141. 8 3
Range (Min, Max): (24.8, 477.8)
Copollutant (correlation):
N02
S02
03
Increment: 10|jg/m3
% Increase (Lower Cl, Upper Cl) lag:
Total (nonaccidental)
-0.23% (-1.47, 1.01)0
0.88% (-0.32, 2.08) 1
0.11 (-0.51, 0.73) 0-3 avg
Cardiovascular
-1.22% (-3.00, 0.56)0
0.63% (-1.09, 2.35)1
0.08% (-0.96, 0.81) 0-3 avg
Respiratory
3.94% (0.01, 7.87)0
3.96% (0.11, 7.81)1
2. 13% (0.03, 4. 22) 0-3 avg
Increment: 10|jg/m3
% Increase (Lower Cl, Upper Cl) lag:
Nonaccidental
0.36% (0.19, 0.53)0
0.28% (0.12, 0.45)1
0.43% (0.24, 0.62) 0-1
0.08% (-0.15, 0.31)0-4
<45yr
0.28% (-0.26, 0.82 0
0.45% (-0.06, 0.96 1
0.53% (-0.08, 1.13)0-1
0.41% (-0.31, 1.13)0-4
>45yr
0.36% (0.19, 0.54)0
0.27% (0.10, 0.44)1
0.42% (0.22, 0.62) 0-1
0.05% (-0.18, 0.29)0-4
<65yr
December 2009
E-312
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
~ 0.20% (-0.08, 0.49) 0
0.25% (-0.03, 0.52) 1
>45yr 0.33% (0.01, 0.66) 0-1
0.01% (-0.38, 0.39)0-4
<65 yr > 65 yr
0.41% (0.21, 0.61)0
Ł65y 0.30% (0.10,0.49)1
0.46% (0.24, 0.69) 0-1
0.10% (-0.16, 0.37)0-4
Cardiovascular
0.51% (0.28, 0.75)0
0.35% (0.12, 0.58)1
0.58% (0.31, 0.84) 0-1
0.35% (0.05, 0.66) 0-4
<45yr
0.59% (-0.62, 1.82 0
0.93% (-0.22, 2.08 1
1.07% (-0.27, 2.42)0-1
1.15% (-0.40, 2.72)0-4
>45yr
0.51% (0.27, 0.75)0
0.33% (0.10, 0.56)1
0.56% (0.30, 0.83) 0-1
0.33% (0.02, 0.63) 0-4
<65yr
0.27% (-0.23, 0.76) 0
0.30% (-0.16, 0.77)1
0.42% (-0.12, 0.97
0-1
0-4
0.43% (-0.19, 1.06
>65yr
0.57% (0.31, 0.83)0
0.36% (0.11, 0.61)1
0.61% (0.32, 0.90)0-1
0.33% (0.00, 0.67) 0-4
Stroke
0.44% (0.16, 0.72)0
0.41% (0.14, 0.68)1
0.58% (0.27, 0.89) 0-1
0.45% (0.09, 0.81)0-4
<45yr
1.18% (-0.45, 2.83)0
1.66% (0.11, 3.24)1
1.91% (0.10, 3.75)0-1
2.72% (0.58, 4.89) 0-4
>45yr
0.42% (0.14, 0.70)0
0.37% (0.10, 0.65)1
0.55% (0.23, 0.86) 0-1
0.39% (0.03, 0.76) 0-4
<65yr
0.26% (-0.35, 0.87 0
0.38% (-0.20, 0.96 1
0.48% (-0.19, 1.16)0-1
0.57% (-0.21, 1.35)0-4
>65yr
0.49% (0.17, 0.80)0
0.41% (0.11, 0.72)1
0.61% (0.26, 0.96)0-1
0.42% (0.02, 0.83) 0-4
Cardiac
0.49% (0.08, 0.89) 0
0.28% (-0.11, 0.67)1
0.49% (0.04, 0.94) 0-1
0.22% (-0.29, 0.74) 0-4
<45yr
0.25% (-1.64, 2.17 0
0.56% (-1.22, 2.38 1
0.61% (-1.47, 2.74)
0-1
-0.42% (-2.80, 2.02) 0-4
>45yr
0.49% (0.09, 0.91)0
0.27% (-0.12, 0.66)1
0.48% (0.03, 0.94) 0-1
0.25% (-0.27, 0.77) 0-4
December 2009 E-313
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
<65yr
0.00% (-0.89, 0.90 0
0.12% (-0.73, 0.98 1
0.13% (-0.86, 1.13)0-1
0.05% (-1.08, 1.20)0-4
>65yr
0.60% (0.17, 1.03)0
0.32% (-0.10, 0.74)1
0.57% (0.09, 1.06)0-1
0.26% (-0.29, 0.82) 0-4
Respiratory
0.71% (0.20, 1.23)0
0.63% (0.13, 1.13)1
0.86% (0.28, 1.44)0-1
0.19% (-0.48, 0.87)0-4
<45yr
1.74% (-1.28, 4.86)0
2.52% (-0.30, 5.42) 1
2.95% (-0.41, 6.42
0-1
0-4
3.47% (-0.61, 7.73
>45yr
0.69% (0.18, 1.21)0
0.58% (0.09, 1.08)1
0.81% (0.23, 1.39)0-1
0.13% (-0.54, 0.80)0-4
<65yr
0.06% (-1.30, 1.43)0
-0.53%
-0.32%
-1.83,0.79)1
-1.84,1.22)0-1
-0.72% (-2.47, 1.05)0-4
>65yr
0.79% (0.27, 1.31)0
0.76% (0.26, 1.26)1
0.99% (0.41, 1.57)0-1
0.30% (-0.38, 0.98) 0-4
Cardiopulmonary
0.46% (0.23, 0.69) 0
0.35% (0.13, 0.57)1
0.53% (0.28, 0.79) 0-1
0.11% (-0.19, 0.42)0-4
<45yr
0.71% (-0.48, 1.92)0
1.26% (0.14, 2.4)1
1.39% (0.06, 2.74)0-1
1.41% (-0.18, 3.03)0-4
>45yr
0.45% (0.23, 0.68) 0
0.32% (0.10, 0.54)1
0.51% (0.25, 0.77)0-1
0.08% (-0.23, 0.38) 0-4
<65yr
0.14% (-0.34, 0.61 0
0.15% (-0.30, 0.61 1
0.23% (-0.30, 0.76) 0-1
0.11% (-0.52, 0.74)0-4
>65yr
0.53% (0.28, 0.78) 0
0.39% (0.15, 0.63)1
0.60% (0.32, 0.88) 0-1
0.11% (-0.22, 0.45)0-4
Two-pollutant Models
Nonaccidental
PM10+N02: 0.14% (-0.07, 0.36) 0
PM10+S02: 0.37% (0.20, 0.55) 0
PM,o+03: 0.34% (0.17, 0.51)0
Cardiovascular
PM,o+N02: 0.34% (0.04, 0.63) 0
PM10+S02: 0.53% (0.28, 0.77) 0
PM10+03: 0.50% (0.26, 0.74) 0
Stroke
PM,o+N02: 0.28% (-0.07, 0.63) 0
PM,o+S02: 0.49% (0.21, 0.78)0
PM10+03: 0.44 (0.16, 0.72)0
December 2009 E-314
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Cardiac
PM10+N02: 0.24% (-0.27, 0.75) 0
PM10+S02: 0.43 (0.01, 0.84)0
PM,o+03: 0.44% (0.03, 0.85) 0
Respiratory
PM,o+N02: 0.46% (-0.19, 1.12)0
PM10+S02: 0.64% (0.11, 1.18)0
PM10+03: 0.67% (0.15, 1.20)0
Cardiopulmonary
PM,o+N02: 0.26% (-0.02, 0.55) 0
PM,o+S02: 0.46% (0.23, 0.70) 0
PM10+03: 0.44% (0.21, 0.67)0
Reference: Qian et al. (2008,1568941
Period of Study: Jul 2001-Jun 2004
Location: Wuhan, China
Outcome: Mortality:
Total (nonaccidental) (<800)
Cardiovascular (390-459)
Stroke (430-438)
Cardiac diseases (390-398, 410-429)
Respiratory (460-519)
Cardiopulmonary (390-459, 460-519)
Study Design: Time-series
Statistical Analyses: Poisson GLM,
natural splines and penalized splines
Age Groups: All ages
<65yr
>65yr
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD):
Normal temperature: 145.7 (64.6)
Low temperature: 117.3 (49.5)
High temperature: 96.3 (27.9)
Range (Min, Max): NR
Copollutant (correlation):
Normal temperature:
N02:r = 0/72
S02:r = 0.59
03:r = 0.06
Low temperature:
N02:r = 0.83
S02:r = 0.74
03:r = 0.19
High temperature:
N02:r = 0.68
S02:r = 0.15
03:r = 0.65
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl) lag:
Nonaccidental:
Normal:
All ages: 0.36 (0.17, 0.56)0-1
<65: 0.23 (-0.10, 0.56)0-1
> 65: 0.51 (0.18, 0.64)0-1
PM,o+N02: 0.07 (-0.17, 0.30)0-1
PM10+S02: 0.27 (0.06, 0.47) 0-1
PM10+03: 0.38 (0.18, 0.58)0-1
Low:
All ages: 0.62 (-0.09,1.34)0-1
<65:1.78 (0.52, 3.05)0-1
> 65: 0.22 (-0.61, 1.05)0-1
PM10+N02: 0.24 (-0.49, 0.97) 0-1
PM10+S02:0.45 (-0.27, 1.17)0-1
PM,o+03: 0.72 (0.00, 1.44)0-1
High:
All ages: 2.20 (0.74, 3.68) 0-1
<65: 2.34 (-0.09, 4.83) 0-1
> 65: 2.14 (0.42, 3.89)0-1
PM,o+N02:1.87 (0.42, 3.35)0-1
PM,o+S02: 2.12 (0.67, 3.60)0-1
PM10+03: 2.15 (0.55, 3.77)0-1
Cardiovascular:
Normal:
All ages: 0.39 (0.11, 0.66) 0-1
<65: 0.17 (-0.40, 0.73)0-1
> 65: 0.44 (0.14, 0.74)0-1
PMio+N02:0.11 (-0.23,0.45)0-1
PM,o+S02: 0.27 (-0.02, 0.55) 0-1
PM10+03: 0.42 (0.15, 0.70)
Low:
All ages: 0.72 (-0.25,1.70)0-1
<65: 2.63 (0.67, 4.63) 0-1
> 65: 0.24 (-0.84, 1.32)0-1
PM,o+N02: 0.37 (-0.62, 1.38)0-1
PM10+S02:0.50 (-0.47, 1.49)0-1
PM10+03: 0.82 (-0.16, 1.80)0-1
High:
All ages: 3.28 (1.24, 5.37)0-1
<65: 4.32 (0.10, 8.71)0-1
> 65: 3.03 (0.77, 5.34) 0-1
PM10+N02: 3.00 (0.95, 5.09) 0-1
PM10+S02: 3.20 (1.16, 5.29)0-1
PMio+03:3.71 (1.50,5.96)0-1
Stroke:
Normal:
All ages: 0.38 (0.06, 0.70)
<65: 0.17 (-0.53, 0.88)0-1
> 65: 0.43 (0.07, 0.79) 0-1
PM10+N02: 0.09 (-0.31, 0.49) 0-1
PM10+S02:0.31 (-0.03, 0.64) 0-1
PM,o+03: 0.38 (0.05, 0.71) 0-1
Low:
All ages: 0.67 (-0.50,1.85)0-1
<65: 2.85 (0.34, 5.42) 0-1
> 65: 0.11 (-1.22, 1.45)0-1
PM,o+N02: 0.29 (-0.90, 1.51)0-1
PMio+S02:0.53 (-0.65, 1.73)0-1
December 2009
E-315
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
PM10+03: 0.69 (-0.48, 1.87)0-1
High:
All ages: 2.35 (-0.03, 4.78) 0-1
<65: 4.54 (-0.79, 10.16)0-1
> 65:1.83 (-0.83, 4.57)
PM10+N02: 2.05 (-0.34, 4.49) 0-1
PM10+S02:2.31 (-0.07, 4.74) 0-1
PM,o+03: 2.77 (0.25, 5.35) 0-1
Cardiac:
Normal:
All ages: 0.32 (-0.14, 0.79)0-1
<65:-0.04 (-1.07, 1.01)0-1
> 65: 0.40 (-0.10,0.91)0-1
PM10+N02: 0.02 (-0.57, 0.60) 0-1
PM10+S02:0.11 (-0.38,0.61)0-1
PM,o+03: 0.41 (-0.06, 0.89) 0-1
Low:
All ages: 0.50 (-1.10, 2.13)0-1
<65:1.79 (-1.65, 5.35)0-1
> 65: 0.19 (-1.55, 1.95)0-1
PM,o+N02: 0.12 (-1.53, 1.80)0-1
PM,o+S02:0.14 (-1.48, 1.78)0-1
PM10+03: 0.72 (-0.90, 2.37) 0-1
High:
All ages: 3.31 (-0.22, 6.97) 0-1
<65: 2.71 (-4.58, 10.56)0-1
> 65: 3.45 (-0.41, 7.46) 0-1
PM,o+N02: 3.01 (-0.54, 6.69) 0-1
PM10+S02:3.17 (-0.37, 6.84)0-1
PM10+03: 4.92 (0.96, 9.03) 0-1
Respiratory:
Normal:
All ages: 0.80 (0.25,1.35)0-1
<65:-0.35 (-1.85, 1.18)0-1
> 65: 0.93 (0.38, 1.50)0-1
PM,o+N02: 0.30 (-0.39, 0.99) 0-1
PM10+S02: 0.64 (0.07, 1.22)0-1
PM10+03: 0.84 (0.28, 1.41)0-1
Low:
All ages: 1.07 (-0.76, 2.95)0-1
<65:-1.13 (-6.33, 4.35)0-1
> 65:1.30 (-0.57, 3.20)0-1
PM10+N02: 0.44 (-1.46, 2.36)0-1
PM10+S02:0.80 (-1.05, 2.69)0-1
PMio+03:1.11 (-0.73,2.99)0-1
High:
All ages: 1.15 (-3.54, 6.07)0-1
<65:-3.42 (-15.82, 10.80)0-1
> 65:1.76 (-3.03, 6.78)0-1
PM,o+N02: 0.63 (-4.07, 5.55) 0-1
PM,o+S02:1.03 (-3.66, 5.94)0-1
PM10+03: 2.66 (-2.44, 8.02) 0-1
Cardiopulmonary:
Normal:
All ages: 0.45 (0.19, 0.70)0-1
<65: 0.07 (-0.47, 0.61)0-1
> 65: 0.53 (0.25, 0.81)0-1
PM,o+N02: 0.15 (-0.17, 0.47)0-1
PM,o+S02: 0.34 (0.07, 0.61)0-1
PM10+03: 0.43 (0.17, 0.70)0-1
Low:
All ages: 0.69 (-0.22,1.61)0-1
<65:1.95 (0.04, 3.90)0-1
> 65: 0.43 (-0.57, 1.44)0-1
PM,o+N02: 0.33 (-0.61, 1.27)0-1
PM10+S02:0.50 (-0.42, 1.43)0-1
PM10+03: 0.76 (-0.16, 1.68)0-1
High:
All ages: 3.02 (1.03, 5.04)0-1
<65: 3.49 (-0.66, 7.81)0-1
> 65: 2.91 (0.74,5.12)0-1
PM10+N02: 2.70 (0.72, 4.73) 0-1
PM10+S02: 2.95 (0.96, 4.97) 0-1
PMio+03: 3.32 (1.16, 5.53)0-1
December 2009 E-316
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ren et al. (2006, 0928241
Period of Study: Jan 1996-Dec 2001
Location: Brisbane, Australia
Outcome: Mortality:
Nonaccidental
Cardiovascular (390-448)
Study Design: Time-series
Statistical Analyses: Poisson GAM,
cubic spline
Age Groups: All ages
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): 15.84
Range (Min, Max): (2.5, 60)
Copollutant: 0;
The study presents quantitative results
associated with an incremental increase
in temperature, notPM10.
Reference: Roberts (2004, 0879241
Period of Study: 1987-1994
Location: Cook County, Illinois
Allegheny County, Pennsylvania
Outcome: Mortality:
Nonaccidental (<800)
Study Design: Time-series
Statistical Analyses: Poisson GAM,
smooth splines
Poisson GLM, natural cubic splines
Age Groups: > 65 yr
Pollutant: PM10
Averaging Time: 24-h avg
Median (SD) unit:
Cook County
Lower Temp.: 29.24
Middle Temp.: 30.03
Upper Temp.: 52.76
Allegheny County
Lower Temp.: 16.50
Middle Temp.: 24.97
Upper Temp.: 55.42
Range (10th, 90th):
Cook County
Lower Tern.: (16.42, 46.42)
Middle Temp.: (14.79, 56.33)
Upper Temp.: (30.81, 82.81)
Allegheny County
Lower Temp.: (5.14, 34.54)
Middle Temp.: (8.91, 57.91)
Upper Temp.: (30.91, 88.99)
Increment: 10|jg/m
% Increase (SE) lag:
GLM
Cook
a = 0.5
No Interaction: 0.288% (0.157)0
Low Temp.:-0.272% (0.380)0
Middle Temp.: 0.344% (0.165)0
Upper Temp.: 0.281% (0.239)0
No Interaction: 0.359% (0.149)1
Low Tern p.:-0.168% (0.372)1
Middle Temp.: 0.361% (0.156)1
Upper Temp.: 0.616% (0.250)1
No Interaction: 0.465% (0.176) 0-1 ma
Low Temp.: 0.043% (0.397) 0-1 ma
Middle Temp.: 0.506% (0.184) 0-1 ma
Upper Temp.: 0.464% (0.256) 0-1 ma
No Interaction: 0.633% (0.214) 0-3 ma
Low Temp.: 0.365% (0.419) 0-3 ma
Middle Temp.: 0.638% (0.222) 0-3 ma
Upper Temp.: 0.718% (0.295) 0-3 ma
a=1
No Interaction: 0.117% (0.157)0
Low Tern p.:-0.351% (0.406)0
Middle Temp.: 0.161% (0.165)0
Upper Temp.: 0.096% (0.264)0
No Interaction: 0.141% (0.150)1
Low Tern p.:-0.366% (0.397)1
Middle Temp.: 0.161% (0.156)1
Upper Temp.: 0.301% (0.278)1
No Interaction: 0.260% (0.181) 0-1 ma
Low Temp.: -0.163% (0.431) 0-1 ma
Middle Temp.: 0.305% (0.188) 0-1 ma
Upper Temp.: 0.207% (0.291) 0-1 ma
No Interaction: 0.289% (0.225) 0-3 ma
Low Temp.: 0.014% (0.459) 0-3 ma
Middle Temp.: 0.311% (0.231) 0-3 ma
Upper Temp.: 0.301% (0.334) 0-3 ma
a = 2
No Interaction: 0.060% (0.158)
0
0
Low Tern p.:-0.464% (0.486)
0
0
Middle Temp.: 0.115% (0.168)
0
0
Upper Temp.:-0.022% (0.319)
0
0
No Interaction: 0.101% (0.152)1
Low Temp.:-0.432% (0.484)1
Middle Temp.: 0.089% (0.160)1
Upper Temp.: 0.455% (0.327)1
No Interaction: 0.129% (0.184) 0-1 ma
Low Temp.: -0.320% (0.546) 0-1 ma
Middle Temp.: 0.157% (0.193) 0-1 ma
Upper Temp.: 0.130% (0.346) 0-1 ma
No Interaction: 0.090% (0.236) 0-3 ma
Low Temp.: -0.319% (0.572) 0-3 ma
Middle Temp.: 0.105% (0.244)
0-3 ma
Upper Temp.: 0.193% (0.412)
0-3 ma
December 2009
E-317
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
Allegheny
a = 0.5
No Interaction: 0.078% (0.209) 0
Low Tern p.:-0.759% (0.643)0
Middle Temp.: 0.207% (0.216)0
High Temp.:-0.367% (0.364)0
No Interaction: 0.189% (0.206)1
Low Temp.:-0.335% (0.691)1
Middle Temp.: 0.293% (0.215)1
High Temp.:-0.171% (0.349)1
No Interaction: 0.224% (0.246) 0-1 ma
Low Temp.:-0.753% (0.763) 0-1 ma
Middle Temp.: 0.353% (0.253) 0-1 ma
High Temp.: -0.142% (0.382) 0-1 ma
No Interaction: 0.526% (0.300) 0-3 ma
Low Temp.: 0.050% (0.733) 0-3 ma
Middle Temp.: 0.688% (0.310) 0-3 ma
High Temp.: -0.043% (0.436) 0-3 ma
a=1
No Interaction: 0.078% (0.211)0
Low Tern p.:-0.694% (0.656)0
Middle Temp.: 0.214% (0.219)0
High Temp.:-0.533% (0.430)0
No Interaction: 0.179% (0.207)1
Low Temp.:-0.283% (0.718)1
Middle Temp.: 0.273% (0.217)1
High Temp.:-0.221% (0.396)1
No Interaction: 0.221% (0.249) 0-1 ma
Low Temp.: -0.731% (0.794) 0-1 ma
Middle Temp.: 0.348% (0.258) 0-1 ma
High Temp.: -0.253% (0.447) 0-1 ma
No Interaction: 0.464% (0.309) 0-3 ma
Low Temp.: 0.056% (0.780) 0-3 ma
Middle Temp.: 0.626% (0.319) 0-3 ma
High Temp.: -0.356% (0.516) 0-3 ma
a = 2
No Interaction: 0.034% (0.217)0
Low Temp.:-1.059% (0.715)0
Middle Temp.: 0.162% (0.230)0
High Temp.:-0.233% (0.489)0
No Interaction: 0.130% (0.214)1
Low Tern p.:-0.189% (0.800)1
Middle Temp.: 0.157% (0.226)1
High Temp.: 0.070% (0.471)1
No Interaction: 0.183% (0.260) 0-1 ma
Low Temp.: -0.918% (0.907) 0-1 ma
Middle Temp.: 0.279% (0.273) 0-1 ma
High Temp.: -0.001% (0.526) 0-1 ma
No Interaction: 0.270% (0.331) 0-3 ma
Low Tern p.: -0.105% (0.898) 0-3 ma
Middle Temp.: 0.394% (0.346) 0-3 ma
High Temp.: -0.287% (0.615) 0-3 ma
GAM
Cook
a = 0.5
No Interaction: 0.438% (0.151)0
Low Tern p.:-0.178% (0.364)0
Middle Temp.: 0.439% (0.163)0
Upper Temp.: 0.627% (0.197)0
No Interaction: 0.495% (0.144)1
Low Temp.:-0.114% (0.361)1
Middle Temp.: 0.460% (0.151)1
Upper Temp.: 0.938% (0.208)1
No Interaction: 0.710% (0.169) 0-1 ma
Low Temp.: 0.151% (0.379) 0-1 ma
Middle Temp.: 0.686% (0.180) 0-1 ma
Upper Temp.: 0.952% (0.214) 0-1 ma
No Interaction: 0.923% (0.203) 0-3 ma
Low Temp.: 0.532% (0.402) 0-3 ma
Middle Temp.: 0.855% (0.210) 0-3 ma
Upper Temp.: 1.289% (0.251) 0-3 ma
a=1
No Interaction: 0.190% (0.154)0
December 2009 E-318
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
Low Tern p.:-0.338% (0.414)0
Middle Temp.: 0.242% (0.162)0
Upper Temp.: 0.161% (0.230)0
No Interaction: 0.239% (0.146)1
Low Tern p.:-0.283% (0.406)1
Middle Temp.: 0.248% (0.152)1
Upper Temp.: 0.453% (0.244)1
No Interaction: 0.353% (0.174) 0-1 ma
Low Temp.: -0.074% (0.437) 0-1 ma
Middle Temp.: 0.388% (0.182) 0-1 ma
Upper Temp.: 0.345% (0.251) 0-1 ma
No Interaction: 0.453% (0.213) 0-3 ma
Low Temp.: 0.190% (0.460) 0-3 ma
Middle Temp.: 0.455% (0.219) 0-3 ma
Upper Temp.: 0.557% (0.294) 0-3 ma
a = 2
No Interaction: 0.071% (0.157)
0
0
Low Tern p.:-0.534% (0.478)
0
0
Middle Temp.: 0.132% (0.165)
0
0
Upper Temp.: 0.011% (0.264)
0
0
No Interaction: 0.099% (0.150)1
Low Temp.:-0.467% (0.472)1
Middle Temp.: 0.109% (0.156)1
Upper Temp.: 0.329% (0.278)1
No Interaction: 0.168% (0.180) 0-1 ma
Low Temp.: -0.371% (0.525) 0-1 ma
Middle Temp.: 0.216% (0.188) 0-1 ma
Upper Temp.: 0.116% (0.290) 0-1 ma
No Interaction: 0.149% (0.227) 0-3 ma
Low Temp.: -0.291% (0.557) 0-3 ma
Middle Temp.: 0.174% (0.233) 0-3 ma
Upper Temp.: 0.210% (0.340) 0-3 ma
Allegheny
a = 0.5
No Interaction: 0.245% (0.203) 0
Low Temp.:-0.727% (0.648)0
Middle Temp.: 0.314% (0.216)0
High Temp.: 0.308% (0.287)0
No Interaction: 0.446% (0.199)1
Low Temp.:-0.307% (0.701)1
Middle Temp.: 0.469% (0.211)1
High Temp.: 0.556% (0.285)1
No Interaction: 0.522% (0.237) 0-1 ma
Low Temp.: -0.646% (0.761) 0-1 ma
Middle Temp.: 0.567% (0.251) 0-1 ma
High Temp.: 0.640% (0.307) 0-1 ma
No Interaction: 0.977% (0.282) 0-3 ma
Low Temp.: 0.307% (0.733) 0-3 ma
Middle Temp.: 1.027% (0.296) 0-3 ma
High Temp.: 1.001% (0.352) 0-3 ma
a=1
No Interaction: 0.107% (0.209)0
Low Temp.:-0.819% (0.699)0
Middle Temp.: 0.229% (0.219)0
High Temp.:-0.214% (0.350)0
No Interaction: 0.223% (0.205) 1
Low Temp.:-0.316% (0.751)1
Middle Temp.: 0.295% (0.216)1
High Temp.: 0.002% (0.341)1
No Interaction: 0.267% (0.246) 0-1 ma
Low Temp.: -0.797% (0.840) 0-1 ma
Middle Temp.: 0.372% (0.257) 0-1 ma
High Temp.: 0.035% (0.372) 0-1 ma
No Interaction: 0.534% (0.302) 0-3 ma
Low Temp.: 0.029% (0.810) 0-3 ma
Middle Temp.: 0.660% (0.314) 0-3 ma
High Temp.: 0.071% (0.431) 0-3 ma
December 2009 E-319
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
a = 2
No Interaction: 0.061% (0.214)0
Low Temp.:-1.048% (0.749)0
Middle Temp.: 0.206% (0.226)0
High Temp.:-0.332% (0.419)0
No Interaction: 0.145% (0.211)1
Low Tern p.:-0.278% (0.816)1
Middle Temp.: 0.210% (0.223)1
High Temp.:-0.105% (0.394)1
No Interaction: 0.180% (0.256) 0-1 ma
Low Temp.: -1.028% (0.931) 0-1 ma
Middle Temp.: 0.298% (0.269) 0-1 ma
High Temp.:-0.114% (0.441) 0-1 ma
No Interaction: 0.275% (0.324) 0-3 ma
Low Tern p.: -0.384% (0.915) 0-3 ma
Middle Temp.: 0.436% (0.338) 0-3 ma
High Temp.: -0.366% (0.513) 0-3 ma
Reference: Roberts (2004, 0879241
Period of Study: 1987-1994
Location: Cook County, Illinois
Allegheny County, Pennsylvania
Outcome: Mortality:
Nonaccidental
Study Design: Time-series
Statistical Analyses: Poisson GLM
Age Groups: 2 65 yr
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max):
Max = 89
The study does not present quantitative
results.
Reference: Roberts (Roberts, 2005,
0879921
Period of Study:
Cook County: 1987-2000.
Allegheny County: 1987-1998
Location: Cook County, Illinois
Allegheny County, Pennsylvania
Outcome: Mortality:
Nonaccidental
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups: 2 65 yr
Pollutant: PM10
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: NR
p(SE)lag:
Standard Model
Cook County
0.000127(0.000264) 0
-0.000042 (0.000249) 1
-0.000441 (0.000246) 2
Allegheny County
0.000693
0.000356
0.000437) 0
0.000423) 1
0.000524(0.000415)2
Moving Total Model
Cook County
0.000150 (0.000187) k= 2
-0.000047 (0.000153) k = 3
0.000009 (0.000133) k= 4
Allegheny County
0.000633 (0.000310) k= 2
0.000542
0.000598
0.000255) k= 3
0.000351) k= 4
Reference: Roberts (2006, 0897621 Outcome: Mortality:
Period of Study: 1987-2000
Location: Cook County, Illinois
Suffolk County, Massachusetts
(NMMAPS)
Nonaccidental
Study Design: Time-series
Statistical Analyses: Poisson GLM
Age Groups: > 65 yr
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD):
Cook County: 33.7 (19.4)
Suffolk County: 25.9 (11.8)
Range (10th, 90th):
Cook County: (13.4, 58.1)
Suffolk County: (14.0, 41.7)
Copollutant (correlation):
Cook County
CO: r = 0.30
N02:r = 0.53
S02:r = 0.45
03:r = 0.44
Suffolk County
CO: r = 0.33
N02:r = 0.43
S02:r = 0.23
03:r = 0.36
Increment:
Cook County: 19.4 pg/m3
Suffolk County: 14.0|jg/m3
% Increase (SD) lag:
Cook County
Standard Model: 0.49% (0.25) 0
Proposed Model: 0.29% (0.16)0
Standard Model: 0.67% (0.25) 0-2 avg
Proposed Model: 0.49% (0.25) 0-2 avg
Suffolk County
Standard Model: 0.88% (1.27)0
Proposed Model: 0.85% (0.84) 0
Standard Model: 1.60% (0.71) 0-2 avg
Proposed Model: 1.35% (0.73) 0-2 avg
December 2009
E-320
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Roberts and Martin (2006,
0977991
Period of Study: 1987-2000
Location: Cook County, Illinois
(NMMAPS)
Outcome: Mortality: Nonaccidental
Study Design: Time-series
Statistical Analyses: Dose-response
1. Piecewise linear relationship (no-
threshold) with change point at
25 pg/m and 50 pg/m
2. Piecewise linear relationship
(threshold), exposure below 25 pg/m3
no effect and exposures above
50 pg/m having a different effect then
exposures between 25 pg/m3 and
50 pg/m3
Age Groups: > 65 yr
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): NR
IQR (26th, 76th):
(23.9, 45.4)
Suffolk County: (14.0, 41.7)
Copollutant (correlation): NR
The study does not present quantitative
results.
Reference: Roberts and Martin (2006,
0886701
Period of Study: 1987-2000
Location: 109 U.S. cities (NMMAPS)
Outcome: Mortality: Nonaccidental
Cardiorespiratory
Study Design: Time-series
Statistical Analyses: Poisson
2-stage Bayesian hierarchical model
Age Groups: All ages
Pollutant: PM10
Averaging Time: 24-h avg
Mean (SD): NR
IQR (26th, 76th): NR
Copollutant (correlation): NR
Increment: NR
P x 1000 (SEx 1000) lag:
Nonaccidental
Model 1
Base df: 0.079 (0.050) 0
Double df: 0.044 (0.046)0
Half df: 0.107 (0.052)0
Base df: 0.180 (0.044)1
Double df: 0.149 (0.047)1
Half df: 0.254 (0.048)1
Base df: 0.059 (0.056) 2
Double df: 0.024 (0.056) 2
Half df: 0.143 (0.054) 2
Model 2
Base df: 0.115 (0.037) 0-2 ma
Double df: 0.107 (0.034) 0-2 ma
Half df: 0.145 (0.039) 0-2 ma
Cardio-respiratory
Model 1
Base df: 0.103 (0.068)0
Double df: 0.056 (0.067)0
Half df: 0.134 (0.066)0
Base df: 0.232 (0.060) 1
Double df: 0.179 (0.067)1
Half df: 0.309 (0.059)1
Base df: 0.210 (0.078) 2
Double df: 0.144 (0.075) 2
Half df: 0.305 (0.079) 2
Model 2
Base df: 0.168 (0.047) 0-2 ma
Double df: 0.140 (0.044) 0-2 ma
Half df: 0.196 (0.051) 0-2 ma
Notes: Model 1 uses current day's
mortality count, while Model 2 uses a 3-
day moving total mortality count.
December 2009
E-321
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Roberts and Martin (2007, Outcome: Mortality: Total
(nonaccidental)
156917'
Period of Study: 1987-2000
Location: 8 U.S. cities and >100 U.S.
cities (NM MAPS)
Cardiorespiratory
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups: All ages
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Increment: 10|jg/m
P x 1000 (SEx 1000) lag:
8 U.S. cities
Distributed Lag Model: 0.229
0-2
Weighted Model: 0.315
0-2
Standard Model:
0.276
0
-0.062
1
0.476
2
90 U.S. cities
Total (nonaccidental)
Standard Model:
0.078 (0.039)
0
0.182(0.037)
1
0.108(0.036)
2
Moving Total Model: 0.131 (0.023)
0-2
Weighted Model: 0.274 (0.075)
0-2
Cardio-respiratory
Standard Model:
0.096 (0.055)
0
0.232 (0.053)
1
0.226(0.051)
2
Moving Total Model:
0.174(0.032)
0-2
Weighted Model:
0.389(0.105)
0-2
Notes: The 8 U.S. cities consist of
Chicago, Cleveland, Denver, El Paso,
Houston, Nashville, Pittsburgh, and Salt
Lake City.
December 2009
E-322
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Roberts and Martin (2007,
1569161
Period of Study: 1987-2000
Location: 10 U.S. cities (NMMAPS)
Outcome: Mortality: Nonaccidental
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups: > 65 yr
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD):
Anchorage: 27.32
Chicago: 36.95
Cleveland: 39.83
Detroit: 40.78
El Paso: 40.14
Minneapolis/a. Paul: 28.01
Pittsburgh: 35.09
Salt Lake City: 37.40
Seattle: 28.72
Spokane: 34.52
Range (Min, Max): NR
Increment: NR
P Coefficient (SE) lag:
Pooled Estimates
Combined Model (Unconstrained
Distributed Lag Model + Piecewise
Linear Dose-Response Function)
Change-point: 60 pg/m3
Slope below: 0.00130 (0.00016)
0-5
Slope above:-0.00163 (0.00026)
0-5
Change-point: 30 pg/m
Slope below: 0.00014 (0.00039)
0-5
Slope above:-0.00003 (0.00015)
0-5
Piecewise Linear Dose-Response
Model
Change-point: 60 pg/m3
Slope below: 0.00044 (0.00011)
3-day ma
Slope above:-0.00077 (0.00020)
3- day ma
Change-point: 30 pg/m3
Slope below: 0.00022 (0.00026)
3-day ma
Slope above:-0.00004 (0.00011)
3-day ma
Polynomial Distributed Lag Model
(degree 2)
0.00046(0.00011)
0-5
Reference: Samoli et al. (2005,
0874361
Period of Study: 1990-1997
Location: 22 European cities (APHEA-
2)
Outcome: Mortality:
All-cause (nonaccidental) (<800)
Cardiovascular (390-459)
Respiratory (460-519)
Study Design: Time-series
Statistical Analyses: Hierarchical
modeling:
1. Poisson GAM, penalized splines
2. Multivariate modeling
Pollutant: PM,0
Averaging Time: 24-h avg
Median (SD) unit:
Range: (Stockholm: 14 pg/m3 to Torino:
65 |jg/ms)
Percentile (90th):
Range: (Stockholm: 27 pg/m3 to Torino:
129 pg/m3)
The study does not present quantitative
results.
«ge oroup*. «,, dya, Copollutant (correlation): BS
Reference: Schwartz (2004, 0789981
Period of Study: 1986-1993
Location: 14 U.S. cities
Outcome: Mortality:
Nonaccidental (<800)
Study Design: Case-crossover
Time-series
Statistical Analyses: Conditional
logistic regression
Poisson
Age Groups: All ages
Notes: Case days matched to referent
days that had the same temperature.
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: 10 pg/m3
% Increase (Lower Cl, Upper Cl) lag:
Overall:
Two stage: 0.36% (0.22, 0.50)1
Single stage: 0.33% (0.19, 0.46)1
More winter temperature lags:
Two Stage: 0.39% (0.23, 0.56)1
One stage: 0.32% (0.19, 0.46)1
Time stratified with temperature
matching:
Two Stage: 0.39% (0.19, 0.58)1
One Stage: 0.53% (0.34, 0.72) 1
Poisson regression:
0.40% (0.18, 0.62)1
December 2009
E-323
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Schwartz (2004, 0535061
Period of Study: 1986-1993
Location: 14 U.S. cities
Outcome: Mortality:
Nonaccidental (<800)
Study Design: Case-crossover
Statistical Analyses: Time-stratified
conditional logistic regression
Age Groups: All ages
Notes: Case days matched to referent
days based on concentration of
gaseous air pollutants. Matched on the
following conditions:
1. 24-h avg S02 within 1 ppb
2. Daily-maximum 0? within 2 ppb
3. 24-h avg N02 within 1 ppb
4. 24-h avg CO within 0.03 ppm
Pollutant: PM,0
Averaging Time: 24-h avg
Median (SD) unit: Range: 23-36 pg/m3
IQR (25th, 75th):
Range 25th: 17-24 pg/m3
Range 75th: 31-57 pg/m3
Copollutant (correlation): CO
S02
N02
03
Increment: 10|jg/m
P x 1000 (SEx 1000) lag:
Matched on CO: 0.527 (0.251)
0-1 avg
Matched on 03: 0.451 (0.170)
0-1 avg
Matched on N02: 0.784 (0.185)
0-1 avg
Matched on S02: 0.811 (0.175)
0-1 avg
Reference: Sharovsky et al. (2004,
1569761
Period of Study: Jul 1996-Jun 1998
Location: Sao Paulo, Brazil
Outcome: Mortality:
Myocardial infarction
Study Design: Time-series
Statistical Analyses: Poisson GAM
Age Groups: 2 35 yr
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): 58.2 (25.8)
Range (Min, Max): (23,186)
Copollutant (correlation):
CO: r = 0.73
S02:r = 0.72
Increment: 10|jg/m
p(SE)lag:
PM10: 0.001 (0.001)
PMio+CO+S02: 0.0004 (0.0008)
Reference: Simpson et al. (2005,
0874381
Period of Study: 1/1996-12/1999
Location: 4 Australian cities
Reference: Slaughter et al. (2005,
0738541
Period of Study: Jan 1995-Dec 1999
Location: Spokane, Washington
Reference: Staniswalis et al. (2005,
0874731
Period of Study: 1992-1995
Location: El Paso, Texas
Reference: Stafoggia et al. (2008,
1570051
Period of Study: 1997-2004
Outcome: Mortality:
Nonaccidental (<800)
Cardiovascular (390-459)
Respiratory (460-51 9)
Study Design: Time-series
meta-analysis
Statistical Analyses: Poisson GAM,
natural splines
Poisson GLM, natural splines
Age Groups: All ages
Outcome: Mortality:
Nonaccidental (<800)
Study Design: Time-series
Statistical Analyses: Poisson GLM,
natural splines
Age Groups: All ages
Outcome: Mortality:
Nonaccidental (<800)
Study Design: Time-series
Statistical Analyses: Poisson
Principal component analysis (PCA)
Age Groups: All ages
Outcome:
Mortality:
Total (nonaccidental) (<800)
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD):
Brisbane: 16.60
Sydney: 16.30
Melbourne: 18.20
Range (Min, Max):
Brisbane: (2.6, 57.6)
Sydney: (3.7, 75.5)
Melbourne: (3.3, 51.9)
Copollutant:
PM25
CO
N02
Pollutant: PM10
Averaging Time: 24-h avg
Mean (SD): NR
Range (9th, 95th): (7.9, 41.9) pg/m3
Copollutant (correlation):
DM
PMio
PM10.25:r = 0.94
CO: r = 0.32
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max):
(0.2, 133.4)
Notes: The chemical composition and
size distribution of PM was not
available, therefore, the study used
wind speed as a surrogate variable for
the PMio composition.
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD) unit:
Bologna: 50.4 (31. 7)
Increment: 10|jg/m3
% Increase (Lower Cl, Upper Cl)
lag:
0.2% (-0.8, 1.2)
Increment: : 25 pg/m3
Relative Risk (Lower Cl, Upper Cl)
lag:
1.00(0.97,1.03) 1
0.98(0.95, 1.01) 2
1.00(0.97, 1.03) 3
Increment: 10|jg/m3
% Increase (Lower Cl, Upper Cl)
lag:
Poisson regression: 1.7% 3
PCA:
24-hly measurements: 2.06% 3
Daily avg: 1.7% 3
Increment: 10|jg/m3
% Increase (Lower Cl, Upper Cl) lag:
Cardiovascular
All yr: 0.63% (0.31, 1.38) 0-1
December 2009
E-324
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Location: 9 Italian cities
Cardiovascular (390-459)
Respiratory (460-519)
Other natural causes
Study Design: Time-stratified case-
crossover
Statistical Analyses:
Conditional logistic regression
Age Groups: > 35 yr
Florence: 37.5 (16.6)
Mestre:48.1 (26.8)
Milan: 57.9 (38.0)
Palermo: 36.2 (21.7)
Pisa: 35.1 (14.9)
Rome: 47.3 (19.9)
Taranto: 59.8 (18.9)
Turin: 71.5 (38.1)
Range (Min, Max): NR
Copollutant (correlation): NR
Winter: 0.15% (-0.29, 0.59)0-1
Spring: 0.72% (-0.07,1.52)0-1
Summer: 2.90% (1.14, 4.69)0-1
Fall: 1.37% (0.43, 2.32)0-1
Apparent Temperature
<50th Percentile:
0.31% (-0.06, 0.67)0-1
50th-75th Percentile:
2.05% (0.47, 3.66) 0-1
>75th Percentile: 2.68% (1.20, 4.17) 0-1
Respiratory
All yr: 0.98% (0.27, 1.70)0-1
Winter: 0.41% (-0.67,1.51)0-1
Spring: 2.99% (1.18, 4.83)0-1
Summer: 3.89% (0.19, 7.73)0-1
Fall: 0.45% (-1.11, 2.03) 0-1
Apparent Temperature
<50th Percentile:
0.54% (-0.47, 1.57)0-1
50th-75th Percentile:
3.15% (0.64, 5.73)0-1
>75th Percentile:
4.12% (0.44, 7.93)0-1
Other natural causes
All yr: 0.37% (0.09, 0.66) 0-1
Wnter: 0.14% (-0.36, 0.63)0-1
Spring: 0.29% (-0.47,1.05)0-1
Summer: 2.15% (0.90, 3.42)0-1
Fall: 0.70% (-0.41, 1.83) 0-1
Apparent Temperature
<50th Percentile:
0.07% (-0.27, 0.41)0-1
50th-75th Percentile:
1.08% (-0.02, 2.19)0-1
>75th Percentile:
2.30% (1.06, 3.56)0-1
Total (nonaccidental)
All yr: 0.53% (0.25, 0.80) 0-1
Wnter: 0.20% (-0.08, 0.49) 0-1
Spring: 0.62% (0.14,1.10)0-1
Summer: 2.54% (1.31, 3.78) 0-1
Fall: 1.21% (0.37, 2.06)0-1
Apparent Temperature
<50th Percentile:
0.21% (-0.06, 0.47)0-1
50th-75th Percentile:
1.60% (0.64, 2.57)0-1
>75th Percentile:
2.55% (1.58, 3.52)0-1
P coefficient (SE) lag:
Linear interaction PM10 and Apparent
Temperature
Cardiovascular
<50th Percentile:
-0.000117(0.000415)0-1
50th-75th Percentile:
0.003445(0.001407)0-1
>75th Percentile:
0.002764(0.001795)0-1
Respiratory
<50th Percentile:
0.001119(0.000943)0-1
50th-75th Percentile:
-0.001120(0.003480)0-1
>75th Percentile:
0.005306 (0.004350) 0-1
Other natural causes
<50th Percentile:
0.000411(0.000383)0-1
50th-75th Percentile:
-0.001526(0.001207)0-1
>75th Percentile:
December 2009
E-325
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
0.002564(0.001958)0-1
Total (nonaccidental)
<50th Percentile:
0.000246 (0.000269) 0-1
50th-75th Percentile:
0.000584 (0.000880) 0-1
>75th Percentile:
0.002396(0.001629)0-1
Reference: Stolzel et al. (2007,
0913741
Period of Study: Sep 1995-Aug 2001
Location: Erfurt, Germany
Outcome:
Mortality:
Total (nonaccidental) (<800)
Cardio-respiratory (390-459, 460-519,
785, 786)
Study Design: Time-series
Statistical Analyses:
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD) unit:: 31.9 (23.2)
IQR (25th, 75th):
(16.5, 39.5)
Copollutant (correlation):
Increment: 23 pg/m
Relative Risk (Lower Cl, Upper Cl)
lag:
Total (nonaccidental)
1.004(0.980
1.029)
0
1.004(0.981
1.027)
1
Reference: Sullivan et al. (2003,
0431561
Period of Study:
1985-1994
Location: Western Washington
Poisson GAM
Age Groups: All ages
Outcome:
Out-of-hospital cardiac arrest
Study Design: Case-crossover
Statistical Analyses:
Conditional logistic regression
Age Groups: 19-79
Study Population: Out-of-hospital
cardiac arrests: 1,206
IVIUJ. I-U.3. r = U.03
MC0.01-2.5:r = 0.84
NO: r = 0.54
N02:r = 0.62
CO: r = 0.50
Pollutant: PM,0
Averaging Time: 24-h avg
Median (SD) unit:
Lag 0: 28.05
Lag 1:27.97
Lag 2: 28.40
Range (Min, Max): (7.38, 89.83)
Copollutant (correlation):
S02
CO
Notes: Study used nephelometry to
measure particles and equated the
measurements to PM2 5 concentrations.
0.998 (0.976
1.021)
2
0.984 (0.962
1 006)
3
0.993 (0.972
1.015)
4
0.990 (0.969
1.012)
5
Cardio-respiratory
1.007 (0.981
1.034)
0
1.006(0.981
1.032)
1
0.996 (0.971
1.021)
2
0.977 (0.953
1.002)
3
0.994 (0.970
1.018)
4
0.993 (0.969
1.017)
5
Increment: : 16.51 pg/m3
Odds Ratio (Lower Cl, Upper Cl) lag:
Overall
1.05(0.87, 1.27)
0
0.91 (0.75, 1.11)
1
1.03(0.82,1.28)
9
z
December 2009
E-326
-------
Study
Reference: Sunyer et al. (2002,
0348351
Period of Study: 1985-1995
Location: Barcelona, Spain
Reference: Touloumi et al. (2005,
0874771
Period of Study: 1990-1997
Location: 7 European cities (London,
Budapest, Stockholm, Zurich, Paris,
Lyon, Madrid) (APHEA2)
Design & Methods
Outcome: Mortality:
Respiratory mortality
Study Design: Case-crossover
Statistical Analyses: Condition logistic
regression
Age Groups: >14
Study population: Asthmatic individuals:
5,610
Outcome: Mortality:
Total (nonaccidental) (<800)
Cardiovascular (390-459)
Study Design: Time-series
Statistical Analyses: Poisson GAM,
LOESS
Age Groups: All ages
Concentrations1
Pollutant: PM,0
Averaging Time: 24-h avg
Median (SD) unit: 61. 2
Range (Mm, Max): (17.3, 240.7)
Copollutant:
BS
N02
03
S02
CO
Pollutant: PM,0
Averaging Time: 24-h avg
Median (SD) unit:
London: 25.1
Budapest: 40.2
Stockholm: 13.7
Zurich: 27.5
Paris: 22.2
Lyon: 38.5 p
Madrid: 33.4
IQR (25th, 75th):
London: (20.3, 33.9)
Budapest: (34.3, 45.8)
Stockholm: (10.3, 19.1)
Zurich: (19.2, 38.5)
Paris: (16.0, 33.0)
Lyon: (29.7, 50.4)
Madrid: (27.6, 41.0)
Copollutant (correlation): NR
Effect Estimates (95% Cl)
Increment: 32.7 pg/m3
Odds Ratio (Lower Cl, Upper Cl) lag:
Asthmatic individuals with 1 ED visit
0.884(0.672, 1.162) 0-2 avg
Asthmatic individuals with >1 ED visit
1.084(0.661, 1.778) 0-2 avg
Asthma/COPD individuals with >1 ED
visit
1.011 (0.746, 1.368) 0-2 avg
Increment: 10|jg/m3
P(x 1000) (SE(x 1000)):
Total (nonaccidental)
No control: 0.4834 (0.1095)
Reported Influenza Data
Count ID: 0.4967 (0.1089)
11 10:0.4740(0.1090)
Ml ID: 0.5019 (0.1096)
RI-ID: 0.4735 (0.1091)
SF ID: 0.6714 (0.1080)
Estimated Influenza Data
APHEA-2: 0.5550 (0.1076)
11 El 0:0.5640 (0.1 073)
Ml EID: 0.5872 (0.1100)
RIEID: 0.5872 (0.1 074)
SF EID: 0.6641 (0.1073)
Cardiovascular
No control: 0.8432 (0.1665)
PpnnrtpH Inflnpnya Plata
Count ID: 0.8896 (0.1662)
11 10:0.8545(0.1661)
Ml ID: 0.8693 (0.1674)
RI-ID: 0.8649 (0.1665)
SF ID: 1.0107 (0.1659)
Estimated Influenza Data
APHEA-2: 0.9389 (0.1654)
11 El 0:0.9485 (0.1648)
Ml EID: 1.0440 (0.1686)
RIEID: 0.9718 (0.1653)
SF EID: 1.0585 (0.1652)
Notes: 11 = one indicator for all
epidemics
M1 = multiple indicators, one per
epidemic
R1 = indicators for intervals indicating
the range of influenza counts
SF = separate smooth function during
epidemic periods.
Reference: Tsai et al. (2003, 0504801
Period of Study: 1994-2000
Location: Kaohsiung, Taiwan
Outcome: Mortality:
Total (nonaccidental) (<800)
Respiratory (460-519)
Circulatory (390-459)
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): 81.45
Range (Min, Max): (20.50, 232.00)
Increment: 67.00 pg/m
Odds Ratio (Lower Cl, Upper Cl) lag:
Total (nonaccidental)
1.000(0.947, 1.056) 0-2 avg
Study Design: Bidirectional case-
crossover
Statistical Analyses: Conditional
logistic regression
Age Groups: All ages
Copollutant:
S02
N02
CO
03
Respiratory
1.023(0.829, 1.264) 0-2 avg
Circulatory
0.971 (0.864, 1.092) 0-2 avg
December 2009
E-327
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Vajanapoom et al. (2002,
0425421
Period of Study: 1992-1997
Location: Bangkok, Thailand
Outcome: Mortality:
Total (nonaccidental) (<800)
Respiratory (460-519)
Cardiovascular (390-459)
Other-causes
Study Design: Time-series
Statistical Analyses: Poisson GAM,
LOESS
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): 68.0 (23.9)
IQR (25th, 75th):
(50.1,80.7)
Copollutant (correlation): NR
Increment: 30 pg/m
% Increase (Lower Cl, Upper Cl) lag:
Total (nonaccidental)
All ages: 2.3% (1.3, 3.3) 0-4 ma
55-64:1.5% (-0.8, 3.9) 0-4 ma
65-74: 4.2% (2.0, 6.3) 0-4 ma
> 75: 3.9% (2.1, 5.6) 0-4 ma
Cardiovascular
All ages: 0.8% (-0.9, 2.4) 0
55-64:-2.5% (-6.3, 1.3)0
65-74: 2.9% (-0.7, 6.5) 0
Reference: Vedal et al. (2003, 0390441
Period of Study: Jan 1994-Dec 1996
Location: Vancouver, British Columbia,
Canada
Age Groups:
All ages
rr c/\ wr
ob-04 yr
65-74 yr
>75yr
Outcome: Mortality:
Total (nonaccidental) (<800)
Respiratory (460-51 9)
Cardiovascular (390-459)
Study Design: Time-series
Statistical Analyses: Poisson GAM,
LOESS
Age Groups: All ages
Pollutant: PM10
Averaging Time: 24-h avg
Mean (SD): 14.4 (5.9)
Range (Min, Max): (4.1, 37.2)
Copollutant (correlation): 0; r = 0.48
S02:r = 0.76
N02:r = 0.84
CO: r = 0.71
> 75: 1.6% (-1.8, 5.0)0
Respiratory
All ages: 5.1% (0.6, 9.6) 0-2 ma
55-64: 1.4% (-11. 3, 14.2) 0-2 ma
65-74: 2.8% (-9.5, 15.2) 0-2 ma
> 75: 10.2% (-0.1, 20.5) 0-2 ma
Other-causes
All ages: 2.4% (1.3, 3.5) 0-4 ma
55-64: 1.7% (-1.1, 4.5) 0-4 ma
65-74: 5.6% (3. 1,8. 1)0-4 ma
> 75: 3.7% (1.8, 5.6) 0-4 ma
The study does not present quantitative
results
Reference: Venners et al. (2003,
0899311
Period of Study: Jan 1995-Dec 1995
Location: Chongqing, China
Outcome: Mortality:
Total (nonaccidental) (<800)
Study Design: Time-series
Statistical Analyses: Poisson GAM,
cubic spline
Age Groups: All ages
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): 146.8
Range (Min, Max): (44.7, 666.2)
Copollutant: S02
Notes: PMi0 was measured for only 7
mo of the study period.
Increment: 100|jg/m
Relative Risk (Lower Cl, Upper Cl)
lag:
1.00(0.93,1.07)0
0.98(0.91, 1.04)1
1.00(0.93,1.07)2
0.96(0.90,1.03)3
0.97(0.90, 1.03)4
0.99(0.93,1.06)5
Reference: Vichit-Vadakan et al. (2008, P,utcome : M°rtQf^
. i ' NL-M-i-^^^iH^nt-^ /Ann DQQ\
Period of Study: Jan 1999-Dec 2003
Location: Bangkok, Thailand
Nonaccidental (AOO-R99)
Cardiovascular (IOO-I99)
Ischemic heart diseases (I20-I25)
Stroke (I60-I69)
Conduction disorder (I44-I49)
Respiratory (JOO-J98)
Lower Respiratory Infection (J10-J22)
COPD (J40-J47)
Asthma (J45-J46)
Senility (R54)
Study Design: Time-series
Statistical Analyses: Poisson, natural
cubic spline
Age Groups: All ages
0-4 yr
5-44 yr
18-50yr
45-64 yr
>50yr
>65yr
>75yr
Pollutant: PM10
Averaging Time: 24-h avg
Mean (SD): 52.1 (20.1)
Range (Min, Max): (21.3,169.2)
Copollutant (correlation): NR
Increment: 10|jg/m3
% Excess Risk (Lower Cl, Upper Cl)
lag:
Cause-specific mortality:
Nonaccidental: 1.3% (0.8,1.7)0-1
Cardiovascular: 1.9% (0.8, 3.0) 0-1
Ischemic heart disease:
1.5% (-0.4, 3.5)0-1
Stroke: 2.3% (0.6, 4.0) 0-1
Conduction disorders:
-0.%3(-5.9, 5.6)0-1
Cardiovascular:
> 65 1.8 (0.2, 3.3)0-1
Respiratory:
All ages: 1.0 (-0.4, 2.4)0-1
Ł1:14.6(2.9, 27.6)0-1
> 65:1.3 (-0.8, 3.3)0-1
LRI: <5: 7.7 (-3.6, 20.3) 0-1
COPD: 1.3 (-1.8, 4.4)0-1
Asthma: 7.4 (1.1,14.1)0-1
Senility: 1.8 (0.7, 2.8)0-1
Age-specific for nonaccidental
0-4: 0.2 (-2.0, 2.4) 0-1
5-44:0.9(0.2, 1.7)0-1
18-50:1.2(0.5, 1.9
45-64:1.1(0.4,1.9
0-1
0-1
> 50:1.4 (0.9, 1.9)0-1
December 2009
E-328
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
>65:1.5(0.9,2.1)0-1
> 75: 2.2 (1.3, 3.0)0-1
Sex-specific for nonaccidental
Male: 1.2 (0.7,1.7)0-1
Female: 1.3 (0.7,1.9)0-1
Nonaccidental
1.2(0.8,1.6)0
0.9(0.6,1.3)1
0.9
0.8
0.5, 1.3)2
0.4,1.2)3
0.3 (-0.1, 0.7) 4
1.3
1.4
0.8, 1.7)0-1
0.9,1.9)0-4
Cardiovascular
1.5(0.5,2.6)0
1.7(0.7,2.7)1
1.6
0.8
0.6, 2.6) 2
-0.1,1.8)3
-0.1 (-1.1,0.9)4
1.9
1.9
0.8, 3.0) 0-1
0.6, 3.2) 0-4
Respiratory
1.0 (-0.3, 2.3)0
0.8 (-0.5, 2.0) 1
-0.1,2.3)2
0.1,2.6)3
0.7 (-0.6, 1.9)4
1.3
1.0
1.9
a 65
1.5
-0.4, 2.4) 0-1
1.2,2.6)0-4
0.9, 2.0) 0
0.6,1.7)1
1.1(0.6,1.6)2
1.2
0.7
0.6, 1.7)3
0.2,1.2)4
1.5(0.9,2.1)0-1
1.9(1.2,2.6)0-4
Sensitivity analysis:
Nonaccidental (df):
3:1.3(0.9,1.8)
4:1.2(0.8,1.7)
6:1.3(0.8, 1.7)
6, with S02:1.2 (0.8, 1.7)
6, with N02:1.0 (0.2,1.8)
6, with Os: 1.1 (0.6, 1.7)
9:1.1(0.7,1.6)
12:1.1 (0.6, 1.5)
15:1.2(0.7, 1.6)
Cardiovascular (df):
" ' " 0.8, 2.7)
0.7, 2.6)
4:1.6
6:1.7(0.7,2.7)
6, with S02: 2.0 (0.9, 3.3)
6, with N02:2.3 (0.2, 4.3)
6, with 03:1.8 (0.5, 3.2)
9:1.7(0.6,2.8)
12:1.8(0.7(03.0)
15:2.2(0.9,3.4)
December 2009
E-329
-------
Study
Reference: Villeneuve et al. (2003,
0550511
Period of Study: 1986-1999
Location: Vancouver, Canada
Reference: Welty et al. (2008, 1571341
Period of Study: 1987-2000
Location: Chicago, Illinois
Reference: Welty and Zeger (2005,
0874841
Period of Study: 1987-2000
Location: 100 U.S. cities (NMMAPS)
Design & Methods
Outcome: Mortality:
Nonaccidental (<800)
Cardiovascular (401 -440)
Respiratory (460-51 9)
Cancer (140-239)
Study Design: Time-series
Statistical Analyses: Poisson, natural
solinss
Age Groups: 2 65 yr
Outcome: Mortality:
Total (nonaccidental)
Study Design: Time-series
Statistical Analyses: Poisson-Gibbs
Sampler
Bayesian Distributed Lag Model
Age Groups: All ages
Outcome: Mortality:
Total (nonaccidental) (<800)
Study Design: Time-series
Statistical Analyses: Bayesian
hierarchical model
Age Groups: All ages
Concentrations1
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD):
Daily 14.0
Every 6th Day 19.6
Range (Min, Max):
Daily (3.8, 52.2)
Every 6th Day (3.5, 63.0)
Copollutant:
S02
CO
N02
n
v-i
PM25
PM.n,r
r IVI10-2.5
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant (correlation): NR
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant (correlation): NR
Effect Estimates (95% Cl)
Increment: 15.4 pg/m3
% Increase (Lower Cl, Upper Cl) lag:
Nonaccidental
3.7% (-0.5, 8.0) 0-2 avg
2.6% -0.9,6.1)0
2.7% -0.7, 6.2) 1
1.9% (-1.4, 5.3)2
Cardiovascular
3.4% (-2.7, 9.8) 0-2 avg
5.1% 0.0, 10.4)0
1.3% -3.8, 6.7) 1
0.6% (-4.3, 5.7) 2
Respiratory
PM10
0.1% -9.5, 10.8 0-2 avg
1.0% -7.5, 10.4 0
0.4% (-7.7, 9.3) 1
-1.3% (-8.9, 7.1)2
Cancer
1.2% (-6.9, 10.1) 0-2 avg
-2.5% (-8.8, 4.3) 0
2.3% (-4.6, 9.6) 1
3.3% (-3.7, 10.8)2
Increment: 10|jg/m3
% Excess Risk (Lower Cl, Upper Cl)
i_ „.
lag.
Poisson-Gibbs Sampler
0.17% (0.01, 0.34) 3
-0.24% (-0.73, 0.23) 0-14
Unconstrained:
-0.19% (-0.86, 0.48)0-14
Bayesian Distributed Lag Model
-0.21% (-0.86, 0.41)0-14
Increment: 10|jg/m3
% Increase (SE) lag:
Distributed Lag Model:
Seasonally-Temporally Varying
Temperature variables: 0, 1-2, 1-7, 1-14
S(t,1xyr): 0.229 (0.053)1
S(t, 2 xyr): 0.220 (0.053)1
S(t, 4 xyr): 0.1 87 0.050 1
S(t, 8 xyr): 0.1 78 0.049 1
Temperature variables: 0, 1-2, 1-7,
1-14, 0x1-2, 0x1-7
1-2 x 1-7
S(t, 1 xyr): 0.1 95 0.048 1
S(t, 2 xyr): 0.200 0.051 1
Sft, 4 xyr): 0.176 (0.050)1
S(t, 8 xyr): 0.149 (0.050)1
Distributed Lag Model: Nonlinear
Temperature variables: 0,1-2,1-7,1-14
S(t, 4 xyr): 0.239 (0.053)1
Temperature variables: 0,1-2,1-7,
1-14,0x1-2,0x1-7,1-2x1-7
S(t, 4 xyr): 0.172 (0.045)1
Temperature variables: S(0,2), 3(1-2,2),
3(1-7,2), 3(1-14,2)
S(t, 4 xyr): 0.186 (0.046)1
Temperature variables: 3(0,2), 3(1-2,2),
3(1-7,2), 3(1-14,2), 3(0x1-2,2),
3(0x1-7,2), 3(1-2x1-7,2)
S(t, 4 xyr): 0.189 (0.047)1
December 2009
E-330
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
Temperature variables: S(0,4), 3(1-2,4),
3(1-7,4), 3(1-14,4)
S(t, 4 xyr): 0.175 (0.046)1
Temperature variables: 3(0,4), 3(1-2,4),
3(1-7,4), 3(1-14,4), 3(0x1-2,4),
3(0x1-7,4), 3(1-2x1-7,4)
S(t, 4 xyr): 0.190 (0.048)1
Temperature variables: 0,1-2,1-7
S(t, 4 xyr): 0.252 (0.053)1
Temperature variables: 0,1-2,1-7,
0x1-2,0x1-7, 1-2x1-7
S(t, 4 xyr): 0.186 (0.044)1
Temperature variables: 3(0,2), 3(1-2,2),
3(1-7,2)
S(t, 4 xyr): 0.198 (0.046)1
Temperature variables: 3(0,2), 3(1-2,2),
3(1-7,2), 3(0x1-2,2), 3(0x1-7,2),
3(1-2x1-7,2)
S(t, 4 xyr): 0.201 (0.047)1
Temperature variables: 3(0,4), 3(1-2,4),
3(1-7,4)
S(t, 4 xyr): 0.189 (0.045)1
Temperature variables: 3(0,4), 3(1-2,4),
3(1-7,4), 3(0x1-2,2), 3(0x1-7,4),
3(1-2x1-7,2)
S(t, 4 xyr): 0.205 (0.047)1
Temperature variables: 3(0,4), 3(1-2,4)
S(t, 4 xyr): 0.250 (0.045)1
Temperature variables: 3(0,4), 3(1-2,4),
3(0x1-2,4)
S(t, 4 xyr): 0.253 (0.044)1
Temperature variables: 3(0,4)
S(t, 4 xyr): 0.220 (0.045)1
Notes: 0 indicates current-day
temperature
1-r indicates avg of lag 1 through lag r
temperature
3 (, p) indicates a natural spline smooth
with p degrees of freedom.
3 (t, a x yr) indicates the natural spline
smooth of time with degrees of freedom
equal to a x (number of yr of data).
December 2009 E-331
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Wong et al. (2007, 0983911
Period of Study: Jan 1998-Dec 1998
Location: Hong Kong, China
Outcome: Mortality:
Total (nonaccidental) (<800)
Cardiorespiratory (390-519)
Study Design: Main analysis: Time-
series
Sensitivity analysis: Case-crossover,
case-only
Statistical Analyses: Main analysis:
Poisson GAM
Sensitivity analysis: Conditional
logistic regression
Age Groups: 2 30 yr; 2 65 yr
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD):
48.1 (24.3)
Range (Min, Max):
(15.5, 140.5)
Copollutant:
N02
S02
03
Increment: 10|jg/m
% Excess Risk (Lower Cl, Upper Cl)
lag:
Main Analysis
Nonaccidental
Smokers:
> 301:.80% (0.35, 3.26)0
1.77% (0.46, 3.11)2
> 65: 3.20% (1.36, 5.07)0
2.42% (0.73, 4.13)2
Never-smokers
> 30:-0.37% (-2.23, 1.52)0
-0.03% (-1.72, 1.66)2
>65P-0.70% (-2.81, 1.46)0
-0.13% (-2.04, 1.80)2
Cardiorespiratory
Smokers
> 30:1.43% (-0.86, 3.78)0
2.32% (0.24, 4.44) 2
> 65: 2.98% (0.47, 5.55) 0
2.61% (0.31, 4.95) 2
Never-smokers
> 30: 0.02% (-2.75, 2.87)0
-0.79% (-3.33, 1.82)2
> 65: 0.25% (-2.62, 3.19)0
-0.66% (-3.29, 2.04) 2
Sensitivity Analysis
Poisson Regression
Nonaccidental
> 30:1.81% (0.21, 3.44)0
1.93% (0.32, 3.56)2
1.99% (0.14, 3.87)0-3
> 65: 2.31% (0.37, 4.29)0
2.16% (0.20, 4.15)2
2.57% (0.30, 4.89) 0-3
Cardiorespiratory
> 30:1.04% (-1.45, 3.59)0
2.18% (-0.35, 4.77)2
1.66% (-1.24, 4.64)0-3
> 65:1.69% (-0.93, 4.37)0
2.44% (-0.23, 5.18)2
2.30% (-0.80, 5.50) 0-3
Case-only: Logistic Regression
Nonaccidental
> 30:1.79% (0.21, 3.37)0
1.94% (0.33, 3.56)2
> 65: 2.30% (0.42, 4.17)0
2.16% (0.26, 4.07)2
Cardiorespiratory
> 30:1.01% (-1.37, 3.40)0
2.16% (-0.28, 4.61)2
> 65:1.65% (-0.96, 4.27)0
2.42% (-0.27, 5.12)2
Case-crossover
Nonaccidental
> 30: 2.54% (0.35, 4.78) 0
1.35% (-0.81, 3.56) 2
> 65: 3.96% (1.37, 6.63)0
2.20% (-0.35, 4.81)2
Cardiorespiratory
> 30: 0.48% (-2.74, 3.80)0
3.24% (-0.03, 6.61)2
> 65: 2.17% (-1.40, 5.86)0
3.43% (-0.13, 7.13)2
Reference: Wong et al. (2007, 0932781 Outcome: Mortality:
Period of Study: Jan 1998-Dec 1998 Total (nonaccidental) (<800)
Pollutant: PMi0
Averaging Time: 24-h avg
Mean (SD):
48.1 (24.3)
Increment: 10|jg/m
% Excess Risk (Lower Cl, Upper Cl)
lag:
Nonaccidental
December 2009
E-332
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Location: Hong Kong, China
Cardiorespiratory (390-519)
Study Design: Main analysis: Time-
series
Sensitivity analysis: Case-only
Statistical Analyses: Main analysis:
Poisson GAM, natural cubic spline
Sensitivity analysis: Logistic
regression
Age Groups: 2 30 yr; 2 65 yr
Range (Min, Max): (15.5,140.5)
Copollutant:
N02
S02
03
Exercise
> 30: 0.13% (-1.16, 1.44 1
> 65: 0.24% (-1.16, 1.67 1
Never-exercise
> 30:1.04% (0.07, 2.02)1
> 65:1.26% (0.27, 2.27)1
Cardio-respiratory
Exercise
> 30: 0.46% (-1.43, 2.39 1
> 65: 0.30% (-1.65, 2.29 1
Never-exercise
> 30: 0.97% (-0.36, 2.32)1
> 65: 0.98% (-0.45, 2.43)1
Difference in % Excess Risk (Exercise
vs. Never-Exercise)
Nonaccidental
Poisson Regression
> 30:-2.86% (-4.03 to-1.67)1
> 65:-3.06% (-4.37 to-1.74)1
Case-only
>30:-2.91% -4.04to-1.77 1
> 65:-3.12% -4.38 to-1.84 1
Cardiorespiratory
Poisson regression
> 30:-2.55% (-4.32 to-0.75)1
> 65:-2.64% (-4.48 to-0.76)1
Case-only
> 30:-2.63% -4.32 to-0.92 1
> 65:-2.73% -4.50 to-0.92 1
Adjusted Case-only
Nonaccidental
Sex
> 30: -2.88%
> 65: -3.09%
-1.73 to-4.01
-1.82 to-4.35
Education
> 30:-2.94% (-1.80 to-4.07)1
> 65:-3.18% (-1.90 to -4.44)1
Job
> 30:-2.88% (-1.74 to-4.02)1
> 65:-3.11% (-1.83 to -4.37)1
Smoking
> 30:-2.82% (-1.66 to-3.96)1
> 65:-2.97% (-1.68 to-4.25)1
Illness time
> 30:-2.94% (-1.80 to-4.07)1
> 65:-3.16% (-1.88 to -4.42)1
Cardiorespiratory
Sex
> 30:-2.61% (-0.89 to-4.29)1
> 65:-2.71% (-0.90 to-4.48)1
Education
> 30:-2.58% (-0.85 to-4.27)1
> 65:-2.77% (-0.95 to-4.54)1
Job
> 30:-2.68% -0.96 to-4.37 1
> 65:-2.68% -0.88 to-4.46 1
Smoking
> 30:-2.46% (-0.73 to-4.17)1
> 65:-2.50% (-0.68 to-4.29)1
Illness Time
> 30:-2.63% (-0.91 to-4.32)1
> 65:-2.73% (-0.92 to-4.51)1
December 2009
E-333
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Case-only by Exercise Group (Never as
Reference)
Nonaccidental
>30
Low:-3.34% (-5.77 to-0.85)1
Moderate:-6.32% (-8.55 to-4.03)1
High:-1.74% (-3.06 to-0.40)1
265
Low:-3.79% (-6.67 to-0.82)1
Moderate: -7.78% (-10.39 to -5.10) 1
High:-1.77% (-3.21 to-0.31)1
Cardiorespiratory
>30
Low: -3.95% (-7.77, 0.04) 1
Moderate: -8.50% (-11.84 to -5.02) 1
High:-0.62% (-2.58, 1.38)1
>65
Low:-3.97% (-8.17, 0.43)1
Moderate:-9.42% (-13.00 to-5.69)1
High:-0.68% (-2.71, 1.38)1
Reference: Wong et al. (2002, 0254361 Outcome: Mortality:
Period of Study: 1995-1998
Location: Hong Kong, China
Respiratory (461-519)
COPD (490-496)
Pneumonia & Influenza (480-487)
Cardiovascular (390-459)
IHD (410-414)
Cerebrovascular (430-438)
Study Design: Time-series
Statistical Analyses: Poisson
Age Groups: 2 30 yr; 2 65 yr
Pollutant: PM10
Averaging Time: 24-h avg
Mean (SD):
51.53(24.79)
Range (Min, Max):
(14.05, 163.79)
Copollutant (correlation):
N02:r = 0.780
S02:r = 0.344
03:r = 0.538
Increment: 10|jg/m3
Relative Risk (Lower Cl, Upper Cl)
lag:
Respiratory
1.008(1.001to1.014)1
COPD
1.017(1.002, 1.033)0-3
Pneumonia & Influenza
1.007(0.999,1.015)2
Cardiovascular
1.003(0.998,1.016)2
IHD
1.013(1.001, 1.025)0-3
Cerebrovascular
1.007(0.998,1.016)2
Respiratory
PMio+S02+03+N02:
1.005(0.992, 1.010)1
COPD
PMio+S02+03+N02:
0.991 (0.968, 1.015)0-3
PMio+03+N02:
0.993(0.970,1.016)0-3
Pneumonia & Influenza
PMio+S02+03+N02:
1.002(0.991,1.013)2
IHD
0.994(0.978, 1.009)0-3
December 2009
E-334
-------
Study
Reference: Wong et al. (2008, 1571521
Period of Study:
Bangkok: 1999-2003
Hong Kong: 1996-2002
ShanghaiS Wuhan: 2001-2004
Location: Bangkok, Thailand
Design & Methods
Outcome (ICD10): Mortality:
Natural causes (AOO-R99)
Cardiovascular (IOO-I99)
Respiratory (JOO-J98)
Study Design: Time-series
Concentrations1
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD):
Bangkok: 52.0
Hong Kong: 51. 6
Shanghai: 102.0
Wuhan: 141.8
Effect Estimates (95% Cl)
Increment: 10|jg/m3
% Excess Risk (Lower Cl, Upper Cl)
lag:
Random Effects (4 cities)
Natural causes: 0.55% (0.26, 0.85 0-1
Cardiovascular: 0.58% (0.22, 0.93 0-1
Respiratory: 0.62% (0.22, 1.02)0-1
Random Effects (3 Chinese cities)
Hong Kong, Shanghai, and Wuhan,
China
natural splines
Age Groups: All ages
>65yr
>75yr
Range (Min, Max):
Bangkok: (21.3,169.2)
Hong Kong: (13.7,189.0)
Shanghai: (14.0, 566.8)
Wuhan: (24.8, 477.8)
Copollutant:
N02
S02
03
Natural causes: 0.37% (0.21, 0.54) 0-1
Cardiovascular: 0.44% (0.19, 0.68) 0-1
Respiratory: 0.60% (0.16,1.04) 0-1
Sensitivity Analysis
Random Effects (4 cities)
Omit PM,o>95th: 0.53% (0.27, 0.78) 0-1
Omit PM10>75th: 0.53% (0.29, 0.78) 0-1
Omit PM10>180 pg/m3:
0.65% (0.24, 1.06)0-1
Omit stations with high traffic source:
0.55% (0.26, 0.85) 0-1
Warm season-dichotomous variables:
0.86% (0.11,1.60) 0-1
Add temperature at lag 1-2 days: 0.51%
(0.23, 0.79) 0-1
Add temperature at lag 3-7 days: 0.35%
(0.14,0.57)0-1
Daily PMio defined by centering: 0.54%
(0.26, 0.82) 0-1
Natural spline with (8, 4, 4f: 0.54%
(0.26,0.81)0-1
Penalized spline:
0.52% (0.26, 0.77) 0-1
Random Effects (3 Chinese cities)
Omit PM,o>95th: 0.47% (0.21, 0.73) 0-1
Omit PM10>75th: 0.55% (0.24, 0.85) 0-1
Omit PM10>180 pg/m3:
0.46% (0.15, 0.76)0-1
Omit stations with high traffic source:
0.38% (0.20, 0.57) 0-1
Warm season-dichotomous variables:
0.43% (0.10, 0.76)0-1
Add temperature at lag 1-2 days:
0.36% (0.18, 0.53)0-1
Add temperature at lag 3-7 days:
0.25% (0.10, 0.40)0-1
Daily PM10 defined by centering:
0.37% (0.21, 0.53) 0-1
Natural spline with (8, 4, 4f:
0.36% (0.23, 0.49) 0-1
Penalized spline:
0.34% (0.23, 0.45) 0-1
Reference: Wong et al. (2008,1571511
Period of Study: Jan 1996-Dec 2002
Location: Hong Kong
Outcome (ICD10): Mortality:
Nonaccidental (AOO-T99
ZOO-Z99)
Cardiovascular (IOO-I99)
Respiratory (JOO-J98)
Study Design: Time-series
Statistical Analyses: Poisson GLM,
natural splines
Age Groups: All ages
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD): 51.6 (25.3)
Range (Min, Max): (13.5,188.5)
Copollutant:
N02
S02
03
Increment: 10 pg/m
% Excess Risk (Lower Cl, Upper Cl)
lag:
Nonaccidental:
Low SDI
0.37 (-0.10,0.84)0
0.40 (-0.04, 0.84) 1
0.14 (-0.28, 0.57)2
-0.12 (-0.55, 0.30)3
-0.14 (-0.56, 0.28)4
Middle SDI
0.70(0.34, 1.07)0
0.48 0.14, 0.82 1
0.35 (0.02, 0.68) 2
0.18 (-0.14, 0.51)3
0.17 (-0.16, 0.50)4
High SDI
0.22 (-0.29, 0.73
0.46 (-0.01, 0.94
0.29 (-0.17, 0.75)2
-0.05
-0.06
-0.51,0.40)3
-0.51,0.40)4
All areas
December 2009
E-335
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
0.45(0.19,0.72)0
0.40 0.15, 0.64 1
0.22 (-0.02, 0.45) 2
0.00 (-0.24, 0.23) 3
0.03 (-0.20, 0.26) 4
Cardiovascular:
Low SDI
0.14 (-0.77, 1.06)0
0.64 (-0.21, 1.49)1
0.24 (-0.58, 1.07)2
-0.27 (-1.09, 0.55)3
0.01 (-0.80, 0.83) 4
Middle SDI
0.66(0.00,1.34)0
0.49 (-0.13, 1.12)1
0.80(0.20,1.40)2
0.65(0.06,1.25)3
0.52 (-0.07, 1.12)4
High SDI
0.83 (-0.08, 1.75)0
0.89(0.04,1.75)1
0.12 (-0.70, 0.95)2
-0.09 (-0.91, 0.73) 3
0.04 (-0.77, 0.86) 4
All areas
0.52(0.05,1.00)0
0.58(0.14,1.03)1
0.43 (0.00, 0.86) 2
0.14 (-0.28, 0.57)3
0.23 (-0.20, 0.65) 4
Respiratory:
Low SDI
00.69 (-0.44, 1.82
10.55 (-0.50, 1.61
2 0.36 (-0.66, 1.39)2
3 -0.24
4-0.17
-1.25,0.78)3
-1.17,0.85)4
Middle SDI
0.31 (-0.50, 1.13)0
0.77(0.01,1.53)1
0.85(0.12, 1.59)2
0.66 (-0.07, 1.39 3
0.69 (-0.03, 1.42 4
High SDI
0.27 (-0.85, 1.40)0
0.72 (-0.32, 1.78)1
1.46 0.45, 2.47)2
0.70 (-0.30, 1.71)3
0.48 (-0.52, 1.48)4
All areas
0.39 (-0.20, 0.99) 0
0.70(0.15, 1.26)1
0.89(0.36,1.42)2
0.45 (-0.08, 0.98) 3
0.43 (-0.10,0.96)4
High SDI vs. Middle SDI
Nonaccidental: 0.23 (-0.25, 0.72) 0-1
Cardiovascular: 0.49 (-0.40,1.40) 0-1
Respiratory: 0.49 (-0.58,1.58)0-1
High SDI vs. Low SDI
Nonaccidental: 0.12 (-0.42, 0.67) 0-1
Cardiovascular: 0.82 (-0.20,1.86) 0-1
Respiratory:-0.15 (-1.39,1.10)0-1
Trend Test
Nonaccidental: 0.04 (-0.15, 0.22) 0-1
Cardiovascular: 0.27 (-0.07, 0.61) 0-1
Respiratory: -0.04 (-0.46, 0.37)
December 2009 E-336
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
0-1 SDI = Social Deprivation Index. The
higher the SDI the lower the SES of the
individual.
Reference: Yang et al. (2004, 0556031 Outcome: Mortality: Pollutant: PMi0 Increment: 31.43 pg/m3
Period of Study: 1994-1998 Nonaccidental (<800) Averaging Time: 24-h avg Odds Ratio (Lower Cl, Upper Cl) lag:
Location: Taipei, Taiwan Circulatory (390-459) Mean (SD): 51. 99 Nonaccidental
Respiratory (460-51 9) Range (Min, Max): (13.71, 211.30) 0.995(0.971, 1.020)0
Study Design: Bi-directional case- Copollutant: Respiratory
2
N02 0.986(0.906,1.074)0
Statistical Analyses: Conditional CO
logistic regression 03 Circulatory
Age Groups: All ages 0.988(0.942,1.035)
December 2009 E-337
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Zanobetti et al. (2003,
0428121
Period of Study: 1990-1997
Location: 10 European cities
(APHEA2)
Outcome: Mortality:
Nonaccidental (<800)
Circulatory (390-459)
Respiratory (460-519)
Study Design: Time-series
Statistical Analyses: Poisson GAM
Age Groups:
15-64yr
65-74 yr
>75yr
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD):
Athens: 42.7 (12.9)
Budapest: 41 (9.1)
Lodz: 53.5 (15.5)
London: 28.8 (13.7)
Madrid: 37.8 (17.7)
Paris: 22.5 (11.5)
Prague: 76.2 (45.7)
Rome: 58.7 (17.4)
Stockholm: 15.5 (7.9)
Tel Aviv: 50.3 (57.5)
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl) lag:
Cardiovascular
0.69% (0.31, 1.08)0-1 avg
40-day distributed lag
1.99% (1.44, 2.54)
4th degree
1.97% (1.38, 2.55)
Unrestricted
Respiratory
0.74% (-0.17, 1.66) 0-1 avg
40-day distributed lag
4.21% (1.70, 6.79)
4th degree
4.20% (1.08, 7.42)
Unrestricted
Unrestricted distributed lags
Cardiovascular
1.34% (0.89, 1.79)20
1.72% (1.20, 2.25)30
1.97% (1.38, 2.55)40
Respiratory
1.71% (-0.65, 4.12)20
2.62% (0.19, 5.11)30
4.20% (1.08, 7.42)40
40-day lags
Nonaccidental
15-64
-0.25% (-0.87, 0.36)
4th degree
-0.01 (-0.76, 0.75)
Unrestricted
65-74
0.78% (0.23, 1.33)
4th degree
0.74% (0.02, 1.45)
Unrestricted
>75
1.84% (0.92, 2.78)
4th degree
1.94% (1.07, 2.81)
Unrestricted
Cardiovascular
65-74
2.06% (1.05, 3.09)
4th degree
1.62(0.54,2.70)
Unrestricted
>75
2.35% (1.42, 3.29)
4th degree
2.52% (1.57, 3.48)
Unrestricted
Respiratory
>75
4.57% (1.25, 7.99)
4th degree
4.52% (0.89, 8.28)
Unrestricted
December 2009
E-338
-------
Study
Reference: Zeka et al. (2005, 0880681
Period of Study: Jan 1989-Dec 2000
Location: 20 U.S. cities
Reference: Zeka et al. (2006, 0887491
Period of Study: Jan 1989-Dec 2000
Location: 20 U.S. cities
Design & Methods
Outcome (ICD10): Mortality:
All-cause (nonaccidental) (V01-Y98)
Heart Disease (101 -151)
IHD(I20-I25)
Myocardial infarction (121, 122)
Dysrhythmias (I46-I49)
Heart failure (ISO)
Stroke (I60-I69)
Respiratory (JOO-J99)
Pneumonia (J12-J18)
COPD (J40-J44, J47)
Study Design: Time-stratified case-
crossover
Statistical Analyses: Conditional
logistic regression
Age Groups: All ages
Outcome (ICD10): Mortality:
All-cause (nonaccidental) (V01-Y98)
Heart Disease (101 -151)
Myocardial infarction (121, 122)
Stroke (I60-I69)
Respiratory (JOO-J99)
Study Design: Time-stratified case-
Concentrations1
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD):
Birmingham: 31.9 (18.0) pg/m3
Boulder: 22.1 (11.3)
Caton:26.6(11.5)
Chicago: 33.7 (16.4)
Cincinnati: 31. 4 (13.9)
Cleveland: 37.5 (18.7)
Colorado Springs: 24.0 (13.2)
Columbus: 28.5 (12.5)
Denver: 28.5 (12.8)
Detroit: 32. 1(1 7.7)
Honolulu: 15.9 (6.8)
Minneapolis: 24.7 (12.3)
Nashville: 30.1 (12.1)
New Haven: 25.4 (14.4)
Pittsburgh: 30.2 (18.5)
Provo: 33.7 (22.2)
Seattle: 26.4 (14.7)
Salt lake City: 35.0 (20.8) u
Terra Haute: 29.2 (14.6) u
Youngstown: 30.8 (13.9)
Range (Min, Max): NR
Copollutant (correlation): NR
Pollutant: PM,0
Averaging Time: 24-h avg
Mean (SD):
Birmingham: 31. 9 (18.0) pg/m3
Boulder: 22.1 (11.3)
Caton: 26.6(11.5)
Chicago: 33.7 (16.4)
Cincinnati: 31. 4 (13.9)
Cleveland: 37.5 (18.7)
Colorado Springs: 24.0 (13.2)
Effect Estimates (95% Cl)
Increment: 10|jg/m3
% Increase (Lower Cl, Upper Cl) lag:
Single-lag model
All-Cause (nonaccidental)
0.20% (0.08, 0.32) 0
0.35% (0.21, 0.49)1
0.24% (0.14, 0.34)2
Respiratory
0.34% (-0.07, 0.75) 0
0.52% (0.15, 0.89)1
0.51% (0.16, 0.86)2
COPD
-0.06% (-0.63, 0.51)0
0.43% (-0.14, 1.00 1
0.39% (-0.16, 0.94 2
Pneumonia
0.50% (0.09, 1.09)0
0.59% (-0.12, 1.30)1
0.82% (0.25, 1.39)2
Heart disease
0.12% (-0.06, 0.30)0
0.30% (0.12, 0.48)1
0.37% (0.17, 0.57)2
IHD
0.19% (-0.03, 0.41)0
0.41% (0.19, 0.63)1
0.43% (0.10, 0.76)2
Myocardial Infarction
0.36% (-0.05, 0.77 0
0.17% (-0.18, 0.52 1
0.13% (-0.22, 0.48)2
Heart Failure
0.17% (-0.63, 0.97)0
-0.01% (-0.81, 0.79)1
0.78% (-0.004, 1.56)2
Dysrhythmias
-0.23% (-1.41, 0.95)0
0.37% (-0.47, 1.21 1
0.33% (-0.55, 1.21 2
Stroke
0.09% (-0.49, 0.60 0
0.41% (-0.02, 0.84 1
0.14% (-0.27, 0.55)2
Unconstrained distributed lag model
All-cause (nonaccidental)
0.45% (0.25, 0.65) 0-3
Respiratory
0.87% (0.38, 1.36)0-3
COPD
0.43% (-0.35, 1.21)0-3
Pneumonia
1.24% (0.46, 2.02)0-3
Heart Disease
0.50% (0.25, 0.75) 0-3
IHD
0.65% (0.32, 0.98)
Myocardial Infarction
0.36% (-0.25, 0.97) 0-3
Heart Failure
0.60% (-0.50, 1.70)0-3
Dysrhythmias
0.20% (-1.03, 1.43)0-3
Stroke
0.46% (-0.13, 1.05)0-3
Increment: 10|jg/m3
% Increase (Lower Cl, Upper Cl)
lag: All-cause (nonaccidental)
Male: 0.46% (0.28 0.64) 1-2 avg
Female: 0.37% (0.17, 0.57) 1-2 avg
White: 0.40% (0.22, 0.58) 1-2 avg
Black: 0.37% (-0.02, 0.76) 1-2 avg
Age:
<65: 0.25% (0.01, 0.49) 1-2 avg
75: 0.23% (-0.06, 0.52) 1-2 avg
December 2009
E-339
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
crossover
Statistical Analyses: Conditional
logistic regression
Age Groups:
All ages
<65yr
65-75 yr
>75yr
Columbus: 28.5 (12.5)
Denver: 28.5 (12.8)
Detroit: 32.1(17.7)
Honolulu: 15.9 (6.8)
Minneapolis: 24.7 (12.3)
Nashville: 30.1 (12.1)
New Haven: 25.4 (14.4)
Pittsburgh: 30.2 (18.5)
Provo: 33.7 (22.2)
Seattle: 26.4 (14.7)
Salt lake City: 35.0 (20.8)
Terra Haute: 29.2 (14.6)
Youngstown: 30.8 (13.9)
Range (Min, Max): NR
Copollutant (correlation): NR
>75: 0.64% (0.44, 0.84) 1-2 avg
Educational Attainment:
Low (<8 yr):
0.62% (0.29, 0.95) 1-2 avg
Medium (8-12 yr):
0.36% (0.12, 0.60) 1-2 avg
High(>12yr):
0.27% (-0.004, 0.54) 1-2 avg
Location of Death:
In hospital: 0.22% (0.04, 0.40) 1-2 avg
Out of hospital:
0.71% (0.51, 0.91) 1-2 avg
Season:
Winter: 0.28% (0.04, 0.52) 1-2 avg
Summer: 0.19% (-0.22, 0.60) 1-2 avg
Transition (spring/fall):
0.49% (0.25, 0.73) 1-2 avg
Respiratory
Male: 0.71% (0.004, 1.42)0-3
Female: 1.04% (0.33,1.75)0-3
White: 0.88% (0.33, 1.43)0-3
Black: 0.71% (-0.56,1.98)0-3
Age:
<65: 0.94% (-0.31, 2.19) 0-3
65-75: 0.87% (-0.25, 1.99)0-3
>75: 0.88% (0.17, 1.59)0-3
Educational Attainment:
Low (<8 yr):
0.82% (-0.32, 1.96)0-3
Medium (8-12 yr):
0.88% (0.12, 1.64)0-3
High(>12yr):
0.88% (-0.04, 1.80)0-3
Location of Death:
In hospital: 0.78% (0.17,1.39)0-3
Out of hospital: 1.09% (0.25,1.93) 0-3
Season:
Wnter: -0.007% (-0.87, 0.86) 0-3
Summer: 0.69% (-0.68, 2.06) 0-3
Transition (spring/fall):
1.57% (0.86, 2.28)0-3
Heart Disease
Male: 0.54% (0.23, 0.85) 2
Female: 0.46% (0.15, 0.77)2
White: 0.50% (0.25, 0.75) 2
Black: 0.64% (0.13,1.15)2
<65: 0.04% (-0.45, 0.53) 2
65-75: 0.60% (0.13, 1.07)2
>75: 0.65% (0.30, 1.00)2
Educational Attainment:
Low (<8 yr):
0.72% (0.23, 1.21)2
Medium (8-12 yr):
0.38% (0.07, 0.69) 2
High(>12yr):
0.54% (0.13, 0.95)2
Location of Death:
In hospital: 0.15% (-0.14, 0.44)2
Out of hospital: 0.93% (0.60,1.26) 2
Season:
Wnter: 0.41% (-0.002, 0.82)2
Summer: 0.52 (0.03,1.01)2
Transition (spring/fall):
0.56% (0.13, 0.99)2
December 2009
E-340
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
Myocardial Infarction
Male: 0.21% (-0.40, 0.82)0
Female: 0.59% (0.08,1.10)0
White: 0.24% (-0.27, 0.75) 0
Black: 0.99% (0.05,1.93)0
<65: 0.12% (-0.76, 1.00)0
65-75: 0.92% (0.21, 1.63)0
>75: 0.16% (-0.58, 0.90)0
Educational Attainment:
Low (<8yr): 0.33% (-0.83, 1.49)0
Medium (8-12 yr): 0.79% (0.28,1.30) 0
High (>12yr):-0.13% (-0.82, 0.56)0
Location of Death:
In hospital: 0.34% (-0.11, 0.79)0
Out of hospital: 0.48% (-0.23,1.19) 0
Season:
Winter: 0.32% (-0.37,1.01)0
Summer: 0.30% (-0.82,1.42)0
Transition (spring/fall):
0.38%-0.31,1.07)0
Stroke
Male: 0.11% (-0.58, 0.80)1
Female: 0.59% (-0.04,1.22)1
White: 0.48% (0.01, 0.95)1
Black: 0.13% (-0.87,1.13)1
(-1.09,1.27)1
65-75:-0.46% (-1.42, 0.50)1
>75: 0.80% (0.27, 1.33)1
Educational Attainment:
Low (<8yr): 0.07% (-1.44,1.58)1
Medium (8-12 yr): 0.29% (-0.32, 0.90)1
High (>12yr): 0.52% (-0.28,1.32)1
Location of Death:
In hospital: 0.06% (-0.49, 0.61)1
Out of hospital: 0.87% (0.05,1.69)1
Season:
Wnter: -0.09% (-0.93, 0.75) 1
Summer: 0.67% (-0.31,1.65)1
Transition (spring/fall):
0.51% (-0.20,1.22)1
Contributing causes of disease: All-
cause
Secondary pneumonia present:
0.67% (0.16, 1.18)1-2avg
Secondary pneumonia absent:
0.34% (0.16, 0.52) 1-2 avg
Secondary heart failure present:
0.42% (0.01, 0.83) 1-2 avg
Secondary heart failure absent:
0.37% (0.19, 0.55) 1-2 avg
Secondary stroke present:
0.85% (0.30, 1.40) 1-2 avg
Secondary stroke absent:
0.32% (0.14, 0.50) 1-2 avg
Diabetes present:
0.57% (0.02, 1.12) 1-2 avg
Diabetes absent:
0.34% (0.14, 0.54) 1-2 avg
Respiratory
Secondary pneumonia present:
1.28% (-0.33, 2.89)0-3
Secondary pneumonia absent:
0.78% (0.15,1.41)0-3
Secondary heart failure present:
1.48% (0.07, 2.89)0-3
Secondary heart failure absent:
December 2009 E-341
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
0.79% (0.26, 1.32)0-3
Secondary stroke present:
1.95% (-0.11,4.01) 0-3
Secondary stroke absent:
0.80% (0.29, 1.31)0-3
Diabetes present:
1.96% (-0.22, 4.14)0-3
Diabetes absent:
0.82% (0.31, 1.33) 0-3
Heart Disease
Secondary pneumonia present:
0.66% (-0.63, 1.95)2
Secondary pneumonia absent:
0.49% (0.27, 0.71)2
Secondary stroke present:
0.73% (-0.05, 1.51)2
Secondary stroke absent:
0.48% (0.24, 0.72) 2
Diabetes present:
0.34% (-0.42,1.10)2
Diabetes absent: 0
.52% (0.28, 0.76) 2
Myocardial Infarction
Secondary pneumonia present:
1.54% (-1.05, 4.13)0
Secondary pneumonia absent:
0.42% (0.05, 0.79) 0
Secondary stroke present:
0.50% (-1.38, 2.38)0
Secondary stroke absent:
0.36% (-0.05, 0.77) 0
Diabetes present:
0.70% (-0.38, 1.78)0
Diabetes absent:
0.41% (0.04, 0.78)0
Stroke
Secondary pneumonia present:
1.74% (0.35, 3.13) 1
Secondary pneumonia absent:
0.29% (-0.16, 0.74)1
Secondary heart failure present:
1.01% (-0.77,1.79)1
Secondary heart failure absent:
0.38% (-0.05, 0.81) 1
Diabetes present:
1.02% (-0.53, 2.57) 1
Diabetes absent:
0.37% (-0.08, 0.82) 1
1AII units expressed in ug/m3 unless otherwise specified.
December 2009
E-342
-------
Table E-17. Short-term
Study
Reference: Burnett et al. (2004,
0862471
Period of Study: 1981-1999
Location: 12 Canadian cities
Reference: Kan et al. (2007, 0912671
Period of Study: Mar 2004-Dec 2005
Location: Shanghai, China
exposure-mortality - PMir>25.
Design & Methods
Outcome: Mortality:
Nonaccidental (<800)
Study Design: Time-series
Statistical Analyses:
1. Poisson, natural splines
2. Random effects regression
model
Age Groups: All ages
Outcome (ICD10): Mortality:
Total (nonaccidental) (AOO-R99)
Cardiovascular (IOO-I99)
Respiratory (JOO-J98)
Study Design: Time-series
Statistical Analyses: Poisson
GAM, penalized splines
Age Groups: All ages
Concentrations'!
Pollutant: P1 0-2.5
Averaging Time: 24-h avg
Mean (SD): 1 1.4
Range (Min, Max): NR
Co pollutant:
hif~\
NU2
f~\
U3
S02
CO
PM,o
PM25
Note: PM10 measurement
calculated as the sum of PM25
and PM10.25 measurements.
Pollutant: PM^.s
Averaging Time: 24-h avg
Mean (SD): 56.4 (1.34)
Range (Min, Max): (8.3, 235.0)
Copollutant (correlation):
PM10:r = 0.88
PM25:r = 0.48
03:r = 0.07
Effect Estimates (95% Cl)
Increment: 10 pg/m3
% Increase (Lower Cl, Upper Cl) lag:
1981-1999
PM10.25: 0.31% (-0.66, 1.33)1
PM10.25+N02: 0.65% (-0.23, 1.59)1
Increment: 10 pg/m3
% Increase (Lower Cl, Upper Cl)
lag: Total: 0.12% (-0.13, 0.36)
0-1
Cardiovascular: 0.34% (-0.05, 0.73)
0-1
Respiratory: 0.40% (-0.34, 1.13)
0-1
Reference: Kettunen et al. (2007,
0912421
Period of Study: 1998-2004
Location: Helsinki, Finland
Outcome (ICD10): Mortality:
Stroke (160-161,163-164)
Study Design: Time-series
Statistical Analyses: Poisson
GAM, penalized thin-plate
splines
Age Groups: > 65 yr
Pollutant: PMi0.2.5
Averaging Time: 24-h avg
Median (SD) unit: Cold Season:
6.7
Warm Season: 8.4
Range (Min, Max):
Cold Season: (0.0,101.4)
Warm Season: (0.0, 42.0)
Copollutant: 03, CO, N02
PM10
PM25
UFP
Increment:
Cold Season: 8.3 pg/m3
Warm Season: 5.7 pg/m3
% Increase (Lower Cl, Upper Cl) lag:
Cold Season: -1.04% (-6.63, 4.89) 0
-2.49% (-7.57, 2.88)
1-4.93% (-9.99, 0.41)2
-4.33% (-9.32, 0.93) 3
Warm Season: 7.05% (-1.88,16.80) 0
4.38% (-4.26, 13.81)
1:-1.19% (-9.45, 7.84)2
1.42% (-6.79, 10.34)3
Reference: Klemm et al. (2004,
0565851
Period of Study: Aug 1998-Jul 2000
Location: Fulton and DeKalb counties,
Georgia (ARIES)
Outcome: Mortality:
Nonaccidental (<800)
Cardiovascular (390-459)
Respiratory (460-519)
Cancer (140-239)
Study Design: Time-series
Statistical Analyses: Poisson
GLM, natural cubic splines
Age Groups: <65 yr, 2 65 yr
Pollutant: PMi0.25
Averaging Time: 24-h avg
Mean (SD): 9.69 (3.94)
Range (Min, Max): (1.71, 25.17)
Copollutant: PM25
03
N02
CO
S02
Acid
EC
OC
S04
Oxygenated Hydrocarbons
Nonmethane hydrocarbons
N03
Increment: NR
P(SE)
lag:
Quarterly Knots:
0.00433 (0.00333) 0-1
Monthly Knots:
0.00617(0.00360)0-1
Biweekly Knots:
0.00516(0.00381)0-1
December 2009
E-343
-------
Study
Reference: Perez et al. (2008, 1560201
Period of Study: Mar 2003-Dec 2005
Location: Barcelona, Spain
Reference: Perez et al. (2008, 1560201
Period of Study: Mar 2003-Dec 2005
Location: Barcelona, Spain
Reference: Perez et al. (2008, 1560201
Period of Study: Mar 2003-Dec 2005
Location: Barcelona, Spain
Reference: Slaughter et al. (2005,
0738541
Period of Study: Jan 1995-Dec 1999
Location: Spokane, Washington
Design & Methods
Outcome: Respiratory mortality
Study Design: Cohort
Covariates: Temperature,
humidity
Statistical Analysis:
autoregressive Poisson
regression models
Statistical Package: NR
Age Groups: All deaths
Outcome:
Cardiovascular mortality
Study Design: Cohort
Covariates: Temperature,
humidity
Statistical Analysis:
Autoregressive Poisson
regression models
Statistical Package: NR
Age Groups: All deaths
Outcome:
Cerebrovascular mortality
Study Design: Cohort
Covariates: Temperature,
humidity
Statistical Analysis:
Autoregressive Poisson
regression models
Statistical Package: NR
Age Groups: All deaths
Outcome: Mortality:
Nonaccidental (< 800)
Study Design: Time-series
Statistical Analyses: Poisson
GLM, natural splines
Age Groups: All ages
Concentrations'!
Pollutant: PM,O., 5
Averaging Time: 24 h
Mean (SD) Unit: 14.0 (9.5)
pg/m
Range (Min, Max): 0.1, 93.1
Copollutant: PM25-1,PM1
Pollutant: PM,O., 5
Averaging Time: 24 h
Mean (SD) Unit: 14.0 (9.5)
pg/m
Range (Min, Max): 0.1, 93.1
Copollutant: PM25-1,PM1
Pollutant: PM10., 5
Averaging Time: 24 h
Mean (SD) Unit: 14.0 (9.5)
pg/m3
Range (Min, Max): 0.1, 93.1
Copollutant: PM25-1,PM1
Pollutant: PMi0.25
Averaging Time: 24-h avg
Mean (SD) unit: NR
Range (9th, 96th): NR
Copollutant (correlation):
PM1:r = 0.19
PM25:r = 0.31
PM,0:r = 0.94
CO: r = 0.32
Effect Estimates (95% Cl)
Increment: 10 pg/m3
Odds Ratio (96%CI) Lag
Single Pollutant Model
Avg LO-1: 1.000(0.944-1.060), p = 0.991
L1: 1.002 0.955-1.052), p = 0.931
L2: 1.070 1.023-1.118), p = 0.003
Multi-pollutant Model
Avg LO-1: 1.002 (0.937-1. 071), p = 0.958
L1 : 0.998 (0.943-1. 056), p = 0.0.936
L2: 1.033 (0.980-1. 089), p = 0.226
Increment: 10 pg/m3
Odds Ratio (96%CI) Lag
Avg LO-1: 1.054 (1.019-1.089), p = 0.002
L1: 1.059 (1.031-1.072), p = 0.000
L2: 1.044 (1.017-1.072), p = 0.001
Multi-pollutant Model
Avg LO-1: 1.053 (1.013-1.094), p = 0.009
L1 : 1.059 (1.026-1. 094), p = 0.001
L2: 1.044 (1.012-1.078), p = 0.007
Increment: 10 pg/m3
Odds Ratio (96%CI) Lag
Avg LO-1: 1.087 (1.018-1. 161), p = 0.013
L1: 1.086 1.030-1.145), p = 0.002
L2: 1.051 0.997-1.108), p = 0.064
Multi-pollutant Model
Avg LO-1: 1.103(1. 022-1. 191), p = 0.011
L1: 1.098 (1.030-1. 171), p = 0.004
L2: 1.076 (1.010-1. 146), p = 0.023
This study does not present quantitative results for
PM,0-2.5.
Reference: Stieb et al. (2002, 0252051
Period of Study:
Publication dates of studies: 1985-Dec
2000
Mortality series: 1958-1999
Location: 40 cities (11 Canadian cities,
19 U.S. cities, Santiago, Amsterdam,
Erfurt, 7 Korean cities)
Outcome: Mortality: All-cause
(nonaccidental)
Study Design: Meta-analysis
Statistical Analyses: Random
effects model
Age Groups: All ages
Pollutant: PMi0.25
Averaging Time: NR
Mean (SD): NR
Range (Min, Max): NR
Copollutant:
Varied between studies:
PM25,03, S02, N02, CO
Increment: 13.0 pg/m
% Increase (Lower Cl, Upper Cl) lag:
Single-pollutant models: 10 studies
PMi0.25:1.2%(0.5, 1.9)
Multipollutant models: 6 studies
PM10.25: 0.9% (-0.3, 2.0)
December 2009
E-344
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Villeneuve et al. (2003,
0550511
Period of Study: 1986-1999
Location: Vancouver, Canada
Outcome: Mortality:
Nonaccidental (<800)
Cardiovascular (401-440)
Respiratory (460-519)
Cancer (140-239)
Study Design: Time-series
Statistical Analyses: Poisson,
natural splines
Age Groups: 2 65
Pollutant: PMi0.2.5
Averaging Time: 24-h avg
Mean (SD):
Daily: 6.1
Every 6th Day
8.3
Range (Min, Max):
Daily: (0.0, 72.0)
Every 6th Day: (0.7, 35.0)
Co pollutant:
PM25
PM10
S02
CO
N02
03
Increment: 11.0|jg/m
% Increase (Lower Cl, Upper Cl) lag:
Nonaccidental
1.4% (-2.5, 5.4) 0-2 avg
1.0% (-1.9, 4.0)0
-1.1% (-4.0, 1.8)1
2.0% (-1.0, 5.1)2
Cardiovascular
5.9% (-0.2, 12.4) 0-2 avg
5.9% 1.1, 10.8)0
1.4% -3.3,6.4)1
2.2% (-2.0, 6.7) 2
Respiratory
-1.0% (-9.8, 8.8) 0-2 avg
-1.5% (-9.4, 7.1 "
-1.5% (-8.4, 6.0
0.1% (-6.4, 6.9)2
Cancer
4.4% (-3.6, 13.1) 0-2 avg
3.1% (-2.9, 9.4)0
-1.0% (-6.9, 5.3)1
4.0% (-2.1, 10.4) 2
Reference: Wilson et al. (2007,
1571491
Period of Study: 1995-1997
Location: Phoenix, Arizona
Outcome: Mortality:
Cardiovascular
Study Design: Time-series
Statistical Analyses: Poisson
GAM, nonparametric smoothing
spline
Age Groups: >25
Pollutant: PM10.25
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: 10 pg/m
% Excess Risk (Lower Cl, Upper Cl) lag:
Central Phoenix:
2.4% (-1.2, 6.1)0-5 ma
Middle Phoenix:
3.8%
3.4%
0.3, 7.5) 0-5 ma
1.0,5.8)1
3.0% (0.7, 5.4) 2
Outer Phoenix:
1.6% (-1.9, 5.2) 0-5 ma
1AII units expressed in ug/m3 unless otherwise specified.
December 2009
E-345
-------
Table E-18. Short-term exposure-mortality • PM2.5 (including PM components/sources).
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Basu et al. (2008, 0987161 Outcome (ICD10): Mortality:
Period of Study: May 1999-Sept 2003 Nonaccidental (V01-Y98)
Location: 9 California counties
Study Design:
(1) Main analysis: Case-crossover
(2) Sensitivity analysis: Time-series
Statistical Analyses:
(1) Main analysis: conditional logistic
regression
(2) Sensitivity analysis: Poisson GAM
Age Groups: All ages
Pollutant: PM25
Averaging Time: 24-h avg
Mean (SE) unit:
Contra Costa: 8.6
Fresno: 7.6
Kern: 11.3
Los Angeles: 19.8
Orange: 17.0
Riverside: 28.4
Sacramento: 8.8
San Diego: 13.4
Santa Clara: 10.8
IQR (25th, 75th):
Contra Costa: (5.8,10.1)
Fresno: (3.8, 9.8)
Kern: (8.0,13.5)
Los Angeles: (14.7, 23.3)
Orange: (11.8, 21.0)
Riverside: (17.9, 36.1)
Sacramento: (5.8,10.1)
San Diego: (10.3,15.8)
Santa Clara: (7.2,13.8)
Copollutant (correlation):
PM,or = 0.45
The study does not provide results
quantitatively.
Reference: Dominici et al. (2007,
097361)
Period of Study:
PM,0: 1987-2000
DM • IQQQ onnn
rivi25. iyyy-zuuu
Location: 100 U.S. counties
(NMMAPS)
Reference: Dominici et al. (2007,
0991351
Period of Study: 2000-2005
Location: 72 U.S. counties
representing 69 communities
Reference: Franklin et al. (2007,
0912571
Period of Study: 1997-2002
Location: 27 U.S. communities
Outcome: Mortality:
All-cause (nonaccidental)
Cardiorespiratory
Other-cause
Study Design: Time-series
Statistical Analyses: 2-stage Bayesian
hierarchical model
Age Groups: All ages
Outcome: Total mortality
Study Design: Time-series
Statistical Analyses: 2-stage Bayesian
hierarchical model
Age Groups: All ages
Outcome: Mortality:
All-cause (nonaccidental (<800)
Cardiovascular (390-429)
Respiratory (460-51 9)
Stroke (430-438)
Study Design: Time-stratified case-
crossover
Statistical Analyses: Conditional
logistic regression
Age Groups: All ages
03 1hr r = 0.28
03 8hr r = 0.22
CO r = 0.45
N02r = 0.43
Pollutant: PM25
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant (correlation): NR
Pollutant: PM25, Nickel, speciated fine
PM and Vanadium
Averaging Time: Annual avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant (correlation): NR
Pollutant: PM25
Averaging Time: 24-h avg
Mean (SD): 15.7 pg/m3
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: 10|jg/m3
% Increase (Lower Cl, Upper Cl) lag:
1999-2000:
All-cause: 0.29% (0.01, 0.57)1
Cardiorespiratory: 0.38% (-0.07, 0.82) 1
The study does not provide results
quantitatively.
Note: The study investigated whether
county-specific short-term effects of
PM10 on mortality are modified by long-
term county-specific nickel or vanadium
PM2 5 concentrations.
Increment: 10|jg/m3
% Increase (Lower Cl, Upper Cl) lag:
All-cause (nonaccidental):
0.67% (-0.12, 1.46)0
1.21% (0.29, 2.14)
10.82% (0.02, 1.63)0-1
Respiratory:
1.31% (-0.10, 2.73)0
1.78% (0.20, 3.36) 1
1.67% (0.19, 3.16)0-1
Cardiovascular:
0.34% (-0.61, 1.28 0
0.94% (-0.14, 2.02 1.
0.54% (-0.47, 1.54)0-1
Stroke:
0.62% (-0.69, 1.94)0
1.03% (0.02, 2.04)1.
0.67% (-0.23, 1.57)0-1
December 2009
E-346
-------
Study Design & Methods Concentrations! Effect Estimates (95% Cl)
Age> 75:
All cause: 1.66% (0.62, 2.70)1
Respiratory: 1.85% (0.27, 3.44) 1
Cardiovascular: 1.29% (0.15, 2.42) 1
Stroke: 1.52% (0.37, 2.67)1
Age<75:
All cause: 0.62% (-0.30,1.55)1
Respiratory: 1.53% (-0.67, 3.74) 1
Cardiovascular: 0.26% (-1.04,1.56) 1
Stroke: -0.78% (-2.32, 0.76) 1
Male:
All cause: 1.06% (0.07, 2.06)1
Respiratory: 1.90% (0.14, 3.65)1
Cardiovascular: 0.52% (-0.63,1.66) 1
Stroke: 0.79% (-0.42, 2.02) 1
Female:
All cause: 1.34% (0.40, 2.27)1
Respiratory: 1.57% (-0.22, 3.35) 1
Cardiovascular: 1.30% (0.14, 2.46) 1
Stroke: 0.79% (-0.51, 2.09)1
East:
II cause: 1.95% (0.50, 3.40)1
Respiratory: 2.66% (0.33, 5.00) 1
Cardiovascular: 1.52% (0.06, 2.98) 1
Stroke: 1.16% (-0.40, 2.73)1
V\fest:
All cause: 0.05% (-1.80,1.89)1
Respiratory: 0.67% (-2.00, 3.34)11
Cardiovascular: 0.11% (-2.03, 2.24) 1|
Stroke: 0.94% (-0.38, 2.26) 1
PM25>15|jg/m3:
All cause: 1.10% (-0.43, 2.64)1
Respiratory: 1.42% (-0.84, 3.68) 1
Cardiovascular: 0.88% (-0.87, 2.62) 1
Stroke: 0.91% (-0.28, 2.10)1
PM25<15|jg/m3:
All cause: 1.41% (-0.49, 3.30)1
Respiratory: 2.46% (-0.49, 5.42) 1
Cardiovascular: 1.09% (-1.15, 3.32) 1
Stroke: 1.36% (-0.56, 3.27)1
Effect of A/C at percentile of air
conditioning prevalence:
25th percentile (45% prevalence of
A/C):
All cause: 1.50% (0.13, 2.88)1
Respiratory: 2.27% (0.27, 4.27) 1
Cardiovascular: 1.04% (-0.54, 2.63) 1
Stroke: 1.04% (-0.44, 2.53)1
75th percentile (80% prevalence of
A/C):
All cause: 0.85% (-0.64, 2.35) 1
Respiratory: 1.04% (-1.29, 3.37)1
Cardiovascular: 0.81% (-0.93, 2.61) 1
Stroke: 1.03% (-0.76, 2.83)1
Effect of A/C at percentile of air
conditioning prevalence in cities with
summer peaking PM25 concentrations:
25th percentile (45% prevalence of
A/C):
All cause: 1.01% (-0.30, 2.32)1
Respiratory: 0.76% (-1.38, 2.90)1
Cardiovascular: 0.43% (-0.86,1.72) 1
Stroke:-0.18% (-2.08,1.73)1
75th percentile (77% prevalence of
A/C):
All cause:-0.55% (-1.95, 0.85)1
December 2009 E-347
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Respiratory: -2.08% (-4.47, 0.31) 1
Cardiovascular: -1.02% (-2.44, 0.41) 1
Stroke: 0.69% (-1.19" 2.57)1
Reference: Franklin et al. (2008,
0974261
Period of Study: 2000-2005
Location: 25 U.S. communities
Outcome (ICD10): Mortality:
Nonaccidental (V01-Y98)
Respiratory (JOO-J99)
Cardiovascular (101-152)
Stroke (I60-J69)
Study Design: Time-series
Statistical Analyses:
1st stage: Poisson, cubic spline
2nd stage: Random effects meta-
analysis
Age Groups: All ages
Pollutant: PM25
Averaging Time: 24-h avg
Range Mean (SD):
Winter: 9.6-34.4
Spring: 6.7-27.6
Summer: 7.6-26.0
Fall: 9.5-32.1
Range (Min, Max): NR
Copollutant:
Al, As, Br, Cr, EC, Fe, K, Mn, Na*, Ni,
N03", NH4i OC, Pb, Si, S042", V, Zn
Increment: 10|jg/rrf
% Increase (Lower Cl, Upper Cl) lag:
Nonaccidental: 0.74% (0.41,1.07) 0-1
Cardiovascular: 0.47% (0.02, 0.92) 0-1
Respiratory: 1.01% (-0.03, 2.05) 1-2
Stroke: 0.68% (-0.21,1.57) 0-1
Winter: 0.15% (-0.42, 0.72)0-1
Spring: 1.88% (1.29, 2.48)0-1
Summer: 0.99% (0.35,1.68)0-1
Fall: 0.19% (-0.25, 0.64)0-1
West: 0.51% (0.10, 0.92)0-1
Easts Central:
0.92% (0.44, 1.39)0-1
% Increase per 10 pg/m3 increase in
PM25 for an IQR increase in species to
PM25 mass proportion
Univariate analysis
Al: 0.58%
As: 0.55%
Br: 0.38
Cr: 0.33%
EC: 0.06%
Fe:0.12%
K:0.41%
Mn:0.14%
Na+: 0.20%
Ni: 0.37%
N0r: -0.49%
NH4: 0.04%
OC: -0.02%
Pb:0.17%
Si: 0.41%
S042":0.51%
V: 0.30%
Zn: 0.23%
Multivariate(l)
Al: 0.79%
Ni: 0.34%
S042": 0.75%
Multivariate (2)
Al:0.61%
Ni: 0.35%
As: 0.58%
Reference: Holloman et al. (2004,
0873751
Period of Study: 1999-2001
Location: 7 North Carolina counties
Outcome (ICD10): Mortality:
Cardiovascular (IOO-I99)
Study Design: Time-series
Pollutant: PM25
Averaging Time: 24-h avg
Mean (SD): NR
Statistical Analyses: 3-stage Bayesian Range (Min, Max): NR
hierarchical model
Copollutant (correlation): NR
Age Groups: >16
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl)
lag:
2.5% (-3.9 to 9.6)
0
4.0% (-3.3 to 12.2)
1
11.4% (2.8-19.8)
2
-1.1% (-7.5 to 5.2)
3
December 2009
E-348
-------
Study
Reference: Hopke et al. (2006,
088390)
Period of Study: Washington, DC: Aug
1988-Dec 1997. Phoenix, Arizona: Mar
1995-Jun 1998
Location: Washington, DC and
surrounding counties
Phoenix, Arizona
Reference: Ito et al. (2006, 0883911
Period of Study: Aug 1988-Dec 1997
Location: Washington, DC and
surrounding counties
Design & Methods
Outcome: Mortality:
Total (nonaccidental)
Cardiovascular
Cardiovascular-Respiratory
Study Design: Source-apportionment
Statistical Analyses: Receptor
modeling
Age Groups: All ages
Outcome: Mortality:
Total (nonaccidental)
Cardiovascular
Cardiovascular-Respiratory
Study Design: Time-series
Source-apportionment
Statistical Analyses: Poisson GLM,
natural splines
Age Groups: All ages
Concentrations!
Pollutant:
Source-apportioned PM25:
Washington, DC: Soil
Traffic
Secondary Sulfate
Nitrate
Residual Oil
Wood Smoke
Sea Salt
Incinerator
Primary Coal
Phoenix, Arizona: Crustal
Traffic
Vegetation and Wood Burning
Secondary Sulfate
Metals
Sea Salt
Primary Coal
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant (correlation): NR
Pollutant:
Source-apportioned PM25:
Soil
Traffic
Secondary Sulfate
Nitrate
Residual Oil
Wood Smoke
Sea Salt
Incinerator
Primary Coal
Averaging Time: 24-h avg
Mean (SD): 17.8 (8.7)
Range (Min, Max): NR
Copollutant (correlation): NR
Effect Estimates (95% Cl)
The study does not present quantitative
results
Increment:
PM25 = 28.7|jg/m3
PM25 Sources 5-95th = Not reported
% Increase (Lower Cl, Upper Cl) lag:
Secondary sulfate (variance-weighted
mean percent excess mortality)
6.7% (1.7, 11.7)3
Primary coal-related PM2 5 (mean
percent excess mortality)
5.0% (1.0, 9.1)3
Residual oil (mean percent excess
mortality)
2.7% (-1.1, 6.5) 2
Traffic-related PM25 (mean percent
excess mortality)
2.6% (-1.6, 6.9) NR
Soil-related PM25 (mean percent
excess mortality)
2.1% (-0.8, 4.9) NR
PMz; Sensitivity analysis:
2 df/yr: 7.9% (3.3,12.6)3
4 df/yr: 8.3% (3.7,13.1)3
8 df/yr: 8.3% (3.7,13.2)3
16 df/yr: 8.1% (3.1,13.2) 3
Reference: Kan et al. (2007, 0912671
Period of Study: Mar 2004-Dec 2005
Location: Shanghai, China
Outcome (ICD10): Mortality:
Total (nonaccidental) (AOO-R99)
Cardiovascular (IOO-I99)
Respiratory (JOO-J98)
Study Design: Time-series
Statistical Analyses: Poisson GAM,
penalized splines
Age Groups: All ages
Pollutant: PM25
Averaging Time: 24-h avg
Mean (SD): 52.3 (1.57)
Range (Min, Max): (2.0, 330.3)
Copollutant (correlation):
PM10:r = 0.84
PMi0.25:r = 0.48
03:r = 0.31
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl)
lag:
Total: 0.36% (0.11, 0.61) 0-1
Cardiovascular: 0.41% (0.01, 0.82) 0-1
Respiratory: 0.95% (0.16,1.73)0-1
December 2009
E-349
-------
Study
Reference: Kettunen et al. (2007,
0912421
Period of Study: 1998-2004
Location: Helsinki, Finland
Reference: Klemm et al. (2004,
0565851
Period of Study: Aug 1998-Jul 2000
Location: Fulton and DeKalb counties,
Georgia (ARIES)
Reference: Lippmann et al. (2006,
0911651
Period of Study: 2000-2003
Location: 60 U.S. cities (NMMAPS)
Reference: Mar et al. (2005, 0875661
Period of Study: 1995-1997
Location: Phoenix, Arizona
Design & Methods
Outcome (ICD10): Mortality:
Stroke (160-161, 163-I64)
Study Design: Time-series
Statistical Analyses: Poisson GAM,
penalized thin-plate splines
Age Groups: > 65
Outcome: Mortality:
Nonaccidental (<800)
Cardiovascular (390-459)
Respiratory (460-51 9)
Cancer (140-239)
Study Design: Time-series
Statistical Analyses: Poisson GLM,
natural cubic splines
Age Groups: <65
>65
Outcome: Mortality:
Nonaccidental (<800)
Study Design: Time-series
Statistical Analyses: Poisson GLM
Age Groups: All ages
Outcome: Mortality:
Nonaccidental (<800)
Cardiovascular (390-448)
Study Design: Time-series
Statistical Analyses: Poisson GLM
Age Groups: 2 65
Concentrations!
Pollutant: PM25
Averaging Time: 24-h avg
Median (SD) unit:
Cold Season: 8.2
\Narm Season: 7.8
Range (Min, Max):
Cold Season: (1.1, 69.5)
Warm Season: (1.1, 41.5)
Copollutant:
03
CO
N02
PM,o
PM,0-2.5
UFP
Pollutant: PM25
Averaging Time: 24-h avg
Mean (SD): 19.62 (8.32)
Range (Min, Max): (5.29, 48.01)
Copollutant:
PM,0-2.5
03
N02
CO
S02
Acid
EC
OC
S04
Oxygenated Hydrocarbons
Nonmethane hydrocarbons
N03
Pollutant: Speciated Fine PM:
Al, Ar, Cr, Cu, EC, Fe, Mn, Ni, Nitrate,
OC, Pb, Se, Si, Sulfate, V, Zn
Averaging Time: Annual avg
Mean (SD): R
Range (Min, Max): NR
Pollutant:
Source-apportioned PM25:
Soil
Traffic
Secondary Sulfate
Nitrate
Residual Oil
V\food Smoke
Sea Salt
Incinerator
Primary Coal
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Effect Estimates (95% Cl)
Increment:
Cold Season: 6.7 pg/m3
\Narn Season: 5.7 pg/m3
% Increase (Lower Cl, Upper Cl) lag:
Cold Season
-0.19% (-3.77, 3.51)0
-0.17% (-3.73, 3.52)1
0.59% (-2.95, 4.26 2
0.46% (-3. 10, 4.15 3
\Narm Season
6.86% (0.37, 13.78)0
7.40% (1.33, 13.84)1
4.01% (-1.79, 10.14)2
-1.72% (-7.38, 4.29)3
Increment: NR
p(SE)lag:
Quarterly Knots:
PM25: 0.00398 (0.00161)
0-1
Monthly Knots:
PM25: 0.00544 (0.00184)
0-1
Biweekly Knots:
PM25: 0.00369 (0.00201)
0-1
The study does not present quantitative
results.
Increment: PM2 5 Sources 5-95th = NR
% Increase (median percent excess
risk) lag:
Secondary sulfate: 16.0% 0
Traffic: 13.2% 1
Copper (Cu) smelter: 12.0%0
Sea salt: 10.2% 5
Biomass/wood combustion: 8.6% 3
December 2009
E-350
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Ostro et al. (2006, 0879911
Period of Study: Jan 1999-Dec 2002
Location: 9 California counties
(CALFINE)
Outcome (ICD10): Mortality:
Total mortality (respiratory,
cardiovascular, ischemic heart disease,
diabetes)
Respiratory (JOO-J98)
Cardiovascular (IOO-I99)
Ischemic heart disease (I20-I25)
Diabetes (E10-E14)
Study Design: Time-series
Statistical Analyses: Poisson, natural
splines and penalized splines
Age Groups: All ages
>65yr
Pollutant: PM25
Averaging Time: 24-h avg
Mean (SD):
Contra Costa: 14
Fresno: 23
Kern: 22
Los Angeles: 21
Orange: 21
Riverside: 29
Sacramento: 14
Santa Clara: 15
San Diego: 16
Range (Min, Max):
Contra Costa: (1,77)
Fresno: (1,160)
Kern: (1,155)
Los Angeles: (4, 85)
Orange: (4,114)
Riverside: (2,120)
Sacramento: (1,108)
Santa Clara: (2, 74)
San Diego: (0, 66)
Copollutant (correlation):
N02 r = 0.56
CO r = 0.60
03(1h) r = -0.14
03(8h) r = -0.22
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl) lag:
Penalized splines
All ages:
All-cause:
0.2% (-0.2, 0.7) 2
0.6% (0.2, 1.0)0-1
Cardiovascular:
0.3% (-0.1, 0.7) 2
0.6% (0.0, 1.1)0-1
Respiratory:
1.3%
2.2%
0.1,2.6)2
0.6, 3.9) 0-1
>65:
All-cause:
0.2% (-0.2, 0.7) 2
0.7% (0.2, 1.1)0-1
Ischemic heart disease:
0.3% (-0.5, 1.0)0-1
Males: 0.5% (-0.2,1.2)0-1
Females: 0.8% (0.3,1.3)0-1
Whites: 0.8% (0.2,1.3)0-1
Blacks: 0.1% (-0.9,1.2)0-1
Hispanics:0.8%(-0.1,1.6)0-1
In hospital: 0.6% (-0.1,1.3)0-1
Out of hospital: 0.6% (0.1,1.1)0-1
High school graduates:
0.4% (0.0, 0.8) 0-1
Non-high school graduates:
0.9% (-0.1,1.9) 0-1
Natural splines
All cause
4 df: 0.5% (-0.1,1.1)0-1
8 df: 0.4% (-0.1, 0.9) 0-1
12 df: 0.3% (-0.1, 0.7) 0-1
Cardiovascular
4 df: 0.4% (-0.2, 0.9) 0-1
8 df: 0.1% (-0.5, 0.6)0-1
12 df: 0.0% (-0.6, 0.6) 0-1
Respiratory
4 df: 2.1% (0.2, 4.1)0-1
8 df: 1.6% (-0.5, 3.6)0-1
12 df: 1.3% (-0.3, 2.9)0-1
>65
All cause
4 df: 0.7% (0.0, 1.3)0-1
8 df: 0.4% (-0.1, 0.9) 0-1
12 df: 0.3% (-0.1, 0.8) 0-1
December 2009
E-351
-------
Study
Reference: Ostro et al. (2007, 0913541
Period of Study: PM2 5 speciation
analysis: Jan 2000-Dec 2003. PM25
analysis: Jan 1999-Dec 2003
Location: 6 California counties
(2000-2003). 9 California counties
(1999-2003) (CALFINE)
Reference: Ostro et al. (2008, 0979711
Period of Study: Jan 2000-Dec 2003
Location: 6 California counties
Reference: Perez et al. (2008, 1560201
Period of Study: Mar 2003-Dec 2005
Location: Barcelona, Spain
Design & Methods
Outcome (ICD10): Mortality:
Total (nonaccidental) mortality
Respiratory (JOO-J98)
Cardiovascular (IOO-I99)
Study Design: Time-series
Statistical Analyses: Poisson, natural
splines
Age Groups: >65
Outcome (ICD10): Mortality:
Cardiovascular (IOO-I99)
Study Design: Time-series
Statistical Analyses: Poisson, natural
cubic splines and natural splines
Age Groups:
Outcome: Respiratory mortality
Study Design: Cohort
Covariates: Temperature, humidity
Statistical Analysis: Autoregressive
Poisson regression models
Concentrations!
Pollutant: PM25
Averaging Time: 24-h avg
Mean (SD):
2000-2003: 19.28
1999-2003: 18.6
Range (Min, Max): NR
Copollutant (correlation):
EC: r = 0.53
OC:r = 0.62
N03:r = 0.65
S04:r = 0.32
Al:r = 0.02
Br: r = 0.54
Ca:r = 0.23
Cl:r=0.15
Cu:r = 0.23
Fe:r = 0.38
K:r = 0.52
Mn:r = 0.21
Ni:r = 0.11
Pb:r = 0.27
S:r = 0.35
Si:r = 0.16
Ti:r = 0.24
V:r = 0.20
Zn:r = 0.51
Pollutant: PM25, EC, OC, N03, S04,
Ca, Cl, Cu, Fe, K, S, Si, Ti, Zn
Averaging Time: 24-h avg
Mean (SD):
PM25: 19.28
EC: 0.966
OC:7.129
N03: 5.415
S04: 1.908
Ca: 0.080
Cl: 0.094
Cu: 0.007
Fe:0.124
K: 0.117
S: 0.648
Si: 0.168
Ti: 0.009
Zn: 0.012
Range (96th): PM25: 46.91
EC: 2.57
OC: 15.91
N03: 17.46
S04:5.18
Ca: 0.20
Cl: 0.41
Cu: 0.02
Fe: 0.34
K: 0.26
S:1.70
Si: 0.43
Ti: 0.02
Zn: 0.04
Pollutant: PM25-i
Averaging Time: 24 h
Mean (SD) Unit: 5.5 (3.8) pg/m3
Range (Min, Max): 0.6, 45.5
Copollutant: PM10.2.5, PM,
Effect Estimates (95% Cl)
Increment: 14.6 pg/m3
% Increase (Lower Cl, Upper Cl)
lag:
Cardiovascular
1.6% (0.0, 3.1)
3
Notes: The study does not present all
estimates quantitatively.
The study does not present quantitative
results.
Increment: 10|jg/m3
Odds Ratio (96%CI) lag
Avg 10-1:0.998 (0.849-1. 174),
p = 0.981
L1: 1.014 (0.886-1. 161), p = 0.838
12:1.295(1.141-1.470), p = 0.000
Statistical Package: NR
Age Groups: All deaths
Iti-pollutant Model
Avg 10-1:0.987 (0.806-1.208),
p = 0.898
11:1.022(0.859-1.214), p = 0.
12:1.206(1.028-1.416), p = 0.022
December 2009
E-352
-------
Study
Reference: Perez et al. (2008, 1560201
Period of Study: Mar 2003-Dec 2005
Location: Barcelona, Spain
Reference: Perez et al. (2008, 1560201
Period of Study: Mar 2003-Dec 2005
Location: Barcelona, Spain
Reference: Rainham et al. (2005,
0886761
Period of Study: 1981-1999
Location: Toronto Canada
Design & Methods
Outcome: Cardiovascular mortality
Study Design: Cohort
Covariates: Temperature, humidity
Statistical Analysis: Autoregressive
Poisson regression models
Statistical Package: NR
Age Groups: All deaths
Outcome: Cerebrovascular mortality
Study Design: Cohort
Covariates: Temperature, humidity
Statistical Analysis: Autoregressive
Poisson regression models
Statistical Package: NR
Age Groups: All deaths
Outcome: Mortality:
Total (nonaccidental) (<800)
Cardiorespiratory (390-459
480-519)
Other-causes
Study Design: Time-series
Statistical Analyses: Poisson GLM,
natural splines
Age Groups: All ages
Concentrations!
Pollutant: PM25.i
Averaging Time: 24 h
Mean (SD) Unit: 5.5 (3.8) pg/m3
Range (Min, Max): 0.6, 45.5
Copollutant: PM10.2.5, PM,
Pollutant: PM2.5.,
Averaging Time: 24 h
Mean (SD) Unit: 5.5 (3.8) pg/m3
Range (Min, Max): 0.6, 45.5
Copollutant: PM10.2.5, PM,
Pollutant: PM25
Averaging Time: 24-h avg
Mean (SD):
All yr: 17.0 (8.7)
Winters (Dec-Feb): 17.2 (6.8)
Summers (Jun-Aug): 18.8 (10.2)
Range (Min, Max): NR
Copollutant:
CO
N02
S02
03
Effect Estimates (95% Cl)
Increment: 10|jg/m3
Odds Ratio (96%CI) lag
Avg 10-1:1.100 (1.002-1.207),
p = 0.046
11:1.112(1.031-1.200), p = 0.006
12:1.078(0.999-1.163), p = 0.052
Multi-pollutant Model
Avg 10-1:0.994 (0.885-1. 116),
p = 0.920
L1:0.984 0.892-1.086), p = 0.754
L2: 0.981 0.891-1.079), p = 0.688
Increment: 10|jg/m3
Odds Ratio (96%CI) lag
Avg 10-1:1.083 (0.897-1.307),
p = 0.406
L1: 1.121 0.964-1.303), p = 0.140
L2:0.984 0.841-1.152), p = 0.839
Multi-pollutant Model
Avg 10-1:0.899 (0.712-1. 135),
p = 0.371
L1: 0.905 0.743-1.102), p = 0.321
L2:0.868 0.711-1.060), p = 0.165
Increment: NR
% Increase (Lower Cl, Upper Cl) lag:
Winter and Wnter Synoptic Events
\A/]ni.r
winter
Total: 0.998% (0.997, 1.000)2
Cardiorespiratory:
0.998(0.996,1.000)2
Other: 0.998% (0.996, 1.000)2
Dry Moderate
Total: 1.001% (0.996, 1.007)1
Cardiorespiratory:
1.005(0.998,1.011)1
Other: 0.997% (0.989, 1.006)0
Dry Polar
Total: 0.998% (0.995, 1.001)2
Cardiorespiratory:
0.995(0.991,0.999)2
Other: 1.002% (0.998,1.005)1
Moist Moderate
Total: 0.998% (0.993, 1.002)2
Cardiorespiratory:
1.003(0.995,1.010)1
Other: 0.997% (0.991,1.004)1
Moist Polar
Total: 1.001% (0.998, 1.005)1
Cardiorespiratory:
1.002(0.997,1.007)2
Other: 1.003% (0.999,1.007)0
Moist Tropical
Total: 1.007% (0.965, 1.203)0
Cardiorespiratory:
1.123(1.031,1.224)2
Other: 1.248% (1.123, 1.387)0
Transition
Total: 1.003% (0.996, 1.009)1
Cardiorespiratory:
0.996(0.987,1.004)0
Other: 0.997% (0.990, 1.004)0
Summer and summer Synoptic Events
Summer
Total: 1.000% (1.000,1.001)0
Cardiorespiratory:
1.001 (1.000, 1.002)0
December 2009
E-353
-------
Study Design & Methods Concentrations! Effect Estimates (95% Cl)
Other: 1.001% (1.000,1.002)0
Dry Moderate
Total: 1.001% (0.999, 1.002)2
Cardiorespiratory:
1.002(0.999,1.004)2
Other: 0.999% (0.997, 1.002)0
Dry Polar
Total: 1.002% (0.999, 1.005)2
Cardiorespiratory:
0.996(0.991,1.000)0
Other: 1.003% (0.999, 1.007)2
Dry Tropical
Total: 1.016% (1.006, 1.027)0
Cardiorespiratory:
1.017(1.005,1.030)2
Other: 1.017% (1.003,1.031)0
Moist Moderate
Total: 1.002% (1.000, 1.004)2
Cardiorespiratory:
1.003(0.999,1.006)2
Other: 1.004% (1.001,1.006)0
Moist Polar
Total: 1.005% (0.998, 1.011)1
Cardiorespiratory:
1.008(0.997,1.018)0
Other: 1.003% (0.995,1.011)1
Moist Tropical
Total: 0.999% (0.997, 1.001)2
Cardiorespiratory:
0.996(0.993,1.000)2
Other: 0.998% (0.995,1.001)1
Transition
Total: 1.005% (0.996, 1.014)1
Cardiorespiratory:
1.007(0.994,1.020)1
Other: 1.002% (0.996, 1.008)2
December 2009 E-354
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Rosenthal et al. (2008,
1569251
Period of Study: Jul 2002-Jul 2006
Location: Indianapolis, Indiana
Outcome: Non-Dead on Arrival (DOA) Pollutant: PM
Out-of-Hospital Cardiac Arrests (OHCA)
Witnessed non-DOAOHCA
Study Design: Case-crossover
Statistical Analyses: Time-stratified
conditional logistic regression
Age Groups: All ages
Study Population: Non-DOAOHCA:
1,374
Witnessed non-DOAOHCA: 511
Averaging Time: 24-h avg
Hourly
Mean (SD):
NR
IQR (25th, 75th):
All non-DOA
All heart rhythms: (9.4,19.5)
OHCA: (9.6, 19.5)
Referents: (9.3,19.5)
Asystole: (9.2,19.4)
OHCA: (9.2, 19.7)
Asystole: (9.2,19.2)
Wtnessed non-DOA hourly
All heart rhythms: (8.8, 20.7)
OHCA: (8.8, 21.9)
Referents: (8.8, 20.4)
Asystole: (8.5,19.8)
OHCA: (9.4, 21.3)
Referents: (8.3,19.1)
Copollutant (correlation): NR
Increment: 10|jg/m
Hazard Ratio (Lower Cl, Upper Cl)
lag:
Out-of-Hospital non-DOA Cardiac
Arrests
1.02(0.94,
1.00(0.92,
0.98 (0.90,
1.00(0.92,
1.02(0.92,
1.01 (0.91,
1.02(0.91,
Asystole
1.03(0.91,
1.00(0.89,
1.01 (0.90,
0.98 (0.87,
1.03(0.90,
1.05(0.90,
1.04(0.88,
Vfib
1.08(0.92,
1.02(0.87,
0.96 (0.80,
1.10(0.93,
1.06(0.88,
1.01 (0.82,
1.05(0.83,
PEA
0.92 (0.77,
0.98 (0.83,
0.96 (0.82,
0.95 (0.82,
0.96 (0.80,
0.98 (0.80,
0.98 (0.78,
1.11)0
1.08)1
1.06)2
1.08)3
1.12)0-1 avg
1.12) 0-2 avg
1.14) 0-3 avg
1.17)0
1.13)1
1.13)2
1.10)3
1.18)0-1 avg
1.22) 0-2 avg
1.22) 0-3 avg
1.28)0
1.21)1
1.14)2
1.31)3
1.28)0-1 avg
1.25) 0-2 avg
1.32) 0-3 avg
1.0
1.15)1
1.14)2
1.10)3
1.17)0-1 avg
1.21) 0-2 avg
1.21) 0-3 avg
Wtnessed Out-of-Hospital non-DOA
Cardiac Arrests (lag represents h in
which or h before OHCA occurred)
All: 1.12 (1.01, 1.25)0
White: 1.18 (1.03,1.35)0
60-75:1.25(1.05, 1.49)0
Asystole: 1.22 (1.01,1.59)0
Reference: Schwartz et al. (2002,
0253121
Period of Study: 1979-Late 1980's
Location: 6 U.S. cities
Outcome: Mortality:
Total (nonaccidental) (<800)
Study Design: Time-series
Statistical Analyses: Hierarchical
modeling:
LPoisson GAM, LOESS
2. Multivariate modeling
Age Groups: All ages
Pollutant: PM25, PM25 sources (Traffic,
Coal, Residual Oil)
Averaging Time: 24-h avg
Mean (SD):
PM25 Range: (Madison: 11.3 to
Steubenville: 30.5)
Traffic Range: (Steubenville: 1.5 to
Boston: 4.8)
Coal Range: (Madison: 4.9 to
Steubenville: 19.2)
Residual Oil Range: (Boston: 0.5 to
Steubenville: 0.9)
Range (Min, Max): NR
The study does not present quantitative
results.
December 2009
E-355
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Simpson et al. (2005,
0874381
Period of Study: Jan 1996-Dec 1999
Location: 4 Australian cities
Outcome: Mortality:
Nonaccidental (<800)
Cardiovascular (390-459)
Respiratory (460-519)
Study Design: Time-series
meta-analysis
Statistical Analyses: Poisson GAM,
natural splines
Poisson GLM, natural splines
Age Groups: All ages
Pollutant: PM25
Averaging Time: 24-h avg
Mean (SD):
Brisbane: PM25: 7.50
Sydney: PM25:9.00
Melbourne: PM25: 9.30
Perth: PM25: 9.0 pg/m3
Range (Min, Max):
Brisbane: PM25: (1.9,19.7)
Sydney: PM25: (2.4, 35.3)
Melbourne: PM25: (2.7, 35.1)
Perth: PM25: (2.8, 37.3)
Copollutant: CO, N02
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl) lag:
PM25
0.9% (-0.7, 2.5)
Reference: Slaughter et al. (2005,
0738541
Period of Study: Jan 1995-Dec 1999
Location: Spokane, Washington
Reference: Stieb et al. (2002, 0252051
Period of Study: Publication dates of
studies: 1985-Dec 2000 Mortality
series: 1958-1999
Location: 40 cities (11 Canadian cities,
19 U.S. cities, Santiago, Amsterdam,
Erfurt, 7 Korean cities)
Reference: Sullivan et al. (2003,
0431561
Period of Study: 1985-1994
Location: Western Washington
Outcome: Mortality:
Nonaccidental (<800)
Study Design: Time-series
Statistical Analyses: Poisson GLM,
natural splines
Age Groups: All ages
Outcome: Mortality:
All-cause (nonaccidental)
Study Design: Meta-analysis
Statistical Analyses: Random effects
model
Age Groups: All ages
Outcome: Out-of-hospital cardiac
arrest
Study Design: Case-crossover
Statistical Analyses: Conditional
logistic regression
Age Groups: 19-79
Study Population: Out-of-hospital
cardiac arrests: 1,206
Pollutant: PM25
Averaging Time: 24-h avg
Mean (SD): NR
Range (9th, 95th): PM25: (4.2, 20.2)
Copollutant (correlation):
PM25:r = 0.95
PM10:r = 0.62
PMi0.25:r = 0.31
CO: r = 0.62
Pollutant: PM25
Averaging Time: NR
Mean (SD): NR
Range (Min, Max): NR
Copollutant: Varied between studies:
03
S02
N02
CO
Pollutant: PM25
Averaging Time: 24-h avg
Median (SD) unit:
PM,o
Lag 0: 28.05
Lag 1:27.97
Lag 2: 28.40
Range (Min, Max): PM10: (7.38, 89.83)
Copollutant (correlation): S02, CO
Notes: Study used nephelometry to
measure particles and equated the
measurements to PM2 5 concentrations.
Increment:
PM25: 10|jg/m3
PM10: 25 pg/m3
Relative Risk (Lower Cl, Upper Cl)
lag:
PM25
(0.97,1.04)1
0.99(0.96, 1.03)2
1.00(0.97, 1.03)3
Increment: PM25: 18.3 pg/m3
% Increase (Lower Cl, Upper Cl) lag:
Single-pollutant models
18 studies
PM25:2.0%(1.2, 2.7)
Multipollutant models
8 studies
PM25:1.3%(0.6, 1.9)
Increment:
PM10: 16.51 pg/m3
PM25: 13.8 pg/m3
Odds Ratio (Lower Cl, Upper Cl) lag:
Overall
PM10
1.05(0.87, 1.27)0
0.91 (0.75, 1.11) 1
1.03(0.82,1.28)2
PM25
0.94(0.88,1.01)0
0.94(0.88,1.02)1
1.00(0.93, 1.08)2
PM2 5: Stratified by subject
characteristics
<55
0.95(0.76,1.18)0
0.89(0.71,1.12)1
0.95(0.75, 1.20)2
>55
0.94(0.88,1.02)0
0.95, (0.88, 1.03)1
1.01(0.93,1.10)2
Male
0.95(0.87, 1.03)0
0.96(0.88,1.04)1
1.01(0.93,1.10)2
Female
0.93(0.82,1.06)0
0.92(0.80,1.07)1
0.98(0.83, 1.15)2
White
December 2009
E-356
-------
Study Design & Methods Concentrations! Effect Estimates (95% Cl)
0.93(0.86, 1.01)0
0.95(0.88, 1.03)1
1.03(0.95,1.12)2
Non-White
1.09(0.88, 1.36)0
0.96(075,1.22)1
0.88(0.68,1.14)2
Current Smoker
1.05(0.92,1.19)0
0.98(0.86,1.12)1
1.06(0.92, 1.22)2
Nonsmoker
0.93(0.85,1.01)0
0.93(0.85, 1.02)1
0.97(0.89,1.07)2
Drinker
1.13(0.92, 1.39)0
1.15(0.94, 1.41)1
1.16(0.92,1.45)2
Nondrinker
0.94(0.86,1.03)0
0.93(0.85,1.02)1
1.00(0.92, 1.10)2
Activity Level-Unrestricted
0.96(0.89,1.03)0
0.96(0.89, 1.04)1
1.01(0.93,1.10)2
Activity Level-Limited
0.82(0.56, 1.20)0
0.70(0.45,1.09)1
0.97(0.65,1.43)2
PM25: Stratified by disease state
Heart disease
0.95(0.87, 1.04)0
0.97(0.89, 1.07)1
1.06(0.96,1.16)2
Ischemic Heart Disease
0.91 (0.80, 1.04)0
0.97(0.84, 1.11)1
1.09(0.95,1.26)2
Active Angina
0.98(0.81,1.20)0
1.07(0.88,1.31)1
1.08(0.89, 1.32)2
Congestive Heart Failure
0.91(0.80,1.03)0
0.99(0.87, 1.13)1
1.11(0.97,1.26)2
Supraventricular tachycardia
1.41 (0.97,2.04)0
1.55(1.07,2.25)1
1.23(0.84,1.82)2
Bradycardia
0.97(0.64,1.46)0
1.29(0.85,1.96)1
1.30(0.84,2.01)2
Asthma
(0.80,1.27)0
0.92(0.71, 1.19)1
0.93(0.71,1.22)2
COPD
1.00(0.86, 1.17)0
1.04(0.88,1.23)1
1.08(0.92,1.28)2
PM2 5: Persons with prior recognized
heart disease stratified by smoking
status
All heart disease
Current smoker
1.08(0.92, 1.26)0
1.06(0.89,1.26)1
1.29(1.06,1.55)2
Nonsmoker
0.91(0.82,1.02)0
0.94(0.84, 1.05)1
0.99(0.88, 1.11)2
Ischemic Heart Disease
December 2009 E-357
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Current smoker
1.06(0.84,1.34)0
0.99(075,1.30)1
1.39(1.04,1.86)2
Nonsmoker
0.86(073,1.02)0
0.93(078,1.11)1
0.99(0.83, 1.18)2
Active Angina
Current smoker
1.28(0.88, 1.86)0
1.26(079,2.01)1
1.57(0.99,2.48)2
Nonsmoker
0.87(0.68,1.12)0
0.93(072,1.21)1
0.91 (0.70, 1.17)2
Congestive Heart Failure
Current smoker
1.00(079, 1.28)0
1.03(078,1.35)1
1.46(1.10,1.96)2
Nonsmoker
0.88(076,1.03)0
0.96(0.82,1.12)1
0.99(0.84, 1.17)2
Supraventricular tachycardia
Current smoker
12.80(1.05, 156.57)0
2.56 (0.82, 7.99) 1
1.15(0.46,2.86)2
Nonsmoker
1.19(074,1.90)0
1.35(0.87,2.10)1
1.15(073, 1.82)2
Bradycardia
Nonsmoker
0.84(0.14,4.95)0
0.42 (0.03, 5.34) 1
0.51 (0.05, 5.79) 2
Nonsmoker
0.99(0.63,1.55)0
1.42(0.90,2.24)1
1.39(0.88,2.20)2
Reference: Thurston et al. (2005,
0979491
Period of Study: Washington, DC: Aug
1988-Dec 1997. Phoenix, Arizona:
1995-1997
Location: Washington, DC and
surrounding counties
Phoenix, Arizona
Outcome: Mortality:
Total (nonaccidental) (<800)
Cardiovascular (390-448)
Study Design: Time-series
Source-apportionment
Statistical Analyses: Poisson GLM,
natural splines
Age Groups: Washington, DC: All ages
Phoenix, Arizona: 2 65
Pollutant:
PM25, and source apportioned PM25:
Crustal
Traffic
Secondary S04
Secondary N03
Wood
Oil
Salt
Incinerator
Averaging Time: 24-h avg
Median (SD) unit: NR
Range (Min, Max): NR
Increment: 10|jg/m3
% Increase:
Total (nonaccidental):
Secondary sulfate:
Phoenix: 5.2%
Washington, DC: 3.8%
Motor vehicles:
Phoenix: 0.9%
Washington, DC: 4.2%
Copollutant: PM25 species (Na, Mg, Al,
Si, P, S, Cl, K, Ca, Sc, Ti, V, Cr, Mn, Fe,
Co, Ni, Cu, Zn, Ga, Ge, As, Se, Br, Rb,
Sr, Y, Zr, Mo, Rh, Pd, Ag, Cd, Sn, Sb,
Te, I, Cs, Ba, La, W, Au, Hg, Pb, OC,
EC)
December 2009
E-358
-------
Study
Design & Methods
Concentrations!
Effect Estimates (95% Cl)
Reference: Villeneuve et al. (2003,
0550511
Period of Study: 1986-1999
Location: Vancouver, Canada
Outcome: Mortality:
Nonaccidental (<800)
Cardiovascular (401-440)
Respiratory (460-519)
Cancer (140-239)
Study Design: Time-series
Statistical Analyses: Poisson, natural
splines
Age Groups: 2 65
Pollutant: PM25
Averaging Time: 24-h avg
Mean (SD):
Daily
PM25:7.9
Every 6th Day
PM25:11.6
Range (Min, Max):
Daily
PM25: (2.0, 32.0)
Every 6th Day
PM25: (1.8, 43.0)
Copollutant:
S02
CO
N02
03
Increment:
PM25(Daily):9.0|jg/m3
PM25(6thDay):15.7|jg/m3
% Increase (Lower Cl, Upper Cl) lag:
Nonaccidental
PM25(Daily)
-0.1% (-5.1,5.2) 0-2 avg
-0.1% (-4.1,4.1 0
-0.3% (-4.2, 3.7 1
0.5% (-3.3, 4.4) 2
PM25(6thDay)
-2.8% (-7.5, 2.1)0
2.0% (-2.6, 7.0) 1
4.5% (-0.3, 9.5) 2
Cardiovascular
PM25(Daily)
1.5% (-6.1,9.7) 0-2 avg
4.3% (-1.7, 10.7)0
-1.0% (-7.0, 5.4 1
-0.5% (-6.5, 5.9 2
PM25(6thDay)
-1.5% (-8.9, 6.5 0
-2.0% (-9.3, 5.8 1
3.0% (-4.2, 10.8)2
Respiratory
PM25(Daily)
-0.7% (-13.1, 13.4) 0-2 avg
6.7% (-3.7, 18.3)0
-3.0% (-12.8, 7.9)1
-5.8% (-15.2, 4.7)2
PM25(6thDay)
10.0% (-4.7, 26.8) 0
8.3% (-5.4, 24.0) 1
0.3% (-12.4, 14.9)2
Cancer
PM25(Daily)
-0.3% (-9.4, 9.8) 0-2 avg
-4.5% (-11.2, 2.8)0
2.7%
2.5%
-5.0, 11.0)1
-5.1,10.7)2
PM25(6thDay)
-5.1% (-13.8, 4.5)0
-0.3% (-9.7, 11.0)1
0.2% (-9.1, 10.4) 2
Reference: Wilson et al. (2007,
1571491
Period of Study: 1995-1997
Location: Phoenix, Arizona
Outcome: Cardiovascular
Study Design: Time-series
Statistical Analyses: Poisson GAM,
nonparametric smoothing spline
Age Groups: >25
Pollutant: PM25
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: 10|jg/m3
% Excess Risk (Lower Cl, Upper Cl)
lag:
Central Phoenix:
11. 5% (2. 8, 20. 9) 0-5 ma
6.6% (1.1, 12.5)1
2.0% (-3.2, 7.5) 2
Middle Phoenix:
2.9%
6.4%
-4.9, 11. 4) 0-5 ma
1.1, 11.9)2
Outer Phoenix:
1.6% (-6.2, 10.0) 0-5 ma
1AII units expressed in ug/m3 unless otherwise specified.
December 2009
E-359
-------
Table E-19. Short-term exposure-mortality - other PM size fractions.
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Perez et al. (2008,1560201 Outcome: Respiratory mortality
Period of Study: Mar 2003-Dec 2005 Study Design: Cohort
Location: Barcelona, Spain Covariates: Temperature, humidity
Statistical Analysis: Autoregressive
Poisson regression models
Statistical Package: NR
Age Groups: All deaths
Pollutant: PMi
Averaging Time: 24 h
Mean (SD) Unit: 20.0 (10.3) pg/m3
Range (Min, Max): 1.9, 80.1
Copollutant: PM10.2.5, PM2.5.i
Increment: 10|jg/m
Odds Ratio (96%CI) lag
Avg 10-1:1.005 (0.960-1.053),
p = 0.824
11:1.012(0.969-1.056), p = 0.599
L2:1.042 (0.998-1.087), p = 0.063
Multi-pollutant Model
Avg 10-1:1.007 (0.957-1.059),
p = 0.799
11:1.008(0.961-1.058), p = 0.739
L2:1.010 (0.963-1059), p = 0.678
Reference: Perez et al. (2008,1560201 Outcome: Cardiovascular mortality
Period of Study: Mar 2003-Dec 2005 Study Design: Cohort
Location: Barcelona, Spain Covariates: temperature, Humidity
Statistical Analysis: Autoregressive
Poisson regression models
Statistical Package: NR
Age Groups: All deaths
Pollutant: PMi
Averaging Time: 24 h
Mean (SD) Unit: 20.0 (10.3) pg/m3
Range (Min, Max): 1.9, 80.1
Copollutant: PM10.2.5, PM2.5.i
Increment: 10|jg/m
Odds Ratio (96%CI) lag
Avg 10-1:1.028 (1.000-1.057),
p = 0.054
11:1.029(1.003-1.056), p = 0.030
L2:1.023 (0.996-1.050), p = 0.091
Multi-pollutant Model
Avg 10-1:1.025 (0.995-1.057),
p = 0.688
11:1.028(1.000-1.058), p = 0.053
L2:1.024 (0.995-1053), p = 0.110
Reference: Perez et al. (2008,1560201 Outcome: Cerebrovascular mortality
Period of Study: Mar 2003-Dec 2005 Study Design: Cohort
Location: Barcelona, Spain
Covariates: Temperature, humidity
Statistical Analysis: Autoregressive
Poisson regression models
Statistical Package: NR
Age Groups: All deaths
Pollutant: PM,
Averaging Time: 24 h
Mean (SD) Unit: 20.0 (10.3) pg/m3
Range (Min, Max): 1.9, 80.1
Copollutant: PM10.2.5, PM2.5.,
Increment: 10|jg/m
Odds Ratio (96%CI) lag
Avg 10-1:1037(0.981-1097),
p = 0.202
L1:1.056 (1.003-1.113), p = 0.039
L2:1.020 (0.968-1.075), p = 0.460
Multi-pollutant Model
Avg 10-11042(0.981-1.107),
p = 0.179
L1:1.063 (1.004-1.124), p = 0.035
L2:1.034 (0.976-1095), p = 0.255
Reference: Slaughter et al. (2005,
0738541
Period of Study: Jan 995-Dec 1999
Location: Spokane, Washington
Outcome: Mortality: Nonaccidental
(<800)
Study Design: Time-series
Statistical Analyses: Poisson GLM,
natural splines
Age Groups: All ages
Pollutant: PM,
Averaging Time: 24-h avg
Mean (SD): NR
Range (9th, 95th)
PM,: (3.3, 17.6)
Copollutant (correlation):
PM,
PM25:r = 0.95
PM10:r = 0.50
PM,0.25:r = 0.19
CO: r = 0.63
This study does not present quantitative
results for PM,.
Reference: Stolzel et al. (2007, 0913741
Period of Study: Sept 1995-Aug 2001
Location: Erfurt, Germany
Outcome: Mortality:
Total (nonaccidental) (<800)
Cardio-respiratory (390-459, 460-519,
785, 786)
Study Design: Time-series
Statistical Analyses: Poisson GAM
Age Groups: All ages
Pollutant: MC0.,.o.5, MC0.o,-2.5
Averaging Time: 24-h avg
Mean (SD):
MC01.o5: 17.6 (14.8)
MC00,-25: 22.3 (19.2)
IQR (25th, 75th):
MCo,.o5: (8.4, 21.5)
MCo.oMs: (10.5, 27.3)
Copollutant (correlation):
MCo,-o5
NO: r = 0.52
N02:r = 0.60
CO: r = 0.58
MCo.o,-2.s
NO: r = 0.51
N02:r = 0.58
CO: r = 0.57
Increment:
MQn-os: 13.1 pg/m3
MQun.2.5: 16.8 pg/m3
Relative Risk (Lower Cl, Upper Cl)
lag:
Total (nonaccidental)
MCo.,.0.5
1010(0.986
1.034)
0
1006(0.983
1.029)
1
1007(0.985
1.029)
2
0.994 (0.973
1.016)
3
December 2009
E-360
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
1.002(0.981
1.023)
4
0.997 (0.976
1.018)
0
MCooi-25
1.007(0.985
1.030)
0
1.005(0.984
1.026)
1
1.003(0.983
1.023)
2
0.989 (0.970
1.009)
3
1.002(0.982
1.022)
4
0.998 (0.979
1.018)
5
Cardio-respiratory
MCO.1-0.5
1.004(0.977
1.031)
0
1.004(0.979
1.029)
1.001 (0.978
1.026)
2
0.991 (0.967
1.014)
1.000(0.977
1.023)
4
1.000(0.976
1.023)
MCO.01-2.5
1.001 (0.977
1.026)
0
0.999 (0.976
1.022)
1
0.998 (0.976
1.021)
2
0.985 (0.964
1.007)
1.001 (0.980
1.022)
4
1.003(0.981
1.024)
0
December 2009 E-361
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Yamazaki et al. (2007,
0907481
Period of Study: 1995-1998
Location: Hong Kong, China
Outcome: Mortality:
Intracerebral hemorrhage (431)
Ischaemic stroke (434)
Study Design: Time-stratified case-
crossover
Statistical Analyses: Conditional
logistic regression
Age Groups: 2 65
Pollutant: PM7
Averaging Time: 1-h avg
Mean (SD):
Warmer Months (Apr-Sep):
40.3
Colder Months (Oct-Mar):
39.4
Range (Min, Max): NR
Copollutant (correlation):
Warmer Months
N02:r = 0.46-0.63
Ox: r = -0.14 to 0.20
Colder Months
N02: 0.42-0.79
Ox: r = -0.36 to -0.14
Increment: 30 pg/m
Odds Ratio (Lower Cl, Upper Cl) lag:
24-h avg concentrations
Intracerebral hemorrhage
Warmer months: 1.041 (0.984,1.102)0
Colder months: 1.005(0.951,1.061)0
Ischaemic stroke
Warmer months: 1.027 (0.993,1.062)0
Colder months: 1.005 (0.973,1.039) 0
Exposure measured jointly as 24-h and
1-h mean concentrations
Warmer months
Intracerebral hemorrhage
1-h with 200 pg/m3 threshold:
2.397(1.476, 3.892) 2 h
24-h: 1.019 (0.960, 1.082)0
Ischaemic stroke
1-h with 200 pg/m3 threshold:
1.051 (0.750, 1.472) 2 h
24-h: 1.018 (0.983, 1.055)0
Warmer months
Intracerebral hemorrhage
1-h with 200 pg/m3 threshold:
0.970(0.712, 1.322) 2 h
24-h: 1.015 (0.958, 1.075)0
Ischaemic stroke
1-h with 200 pg/m threshold:
1.040(0.855, 1.265) 2 h
24-h: 1.003 (0.968, 1.039)0
1AII units expressed in ug/m3 unless otherwise specified.
December 2009
E-362
-------
E.4. Long-Term Exposure and Cardiovascular Outcomes
Table E-20. Long-term exposure - cardiovascular morbidity outcomes - PM10.
Study
Design & Methods
Concentrations
Effect Estimates (95% Cl)
Reference: Baccarelli et al. (2008,
1579841
Period of Study: 1995-2005
Location: Italy (Lombardy region)
Outcome (ICD9 and ICD10): Deep
Vein Thrombosis (DVT)
Prothrombin time (PT)
Activated partial thromboplastin time
(aPTT)
Age Groups: 18-84yrs
Study Design: Case-control (DVT
outcome)
Cross-sectional (PT and aPTT
outcomes)
N: 871 cases
1210 controls (randomly selected from
friends and nonblood relatives of cases
Frequency matched by age to cases)
Statistical Analyses: Unconditional
logistic regression (DVT outcome)
Linear regression (PT and aPTT
outcomes)
Covariates: Sex, area of residence,
education, factor V Leiden or G20210A
prothrombin mutation, current use of
oral contraceptives or hormone therapy
(Variables controlled using penalized
regression splines with 4 df) age, BMI,
day of yr (for seasonality), index date,
ambient temperature
Season: covariate
Dose-response Investigated? Yes
Statistical Package: STATAv9.0 and R
V2.2.0
Pollutant: PM10
Averaging Time: 1 yr (immediately
preceding the diagnosis date for cases
or the date of examination for controls)
assessed other averaging periods
presented in supplements (90 days, 180
days, 270 days, 2 yr)
Mean (SD): NR
Percentiles: NR
Range (Min, Max):
Range for tertiles of exposure:
1:12.0-44.2
2:44.3-48.1
3:48.2-51.5
Monitoring Stations: Monitors from 53
sites
exposure assigned by dividing area into
9 regions
Copollutant (correlation): NR
PM Increment: 10 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
Estimated changes of PT associated
with PMio:
Among DVT cases: -0.12 (-0.23, 0.00),
p = 0.04
Among Controls: -0.06 (-0.11, 0.00),
p = 0.04
Estimated changes of a PIT
associated with PMio:
Among Controls: -0.09 (-0.19, 0.01),
p = 0.07
Among DVT cases: 0.01 (-0.03, 0.04),
p = 0.78
Risk of DVT associated with PMio
(avg of 1 yr preceding
diagnosis/exam date): |
All subjects:
1.70(1.30, 2.23), p< 0.001
Sex:
Male: 2.07 (1.50, 2.84), p< 0.001
Female: 1.40 (1.02,1.92),p = 0.04
P for interaction: p = 0.02
Age:
18-35yrs: 1.57 (1.11, 2.24), p = 0.01
36-50yrs: 1.97 (1.41, 2.77), p< 0.001
51-84yrs: 1.54 (0.90, 2.63), p = 0.12
P for interaction: p =
0.99Premenopausal women with
current use of oral contraceptives:
No: 1.53 (0.86, 2.72), p = 0.14
Yes: 0.87 (0.46, 1.67), p = 0.68
P for interaction: p =
0.11 Postmenopausal women with
current use of hormone therapy:
No: 1.60 (0.72, 3.54), p = 0.24
Yes: 0.85 (0.29, 2.45), p = 0.76
P for interaction: p = 0.27Current use
of oral contraceptive or hormone
replacement therapy:
No: 1.64 (1.05, 2.57), p = 0.03
Yes: 0.97 (0.58, 1.61), p = 0.89
P for interaction: p = 0.048
Body Mass Index:
13.3-22.0:1.47(0.97, 2.23), p = 0.07;
22.1-24.9:1.72 1.17, 2.54), p = 0.006
25.0-53.3:1.83(1.03, 3.24), p = 0.04
P for interaction: p = 0.37
Education: Elementary/middle school:
1.93(1.35, 2.76), p< 0.001
High school: 1.72 (1.24, 2.39),
p = 0.001
College: 1.35 (0.74, 2.45), p = 0.33
P for interaction: p = 0.21
Deficiencies of natural anticoagulant
proteins:
None: 1.66 (1.26, 2.18), p< 0.001
Any: 2.56 (0.91, 7.18), p = 0.07
P for interaction: p = 0.41
Factor V Leiden or G20210A
prothrombin mutation:
None: 1.69 (1.27, 2.23), p< 0.001
Any: 1.79 (1.05, 3.05), p = 0.03
P for interaction: p = 0.83
Hyperhomocysteinemia:
December 2009
E-363
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
No: 1.66(1.26,2.19), p< 0.001
Yes: 2.19 (1.33, 3.61), p = 0.002
P for interaction: p = 0.25
Any cause of thrombophilia:
No: 1.59 (1.19, 2.13), p = 0.002
Yes: 1.96 (1.34, 2.87), p< 0.001
P for interaction: p = 0.27
Year of diagnosis:
1995-97:1.61(1.06, 2.46), p = 0.03
1998-00:1.34(0.90, 1.99), p = 0.15
2001-05:2.14(1.04, 4.39), p = 0.04
P for interaction: p = 0.12
Risk of DVT associated with PM10
over varying averaging times:
90 days: 0.91 (0.80,1.03), p = 0.12
180 days: 0.96 (0.82, 1.13), p = 0.63
270 days: 1.26
365 days: 1.70
1.01, 1.57
1.30,2.23
p = 0.04
p = 0.0001
2 yr: 1.47 (1.01, 2.14), p = 0.04
Risk of DVT associated with PMio (yr
preceding diagnosis/exam date)
sensitivity analysis to evaluate the
effect of different methods for
adjusting for long-term trends:
Handling of long-term time trends:
Ignored: 1.13 (0.89,1.42), p = 0.31
Dummy variable for each yr:
1.78(1.31,2.44), p = 0.0003
Linear term: 1.32 (1.02,1.69), p = 0.03
Penalized spline, 2 df: 1.54 (1.19, 2.00),
p = 0.001
Penalized spline, 3 df: 1.64(1.26, 2.14),
p = 0.0002
Penalized spline, 4 df: 1.70 (1.30, 2.23),
p = 0.0001
Penalized spline, 5 df: 1.70 (1.29, 2.22),
p = 0.0002
Penalized spline, 6 df: 1.66(1.26, 2.19),
p = 0.0003
Penalized spline, 7 df: 1.60 (1.21, 2.13),
p = 0.001
Penalized spline, 8 df: 1.55(1.15, 2.10),
p = 0.004
Reference: Baccarelli et al. (2009,
1881831
Period of Study: Jan 1995-Sept 2005
Location: Lombardia Region, Italy
Outcome: Deep Vein Thrombosis
Study Design: Case-control
Covariates: Age, Sex, area of
residence, BMI, education, medication
use
Statistical Analysis: Logistic
regression
Statistical Package: Stata
Pollutant: PM10
Risk of DVT measured with regards to
distance of residence from major road.
Specific levels of PM10 not given.
Increment: NA
Relative Risk (96%CI) of DVT
All subjects, age-adjusted:
1.33(1.03-1.71), p = 0.03
All subjects, adjusted for covariates:
1.47(1.10-1.96), p = 0.008
All subjects, adjusted for covariates and
background PM10 exposure: 1.47 (1.11-
1.96), p = 0.008
December 2009
E-364
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Calderon-Garciduenas et
al. (2008, 1563171
Period of Study: Children recruited
between Jul 2003 and Dec 2004
Location: Mexico (northeast or
southwest Mexico city or Polotitlan)
Outcome (ICD9 and ICD10): Plasma
Endothelin-1 (ET-1) and pulmonary
arterial pressure (PAP)
Age Groups: 6-13 yr
7.9+1.3 yr
Study Design: Cross-sectional
N: 81 children
Statistical Analyses: Analysis of
variance by parametric one-way
analysis of variance and the Newman-
Keuls multiple comparison post test,
Pearson's correlation
Covariates: Doesn't appear to have
performed multivariable analyses
However, collected information on age,
place and length of residency, daily
outdoor time, household cooking
methods, parents' occupational history,
family history of atopic illnesses and
respiratory disease, and personal
history of otolaryngologic and
respiratory symptoms
Season: No
Dose-response Investigated? No
Statistical Package: STATAvS.3, or
GraphPad Software, Inc.
Pollutant: PM,0 ftjg/rri)
Exposures assessed quantitatively in
Mexico City only
No monitors in Polotitlan
Averaging Time: 1, 2, and 7 days
before the exam
Pollutant concentrations between 0700
and 1900 h were used for the estimates
Mean (SD): Presented only in figures
Percentiles: NR
Range (Min, Max): Presented only in
figures
Monitoring Stations: 4 (2 in northeast
and 2 in southwest Mexico City)
Residence and school within 5 mi of
one of these monitors)
Copollutant (correlation): 0;
PM Increment: NA
Effect Estimate [Lower Cl, Upper Cl]:
No health effects models with measured
PM concentrations were presented
Used city of residence to assign
exposure
No multivariable analyses presented
Authors presented (statistically
significantly) elevated ET-1 levels
among children residing in both areas
of Mexico City as compared to
Polotitlan (control city):
Mean + SE (pg/mL)
Control: 1.23 ±0.06
Southwest Mexico City: 2.40 ±0.14
Northeast Mexico City: 2.09 + 0.10
Mexico City (overall): 2.24 ±0.12
Authors presented (statistically
significantly) elevated PAP levels
among children residing in both areas
of Mexico City as compared to
Polotitlan (control city):
Mean + SE (mmHg)
Control: 14.6 ±0.4
Southwest Mexico City: 16.7 + 0.6
Northeast Mexico City: 18.6 + 0.9
Mexico City (overall): 17.3 ±0.5
Correlation between ET-1 and time
spent outdoors: r = 0.31, p = 0.0012
Correlation between PAP and time
spent outdoors: r = 0.42, p = 0.0008
December 2009
E-365
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Diez Roux et al. (2008,
1564011
Period of Study: Baseline data
collected Jun 2000-Aug 2002
Exposure assessed retrospectively
between Aug 1982 and baseline date
Location: USA (6 field centers:
Baltimore, MD
Chicago, IL
Forsyth Co, NC
Los Angeles, CA
New York, NY
St. Paul, MN
Outcome (ICD9 and ICD10): Three
measures of subclinical atherosclerosis
(common carotid intimal-medial
thickness (CIMT), coronary artery
calcification, and ankle-brachial index
(ABI))
Age Groups: 44-84 yr (MESA cohort)
Study Design: Cross-sectional
retrospective cohort
N: 5172 for coronary calcium analysis
5037 for CIMT analysis
5110 for ABI analysis
Statistical Analyses: Generalized
Additive Models (Binomial regression:
presence of calcification
Linear regression: CIMT, ABI, amount of
calcium among persons with non-zero
calcification)
Covariates: Age, sex, race/ethnicity,
socioeconomic factors, cardiovascular
risk factors (BMI, hypertension, high
density lipoprotein and low density
lipoprotein cholesterol, smoking,
diabetes, diet, physical activity
models presented with and without
adjustment for cardiovascular RFs)
Season: NA
Dose-response Investigated? No
Statistical Package: NR
Pollutant: PM,0 (fjg/nr)
Averaging Time: 20-yr imputed mean
Mean (SD): 34.1 (7.5)
Percentiles: NR
Range (Min, Max): NR
Monitoring Stations: A spatio-temporal
model was used to predict monthly
PM2s exposures based on the
geographic location of each
participant's residence.
Copollutant (correlation with 20-yr
imputed mean):
PMio 20-yr observed mean
r = 0.93
PM25 20-yr imputed mean
r = 0.73
PM10 2001 imputed mean
r = 0.75
PM10 2001 observed mean
r = 0.80
PM2.5 2001 mean
PM Increment: 21.0 pg/m (approx.
10th-90th percentile)
Effect Estimate [Lower Cl, Upper Cl]:
CIMT:
Relative difference (95% Cl):
1.01 (1.00, 1.02)
Adj. for additional CVD RFs:
1.02(1.00, 1.03)
ABI:
Mean difference (95% Cl):
0.002 (-0.005, 0.009)
Adj. for additional CVD RFs:
0.001 (-0.006, 0.009)
Coronary calcium:
Relative prevalence (95% Cl):
1.02(0.96,1.07)
Adj. for additional CVD RFs:
1.02(0.96,1.08)
Coronary calcium (in those with
calcium):
Relative difference (95% Cl):
0.98(0.84, 1.13)
Adj. for additional CVD RFs:
1.01 (0.86, 1.18)
Found no clear heterogeneity by age,
sex, lipid status, smoking status,
diabetes status, BMI, education or study
site.
December 2009
E-366
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Maheswaran et al. (2005,
0886831
Period of Study: 1994-1998
Location: Sheffield, United Kingdom
Outcome (ICD9 and ICD10): Stroke
mortality (ICD9: 430-438) and
Emergency hospital admissions (ICD10:
I60-I69)
Age Groups: 2 45 yr
Study Design: Small area ecological
cross-sectional
N: 1030 census enumeration districts
(CEDs)
108 CEDs excluded from PM analyses
due to artifacts in the modeled
emissions data. The analysis was
based on 2979 deaths, 5122
admissions and a population of 199,682
Statistical Analyses: Poisson
regression
Covariates: Age, sex, socioeconomic
deprivation, and smoking prevalence
(some models also included age-by-
deprivation interaction)
Season: NA
Dose-response Investigated? Yes,
examined quintiles of exposure
Statistical Package: SAS
Pollutant: PM,o (fjg/nr)
Averaging Time: 5-yr avg
Mean (SD): Presented mean values
and ranges for each quintile of
exposure:
2: 17.5 (> 16.8, <18.2)
3: 18.8 (> 18.2, <19.3)
4: 19.8 (> 19.3, <20.6)
5: 23.3 (> 20.6)
Monitoring Stations: Used air pollution
model incorporating point, line and grid
sources of pollution and meteorological
data.
Copollutant (correlation):
CO (r = 0.82)
N0x(r = 0.87)
PM Increment: NA
Effect Estimate [Lower Cl, Upper Cl]:
Rate Ratios (96%CI) for stroke
mortality adjusted for overdispersion
by quintile of PMio level
Adjusted for sex and age:
1:1 (ref)
2:0.95(0.84,1.08)
3:1.12(0.99, 1.27)
4:1.16 1.03, 1.32)
5:1.39(1.23,1.58)
Adjusted for sex, age, deprivation, and
smoking:
1:1 (ref)
2:0.94(0.83, 1.07)
3:1.08 0.94, 1.24)
4:1.12(0.97,1.29)
5:1.33(1.14, 1.56)
Rate Ratios (96%CI) for emergency
hospital admissions because of
stroke by quintile of PMio level
Adjusted for sex and age:
1:1 (ref)
2:1.06(0.95,1.17)
3:1.10(0.99,1.23)
4:1.25 1.12, 1.38)
5:1.40(1.26,1.55)
Adjusted for sex, age, deprivation, and
smoking:
1:1 (ref)
2:1.01(0.91,1.13)
3:0.98(0.87, 1.10)
4:1.08(0.96,1.22)
5:1.13(0.99,1.29)
Rate Ratios (96%CI) for stroke
mortality in relation to spatially
smoothed (using a 1-km radius)
modeled outdoor air pollution
quintiles
Adjusted for sex, age, socioeconomic
deprivation, age by deprivation
interaction, and smoking prevalence:
1:1 (ref)
2: 0.86 (0.75, 0.98)
3:1.05(0.92, 1.21)
4:1.03 0.89, 1.19)
5:1.24(1.05,1.47)
Rate Ratios (96%CI) for emergency
hospital admissions because of
stroke in relation to spatially
smoothed modeled outdoor air
pollution quintiles
Adjusted for sex, age, socioeconomic
deprivation, age by deprivation
interaction, and smoking prevalence:
1:1 (ref)
2:1.05(0.94,1.17)
3:1.07(0.95,1.20)
4:1.06 0.94, 1.20)
5:1.15(1.01,1.31)
December 2009
E-367
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Maheswaran et al. (2005,
0907691
Period of Study: 1994-1998
Location: Sheffield, United Kingdom
Outcome (ICD9 and ICD10): Coronary
Heart Disease (CHD) mortality (ICD9:
410-414) and Emergency hospital
admissions (ICD10:120-125)
Age Groups: 2 45 yr
Study Design: Small area ecological
cross-sectional
N: 1030 census enumeration districts
(CEDs)
108 CEDs excluded from PM analyses
due to artifacts in the modeled
emissions data. Results based on 6857
deaths, 11407 hospital admissions and
199,682 people aged > 45 yr
Statistical Analyses: Poisson
regression
Covariates: Age, sex, socioeconomic
deprivation, and smoking prevalence
(some models also included age-by-
deprivation interaction)
Season: NA
Dose-response Investigated? Yes,
examined quintiles of exposure
Statistical Package: SAS
Pollutant: PM,0 ftjg/rri )
Averaging Time: 5-yr avg
Mean (SD): Presented mean values
and ranges for each quintile of
exposure:
2: 17.5 (> 16.8, <18.2)
3: 18.8 (> 18.2, <19.3)
4: 19.8 (> 19.3, <20.6)
5: 23.3 (> 20.6)
Monitoring Stations: Study used an air
pollution model incorporating points,
lines, and grids as sources of pollution,
and meteorological data.
Copollutant (correlation):
CO (r = 0.82)
N0x(r = 0.87)
PM Increment: NA
Effect Estimate [Lower Cl, Upper Cl]:
Rate Ratios (96%CI) for CHD
mortality in relation to modeled
outdoor air pollution quintiles,
adjusted for overdispersion
Adjusted for sex and age:
1:1 (ref)
2:1.06(0.98, 1.16)
3:1.10 1.01,1.21)
4:1.23(1.13,1.35)
5:1.30(1.19, 1.43)
Adjusted for sex, age, deprivation, and
smoking:
1:1 (ref)
2:1.03(0.94, 1.12)
3:1.00(0.90,1.11)
4:1.08(0.98, 1.20)
5:1.08(0.96, 1.20)
Adjusted for sex, age, deprivation, and
smoking (spatially smoothed using a
1km radius):
1:1 (ref)
2:0.97(0.89, 1.07)
3:1.00 0.90, 1.10)
4:1.03(0.93,1.15)
5:1.07(0.96, 1.21)
Rate Ratios (96%CI) for emergency
hospital admissions from CHD in
relation to modeled outdoor air
pollution quintiles
Adjusted for sex and age:
1:1 (ref)
2:1.08(0.98, 1.19)
3:1.11 (1.01,1.22)
4:1.17(1.07,1.29)
5:1.36(1.23,1.50)
Adjusted for sex, age, deprivation, and
smoking:
1:1 (ref)
2:1.03(0.93, 1.13)
3:0.96(0.86,1.07)
4:0.97(0.87,1.08)
5:1.01 (0.90, 1.14)
Adjusted for sex, age, deprivation, and
smoking (spatially smoothed using a
1km radius):
1:1 (ref)
2:1.01(0.92,1.11)
3:1.04 0.93, 1.15)
4:0.97(0.87,1.08)
5:1.07(0.95,1.20)
December 2009
E-368
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: O'Neill et al. (2007,
1560061
Period of Study: 2000-2004
Location: USA (6 field centers:
Baltimore, MD
Chicago, IL
Forsyth Co, NC
Los Angeles, CA
New York, NY
St. Paul, MN
Outcome (ICD9 and ICD10):
Creatinine adjusted urinary albumin
excretion
Assessed 2 ways: continuous log
urinary albumin/creatine ration (UACR)
and clinically defined micro- or macro-
albuminuria (UACR > 25 mg/g) vs.
normal levels
Age Groups: 44-84 yr
Study Design: Cross-sectional
analyses and prospective cohort
analyses
N: 3901 participants free of clinical CVD
at baseline
Statistical Analyses: At baseline:
multiple linear regression (continuous
outcome)
Binomial regression (dichotomous
outcome)
3-yr change: repeated measures model
with random subject effects (estimate 3-
yr change in log UACR by levels of
exposure)
Covariates: Age, gender, race, BMI,
cigarette status, ETS, percent dietary
protein
For repeated measures models: time
Timex PMio
Season: NA
Dose-response Investigated? Yes,
examined quartiles of exposure
Statistical Package: SAS
Pollutant: PM,0
Averaging Time: Avg of previous
month, avg of previous 2 mo (recent
exposures)
20-yr directly monitored PM10 avg,
20-yr imputed PM10 avg (longer-term
exposures)
Mean (SD):
Previous 20 yr: 34.7 (7.0)
Previous month: 27.5 (7.9)
Percentiles: NR
Range (Min, Max): NR
Monitoring Stations: NR (used closest
monitor to residence to assign exposure
20-yr imputed PMio was derived using a
space-time model)
Copollutant (correlation): PM25
PM Increment: 10 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
Adjusted mean differences in log
UACR (mg/g) per increase in PIvl-;
among participants seen at baseline
Previous 30 days
Full sample: -0.42 (-0.085, 0.002)
Within 10 km:-0.023 (-0.079, 0.034)
Previous 60 days
Full sample:-0.056 (-0.106 to-0.005)
Within 10 km:-0.040 (-0.106, 0.025)
20 yr PMio (nearest monitors)
Full sample: -0.019 (-0.072, 0.033)
Wthin 10 km: 0.009 (-0.067, 0.085)
Imputed 20 yr exposure
Full sample: -0.002 (-0.038, 0.035)
Wthin 10 km: 0.016 (-0.033, 0.066)
Adjusted relative prevalence of
microalbuminuria vs. high-normal
and normal levels (below 26 mg/g)
per increase in PIvl-: among
participants without
macroalbuminuria during the
baseline visit
Previous 30 days: 0.88 (0.76,1.02)
Previous 60 days: 0.83 (0.70, 0.99)
20 yr PM10 (nearest monitors):
0.92(0.77, 1.08)
Imputed 20 yr exposure:
0.98(0.87,1.10)
Adjusted mean 3-yr change (SE) in
log UACR (mg/g) by quartiles of
1982-2002 exposure to PM10 from
ambient monitors among
participants seen in 2000-20004
Full sample
Quartile:
18.5 to <29.3: 0.147 (0.024)
29.3 to <33.1: 0.159 (0.024)
33.1to<36.3:0.163(0.024)
36.3(055.7:0.174(0.023)
p-trend: 0.42
Within 10 km
Quartile:
18.5 to <29.3: 0.159 (0.030)
29.3 to <33.1: 0.155 (0.031)
33.1to<36.3:0.167(0.028)
36.3to55.7:0.152(0.036)
p-trend: 0.99
Interactions with either 20 yr or shorter-
term PM exposure were not significant
(p< 0.01) by gender, age, city,
race/ethnicity or study site.
Reference: Puett et al, (2008,1568911
Period of Study: 1992-2002
Location: Northeastern metropolitan
U.S.
Outcome: Nonfatal myocardial
infarction
Study Design: Cohort
Covariates: Age in months, state of
residence, yr and season
Statistical Analysis: Cox proportional
hazard
Statistical Package: SAS
Age Groups: 30-55
Pollutant: PM,0
Averaging Time: 3-, 12-, 24-, 36- and
48-mo ma
Mean (SD) Unit: NR
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: 10|jg/m
Hazard Ratio, 96% Cl, 12 month ma
0.94(0.77-1.15)
December 2009
E-369
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Rosenlund et al. (2006,
1146781
Period of Study: 1992-1994
Location: Stockholm County, Sweden
Outcome (ICD9 and ICD10):
Myocardial infarction (Ml)
Age Groups: 45-70 yr
Study Design: Case-control
N: 1397 cases
1870 controls
Statistical Analyses: Logistic
regression (main analysis)
Also performed multinomial logistic
regression to assess cases as nonfatal,
fatal in the hospital within 28 days, and
out-of-hospital death within 28 days with
all controls as reference
Covariates: Age, sex, and hospital
catchment area (frequency matched
variables)
Smoking, physical inactivity, diabetes,
SES
Also assessed but did not include
hypertension, BMI, job strain, diet,
passive smoking, alcohol consumption,
coffee intake, and occupational
exposure to motor exhaust and other
combustion products
Season: NA
Dose-response Investigated? No
Statistical Package: STATAv8.2
Pollutant: PM10
(modeled traffic-related pollution; also
modeled PM25, but since the PM
correlation was high (r = 0.998) only
PM10 results were presented) (|jg/m3)
Averaging Time: 30 yr (PM only
assessed during 2000, thus assumed
constant levels during 1960-2000)
Median (6th-96th percentile):
Cases: 2.6 (0.5-6.0)
Controls: 2.4 (0.6-5.9)
Range (Min, Max): NR
Monitoring Stations: NR
Copollutant (correlation):
N02(r = 0.93)
CO (r = 0.66)
SO,
PM Increment: 5 pg/rri (5th to 95th
percentile distribution among controls)
Effect Estimate [Lower Cl, Upper Cl]:
Association of 30-yr avg exposure to air
pollution from traffic with Ml
Logistic regression
All cases: 1.00 (0.79,1.27)
Multinomial logistic regression
Nonfatal cases: 0.92 (0.71,1.19)
Fatal cases: 1.39 (0.94, 2.07)
In-hospital death: 1.21 (0.75,1.94)
Out-of-hospital death: 1.84 (1.00, 3.40)
After adjustment for heating-related
S02, the estimate for fatal Ml was 1.40
(0.86-2.26) for PM10.
December 2009
E-370
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Zanobetti & Schwartz
(2007, 0912471
Period of Study: 1985-1999
Location: 21 U.S. cities (Birmingham,
Alabama
Boulder, Colorado
Canton, Ohio
Chicago, Illinois
Cincinnati, Ohio
Cleveland, Ohio
Colorado Springs, Colorado
Columbus, Ohio
Denver, Colorado
Detroit, Michigan
Honolulu, Hawaii
Houston, Texas
Minneapolis-St. Paul, Minnesota
Nashville, Tennessee
New Haven, Connecticut
Pittsburgh, Pennsylvania
Provo-Orem, Utah
Salt Lake City, Utah
Seattle, Washington
Steubenville, Ohio
and Youngstown, Ohio)
Outcome (ICD9 and ICD10): Death,
subsequent myocardial infarction (Ml
ICD9 codes 410.0-410.9), and a first
admission for congestive heart failure
(CHF
ICD9 code 428)
Age Groups: > 65 yr
Study Design: Cohort
N: 196,000 persons discharged alive
following an acute Ml
Statistical Analyses: Cox's
Proportional Hazards Regression
Pollutant: PM,0
Averaging Time: Yearly avg of
pollution for that yr and lags up to the 3
previous yr (distributed lag)
Mean (SD): 28.8 (all cities
SD not reported)
Percentiles: 10, 50, and 90 percentiles
listed individually for each city (Table 2)
Range (Min, Max): NR
Monitoring Stations: NR (obtained
data from the U.S. EPAAerometric
Information Retrieval System)
Meta-regression for city-specific results Copollutant (correlation): None
Covariates: Age, sex, race, type of Ml,
number of days of coronary care and
intensive care, previous diagnoses for
atrial fibrillation, and secondary or
previous diagnoses for COPD,
diabetes, and hypertension, and for
season of initial event (time period, and,
sex, race, and type of Ml were treated
as stratification variables)
Season: Assessed as a confounder
Dose-response Investigated? No
Statistical Package: NR
PM Increment: 10 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
Hazard ratio (95%CI) for an increase in
PM for the yr of failure and for the
distributed lag from the yr of failure up
to 3 previous yr
Death
PM,o annual: 1.11 (1.05,1.19),
p = 0.001
Distributed lag model
Lag 0:1.04 (0.96, 1.14), p = 0.336
Lag 1:1.07 (0.99, 1.14), p = 0.070
Lag2:1.14 1.10,1.18, p = 0.000
Lag 3:1.06 0.99, 1.12 , p = 0.077
Sum lags 0-3:1.34 (1.14,1.52),
p = 0.000
CHF
PM,o annual: 1.11 (1.03,1.21),
p = 0.009
Distributed lag model
Lag 0:1.09 (1.01,1.18), p = 0.030
Lag 1:1.09 1.01,1.19, p = 0.038
Lag2:1.13 1.02, 1.25, p = 0.014
Lag 3:1.04 (0.97, 1.12), p = 0.260
Sum lags 0-3:1.41 (1.19,1.66),
p = 0.000
2nd Ml
PM10 annual: 1.17 (1.05,1.31),
p = 0.003
Distributed lag model
Lag 0:1.09
Lag 1:1.12
0.92,1.30
p = 0.325
p = 0.108
__a _ 0.97, 1.30, r
Lag 2:1.15 (1.08, 1.23), p = 0.000
Lag 3:1.01 (0.94, 1.09), p = 0.783
Sum lags 0-3:1.43 (1.12,1.82),
p = 0.005
Hazard Ratio (95%CI) for an increase in
PM (sum of the previous 3 yr distributed
lag) for the sensitivity analyses
Death
Subjects with follow-up starting after
2nd Ml:
1.33(1.15, 1.55), p = 0.000
Subjects admitted between 1985-1996:
1.45(1.26, 1.68), p = 0.000
2nd cohort definition (yr defined at time
of Ml):
1.29(1.15, 1.44), p = 0.000
CHF
Subjects with follow-up starting after
2nd Ml:
1.42(1.22, 1.65), p = 0.000
Subjects admitted between 1985-1996:
1.51 (1.26, 1.81 ,p = 0.000
2nd Ml
Subjects admitted between 1985-1996:
1.62(1.23, 2.13), p = 0.001
Note: Age and sex effect modification
results presented in Fig 1
Used meta-regression to examine
predictors of heterogeneity across city
and found that most predictors were not
significant modifiers of PM (Table 7)
All units expressed in pg/m unless otherwise specified.
December 2009
E-371
-------
Table E-21. Long-term effects-cardiovascular- PIVhs (including PM components/sources).
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Allen et al. (2009,1562091
Period of Study: Oct 2000-Sep 2002
(exposure averaging period)
outcome assessed in 2002
Location: 5 U.S. communities
(Chicago, Illinois
Forsyth County, North Carolina
Los Angeles, California
Northern Manhattan and the Bronx,
New York
and St. Paul, Minnesota)
part of MESA (Multi-ethnic Study of
Atherosclerosis)
Outcome (ICD9 and ICD10):
Abdominal aortic calcium (AAC), a
marker of systemic atherosclerosis
(quantitative measure of interest was
the Agatston score)
Age Groups: 46-88 yr
Study Design: Cross-sectional
N: 1,147 participants (sensitivity
analysis among 1,269 participants)
Statistical Analyses: 2-part modeling
approach:
1) Modeled relative risk of having any
AAC using a log link and a Gaussian
error model
Sensitivity analysis used modified
Poisson regression with robust error
variance
2) Multiple linear regression of the log-
transformed AAC Agatston score
(among those with AAC>0)
Sensitivity analysis modeled all
participants by adding 1 prior to log-
transforming
Covariates: Age, gender, race/ethnicity,
BMI, smoking status, pack-yr of
smoking, diabetes, education, annual
income, blood lipid concentration, blood
pressure, and medications
Assessed impact of gender, age,
diabetes, obesity, use of lipid-lowering
medications, education, income,
race/ethnicity, and employment status
on heterogeneity of effects (or in
sensitivity analyses)
Season: NA
Dose-response Investigated? NR
Statistical Package: SASv9.1
Pollutant: PM25
Averaging Time: 2-yr averaging period
(Oct 2000-Sep 2002)
Mean (SD): 15.8 (3.6) pg/m3
Percentiles: NR
Range (Min, Max): 10.6-247 pg/m3
Monitoring Stations: All monitors with
1) the objective of "population
exposure," "regional transport," or
"general/background;) and 2) at least
50% data reporting in each of 8 3-
month periods over the averaging time
Used monitors located within 50 km of a
study participant's residence
Copollutant (correlation):
Assessed traffic by roadway proximity
PM Increment: 10 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
Results for fully adjusted models under
different participant inclusion,
employment status, and roadway
proximity criteria.
Presence/Absence of Calcium RR
(95% Cl)
Inclusion criteria: <10yrs at address:
1.04(0.89,1.22)
>10yrs at address: 1.06 (0.96,1.16)
2 10yrs at address & <10km from
monitor: 1.08 (0.98,1.18)
>20yrs at address: 1.10(0.99,1.22)
2 20yrs at address & <10km from
monitor: 1.11 (1.00,1.24)
<10yrs at address & employed:
1.02(0.87,1.20)
> 20yrs at address & employed:
1.07(0.89, 1.27)
<10yrs at address & not employed:
1.10(1.00,1.22)
2 20yrs at address & not employed:
1.16(1.02,1.31)
<10yrs at address & near major road:
0.85(0.69, 1.05)
2 20yrs at address & not near major
road: 1.10 (0.99,1.23)
Log-transformed Agatston Score
(Agatston >0)
% Change (96% Cl)
Inclusion criteria: <10yrs at address:
-6.6 (-64.0, 50.9)
2 10yrs at address: 8.0 (-29.7, 45.7)
2 10yrs at address & <10km from
monitor: 19.7 (-19.6, 58.9)
2 20yrs at address: 14.4 (-32.8, 61.7)
2 20yrs at address & <10km from
monitor: 24.6 (-24.6, 73.8)
<10yrs at address & employed:
29.1 (-25.7,83.8)
2 20yrs at address & employed:
43.8 (-32.4, 119.9)
<10yrs at address & not employed:
-15.1 (-66.3,36.1)
> 20yrs at address & not employed:
-14.1 (-72.6,44.4)
<10yrs at address & near major road:
34.0 (-44.2,112.1)
2 20yrs at address & not near major
road: 3.9 (-39.9, 47.8)
Log-transformed Agatston Score (all)
% Change (95% Cl)
Inclusion criteria: <10yrs at address: -
8.5 (-81.3, 64.2)
2 10yrs at address: 40.7 (-11.5, 92.8)
2 10yrs at address & <10km from
monitor: 60.7 (5.9,115.4)
2 20yrs at address: 64.1 (-1.73,129.9)
2 20yrs at address & <10km from
monitor: 79.2 (10.1,148.3)
<10yrs at address & employed:
33.5 (-35.9, 102.9)
> 20yrs at address & employed:
55.8 (-37.2, 148.7)
<10yrs at address & not employed:
54.8 (-23.8, 133.4)
2 20yrs at address & not employed:
89.3 (-3.7, 182.3)
<10yrs at address & near major road:
-30.6 (-141.3, 80.1)
2 20yrs at address & not near major
December 2009
E-372
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
road: 51.3 (-8.3,110.8)
Exploratory/sensitivity analyses
(also presented in figures):
Detectable AAC RR (96%CI):
Among women: 1.14(1.00,1.30)
Among persons >65yrs:
1.10(1.01,1.19)
Among users of lipid-lowering
medications: 1.14(1.00,1.30)
Among Hispanics: 1.22 (1.03,1.45)
Imputing missing covariates among
residentially stable participants:
1.08(0.98, 1.19)
Agatston score % change (96%CI):
Among Hispanics: 64 (-4,133)
Among persons earning >$50,000: 72
(5, 139)
Agatston score including those with
Agatston = 0
% change (96%CI):
Fully adjusted model: 41 (-12,93)
Among persons >65yrs: 75 (8,143)
Among diabetics: 149(29,270)
Among users of lipid-lowering
medications: 121 (25, 217)
Among Hispanics: 141 (45,236)
Imputing missing
Covariates: 49 (1.3,100.1)
December 2009 E-373
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Auchincloss et al. (2008,
1562341
Period of Study: Jul 2000-Aug 2002
Location: 6 U.S. communities
(Baltimore City and Baltimore County,
Maryland
Chicago, Illinois
Forsyth County, North Carolina
Los Angeles, California
Northern Manhattan and the Bronx,
New York
and St. Paul, Minnesota)
part of MESA (Multi-ethnic Study of
Atherosclerosis)
Outcome (ICD9 and ICD10): Blood
pressure: systolic (SBP), diastolic
(DBP), mean arterial (MAP), pulse
pressure (PP)
Avg of 2nd and 3rd BP measurement
used for analyses
Age Groups: 45-84 yr
Study Design: Cross-sectional (Multi-
Ethnic Study of Atherosclerosis baseline
examination)
N: 5,112 persons (free of clinically
apparent cardiovascular disease)
Statistical Analyses: Linear regression
Secondary analyses used log binomial
models to fit a binary hypertension
outcome
Covariates: Age, sex, race/ethnicity,
per capita family income, education,
BMI, diabetes status, cigarette smoking
status, exposure to ETS, high alcohol
use, physical activity, BP medication
use, meteorology variables, and
copollutants
Examined site as a potential
confounder and effect modifier
Heterogeneity of effects also examined
by traffic-related exposures, age, sex,
type 2 diabetes, hypertensive status,
cigarette use
Season: Adjusted for temperature and
barometric pressure to adjust for
seasonality (because seasons vary by
the study sites)
Also performed sensitivity analyses
adjusting for season to examine the
potential for residual confounding not
accounted for by weather variables
Dose-response Investigated?
Assessed nonlinear relationships-no
evidence of strong threshold/nonlinear
effects for PM25
Statistical Package: NR
Pollutant: PM25
Averaging Time: 5 exposure metrics
constructed: prior day, avg of prior 2
days, prior 7 days, prior 30 days, and
prior 60 days
Mean (SD): Prior day: 17.0 (10.5)
Prior 2 days: 16.8 (9.3)
Prior 7 days: 17.0 (6.9)
Prior 30 days: 16.8 (5.0)
Prior 60 days: 16.7 (4.4)
Percentiles: NR
Range (Min, Max): NR
Monitoring Stations: Used monitor
nearest the participant's residence to
calculate exposure metrics
Copollutant (correlation):
S02
N02
CO
Traffic-related exposures (straight-line
distance to a highway; total road length
around a residence)
PM Increment: 10 pg/m (approx.
equivalent to difference between 90th
and 10th percentile for prior 30 day
mean)
Effect Estimate [Lower Cl, Upper Cl]:
Adjusted mean difference (95% CIJ in
PP and SBP (mmHg) per 10 pg/m
increase in PM2 5 (avg for the prior 30
days)
Pulse Pressure
Adjustment variables: Person-level
Covariates: 1.04 (0.25,1.84)
Person-level cov, weather:
1.12(0.28,1.97)
Person-level cov, weather, gaseous
copollutants: 2.66 (1.61, 3.71)
Person-level cov, study site:
0.93 (-0.04,1.90)
Person-level cov, study site, weather:
1.11 (0.01,2.22)
Person-level cov, study site, weather,
gaseous copollutants: 1.34 (0.10, 2.59)
Systolic Blood Pressure
Adjustment variables: Person-level
Covariates: 0.66 (-0.41,1.74)
Person-level cov, weather:
0.99 (-0.15, 2.13)
Person-level cov, weather, gaseous
copollutants: 2.8 (1.38, 4.22)
Person-level cov, study site:
0.86 (-0.45, 2.17)
Person-level cov, study site, weather:
1.32 (-0.18, 2.82)
Person-level cov, study site, weather,
gaseous copollutants: 1.52 (-0.16, 3.21)
Additional results: Associations
became stronger with longer averaging
periods up to 30 days. For example:
Adjusted (personal covariates and
weather) mean differences in PP: Prior
day: -0.38 (-0.76, 0.00)
Prior 2 days:-0.22 (-0.65, 0.21)
Prior 7 days: 0.52 (-0.08,1.11)
Prior 30 days: 1.12 (0.28,1.97)
Prior 60 days: 1.08 (0.11, 2.05)
(Pattern held for additional adjustments
and for SBP results. Therefore, only
results for 30-day mean differences
were presented)
Additional results (not presented):
None of DBP results were statistically
significant. Rresults for MAP were
similar to SBP, though weaker and
generally not significant
Effect modification: associations
between PM2 5 and BP were stronger
for persons taking medications, with
hypertension, during warmer weather,
in the presence of high N02, residing <
300m from a highway, and surrounded
by a high density of roads (Fig 1)
Associations were not modified for age,
sex, diabetes, cigarette smoking, study
site, high levels of CO or S02, season ,
nor residence < 400m fro a highway
Note: supplementary material available
on-line
December 2009
E-374
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Calderon-Garciduenas et
al. (2009, 1921071
Period of Study: Sept 2004-Jan 2005
Location: Mexico City and Polotitlan,
Mexico
Outcome: Flow cytometry
Study Design: Panel
Covariates: NR
Statistical Analysis: Pearson's
Correlation
Statistical Package: Stata
Age Groups: 9.7 + 1.2yr
Pollutant: PM25
Averaging Time: 1-, 2- and 7-day avg
Mean (SD) Unit: 35.89 + 0.93 pg/m3
Range (Min, Max): NR
Copollutant: PM10, 03
Increment: NR
Flow cytometry results and their
statistical significance in control vs.
exposed children
CDS
Ex posed: 62.9+1.8
Control: 67.1+1.7
P = 0.1
CD4
Exposed: 39.3+1.3
Control: 38.2+1.4
P = 0.57
CDS
Exposed: 24.0+0.95
Control: 27.3+1.0
P = 0.02
CD4/CD8
Exposed: 1.7+0.14
Control: 1.4+0.07
P = 0.09
CD3-/CD19+
Exposed: 11.8+1.0
Control: 14.8+1.0
P = 0.04
CD56+
Exposed: 11.5+1.2
Control: 12.4+1.5
P = 0.63
CD56+/CD3-NK
Exposed: 14.0+9.5
Control: 7.0+2.7
P = 0.003
HLA-DR+
Exposed: 27.5+4.2
Control: 17.0+2.4
P = 0.04
mCD14+
Exposed: 66.5+2.3
Control: 80.6+1.8
P = <0.001
CD14/CD69
Exposed: 0.20+0.07
Control: 1.0+0.26
P = O.001
CD4/CD69
Exposed: 0.08+0.03
Control: 3.1+0.65
P = <0.001
December 2009
E-375
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Diez Roux et al. (2008,
1564011
Period of Study: Baseline data
collected Jun 2000-Aug 2002
Exposure assessed retrospectively
between Aug 1982 and baseline date
Location: USA (6 field centers:
Baltimore, MD
Chicago, IL
Forsyth Co, NC
Los Angeles, CA
New York, NY
St. Paul, MN
Outcome (ICD9 and ICD10): Three
measures of subclinical atherosclerosis
(common carotid intimal-medial
thickness (CIMT), coronary artery
calcification, and ankle-brachial index
(ABI))
Age Groups: 44-84 yr
Study Design: Cross-sectional
N: 5172 for coronary calcium analysis
5037 for CIMT analysis
5110 for ABI analysis
Statistical Analyses: Generalized
Additive Models (Binomial regression:
presence of calcification
Linear regression: CIMT, ABI, amount of
calcium)
Covariates: Age, sex, race/ethnicity,
socioeconomic factors, cardiovascular
risk factors (BMI, hypertension, high
density lipoprotein and low density
lipoprotein cholesterol, smoking,
diabetes, diet, physical activity
Models presented with and without
adjustment for cardiovascular RFs)
Season: NA
Dose-response Investigated? No
Statistical Package: NR
Pollutant: PM25
Averaging Time: 20-yr imputed mean
Mean (SD): 21.7 (5.0)
Percentiles: NR
Range (Min, Max): NR
Monitoring Stations: NR
Long-term exposure to PM estimated
based on residential history reported
retrospectively
all addresses geocoded
ambient AP obtained from U.S. EPA
Copollutant (correlation):
PM10 20-yr observed mean
r = 0.64
PMio 20-yr imputed mean
r = 0.73
PM,o 2001 mean
r = 0.43
PM252001 mean
r = 0.64
Due to high correlation among PM
exposures, only results of mean 20-yr
exposures are reported.
PM Increment: 12.5 pg/m (approx.
10th-90th percentile)
Effect Estimate [Lower Cl, Upper Cl]:
CIMT:
Relative difference (95% Cl):
1.01 (1.00,1.01)
Adj. for additional CVD RFs:
1.01 (1.00, 1.02)
1.02
ABI:
Mean difference (95% Cl):
0.000 (-0.006, 0.006)
Adj. for additional CVD RFs:
-0.001 (-0.006, 0.006)
Coronary calcium:
Relative prevalence (95% Cl):
1.01 (0.96,1.05)
Adj. for additional CVD RFs:
1.01 (0.96, 1.06)
1.02
Coronary calcium (in those with
calcium):
Relative difference (95% Cl):
0.99(0.88, 1.12)
Adj. for additional CVD RFs:
1.01 (0.89, 1.14)
December 2009
E-376
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Hoffman et al. (2007,
0911631
Period of Study: 2000-2003
Location: Ruhr area of Germany (3
large cities: Essen, Mulheim, and
Bochum)
Outcome (ICD9 and ICD10): Coronary Pollutant: PM
artery calcification (CAC)
Age Groups: 45-74 yr
Study Design: Cross-sectional
N: 4494 participants
Statistical Analyses: Linear regression
(outcome = natural logarithm of CAC
score + 1)
Logistic regression (outcome = CAC
score above/below the age- and
gender-specific 75th percentile)
Covariates: City and area of residence,
age, sex, education, smoking, ETS,
physical inactivity, waist-to-hip ratio,
diabetes, blood pressure, and lipids
(and household income in a subset)
Season: NA
Dose-response Investigated? Yes,
PM was also categorized into quartiles
for analyses
Statistical Package: NR
Averaging Time: 1 yr (2002, midpoint
of the study)
Mean (SD): Total:
22.8(1.5)
High traffic exposure (< 100m):
22.9(1.4)
Low traffic exposure (>100m):
22.8(1.5)
Percentiles:
Q1:21.54
Q2: 22.59
Q31 23 75
10th-90th percentile: 3.91
Range (Min, Max): NR
Monitoring Stations: Daily mean PM25
values for 2002 were estimated with the
EURAD model using data from official
emission inventories, meteorological
information, and regional topographical
data.
Copollutant (correlation): None
(Traffic was assessed using distance to
roadways)
Correlation between modeled daily avg
of PM2.5 and measured PM25: 0.86-
0.88, depending on season.
PM Increment: 3.91 pg/rri (10th-90th
percentile)
Effect Estimate [Lower Cl, Upper Cl]:
Percent change (96%CI) in CAC
associated with an increase in PM25
Unadjusted: 12.7 (-7.0, 36.4)
Model 1 (adjusted for distance to major
road): 12.3 (-7.3, 35.9)
Model 2 (model 1 + city and area of
residence): 29.7 (0, 68.3)
Model 3 (model 2 + age, sex,
education): 24.2 (0,55.1)
Model 4 (model 3 + smoking, ETS,
physical inactivity, waist-to-hip ratio):
17.9 (-5.3, 46.7)
Model 5 (model 4 + diabetes, blood
pressure, LDL, HDL, triglycerides):
17.2 (-5.6, 45.5)
Adjusted ORs(96%CI) for the
association between the top quarter
of PM exposure vs. the low quarter
of PM exposure and a CAC score
above the age- and sex-specific 76th
percentiles
All: 1.22 (0.96, 1.54)
No CHD: 1.22 (0.95, 1.57)
Men: 1.09 (0.78,1.53)
Women: 1.34 (0.97,1.87)
Age<60yr:1.18(0.83,1.68
Age >60yr: 1.27 (0.93,1.75
Nonsmokers:1.17(0.89,1.53)
Current smokers: 1.30 (0.83, 2.05)
Educational level
Low: 1.16 (0.86, 1.57)
Medium: 1.30 (0.83, 2.05)
High: 1.62 (0.81, 3.25)
Additional notes:
No clear dose-response relationship
demonstrated when exposure assessed
in quartiles (Fig 2)
Participants who had not been working
full-time during the last 5 yr showed
stronger effects, with possible dose-
response between PM25 and CAC
(results presented in Fig 3)
Reference: Hoffman et al. (2006,
0911621
Period of Study: Dec 2000-Jul 2003
Location: Ruhr area of Germany (2
large cities: Essen, Mulheim)
Outcome (ICD9 and ICD10): Clinically
manifest CHD (defined as self-reported
history of a 'hard' coronary event, i.e.
myocardial infarction or application of a
coronary stent or angioplasty or bypass
surgery)
Age Groups: 45-75 yr
Study Design: Cross-sectional
(German Heinz Nixdorf RBCALL study)
N: 3399 participants
Statistical Analyses: Multivariable
logistic regression
Covariates: Sex, diabetes,
hypertension, smoking status, ETS,
educational level, physical activity, BMI,
triglycerides, age, cigarettes smoked
per day, WHR, LDL, HDL, HbAlc,
indicator variable for cities, indicator
variable for living in northern part of
cities.
Statistical Package: SASv8.2
Pollutant: PM25 (pg/m3)
Averaging Time: Yearly mean
estimated with model for yr 2002 (on a
spatial scale of 5 km)
Mean (SD):
Total: 23.3 (1.4)
High traffic: 23.4 (1.4)
Low traffic: 23.3 (1.4)
Percentiles: NR
Range (Min, Max): NR
Monitoring Stations: NR
Copollutant (correlation): None
(Traffic was assessed using distance to
roadways)
Effect Estimate [Lower Cl, Upper Cl]:
Model 1: PM25 + high traffic exposure
0.92 (0.36, 2.39)
Model 2: model 1 + age, sex
0.83(0.31,2.27)
Model 3: model 2 + education, diabetes,
HbAlc, BMI, WHR, smoking status,
ETS, physical activity, city, area of
residence
0.56(0.16,2.01)
Model 4: model 3 + hypertension, lipids
0.55(0.14,2.11)
Modeled vs. Measured: r = 0.86-0.88,
depending on season
December 2009
E-377
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Hoffmann et al, (2009,
1903761
Period of Study: 2000-2003
Location: Ruhr area, Germany
Outcome: Peripheral Arterial Disease
Study Design:
Covariates: Height, weight, medication
use, diabetes, physical activity level,
smoking, socioeconomic status,
education, population density
Statistical Analysis: NR
Statistical Package: NR
Age Groups: 45-75 yr
Pollutant: PM25
Averaging Time: Daily
Mean (SD) Unit: 22.96 (0.85)
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: 3.91 pg/m
Odds Ratio (96%CI) for prevalence of
peripheral arterial disease
0.87(0.57-1.34)
Reference: Kunzli et al. (2005, 0873871
Period of Study: 1998-2003
Location: Los Angeles Basin
Outcome (ICD9 and ICD10): Carotid
intima-media thickness (CIMT)
Age Groups: Less than 40 yr excluded
Mean age = 59.2 + 9.8
Study Design: Cross-sectional
N: 798 participants
Statistical Analyses: Linear regression
Covariates: Age, sex, education,
income, smoking, ETS, blood pressure,
LDL cholesterol, treatment with
antihypertensives or lipid-lowering
medications
Season: NA
Dose-response Investigated? Yes,
assessed PM25 in quartiles
Statistical Package: NR
Pollutant: PM25 (pg/rri)
Averaging Time: GIS/geostatics model
to estimate long-term mean ambient
concentrations of PM25' derived from
data collected in 2000, including data
from 23 state and local monitoring
stations.
Mean (SD): 20.3 ± 2.6
Percentiles: NR
Range (Min, Max): 5.2, 26.9
Monitoring Stations: 23 monitors
Copollutant (correlation): None
PM Increment: 10 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
Percent change (95%CI) in CIMT
associated with an increase in PM25
concentration
Based on a linear model with log intima-
media thickness as dependent variable
Total population:
Unadjusted: 5.9 (1.0, 10.9), p = 0.018
Adjusted for age, sex, education,
income:
4.4 (0.0, 9.0), p = 0.056
Adjusted for above + smoking, ETS,
multivitamins, alcohol:
4.2 (-0.2, 8.9), p = 0.064
Among Females Ł 60 yr:
Unadjusted: 19.2 (8.8, 30.5), p = 0.001
Adjusted for age, sex, education,
income:
15.7(5.7, 26.6), p = 0.002
Adjusted for above + smoking, ETS,
multivitamins, alcohol:
13.8(4.0, 24.5), p = 0.002
Among those taking lipid-lowering
therapy:
Unadjusted: 15.8(2.1, 31.2), p = 0.024
Adjusted for age, sex, education,
income:
13.3(0, 28.5), p = 0.031
Adjusted for above + smoking, ETS,
multivitamins, alcohol:
13.3 (-0.3, 28.8), p = 0.060
For the observed contrast between
lowest and highest exposure:
Approximately 20 pg/ m -> 12.1% (2.0-
231%) increase in CIMT.
Among nonsmokers: 6.6% (1.0-12.3%).
The estimate was small and not
significant in current and former
smokers.
Wfomen: In the range of 6-9% per 10
pg/m3
Unadjusted means of CIMT across
quartiles of exposure were 734, 753,
758, and 774 pm
Adjusted means trend across exposure
groups, p = 0.041
Stratified results presented in figures
Reference: Miller et al. (2007, 0901301
Period of Study: 1994-2003
Location: 36 U.S. metropolitan areas
(Women's Health Initiative)
Outcome (ICD9 and ICD10): First
cardiovascular event (myocardial
infarction, coronary revascularization,
stroke, and death from either coronary
heart disease [categorized as "definite"
or "possible"] or cerebrovascular
disease)
Age Groups: 50-79 yr (median age at
Pollutant: PM25 (pg/rri)
Averaging Time: Annual avg
concentration in 2000 (used to
represent long-term exposure)
Mean (SD):
Individual exposure: 13.5 (3.7)
Citywide avg exposure: 13.5 (3.3)
PM Increment: 10 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
Estimated Hazards Ratio (95%CI) for
the time to the first cardiovascular event
or death associated with an increase in
PM25
Any cardiovascular event (first event)
Overall: 1.24 (1.09,1.41)
December 2009
E-378
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
enrollment: 63)
Study Design: Cohort (median follow-
up of 6 yr)
N: 65,893 postmenopausal women
without previous cardiovascular disease
Statistical Analyses: Cox-proportional
hazards regression
Covariates: Age, race/ethnicity,
smoking status, the number of
cigarettes smoked per day, the number
of yr of smoking, systolic blood pres-
sure, education level, household
income, BMI, and presence or absence
of diabetes, hypertension, or hyper-
cholesterolemia (also evaluated ETS,
occupation, physical activity, diet,
alcohol consumption, waist circum-
ference, waist-to-hip ratio, medical
history, medications, and presence or
absence of a family history of cardio-
vascular disease as possible
confounders in extended models)
Season: NA
Dose-response Investigated?
Statistical Package: SAS v8.0, STATA
v8.0
Median: 13.4
Percentiles:
Quintile ranges:
1:3.4,10.9
2:11.0,12.4
3:12.5, 14.2
4:14.3, 16.4
5:16.5, 28.3
IQR: 11.6-18.3
10th-90th
Personal: 9.1-18.3
City-wide: 9.3-17.8
Range (Min, Max):
Personal exposure: 3.4, 28.3
Citywide exposure: 4.0,19.3
Monitoring Stations: 573 monitors
The nearest monitor to the location of
each residence was used to assign
exposure (monitor within 30 mi of
residence
Median of 20 monitors per city (range:
4-78))
Copollutant (correlation):
PM10
S02
N02
CO
03
Between cities: 1.15(0.99,1.32)
Within cities: 1.64 (1.24, 2.18)
Coronary heart disease (first event):
Overall: 1.21 (1.04,1.42)
Between cities: 1.13 (0.95,1.35)
Within cities: 1.56 (1.11,2.19)
Cerebrovascular disease (first event):
Overall: 1.35 (1.08,1.68)
Between cities: 1.20 (0.94,1.54)
Wthin cities: 2.08 (1.28, 3.40)
Ml (first event):
Overall: 1.06 (0.85,1.34)
Between cities: 0.97 (0.75,1.25)
Wthin cities: 1.52 (0.91, 2.51)
Coronary revascularization (first event):
Overall: 1.20 (1.00,1.43)
Between cities: 1.14 (0.93,1.39)
Wthin cities: 1.45 (0.98, 2.16)
Stroke (first event):
Overall: 1.28 (1.02,1.61)
Between cities: 1.12 (0.87,1.45)
Wthin cities: 2.08 (1.25, 3.48)
Any death from cardiovascular cause:
Overall: 1.76 (1.25, 2.47)
Between cities: 1.63 (1.10, 2.40)
Wthin cities: 2.28 (1.10, 4.75)
Coronary heart disease death (definite
diagnosis):
Overall: 2.21 (1.17, 4.16)
Between cities: 2.22 (1.06, 4.62)
Wthin cities: 2.17 (0.60, 7.89)
Coronary heart disease death (possible
diagnosis): Overall: 1.26 (0.62, 2.56)
Between cities: 1.20(0.54,2.63)
Wthin cities: 1.57 (0.29, 8.51)
Cerebrovascular disease death:
Overall: 1.83 (1.11, 3.00)
Between cities: 1.58 (0.90, 2.78)
Wthin cities: 2.93 (1.03, 8.38)
Estimated Hazard Ratios for
cardiovascular events associated with
an increase in PM25 according to
selected characteristics (presented
adjusted H and adjusted H including
adjustment for city)
Any cardiovascular event:
H: 1.24 (1.09, 1.41)
H (city): 1.69 (1.26, 2.27)
Household income <$20,000:
H: 1.30 (1.10, 1.53)
H (city): 1.75 (1.28, 2.40)
Household income $20,000-49,999:
H: 1.23 (1.08, 1.41)
H (city): 1.69 (1.25, 2.27)
Household income > $50,000:
H: 1.20 (1.02, 1.40)
6
H (city): 1.66 (1.22, 2.26)
P for trend: HR:p = 0.34
HR (city): p = 0.54
Education: Not high-school graduate:
H: 1.40 (1.11, 1.75)
H (city): 1.88 (1.32, 2.67)
Education: High school grad/trade
school/GED: H: 1.33 (1.14, 1.55)
H (city): 1.79 (1.32, 2.44)
Education: Some college or associate
degree: H: 1.26 (1.09,1.44)
December 2009
E-379
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
H(city): 174(1.29,2.34)
Education: Bachelor's degree or higher:
H: 1.11 (0.94, 1.31)
H (city): 1.54 (1.13, 2.10)
P for trend: H:p = 0.07
H (city): p = 0.15
Age <60yr:H: 1.21 (0.84,1.73)
H (city): 1.66 (1.05, 2.61)
Age 60-69 yr: H: 1.14 (0.93,1.39)
H (city): 1.53 (1.09, 2.14)
Age > 70 yr: H: 1.34 (1.11,1.63)
H (city): 1.85 (1.34, 2.56)
P for trend: H:p = 0.20
H (city): p = 0.20
Current smoker: H: 1.68 (1.06, 2.66)
H (city): 2.28 (1.33, 3.92)
Former smoker: H: 1.24 (1.01,1.52)
H (city): 1.71 (1.23, 2.39)
Never smoked: H: 1.18 (0.99,1.40)
H (city): 1.60 (1.16, 2.21)
Living with smoker currently:
H: 1.28 (0.84, 1.97)
H (city): 1.65 (0.99, 2.76)
Living with smoker formerly:
H: 1.18 (1.00, 1.38)
H (city): 1.59 (1.16, 2.16)
Living with smoker never:
H: 1.39 (1.07, 1.80)
H (city): 1.90 (1.31, 2.78)
BMK22.5:H: 0.99 (0.80, 1.21)
H (city): 1.35 (0.96, 1.88)
BMI 22.5-24.7: H: 1.16 (0.96, 1.40)
H (city): 1.58 (1.14, 2.19)
BMI 24.8-27.2: H: 1.24 (1.05, 1.45)
H (city): 1.69 (1.24, 2.30)
BMI 27.3-30.9: H: 1.38 (1.18, 1.61)
H (city): 1.88 (1.38, 2.56)
BMI >30.9:H: 1.35 (1.12, 1.64)
H (city): 1.84 (1.33, 2.55)
P for trend: H:p = 0.003
H (city): p = 0.007
Waist-to-hip ratio <0.74:
H: 1.07 (0.90, 1.29)
H (city): 1.45 (1.05, 2.00)
Waist-to-hip ratio 0.74-0.77:
H: 1.12 (0.95, 1.31)
H (city): 1.51 (1.11,2.06)
Waist-to-hip ratio 0.78-0.80:
H: 1.24 (1.07, 1.44)
H (city): 1.68 (1.23, 2.27)
Waist-to-hip ratio 0.81-0.86:
H: 1.30 (1.13, 1.50)
H (city): 1.76 (1.30, 2.38)
Waist-to-hip ratio >0.86:
H: 1.29 (1.11, 1.50)
H (city): 1.75 (1.29, 2.37)
Waist circumference <73 cm:
H: 1.05 (0.86, 1.27)
H (city): 1.43 (1.02, 1.99)
Waist circumference 73-78 cm:
H: 1.20 (1.02, 1.41)
H (city): 1.63 (1.19, 2.23)
Waist circumference 79-85 cm:
H: 1.22 (1.05, 1.41)
H (city): 1.66 (1.22, 2.24)
Waist circumference 86-95 cm:
H: 1.33 (1.15, 1.53)
H (city): 1.80 (1.33, 2.43)
Waist circumference >95 cm:
H: 1.27 (1.07, 1.51)
H (city): 1.73 (1.26, 2.36)
P for trend: H:p = 0.06
H (city): p = 0.07
Norm one-replacement therapy-Current
Use: H: 1.33 (1.09,1.61)
H (city): 1.85 (1.32, 2.58)
Hormone-replacement therapy-No
Current Use: H: 1.16 (0.98,1.39)
December 2009 E-380
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
H(city): 1.57(1.14,2.17)
Diabetes-yes: H: 0.96 (0.67,1.37)
H (city): 1.24 (0.78, 1.96)
Diabetes-no: H: 1.28 (1.12,1.47)
H (city): 1.75 (1.30, 2.36)
Hypertension-yes: H: 1.22 (1.02,1.45)
H (city): 1.65 (1.09, 2.27)
Hypertension-no: H: 1.26 (1.05,1.51)
H (city): 1.74 (1.25, 2.40)
Hypercholesterolemia-yes:
H: 1.25 (0.94, 1.67)
H (city): 1.71 (1.15, 2.54)
Hypercholesterolemia-no:
H: 1.23 (1.07, 1.42)
H (city): 1.69 (1.25, 2.28)
Family history of CVD-yes:
H (city): 1.80 (1.32, 2.44)
Family history of CVD-no:
H: 1.07 (0.83, 1.37)
H (city): 1.46 (1.00, 2.12)
Time lived in current state: > 20 yr:
H: 1.21 (1.06, 1.39)
H (city): 1.66 (1.23, 2.23)
Time lived in current state: 10-19 yr:
H: 1.39 (1.12, 1.72)'
H (city): 1.97 (1.40, 2.79)
Time lived in current state: < 9 yr:
H: 1.54 (1.06, 2.26)
H (city): 2.24 (1.39, 3.59)
Health insurance coverage-yes:
H: 1.22 (1.07, 1.39)
H (city): 1.71 (1.27,2.30)
Health insurance coverage-no:
H: 1.82 (0.81, 4.10)
H (city): 2.65 (1.12, 6.28)
Time spent outdoors: <30 min:
H: 1.09 (0.86, 1.39)
H (city): 1.56 (1.05, 2.31)
Time spent outdoors: > 30 min
H: 1.26 (1.05, 1.50)
H (city): 1.82 (1.29, 2.57)
Reference: O'Neill etal. (2007,
1560061
Period of Study: 2000-2004
Location: USA (6 field centers:
Baltimore, MD
Chicago, IL
Forsyth Co, NC
Los Angeles, CA
New York, NY
St. Paul, MN
Outcome (ICD9 and ICD10):
Creatinine adjusted urinary albumin
excretion
Assessed 2 ways: continuous log
urinary albumin/creatine ration (UACR)
and clinically defined micro- or macro-
albuminuria (UACR 2 25 mg/g) vs.
normal levels
Age Groups: 44-84 yr
Study Design: Prospective cohort
analyses (MESA cohort)
N: 3901 participants, free of clinical
CVD at baseline
Statistical Analyses: Multiple linear
regression (continuous outcome)
Binomial regression (dichotomous
outcome)
Covariates: Age, gender, race, BMI,
cigarette status, ETS, percent dietary
protein
Season: NA
Dose-response Investigated? Yes,
examined quartiles of exposure
Statistical Package: SAS
Pollutant: PM25 (|jg/m3)
Averaging Time: Avg of previous
month, avg of previous 2 mo (recent
exposures)
20-yr imputed PM25 avg (longer-term
exposures)
Mean (SD): Previous month:
16.5(4.8)
Percentiles: NR
Range (Min, Max): NR
Monitoring Stations: NR (used closest
monitor to residence to assign value for
recent exposures
20-yr PM2 5 exposures were imputed
using a space-time model.)
Copollutant (correlation): PIvl-;
PM Increment: 10 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
Adjusted mean differences in log
UACR (mg/g) per increase in PM25
among participants seen at baseline
Previous 30 days
Full sample: -0.017 (-0.087, 0.052)
Within 10 km: 0.026 (-0.067, 0.119)
Previous 60 days
Full sample:-0.040 (-0.121, 0.042)
Within 10 km:-0.013 (-0.122, 0.097)
Imputed 20 yr exposure
Full sample: 0.002 (-0.048, 0.052)
Wthin 10 km:-0.012 (-0.076, 0.053)
Adjusted relative prevalence of
microalbuminuria vs. high-normal
and normal levels (below 26 mg/g)
per increase in PM2 5 among
participants without
macroalbuminuria during the
baseline visit
Previous 30 days: 0.94 (0.77,1.16)
Previous 60 days: 0.90 (0.71,1.14)
Imputed 20 yr exposure:
0.98(0.84, 1.14)
December 2009
E-381
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Solomon et al. (2003,
1569941
Period of Study: Exposures measures
1966-1969
Health endpoints assessed via
questionnaire, yr not reported but
apparently 30 yr after exposure
assessment (given the 30 yr residency
requirement)
Location: United Kingdom
Outcome (ICD9 and ICD10): Ischemic
heart disease (a self-reported history of
medically diagnosed angina or heart
attack)
Age Groups: 45 yr and older
Study Design: Cross-sectional
N: 1,166 women
Statistical Analyses: Log linear
modeling
Covariates: Smoking, passive smoking
in childhood, tenancy, social class,
worked in industry with respiratory
hazard, childhood hospital admission
for chest problem, diabetes, BMI
Season: NA
Dose-response Investigated? No
Statistical Package: STATA
Pollutant: Black smoke (pg/m )
Averaging Time: Exposure measures
performed 1966-1969
women had to live within 5 miles of their
current address for the past 30 yr to be
included
Mean(SD): 11 wards with pollution
measures were categorized into high
(mean >120 pg/m ) and low (mean
<50 pg/m3) exposure categories when
classified according to their black
smoke levels during 1966-69
SD not reported
Percentiles: NR
Range (Min, Max): NR
Monitoring Stations: NR
Copollutant (correlation): S0; (health
results not presented)
PM Increment: Categorical
Effect Estimate [Lower Cl, Upper Cl]:
Association of particulate pollution in
place of residence and ischemic heart
disease
Low(ref): 1.0
High: 1.0 (0.7, 1.4)
'All units expressed in ug/m3 unless otherwise specified.
December 2009
E-382
-------
E.5. Long-Term Exposure and Respiratory Outcomes
Table E-22. Long-term exposure - respiratory morbidity outcomes - PMio.
Study
Design & Methods
Concentrations
Effect Estimates (95% Cl)
Reference: Ackermann-Liebrich et al.
(1997, 0775371
Period of Study: 1991-1993
Location: Switzerland (Aarau, Basel,
Davos, Geneva, Lugano, Montana,
Payerne, Vteld)
Outcome: Pulmonary function
Age Groups: 18-60yr
Study Design: Cross-sectional
N: 9651 people
Statistical Analyses: Regression
analysis
Covariates: Age, sex, height, weight,
education level, nationality, workplace
exposure
Season: NR
Dose-response Investigated? No
Statistical Package: NR
Pollutant: PM10
Averaging Time: Continuously
measured, 12-mo. avg. used
Mean (SD): 21.2 (7.4)
Range: (10.1-33.4)
Copollutant (correlation):
S02:r = 0.93
N02:r = 0.91
03:r = -0.55
Summer Daytime
03: r = 0.31
Excess 03: r = 0.67
Altitude: r =-0.77
PM Increment: 10 pg/m
Regression Coefficient p (Lower Cl,
Upper Cl) for air pollutants as
predictors of pulmonary function
FVC:-0.0345 (-0.0407 to-0.0283)
p < 0.001
FEV,: -0.0160 (-0.0225 to -0.0095)
p < 0.001
Percent Change (Lower Cl, Upper Cl)
associated with increase in avg
annual air pollution concentration
Healthy Never-smokers
FVC: -3.39
p < 0.001
FEV,:-1.59
p < 0.001
All Never-smokers
FVC:-3.14
p < 0.001
FEV,:-1.06
p < 0.001
Former Smokers
FVC: -3.03
p < 0.001
FEV,:-0.42
Current Smokers
FVC: -3.21
p < 0.001
FEV,:-1.35
p < 0.001
All
FVC:-3.14
p < 0.001
FEV,:-1.03
p < 0.001
Long-term Residents
FVC:-3.16
p < 0.001
FEV,:-0.96
p < 0.001
Reference: Avol et al. (2001, 0205521
Period of Study: 1993-1998
Location: Southern California
Outcome: FVC, FEV,, MMEF, PEFR Pollutant: PM
Age Groups: 10yr
Study Design: cohort
N:110
Statistical Analyses: Linear regression
Covariates: Sex, race, cohort entry yr,
annual avg change in height, weight,
BMI
Dose-response Investigated? No
Averaging Time: 24-h PM,0 avgd over
1994
Mean (SD): 15.0-66.2
PM Increment: 10 pg/m
Mean Change (Lower Cl, Upper Cl)
FVC:-1.8 (-9.1,5.5)
FEV,:-6.6 (-13.5, 0.3)
MMEF: -16.6 (-32.1 to-1.1)
PEFR:-34.9 (-59.8 to-10.0)
December 2009
E-383
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Bayer-Oglesby et al. (2005,
0862451
Period of Study: 1992-2001
Location: Switzerland (Lugano, Zurich,
Bern, Geneva, Anieres, Biel, Langnau,
Payerne, & Montana)
Outcome: Respiratory symptoms
(chronic cough, bronchitis, cold, dry
cough, conjunctivitis, wheeze,
sneezing, asthma, & hay fever)
Age Groups: 6-15 yr
Study Design: Cross-sectional
N: 9,591 children
Statistical Analyses: Logistic
regression models
Covariates: Age, sex, nationality,
parental education, number of siblings,
farming status, low birth weight, breast
feeding, smoking, family history of
asthma, bronchitis and/or atopy, mother
who smokes, indoor humidity, mode of
cooking & heating, carpeting, pets,
removal of carpets/pets for health
reasons, completed questionnaire &
month, days max temperature <0°C,
mother's belief of association between
environmental exposures & respiratory
health
Dose-response Investigated? Yes
Statistical Package: STATA
Pollutant: PM,0
Averaging Time: 12-mo avg
Mean (SD): NR
Range (Min, Max): NR
Monitoring Stations: 9
Copollutant (correlation): NR
PM Increment: 10 pg/m
"Fig 2 shows that declining levels of
PMio were associated with declining
prevalence of chronic cough, bronchitis,
common cold, nocturnal dry cough, and
conjunctivitis symptoms. For wheezing,
sneezing, asthma, and hay fever, no
significant association could be seen
with declining PMio levels."
"Fig 3 illustrates that, on an aggregate
level, across regions the mean change
in PM10 levels (r pearson = 0.81,
p = 0.008). The strongest decline of
adjusted prevalence of nocturnal dry
cough was observed in Geneva,
Lugano, and Anieres, where the
strongest reduction of PMio had also
been achieved."
Reference: Burr et al. (2004, 0878091
Period of Study: 3 wk in Jul and Jan
1997 and 2 wk in Nov 1996 and Apr
1997
Location: North Wfeles, England
Outcome: Self-report of symptoms only Pollutant: PM
for wheeze, cough, phlegm, rhinitis, and
itchy eyes.
Age Groups: all
Study Design: Repeated measures
N: 386 persons in congested streets
and 425 in the uncongested streets in
1996/1997. Of these, 165 and 283
completed the second phase of the
study.
Averaging Time: Mean hourly
concentrations
Mean (SD): SD NR
Congested streets -
1996-9735.2
1998-9927.2
Uncongested Streets
1996-9711.6
1998-998.2
Monitoring Stations: 1 in congested
street and 1 in uncongested
Percent change PM10 in congested
streets: 22.7
Percent change PM10 in
uncongested streets: 28.9
Uncongested street sampling site was
20 m from the congested street
sampler.
The opening of the by-pass produced a
reduction in pollution in the congested
streets. The health effects of these
changed is likely to be greater for nasal
and ocular symptoms than for lower
respiratory symptoms. Uncertainty
about the causality arises from low
response rates and conflicting trends in
respiratory and nasal symptoms.
December 2009
E-384
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Calderon-Garciduenas et
al. (2006, 0912531
Period of Study: 1999, 2000
Location: Southwest Mexico City &
Tlaxcala, Mexico
Outcome: Hyperinflation, interstitial
markings-measured by chest
radiograph, and lung function-FVC,
FB/1, PEF, FEF25-75, measured using
spirometry tests
Age Groups: 5-13 yr
Study Design: Cohort
N: 249 (total), 230 (Southwest Mexico
City), 19 (Tlaxcala)
Statistical Analyses: Bayes test,
Spearman rank correlation, multiple
regression
Covariates: Age, sex
Dose-response Investigated? No
Statistical Package: SAS 8.2
Pollutant: PM,0
Averaging Time: 1 yr
Mean (SD):
Mexico City
1999-48
2000-45
Tlaxacala:
1994-2000:
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Downs et al. (2007,
0928531
Period of Study: 1991, 2002
Location: Switzerland
Outcome: FEV,, FEV, as % of FVC,
FEF25-75
Age Groups: 18-60yr
Study Design: Prospective Cohort
N: 4742 people
Statistical Analyses: Linear random
effects models
Pollutant: PM,0
Averaging Time: Annual
Mean:
Mean interval exposure: 238 pg/m3/yr
Percentiles:
25th: 197
Covariates: Age, sex, height, parental 75th: 287
smoking, season, education, nationality,
occupational exposure, smoking
(status, pack-yr), atopy, BMI
Dose-response Investigated?
Yes-linear fit best
Statistical Package: SAS 9.1, STATA
8.2, R 2.4
PM Increment: 10 pg/m reduction in
annual mean
Percent / absolute reduction in annual
decline in lung function over 11-yr
period (95% Cl):
Annual decline in FEV, reduced by 9% /
3.1 ml (0.03-6.2)
Annual decline in FEF25.75 reduced by
16%/11.3 ml/second (4.3-18.2)
Annual decline in FEVi as a percentage
of FVC of 0.06 (0.01-0.12)
A reduction in interval exposure of 109
pg per m3 cubic meter-yr (equivalent to
a reduction of 10 pg/m in the annual
avg during the mean follow-up time of
10.9 yr) was associated with:
A reduction of 6.9 ml (95% Cl, 2.1 to
11.7) in the annual decline in FEV,
A 22% reduction in the annual decline
in FEF25-75 (i.e., by 14.0 ml per
second 95% Cl, 3.1 to 24.8)
Reference: Gauderman et al. (2000,
0125311
Period of Study: 1993-1997
Location: Southern California
Outcome: FVC, FEV,, MMEF, FEF75 Pollutant: PM10
Age Groups: Fourth, seventh, or tenth Averaging Time: 24-h avg PM10
araders
Mean(SD):PMi051.5
Study Design: Cohort
Copollutant (correlation):
N: 3035 subjects PM25r = 0.96
Statistical Analyses: Linear regression 03 r = -0.32
Covariates: Height, weight, BMI, PM10.25 r = 0.92
asthma, smoking, exercise, room
temperature, barometric pressure N°2r - °'65
Dose-response Investigated? Yes Inorg. Acid r = 0.68
Statistical Package: SAS
PM10 Increment: 51.5 pg/m
% Change (Lower Cl, Upper Cl)
PM10-4th grade
FVC -0.58 (-1.14to-0.02)
FEV,-0.85 (-1.59 to-0.10)
MMEF-1.32 (-2.43 to-0.20)
FEF75-1.63(-3.14to-0.11)
PM10-7th grade
FVC-0.45 (-1.03, 0.13)
FEV,-0.44 (-1.10, 0.23)
MMEF-0.48 (-2.51, 1.59)
FEF75-0.50 (-2.26, 1.29)
PM10-10th grade
FVC 0.07 (-0.99, 1.13)
FEV,-0.46 (-1.84, 0.94)
MMEF-0.71 (-4.87,3.63)
FEF75-1.54 (-5.61, 2.71)
Reference: Gauderman et al. (2002,
0260131
Period of Study: 1996-2000
Location: Southern California
Outcome: Lung function development:
FEV,, maximal midexpiratory flow
(MMEF)
Age Groups: Fourth grade children
(avg age = 9.9 yr)
Study Design: Cohort study
N: 1678 children, 12 communities
Statistical Analyses: Mixed model
linear regression
Covariates: Height, BMI, doctor-
diagnosed asthma and cigarette
smoking in previous yr, respiratory
illness and exercise on day of test,
interaction of each of these variables
with sex, barometric pressure,
temperature at test time, indicator
variables for field technician and
spirometer
Dose-response Investigated? Yes
Statistical Package: SAS (10)
Pollutant: PM,0
Averaging Time: Annual 24-h avg
Mean (SD): The avg levels were
presented in an online data supplement
(FigE1)
Monitoring Stations: 12
Copollutant (correlation):
03(10AMto6PM)r = 0.13
03r = -0.37
N02r = 0.64
Acid vapor r = 0.79
PM25r = 0.95
PM,0.25r = 0.95
EC r = 0.86
OCr = 0.97
PM Increment: 51.5 pg/m
Association Estimate:
None of the pulmonary function tests
had a statistically significant correlation
with PM10
FEV,r = -0.12p = 0.63
MMEF r = -0.22 p = 0.30
December 2009
E-386
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Gauderman et al. (2004,
0565691
Period of Study: Air pollution data
ascertainment: 1994-2000. Spirometry
testing: Spring 2001-Spring 2003
Location: 12 Communities in Southern
California
Outcome: Lung function
FVC, FEV,, MMEF (Maximal
midexpiratory flow rate)
Age Groups: Children, Avg age 10 yr
Study Design: Prospective Cohort
Study
N: 12 Communities
2,034 Children
24,972 child-months
Statistical Analyses: Linear regression
of changes in sex-and-community
specific lung growth function and PM
Covariates: Random effect for
communities
Season: ALL (except for PM25)
Dose-response Investigated? No
Statistical Package: SAS
Pollutant: PM,0
Averaging Time: 24-h measurements
over each yr used to create annual avg
Mean: Means are presented in figures
only.
Range (Min, Max):-15,-65
Monitoring Stations: 12
Copollutant (correlation):
03:r = 0.18
N02:r = 0.67
PM25:r = 0.95
EC: r = 0.85
OC:r = 0.97
PM Increment: Most to least polluted
community
Range:
PMi0:51.4|jg/m3
EC:1.2|jg/nr
OC: 10.5 pg/m3
Difference in Lung Growth [Lower Cl,
Upper Cl];
FVC-60.2 (-190.6 to 70.3)
FEV,-82.1 (-176.9(012.8)
MMEF-154.2 (-378.3 to 69.8)
EC:
FVC-77.7 (-166.7to 11.3)
FEV,-87.9 (-146.4 to-29.4)
MMEF-165.5 (-323.4 to-7.6)
OC:
FVC-58.6 (-196.1 to 78.8)
FEV,-86.2 (-185.6 to 13.3)
MMEF-151.2 (-389.4 to 87.1)
Correlation with % below 80% predicted
Lung function (p-value)
PM,0: 0.66 (0.02)
EC: 0.74 (0.006)
Reference: Gauderman et al. (2007,
0901211
Period of Study: 1993-2004
Location: 12 Southern California
Communities
Outcome: pulmonary function tests Pollutant: PM,0
FVC, FEV,, MMEF/FEF25.75
Monitoring Stations: 1 in each
Age Groups: Children (mean age 10 at community
recruitment, followed for 8 yr)
Study Design: Cohort Study
(Children's Health Study)
N: 3677 children
(1718 in cohort 1 recruited 1993 and
1959 in cohort 2 recruited 1996)
22686 pulmonary function tests.
Statistical Analyses: Hierarchical
mixed effects model with linear splines
Covariates: Adjustments for height,
height squared, BMI, BMI squared,
present asthma status, exercise or
respiratory illness on day of test,
smoking in previous yr, field technician,
traffic indicator (distance from freeway,
distance from major roads), random
effects for participant and community.
Dose-response Investigated? no
Statistical Package: SAS
PM Increment: 51.4 pg/m
Pollutant effect reported as difference in
8 yr lung function growth from least to
most polluted community. Negative
difference indicates growth deficits
associated with exposure. For PM10
FEV growth deficit is-111
December 2009
E-387
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Goss et al. (2004, 0556241
Period of Study: 1999-2000
Location: USA
Outcome: Cystic Fibrosis pulmonary Pollutant:
exacerbations, FEVi
Age Groups: > 6
Study Design: Cohort
N: 11484 patients
Statistical Analyses: Logistic
regression, t-tests, Mann-Whitney tests,
Chi-squared tests, polytomous
regression, multiple linear regression
Covariates: Age, sex, lung function,
weight, insurance status, pancreatic
insufficiency, airway colonization,
genotype, median household income by
census tract, zipcode.
Dose-response Investigated? No
Statistical Package: STATA, SAS
PM Increment: 10 pg/m
Averaging Time: Annual mean of 24-h Odds Ratio Estimate [Lower Cl,
avg
Mean (SD): 24.8(7.8) mg/m3
Percentiles: 25th: 20.3
SOth(Median): 24.0
75th: 28.9
Monitoring Stations: 626
Upper Cl]:
Odds of having 2 or more pulmonary
exacerbations as compared to 1 or less
in 2000
1.08(1.02-1.15)
Odds of having 2 or more pulmonary
exacerbations as compared to no
exacerbations in 2000
1.09(1.02-1.17)
Decrease in FEV, 38ml(18-58)
Reference: Hanigan et al, (2008,
1565181
Period of Study: Fire Season (Apr-
Nov) from 1996-2005
Location: Darwin, Australia
Outcome: Respiratory admissions
Study Design: Time-series
Covariates: Race, age
Statistical Analysis: Over-dispersed
Poisson generalized linear models
Statistical Package: R
Age Groups: All
Pollutant: PM10
Averaging Time: Daily levels estimated
from visibility data
Mean Unit: "Only reported for 2005*
15.31 pg/m3
Range (Min, Max): 6.93, 31.12
Copollutant (correlation): NR
Increment: 10 pg/m
Percent Increase (96% Cl)
*Full results reported visually in Fig 3*
Total Respiratory Admissions
4.81% (-1.04-11.01)
Indigenous Respiratory Admissions, No
9.40% (1.04-18.46)
Non-Indigenous Respiratory
Admissions, No Lag
3.14% (-2.99-9.66)
Indigenous Respiratory Admissions,
Lag3
15.02% (3.73-27.54)
Non-Indigenous Respiratory
Admissions, Lag 3
0.67% (-7.55-9.61)
Indigenous Asthma Admissions, Lag 1
16.27% (3.55-40.17)
Non-Indigenous Asthma Admissions,
Lag1
8.54% (-5.60-24.80)
Reference: Ho et al. (2007, 0932651
Period of Study: Oct 1995-Mar 1996
Location: Taiwan, Republic of China
Outcome: Asthma
Age Groups: 10-17 yr
Study Design: Screened junior high
students for asthma, collected
meteorological data to determine the
relationship.
N: 69,367
Statistical Analyses: Logistic
regression model, the maximum
likelihood estimation with Fisher's
scoring algorithm, stepwise regression
model, Wald statistic, Akaike criteria.
GEE, GENMOD
Covariates: Wind, barometric pressure,
temperature, rain, humidity
Season: Fall-spring
Dose-response Investigated? No
Statistical Package: SAS
Pollutant: PM,0
Averaging Time: Monthly
Monitoring Stations: 72
Odds Ratio from stepwise regression
model:
Females (n = 32, 648)
0.993 [0.990-0.997]
Males: NS
Higher PM10 concentration resulted in
less asthma prevalence. However, a
higher number of rain days seemed to
reduce asthma prevalence
Rain days might interact with PMi0.
December 2009
E-388
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Hong et al. (2004, 1565651 Outcome: Respiratory symptoms
„ . J .^ J „„„., . „ ,„
Period of Study: 2001 Age Groups: <12 yr
Location: Kerinci, SP7, and Pelalawan, Study Design: Disproportionate
Indonesia random sampling was used to select
100 households from each village. An
interviewer interviewed all children
through the caregiver/parent to obtain
symptoms in the past 2 wk (cough, cold,
nhlenmlanrithpla«M9mn '
phlegm) and the last 12 mo.
N: 382 children
Statistical Analyses: Chi-square test,
analysis of variance, prevalence rates,
adjusted odds ratios, multivariate
adjusted odds ratios from multiple
logistic regression models, allowing for
clustering.
_ .. , , ,
Covanates: Age, gender, no. of
Pollutant: PMi0
. „ „„, t
Averaging Time: 24-h measurements
were taken daily from 2 wk before he
fie|d survey to 1 mo after the survey
Mean(SD):
v . '„„„„ , ,
Kerinci 102.9 (49.6) pg/m3
CD7 7, -, ,M 7>
SP7 73.7 41.7
-. , M, „
Pelalawan 26'1 <145'
p
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Horaket al. (2002, 0347921 Outcome:
Period of Study: 1994-1997
Location: Lower Austria
Lung function growth measured by
changes in:
1. FVC (forced vital capacity)
2. FEV,
3. MEF25-75 (midexpiratory flow
between 25-75% of the forced vital
capacity)
Age Groups: 2-3 grade schoolchildren
(mean age = 8)
Study Design: Prospective cohort with
repeated measures
N: 975 children
Statistical Analyses: Linear regression
GEE, nonstationary M-dependent
correlation structure
Covariates: Gender, atopy, ETS
exposure, baseline lung function, first
height, height difference, school site
Season: Winter, summer
Dose-response Investigated? No
Pollutant: PM,0
Mean (SD):
Winter: 21.0(4.8)
Summer: 17.4 (2.8)
Range (Min, Max):
Wnter: 9.4-30.5
Summer: 11.7-28.9
Monitoring Stations:
NR, stations were located in the
immediate vicinity of each of the 8
elementary schools
Copollutant (correlation):
Wnter
03:(r =-0.581)
S02(r = 0.520)
N02(r = 0.595)
Summer
03(r =-0.429)
S02(r = 0.335)
N02(r = 0.412)
PM Increment: 1 pg/m
Mean per unit increase in PM (p-value)
Outcome: difference per day of FVC
(mL/day)
Summer: 0.001 (0.938)
Wnter: 0.008 (0.042)
Controlling for temperature:
Summer:-0.007 (0.417)
Wnter: -0.003 (0.599)
Controlling for 03:
Summer: 0.001 (0.911)
Wnter: 0.010 (0.019)
Controlling for N02:
Summer:-0.018 (0.056)
Wnter: 0.015 (0.000)
Controlling for S02:
Summer: 0.005 (0.575)
Wnter: 0.004 (0.492)
In non-asthmatic children:
Summer:-0.003 (0.710)
Wnter: 0.009 (0.030)
In group not exposed to ETS:
Summer: 0.014 (0.154)
Wnter: 0.012 (0.0018)
In group exposed to ETS:
Summer: 0.022 (0.088)
Wnter: 0.003 (0.656)
Outcome: difference per day of FEV,
(mL/day)
Summer: -0.023 (0.003)
Wnter: 0.001 (0.885)
Controlling for temperature:
Summer: -0.034 (0.000)
Wnter: -0.011 (0.016)
Controlling for 03:
Summer: -0.022 (0.008)
Wnter: 0.004 (0.338)
Controlling for N02:
Summer: -0.038 (0.000)
Wnter: 0.011 (0.005)
Controlling for S02:
Summer:-0.022 (0.010)
Wnter: -0.005 (0.358)
Outcome: difference per day MEF25.75
(mL/day)
Summer: -0.090 (0.000)
Wnter: -0.008 (0.395)
Controlling for temperature:
Summer:-0.112 (0.000)
Wnter:-0.013 (0.295)
Controlling for 03:
Summer: -0.087 (0.000)
Wnter: -0.008 (0.434)
Controlling for N02:
Summer:-0.102 (0.000)
Wnter: 0.005 (0.610)
Controlling for S02:
Summer: -0.095 (0.000)
Wnter: -0.011 (0.474)
December 2009
E-390
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Hwang et al. (2006,
0889711
Period of Study: 2001
Location: Taiwan
Outcome: Peak expiratory flow rate
(PEFR), Forced Expiratory Volume in 1
second (FB/i), Forced Vital Capacity
(FVC), Self reported "frequent
coughing," Self reported "shortness of
breath," Self reported" irritation of
respiratory tract"
Age Groups: 24-55 yr (mean = 40)
Study Design: Cohort
N: 120 men (60 traffic policemen and
60 controls)
Statistical Analyses: ANOVA, odds
ratios calculated from 2X2 table
Dose-response Investigated? No
Pollutant: PM,0
Mean (SD): 55.58 (16.57)
Percentiles: 25th: 42.96
SOth(Median): 53.81
75th: 70.37
Range (Min, Max): 29.36, 99.58
Monitoring Stations: 22
Copollutant (correlation):
N0x(r = 0.34)
S02(r = 0.58)
CO (r = 0.27)
03(r = 0.28)
PM Increment: 10 pg/m
RR Estimate [Lower Cl, Upper Cl]
Single pollutant model: 1.00 [0.99,1.02]
Controlling for NOX: 0.99 [0.97,1.00]
Controlling for CO: 1.00 [0.99,1.01]
Controlling for 03:1.00 [0.99,1.02]
Reference: Hwang et al, (2008,
1344201
Period of Study: 2001-2003
Location: Taiwan
Outcome: Oral Cleft
Study Design: Case-control
Covariates: Maternal age, plurality,
gestational age, population density and
season of conception
Statistical Analysis: Logistic
regression
Age Groups: Infants
Pollutant: PM,0
Averaging Time: hourly
Mean (SD) Unit:
Avg: 54.83+13.07 pg/m3
Spring: 64.44+ 16.21 pg/m3
Summer: 39.11 ±8.31 pg/m3
Fall: 47.76 ±11.77 pg/m*
Winter: 68.00 ±21.88 pg/m3
Range (Min, Max):
Avg: 20.75-78.05 pg/m3
Spring: 23.33-94.33 pg/m3
Summer: 17.33-60.00 ug/m3
Fall: 21.00-72.00 pg/m3
Winter: 21.33-116.00 pg/m3
Copollutant (correlation):
CO:-0.19
NOX: 0.56
03: 0.39
SO,: 0.50
Increment: 10|jg/m
Odds Ratio (Min Cl, Max Cl);
Single Pollutant Model
Month 1:1.01 (0.96-1.06)
Month 2:1.00 (0.95-1.05)
Month 3: 0.99 (0.95-1.05)
Two Pollutant Model (03 + PM,C
Month 1:0.99 (0.94-1.04)
Month 2: 0.99
Month 3: 0.98
0.94-1.04
0.93-1.04
Two Pollutant Model (CO+PM,o)
Month 1:1.01 0.96-1.06
Month 2:1.00 0.95-1.05
Month 3: 0.99 (0.95-1.05)
Two Pollutant Model (NOx+PM,o)
Month 1:1.02 (0.97-1.08)
Month 2:1.01 (0.95-1.07)
Month 3:1.01 (0.95-1.07)
Three Pollutant Model (03 + CO +
PM,0)
Month 1:0.99
Month 2: 0.99
0.94-1.04
0.94-1.04
Month 3: 0.99 (0.93-1.04)
Three Pollutant Model (03 + NOX +
PM,0)
Month 1:1.00 (0.94-1.06)
Month 2: 0.98 0.92-1.05
Month 3:1.00 0.93-1.06
December 2009
E-391
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ingle et al. (2005, 0890141
Period of Study: May 2003-Apr 2004
Location: Jalgaon City, India
Outcome: Peak expiratory flow rate
(PEFR), Forced Expiratory Volume in 1
second (FB/i), Forced Vital Capacity
(FVC), Self reported "frequent
coughing," Self reported "shortness of
breath," Self reported" irritation of
respiratory tract" Age Groups: 24-55 yr
(mean = 40)
Study Design: Cohort
N: 120 men (60 traffic policemen and
60 controls)
Statistical Analyses: ANOVA, odds
ratios calculated from 2X2 table
Dose-response Investigated? No
Pollutant: PM,0
Mean (SD): Location-specific means:
Prabhat: 224 (27)
Ajanta:269(41)
Icchdevi: 229 (24)
Monitoring Stations: 3
OR Estimate [p-value]
Self reported frequent coughing
2.96 [p < 0.05]
Self reported shortness of breath
1.22 [p< 0.05]
Self reported irritation in respiratory
tract
7.5 [p < 0.05]
Observed/expected lung function
p-value for difference between groups:
FVC(L)
Traffic policemen: 0.82
Controls: 0.99
Traffic policemen:
Obs = 3.03±1.7Exp = 3.70±2.8
Controls:
Obs = 3.18 +0.91 Exp = 3.19 ±1.71
FEV, (L)
Traffic policemen: 0.73
Controls: 1.18
Traffic policemen:
Obs = 2.27 ± 1.05 Exp = 3.08 ±2.7
Controls:
Obs = 3.61 ± 0.90 Exp = 3.06 ±0.91
PEFR (L/s)
Traffic policemen: 0.66
Controls: 0.92
Traffic policemen:
Obs = 6.05 ± 2.15 Exp = 9.21 ±0.47
Controls:
Obs = 5.54 ±1.85 Exp = 6.11 ±2.31
Reference: Islam et al. (2007, 0906971 Outcome: Respiratory symptoms,
Period of Study: 2006
Study Design: Longitudinal study
Location: 12 California communities cohort
Statistical Analyses: Cox proportional
hazards regression
Age Groups:
7-9
10-11
Pollutants: PM10
Averaging Time: 24-h avg
Copollutants (correlation):
03
N02
EC
OC
The study doesn't present quantitative
results on PMi0.
December 2009
E-392
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Janssen et al. (2003,
1335551
Period of Study: Apr 1997-Jul 1998
Location: Netherlands-24 schools
Outcome: Symptoms of asthma and
allergic disease (asthma, conjunctivitis,
hay fever, itchy rash, eczema, phlegm,
bronchitis), skin prick test (SPT)
reaction to allergens, lung function
(forced vital capacity [FVC], forced
expiratory volume in 1 second [FEVi],
and positive test for fall in FEV, > 15%
after inhalation of maximal 23 ml
hypertonic saline [BHR = bronchial
hyper-responsiveness])
Age Groups: 7-12 yr old
Study Design: Cohort
N: 24 schools (see notes)
Statistical Analyses: Multilevel model
Covariates: Age, sex, non-Dutch
nationality, cooking on gas, current
parental smoking, current pet
possession, parental education level,
number of persons in the household,
presence of an unvented water heater
in kitchen, questionnaire not filled out
by the mother, presence of mold stains
in kitchen or living room or bedroom,
parental respiratory symptoms, distance
of home to motorway, cough or cold at
time of lung function measurement,
bronchitis or severe cold or flu in 3 wk
preceding measurement, season
Dose-response Investigated? No
Statistical Package: MLwiN
Pollutant: PM25
Averaging Time: Annual
Mean (SD): 20.5 pg/m3 (2.2)
Percentiles:
25th: 18.6
50th (Median): 20.4
75th: 22.1
Range (Min, Max):
17.3, 24.4
PM Increment: 'Difference between the
maximum and the minimum of the
exposure indicator' (3.5 pg/m3)
RR Estimate [Lower Cl, Upper Cl]
lag:
Current wheeze 1.51 (0.90, 2.53)
Asthma ever 1.03 (0.59,1.82)
Current conjunctivitis 2.08 (1.17, 3.71)
Hay fever ever 2.28 (1.13, 4.57)
Current itchy rash 1.63 (0.91, 2.89)
Eczema ever 1.31 (0.94,1.83)
Current phlegm 1.53 (0.74, 3.19)
Current bronchitis 1.71 (0.84, 3.50)
Elevated total IgE 1.45 (0.74, 2.84)
Any allergen (spt reactivity) 1.33 (0.83,
2.11)
Indoor allergens (spt reactivity) 1.17
(0.70,1.94)
Outdoor allergens (spt reactivity) 1.90
(1.06, 3.40)
FVC < 85% predicted 0.54 (0.29,1.00)
FEV, < 85% predicted 0.88 (0.37, 2.09)
BHR 0.93 (0.51, 1.68)
Notes:
Fig 1 of the article illustrates the
association between exposures,
including PM25, and various respiratory
symptoms among children with and
without a positive SPT and positive
BHR. In general, the association
between PM2 5 and respiratory
symptoms were higher for children with
a positive SPT or BHR, except for the
outcome of current phlegm. This effect
appeared to be the strongest for
children with a positive BHR,
particularly for current wheeze and
current bronchitis.
The authors also reported separate
analyses for children with SPT reactivity
for indoor and outdoor allergens, but did
not report any clear differences
between the two groups. The authors
did report, in the text, that the OR of
PM25 exposure for children sensitized
for outdoor allergens was 7.64 for
current itchy rash (p < 0.05).
December 2009
E-393
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Kan, et al. (2007, 0913831
Period of Study: 1987-1992
Location: Four Communities in the
U.S.: Forsyth County, North Carolina
Jackson, Mississippi
northwest suburbs of Minneapolis,
Minnesota
and Washington County, Maryland.
Outcome: FEV, and FVC
Pollutant: PM,(
Age Groups: Middle-aged (mean age Averaging Time: 24-h PM10 averaged
was 54.2 yr) over study period
Study Design: Hierarchical regression PM Component: Vehicle emissions
N: 15,792 Monitoring Stations: 0
Copollutant:
Statistical Analyses: SAS PROC
MIXED
Covariates: Distance to major roads,
traffic exposure, age, ethnicity, sex,
smoking, environmental tobacco smoke
exposure, occupation, education,
medical history, BMI.
Dose-response Investigated? No
Statistical Package:
SPSS Version 11 for traffic density,
SAS Version 9.1.2 for statistical
analysis
N02
03
RR Estimate (Lower Cl, Upper Cl):
(Note: for ARIC participants living <150
meters from major roads)
Wfomen
FEV,(mL)
Age-adjusted model
-29.5 (-52.2 to -6.9)
Multivariate model
-15.7 (-34.4 to-2.9)
FVC (ml)
Age-adjusted model
-33.2 (-60.4 to -5.9)
Multivariate model
-24.2 (-46.2,-2.3)
FEV,/FVC (%)
Age-adjusted model
-0.1(-0.5,0.2)
Multivariate model
0.1 (-0.3,0.4)
Men
FEV,(mL)
Age-adjusted model
-38.4 (-76.7,0.6)
Multivariate model
-6.4 (-38.1,25.3)
FVC (ml)
Age-adjusted model
-17.0(-62.0,28.0)
Multivariate model
10.9(-24.7,46.5)
FEV,/FVC (%)
Age-adjusted model
-0.05 (-0.9,0.0)
Multivariate model
-0.3 (-0.7,0.2)
Reference: Kim et al. (2005, 0874181
Period of Study: Mar and Dec 2000
Location: Incheon & Ganghwa, Korea
Outcome: Lung function (FEV,, FVC)
Age Groups: Middle school students
Study Design: Panel
N: 368 children
Statistical Analyses: Generalized liner
model
Covariates: Gender, grade
Season: Spring and fall
Dose-response Investigated? No
Statistical Package: SAS
Pollutant: PM,0
Averaging Time: Monthly
Mean (SD):
Incheon
Mar 64
Dec 54
Ganghwa
Mar 64
Dec 53
Range (Min, Max): NR
PM Increment: NR
OR Estimate [Lower Cl, Upper Cl]:
"The present study showed that the
values of FEVi and FVC were greater in
Dec than in Mar for both male and
female students at all academic
yr... Because only the level of PMi0 was
significantly higher for Mar than for Dec
in both areas, the authors suggest that
decrements of pulmonary function in
Mar for both areas are associated with
the increased level of PM10"
Reference: Kim et al. (2004, 0873831
Period of Study: Mar-Jun (spring)
2001
Sep-Nov (fall) 2001
Location: Alameda County, CA
Outcome: Asthma, bronchitis
Age Groups: Children (in grades 3-5)
Study Design: Cross-sectional
N: 1109 children, 871 (long term
resident children), 462 (long term
related females), 403 (long term related
males)
Statistical Analyses: 2-stage multiple
logistic regression model
Covariates: Respiratory illness before
age of 2, household mold/moisture,
pests, maternal history of asthma (for
asthma) Season: Spring and fall
Dose-response Investigated? Yes
Statistical Package: SAS 8.2
Pollutant: PM,0
Averaging Time: 9 wk
Mean (SD): Study Avg 30
Monitoring Stations: 10
Copollutant (correlation): r2 is
approximately 0.9 for all copollutants
BC, PM25, NOX, N02, NO (NOX-N02)
PM Increment: 1.4 (IQR)
OR Estimate [Lower Cl, Upper Cl]:
Bronchitis
All subjects: 1.03 [0.99,1.07]
LTR subjects: 1.02 [0.98, 1.07]
LTR females: 1.04 [1.01,1.09]
LTR males: 1.01 [0.95,1.06]
Asthma
All subjects: 1.02 [0.96,1.09]
LTR subjects: 1.04 [0.97,1.12]
LTR females: 1.09 [0.92,1.29]
LTR males: 1.02 [0.94,1.10]
Asthma excluding outlier school having
a larger proportion of Hispanics
All subjects: 1.06 [0.97,1.16]
LTR subjects: 1.08 [0.98,1.19]
LTR females: 1.09 [0.96,1.24]
LTR males: 1.08 [0.97,1.19]
December 2009
E-394
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Kumar et al. (2004,
0898731
Period of Study: 1999-2001
Location: Mandi Gobindgarh and
Morinda, Punjab State, northern India
Outcome: Chronic respiratory
symptoms & Spirometric ventilatory
defect
Age Groups: >1 Syr
Study Design: Cross-sectional
N: 3603 individuals
Statistical Analyses: Logistic
regression
Covariates: Age, gender, migration,
SES, smoking, type of cooking fuel use
Dose-response Investigated? No
Pollutant: PM,0
Mean(SD): Study town 112.8 (17.9)
Reference town 75.8 (2.9)
PMio Increment:
Low vs. High
OR (Lower Cl, Upper Cl)
p-value
Chronic respiratory symptoms
LowlOO(ref)
High 1.5 (1.2, 1.8)
<0.001
Spirometric ventilatory defect
LowlOO(ref)
High 2.4 (2.0-2.9)
O.001
Reference: Leonard! et al. (2000,
0102721
Period of Study: 1996
Location: 17 cities of Central Europe
(Bulgaria, Czech Republic, Hungary,
Poland, Romania, Slovakia)
Outcome: Immune biomarkers
Age Groups: 9-11
Study Design: Cross-sectional
N: 366 school children
Statistical Analyses: Linear regression
Covariates: Age, gender, parental
smoking, laboratory of analysis, recent
respiratory illness
Dose-response Investigated? No
Statistical Package: STATA
Pollutant: PM,0
Averaging Time: Annual PMio
Mean(SD):PM10:65(14)
Range (Min, Max):
PM10:(41,96)
5th, median, & 95th percentile
PMi0:41,63, 90
% Change (Lower Cl, Upper Cl)
p-value
PM10
Neutrophils -5 (-33, 36)
>20
Total lymphocytes 20 (-6, 54);. 150
B lymphocytes 42 (-3,107); .067
Total T lymphocytes 30 (-2, 73); .072
CD4+28(-10, 82);.177
CD8+ 29 (-5, 75); .097
CD4/CD8 7 (-20, 43)
>20
NK 33 (-10, 97); .157
Total IgG 11 (-10, 38)
>20
Total IgM 5 (-21, 39)
>20
Total lgA11 (-16, 46)
>20
Total IgE-8 (-62, 123)
>20
Reference: Lichtenfels et al, (2007,
0970411
Period of Study: 2001 -2003
Location: Sao Paulo, Brazil
Reference: Lubinski, et al. (2005,
087563)
Period of Study: 1993-1997
Location: Poland
Outcome: Secondary sex ratio
Study Design: Retrospective Cohort
Covariates: NR
Statistical Analysis: Correlation
Coefficient
Age Groups: Infants
Outcome: Pulmonary function
TLC: total lung capacity
ITGV: interthoracic gas volume
ITGV%TLC: ITGV percent total lung
cspscity
Raw: airway resistance
FVC: forced vital capacity
FEV,: forced expiratory volume, 1
sscond
FEVi%FVC: FEV, percent forced vital
cspscity
PEF: peak expiratory flow
FEF50: forced expiratory flow
Age Groups: 18-23 males, healthy
Study Design: Ecological cross-
sectional study
N: 1278 subjects
Statistical Analyses: Multiple linear
regression, ANOVA
Covariates: Report unclear on whether
or not there was covariate control, but
may include N02 and S02
Dose-response Investigated? No
Pollutant: PM,0
Averaging Time: Annual
Mean (SD) Unit:
2001: 49.8 (10.5) pg/m3
2002: 48.5 (11.4) pg/m3
2003: 49.4 (14.4) pg/m3
Range (Min, Max): 31.71-60.96 pg/m3
Copollutant (correlation): NR
Pollutant: PM10
Averaging Time: 12 mo
Mean (SD):
A: Highest Pollution Region
Katowice 67-1 25
Krakow 41 -49
B: Moderate Pollution Region
Bielsko-Biala 29-48
Opole 18-45
Lodz 23-38
Warsaw 35-45
Wroclaw 28-76
Zagan 5-35
C: Lowest Pollution Region
Gizycko5-18
Hel 12-18
Ostroda 23-33
Swinoujscie7-16
Ustka 12-26
Copollutant: N02, S02
Increment: NR
Correlation Coefficient:
R2 = 0.7642, P = 0.1 3
PM Increment: 1 pg/m3
Slope, multiple regression
TLC FEV,
PM10: -0.05 PM10: 0.031
+S02: 0.03 +S02: -0.08
+N02:-0.06 +N02:-0.12
ITGV FEV,%FVC
PM10:0.01 PM10:0.00
+S02:-0.07 +S02:-0.14
+N02: -0.07 +N02: -0.048
ITGV%TLC PEF
PMi0:-0.06 PM,0:-0.18
+S02: 0.08 +S02: 0.056
+N02: 0.00 +N02: -0.09
Raw FEF50
PM10: 0.075 PM10: 0.031
+S02:-0.08 +S02:-0.11
+N02: 0.127 +N02:-0.04
FVC
PM,0: 0.045
+S02: 0.045
+N02:-0.14
December 2009
E-395
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: McConnell et al. (1999,
0070281
Period of Study: 1993
Location: Southern California
Outcome: Bronchitis, chronic cough,
phlegm
Age Groups: Children: 4th, 7th, & 10th
graders
Study Design: Cross-sectional
N: 3676 people
Statistical Analyses: Logistic
regression
Covariates: Age, sex, race, grade,
health insurance
Dose-response Investigated? Yes
Pollutant: PM,0
Averaging Time: Yearly avg 24 h PM10
Mean (SD): 34.8
Range (Min, Max): 13.0, 70.7
Copollutant (correlation):
N02r = 0.74
03r = 0.32
Acid r = 0.54
PM25r = 0.90
N02r = 0.83
03r = 0.50
Acid r = 0.71
PMio Increment: 19 pg/m
Children w/ asthma
Bronchitis: 1.4 (1.1,1.8)
Phlegm: 2.1 (1.4,3.3)
Cough: 1.1 (0.8,1.7)
Children w/wheeze, no asthma
Bronchitis: 0.9 (0.7,1.3)
Phlegm: 0.9 (0.6,1.4)
Cough: 1.2 (0.9,1.8)
Children w/ no wheeze, no asthma
Bronchitis: 0.7 (0.4,1.0)
Phlegm: 0.8 (0.6,1.3)
Cough: 0.9 (0.7,1.2)
Reference: McConnell et al. (2003,
0494901
Outcome: Bronchitis symptoms
Age Groups: 9-19
Period of Study: 1993-1999
Study Design: Communities selected
Location: 12 Southern CA communities on basis of historic levels of criteria
pollutants and low residential mobility.
N: 475 children
Statistical Analyses: 3 stage
regression combined to give a logistic
mixed effects model
Covariates: Sex, ethnicity, allergies
history, asthma history, SES, insurance
status, current wheeze, current
exposure to ETS, personal smoking
status, participation in team sports, in
utero tobacco exposure through
maternal smoking, family history of
asthma, amount of time routinely spent
outside by child during 2-6 pm.
Dose-response Investigated? No
Statistical Package: SAS Glimmix
macro
Pollutant: PM,0
Averaging Time: 4-yr avg
Mean (SD):.30.8(13.4) pg/m3
Range (Min, Max): 15.7-63.5
PM Component: particulate OC and EC
Copollutant (correlation):
PM25:r = 0.79
PM10.25:r = 0.79
Inorganic acid: r = 0.72
Organic Acid: r = 0.59
EC: r = 0.71
OC:r = 0.70
N02:r = 0.20
03:r = 0.64
PM Increment:
Between community range 47.8 pg/m3
Between community unit 1 pg/m3
Within community 1 pg/m3
OR Estimate [Lower Cl, Upper Cl]
Between community per range
1.72(0.93-3.20)1
Between Community per unit 1.01(1.00-
1.02)|
Within community per unit 1.04(0.99-
1.10)
Reference: McConnell et al. (2002,
0231501
Period of Study: 1993-1998
Location: 12 communities in Southern
California (grouped into either high and
low pollution communities)
Outcome: Asthma (new diagnosis)
Age Groups: 9-12 yr, 12-13 yr, 15-16 yr
Study Design: Cohort
N:3535
Statistical Analyses: Multivariate
proportion hazard model
Covariates: Sex, age, ethnic origin,
BMI, child history of allergies and
asthma history, SES, maternal smoking,
time spent outside, history of wheezing,
ownership of insurance (yes/no),
number and type of sports played
Dose-response Investigated? Yes
Statistical Package: SAS 8.1
Pollutant: PM,0
Averaging Time: 4 yr
Mean (SD): Low pollution communities:
21.6(3.8)
High pollution communities: 43.3
(12.0)
Percentiles: Low pollution
communities: SOth(Median): 20.8
High pollution communities:
SOth(Median): 43.3
Range (Min, Max): Low pollution
communities: 16.62, 27.3
High pollution communities: 33.5, 66.9
Monitoring Stations: 12
Copollutant (correlation):
PM25:r = 0.96
N02:r = 0.65
03
RR Estimate [Lower Cl, Upper Cl]
lag:
Low PM communities: 1.0 [ref] 0 sport
1.5 [1.0, 2.2] 1 sport
1.2 [0.7, 1.9] 2 sports
1.7 [0.9, 3.2] a 3 sports
High PM communities: 1.0 [ref] 0 sport
1.1 [0.7,1.7] 1 sport
0.9
2.0
0.5,1.7] 2 sports
1.1,3.6] >3 sports
High vs. Low PMio communities: 0.8
(0.6,1.0)
Incidence-N (incidence) number of
sports:
Low PM communities: 49 (0.023) 0
54 (0.032) 1
22 (0.024) 2
13 (0.033) >3
High PM communities: 55 (0.021) 0
36(0.021)1
14(0.018)2
16 (0.033) >3
December 2009
E-396
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: McConnell, et al. (2006,
1802261
Period of Study: 1996-1999
Location: 12 Southern California
communities
Outcome: Prevalence of bronchitic
symptoms (yrly).
Age Groups: 10-15-yr-old
Study Design: Longitudinal cohort
N: 475 asthmatic children
Statistical Analyses: Multilevel logistic
mixed effects models.
Covariates: Age, second-hand smoke
Personal smoking history
Sex, race.
Dose-response Investigated? No
Statistical Package: SAS with
GLIMMIX macro
Pollutant: PM,0
Averaging Time: 365 days
Percentiles: Community byyr
(n = 48 = 12 communities • 4 yr)
25th: NR
SOth(Median): 3.4
75th: NR
Range (Min, Max):
Community byyr (n = 48= 12
communities' 4yr):
(0.89, 8.7)
Monitoring Stations: 12
Copollutant: 03, N02, EC, OC,
Acid vapor (acetic and formic acid)
PM Increment: 6.1 pg/m
OR Estimate [Lower Cl, Upper Cl]
PM10
Dog (n = 292): 1.60 [1.12: 2.30]
No dog (n = 183): 0.89 [0.57:1.39]
PMio'Dog interaction p-value: 0.02
Cat (n = 202): 1.47 [0.96: 2.24]
No Cat (n = 273): 1.20 [0.83:1.73]
PMio'Cat interaction p-value: 0.41
Neither pet (n = 112): 0.91 [0.53:1.56]
Cat only (n = 71): 0.84 [0.42:1.66]
Dog only (n = 161): 1.41 [0.91:2.19]
Both pets (n = 131): 1.89 [1.15: 3.10]
Results suggest that dog ownership, a
source of residential exposure to
endotoxin, may worsen the severity of
respiratory symptoms from exposure to
air pollutants in asthmatic children.
Reference: Meng et al. (2007, 0932751
Period of Study: Nov 2000 and Sep
2001 (collection of health data)
Location: Los Angeles and San Diego
counties
Outcome: Poorly controlled asthma vs.
controlled asthma
Age Groups: 18-64, 65+
Study Design: Long-term exposure
study
Comparison of cases and controls
N: 1,609 adults (represented individuals
age 18+ who reported ever having been
diagnosed as having asthma by a
physician and had their address
successfully geocoded)
Statistical Analyses: Logistic
regression to evaluate associations
between TD (traffic density) and annual
avg air pollution concentrations and
poorly controlled asthma. Used sample
weights that adjusted for unequal
probabilities of selection into the CHIS
sample.
Covariates: Age, sex, race/ethnicity,
family federal poverty level, county,
insurance status, delay in care for
asthma, taking medications, smoking
behavior, self-reported health status,
employment, physical activity
Dose-response Investigated? yes
Pollutant: PM,0
Averaging Time: 24 h over 1 yr
Copollutant (correlation):
03:r = -0.72
N02:r = 0.83
PM25:r = 0.84
CO: r = 0.42
TD:r = 0.14
PM Increment: Continuous data: per
10 pg/m3
OR Estimate [Lower Cl, Upper Cl]
lag:
All Adults: 1.08 [0.82, 1.43]
Non-Elederly Adults: 1.14 [0.84,1.55]
Elderly: 0.84 [0.41,1.73]
Vtomen: 1.38 [0.99,1.94]
December 2009
E-397
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Millstein et al. (2004,
0886291
Period of Study: Mar-Aug, 1995, and
Sep1995-Feb1996
Data were taken from the Children's
Health Study
Outcome: Wheezing & asthma
medication use (ICD9 NR)
Age Groups: 4th grade students,
mostly 9 yr at the time of the study
Study Design: Cohort Study, stratified
into 2 seasonal groups/
Location: Alpine, Atascadero, Lake
Arrowhead, Lake Elsinore, Lancaster,
Lompoc, Long Beach, Mira Loma,
Riverside, San Dimas, Santa Maria, and Statistical Analyses: Multilevel, mixed-
N: 2081 enrolled, 2034 provided parent-
completed questionnaire.
Upland, CA
effects logistic model.
Covariates: Contagious respiratory
disease, ambient airborne pollen and
other allergens, temperature, sex, age
race, allergies, pet cats, carpet in home,
environmental tobacco smoke, heating
fuel, heating system, water damage in
home, education level of questionnaire
signer, physician diagnosed asthma.
Season: Mar-Aug, 1995, and Sep,
1995toFeb, 1996
Statistical Package: GLIMMIX SAS
8.00 macro for generalized linear mixed
models.
Lags Considered: 14
Pollutant: PM,0
Averaging Time: Monthly means for
PM,O.
PM Component: Nitric acid, formic
acid, acetic acid
Monitoring Stations:
1 central location in each community
Copollutant (correlation):
03:r = 0.76
N02:r = 0.39
PM25:r = 0.91
PM Increment: IQR 13.39 pg/m
Odds Ratio [lower Cl, Upper Cl]
Annual
PM10: 0.93 [0.67, 1.27]
Mar-Aug
PM,0:0.91 [0.46, 1.80]
Sep-Feb
PM10: 0.65 [0.40, 1.06]
Reference: Neuberger et al. (2004,
0932491
Period of Study: Jun 1999-Jun 2000
Location: Austria (Vienna and a rural
area near Linz)
Outcome: Questionnaire derived
asthma score, and a 1-5 point
respiratory health rating by parent
Age Groups: 7-10 yr
Study Design: Cross-sectional survey
N: about 2000 children
Statistical Analyses: Mixed models
linear regression-used factor analysis to
develop the "asthma score"
Covariates: Pre-existing respiratory
conditions, temperature, rainy days, #
smokers in household, heavy traffic on
residential street, gas stove or heating,
molds, sex, age of child, allergies of
child, asthma in other family members
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 4 week avg
(preceding interview)
Pollutant: PM,0
Averaging Time: 24 h
Copollutant (correlation):
PM2.5 (r = 0.94) in Vienna
PM Increment: 10 pg/m
Change in mean associated unit
increase in PM (p-value) lag
Respiratory Health score
Vienna: 0.005 (p>0.05)
lag 4 week avg
Rural area: 0.008 (p>0.05)
lag 4 week avg
Asthma score
Vienna: 0.006 (p>0.05)
lag 4 week avg
Rural area: -0.001 (p>0.05)
lag 4 week avg
Reference: Oftedal et al. (2008,
0932021
Period of Study: 2001 -2002
Location: Oslo, Norway
Outcome: Lung function (PEF, Pollutant: PM10
FEF25%, FEF50%, FEV,, FVC)
IQR:
Age Groups: 9-10 yr PMlo in 1st yr of life: 10.3
Study Design: Cross-sectional pMio |ifetime: 58
N: 1847 children
Statistical Analyses: Linear regression
Covariates: Height, age, BMI, birth
weight, temperature, maternal smoking,
sex
Dose-response Investigated? Yes
Statistical Package:
SPSS, STATA, S-Plus
Lags Considered: 1-3
PM Increment: Per IQR
P (Lower Cl, Upper Cl)
PM10in 1st yr of life
PEF -72.5 (-122.3 to -22.7)
FEF25% -77.4 (-133.4 to -21.4)
FEF50% -53.9 (-102.6. to -5.2)
FEV, -6.7 -24.1, 10.7)
FVC 0.5 (-18.5, 19.6)
PM,0 lifetime exposure
PEF -66.4 (-109.5 to -23.3)
FEF25% -61. 5 (-11 0.0 to -13.1)
FEF50% -45.6 (-87.7 to -3.5)
FEV, -7.3 (-22.4, 7.7)
FVC -2.1 (-18.6, 14.4)
December 2009
E-398
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Parker et al. (2009,
1923591
Period of Study: 1999-2005
Location: U.S.
Outcome: Respiratory allergy/hayfever Pollutant:
Study Design: Cohort
Covariates: Survey yr, age, family
structure, usual source of care, health
insurance, family income relative to
federal poverty level, race/ethnicity
Statistical Analysis: Logistic
regression
Statistical Package: SUDAAN
Age Groups: 73,198 children aged
3-17 yr
Averaging Time: NR
Median: 24.1 pg/m3
IQR: 20.8-28.7
Copollutant (correlation):
Summer 03: 0.26
S02:-0.19
N02: 0.48
PM25:0.51
PM10.25:0.86
Increment: 10|jg/m
Odds Ratio (96% Cl)
Single Pollutant Model, variable N
Adjusted: 1.04 (0.99-1.09)
Reference: Penard-Morand et al.
(2005, 0879511
Period of Study: Mar 1999-Oct 2000
Mean concentrations of N02, S02,
PMio, and 03 were taken from Jan
1998-Dec2000
Location: 6 French cities: Bordeaux,
Clermont-Ferrand, Creteil, Marseille,
Strasbourg, Reims.
Outcome:
Flexural dermatitis
Asthma (493)
Rhinoconjunctivitis
Atopic dermatitis
Wheeze
Allergic rhinitis
Atopy
EIB (exercise-induced bronchial
reactivity)
Age Groups: 9-11 yr
Study Design: Cross-sectional
N: 9615 Children (6672 complete
examination and questionnaire info)
Statistical Analyses: Logistic
regression
Marginal Model (GENMOD)
Covariates: Age, Sex, Family history of
allergy, Passive smoking
Parental education
Season: All
Excluding end of spring and during
summer for clinical examinations
Dose-response Investigated? No
Statistical Package: SAS
Pollutant: PM,0
Averaging Time: 3 yr
Mean (SD): Low concentrations: 26.9
High Concentrations: 23.8
Range (Min, Max):
Low concentrations: 10-20
High concentrations: 21.5-29.5
Copollutant (correlation):
\N02:r = 46
S02:r=.76
03:r = -.02
Monitoring Stations: 16
PM Increment: 10 pg/m3 (IQR)
OR Estimate [Lower Cl, Upper Cl]:
EIB (during exam): 1.43 (1.02-2.01)
Flexural dermatitis (during exam):
0.79(0.59-1.07)
Wheeze (past yr): 1.05 (0.72-1.54)
Asthma (past yr): 1.23 (0.77-1.95)
Rhinoconjunctivitis (past yr):
1.17(0.86-1.59)
Atopic dermatitis (past yr):
1.28(0.96-1.71)
Asthma (lifetime): 1.32 (0.96-1.81)
Allergic rhinitis (lifetime):
1.32(1.04-1.68)
Atopic dermatitis (lifetime):
1.09(0.88-1.36)
Atopy (lifetime): 0.98(0.80-1.22)
Pollen: 1.14 (0.85-1.53)
Indoor: 0.91 (0.72-1.15)
Moulds: 1.00 (0.53-1.88)
Highest correlated pollutant
adjustments:
EIB (during exam): 1.16 (0.72-1.85)
Flexural dermatitis (during exam):
0.930(.60-1.43)
Wheeze (past yr): 1.31 (0.71-2.36)
Asthma (past yr): 1.25 (0.66-2.37)
Rhinoconjunctivitis (past yr):
1.22(0.98-1.68)
Atopic dermatitis (past yr):
1.63(1.07-2.49)
Asthma (lifetime): 1.11 (0.70-1.74)
Allergic rhinitis (lifetime):
1.19(0.94-1.59)
Atopic dermatitis (lifetime):
1.47(1.07-2.00)
Atopy (lifetime): 0.93(0.69-1.26)
Pollen: 1.30 (0.98-1.57)
Indoor: .83 (0.63-1.12)
Molds: 1.62 (0.64-4.09)
Reference: Peters et al., (1999,
0872371
Period of Study: 1986-1990,1994
Location: Southern California
Outcome: Asthma, cough, bronchitis,
wheeze
Age Groups: 4th, 7th, & 10th graders
Study Design: Cohort
N: 3676 children
Statistical Analyses: Stepwise logistic
regression
Covariates: Community, grade, race,
sex, height, BMI, asthma in parents,
hay fever, health insurance, plants in
home, mildew in home, passive smoke
exposure, pest infestation, carpet,
vitamin supplements, active smoking,
pets, gas stove, air conditioner
Dose-response Investigated? Yes
Pollutant: PM,0
Averaging Time:
24-h PM10 averaged over 1994
Mean based on data collected during
1986-1990,1994:
Alpine 37.4, 21.3
Atascadero 28.0, 20.7
Lake Elsinore 59.5, 34.7
Lake Gregory 38.3, 24.2
Lancaster 47.0, 33.6
LompocSO.0,13.0
Long Beach 49.5, 38.8
Mira Loma 84.9, 70.7
Riverside 84.9, 45.2
San Dimas67.0, 36.7
Santa Maria 28.0, 29.2
Upland 75.6, 49.0
PM Increment: 25 pg/m
OR (Lower Cl, Upper Cl) for
respiratory illness
Based on 1986-1990 pollutant levels
Ever asthma 0.93 (0.76,1.13)
Current asthma 1.09 (0.86,1.37)
Bronchitis 0.94 (0.74,1.19)
Cough 1.06 (0.93, 1.21)
Wheeze 1.05 (0.89,1.25)
Based on 1994 pollutant levels
Ever asthma 0.87 (0.67,1.14)
Current asthma 1.11 (0.81,1.54)
Bronchitis 0.90 (0.65,1.26)
Cough 1.14 (0.96, 1.35)
Wheeze 1.01 (0.79,1.29)
December 2009
E-399
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Pierse, et al. (2006,
0887571
Period of Study: 2 yr (once in 1998
and once in 2001—surveys)
Location: Leicestershire, UK
Outcome: Cough without a cold
Night time cough
Current wheeze
Age Groups: 1-5yr
Study Design: Cross-sectional
(cohorts)
N: 4400 children
Statistical Analyses: Binomial
generalized linear models (compared
with likelihood ratio tests)
Spatial variograms (due to the spatial
concerns)
Covariates: Age, Gender
Mother/father has asthma
Coal heating the home, Smoking by
household member in the home, Either
parent continued education past 16 yr
of age, Pre-term birth, Breastfeeding,
Gas cooking, Presence of pets, Number
of cigarettes smoked by mother,
Overcrowding, Single parenthood, Diet
Dose-response Investigated? Yes
(Fig. 2 shows evidence of dose-
response effect based on surveys,
states in discussion).
Statistical Package: SAS 8.2
S-Plus6.1
Pollutant: PM,0
Averaging Time: Annual PM10
Mean (SD):
1998:1.47
2001:1.33
Percentiles:
25th: 1998 (.73) and 2001 (.8)
75th: 1998 (1.93) and 2001 (1.84)
PM Increment: 10 pg/rri (IQR)
Unadjusted OR estimates [Lower Cl,
Upper Cl]:
Cough without cold (1998):
1.22(1.10(01.36)
Cough without cold (2001):
1.46(1.27to1.68)
Night-time cough (1998):
1.11 (1.01 to 1.23)
Night-time cough (2001):
1.25 (1.09 to 1.43)
Current wheeze (1998):
0.99(0.89to1.10)
Current wheeze (2001):
1.09 (0.93 to 1.30)
Adjusted OR Estimate [Lower Cl,
Upper Cl]:
Cough without cold (1998):
1.21(1.07(01.38)
Cough without cold (2001):
1.56 (1.32 to 1.84)
Night-time cough (1998):
106 (0.94 to 1.19)
Night-time cough (2001):
1.25 (1.06 to 1.47)
Current wheeze (1998):
0.99(0.88to1.12)
Current wheeze (2001):
1.28 (1.04 to 1.58)
When the child was originally
asymptomatic in 1998:
Unadjusted OR estimates [Lower Cl,
Upper Cl]:
Cough without cold (2001):
1.68 (1.39 to 2.03)
Night-time cough (2001):
1.21 (1.00 to 1.46)
Current wheeze (2001):
1.22 (0.92 to 1.62)
Adjusted OR Estimate [Lower Cl,
Upper Cl]:
Cough without cold (2001):
1.62(1.31to2.00)
Night-time cough (2001):
1.19 (0.96 to 1.47)
Current wheeze (2001):
1.42 (1.02 to 1.97)
December 2009
E-400
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Qian et al. (2005, 0932831
Period of Study: 1990-1992
Location: Forsythe, NC
Minneapolis, MN
Jackson, MS.
Outcome: FVC, FEV,, FEV,/FVC
Age Groups: Middle aged (avg 56.8 yr)
Study Design: Cross-sectional
N: 10,240 people
Statistical Analyses: Regression
equations, multiple linear regression
analyses
Covariates: Smoking status, recent use
of respiratory medication, current
respiratory symptoms, chronic lung
diseases, field center
Dose-response Investigated? No
Statistical Package: SAS software,
version 9.1
Pollutant: PM,0
Averaging Time: Annual
Mean (SD): 27.9 (2.8)
Percentiles: 25th: 25.8
SOth(Median): 27.5
75th: 30.2
Range (Maximum-Minimum): 12.2
Monitoring Stations:
3 (Minneapolis, MN)
5 (Jackson, MS)
and 9 (Forsythe, NC)
Copollutant: 0;
PM Increment: 2.8 pg/rri (1 SD)
Effect Estimate:
In Never Smokers
FVC IS = -0.0108, SE = 0.0026,
p = 0001
FEV, IS = -0.0082, SE = 0.0029,
p = 0047
FEV,/FVC IS = -0.0024, SE = 0.0023,
p = 2787
Smoking status
Current n = 2377, FVC = -1.96,
FEV, = -2.23, FEV,/FVC = -0.94
Former n = 3858, FVC = -1.25,
FEV, = -1.10, FEV,/FVC = -0.30
Never n = 4005, FVC = -1.12,
FEV, = -0.63, FEV,/FVC = 0.06
Recent Use of Respiratory Medication
Yes n = 424, FVC = -2.65,
FEV, = -3.89, FEV,/FVC = -3.00
No n = 9816, FVC = -1.41,
FEV, = -1.20, FEV,/FVC =-0.24
Current Respiratory Symptoms
Yes n = 4340, FVC = -1.68,
FEV, = -1.70, FEV,/FVC =-0.63
No n = 5900, FVC = -1.05,
FEV, = -0.63, FEV,/FVC = 0.05
Chronic Lung Diseases
Yes n = 1374, FVC = -1.95,
FEV, = -2.31,FEV,/FVC = -1.18
No n = 8866, FVC = -1.35,
FEV, = -1.10, FEV,/FVC = -0.19
Field Center
Forsythe, NC n = 3504, FVC = -0.03,
FEV, = 0.05, FEV,/FVC =-0.33
Minneapolis, MN n = 3793, FVC = 0.50,
FEV, = 0.54, FEV,/FVC =-0.30
Jackson, MS n = 2943, FVC = -0.01,
FEV, = 0.17, FEV,/FVC = -0.32
Reference: Rios et al. (2004, 0878001
Period of Study: 1998-2000
Location: the metropolitan area of Rio
de Janiero, Brazil, Duque de Caxias
(DC) and Seropedica (SR)
Outcome: Wheezing, asthma, cough at Pollutant: PM
night
Age Groups: 13-14 yr
Study Design: Cohort
N: 4064 students
Statistical Analyses: Cchi-squared
Covariates: Sex, type of school, time of
residence, domestic smoking, residents
per home
Dose-response Investigated? Yes
Statistical Package: Epilnfo
Averaging Time: Weekly
measurements used to create annual
PM estimate
Mean (SD):
DC
1998:147
1999:115
2000:110
Total: 124
SR
1998:37
1999:31
2000: 37
Total: 35
Monitoring Stations: NR
PM Increment: High vs. Low
Global Cut-Off Score %, p-val:
DC
Male: 15.0
Female: 22.3, p < .05t
Private School: 16.6
Public School: 19.4, p<.05*
<5yr residence: 20.9
>5yr residence: 16.8
No domestic smoking exposure: 17.6
Domestic smoking exposure: 20.4, p <
.05T
<5 residents per home: 18.4
5+ residents per home: 19.5
SR
Male: 12.3
Female: 19.7, p<.05t
Private School: 28.3, p<.05*t
Public School: 14.7
<5yr residence: 10.8
>5yr residence: 16.5
No domestic smoking exposure: 14.8
Domestic smoking exposure: 18.3
<5 residents per home: 15.6
5+ residents per home: 17.4
Notes: The Global Cut-off Score
encompasses replies to the asthma
component of ISAAC'S written
questionnaire that establishes a cut-off
from which is defined the presence of
asthma for the Brazilian population.
'Comparing the cities in the same
control led variable
tComparing the controlled variable in
the same city
December 2009
E-401
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Rojas-Martinez et al. (2007, Outcome: Lung function: FEVi, FVC,
0910641 FEF25-75%
Period of Study: 1996-1999
Location: Mexico City, Mexico
Age Groups: Children 8 yr old at time
of cohort recruitment
Study Design: School-based "dynamic"
cohort study
N: 3170 children
14,545 observations
Statistical Analyses: Three-level
generalized linear mixed models with
unstructured variance-covariance matrix
Covariates: Age, body mass index,
height, height by age, weekday spent
outdoors, environmental tobacco
smoke, previous-day mean air pollutant
concentration, time since first test
Dose-response Investigated? No
Statistical Package: SA
Pollutant: PM,0
Averaging Time: 6 mo
Mean (SD): 6-mo averaging
SD:NR
Mean: 75.6
Percentiles: 6-mo averaging
25th: 55.8
SOth(Median): 67.5
75th: 92.2
Monitoring Stations: 5 sites for PM10,
10 for other pollutants
Copollutant:
03
NO,
PM Increment: IQR 6-LC: 36.4
Slope [Lower Cl, Upper Cl]
Girls
One-pollutant model
FVC:-39 [-47:-31]
FEV: -29 [-36: -21]
FEF25-75%:-17[-36:1]
FEV,/FVC: 0.12 [0.07: 0.17]
Two-pollutant model: PM10, 6-LCS 03
FVC: -30 [-39: -22]
FEV:-24 [-31:-16]
FEF25-75%: -9 [-26: 9]
FEV,/FVC: 0.10 [0.06: 0.15]
PM10, 6-LCS N02
FVC:-21 [-30:-13]
FEV:-17 [-25:-8]
FEF25-75%: -23 [-43: -4]
FEV,/FVC: 0.07 [0.02: 0.13]
Multipollutant model: PM10, 6-LC, 03, &
N02
FVC: -14 [-23: -5]
FEV:-11 [-20:-3]
FEF25-75%: -7 [-27: 12]
FEV,/FVC: 0.08 [0.03: 0.13]
Boys
One-pollutant model
FVC:-33 [-41:-25]
FEV:-27 [-34:-19]
FEF25-75%: -18 [-34: -2]
FEV,/FVC: 0.04 [-0.01: 0.09]
Two-pollutant model: PM10, 6-LCS 03
FVC:-28 [-36:-19]
FEV:-22 [-30:-15]
FEF25-75%:-10[-27:7]
FEV,/FVC: 0.04 [-0.01: 0.09]
FEV,/FVC: 0.24 [0.13: 0.34]
PM10, 6-LCS N02
FVC: -16 [-26: -7]
FEV:-19 [-27:-10]
FEF25-75%: -26 [-44: -9]
FEVi/FVC: 0.005 [-0.06: 0.05]
Multipollutant model PM10, 6-LC, 03, &
N02
FVC: -12 [-22: -3]
FEV:-15 [-23:-6]
FEF25-75%:-12[-30:6]
FEV^FVC: -0.002 [-0.06: 0.05]
Reference: Schikowski et al. (2005,
0886371
Period of Study: 1985-1994
Location: Rhine-Ruhr Basin of
Germany [Dortmund (1985,1990),
Duisburg (1990), Gelsenkirchen (1986,
1990), and Herne (1986)]
Outcome: Respiratory symptoms &
pulmonary function
Age Groups: Age 54-55
Study Design: Cross-sectional
N: 4757 women
Statistical Analyses: Linear & Logistic
regressions, including random effects
model
Covariates: Age, smoking, SES,
occupational exposure, form of heating,
BMI, height
Season: NR
Dose-response Investigated? No
Statistical Package: SAS
Lags Considered: NR
Pollutant: PM,0
Averaging Time: NR
Min, P25, Median, Mean, P75, Max
Annual Mean
35, 40, 43, 44, 47, 53
5-yr Mean
39, 43, 47, 48, 53, 56
Monitoring Stations: 7
Copollutant (correlation): NR
PM Increment: 7 pg/m
OR (Lower Cl, Upper Cl) for asthma
symptoms
Annual means
Chronic bronchitis 1.00 (0.85,1.18)
Chronic cough 1.03 (0.87,1.23)
Frequent cough 1.01 (0.93,1.10)
COPD 1.37 (0.98, 1.92)
p<0.1
FEV, 0.953 (0.916, 0.989)
p<0.1
FVC 0.966 (0.940, 0.992)
p<0.1
FEV,/FVC 0.989 (0.978, 1.000)
p<0.1
Five yr means
Chronic bronchitis 1.13 (0.95,1.34)
Chronic cough 1.11 (0.93,1.31)
Frequent cough 1.05 (0.94,1.17)
COPD 1.33 (1.03, 1.72)
p<0.1
FEV, 0.949 (0.923, 0.975)
p < 0.05
FVC 0.963 (0.945, 0.982)
p < 0.05
FEV,/FVC 0.989 (0.980, 0.997)
p<0.1
Reference: Schindler et al, (2009, Outcome: Respiratory Symptoms
Pollutant: PM,(
Increment: 10|jg/m
December 2009
E-402
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
1919501
Period of Study: 1991-2002
Location: Switzerland
Study Design: Prospective Cohort Averaging Time: Annual
Statistical Analysis: Logistic
Regression Model
Age Groups: Adults, 18-60 yr of age at
start of study
Covariates: Sex, age, level of
education, Swiss citizenship, BMI,
parental smoking, parental history of
asthma/atopy, early respiratory
infection, smoking status, pack yr, daily
number of cigarettes, yr since smoking
cessation, passive smoking in
general/at work, occupational exposure
to airbourne irritants
Mean (SD) Unit:
Range (Min, Max):
Copollutant (correlation): NR
Odds Ratio (96%CI) of reporting
symptoms at second interview
Entire Sample, New Reports
Regular Cough: 0.77 (0.62-0.97)
Regular Phlegm: 0.74 (0.56-0.99)
Chronic Cough or Phlegm:
0.78 (0.62-0.98)
Wheezing: 1.01 (0.74-1.39)
Wheezing with Dyspnea:
0.70(0.49-1.01)
Wheezing without Cold:
1.06(0.76-1.50)
Entire Sample, Persistent Reports
Regular Cough: 0.55 (0.39-0.78)
Regular Phlegm: 0.82 (0.52-1.33)
Chronic Cough or Phlegm:
0.67(0.40-1.15)
Wheezing: 0.50 (0.32-0.80)
Wheezing with Dyspnea:
0.59(0.30-1.23)
Wheezing without Cold: 0.61- (0.35-
1.12)
Persistent Non-Smokers, New Reports
Regular Cough: 0.86 (0.63-1.19)
Regular Phlegm: 0.70 (0.49-0.99)
Chronic Cough or Phlegm:
0.71 (0.52-0.99)
Wheezing: 0.93 (0.60-1.46)
Wheezing with Dyspnea:
0.77(0.50-1.20)
Wheezing without Cold:
1.11(0.66-1.92)
Persistent Non-Smokers,
Persistent Reports
Regular Cough: 0.28 (0.14-0.60)
Regular Phlegm: 0.87 (0.43-1.84)
Chronic Cough or Phlegm:
0.35(0.16-0.81)
Wheezing: 0.53 (0.28-1.08)
Wheezing with Dyspnea:
0.76(0.30-2012)
Wheezing without Cold:
0.61 (0.26-1.52)
Gender-specific odds ratio (96%CI)
of reporting symptoms at second
interview
New Reports
Regular Cough, p = 0.73
Men: 0.75 (0.53-1.06)
Women: 0.81 (0.58-1.15)
Regular Phlegm, p = 0.41
Men: 0.85 (0.60-1.20)
Women: 0.68 (0.46-1.00)
Chronic Cough or Phlegm: 0.36
Men: 0.87 (0.63-1.21)
Women: 0.71 (0.51-0.97)
Wheezing, p = 0.20
Men: 0.83 (0.57-1.20)
Women: 1.20 (0.78-1.87)
Wheezing with Dyspnea, p = 0.11
Men: 0.56 (0.36-0.87)
Women: 1.00.57-1.842
Wheezing without Cold, p = 0.43
Men: 0.95 (0.63-1.42)
Women: 1.25 (0.72-2.17)
Persistent Reports
Regular Cough, p = 0.02
Men: 0.75 (0.48-1.18)
Women: 0.31 (0.17-0.56)
Regular Phlegm, p = 0.33
Men: 0.65 (0.37-1.12)
Women: 1.04 (0.47-2.34)
Chronic Cough or Phlegm: 0.47
Men: 0.68 (0.39-1.20)
Women: 0.47 (0.20-1.11)
Wheezing, p = 0.29
Men: 0.34 (0.17-0.72)
December 2009
E-403
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Women: 0.57 (0.32-1.01)
Wheezing with Dyspnea, p = 0.63
Men: 0.56 (0.16-1.95)
Women: 0.37 (0.13-1.05)
Wheezing without Cold, p = 0.57
Men: 0.34 (0.12-0.91)
Women: 0.49 (0.21-1.15)
Odds Ratio (96%CI) of reporting
symptoms at second interview with
additional adjustment for annual
outdoor PM exposure at baseline
Entire Sam pie
Regular Cough, p = 0.0003
New Reports: 0.77 (0.61-0.97)
Persistent Reports: 0.55 (0.39-0.78)
Regular Phlegm, p = 0.13
New Reports: 0.77 (0.59-1.02)
Persistent Reports: 0.79 (0.46-1.33)
Chronic Cough or Phlegm, p = 0.02
New Reports: 0.78 (0.62-0.98)
Persistent Reports: 0.64 (0.40-1.02)
Wheezing, p = 0.002
New Reports: 0.91 (0.63-1.33)
Persistent Reports: 0.47 (0.31-0.72)
Wheezing with Dyspnea, p = 0.03
New Reports: 0.65 (0.43-0.98)
Persistent Reports: 0.55 (0.28-1.10)
Severe Wheezing, p = 0.28
New Reports: 0.96 (0.66-1.40)
Persistent Reports: 0.62 (0.34-1.12)
Non-Smokers
Regular Cough, p< 0.001
New Reports: 0.87 (0.63-1.19)
Persistent Reports: 0.29 (0.16-0.52)
Regular Phlegm, p = 0.07
New Reports: 0.70 (0.50-0.99)
Persistent Reports: 0.67 (0.34-1.33)
Chronic Cough or Phlegm, p = 0.008
New Reports: 0.72 (0.52-0.99)
Persistent Reports: 0.38 (0.17-0.84)
Wheezing, p = 0.07
New Reports: 0.87 (0.52-1.48)
Persistent Reports: 0.48 (0.25-0.91)
Wheezing with Dyspnea, p = 0.36
New Reports: 0.76 (0.48-1.19)
Persistent Reports: 0.70 (0.27-1.82)
Severe Wheezing, p = 0.57
New Reports: 1.11 (0.64-1.93)
Persistent Reports: 0.64 (0.26-1.54)
Reference: Sharma et al. (2004,
1569741
Period of Study: Nov 2002-Apr 2003
Location: 3 sections in Kanpur City,
India:
1) Indian Institute of Technology Kanpur
(IITK)
2) Vikas Nagar (VN)
3) Juhilal Colony (JC)
Outcome: Lung function
Age Groups: 20-55 yr
Study Design: Cohort
N: 91 people
Statistical Analyses: Linear regression
Covariates: NR
Season: Fall, Wnter, spring
Dose-response Investigated? No
Statistical Package: Microsoft Excel
Lags Considered: 1-day lag & 5-day
ma
Pollutant: PM10
Averaging Time: 24 h
Mean (SD):
IITK 184 (40)
VN 295 (58)
JC 293 (90)
PM Component: Lead, Nickel,
Cadmium, Chromium, Iron, Zinc
Benzene soluble fraction (includes
polycyclic aromatic hydrocarbons
[PAHs])
Copollutant (correlation):
APEF = mean daily deviations in PEF
PM10-APEF: (-0.52)
PMio-PM25:(0.67)
PM10-PM10 (1-day lag): (0.45)
PMio-PM25 (1-day lag): (0.46)
PM Increment: 1 pg/m
APEF (difference or change in peak
expiratory flow)
-0.0318 L/min
December 2009
E-404
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Tager et al. (2005, 0875381
Period of Study: Apr 2000-Jun 2000,
Feb2001-Jun2001,
Feb2002-Jun2002
Location:
Los Angeles, California
San Francisco, California
Outcome: Lung Function (FEVi, FVC,
PEFR, FEF75, FEF25-75,
FEF25-75/FVC ratio)
Age Groups: 16-21+ y/o
College Freshman
Study Design: Retrospective cohort
N: 255 students
108 Men(M)
147V\fomen(W)
Statistical Analyses: Multivariate
Linear Regression
Covariates: Sex, height, weight, area
of residence, age, race, ETS exposure,
respiratory disease history
Dose-response Investigated? No
Pollutant: PM,0
Averaging Time: Cumulative lifetime
exposure
Median:
Prior to 1987: M: 73
W:71
1987 and later: M: 36
W:34
Lifetime: M: 48
W:45
Range (Min, Max):
Prior to 1987: M: 34,117
W:31,124
1987 and later: M: 18, 68
W: 20, 61
Lifetime: M: 21, 80
W: 18, 71
Monitoring Stations: Between 1 and 3
Copollutant (correlation):
03 prior to 1987: r = 0.68
031987 and later: r = 0.81
03-Lifetime:r = 0.57
PM Increment: 1 pg/m
Parameter Estimates (SD)
(Lifetime PM10, Interaction PM10
FEF25-75/FVC)
LnFEF75:
M: -0.009 (0.0009), 0.009 (0.007)
W:-0.010 (0.0007), 0.008(0.0005)
Reference: Tamura et a. (2003,
0874451
Period of Study: 1998-1999
Location: Bangkok, Thailand
Outcome: Non-specific respiratory
disease (Chronic bronchitis, acute
bronchitis, bronchial asthma, dyspnea
and wheezing)
Age Groups: Adults
Study Design: Cross-sectional
N: 1603 policemen
Statistical Analyses: Multiple logistic
regression
Covariates: Age, smoking status
Dose-response Investigated? Yes
Statistical Package: SPSS
Pollutant: PM10
Averaging Time: 24 h
Mean (SD):
Heavily Polluted 80-190
Moderately Polluted 60-69
Control 59
Monitoring Stations: 13
PM Increment: Heavily Polluted vs.
Moderately Polluted vs. Control
Number and Prevalence (%) of
respiratory disease among heavily
polluted, moderately polluted, and
control areas.
Heavily Polluted
Chronic bronchitis 16 (3.0)
Acute bronchitis 19 (3.5)
Bronchial asthma 5 (0.9)
Dyspnea & wheezing 49 (9.2)
Anyl of above 69 (13.0)
Persistent cough 11 (2.1)
Persistent phlegm 27 (1.3)
CoughS phlegm 6 (1.1)
Moderately Polluted
Chronic bronchitis 8 (2.4)
Acute bronchitis 12 (9.0)
Bronchial asthma 2 (0.6)
Dyspnea & wheezing 23 (6.8)
Any 1 of above 37 (10.9)
Persistent cough 1 (0.3)
Persistent phlegm 11 (3.3)|
Cough & phlegm 1 (0.3)
Control
Chronic bronchitis 6 (1.9)
Acute bronchitis 11 (3.3)
Bronchial asthma 0 (0.0)
Dyspnea S wheezing 23 (7.2)
Anyl of above 31 (9.4)
Persistent cough 1 (0.3)
Persistent phlegm 8 (2.4)
Cough & phlegm 1 (0.3)
December 2009
E-405
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Wheeler and Ben-Schlomo
(2005, 1887661
Period of Study: 1995-1997
Location: England
Outcome: FEVi
Age Groups: 16-79yr
Study Design: Data from Health
Survey for England were coupled
geographically with air pollution
measurements on a 1 km grid.
N: 26,426 households with 39,251
adults
Statistical Analyses: Logistic
regression, least squares regression
Covariates: Age, sex, height, body
mass index, smoking status, household
passive smoke exposure, inhaler use in
the previous 24-h, doctor diagnosis of
asthma.
Dose-response Investigated? No
Pollutant: PM,0
Averaging Time: 1996 annual mean
Mean (SD): 23.95 (3.58)
Range (Min, Max): 17.87-43.37
P (96%CI) for Height-age
standardized FEV1 by ambient air
quality index
p-value
Male
Good (ref)
Poor-0.023 (-0.030 to-0.016)
O.001
Female
Good (ref)
Poor-0.019 (-0.026 to-0.013)
O.001
Reference: Zhang et al., (2002,
0348141
Period of Study: 1993-1996
Location: 4 Chinese cities (urban and
suburban location in each city):
Guangzhou, Wuhan, Lanzhou,
Chongqing
Outcome: Interview-self reports of
symptoms: Wheeze (ever wheezy when
having a cold)
Asthma (diagnosis by doctor)
Bronchitis (diagnosis by doctor),
Hospitalization due to respiratory
disease (ever)
Persistent cough (coughed for at least 1
month per yr with or apart from colds)
Persistent phlegm (brought up phlegm
or mucus from the chest for at least 1
month per yr with or apart from colds)
Age Groups: Elementary school
students
age range: 5.4-16.2
Study Design: Cross-sectional
N: 7,557 returned questionnaires
7,392 included in first stage of analysis
Statistical Analyses: 2-stage
regression approach: Calculated odds
ratios and 95% CIs of respiratory
outcomes and covariates Second stage
consisted of variance-weighted linear
regressions that examined associations
between district-specific adjusted
prevalence rates and district-specific
ambient levels of each pollutant.
Covariates: Age, gender, breast-fed,
house type, number of rooms, sleeping
in own or shared room, sleeping in own
or shared bed, home coal use,
ventilation device used, homes
smokiness during cooking, eye irritation
during cooking, parental smoking,
mother's education level, mother's
occupation, father's occupation,
questionnaire respondent, yr of
questionnaire administration, season of
questionnaire administration, parental
asthma prevalence
Pollutant: PM10
Averaging Time: 2 yr
Mean (SD): 151 (56)
IQR: 87
Range (Min, Max):
Gives range (max.-min.):
80
Monitoring Stations:
2 types: municipal monitoring stations
over a period of 4 yr (1993-1996)
Schoolyards of participating children
over a period of 2 yr (1995-1996)
PM Increment: Interquartile range
corresponded to 1 unit of change.
RR Estimate [Lower Cl, Upper Cl]
lag:
Association between persistent phlegm
and PM10: 3.21 (1.55, 6.67)
p < 0.05
Between and within city modeled ORs,
scaled to interquartile range of
concentrations for each pollutant.
No associations between any type of
respiratory outcome and PMi0
When scaled to an increment of
50 pg/m3 of PM10, ORs were:
Wheeze: 1.07
Asthma: 1.18
Bronchitis: 1.53
Hospitalization: 1.17
Persistent cough: 1.20
Persistent phlegm: 1.95
1AII units expressed in ug/m3 unless otherwise specified.
December 2009
E-406
-------
Table E-23. Long-term exposure - respiratory morbidity outcomes - PMio2s-
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Chattopadhyay et al. (2007, Outcome: pulmonary function tests
(respiratory impairments)
147471
Period of Study: NR
Location: Three different points in
Kolkata, India: North, South, and
Central
Age Groups: All ages
Study Design: Cross-sectional
N: 505 people studied for PFT
total population of Kolkata not given
Statistical Analyses:
Frequencies
Covariates: Meteorologic data (i.e.
temperature, wind direction, wind
speed, and humidity)
Dose-response Investigated? No
Pollutant: PM<3.3-0.4
Averaging Time: 8 h
Mean (SD):
North Kolkata: 266.1
Central Kolkata: 435.3
South Kolkata: 449.1
Unit (i.e. pg/m3): pg/m3
Monitoring Stations: 1
Copollutant (correlation):
PM10
PM<10-3.3
PM Increment: NR
Respiratory impairments (SD):
North Kolkata
Male(n=137)
Restrictive: 4 (2.92)
Obstructive: 5 (3.64)
Combined Res. And Obs.: 6 (4.37)
Total: 15 (10.95)
Female (n=152)
Restrictive: 3 (1.97)
Obstructive: 5 (3.28)
Combined Res. And Obs.: 0
Total: 8 (5.26)
Total (n=289)
Restrictive: 7 (2.42)
Obstructive: 10 (3.46)
Combined Res. And Obs.: 6 (2.07)
Total: 23 (7.96)
Central Kolkata
Male (n=44)
Restrictive: 6 (13.63)
Obstructive: 1 (2.27)
Combined Res. And Obs.:1 (2.27)
Total: 8 (18.18)
Female (n=50)
Restrictive: 3 (6.00)
Obstructive: 2 (4.00)
Combined Res. And Obs.: 0
Total: 5 (10.00)
Total (n=94)
Restrictive: 9 (9.57)
Obstructive: 3 (3.19)
Combined Res. And Obs.: 1 (1.06)
Total: 13 (13.82)
South Kolkata
Male (n=52)
Restrictive:! (1.92)
Obstructive: 2 (3.84)
Combined Res. And Obs.: 3 (5.76)
Total: 6 (11.53)
Female (n=70)
Restrictive: 2 (2.85)
Obstructive:! (1.42)
Combined Res. And Obs. :0
Total: 3 (4.28)
Total (n=122)
Restrictive: 3 (2.45)
Obstructive: 3 (2.45)
Combined Res. And Obs.: 3 (2.45)
Total: 9 (7.37)
December 2009
E-407
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Chattopadhyay et al. (2007, Outcome: Pulmonary function tests
' (respiratory impairments)
147471
Period of Study: NR
Location: Three different points in
Kolkata, India: North, South, and
Central
Age Groups: All ages
Study Design: Cross-sectional
N: 505 people studied for PFT
Total population of Kolkata not given
Statistical Analyses: Frequencies
Covariates: Meteorologic data (i.e.
temperature, wind direction, wind
speed, and humidity)
Dose-response Investigated? No
Pollutant: PM<10-3.3
Averaging Time: 8 h
Mean (SD):
North Kolkata: 269.8
Central Kolkata: 679.2
South Kolkata: 460.1
Monitoring Stations: 1
Copollutant (correlation):
PM10
PMO.3-0.
PM Increment: NR
Respiratory impairments (SD):
North Kolkata
Male(n=137)
Restrictive: 4 (2.92)
Obstructive: 5 (3.64)
Combined Res. And Obs.: 6 (4.37)
Total: 15 (10.95)
Female (n=152)
Restrictive: 3 (1.97)
Obstructive: 5 (3.28)
Combined Res. And Obs. :0
Total: 8 (5.26)
Total (n=289)
Restrictive: 7 (2.42)
Obstructive: 10 (3.46)
Combined Res. And Obs.: 6 (2.07)
Total: 23 (7.96)
Central Kolkata
Male (n=44)
Restrictive: 6 (13.63)
Obstructive: 1 (2.27)
Combined Res. And Obs.: 1 (2.27)
Total: 8 (18.18)
Female (n=50)
Restrictive: 3 (6.00)
Obstructive: 2 (4.00)
Combined Res. And Obs.: 0
Total: 5 (10.00)
Total (n=94)
Restrictive: 9 (9.57)
Obstructive: 3 (3.19)
Combined Res. And Obs.: 1 (1.06)
Total: 13 (13.82)
South Kolkata
Male (n=52)
Restrictive:! (1.92)
Obstructive: 2 (3.84)
Combined Res. And Obs.: 3 (5.76)
Total: 6 (11.53)
Female (n=70)
Restrictive: 2 (2.85)
Obstructive:! (1.42)
Combined Res. And Obs. :0
Total: 3 (4.28)
Total (n=122)
Restrictive: 3 (2.45)
Obstructive: 3 (2.45)
Combined Res. And Obs.: 3 (2.45)
Total: 9 (7.37)
Reference: Dales et al., (2008,
1563781
Period of Study: Location: Windsor,
ON
Outcome: Pulmonary function and
inflammation
Age Groups: Grades 4-6
Pollutant: PM10.25
Averaging Time: Annual
Mean: 7.25
5th: 6.02
Study Design: Cross-sectional
prevalence design
Statistical Analyses: Multivariate linear 95th: 8.23
re9ression Copollutant:
Covariates: Ethnic background, SQ
smokers at home, pets at home, acute 2
respiratory illness, medication use NO,
Increment: Tertiles of exposure
FEV,:
<7.04: 2.18 ±0.01
7.04-7.53:2.19 + 0.02
>7.53: 2.14 ±0.01
FVC:
<7.04: 2.52 ± 0.02
7.04-7.53: 2.53 ±0.02
>7.53: 2.48 ±0.02
eNO:
<7.04:15.48±0.63
7.04-7.53:16.73 ±0.76
>7.53:16.59 ±0.79
December 2009
E-408
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Gauderman et al. (2000,
0125311
Period of Study: 1993-1997
Location: Southern California
Outcome: FVC, FEV,, MMEF, FEF75 Pollutant: PMi0.2.5
Age Groups: Fourth, seventh, or tenth Averaging Time: 24-h avg PM10 &
graders annual avg of 2-wk avg PM2 5
Study Design: Cohort Mean (SD): PM10.2.5 25.6
N: 3035 subjects Copollutant (correlation):
Statistical Analyses: Linear regression 03 r = -0.29
Covariates: Height, weight, BMI, N02 r = 0.44
asthma, smoking, exercise, room
temperature, barometric pressure Inor9'Acld ' = a43
Dose-response Investigated? Yes
Statistical Package: SAS
Increment: 25.6 pg/m
% Change (Lower Cl, Upper Cl)
PM10.25-4th grade
FVC-0.57 (-1.20 to-0.06)
FEV,-0.90 (-1.71 to-0.09)
MMEF-1.37 (-2.57 to-0.15)
FEF75-1.62 (-3.24, 0.04)
PMio.25-7th grade
FVC-0.35 (-1.02, 0.31)
FEV,-0.49 (-1.21, 0.24)
MMEF-0.64 (-2.83, 1.60)
FEF75-0.74 (-2.65, 1.20)
PMi0.25-10th grade
FVC-0.17 (-1.32, 0.99)
FEV,-0.68 (-2.15, 0.81)
MMEF-1.41 (-5.85,3.25)
FEF75-2.32 (-6.60, 2.17)
Reference: Gauderman et al. (2002,
0260131
Period of Study: 1996-2000
Location: Southern California
Outcome: Lung function development:
FEV,, maximal mid-expiratory flow
(MMEF)
Age Groups: Fourth grade children
(avg age = 9.9 yr)
Study Design: Cohort study
N: 1678 children, 12 communities
Statistical Analyses: Mixed model
linear regression
Covariates: Height, BMI, doctor-
diagnosed asthma and cigarette
smoking in previous yr, respiratory
illness and exercise on day of test,
interaction of each of these variables
with sex, barometric pressure,
temperature at test time, indicator
variables for field technician and
spirometer
Dose-response Investigated? Yes
Statistical Package: SAS (10)
Pollutant: PMi0.25
Averaging Time: Annual 24-h avg
Mean (SD): The avg levels were
presented in an online data supplement
(FigE1)
Monitoring Stations: 12
Copollutant (correlation):
03(10AMto6PM)r = 0.10
03r = -0.31
N02r = 0.46
Acid vapor r = 0.63
PM10r = 0.95
PMi0.25r = 0.81
EC r = 0.71
OCr = 0.96
PM Increment: 29.1 pg/m
Association Estimate:
PM10.2 5 was not correlated with any of
the pulmonary function tests that were
analyzed
Reference: Leonard! et al. (2000,
0102721
Period of Study: 1996
Location: 17 cities of Central Europe
(Bulgaria, Czech Republic, Hungary ,
Poland, Romania, Slovakia)
Outcome: Immune biomarkers
Age Groups: 9-11
Study Design: Cross-sectional
N: 366 school children
Statistical Analyses: Linear regression
Covariates: Age, gender, parental
smoking, laboratory of analysis, recent
respiratory illness
Dose-response Investigated? No
Statistical Package: STATA
Pollutant: PM10.25
Averaging Time: Subtracting PM25
from PM10 provides avg PM10.25
Mean(SD):PM10.25:20(5)
Range (Min, Max):
DM • MO QQ\
PMio-2.5. (12, 38)
5th, median, & 95th percentile
PM10.25: 12, 19, 29
% Change (Lower Cl, Upper Cl)
p-value
PM,0-2.5
Neutrophils 1 (-27, 38)
>20
Total lymphocytes 8 (-15, 38)
>20
B lymphocytes 22 (-16, 76)
>20
Total T lymphocytes 2 (-25, 37)
-> on
'".ZU
CD4+-1 (-30,41)
>20
CD8+3J-25, 41)
>20
CD4/CD8 0 (-23, 30)
>20
NK1 (-33,51)
>20
TotallgG-3(-21, 18)
>20
Total IgM 19 (-9, 55)
>20
Total IgA 16 (-12, 52)
>20
Total IgE -29 (-70, 70)
>20
December 2009
E-409
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: McConnell et al. (2003,
0494901
Outcome: Bronchitic symptoms
Age Groups: 9-19
Period of Study: 1993-1999
Study Design: Communities selected
Location: 12 Southern CA communities on basis of historic levels of criteria
pollutants and low residential mobility.
N: 475 children
Statistical Analyses: 3 stage
regression combined to give a logistic
mixed effects model
Covariates: Sex, ethnicity, allergies
history, asthma history, SES, insurance
status, current wheeze, current
exposure to ETS, personal smoking
status, participation in team sports, in
utero tobacco exposure through
maternal smoking, family history of
asthma, amount of time routinely spent
outside by child during 2-6 pm.
Dose-response Investigated? No
Statistical Package: SAS Glimmix
macro
Pollutant: PM10.2.5
Averaging Time: 4-yr avg
Mean (SD): 17.0(6.4)
Range (Min, Max): 10.2-35.0
Copollutant (correlation):
PM25:r = 0.24
PMi0:r = 0.79
Inorganic acid: r = 0.38
Organic Acid: r = 0.35
EC: r = 0.30
OC:r = 0.27
N02:r = -0.22
03:r = 0.29
PM Increment: Between community
range 24.8 pg/m3
Between community unit 1 pg/m3
Within community 1 pg/m3
OR Estimate [Lower Cl, Upper Cl]
Between community per range
1.38(0.65-2.92)
Between Community per unit
1.01(0.98-1.04)
Within community per unit
1.02(0.95-1.10)
Reference: Millstein et al. (2004,
0886291
Period of Study: Mar-Aug, 1995, and
Sep1995-Feb1996
Data were taken from the Children's
Health Study
Outcome: Wheezing & asthma
medication use
Age Groups: 4th grade students,
mostly 9 yr at the time of the study
Study Design: Cohort Study, stratified
into 2 seasonal groups/
Pollutant: PM10.2.5
Averaging Time: monthly
PM Component: Nitric acid, formic
acid, acetic acid
PM Increment: IQR 11.44 pg/m3
Odds Ratio [lower Cl, Upper Cl]
Annual
PM10.25: 0.96 [0.74, 1.25]
Location: Alpine, Atascadero, Lake
Arrowhead, Lake Elsinore, Lancaster,
Lompoc, Long Beach, Mira Loma,
Riverside, San Dimas, Santa Maria, and Statistical Analyses: Multilevel, mixed-
N: 2081 enrolled, 2034 provided parent-
completed questionnaire.
Upland, CA
effects logistic model.
Covariates: Contagious respiratory
disease, ambient airborne pollen and
other allergens, temperature, sex, age
race, allergies, pet cats, carpet in home,
environmental tobacco smoke, heating
fuel, heating system, water damage in
home, education level of questionnaire
signer, physician diagnosed asthma.
Season: Mar-Aug, 1995, and Sep,
1995toFeb, 1996
Statistical Package: SAS 8.00
Lags Considered: 14
Monitoring Stations: 1 central location
in each community Mar-Aug
PMio.25: 0.93 [0.54, 1.59]
Sep-Feb
PM10.25: 0.68 [0.46, 1.01]
Copollutant (correlation):
N02:r = 0.29
03:r = 0.77
PM25:r = -0.08
Reference: (Parker et al., 2009,
192369)
Period of Study: 1999-2005
Location: U.S.
Outcome: Respiratory allergy/hayfever
Study Design: Cohort
Covariates: Survey yr, age, family
structure, usual source of care, health
insurance, family income relative to
federal poverty level, race/ethnicity
Statistical Analysis: Logistic
regression
Statistical Package: SUDAAN
Age Groups: 73,198 children aged
3-17 yr
Pollutant: PM10.25
Averaging Time: NR
Median:11.2|jg/m3
IQR: 8.2-15.2
Copollutant (correlation):
Summer
03:0.16
S02: -0.33
N02: 0.29
PM25:0.02
PMi0:0.86
Increment: 10|jg/m
Odds Ratio (96% Cl)
Single Pollutant Model, variable N
Adjusted: 1.01 (0.95-1.07)
Single Pollutant Model, constant N
Adjusted: 1.13 (1.04-1.46)
Multi-pollutant Model: 1.16 (1.06-1.24)
December 2009
E-410
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Zhang et al. (2002,
0348141
Period of Study: 1993-1996
Location: 4 Chinese cities (urban and
suburban location in each city):
Guangzhou, Wuhan, Lanzhou,
Chongqing
Outcome: Interview-self reports of
symptoms: Wheeze (ever wheezy when
having a cold)
Asthma (diagnosis by doctor)
Bronchitis (diagnosis by doctor),
Hospitalization due to respiratory
disease (ever)
Persistent cough (coughed for at least 1
month per yr with or apart from colds)
Persistent phlegm (brought up phlegm
or mucus from the chest for at least 1
month per yr with or apart from colds)
Age Groups: Elementary school
students
age range: 5.4-16.2
Study Design: Cross-sectional
N: 7,557 returned questionnaires
7,392 included in first stage of analysis
Statistical Analyses: 2-stage
regression approach: Calculated odds
ratios and 95% CIs of respiratory
outcomes and covariates Second stage
consisted of variance-weighted linear
regressions that examined associations
between district-specific adjusted
prevalence rates and district-specific
ambient levels of each pollutant.
Covariates: Age, gender, breast-fed,
house type, number of rooms, sleeping
in own or shared room, sleeping in own
or shared bed, home coal use,
ventilation device used, homes
smokiness during cooking, eye irritation
during cooking, parental smoking,
mother's education level, mother's
occupation, father's occupation,
questionnaire respondent, yr of
questionnaire administration, season of
questionnaire administration, parental
asthma prevalence
Pollutant: PM10.2.5
Averaging Time: 2 yr
Mean (SD): 59 (28)
Percentiles:
25th: NR
SOth(Median): NR
75th: NR
IQR: 42
Range (Min, Max):
Gives range (max.-min.): 80
Monitoring Stations:
2 types: municipal monitoring stations
over a period of 4 yr (1993-1996)
Schoolyards of participating children
over a period of 2 yr (1995-1996)
PM Increment: Interquartile range
corresponded to 1 unit of change.
RR Estimate [Lower Cl, Upper Cl]
Association between bronchitis and
PM10.25: 2.20 (1.14, 4.26)
p < 0.05
Association between persistent cough
and PM10.25:1.46 (1.12,1.90)
p < 0.05
Between and within city associations:
Bronchitis: 3.18 (between city)
Persistent phlegm (between city): 2.78
When scaled to an increment of
50 pg/m3 of PMi0.2 5 associations (ORs)
between respiratory outcome and PM10.
25 were:
Wheeze: 1.14
Asthma: 1.34
Bronchitis: 2.56
Hospitalization: 1.58
Persistent cough: 1.57
Persistent phlegm: 3.45
All units expressed in pg/m unless otherwise specified.
December 2009
E-411
-------
Table E-24. Long-term exposure - respiratory morbidity outcomes - PIVhs (including PM
components/sources).
Study
Reference: Annesi-
Maesanoetal.(2007,
0913481
Period of Study: Mar
1999-0ct2000
Location: France
(Bordeaux, Clermont-
Ferrand Creteil
Marseille, Strasbourg,, &
Reims)
Design & Methods Concentrations1
Outcome: EIB, Flexural atopic Pollutant: PM25
dermatitis, asthma, rhiniconjuctivitis,
allergic rhinitis Averaging Time: 5-day mean
(Mon.-Fri.) over a 13-wkto 24-wk
Age Groups: Children mean span
10.4 + 0.7 yr
Residential Proximity Level
Study Design: Semi-individual
design Mean (SD):
Low cone: 8. 7
N: 5338
High cone: 20.7
Statistical Analyses: Logistic
regression Range (Min, Max):
Covariates: Age, sex, family history Low conc: (1-6, 12.2)
of allergy, passive smoking High cone: (12.5, 54.0)
Season1 NR
season. INK Qty Leve|
Dose-response Investigated? No .. ,g-,.
Statistical Package: SAS Low conc: 9 6
High cone: 23.0
Range (Min, Max):
Low cone: (4.7, 12.7)
High cone: (13.0, 54.5)
Effect Estimates (95% Cl)
PM Increment: High vs. Low
Allergic and respiratory morbidity OR Estimate (Lower Cl,
Upper Cl)
Proximity Level
EIB (0)1.35(1.10, 1.67)
Fl. Atopic dermatitis (C) 2.51 (2.06, 3.06)
Asthma (P) 1.11 (0.88, 1.39)
Atopic asthma (P) 1.43 (1.07, 1.91)
Non-atopic asthma (P) 0.73 (0.49, 1.07)
Rhiniconjunctivitis (P) 0.94 (0.77, 1.15)
Atopic dermatitis (P) 1.05(0.88, 1.27)
Asthma (L) 1.00 (0.82, 1.22)
Allergic Rhinitis (L) 1.09 (0.93, 1.27)
Atopic dermatitis (L) 0.94 (0.82, 1.09)
City Level
EIB(C) 1.43(1.15, 1.78)
Fl. Atopic dermatitis (C) 2.06 (1.69, 2.51)
Asthma (P) 1.31 (1.04, 1.66)
Atopic asthma (P) 1.58 (1.1 7, 2.14)
Non-atopic asthma (P) 1.00 (0.68, 1.49)
Rhiniconjunctivitis (P) 0.98 (0.80, 1.20)
Atopic dermatitis (P) 1.08(0.90, 1.30)
Asthma (L) 1.09 (0.89, 1.33)
Allergic Rhinitis (L) 1.13 (0.97, 1.33)
Atopic dermatitis (L) 0.95 (0.82, 1.09)
Notes: C = Current
P = Past yr
L = Lifetime
Allergic sensitization OR Estimate (Lower Cl, Upper Cl)
Proximity Level
All allergens 1.19 (1.04,1.36)
Indoor allergens 1.29 (1.11,1.50)
Outdoor allergens 1.02 (0.85,1.23)
Moulds 1.13 (0.78, 1.65)
City Level
All allergens 1.32 (1.15,1.51)
Indoor allergens 1.51 (1.29,1.76)
Outdoor allergens 1.06 (0.88,1.28)
Molds 1.00 (0.69, 1.46)
Reference: Bakke et al.
(2004, 1562461
Period of Study: Jan
1989-Jun2002
Location: One of
Norway's major
construction companies
Outcome: Spirometric
measurements
Age Groups: All ages, mean = 39
yr
Study Design: Cohort
N: 651 male construction workers
Statistical Analyses: Multiple
linear regression models
Covariates: Age, yr for non-
smokers and ever smokers
Dose-response Investigated? No
Statistical Package: SYSTAT 10.0
and SPSS 11.0
Pollutant: Respirabledust
Averaging Time: 5-8 h
Mean (SD):
Drill and blast workers: 6.3 (2.8)
Tunnel concrete workers: 6.1 (3.1)
Shotcreting operators: 19 (11)
IBM workers: 16 (6.6)
Outdoor concrete workers: 1.4
(0.73)
Foremen: 0.28 (0.48)
Engineers: 0.09 (0.28)
Unit (i.e. ug/m3): mg y/m3
Monitoring Stations: 16 tunnel
sites visited with sampling
equipment
Copollutant (correlation):
Total dust: r = 0.99
a quartz: r = 0.48
N02:r = 0.75
CO: r = 0.61
Oil mist: r = 0.83
Oil vapor: r = 0.68
VOC:r = 0.89
PM Increment: NR-exposure respirable dust
Effect Estimate (Lower Cl, Upper Cl):
Lung function changes predicted by multiple linear regression
models using one exposure variable adjusted for age and
observation time by non-smokers and ever smokers
Non-smokers: IS = -16.0
(-24- -6.8)
SE = 4.5
Ever smokers: IS = -9.3
(-17--1.6)
SE = 4.0
December 2009
E-412
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Bakke et al.
(2004, 1562461
Period of Study: Jan
1989-Jun2002
Location: One of
Norway's major
construction companies
Outcome: Spirometric
measurements
Age Groups: All ages, mean = 39
yr
Study Design: Cohort
N: 651 male construction workers
Statistical Analyses: Multiple
linear regression models
Covariates: Age, yr for non-
smokers and ever smokers
Dose-response Investigated?
No
Statistical Package:
SYSTAT 10.0 and SPSS 11.0
Pollutant: Total dust
Averaging Time: 5-8 h
Mean (SD):
Drill and blast workers: 18 (7.8)
Tunnel concrete workers: 21 (11)
Shotcreting operators: 73 (41)
TBM workers: 48 (20)
Outdoor concrete workers: 6.5 (3.4)
Foremen: 0.78 (1.3)
Engineers: 0.27 (0.78)
Unit (i.e. ug/m3): mg y/m3
Monitoring Stations: 16 tunnel
sites visited with sampling
equipment
Copollutant (correlation):
Respirable dust: r = 0.99
a quartz: r = 0.42
N02:r = 0.67
CO: r = 0.49
Oil mist: r = 0.81
Oil vapor: r = 0.64
VOC:r = 0.91
PM Increment: NR-exposure expirable dust
Lung function changes predicted by multiple linear regression
models using one exposure variable adjusted for age and
observation time by non-smokers and ever smokers
Non-smokers: IS = -4.0 (-6.5-1.4)
SE=1.3
Ever smokers: IS = -2.0 (-4.2-0.23)
SE=1.1
Reference: Bennett et al. Outcome: Respiratory symptoms
(2007, 1562681
Period of Study:
1992-2005
Location: Melbourne,
Australia
(from questionnaire)
Age Groups: All ages, mean = 37.2
yr
Study Design: Cohort
N:1446
Statistical Analyses: Logistic
regression models
Covariates: Age, gender, use of
(S2-agonists, use of inhaled
corticosteroids, smoking, yr of data
collection, and avg daily exposure
to PM25 in the 12 mo corresponding
to the time frame of symptoms
Dose-response Investigated? No
Statistical Package: STATA,
version 9
Pollutant: PM25
Averaging Time: 24 h
Mean (SD): 6.8
Range (Min, Max): (1.8-73.3)
Monitoring Stations: up to 3
PM Increment: NR
Effect Estimate [Lower Cl, Upper Cl]:
Respiratory symptoms in last 12 mo and exposure to ambient
PM25 over the same period
Within-person (longitudinal) effects
Wheeze: OR = 1.08 (0.79-1.48), p = 0.62
SOB on waking: OR = 1.34 (0.84-2.16), p = 0.22
Cough (AM): OR = 0.74 (0.47-1.15), p = 0.18
Phlegm (AM): OR = 1.55 (0.95-2.53), p = 0.08
Cough w/ phlegm (AM): OR = 1.28 (0.70-2.33), p = 0.42
Asthma attack: OR = 0.91 (0.55-1.49), p = 0.69
Between-person (cross-sectional) effects
Wheeze: OR = 1.32 (0.82-2.10), p = 0.25
SOB on waking: OR = 1.29 (0.46-3.60), p = 0.63
Cough (AM): OR = 0.21 (0.07-0.62), p = 0.01
Phlegm (AM): OR = 0.49 (0.16-1.44), p = 0.19
Cough w/ phlegm (AM): OR = 0.28 (0.08-0.97), p = 0.05
Asthma attack: OR = 0.52 (0.17-1.59), p = 0.26
December 2009
E-413
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: (Brauer et
al, 2007, 0906911
Period of Study:
1999-2000
Location: The
Netherlands
Outcome:
Allergen sensitivity (any, indoor,
outdoor, food, total) lgE>100 lU/mL
Asthma (probable, MD-diagnosed,
ever MD-diagnosed)
Bronchitis (MD-diagnosed, ever
MD-diagnosed)
Dry cough at night
Itchy rash
Itchy rash/eczema
Ear/Nose/Throat (ENT) infection
Eczema, MD-diagnosed
Eczema, ever MD-diagnosed
Flu/serious cold, MD-diagnosed
Wheeze (ever, early, early frequent,
persistent)
Age Groups: Very young children
(<4-yr-old) enrolled prenatally
Study Design: Prospective birth
cohort study
N: -4000 subjects
Statistical Analyses: Multiple
logistic regression
Dose-response Investigated? No
Pollutant: PM25
Averaging Time: 12 mo
Mean (SD): SD: NR
16.9
Percentiles: 25th: 14.8
SOth(Median): 17.3
75th: 18.1
Range (Min, Max): (13.5, 25.2)
Monitoring Stations: 40
Copollutant (correlation): Soot: r
= 0.97
N02:r = 0.93
PM Increment: IQR 3.3 |ig/m
Notes: Traffic-related pollution (PM25, soot, N02) was
associated with respiratory infections, asthma, and allergic
sensitization in children during the first 4 yr of life.
Symptom At 4-Yr-Old
Wheeze
4-yr-old: 1.23 [1.00:1.51]
Early-life: 1.20 [0.99:1.46]
Asthma, MD-diagnosed
4-yr-old: 1.15 [0.82:1.62]
Early-life: 1.32 [0.96:1.83]
Dry cough at night
4-yr-old: 1.11 [0.94:1.31]
Early-life: 1.14 [0.98:1.33]
Bronchitis, MD-diagnosed
4-yr-old: 0.88 [0.66:1.18]
Early-life: 0.86 [0.66:1.11]
ENT infection
4-yr-old: 1.13 [0.98:1.31]
Early-life: 1.17 [1.02:1.34]
Flu/serious cold, MD-diagnosed
4-yr-old: 1.21 [1.02:1.42]
Early-life: 1.25 [1.07:1.46]
Itchy rash
4-yr-old: 0.96 [0.82:1.11]
Early-life: 0.98 [0.85:1.14]
Eczema, MD-diagnosed
4-yr-old: 1.00 [0.88:1.21]
Early-life: 0.98 [0.82:1.17]
Allergen Sensitivity At 4-Yr-Old
Allergen, any: 1.55 [1.13: 2.11]
Allergen, indoor: 1.03 [0.69:1.55]
Allergen, outdoor: 0.93 [0.54:1.58]
Allergen, food: 1.75 [1.23: 2.47]
Allergen, total lgE>100 lU/mL: 0.84 [0.59:1.18]
Cumulative Allergy/Asthma Symptoms At 4-Yr-Old
Wheeze, ever: 1.22 [1.06:1.41]
Asthma, ever MD-diagnosed: 1.32 [1.04:1.69]
Asthma, probable: 1.08 [0.90:1.30]
Wheeze, early: 1.16 [1.00:1.34]
Wheeze, persistent: 1.19 [0.96:1.48]
Wheeze, early frequent: 1.19 [0.96:1.47]
Bronchitis, ever MD-diagnosed: 0.96 [0.81:1.13]
Itchy rash/eczema: 0.99 [0.88:1.13]
Eczema, ever MD-diagnosed: 0.98 [0.85:1.13]
December 2009
E-414
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: (Brauer et
al, 2007, 0906911
Period of Study:
1999-2000
Location: The
Netherlands
Outcome: Allergen sensitivity (any, Pollutant: Soot (as PM25
indoor, outdoor, food, total) lgE>100 absorbance)
lu/mL . • ,- „,
Averaging Time: 12 mo
Asthma (probable, MD-diagnosed,
ever MD-diagnosed) Mean (SD): 171
Bronchitis (MD-diagnosed, ever Percentiles:
MD-diagnosed) 25th: 1.33
PM Increment: IQR 0.58 E-5/m
Notes: Traffic-related pollution (PM25, soot, N02) was
associated with respiratory infections, asthma, and allergic
sensitization in children during the first 4 yr of life.
Symptom At 4-Yr-Old
Wheeze
4-yr-old: 1.18 [0.98: 1.41]
Earlv-life: 1.18 M.00:1. 401
Dry cough at night
Itchy rash
Itchy rash/eczema
Ear/Nose/Throat (ENT) infection
Eczema, MD-diagnosed
Eczema, ever MD-diagnosed
Flu/serious cold, MD-diagnosed
Wheeze (ever, early, early frequent,
persistent)
Age Groups: Very young children
(<4-yr-old) enrolled prenatally
Study Design: Prospective birth
cohort study
N: -4000 subjects
Statistical Analyses: Multiple
logistic regression
Dose-response Investigated? No
SOth(Median): 1.78
75th: 1.91
Range (Min, Max): (0.77, 3.68)
Unit (i.e. ug/m3): 1 E-5/m
Monitoring Stations: 40
Copollutant (correlation):
N02:r = 0.96
PM25:r = 0.97
Asthma, MD-diagnosed
4-yr-old: 1.15 [0.85:1.55]
Early-life: 1.30 [0.98:1.71]
Dry cough at night
4-yr-old: 1.13 [0.97:1.30]
Early-life: 1.14 [1.00:1.31]
Bronchitis, MD-diagnosed
4-yr-old: 0.90 [0.69:1.16]
Early-life: 0.88 [0.69:1.11]
ENT infection
4-yr-old: 1.15 [1.01:1.31]
Early-life: 1.16 [1.03:1.31]
Flu/serious cold, MD-diagnosed
4-yr-old: 1.18 [1.02:1.36]
Early-life: 1.19 [1.04:1.37]
Itchy rash
4-yr-old: 0.94 [0.82:1.08]
Early-life: 0.97 [0.85:1.10]
Eczema, MD-diagnosed
4-yr-old: 0.99 [0.84:1.17]
Early-life: 0.97 [0.83:1.14]
Allergen Sensitivity At 4-Yr-Old
Allergen, any: 1.45 [1.11:1.91]
Allergen, indoor: 1.02 [0.71:1.46]
Allergen, outdoor: 0.95 [0.59:1.52]
Allergen, food: 1.64 [1.21: 2.23]
Allergen, total lgE>100 lU/mL: 0.80 [0.59:1.09]
Cumulative Allergy/Asthma Symptoms At 4-Yr-Old
Wheeze, ever: 1.18 [1.04:1.34]
Asthma, ever MD-diagnosed: 1.26 [1.02:1.56]
Asthma, probable: 1.06 [0.90:1.24]
Wheeze, early: 1.11 [0.97:1.26]
Wheeze, persistent: 1.18 [0.98:1.42]
Wheeze, early frequent: 1.14 [0.95:1.37]
Bronchitis, ever MD-diagnosed: 0.95 [0.82:1.10]
Itchy rash/eczema: 0.99 [0.89:1.11]
Eczema, ever MD-diagnosed: 0.99 [0.87:1.12]
Reference: Brauer et al.
(2002, 0351921
Period of Study: NR
Location: The
Netherlands
Outcome: Questionnaire derived Pollutant: PM25
wheezing, dry nighttime cough, ear,
nose and throat infections, skin rash Averaging Time: 4 2-wk periods
dispersed throughout 1 yr, adjusted
Physician diagnosed asthma,
bronchitis, influenza, eczema
Age Groups: age 2
Study Design: Prospective cohort
N: 4146 children
Statistical Analyses: Logistic
regression
for temporal trend
Mean (SD): 16.9
Percentiles:
10th: 14.0
25th: 15.0
SOth(Median): 17.3
75th: 18.2
Covariates: Maternal age, maternal
smoking, mattress cover (allergen- 90tn: la1
free), maternal education, paternal
education, gender, gas stove, gas
water heater, any other siblings,
ethnicity, breastfeeding, mold at
home, pets, allergies in mother,
allergies in father
Range (Min, Max): 13.5, 25.2
Monitoring Stations: 40
Dose-response Investigated? No
Copollutant (correlation):
Soot: r = 0.99
N02:r = 0.97
PM Increment: 3.2 pg/m
OR Estimate [Lower Cl, Upper Cl];
Unadjusted
Wheeze 1.14 (0.99-1.30)
Asthma 1.08 (0.84-1.37)
Dry cough at night 1.10 (0.95-1.27)
Bronchitis 1.00 (0.85-1.18)
E, N, T infections 1.14 (0.99-1.33)
Flu 1.15 (1.03-1.28)
Itchy rash 1.07 (0.95-1.20)
Eczema 1.02 (0.90-1.16)
Adjusted
Wheeze 1.14 (0.98-1.34)
Asthma 1.12 (0.84-1.50)
Dry cough at night 1.04 (0.88-1.23)
Bronchitis 1.04 (0.85-1.26)
E, N, T infections 1.20 (1.01-1.42)
Flu 1.12 (1.00-1.27)
Itchy rash 1.01 (0.88-1.16)
Eczema 0.95 (0.83-1.10)
December 2009
E-415
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Brauer et al.
(2002, 0351921
Period of Study: NR
Location: The
Netherlands
Outcome: Questionnaire derived
wheezing, dry nighttime cough, ear,
nose and throat infections, skin rash
Physician diagnosed asthma,
bronchitis, influenza, eczema
Age Groups: Age 2
Study Design: Prospective cohort
N: 4146 children
Statistical Analyses: Logistic
regression
Covariates: Maternal age, maternal
smoking, mattress cover (allergen-
free), maternal education, paternal
education, gender, gas stove, gas
water heater, any other siblings,
ethnicity, breastfeeding, mold at
home, pets, allergies in mother,
allergies in father
Dose-response Investigated? No
Pollutant: PM25 "soot"
Averaging Time: 4 2-wk periods
dispersed throughout 1 yr, adjusted
for temporal trend
Mean (SD): 16.9 10-5/m
Percentiles: 10th: 1.16
25th: 1.38
SOth(Median): 1.78
75th: 1.92
90th: 2.19
Range (Min, Max): 0.77, 3.68
Unit (i.e. ug/m3): 10-5/m
Monitoring Stations: 40
Copollutant (correlation):
PM25(r = 0.99)
N02(r = 0.96)
PM Increment: 0.54 x 10-5/m (equivalent to 0.8 pg/m EC)
OR Estimate [Lower Cl, Upper Cl]
Unadjusted
Wheeze 1.11 [0.99-1.24]
Asthma 1.07 [0.87-1.31]
Dry cough at night 1.08 [0.95-1.21]
Bronchitis 0.98 [0.85-1.12]
E,N,T infections 1.12 [0.99-1.27]
Flu 1.13 [1.03-1.23]
Itchy rash 1.07 [0.97-1.19]
Eczema 1.01 [0.91-1.13]
Adjusted
Wheeze 1.11 [0.97-1.26]
Asthma 1.12 [0.88-1.43]
Dry cough at night 1.02 [0.88-1.17]
Bronchitis 0.99 [0.84-1.17]
E,N,T infections 1.15 [1.00-1.33]
Flu 1.09 [0.98-1.21]
Itchy rash 1.02 [0.91-1.15]
Eczema 0.96 [0.85-1.08]
Reference: Brauer et al.
(2006, 0907571
Period of Study:
1997-2001
Location: Germany
The Netherlands
Reference: Burr etal.
(2004, 0878091
Period of Study: 3 wk in
Jul and Jan 1997 and 2
wkinNov1996andApr
1997
Location: North Wales,
England
Outcome: Otitis Media (parental
report of doctor's diagnosis prior to
age 2 yr)
Age Groups: 0-2 yr
Study Design: Prospective Cohort
Study
N: 4,379 children total
The Netherlands: 3,714
Germany: 665
Statistical Analyses: Logistic
regression
Covariates: Sex, parental atopy,
maternal education, siblings,
maternal smoking during
pregnancy, ETS exposure at home,
use of gas for cooking, indoor
moulds and dampness, number of
siblings, breast-feeding, and
presence of pets in the home
Season: All
Dose-response Investigated? No
Outcome: Self-report of symptoms
only for wheeze, cough, phlegm,
rhinitis, and itchy eyes.
Age Groups: All
Study Design: Repeated measures
N: 386 persons in congested
streets and 425 in the uncongested
streets in 1996/1997. Of these, 165
and 283 completed the second
phase of the study.
Pollutant: PM25
PM Component: EC (EC)
Averaging Time: 8 wk (4 2-week
periods dispersed throughout 1 yr,
adjusted for temporal trends)
Mean:
The Netherlands:
P M • 1 R Q
r IVI2 5. I u.y
EC: 1.72
fnprm3ny
oci 1 1 ioi ly.
PM25:13.4
EC: 1.76
Range (Min, Max):
The Netherlands:
PM25: 13.5, 25.2
EC: 0.77, 3.68
Germany:
PM25: 12.0, 21.9
EC: 1.40, 4.39
Monitoring Stations: 80 (40 for
each cohort)
Pollutant: PM25
Averaging Time: Mean hourly
concentrations
Mean (SD):
Congested Streets
1996-9721.2
1998-9916.2
Uncongested Streets
1996-976.7
1998-994.9
Monitoring Stations: 1 in
congested street and 1 in
uncongested
PM Increment: PM25: 3 pg/m3 (~ IQR)
EC: -0.5 pg/m3 (~ IQR)
OR Estimate [Lower Cl, Upper Cl]
The Netherlands:
PM25:
At agel: 1.13(0.98-1.32
At age 2: 1.13 (1.00-1.27
EC'
At age 1:1. 11 (0.98-1.26)
At age 2: 1.10 (1.00-1.22)
Germany:
DM
PM25.
At agel: 1.19(0.73-1.92)
At age 2: 1.24 (0.84-1. 83)
EC1
At age 1:1. 12 (0.83-1.51)
At age 2: 1.10 (0.86-1.41)
% change PM10 in congested streets: 23.6
% change PMi0 in uncongested streets: 26.6
Uncongested street sampling site was 20 m from the
congested street sampler.
The opening of the by-pass produced a reduction in pollution
in the congested streets. The health effects of these changed
are likely to be greater for nasal and ocular symptoms than for
lower respiratory symptoms. Uncertainty about the causality
arises from low response rates and conflicting trends in
respiratory and nasal symptoms.
December 2009
E-416
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Calderon-
Garciduefias et al. (2006,
0912531
Period of Study:
1999-2000
Location: Southwest
Mexico City & Tlaxcala,
Mexico
Outcome: Hyperinflation, interstitial
markings-measured by chest
radiograph, and lung function-FVC,
FEV,, PEF, FEF25-75, measured
using spirometry tests
Age Groups: 5-13 yr
Study Design: Cohort1999-
N: 249 (total), 230 (Southwest
Mexico City), 19 (Tlaxcala)
Statistical Analyses: Bayestest,
Spearman rank correlation, multiple
regression
Covariates: Age, sex
Dose-response Investigated? No
Statistical Package: SAS 8.2
Pollutant: PM25
Averaging Time: 1 yr
Mean (SD): 21
2000-19
Tlaxacala:
1994-2000:
-------
Study
Reference: Dales etal.,
(2008, 1563781
Period of Study:
Location: Windsor, ON
Reference: Gauderman
et al. (2000, 0125311
Period of Study:
1993-1997
Location: Southern
California
Reference: Gauderman
et al. (2002, 0260131
Period of Study:
1996-2000
Location: Southern
California
Design & Methods
Outcome: Pulmonary function and
inflammation
Age Groups: Grades 4-6
Study Design: Cross-sectional
prevalence design
Statistical Analyses: Multivariate
linear regression
Covariates: Ethnic background,
smokers at home, pets at home,
acute respiratory illness, medication
use
Outcome: FVC, FEV,, MMEF,
FEF75
Age Groups: Fourth, seventh, or
tenth graders
Study Design: Cohort
N: 3035 subjects
Statistical Analyses: Linear
regression
Covariates: Height, weight, BMI,
asthma, smoking, exercise, room
temperature, barometric pressure
Dose-response Investigated? Yes
Statistical Package: SAS
Outcome: Lung function
development: FEV,, maximal
midexpiratory flow (MMEF)
Age Groups: Fourth grade children
(avg age = 9.9yr)
Study Design: Cohort study
N: 1678 children, 12 communities
Concentrations1
Pollutant: PM25
Averaging Time: Annual
Mean: 15.4
5th: 14.2
95th: 17.2
Copollutant:
en
0^2
N02
Pollutant: PM25
Averaging Time: Annual avg of
2-wkavgPM25
Mean (SD):PM25 25.9
Copollutant (correlation):
03: r = -0.32
DUrt • i- — n 7C
rM,o-2.5. r- U. 10
N02:r = 0.74
Inorg.Acid: r = 0.79
Pollutant: PM25
Averaging Time: Annual 24-h avg
Mean (SD): The avg levels were
presented in an online data
supplement (Fig E1)
PM Component: EC and OC.
Effect Estimates (95% Cl)
Increment: Tertiles of exposure
FEV,:
<15.19: 2.16 + 0.01
15.19-15.96:2.17 + 0.02
>15.96: 2.18 + 0.01
FVC:
<15.19: 2.51 +0.02
15.19-15.96:2.50 + 0.02
>15.96: 2.52 + 0.02
eNO:
<15.19: 16.08 + 0.70
15.19-15.96: 15.80 + 0.76
>15.96: 16.79 + 0.72
Increment: 25.9 pg/m3
% Change (Lower Cl, Upper Cl)
PM25-4th grade
FVC -0.47 (-0.94, 0.01)
FEV, -0.64 (-1.28, 0.01)
MMEF -1.03 (-1.95 to -0.09)
FEF75-1.31 (-2.57 to -0.03)
PM25-7th grade
FVC -0.42 (-0.89, 0.05)
FEV, -0.32 (-0.88, 0.24)
MMEF -0.29 (-1.99, 1.44)
FEF75 -0.26 (-1.75, 1.25)
PM25-10th grade
FVC 0.19 (-0.68, 1.07)
FEV, -0.25 (-1.41, 0.93)
MMEF -0.17 (-3.66, 3.46)
FEF75 -0.79 (-4.27, 2.82)
PM Increment: 22.2 pg/m3
Association Estimate:
Non-statistically significant negative correlation between PM2 5
and FEV,and FVC growth rates were observed. MMEF growth
rates had a negative correlation with PM25 (r = -0.43 p =
PM2 5 was not significantly correlated to FEV, (r = -0.31
p = 0.25)
0.05).
Statistical Analyses: Mixed model
linear regression
Covariates: Height, BMI, doctor-
diagnosed asthma and cigarette
smoking in previous yr, respiratory
illness and exercise on day of test,
interaction of each of these
variables with sex, barometric
pressure, temperature at test time,
indicator variables for field
technician and spirometer
Dose-response Investigated? Yes
Statistical Package: SAS (10)
Monitoring Stations: 12
Copollutant (correlation):
03:(10AMto6PM)r = 0.14
03:r = -0.39
N02:r = 0.77
Acid vapor: r = 0.87
PM,0:r = 0.95
PM,0.25:r = 0.81
EC: r = 0.93
OC:r = 0.89
December 2009
E-418
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: (Gauderman
etal., 2004, 0565691
Period of Study: Air
pollution data
ascertainment:
1994-2000. Spirometry
testing: spring
2001-spring 2003
Location: 12
Communities in Southern
California
Outcome: Lung function
FVC, FEV,, MMEF (Maximal
midexpiratory flow rate)
Age Groups: Children, Avg age 10
yr
Study Design: Prospective Cohort
Study
N: 12 Communities
2,034 children
24,972 child-mo
Statistical Analyses: Linear
regression of changes in sex-and-
community specific lung growth
function and PM
Correlation between % with low
attained FEV, and PM.
Covariates: Random effect for
communities
Dose-response Investigated? No
Statistical Package: SAS
Pollutant: PM25
Averaging Time: 2-wk
measurements used to create
annual avg
Mean: Means are presented in
figures only.
Range (Min, Max): ~6, -27
Monitoring Stations: 12
Copollutant (correlation):
PM10:r = 0.95
03:r = 0.18
N02:r = 0.79
EC: r = 0.91
OC:r = 0.91
PM Increment: Most to least polluted community Range:
22.8 pg/m3
Difference in Lung Growth [Lower Cl, Upper Cl];
FVC-60.1 (-166.1 to 45.9)
FEV, -79.7 (-153.0 to j6.4)
MMEF-168.9 (-345.5 to 7.8)
Correlation with % below 80% predicted Lung function (p-
value)
PM25: 0.79 (0.002)
Reference: Gauderman
et al. (2007, 0901211
Period of Study:
1993-2004
Location: 12 Southern
California Communities
Outcome: Pulmonary function tests Pollutant: PM
FVC, FEV,, MMEF/FEF25.75
Age Groups: Children (mean age
10 at recruitment, followed for 8 yr)
Study Design: Cohort Study
(Children's Health Study)
N: 3677 children (1718 in cohort 1
recruited 1993 and 1959 in cohort 2
recruited 1996)
22686 pulmonary function tests.
Statistical Analyses: Hierarchical
mixed effects model with linear
splines
Covariates: Adjustments for height,
height squared, BMI, BMI squared,
present asthma status, exercise or
respiratory illness on day of test,
smoking in previous yr, field
technician, traffic indicator (distance
from freeway, distance from major
roads), random effects for
participant and community.
Dose-response Investigated? No
Statistical Package: SAS
Monitoring Stations: 1 in each
community
PM Increment: 22.8 pg/m
Pollutant effect reported as difference in 8 yr lung function
growth from least to most polluted community. Negative
difference indicate growth deficits associated with exposure.
For PM2 5 FEV growth deficit is -100
Reference: Gehring et
al. (2002, 0362501
Period of Study:
1995-2002
Location: Munich,
Germany
Outcome: Wheezing, cough
without infection, dry cough at night,
obstructive, spastic or asthmoid
bronchitis, respiratory infections,
sneezing, runny/stuffed nose
Age Groups: 0-2 yr
Study Design: Prospective cohort
N: 1756 infants
Statistical Analyses: Logistic
regression
Covariates: Sex, parental atopy
(yes/no), maternal education,
siblings (y/n), environmental
tobacco smoke at home (y/n), use
of gas for cooking (y/n), home
dampness (y/n), indoor moulds
Pollutant: PM25
Mean (SD):PM25 mass: 13.4
PM25 absorb. 1.77*10-5/m
Percentiles: PM25 mass:
10th: 12.2
25th: 12.5
50th(Median):13.1
75th: 14.0
90th: 14.9
PM25 absorbance:
10th: 1.47* 10-5
25th: 1.54* 10-5
PM Increment:
PM25 mass: 1.5 pg/m
PM25 absorb. 0.4* 10-5/m(IQR)
RR Estimate [Lower Cl, Upper Cl]
Wheeze (PMu mass)
Age of 1 yr: All: 0.91 (0.76-1.09)
Males: 0.91 (0.72-1.16)
Females: 0.94 (0.70-1.27)
Age of 2 yr: All: 0.96 (0.83-1.12)
Males: 0.93 (0.76-1.14)
Females: 1.04 (0.83-1.30)
Cough W/0 Infection (PM25 mass)
Age of 1 yr: All: 1.34 (1.11-1.61)
Males: 1.43 (1.14-1.80)
Females: 1.19 (0.84-1.70)
Dry Cough At Night (PM25 mass)
Age of 1yr: All: 1.31 (1.07-1.60)
Males: 1.39 (1.08-1.78)
Females: 1.17 (0.81-1.68)
December 2009
E-419
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
(y/n), keeping of dogs (y/n) and cats
(y/n) study (GINI or LISA)
Dose-response Investigated? No
SOth(Median): 1.70* 10-5
75th: 1.88* 10-5
90th: 2.13* 10-5
Range (Min, Max):
PM2.5 mass: 11.9, 21.9
PM25 absorbance:
1.38to4.39*10-5
PM25 mass:
PM25 absorbance: 1/m
PM Component: PM25 mass
PM2 5 absorbance (as a marker of
diesel soot)
Monitoring Stations: 40
Copollutant (correlation):
N02:r = 0.99
PM25 absorbance and N02: r = 0.95
PM25 mass and PM25 absorbance:
r = 0.96
Age of 2 yr: All: 1.20 (1.02-1.42)
Males: 1.25 (1.01-1.55)
Females: 1.13 (0.86-1.48)
Bronchitis (PM25 mass)
Age of 1 yr: All: 0.98 (0.80-1.20)
Males: 0.97 (0.76-1.25)
Females: 0.98 (0.68-1.41)
Age of 2 yr: All: 0.92 (0.78-1.09)
Males: 0.92 (0.74-1.14)
Females: 0.91 (0.68-1.21)
Resp Infections (PM25 mass)
Age of 1yr: All: 1.04 (0.91-1.19)
Males: 1.04 (0.87-1.25)
Females: 1.06 (0.87-1.31)
Age of 2 yr: All: 0.98 (0.80-1.20)
Males: 0.99 (0.74-1.31): Females: 0.98 (0.73-1.31)
Sneezing/Runny Nose (PM25 mass)
Age of 1yr: All: 1.01 (0.85-1.20)
Males: 0.97 (0.77-1.24)
Females: 1.08 (0.84-1.41)
Age of 2 yr: All: 0.96 (0.82-1.12)
Males: 0.91 (0.73-1.12)
Females: 1.04 (0.83-1.31)
Wheeze (PM25 absorbance)
AgeoH yr: All: 0.93(0.78-1.12)
Males: 0.91 (0.71-1.15)
Females: 1.01 (0.74-1.37)
Age of 2 yr: All: 0.98 (0.84-1.14)
Males: 0.92 (0.75-1.13)
Females: 1.07 (0.85-1.36)
Cough W/0 Infection (PM25 absorbance)
AgeoH yr: All: 1.32 (1.10-1.59)
Males: 1.38 (1.11-1.71)
Females: 1.25 (0.87-1.78)
Dry Cough At Night (PM25 absorbance)
Age oHyr: All: 1.27 (1.04-1.55)
Males: 1.31 (1.04-1.67)
Females: 1.16 (0.79-1.71)
Age of 2 yr: All: 1.16 (0.98-1.37)
Males: 1.17 (0.95-1.44)
Females: 1.12 (0.84-1.48)
Bronchitis (PM25 absorbance)
Age of 1 yr: All: 0.99 (0.81-1.22)
Males: 1.00 (0.78-1.27)
Females: 0.94 (0.63-1.39)
Age of 2 yr: All: 0.94 (0.79-1.12)
Males: 0.91 (0.72-1.13)
Females: 0.95 (0.71-1.28)
Resp Infections (PM25 absorbance)
Age of 1 yr: All: 1.03 (0.90-1.18)
Males: 1.03 (0.86-1.23)
Females: 1.05 (0.85-1.30)
Age of 2 yr: All: 0.99 (0.80-1.22)
Males: 0.96 (0.73-1.26)
Females: 1.04 (0.75-1.43)
Sneezing/Runny Nose (PM25 absorbance)
Age oHyr: All: 0.95 (0.79-1.14)
Males: 0.90 (0.70-1.16)
Females: 1.06 (0.80-1.39)
Age of 2 yr: All: 0.92 (0.78-1.09)
Males: 0.83 (0.66-1.05)
Females: 1.06 (0.83-1.34))
December 2009
E-420
-------
Study
Reference: Goss et al.
(2004, 0556241
Period of Study:
1999-2000
Location: USA
Reference: Hertz-
Picciotto et al. (2005,
0886781
Period of Study: May
1994-Mar1999
Location: Teplice and
Prachatice, Czech
Republic
Design & Methods
Outcome: Cystic Fibrosis
pulmonary exacerbations, FEVi
Age Groups: Children and adults
over the age of 6
Study Design: Ccohort
N: 11484 patients
Statistical Analyses: Logistic
regression, t-tests, Mann-Whitney
tests, Chi-squared tests,
polytomous regression, multiple
linear regression
Covariates: Age, sex, lung
function, weight, insurance status,
pancreatic insufficiency, airway
colonization, genotype, median
household income by census tract,
zipcode.
Dose-response Investigated? No
Statistical Package: STATA, SAS
Outcome: Developmental
immunotoxicity as assessed by
neonatal immunophenotypes
Age Groups: Not specified: every
woman who delivered in the two
aforementioned districts were asked
to participate
Study Design: Cohort study
N: 1397 mother-infant pairs
Statistical Analyses: Multiple
linear regression with lymphocyte
percentage as responding variable
and pollutant exposure to 14day
averaging period before the date of
cord blood collection
Concentrations1
Pollutant: PM25
Averaging Time: Annual mean of
24-h avg
Mean (SD): 13.7(4.2)
Percentiles:25th:11.8
SOth(Median): 13.9
75th: 15.9
Monitoring Stations: 713
Pollutant: PM25
Averaging Time: 24 h
14 day avg
Mean (SD): Overall 24 h: 24.8
14-day avg:
Teplice: 30.1
Prachatice 19.8
PM Component: PAHs
Monitoring Stations: 2 stations:
Teplice and Prachatice
Effect Estimates (95% Cl)
PM Increment: 10 pg/m3
Odds Ratio Estimate [Lower Cl, Upper Cl]:
Odds of having 2 or more pulmonary exacerbations as
compared to 1 or less in 2000
1.21 (1.07-1.33)
Odds of having 1 pulmonary exacerbation as compared to no
exacerbations in 2000
0.70 (0.59-0.98)
Decrease in FEV, 155ml(115-194)
Decrease in FEV, in 2000 after adjusting for FEV, in 1999
24ml(7-40)
PM Increment: 25 pg/m3
Adjusted for 3-day temperature and season, PM2 5 exposure
during the 14 days before birth was associated with reduced
T-lymphocyte fractions CD4+, CD3+ and an increase in B-
lymphocyte fraction (CD19+).
The associations were not quantitatively reported anywhere
else in the paper other than in Fig 2 and Table 3
Covariates: Season, length of
labor, parity, number of previous
stillbirths, medication during
delivery, working status of mother,
maternal education, exposure to
active and secondhand smoke,
family history of allergy, self-reports
of workplace exposure to dust
during pregnancy, self-reported
maternal chronic or severe
respiratory diseases during
pregnancy. Ambient temperature
and season were controlled for.
Dose-response Investigated? Yes
Statistical Package: SUDAAN
(version 8)
December 2009
E-421
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: (Hertz-
Picciotto et al, 2007,
1359171
Period of Study:
1994-98 +follow-ups at
up to 4.5 yr of age for
child
Location: Czech
Republic districts of
Teplice and Prachatice
Outcome: Lower respiratory
illnesses, majority being acute
laryngitis, tracheitis, bronchitis.
ICD10 codes J04 and J20
Age Groups: Birth-4.5 yr of age.
Study Design: Longitudinal follow
up of a stratified random sample of
mother-infant pairs from previous
Pregnancy Outcome Study. Low
birth weight and preterm births
sampled at higher fractions.
N: 1133 children
Statistical Analyses: Generalized
linear longitudinal models, GEE to
adjust for within subject
correlations, robust variance
estimates were obtained. Model fit
judged using Akaike Information
criterion.
Covariates: Age of child, breast
feeding, environmental tobacco
smoke, season, day of week, yrof
birth, gender, birth weight,
pregnancy data including age at
delivery, length of gestation,
maternal hypertension and
diabetes, infant APGAR score,
maternal work history,
demographics, lifestyle,
reproductive and medical histories,
temperature, fuel type, other
children in household
Dose-response Investigated? No
Statistical Package: SUDAAN
version 8
Pollutant: PM25
Averaging Time: Used 3-, 7-, 14-,
30- and 45-day avg
Mean (SD): Daily mean 22.3
(sd 16 for 3-day avg, 11 for 45-day
avg)
PM Increment: 25 pg/m
RR Estimate [Lower Cl, Upper Cl] lag:
Bronchitis, birth-23 mo of age
Categorical model
High 30-day avg PM2 5 (greater than 50 pg/m3)
2.26(1.81-2.82)
Medium 30-day avg PM25 (between 25 and 50 pg/m3)
1.48(1.32-1.65)
Continuous model
1.30(1.08-1.58)
Bronchitis, 2-4.5 yrof age
Categorical model
High 30-day avg PM25 (greater than 50 pg/m3)
3.66(2.07-6.48)
Medium 30-day avg PM25 (between 25 and 50 pg/m3)
1.60(1.41-1.82)
Continuous model
1.23(0.94-1.62)
Notes: Results of other averaging periods shown in plots.
Reference: (Hogervorst
et al, 2006, 156559)
Period of Study: NR
Location: Maastricht, the
Netherlands (six schools
selected)
Outcome:
Decreased lung function
Age Groups: 8-1 Syr old
Study Design: Multivariate linear
regression (enter method) analysis
N: 342 children
Statistical Analyses: ANOVA, Chi
square
Covariates: Independent variables:
Age, height, gender, smoking at
home by parents, pets, use of
ventilation hoods during cooking,
presence of unvented geysers,
tapestry in the home, indoor/outdoor
time, education level of parents.
Dependent variables: lung function
indices
Dose-response Investigated? No
Pollutant: PM25
Averaging Time: Daily
Mean (SD): 19.0 (3.2)
Monitoring Stations: 6
Copollutant:
PM10
TSP
PM Increment: 10|jg/m
RR Estimate [Lower Cl, Upper Cl] lag:
FEV
3.62 [0.50,7.63]
FVC
1.80 [-2.10, 5.80]
FEF
5.93 [-2.34, 14.89]
December 2009
E-422
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Islam et al.
(2007, 0906971
Period of Study:
1993-2001
Location: 12
communities in Southern
California, U.S.
Outcome: New onset asthma
Age Groups: 9-1 Oyr
Study Design: Cohort
N:2057
Statistical Analyses: Cox
proportional hazard model
Covariates: Community, sex,
race/ethnicity
Season: All
Dose-response Investigated? No
Statistical Package: SASV 9.1
Lags Considered: 0-2 yr
Pollutant: PM25
Range (Min, Max):
"Low" PM25 Communities
(5.7-8.5)
"High" PM2 5 Communities
(13.7-29.5)
Monitoring Stations: 12
Copollutant: N02, acid vapor, PM10
and elemental and OC correlated
as a "non-03 package" of pollutants
with a similar pattern relative to
each other across the 12
communities.
PM Increment: NR
IR Estimate [Lower Cl, Upper Cl]
LowPM
FVC Ł90:19.4(7.5, 50.5)
FVC 90-110:16.8 (7.0, 40.1)
FV0110: 7.9 (2.9, 21.9)
FEV, Ł90:23.7(9.4, 59.4)
FEV, 90-110:15.6 (6.5, 37.4)
FEV, >110: 6.5 (2.3, 18.7)
FEF25-75Ł90:21.1 (8.8,50.5)
FEF25-75 90-110:11.9 (4.7, 30.0)
FEF25-75>110: 6.4 (2.3, 18.2)
Overall: 14.2 (7.0, 28.7)
HighPM
FVC Ł90:14.2 (5.1,39.6)
FVC 90-110: 25.6 (11.1,59.2)
FVC >110:16.7 (6.5, 42.9)
FEV, Ł90: 20.8 (8.0, 54.0)
FEV, 90-110: 23.1 (10.0,53.7)
FEV,>110:18.8 (7.5, 47.3)
FEF25-75Ł 90: 23.8 (10.2, 55.6)
FEF25-75 90-110: 23.9 (9.9, 57.7)
FEF25-75>110:15.9 (6.3, 40.5)
Overall: 18.4 (9.4, 35.9)
Reference: Karretal.
(2007, 0907191
Period of Study:
1995-2000
Location: South Coast
Air Basin of southern
California
Outcome: Bronchioloitis
Study Design: Case-control.
Cases included subjects with a
record of a single hospitalization
with a discharge diagnosis of acute
bronchiolitis.10 controls per case
were matched on birth date and
gestational age.
N: 18,595 cases
169,472 controls
Statistical Analyses: Conditional
logistic regression to estimate
relative risk of hospitalization for
bronchiolitis.
Covariates: Confounders included
in the model were: gender, parity,
chronic lung disease, cardiac and
pulmonary anomalies, SES
covariates
Age, Gestational age, and season
of birth were controlled for by
matching
Dose-response Investigated? Yes
Statistical Package: STATA
(Version 8)
Pollutant: PM25
Averaging Time: 24 h (lifetime
monthly avg from birth & 30 days
preceding cases hospitalization)
Mean (SD): 25
Percentiles:25th:19
SOth(Median): 23
75th: 29
Range (Min, Max): 6 to 111
Monitoring Stations: 17
PM Increment: 10|jg/m
RR Estimate [Lower Cl, Upper Cl]
Sub-chronic and chronic exposure: OR = 1.09 (1.04-1.14)
Adjusted for adjusted: Sub-chronic OR = 1.10 (1.04,1.16)
Chronic OR =1.09 (1.03-1.15)
Adjusted for CO and N02: Sub-chronic OR = 1.14 (1.07,1.21)
Chronic OR =1.12 (1.06,1.20)
Adjusted for 03, CO, and N02: Chronic OR = 1.15 (1.08,1.22)
Sub-chronic OR = 1.13 (1.06,1.21)
Reference: (Kim etal.,
2004, 0873831
Period of Study:
Mar-Jun (spring) 2001
Sep-Nov (fall) 2001
Location: Alameda
County, CA
Outcome: Asthma, bronchitis
Age Groups: Children (grades 3-5)
Study Design: Cross-sectional
N: 1109 children, 871 (long term
resident children), 462 (long term
related females), 403 (long term
related males)
Statistical Analyses: 2-stage
multiple logistic regression model
Covariates: Respiratory illness
before age of 2, household
mold/moisture, pests, maternal
history of asthma (for asthma)
Season: spring and fall
Dose-response Investigated? Yes
Statistical Package: SAS 8.2
Pollutant: PM25
Averaging Time: 10 wk
Mean (SD): Study Avg 12
Monitoring Stations: 10
Copollutant (correlation): r2 is
approximately 0.9 for all
copollutants-Black Carbon (BC),
PM10, NOX, N02, NO (NOX-N02)
PM Increment: 0.7 (IQR)
OR Estimate [Lower Cl, Upper Cl]:
Bronchitis
All subjects: 1.02 [1.00,1.08]
LTR subjects: 1.03 [1.01,1.08]
LTR females: 1.04 [1.02,1.05]
LTR males: 1.02 [0.99,1.05]
Asthma
All subjects: 1.00 [0.96,1.12]
LTR subjects: 1.01 [0.97,1.06]
LTR females: 1.06 [0.99,1.15]
LTR males: 0.99 [0.95,1.04]
Asthma excluding outlier school having a larger proportion of
Hispanics
All subjects: 1.04 [0.96,1.12]
LTR subjects: 1.03 [0.94,1.13]
LTR females: 1.03 [0.91,1.17]
LTR males: 1.03 [0.94,1.18]
December 2009
E-423
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Leonard! et
al. (2000, 0102721
Period of Study: 1996
Location: 17 cities of
Central Europe (Bulgaria,
Czech Republic, Hungary
, Poland, Romania,
Slovakia)
Outcome: Immune biomarkers
Age Groups: 9-11
Study Design: Cross-sectional
N: 366 school children
Statistical Analyses: Linear
regression
Covariates: Age, gender, parental
smoking, laboratory of analysis,
recent respiratory illness
Dose-response Investigated? No
Statistical Package: STATA
Pollutant: PM25
Averaging Time: Annual PM25
Mean(SD):PM25:46(10)
Range (Min, Max):
PM25: (29, 67)
5th, median, & 95th percentile
PM25: 29, 44, 67
% Change (Lower Cl, Upper Cl) p-value
PM25
Neutrophils-10(-45, 46)
>20
Total lymphocytes 49 (11,101); .008
B lymphocytes 63 (4,155); .034
Total T lymphocytes 72 (32
123)
<001
CD4+ 80 (34
143)
<001
CD8+61 (17, 119); .003
CD4/CD816(-17, 62)
>20
NK63(3, 158); .035
Total IgG 24 (2, 52); .034
Total IgM -9 (-32, 22)
>20
Total lgA-1 (-25, 32)
>20
TotallgE-4(-61, 137)
>20
Reference: McConnell
(1999, 0070281
Period of Study: 1993
Location: Southern
California
Outcome: Bronchitis, chronic
cough, phlegm
Age Groups: Children: 4th, 7th, &
10th graders
Study Design: Cross-sectional
N: 3676 people
Statistical Analyses: Logistic
regression
Covariates: Age, sex, race, grade,
health insurance
Dose-response Investigated? Yes
Pollutant: PM25
Averaging Time: Yearly 2-wk avg
Mean (SD): 15.3
Range (Min, Max): 6.7, 31.5
Copollutant (correlation):
N02r = 0.83
03r = 0.50
Acid r = 0.71
Child Respiratory symptoms OR Estimate (Lower Cl,
Upper Cl)
PM2.s Increment: 15|jg/m3
Children w/ asthma
Bronchitis: 1.4 (0.9, 2.3)
Phlegm: 2.6 (1.2, 5.4)
Cough: 1.3 (0.7, 2.4)
Children w/wheeze, no asthma
Bronchitis: 0.9 (0.6,1.4)
Phlegm: 1.0 (0.6,1.8)
Cough: 1.1 (0.6,1.9)
Children w/ no wheeze, no asthma
Bronchitis: 0.5 (0.3,1.0)
Phlegm: 0.8 (0.4,1.5)
Cough: 0.9 (0.6,1.3)
Reference: McConnell et Outcome: Bronchitic symptoms
al. (2003, 0494901
Age Groups: 9-19
Period of Study:
1993-1999 Study Design: Communities
selected on basis of historic levels
Location: 12 Southern of criteria pollutants and low
CA communities residential mobility.
N: 475 children
Statistical Analyses: 3 stage
regression combined to give a
logistic mixed effects model
Covariates: Sex, ethnicity, allergies
history, asthma history, SES,
insurance status, current wheeze,
current exposure to ETS, personal
smoking status, participation in
team sports, in utero tobacco
exposure through maternal
smoking, family history of asthma,
amount of time routinely spent
outside by child during 2-6 pm.
Dose-response Investigated? No
Statistical Package: SAS Glimmix
Pollutant: PM25
Averaging Time: 4-yr avg
Mean (SD): 13.8(7.7)
Range (Min, Max): 5.5-28.5
Copollutant (correlation):
PM10:r = 0.79
PMi0.25:r = 0.24
Inorganic acid: r = 0.76
Organic Acid: r = 0.58
EC: r = 0.83
OC:r = 0.84
N02:r = 0.54
03:r = 0.72
PM Increment: Between community range 23 pg/m
Between community unit 1 pg/m3
Within community 1 pg/m3
OR Estimate [Lower Cl, Upper Cl]
Between community per range
1.81(1.14-2.88)
Between Community per unit
1.03(1.01-1.05)
Within community per unit
1.09(1.01-1.17)
December 2009
E-424
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: McConnell et Outcome: Bronchitic symptoms
al. (2003, 0494901
Age Groups: 9-19
Period of Study:
1993-1999 study Design: Communities
selected on basis of historic levels
Location: 12 Southern of criteria pollutants and low
CA communities residential mobility.
N: 475 children
Statistical Analyses: 3 stage
regression combined to give a
logistic mixed effects model
Covariates: Sex, ethnicity, allergies
history, asthma history, SES,
insurance status, current wheeze,
current exposure to ETS, personal
smoking status, participation in
team sports, in utero tobacco
exposure through maternal
smoking, family history of asthma,
amount of time routinely spent
outside by child during 2-6 pm.
Dose-response Investigated? No
Statistical Package: SAS Glimmix
macro
Pollutant: EC
Averaging Time: 4-yr avg
Mean (SD): 0.71(0.41)
Range (Min, Max): 0.1-1.2
Copollutant (correlation):
PM25:r = 0.83
PM10:r = 0.71
PM10.25:r = 0.30
Inorganic acid: r = 0.82
Organic Acid: r = 0.66
OC:r = 0.88
N02:r = 0.54
03:r = 0.68
PM Increment: Between community range 1.1 pg/m
Between community unit 1 pg/m3
Within community 1 pg/m3
OR Estimate [Lower Cl, Upper Cl]
Between community per range
1.64(1.06-2.54)
Between Community per unit
1.55(1.05-2.30)
Within community per unit
2.63(0.83-8.33)
Reference: McConnell et Outcome: Bronchitic symptoms
al. (2003, 0494901
Age Groups: 9-19
Period of Study:
1993-1999 study Design: Communities
selected on basis of historic levels
Location: 12 Southern of criteria pollutants and low
CA communities residential mobility.
N: 475 children
Statistical Analyses: 3 stage
regression combined to give a
logistic mixed effects model
Covariates: Sex, ethnicity, allergies
history, asthma history, SES,
insurance status, current wheeze,
current exposure to ETS, personal
smoking status, participation in
team sports, in utero tobacco
exposure through maternal
smoking, family history of asthma,
amount of time routinely spent
outside by child during 2-6 pm.
Dose-response Investigated? No
Statistical Package: SAS Glimmix
macro
Pollutant: OC
Averaging Time: 4-yr avg
Mean (SD): 4.5(2.7)
Range (Min, Max): 1.4-11.6
Copollutant (correlation):
PM25:r = 0.84
PM,0:r=.70
PM10.25:r = 0.27
Inorganic acid: r = 0.83
Organic Acid: r = 0.69
EC: r = 0.88
N02:r = 0.67
03:r = 0.81
PM Increment: Between community range 10.2 pg/m
Between community unit 1 pg/m3
Wthin community 1 pg/m3
OR Estimate [Lower Cl, Upper Cl]
Between community per range
1.74(0.89-3.4)
Between Community per unit
1.06(0.99-1.13)
Wthin community per unit
1.41(1.12-1.78)
December 2009
E-425
-------
Study
Reference: McConnell,
et al. (2006, 1802261
Period of Study:
^ one ^ nnn
1990-1999
Location: 12 Southern
California communities
Design & Methods
Outcome:
Prevalence of bronchitic symptoms
(yrly).
Age Groups: 10-15-yr-old
Study Design: Longitudinal cohort
N: 475 asthmatic children
Ct-atiotirt'al AnalwoAO1 Mi iltllawal
Concentrations1
Pollutant: PM25
Averaging Time: 365 days
Percentiles: Community byyr
(n = 48 = 12 communities • 4 yr)
25th: NR
SOth(Median): 3.4
75th: NR
Range (Min, Max): Community by
Effect Estimates (95%
PM Increment: 3.4 pg/m3
OR Estimate [Lower Cl, Upper Cl]
PM25
Dog (n = 292): 1.56 [1.15: 2.12]
No dog (n = 183): 1.03 [0.71: 1.49]
PM25*Dog interaction p-value: 0.06
Cat (n = 202): 1.30 [0.90: 1.88]
No Cat (n = 273): 1.36 [0.99: 1.83]
PM25*Cat interaction p-value: 0.87
Cl)
Statistical Analyses: Multilevel
logistic mixed effects models.
Covariates: Age, second-hand
smoke
Personal smoking history
Sex, race.
Dose-response Investigated? No
Statistical Package: SAS
yr (n = 48 = 12 communities • 4 yr):
(0.89, 8.7)
Monitoring Stations: 12
Copollutant:
03
N02
EC
OC
Acid vapor (acetic and formic acid)
Neither pet (n = 112): 1.11 [0.71:1.74]
Cat only (n = 71): 0.85 [0.46:1.57]
Dog only (n = 161): 1.53 [1.04: 2.25]
Both pets (n = 131): 1.58 [1.02: 2.46]
Results suggest that dog ownership, a source of residential
exposure to endotoxin, may worsen the severity of respiratory
symptoms from exposure to air pollutants in asthmatic
children.
Although PM25 was associated at a statistically significant
level with ownership of both cats and dogs, it appears that dog
ownership (with or without a cat) specifically worsens the
association between PM25 and respiratory symptoms in
asthmatic children.
Reference: (Mengetal.,
2007, 0932751
Period of Study: Nov
2000 and Sep 2001
Location: Los Angeles
and San Diego counties
Outcome: Poorly controlled asthma Pollutant: PM25
vs. controlled asthma
Averaging Time: 24 h
ICD9NR
Copollutant (correlation):
Age Groups: 18-64, 65+ ^^
Study Design: Long-term exposure ^^
comparison of cases and controls PMio: r = 0.84
N: 1,609 adults (represented CO: r = 0.52
individuals age 1 8+ who reported jQ' r - 0 1 3
Results for PM25 were nonsignificant and not reported
quantitatively.
having asthma by a physician and
had their address successfully
geocoded)
Statistical Analyses: Logistic
regression to evaluate associations
between TD (traffic density) and
annual avg air pollution
concentrations and poorly
controlled asthma. Used sample
weights that adjusted for unequal
probabilities of selection into the
CHIS sample.
Covariates: Age, sex,
race/ethnicity, family federal poverty
level, county, insurance status,
delay in care for asthma, taking
medications, smoking behavior,
self-reported health status,
employment, physical activity
Dose-response Investigated? yes
December 2009
E-426
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Millstein, J et Outcome: Wheezing & asthma
al. (2004, 0886291 medication use (ICD 9 NR)
Period of Study:
Mar-Aug 1995, and Sep
1995-Feb1996
Data were taken from the
Children's Health Study
Location: Alpine,
Atascadero, Lake
Arrowhead, Lake
Elsinore, Lancaster,
Lompoc, Long Beach,
Mira Loma, Riverside,
San Dimas, Santa Maria,
and Upland, CA
Age Groups: 4th grade students,
mostly 9 yr at the time of the study
Study Design: Cohort Study,
stratified into 2 seasonal groups/
N: 2081 enrolled, 2034 provided
parent-completed questionnaire.
Statistical Analyses: Multilevel,
mixed-effects logistic model.
Covariates: Contagious respiratory
disease, ambient airborne pollen
and other allergens, temperature,
sex, age race, allergies, pet cats,
carpet in home, environmental
tobacco smoke, heating fuel,
heating system, water damage in
home, education level of
questionnaire signer, physician
diagnosed asthma.
Season: Mar-Aug, 1995, and Sep,
1995toFeb, 1996
Statistical Package: GLIMMIX
SAS 8.00 macro for generalized
linear mixed models.
Lags Considered: 14
Pollutant: PM25
Averaging Time: Integrated values
for successive 2-wk periods
PM Component: Nitric acid, formic
acid, acetic acid
Monitoring Stations: 1 central
location in each community
Copollutant (correlation):
03:r = 0.09
N02:r = 0.28
PM10:r = 0.33
PMi0.25:r = -0.08
PM Increment: IQR: 5.24 pg/rri
Odds Ratio [lower Cl, Upper Cl]
Annual
PM25:1.04 [0.83, 1.29]
Mar-Aug
PM25:0.91 [0.64, 1.30]
Sep-Feb
PM25:1.18 [0.89, 1.58]
Reference: Morgenstern
et al. (2007, 0907471
Period of Study: Mar
1999-Jul2000
Location: Munich,
Germany
Outcome: Asthma, wheezing,
spastic/obstructive bronchitis. Dry
cough at night, respiratory
infections, sneezing, runny/stuffed
nose without a cold.
Age Groups: at 1 yr & at 2 yr
Study Design: Cohort
N: 3577 children for the prediction
models. Respiratory data available
for 3129 children at 1 yr.
Statistical Analyses: Pearson's
correlation coefficient, prediction
error expressed as root mean
squared error (RMSE), multiple
logistic regression with confounding
factors, odds ratios
Covariates: Sex, Parental atopy
(genetic predisposition to allergies),
environmental tobacco smoke at
home, maternal education >or <12
yr, sibling, gas stove, home
dampness, indoor mold, pets. Since
it was not feasible to measure
personal exposure to N02, PM25,
and PM25 absorbance, exposure
modeling was used.
Statistical Package: SAS V.8.02
Pollutant: PM25
Averaging Time: Annual
Mean (SD): 12.8
Percentiles: 25th: 12.5
SOth(Median): 12.9
75th: 13.3
Range (Min, Max): 6.8,15.3
Monitoring Stations: 40: traffic,
n = 17 and background, n = 23.
Copollutant (correlation):
P M2 5 absorbance r = 0.49
NO, r = 0.45
PM Increment: 1.04 pg/m
Odds Ratio [Lower Cl, Upper Cl]
Adjusted OR for PM25 and: sneezing, runny/stuffed nose
during the first yr of life was 1.16 [1.01,1.34]
At age 1 yr
For wheezing 1.01 [0.87,1.18]
For cough without infection 1.05 [0.88,1.25]
For dry cough at night! .08 [0.86,1.27]
For asthmatic, spastic, or obstructive bronchitis
1.04 [0.90, 1.29]
For respiratory infectionl.05 [0.88,1.22]
For sneezing, runny or stuffed nose 1.16 [1.01,1.34]
At age 2 yr
For wheezing 1.10 [0.96,1.25]
For cough without infection NA, insufficient sample
For dry cough at night 1.03 [0.86,1.19]
For asthmatic, spastic, or obstructive bronchitis
1.05 [0.92,1.20]
For respiratory infection 1.09 [0.94,1.07]
For sneezing, runny or stuffed nose 1.19 [1.04,1.36]
December 2009
E-427
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Morgenstern
et al. (2007, 0907471
Period of Study: May
1999-Jul2000
Location: Munich,
Germany
Outcome: Asthma, wheezing,
spastic/obstructive bronchitis. Dry
cough at night, respiratory
infections, sneezing, runny/stuffed
nose without a cold.
Age Groups: at 1 yr & at 2 yr
Study Design: Cohort
N: 3577 children for the prediction
models. Respiratory data were
available for 3129 children at 1 yr.
Statistical Analyses: Pearson's
correlation coefficient, prediction
error expressed as root mean
squared error (RMSE), multiple
logistic regression with confounding
factors, odds ratios
Covariates: Sex, Parental atopy
(genetic predisposition to allergies),
environmental tobacco smoke at
home, maternal education >or <12
yr, sibling, gas stove, home
dampness, indoor mold, pets. Since
it was not feasible to measure
personal exposure to N02, PM25,
and PM25 absorbance, exposure
modeling was used.
Statistical Package: SAS V.8.02
Pollutant: PM25 Absorbance (PM25
ab)
Averaging Time: Annual
Mean(SD):1.710-5m-1,
Percentiles:25th:1.610-5m-1
50th(Median):1.710-5m-1
75th: 1.8 10 -5 m-1
Range (Min, Max):
1.3, 3.2 10-5 m-1
Unit(i.e. ug/m3): 10-5 m-1
Monitoring Stations: 40: traffic,
n = 17 and background, n = 23.
PM Increment: 0.22x10-5
Odds Ratio [Lower Cl, Upper Cl]
no lag
At age 1 yr
For wheezing 0.97 [0.77,1.23]
For cough without infection 1.16[0.87,1.54]
For dry cough at night! .09 [0.78,1.51]
For asthmatic, spastic, or obstructive bronchitis
1.14 [0.88,1.48]
For respiratory infections'! .03 [0.86,1.24]
For sneezing, runny or stuffed nose 1.30 [1.03,1.1
At age 2 yr
For wheezing 1.09 [0.90,1.33]
For cough without infection NR insufficient data
For dry cough at night!. 18 [0.93,1.50]
For asthmatic, spastic, or obstructive bronchitis
0.85 [0.30, 2.34]
For respiratory infections'! .05 [0.79,1.39]
For sneezing, runny or stuffed nose
1.27 [1.04,1.56]
Reference: Oftedal et al.
(2008, 0932021
Period of Study:
2001-2002
Location: Oslo, Norway
Outcome: Lung function (PEF,
FEF25%, FEF50%, FEV,, FVC)
Age Groups: 9-1 Oyr
Study Design: Cross-sectional
N: 1847 children
Pollutant: PM25
IQR:
PM2 5 in 1st yr of life: 6.2
PM25 lifetime: 3.6
PM Increment: Per IQR
P (Lower Cl, Upper Cl)
PM25in1styroflife
PEF -76.1 (-122.2(0-30.0)
Statistical Analyses: Linear
regression
Covariates: Height, age, BMI, birth
weight, temperature, maternal
smoking, se
Dose-response Investigated? Yes
Statistical Package: SPSS,
STATA, S-Plus
Lags Considered: 1-3
FEF25%-75.6 (-127.4 to-23.8)
FEF 50%-62.4 (-107.4 to-17.4)
FEV, -12.7 (-28.8, 3.4)
FVC-2.9 (-20.5, 14.7)
PM25 lifetime exposure
PEF-57.7 (-94.4(0-21.1)
FEF25%-51.8 (-93.1 to-10.6)
FEF 50%-48.4 (-84.2 to-12.6)
FEV,-10.4 (-23.2, 2.4)
FVC-3.9 (-17.9, 10.1)
Reference: (Parker et
al., 2009, 1923591
Period of Study: 1999-
2005
Location: U.S.
Outcome: Respiratory
allergy/hayfever
Study Design: Cohort
Covariates: Survey yr, age, family
structure, usual source of care,
health insurance, family income
relative to federal poverty level,
race/ethnicity
Statistical Analysis: Logistic
regression
Statistical Package: SUDAAN
Age Groups: 73,198 children aged
3-17 yr
Pollutant: PM25
Averaging Time: NR
Median: 13.1
IQR: 10.9-15.2
Copollutant (correlation):
Summer 03: 0.10
S02: 0.21
N02: 0.53
PM,0.2.5: 0.02
PM10:0.51
Increment: 10 pg/m3
Odds Ratio (96% Cl)
Single Pollutant Model, variable N
Adjusted: 1.16 (1.04-1.30)
Single Pollutant Model, constant N
Adjusted: 1.23 (1.04-1.46)
Multi-pollutant Model: 1.29 (1.07-1.56)
December 2009
E-428
-------
Study
Reference: Sekine et al.
(2004, 0907621
Period of Study:
1987-1994
Location: Nine districts
in the Tokyo, Japan
metropolitan area: Chuo
ward, Ohta ward,
Shibuya ward, Itabashi
ward, Hachioji City,
Tachikawa City, Om e
City, Machida City,
Tanashi City
Reference: Sharma et
al. (2004, 1569741
Period of Study: Nov
2002-Apr2003
Location: 3 sections in
KanpurCity, India
1) Indian Institute of
Technology Kanpur (I ITK)
2) Vikas Nagar (VN)
3) Juhilal Colony (JC)
Reference: (Singh etal.,
2003, 0526861
Period of Study: NR
Location: Jaipur, India
Design & Methods
Outcome: Pulmonary function tests
Age Groups: 30-59 yr
Study Design: Cross-sectional and
longitudinal
N: 500 females
Statistical Analyses: Multiple
logistic regression analysis
Covariates: Group (classification
by air pollution level), pulmonary
function at initial test, age and
height at the time of the initial test,
number of yr investigated, yr of
residence in the area, type of
heater, housing structure, and job
status
Dose-response Investigated? No
Statistical Package: SAS
Outcome: Lung function
Age Groups: 20-55 yr
Study Design: Cohort
N: 91 people
Statistical Analyses: Linear
regression
Covariates: NR
Season: Fall, Winter, spring
Dose-response Investigated? No
Statistical Package:
Microsoft Excel
Lags Considered: 1 day lag &
5-day ma
Outcome: Lung function (peak
expiratory flow variability)
Age Groups: Medical school-aged
students
Study Design: Cross sectional
N: 313 nonsmoker students
Concentrations1
Pollutant: Suspended PM (SPM)
Averaging Time: Measured each
month for three consecutive days
(72 h)
Mean (SD): 28. 1-63.3
Range (Min, Max): 3.4-140.6
Copollutant (correlation): N0;
Pollutant: PM25
Averaging Time: 24 h
Mean(SD):IITK158(22)
VN 85 (30)
JC 59 (9)
PM Component: Lead, Nickel,
Cadmium, Chromium, Iron, Zinc
Benzene soluble fraction (includes
polycyclic aromatic hydrocarbons
[PAHs])
Copollutant (correlation):
APEF = mean daily deviations in
PEF
PM25-APEF: -0.30
PM25-PM10: 0.67
PM25-PMio (1-day lag): 0.49
PM25-PM25 (1-day lag): 0.88
Pollutant: Respirable suspended
PM (RSPM)
Averaging Time: 8 h
Mean (SD): Roadside: 1,666
Campus: 177
Monitoring Stations: 2
Effect Estimates (95% Cl)
Results of multiple logistic regression analysis for respiratory
symptoms
Persistent cough
Group 3: OR = 1.00
Group 2: OR =1.02 (0.70-1. 48)
Group 1: OR =1.07 (0.67-1. 70)
Persistent phlegm
Group 3: OR =1.00
Group 2: OR =1.51 (1.11-2.04)
Group 1:OR= 1.78 (1.26-2.53)
Asthma
Group 3: OR = 1.00
Group 2: OR =1.99 (0.82-4.83)
Group 1: OR = 2.66 (0.98-7.19)
Wheeze
Group 3: OR = 1.00
Group 2: OR =1.39 (0.95-2.01)
Group 1:OR= 1.34 (0.85-2.11)
Breathlessness
Group 3: OR =1.00
Group 2: OR = 0.84 (0.47-1.50)
Group 1: OR = 2.70 (1.48-4.91)
PM Increment: 1 pg/m3
APEF (difference or change in peak expiratory flow)
-0.0297 L/min
It appears that no associations between particulates and the
outcome of interest were calculated and reported in this study
Statistical Analyses: Amplitude %
mean was used as the measure of
PEF variability. Mean value of
amplitude % mean of peak flow
variability were compared for in the
two groups by application of
Student's t-test. The two groups
were: living on campus and
commuters.
Dose-response Investigated? Yes
December 2009
E-429
-------
Study
Reference: (Solomon et
al, 2003, 0874411
Period of Study:
1966-1997
Location: United
Kingdom: Northern
England, North-West
Midlands, and V\feles.
Design & Methods
Outcome: Cardio-respiratory
morbidity
Age Groups: 45 yr and older
Study Design: Cross-sectional
N: 1,1 66 women
Statistical Analyses: Prevalence
Concentrations1
Pollutant: Black Smoke
Averaging Time: Annual
Effect Estimates (95% Cl)
RR Estimate [Lower Cl, Upper Cl]
The findings provide no indication that prolonged residence in
places that have had relatively high levels of particulate air
pollution causes an important increase in cardio-respiratory
morbidity.
Prevalence ratios are based on high vs. low pollution with low
as referent.
heart disease, asthma, productive
cough, wheeze, and use of an
inhaler for asthma or other breathing
problems.
Covariates: Smoked, passive
smoking in childhood, tenancy, SES,
worked in industry with respiratory
hazards, childhood admission to
hospital for chest problem, diabetes,
BMI were all controlled for as
potential confounders.
Dose-response Investigated? yes
Statistical Package: STATA
Particulate pollution in place of residence:
Rr = 1.0 (0.7-1.4) for ischemic heart disease;
Rr = 0.7 (0.5-1.0) for asthma
Rr = 1.0 (0.7 -1.5) for productive cough
Reference: Suglia et al.
(2008, 1570271
Period of Study: Mar
1986-0ct1992
Location: Boston, MA
Outcome: Lung function Pollutant: Black Carbon (BC)
Age Groups: 18-42 Averaging Time: Annual
Study Design: Prospective cohort Mean (SD): 0.62 (0.15)
N: 272 women of childbearing age
Statistical Analyses: Linear
regression
Covariates: Height, age, weight,
race/ethnicity, yr, education
Dose-response Investigated?
yes-tertiles of exposure
Statistical Package: SASv. 9.0
PM Increment: 0.22 pg/m3 (IQR)
Effect Estimate [Lower Cl, Upper Cl]
FEV,: -1.08 (-2.5, 0.3)
FVC: -0.62 (-1.9, 0.6)
FEF25-75%:-2.97(-5.8to-0.2)
Current Smokers:
FEV,: 0.62 (-2. 1,3.4)
FVC: 0.64 (-2.0, 3.3)
FEF25-75%: -2.63 (-3.7, 8.9)
Former Smokers1
FEVi:-4.40(-7.8to-1.0)
FVC: -3. 11 (-6.1(0-0.2)
FEF25-75%: -8.78 (-14.7 to -2.9)
Nonsmokers:
FEV,: -0.98 (-2.9, 0.9)
FVC: -0.32 (-2.0, 1.4)
FEF25-75%:-4.39(-8.1to-0.6)
Exposure-response relationship presented graphically in Fig
1: the highest BC exposure group had decreases in FEV,,
FVC, and FEF25-75% compared with the lowest fertile group,
although these differences were not statistically significant.
Reference: (Sunyer et
al., 2006, 0897711
Period of Study: initial
selection: 1991-1993,
follow-up Jun 2000-Dec
2001
Location: 21 centers in
10 European countries
Outcome: Chronic bronchitis
Age Groups: Mean age (range)
Males-42.62 (38.12-45.62)
Females- 42.57 (39.92-45.69)
Study Design: Hierarchical models
N:6924
Statistical Analyses: General
additive models (GAM)
Covariates: Smoking, age at end of
education, occupational group,
occupational exposures, respiratory
infections during childhood, rhinitis,
asthma, traffic intensity at household
level.
Statistical Package: STATA-8
Pollutant: PM25
Averaging Time: 18 mo
Mean (SD): 3.7-44.9
Copollutants: N02, S02
PM Increment: NR
Odds ratio [Lower Cl, Upper Cl]
Chronic phlegm prevalence at follow up
Males: 0.97 [0.70,1.35]
December 2009
E-430
-------
Study
Reference: Zhang et al.
(2002, 0348141
Period of Study:
1993-1996
Location: 4 Chinese
cities (urban and
suburban location in
each city): Guangzhou,
Wuhan, Lanzhou,
Chongqing
Design & Methods
Outcome: Interview-self reports of
symptoms: Wheeze (ever wheezy
when having a cold)
Asthma (diagnosis by doctor)
Bronchitis (diagnosis by doctor)
Hospitalization due to respiratory
disease (ever)
Persistent cough (coughed for at
least 1 month per yr with or apart
from colds)
Persistent phlegm (brought up
phlegm or mucus from the chest for
at least 1 month per yr with or apart
from colds).
Age Groups: Elementary school
students
age range: 5.4-16.2
Study Design: Cross-sectional
N: 7,557 returned questionnaires
Concentrations1
Pollutant: PM25
Averaging Time: 2 yr
Mean (SD): 92 (31)
Percentiles:
25th: NR
SOth(Median): NR
75th: NR
IQR: 39
Range (Min, Max):
Gives range (max.-min.):
PM25-98
Monitoring Stations: 2 types:
municipal monitoring stations over
a period of 4 yr (1993-1996)
schoolyards of participating
children over a period of 2 yr
(1995-1996)
Effect Estimates (95% Cl)
PM Increment: Interquartile range corresponded to 1 unit of
change.
RR Estimate [Lower Cl, Upper Cl]
lag:
No association between PM25 and any type of respiratory
morbidity.
No between or within city association between PM25 and any
type of respiratory morbidity.
When scaled to an increment of 50 pg/m3 increase in PM25,
association (ORs) between respiratory outcome and PM25
was:
Wheeze: 1.06
Asthma: 1.29
Bronchitis: 1.68
Hospitalization: 1.08
Persistent cough: 1.24
Persistent phlegm: 3.09
7,392 included in first stage of
analysis
Statistical Analyses: 2-stage
regression approach:
Calculated odds ratios and 95% CIs
of respiratory outcomes and covari-
ates Second stage consisted of vari-
ance-weighted linear regressions
that examined associations between
district-specific adjusted prevalence
rates and district-specific ambient
levels of each pollutant.
Covariates: Age, gender, breast-fed,
house type, number of rooms,
sleeping in own or shared room,
sleeping in own or shared bed, home
coal use, ventilation device used,
homes smokiness during cooking,
eye irritation during cooking, parental
smoking, mother's education level,
mother's occupation, father's
occupation, questionnaire
respondent, yr of questionnaire
administration, season of question-
naire administration, parental asthma
prevalence.
All units expressed in pg/m unless otherwise specified.
December 2009
E-431
-------
Table E-25. Long-term exposure - respiratory morbidity outcomes - other PM size fractions.
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: El-Zein et al. (2007,
0930431
Period of Study: 2000-2004
Location: Beirut, Lebanon
ED Admissions
Outcome: Acute respiratory symptoms:
asthma, URTI, pneumonia, bronchitis
Age Groups: <17
Study Design: Ecological (natural
experiment comparing admissions
before and after ban on diesel fuel)
N: 5 hospitals, 7573 admissions Oct-
Feb, 4303 admissions Oct-Dec
Statistical Analyses: T-test, Poisson
regression
Covariates: Month of Year,
temperature, humidity, orthogonalized
rainfall
Season: Oct-Dec (excluding flu
season) and Oct-Feb
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: 1-2 yr before the
ban compared to 1-2 yr after the ban
Pollutant: PM from diesel
Range (Min, Max): NR
PM Component: NR
Monitoring Stations: 1
Notes: Did not look at specific
exposure data
looked at outcome with respect to a
timeline that plotted admissions before
and after a ban on diesel fuel.
Copollutant: NR
PM Increment: NA
|3 (p-value):
2-yr pre-ban vs. 2-yr post-ban
Oct to Feb
AIIResp:0.128
0.32
0.16
Asthma:-0.176
Bronchitis: 0.505 (0.02)
Pneumonia: 0.287 (0.17)
URTI:-0.265 (0.41)
Oct to Dec
All Resp: -0.022 (0.87)
Asthma: -0.21 (0.07)
Bronchitis: 0.2 (0.35)
Pneumonia: -0.065 (0.78)
URTI: -0.628 (0.05)
2-yr pre-ban vs. 1-yr post-ban
Oct-Feb
All Resp: -0.093 (0.45)
Asthma: -0.208 (0.05)
Bronchitis: 0.286 (0.32)
Pneumonia: -0.07 (0.76)
URTI:-0.715 (0.11)
Oct to Dec
All Resp:-0.147 (0.02)
Asthma:-0.147 (0.00)
Bronchitis: -0.011 (0.96)
Pneumonia:-0.214 (0.15)
URTI: -0.885 (0.06)
1-yr pre-ban vs. 1-yr post-ban
Oct-Feb
All Resp:-0.165 (0.04)
Asthma: -0.212 (0.09)
Bronchitis: 0.059 (0.85)
Pneumonia: -0.034 (0.84)
URTI:-1.023 (0.00)
Oct to Dec
All Resp:-0.17 (0.00)
Asthma:-0.131 (0.00)
Bronchitis:-0.145 (0.001)
Pneumonia:-0.168 (0.12)
URTI:-1.036 (0.00)
Reference: Kasamatsu et al. (2006,
1566271
Period of Study: 2001-2002
Location: Shenyang, China
Outcome: FVC, FEV,, PEF, FEF75 Pollutant: PM7
Age Groups: School Children aged 8-
10
Study Design: Children in three
schools in three types of areas
(commercial city area, residential city
area, residential suburban area) invited
to participate
N: 322 children participated, 244 have
complete data.
Statistical Analyses: Genralized
estimating equations
Covariates: Age, height,
Dose-response Investigated? No
Statistical Package: SAS
Lags: Considered: previous quarter.
Averaging Time: Avg of 4 separate 2-7
consecutive day measurements within
each designated measurement month
of the quarter
Mean (SD):
School A
7/2001 86.4(14.2)
10/2001 114.1(35.1)
1/2002 118.2(28.2)
4/2002 182.7(102.1)
School B
7/2001 90.1(8.3)
10/2001 161.5(45.7)
1/2002 118.8(28.2)
4/2002152.0(31.3)
School C
7/2001 78.1(16.9)
10/2001 131.2(29.6)
1/2002 142.2(37.6)
4/2002173.6(121.5)
PM Component: mainly pollutants
associated with coal heating
Monitoring Stations: 1 at each location
PM Increment: 63.0 pg/m
Mean change of pulmonary function
value [Lower Cl, Upper Cl] at lag 0
Boys
FVC-0.095(-0.170,-0.019)
FEVi-0.088(-0.158,-0.019)
PEF-0.170(-0.365,0.032)
FEF75-0.063(-0.183,0.050)
Girls
FVC-0.082(-0.145,-0.019)
FEV,-0.069(-0.126,-0.006)
PEF0.095(-0.095,0.290)
FEF75-0.032(-0.151,0.082)
Mean change of pulmonary function
value [Lower Cl, Upper Cl] at lag
1 (previous quarter)
Boys
FVC-0.145(-0.189,-0.095)
FEV1-0.095(-0.139,-0.057)
PEF-0.082(-0.208,0.050)
FEF750.013(-0.063,0.088)
Girls
FVC-0.126(-0.170,-0.088)
FEVi-0.101(-0.139,-0.063)
PEF-0.101(-0.227,0.025)
FEF75-0.057(-0.132,0.019)
December 2009
E-432
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Kasamatsu et al.(2006,
1566271
Period of Study: 2001-2002
Location: Shenyang, China
Outcome: FVC, FEV,, PEF, FEF75
Age Groups: School Children aged 8-
10
Study Design: Children in three
schools in three types of areas
(commercial city area, residential city
area, residential suburban area) invited
to participate
N: 322 children participated, 244 have
complete data.
Statistical Analyses: Genralized
estimating equations
Covariates: Age, height,
Dose-response Investigated? no
Statistical Package: SAS
Lags: Considered: previous quarter.
Pollutant: PM2,
Averaging Time: Avg of 4 separate 2-7
consecutive day measurements within
each designated measurement month
of the quarter
Mean (SD):
School A
7/2001 47.6(6.4)
10/2001 54.2(20.5)
1/200268.9(15.8)
4/2002 115.8(76.7)
School B
7/2001 45.6(6.5)
10/2001 74.4(27.1)
1/200263.3(17.9)
4/2002 96.3(27.6)
School C
7/2001 42.5(9.5)
10/2001 59.7(13.1)
1/200276.4(22.1)
4/2002 123.0(100.9)
PM Component: mainly pollutants
associated with coal heating
Monitoring Stations: 1 at each location
PM Increment: 42.1 pg/m
Mean change of pulmonary function
value [Lower Cl, Upper Cl] at lag 0
Boys
FVC-0.126(-0.181,-0.076)
FEV1-0.122(-0.173,-0.076)
PEF-0.164(-0.303,-0.025)
FEF75-0.046(-0.131,0.038)
Girls
FVC-0.110(-0.156,-0.067)
FEVi-0.101(-0147,-0.059)
PEF0.008(-0.131,0.147)
FEF75-0.055(-0.139,0.030)
Mean change of pulmonary function
value [Lower Cl, Upper Cl] at lag
1 (previous quarter)
Boys
FVC-0.099(-0.145,-0.053)
FEV1-0.059(-0.106,-0.020)
PEF-0.040(-0.158,0.086)
FEF750.026(-0.046,0.092)
Girls
FVC-0.086(-0.125,-0.046)
FEVi-0.066(-0.106,-0.026)
PEF-0.079(-0.198,0.040)
FEF75-0.033(-0.106,0.040)
All units expressed in pg/m unless otherwise specified.
December 2009
E-433
-------
E.6. Long-Term Exposure and Cancer
Table E-26. Long-term exposure - cancer outcomes - PMio.
Study
Design & Methods
Concentrations
Effect Estimates (95% Cl)
Reference: (Abbey et al, 1999,
0475591
Period of Study: 1977-1992
Location: California
Outcome (ICD9): Lung Cancer
Mortality (162)
Age Groups: 27-95 at baseline
Study Design: Cohort (AHSMOG)
N: 6,338 nonsmoking CA Seventh-Day
Adventists
Statistical Analyses: Time-dependent,
gender-specific, Cox proportional
hazards regression models
Covariates: Age, smoking, education,
occupation, BMI
Pollutant: PM10
Averaging Time: Monthly estimates
from 1966-1992
Mean (SD): 51.24 (16.63)
Percentiles: IQR: 24.08
Range (Min, Max): 0, 83.9
Correlations:
S04:r = 0.68)
S02:r = 0.31
03:r = 0.77
N02:r = 0.56
Lag:3 yr
PM Increment: 24.08 (IQR)
RR, males: 3.36 [1.57, 7.19]
RR, females: 1.33 [0.60, 2.96]
PM10 above 100ug/m3 (days peryr)
IQR: 43 days/yr
Males: 2.38 (1.42, 3.97)
Females: 1.08 (0.55, 2.13)
Reference: Beeson et al. (1998,
0488901
Period of Study: 1977-1992
Location: California
Outcome (ICD9: Lung Cancer Mortality Pollutant: PM
(ICDO-1:162, ICDO-2: C34.0-C34.9)
Age Groups: 27-95 at baseline
Study Design: Cohort (AHSMOG)
N: 6,338 nonsmoking CA Seventh-Day
Adventists (non-Hispanic white)
Statistical Analyses: Time-dependent,
gender-specific, Cox proportional
hazards regression models
Covariates: Smoking, Education, Age,
Alcohol
Statistical Package: SAS
Lags Considered: 3 yr
Averaging Time: Averaged monthly
estimates from 1966-1992
Mean (SD): 51 (16.52)
Percentiles: IQR: 24
Range (Min, Max): 0, 84
PM Increment: 24 (IQR)
RR, males: 5.21 [1.94,13.99]
RR, females: Positive, but not
statistically significant
Reference: Binkova et al. (2007,
1562731
Period of Study: Feb 2001
Location: Prague, Czech Republic
Outcome: Total DNAadducts (bulky
aromatic PAH-DNAadducts and ...
Age Groups: 22-50 yr
Study Design: Case Control
N: 53 occupationally exposed
policemen and 52 control policemen
Statistical Analyses: Multivariate
logistic regression, Mann-Whitney u-test
Covariates: Smoking. Vitamin C,
polymorphisms of XPD repair gene in
exon 23 and 6 and GSTM 1 and
XRCC1 genes
Season: Winter
Pollutant: PM,0
Range (Min, Max): 32-55
Monitoring Stations: 2 (and personal
monitors)
No relationship between short term
exposure to C-PAHs evaluated by
personal monitors and DMA adduct
level. Genetic damage was observed in
city policemen working in winter
outdoors in the Prague downtown area
they had slightly elevated aromatic DMA
adduct levels, which was statistically
significant for a distinct DMA adduct
spot that could originate from ambient
exposure to B[a]P
Total PAH-DNA adducts: p = 0.065
Exposed: 0.92 ± 0.28 adducts/108
nucleotids
Control: 0.82 ±0.23 adducts/108
nucleotids
B[o]P-like adducts:
Exposed: 0.122 ± 0.36 adducts/108
nucleotids
Control: 0.099 ± 0.035 adducts/108
nucleotids
Multiple regression "like" B[a]P-DNA
adduct for air pollution exposure group:
6 = 0.016, p = 0.01
December 2009
E-434
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: (Liu et al, 2009,1902921
Period of Study: 1995-2005
Location: Taiwan
Outcome: Bladder Cancer Mortality
(ICD-9 188)
Age Groups: 50-69
Study Design: Case-crossover
Statistical Analysis: Multiple Logistic
Regression
Statistical Package: NR
Covariates: none
Dose-response Investigated? No
Pollutant: PM,0
Averaging Time: Annual mean of 24-h
avg
Tertiles (median):
T1:<52.80
72:53.04-71.72
T3: 72.24-90.29
Copollutant: 03, CO, N02, S02
Copollutant (correlation): NR
Monitoring Sattions: 64
Increment:
Odds Ratio (M in Cl, Max Cl)
Lag
T1vs. 71:1.00 (ref)
T2 vs. 71:1.08 (0.83-1.41)
T3 vs. 71:1.39(1.06-1.83)
P for trend = .020
Reference: (Pope et al., 2002, 0246891
Period of Study: 1982-1998
Location: 50 U.S. states, District of
Columbia, and Puerto Rico
Outcome (ICD9): Lung cancer mortality
(162)
Age Groups: Ages >30 yr Study
Design: Longitudinal cohort (Cancer
Prevention Study II)
N: 1.2 million people
Statistical Analyses: Cox proportional
hazard, generalized additive
Covariates: Age, sex, race, education,
smoking status, marital status,
occupational exposure, diet, body-mass
index, alcohol consumption
Pollutant: PM,0
Mean (SD): 1982-1998: 28.8(5.9)
Effect estimates: Effect estimates were
recorded in Fig 5 and not presented
quantitatively anywhere else
Reference: Sram et al, (2007, 1884571
Period of Study: Jan and Mar of 2004
Location: Prague, Czech Republic
Outcome: Chromosomal aberrations
Study Design: Panel
Covariates: Urinary cotinine, plasma
levels of vitamins A, E and C, folate,
total cholesterol, HDL and LDL
cholesterols, and triglycerides
Statistical Analysis: Bivariate
correlations, ANOVA, Mann-Whitney,
Kruskal-Wallis and Spearman rank
correlation
Statistical Package: S7A7IS7ICA
Age Groups: 61 city policemen, aged
34 + 8 yr, spending 8+ h outdoors
Pollutant: PM,0
Averaging Time: NR
Mean (SD) Unit:
Jan: 55.6 pg/m3
Mar: 36.4 pg/m3
Copollutant: PM25
Results not given by PM increment.
Reference: Sram et al, (2007, 1884571
Period of Study: Jan and Mar of 2004
Location: Prague, Czech Republic
Outcome: Chromosomal aberrations
Study Design: Panel
Covariates: Urinary cotinine, plasma
levels of vitamins A, E and C, folate,
total cholesterol, HDL and LDL
cholesterols, and triglycerides
Statistical Analysis: Bivariate
correlations, ANOVA, Mann-Whitney,
Kruskal-Wallis and Spearman rank
correlation
Statistical Package: S7A7IS7ICA
Age Groups: 61 city policemen, aged
34 + 8 yr, spending 8+ h outdoors
Pollutant: PM25
Averaging Time: NR
Mean (SD) Unit:
Jan: 44.4 pg/m3
Mar: 24.8 pg/m3
Copollutant: PM,0
Results not given by PM increment.
December 2009
E-435
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: (Tarantini et al., 2009,
192010)Period of Study: NR
Location: Brescia, Italy
Outcome: DMA methylation content
estimated byAlu, LINE-1 and iNOS
analysis
Study Design: Panel
Covariates: age, BMI, smoking,
number of cigarettes/day
Statistical Analysis: Mixed effects
models
Statistical Package: NR
Age Groups: 63 male workers between
27 and 55 yr, mean age 44.
Pollutant: PM,0
Averaging Time: NR
Mean (SD) Unit: NR
Individual Exposure Range: 73.4-
1220|jg/m3
Copollutant (correlation): NR
Difference in DMA Methylation before
and after work exposure, mean (SE)
Alu (%5mC): 0.00 (0.08), p = 0.99
LINE-1 (%5mC): 0.02 (0.11), p = 0.89
iNOS (%5mC): -0.61 (0.26), p = 0.02
Reference: (Vineis et al, 2006,
1920891
Period of Study: 1990-1999
Location: 10 European countries
Reference: (V\fei et al, 2009, 1923611
Period of Study: Nov 2006-Jan 2007
Location: Peking, China
Outcome: Lung cancer
Study Design: Nested case-control
Covariates: Age, sex, country, smoking
status, time since recruitment,
education, BMI, physical activity, intake
of fruit, vegetables, meat, alcohol and
energy
Statistical Analysis: Conditional
logistic regression models
Statistical Package: NR
Age Groups: 35-74 at recruitment
Outcome: Urinary 8-OHdG increase
Study Design: Panel
Covariates: NR
Statistical Analysis: Analysis of
variance model with autoregressive
terms
Pollutant: PM,0
Averaging Time: NR
Mean by Country (ug/m3):
France
lle-de-France
1990-1994:22.3
1995-1999: 19.9
Northeast France
1990-1994:30.2
1995-1999:29.5
Italy
Turin
1990-1 994' 73 4
1995-1999:61.1
Florence
1990-1994:40.4
1995-1999:33.3
United Kingdom
Oxford
1990-1994:29.0
1995-1999:25.5
Cambridge
1990-1994: NR
1995-1999:25.4
The Netherlands
Utrecht
1990-1994:42.8
1995-1999:40.0
Bilthoven
1990-1994:39.0
1995-1999:37.2
Germany
Heidelberg
1990-1994: NR
1995-1999:27.0
Potsdam
1990-1994:32.0
1995-1999:28.9
Range (Min, Max): NR
Copollutant: N02, 03, S02
Pollutant: PM25
Averaging Time: 24 h
Median: 154. 87 pg/m3
IQR: 166.29
Copollutant (correlation): NA
Increment: 10|jg/m3
Odds Ratios (Min Cl, Max Cl) for
increase in lung cancer per
increment increase in PM-:
0.91 (0.70-1.18)
Increment: 1 66.29 pg/m3
8-OHdG Concentrations, pre and
post-work shift, subjects avgd
Pre-work: 1.83
Post-work: 6.92
Statistical Package: SAS
Age Groups: Two nonsmoking security
guards, ages 18 and 20
Concentration Changes (96%CI) of 8-
OHdG per IQR Increase
Pre-work: 0.256 (0.040, 0.472), p =
0.021
Post-work: 2.370 (0.907, 3.833), p =
0.002
All units expressed in pg/m unless otherwise specified.
December 2009
E-436
-------
Table E-27. Long-term exposure - cancer outcomes - PIVhs (including PM components/sources).
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Baccarelli et al, (2009,
1881831
Outcome: DMA methylation of LINE-1
andAlu
Period of Study: Jan 1999-Jun 2007 Study Design: Panel
Location: Boston, Massachusetts
Covariates: age, BMI, smoking status,
pack-yr, statin use, fasting blood
glucose, diabetes mellitus, percent
lymphocytes and neutrophils in
differential blood count, day of the
week, season, temperature
Statistical Analysis: Mixed effects
models
Statistical Package: SAS
Age Groups: 719 elderly individuals,
mean age 73.3, range 55-100 yr
Pollutant: PM25
Averaging Time: NR
Mean (SD) Unit:
4h: 12.2 (7.7) pg/m3
1 day: 10.9 (6.3) pg/m3
2 day: 10.6 (5.2 pg/m3
3 day: 10.4(4.8 pg/m3
4day:10.3(4.3)|jg/m3
ZA- ^n 0 Q Q I in/r^
Increment: SD for each lag
Correlation Coefficient (96% Cl)
Lag for LINE-1 Methylation
4h: -0.07 (-0.13, -0.01), p = 0.03
1 day: -0.09 (-0.16, -0.02), p = 0.008
2 day: -0.10 (-0.17, -0.03), p = 0.003
pg/m3
3 day:-0.10
4 day:-0.10
-0.17, -0.04), p = 0.003
-0.16, -0.03), p = 0.004
7d: 10.3 (3.3) pg/m3
Copollutants: Black carbon, Sulfate
5d: -0.10 (-0.16, -0.03), p = 0.004
6d:-0.11 (-0.17, -0.04), p = 0.001
7d:-0.13 (-0.19,-0.06), p< 0.001
Correlation Coefficient (96% Cl)
Lag for Alu Methylation
4h: 0.03 (-0.03, 0.09), p = 0.28
1 day: -0.01 (-0.07, 0.05), p = 0.74
2 day: -0.01
3 day:-0.01
-0.07, 0.05), p = 0.82
-0.07, 0.05), p = 0.78
4 day: -0.01 (-0.07, 0.05), p = 0.75
5d:-0.01
6d:-0.01
-0.07, 0.05
-0.07, 0.05
, p = 0.84
, p = 0.74
7d: -0.01 (-0.07, 0.05), p = 0.71
Correlation Coefficient (96% Cl)
LINE-1 Methylation and ma of
pollutant levels
4h: -0.04 (-0.11, 0.03), p = 0.24
7d: -0.11 (-0.18, -0.05), p = 0.001
Reference: Binkova et al. (2007,
1562731
Period of Study: Feb 2001
Location: Prague, Czech Republic
Outcome: Bulky aromatic PAH-DNA
adducts
Age Groups: 22-50 yr
Study Design: Case Control
N: 53 exposed policemen and 52
control policemen
Statistical Analyses: Multivariate
logistic regression, Mann-Whitney,
Rank-Sum U-test
Covariates: Smoking. Vitamin C,
polymorphisms of XPD repair gene in
exon 23 and 6 and GSTM 1 gene
Season: Winter
Pollutant: PM25
Range (Min, Max): 27-38
c-PAHs: range = 18-22 ng/m3
B[a]P: range = 2.5-3.1 ng/m3
Monitoring Stations: 2
Genetic damage was observed in city
policemen working in winter outdoors in
the Prague downtown area
They had slightly elevated aromatic
DMA adduct levels, which was more
pronounced for a distinct DMA adduct
spot that could originate from ambient
exposure to B[a]P
Total DNA-adduct level
Exposed: 0.92+0.28 adducts/108
nucleotides
Control: 0.82+0.23 adducts/108
nucleotides
p = 0.065
"Like" B[a]P-derived DNA adducts
Ex posed: 0.122+0.036
Control: 0.101+0.035
p < 0.01
Multiple Regression (exposed vs.
control)
B = 0.016, p = 0.011
December 2009
E-437
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Brunekreef et al, (2009,
1919471
Period of Study: 1987-1996
Location: The Netherlands
Outcome: Air pollution related lung
cancer deaths (ICD-9 162)
Study Design: Case-cohort
Covariates
Individual: Sex, age, Quetelet index,
smoking status, passive smoking
status, educational level, occupation,
occupational exposure, marital status,
alcohol use, intake of vegetables, fruits,
energy, saturatured and
monounsaturated fatty acids, trans fatty
acids, total fiver, folic acid and fish
Area-level: Percent of population with
income below the 40th percentile and
above the 80th percentile
Statistical Analysis: Cox proportional
hazards
Statistical Package: State, SPSS, R
Age Groups: 120,000 adults aged 55-
69 yr at enrollment
Pollutant: PM25, estimated from PM10
levelsf
Averaging Time: 24 h
60th Percentile: 28 pg/m3
Range (Min, Max): 23-37
Copollutant (correlation):
N02: 0.75
Black Smoke: 0.84
NO: 0.69
SO,: 0.43
Increment: 10|jg/m
Relative Risk (96% Cl) for
associations between PM25 and lung
cancer incidence
Case Cohort
Unadjusted: 0.93 (0.71-1.22)
Adjusted: 0.67 (0.41-1.10)
Unadjusted Complete: 0.87 (0.60-1.25)
Full Cohort
Unadjusted: 0.96 (0.79-1.18)
Adjusted: 0.81 (0.63-1.04)
Unadjusted Complete: 0.92 (0.74-1.15)
Reference: Liu et al. (2008,1567081
Period of Study: 1995-2005
Location: Taiwan
Outcome: Brain cancer deaths
ICD9:191
Age Groups: 29 yr of age or younger
Study Design: Matched case-control
by sex, yr of birth and death
N: 340 matched pairs
Statistical Analyses: Conditional
logistic regression
Covariates: Age, gender, urbanization
level of residence, nonpetrochemical air
pollution exposure level
No direct measures of pollutants
used an index to assign petrochemical
air pollution exposure (each municipality
was assigned an exposure by dividing
the number of workers per municipality
employed in the petrochemical industry
by the municipalities total population).
Study participants divided into tertiles
based on this index.
People who lived in the group of
municipalities with the highest levels of
air pollutants arising from petrochemical
sources were at a statistically significant
increased risk for brain cancer
development compared to the group
living in municipalities with the lowest
petrochemical air pollution exposure
index.
Effect Measure: OR (95%CI)
Tertile1:1.?0
Tertile 2:1.54 (0.98-2.42)
Tertile 3:1.65 (1.00-2.73)
P for trend <0.01
Reference: Nafstad et al. (2004,
0879491
Period of Study: 1972-1998
Location: Oslo, Norway
Outcome: Lung cancer
ICD7 162.1-162.9
Age Groups: 40-49 yr old men
Study Design: Cohort
N: 16,209 males
Statistical Analyses: Cox regression
models (proportional hazards)
Covariates: Age at inclusion, smoking
habits, education
Season: all yr
PM values had small variations in
exposure level, and strong correlations
with another pollutant of interest (S02)
and were not considered in analyses.
Co pollutants:
S02
NOX
No effect estimates for PM
Reference: (Pope and Burnett, 2007,
0909281
Period of Study: 1982-1998
Location: 50 U.S. states, District of
Columbia, and Puerto Rico
Outcome: Lung cancer mortality (162)
Age Groups: >30 yr
Study Design: Longitudinal cohort
(Cancer Prevention II Study)
N: 415,000 CPS 11 patients with
information involving PM10
Statistical Analyses: Cox proportional
hazard, incorporating a spatial random-
effects component
Covariates: Age, sex, race, education,
ETS, smoking status, marital status,
occupational exposure, diet, body-mass
index, alcohol consumption
Pollutant: PM25
Mean (SD): 1979-1983: 21.1(4.6)
1999-2000:14.0(3.0)
Avg: 17.7(3.7)
Averaging time: 1982-1998
PM Increment: 10 pg/m
RR Estimate [Lower Cl, Upper Cl]
Lung Cancer: 1979-1983:1.08[1.01,
1.16]
1999-2000:1.13[1.04, 1.22]
Avg:1.14[1.04,1.23]
RR results were also presented in Fig
2-5. Authors found that PM25 had the
strongest association with increased
risk of all-cause, cardiopulmonary, and
lung cancer mortality.
December 2009
E-438
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Sram et al, (2007, 1884571 Outcome: Chromosomal aberrations
Period of Study: Feb 2001 Study Design: Panel
Location: Prague, Czech Republic
Covariates: Urinary cotinine, plasma
levels of vitamins A, E and C
Statistical Analysis: Bivariate
correlations, ANOVA, Mann-Whitney,
Kruskal-Wallis and Spearman rank
correlation
Statistical Package: STATISTICA,
SAS
Age Groups: 53 city policemen, aged
22-50 yr, spending 8+ h outdoors
Pollutant: PM,0
Averaging Time: NR
Range: 32-55|jg/m3
Copollutant: PM25
Results not given by PM increment.
Reference: Sram et al, (2007, 1884571 Outcome: Chromosomal aberrations
Period of Study: Feb 2001 Study Design: Panel
Location: Prague, Czech Republic
Covariates: Urinary cotinine, plasma
levels of vitamins A, E and C
Statistical Analysis: Bivariate
correlations, ANOVA, Mann-Whitney,
Kruskal-Wallis and Spearman rank
correlation
Statistical Package: STATISTICA,
SAS
Age Groups: 53 city policemen, aged
22-50 yr, spending 8+ h outdoors
Pollutant: PM25
Averaging Time: NR
Range: 27-38|jg/m3
Copollutant: PM,0
Results not given by PM increment.
Reference: Tovalin et al. (Tovalin et al., Outcome: DMA damage (comet tail
length)
2006, 091322'
Period of Study: Apr-May 2002
Location: Mexico City and Puebla
Age Groups: 18-60
Study Design: Panel Study
N: 55 male workers
Statistical Analyses: Mann-Whitney
test, Chi-square, Spearman's
correlation, logistic regression
Statistical Package: SPSS and STATA
Pollutant: PM25
Personal monitoring values observed in
this study reported in Tovalin et al. 2003
Median Personal Exposure to PM;::
Mexico City
Outdoor Worker: 133|jg/m3
Indoor Worker: 86.6 pg/m3
Puebla
Outdoor Worker: 122 pg/m3
Indoor Worker: 78.3 pg/m3
OR for being a highly damaged worker:
1.02(1.01-1.04), p = 0.03
Correlation between comet tail length
and PM 2.5: 0.57, p = 0.000
OR for being a highly damaged worker:
1.03, p< 0.07
Comet Tail Length
Outdoor Worker: 46.80 pm
Indoor Worker: 30.11 pm
p<0.01
Percent Highly DMA Damaged Cells
Outdoor Worker: 68%
Indoor Worker: 20%
All units expressed in pg/m unless otherwise specified.
December 2009
E-439
-------
Table E-28. Long-term exposure - cancer outcomes - other PM size fractions.
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: (Pope et al., 2002, 0246891 Outcome: Lung cancer mortality (162)
Period of Study: 1982-1998
Location: 50 U.S. states, District of
Columbia, and Puerto Rico
Age Groups: Ages >30 yr who were
members of a household with at least 1
individual >45yrs.
Study Design: Longitudinal cohort
(Cancer Prevention Study II)
N: 359,000 CPS II participants with
information regarding PM15 and
PM15-PM25
Statistical Analyses: Cox proportional
hazard, incorporating a spatial random-
effects component
Covariates: Age, sex, race, education,
ETS, smoking status, marital status,
occupational exposure, diet, body-mass
index, alcohol consumption
Smoking covariates adjusted for:
Indicator: current smoker, former
smoker, pipe or cigar smoker, started
smoking before or after age 18
Continuous, current and former
smokers: yr smoked, yr smoked
squared, cigarettes per day, cigarettes
per day squared, number of h per day
exposed to passive cigarette smoke.
Pollutant: PM,5
Mean (SD): 1979-1983: 40.3(7.7)
Pollutant: PM15-2.5
Mean (SD): 1979-1983:19.2(6.1)
Averaging Time: 1979-1983
Relative risks effect estimates were
recorded in Fig 5 and not presented
quantitatively anywhere else.
All units expressed in pg/m unless otherwise specified.
December 2009
E-440
-------
E.7. Long-Term Exposure and Reproductive Effects
Table E-29. Long-term exposure - reproductive outcomes - PMio.
Study
Design & Methods
Concentrations
Effect Estimates (95% Cl)
Reference: Bell at al. (2007, 0910591
Period of Study: 1999-2002
Location: Connecticut-Fairfield,
Hartford, New Haven, New London,
Windham, Massachusetts-Barnstable,
Berkshire, Bristol, Essex, Hampden,
Middlesex, Norfolk, Plymouth, Suffolk,
Worcester
Outcome: Low birth weight
Age Groups: Neonates
Study Design: Cross-sectional
N: 358,504 births
Statistical Analyses: Multiple logistic
and linear regressions
Covariates: Child's sex, mother's
education, tobacco use, mother's
marital status, mother's race, time
prenatal care began, mother's age, birth
order, gestation length
Dose-response Investigated? No
Statistical Package: NR
Pollutant: PM10
Averaging Time: 24 h
Mean (SD): 22.3 (5.3)
Monitoring Stations: NR
Copollutant: N02, CO, S02
Gestation exposure correlation:
PM25:r = 0.77
NO,: r = 0.55
PM Increment: 7.4 pg/rri (IQR)
Difference in birth weight [Lower Cl,
Upper Cl]
per IQR for the gestational period:
-8.2 [-11.1to-5.3]
Difference in birth weight by race of
mother [Lower Cl, Upper Cl]:
Black:-7.9 [-16.0, 0.2]
White:-9.0 [-12.2 to-5.9]
Range among trimester models for
change in birth weight per IQR
increase (min, max)
trimester: -6.6 to -4.7
3rd
OR Estimate for birth weight <2600 g
[Lower Cl, Upper Cl]
per IQR for the gestational period:
1.027 [0.991, 1.064]
Notes: Analyses using first births alone
yielded similar results. Two pollutant
models for uncorrelated pollutants were
analyzed but not presented
quantitatively.
Reference: Brauer et al. (2008,
1562921
Period of Study: 1999-2002
Location: Vancouver, BC
Outcome: Preterm birth, SGA, LBW
Age Groups: Study Design: Cross-
sectional
N: 70,249 births
Statistical Analyses: Logistic
regression
Covariates: Sex, parity, month and yr
of birth, maternal age and smoking,
neighborhood level income and
education
Statistical Package: SAS
Pollutant: PM,0
Averaging Time: 24-h
Mean (SD): 12.7
Range (Min, Max): 5.6, 35.4
Monitoring Stations: 19
Copollutant:
NO
N02
CO
S02
03
PM Increment: 1 pg/m
Effect Estimate [Lower Cl, Upper Cl]
pollutant assessed for entire
pregnancy period:
SGA: 1.02 (0.99, 1.05)
LBW: 1.01 (0.95, 1.08)
Preterm (<30wk): 1.13 (0.95,1.35)
December 2009
E-441
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Chen et al. (2002, 0249451
Period of Study: 1991-1999
Location: Washoe County, Nevada
Outcome: Birth weight
Age Groups: Sngle births with
gestationai age between 37-44 wk and
maternal all ages
Study Design: Cross-sectional
N: 33,859 single births
Statistical Analyses: multiple linear
and logistic regression
Covariates: infant sex, maternal
residential city, education, medical risk
factors, active tobacco use, drug use,
alcohol use, prenatal care, mother's
age, race and ethnicity of mothers and
weight gain of mothers
Dose-response Investigated? No
Statistical Package: SPSS 10.0
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 31.53 (22.32)
Percentiles: 26th: 16.80
60th(Median): 26.30
76th: 39.35
Range (Min, Max): (0.97-157.32)
Monitoring Stations: 4
Copollutant: CO
03
PM Increment: 10 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
Using continuous pollutant variables
Model 1-PM10
1 trimester
Crude model: IS = -0.186 (0.225)
Adjusted model: IS = -0.082 (0.221)
2 trimester
Crude model: 6 = 0.045(0.223)
Adjusted model: IS = -0.020 (0.221)
3 trimester
Crude model: IS = -0.509 (0.231)
Adjusted model: IS = -0.395 (0.227)
Whole
Crude model: IS = -0.823 (0.459)
Adjusted model: IS = -0.726 (0.483)
Model 2
COandPMio
3 trimester
Crude model: IS = -1.044 (0.457)
Adjusted model: IS = -1.078 (0.445)
03andPM10
3 trimester
Crude model: IS = -1.035 (0.385)
Adjusted model: IS = -0.966 (0.378)
Model 3
PMio, 03, and CO
3 trimester
Crude model: IS = -1.070 (0.458)
Adjusted model: IS = -1.102 (0.446)
Whole
Crude model: IS = -1.413 (0.733)
Adjusted model: IS = -1.332 (0.738)
Using categorical pollutant variables-3
trimester
Model 1-PM10
Adjusted model: IS = -10.243 (5.235)
Model 2
PM,o and CO
Adjusted model: IS = -11.883 (6.108)
PMio and 03 Adjusted model:
IS = -9.144 (5.860)
Model 3
PM10, CO, and 03 Adjusted model:
IS = -10.937 (6.222)
Using logistic regression
(ref value = <19.72 pg/m3
Exposure to PM10 at 3 trimester at
>44.74 pg/m3: OR =
1.105(0.714-1.709)
Between 19.72-44.74 pg/m3:
OR = 1.050 (0.811-1.360)
Notes: Crude model: model with air-
pollutant variables controlled with
gestationai age only. Adjusted model:
model with air-pollutant variables
controlled with confounding variables
including gestationai age, infant sex,
maternal residential city, education,
medical risk factors, active tobacco use,
drug use, alcohol use, the trimester
begins prenatal visits, total prenatal
visits, mother's age, race and ethnicity
of mother, and weight gain of mother.
December 2009
E-442
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Dales et al. (2004, 0873421 Outcome: SIDS (a sudden,
„ • . ,~ , , ^n^ r» -.000 unexplained death of a child <1 yr of
Period of Study: Jan 1984-Dec 1999 - ••
Location: Canada (12 cities)
age for which a clinical investigation
and autopsy fail to reveal a cause of
death)
Age Groups: Infants <1 yr
Study Design: Time-series
N: Total population of 12 cities:
10,310,309
1556 cases of SIDS over study period
Statistical Analyses: Random-effects
regression model for count data (a
linear association between air pollution
and the incidence of SIDS was
assumed on the logarithmic scale)
Covariates: Weather factors (daily
mean temp, daily mean relative
humidity, maximum change in
barometric pressure, all measured on
the day of death), length of time-period
adjustment, seasonal indicator
variables, and size-fractionated PM
Season: Used piece-wise constant
functions in time that varied by 3, 6, or
12 mo
Dose-response Investigated? No
Statistical Package: NR
Pollutant: PM,0
Averaging Time: 24-hs (PM measures
every 6 days
gaseous pollutants every day)
Mean (IQR): PM10: 23.43 (15.56)
Range (Min, Max): IQR presented
above
Monitoring Stations: When data were
available from more than 1 monitoring
site, they were avgd
Copollutant:
PM25
PM10
CO
N02
03
SO,
Notes: The abstract reports no
association between increased daily
rates of SIDS and fine particles
measured every sixth day. However, no
effect estimates presented for PM (only
gaseous pollutants adjusted for PM).
December 2009
E-443
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Dugandzic et al. (2006,
0886811
Period of Study: Jan 1988-Dec 2000
Location: Nova Scotia, Canada
Outcome: Low birth weight (LBW)
(<2500 grams)
Age Groups: Babies born 2 37 wk (full
term)
Study Design: Cross-sectional
N: 74,284 births
Statistical Analyses: Logistic
regression
Covariates: Maternal age, parity, prior
fetal death, prior neonatal death, prior
low birth weight infant, smoking during
pregnancy, neighborhood family
income, infant gender, gestational age,
weight change, yr of birth
Season: All
Dose-response Investigated? Yes
Statistical Package: SAS
Pollutant: PM,0
Averaging Time: 24-h
Mean (SD):
Percentiles:26th:14
60th(Median): 16
76th: 19
Range (Min, Max): Max: 53
Monitoring Stations: 18
Copollutant: S02, 03
Notes: Only 3 stations monitored more
than 1 pollutant. Daily data were
available for gaseous pollutants while
particulate levels were measured every
sixth day.
PM Increment:
1) IQR (5 pg/m3)
2) Quartiles (first quartile is the
reference)
Exposure period: first trimester
Unadjusted model
2nd quartile: 1.24 (0.95,1.62)
3rd quartile: 1.25 (0.96,1.62)
4th quartile: 1.28 (1.00,1.65)
Per IQR: 1.09 (1.00,1.18)
Adjusted model
2nd quartile: 1.24 (0.94,1.64)
3rd quartile: 1.24 (0.95,1.64)
4th quartile: 1.33 (1.02,1.74)
Per IQR: 1.09 (1.00,1.19)
Adjusted for Birth Year model
2nd quartile: 1.14 (0.86,1.52)
3rd quartile: 1.08 (0.82,1.44)
4th quartile: 1.11 (0.84,1.48)
Per IQR: 1.03 (0.94,1.14)
Exposure period: second trimester
Unadjusted model
2nd quartile: 0.98 (0.76,1.28)
3rd quartile: 1.09 (0.84,1.40)
4th quartile: 1.00 (0.77,1.28)
Per IQR: 1.00 (0.91,1.09)
Adjusted model
2nd quartile: 1.02 (0.77,1.34)
3rd quartile: 1.16 (0.89,1.51)
4th quartile: 1.09 (0.83,1.42)
Per IQR: 1.02 (0.93,1.12)
Adjusted for Birth Year model
2nd quartile: 0.99 (0.75,1.31)
3rd quartile: 1.10(0.84,1.45)
4th quartile: 1.01 (0.76,1.34)
Per IQR: 1.00 (0.90,1.10)
Exposure period: third trimester
Unadjusted model
2nd quartile: 0.93 (0.72,1.20)
3rd quartile: 1.07 (0.83,1.37)
4th quartile: 0.92 (0.71,1.18)
Per IQR: 0.95 (0.87, 1.05)
Adjusted model
2nd quartile: 0.96 (0.73,1.26)
3rd quartile: 1.14 (0.88,1.48)
4th quartile: 1.03 (0.79,1.35)
Per IQR: 0.99 (0.89, 1.09)
Adjusted for Birth Year model
2nd quartile: 0.92 (0.70,1.21)
3rd quartile: 1.04 (0.80,1.36)
4th quartile: 0.92 (0.69,1.22)
Per IQR: 0.94 (0.85, 1.05)
Reference: Gilboa, et al. (2005,
0878921
Period of Study: Jan 1996-Dec 2000
Location: Seven Counties in Texas,
USA: (Bexar, Dallas, El Paso, Harris,
Hidalgo, Tarrant, Travis)
Outcome: Birth defects
Age Groups: Newborn babies
Study Design: Case-control
N: 5,338 newborn babies
4574 controls
Statistical Analyses: Logistic
regression
Covariates: Alcohol consumption
during pregnancy, attendant of delivery
(i.e., the person who delivered the baby
(physician/nursemaid-wife vs. other)),
gravidity, marital status, maternal age,
maternal education, maternal illness,
maternal race/ethnicity, parity, place of
delivery, plurality, prenatal care, season
of conception, and tobacco use during
pregnancy
Control frequency matched to cases by
vital status, yr and maternal county of
Pollutant: PM10
Averaging Time: NR
Percentiles:26th:<19.5
50th(Median): 19.5-<23.8
76th: 23.8-<29.0
100th: > 29.0
Monitoring Stations: The
Environmental Protection Agency
provided raw data or hourly (for gases)
or daily (for PM) air pollution
concentrations for the seven study
counties
Copollutant: CO, N02,03, S02
PM Increment: calculated as quartiles
of avg concentration during wk3-8 of
pregnancy
Isolated Cardiac Defects
Aortic artery and valve defects:
25th: 0.40 (0.15, 1.03)
50th: 0.45 (0.18, 1.13)
75th: 0.68 (0.28, 1.65)
Atrial Sepal defects:
25th: 1.41 (0.86, 2.31)
50th: 2.13
75th: 2.27
1.34,3.37
1.43,3.60
Pulmonary artery and valve defects:
25th: 1.14
50th: 0.79
0.62,2.10
0.41, 1.55
75th: 0.68 (0.33, 1.40)
Ventricular Sepal defects:
25th: 0.83 (0.61, 1.11)
50th: 1.12 (0.85, 1.48)
75th: 0.98 (0.73, 1.32)
Multiple Cardiac Defects
Conotruncal defects:
25th: 1.13 0.79, 1.62
50th: 1.20 0.84,1.72
December 2009
E-444
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
residence
Season: Covariate in model
Dose-response Investigated? Yes
Statistical Package: SASv 8.2
75th: 1.26 (0.86, 1.84)
Endocardial cushion and mitral valve
defects:
25th: 0.82 (0.54, 1.25)
50th: 0.66
75th: 0.63
0.42, 1.05
0.38, 1.03
Isolated Oral Clefts
Cleft lip with or without palate:
25th: 1.29 (0.90, 1.85)
50th: 1.45 (1.01, 2.07)
75th: 1.37 (0.94,2.00)
Cleft palate:
25th: 0.99 (0.55, 1.78)
50th: 1.14 (0.64, 2.03)
75th: 1.11 (0.60, 2.06)
Individual Birth Defects
Aortic valve stenosis:
25th: 0.91 (0.53, 1.57)
50th: 0.86 (0.50, 1.50)
75th: 1.12 (0.63, 1.99)
Atrial Sepal defects:
25th: 1.10 (0.89, 1.35)
50th: 1.28
75th: 1.26
1.04, 1.57
1.03, 1.55
Coarctation of the aorta:
25th: 0.78
50th: 0.68
0.53, 1.15
0.45, 1.02
75th: 0.75 (0.48, 1.15)
Endocardial cushion defects:
25th: 0.87 (0.49, 1.55)
50th: 1.12 (0.64, 1.96)
75th: 0.89 (0.47, 1.65)
Ostium secundum:
25th: 1.15 (0.85, 1.55)
50th: 1.13
75th: 1.06
0.83, 1.53
0.77, 1.4
Pulmonary artery atresia without
ventricular Sepal defects:
25th: 1.93 (1.08, 3.45)
50th: 2.01 (1.11,3.64)
75th: 0.86 (0.41, 1.83)
Pulmonary valve stenosis:
25th: 1.16 (0.88, 1.55)
50th: 1.25
75th: 1.27
0.94, 1.66
0.94, 1.71
Tetralogy of Fallot:
25th: 1.21
50th: 1.40
0.72, 2.01
0.84, 2.33
75th: 1.450.85, 2.48)
Ventricular Sepal defects:
25th: 1.06 (0.90, 1.24)
50th: 1.10 (0.94, 1.29)
75th: 1.08 (0.92, 1.27)
December 2009
E-445
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Gouveia et al. (2004,
0556131
Period of Study: 1997
Location: Sao Paulo, Brazil
Outcome: Birth weight
Age Groups: Singleton full term live
births within 1000 g to 5500 g
Study Design: Cross sectional study
N: 179,460 live births
Statistical Analyses: GAM and
Logistic regression models
Covariates: Maternal age, length of
gestation, season, infant gender,
maternal education, number of
antenatal care visits, parity, and the
type of delivery
Season: All seasons
Dose-response Investigated? Yes
Statistical Package: S-Plus 2000
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 60.3 (25.2)
Range (Mm, Max): (25.5-153.0)
Monitoring Stations: maximum of 12
sites
Copollutant (correlation):
C0:r = 0.9
S02
N02
03
PM Increment: 10 pg/m
Mean [Lower Cl, Upper Cl]:
Changes in birth weight (in g)
First trimester = -13.7 (-27.0,-0.4)
Second trimester = -4.4 (-18.9,10.1)
Third trimester =14.6 (0.0, 29.2)
RR Estimate [Lower Cl, Upper Cl]:
(RR estimates are adjusted odds ratios
for low birth weight according to
quartiles of air pollution in each
trimester of pregnancy.)
Istquartile
First trimester = 1 (REF)
Second trimester = 1 (REF)
Third trimester = 1 (REF)
2nd quartile
First trimester =1.105 (0.994,1.229)
Second trimester = 1.003(0.904,1.113)
Third trimester =1.004 (0.914,1.104)
3rd quartile
First trimester =1.049 (0.903,1.219)
Second trimester = 1.074(0.920,1.254)
Third trimester =1.003 (0.861,1.169)
4th quartile
First trimester =1.144 (0.878,1.491)
Second trimester = 1.252 (1.028,1.525)
Third trimester = 0.970 (0.780,1.205)
Multiple linear regression coefficients
(SE) obtained from single, dual, and
three pollutant models
Single pollutant model = -1.37 (0.68)
Two pollutant (PM,0 and CO) = -0.51
(0.87)
Two pollutant (PM10 and S02) = -0.94
(0.75)
Three pollutant = -0.47 (0.88)
Reference: Ha et al. (2003, 0425521
Period of Study: Jan 1995-Dec 1999
Location: Seoul, South Korea
Outcome: Post-neonate total and
respiratory mortality
Age Groups: 1 month-1 yr
2 yr-65 yr, >65 yr
Study Design: Time-series
N: 1045 post-neonate deaths, 67,597 2-
65 yr old deaths, 100,316 >65yr old
deaths
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD): 69.2 (31. 6)
Percentiles: 26th: 44.8
SOth(Median): 64.2
75th: 87.7
Ranne/Min Mav\- 1 n -, I in/m ;
PM Increment: 42.9 pg/m3
RR Estimate [Lower Cl, Upper Cl]
lag:
Total Mortality:
1 month-1 yr (post-neonates):
1.142 [1.096, 1.190] lag 0
2 yr-65 yr:
1. 008 [1.006, 1.010] lag 0
>65yr (elderly):
1.023 [1.023, 1.024] lag 0
Statistical Analyses: Generalized
additive model
Covariates: Seasonality, temperature,
relative humidity, day of the week
Dose-response Investigated? No
Statistical Package: S Plus
Lags Considered: 0,1,2,3,4,5,6,7,
ma from 1-5 days
245.4 pg/m3
Monitoring Stations: 27
Copollutant (correlation):
N02:r = 0.73
S02:r = 0.62
03:r = -0.02
CO: r = 0.63
Respiratory Mortality:
1 month-1 yr (post-neonates):
2.018 [1.784, 2.283] lag 0
2 yr-65 yr:
1.066 [1.044, 1.090] lag 0
>65yr (elderly):
1.063 [1.055, 1.072] lag 0
December 2009
E-446
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Hansen, et al. (2006,
0898181
Period of Study: Jul 2000-Jun 2003
Location: Brisbane, Australia
Outcome: Pre-term birth (<37 wk)
Age Groups: Newborn babies
Study Design: Cross-sectional
N: 1583 live pre-terms births
28,200 singleton live births
Statistical Analyses: Multiple logistic
regression models
Covariates: Neonate gender, mother's
age, parity, indigenous status, number
of antenatal visits, marital status,
number of previous
abortions/miscarriages, type of delivery,
and index of SES
Season: all
Dose-response Investigated? Yes
Statistical Package: SAS version 8.2
Pollutant: PM,0
Averaging Time: recorded hourly, avi
daily
Mean (SD): 19.6 (9.4)
Range (Min, Max): 4.9, 171.7
Monitoring Stations: 5
Copollutant (correlation):
Fine PM or bsp, 0.1 to <2.5 pg in
diameter (0.58 to 0.76)
03 (0.54 to 0.83)
N02 (0.54 to 0.75)
PM10 (0.80 to 0.93)
Note: Correlations presented are for
the individual pollutant across
monitoring stations (not correlations
between PM10 and the pollutant.)
PM Increment: Trimester One
4.5 pg/m3
Last 90 days prior to birth
5.7 pg/m3
Odds Ratio [Lower Cl, Upper Cl]:
Trimester 1
1.15 [1.06,1.25]
Last 90 days prior to birth
1.04 [0.92,1.16]
Reference: Hansen et al. (2007,
0907031
Period of Study: Jul 2000-Jun 2003
Location: Brisbane, Australia
Outcome: Birth weight and Small for
Gestational Age (SGA
<10th percentile for age and gender)
Head circumference (HC) and crown-
heel length (CHL) among subsample
Study Design: Cross-sectional
N: 26,617 births (birth weight analysis)
and 21,432 (HC and CHL analyses)
Statistical Analyses: Logistic (SGA)
and linear (birth weight, HC, CHL)
regressions
Covariates: Gender, gestational age
(with a quadratic term), maternal age,
parity, number of previous
abortions/miscarriages, marital status,
indigenous status, number of antenatal
visits, type of delivery, an index of SES,
and season of birth
Season: Assessed as a covariate
Dose-response Investigated? Yes,
assessed exposures as quartiles
Statistical Package: SAS v8.2
Pollutant: PM10
Averaging Time: Trimester and
monthly avg were used in analyses
(calculated as the mean of daily values
Hourly data was use to calculate daily
means
City-wide avg used)
Mean (SD): 19.6 (9.4)
Percent! les:
25th: 14.6
50th: 18.1
75th: 22.7
Range (Min, Max): (4.9,171.7)
Monitoring Stations: 5
Copollutant (correlation):
By trimesters:
PM10T1:
PM10T2:r = 0.12
PM,oT3:r = -0.55
03T1:r = 0.77
03T2:r = 0.28
03T3:r = -0.61
N02T1:r = 0.32
N02T2:r = -0.65
N02T3:r = -0.17
visibility reducing particles (bsp)
T1:r = 0.82
visibility reducing particles (bsp)
T2:r = -.15
visibility reducing particles (bsp)
T3:r = -0.50
PM10T1:r = 0.12
PM10T2:
PM,0T3:r = 0.04
03T1:r = -0.11
03T2:r = 0.80
03T3:r = 0.18
N02T1:r = 0.77
N02T2:r = 0.25
N02T3:r = -0.72
visibility reducing particles (bsp)
T1:r = 0.23
visibility reducing particles (bsp)
T2:r = 0.80
visibility reducing particles (bsp)
T3:r = -0.24
PM10T1:r = -0.55
PM10T2:r = 0.04
PMi0T3:
PM Increment: IQR (8.1 fjg/nr)
Effect Estimate [Lower Cl, Upper Cl]:
Change (P) in mean birth weight (g)
associated with trimester-specific
exposures
Trimester 1:
Continuous exposure: -3.2 (-11.9, 5.5)
Quartiles of exposure:
1:Ref
2:-4.7 (-19.7, 10.2)
3: 4.2 (-12.9, 21.3)
4:-0.2 (-19.2, 18.8)
p-trend: 0.864
Trimester 2:
Continuous exposure: 0.4 (-9.4,10.2)
Quartiles of exposure:
1:Ref
2:12.7 (-2.3, 27.6)
3: 7.6 (-10.6, 25.7)
4:1.0 (-18.7, 20.7)
p-trend: 0.922
Trimester 3:
Continuous exposure: 3.6 (-6.9,14.0)
Quartiles of exposure:
1:Ref
2: 2.9 (-12.8, 18.7)
3:18.5(0.0,36.9)
4: 4.3 (-15.8, 24.4)
p-trend: 0.524
ORs for SGA associated with
trimester-specific exposures
Trimester 1:
Continuous exposure: 1.04(0.96,1.12)
Quartiles of exposure:
1:Ref
2:1.23(1.07, 1.42)
3:1.12(0.95,1.31)
4:1.12(0.94, 1.34)
p-trend: 0.361
Trimester 2:
Continuous exposure: 0.95 (0.88,1.04)
Quartiles of exposure:
1:Ref
2:0.96(0.83, 1.11)
3:1.06 0.89, 1.25)
4:0.98(0.81,1.18)
p-trend: 0.962
3: -0.02 (-0.08, 0.04)
4: -0.02 (-0.08, 0.05)
p-trend: 0.605
Trimester 2:
Continuous exposure:-0.01 (-0.04,
December 2009
E-447
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
03T1:r = -0.56
03T2:r = -0.18
03T3:r = 0.81
N02T1:r = -0.20
N02T2:r = 075
N02T3:r = 0.22
visibility reducing particles (bsp)
T1:r = -0.62
visibility reducing particles (bsp)
T2:r = 0.19
visibility reducing particles (bsp)
T3:r = 0.79
0.02)
Quartiles of exposure:
1:Ref
Trimester 3:
Continuous exposure: 0.93 (0.85,1.03)
Quartiles of exposure:
1:Ref
2:0.90(0.78, 1.04)
3: 0.81 (0.68, 0.96)
4:0.86(0.71,1.04)
p-trend: 0.098
Change (P) in mean head
circumference (HC
cm) associated with trimester-
specific exposures
Trimester 1:
Continuous exposure:-0.01 (-0.04,
0.02)
Quartiles of exposure:
1:Ref
2: -0.02 (-0.07, 0.04)
2: 0.03 (-0.02, 0.08)
3: 0.00 (-0.06, 0.06)
4: -0.01 (-0.08, 0.05)
p-trend: 0.538
Trimester 3:
Continuous exposure: 0.02 (-0.02, 0.05)
Quartiles of exposure:
1:Ref
2: 0.02 (-0.04, 0.07)
3:0.07(0.01,0.13)
4: 0.04 (-0.03, 0.11)
p-trend: 0.171
Change (P) in mean crown-heel
length (CHL
cm) associated with trimester-
specific exposures
Trimester 1:
Continuous exposure: 0.00 (-0.05, 0.05)
Quartiles of exposure:
1:Ref
2: 0.02 (-0.07, 0.11)
3: 0.01 (-0.10, 0.11)
4: 0.04 (-0.07, 0.16)
p-trend: 0.511
Trimester 2:
Continuous exposure: 0.07 (0.01, 0.13)
Quartiles of exposure:
1:Ref
2:0.10(0.01,0.18)
3:0.11(0.00,0.21)
4:0.13(0.01,0.24)
p-trend: 0.049
Trimester 3:
Continuous exposure:-0.01 (-0.07,
0.05)
Quartiles of exposure:
1:Ref
2:-0.02 (-0.11,0.07)
3: 0.10 (-0.01, 0.21)
4:-0.01 (-0.13, 0.10)
p-trend: 0.883
December 2009
E-448
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: (Hansen et al., 2009,
1923621
Period of Study: Jan 1997-Dec 2004
Location: Brisbane, Australia
Outcome: Birth defects- artery and
valve, atrial and ventricular Sepal,
conotruncal, endocardial cushion and
mitral valve, cleft lip and palate
Study Design: Case-control
Covariates: Mother's age, marital
status, indigenous status, previous
pregnancies, last menstrual period,
area-level socioeconomic status,
distance to a pollution monitor
Statistical Analysis: Conditional
logistic regression
Statistical Package: R
Age Groups: Neonates
Pollutant: PM,0
Averaging Time: daily
Mean (SD) Unit: 18.0 pg/m3
Range (Mm, Max): (4.4,151.7)
Copollutant (correlation): NR
Increment: 4pg/m
Odds Ratios (96% Cl) for risk of
defect
Aortic Artery and Valve Defects
All Births, Matched: 1.10 (0.76-1.56)
Births < 12km to Monitor:
1.83(1.16-2.98)
Births Ł 6km to Monitor:
1.43(0.73-2.90)
All Births, Unmatched: 1.09 (0.84-1.39)
Atrial Sepal Defects
All Births, Matched: 1.06 (0.86-1.30)
Births < 12km to Monitor:
1.07(0.84-1.37)
Births < 6km to Monitor:
0.88(0.60-1.27)
All Births, Unmatched: 1.14 (0.98-1.33)
Pulmonary Artery and Valve Defects
All Births, Matched: 0.90 (0.61-1.29)
Births < 12km to Monitor: 0.69
(0.43-1.08)
Births < 6km to Monitor:
1.46(0.76-2.73)
All Births, Unmatched: 0.99
(0.78-1.24)
Ventricular Sepal Defects
All Births, Matched: 0.87 (0.73-1.04)
Births < 12km to Monitor:
0.85(0.69-1.03)
Births < 6km to Monitor:
0.90(0.68-1.18)
All Births, Unmatched: 1.15(1.02-1.30)
Conotruncal Defects
All Births, Matched: 0.80 (0.54-1.19)
Births < 12km to Monitor: 0.94 (0.55-
1.49)
Births < 6km to Monitor: 0.66 (0.27-
1.45)
All Births, Unmatched: 0.97 (0.74-1.24)
Endocardial Cushion and Mitral Valve
Defects
All Births, Matched: 1.29 (0.82-2.04)
Births < 12km to Monitor:
1.28(0.75-2.19)
Births < 6km to Monitor:
0.90(0.44-1.86)
All Births, Unmatched: 0.94 (0.68-1.26)
Cleft Lip
All Births, Matched: 1.05 (0.72-1.51)
Births < 12km to Monitor:
1.16(0.72-1.82)
Births < 6km to Monitor:
1.03(0.56-1.82)
All Births, Unmatched: 1.01 (0.79-1.27)
Cleft Palate
All Births, Matched: 0.69 (0.50-0.93)
Births < 12km to Monitor:
0.53 (0.29-0.87)
Births Ł 6km to Monitor:
0.71 (0.49-1.00)
All Births, Unmatched: 0.89 (0.72-1.10)
Cleft Lip with or without Cleft Palate
All Births, Matched: 1.05 (0.84-1.30)
Births < 12km to Monitor:
1.03(0.79-1.34)
Births Ł 6km to Monitor:
0.83(0.58-1.19)
All Births, Unmatched: 1.04 (0.89-1.21)\
December 2009
E-449
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Jalaludin et al. (2007,
1566011
Period of Study: 1998-2000
Location: Sydney, Australia
Outcome: Gestational age
(categorized: preterm birth: <37wk
term birth: > 37 wk but <42 wk)
Age Groups: Infants
Study Design: Cross-sectional
N: 123,840 singleton births of >20 wk
gestation
Statistical Analyses: Logistic
regression
Covariates: Sex of child, maternal age,
maternal smoking during pregnancy,
gestational age at first antenatal visit,
whether mother identifies as being
Aboriginal or Torres Strait Islander,
whether first pregnancy, season of
conception, SES, (temperature and
relative humidity were not significant in
single variable models and therefore,
were not included)
Season: Examined as covariate and
effect modifier
Dose-response Investigated? No
Statistical Package: SASvS
Pollutant: PM,0
Averaging Time: 24 h avg used to
calculate the mean concentration over
the first trimester, the 3 mo preceding
birth, the first month after the estimated
date of conception, and the month prior
to delivery
Mean (SD): (24 h avg)
All yr: 16.3 (6.38)
Summer: 18.2 (7.20)
Fall: 17.0 (6.23)
Winter: 14.5 (5.57)
Spring: 15.7 (5.82)
Monitoring Stations: 14 stations within
the Sydney metropolitan area (levels
avgd to provide 1 estimate for the entire
study area)
Copollutant (correlation):
PM10
PM25(r = 0.83)
CO (r = 0.28)
N02(r = 0.48)
03(r = 0.50)
S02(r = 0.42)
Notes: Correlations between
monitoring stations measuring PM10
ranged from 0.67 to 0.91
PM Increment: 1 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
ORs (air pollutant concentration during
the 1st trimester and preterm birth by
season)
Fall: 1.462 (1.267, 1.688)
Winter: 1.343 (1.190,1.516)
Spring: 1.119 (0.973,1.288)
Summer: 0.913 (0.889, 0.937)
ORs (air pollutant concentrations during
different exposure periods and preterm
birth
for all of Sydney and among only those
residing within 5 km of a monitoring
station)
1 month preceding birth
Sydney: 0.991 (0.979,1.003)
5km: 1.008 (0.993, 1.022)
3 mo preceding birth
Sydney: 0.989 (0.975,1.004)
5km: 1.012 (0.995,1.030)
1st month of gestation
Sydney: 0.983 (0.973, 0.993)
5km: 0.957 (0.914, 1.002)
1st trimester
Sydney: 0.987 (0.973,1.001)
5km: 1.009 (0.978,1.041)
Notes: Authors note that effect of PM10
on preterm birth for infants conceived
during the fall did not remain in 2
pollutant models (ORs between 0.77
and 1.04)
Reference: Kaiser et al. (2004,
0766741
Period of Study: 1995-1997
Location: 25 U.S. counties (23
metropolitan areas): Jackson, AL
Fresno, CA
Los Angeles, CA
Sacramento, CA
San Diego, CA
San Francisco, CA
Denver, CO
Hartford, CT
Cook, IL
Baltimore, MD
Wayne, Ml
St. Louis, MO
Rrrinv MV
DlONX, IN I
l/lnne MV
r\ings, NY
New York, NY
Philadelphia, PA
El Paso, TX
Harris TX
Dallac TY
Udlldo, I A
Oklahoma, OK
Tulsa, OK
Providence, Rl
Salt Lake City, UT
King, WA
Milwaukee, Wl
Outcome: Postneonatal death:
All cause, SIDS (798.0)
Respiratory disease (460-519)
Age Groups: Infants between 1-12 mo
Study Design: Attributable risk
assessment
N: 700,000 infants (# deaths NR)
Statistical Analyses: Risk assessment
methods described in: Kunzli et al.
Public-health impact of outdoor and
traffic-related air pollution: a European
assessment. Lancet 2000, 356: 795-
Qf"M
OU1.
Covariates: Maternal education,
maternal ethnicity, parental marital
status, maternal smoking during
pregnancy, infant's month and yr of
birth, avg temperature in the first 2 mo
of life
Season: All
adjusted for month/yr of birth
Dose-response Investigated? NR
Statistical Package: NR
Lags Considered: Annual, county-level
mean
Pollutant: PM,0
Averaging Time: "annual mean levels"
in each county
Mean (SD): 28.4
Range (Min, Max):
County range: 18.0,44.8
Monitoring Stations: NR
Notes: 14 out of 25 counties had PMi0
levels >25 pg/m3
PM Increment: Analysis 1 :
16.4 pg/m3 (difference between
reference level of 12 pg/m3 and
observed mean level of 28.4 pg/m )
Analysis 2:
13 pg/m3 (difference between reference
level of 12 pg/m3 and 25 pg/m3)
AR Estimate [Lower Cl, Upper Cl]:
Analysis 1 :
All cause 6% [3, 11]
SIDS 16% [9, 23]
Respiratory 24% [7, 44]
Attributable* deaths per 100,000
All cause 14.7 [7.3, 25.6]
SIDS 11. 7 [6.8, 16.6]
Respiratory 2.3 [0.7, 4.1]
Analysis 2:
All cause 5% [2, 8]
SIDS 12% [7, 18]
Respiratory 19% [6, 34]
Attributable* deaths per 100,000
All cause 10.9 [5.5, 19.1]
SIDS 9.0 [5.3, 12.8]
Respiratory 1.8 [0.5, 3.2]
Notes: -Authors did not extrapolate
attributable cases below 12 pg/m3 (i.e.,
reference level was set at 12 pg/m )
-Attrihiit^hla rielxe •am K^eaH An tha DDe
reported by Woodruff et al, 1997 for a
10 pg/m increase:
All cause 1.04 [1.02-1.07]
SIDS 1.12 [1.07, 1.17]
Respiratory 1.20 [1.06,1.36]
December 2009
E-450
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: (Kim et al, 2007,1566421
Period of Study: May 2001-May 2004
Location: Seoul, Korea
Outcome (ICD9 and ICD10): LBW (low
birth weight, less than 2500 g at later
than gestational week 37), premature
delivery (birth before the completion of
the 37th week), stillbirth (intrauterine
fetal death), IUGR (birth weight lower
than the 10th percentile for the given
gestational age), and congenital
anomaly (a defect in the infant's body
structure)
Age Groups: Infants
Study Design: Cross-sectional (women
visiting the clinic for prenatal care were
recruited with follow-up until discharge
after delivery)
N: 1514 observations (births)
Statistical Analyses: Multiple logistic
and linear regression (in addition, for
birth weight, used generalized additive
model to account for long-term trends
and nonlinear relationships between the
response variable and the predictors,
and to produce smoothed plots of the
relationship between PM and birth
weight)
Covariates: Adjustment 1: infant sex,
infant order, maternal age and
education, paternal education, season
of birth
Adjustment 2: adjustment 1 factors plus
alcohol, maternal BMI, maternal weight
prior to delivery
(collected information on smoking, ETS,
parity, past history of illnesses, history
of illnesses during pregnancy but did
not use in analyses due to small
numbers or non-significance)
Season: Adjusted for season of
delivery
Dose-response Investigated? Yes
Statistical Package: SAS 8.01, S-Plus
2000
Pollutant: PM,0
Averaging Time: Used hourly
exposure levels to calculate avg
exposure levels at each trimester, each
month of pregnancy, and 6 wk before
delivery from the nearest monitoring
station (based on home address of
mother)
Also created categories within each
pregnancy period (<25th percentile
[referent], 25th to 50th percentile, and
>50th percentile)
Mean (SD): Range of PM means
across pregnancy periods: 88.7-89.7
Monitoring Stations: 27 stations
PM increment: 10 pg/m
Preterm:
1st Trimester Odds Ratios:
Crude: 0.95 (0.90,1.01)
Adj 1:0.93 (0.87, 1.00)
Adj 2: 0.93 (0.85, 1.01)
2nd Trimester Odds Ratios:
Crude: 0.99 (0.94,1.06)
Adj 1:0.98 (0.92, 1.04)
Adj 2:1.00 (0.93, 1.07)
3rd Trimester Odds Ratios:
Crude: 1.02 (0.98,1.06)
Adj 1:1.05 (1.00, 1.10)
Adj 2:1.05 (0.99, 1.11)
LBW:
1st Trimester Odds Ratios:
Crude: 1.02 (0.93,1.12)
Adj 1:1.03 (0.93, 1.14)
Adj 2:1.07 (0.96, 1.19)
2nd Trimester Odds Ratios:
Crude: 1.03 (0.94,1.14)
Adj 1:1.04 (0.93, 1.17)
Adj 2:1.07 (0.94, 1.22)
3rd Trimester Odds Ratios:
Crude: 1.04 (0.97,1.11)
Adj 1:1.05 (0.97, 1.14)
Adj 2:1.05 (0.96, 1.16)
IUGR:
1st Trimester Odds Ratios:
Crude: 1.07 (0.97,1.19)
Adj 1:1.07 (0.95, 1.21)
Adj 2:1.14 (0.99, 1.31)
2nd Trimester Odds Ratios:
Crude: 0.97 (0.85,1.12)
Adj 1:0.97 (0.82, 1.13)
Adj 2: 0.93 (0.77, 1.13)
3rd Trimester Odds Ratios:
Crude: 0.82 (0.68, 0.99)
Adj 1:0.88 (0.72, 1.08)
Adj 2: 0.85 (0.67, 1.08)
Birth defect:
1st Trimester Odds Ratios:
Crude: 1.08 (0.98,1.20)
Adj 1:1.12 (1.00, 1.25)
Adj 2:1.08 (0.95, 1.22)
2nd Trimester Odds Ratios:
Crude: 1.09 (0.99,1.21)
Adj 1:1.11 (0.98, 1.26)
Adj 2:1.16 (1.00, 1.34)
3rd Trimester Odds Ratios:
Crude: 1.00 (0.90,1.11)
Adj 1:0.97 (0.86, 1.08)
Adj 2: 0.97 (0.87, 1.10)
Stillbirth:
1st Trimester Odds Ratios:
Crude: 0.83 (0.76, 0.90)
Adj 1:0.93 (0.85, 1.02)
Adj 2: 0.95 (0.85, 1.02)
2nd Trimester Odds Ratios:
Crude: 0.99 (0.93,1.05)
Adj 1:1.03 (0.95, 1.11)
Adj 2:1.07 (0.98, 1.17)
3rd Trimester Odds Ratios:
Crude: 1.14 (1.10,1.18)
Adj 1:1.09 (1.04, 1.15)
Adj 2:1.08 (1.02, 1.14)
LBW (categorical PM exposure):
1st Trimester ORs:
<25th:1.0
25th-50th:0.5(0.1,3.2)
>50th:1.0(0.3, 3.8)
3rd Trimester ORs:
<25th:1.0
25th-50th:1.3(0.2, 10.4)
>50th: 3.0 (0.5, 18.5)
6 wk before birth ORs:
<25th:1.0
December 2009
E-451
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
25th-50th: 3.2 (0.3, 33.7)
>50th: 5.2 (0.6, 47.6)
Changes in Birth Weight (96%CI) per
10 ug/m3 increase in PM
concentration:
1st trimester: 7.8 (1.2,14.5)
2nd trimester:-0.3 (-7.3, 6.8)
3rd trimester:-2.1 (-7.5,3.4)
1st month: 4.4 (-1.0, 9.8)
2nd month: 6.4 (0.6,12.2)
3rd month: 4.3 (-1.5, 10.2)
4th month: 3.0 (-3.7, 9.6)
5th month:-3.9 (-10.5, 2.7)
6th month: 0.1
(-5.7, 5.8)
7th month: 0.1 (-5.1,5.3)
8th month: 0.0
9th month: 1.8
-4.5, 4.5)
-2.3, 5.9)
Last 6 wk: -4.8 (-9.9, 0.4)
Reference: Lee et al. (2003, 0432021
Period of Study: Jan 1996-Dec 1998
Location: Seoul, South Korea
Outcome: Low birth weight (LBW),
<2500 g
Age Groups: Child-bearing age women
and their newborn children-delivered at
37-44 gestational wk
Study Design: Cross-section
N: 388,905 full-term single births
Statistical Analyses: Generalized
additive model, LOESS, Akaike's
criterion,
Covariates: Infant sex, birth order,
maternal age, parental education level,
time trend and gestational age.
Season: All
Dose-response Investigated? Yes
Statistical Package: NR
Pollutant: PM10
Averaging Time: Arithmetic avg of
hourly measurements at 20 stations
Mean (SD): 71.1(30.1)
Percent! les:
25th: 47.4
60th(Median): 67.6
76th: 89.3
Range (Mm, Max): 18.4, 236.9
Monitoring Stations: 20
Copollutant (correlation):
1st trimester:
PM10-CO: 0.47
PMio-S02: 0.78
PM10-N02: 0.66
2nd trimester:
PMio-CO: 0.68
PM10-S02: 0.82
PM10-N02: 0.81
3rd trimester:
PM10-CO: 0.69
PM10-S02: 0.85
PMio-N02: 0.80
PM Increment: IQR, 41.9
RR Estimate [Lower Cl, Upper Cl]:
1st trimester: 1.03 [1.00,1.07]
2nd trimester: 1.04 [1.00,1.08]
3rd trimester: 1.00 [0.95,1.04]
All trimesters: 1.06 [1.01,1.10]
Low exposure in last 5 mo using IQR
during last 5 mo: 0.94 [0.85,1.05]
Low exposure in first 5 mo using IQR
during first 5 mo: 1.04 [1.01,1.08]
Notes: Birth weight was decreased by
19.6 g for an IQR increase in the 2nd
trimester.
The OR for LBW increased for female
children, fourth or higher order child,
mother <20 yr of age, and low parental
education level.
December 2009
E-452
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Leem et al. (2006, 0898281
Period of Study: 2001-2002
Location: Incheon, Korea
Outcome (ICD9 and ICD10): Age
Groups: Pre-term delivery
Study Design: Cross-sectional
N:Cases: 2,082
Controls: 50,031
Statistical Analyses: Log-binomial
regression (corrected for overdispersion
Used the log link function)
Covariates: Maternal age, parity, sex,
season of birth, and education level of
each parent
Season: Controlled as a covariate
Dose-response Investigated? Yes,
assessed quartiles of exposure
Statistical Package: NR
Pollutant: PM,0
Averaging Time: Trimesters (daily
hourly data used to calculate)
Range (Min, Max): Reported ranges
within quartiles by trimester:
1st Trimester:
4:64.57-106.39
3: 53.84-64.56
2: 45.95-53.83
1:26.99-45.94
3rd Trimester:
4: 65.63-95.91
3: 56.07-65.62
2: 47.07-56.06
1:33.12-47.06
Monitoring Stations: 27 monitoring
stations
Pollutant levels for each area were
predicted from the levels recorded at
the monitors using ordinary block
kriging
Copollutant (correlation):
S02(r = 0.13)
N02(r = 0.37)
CO (r = 0.27)
Effect Estimate [Lower Cl, Upper Cl]:
Crude and Adjusted RRfor preterm
delivery and exposure during the 1st
trimester
Crude
Quartiles of exposure:
4:1.07(0.95, 1.21)
3:1.02 0.90, 1.15)
2:1.06(0.94,1.20)
1:1.00
Adjusted
Quartiles of exposure:
4:1.27(1.04,1.56)
3:1.13 0.94, 1.37)
2:1.14(0.97,1.34)
1:1.00
p-trend: 0.39
Crude and Adjusted RRfor preterm
delivery and exposure during the 3rd
trimester
Crude
Quartiles of exposure:
4:1.06(0.94,1.20)
3:1.06(0.94, 1.19)
2:1.05 0.93, 1.18)
1:1.00
Adjusted
Quartiles of exposure:
4:1.09(0.91,1.30)
3:1.04(0.90, 1.21)
2:1.05 0.91, 1.20)
1:1.00
p-trend: 0.33
Reference: Lin et al. (2004, 0957871
Period of Study: Jan 1998-Dec 2000
Location: Sao Paulo, Brazil
Outcome: Neonatal death
Age Groups: Neonates (infants 0-28
days after birth)
Study Design: Time series
N: 1096 days, 6697 deaths
Statistical Analyses: Poisson
regression (GAM)
Covariates: Non-parametric LOESS
smoothers to control for: time (long term
trend), temperature, humidity, and day
of week
Also controlled for holidays with linear
term
Season: All
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: Lag 0, "ma from 2 to
7 days"
Notes: No explicit control for season
apart from temperature
Pollutant: PM,0
Averaging Time: Daily values
Mean (SD): 48.62 (21.18)
Range (Min, Max): 13.9,157.3
Monitoring Stations: NR (indicated
more than 1)
Copollutant (correlation):
CO r = 0.71
N02r = 0.76
S02r = 0.80
03r = 0.36
PM Increment: 1 pg/m
Log relative rate (standard error) lag
Single pollutant model
0.0017 (0.0008) lag 0
This translates to a 4.0% [95% Cl: 0.3,
7.9] increase in neonatal mortality for a
23.3 pg/m3 increase in PM,0
Two-pollutant model
0.0000 (0.0011) lag 0
Notes: -In two-pollutant model with
PM10 and S02 (which are highly
correlated), effect of PM disappeared
and effect of S02 remained constant
- Results from pollutant ma from
2-7 days not reported, authors indicate
effects only found for lag 0 (same day
levels)
- Confidence intervals reported in
abstract are incompatible with
ps/standard errors and plotted results in
text: abstract indicates a 4% increase in
mortality with 95% Cl: 2-6 for a
23.3 pg/m3 increase in PM10
December 2009
E-453
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: (Lin et al, 2004, 0898271
Period of Study: 1995-1997
Location: Taipei and Kaoshiung,
Taiwan
Outcome: Low birth weight (<2500
grams)
Age Groups: Newborns
Study Design: Cross-sectional
N: 92,288 infants
Statistical Analyses: Logistic
regression
Covariates: Gender, birth order,
gestational weeks, season of birth,
maternal age, maternal education,
copollutants
Season: All
Dose-response Investigated? Yes
Statistical Package: NR
Lags Considered: The 9-month
pregnancy period for each infant, and
each trimester
Pollutant: PM,0
Averaging Time: NR, "daily
measurements"
Mean (SD): Reported by monitoring
station: Taipei:
1.48.78
2. 46.29
3. 48.79
4. 50.80
5. 52.54
Kaohsiung
1.69.99
2. 63.39
3. 64.89
4. 75.79
5. 77.27
Monitoring Stations:
10 (5 in each city)
Notes: All pregnant women/infants
included in study lived within 3 km of an
air quality monitoring station
Pollution assigned based on nearest air
quality station to the maternal residence
Co-pollutant: CO, S02, 03, N02
PM Increment: Tertiles
Entire pregnancy
T1:<46.4ppb
T2: 46.4-63.1 ppb
T3: >63.1 ppb
First trimester
T1:<45.8ppb
T2: 45.8-67.6 ppb
T3: >67.6 ppb
Second trimester
T1:<44.6ppb
T2: 44.6-64.2 ppb
T3: >64.2 ppb
Third trimester
T1:<43.7ppb
T2: 43.7-63.7 ppb
T3: >63.7 ppb
RR Estimate [Lower Cl, Upper Cl]
Entire pregnancy
T1:1.00
T2: 0.96 [0.83, 1.11]
T3: 0.87 [0.71, 1.05]
First trimester
T1:1.00
T2: 0.96 [0.84, 1.09]
T3: 0.97 [0.80, 1.17]
Second trimester
T1:1.00
T2:1.03 [0.90, 1.17]
T3:1.00 [0.83, 1.21]
Third trimester
T1:1.00
T2:1.02 [0.90, 1.16]
T3: 0.97 [0.81, 1.17]
Notes: RR for births in Kaoshiung vs.
Taipei: 1.13 [1.03,1.24]
Reference: Lipfert et al. (2000, 0041031
Period of Study: 1990
Location: U.S.
Outcome: Infant mortality
Including respiratory mortality
(traditional definition, ICD9 460-519),
expanded definition (adds ICD9 769
and 770)
Age Groups: Infants
Study Design: Cross-sectional
N: 2,413,762 infants in 180 counties
(Ns differ for various models)
Statistical Analyses: Logistic
regression
Covariates: Mother's smoking,
education, marital status, and race
Month of birth
And county avg heating degree days
Dose-response Investigated? NR
Statistical Package: NR
Pollutant: PM,0
Averaging Time: Yearly avg used
Mean (SD): 33.1 (9.17) (based on 180
counties)
Range (Min, Max): (16.9, 59)
Monitoring Stations: NR
Copollutant (correlation):
PM10
S042"(r = 0.10)
NSPMio-non-sulfate portion of PMi0
(r = 0.91)
CO (r = 0.27)
S02(r = 0.04)
Notes: TSP-based sulfate was adjusted
for compatibility with the PMio-based
data
PM Increment: NR (present regression
coefficients)
Effect Estimate [Lower Cl, Upper Cl]:
Presented regression coefficients
standard errors)
3 PM exposures regressed jointly)
bold = p <0.05
Cause of death: All
Birth weight: All
PM,,: 0.0114 (0.0015)
S04:-0.0002 (0.0061)
NSPM10: 0.0115 (0.0014)
Cause of death: All
Birth weight: LBW
PM,,: 0.0088 (0.0019)
S04 : 0.0265 (0.0080)
NSPM,0: 0.0086 (0.0020)
Cause of death: All
Birth weight: normal
PM,n: 0.0092 (0.0024)
S04 : -0.0488 (0.0098)
NSPM10: 0.0096 (0.0024)
Cause of death: All neonatal
Birth weight: All
PM,,: 0.0126 (0.0018)
S04 :0.0267 (0.0076)
NSPM10: 0.0126 (0.0018)
Cause of death: All neonatal
Birth weight: LBW
PM,,: 0.0086 (0.0022)
S04 :0.0388 (0.0088)
NSPM,0: 0.0093 (0.0022)
Cause of death: All neonatal
Birth wt: normal
PM,n: 0.0123 (0.0041)
SO/':-0.0334 (0.0169)
NSPM10: 0.0125 (0.0040)
Cause of death: All post neonatal
Birth wt: All
December 2009
E-454
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
PM,,: 0.0091 (0.0024)
S04:-0.0474 (0.0100)
NSPM10: 0.0096 (0.0024)
Cause of death: All post neonatal
Birth wt: LBW
PM1fl: 0.0096 (0.0043)
S04:-0.0247 (0.0173)
NSPM,o: 0.0101 (0.0042)
Cause of death: All post neonatal
Birth wt: normal
PM,n: 0.0074 (0.0030)
S04 : -0.0569 (0.0121)
NSPM10: 0.0080 (0.0029)
Cause of death: SIDS
Birth weight: All
PM,n: 0.0138 (0.0038)
S04:-0.1078 (0.0151)
NSPM10: 0.0149 (0.0037)
Cause of death: SIDS
Birth weight: LBW
PIVU 0.0115 (0.0088)
S04:-0.1378 (0.0337)
NSPM,0: 0.0146 (0.0085)
Cause of death: SIDS
Birth weight: normal
PM,n: 0.0137 (0.0042)
S04:-0.0995 (0.0168)
NSPM10: 0.0147 (0.0041)
Cause of death: All respiratory (ICD9:
460-519, 769, 770)
Birth weight: All
PM,n: 0.0168 (0.0034)
S04: 0.0706 (0.0146)
NSPM10: 0.0166 (0.0034)
Cause of death: All respiratory (ICD9:
460-519, 769, 770)
Birth weight: LBW
PM,n: 0.0144 (0.0038)
S04: 0.0821 (0.0158)
NSPM10: 0.0139 (0.0038)
Cause of death: All respiratory (ICD9:
460-519, 769, 770)
Birth weight: normal
PM,n: 0.0177 (0.0091)
S04 :0.0001 (0.0392)
NSPM10: 0.0118 (0.0090)
Cause of death: Respiratory disease
(ICD9: 460-519)
Birth weight: All
PM,n: 0.0133 (0.0089)
S04 :0.0093 (0.0384)
NSPM10: 0.0134 (0.0089)
Cause of death: Respiratory disease
(ICD9: 460-519)
Birth weight: LBW
PM,n: 0.0092 (0.0137)
S04 :0.0434 (0.0580)
NSPM10: 0.0089 (0.0138)
Cause of death: Respiratory disease
(ICD9: 460-519)
Birth weight: normal
PM,n: 0.0126 (0.0120)
S04:-0.0177 (0.0509)
NSPMy 0.0128 (0.0119)
Associations with SIDS by smoking
status
Smoking status: Yes
Birth weight: Normal
PM1fl: 0.0202 (0.0073)
S04 : -0.0722 (0.0284)
NSPM,0: 0.0206 (0.0071)
Smoking status: No
Birth weight: Normal
PMin:0.0104
0.0051)
0.021)
SCV':-0.114
NSPM10: 0.0117 (0.005)
Smoking status: Yes
Birth weight: LBW
December 2009 E-455
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
PM,,: 0.0322 (0.0130)
S04 : -0.0958 (0.0483)
NSPM10: 0.0345 (0.0125)
Smoking status: No
Birth weight: LBW
PM,,:-0.0044 (0.012)
S04 : -0.0172 (0.047)
NSPM,0: -0.0007 (0.012)
Mean risks (95%CI) between post
neonatal SIDS among normal birth
weight babies
pollutants regressed one at a time
PM,,: 1.20 (1.02, 1.42)
S04 : 0.43 (0.37, 0.51)
NSPM10:1.33 (1.18, 1.50)
Reference: Maisonet et al. (2001,
0166241
Period of Study: 1994-1996
Location: Northeastern U.S. (6 cities:
Boston, Hartford, Philadelphia,
Pittsburgh, Springfield, Washington DC)
Outcome: Low birth weight (LBW):
infants with a birth weight <2,500 g and
having a gestational age between 37
and 44 wk
Age Groups: Term live births
(singleton)
Study Design: Cross-sectional
N: 89,557 infants
Statistical Analyses: Logistic
regression (LBW) and linear regression
(for reductions in birth weight)
Covariates: Gestational age, gender,
birth order, maternal age, race/ethnicity,
yr of education, marital status,
adequacy of prenatal care, previous
induced or spontaneous abortions,
weight gain during pregnancy, maternal
prenatal smoking, and alcohol
consumption
Season
Season: Yes, as covariate
Dose-response Investigated? Yes,
categorical exposure variables
assessed
Statistical Package: STATA
Pollutant: PM,0
Averaging Time: Trimester avg
calculated using 24-h measurements
taken every 6 days
Range (Min, Max): Ranges for
categories of exposure:
1st Trimester
<25th: <24.821
25 to <50th: 24.821, 30.996
50 to <75th: 30.997, 36.142
75 to <95th: 36.143, 46.547
> 95th :> 46. 548
2nd Trimester
<25th: <24.702
25 to <50th: 24.702, 30.294
50 to <75th: 30.295, 35.410
75 to <95th: 35.411, 43.928
> 95th :> 43. 929
3rd Trimester
<25th: <24.702
25 to <50th: 24.702, 30.162
50 to <75th: 30.163, 35.642
75 to <95th: 35.643, 43.588
> 95th :> 43. 589
Monitoring Stations: 3-4 per city
Copollutants: CO, S02
PM Increment: 10 pg/m3 for analyses
assessing exposures continuously
Effect Estimate [Lower Cl, Upper Cl]:
ORs for term LBW by trimester
1st Trimester Crude
<25th:1.00
25 to <50th: 1.02 (0.90, 1.14)
50to<75th:0.90 0.65, 1.24
75to<95th:0.87 0.58,1.30
> 95th: 0.89 (0.60 1.33)
Continuous: 0.93 (0.77, 1.13)
1st Trimester Adjusted
<25th:1.00
25to<50th' 1 02 094 1 11
50to<75th'090 073 1 03
75 to <95th: 0.85 (0.73, 1.00
> 95th: 0.83 (0.70, 0.97)
Continuous: 0.93 (0.85, 1.00)
2nd Trimester Crude
<25th' 1 00
25to<50th:1.01 (0.93, 1.10)
50 to <75th: 0.90 (0.66, 1.21)
75to<95th:0.92(0.62, 1.34)
> 95th: 0.90 (0.61 1.33)
Continuous: 0.95 (0.78, 1.16)
2nd Trimester Adjusted
<25th:1.00
25 to <50th: 1.06 (0.97, 1.15)
50to<75th:0.95 0.85, 1.07
75to<95th'091 079 1 05
> 95th: 0.77 (0.63 0.95)
Continuous: 0.93 (0.85, 1.02)
3rd Trimester Crude
<25th:1.00
25to<50th:0.94 0.85, 1.05
50to<75th:0.86 0.58,1.25
75 to <95th: 0.86 (0.57, 1.29)
> 95th: 0.92 (0.61, 1.38)
Continuous: 0.95 (0.75,1.20)
3rd Trimester Adjusted
<25th:1.00
25 to <50th: 0.98 (0.87, 1.10)
50 to <75th: 0.92 (0.76, 1.11)
75to<95th:0.88(0.75, 1.04)
> 95th: 0.91 (0.77, 1.07)
Continuous: 0.96 (0.88,1.06)
Adjusted ORs by race/ethnicity
Whites:
1st Trimester
<25th:1.00
25 to <50th: 1.13 (0.96, 1.33)
50 to <75th: 1.00 (0.92, 1.08)
75to<95th:1.00 0.91, 1.09)
> 95th: 0.92 (0.81, 1.04)
Continuous: 0.94 (0.90, 0.98)
2nd Trimester
<25th:1.00
25 to <50th: 0.88 (0.77, 1.02)
50to<75th:0.95 0.89, 1.02
75to<95th:0.95 0.84,1.07
> 95th: 0.89 (0.64, 1.26)
Continuous: 0.96 (0.89,1.04)
December 2009
E-456
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
3rd Trimester
<25th:1.00
25 to <50th: 0.84 (0.64, 1.11)
50 to <75th: 0.91 (0.83, 1.01)
75 to <95th: 0.80 0.71,0.90)
> 95th: 1.03 (0.86, 1.24)
Continuous: 0.95 (0.90,1.00)
African Americans:
1st Trimester
<25th:1.00
25to<50th:1.01 0.98, 1.05
50to<75th:0.88 0.79,0.98
75 to <95th: 0.83 (0.70, 0.97)
> 95th: 0.81 (0.67,0.99)
Continuous: 0.93 (0.85,1.01)
2nd Trimester
<25th:1.00
25 to <50th: 1.10(0.93, 1.30)
50 to <75th: 0.95 (0.80, 1.12)
75to<95th:0.88(0.69, 1.11)
> 95th: 0.75 (0.54, 1.03)
Continuous: 0.92 (0.80,1.05)
3rd Trimester
<25th:1.00
25 to <50th: 1.08 (0.92, 1.27)
50to<75th:0.89 0.70, 1.12
75to<95th:0.94 0.75,1.18
> 95th: 0.83 (0.71, 0.97)
Continuous: 0.99 (0.87,1.11)
Hispanics:
1st Trimester
<25th:1.00
25 to <50th: 0.83 (0.64, 1.06)
50 to <75th: 0.86 (0.70, 1.05)
75 to <95th: 0.79 (0.68, 0.93)
> 95th: 1.36 (1.06, 1.75)
Continuous: 0.96 (0.84,1.09)
2nd Trimester
<25th:1.00
25 to <50th: 1.16 (0.84, 1.61)
50to<75th:0.86 0.63, 1.19
75to<95th:0.98 0.71,1.34
> 95th: 0.68 (0.38, 1.21)
Continuous: 0.92 (0.81,1.05)
3rd Trimester
<25th:1.00
25to<50th:0.77(0.55, 1.07)
50to<75th:1.12 0.76, 1.66
75to<95th:0.93 0.65,1.31
> 95th: 0.90 (0.55, 1.47)
Continuous: 0.96 (0.80,1.15)
December 2009 E-457
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Mannes et al.(2005,
0878951
Period of Study: Jan 1998-Dec 2000
Location: Metropolitan Sydney,
Australia
Outcome: Risk of SGA and birth weight Pollutant: PM10
Averaging Time: 24 h
Mean (SD): 16.8 (7.1)
25th: 12.3
60th(Median): 15.7
76th: 19.9
Range (Min, Max): (3.8-104.0)
Monitoring Stations: up to 14
Age Groups: All singleton births >20
wk and 2 400 grams birth weight and
maternal all ages
Study Design: Cross-sectional
N: 138,056 singleton births
Statistical Analyses: Logistic and
linear regression models
Covariates: Sex of child, maternal age,
gestational age, maternal smoking,
gestational age at first antenatal visit,
maternal indigenous status, whether
first pregnancy, season of birth,
socioeconomic status
Season: All seasons
Included as covariate
Dose-response Investigated? No
Statistical Package: SASv8.02
Copollutants (correlations):
CO: r = 0.26
N02:r = 0.47
03:r = 0.52
PM25:r = 0.81
PM Increment: 1 pg/m
Risk of SGA
All births
One month before birth:
OR = 1.01 (1.00-1.03)
Third trimester: OR = 1.00 (0.99-1.013)
Second trimester:
OR = 1.01 (1.00-1.04)
First trimester: OR = 1.00 (0.98-1.02)
5 km births
One month before birth: OR = 1.00
(0.99-1.02)
Third trimester: OR = 1.01 (0.99-1.02)
Second trimester:
OR = 1.02 (1.01-1.03)
First trimester: OR = 1.01 (0.99-1.02)
Change in birth weight
All births
One month before birth:
IS = -1.21 (-2.31--0.11)
Third trimester: IS = -0.95 (-2.30-0.40)
Second trimester:
IS = -2.05 (-3.36--0.74)
First trimesters = -0.14 (-1.37-1.09)
5 km births
One month before birth:
IS = -2.98(-4.25--1.71)
Third trimester: IS = -3.84 (-5.35- -2.33)
Second trimester:
IS = -4.28 (-5.79--2.77)
First trimesters = -2.57 (-4.04--1.10)
Key second trimester findings
Single pollutant model:
IS = -4.28 (-5.79--2.77)
2 pollutant (PM10 and CO):
IS = -3.72 (-6.29--1.15)
2 pollutant (PM10 and N02):
IS = -2.65 (-4.32--0.98)
2 pollutant (PM,0 and 03):
IS = -5.47 (-7.06--3.88)
4 pollutant (PM10, N02, CO and 03):
IS = -3.27 (-7.05-0.51)
Controlling for exposures in other
pregnancy periods:
IS = -3.03(-4.85--1.21)
Reference: Pereira et al. (1998,
0072641
Period of Study: Jan 1991-Dec 1992
Location: Sao Paulo, Brazil
Notes: Paper does not focus on PM as
a pollutant of interest.
Outcome: Intrauterine mortality
(fetuses over 28 wk of pregnancy)
Study Design: Time-series
N: 730 days with PM measures
Statistical Analyses: Poisson
regression
Covariates: Season, day of the week
and weather (temperature and relative
humidity)
Season: Assessed by including 24
indicator variables for month and yr
Dose-response Investigated? No
Statistical Package: NR
Lags Considered: Paper focuses on
other pollutants (lags for PM not
reported)
Pollutant: PM,0
Averaging Time: 24 h mean
Mean (SD): 65.04 (27.28)
Range (Min, Max): (14.80,192.80)
Monitoring Stations: 13 (avgd to
provide city-wide pollutant level)
Copollutants (correlation):
N02(r = 0.45)
S02(r = 0.74)
CO (r = 0.41)
03(r = 0.25)
PM Increment: NR (reported only
regression coefficients for PM)
Effect Estimate [Lower Cl, Upper Cl]:
Regression coefficients (standard
errors) for pollutants when considered
separately and simultaneously in the
completed model:
Separately: 0.0008 (0.0006)
Simultaneously: -0.0005 (0.0010)
December 2009
E-458
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ritz et al. (2000, 0120681
Period of Study: 1989-1993
Location: Southern California
Outcome: Preterm birth (treated
dichotomously as birth at <37 wk
gestation
Also analyzed continuously)
Age Groups: Infants (born vaginally
between 26-44 wk of gestation)
Study Design: Cross-sectional
N: 97,158 births
Statistical Analyses: Logistic and
linear regression
Covariates: Maternal age, race,
education, parity, interval since the
previous live birth, access to prenatal
care, infant sex, previous low weight or
preterm births, smoking (reported as
"pregnancy complications")
To examine effect modification, authors
conducted stratified analysis by region,
birth and conception seasons, maternal
age, race, education, and infant gender
Season: Some models included
season of birth or conception
Also assessed as effect modifier in
stratified analyses
Dose-response Investigated?
Examined adequacy of linear or log-
linear relation using indicator terms for
pollutant-avg quartiles
Results presented in Fig 2 (dose-
response demonstrated for last 6 wk
exposure period)
Statistical Package: NR
Pollutant: PM,0
Averaging Time: 24-h avg at 6 day
intervals
avgd pollutant measures for 1, 2, 4, 6,
8,12, and 26 wk before birth and the
whole pregnancy period
Mean (SD): 6 wk before birth: 47.5
(15.0)
1st month of pregnancy: 49.3 (16.9)
Range (Min, Max): 6 wk before birth:
12.3-152.3
1st month of pregnancy: 9.5-178.8
Monitoring Stations: 17 stations (PM
measured at only 8 stations)
Copollutants (correlations):
6 wk before birth:
CO (r = 0.43)
N02(r = 0.74)
03(r = 0.20)
1st month of pregnancy:
CO (r = 0.37)
N02(r = 0.71)
03(r = 0.23)
Notes: Avgd pollutant measures taken
at the air monitoring station closest to
the residence
PM Increment: 50 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
All 8 stations
6 wk before birth
Crude: 1.20 (1.09,1.33)
2 exposure periods: 1.18 (1.07,1.31)
Other riskfactors: 1.15 (1.04,1.26)
Other RFs plus season: 1.15(1.03,
1.29)
Multipollutant model: 1.19 (1.01,1.40)
1st month of pregnancy
Crude: 1.16 (1.06,1.26)
2 exposure periods: 1.13 (1.04,1.24)
Other riskfactors: 1.09 (1.00,1.19)
Other RFs plus season: 1.09 (0.99,
1.20)
Multipollutant model: 1.12 (0.97,1.29)
Coastal stations only
6 wk before birth
Crude: 1.22 (1.00,1.49)
2 exposure periods: 1.28 (1.04,1.56)
Other risk factors: 1.13 (0.93,1.38)
Other RFs plus season: 1.18(0.92,
1.51)
Multipollutant model: 1.42 (097, 2.01)
1st month of pregnancy
Crude: 1.28 (1.06,1.54)
2 exposure periods: 1.32 (1.09,1.59)
Other risk factors: 1.17 (0.97,1.40)
Other RFs plus season: 0.99 (0.79,
1.24)
Multipollutant model: 1.09(0.83,1.41)
Inland stations only
6 wk before birth
Crude: 1.27 (1.12,1.44)
2 exposure periods: 1.27 (1.11,1.44)
Other riskfactors: 1.19 (1.05,1.35)
Other RFs plus season: 1.27(1.10,
1.48)
Multipollutant model: 1.18 (0.97,1.43)
1st month of pregnancy
Crude: 1.16 (1.04,1.29)
2 exposure periods: 1.16 (1.04,1.29)
Other risk factors: 1.09 (0.98,1.21)
Other RFs plus season: 1.09 (0.97,
1.24)
Multipollutant model: 1.11 (0.93,1.33)
Crude estimates for last 6 wk
exposure by season
Fall: 1.08 (0.88, 1.31)
Summer: 1.06 (0.87,1.29)
Winter: 1.33 (1.07,1.65)
Spring: 1.81 (1.41,2.31)
Reduction in mean gestation length
for each increase in PM10 during last
6 wk before birth (linear regression
analysis)
Crude: 0.66 (± 0.24) days
Adj: 0.90 (± 0.27) days
Notes: Effect estimates remain stable
when excluding SGA or LBW children
or when restricting preterm births to
SGA or LBW children only (results not
presented)
December 2009
E-459
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ritz, et al. (2002, 0232271
Period of Study: 1987-1993
Location: Southern California
(Jul1990-Jul 1993 for Los Angeles,
1989 for Riverside, 1988-1989 for San
Bernardino, and 1987-1989 for Orange
counties
Outcome:
1 Aortic defects
2 Defects of the atrium and atrium
Sepum
3 Endocardial and mitral valve defects
4 Pulmonary artery and valve defects
5) Conotruncal defects including
tetralogy of Fallot, transposition of great
vessels, truncus arteriosus communis,
double outlet right ventricle, and
aorticopulmonary window
and 6) Ventricular Sepal defects not
included in the conotruncal category.
Age Groups: All live born infants and
fetal deaths diagnosed between 20 wk
of gestation and 1 yr after birth
Study Design: Case-control
N: 10,649 infants and fetuses
Statistical Analyses: Hierarchical (two-
level) regression model, polytomous
logistic regression, linear model
Covariates: Gender, no prenatal care,
multiple births, no siblings, maternal
race, maternal age, maternal education,
born before 1990, season of
conception,
Season: All
Dose-response Investigated? Yes, for
03 and CO, study found a clear dose-
response pattern for aortic Sepum and
valve and ventricular Sepal defects and
possibly for conotruncal and pulmonary
artery and valve defects
Statistical Package: SAS
Pollutant: PM,0
Averaging Time: 24 h (every 6 days)
PM Component: vehicle emissions
Monitoring Stations: 11 (for PM10)
Copollutants (correlations):
CO: r = 0.32
N02(NR)
03 (NR)
Notes: The authors did not observe
consistently increased risks and dose-
response patterns for PM10 after
controlling for the effects of CO and 03
on these cardiac defects. (Quantitative
results not shown).
December 2009
E-460
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ritz et al. (2006, 0898191
Period of Study: 1989-2000
Location: 389 South Coast Air Basin
(SoCAB) zip codes
Outcome: Total infant deaths during the
first yr of life as well as all respiratory
causes of death (ICD-9 codes 460-519,
769, 770.4, 770.7, 770.8, and 770.9
and ICD-10 codes JOO-J98, P22.0,
P22.9, P27.1.P27.9, P28.0, P28.4,
P28.5, and P28.9) and sudden infant
death syndrome (SIDS) (ICD-9 code
798.0 and ICD-10 code R95).
Age Groups: Infants 0-1 yr
Study Design: Case-control
N: 2,975,059 births and 19,664 infant
deaths
Cases, n = 13,146
Controls, n = 151,015
Statistical Analyses: Conditional
logistic regression analysis
Covariates: Risk factors available on
birth and/or death certificates (maternal
age, race/ethnicity, and education, level
of prenatal care, infant gender, parity,
birth country, and death season)
Season: Death season (spring,
summer, fall, winter)
Dose-response Investigated? Yes
Statistical Package: NR
Pollutant: PMi0
Averaging Time: 24 h
Mean (SD):
2 wk before death: 46.2
1 month before death: 46.3
2 mo before death: 46.3
6 mo before death: 46.3
Range (Min, Max):
2 wk before death: (21.0-83.5)
1 month before death: (25.0-77.2)
2 mo before death: (27.6-74.2)
6 mo before death: (31.3-69.5)
Monitoring Stations: maximum of 31
Copollutants (correlation):
2 wk before death
CO: r = 0.33
N02:r = 0.48
03:r = 0.12
1 month before death
CO: r = 0.33
N02:r = 0.48
03:r = 0.12
2 mo before death
CO: r = 0.32
N02:r = 0.48
03:r = 0.12
6 mo before death
CO: r = 0.29
N02:r = 0.44
03:r = 0.16
PM Increment: 10 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
All-cause death
2 mo before death
Single-pollutant model:
<25th = 1.04 (1.01-1.06)
25th-75th = 0.96 (0.89-1.04)
>75th = 1.14 (1.03-1.27)
Multiple-pollutant model:
<25th = 1.02 (0.99-1.05)
25th-75th = 0.92 (0.84-1.00)
>75th = 1.07 (0.95-1.20)
SIDS
2 mo before death:
Single-pollutant model:
<25th = 1.03 (0.99-1.08)
25th-75th = 0.94 (0.81-1.08)
>75th = 1.13 (0.93-1.36)
Multiple-pollutant model:
<25th = 1.01 (0.95-1.07)
25th-75th = 0.90 (0.76-1.06)
>75th = 0.99 (0.80-1.24)
Respiratory death
2 wk before death
Postneonatal deaths (28 days to 1 y)
Single-pollutant model:
<25th = 1.05 (1.01-1.10)
25th-75th = 1.13 (1.01-1.10)
>75th = 1.46 (1.13-1.88)
Multiple-pollutant model:
<25th = 1.04 (0.98-1.09)
25th-75th = 1.09 (0.86-1.38)
>75th = 1.40 (1.03-1.89)
Postneonatal deaths (28 days to 3
mo)
Single-pollutant model:
<25th = 1.01 (0.95-1.08)
25th-75th = 1.16(0.82-1.63)
>75th = 1.44 (0.96-2.17)
Multiple-pollutant model:
<25th = 1.00 (0.92-1.09)
25th-75th = 0.97 (0.67-1.42)
>75th = 1.23 (0.76-2.00)
Post neonatal deaths (4-12 mo)
Single-pollutant model:
<25th = 1.12 (1.02-1.23)
25th-75th = 1.08 (0.81-1.44)
>75th = 1.41 (1.02-1.96)
Multiple-pollutant model:
<25th = 1.07 (1.00-1.15)
25th-75th = 1.02 (0.75-1.40)
>75th = 1.36 (0.92-2.01)
December 2009
E-461
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Rogers et al. (2006,
0912321
Period of Study: 1986-1988
Location: Georgia, USA
Outcome: VLBW
Term, AGA, Preterm AGA, Preterm,
SGA
Age Groups: Newborns and their
mothers (< 19 to >35-yr-old)
Study Design: Case-control
N: 325 infants (69 preterm SGA
59 preterm AGA
197 term AGA) and their mothers
Statistical Analyses: Logistic
regression
Covariates: Maternal age, maternal
race, maternal education, active and
passive smoking, birth season,
prepregnancy weight, pregnancy weight
gain, maternal toxemia, anemia,
asthma
Dose-response Investigated? Yes,
used
Statistical Package: SUDAAN
Cochran-Armitage test for trend to
determine whether the observed
proportions of cases and controls
differed in a linear manner across
exposure categories.
Pollutant: PM,0
Averaging Time: annual
Preterm SGA:
60th(Median): 3.38
Preterm AGA:
60th(Median): 7.84
Term AGA:
60th(Median): 3.23
Monitoring Stations: NR
Percent Mothers Residing In County
With Industrial Point Source
Preterm SGA: 60.9%
Preterm AGA: 79.7%
Term AGA: 60.4%
Percent Mothers Residing In PM10
Quartile (based on environmental
transport model)
Preterm SGA
1st quartile(<1.48): 31.9%
2nd quartile (1.48-3.74): 18.8%
3rd quartile (3.75-15.07): 26.1%
4th quartile (>15.07): 23.2%
Preterm AGA
1st quartile (<1.48): 16.9%
2nd quartile (1.48-3.74): 22.1%
3rd quartile (3.75-15.07): 28.8%
4th quartile (>15.07): 32.2%
Term AGA
1st quartile (<1.48): 24.7%
2nd quartile (1.48-3.74): 28.4%
3rd quartile (3.75-15.07): 27.9%
4th quartile (>15.07): 19.3%
PM Increment: Quartile
Notes: Statistically significant increases
in the odds of VLBW and preterm AGA
births are associated with living in a
county with a PMi0 point source.
Preterm AGA births are also associated
with living in an area with very high (4th
quartile) estimated PM10 exposure.
Delivery of VLBW vs. Term AGA infant
County with point source
2.54 [1.46, 4.22]
PM10 quartile
1st quartile: reference
2nd quartile:
0.81 [0.42, 1.55]
3rd quartile:
0.85 [0.45, 1.16]
4th quartile:
1.94 [0.98, 3.83]
Delivery of Preterm AGA vs. Term AGA
infant
County with point source
4.31 [1.88:9.87]
PM10 quartile
1st quartile: reference
2nd quartile:
1.56 [0.56: 4.35]
3rd quartile:
1.19 [0.44: 3.23]
4th quartile:
3.68 [1.44: 9.44]
Delivery of Preterm AGA vs. Preterm
SGA infant
County with point source
2.07 [0.83: 5.16]
PM10 quartile
1st quartile: reference
2nd quartile:
1.96 [0.59: 6.43]
3rd quartile:
2.10 [0.66: 6.73]
4th quartile:
2.58 [0.78: 8.51]
December 2009
E-462
-------
Study
Reference: Romieu et al. (2004,
0930741
Period of Study: 1997-2001
Location: Ciudad Juarez, Mexico
Design & Methods
Outcome: Respiratory-related infant
mortality ICD9 (460-51 9)
ICD10(JOO-J99)
Age Groups: 1 month to 1 yr
Study Design: Case crossover
N: 216 respiratory-related deaths
N = 412 other causes and N = 628 total
deaths
Statistical Analyses: The acute effects
of air pollution was modeled on both
total and respiratory-related mortality as
a function of the pollution levels on the
same day and preceding days and over
2- and 3-day avg before the date of
death. Case-crossover with semi-
symmetric bidirectional referent
selection was the approach used. Data
were stratified by day of the week and
calendar month. Data were analyzed
with conditional logistic regression.
Second and third polynomial distributed
lag models were used to study lag
structure. BIC was used to determine
Igq ignqth
Covariate: Temperature, season
Concentrations1
Pollutant: PM,0
Averaging Time: 24 h
Mean (SD):
1997: 33.04 (20.67) pg/m3
1998: 35.25 (17.32) pg/m3
1999: 45.92 (28.69) pg/m3
2000: 43.38 (23.77) pg/m3
2001: 39.46 (29.43) pg/m3
Monitoring Stations: 5 stations in
Ciudad Juarez
2 stations in El Paso (close to U.S.-
Mexico border)
Copollutant (correlation): 0; r = 0.01
Notes: Ciudad Juarez monitors
measured PM10 every 6 days while El
Paso monitors measured on a daily
basis.
Effect Estimates (95% Cl)
PM Increment: 20 pg/m3
RR Estimate [Lower Cl, Upper Cl]
lag:
Total mortality:
OR = 1.02 (0.94-1. 11) lag 1
OR = 1.03 (0.95-1. 12) lag 2
OR = 1.03 (0.94-1.1 3 ac2
OR = 1.04 (0.95-1.1 5) ac3
Respiratory mortality
OR = 0.95 (0.83-1. 09) lag 1
OR = 1.04 (0.91 -1.1 9 lag 2
OR = 0.98 (0.81 -1.1 9 ac2
OR = 0.97 (0.74-1. 26) ac3
Higher SES
OR = 0.82 (0.59, 1.1 4) lag 1
OR = 1.08 (0.84, 1.40) lag 2
OR = 0.89 (0.58, 1.35 ac2
OR = 0.97 (0.52, 1.82 ac3
Medium SES
OR = 0.99 (0.79, 1.27) lag 1
OR = 1.11 (0.86, 1.43) lag 2
OR = 1.03 (0.73, 1.45)ac2
OR = 1.1 7 (0.72, 1.90)ac3
Lower SES
OR = 1.61 (0.97-2.66) lag 1
OR = 1.07 (0.65, 1.75) lag 2
OR = 2. 56 (1.06-6.1 7) ac2
OR = 1.76 (0.59, 5.23) ac3
Notes:
Dose-response Investigated? Yes
Statistical Package: STATA7.0
Lags Considered: 1-15 days
ac2 and ac3 represent cumulative PMi0
ambient levels over 2 or 3 days before
death.
Reference: Sagiv et al. (2005, 0874681
Period of Study: Jan 1997-Dec 2001
Location: Allegheny county, Beaver
county, Lackawanna county,
Philadelphia county, Pennsylvania,
U.S.A.
Outcome: Preterm birth (<36 wk)
Age Groups: Babies born between 20
and 44 wk
Study Design: Time series
N: 3704 observation days, 187,997
births
Statistical Analyses: Poisson
regression
Multivariable mixed-effects model with a
random intercept for each county to
incorporate count-level information.
Covariates: Temperature, dew point
temperature, mean 6-week level of
copollutants (CO, N02, and S02), long-
term preterm birth trends
Season: All
Dose-response Investigated? Yes
Statistical Package: NR
Lags Considered: 1,2, 3, 4, 5, 6, 7
Pollutant: PM10
Averaging Time: Daily used to
calculate 6-week period
Mean (SD): 6-week period
27.1 (8.3)
Daily
25.3(14.6)
Percentiles: 6-week period
60th (Median): 26.0 Daily
60th (Median): 21.6
Range (Min, Max): 6-week period: 8.7,
68.9
Daily: 2.0,156.3
Monitoring Stations: NR
Copollutant (correlation): Daily
PMio-dailyS02:r = 0.46
Also considered CO, N02 and 03 as
copollutants.
PM Increment: 1) 50 pg/m 2) Quartiles
(first quartile is the reference)
Exposure period: 6 wk before birth
Per 50 pg/m3:1.07 (0.98,1.18)
2nd quartile: 1.00 (0.95,1.05)
3rd quartile: 1.04 (0.99,1.09)
4th quartile: 1.03 (0.98,1.08)
Exposure period: 1-day acute time
windows Per 50 pg/m 2-day lag: 1.10
(1.00,1.21)
5-day lag: 1.07 (0.98,1.18)
Notes: Within the article, authors
provide a Fig 1 displaying a graph of
the relative risk (RR) and 95%
confidence i
ntervals (Cl) for 1- to 7-day lags. While
the authors report the 2- and 5-day lag
RRs and 95% CIs in the text, the others
are not specifically reported. However,
the Fig shows the approximate RRs per
50 pg/m3 as indicated below:
1-day lag: 1.05
3-day lag: 1.05
4-day lag: 1.00
6-day lag: 0.97
7-day lag: 1.03
December 2009
E-463
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: (Salam et al., 2005,
0878851
Period of Study: 1975-1987
Location: Southern California
Outcome: Birth weight
Low birth weight (LBW
<2500 g)
Intrauterine growth retardation (IUGR)
Age Groups: Children born full-term
(between 37 and 44 wk)
Study Design: Cohort study
N: 3901 children
Statistical Analyses: Linear mixed-
effects
Logistic regression
Covariates: Maternal age, months
since last live birth, parity, maternal
smoking during pregnancy, SES, marital
status at childbirth, gestational diabetes,
child's sex, child's race/ethnicity, child's
grade in school (4th, 7th, and 10th),
Julian day of birth
Season: All
Dose-response Investigated? Yes
Statistical Package: SAS
Pollutant: PM,0
Averaging Time: Monthly
Mean (SD):
Entire pregnancy: 45.8 (12.9)
First trimester: 46.6 (15.9)
Second trimester: 45.4 (14.8)
Third trimester: 45.4 (15.5)
Monitoring Stations: 1 or 3 (See
notes)
Copollutant (correlation):
Entire pregnancy
PMio-03[10-6]:r = 0.54
PMio-03[24h):r = 0.20
PM10-N02:r = 0.55
PMio-CO:r = 0.41
First trimester
PM10-03[10-6]:r = 0.54
PMio-03[24h]:r = 0.34
PMio-N02:r = 0.48
PM10-CO:r = 0.29
Second trimester
PM10-03[10-6):r = 0.50
PM10-03(24h):r = 0.27
PMio-N02:r = 0.53
PM10-CO:r = 0.35
Third trimester
PMio-03[10-6]:r = 0.52
PMio-03[24h]:r = 0.31
PM10-N02:r = 0.52
PMio-CO:r = 0.37
Notes: Exposure estimates were
calculated by spatially interpolated
monthly avg which were based off of
three monitoring stations located within
50 km of the ZIP code region of
maternal birth residences.
PM Increment: IQR (interquartile
range)
Outcome: birth weight (g)
Single-pollutant model
Entire pregnancy
18 pg/m3: -19.9 (-43.6, 3.8)
First trimester
20 pg/m3: -3.0 (-22.7, 16.7)
Second trimester
19|jg/m3:-15.7(-36.1,4.7)
Third trimester
20|jg/m3:-21.7(-42.2to-1.1)
Multipollutant model
(Included 03
(24 h) in model
Third trimester exposure)
20|jg/m3:-10.8(-31.8,10.2)
Outcome: IUGR (ORs)
Single-pollutant model
Entire pregnancy
18|jg/m3:1.1 (0.9,1.3)
First trimester
20|jg/m3:1.0(0.9,1.2)
Second trimester
19|jg/m3:1.0(0.9,1.2)
Third trimester
20|jg/m3:1.1 (0.9,1.3)
Outcome: LBW
Single-pollutant model
Entire pregnancy
18|jg/m3:1.3(0.8, 2.2)
First trimester
20|jg/m3:1.0(0.7,1.5)
Second trimester
19|jg/m3:1.2(0.8,1.7)
Third trimester
20|jg/m3:1.3(0.9,1.9)
Notes: Numbers reported for birth
weight outcome are the effects on birth
weight outcome (the change in birth
weight in grams) across the IQR (which
vary depending on air pollutant and
duration of exposure measurement).
Reference: (Sokol et al, 2006, 0985391 Outcome: Semen Quality
Period of Study: Jan 1996-Dec 1998 Study Design: Panel
Location: Los Angeles, California Statistical Analysis: Univariate and
Multivariate Regression
Statistical Package: SAS
Age Groups: Males ranging 19-35 in
age
Pollutant: PM10
Averaging Time: 0-9d, 10-14d and 70-
90d
Mean (SD) Unit: 35.74 + 13.83 pg/m3
Copollutant (correlation):
03, N02, CO
PM10 specific results are given in Fig 3-.
PM10 was not significantly correlated
with sperm quality.
Reference: (Suh et al, 2007,1570281
Period of Study: 2001-2004
Location: Seoul, Korea
Outcome: Birth weight
Age Groups: Prenatal follow-up for
newborns
Study Design: based prospective
cohort study
N: 199 pregnant mothers
Statistical Analyses: ANCOVA,
generalized linear models
Covariates: infant's sex, maternal age,
maternal and paternal education, parity,
presence of illness during pregnancy,
delivery month, gestational age
(squared)
Dose-response Investigated? Yes
Statistical Package: SAS
Pollutant: PM,0
Averaging Time: 24-h
Mean (SD): 1st trimester: 76.41
2nd trimester: 77.84 (31.63)
3rd trimester: 95.61 (26.15)
Percentiles: 1st trimester
26th: 55.28
60th(Median):71.09
76th: 92.38
2nd trimester
26th: 48.65
60th(Median): 72.36
76th: 108.00
3rd trimester
26th:77.10
60th(Median): 96.35
76th: 116.68
Range (Min, Max):
1st trimester (21.00,151.65)
PM Increment: Trimester Ł 90th
percentile compared to <90thpercentile
Least-square (ANCOVA) mean (SE)
All Genotypes
1st trimester
<90th, N(%):
158 (90.3%): 3253 (37)
> 90th percentile, N(%): 17 (9.7%):
2841 (145)
P-Value for mean birth weight for > 90th
percentile PM10 vs. for <90th percentile
PM10
Adjusted: 0.009
Adjusted, with CO: 0.041
Adjusted, with N02: 0.092
Adjusted, with S02: 0.012
2nd trimester
<90th percentile, N(%):
153 (89.5%): 3253 (39)
> 90th percentile, N(%):
18 (10.5%): 3026 (157)
December 2009
E-464
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
2nd trimester (31.45,139.13)
3rd trimester (23.45,172.75)
Monitoring Stations: 27
Copollutant:
CO
S02
N02
P-Value for mean birth weight for 2 90th
percentile PMi0 vs. for <90th percentile
PM10
Adjusted: 0.177
Adjusted, with CO: 0.203
Adjusted, with N02: 0.151
Adjusted, with S02: 0.151
3rd trimester
<90th percentile, N(%):
162 (90.5%): 3226 (38)
> 90th percentile, N(%): 17 (9.5%):
3122(140)
P-Value for mean birth weight for > 90th
percentile PMi0 vs. for <90th percentile
PM10
Adjusted: 0.487
Adjusted, with CO: 0.748
Adjusted, with N02: 0.420
Adjusted, with S02: 0.466
Genotype Mspl TT
1st trimester
<90th percentile, N(%): 60 (34.3%):
3350 (64)
> 90th percentile, N(%): 5 (2.9%): 3001
(229)
P-Value for mean birth weight for 2 90th
percentile PM10 vs. for <90th percentile
PM10
Adjusted: 0.147
Adjusted, with CO: 0.186
Adjusted, with N02: 0.430
Adjusted, with S02: 0.155
2nd trimester
<90th percentile, N(%): 59 (34.5%):
3335 (66)
> 90th percentile, N(%): 6 (3.5%): 3281
(249)
P-Value for mean birth weight for
> 90th percentile PM,o vs. for <90th
percentile PM10
Adjusted: 0.833
Adjusted, with CO: 0.833
Adjusted, with N02: 0.778
Adjusted, with S02: 0.806
3rd trimester
<90th percentile, N(%): 61 (34.1%):
3327 (65)
> 90th percentile, N(%): 6 (3.4%): 3227
(300)
p-Value for mean birth weight for
> 90th percentile PM,o vs. for <90th
percentile PM10
Adjusted: 0.749
Adjusted, with CO: 0.980
Adjusted, with N02: 0.635
Adjusted, with S02: 0.687
Genotype Mspl TC/CC
1st trimester
<90th percentile, N(%): 98 (56.0%):
3193 (48)
> 90th percentile, N(%): 12(6.9%):
2799(169)
P-Value for mean birth weight for
> 90th percentile PM,0 vs. for <90th
percentile PMi0
Adjusted: 0.033
Adjusted, with CO: 0.073
Adjusted, with N02: 0.150
Adjusted, with S02: 0.036
2nd trimester
<90th percentile, N(%): 94 (55.0%):
3200 (52)
> 90th percentile, N(%): 12 (7.0%):
2933(176)
P-Value for mean birth weight for
> 90th percentile PM,0 vs. for <90th
percentile PMi0
Adjusted: 0.161
December 2009
E-465
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
Adjusted, with CO: 0.172
Adjusted, with N02: 0.152
Adjusted, with S02: 0.158
3rd trimester
<90th percentile, N(%): 101 (56.4%):
3165(49)
> 90th percentile, N(%): 11 (6.2%):
3087(147)
P-Value for mean birth weight for
> 90th percentile PMi0 vs. for <90th
percentile PMio
Adjusted: 0.626
Adjusted, with CO: 0.978
Adjusted, with N02: 0.551
Adjusted, with S02: 0.614
Genotype Ncol llelle
1st trimester
<90th percentile, N(%): 87 (49.7%):
3244 (52)
> 90th percentile, N(%): 7 (4.0%): 2983
(232)
P-Value for mean birth weight for
> 90th percentile PM10 vs. for <90th
percentile PM10
Adjusted: 0.289
Adjusted, with CO: 0.344
Adjusted, with N02: 0.641
Adjusted, with S02: 0.293
2nd trimester
<90th percentile, N(%): 82 (48.0%):
3243 (55)
> 90th percentile, N(%): 11 (6.4%):
3185(207)
p-Value for mean birth weight for
> 90th percentile PM10 vs. for <90th
percentile PM10
Adjusted: 0.790
Adjusted, with CO: 0.783
Adjusted, with N02: 0.707
Adjusted, with S02: 0.733
3rd trimester
<90th percentile, N(%): 90 (50.3%):
3239 (53)
> 90th percentile, N(%): 9 (5.0%): 2944
(198)
P-Value for mean birth weight for
> 90th percentile PM10 vs. for <90th
percentile PM10
Adjusted: 0.161
Adjusted, with CO: 0.279
Adjusted, with N02: 0.134
Adjusted, with S02: 0.150
Genotype Ncol HeVal/ValVal
1st trimester
<90th percentile, N(%): 71 (40.6%):
3262 (56)
> 90th percentile, N(%): 10(5.7%):
2773(171)
P-Value for mean birth weight for
> 90th percentile PM,o vs. for <90th
percentile PM10
Adjusted: 0.009
Adjusted, with CO: 0.031
Adjusted, with N02: 0.058
Adjusted, with S02: 0.010
2nd trimester
<90th percentile, N(%): 71 (41.5%):
3264(61)
> 90th percentile, N(%): 7 (4.1%): 2862
(208)
P-Value for mean birth weight for
> 90th percentile PM,o vs. for <90th
percentile PM10
Adjusted: 0.076
Adjusted, with CO: 0.093
Adjusted, with N02: 0.063
Adjusted, with S02: 0.061
3rd trimester
December 2009 E-466
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
<90th percentile, N(%): 72 (40.2%):
3207 (58)
> 90th percentile, N(%): 8 (4.5%): 3262
(180)
P-Value for mean birth weight for
> 90th percentile PMi0 vs. for <90th
percentile PM10
Adjusted: 0.777
Adjusted, with CO: 0.607
Adjusted, with N02: 0.843
Adjusted, with S02: 0.791
Reference: Tsai et al. (2006, 0907091
Period of Study: 1994-2000
Location: Kaohsiung, Taiwan
Outcome: Post neonatal mortality
Age Groups: Infants more than 27
days and less than 1 yr
Study Design: Case-crossover study
N: 207 deaths
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature, humidity
Dose-response Investigated? No
Pollutant: PM,0
Averaging Time: 24 h
Mean(SD):81.45|jg/m3
Percentiles: 26th: 44.50
60th(Median): 79.20
76th:111.50
Range (Mm, Max): (20.50-232.00)
Monitoring Stations: 6
PM Increment: 67.00 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
OR = 1.040 (0.340-3.177)
Note: Air pollution levels at the dates of
infant death were compared with air
pollution levels 1 week before and 1
week after death
A cumulative lag up to 2 previous days
was used to assign exposure.
Reference: Wilhelm and Ritz (2005,
088668)
Period of Study: 1994-2000
Location: Los Angeles County,
California, U.S.
Statistical Package: SAS, version 8.2
Outcome: Term low birth weight (LBW)
(<2500gat>37completedwk
gestation), Vaginal birth <37 completed
wk gestation
Age Groups: LBW: 2 37 completed wk
Preterm births: <37 completed wk
Study Design: Cross-sectional
N: For LBW: 136,134
Forpreterm birth:
106,483
Copollutant:
S02
N02
CO
03
Pollutant: PM,0
Averaging Time:
24 h (every 6 days)
Entire pregnancy
Trimesters of pregnancy
Months of pregnancy
6 wk before birth
Mean (SD):
First trimester: 42.2
Third trimester: 41. 5
6 wk before birth: 39.1
Range (Min, Max):
PM Increment:
1)10|jg/m3
2) 3 levels:
a) <25 percentile (reference)
b) 25%-75 percentile
c) > 75 percentile
Incidence of LBW (third trimester
exposure)
<32.8|jg/m3:2.0(1.8, 2.2)
32.8to<43.4|jg/m3:2.0(1.9, 2.1)
> 43.4 pg/m3: 2.2 (2.0, 2.4)
Incidence of preterm birth (first
trimester exposure)
Statistical Analyses: Logistic
regression
Covariates: Maternal age, maternal
race, maternal education, parity, interval
since previous live birth, level of
prenatal care, infant sex, previous LBW
or preterm infant, birth season, other
pollutants (CO, N02, 03, PM10),
gestational age (in birth weight analysis)
Dose-response Investigated? Yes
Statistical Package: NR
First trimester: 26.3, 77.4
Third trimester: 25.7, 74.6
6 wk before birth: 13.0, 103.7
Monitoring Stations:
Zip-code-level analysis: 8
Address-level analysis: 6
Copollutant (correlation):
First trimester:
PM,o-CO:r = 0.12
PM10-N02:r = 0.29
PMio-03:r = -0.01
PMio-PM25:r = 0.43
Third trimester:
PMio-CO:r = 0.32
PMio-N02:r = 0.45
PMio-03:r = -0.08
PMio-PM25:r = 0.52
6 wk before birth:
PM10-CO:r = 0.36
PM,o-N02:r = 0.49
PMio-03:r = -0.16
PM10-PM25:r = 0.60
<32.9 pg/rri : 8.7 (8.3, 9.2)
32.9to<43.9|jg/m3:8.8(8.5, 9.1)
> 43.9 pg/m3: 8.6 (8.1, 9.0)
Incidence of preterm birth (6 wk
before birth exposure)
<31.8|jg/m3:8.8(8.4, 9.3)
31.8to<44.1 pg/m3:8.6 (8.3, 8.9)
a44.1 pg/m3: 8.8 (8.4, 9.2)
Outcome: LBW
Exposure Period: Third trimester
Address-level analysis:
Single-pollutant model:
Distance < 1 mile
Per 10 pg/m3:1.22 (1.05,1.41)
33.4 to <44.7|jg/m3:1.08 (0.76, 1.52)
> 44.7 pg/m3:1.48 (1.00, 2.19)
Multipollutant model:
Distance Ł 1 mile
Per 10 pg/m3:1.36 (1.12,1.65)
33.4to<44.7|jg/m3:1.16 (0.77, 1.74)
> 44.7 pg/m3:1.58 (0.95, 2.62)
Single-pollutant model:
1 44.7 pg/m3: 0.96 (0.78, 1.18)
Multipollutant model:
1 44.7 |jg/m3:1.02 (0.79, 1.32)
December 2009
E-467
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
Single-pollutant model:
2 45.0 pg/m3:1.08 (0.97, 1.20)
Multipollutant model:
2 45.0 pg/m3:1.06 (0.93,1.21)
Zip-code-level analysis
Single-pollutant model:
Per 10 pg/m3:1.03 (0.97,1.09)
33.2 to <43.6|jg/m3:0.98 (0.86, 1.11)
> 43.6 pg/m3:1.03 (0.88, 1.21)
Multipollutant model:
Per 10 pg/m3:1.07 (0.99,1.15)
33.2 to <43.6|jg/m3: 0.97 (0.85, 1.12)
> 43.6 pg/m3:1.09 (0.90,1.31)
Outcome: LEW
Exposure Period: Entire pregnancy
period
Address-level analysis:
Multipollutant model:
Per 10 pg/m3:1.24 (0.91,1.70)
Outcome: Preterm Birth
Exposure Period: First trimester of
pregnancy
Address-level analysis:
Single-pollutant model:
Distance Ł 1 mile
Per 10 pg/m3:1.00 (0.93,1.09)
33.3to<45.1|jg/m3:1.07 (0.90, 1.26)
a45.1 fjg/m3:1.12 (0.91,1.38)
Multipollutant model:
Distance < 1 mile
Per 10 pg/m3:1.00 (0.90,1.12)
33.3to<45.1 |jg/m3:1.12(0.92, 1.36)
> 45.1 pg/m3:1.17 (0.90,1.50)
Single-pollutant model:
1 45.3 pg/m3:1.07 (0.97, 1.19)
Multipollutant model:
1 45.3 pg/m3:1.13 (1.00,1.27)
Single-pollutant model:
2 45.5 pg/m3:1.02 (0.96, 1.07)
Multipollutant model:
2 45.5 pg/m3: 0.94 (0.89, 1.01)
Zip-code-level analysis
Single-pollutant model:
Per 10 pg/m3: 0.99 (0.96,1.01)
33.3 to <44.2pg/m3:1.01 (0.95, 1.08)
> 44.2 pg/m3: 0.98 (0.90, 1.05)
Multipollutant model:
D^r ^ n i i^/fvi^1 n oo /n
Per 10 pg/rri : 0.99 (0.96,1.03)
33.3 to <44.2pg/m3:1.03 (0.97,
> 44.2 pg/m3:1.01 (0.92,1.11)
Outcome: Preterm birth
Exposure Period: 6 wk before birth
Address-level analysis:
Single-pollutant model:
Distance Ł 1 mile
December 2009 E-468
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
PeMOfjg/m°:1.02(0.95,1.10)
32.5 to <44.8|jg/m3:1.09 (0.92, 1.29)
> 44.8 pg/m3:1.12 (0.92, 1.37)
Multipollutant model:
Distance Ł 1 mile
Per 10 pg/m3:1.06 (0.97,1.16)
32.5 to <44.8|jg/m3:1.09 (0.90, 1.31)
> 44.8 pg/m3:1.17 (0.91,1.49)
Single-pollutant model:
1 45.3 pg/m3: 0.99 (0.89, 1.10)
Multipollutant model:
1 45.3 pg/m3:1.02 (0.91,1.16)
Single-pollutant model:
2 45.3 pg/m3: 0.98 (0.93, 1.03)
Multipollutant model:
2 45.3 pg/m3: 0.98 (0.92, 1.04)
Zip-code-level analysis
Single-pollutant model:
Per 10 pg/m3:1.02 (0.99,1.04)
32.1 to <44.3|jg/m3:1.01 (0.95, 1.07)
> 44.3 pg/m3:1.04 (0.96, 1.12)
Multipollutant model:
Per 10 pg/m3:1.02 (0.99,1.06)
32.1 to <44.3|jg/m3:1.02 (0.95, 1.09)
> 44.3 pg/m3:1.04 (0.95, 1.14)
Notes: multipollutant model adds
CO,N02, and 03 in addition to the main
pollutant of interest, PMi0.
December 2009 E-469
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Woodruff etal. (1997,
0842711
Period of Study: 1989-1991
Location: 86 Metropolitan Statistical
Areas in the U.S. (counties with
populations less than 100,000 were
excluded)
Outcome: Postneonatal mortality
(death of an infant between 1 month
and 1 yr of age)
1 All post neonatal deaths
2 Normal birth weight (NBW, > 2500 g)
SIDS deaths
3 NBW respiratory deaths
4 Low birth weight (LBW) respiratory
death
Respiratory deaths: ICD9 codes 460-
519
SIDS: ICD9 code 798.0
Age Groups: Infants (1 month-1yrof
age)
Study Design: Cross-sectional
N: 3,788,079 infants
Statistical Analyses: Logistic
regression
Covariates: Maternal education,
maternal race, parental marital status,
maternal smoking during pregnancy
Avg temperature during the first 2 mo of
life
Infant's month and yr of birth
Assessed race as an effect modifier (p-
val for interaction terms >0.2)
Dose-response Investigated? Yes
Statistical Package: NR
Pollutant: PM,0
Averaging Time: Mean of 1st 2 mo of
life
analyzed as tertiles of exposure and as
continuous exposure
Mean (SD): 31.4 (7.8)
Range (Min, Max):
Overall: 11.9-68.8
Low category: <28.0
Medium category: 28.1-40.0
High category: >40.0
Monitoring Stations: NR
PM Increment: 10 pg/m (for
continuous exposure analysis)
Adjusted ORs for cause-specific
post neonatal mortality by pollution
category (tertiles)
All causes
Low: Ref
Medium: 1.05 (1.01,1.09)
High: 1.10 (1.04,1.16)
SIDS, NBW:
Low: Ref
Medium: 1.09 (1.01,1.17)
High: 1.26 (1.14,1.39)
Respiratory death, NBW:
Low: Ref
Medium: 1.08 (0.87,1.33)
High: 1.40 (1.05,1.85)
Respiratory death, LBW:
Low: Ref
Medium: 0.93 (0.73,1.18)
High: 1.18 (0.86,1.61)
All other causes:
Low: Ref
Medium: 1.03 (0.97,1.08)
High: 0.97 (0.90, 1.04)
Adjusted ORs for a continuous
10 ug/m3 change in exposure
All causes: 1.04 (1.02,1.07)
SIDS, NBW: 1.12 (1.07, 1.17)
Respiratory death
NBW: 1.20 (1.06, 1.36)
Respiratory death
LBW: 1.05 (0.91, 1.22)
All other causes: 1.00 (0.99,1.00)
December 2009
E-470
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Woodruff etal. (2008,
0983861
Period of Study: 1999-2002
Location: U.S. counties with >250,000
residents (96 counties)
Outcome: Postneonatal deaths
Respiratory mortality (ICD10: JOOO-99,
plus bronchopulmonary dysplasia [BPD]
P27.1)
SIDS(ICD10:R95)
Ill-defined causes (R99);
All other deaths evaluated as a control
category
Age Groups: Infants aged >28 days
and <1 yr
Study Design: Cross-sectional
N: 3,583,495 births (6,639 post
neonatal deaths)
Statistical Analyses: Logistic GEE
(exchangeable correlation structure)
Covariates: Maternal race/ethnicity,
marital status, age, education,
primiparity, county-level poverty and per
capita income levels, yr and month of
birth dummy variables to account for
time trend and seasonal effects, and
region of the country
Sensitivity analyses performed among
only those mothers with smoking
information (adjustment for smoking
had no effect on the estimates)
Season: Adjusted for yr and month of
birth dummy variables to account for
time trend and seasonal effects
Dose-response Investigated?
Evaluated the appropriateness of a
linear form from analysis based on
quartiles of exposure and concluded
that linear form was appropriate (data
not shown)
Statistical Package: SAS
Pollutant: PM,0
Averaging Time: Measured
continuously for 24 h once every 6 days
exposure assigned by calculating avg
concentration of pollutant during first 2
mo of life
Median and IQR (25th-75th
percentile): Survivors: 28.9 (23.3-34.4)
All causes of death: 29.1 (23.9-34.5)
Respiratory: 29.8 (24.3-36.5)
SIDS: 28.6 (23.5-33.8)
SIDS+ill-defined: 28.8 (23.9-33.9)
Other causes: 29.2 (23.9-34.5)
Percent! les: see above
PM Component: Not assessed, but
controlled for region of the country to
account for PM composition variation
Monitoring Stations: NR
Copollutant (correlation):
PM10
PM25(r = 0.34)
CO (r = 0.18)
S02(r = 0.00)
03(r = 0.20)
Notes: Monthly avg calculated if there
were at least 3 available measures for
PM
Assigned exposures using the avg
concentration of the county of residence
PM Increment: IQR (11 pg/rri)
Effect Estimate [Lower Cl, Upper Cl]:
Adjusted ORs for single pollutant
models
All causes: 1.04 (0.99,1.10)
Respiratory: 1.18 (1.06,1.31)
SIDS: 1.02 (0.89, 1.16)
Ill-defined + SIDS: 1.06 (0.97,1.16)
Other causes: 1.02 (0.96,1.07)
Adjusted ORs for multipollutant models
(including CO, 03, S02)
Respiratory: 1.16 (1.04,1.30)
SIDS: 1.02 (0.90, 1.16)
OR for deaths coded as BPD per
increase in IQR: 1.19 (0.85,1.65)
OR for respiratory post neonatal death
stratified by birth weight
NBW only: 1.19 (1.05,1.36)
LBWonly: 1.12 (0.95,1.31)
OR for respiratory deaths removing
region of U.S. as a confounding
variable: 1.30 (1.04,1.61)
OR for respiratory deaths assessing
exposure as quartiles
Highest vs. Lowest quartile:
1.31 (1.00,1.71)
OR for respiratory deaths among only
those deaths that occurred during the
first 90 days (most closely matched
exposure metric of the avg over the first
2 mo of life): 1.25 (1.06,1.47)
Reference: (Suh et al, 2007,1570281
Period of Study: 2001-2004
Location: Seoul, Korea
Outcome: Birth weight
Age Groups: Prenatal follow-up for
newborns
Study Design: Based prospective
cohort study
N: 199 pregnant mothers
Statistical Analyses: ANCOVA,
generalized linear models
Covariates: Infant's sex, maternal age,
maternal and paternal education, parity,
presence of illness during pregnancy,
delivery month, gestational age
(squared)
Dose-response Investigated? Yes
Statistical Package: SAS
Pollutant: PM,0
Averaging Time: 24-h
Mean (SD):
1st trimester: 76.41 (28.80)
2nd trimester: 77.84 (31.63)
3rd trimester: 95.61 (26.15)
Percent! les:
1st trimester
25th: 55.28
60th(Median):71.09
76th: 92.38
2nd trimester
26th: 48.65
60th(Median): 72.36
76th: 108.00
3rd trimester
26th:77.10
60th(Median): 96.35
76th: 116.68
Range (Min, Max):
1st trimester (21.00,151.65)
2nd trimester (31.45,139.13)
3rd trimester (23.45,172.75)
Monitoring Stations: 27
Copollutant:
CO
S02
PM Increment: Trimester Ł 90th
percentile compared to <90th percentile
Least-square (ANCOVA) mean (SE)
All Genotypes
1st trimester
<90th percentile, N(%):
158 (90.3%): 3253 (37)
> 90th percentile, N(%): 17 (9.7%):
2841 (145)
P-Value for mean birth weight for
> 90th percentile PM10 vs. for <90th
percentile PM10
Adjusted: 0.009
Adjusted, with CO: 0.041
Adjusted, with N02: 0.092
Adjusted, with S02: 0.012
2nd trimester
<90th percentile, N(%):
153 (89.5%): 3253 (39)
> 90th percentile, N(%):
18 (10.5%): 3026 (157)
p-Value for mean birth weight for
> 90th percentile PM10 vs. for <90th
percentile PM10
Adjusted: 0.177
Adjusted, with CO: 0.203
Adjusted, with N02: 0.151
Adjusted, with S02: 0.151
3rd trimester
<90th percentile, N(%):
December 2009
E-471
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
~162 (90.5%): 3226 (38)
> 90th percentile, N(%): 17 (9.5%):
3122(140)
P-Value for mean birth weight for
> 90th percentile PM10 vs. for <90th
percentile PM10
Adjusted: 0.487
Adjusted, with CO: 0.748
Adjusted, with N02: 0.420
Adjusted, with S02: 0.466
Genotype Mspl TT
1st trimester
<90th percentile, N(%): 60 (34.3%):
3350 (64)
> 90th percentile, N(%): 5 (2.9%): 3001
(229)
P-Value for mean birth weight for
> 90th percentile PM,o vs. for <90th
percentile PM10
Adjusted: 0.147
Adjusted, with CO: 0.186
Adjusted, with N02: 0.430
Adjusted, with S02: 0.155
2nd trimester
<90th percentile, N(%): 59 (34.5%):
3335 (66)
> 90th percentile, N(%): 6 (3.5%): 3281
(249)
p-Value for mean birth weight for
> 90th percentile PM,o vs. for <90th
percentile PM10
Adjusted: 0.833
Adjusted, with CO: 0.833
Adjusted, with N02: 0.778
Adjusted, with S02: 0.806
3rd trimester
<90th percentile, N(%): 61 (34.1%):
3327 (65)
> 90th percentile, N(%): 6 (3.4%): 3227
(300)
P-Value for mean birth weight for
> 90th percentile PM,o vs. for <90th
percentile PM10
Adjusted: 0.749
Adjusted, with CO: 0.980
Adjusted, with N02: 0.635
Adjusted, with S02: 0.687
Genotype Mspl TC/CC
1st trimester
<90th percentile, N(%): 98 (56.0%):
3193 (48)
> 90th percentile, N(%): 12(6.9%):
2799(169)
P-Value for mean birth weight for
> 90th percentile PM,0 vs. for <90th
percentile PMi0
Adjusted: 0.033
Adjusted, with CO: 0.073
Adjusted, with N02: 0.150
Adjusted, with S02: 0.036
2nd trimester
<90th percentile, N(%): 94 (55.0%):
3200 (52)
> 90th percentile, N(%): 12 (7.0%):
2933(176)
P-Value for mean birth weight for
> 90th percentile PM,0 vs. for <90th
percentile PMi0
Adjusted: 0.161
Adjusted, with CO: 0.172
Adjusted, with N02: 0.152
Adjusted, with S02: 0.158
3rd trimester
<90th percentile, N(%):
101 (56.4%): 3165 (49)
> 90th percentile, N(%): 11 (6.2%):
3087(147)
P-Value for mean birth weight for
December 2009 E-472
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
> 90th percentile PM10 vs. for <90th
percentile PMi0
Adjusted: 0.626
Adjusted, with CO: 0.978
Adjusted, with N02: 0.551
Adjusted, with S02: 0.614
Genotype Ncol llelle
1st trimester
<90th percentile, N(%): 87 (49.7%):
3244 (52)
> 90th percentile, N(%): 7 (4.0%): 2983
(232)
P-Value for mean birth weight for
> 90th percentile PM10 vs. for <90th
percentile PM10
Adjusted: 0.289
Adjusted, with CO: 0.344
Adjusted, with N02: 0.641
Adjusted, with S02: 0.293
2nd trimester
<90th percentile, N(%): 82 (48.0%):
3243 (55)
> 90th percentile, N(%): 11 (6.4%):
3185(207)
P-Value for mean birth weight for
> 90th percentile PM10 vs. for <90th
percentile PM10
Adjusted: 0.790
Adjusted, with CO: 0.783
Adjusted, with N02: 0.707
Adjusted, with S02: 0.733
3rd trimester
<90th percentile, N(%): 90 (50.3%):
3239 (53)
> 90th percentile, N(%): 9 (5.0%): 2944
(198)
P-Value for mean birth weight for
> 90th percentile PM10 vs. for <90th
percentile PM10
Adjusted: 0.161
Adjusted, with CO: 0.279
Adjusted, with N02: 0.134
Adjusted, with S02: 0.150
Genotype Ncol HeVal/ValVal
1st trimester
<90th percentile, N(%): 71 (40.6%):
3262 (56)
> 90th percentile, N(%): 10(5.7%):
2773(171)
P-Value for mean birth weight for
> 90th percentile PM,o vs. for <90th
percentile PM10
Adjusted: 0.009
Adjusted, with CO: 0.031
Adjusted, with N02: 0.058
Adjusted, with S02: 0.010
2nd trimester
<90th percentile, N(%): 71 (41.5%):
3264(61)
> 90th percentile, N(%): 7 (4.1%): 2862
(208)
P-Value for mean birth weight for
> 90th percentile PM,o vs. for <90th
percentile PM10
Adjusted: 0.076
Adjusted, with CO: 0.093
Adjusted, with N02: 0.063
Adjusted, with S02: 0.061
3rd trimester
<90th percentile, N(%): 72 (40.2%):
3207 (58)
> 90th percentile, N(%): 8 (4.5%): 3262
(180)
P-Value for mean birth weight for
> 90th percentile PM,o vs. for <90th
percentile PM10
Adjusted: 0.777
Adjusted, with CO: 0.607
December 2009 E-473
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Adjusted, with N02: 0.843
Adjusted, with S02: 0.791
Reference: Tsai et al. (2006, 0983121
Period of Study: 1994-2000
Location: Kaohsiung, Taiwan
Outcome: Post neonatal mortality
Age Groups: Infants more than 27
days and less than 1 yr
Study Design: Case-crossover study
N: 207 deaths
Statistical Analyses: Conditional
logistic regression
Covariates: Temperature, humidity
Dose-response Investigated? No
Statistical Package: SAS, version 8.2
Pollutant: PMi0
Averaging Time: 24 h
Mean(SD):81.45|jg/m3
Percentiles: 26th: 44.50
60th(Median): 79.20
76th:111.50
Range (Mm, Max): (20.50-232.00)
Monitoring Stations: 6
Copollutant:
S02
N02
CO
0,
PM Increment: 67.00 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
OR = 1.040 (0.340-3.177)
Note: Air pollution levels at the dates of
infant death were compared with air
pollution levels 1 week before and 1
week after death
A cumulative lag up to 2 previous days
was used to assign exposure.
Reference: Wilhelm and Ritz (2006,
Period of Study: 1994-2000
Location: Los Angeles County,
California, U.S.
Outcome: Term low birth weight (LBW)
(<2500 g at > 37 completed wk
gestation), Vaginal birth <37 completed
wk gestation
Age Groups: LBW: 2 37 completed wk
Preterm births: <37 completed wk
Study Design: Cross-sectional
N: For LBW: 136,134
Forpreterm birth:
106,483
Statistical Analyses: Logistic
regression
Covariates: Maternal age, maternal
race, maternal education, parity, interval
since previous live birth, level of
prenatal care, infant sex, previous LBW
or preterm infant, birth season, other
pollutants (CO, N02, 03, PM10),
gestational age (in birth weight analysis)
Dose-response Investigated? Yes
Statistical Package: NR
Pollutant: PM,0
Averaging Time:
24 h (every 6 days)
Entire pregnancy
Trimesters of pregnancy
Months of pregnancy
6 wk before birth
Mean (SD): First trimester: 42.2
Third trimester: 41.5
6 wk before birth: 39.1
Range (Min, Max):
First trimester: 26.3, 77.4
Third trimester: 25.7, 74.6
6 wk before birth: 13.0, 103.7
Monitoring Stations:
Zip-code-level analysis: 8
Address-level analysis: 6
Copollutant (correlation):
First trimester: PMio-CO:r = 0.12
PM,o-N02:r = 0.29
PMio-03:r = -0.01
PMio-PM25:r = 0.43
Third trimester: PM10-CO: r = 0.32
PMio-N02:r = 0.45
PMio-03:r = -0.08
PMio-PM25:r = 0.52
6 wk before birth:
PM10-CO:r = 0.36
PM,o-N02:r = 0.49
PMio-03:r = -0.16
PM10-PM25:r = 0.60
PM Increment:
1)10|jg/m3
2) 3 levels:
a) <25 percentile (reference)
b) 25%-75 percentile
c) > 75 percentile
Incidence of LBW (third trimester
exposure)
<32.8|jg/m3:2.0(1.8, 2.2)
32.8to<43.4|jg/m3:2.0(1.9, 2.1)
> 43.4 pg/m3: 2.2 (2.0, 2.4)
Incidence of preterm birth (first
trimester exposure)
<32.9 pg/m3:8.7 (8.3, 9.2)
32.9to<43.9|jg/m3:8.8(8.5, 9.1)
> 43.9 pg/m3: 8.6 (8.1, 9.0)
Incidence of preterm birth (6 wk
before birth exposure)
<31.8|jg/m3:8.8(8.4, 9.3)
31.8to<44.1 pg/m3:8.6 (8.3, 8.9)
a44.1 pg/m3: 8.8 (8.4, 9.2)
Outcome: LBW
Exposure Period: Third trimester
Address-level analysis:
Single-pollutant model:
Distance Ł 1 mile
Per 10 pg/m3:1.22 (1.05,1.41)
33.4 to <44.7|jg/m3:1.08 (0.76, 1.52)
> 44.7 pg/m3:1.48 (1.00, 2.19)
Multipollutant model:
Distance < 1 mile
Per 10 pg/m3:1.36 (1.12,1.65)
33.4to<44.7|jg/m3:1.16 (0.77, 1.74)
> 44.7 pg/m3:1.58 (0.95, 2.62)
Single-pollutant model:
1 44.7 pg/m3: 0.96 (0.78, 1.18)
Multipollutant model:
1 44.7 pg/m3:1.02 (0.79, 1.32)
Single-pollutant model:
2 45.0 pg/m3:1.08 (0.97, 1.20)
Multipollutant model:
2 45.0 pg/m3:1.06 (0.93,1.21)
Zip-code-level analysis
Single-pollutant model:
December 2009
E-474
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
PeMOfjg/m°:1.03(0.97,1.09)
33.2 to <43.6|jg/m3:0.98 (0.86, 1.11)
> 43.6 pg/m3:1.03 (0.88, 1.21)
Multipollutant model:
Per 10 pg/m3:1.07 (0.99,1.15)
33.2 to <43.6|jg/m3: 0.97 (0.85, 1.12)
> 43.6 pg/m3:1.09 (0.90,1.31)
Outcome: LEW
Exposure Period: Entire pregnancy
period
Address-level analysis:
Multipollutant model:
Per 10 pg/m3:1.24 (0.91,1.70)
Outcome: Preterm Birth
Exposure Period: First trimester of
pregnancy
Address-level analysis:
Single-pollutant model:
Distance < 1 mile
Per 10 pg/m3:1.00 (0.93,1.09)
33.3to<45.1 pg/m3:1.07 (0.90, 1.26)
> 45.1 pg/m3:1.12 (0.91,1.38)
Multipollutant model:
Distance Ł 1 mile
Per 10 pg/m3:1.00 (0.90,1.12)
33.3to<45.1 |jg/m3:1.12(0.92, 1.36)
a45.1 pg/m3:1.17(0.90,1.50)
Single-pollutant model:
1 45.3 pg/m3:1.07 (0.97, 1.19)
Multipollutant model:
1 45.3 pg/m3:1.13 (1.00,1.27)
Single-pollutant model:
2 45.5 pg/m3:1.02 (0.96, 1.07)
Multipollutant model:
2 45.5 pg/m3: 0.94 (0.89, 1.01)
Zip-code-level analysis
Single-pollutant model:
Per 10 pg/m3: 0.99 (0.96,1.01)
33.3 to <44.2pg/m3:1.01 (0.95, 1.08)
> 44.2 pg/m3: 0.98 (0.90, 1.05)
Multipollutant model:
Per 10 pg/m3: 0.99 (0.96,1.03)
33.3 to <44.2pg/m3:1.03 (0.97, 1.11)
> 44.2 pg/m3:1.01 (0.92,1.11)
Outcome: Preterm birth
Exposure Period: 6 wk before birth
Address-level analysis:
Single-pollutant model:
Distance Ł 1 mile
Per 10 pg/m3:1.02 (0.95,1.10)
32.5to<44.8|jg/m3:1.09 (0.92, 1.29)
> 44.8 pg/m3:1.12 (0.92, 1.37)
Multipollutant model:
Distance Ł 1 mile
Per 10 pg/m3:1.06 (0.97,1.16)
32.5 to <44.8|jg/m3:1.09 (0.90, 1.31)
> 44.8 pg/m3:1.17 (0.91,1.49)
Single-pollutant model:
1 45.3 pg/m3: 0.99 (0.89, 1.10)
Multipollutant model:
December 2009 E-475
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
1 45.3 pg/m3:1.02 (0.91,1.16)
Single-pollutant model:
2 45.3 pg/m3: 0.98 (0.93, 1.03)
Multipollutant model:
2 45.3 pg/m3: 0.98 (0.92, 1.04)
Zip-code-level analysis
Single-pollutant model:
Per 10 pg/m3:1.02 (0.99,1.04)
32.1 to <44.3|jg/m3:1.01 (0.95, 1.07)
> 44.3 pg/m3:1.04 (0.96, 1.12)
Multipollutant model:
Per 10 pg/m3:1.02 (0.99,1.06)
32.1 to <44.3|jg/m3:1.02 (0.95, 1.09)
> 44.3 pg/m3:1.04 (0.95, 1.14)
Notes: multipollutant model adds
CO,N02, and 03 in addition to the main
pollutant of interest, PM10.
Reference: Woodruff etal. (1997,
0842711
Period of Study: 1989-1991
Location: 86 Metropolitan Statistical
Areas in the U.S. (counties with
populations less than 100,000 were
excluded)
Outcome: Postneonatal mortality
(death of an infant between 1 month
and 1 yr of age
1)AII post neonatal deaths
2) Normal birth weight (NBW, > 2500 g)
SIDS deaths
3) NBW respiratory deaths
4) Low birth weight (LBW) respiratory
death
Respiratory deaths: ICD9 codes
460-519
SIDS: ICD9 code 798.0
Age Groups: Infants (1 month-1yrof
age)
Study Design: Cross-sectional
N: 3,788,079 infants
Statistical Analyses: Logistic
regression
Covariates: Maternal education,
maternal race, parental marital status,
maternal smoking during pregnancy
Avg temperature during the first 2 mo of
life
Infant's month and yr of birth
Assessed race as an effect modifier
(p-val for interaction terms >0.2)
Dose-response Investigated? Yes
Statistical Package: NR
Pollutant: PM,0
Averaging Time: Mean of 1st 2 mo of
life
analyzed as tertiles of exposure and as
continuous exposure
Mean (SD): 31.4 (7.8)
Range (Min, Max):
Overall: 11.9-68.8
Low category: <28.0
Medium category: 28.1-40.0
High category: >40.0
Monitoring Stations: NR
PM Increment: 10 pg/m (for
continuous exposure analysis)
Adjusted ORs for cause-specific
post neonatal mortality by pollution
category (tertiles)
All causes
Low: Ref
Medium: 1.05 (1.01,1.09)
High: 1.10 (1.04,1.16)
SIDS, NBW:
Low: Ref
Medium: 1.09 (1.01,1.17)
High: 1.26 (1.14,1.39)
Respiratory death, NBW:
Low: Ref
Medium: 1.08 (0.87,1.33)
High: 1.40 (1.05,1.85)
Respiratory death, LBW:
Low: Ref
Medium: 0.93 (0.73,1.18)
High: 1.18 (0.86,1.61)
All other causes:
Low: Ref
Medium: 1.03 (0.97,1.08)
High: 0.97 (0.90, 1.04)
Adjusted ORs for a continuous
10 ug/m3 change in exposure
All causes: 1.04 (1.02,1.07)
SIDS, NBW: 1.12 (1.07, 1.17)
Respiratory death, NBW: 1.20 (1.06,
1.36)
Respiratory death, LBW: 1.05(0.91,
1.22)
All other causes: 1.00 (0.99,1.00)
December 2009
E-476
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Woodruff etal. (2008,
0983861
Period of Study: 1999-2002
Location: U.S. counties with >250,000
residents (96 counties)
Outcome: Postneonatal deaths
Respiratory mortality (ICD10: JOOO-99,
plus bronchopulmonary dysplasia [BPD]
P27.1)
SIDS(ICD10:R95)
Ill-defined causes (R99);
All other deaths evaluated as a control
category
Age Groups: Infants aged >28 days
and <1 yr
Study Design: Cross-sectional
N: 3,583,495 births (6,639 post
neonatal deaths)
Statistical Analyses: Logistic GEE
(exchangeable correlation structure)
Covariates: Maternal race/ethnicity,
marital status, age, education,
primiparity, county-level poverty and per
capita income levels, yr and month of
birth dummy variables to account for
time trend and seasonal effects, and
region of the country
Sensitivity analyses performed among
only those mothers with smoking
information (adjustment for smoking
had no effect on the estimates)
Season: Adjusted for yr and month of
birth dummy variables to account for
time trend and seasonal effects
Dose-response Investigated?
Evaluated the appropriateness of a
linear form from analysis based on
quartiles of exposure and concluded
that linear form was appropriate (data
not shown)
Statistical Package: SAS
Pollutant: PM,0
Averaging Time: Measured
continuously for 24 h once every 6 days
exposure assigned by calculating avg
concentration of pollutant during first 2
mo of life
Median and IQR (25th-75th
percentile):
Survivors: 28.9 (23.3-34.4)
All causes of death: 29.1 (23.9-34.5)
Respiratory: 29.8 (24.3-36.5)
SIDS: 28.6 (23.5-33.8)
SIDS+ill-defined: 28.8 (23.9-33.9)
Other causes: 29.2 (23.9-34.5)
Percent! les: see above
PM Component: Not assessed, but
controlled for region of the country to
account for PM composition variation
Monitoring Stations: NR
Copollutant (correlation):
PM10
PM25(r = 0.34)
CO (r = 0.18)
S02(r = 0.00)
03(r = 0.20)
Notes: Monthly avg calculated if there
were at least 3 available measures for
PM
Assigned exposures using the avg
concentration of the county of residence
PM Increment: IQR (11 pg/rri)
Effect Estimate [Lower Cl, Upper Cl]:
Adjusted ORs for single pollutant
models
All causes: 1.04 (0.99,1.10)
Respiratory: 1.18 (1.06,1.31)
SIDS: 1.02 (0.89, 1.16)
Ill-defined + SIDS: 1.06 (0.97,1.16)
Other causes: 1.02 (0.96,1.07)
Adjusted ORs for multipollutant models
(including CO, 03, S02)
Respiratory: 1.16 (1.04,1.30)
SIDS: 1.02 (0.90, 1.16)
OR for deaths coded as BPD per
increase in IQR: 1.19 (0.85,1.65)
OR for respiratory post neonatal death
stratified by birth weight
NBW only: 1.19 (1.05,1.36)
LBWonly: 1.12 (0.95,1.31)
OR for respiratory deaths removing
region of U.S. as a confounding
variable: 1.30 (1.04,1.61)
OR for respiratory deaths assessing
exposure as quartiles
Highest vs. Lowest quartile: 1.31 (1.00,
1.71)
OR for respiratory deaths among only
those deaths that occurred during the
first 90 days (most closely matched
exposure metric of the avg over the first
2 mo of life): 1.25 (1.06,1.47)
December 2009
E-477
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Jedrychowski, et al., (2007,
1566071
Period of Study: Jan 2001-Feb 2004
Location: Krakow, Poland
Outcome: Birth weight (grams), birth
length (cm)
Age Groups: Pregnant women 18-35
yr
Study Design: Prospective cohort
N: 493 women
Statistical Analyses: Linear regression
Covariates: Environmental tobacco
smoke (# cigarettes smoked daily in
presence of pregnant woman), season
of birth, size of mother, parity,
gestational age, gender of child, vitamin
A intake
Season: All
Dose-response Investigated? Yes
Statistical Package: NR
Lags Considered: Two consecutive
days in the second trimester
Pollutant: PM25
Averaging Time: 48 h period
Percentiles: 60th(Median): 35.3
Range (Mm, Max): 10.3, 294.9
Monitoring Stations: No stations,
personal monitoring
Notes: PM measured during a 2 day
period in the second trimester by
Personal Environmental Monitoring
Sampler (PEMS)
PM Increment: in 1 pg/rrf and tertiles
T1:<27.0|jg/m3
T2: 27.0-46.2 pg/m3
T3:a46.2fjg/nr
Mean [Lower Cl, Upper Cl]:
Birth weight (g)
For In unit PM:|3 = -172.39 (p = 0.02)
Tertiles:
T1:ref
T2:|3 =-16.510 [-94.630, 61.610]
T3:|3 =-109.956 [-196.649 to-23.263]
In low Vitamin A group (<1,378 pg)
T1:ref
T2:|3 =-68.354 [-165.643, 28.935]
T3: (3 = -185.070 [-293.393 to -76.747]
In high Vitamin A group (>1,378 pg)
T1: ref
T2: (3 = 64.262 [-70.464, 198.988]
T3: (3 = 38.593 [-109.853, 187.039]
Birth length (cm)
For In unit PM: (3 = -1.39 (p = 0.00)
Tertiles:
T1:ref
T2: (3 = -0.288 [-0.790, 0.214]
T3: (3 = -0.810 [-1.367 to-0.253]
In low Vitamin A group (<1,378 |jg)
T1:ref
T2: (3 = -0.514
T3: (3 = -1.100
-1.114,0.086]
-1.768 to-0.432]
In high Vitamin A group (>1,378 |jg)
T1:ref
T2: (3 = 0.039 [-0.896, 0.974]
T3: (3 = -0.301 [-1.326,0.724]
Reference: (Lipfert et al., 2000,
0041031
Period of Study: 1990
Location: U.S.
Outcome (ICD9 and ICD10): Infant
mortality
Including respiratory mortality
(traditional definition, ICD9 460-519),
expanded definition (adds ICD9 769
and 770)
Age Groups: Infants
Study Design: Cross-sectional
N: 2,413,762 infants in 180 counties
(Ns differ for various models)
Statistical Analyses: Logistic
regression
Covariates: Mother's smoking,
education, marital status, and race
Month of birth
And county avg heating degree days
Dose-response Investigated? NR
Statistical Package: NR
Pollutant: S04 7NSPM,0 (regressed
jointly)
Averaging Time: Yearly avg used
Mean (SD): 33.1 (9.17) (based on 180
counties)
Range (Min, Max): (16.9, 59)
Monitoring Stations: NR
Copollutant:
PM10
NSPM,o
CO
S02
Notes: TSP-based sulfate was adjusted
for compatibility with the PM10-based
data
PM Increment: NR (present regression
coefficients)
Effect Estimate [Lower Cl, Upper Cl]:
Presented regression coefficients
(standard errors)
(3 PM exposures regressed jointly)
bold = p O.05
Cause of death: All
Birth weight: All
S042":-0.0002 (0.0061)
NSPM10: 0.0115 (0.0014)
Cause of death: All
Birth weight: LBW
S042": 0.0265 (0.0080)
NSPM,0: 0.0086 (0.0020)
Cause of death: All
Birth weight: normal
S042": -0.0488 (0.0098)
NSPM10: 0.0096 (0.0024)
Cause of death: All neonatal
Birth weight: All
S042": 0.0267 (0.0076)
NSPM10: 0.0126 (0.0018)
Cause of death: All neonatal
Birth weight: LBW
S042": 0.0388 (0.0088)
NSPM,0: 0.0093 (0.0022)
Cause of death: All neonatal
Birth wt: normal
S042":-0.0334 (0.0169)
NSPM10: 0.0125 (0.0040)
Cause of death: All post neonatal
Birth wt: All
PM,,: 0.0091 (0.0024)
S04 : -0.0474 (0.0100)
NSPM,0: 0.0096 (0.0024)
Cause of death: All post neonatal
Birth wt: LBW
S042":-0.0247 (0.0173)
NSPM10: 0.0101 (0.0042)
Cause of death: All post neonatal
Birth wt: normal
S042":-0.0569 (0.0121)
December 2009
E-478
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
NSPM10: 0.0080 (0.0029)
Cause of death: SIDS
Birth weight: All
S042":-0.1078 (0.0151)
NSPM,0: 0.0149 (0.0037)
Cause of death: SIDS
Birth weight: LBW
S042":-0.1378 (0.0337)
NSPM10: 0.0146 (0.0085)
Cause of death: SIDS
Birth weight: normal
PIVU 0.0137 (0.0042)
S04:-0.0995 (0.0168)
NSPM,0: 0.0147 (0.0041)
Cause of death: All respiratory (ICD9:
460-519, 769, 770)
Birth weight: All
S042": 0.0706 (0.0146)
NSPM10: 0.0166 (0.0034)
Cause of death: All respiratory (ICD9:
460-519, 769, 770)
Birth weight: LBW
S042": 0.0821 (0.0158)
NSPM10: 0.0139 (0.0038)
Cause of death: All respiratory (ICD9:
460-519, 769, 770)
Birth weight: normal
PM,n: 0.0177 (0.0091)
S04 :0.0001 (0.0392)
NSPM10: 0.0118 (0.0090)
Cause of death: Respiratory disease
(ICD9: 460-519)
Birth weight: All
PM,,: 0.0133 (0.0089)
S04 :0.0093 (0.0384)
NSPM10: 0.0134 (0.0089)
Cause of death: Respiratory disease
(ICD9: 460-519)
Birth weight: LBW
PM,n: 0.0092 (0.0137)
S04 :0.0434 (0.0580)
NSPM10: 0.0089 (0.0138)
Cause of death: Respiratory disease
(ICD9: 460-519)
Birth weight: normal
S042":-0.0177 (0.0509)
NSPMy 0.0128 (0.0119)
Associations with SIDS by smoking
status
Smoking status: Yes
Birth weight: Normal
S042": -0.0722 (0.0284)
NSPM,0: 0.0206 (0.0071)
Smoking status: No
Birth weight: Normal
S042":-0.114 (0.021)
NSPM10: 0.0117 (0.005)
Smoking status: Yes
Birth weight: LBW
S042": -0.0958 (0.0483)
NSPM10: 0.0345 (0.0125)
Smoking status: No
Birth weight: LBW
S042": -0.0172 (0.047)
NSPM,0: -0.0007 (0.012)
Mean risks (95%CI) between post
neonatal SIDS among normal birth
weight babies
pollutants regressed one at a time
S042": 0.43 (0.37, 0.51)
NSPM10:1.33 (1.18, 1.50)
December 2009 E-479
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: (Liu et al, 2007, 0904291
Period of Study: 1985-2000
Location: 3 Canadian cities: Calgary,
Edmonton, and Montreal
Outcome: Intrauterine growth
restriction (IUGR)
Age Groups: Singleton term live births
(37-42 wks gestation)
Study Design: Retrospective cohort
N: 386,202 singleton live births
Statistical Analyses: Multiple logistic
regression
Covariates: Maternal age, parity, infant
gender, season, and city of residence at
time period of birth
Season: All seasons
Dose-response Investigated? No
Statistical Package: NR
Pollutant: PM,
PM Increment: 10 pg/m
Averaging Time: 24 h (6-day schedule) Effect Estimate
Mean (SD): 12.2
Percentiles: 26th: 6.3
60th(Median): 9.7
76th:15
PM Component: metals and organic
matter such as polycyclic aromatic
hydrocarbons
Monitoring Stations: Calgary (4),
Edmonton (2), and Montreal (8)
Copollutant (correlation):
S02: r = 0.44, p< 0.0001
N02: r = 0.41, p< 0.0001
CO: r = 0.31, p< 0.0001
03: r =-0.14, p< 0.0001
Single-pollutant model [Lower Cl,
Upper Cl]:
1st trimester
OR = 1.07 (1.03-1.10)
2nd trimester
OR = 1.06 (1.03-1.10)
3rd trimester
OR = 1.06 (1.03-1.10)
Effect Estimate
multi-pollutant model [Lower Cl,
Upper Cl]:
1st trimester
OR= 1.03 (0.99-1.06)
2nd trimester
OR=1.01 (0.98-1.05)
3rd trimester
OR= 1.03 (0.99-1.06)
Note: ORs and CIs estimated from Fig.
6 and 7
Reference: Loomisetal. (1999,
0872881
Period of Study: Jan 1993-Jul 1995
Location: Mexico City (southwestern
section)
Outcome (ICD9 and ICD10): Infant
mortality (daily counts of deaths)
All ICD9 codes, excluding accidents,
poisoning, and violence (ICD9 >800)
Age Groups: Children <1 yrof age
Study Design: Time-series
N: 942 deaths (days were the unit of
observation)
Statistical Analyses: Poisson
regression (generalized additive model)
Covariates: Final models controlled for
mean temp of 3 days before death and
nonparametrically smoothed periodic
cycles
Season: Yes (considered)
Dose-response Investigated? Loess
smoother
Statistical Package: NR
Lags Considered: 0-5 (also
considered lags with avg exposure
levels during "windows" of 2 to 4 days)
Pollutant: PM25
Averaging Time: 24-h
Mean (SD): 27.4 (10.5)
Percentiles: Lower quartile: 20
Median: 26
Upper quartile: 34
Range (Min, Max): 4, 85
Monitoring Stations: 1
Copollutant:
03
N02
NO
NOX
S02
Notes: Pearson correlation coefficients
ranging from 0.52 to 0.71
PM Increment: 10 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
%Change in infant mortality
Lags 0-5 (single day) presented in Fig
LagO,1,2: No association (results not
presented)
Lag3: 4.8 (0.97, 8.61)
Lag4: 4.2 (0.37, 7.93)
%Change in mortality when avg
exposure levels during "windows" of 2
to 4 days were considered
2 Days:
No lag:-1.36 (-5.51, 2.8)
Lag 1:-0.95 (-5.10, 3.20)
Lag2: 2.78 (-1.33, 6.89)
Lag3: 4.93 (0.86, 9.01)
3 Days:
No lag: -0.81 (-5.29, 3.67)
Lag 1:1.99 (-2.46, 6.45)
Lag2: 4.54 (0.12, 8.96)
Lag3: 6.87 (2.48, 11.26)
4 Days:
No lag: 1.95 (-2.76, 6.66)
Lag 1:3.74 (-0.95, 8.42)
Lag2: 5.87 (1.21, 10.53)
Multipollutant models (3-day mean w/ 3-
day lag)
1 pollutant model:
6.87(2.48,11.26)
2 pollutant models:
w/03: 6.24 (1.35, 11.14)
w/N02:5.91 (-0.76, 12.59)
3 Pollutant model (w/ 03 and N02):
6.30 (-0.54, 13.15)
December 2009
E-480
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Mannes et al. (2005,
0878951
Period of Study: Jan 1998-Dec 2000
Location: metropolitan Sydney,
Australia
Outcome: Risk of small for gestational
age (SGA) and birth weight
Age Groups: All singleton births >20
wk and > 400 grams birth weight and
maternal all ages
Study Design: Cross-sectional
N: 138,056 singleton births
Statistical Analyses: Logistic and
linear regression models
Covariates: Sex of child, maternal age,
gestational age, maternal smoking,
gestational age at first antenatal visit,
maternal indigenous status, whether
first pregnancy, season of birth, and
socioeconomic status (SES)
Season: All seasons
included as covariate.
Dose-response Investigated? No
Statistical Package: SAS System for
Windows v8.02
Pollutant: PM25
Averaging Time: 24 h
Mean (SD): 9.4 (5.1)
Percentiles: 26th: 6.5
60th(Median): 8.4
76th: 11.2
Range (Min, Max): (2.4-82.1)
Monitoring Stations: up to 14
Copollutant (correlation):
CO: r = 0.53
N02:r = 0.66
03:r = 0.36
PM,o:r = 0.81
PM Increment: 1 pg/m
Risk of SGA
All births
1 month before birth:
OR = 1.01 (0.99-1.03)
Third trimester: OR = 0.99 (0.97-1.02)
Second trimester:
OR = 1.03 (1.01-1.05)
First trimester: OR = 0.99 (0.97-1.01)
5 km births
1 month before birth:
OR = 1.01 (0.97-1.04)
Third trimester: OR = 1.00 (0.95-1.05)
Second trimester:
OR = 1.00 (0.96-1.05)
First trimester: OR = 0.99 (0.94-1.04)
Change in birth weight
All births
1 month before birth:
IS = -2.48 (-4.58--0.38)
Third trimester: IS = -0.98 (-3.74-1.78)
Second trimester:
IS = -4.10 (-6.79--1.41)
First trimesters = 0.36 (-2.29-3.01)
5 km births
1 month before birth:
IS = -2.70 (-6.80-1.40)
Third trimester: IS = -2.83 (-9.00-3.34)
Second trimester: IS = 1.54 (-4.59-7.67)
First trimester: IS =1.89 (-1.99-5.77)
Reference: Parker et al. (2005,
0874621
Period of Study: 1999-2000
Location: California
Outcome: Small for gestational age
(SGA) and birth weight
Age Groups: Infants delivered at 40 wk
gestation
maternal all ages
Study Design: Cross-sectional
N: 18,247 singleton births
Statistical Analyses: Linear and
logistic regression models
Covariates: Maternal race, maternal
Hispanic origin, marital status, parity,
maternal education, and maternal age
Season: Season of delivery (covariate)
Dose-response Investigated? Yes
Statistical Package: STATA
Pollutant: PM25
Averaging Time: NR (measurement
taken every 6 days)
Mean (SD): 15.42 (5.08)
PM Component: metals, polycyclic
aromatic hydrocarbons
Monitoring Stations: 40
Copollutant (correlation):
PM25-CO:r = 0.6
Notes: Mean calculated for 9-month
exposure. The following means (SDs)
are calculated for trimester:
First: 15.70 (6.26)
Second: 15. 40 (6. 53)
Third: 14.29 (6.35)
PM categorized into quartiles:
02:11.9-13.9
03: 13.9-18.4
04: >18.4
PM Increment: <11.9 pg/m |
Referent PM Increment: 11.9-
13.9|jg/m3
Effect Estimate [Lower Cl, Upper Cl]:
First Trimester
Birth weight: IS = -5.7 (-27.9-16.5)
SGA: OR =1.02 (0.84-1.23)
Second Trimester
Birth weight: IS =11.3 (-12.2-34.9)
SGA: OR = 0.89 (0.73-1.09)
Third Trimester
Birth weight: IS = 8.3 (-13.1-29.8)
SGA: OR =1.00 (0.83-1.19)
PM Increment: 13.9-18.4 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
First Trimester
Birth weight: IS = -2.5 (-24.5-19.5)
SGA: OR =1.12 (0.93-1.34)
Second Trimester
Birth weight: IS = -17.2 (-39.4-4.9)
SGA: OR =1.05 (0.88-1.26)
Third Trimester
Birth weight: IS = -8.1 (-30.2-13.9)
SGA: OR = 0.98 (0.82-1.181
PM Increment: >18.4 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
First Trimester
Birth weight: IS = -35.8 (-58.4--13.3)
SGA: OR =1.26 (1.04-1.51)
Second Trimester
Birth weight: IS = -46.6 (-68.6- -24.6)
SGA: OR =1.24 (1.04-1.49)
Third Trimester
Birth weight: IS = -31.6 (-52.0- -11.1)
SGA: OR =1.21 (1.02-1.43)
December 2009
E-481
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Parker and Woodruff
(2008,156846)
Period of Study: 2001-2003
Location: U.S.
Outcome: Low birth weight
Study Design: Cohort
Pollutant: PM25
Averaging Time: 9 mo
N: 785,965 Singleton births delivered at Mean (SD): 14.5
40 wk gestation
a 25th: 12.1
Statistical Analyses: GEE regression
models
linear and logistic regression
Covariates: Race/ethnicity, parity,
maternal age
Season: Season of delivery
Statistical Package: SUDAAN
75th: 17.6
Copollutant (correlation):
S02, N02 CO 03
PM Increment: 10 pg/m
Change in Birth weight (9 month
exposure):
Unadjusted: 19.4 (9.8, 29.0)
Adjusted for maternal factors:
18.4(9.2,27.7)
Stratified by region:
Industrial Midwest: -15.3 (-43.4,12.9)
Northeast:-9.8 (-11.9, 26.6)
Northwest: 27.5 (5.5, 49.4)
Southern CA: 5.5 (-9.6, 20.5)
Southeast: 7.3 (-11.9, 26.6)
Southwest: 72.3 (34.0,110.5)
Upper Midwest: -0.7 (-62.0, 60.6)
Multipollutant models:
PM25+PMio.25:14.2 (4.3, 24.1)
PM25+PM10.25+S02+CO+N02+03: 28.6
(14.2, 43.0)
December 2009
E-482
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Rich et al. (2009,1801221
Period of Study: 1999-2003
Location: New Jersey, United States
Outcome: Small for gestational age
Study Design: Retrospective Cohort
Covariates: Month and calendar yr of
birth, apparent temperature, pregnancy
complications
Statistical Analysis: Polytomous
logistic regression
Statistical Package: SAS
Age Groups: Gestational age 37-42
wks
Pollutant: PM25
Averaging Time: 24 h
Mean (SD) Unit:
*AII values are for first trimester, other
trimesters are available in paper
Reference Births: 13.8(2.5)
SGA Births: 13.9 (2.5)
VSGA Births: 13.9 (2.4)
Range (Min, Max): 2.0, 29.0
Copollutant (correlation):
*AII values are for first trimester, other
trimesters are available in paper
N02: 0.01
S02:0.17
CO: 0.25
*AII values are for first trimester, other
trimesters are available in paper
Increment: 4 pg/m3
Percent Change in Risk (96% Cl)
SGA: 4.5 (0.5-8.7)
VSGA: 2.6 (-4.4-10.0)
Percent Change in Risk (96% Cl) for
single and two-pollutant models
Single, SGA: 4.6 (-0.3-9.8)
Single, VSGA: 4.5 (-4.0-13.4)
Two (PM25 SN02), SGA: 4.5 (-0.4-9.7)
Two (PM25 & N02), VSGA: 3.2 (-5.2-
12.4)
Percent Change in Risk (96% Cl) by
pregnancy complication in third
trimester
SGA
Any Complication
No: 4.7 (0.6-9.0)
Yes: 2.2 (-6.1-11.3)
Placental Abruption
No: 4.0 (0.3-7.9)
Yes: 11.7 (-21.7-59.5)
Placental Praevia
No: 3.9 (0.2-7.8)
Yes: 23.2 (-20.9-91.9)
Pre-eclampsia
No: 4.2 (0.4-8.2)
Yes: 2.7 (-13.8-22.3)
Gestational Hypertension
No: 4.3 (0.4-8.4)
Yes: 3.9 (-7.8-17.1)
Premature Rupture of the Membrane
No: 3.7 (-0.1-7.7)
Yes: 14.6 (-3.3-35.9)
Gestational Diabetes
No: 4.6 (0.8-8.6)
Yes: -9.3 (-24.7-9.3)
VSGA
Any Complication
No: 1.5 (-6.1-9.7)
Yes: 12.6 (0.1-26.7)
Placental Abruption
No: 4.1 (-2.6-11.2)
Yes: 7.6 (-29.8-64.9)
Placental Praevia
No: 4.1 (-2.5-11.2)
Yes: 3.2 (-43.0-86.9)
Pre-eclampsia
No: 4.4 (-2.6-11.9)
Yes: 3.9 (-15.7-28.1)
Gestational Hypertension
No: 3.2 (-4.0-10.9)
Yes: 12.9 (-3.3-31.9)
Premature Rupture of the Membrane
No: 3.3 (-3.5-10.5)
Yes: 21.9 (-3.6-54.2)
Gestational Diabetes
No: 4.3 (-2.5-11.5)
Yes: 1.4 (-27.0-40.9)
December 2009
E-483
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Ritz et al. (2007, 0961461
Period of Study: Jan 2003-Dec 2003
Location: Los Angeles, California
Outcome: Preterm births (infants
delivered before 37 wk)
Age Groups: Births
Study Design: Case-control nested
within a birth cohort (cases and controls
matched on zip code and birth month)
Phase 1: cross-sectional including all
birth cohort
Phase 2: nested case-control of survey
respondents
N: Phase 1: Birth cohort consisted of
58,316 eligible births. Phase II: 2,543
Statistical Analyses: Logistic
regression
Covariates: Birth certificate
information: maternal age,
race/ethnicity, parity, education, season
of birth
survey information: maternal smoking,
alcohol consumption, living with a
smoker, and marital status during
pregnancy
income (imputed)
occupation and pregnancy weight gain
considered but not included in final
models
Season: Yes
Dose-response Investigated? Yes,
examined categories of exposure
Statistical Package: NR
Pollutant: PM25
Averaging Time: daily or every 3rd day
used to calculate the entire pregnancy,
the first trimester, and the last 6 wk
before delivery
Only reported first trimester exposures
forPM
Range (Min, Max): NR
Ranges for 3 categories reported:
Low (ref):< 18.63
Mid: 18.64-21.36
High: >21.36
Monitoring Stations: Each zip code
was linked to the nearest monitoring
station (number not reported)
Copollutant (correlation):
CO
N02
03
Notes: Daily or every 3rd day
measurements used for mean
calculations
PM Increment: Reported analyses
using exposure categories
Effect Estimate [Lower Cl, Upper Cl]:
Birth cohort (phase I)
Crude: Low: 1.0
Mid: 0.96 (0.90, 1.03)
High: 1.05 (0.99,1.12)
Adj for birth cert Covariates: Low: 1.0
Mid: 1.01 (0.93, 1.09)
High: 1.10 (1.01,1.20)
Survey respondents (phase II)
Crude: Low: 1.0'Mid: 1.11 (0.90,1.36)
High: 1.27 (1.06, 1.53)
Adj for birth cert Covariates: Low: 1.0
Mid: 1.14 (0.90, 1.46)
High: 1.27 (0.99, 1.64)
Adj for all Covariates: Low: 1.0
Mid: 1.15 (0.90, 1.47)
High: 1.29 (1.00,1.67)
Two-phase model: * Low: 1.0
Mid: 0.98 (0.84, 1.15)
High: 1.07 (0.85, 1.35)
'Method to reduce potential selection
bias and increase statistical efficiency
Reference: Slama et al. (2007, 0932161
Period of Study: Jan 1998-Jan 1999
Location: Munich, Germany
Outcome: Birth weight offspring at term
Study Design: Cohort study
N: 1016 births
Statistical Analyses: Poisson model
Covariates: Maternal passive smoking,
maternal age, gestational duration, sex
of child, parity, maternal education,
maternal size, prepregnancy weight,
other pollutants (PM25, PM25
absorbance, N02), season of conception
Dose-response Investigated? Yes
Statistical Package: STATA
Pollutant: PM25 (estimated based on
larger PM size fractions)
Averaging Time: Entire pregnancy
period and trimesters
Mean (SD): 14.4
Percentiles: 26th: 13.5
60th(Median): 14.4
76th: 15.4
Monitoring Stations: Spatial
component: 40
Temporal component: 1
Copollutant (correlation):
p.a. = pregnancy avg
trim. = trimester
PM25(p.a.)-PM25 (1sttrim.): 0.85
PM25(p.a.)-PM25 (2nd trim.): 0.77
PM25 p.a.
PM25 p.a.
PM25 p.a.)-NQ, (1sttrim.): 0.18
-PM25 (3rd trim.): 0.87
-N02 (p.a.): 0.45
PM25 p.a.
PM25 p.a.
PM25 1sttrim.)-PM25 (2nd trim.): 0.40
-N02 (2nd trim.): 0.32
-N02 (3rd trim.): 0.37
1st trim.
PM25
PM25
PM25 (1sttrim.)-N02 (1sttrim.): 0.15
1st trim.
-PM25 (3rd trim.): 0.68
-N02 (p.a.): 0.48
1st trim.
1st trim.
-N02
2nd trim.): 0.41
PM25
PM25
PM25 (2ndtrim.)-PM25 (3rd trim.): 0.51
-N02
3rd trim.): 0.39
PM25 2nd trim.)-N02 (p.a.): 0.23
PM25 2nd trim.)-N02 (1sttrim.):-0.03
PM25 (2nd trim.)-N02 (2nd trim.): 0.17
PM25 (2ndtrim.)-N02 (3rd trim.): 0.30
PM Increment:
1 1 pg/m3
2 Quartiles:
a) 1st (reference) (7.2-13.5 pg/m3)
b) 2nd (13.5-14.4 pg/m3)
c) 3rd (14.4-15.4 pg/m3)
day) 4th (15.41-17.5 pg/m3)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
the whole pregnancy
Single-pollutant models
Unadjusted models
2nd quartile: 1.07 (0.65,1.73); 3rd
quartile: 1.38 (0.91, 2.09)
4th quartile: 1.45 (0.92, 2.25)
Perl pg/m3:1.06 (0.95,1.19)
Adjusted models
2nd quartile: 1.08 (0.63,1.82); 3rd
quartile: 1.34 (0.86, 2.13)
4th quartile: 1.73 (1.15, 2.69); Per
1 pg/m3:1.13 (1.00,1.29)
Multipollutant models
Adjusted models
2nd quartile: 1.01 (0.57,1.85)
3rd quartile: 1.12 (0.64,1.87)
4th quartile: 1.36 (0.72, 2.45); Per
1 pg/m3:1.07 (0.91,1.26)
Single-pollutant models (restricted
analysis to PM25 absorbance below the
median)
Perl pg/m3:1.15 (0.89,1.52)
Prevalence ratios (PRs) of birth
weight <3000 g
Multipollutant models (simultaneous
December 2009
E-484
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
PM25
PM25
PM25
PM25
PM25
0.69
PM25
PM25
PM25
PM25
PM25
0.27
PM25
0.53
PM25
0.51
PM25
PM25
0.08
PM25
0.29
PM25
0.41
PM25
PM25
0.48
PM25
0.36
PM25
0.37
(3rd trim.)-N02 (p.a): 0.39
3rd trim.)-N02 (1st trim.): 0.33
3rd trim.)-N02 (2nd trim.): 0.21
(3rd trim.)-N02 (3rd trim.): 0.23
(p.a.)- PM25 absorbance (p.a.):
(p.a.)-PM2.5abs (1st trim.): 0.33
-PM25abs(2ndtrim.):0.48
- PM25 abs (3rd trim.): 0.52
(1sttrim.)-PM25abs(p.a.j: 0.68
(1sttrim.)-PM25 abs (1sttrim.):
(1st trim.)-PM25 abs (2nd trim.):
(1sttrim.)-PM25 abs (3rd trim.):
2ndtrim.)-PM25abs(p.a.):0.41
2nd trim.)-PM25 abs (1st trim.):
(2nd trim.)- PM25 abs (2nd trim.):
(2nd trim.)- PM25 abs (3rd trim.):
(3rd trim.)-PM25 abs (p.a.): 0.62
(3rd trim.)-PM25 abs (1sttrim.):
(3rd trim.)-PM25 abs (2nd trim.):
(3rd trim.)- PM25 abs (3rd trim.):
Adjustment of 3rd trimester Plfeand
whole pregnancy PMu)
PM25 (whole pregnancy)
Perl pg/m3: 0.96 (0.75,1.19)
PM25 (3rd trimester)
Perl pg/m3:1.17 (0.98,1.40)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
the whole pregnancy (adjustment for
season of conception)
4th quartile: 1.68 (1.05, 2.75); Per
1 |jg/m3:1.12(0.97,1.28)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
first trimester of pregnancy
Each trimester separately
2nd quartile: 1.14 (0.74,1.96); 3rd
quartile: 1.28 (0.84, 2.10)
4th quartile: 1.65 (1.02, 2.60)
Perl pg/m3:1.10 (0.99,1.20)
All trimesters adjusted simultaneously
2nd quartile: 0.97 (0.60,1.73); 3rd
quartile: 0.98 (0.57,1.75)
4th quartile: 1.22 (0.71, 2.18)
Perl pg/m3:1.03 (0.90,1.17)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
second trimester of pregnancy
Each trimester separately
2nd quartile: 0.83 (0.52,1.32); 3rd
quartile: 1.08 (0.71,1.60)
4th quartile: 0.94 (0.61,1.47)
Perl pg/m3:1.01 (0.92,1.12)
All trimesters adjusted simultaneously
2nd quartile: 0.75 (0.46,1.24)
3rd quartile: 0.86 (0.56,1.30);
4th quartile: 0.75 (0.48,1.23)
Perl pg/m3: 0.94 (0.84,1.06)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
third trimester of pregnancy
Each trimester separately
2nd quartile: 1.30 (0.80, 2.17)
3rd quartile: 1.44 (0.85, 2.27)
4th quartile: 1.90 (1.20, 2.82)
Perl pg/m3:1.14 (1.02,1.24)
All trimesters adjusted simultaneously
2nd quartile: 1.34 (0.79, 2.30)
3rd quartile: 1.48 (0.86, 2.58)
4th quartile: 1.91 (1.00,3.20)
Perl pg/m3:1.14 (0.99,1.29)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
third trimester of pregnancy
(adjustment for season of
conception)
All trimesters adjusted simultaneously
Perl pg/m3:1.25 (1.04,1.50)
Sensitivity analysisjbootstrapped PR)
2nd quartile: 0.98 (0.63,1.61); 3rd
quartile: 1.22 (0.82, 2.02)
4th quartile: 1.57 (1.02, 2.57)
Perl |jg/m3:1.11 (0.98,1.27)
Estimated increments in prevalence
of birth weight of <3000 g during
exposure 9 mo after birth
Per 1 pg/m3: 7% (-7%, 22%)
Reference: (Slama et al., 2007,
0932161
Period of Study: Jan 1998-Jan 1999
Outcome: Birth weight offspring at term Pollutant: PM25 absorbance (estimated) ™ln*re(m|nt:
Study Design: Cohort study
Averaging Time: Entire pregnancy
period and trimesters
b
1st (reference) (1.29-1.61)
2nd (1.61-1.72)
December 2009
E-485
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Location: Munich, Germany
N: 1016 births
Statistical Analyses: Poisson model
Covariates: Maternal passive smoking,
maternal age, gestational duration, sex
of child, parity, maternal education,
maternal size, prepregnancy weight,
other pollutants (PM25, PM25
absorbance, N02), season of conception
Dose-response Investigated? Yes
Statistical Package: STATA
Mean (SD): 1.76*
Percentiles: 26th: 1.61*
50th(Median): 1.72*
75th: 1.89*
Unit(i.e. ug/m3): 10-5/m
Monitoring Stations:
Spatial component: 40
Temporal component: 1
Copollutant (correlation):
p.a. = pregnancy avg
trim. = trimester
abs = absorbance
PM25 abs (p.a.)-PM25 abs (1st trim.):
0.54
PM25 abs (p.a.)-PM25 abs (2nd trim.):
0.84
PM25 abs (p.a.)-PM25 abs (3rd trim.):
0.55
PM25abs(p.a.)-PM25(p.a.):0.69
PM25abs
PM2.5 abs
-PM2
-PM2;
1st trim.): 0.68
2nd trim.): 0.41
PM2.5 abs (p.a.)-PM2.5 (3rd trim.): 0.62
PM25abs
PM2.5 abs
-N02
-N02
p.a.): 0.67
1st trim.): 0.34
PM2.5 abs (p.a.)-N02 (2nd trim.): 0.63
p.a.)-N02 (3rd trim.): 0.36
1sttrim.)-PM25abs(2nd
PM25abs
PM25abs
trim.): 0.32
PM25abs(1sttrim.)-PM25abs(3rd
trim.): -0.26
PM25abs(1sttrim.)-PM25(p.a.):0.33
PM25abs 1st trim. -PM25 1st trim.):
0.27
PM25abs(1sttrim.)-PM25(2ndtrim.):
0.08
PM25abs(1sttrim.)-PM25(3rdtrim.):
0.48
PM25abs 1sttrim.)-N02 p.a.): 0.29
PM25abs 1sttrim.)-N02 1st trim.): 0.84
PM25abs(1sttrim.)-N02(2ndtrim.):
0.16
PM25abs(1sttrim.)-N02(3rdtrim.):-
0.39
PM25abs(2ndtrim.)-PM25abs(3rd
trim.): 0.31
PM25 abs (2nd trim.)-PM25 (p.a.): 0.48
PM25abs(2ndtrim.)-PM25(1sttrim.):
0.53
PM25 abs (2nd trim.)-PM25 (2nd trim.):
0.29
PM25 abs (2nd trim.)-PM25 (3rd trim.):
0.36
PM25abs
PM25abs
0.19
2nd trim,
2nd trim.
-N02 (p.a.): 0.61
-N02 (1st trim.):
PM25abs(2ndtrim.)-N02(2ndtrim.):
0.85
PM25abs(2ndtrim.)-N02(3rdtrim.):
0.17
PM25abs(3rdtrim.)-PM25(p.a.):0.52
PM25abs(3rdtrim.)-PM25(1sttrim.):
0.51
PM25abs(3rdtrim.)-PM25(2ndtrim.):
0.41
PM25abs(3rdtrim.)-PM25(3rdtrim.):
0.37
PM25abs(3rdtrim.)-N02(p.a.):0.40
PM25abs(3rdtrim.)-N02(1st
trim.):-0.34
PM25abs(3rdtrim.)-N02(2ndtrim.):
0.21
PM25abs(3rdtrim.)-N02(3rdtrim.):
0.88
c) 3rd (1.72-1.89)
day) 4th (1.89-3.10)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
the whole pregnancy
Single-pollutant models Unadjusted
models
2nd quartile: 1.19 (0.74,1.99)
3rd quartile: 1.56 (0.98, 2.50);
4th quartile: 1.52 (0.96, 2.46)
Per 0.5*10-5/m: 1.25 (0.90,1.70)
Adjusted models
2nd quartile: 1.21 (0.73,1.97)
3rd quartile: 1.63 (0.98, 2.57);
4th quartile: 1.78 (1.10,2.70)
Per0.5*10-5/m: 1.45(1.06,1.87)
Multipollutant models Adjusted models
2nd quartile: 1.19 (0.70, 2.01)
3rd quartile: 1.55 (0.80, 2.80);
4th quartile: 1.46 (0.67, 2.90)
Per0.5*10-5/m: 1.33(0.76, 2.38)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
the whole pregnancy (adjustment for
season of conception)
4th quartile: 1.72 (1.08, 2.73)
Per 0.5*10-5/m: 1.38 (0.96,1.86)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
the whole pregnancy
Single-pollutant models
(Restricted analysis to PM25 below the
median)
Per0.5* 10-5/m: 1.67 (0.66, 3.73)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
first trimester of pregnancy
Each trimester separately
2nd quartile: 1.15 (0.73,1.80)
3rd quartile: 1.01 (0.61,1.53);
4th quartile: 1.04 (0.70,1.57)
Per 0.5* 10-5/m: 1.03 (0.82,1.28)
All trimesters adjusted simultaneously
2nd quartile: 0.90 (0.52,1.58)
3rd quartile: 0.82 (0.45,1.31);
4th quartile: 0.88 (0.53,1.42)
Per 0.5* 10-5/m: 1.02 (0.77,1.29)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
second trimester of pregnancy
Each trimester separately
2nd quartile: 1.33 (0.85, 2.22)
3rd quartile: 1.76 (1.07, 2.91);
4th quartile: 1.83 (1.11,2.81)
Per 0.5* 10-5/m: 1.27 (1.04,1.54)
All trimesters adjusted simultaneously
2nd quartile: 1.30 (0.77, 2.16)
3rd quartile: 1.63 (0.93, 2.73);
4th quartile: 1.99 (1.12, 3.33)
Per 0.5*10-5/m: 1.21 (0.93,1.54)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
third trimester of pregnancy
Each trimester separately
2nd quartile: 1.30 (0.85, 2.09)
3rd quartile: 0.92 (0.55,1.50);
4th quartile: 1.50 (1.00, 2.27)
Per 0.5*10-5/m: 1.20 (0.98,1.44)
All trimesters adjusted simultaneously
2nd quartile: 0.99 (0.64,1.62)
3rd quartile: 0.71 (0.40,1.20);
December 2009
E-486
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
4thquartile: 1.14(0.68,1.91)
Per0.5*10-5/m: 1.15(0.92,1.42)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
first trimester of pregnancy
(adjustment for season of
conception)
All trimesters adjusted simultaneously
4th quartile: 0.73 (0.38,1.38)
Per0.5*10-5/m: 0.93 (0.41,1.32)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
second trimester of pregnancy
(adjustment for season of
conception)
All trimesters adjusted simultaneously
4th quartile: 2.45 (1.22, 4.77)
Per0.5*10-5/m: 1.14(0.70,1.64)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
third trimester of pregnancy
(adjustment for season of
conception)
All trimesters adjusted simultaneously
4th quartile: 1.19 (0.60, 2.48)
Per0.5*10-5/m: 1.29 (0.90,1.75)
Sensitivity analysis (bootstrapped PR)
2nd quartile: 1.19 (0.76,1.91)
3rd quartile: 1.52 (0.99, 2.34);
4th quartile: 1.62 (1.06, 2.55)
Per0.5*10-5/m: 1.35(1.01,1.83)
Estimated increments in prevalence
of birth weight <3000 g during
exposure 9 mo after birth
Per0.5*10-5/m:18%(-16%, 57%)
Reference: Wilhelm et al. (2005,
0886681
Period of Study: 1994-2000
Location: Los Angeles County,
California, U.S.
Outcome: Term low birth weight (LBW)
(<2500 g at > 37 completed wk
gestation)
Vaginal birth <37 completed wk
gestation
Pollutant: PM25 Ij*
Averaging Time: 24 h (every 3 days) 2
Entire pregnancy a
Trimesters of pregnancy ^
Months of pregnancy c
6 wk before birth
M Increment:
10|jg/m3
3 levels:
<25 percentile (reference)
25%-75 percentile
2 75 percentile
Age Groups: LBW: 2 37 completed wk
Preterm births: <37 completed wk
Study Design: Cross-sectional study
N: For LBW: 136,134
For preterm birth:
106,483
Statistical Analyses: Logistic
regression
Covariates: Maternal age, maternal
race, maternal education, parity, interval
since previous live birth, level of prenatal
care, infant sex, previous LBW or
preterm infant, birth season, other
pollutants (not specified in birth weight
analyses, also adjusted for gestational
age)
Dose-response Investigated? Yes
Statistical Package: NR
Mean (SD):
First trimester: 21.9
Third trimester: 21.0
6 wk before birth: 21.0
Range (Min, Max):
First trimester: 11.8-38.9
Third trimester: 11.8-.38.9
6 wk before birth: 9.9-48.5
Monitoring Stations:
Zip-code-level analysis: 9
Address-level analysis: 8
Copollutant (correlation):
First trimester
PM25-CO:0.57
PM25-N02:0.73
PM25-03:-0.55
PM25-PM10: 0.43
Third trimester:
PM25-CO:0.67
PM25-N02:0.78
PM25-03:-0.60
PM2.5-PM10: 0.52
6 wk before birth:
PM25-CO:0.63
PM25-N02:0.74
PM25-03:-0.60
PM25-PM10:0.60
Incidence of LBW (third trimester
exposure)
<17.1|jg/m3:2.4(2.0, 2.8)
17.1to<24.0|jg/m3:2.2(2.0, 2.5)
> 24.0 pg/m3: 2.1 (1.7,2.4)
Incidence of preterm birth (first
trimester exposure)
<18.0|jg/m3:10.6 (9.6,11.7)
18.0ttx25.4fjg/rn3: 8.8 (8.1, 9.5)
> 25.4 pg/m3: 9.0 (8.1,10.0)
Incidence of preterm birth (6 wk
before birth exposure)
<16.5 pg/m3: 8.2 (7.4, 9.1)
16.5to<24.7|jg/m3:8.8(8.2, 9.4)
> 24.7 pg/m3: 9.6 (8.7, 10.5)
Outcome: Preterm birth
Exposure Period: First trimester of
pregnancy
Address-level analysis:
Single-pollutant model: Distance Ł 1
mile
Per 10 pg/m3: 0.85 (0.70,1.02)
18.1 to <25.2|jg/m3: 0.91 (0.72, 1.16)
> 25.2 pg/m3: 0.83 (0.60, 1.14)
Single-pollutant model:
1
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
> 25.2 pg/nf: 0.79 (0.65, 0.97)
Multipollutant modell 24.9 pg/m3: 0.76 (0.70, 0.84)
Zip-code-level analysis:
Single-pollutant model:
Per 10 pg/m3: 0.73 (0.67, 0.80)
18.0to<25.4|jg/m3: 0.70 (0.61, 0.80)
> 25.4 pg/m3: 0.64 (0.53, 0.76)
Outcome: Preterm birth
Exposure Period: 6 wk before birth
Address-level analysis:
Single-pollutant model:
Distances 1 mile
Per 10 pg/m3:1.09 (0.91,1.30)
16.8to<24.1 pg/m3:1.21 (0.97, 1.51)
a24.1 pg/m3:1.25 (0.93, 1.68)
Single-pollutant model:
1 24.5 pg/m3:1.04 (0.87, 1.24)
Single-pollutant model:
2 24.6 pg/m3:1.08 (0.99, 1.17)
Zip-code-level analysis
Single-pollutant model: Per 10 pg/m3:
1.10(1.00,1.21)
16.5 to <24.7|jg/m3:1.06 (0.94, 1.20)
> 24.7 pg/m3:1.19 (1.02,1.40)
(See Notes)
Multipollutant model
Per 10 pg/m3:1.12 (0.90,1.40)
> 24.6 pg/m3:1.12 (0.82, 1.52)
Notes: In the table, the 75 percentile is
noted as 24.7 pg/m3. However, the text
notes the 75 percentile as 24.3 pg/m .
December 2009 E-488
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Woodruff etal. (2006,
0887581
Period of Study: 1999-2000
Location: California
Outcome (ICD10): SIDS (R95)
Respiratory mortality (JOO-J99)
Bronchopulmonary dysplasia (P27.1)
External accidents (V01-Y98)
Ill-defined and unspecified causes of
mortality (R99)
Age Groups: >28 days old
Study Design: Matched case-control
(matched on date of birth and birth
weight)
N: 3877 infants
Statistical Analyses: Conditional
logistic regression
Covariates: Maternal race, education,
parity, age, marital status
Dose-response Investigated? Yes
Statistical Package: STATA
Pollutant: PM25
Averaging Time: 24 hrs (every 6 days)
(time period between birth and post
neonatal death for the infant who died
and the same period for its four matched
surviving infants)
Percentiles: Infants who died of all
causes (cases)
26th: 13.4
60th(Median): 19.2
76th: 23.6
Matched controls
26th: 13.5
60th(Median): 18.4
76th: 22.7
Monitoring Stations:
73 (from 39 counties)
PM Increment: 10|jg/m
RR Estimate [Lower Cl, Upper Cl] lag:
All-cause mortality:
Unadjusted: 1.15 (1.00,1.32)
Adjusted: 1.07 (0.93,1.24)
Cause-specific mortality:
Respiratory (all):
Unadjusted: 2.15 (1.15, 4.02)
Adjusted: 2.13 (1.12, 4.05)
Respiratory (excluding deaths due to
BPD):
Adjusted: 1.42 (0.66, 3.03)
Respiratory (BPD alone):
Unadjusted: 6.00 (1.40, 27.76)
Respiratory (low birth weight infants
only):
Unadjusted: 3.09 (1.14, 8.40)
Respiratory (normal birth weight infants
only):
Unadjusted: 1.66 (0.74, 3.70)
Respiratory (with matched PM25 avgd
over all monitors in county)
Adjusted: 2.28 (0.94, 5.52)
Respiratory (averaging all PM25
measurements in county over the 2-yr
study period):
Adjusted: 2.26 (0.83, 6.21)
SIDS:
Unadjusted: 0.86 (0.61,1.22)
Adjusted: 0.82 (0.55,1.23)
SIDS (includes ICD10 code R99: ill-
defined and unspecified causes of
mortality):
Adjusted: 1.03 (0.79, 1.35)
External causes:
Unadjusted: 0.91 (0.56,1.47)
Adjusted: 0.83 (0.50, 1.39)
Compare against the lowest quartile,
estimates for respiratory-specific
mortality were provided:
2nd quartile: 1.28 (0.47, 3.51)
3rd quartile: 1.75 (0.65, 4.72)
4th quartile: 2.35 (0.85, 6.54)
December 2009
E-489
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Woodruff etal. (2008,
0983861
Period of Study: 1999-2002
Location: U.S. counties with >250,000
residents (96 counties)
Outcome (ICD10): Postneonatal
deaths: Respiratory mortality (JOOO-99,
plus bronchopulmonary dysplasia [BPD]
P27.1)
SIDS (R95)
Ill-defined causes (R99)
All other deaths evaluated as a control
category
Age Groups: Infants aged >28 days
and <1 yr
Study Design: Cross-sectional
N: 3,583,495 births (6,639 post neonatal
deaths)
Statistical Analyses: Logistic GEE
(exchangeable correlation structure)
Covariates: maternal race/ethnicity,
marital status, age, education,
primiparity, county-level poverty and per
capita income levels, yr and month of
birth dummy variables to account for
time trend and seasonal effects, and
region of the country
sensitivity analyses performed among
only those mothers with smoking
information (adjustment for smoking had
no effect on the estimates)
Season: Adjusted for yr and month of
birth dummy variables to account for
time trend and seasonal effects
Dose-response Investigated?
Evaluated the appropriateness of a
linear form from analysis based on
quartiles of exposure and concluded that
linear form was appropriate (data not
shown)
Statistical Package: SAS
Pollutant: PM25
Averaging Time: Measured
continuously for 24 h once every 6 days
exposure assigned by calculating avg
concentration of pollutant during first 2
mo of life
Median and IQR (25th-75th
percentile):
Survivors: 14.8 (11. 7-18.7)
All causes of death: 14.9 (12.0-18.6)
Respiratory: 14.8 (11.5-18.5)
SIDS: 14.5 (12.0-17.5)
SIDS + ill-defined: 14.8 (12.1-18.5)
Other causes: 14.9 (12.0-1 8.6)
Percentiles: See above
PM Component: Not assessed, but
controlled for region of the country to
account for PM composition variation
Monitoring Stations: NR
Copollutant (correlation):
PM10(r = 0.34)
PM25
CO (r = 0.35)
S02(r = 0.21)
03(r = -
PM Increment: IQR (7 pg/rri)
Effect Estimate [Lower Cl, Upper Cl]:
Adjusted ORs for single pollutant models
All causes: 1.04 (0.98,1.11)
Respiratory: 1.11 (0.96,1.29)
SIDS: 1.01 (0.86, 1.20)
Ill-defined + SIDS: 1.06 (0.97,1.17)
Other causes: 1.03 (0.96,1.12)
Adjusted ORs for multipollutant models
(including CO, 03, S02)
Respiratory: 1.05 (0.89,1.24)
SIDS: 1.04 (0.87, 1.23)
OR for respiratory deaths assessing
exposure as quartiles
Highest vs. Lowest quartile: 1.39(1.04,
1.85)
Notes: Monthly avg calculated if there
were at least 3 available measures for
PM
Assigned exposures using the avg
concentration of the county of residence
All units expressed in pg/m unless otherwise specified.
December 2009
E-490
-------
E.8. Long-Term Exposure and Mortality
Table E-30. Long-term exposure-mortality - PMio.
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: (Breitner et al., 2009,
1884391
Period of Study: Oct 1991-Mar 2002
Location: Efurt, Germany
Outcome: Mortality, excluding infants
and ICD-9 > 800
Study Design: Time-series
Covariates: Seasonal and weekday
variations, influenza epidemics, air
temperature, relative humidity
Statistical Analysis: Semiparametric
Poisson regression, polynomial
distributed lag (PDL)
Statistical Package: R
Age Groups: All
Pollutant: PM10
Averaging Time: Daily
Mean (SD) Unit:
1 (10/1/1991-8/31/1995):
50.6 ± 32.2 pg/m3
2(9/1/1995-2/28/1998):
41.1 ±28.4|jg/m3
3(3/1/1998-3/31/2002):
24.3 + 15.4 pg/m3
Total: 38.0 ±28.3 pg/m3
Range (Min, Max): NR
Copollutant: N02, CO, UFP
Increment: IQR
Relative Risk (96% Cl) Lag
New City Limits
6-day IQR: 17.2
PDL: 0.997 (0.972-1.022)
Mean of lags 0-5: 0.995 (0.971-1.019)
Old City Limits
6-day IQR: 17.2
PDL: 1.004 (0.978-1.031)
Mean of lags 0-5:1.001 (0.976-1.027)
New City Limits
15-day IQR: 14.5
PDL: 1.008 (0.982-1.036)
Mean of lags 0-14:1.006 (0.981 -1.032)
Old City Limits
15-day IQR: 14.5
PDL: 1.019 (0.991-1.048)
Mean of lags 0-14:1.017 (0.990-1.044)
Multiday Ma, 6-day
Overall IQR: 24.2
Overall RR (95% Cl):
0.998(0.976-1.021)
Period 1:0.996 (0.969-1.024)
Period 2:1.013 (0.972-1.056)
Period 3: 0.949 (0.897-1.004)
Multiday Ma, 15-day
Overall IQR: 22.3
Overall RR (95% Cl):
1.020(0.993-1.093)
Period 1:1.017 (0.984-1.051)
Period 2:1.012 (0.973-1.071)
Period 3: 0.978 (0.911-1.051)
Reference: (Slama et al, 2007,
0932161
Period of Study: Jan 1998-Jan 1999
Location: Munich, Germany
Outcome: Birth weight offspring at term
Study Design: Cohort study
N: 1016 births
Statistical Analyses: Poisson model
Covariates: Maternal passive smoking,
maternal age, gestational duration, sex
of child, parity, maternal education,
maternal size, prepregnancy weight,
other pollutants (PM25, PM25
absorbance, N02), season of conception
Dose-response Investigated? Yes
Statistical Package: STATA
Pollutant: PM25 (estimated based on
larger PM size fractions)
Averaging Time: Entire pregnancy
period and trimesters
Mean (SD): 14.4
Percentiles: 26th: 13.5
60th(Median): 14.4
76th: 15.4
Monitoring Stations:
Spatial component: 40
Temporal component: 1
Copollutant (correlation):
p.a. = pregnancy avg
trim. = trimester
PM25(p.a.)-PM25 (1sttrim.): 0.85
PM25(p.a.)-PM25 (2nd trim.): 0.77
PM25(p.a.)-PM25 (3rd trim.): 0.87
PM25 (p.a.)-N02 (p.a.): 0.45
PM25(p.a.)-N02 (1sttrim.): 0.18
PM2.5 (p.a.J-NQ, (2nd trim.): 0.32
PM Increment:
1) 1 pg/m3
2) Quartiles:
a) 1st (reference) (7.2-13.5 pg/m3)
b) 2nd (13.5-14.4 pg/m3)
c) 3rd (14.4-15.4 pg/m3)
day) 4th (15.41-17.5 pg/m3)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
the whole pregnancy
Single-pollutant models
Unadjusted models
2nd quartile: 1.07 (0.65,1.73); 3rd
quartile: 1.38 (0.91, 2.09)
4th quartile: 1.45 (0.92, 2.25)
Perl pg/m3:1.06 (0.95,1.19)
Adjusted models
2nd quartile: 1.08 (0.63,1.82); 3rd
quartile: 1.34 (0.86, 2.13)
4th quartile: 1.73 (1.15, 2.69); Per
1 pg/m3:1.13 (1.00,1.29)
Multipollutant models
Adjusted models
2nd quartile: 1.01 (0.57,1.85)
3rd quartile: 1.12 (0.64,1.87)
4th quartile: 1.36 (0.72, 2.45); Per
1 pg/m3:1.07 (0.91,1.26)
Single-pollutant models (restricted
December 2009
E-491
-------
Study Design & Methods
Concentrations1
PM2.5(p.a.)-N02 (3rd trim.): 0.37
PM25 (1sttrim.)-PM2.5 (2nd trim.): 0.40
Effect Estimates (95% Cl)
Analysis to PM25 absorbance below the
median)
Perl ug/m3: 1.15 (0.89, 1.52)
PM2.5 (1st trim.)-PM2.s (3rd trim.): 0.68
PM2.5 (1st trim.)-N02 (p.a): 0.48
PM2.5 (1st trim.)-N02 (1st trim.): 0.15
PM2.5 (1st trim.)-N02 (2nd trim.): 0.41
PM2.5 (1st trim.)-N02 (3rd trim.): 0.39
PM2.5 (2nd trim.)-PM2.s (3rd trim.): 0.51
PM2.5 (2nd trim.)-N02 (p.a): 0.23
PM2.5 (2ndtrim.)-N02 (1st trim.): -0.03
PM2.5 (2nd trim.)-N02 (2nd trim.): 0.17
PM25 (2nd trim.)-N02 (3rd trim.): 0.30
PM2.5 (3rd trim.)-N02 (p.a.): 0.39
PM25 (3rd trim.)-N02 (1st trim.): 0.33
PM2.5 (3rd trim.)-N02 (2nd trim.): 0.21
PM2.5 (3rd trim.)-N02 (3rd trim.): 0.23
PM2.s (p.a.)- PM25 absorbance (p.a.):
nfiQ
\j.\jy
PM2.5 (p.a. )-PM2.5abs (1st trim.): 0.33
PM25 (p.a.)- PM25 abs (2nd trim.): 0.48
PM25 (p.a.)- PM25 abs (3rd trim.): 0.52
PM25 (1st trim.)- PM25 abs (p.a.): 0.68
PM25 (1st trim.)- PM25 abs (1st trim.):
0.27
PM25(1st trim.)- PM25abs (2nd trim.):
0.53
PM25(1st trim.)- PM25abs (3rd trim.):
0.51
PM25(2ndtrim.)-PM25abs(p.a.):0.41
PM25(2nd trim.)- PM25abs(1st trim.):
0.08
PM2 5 (2nd trim.)- PM2 5 abs (2nd trim.):
0.29
PM25 (2nd trim.)- PM25 abs (3rd trim.):
0.41
PM25
PM25
0.48
3rdtrim.)-PM25abs p.a.): 0.62
3rd trim.)- PM2 5 abs 1st trim.):
PM2 5 (3rd trim.)- PM2 5 abs (2nd trim.):
0.36
PM2 5 (3rd trim.)- PM2 5 abs (3rd trim.):
0.37
Prevalence ratios (PRs) of birth
weight <3000 g
Multipollutant models (simultaneous
adjustment of 3rd trimester PM2 5 and
whole pregnancy PM25)
PM25 (whole pregnancy)
Perl pg/m3: 0.96 (0.75, 1.19)
PM2 5 (3rd trimester)
Perl pg/m3: 1.17 (0.98, 1.40)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
the whole pregnancy (adjustment for
season of conception)
4th quartile: 1.68 (1.05, 2.75); Per
1 |jg/m3:1.12(0.97, 1.28)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
first trimester of pregnancy
Each trimester separately
2nd quartile: 1.14 (0.74, 1.96); 3rd
quartile: 1.28 (0.84, 2.10)
4th quartile: 1.65 (1.02, 2.60)
Perl pg/m3: 1.10 (0.99, 1.20)
All trimesters adjusted simultaneously
2nd quartile: 0.97 (0.60, 1.73); 3rd
quartile: 0.98 (0.57, 1.75)
4th quartile: 1.22 (0.71, 2.18)
Perl pg/m3: 1.03 (0.90, 1.17)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
second trimester of pregnancy
Each trimester separately
2nd quartile: 0.83 (0.52, 1.32); 3rd
quartile: 1.08 (0.71, 1.60)
4th quartile: 0.94 (0.61, 1.47)
Perl pg/m3:1.01 (0.92, 1.12)
All trimesters adjusted simultaneously
2nd quartile: 0.75 (0.46, 1.24)
3rd quartile: 0.86 (0.56, 1.30);
4th quartile: 0.75 (0.48, 1.23)
Perl pg/m3: 0.94 (0.84, 1.06)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
third trimester of pregnancy
Each trimester separately
2nd quartile: 1.30 (0.80, 2.17)
3rd quartile: 1.44 (0.85, 2.27)
4th quartile: 1.90 (1.20, 2.82)
Perl pg/m3: 1.14 (1.02, 1.24)
All trimesters adjusted simultaneously
2nd quartile: 1.34 (0.79, 2.30)
3rd quartile: 1.48 (0.86, 2.58)
4th quartile: 1.91 (1.00,3.20)
Perl pg/m3: 1.14 (0.99, 1.29)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
third trimester of pregnancy
(adjustment for season of
conception)
All trimesters adjusted simultaneously
Perl pg/m3:1.25 (1.04,1.50)
Sensitivity analysisjbootstrapped PR)
2nd quartile: 0.98 (0.63,1.61); 3rd
quartile: 1.22 (0.82, 2.02)
4th quartile: 1.57 (1.02, 2.57)
Perl |jg/m3:1.11 (0.98,1.27)
Estimated increments in prevalence
of birth weight of <3000 g during
exposure 9 mo after birth
December 2009
E-492
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Per 1 |jg/m°: 7% (-7%, 22%)
Reference: (Slama et al., 2007,
0932161
Period of Study: Jan 1998-Jan 1999
Location: Munich, Germany
Outcome: Birth weight offspring at term
Study Design: Cohort study
N: 1016 births
Statistical Analyses: Poisson model
Pollutant: PM25 absorbance (estimated)
Averaging Time: Entire pregnancy
period and trimesters
Mean (SD): 1.76*
Percentiles: 26th: 1.61*
PM Increment:
1
2
Covariates: Maternal passive smoking,
maternal age, gestational duration, sex 60th(Median): 1.72*
of child, parity, maternal education,
maternal size, prepregnancy weight,
other pollutants (PM25, PM25
absorbance, N02), season of conception
75th: 1.89*
Unit(i.e. ug/m3): 10-5/m
Dose-response Investigated? Yes
Statistical Package: STATA
Monitoring Stations:
Spatial component: 40
Temporal component: 1
Copollutant (correlation):
p.a. = pregnancy avg
trim. = trimester
abs = absorbance
PM25 abs (p.a.)-PM25 abs (1st trim.):
0.54
PM25 abs (p.a.)-PM25 abs (2nd trim.):
0.84
PM25 abs (p.a.)-PM25 abs (3rd trim.):
0.55
PM25abs(p.a.)-PM25(p.a.):0.69
PM25abs
PM2.5 abs
-PM2
-PM2;
1st trim.): 0.68
2nd trim.): 0.41
PM2.5 abs (p.a.)-PM2.5 (3rd trim.): 0.62
PM25abs
PM2.5 abs
-N02
-N02
p.a.): 0.67
1st trim.): 0.34
PM2.5 abs (p.a.)-N02 (2nd trim.): 0.63
PM25abs
PM2.5 abs
p.a.)-N02 (3rd trim.): 0.36
1sttrim.)-PM25abs(2nd
trim.): 0.32
PM25abs(1sttrim.)-PM25abs(3rd
trim.): -0.26
PM25abs(1sttrim.)-PM25(p.a.):0.33
PM25abs 1st trim. -PM25 1st trim.):
0.27
PM25abs(1sttrim.)-PM25(2ndtrim.):
0.08
PM25abs(1sttrim.)-PM25(3rdtrim.):
0.48
PM25abs 1sttrim.)-N02 p.a.): 0.29
PM25abs 1sttrim.)-N02 1st trim.): 0.84
PM25abs(1sttrim.)-N02(2ndtrim.):
0.16
PM25abs(1sttrim.)-N02(3rdtrim.):-
0.39
PM25abs(2ndtrim.)-PM25abs(3rd
trim.): 0.31
PM25 abs (2nd trim.)-PM25 (p.a.): 0.48
PM25abs(2ndtrim.)-PM25(1sttrim.):
0.53
PM25 abs (2nd trim.)-PM25 (2nd trim.):
0.29
PM25 abs (2nd trim.)-PM25 (3rd trim.):
0.36
PM25abs
PM25abs
0.19
2nd trim,
2nd trim.
-N02 (p.a.): 0.61
-N02 (1st trim.):
PM25abs(2ndtrim.)-N02(2ndtrim.):
0.85
PM25abs(2ndtrim.)-N02(3rdtrim.):
0.17
PM25abs(3rdtrim.)-PM25(p.a.):0.52
PM25abs(3rdtrim.)-PM25(1sttrim.):
0.51
PM25abs(3rdtrim.)-PM25(2ndtrim.):
0.41
PM25abs(3rdtrim.)-PM25(3rdtrim.):
0.37
PM25abs(3rdtrim.)-N02(p.a.):0.40
PM25abs(3rdtrim.)-N02(1sttrim.):-
0.5* 10-5/m
Quartiles:
a) 1st (reference) (1.29-1.61)
b) 2nd (1.61-1.72)
c) 3rd (1.72-1.89)
day) 4th (1.89-3.10)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
the whole pregnancy
Single-pollutant models
Unadjusted models
2nd quartile: 1.19 (0.74,1.99)
3rd quartile: 1.56 (0.98, 2.50);
4th quartile: 1.52 (0.96, 2.46)
Per 0.5* 10-5/m: 1.25 (0.90,1.70)
Adjusted models
2nd quartile: 1.21 (0.73,1.97)
3rd quartile: 1.63 (0.98, 2.57);
4th quartile: 1.78 (1.10,2.70)
Per 0.5* 10-5/m: 1.45(1.06,1.87)
Multipollutant models Adjusted models
2nd quartile: 1.19 (0.70, 2.01)
3rd quartile: 1.55 (0.80, 2.80);
4th quartile: 1.46 (0.67, 2.90)
Per0.5*10-5/m: 1.33(0.76, 2.38)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
the whole pregnancy (adjustment for
season of conception)
4th quartile: 1.72 (1.08, 2.73)
Per 0.5* 10-5/m: 1.38 (0.96,1.86)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
the whole pregnancy
Single-pollutant models
(Restricted analysis to PM25 below the
median)
Per 0.5* 10-5/m: 1.67 (0.66, 3.73)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
first trimester of pregnancy
Each trimester separately
2nd quartile: 1.15 (0.73,1.80)
3rd quartile: 1.01 (0.61,1.53);
4th quartile: 1.04 (0.70,1.57)
Per 0.5*10-5/m: 1.03 (0.82,1.28)
All trimesters adjusted simultaneously
2nd quartile: 0.90 (0.52,1.58)
3rd quartile: 0.82 (0.45,1.31);
4th quartile: 0.88 (0.53,1.42)
Per 0.5* 10-5/m: 1.02 (0.77,1.29)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
second trimester of pregnancy
Each trimester separately
2nd quartile: 1.33 (0.85, 2.22)
3rd quartile: 1.76 (1.07, 2.91);
4th quartile: 1.83 (1.11,2.81)
Per 0.5* 10-5/m: 1.27 (1.04,1.54)
All trimesters adjusted simultaneously
2nd quartile: 1.30 (0.77, 2.16)
3rd quartile: 1.63 (0.93, 2.73);
4th quartile: 1.99 (1.12, 3.33)
Per0.5*10-5/m:1.21 (0.93,1.54)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
third trimester of pregnancy
Each trimester separately
2nd quartile: 1.30 (0.85, 2.09)
3rd quartile: 0.92 (0.55,1.50);
December 2009
E-493
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
0.34
PM25abs(3rdtrim.)-N02(2ndtrim.):
0.21
PM25abs(3rdtrim.)-N02(3rdtrim.):
0.88
4th quartile: 1.50 (1.00, 2.27)
Per0.5*10-5/m: 1.20 (0.98,1.44)
All trimesters adjusted simultaneously
2nd quartile: 0.99 (0.64,1.62)
3rd quartile: 0.71 (0.40,1.20);
4th quartile: 1.14 (0.68,1.91)
Per 0.5*10-5/m: 1.15 (0.92,1.42)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
first trimester of pregnancy
(adjustment for season of
conception)
All trimesters adjusted simultaneously
4th quartile: 0.73 (0.38,1.38)
Per0.5*10-5/m: 0.93 (0.41,1.32)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
second trimester of pregnancy
(adjustment for season of
conception)
All trimesters adjusted simultaneously
4th quartile: 2.45 (1.22, 4.77)
Per0.5*10-5/m: 1.14(0.70,1.64)
Prevalence ratios (PRs) of birth
weight <3000 g during exposure over
third trimester of pregnancy
(adjustment for season of
conception)
All trimesters adjusted simultaneously
4th quartile: 1.19 (0.60, 2.48)
Per0.5*10-5/m: 1.29 (0.90,1.75)
Sensitivity analysis (bootstrapped PR)
2nd quartile: 1.19 (0.76,1.91)
3rd quartile: 1.52 (0.99, 2.34);
4th quartile: 1.62 (1.06, 2.55)
Per 0.5*10-5/m: 1.35(1.01,1.83)
Estimated increments in prevalence
of birth weight <3000 g during
exposure 9 mo after birth Per 0.5 * 10-
5/m:18%(-16%, 57%)
Reference: Wilhelm et al. (2005,
Period of Study: 1994-2000
Location: Los Angeles County,
California, U.S.
Outcome: Term low birth weight (LBW)
(<2500 g at > 37 completed wk
gestation)
Vaginal birth <37 completed wk
gestation
Age Groups: LBW: 2 37 completed wk
Preterm births: <37 completed wk
Study Design: Cross-sectional study
N: For LBW: 136,134
For preterm birth:
106,483
Statistical Analyses: Logistic
regression
Covariates: Maternal age, maternal
race, maternal education, parity, interval
since previous live birth, level of prenatal
care, infant sex, previous LBW or
preterm infant, birth season, other
pollutants (not specified in birth weight
analyses, also adjusted for gestational
age)
Dose-response Investigated? Yes
Statistical Package: NR
Pollutant: PM25
Averaging Time: 24 h (every 3 days)
Entire pregnancy
Trimesters of pregnancy
Months of pregnancy
6 wk before birth
Mean (SD):
First trimester: 21.9
Third trimester: 21.0
6 wk before birth: 21.0
Range (Min, Max):
First trimester: 11.8-38.9
Third trimester: 11.8-.38.9
6 wk before birth: 9.9-48.5
Monitoring Stations:
Zip-code-level analysis: 9
Address-level analysis: 8
Copollutant (correlation):
First trimester
PM25-CO:0.57
PM25-N02:0.73
PM25-03:-0.55
PM2.5-PM10: 0.43
Third trimester:
PM25-CO:0.67
PM25-N02:0.78
PM25-03:-0.60
PM25-PM10: 0.52
6 wk before birth:
PM Increment:
1)10|jg/m3
2) 3 levels:
a) <25 percentile (reference)
b) 25%-75 percentile
c) > 75 percentile
Incidence of LBW (third trimester
exposure)
<17.1 pg/m3: 2.4 (2.0, 2.8)
17.1to<24.0|jg/m3:2.2(2.0, 2.5)
> 24.0 pg/m3: 2.1 (1.7,2.4)
Incidence of preterm birth (first
trimester exposure)
<18.0|jg/m3:10.6 (9.6,11.7)
18.0to<25.4fjg/nf:8.8(8.1,9.5)
> 25.4 pg/m3: 9.0 (8.1,10.0)
Incidence of preterm birth (6 wk
before birth exposure)
<16.5 pg/m3: 8.2 (7.4, 9.1)
16.5to<24.7|jg/m3:8.8(8.2, 9.4)
> 24.7 pg/m3: 9.6 (8.7, 10.5)
Outcome: Preterm birth
Exposure Period: First trimester of
pregnancy
Address-level analysis:
Single-pollutant model:
Distances 1 mile
Per 10 pg/m3: 0.85 (0.70,1.02)
18.1to<25.2|jg/m3:
December 2009
E-494
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
PM25-CO:0.63 0.91 (0.72, 1.16)
PM25-N02: 0.74 >25.2 pg/m3: 0.83 (0.60, 1.14)
PM25-03: -0.60 Single-pollutant model:
PM25-PM10:0.60 1 25.2 pg/m3: 0.79 (0.65, 0.97)
Multipollutant model! 24.9 pg/m3: 0.76 (0.70, 0.84)
Zip-code-level analysis:
Single-pollutant model:
Per 10 pg/m3: 0.73 (0.67, 0.80)
18.0to<25.4 pg/m3: 0.70 (0.61, 0.80)
> 25.4 pg/m3: 0.64 (0.53, 0.76)
Outcome: Preterm birth
Exposure Period: 6 wk before birth
Address-level analysis:
Single-pollutant model:
Distances 1 mile
Per 10 pg/m3:1.09 (0.91,1.30)
16.8to<24.1 pg/m3:1.21 (0.97, 1.51)
Ł24.1 pg/m3:1.25 (0.93, 1.68)
Single-pollutant model:
1 24.5 pg/m3:1.04 (0.87, 1.24)
Single-pollutant model:
2 24.6 pg/m3:1.08 (0.99, 1.17)
Zip-code-level analysis
Single-pollutant model:
Per 10 pg/m3:1.10 (1.00,1.21)
16.5 to <24.7|jg/m3:1.06 (0.94, 1.20)
> 24.7 pg/m3:1.19 (1.02,1.40)
(See Notes)
Multipollutant model
Per 10 pg/m3:1.12 (0.90,1.40)
> 24.6 pg/m3:1.12 (0.82, 1.52)
Notes: In the table, the 75 percentile is
noted as 24.7 pg/m . However, the text
notes the 75 percentile as 24.3 pg/m3.
December 2009 E-495
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Woodruff etal. (2006,
0887581
Period of Study: 1999-2000
Location: California
Outcome (ICD10): SIDS (R95)
Respiratory mortality (JOO-J99)
Bronchopulmonary dysplasia (P27.1)
External accidents (V01-Y98)
Ill-defined and unspecified causes of
mortality (R99)
Age Groups: >28 days old
Study Design: Matched case-control
(matched on date of birth and birth
weight)
N: 3877 infants
Statistical Analyses: Conditional
logistic regression
Covariates: Maternal race, education,
parity, age, marital status
Dose-response Investigated? Yes
Statistical Package: STATA
Pollutant: PM25
Averaging Time: 24 h (every 6 days)
(time period between birth and post
neonatal death for the infant who died
and the same period for its four matched
surviving infants) Percentiles: Infants
who died of all causes (cases)
25th: 13.4
60th(Median): 19.2
76th: 23.6
Matched controls
25th: 13.5
60th(Median): 18.4
75th: 22.7
Monitoring Stations:
73 (from 39 counties)
PM Increment: 10|jg/m
RR Estimate [Lower Cl, Upper Cl] lag:
All-cause mortality:
Unadjusted: 1.15 (1.00,1.32)
Adjusted: 1.07 (0.93,1.24)
Cause-specific mortality:
Respiratory (all):
Unadjusted: 2.15 (1.15, 4.02)
Adjusted: 2.13 (1.12, 4.05)
Respiratory (excluding deaths due to
BPD):
Adjusted: 1.42 (0.66, 3.03)
Respiratory (BPD alone):
Unadjusted: 6.00 (1.40, 27.76)
Respiratory (low birth weight infants
only): Unadjusted: 3.09 (1.14, 8.40)
Respiratory (normal birth weight infants
only): Unadjusted: 1.66 (0.74, 3.70)
Respiratory (with matched PM25 avgd
over all monitors in county)
Adjusted: 2.28 (0.94, 5.52)
Respiratory (averaging all PM25
measurements in county over the 2-yr
study period):
Adjusted: 2.26 (0.83, 6.21)
SIDS:
Unadjusted: 0.86 (0.61,1.22)
Adjusted: 0.82 (0.55,1.23)
SIDS (includes ICD10 code R99: ill-
defined and unspecified causes of
mortality):
Adjusted: 1.03 (0.79, 1.35)
External causes:
Unadjusted: 0.91 (0.56,1.47)
Adjusted: 0.83 (0.50,1.39)
Compare against the lowest quartile,
estimates for respiratory-specific
mortality were provided:
2nd quartile: 1.28 (0.47, 3.51)
3rd quartile: 1.75 (0.65, 4.72)
4th quartile: 2.35 (0.85, 6.54)
December 2009
E-496
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Woodruff etal. (2008,
0983861
Period of Study: 1999-2002
Location: U.S. counties with >250,000
residents (96 counties)
Outcome (ICD10): Postneonatal
deaths: Respiratory mortality (JOOO-99,
plus bronchopulmonary dysplasia [BPD]
P27.1)
SIDS (R95)
Ill-defined causes (R99)
All other deaths evaluated as a control
category
Age Groups: Infants aged >28 days
and <1 yr
Study Design: Cross-sectional
N: 3,583,495 births (6,639 post neonatal
deaths)
Statistical Analyses: Logistic GEE
(exchangeable correlation structure)
Covariates: Maternal race/ethnicity,
marital status, age, education,
primiparity, county-level poverty and per
capita income levels, yr and month of
birth dummy variables to account for
time trend and seasonal effects, and
region of the country
sensitivity analyses performed among
only those mothers with smoking
information (adjustment for smoking had
no effect on the estimates)
Season: Adjusted for yr and month of
birth dummy variables to account for
time trend and seasonal effects
Pollutant: PM25
Averaging Time: Measured
continuously for 24 h once every 6 days
Exposure assigned by calculating avg
concentration of pollutant during first 2
mo of life
Median and IQR (25th-75th
percentile):
Survivors: 14.8 (11. 7-18.7)
All causes of death: 14.9 (12.0-18.6)
Respiratory: 14.8 (11.5-18.5)
SIDS: 14.5 (12.0-17.5)
SIDS + ill-defined: 14.8 (12.1-18.5)
Other causes: 14.9 (12.0-1 8.6)
Percentiles: See above
PM Component: Not assessed, but
controlled for region of the country to
account for PM composition variation
Monitoring Stations: NR
Copollutant (correlation):
PM,o(r = 0.34)
PM25
CO (r = 0.35)
S02(r = 0.21)
03(r = -
Notes: Monthly avg calculated if there
were at least 3 available measures for
PM
PM Increment: IQR (7 pg/rri)
Effect Estimate [Lower Cl, Upper Cl]:
Adjusted ORs for single pollutant models
All causes: 1.04 (0.98,1.11)
Respiratory: 1.11 (0.96,1.29)
SIDS: 1.01 (0.86, 1.20)
Ill-defined + SIDS: 1.06 (0.97,1.17)
Other causes: 1.03 (0.96,1.12)
Adjusted ORs for multipollutant models
(including CO, 03, S02)
Respiratory: 1.05 (0.89,1.24)
SIDS: 1.04 (0.87, 1.23)
OR for respiratory deaths assessing
exposure as quartiles
Highest vs. Lowest quartile: 1.39(1.04,
1.85)
quartiles of exposure and concluded that
linear form was appropriate (data not
shown)
Statistical Package: SAS
All units expressed in pg/m unless otherwise specified.
December 2009
E-497
-------
E.9. Long-Term Exposure and Mortality
Table E-31. Long-term exposure-mortality - PMio.
Study
Design & Methods
Concentrations
Effect Estimates (95% Cl)
Reference: (Breitner et al., 2009,
1884391
Period of Study: Oct 1991-Mar 2002
Location: Efurt, Germany
Outcome: Mortality, excluding infants
and ICD-9 > 800
Study Design: Time-series
Covariates: Seasonal and weekday
variations, influenza epidemics, air
temperature, relative humidity
Statistical Analysis: Semiparametric
Poisson regression, polynomial
distributed lag (PDL)
Statistical Package: R
Age Groups: All
Pollutant: PM10
Averaging Time: Daily
Mean (SD) Unit:
1 (10/1/1991-8/31/1995):
50.6 + 32.2 pg/m3
2(9/1/1995-2/28/1998):
41.1+ 28.4 pg/m3
3 (3/1/1998-3/31/2002):
24.3+ 15.4 pg/m3
Total: 38.0 ±28.3 pg/m3
Range (Min, Max): NR
Copollutant: N02, CO, UFP
Increment: IQR
Relative Risk (96% Cl) lag
New City Limits
6-day IQR: 17.2
PDL: 0.997 (0.972-1.022)
Mean of lags 0-5: 0.995 (0.971-1.019)
Old City Limits
6-day IQR: 17.2
PDL: 1.004 (0.978-1.031)
Mean of lags 0-5:1.001 (0.976-1.027)
New City Limits
15-day IQR: 14.5
PDL: 1.008 (0.982-1.036)
Mean of lags 0-14:1.006 (0.981-1.032)
Old City Limits
15-day IQR: 14.5
PDL: 1.019 (0.991-1.048)
Mean of lags 0-14:1.017 (0.990-1.044)
Multiday Ma, 6-day
Overall IQR: 24.2
Overall RR (95% Cl): 0.998 (0.976-
1.021)
Period 1:0.996 (0.969-1.024)
Period 2:1.013 (0.972-1.056)
Period 3: 0.949 (0.897-1.004)
Multiday Ma, 15-day
Overall IQR: 22.3
Overall RR (95% Cl): 1.020 (0.993-
1.093)
Period 1:1.017
Period 2:1.012
0.984-1.051
0.973-1.071
Period 3: 0.978 (0.911-1.051)
1AII units expressed in ug/m3 unless otherwise specified.
December 2009
E-498
-------
Table E-32. Long-term exposure-mortality - PMio25.
Study
Reference: (Chen et al, 2005, 0879421
Period of Study: 1973-1998
Location: San Francisco, San Diego,
Los Angeles, CA
Reference: Goss et al. (2004, 0556241
Period of Study: 1999-2000
Location: United States
Reference: Lipert et al. (2009,
190271)
Design & Methods
Outcome: Mortality: CHD
Study Design: Cohort
Statistical Analyses: Cox proportion
hazards model
Age Groups: >25
Outcome: Mortality
Study Design: Cohort Study (Cystic
Fibrosis Cohort)
Statistical Analyses: Logistic
Regression
Age Groups: >6 yr
Outcome: Mortality
Study Design: Retrospective Cohort
Concentrations1
Pollutant: PM10.2.5
Averaging Time: 25 yr
Mean (SD): 25.4
Range (Min, Max): NR
Copollutant:
N02
03
so.
Pollutant: PM25
Averaging Time: Annual avg
Mean (SD) unit: PM25: 13.7 (4.2)
IQR: PM25: 11.8-15.9
Copollutant:
03
N02
S02
CO
Pollutant: PMi0.25
Mean (SD): 16.0 (5.1)
Effect Estimates (95% Cl)
Increment: 10 pg/m3
Relative Risk (Lower Cl, Upper Cl)
lag:
M_|.c
Males
PMio.25: 0.93 (0.68, 1.29)0-1
PMi0.25+N02: 0.86 (0.62, 1.20)0-1
PM10.25+S02: 0.90 (0.64,1. 27) 0-1
PMi0.25+03:1.01 (0.67,1.51)0-1
Females
PMio.25: 1.20 (0.95, 1.53)0-1
PMi0.25+N02: 1.19 (0.92, 1.54)0-1
PM10.25+S02: 1.31 (1.03,1. 68) 0-1
PMi0.25+03: 1.47 (1.10,1. 96) 0-1
Increment: 10|jg/m3
PM25: 1.32 (0.91-1. 93)
Increment: 12
1.07(1.01,1.13)
Period of Study: 1989-1996
Location: Various parts of the Untied
States
Statistical Analyses: Cox proportional
hazards regression
Age Groups: Male U.S. veterans
between ages of 39 and 63 (Avg. age:
51)
Reference: McDonnell et al. (2000,
0103191
Period of Study: 1973-1977
Location: California
Outcome: Mortality
Study Design: Cohort (AHSMOG
airport cohort)
Statistical Analyses: Cox regression
models
Age Groups: Males, 27 yr+
Pollutant: PM10.25
Averaging Time: Monthly avg
Mean (SD): PM10.25: 27.3 (8.6)
IQR: 9.7
Copollutant:
03: 0.70
S02: 0.31
N02: 0.23
S04: 0.47
Increment: IQR
All Cause: 1.05 (0.92-1.20)
Resp: 1.19 (0.88, 1.62)
Lung Cancer: 1.25 (0.63-2.49)
All units expressed in pg/m unless otherwise specified.
December 2009
E-499
-------
Table E-33. Long-term exposure-mortality - PIVhs (including PM components/sources).
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Abrahamowicz et al. (2003, Outcome: Mortality: All-causes
0862921
Study Design: Case-cohort study
Period of Study: 1982-1989
Statistical Analyses: Cox proportion-
Pollutant: PM25
Averaging Time: Annual
Mean (SD): 18.2
Relative Risk (Min Cl, Max Cl)
Estimated from graph (Fig 1): log HR
for a 24.5 pg/m increase in PM25 over
time
spline generalization
Age Groups: >18
Range (Min,
Copollutant:
Max): (9.0, 33.5)
Sulfates
Yr
0:0.5 -1.1, 1.6)
2: 0.6 0.2, 0.9)
4: 0.6 (0.3, 0.8)
6:0.8(0.3, 1.1)
8: -1.0 (-1.5, 1.0)
Reference: Abrahamowicz et al. (2003, Outcome: Mortality: All-causes
0862921
Study Design: Case-cohort study
Period of Study: 1982-1989
Statistical Analyses: Cox proportion-
Pollutant: Sulfates
Averaging Time: Annual
Mean (SD): 18.2
Relative Risk (Min Cl, Max Cl)
Estimated from graph (Fig 1): Log HR
for a 19.9 pg/m3 increase in Sulfates
overtime
^vwauvii. IUIWILICO na^amo inuuci IICAIUIC ic^icoo
spline generalization
Age Groups: >1 8
Range (Min, Max): (9.0, 33.5)
Copollutant: PM25
Yr
0:0.1 (-0.2, 0.7)
2:0.1 (-0.2,0.4)
4:0.0
6:0.3
-0.4, 0.3)
-0.1,0.5)
8: 0.4 (-0.4, 1.6)
Reference: Ballester et al. (2008,
1899771
Period of Study: 2001-2002
Location: Europe
Outcome: Mortality-All-causes
Study Design: Health Impact
Assessment
Statistical Analyses: Aphesis Network
Age Groups: >30
Pollutant: PM25
Averaging Time: Annual
Mean (SD): NR
Range (Min, Max): NR
Potential Reduction in the total
burden of mortality (min Cl, max Cl)
for four different decreases in annual
PMu using a conservative estimate
Reduction to 25 pg/m3 - 0.4 (0.1, 0.8
Reduction to 20 pg/m3 - 0.8 (0.2,1.6
Reduction to 15 pg/m3-1.6 (0.4, 3.1)
Reduction to 10 |jg/m3 - 3.0 (0.8, 5.8)
Reference: Beelen et al. (2008,
1562631
Period of Study: 1987-1996
Location: Netherlands
Outcome: Mortality:
Total (nonaccidental) (<800)
Cardio-respiratory (390-448, 490-496,
487, 480-486, 507)
Pulmonary (460-519)
Cardiovascular (400-440)
Lung Cancer (162)
Other-causes
Study Design: Case-cohort study and
prospective cohort
Statistical Analyses: Cox proportion-
hazards model
Age Groups: 55-69
Pollutant: PM25
Averaging Time: Annual
Mean(SD):28.3(2.1)|jg/m3
Range (Min, Max): (23.0, 36.8)
Copollutant (correlation):
N02: (>0.8)
BS: (>0.8)
S02: (>0.6)
Increment: 11 pg/m
Relative Risk (Min Cl, Max Cl)
RR for the association between
exposures to PM25 and cause
specific mortality
Natural Cause:
Full cohort: 1.06 (0.97,1.16)
Case cohort: 0.86 (0.66,1.13)
Cardiovascular:
Full cohort: 1.04 (0.90,1.21)
Case cohort: 0.83 (0.60,1.15)
Respiratory:
Full cohort: 1.07 (0.75,1.52)
Case cohort: 1.02 (0.56,1.88)
Lung Cancer: Full cohort: 1.06 (0.82,
1.38)
Case cohort: 0.87 (0.52,1.47)
Other cause: F
Ull cohort: 1.08 (0.96,1.23)
Case cohort: 0.85 (0.65,1.12)
RR for the association between
exposures to BS and cause specific
mortality
Natural Cause:
Full cohort: 1.05 (1.00,1.11)
Case cohort: 0.97 (0.83,1.13)
Cardiovascular: Full cohort: 1.04 (0.95,
1.13)
Case cohort: 0.98 (0.81,1.18)
Respiratory:
Full cohort: 1.22 (0.99,1.50)
Case cohort: 1.29 (0.91,1.83)
Lung Cancer:
Full cohort: 1.03 (0.88,1.20)
Case cohort: 1.03 (0.77,1.38)
Other cause:
Full cohort: 1.04 (0.97,1.12)
Case cohort: 0.91 (0.78,1.07)
December 2009
E-500
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Breitner et al. (2009,
1884391
Period of Study: Oct 1991-Mar 2002
Location: Efurt, Germany
Outcome: Mortality, excluding infants
and ICD-9 > 800
Study Design: Time-series
Covariates: Seasonal and weekday
variations, influenza epidemics, air
temperature, relative humidity
Statistical Analysis: Semiparametric
Poisson regression, polynomial
distributed lag (PDL)
Statistical Package: R
Age Groups: All
Pollutant: PM25
Averaging Time: Daily
Mean (SD) Unit:
1 (10/1/1991-8/31/1995):
50.6 + 32.2 pg/m3
2(9/1/1995-2/28/1998):
41.1+ 28.4 pg/m3
3 (3/1/1998-3/31/2002):
24.3+ 15.4 pg/m3
Total: 38.0 ±28.3 pg/m3
Range (Min, Max): NR
Copollutant: N02, CO, UFP
Increment: IQR
Relative Risk (96% Cl) lag
New City Limits
6-day IQR: 13.3
PDL: 1.009 (0.984-1.035)
Mean of lags 0-5:1.004 (0.981-1.027)
Old City Limits
6-day IQR: 13.3
PDL: 1.017 (0.990-1.044)
Mean of lags 0-5:1.010 (0.986-1.035)
New City Limits
15-day IQR: 11.5
PDL: 1.019 (0.988-1.050)
Mean of lags 0-14:1.017 (0.992-1.042)
Old City Limits
15-day IQR: 11.5
PDL: 1.030 (0.997-1.063)
Mean of lags 0-14:1.025 (0.999-1.052)
Multiday Ma, 6-day
Overall IQR: 13.3
Overall RR (95% Cl):
1.004(0.981-1.027)
Period 1:NR
Period 2:1.017 (0.990-1.044)
Period 3: 0.974 (0.937-1.013)
Multiday Ma, 15-day
Overall IQR: 11.5
Overall RR (95% Cl):
1.017(0.992-1.042)
Period 1:NR
Period 2:1.016 (0.988-1.045)
Period 3:1.016 (0.971-1.063)
December 2009
E-501
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Brunekreef et al. (2009,
1919471
Period of Study: 1987-1996
Location: The Netherlands
Outcome: All cause mortality (ICD-9
400-440, 460-519, > 800)
Study Design: Case-cohort
Pollutant: PM25, estimated from PM10
levelsf
Covariates:
Individual: sex, age, Quetelet index,
smoking status, passive smoking
status, educational level, occupation,
occupational exposure, marital status,
alcohol use, intake of vegetables, fruits, |\|o2: 0.75
energy, saturatured and
monounsaturated fatty acids, trans fatty BS: 0.84
acids, total fiver, folic acid and fish
Averaging Time: 24 h
60th Percentile: 28 pg/m3
Range (Min, Max): 23-37
Copollutant (correlation):
Area-level: Percent of population with
income below the 40th percentile and
above the 80th percentile
Statistical Analysis: Cox proportional
hazards
Statistical Package: State, SPSS, R
Age Groups: 120,000 adults aged
55-69 yr at enrollment
NO: 0.69
SO,: 0.43
Increment: 10|jg/m
Relative Risk (96 % Cl) for PM2 5
concentrations and cause specific
mortality
Case Cohort
Natural Cause: 0.86 (0.66-1.13)
Cardiovascular: 0.83 (0.60-1.15)
Respiratory: 1.02 (0.56-1.88)
Lung Cancer: 0.87 (0.52-1.47)
Noncardiopulmonary, non-lung cancer:
0.85(0.65-1.23)
Full Cohort
Natural Cause: 1.06 (0.97-1.16)
Cardiovascular: 1.04 (0.90-1.21)
Respiratory: 1.07 (0.75-1.52)
Lung Cancer: 1.06(0.82-1.38)
Noncardiopulmonary, non-lung cancer:
1.08(0.72-1.19)
Relative Risk (96%CI) for PM2.s
concentrations and cause specific
mortality in full cohort analysis by
confounder model
Natural Cause Mortality
Unadjusted: 1.11 (1.04-1.20)
Smoking: 1.04 (0.96-1.13)
Smoking, area-level income:
1.06(0.97-1.16)
Cardiovascular Mortality
Unadjusted: 1.09 (0.97-1.23)
Smoking: 1.02 (0.90-1.16)
Smoking, area-level income:
1.04(0.90-1.21)
Respiratory Mortality
Unadjusted: 1.23 (0.92-1.65)
Smoking: 1.10 (0.81-1.50)
Smoking, area-level income:
1.07(0.75-1.52)
Lung Cancer Mortality
Unadjusted: 1.17 (0.95-1.46)
Smoking: 1.06 (0.85-1.33)
Smoking, area-level income:
1.06(0.82-1.38)
Noncardiopulmonary, Non-Lung Cancer
Mortality
Unadjusted: 1.10 (1.00-1.22)
Smoking: 1.05 (0.94-1.16)
Smoking, area-level income:
1.08(0.96-1.22)
Reference: Chen et al. (2005, 0879421 Outcome: Mortality: CHD
Period of Study: 1973-1998 Study Design: Cohort
Location: San Francisco, San Diego,
Los Angeles, CA
Statistical Analyses: Cox proportion
hazards model
Age Groups: >25
Pollutant: PM25
Averaging Time: 25 yr
Mean (SD): 29.0
Range (Min, Max): NR
Copollutant: N02, 03, S02
Increment: 10|jg/m
Relative Risk (Lower Cl, Upper Cl)
lag:
Males
PM25: 0.89 (0.69, 1.17)0-1
PM25+N02: 0.82 (0.61, 1.10); 0-1
PM25+S02: 0.86 (0.65,1.14) 0-1
PM25+03: 0.92 (0.65,1.29) 0-1
Females
PM25:1.19 (0.96, 1.47)0-1
PM25+N02:1.18 (0.95, 1.47); 0-1
PM25+S02:1.36 (1.05,1.74) 0-1
PM25+03:1.61 (1.17,2.22) 0-1
December 2009
E-502
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Eftim et al. (2008, 0991041 Outcome (ICD-9): All nonaccidental
causes(<8001
Period of Study: 2000-2002 v '
Location: USA, Same cities as six
cities and ACS cohorts
Study Design: Cross-sectional
Statistical Analyses: Log-linear
regression, Poisson
Age Groups: >65
Pollutant: PM25
Averaging Time: Annual avg
Mean (SD): |
ACS: 13.6 (2.8)
SCS: 14.1 (3.1)
Range (Min, Max):
ACS: (6.0, 25.1); SCS: (9.6, 19.1)
Increment: 10 pg/m
% Increase in Mortality for overall
exposure period and individual yr
(96%CI Min, 96%CI Max):
ACS (adjusted for age, sex)
Overall: 10.8 (8.6,13.0)
2000:10.9(7.3, 14.6)
2001:9.1 (5.3, 12.7)
2002:10.1(6.0,14.3)
SCS (adjusted for age, sex)
Overall: 20.8 (14.8, 27.1)
2000:17.8 (9.8, 26.4)
2001:16.5 7.4, 25.0
2002:33.5(19.2,49.3)
Reference: Enstrom et al. (2005,
0873561
Period of Study: 1973-2002
Location: 25 California Colonies
Outcome: Mortality: Cardiovascular-
respiratory (390-448)
(480-486, 487, 490-496, 507)
Study Design: Retrospective cohort
11 California Colonies (EPA IPN study) Statistical Analyses: Cox proportional
hazards regression model, SAS
PHREG
Age Groups: 35 or older
Pollutant: PM25
Averaging Time: Annual
Mean (SD): 23.4
Range (Min, Max): (13.1 pg/m3, 36.1)
Relative Risk (Lower Cl, Upper Cl)
RR from causes for both sexes by
county from 1973-2002
Alameda: 0.962 (0.926,0.999)
Butte: 0.999 (0.910,1.096)
Contra Costa: 0.999 (0.943,1.058)
Fresno: 0.935 (0.872,1.002)
Humboldt: 0.992 (0.900,1.092)
Kern: 0.944 (0.872,1.023)
Marin: 0.939 (0.867,1.016)
Napa: 0.949 (0.868,1.038)
Orange: 0.990 (0.948,1.034)
Riverside: 0.959 (0.906,1.015)
Sacramento: 0.998 (0.944,1.055)
San Bernardino: 0.992 (0.938,1.049)
San Diego: 0.992 (0.954,1.033)
San Francisco: 0.963 (0.914,1.014)
San Joaquin: 0.925 (0.816,1.049)
San Mateo: 0.949 (0.899,1.003)
Santa Barbara: 0.968 (0.878,1.068)
Santa Clara: 0.955 (0.910,1.003)
Santa Cruz: 0.890 (0.793,0.999)
Solano: 0.901 (0.815,0.995)
Sonoma: 0.968 (0.884,1.060)
Stanislaus: 0.984 (0.904,1.072)
Tulare: 1.047 (0.979,1.119)
Ventura: 0.967 (0.872,1.072)
RR from all causes for 11 counties
for both sexes (EPA IPN study)
Santa Barbara: 0.968 (0.878,1.068)
Contra Costa: 0.999 (0.943,1.058)
Alameda: 0.962 (0.926,0.999)
Butte: 0.999 (0.910,1.096)
San Francisco: 0.963 (0.914,1.014)
Santa Clara: 0.955 (0.910,1.003)
Fresno: 0.935 (0.872,1.002)
San Diego: 0.992 (0.954,1.033)
Kern: 0.944 (0.872,1.023)
Riverside: 0.959 (0.906,1.015)
Reference: Filleul et al. (2005, 0873571 Outcome: Nonaccidental causes
(<800), cardiopulmonary disease (401-
440 and 460-519), lung cancer (162)
Period of Study: 1974-1976
Location: 7 cities in France
Age Groups: 25-59 yr
Study Design: Cohort
N: 14,284 people
Statistical Analyses: Cox proportional
hazard, regression
Covariates: Sex, smoking habits,
educational level, body-mass index
(BMI), occupational exposure
Statistical Package: Proc Phreg SAS
Pollutant: Total suspended particles
(TSP)
Averaging Time: NR
Mean (SD): NR
Range (Min, Max): (45, 243)
PM Component: NR
Monitoring Stations: 1 station
Copollutant (correlation):
BS r = 0.87
S02r = 0.17
NO r = 0.84
N02r = 0.60
Increment: 10|jg/m
Adjusted mortality rate ratios: 24 areas:
All nonaccidental causes:
1.00[0.99,1.01]
Lung cancer: 0.97[0.94,1.01]
Cardiopulmonary disease:
1.01 [0.99,1.03]
18 areas: All nonaccidental causes:
1.05[1.02, 1.08]
Lung cancer: 1.00[0.92,1.10]
Cardiopulmonary disease:
1.06[1.01,1.12]
December 2009
E-503
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Fuentes et al. (2006,
0976471
Period of Study: Jun 2000
Location: Conterminous U.S.
Outcome: Mortality:
Study Design: Time-series
Statistical Analyses: Generalized
Poisson Regression
Age Groups: 0-14,15-64, >65
Covariates: Temperature, pressure,
dew point, wind speed, elevation, age,
ethnicity
Pollutant: PM25
Averaging Time: Monthly
Mean (SD): 6.60 (0.76)
Copollutant: PMi0, 03
Increment: 10|jg/m
PM25:1.066 (1.064, 1.069)
PM10:1.030 (1.028, 1.032)
Reference: Janes et al. (2007, 0909271
Period of Study: 2000-2002
Location: 113 U.S. counties
Outcome: Mortality:
Study Design: Time-series
Statistical Analyses: Cox proportional
hazards model
Age Groups: 65-74
75-84
85+
Pollutant: PM25
Averaging Time: Annual avg
Mean (SD): NR
Range (Min, Max): NR
Increment: 1 pg/m
% Increase (Lower Cl, Upper Cl) lag:
Overall % Increase by age-sex stratum
Age Category
65-74: Male: 1.48 (0.93,2.03)
Female: 0.83 (0.24,1.43)
75-84: Male: 0.85 (0.34,1.35)
Female: 0.77 (0.28,1.27)
85+: Male: 0.70 (0.03,1.38)
Female: 0.59 (0.05,1.12)
National Trend % Increase by age-sex
stratum
Age Category
65-74: Male: 3.55 (2.77,4.34)
Female: 1.97 (1.12,2.83)
75-84: Male: 2.48 (1.83,3.14)
Female: 2.29 (1.66,2.93)
85+: Male: 1.38 (0.52,2.26)
Female: 1.65 (1.01,2.29)
Local Trend % Increase by age-sex
stratum
Age Category
65-74: Male: 0.04 (-0.58,0.67)
Female:-0.03 (-0.71,0.66)
75-84: Male:-0.34 (-0.87,0.19)
Female:-0.31 (-0.82,0.21)
85+: Male: <0.01 (-0.71,0.73)
Female: -0.22 (-0.74,0.31)
'Local trends are county specific
deviations from national trends
Reference: Jerrett et al. (2003,
0873801
Period of Study: 1982
Location: 151 cities from ACS
Outcome: Mortality
Study Design: Multilevel, individual-
ecologic analysis
Statistical Analysis: Cox proportional
hazards model
Covariates: Smoking, education,
occupational exposures, BMI, marital
status, alcohol consumption, gender
Pollutant: Sulfates
Mean (SD): 10.6
Range (Min, Max): 3.6,23.5
Increment: 19.9 (Range)
All Cause: S04:1.17 (1.07,1.27)
S04 +CO: 1.16 (1.10, 1.23)
S04 + N02:1.16 (1.08, 1.24)
S04 + 03:1.17 (1.11,1.24)
S04 + S02:1.05 (0.98, 1.12)
CPD:S04:1.25 (1.16, 1.35)
S04+ 00:1.28(1.18, 1.39)
S04 + N02:1.29 (1.17, 1.42)
S04 + 03:1.27 (1.17, 1.38)
S04 + S02:1.13 (1.03, 1.24)
Lung Cancer: S04:1.31 (1.09,1.58)
S04+ 00:1.26(1.03, 1.53)
S04 + N02:1.31 (1.05, 1.65)
S04 + 03:1.30 (1.07, 1.59)
S04 + S02:1.37 (1.08, 1.73)
December 2009
E-504
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Jerrett et al. (2005,
0876001
Period of Study: 1982-2000
Location: Los Angeles, California
Outcome: Mortality: Non- accidental
(<800)
IHD (410-414)
Cardiopulmonary (400-440, 460-519)
Lung Cancer (162)
OtherCancers (140-149,160,161,163-
239)
Other causes
Study Design: Retrospective Cohort
Statistical Analyses: Cox regression
hazards model
kriging, radial basis function
multiquadric interpolator
Age Groups: All ages
Pollutant: PM25
Averaging Time: Annual avg
Mean (SD): NR
Range (Min, Max): NR
Copollutant: 03
Increment: 10|jg/m
Relative Risk (Lower Cl, Upper Cl)
All Causes-PM2 5 Only:
1.24(1.11,1.37)
44 Ind. Covariatestogether+PM25:
1.17(1.03,1.32)
44 Ind. Covariatestogether PM25+03:
1.20(1.07,1.34)
44 Ind. Covariates together+intersection
within freeways within 500 m+
PM25+03:1.17 (1.05,1.31)
IHD-PM25 Only: 1.49 (1.20,1.85)
44 Ind. Covariates together+PM25:
1.39(1.12,1.73)
44 Ind. Covariates together+PM25+03:
1.45(1.15,1.82)
44 Ind. Covariates together+intersection
within freeways within 500 m+
PM25+03:1.38 (1.11,1.72)
Cardiopulmonary - PM25 Only:
1.20(1.04,1.39)
44 Ind. Covariates together PM25+03:
1.19(1.02,1.38)
44 Ind. Covariates together+intersection
within freeways within 500 m+
PM25+03:1.13 (0.97,1.31)
Lung Cancer-PM25 Only: 1.60
(1.09,2.33)
44 Ind. Covariates together+PM25:
1.44(0.98,2.11)
44 Ind. Covariates together+intersection
within freeways within 500 m+
PM25+03:1.46 (0.99,2.16)
Other Cancers-PM2 5 Only:
1.09(0.85,1.40)
44 Ind. Covariates together PM25+03:
1.08(0.83,1.39)
44 Ind. Covariates together+intersection
within freeways within 500 m+
PM25+03:1.08 (0.83,1.39)
All Other Causes - PM2 5 Only:
1.11 (0.74,1.67)
44 Ind. Covariates together PM25+03:
0.95(0.64,1.39)
44 Ind. Covariates together+intersection
within freeways within 500 m+
PM25+03:1.02 (0.71,1.48)
December 2009
E-505
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Laden et al. (2006, 0876051
Period of Study: 1974-1998
Period 1:1974-1989
Period 2:1990-1998
Location: Nine U.S. Cities
Watertown, MA
Kingston, TN
Harriman, TN
St. Louis, MO
Steubenville, OH
Portage, Wl
|Wyocena, Wl
Pardeeville, Wl
Topeka, KS
Outcome: Total mortality
Nonaccidental (<800)
Cardiovascular (400-440)
Respiratory (485-496)
Lung Cancer (162)
Other
Study Design: Prospective Cohort
Statistical Analyses: Cox proportional
hazards regression
Age Groups: 25-74
Pollutant: PM25
Averaging Time: Annual avg
Mean (SD):
Period 1
Portage: 11.4
Topeka: 12.4
Watertown: 15.4
Harriman: 20.9
St Louis: 19.2
Steubenville: 29.0
Period 2
Portage: 10.2
Topeka: 13.1
Watertown: 12.1
Harriman: 18.1
St. Louis: 13.4
Steubenville: 22.0
Increment: 10|jg/m
Relative Risk (Lower Cl, Upper Cl)
lag:
Period 1:
Portage: 1.00
Topeka: 1.06 (0.86,1.31)
Watertown: 1.06 (0.87,1.28)
Harriman: 1.19 (0.98,1.44)
St Louis: 1.15 (0.96, 1.38)
Steubenville: 1.31 (1.10,1.57)
Period 2:
Portage: NR
Topeka: 1.01 (0.83,1.22)
Watertown: 0.82 (0.67,1.00)
Harriman: 1.10 (0.91,1.33)
St Louis: 0.96 (0.80, 1.15)
Steubenville: 1.06 (0.89,1.27)
Complete Period:
Portage: 1.00
Topeka: 1.03 (0.89,1.19)
Watertown: 0.95 (0.83,1.08)
Harriman: 1.15 (1.01,1.32)
St. Louis: 1.05 (0.93, 1.20)
Steubenville: 1.18 (1.04,1.34)
RR for complete follow up avg PIVhs
Total Mortality: 1.16 (1.07,1.26)
Cardiovascular: 1.28 (1.13,1.44)
Respiratory: 1.08 (0.79,1.49)
Lung Cancer: 1.27 (0.96,1.69)
Other: 1.02 (0.90,1.17)
RR for Period 1 avg PMu
Total Mortality: 1.18 (1.09,1.27)
Cardiovascular: 1.28 (1.14,1.43)
Respiratory: 1.21 (0.89,1.66)
Lung Cancer: 1.20 (0.91,1.58)
Other: 1.05 (0.93,1.19)
Decrease in avg PIVh.s over the 2
periods
Total Mortality: 0.73 (0.57, 0.95)
Cardiovascular: 0.69 (0.46,1.01)
Respiratory: 0.43 (0.16,1.13)
Lung Cancer: 1.06 (0.43, 2.62)
Other: 0.85 (0.56,1.27)
Reference: Lipfert et al. (2006, 0887561
Period of Study: 1989-1996
Location: Various parts of the Untied
States
Outcome: Mortality
Study Design: Retrospective Cohort
Statistical Analyses: Cox proportional
hazards regression
Age Groups: Male U.S. veterans
between ages of 39 and 63 (Avg. age:
51)
Pollutant: Sulfate
Mean (SD) from 1976-81:10.7 (3.6)
Increment: 8
1.045(0.944, 1.157)
Reference: Lipfert et al. (2006, 0887561
Period of Study: 1989-1996
Location: Various parts of the Untied
States
Outcome: Mortality
Study Design: Retrospective Cohort
Statistical Analyses: Cox proportional
hazards regression
Age Groups: Male U.S. veterans
between ages of 39 and 63 (Avg age
51)
Pollutant: PM25
Mean (SD): 14.3 (3.2)
Increment: 8
1.118(1.038,1.203)
December 2009
E-506
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Lipfert et al. (2006, 0882181 Outcome: Mortality: Non- accidental
Period of Study: 1997-2002 ^ '
Study Design: Retrospective cohort ,
Location: Various parts of the Untied Mean (SD): 15.02 (4.80) pg/m3 (2000-
States Statistical Analyses: Cox proportional 2003)
Pollutant: PM25
Averaging Time: Annual avg
Increment: 10|jg/m
% Increase per 10 ug/m increase in
PM2.s
Single-Pollutant Model
As: -5.23%
AIC
Age Groups: Male U.S. veterans
between ages of 39 and 63 (Avg. age:
51)
Range (Min, Max): (3.29, 24.96)
Copollutant (correlation):
As: r = 0.443
Cr:r = 0.379
Cu:r = 0.530
Fe:r = 0.379;
Pb:r = 0.489
Mn:r = 0.389;
Ni:r = 0.140
Se:r = 0.312;
V:r = 0.197
Zn: r = 0.420;
OC:r = 0.620
EC: r = 0.544; |
S04:r = 0.827
N03:r = 0.649
N02:r = 0.641
Peak CO: r = 0.040
Peak 03:r = 0.222
Peak S02:r = 0.714
Cu:2.12%
Fe:2.81%
Pb: -2.40%
Mn:-1.20%
Ni: 3.75%
Se: -0.30%
V: 5.08%
Zn: 1.52%
OC: -0.02%
EC' 9 16%
S04: 3.04%
N03: 6.60%
N02: 6.92%
Peak CO: -0.61%
Peak 03: 4.95%
PeakS02:-4.20%
Multiple Pollutants model- Pollutant
with traffic density
N03: 3.42%
S04: -2.73%
EC: 6.27%
Ni:2.51%
V: 3.27%
Pollutant with NO;
EC: 5.93%
Ni:2.31%
. . 0
Pollutant with Peak Oi
Traffic density: 2.40%
EC: 10.79%
Fe: 5.94%
N03: 7.57%
PM2 5:8.97%
V: 4.93%
Ni: 3.65%
S04: 6.75%
Cu: 1.55%
OC:0.21%
December 2009
E-507
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Krewski et al. (2009,
1911931
Period of Study: 1979-2000
Location: 48 contiguous states U.S.
Outcome: Death
Study Design: Cohort
Covariates: Demographic,
socioeconomic and ecologic
characteristics
Statistical Analysis: Cox proportional-
hazards model
Statistical Package: NR
Age Groups: Adults of at least 30 yr
Pollutant: PM25
Averaging Time: NR
Mean Unit:
1979-1983: 21.20 pg/m3
1999-2000:14.02 pg/m3
Range (Min, Max):
1979-1983:10.77-30.01
1999-2000:5.80-22.20
Copollutant:
S042", S02, PM15, TSP, 03, N02, CO
Increment: 10|jg/m
Hazard Ratio (96% Cl)
MSA&DIFF
Increment Change:
10.78(1.043-1.115}
Change 5-15 pg/m
1.128(1.077-1.183)
Change 10-20 pg/m3:
1.079(1.048-1.112)
HR(96%CI)
Los Angeles
Parsimonious ecologic covariates:
1.126(1.014-1.251)
HR(96%CI)
16-yr time window
GroupA: 0.98 (0.92-1.06)
Group B: 1.01 (0.99-1.02)
HR(96%CI)
Third follow-up, 7 Ecologic Variables
1979-1983:1.044
1999-2000:1.057
1.028-1.060
1.036-1.079
HR(96%CI)
Nationwide analysis, 1999-2000
Standard Cox: 1.03 (1.01-1.05)
Random Effects Cox: 1.06 (1.04-1.0
Increment: 1.5|jg/m3
HR(96%CI)
28 County, 3-yr model
All 7 ecologic covariates:
0.977(0.932-1.025)
Reference: Krewski et al. (2009,
1911931
Period of Study: 1979-2000
Location: 48 contiguous states U.S.
Outcome: Death from cardiopulmonary Pollutant: PM25
disease .
Averaging Time: NR
Study Design: Cohort
Mean Unit:
Covariates: Demographic,
socioeconomic and ecologic
characteristics
Statistical Analysis: Cox proportional-
hazards model
Statistical Package: NR
Age Groups: Adults of at least 30 yr
1979-1983:21.2
1999-2000:14.02 pg/m3
Range (Min, Max):
1979-1983:10.77-30.01
1999-2000:5.80-22.20
Copollutant:
S042", S02, PM15, TSP, 03, N02, CO
Increment: 10|jg/m
Hazard Ratio (96% Cl)
MSA&DIFF
Increment Change: 1.078 (1.077-1.182)
Change 5-15 pg/m3:
1.208(1.132-1.290)
Change 10-20 pg/m3:
1.127(1.081-1.174)
HR(96%CI)
Los Angeles
Parsimonious ecologic covariates:
1.086(0.939-1.285)
HR(96%CI)
16-yr time window
Group A: 1.00 (0.90-1.11)
Group 6:1.05(1.03-1.07)
HR(96%CI)
Third follow-up, 7 Ecologic Variables
1979-1983:1.094(1.070-1.118)
1999-2000:1.138 1.106-1.172)
HR(96%CI)
Nationwide analysis, 1999-2000
Standard Cox: 1.09 (1.06-1.12)
Random Effects Cox: 1.13(1.10-1.16)
Increment: 1.5|jg/m3
HR(96%CI)
28 County, 3-yr model
All 7 ecologic covariates: 0
.940(0.875-1.011)
December 2009
E-508
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Krewski et al. (2009,
1911931
Period of Study: 1979-2000
Location: 48 contiguous states U.S.
Outcome: Death from ischemic heart
disease
Study Design: Cohort
Covariates: Demographic,
socioeconomic and ecologic
characteristics
Statistical Analysis: Cox proportional-
hazards model
Statistical Package: NR
Age Groups: Adults of at least 30 yr
Pollutant: PM25
Averaging Time: NR
Mean Unit:
1979-1983: 21.20 pg/m3
1999-2000:14.02 pg/m3
Range (Min, Max):
1979-1983:10.77-30.01
1999-2000:5.80-22.20
Copollutant:
S042", S02, PM15, TSP, 03, N02, CO
Increment: 10|jg/m
Hazard Ratio (96% Cl)
MSA&DIFF
Increment Change: 1.196 (1.177-1.407)
Change 5-15 pg/m3:
1.484(1.311-1.680)
Change 10-20 pg/m3:
1.283(1.186-1.387)
HR(96%CI)
Los Angeles
Parsimonious ecologic covariates:
1.263(10.22-1.563)
HR(96%CI)
Third follow-up, 7 Ecologic Variables
1979-1983:1.184(1.146-1.222)
1999-2000:1.242(1.191-1.295)
HR(96%CI)
Nationwide analysis, 1999-2000
Standard Cox: 1.15 (1.11-1.20)
Random Effects Cox: 1.24(1.19-1.29)
Increment: 1.5|jg/m3
HR(96%CI)
28 County, 3 yr model
All 7 ecologic covariates:
1.072(0.980-1.172)
Reference: Krewski et al. (2009,
1911931
Period of Study: 1979-2000
Location: 48 contiguous states U.S.
Outcome: Death from lung cancer
Study Design: Cohort
Covariates: Demographic,
socioeconomic and ecologic
characteristics
Statistical Analysis: Cox proportional-
hazards model
Statistical Package: NR
Age Groups: Adults of at least 30 yr
Pollutant: PM25
Averaging Time: NR
Mean Unit:
1979-1983: 21.20 pg/m3
1999-2000:14.02 pg/m3
Range (Min, Max):
1979-1983:10.77-30.01
1999-2000:5.80-22.20
Copollutant:
S042", S02, PM15, TSP, 03, N02, CO
Increment: 10|jg/m
Hazard Ratio (96% Cl)
MSA&DIFF
Increment Change: 1.142 (1.057-1.234)
Change 5-15 pg/m3:
1.236(1.114-1.372)
Change 10-20 pg/m3:
1.143(1.071-1.221)
HR(96%CI)
Los Angeles
Parsimonious ecologic covariates:
1.311 (0.897-1.915)
HR(96%CI)
16-yr time window
GroupA: 1.08 (0.87-1.35)
Group 6:1.07(1.02-1.13)
HR(96%CI)
Third follow-up, 7 Ecologic Variables
1979-1983:1.092
1.033-1.154
1.057-1.225
1999-2000:1.138
HR(96%CI)
Nationwide analysis, 1999-2000
Standard Cox: 1.11 (1.04-1.18)
Random Effects Cox: 1.14(1.06-1.23)
Increment: 1.5|jg/m3
HR(96%CI)
28 County, 3-yr model
All 7 ecologic covariates:
0.985(0.832-1.166)
Reference: Krewski et al. (2009,
1911931
Period of Study: 1979-2000
Location: 48 contiguous states U.S.
Outcome: Death from diabetes
Study Design: Cohort
Covariates: Demographic,
socioeconomic and ecologic
characteristics
Statistical Analysis: Cox proportional-
hazards model
Statistical Package: NR
Age Groups: Adults of at least 30 yr
Pollutant: PM25
Averaging Time: NR
Mean Unit:
1979-1983: 21.20 pg/m3
1999-2000:14.02 pg/m3
Range (Min, Max):
1979-1983:10.77-30.01
1999-2000:5.80-22.20
Copollutant:
S042", S02, PM15, TSP, 03, N02, CO
Increment: 1.5|jg/m
HR(96%CI)
28 County, 3 yr model
All 7 ecologic covariates:
1.083(0.723-1.621)
December 2009
E-509
-------
Study
Reference: Krewski et al. (2009,
1911931
Period of Study: 1979-2000
Location: 48 contiguous states U.S.
Reference: Krewski et al. (2009,
1911931
Period of Study: 1979-2000
Location: 48 contiguous states U.S.
Reference: Krewski et al. (2009,
1911931
Period of Study: 1979-2000
Location: 48 contiguous states U.S.
Reference: Krewski et al. (2009,
1911931
Period of Study: 1979-2000
Location: 48 contiguous states U.S.
Reference: McDonnell et al. (2000,
0103191
Period of Study: 1973-1977
Location: California
Design & Methods
Outcome: Death from endocrine
disease
Study Design: Cohort
Covariates: Demographic,
socioeconomic and ecologic
characteristics
Statistical Analysis: Cox proportional-
hazards model
Statistical Package: NR
Age Groups: Adults of at least 30 yr
Outcome: Death from digestive cancer
Study Design: Cohort
Covariates: Demographic,
socioeconomic and ecologic
characteristics
Statistical Analysis: Cox proportional-
hazards model
Statistical Package: NR
Age Groups: Adults of at least 30 yr
Outcome: Death cancers other than
lung and digestive
Study Design: Cohort
Covariates: Demographic,
socioeconomic and ecologic
characteristics
Statistical Analysis: Cox proportional-
hazards model
Statistical Package: NR
Age Groups: Adults of at least 30 yr
Outcome: Deaths from causes other
than CPD, IHD and lung cancer
Study Design: Cohort
Covariates: Demographic,
socioeconomic and ecologic
characteristics
Statistical Analysis: Cox proportional-
hazards model
Statistical Package: NR
Age Groups: Adults of at least 30 yr
Outcome: Mortality
Study Design: Cohort (AHSMOG
airport cohort)
Statistical Analyses: Cox regression
models
Concentrations1
Pollutant: PM25
Averaging Time: NR
Mean Unit:
1979-1983: 21.20 pg/m3
1999-2000:1 4.02 pg/m3
Range (Min, Max):
1979-1983: 10.77-30.01
1999-2000:5.80-22.20
Copollutant:
S042", S02, PM15, TSP, 03, N02, CO
Pollutant: PM25
Averaging Time: NR
Mean Unit:
1979-1983: 21.20 pg/m3
1999-2000:1 4.02 pg/m3
Range (Min, Max):
1979-1983: 10.77-30.01
1999-2000:5.80-22.20
Copollutant:
S042", S02, PM15, TSP, 03, N02, CO
Pollutant: PM25
Averaging Time: NR
Mean Unit:
1979-1983: 21.20 pg/m3
1999-2000:1 4.02 pg/m3
Range (Min, Max):
1979-1983: 10.77-30.01
1999-2000:5.80-22.20
Copollutant:
S042", S02, PM15, TSP, 03, N02, CO
Pollutant: PM25
Averaging Time: NR
Mean Unit:
1979-1983: 21.20 pg/m3
1999-2000:1 4.02 pg/m3
Range (Min, Max):
1979-1983: 10.77-30.01
1999-2000:5.80-22.20
Copollutant:
S042", S02, PM,5, TSP, 03, N02, CO
Pollutant: PM25
Averaging Time: Monthly avg
Mean (SD): 31. 9 (10.7)
IQR: 24.3
Effect Estimates (95% Cl)
Increment: 1.5|jg/m3
HR(96%CI)
28 County, 3-yr model
All 7 ecologic covariates:
1.143(0.835-1.564)
Increment: 10|jg/m3
HR(96%CI)
Los Angeles
Parsimonious ecologic covariates:
1.199(0.817-1.758)
Increment: 10|jg/m3
HR(96%CI)
Los Angeles
Parsimonious ecologic covariates:
1.012 (0.788-1.299)
Increment: 10|jg/m3
Hazard Ratio (96% Cl)
MSA&DIFF
Increment Change: 1.010(0.968-1.055)
Change 5-1 5 pg/m3:
1.026(0.970-1.085)
Change 1 0-20 pg/m3:
1.016(0.981-1.053)
HR(96%CI)
Third follow-up, 7 Ecologic Variables
1979-1983:0.983(0.960-1.007)
1999-2000: 0.953 (0.923-0.984)
Increment: IQR
All Cause: 1.22 (0.95-1. 58)
Resp: 1.64 (0.93-2.90)
Lung Cancer: 2.23 (0.56-8.94)
Age Groups: Males, 27 yr+
Copollutants (correlation):
03: 0.68
S02:0.18
N02: -0.08;
S04: 0.33
December 2009
E-510
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Miller et al. (2007, 0901301
Period of Study: 1994-1998
Location: 36 U.S. Metropolitan Areas
Outcome: CVD Mortality
Study Design: Prospective Cohort
(WHI)
Statistical Analyses: Cox proportional
hazards regression
Age Groups: Postmenopausal women
ages 50-79
Pollutant: PM25
Averaging Time: Annual avg (2000)
Mean (SD): 13.4
IQR: 11.6, 18.3
Range: 3.4, 28.3
Increment: 10|jg/m
CVD Death: 1.76 (1.25, 2.47)
CHD Death: 2.21 (1.17, 4.16)
CV Death: 1.83 (1.11, 3.00)
Reference: Naess et al. (2007,
0907361
Period of Study: 1992-1998
Location: Oslo, Norway
Outcome: Mortality: Nonaccidental
(<800)
Lung cancer(162)
COPD (490-496)
Cardiovascular (390-459)
Study Design: Prospective Cohort
Statistical Analyses: Cox proportional
hazards regression model
Age Groups: 51-70, 71-90
Pollutant: PM25
Averaging Time: 4-yr avg
Mean(SD):PM25:15
Range (Min, Max): PM25: (7, 22)
Copollutant (correlation):
N02:r = 0.95
Relative Risk (Cl min, Cl max)
RR for deaths from all causes
Men (ages 51-70) PM25 exposure
(in pg/m3)
6.56-11.45:1.00
11.46-14.25:0.96(0.89, 1.04)
14.26-18.43:1.12 1.03, 1.22
18.44-22.34:1.48 1.36, 1.60
Men (ages 71-90) PM25 exposure
(in fjg/m )
6.56-11.45:1.00
11.46-14.25:0.99(0.93, 1.06)
14.26-18.43:1.10 1.03, 1.17
18.44-22.34:1.19 1.12,1.27
Wfomen (ages 51-70) PM25 exposure
(in fjg/m )
6.56-11.45:1.00
11.46-14.25:0.96(0.87, 1.07)
14.26-18.43:1.08 0.98, 1.20
18.44-22.34:1.44 1.30, 1.59
Wfomen (ages 71 -90) PM25 exposure
(in fjg/m )
6.56-11.45:1.00
11.46-14.25:1.03(0.97, 1.09)
14.26-18.43:1.07(1.01, 1.12)
18.44-22.34:1.11(1.05,1.16)
Increment: 10 pg/m3
RR for death from CVD and lung cancer
Men (ages 51-70)
CVD-PM25:1.11 (1.06, 1.16)
COPD-PM25:1.32 (1.17, 1.49)
Lung Cancer-PM25:1.07 (0.98,1.17)
V\fomen(ages51-70)
CVD: PM25:1.16 (1.09, 1.24)
COPD: PM25:1.18 (1.03, 1.34)
Lung Cancer: PM25:1.23 (1.10,1.37)
Men (ages 71-90)
CVD: PM25:1.06 (1.03, 1.09)
COPD: PM,0:1.13 (1.04, 1.24)
PM25:1.14 (1.04, 1.24)
Lung Cancer: PM25:1.08 (0.98,1.19)
V\fomen(ages71-90)
CVD: PM25:1.02 (1.00, 1.05)
COPD: PM25:1.09 (1.00, 1.18)
Lung Cancer: PM25:1.16 (1.03,1.31)
Reference: Naess et al. (2007,
0907361
Period of Study: 1992-1998
Location: Oslo, Norway
Outcome: Mortality: Lung cancer (162)
COPD (490-496)
Cardiovascular (390-459)
Psychiatric causes (290, 292-302, 304,
306-319)
Stomach cancer (151)
Violence (800-999)
Study Design: Multilevel cohort
Statistical Analyses: WinBUGS
Age Groups: 50-74
Pollutant: PM25
Averaging Time: (Mo-yr) avg
Range Mean (SD): 14.2 (3.6)
IQ Range (1st, 4th): (6.6, 22.3)
Copollutant (correlation):
PM10:r = 0.95|
N02:r = 0.87
Relative Risk (Cl min, Cl max)
RR on All-cause mortality of PMz; in
Men Age 60-74
Primary Education:
PM25:1.06 (1.00,1.11)
Individual: 1.34 (1.24,1.43)
Neighborhood: 1.22 (1.16,1.28)
Manual Class: PM25:1.06 (1.01,1.12)
Individual: 1.28 (1.20,1.37)
Neighborhood: 1.20 (1.14,1.26)
Income below median:
PM25:1.05 (1.00, 1.12)
Individual: 1.44 (1.35,1.53)
Neighborhood: 1.16 (1.11,1.21)
Not owner occupied:
PM25:1.06 (1.00, 1.13)
Individual: 1.24 (1.12,1.36)
Neighborhood: 1.11 (1.05,1.17)
Lives in flat dwelling:
December 2009
E-511
-------
Study Design & Methods Concentrations1 Effect Estimates (95% Cl)
PM25:1.04(0.98, 1.11)
Individual: 1.19 (1.09,1.31)
Neighborhood: 1.10 (1.04,1.17)
More than one person per room in
dwelling: PM25:1.10 (1.02,1.18)
Individual: 1.05 (0.98,1.13)
Neighborhood: 1.01 (0.96,1.05)
RR on All-cause mortality of PWh.s in
Women Age 60-74
Primary Education Only:
PM25:1.05 (1.00,1.11)
Individual: 1.32 (1.23,1.42)
Neighborhood: 1.18 (1.12,1.24)
Manual Class: PM25:1.07 (1.01,1.13)
Individual: 1.27 (1.18,1.36)
Neighborhood: 1.18 (1.12,1.24)
Income below median:
PM25:1.05 (1.01,1.10)
Individual: 1.52 (1.41,1.63)
Neighborhood: 1.13 (1.09,1.18)
Not owner occupied:
M25:1.07 (1.01,1.14)
Individual: 1.24 (1.12,1.38)
Neighborhood: 1.08 (1.02,1.14)
Lives in a flat dwelling:
PM25:1.05 (0.99, 1.11)
Individual: 1.21 (1.09,1.34)
Neighborhood: 1.09 (1.02,1.15)
More than one person per room in
dwelling: PM25:1.11 (1.04,1.19)
Individual: 1.07 (0.99,1.14)
Neighborhood: 1.01 (0.96,1.05)
RRfor Interquartile Increase (Ml) in
PM2 5 for different causes of death
CVD:
Age and sex adjusted: 1.11 (1.07,1.15)
Primary education only:
M1+Individual: 1.07 (1.04,1.11)
M1+Neighborhood: 1.03 (1.00,1.07)
Manual Class: M1+ Individual: 1.08
(1.04,1.11)
M1+Neighborhood: 1.06 (1.02,1.10)
Income below Median: M1+ Individual:
1.07(1.03,1.11)
M1+Neighborhood: 1.02 (0.98,1.05)
Not owner occupied:
M1+Individual: 1.05 (1.01,1.09)
M1+ Neighborhood: 1.03(0.99,1.07):
Living in a Flat dwelling
M1+Individual: 1.04 (1.00,1.08)
M1+Neighborhood: 1.01 (0.97,1.05)
Crowded household:
M1+Individual: 1.10 (1.05,1.14)
M1+Neighborhood: 1.10 (1.06,1.15)
Pulmonary Cancer: Age and sex
adjusted: 1.12 (1.05,1.19)
Primary education only:
M1+Individual: 1.09 (1.01,1.17)
M1+Neighborhood: 1.05 (0.98,1.13)
Manual Class:
M1+Individual: 1.09 (1.01,1.17)
M1+Neighborhood: 1.10 (1.06,1.13)
Income below Median:
M1+Individual: 1.09 (1.01,1.17)
M1+Neighborhood: 1.02 (0.95,1.10)
Not owner occupied:
M1+Individual: 1.07 (1.00,1.15)
M1+Neighborhood: 1.04 (0.97,1.12)
Living in a Flat dwelling:
M1+Individual: 1.03 (0.96,1.11)
M1+Neighborhood: 1.00 (0.92,1.08)
Crowded household:
M1+Individual: 1.10 (1.03,1.14)
M1+Neighborhood:1.11 (1.04,1.20)
COPD: Age and sex adjusted:
1.17(1.09,1.25)
December 2009 E-512
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Primary education only:
M1+Individual: 1.13 (1.05,1.22)
M1+Neighborhood: 1.09 (1.01,1.19)
Manual Class:
M1+Individual: 1.14 (1.05,1.23)
M1+Neighborhood: 1.12 (1.04,1.22)
Income below Median:
M1+Individual: 1.13 (1.04,1.22)
M1+Neighborhood: 1.06 (0.97,1.15)
Not owner occupied: M1+ Individual:
1.10(1.02,1.19)
M1+Neighborhood: 1.07 (0.99,1.16)
Living in a Flat dwelling:
M1+Individual: 1.08 (1.00,1.18)
M1+Neighborhood: 1.03 (0.95,1.13)
Crowded household:
M1+Individual: 1.16 (1.07,1.26)
M1+Neighborhood: 1.16 (1.07,1.26)
Estimates for psychiatric diseases,
genetic cancer and violent death
Reference: Nerriere et al. (2005,
0886301
Period of Study:
Grenoble (2001)
Paris (2002)
Rouen (2002-2003)
Strasbourg (2003)
Location: Four French Cities-
Grenoble, Rouen, Paris, and
Strasbourg
Outcome: Mortality: Lung Cancer (162) Pollutant: PM25
Study Design: Time-series Averaging Time: 48-h avg
Statistical Analyses: GIS
Age Groups: 30-71 yr old nonsmoking
adults
Mean Range: 17 to 49 pg/m3
Increment: 10|jg/m
% Increase (Lower Cl, Upper Cl)
% increase in lung cancer deaths
attributable to PM2 5 exposure
France: 8(1,16)
Grenoble: 10 (3,19)
Rouen: 10 (2,19)
Strasbourg: 24 (4, 40)
Reference: Ozkaynak and Thurston
(1987, 0729601
Period of Study: 1980
Location: U.S.
Outcome: Total Mortality
Study Design: Cross-sectional
Statistical Analyses: Multiple
regression analysis
Pollutant: Sulfate
Averaging Time: Annual avg
Mean Range: Sulfate: 11.1 (3.5)
Range of estimated total mortality
effects of air pollutions:
Sulfate: 4-9%
"Sulfate concentration was consistently
found to be a significant predictor of
mortality in the models considered. Fine
particle mass coefficients were also
often found to be statistically significant
in the mortality regressions."
Reference: Pope et al. (2004, 0558801
Period of Study: 1982-2000
Location: Metropolitan areas in all 50
states in the U.S.
Outcome: Mortality: Cardiovascular
Diseases (390-459)
Diabetes (250)
Respiratory Disease (460-519)
Study Design: Prospective Cohort
Statistical Analyses: Cox proportional
hazards regression
Age Groups: >30
Pollutant: PM25
Averaging Time: Annual avg
Mean (SD): 17.1 (3.7)
Range (Min, Max): NR
Increment: 10|jg/m
Relative Risk (Lower Cl, Upper Cl)
All cardiovascular disease plus
diabetes: PM25:1.12 (1.08,1.15)
FormerSmoker: 1.26 (1.23,1.28)
Current Smoker: 1.94 (1.90,1.99)
Ischemic Heart Disease: PM25:1.18
(1.14,1.23)
Former Smoker: 1.33 (1.29,1.37)
Current Smoker: 2.03 (1.96, 2.10)
Diabetes: PM25: 0.99 (0.86,1.14)
Former Smoker: 1.05 (0.94,1.16)
Current Smoker: 1.35 (1.20,1.53)
All other Cardiovascular Diseases:
PM25: 0.84 (0.71, 0.99)
Former Smoker: 1.22 (1.09,1.38)
Current Smoker: 1.78 (1.56, 2.04)
Diseases of the respiratory system:
PM25: 0.92 (0.86, 0.98)
Former Smoker: 2.16 (2.04, 2.28)
Current Smoker: 3.88 (3.66, 4.11)
COPD: PM25: 0.84 (0.77, 0.93)
Former Smoker: 4.93 (4.48, 5.42)
Current Smoker: 9.85 (8.95,10.84)
All other respiratory diseases: PM25:
0.86(0.73, 1.02)
Former Smoker: 1.54 (1.36,1.74)
Current Smoker: 1.83 (1.57, 2.12)
December 2009
E-513
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Pope et al. (2007, 0912561
Period of Study: 1960-1975
Location: New Mexico, Arizona, Utah,
and Nevada
Outcome (ICD7&8):
Mortality: Cardiovascular (ICD 7: 400-
468, 331, 332 ICD 8: 390-458)
Respiratory (ICD 7: 470-527 ICD 8:
460-519)
Influenza/ pneumonia (ICD 7: 480-483,
490-493, ICD 8: 470-474, 480-486)
Study Design: Retrospective Cohort
Statistical Analyses: Poisson
regression model
GAM
SAS
Age Groups: All smelter workers >18
Pollutant: PM25
Averaging Time: 24-h avg
Mean (SD): NR
Range (Min, Max): NR
The study does not present quantitative
results
Results are presented in figures. The
References found that the strike-related
estimated percent decrease in mortality
was 2.5% (1.1-4.0),
Reference: Pope et al. (2009,1901071
Period of Study: 1978-1982,
1997-2001
Location: 211 U.S. counties and 51
metropolitan areas
Outcome: Increased life expectancy
Study Design: Cross-sectional
Statistical Analysis: Cross-sectional
regression
Age Groups: Adults >45 yr
Pollutant: PM25
Averaging Time: Daily, quarterly and
annual
Mean (SD) Unit:
1979-1983: 20.61+ 4.36 pg/m3
1999-2000:14.10 + 2.86 pg/m3
Range (Min, Max): NR
Copollutant (correlation): NR
Increment: 10|jg/m
Regression Coefficient + SD
211 County Units
Intercept: 1.75 ±0.27
Reduction in PM25: 0.61 ±0.20
Change in Income: 0.13 ±0.01
Change in Population: 0.06 ± 0.02
Change in Black Population:
-2.70 ±0.64
Change in Lung Cancer Mortality Rate:
-0.06 ±0.02
Change in COPD Mortality Rate:
-0.08 ±0.02
R: 0.53
51 Metropolitan Areas
Intercept: 2.09 ±0.36
Reduction in PM25: 0.95 ± 0.23
Change in Income: 0.11 ±0.02
Change in Population: 0.05 ± 0.02
Change in Black Population:
-5.98 ±1.99
Change in Lung Cancer Mortality Rate:
0.02 ± 0.03
Change in COPD Mortality Rate:
-0.19 ±0.05
R: 0.74
December 2009
E-514
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Rainham et al. (2005,
0886761
Period of Study: 1981-1999
Location: Toronto, Canada
Outcome: Total deaths (ICD9 <800), Pollutant: PM2
cardiorespiratory
cardiorespiratory
390-459), non-
ICD9-NR)
Study Design: Time-series
Statistical Analyses: Generalized
linear models were used
Season: Winter (Dec-Feb)
Summer (Jun-Aug)
Statistical Package: S-Plus 6.1
Averaging Time: NR
Mean (SD):
All yr: 17.0 (8.7) pg/m3
Winters: 17.2 (6.8)
Summers: 18.8(10.2)
AvgWnter values: Dry Moderate: 17.0
(1.0)
Dry Polar: 17.5 (0.5)
Dry Tropical: No Comparison
Moist Moderate: 17.1 (0.8)
Moist Polar: 17.5 (0.6)
Moist Tropical: 16.5(3.6)
Transition: 16.7 (1.0)
Avg summer values: Dry Moderate:
18.4(0.9)
Dry Polar: 19.0 (1.2)
Dry Tropical: 18.5(2.4)
Moist Moderate: 19.2 (1.2)
Moist Polar: 17.5 (2.0)
Moist Tropical: 19.8 (1.1)
Transition: 17.6 (1.5)
Mortality risk for winter season and
within winter synoptic weather
categories
RR Estimate [Lower Cl, Upper Cl]:
Winter: Total: 0.998(0.997,1.000]
Cardioresp: 0.998(0.996,1.000]
Other: 0.998 [0.996, 1.000]
Dry Moderate:
Total: 1.001(0.996, 1.007]
Cardioresp: 1.005(0.998,1.011]
Other: 1.002 [0.998,1.005]
Dry Polar: Total: 0.998(0.995,1.001]
Cardioresp: 0.995(0.991, 0.999]
Other: 1.002 [0.998,1.005]
Dry Tropical: NA
Moist Moderate:
Total: 0.998(0.993, 1.002]
Cardioresp: 1.003(0.995,1.010]
Other: 0.997 [0.991,1.004]
Moist Polar: Total: 1.001(0.998,1.005]
Cardioresp: 1.002(0.997,1.007]
Other: 1.003 [0.999,1.007]
Moist Tropical:
Total: 1.007(0.965, 1.203]
Cardioresp: 1.123(1.031,1.224]
Other: 1.248 [1.123,1.387]
Transition Total: 1.003(0.996,1.009]
Cardioresp: 0.996(0.987,1.004]
Other: 0.997 [0.990, 1.004]
Mortality risk for summer season and
within summer synoptic weather
categories
RR Estimate [Lower Cl, Upper Cl]:
Summer: Total: 1.000(1.000,1.001]
Cardioresp: 1.001 [1.000,1.002]
Other: 1.001 [1.000,1.002]
Dry Moderate:
Total: 1.001(0.999, 1.002]
Cardioresp: 1.002(0.999,1.004]
Other: 0.999(0.997, 1.002]
Dry Polar: Total: 1.002(0.999,1.005]
Cardioresp: 0.996(0.991,1.000]
Other: 1.003 0.999,1.007]
Dry Tropical: Total: 1.016(1.006,1.027]
Cardioresp: 1.017(1.005,1.030]
Other: 1.017 [1.003,1.031]
Moist Moderate:
Total: 1.002(1.000, 1.004]
Cardioresp: 1.003(0.999,1.006]
Other: 1.004 [1.001,1.006]
Moist Polar:
Total: 1.005(0.998, 1.011]
Cardioresp: 1.008(0.997,1.018]
Other: 1.003 [0.995,1.011]
Moist Tropical:
Total: 0.999(0.997, 1.001]
Cardioresp: 0.996(0.993,1.000]
Other: 0.998 [0.995,1.001]
Transition: Total: 1.005(0.996,1.014]
Cardioresp: 1.007(0.994,1.020]
Other: 1.002 [0.996,1.008]
Reference: Roman et al. (2008,
1569211
Period of Study: 2006
Location: U.S.
Outcome: Mortality
Study Design: Expert Judgment Study
Statistical Analyses: Standard best
practices for expert elicitation
Pollutant: PM25
Averaging Time: Annual avg
Mean (SD): 4-30
Quantitative results are not presented in
the text, but can be found graphically in
Fig 3.
"Most of the experts' central estimates
fall at or above the 2002 ACS median
(0.6% per pg/m3) and below the original
Six Cities median (1.2% per pg/m )."
December 2009
E-515
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Schwartz, et al. (2008,
156921
Outcome: Mortality
Pollutant: PM2
PM Increment: 10 pg/m
Period of Study: 1979-1988
Location: Six U.S. metropolitan areas:
Boston, Massachusetts
Knoxville, Tennessee
St. Louis, Missouri
Steubenville, Ohio
Madison, Wisconsin
and Topeka, Kansas
Study Design: Poisson regression with
GAM
Statistical Analyses: Weighted linear
regression
Season: all
Dose-response Investigated? No
Statistical Package: S-plus
Averaging Time: Daily
Mean (SD):
Boston-16.5
Knoxville-21.1
St. Louis-19.2
Steubenville-30.5
Madison-11.3
Topeka-12.2
SD not reported
Range (Min, Max): (0,35)
Monitoring Stations: 6
The difference between mean PM25
concentrations of 10 pg/m3 and 20
pg/m3 is associated with about a 1.5%
increase in deaths.
Reference: (Schwartz et al, 2008,
1569631
Period of Study: 1979-1998
Location: Watertown, MA
Kingston and Harriman, TN
St Louis, MO
Steubenville, OH
Portage, Wyocena
Pardeeville Wl
Topeka, KS
Outcome: Mortality: Nonaccidental
(<800)
Study Design: Cross-sectional
Statistical Analyses: Cox proportional
hazards regression
penalized splines
Bayesian Model Averaging
Age Groups: >18
Pollutant: PM25
Averaging Time: Annual avg
Mean (SD): 17.5 (6.8)
Range (Min, Max): (8, 40)
Increment: 10 pg/m
Relative Risk (Lower Cl, Upper Cl)
Estimated from Fig 4:
All Cause Mortality - Year before Death
0:1.10(1.00,1.21)
1:1.03(0.98, 1.08)
2:1.01 (1.00, 1.02)
3:1.00(0.99,1.01)
4:1.00(0.99,1.01)
5: 1.00
Lung Cancer Mortality - Year Before
Death
Estimated from Fig 6
0:1.18(1.00,1.48)
1:1.12(0.98, 1.33)
2:1.08 0.92, 1.22)
3:1.02(1.01,1.03)
4:1.01(1.00,1.02)
5: 1.01
RR per 10 pg/m3 increase of PM25
exposure
Level Of Increase
Estimated from Fig 3
10 pg/m3:1.15
20 pg/m3:1.29
30 pg/m3:1.46
40 pg/m3:1.64
Reference: Tainio et al. (2005, 0874441 Outcome (ICD10): Mortality: Pollutant: PM25
Cardiopulmonary (111-170 and J15-J47)
Averaging Time: 24-h avg
Period of Study: 1997-Present
Location: Helsinki, Finland
Study Design: Time-series simulation
Statistical Analyses: Monte Carlo
Simulation
Age Groups: All ages
Mean (SD): 10.7
Range (Min, Max): NR
Estimated Deaths Per Year (Min Cl,
Max Cl) Associated with Primary
PM2.s
Emissions from buses in Helsinki in
2020 for different bus strategies
Cardiopulmonary Mortality
Current Fleet: 15.9 (0,46.6)
Modern Diesel: 7.9 (0, 23.0)
Diesel with particle trap: 3.9 (0,12)
Natural gas bus: 2.3 (0, 6.8)
Lung Cancer Mortality
Current Fleet: 2.2 (0,6.1)
Modern Diesel: 1.1 (0,3.0)
Diesel with particle trap: 0.6 (0,1.6)
Natural gas bus: 0.3 (0, 0.9)
Total Mortality
Current Fleet: 18.1 (0,55.0)
Modern Diesel: 9.0 (0, 27.0)
Diesel with particle trap: 4.4 (0,14.1)
Natural Gas Bus: 2.6 (0, 8.0)
December 2009
E-516
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Villeneuve et al. (2002,
0425761
Period of Study: 1974-1991
Location: Six U.S. Cities: Steubenville,
OH, St. Louis, MO, Portage, Wl,
Topeka, KS, Watertown, MA, Kingston/
Harriman, TN
Outcome (ICD10): Mortality:
Nonaccidental (<800)
Study Design: Prospective Cohort
Statistical Analyses: Poisson,
EPICURE
Age Groups: All ages
<60
>60
Pollutant: PM25
Averaging Time: 24-h avg
Mean (SD): Portage: 10.9 (7.2)
Topeka: 12.1 (7.1)
Harriman: 20.7 (9.4)
Watertown: 14.9 (8.4)
St. Louis: 18.7 (10.6)
Steubenville: 28.6 (21.0)
Overall: 18.6
Range (Min, Max): NR
Increment: 18.6 pg/m
Relative Risk (Min Cl, Max Cl)
RR of all cause mortality for exposure of
PM2 5 by age group
Exposure to PM25 remained fixed over
entire study period
<60:1.89 (1.32, 2.69)
>60:1.21 (1.02, 1.43)
Total: 1.31 (1.12,1.52)
Exposure to PM25 was defined
according to 13 calendar periods* (no
smoothing)
<60:1.52 (1.15, 2.00)
>60:1.11 (0.95, 1.29)
Total: 1.19 (1.04,1.36)
Exposure to PM25 was defined
according to 13 calendar periods*
(smoothed)
<60:1.43 (1.10, 1.85)
>60:1.09 (0.93, 1.26)
Total: 1.16 (1.02,1.32)
Time dependent estimate of PM2 5
received during the previous 2 yr
<60:1.42 (1.09, 1.82)
>60:1.08 (0.94, 1.25)
Total: 1.16 (1.02,1.31)
Time dependent estimate of PM2 5
received 3-5 yr before current yr
<60:1.35 (1.08, 1.67
>60:1.08 (0.95, 1.22
Total: 1.14 (1.02,1.27)
Time dependent estimate of PM2 5
received >5 yr before current yr
<60:1.34 (1.11, 1.59)
>60:1.09 (0.99, 1.20)
Total: 1.14 (1.05,1.23)
* The calendar periods used were:
1970-1978, 1979, 1980, 1981, 1982,
1983, 1984, 1985, 1986, 1987, 1988,
1989, and 1990+.
RR of all cause mortality and PM25
exposure by city
Portage: 1.16 (0.96,1.39)
Topeka: 1.06 (0.89,1.27)
Harriman
Men: 1.04 (0.79,1.36)
Women: 0.96 (0.69,1.31)
All: 1.13 (0.95, 1.35)
Watertown
Men: 1.20 (0.95,1.51)
Women: 1.06 (0.78,1.43)
All: 1.32 (1.11,1.51)
St. Louis
Men: 0.97 (0.76, 1.24)
Women: 1.13 (0.86,1.49)
Steubenville
Men: 1.39 (1.11,1.74)
Women: 1.22 (0.93,1.61)
Reference: Willis et al. (2003, 0899221
Period of Study: 1982-1989
Location: U.S. Metropolitan areas in all
50 states
Outcome: Mortality: All causes
Lung Cancer (162)
Cardiopulmonary (401-440, 460-519)
Study Design: Prospective Cohort
Statistical Analyses: Cox proportional
hazards model
Age Groups: All ages
Pollutant: Sulfates
Averaging Time: Annual avg
Mean (SD): 10.6 pg/m3
Range (Min, Max): 3.6, 23.5
Copollutant: CO, N02, 03, S02
All Cause, Metropolitan Scale: 1.25
(1.13, 1.37)
All Cause, County Scale: 1.50 (1.30,
1.73)
CPD, Metropolitan Scale: 1.29 (1.15,
1.46)
CPD, County Scale: 1.75 (1.48, 2.08)
Reference: Zanobetti and Schwartz Outcome: Mortality, all causes,
Pollutant: PM25
Increment: 10|jg/m
Period of Study: 1999-2005
Location: 112 U.S. Cities
Study Design: Time-series
Covariates: Region, season
Averaging Time: 24 h
Mean (SD)
Birmingham AL- 16.5
Phoenix AZ- 11. 4
LittleRockAR-14.3
Percent Increase (96% Cl) in
mortality by increment of PM2 5,
combined by season
All Cause Mortality
Overall: 0.98 (0.75-1. 22)
December 2009
E-517
-------
Study Design & Methods Concentrations1
Statistical Analysis: Poisson Łres no CA - 19.4
recession BakersfieldCA-217
regresslon LosAngelesCA-19.9
Age Groups1 All Anaheim CA- 16.3
RubidouxCA-24.9
Sacramento CA- 13.0
EICajonCA-13.5
Denver CO -10. 3
Hartford CT-1 1.6
New HavenCT -13. 7
Wilmington DE- 15.1
Davie FL - 8.4
Miami FL- 9.4
Jacksonville FL- 10.6
PensacolaFL-12.4
Tampa FL- 11. 9
Orlando FL-1 0.3
Palm beach FL- 7.9
PinellasFL-10.4
Atlanta GA- 17. 6
Chicago IL- 15.9
Gary IN -15.3
Indianapolis IN -16.3
Cedar Rapids IA- 11.0
DesMoineslA-10.5
Davenport IA- 12.3
Louisville KY- 15.9
Baton Rouge LA -13. 4
AvondaleLA-12.3
New Orleans LA -12.6
Baltimore MD- 15.6
Springfield MA -12.3
Boston MA -12.4
Worcester MA -11. 3
Holland Ml -12.1
Grand Rapids Ml -13.6
Detroit Ml -16.2
Minneapolis MN- 11.1
Kansas MO -12.0
St Louis MO -14.5
Omaha NE- 10.4
Elizabeth NJ - 14.7
Albuquerque NM - 6.7
New York NY -14.8
Bath NY -9.6
Durham NC- 14.3
WnstonNC-14.7
GreensboroughNC-14.2
Charlotte NC- 15. 3
Raleigh NC- 14.3
MiddletownOH-16.4
YoungstownOH-15.6
Cleveland OH -16.4
Columbus OH -16.2
Cincinnati OH -17.1
SteubenvilleOH-17.0
Toledo OH -14.9
Dayton OH -16.2
Akron OH - 16.0
Warren OH -15.3
Oklahoma OK - 9.9
TulsaOK-11.1
Bend OR - 7.8
MedfordOR-9.9
KlamathOR-10.6
Eugene OR - 8.0
Portland OR - 8.8
Gettysburg PA -13. 4
Pittsburgh PA- 15.7
State College PA -13.2
Carlisle PA -15.1
Harrisburg PA- 15.5
Erie PA -13.1
ScrantonPA- 11.8
AllentownPA-14.2
WlkesBarrePA-12.8
Mercer PA -14.1
EastonPA-14.0
Effect Estimates (95% Cl)
Wnter: 0.56 (0.17-0.94)
Spring: 2.57 (1.96-3. 19)
Summer: 0.25 (-0.1 3-0.63)
Fall: 0.95 (0.56-1. 34)
CVD
Overall: 0.85 (0.46-1. 24)
Wnter: 0.70 (0.04-1. 36)
Spring: 2. 18 (1.22-3. 15)
Summer: -0.03 (-0.75-0.69)
Fall: 0.92 (0.17-1. 68)
Ml
Overall: 1.18 (0.48-1. 89)
Wnter: 1.29 (-0.14-2.75)
Spring: 2. 12 (0.53-3.74)
Summer: -0.03 (-1.46-1. 42)
Fall: 1.24 (0.12-2.36)
Stroke
Overall: 1.78 (0.96-2.62)
Wnter: 1.93 (0.34-3.54)
Spring: 2.04 (-0.02-4.13)
Summer: 1.64 (0.05-3.26)
Fall: 1.69 (0.06-3.35)
Respiratory
Overall: 1.68 (1.04-2.33)
Wnter: 0.86 (-0.16-1.88)
Spring: 4.62 (3.08-6. 18)
Summer: 0.78 (-0.49-2.06)
Fall: 1.45 (0.19-2.72)
Percent Increase (96% Cl) in
mortality by increment in PM25
combined by region
All Cause Mortality
Humid Subtropical and Maritime:
1.02(0.65-1.38)
Warm Summer Continental:
1.19(0.73-1.64)
Hot Summer Continental:
1.14(0.55-1.73)
Dry: 1.1 8 (-0.70-3. 10)
Dry, Continental: 1.26 (-0.21-2.76)
Mediterranean: 0.50 (0.00-1.01)
CVD
Humid Subtropical and Maritime:
0.78(0.05-1.51)
Warm Summer Continental:
1.43(0.67-2.19)
Hot Summer Continental:
0.43 (-0.53-1. 40)
Dry: 3.11 (-0.02-6.33)
Dry, Continental: 1.67 (-0.75-4.16)
Mediterranean: 0.16 (-0.46-0.79)
Ml
Humid Subtropical and Maritime:
0.97 (-0.29-2.26)
Warm Summer Continental:
1.50(0.05-2.97)
Hot Summer Continental:
0.64 (-0.96-2.28)
Dry: 4.25 (-2.38-11. 33)
Dry, Continental: 0.60 (-7.42-9.32)
Mediterranean: 1.85 (-0.66-4.41)
Stroke
Humid Subtropical and Maritime:
2.94(1.59-4.32)
Warm Summer Continental:
1.85(0.04-3.69)
Hot Summer Continental:
0.77 (-1.77-3.38)
Dry: 1.82 (-6.98-11. 45)
Dry, Continental: 2.49 (-2.32-7.53)
Mediterranean: 0.95 (-0.66-2.59)
Respiratory
Humid Subtropical and Maritime:
0.91 (-0.25-2.08)
Warm Summer Continental:
2.12(0.89-3.36)
Hot Summer Continental:
December 2009
E-518
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Philadelphia PA-14.5
Washington PA-14.7
Providence Rl-11.5
Charleston SC-12.1
Taylors SC-15.3
Columbia SC-14.0
SpartanburgSC-14.2
Nashville TN -14.0
KnoxvilleTN-16.0
Memphis TN-13.5
San Antonio TX - 9.4
Dallas TX-12.9
El Paso TX-9.2
Houston TX-12.9
Port Arthur TX-11.5
Ft Worth TX-12.2
Austin TX-10.4
Salt Lake UT-11.5
Provo UT - 9.5
WDCVA-15.2
AnnandaleVA-14.0
Dumbarton VA-13.6
Chesapeake VA-12.7
Norfolk VA-12.7
Richmond VA-14.3
Seattle WA-10.1
TacomaWA-11.2
Spokane WA-9.1
Dodge Wl -11.1
Milwaukee Wl -13.2
WaukeshaWI-13.2
Range (Min, Max): NR
Copollutant (correlation): NR
3.36(1.95-4.79)
Dry: 5.81 (-0.04-12.00)
Dry, Continental:-0.31 (-5.89-5.61)
Mediterranean: 1.06 (-0.36-2.50)
Reference: Zeger et al. (2007,1571761 Outcome: Mortality
Period of Study: 2000-2002
Location: 250 largest U.S. counties
Study Design: Retrospective Cohort
(MCAPS)
Pollutant: PM25
Averaging Time: 3-yr avg
Statistical Analyses: Log-linear
regression models (GAM)
Covariates: Age, gender, race, county-
level SES, education and COPD SMR
Age Groups: 65+
65-74, 75-84, 85+
Increment: 10|jg/m
65+: 1.076 (1.044, 1.108)
Eastern U.S.: 1.125 (1.091,1.159)
Central U.S.: 1.196 (1.115,1.277)
Western U.S.: 1.029 (0.994,1.064)
65-74:1.156(1.117, 1.196)
75-84:1.081 (1.042, 1.121)
85+: 0.995 (0.956, 1.035)
December 2009
E-519
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Reference: Zeger et al. (2008,1919511 Outcome: Mortality
Period of Study: 2000-2005
Location: 4568 zip codes in urban
areas
Study Design: Retrospective Cohort
Statistical Analysis: Log-linear
regression model
Age Groups: Ł65
Pollutant: PM25
Averaging Time: Annual
Median (SD) Unit:
Eastern: 14.0|jg/m3
Central: 10.7 pg/m3
Western: 13.1 pg/m3
All: 13.2 pg/m3
Range (IQR):
Eastern: 12.3-15.3
Central: 9.8-12.2
Western: 10.4-18.5
All: 11.1-14.9
Copollutant (correlation): NR
Increment: 10 pg/m
Relative Risk (Min Cl, Max Cl) lag
Risk estimate for increase in mortality
per increase in PM25, all ages
Eastern Region
Age: 1.155(1.130-1.180)
Age+ SES: 1.105 (1.084-1.125)
Age + SES + COPD:
1.068(1.049-1.087)
Central Region
Age: 1.178 (1.133-1.222)
Age + SES: 1.089 (1.052-1.125)
Age + SES + COPD:
1.132(1.095-1.169)
Western Region
Age: 1.003 (0.981-1.025)
Age + SES: 0.997 (0.978-1.016)
Age + SES + COPD:
0.989(0.970-1.008)
Risk estimate for increase in mortality
per increase in PM25, ages 65-74
Eastern Region
Age: 31.1(26.8-35.5)
Age + SES: 17.3 (14.6-20.0)
Age + SES + COPD: 11.4 (8.8-14.1)
Central Region
Age: 39.0 (29.7-48.2)
Age + SES: 16.5 (10.9-22.1)
Age + SES + COPD: 20.4 (15.0-25.8)
Western Region
Age: 6.0 (2.3-9.6)
Age + SES:-2.1 (-5.0-0.8)
Age + SES + COPD:-1.5 (-4.2-1.1)
Risk estimate for increase in mortality
per increase in PM25, ages 75-84
Eastern Region
Age: 17.6 (14.9-20.4)
Age + SES: 12.4 (10.1-14.6)
Age + SES + COPD: 8.9 (6.8-11.0)
Central Region
Age: 17.5 (12.7-22.2)
Age + SES: 8.8 (4.6-13.0)
Age + SES + COPD: 12.0 (7.6-16.4)
Western Region
Age: 0.4 (-2.0-2.7)
Age + SES: 0.3 (-1.8-2.5)
Age + SES + COPD:-0.2 (-2.2-1.9)
Risk estimate for increase in mortality
per increase in PM25, aged >85
Eastern Region
Age:-1.4 (-3.5-0.8)
Age + SES: 1.4 (-0.7-3.5)
Age + SES + COPD: 1.7 (-0.3-3.7)
Central Region
Age:-2.1 (-5.9-1.6)
Age + SES: -0.7 (-4.2-2.8)
Age + SES + COPD: -0.3 (-4.0-3.3)
Western Region
Age: -5.2 (-7.2-3.2)
Age + SES: 0.9 (-0.8-2.7)
Age + SES + COPD:-0.5 (-2.5-1.5)
All units expressed in pg/m unless otherwise specified.
December 2009
E-520
-------
Table E-34. Long-term exposure - central nervous system outcomes - PM.
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Author: Calderon-Garciduenas et al.
(2008, 1923691
Period of Study: NR
Location: Mexico City (polluted city)
and Tlaxcala and Veracruz (control
cities), Mexico
Outcome (ICD9 and ICD10): COX2
(cyclooxygenase), IL-lp, CD14 in lungs,
OB (olfactory bulb), frontal cortex,
hippocampus, substantia nigrae,
periaqueductal gray and vagus nerves
Age Groups Analyzed:
Subjects 2-45 yr of age
mean=25.1 + 1.5yr
Study Design:
Cross-sectional
N: 47 deceased subjects with complete
autopsies and neuropathological
examinations (each subject had to be
considered clinically healthy and
cognitively and neurologically intact
prior to death) (primarily cause of death:
accidents resulting in immediate death)
Statistical Analyses: NR
likely used T-tests
in addition, stated using "parametric
procedure that considers the
differences among variances of the
variables of interest"
Covariates: Age, gender, place of birth,
place of residency, occupation, smoking
habits, clinical histories, cause of death,
and time between death and autopsy
Season: NR
Dose-response Investigated?
(Yes/No): No
Statistical package: Stata
PM Size: No measure of PM
used Mexico City as the "polluted city"
and Tlaxcala and Veracruz as the
"control cities"
Averaging Time: NA
Mean (SD): NA
Percentiles: NA
Range (Min, Max): NA
Unit(i.e. ug/m3): NA
Number of Monitoring Stations: NA
Co-pollutant (correlation):
NA
PM Increment: NA
Effect Estimate [Lower Cl, Upper Cl]:
RT-PCR sample results from Control
and Mexico City (MC) lung, CMS, PNS
(peripheral nervous system) tissues and
p-value for the difference between the
Concentrations are normalized to the
amount of GAPDHcDNA
COX2 (cyclooxygenase-2) lung
Controls: 15.9+6.7x106
MC residents: 42.3+7.4x106
P-value: 0.015
IL-1|3lung
Controls: 3.08+1.87x106
MC residents: 4.51+2.6x106
P-value: 0.60
COX2 OB (olfactory bulb)
Controls: 12.9+.Ox 105
MC residents: 38.7+5.5x105
P-value: 0.0002
IL-1POB
Controls: 3.4+0.8x104
MC residents: 7.7+1.0x104
P-value: 0.003
CD140B
Controls: 0.01+ 0.001
MC residents: 0.04+ 0.01
P-value: 0.04
COX2 frontal
Controls: 2.6+0.4x105
MC residents: 5.0+0.7x105
P-value: 0.008
IL-1P frontal
Controls: 0.6+0.2x104
MC residents: 6.2+1.3x104
P-value: 0.0002
COX2 hippocampus
Controls: 1.9+0.5x105
MC residents: 1.6+8.7x105
P-value: 0.1
IL-lp hippocampus
Controls: 1.8+0.2x104
MC residents: 3.0+0.5x104
P-value: 0.06
COX2 substantia nigrae
Controls: 0.16+0.06
MC residents: 0.97+ 0.2
P-value: 0.03
IL-1|3substanita nigrae
Controls: 0.01+ 0.005
MC residents: 0.09+ 0.03
P-value: 0.06
CD14 substantia nigrae
Controls: 0.02+ 0.005
MC residents: 0.03+ 0.007
P-value: 0.7
COX2 periaqueductal gray
Controls: 0.10+0.03
MC residents: 0.45+0.12
P-value: 0.12
IL-1p periaqueductal gray
Controls: 0.009+ 0.003
MC residents: 0.07+ 0.02
P-value: 0.09
COX2 left vagus
Controls: 0.65+0.18
MC residents: 2.68+ 0.82
P-value: 0.03
COX2 right vagus
Controls: 0.43+ 0.09
December 2009
E-521
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
MC residents: 3.68+ 0.8
P-value: 0.0002
IL-1P left vagus
Controls: 0.1 ±0.03
MC residents: 1.3+0.73
P-value: 0.06
IL-1 p right vagus
Controls: 0.15+0.09
MC residents: 0.87+ 0.53
p-value: 0.66
CD14 left vagus
Controls: 0.07+ 0.01
MC residents: 0.79+ 0.41
P-value: 0.01
CD14 right vagus
Controls: 0.05+ 0.01
MC residents: 0.31+0.1
P-value: 0.02
Distribution of subjects with expression
of A|342 as a function of age and
residency
Groups: No (%) with A|342 expression
Controls <25yrAPOE 3/3
Controls >25yrAPOE 3/3
n=6
n=3
:0(0
:0(0
MC E2 or E3<25yr(n=17):10(58.82)
MCE2orE3>25yr(n=10):8(80)
MCE4(n=8):8(100)
Controls E4(n=3): 2 (66)
Distribution of subjects with expression
of a-synuclein as a function of age and
Residency
Groups: No (%) with a-synuclein
expression
Controls <25yr APOE 3/3 (n=6): 0 (0)
Controls >25yr APOE 3/3 n=3: 0 (0)
MCE2orE3<25yr(n=17):4(23.5)
MCE2orE3>25yr(n=10):3(30)
MC E4 (n=8): 2 (25)
Controls E4 (n=3): 0 (0)
Reference: Chen and Schwartz (2009,
1799451
Period of Study: 1989-1991
Location: U.S.
Outcome: Change in central nervous
system function
Study Design: Panel
Covariates: Age, sex, race/ethnicity,
individual socioeconomic position,
lifestyle factors, household and
neighborhood characteristics,
conventional CVD risk factors
Statistical Analysis: Pearson Chi-
square tests and t-tests, as appropriate
Statistical Package: STATA
Age Groups: 20-59 yr
Pollutant: PM,0
Averaging Time: 1 yr
Mean (SD) Unit: 37.2 + 12.8 pg/m3
Copollutant: 0;
Increment: 10|jg/m
Regression Coefficient p (96% Cl)
Crude
SRTT: 2.14 (-0.08-4.36)
SDST: 0.08 (0.04-0.13)
SDLT Trials: 0.22 (0.13-0.31)
SDLT Total: 0.44 (0.23-0.65)
Model 1: adjusted forage, sex,
race/ethnicity
SRTT: 2.03 (-0.15-4.20)
SDST: 0.10(0.05-0.15)
SDLT Trials: 0.23 (0.14-0.32)
SDLT Total: 0.48 (0.27-0.68)
Model 2: Model 1 + socioeconomic
factors
SRTT:-0.11 (-2.38-2.16)
SDST: 0.01 (-0.04-0.06)
SDLT Trials: 0.01 (-0.08-0.10)
SDLT Total:-0.07 (-0.27-0.13)
Model 3: Model 2 + lifestyle factors
SRTT:-0.36 (-2.58-1.85)
SDST: 0.00 (-0.04-0.05)
SDLT Trials: 0.09 (0.00-0.17)
SDLT Total: 0.12 (-0.07-0.31)
December 2009
E-522
-------
Study
Design & Methods
Concentrations1
Effect Estimates (95% Cl)
Author: Suglia et al. (2008,1570271 Outcome (ICD9 and ICD10):
Period of Study: 1986-2001
Location: Boston, Massachusetts
Cognition:
Kaufman Brief Intelligence Test, K-BIT
(vocabulary and matrices subscales
and composite IQ score)
Wide Range Assessment of Memory
and Learning, WRAML (psychometric
instrument with subscales on verbal
memory, visual memory, learning, and
overall general index scale)
All cognition scores have a mean of 100
andSD=15.
Age Groups Analyzed: Cognitive tests
administered when children were 8-11
yr of age
Study Design: Cross-sectional
N: 202 children
Statistical Analyses: Linear regression
Covariates: Child's age at cognitive
assessment, gender, primary language
spoken at home, and maternal
education (model 1
"Demographic factors")
Sensitivity analyses performed with
further adjustment for in-utero and
postnatal secondhand tobacco smoke
exposure (via questionnaire during
follow-ups and urinary cotinine levels)
(model 2)
Birth weight (model 3) and blood lead
level (model 4)
Season: Separate land-use regression
models were fit for the warm (May-Oct)
and cold (Nov-Apr) seasons
Used avg of two seasons as measure
of avg lifetime BC exposure
Dose-response Investigated?
(Yes/No): No
Statistical package: SAS (v9.0)
PM Size: Black carbon (BC)
Averaging Time: Lifetime exposure
Estimated 24 h measures of traffic
using a spatiotemporal land-use
regression model using data from >80
locations in Greater Boston (6021
pollution measurements from 2127
unique exposure days)
Predictors in the land-use regression
analysis were the BC level at a central
station (to capture avg concentrations
on that day), meteorological conditions,
weekday/weekend, and measure of
traffic activity (CIS-based measures of
cumulative traffic density within 100m,
population density, distance to nearest
major roadway, % urbanization)
Used the avg of the cold and warm
seasons as the measure of avg lifetime
BC exposure
Mean (SD): 0.56 (0.13) pg/m3
Percentiles: NR
Range (Min, Max): NR
Unit(i.e. ug/m3):
Number of Monitoring Stations: >80
locations
Co-pollutant (correlation): NA
PM Increment: 0.4 pg/m
Effect Estimate [Lower Cl, Upper Cl]:
Change in subscale score (95%CI) per
IQR (0.4 pg/m3) increase in log BC level
K-BIT
Vocabulary:
Adj for demographic factors: -2.0 (-5.3,
1.3)
Adj for above factors + secondhand
smoke:-2.0 (-5.3,1.4)
Adj for above factors + birth weight: -
2.0 (-5.4,1.3)
Adj for above factors + blood lead level:
-2.2 (-5.5, 1.1)
Matrices:
Adj for demographic factors:
-4.2 (-7.7, -0.7)
Adj for above factors + secondhand
smoke: -4.0 (-7.5, -0.4)
Adj for above factors + birth weight:
-4.0 (-7.6, -0.5)
Adj for above factors + blood lead level:
-4.0 (-7.6, -0.5)
Composite:
Adj for demographic factors:
-3.4 (-6.6, -0.3)
Adj for above factors + secondhand
smoke:-3.3 (-6.4,-0.1)
Adj for above factors + birth weight:
-3.3 (-6.5, -0.2)
Adj for above factors + blood lead level:
-3.4 (-6.6, -0.3)
WRAML
Verbal:
Adj for demographic factors:
-1.1 (-4.6, 2.3)
Adj for above factors + secondhand
smoke:-1.2 (-4.7, 2.3)
Adj for above factors + birth weight:
-1.3 (-4.7, 2.2)
Adj for above factors + blood lead level:
-1.3 (-4.8, 2.2)
Visual:
Adj for demographic factors:
-5.2 (-8.6,-1.7)
Adj for above factors + secondhand
smoke:-5.3 (-8.8,-1.8)
Adj for above factors + birth weight:
-5.3 (-8.8,-1.8)
Adj for above factors + blood lead level:
-5.4 (-8.9,-1.9)
Learning:
Adj for demographic factors:
-2.7 (-6.5, 1.1)
Adj for above factors + secondhand
smoke:-2.6 (-6.5,1.2)
Adj for above factors + birth weight:
-2.6 (-6.5,1.3)
Adj for above factors + blood lead level:
-2.8 (-6.6, 1.1)
General:
Adj for demographic factors:
-3.7 (-7.2, -0.2)
Adj for above factors + secondhand
smoke:-3.7 (-7.3,-0.1)
Adj for above factors + birth weight:
-3.8 (-7.4, -0.2)
Adj for above factors + blood lead level:
-3.9 (-7.5, -0.3)
All units expressed in pg/m unless otherwise specified.
December 2009
E-523
-------
Annex E References
Abbey DE; Nishino N; McDonnell WF; Burchette RJ; Knutsen SF; Beeson WL; Yang JX. (1999). Long-term inhalable
particles and other air pollutants related to mortality in nonsmokers. Am J Respir Crit Care Med, 159: 373-382.
047559
Abrahamowicz M; Schopflocher T; Leffondre K; Du Berger R; Krewski D. (2003). Flexible modeling of exposure-
response relationship between long-term average levels of particulate air pollution and mortality in the American
Cancer Society study. J Toxicol Environ Health A, 66: 1625-1654. 086292
Ackermann-Liebrich U; Leuenberger P; Schwartz J; Schindler C; Monn C; Bolognini B; Bongard JP; Brandli O;
Domenighetti G; Elsasser S; Grize L; Karrer W; Keller R; Keller-Wossidlo H; Kunzli N; Martin BW; Medici TC;
Pliger B; Wuthrich B; Zellweger JP; Zemp E. (1997). Lung function and long term exposure to air pollutants in
Switzerland. Am J Respir Crit Care Med, 155: 122-129. 077537
Adamkiewicz G; Ebelt S; Syring M; Slater J; Speizer FE; Schwartz J; Suh H; Gold DR. (2004). Association between air
pollution exposure and exhaled nitric oxide in an elderly population. Thorax, 59: 204-209. 087925
Adar SD; Adamkiewicz G; Gold DR; Schwartz J; Coull BA; Suh H. (2007). Ambient and microenvironmental particles and
exhaled nitric oxide before and after a group bus trip. Environ Health Perspect, 115: 507-512. 098635
Adar SD; Gold DR; Coull BA; Schwartz J; Stone PH; Suh H. (2007). Focused exposures to airborne traffic particles and
heart rate variability in the elderly. Epidemiology, 18: 95-103. 001458
Aekplakorn W; Loomis D; Vichit-Vadakan N; Shy C; Plungchuchon S. (2003). Acute effects of SO2 and particles from a
power plant on respiratory symptoms of children, Thailand. Southeast Asian J Trop Med Public Health, 34: 906-
914. 089908
Aga E; Samoli E; Touloumi G; Anderson HR; Cadum E; Forsberg B; Goodman P; Goren A; Kotesovec F; Kriz B;
Macarol-Hiti M; Medina S; Paldy A; Schindler C; Sunyer J; Tittanen P; Wojtyniak B; Zmirou D; Schwartz J;
Katsouyanni K. (2003). Short-term effects of ambient particles on mortality in the elderly: results from 28 cities in
the APHEA2 project. Eur Respir J, 40: 28s-33s. 054808
Agarwal R; Jayaraman G; Anand S; Marimuthu P. (2006). Assessing respiratory morbidity through pollution status and
meteorological conditions for Delhi. Environ Monit Assess, 114: 489-504. 099086
Allen RW; Criqui MH; Diez Roux AV; Allison M; Shea S; Detrano R; Sheppard L; Wong N; Hinckley Stukovsky K;
Kaufman JD. (2009). Fine particulate air pollution, proximity to traffic, and aortic atherosclerosis: The multi-ethnic
study of atherosclerosis. Epidemiology, 20: 254-264. 156209
Allen RW; Mar T; Koenig J; Liu LJ; Gould T; Simpson C; Larson T. (2008). Changes in lung function and airway
inflammation among asthmatic children residing in a woodsmoke-impacted urban area. Inhal Toxicol, 20: 423-433.
156208
Analitis A; Katsouyanni K; Dimakopoulou K; Samoli E; Nikoloulopoulos AK; Petasakis Y; Touloumi G; Schwartz J;
Anderson HR; Cambra K; Forastiere F; Zmirou D; Vonk JM; Clancy L; Kriz B; Bobvos J; Pekkanen J. (2006).
Short-term effects of ambient particles on cardiovascular and respiratory mortality. Epidemiology, 17: 230-233.
088177
Andersen ZJ; Wahlin P; Raaschou-Nielsen O; Ketzel M; Scheike T; Loft S. (2008). Size distribution and total number
concentration of ultrafine and accumulation mode particles and hospital admissions in children and the elderly in
Copenhagen, Denmark. Occup Environ Med, 65: 458-66. 189651
Andersen ZJ; Wahlin P; Raaschou-Nielsen O; Scheike T; Loft S. (2007). Ambient particle source apportionment and daily
hospital admissions among children and elderly in Copenhagen. J Expo Sci Environ Epidemiol, 17: 625-636.
093201
Anderson HR; Atkinson RW; Bremner SA; Marston L. (2003). Particulate air pollution and hospital admissions for
cardiorespiratory diseases: are the elderly at greater risk?. Eur Respir J, 40: 39s-46s. 054820
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http: //epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009 E-524
-------
Anderson ME; Bogdan GM. (2007). Environments, indoor air quality, and children. Pediatr Clin North Am, 54: 295-307.
156214
Annesi-Maesano, I.; Forastiere, R; Kunzli, N.; Brunekref, B.; Environment and Health Committee of the European
Respiratory Society. (2007). Particulate matter, science and EU policy. Eur Respir J, 29: 428-431. 091348
Arbex MA; Martins LC; De Oliveira RC; Pereira LA; Arbex FF; Cancado JE; Saldiva PH; Braga AL. (2007). Air pollution
from biomass burning and asthma hospital admissions in a sugar cane plantation area in Brazil. J Epidemiol
Community Health, 61: 395-400. 091637
Auchincloss AH; Roux AV; Dvonch JT; Brown PL; Barr RG; Daviglus ML; Goff DC; Kaufman JD; O'Neill MS. (2008).
Associations between Recent Exposure to Ambient Fine Particulate Matter and Blood Pressure in the Multi-Ethnic
Study of Atherosclerosis (MESA). Environ Health Perspect, 116: 486-491. 156234
Avol EL; Gauderman WJ; Tan SM; London SJ; Peters JM. (2001). Respiratory effects of relocating to areas of differing air
pollution levels. Am J Respir Grit Care Med, 164: 2067-2072. 020552
Babin SM; Burkom HS; Holtry RS; Tabernero NR; Stokes LD; Davies-Cole JO; DeHaan K; Lee DH. (2007). Pediatric
patient asthma-related emergency department visits and admissions in Washington, DC, from 2001-2004, and
associations with air quality, socio-economic status and age group. Environ Health Global Access Sci Source, 6: 9.
093250
Baccarelli A; Martinelli I; Pegoraro V; Melly S; Grillo P; Zanobetti A; Hou L; Bertazzi PA; Mannucci PM; Schwartz J.
(2009). Living near Major Traffic Roads and Risk of Deep Vein Thrombosis. Circulation, 119: 3118-3124. 188183
Baccarelli A; Martinelli I; Zanobetti A; Grillo P; HouLF; Bertazzi PA; Mannucci PM; Schwartz J. (2008). Exposure to
Particulate Air Pollution and Risk of Deep Vein Thrombosis. Arch Intern Med, 168: 920-927. 157984
Baccarelli A; Zanobetti A; Martinelli I; Grillo P; Hou L; Giacomini S; Bonzini M; Lanzani G; Mannucci PM; Bertazzi PA;
Schwartz J. (2007). Effects of exposure to air pollution on blood coagulation. J Thromb Haemost, 5: 252-260.
090733
Baccarelli A; Zanobetti A; Martinelli I; Grillo P; Hou L; Lanzani G; Mannucci PM; Bertazzi PA; Schwartz J. (2007). Air
pollution, smoking, and plasma homocysteine. Environ Health Perspect, 115: 176-181. 091310
Bakke B; Ulvestad B; Stewart P; Eduard W. (2004). Cumulative exposure to dust and gases as determinants of lung
function decline in tunnel construction workers. Occup Environ Med, 61: 262-269. 156246
Ballester F; Medina S; Boldo E; Goodman P; Neuberger M; Iniguez C; Kunzli N. (2008). Reducing ambient levels of fine
particulates could substantially improve health: a mortality impact assessment for 26 European cities. J Epidemiol
Community Health, 62: 98-105. 189977
Ballester F; Rodriguez P; Iniguez C; Saez M; Daponte A; Galan I; Taracido M; Arribas F; Bellido J; Cirarda FB; Canada A;
Guillen JJ; Guillen-Grima F; Lopez E; Perez-Hoyos S; Lertxundi A; Toro S. (2006). Air pollution and
cardiovascular admisisons association in Spain: results within the EMECAS project. J Epidemiol Community
Health, 60: 328-336. 088746
Ballester F; Saez M; Perez-Hoyos S; Iniguez C; Gandarillas A; Tobias A; Bellido J; Taracido M; Arribas F; Daponte A;
Alonso E; Canada A; Guillen-Grima F; Cirera L; Perez-Boillos MJ; Saurina C; Gomez F; Tenias JM. (2002). The
EMECAM project: a multicentre study on air pollution and mortality in Spain: combined results for particulates
and for sulfur dioxide. Occup Environ Med, 59: 300-308. 030371
Barclay JL; Egred M; Kruszewski K; Nandakumar R; Norton MY; Stirrat C; Redpath TW; Walton S; Hillis GS. (2007).
The relationship between transmural extent of infarction on contrast enhanced magnetic resonance imaging and
recovery of contractile function in patients with first myocardial infarction treated with thrombolysis. Cardiology,
108: 217-22. 192229
Barclay JL; Miller BG; Dick S; Dennekamp M; Ford I; Hillis GS; Ayres JG; Seaton A. (2009). A panel study of air
pollution in subjects with heart failure: negative results in treated patients. Occup Environ Med, 66: 325-334.
179935
Barnett AG; Williams GM; Schwartz J; Neller AH; Best TL; Petroeschevsky AL; Simpson RW. (2005). Air pollution and
child respiratory health: a case-crossover study in Australia and New Zealand. Am J Respir Crit Care Med, 171:
1272-1278. 087394
December 2009 E-525
-------
Barraza-Villarreal A; Sunyer J; Hernandez-Cadena L; Escamilla-Nunez MC; Sienra-Monge JJ; Ramirez-Aguilar M;
Cortez-Lugo M; Holguin F; Diaz-Sanchez D; Olin AC; Romieu I. (2008). Air pollution, airway inflammation, and
lung function in a cohort study of Mexico City schoolchildren. Environ Health Perspect, 116: 832-838. 156254
Bartzokas A; Kassomenos P; Petrakis M; Celessides C. (2004). The effect of meteorological and pollution parameters on
the frequency of hospital admissions for cardiovascular and respiratory problems in Athens. Indoor Built Environ,
13:271-275.093252
Basu R; Feng WY; Ostro BD. (2008). Characterizing temperature and mortality in nine California counties. Epidemiology,
19: 138-45.098716
Bateson TF; Schwartz J. (2004). Who is sensitive to the effects of particulate air pollution on mortality? A case-crossover
analysis of effect modifiers. Epidemiology, 15: 143-149. 086244
Bayer-Oglesby L; Grize L; Gassner M; Takken-Sahli K; Sennhauser FH; Neu U; Schindler C; Braun-Fahrlander C. (2005).
Decline of ambient air pollution levels and improved respiratory health in Swiss children. Environ Health Perspect,
113: 1632-1637. 086245
Beelen R; Hoek G; van den Brandt PA; Goldbohm RA; Fischer P; Schouten LJ; Jerrett M; Hughes E; Armstrong B;
Brunekreef B. (2008). Long-term effects of traffic-related air pollution on mortality in a Dutch cohort (NLCS-AIR
study). Environ Health Perspect, 116: 196-202. 156263
Beeson WL; Abbey DE; Knutsen SF. (1998). Long-term concentrations of ambient air pollutants and incident lung cancer
in California adults: results from the AHSMOG study. Environ Health Perspect, 106: 813-823. 048890
Bell M; Ebisu K; Peng R; Samet J; Dominici F. (2009). Hospital Admissions and Chemical Composition of Fine Particle
Air Pollution. Am J Respir Grit Care Med, 179: 1115-1120. 191997
Bell ML; Ebisu K; Belanger K. (2007). Ambient air pollution and low birth weight in Connecticut and Massachusetts.
Environ Health Perspect, 115: 1118-24. 091059
Bell ML; Ebisu K; Peng RD; Dominici F. (2009). Adverse health effects of particulate air pollution: modification by air
conditioning. Epidemiology, 20: 682-686. 191007
Bell ML; Kim JY; Dominici F. (2007). Potential confounding of particulate matter on the short-term association between
ozone and mortality in multisite time-series studies. Environ Health Perspect, 115: 1591-1595. 093256
Bell ML; Levy JK; Lin Z. (2008). The effect of sandstorms and air pollution on cause-specific hospital admissions in
Taipei, Taiwan. Occup Environ Med, 65:104- 111. 091268
Bellini P; Baccini M; Biggeri A; Terracini B. (2007). The meta-analysis of the Italian studies on short-term effects of air
pollution (MISA): Old and new issues on the interpretation of the statistical evidences. Environmetrics, 18: 219-
229. 097787
Bennett CM; Simpson P; Raven J; Skoric B; Powell J; Wolfe R; Walters EH; Abramson MJ. (2007). Associations between
ambient PM2.5 concentrations and respiratory symptoms in Melbourne, 1998-2005. J Toxicol Environ Health A,
70: 1613-1618. 156268
Binkova B; Chvatalova I; Lnenickova Z; Milcova A; Tulupova E; Farmer PB; Sram RJ. (2007). PAH-DNA adducts in
environmentally exposed population in relation to metabolic and DNA repair gene polymorphisms. Mutat Res Fund
Mol Mech Mutagen, 620: 49-61. 156273
Boezen HM; Van Der Zee SC; Postma DS; Vonk JM; Gerritsen J; Hoek G; Brunekreef B; Rijcken B; Schouten JP. (1999).
Effects of ambient air pollution on upper and lower respiratory symptoms and peak expiratory flow in children.
Lancet, 353: 874-878. 040410
Boezen HM; Vonk JM; Van Der Zee SC; Gerritsen J; Hoek G; Brunekreef B; Schouten JP; Postma DS. (2005).
Susceptibility to air pollution in elderly males and females. Eur Respir J, 25: 1018-1024. 087396
Bourotte C; Curi-Amarante A-P; Forti M-C; APereira LA; Braga AL; Lotufo PA. (2007). Association between ionic
composition of fine and coarse aerosol soluble fraction and peak expiratory flow of asthmatic patients in Sao Paulo
city (Brazil). Atmos Environ, 41: 2036-2048. 150040
Brauer M; Gehring U; Brunekreef B; De Jongste J; Gerritsen J; Rovers M; Wichmann H-E; Wijga A; Heinrich J. (2006).
Traffic-related air pollution and otitis media. Environ Health Perspect, 114: 1414-1418. 090757
December 2009 E-526
-------
Brauer M; Hoek G; Smit HA; De Jongste JC; Gerritsen J; Postma DS; Kerkhof M; Brunekreef B. (2007). Air pollution and
development of asthma, allergy and infections in a birth cohort. Eur Respir J, 29: 879-888. 090691
Brauer M; Hoek G; Van Vliet P; Meliefste K; Fischer PH; Wrjga A; Koopman LP; Neijens HJ; Gerritsen J; Kerkhof M;
Heinrich J; Bellander T; Brunekreef B. (2002). Air pollution from traffic and the development of respiratory
infections and asthmatic and allergic symptoms in children. Am J Respir Crit Care Med, 166: 1092-1098. 035192
Brauer M; Lencar C; Tamburic L; Koehoorn M; Demers P; Karr C. (2008). A cohort study of traffic-related air pollution
impacts on birth outcomes. Environ Health Perspect, 116: 680-686. 156292
Breitner S; Stolzel M; Cyrys J; Pitz M; Wolke G; Kreyling W; Kilchenhoff H; Heinrich J; Wichmann H; Peters A. (2009).
Short-Term Mortality Rates during a Decade of Improved Air Quality in Erfurt, Germany. Environ Health Perspect,
117:448-454. 188439
Briet M; Collin C; Laurent S; Tan A; Azizi M; Agharazii M; Jeunemaitre X; Alhenc-Gelas F; Boutouyrie P. (2007).
Endothelial function and chronic exposure to air pollution in normal male subjects. Hypertension, 50: 970-976.
093049
Brunekreef B; Beelen R; Hoek G; Schouten L; Bausch-Goldbohm S; Fischer P; Armstrong B; Hughes E; Jerrett M; van
den Brandt P. (2009). Effects of long-term exposure to traffic-related air pollution on respiratory and cardiovascular
mortality in the Netherlands: The NLCS-AIR Study. Health Effects Institute. Boston, MA. 139. 191947
Burnett RT; Stieb D; Brook JR; Cakmak S; Dales R; Raizenne M; Vincent R; Dann T. (2004). Associations between short-
term changes in nitrogen dioxide and mortality in Canadian cities. Arch Environ Occup Health, 59: 228-236.
086247
Burr ML; Karani G; Davies B; Holmes BA; Williams KL. (2004). Effects on respiratory health of a reduction in air
pollution from vehicle exhaust emissions. Occup Environ Med, 61: 212-218. 087809
Cakmak S; Dales RE; Vidal CB. (2007). Air pollution and mortality in Chile: Susceptibility among the elderly. Environ
Health Perspect, 115: 524-7. 091170
Calderon-Garciduenas L; Macias-Parra M; Hoffmann HJ; Valencia-Salazar G; Henriquez-Roldan C; Osnaya N; Monte OC;
Barragan-Mejia G; Villarreal-Calderon R; Romero L; Granada-Macias M; Torres-Jardon R; Medina-Cortina H;
Maronpot RR. (2009). Immunotoxicity and environment: immunodysregulation and systemic inflammation in
children. Toxicol Pathol, 37: 161-169. 192107
Calderon-Garciduenas L; Maronpot RR; Torres-Jardon R; Henriquez-Roldan C; Schoonhoven R; Acuna-Ayala H;
Villarreal-Calderon A; Nakamura J; Fernando R; Reed W; Azzarelli B; Swenberg JA. (2003). DNA damage in nasal
and brain tissues of canines exposed to air pollutants is associated with evidence of chronic brain inflammation and
neurodegeneration. Toxicol Pathol, 31: 524-538. 156316
Calderon-Garciduenas L; Mora-Tiscareno A; Fordham LA; Chung CJ; Valencia-Salazar G; Flores-Gomez S; Solt AC;
Gomez-del Campo A; Jardon-Torres R; Henriquez-Roldan C; Hazucha MJ; Reed W. (2006). Lung radiology and
pulmonary function of children chronically exposed to air pollution. Environ Health Perspect, 114: 1432-1437.
091253
Calderon-Garciduenas L; Mora-Tiscareno A; Ontiveros E; Gomez-Garza G; Barragan-Mejia G; Broadway J; Chapman S;
Valencia-Salazar G; Jewells V; Maronpot RR; Henriquez-Roldan C; Perez-Guille B; Torres-Jardon R; Herrit L;
Brooks D; Osnaya-Brizuela N; Monroy M. (2008). Air pollution, cognitive deficits and brain abnormalities: A pilot
study with children and dogs. Brain Cognit, 68: 117-127. 156317
Calderon-Garciduenas L; Solt A; Henriquez-Roldan C; Torres-Jardon R; Nuse B; Herritt L; Stone I. (2008). Long-term air
pollution exposure is associated with neuroinflammation , an altered innate immune response, disruption of the
blood-brain-barrier, ultrafine particle deposition, and accumulation of amyloid P42 and a synuclein in children and
young adults. Toxicol Pathol, 36: 289-310. 192369
Cardenas M; Vallejo M; Romano-Riquer P; Ruiz-Velasco S; Ferreira-Vidal AD; Hermosillo AG (2008). Personal exposure
to PM2.5 air pollution and heart rate variability in subjects with positive or negative head-up tilt test. Environ Res,
108: 1-6. 191900
Cavallari JM; Eisen EA; Chen JC; Fang SC; Dobson CB; Schwartz J; Christiani DC. (2007). Night Heart Rate Variability
and Particulate Exposures among Boilermaker Construction Workers. Environ Health Perspect, 115:1046-1051.
157425
December 2009 E-527
-------
Cavanagh JA; Brown L; Trought K; Kingham S; Epton MJ. (2007). Elevated concentrations of 1 -hydroxypyrene in
schoolchildren during winter in Christchurch, New Zealand. Sci Total Environ, 374: 51-59. 189802
Cesaroni G; Badaloni C; Porta D; Forastiere F; Perucci CA. (2008). Comparison between several indices of exposure to
traffic-related air pollution and their respiratory health impact in adults. Occup Environ Med, 65: 683-690. 156326
Chahine T; Baccarelli A; Litonjua A; Wright RO; Suh H; Gold DR; Sparrow D; Vokonas P; Schwartz J. (2007). Particulate
air pollution, oxidative stress genes, and heart rate variability in an elderly cohort. Environ Health Perspect, 115:
1617-1622. 156327
Chan CC; Chuang KJ; Chen WJ; Chang WT; Lee CT; Peng CM. (2008). Increasing cardiopulmonary emergency visits by
long-range transported Asian dust storms in Taiwan. Environ Res, 106: 393-400. 093297
Chan MN; Chan CK. (2007). Mass transfer effects on the hygroscopic growth of ammonium sulfate particles with a water-
insoluble coating. Atmos Environ, 41: 4423-4433. 147787
Chang SY; Fang GC. (2007). Springtime soluble particles in a suburban area of Taichung in central Taiwan. Atmos Res, 86:
30-41. 147621
Chardon B; Lefranc A; Granados D; Gremy I. (2007). Air pollution and doctors' house calls for respiratory diseases in the
greater Paris area (2000-3). Occup Environ Med, 64: 320-324. 091308
Chattopadhyay BP; Mukherjee A; Mukherjee K; Roychowdhury A. (2007). Exposure to vehicular pollution and assessment
of respiratory function in urban inhabitants. Lung, 185: 263-70. 147471
Chen J-C; Schwartz J. (2009). Neurobehavioral effects of ambient air pollution on cognitive performance in US adults.
Neurotoxicology, 30: 231-239. 179945
Chen JC; Schwartz J. (2008). Metabolic syndrome and inflammatory responses to long-term particulate air pollutants.
Environ Health Perspect, 116: 612-617. 190106
ChenL; Yang W; JennisonBL; Goodrich A; Omaye ST. (2002). Air pollution and birth weight in northern Nevada, 1991-
1999. Inhal Toxicol, 14: 141-157. 024945
Chen LH; Knutsen SF; Shavlik D; Beeson WL; Petersen F; Ghamsary M; Abbey D. (2005). The association between fatal
coronary heart disease and ambient particulate air pollution: Are females at greater risk?. Environ Health Perspect,
113: 1723-1729. 087942
Chen Y; Yang Q; Krewski D; Burnett RT; Shi Y; McGrail KM. (2005). The effect of coarse ambient particulate matter on
first, second, and overall hospital admissions for respiratory disease among the elderly. Inhal Toxicol, 17: 649-655.
087555
Cheng M-F; Tsai S-S; Wu T-N; Chen P-S; Yang C-Y (2007). Air pollution and hospital admissions for pneumonia in a
tropical city: Kaohsiung, Taiwan. J Toxicol Environ Health A, 70: 2021 -2026. 093034
Chimonas MA; Gessner BD. (2007). Airborne particulate matter from primarily geologic, non-industrial sources at levels
below National Ambient Air Quality Standards is associated with outpatient visits for asthma and quick-relief
medication prescriptions among children less than 20 years old enrolled in Medicaid in Anchorage, Alaska. Environ
Res, 103: 397-404. 093261
Chiu H; Tiao M; Ho S; Kuo H; Wu T; Yang C. (2008). Effects of Asian Dust Storm events on hospital admissions for
chronic obstructive pulmonary disease in Taipei, Taiwan. Inhal Toxicol, 20: 777-781. 191989
Chiu HF; Cheng MH; Yang C Y. (2009). Air pollution and hospital admissions for pneumonia in a subtropical city: Taipei,
Taiwan. Inhal Toxicol, 21: 32-37. 190249
Choi JH; Xu QS; Park SY; Kim JH; Hwang SS; Lee KH; Lee HJ; Hong YC. (2007). Seasonal variation of effect of air
pollution on blood pressure. J Epidemiol Community Health, 61: 314-318. 093196
Chuang K-J; Chan C-C; Su T-C; Lee C-T; Tang C-S. (2007). The effect of urban air pollution on inflammation, oxidative
stress, coagulation, and autonomic dysfunction in young adults. Am J Respir Crit Care Med, 176: 370-376. 091063
Chuang KJ; Chan CC; Su TC; Lin LY; Lee CT. (2007). Associations between particulate sulfate and organic carbon
exposures and heart rate variability in patients with or at risk for cardiovascular diseases. J Occup Environ Med, 49:
610-617. 098629
DTppoliti D; Forastiere F; Ancona C; Agabiti N; Fusco D; Michelozzi P; Perucci CA. (2003). Air pollution and myocardial
infarction in Rome: a case-crossover analysis. Epidemiology, 14: 528-535. 074311
December 2009 E-528
-------
Dales R; Burnett RT; Smith-Doiron M; Stieb DM; Brook JR. (2004). Air pollution and sudden infant death syndrome.
Pediatrics, 113: 628-631. 087342
Dales R; Liu L; Szyszkowicz M; Dalipaj M; Willey J; Kulka R; Ruddy TD. (2007). Particulate air pollution and vascular
reactivity: the bus stop study. Int Arch Occup Environ Health, 81: 159-164. 155743
Dales R; Wheeler A; Mahmud M; Frescura AM; Smith-Doiron M; Nethery E; Liu L. (2008). The Influence of Living Near
Roadways on Spirometry and Exhaled Nitric Oxide in Elementary Schoolchildren. Environ Health Perspect, 116:
1423-1427. 156378
Dales RE; Cakmak S; Doiron MS. (2006). Gaseous air pollutants and hospitalization for respiratory disease in the neonatal
period. Environ Health Perspect, 114: 1751-1754. 090744
Daniels MJ; Dominici F; Zeger SL; Samet JM. (2004). The national morbidity, mortality, and air pollution study Part III:
PM10 concentration-response curves and thresholds for the 20 largest US cities. Health Effects Institute.
Cambridge, MA. 087343
Delfino R; Brummel S; Wu J; Stern H; Ostro B; Lipsett M; Winer A; Street D; Zhang L; Tjoa T. (2009). The relationship of
respiratory and cardiovascular hospital admissions to the southern California wildfires of 2003. Occup Environ
Med, 66: 189. 191994
Delfino RJ; Gong H; Linn WS; Hu Y; Pellizzari ED. (2003). Respiratory symptoms and peak expiratory flow in children
with asthma in relation to volatile organic compounds in exhaled breath and ambient air. J Expo Sci Environ
Epidemiol, 13: 348-363. 090941
Delfino RJ; Quintana PJE; Floro J; Gastanaga VM; Samimi BS; Kleinman MT; Liu L-JS; Bufalino C; Wu C-F; McLaren
CE. (2004). Association of FEV1 in asthmatic children with personal and microenvironmental exposure to airborne
particulate matter. Environ Health Perspect, 112: 932-941. 056897
Delfino RJ; Staimer N; Gillen D; Tjoa T; Sioutas C; Fung K; George SC; Kleinman MT. (2006). Personal and ambient air
pollution is associated with increased exhaled nitric oxide in children with asthma. Environ Health Perspect, 114:
1736-1743. 090745
Delfino RJ; Staimer N; Tjoa T; Polidori A; Arhami M; Gillen DL; Kleinman MT; Vaziri ND; Longhurst J; Zaldivar F;
Sioutas C. (2008). Circulating biomarkers of inflammation, antioxidant activity, and platelet activation are
associated with primary combustion aerosols in subjects with coronary artery disease. Environ Health Perspect,
116:898-906. 156390
Delfino RJ; Zeiger RS; Seltzer JM; Street DH. (1998). Symptoms in pediatric asthmatics and air pollution: differences in
effects by symptom severity, anti-inflammatory medication use and particulate averaging time. Environ Health
Perspect, 106: 751-761.051406
Delfino RJ; Zeiger RS; Seltzer JM; Street DH; McLaren CE. (2002). Association of asthma symptoms with peak
particulate air pollution and effect modification by anti-inflammatory medication use. Environ Health Perspect,
110: A607-A617. 093740
DeMeo DL; Zanobetti A; Litonjua AA; Coull BA; Schwartz J; Gold DR. (2004). Ambient air pollution and oxygen
saturation. Am J Respir Grit Care Med, 170: 383-387. 087346
Desqueyroux H; Pujet J-C; Prosper M; Squinazi F; Momas I. (2002). Short-term effects of low-level air pollution on
respiratory health of adults suffering from moderate to severe asthma. Environ Res, 89: 29-37. 026052
De Hartog JJ; Hoek G; Peters A; Timonen KL; Ibald-Mulli A; Brunekreef B; Heinrich J; Tiittanen P; Van Wrjnen JH;
Kreyling W; Kulmala M; Pekkanen J. (2003). Effects of fine and ultrafine particles on cardiorespiratory symptoms
in elderly subjects with coronary heart disease: the ULTRA study. Am J Epidemiol, 157: 613-623. 001061
de Hartog JJ; Lanki T; Timonen KL; Hoek G; Janssen NA; Ibald-Mulli A; Peters A; Heinrich J; Tarkiainen TH; van
Grieken R; van Wijnen JH; Brunekreef B; Pekkanen J. (2009). Associations between PM2.5 and heart rate
variability are modified by particle composition and beta-blocker use in patients with coronary heart disease.
Environ Health Perspect, 117:105-111. 191904
De Leon SF; Thurston GD; Ito K. (2003). Contribution of respiratory disease to nonrespiratory mortality associations with
air pollution. Am J Respir Grit Care Med, 167: 1117-1123. 055688
December 2009 E-529
-------
Diette GB; Hansel NN; Buckley TJ; Curtin-Brosnan J; Eggleston PA; Matsui EC; McCormack MC; Williams DL; Breysse
PN. (2007). Home indoor pollutant exposures among inner-city children with and without asthma. Environ Health
Perspect, 115: 1665-1669. 156399
Diez Roux AV; Auchincloss AH; Astor B; Barr RG; Cushman M; Dvonch T; Jacobs DR Jr; Kaufman J; Lin X; Samson Px.
(2006). Recent exposure to particulate matter and C-reactive protein concentration in the multi-ethnic study of
atherosclerosis x. Am J Epidemiol, 164: 437-448. 156400
Diez Roux AV; Auchincloss AH; Franklin TG; Raghunathan T; Barr RG; Kaufman J; Astor B; Keeler J. (2008). Long-term
exposure to ambient particulate matter and prevalence of subclinical atherosclerosis in the Multi-Ethnic Study of
Atherosclerosis. Am J Epidemiol, 167: 667-675. 156401
Dominici F. (2004). Time-series analysis of air pollution and mortality: a statistical review. Health Effects Institute. Boston.
059158
Dominici F; Daniels M; McDermott A; Zeger SL; Samet JM. (2003). Shape of the exposure-response relation and mortality
displacement in the NMMAPS database. Health Effects Institute. Boston, MA. 042804
Dominici F; Peng RD; Bell ML; Pham L; McDermott A; Zeger SL; Samet JL. (2006). Fine particulate air pollution and
hospital admission for cardiovascular and respiratory diseases. JAMA, 295: 1127-1134. 088398
Dominici F; Peng RD; Ebisu K; Zeger SL; Samet JM; Bell ML. (2007). Does the effect of PM10 on mortality depend on
PM nickel and vanadium content? A reanaly sis of the NMMAPS data. Environ Health Perspect, 115: 1701-1703.
099135
Dominici F; Peng RD; Zeger SL; White RH; Samet JM. (2007). Particulate air pollution and mortality in the United States:
did the risks change from 1987 to 2000?. Am J Epidemiol, 166: 880-888. 097361
Dominici F; Zanobetti A; Zeger SL; Schwartz J; Samet JM. (2004). Hierarchical bivariate time series models: a combined
analysis of the effects of particulate matter on morbidity and mortality. Biostatistics, 5: 341-360. 096951
Downs SH; Schindler C; Liu L-JS; Keidel D; Bayer-Oglesby L; Brutsche MH; Gerbase MW; Keller R; Kunzli N;
Leuenberger P; Probst-Hensch NM; Tschopp J-M; Zellweger J-P; Rochat T; Schwartz J; Ackermann-Liebrich U;.
(2007). Reduced exposure to PM10 and attenuated age-related decline in lung function. J Expo Sci Environ
Epidemiol, 15: 185-204. 092853
Dubowsky SD; Suh H; Schwartz J; Coull BA; Gold DR. (2006). Diabetes, obesity, and hypertension may enhance
associations between air pollution and markers of systemic inflammation. Environ Health Perspect, 114: 992-998.
088750
Dugandzic R; Dodds L; Stieb D; Smith-Doiron M. (2006). The association between low level exposures to ambient air
pollution and term low birth weight: A retrospective cohort study. Environ Health, 5:3. 088681
Ebelt ST; Wilson WE; Brauer M. (2005). Exposure to ambient and nonambient components of particulate matter: a
comparison of health effects. Epidemiology, 16: 396-405. 056907
Eftim SE; Samet JM; Janes H; McDermott A; Dominici F. (2008). Fine particulate matter and mortality: a comparison of
the six cities and American Cancer Society cohorts with a medicare cohort. Epidemiology, 19: 209-216. 099104
El-Zein A; Nuwayhid I; El-Fadel M; Mroueh S. (2007). Did a ban on diesel-fuel reduce emergency respiratory admissions
for children?. Sci Total Environ, 384: 134-140. 093043
Enstrom JE. (2005). Fine particulate air pollution and total mortality among elderly Californians, 1973-2002. Inhal Toxicol,
17: 803-816. 087356
Erbas B; Kelly A-M; Physick B; Code C; Edwards M. (2005). Air pollution and childhood asthma emergency hospital
admissions: estimating intra-city regional variations. Int J Environ Health Res, 15: 11-20. 073849
Fan Z; Meng Q; Weisel C; Laumbach R; Ohman-Strickland P; Shalat S; Hernandez M; Black K. (2008). Acute exposure to
elevated PM (2.5) generated by traffic and cardiopulmonary health effects in healthy older adults. J Expo Sci
Environ Epidemiol, 19: 525-533. 191979
Farhat SCL; Paulo RLP; Shimoda TM; Conceicao GMS; Lin CA; Braga ALF; Warth MPN; Saldiva PHN. (2005). Effect of
air pollution on pediatric respiratory emergency room visits and hospital admissions. Braz J Med Biol Res, 38: 227-
235. 089461
December 2009 E-530
-------
Ferdinands JM; Crawford CA; Greenwald R; Van Sickle D; Hunter E; league WG. (2008). Breath acidification in
adolescent runners exposed to atmospheric pollution: a prospective, repeated measures observational study. Environ
Health Global Access Sci Source, 7: 10. 156433
Filleul L; Rondeau V; Vandentorren S; Le Moual N; Cantagrel A; Annesi-Maesano I; Charpin D; Declercq C; Neukirch F;
Paris C; Vervloet D; Brochard P; Tessier JF; Kauffmann F; Baldi I. (2005). Twenty five year mortality and air
pollution: results from the French PAARC survey. Occup Environ Med, 62: 453-460. 087357
Fischer P; Hoek G; Brunekreef B; Verhoeff A; van Wrjnen J. (2003). Air pollution and mortality in the Netherlands: are the
elderly more at risk?. Eur Respir J, 40: 34S-38S. 043739
Fischer PH; Brunekreef B; Lebret E. (2004). Air pollution related deaths during the 2003 heat wave in the Netherlands.
Atmos Environ, 38: 1083-1085. 055605
Fischer SL; Koshland CP (2007). Daily and peak 1 h indoor air pollution and driving factors in a rural Chinese village.
Environ Sci Technol, 41: 3121-3126. 156435
Folino AF; Scapellato ML; Canova C; Maestrelli P; Bertorelli G; Simonato L; Iliceto S; Lotti M. (2009). Individual
exposure to particulate matter and the short-term arrhythmic and autonomic profiles in patients with myocardial
infarction. Eur Heart J, 30: 1614-1620. 191902
Forastiere F; Stafoggia M; Berti G; Bisanti L; Cernigliaro A; Chiusolo M; Mallone S; Miglio R; Pandolfi P; Rognoni M;
Serinelli M; Tessari R; Vigotti M; Perucci C. (2008). Particulate Matter and Daily Mortality: A Case-Crossover
Analysis of Individual Effect Modifiers. Epidemiology, 19: 571-580. 186937
Forastiere F; Stafoggia M; Picciotto S; Bellander T; DTppoliti D; Lanki T; Von Klot S; Nyberg F; Paatero P; Peters A;
Pekkanen J; Sunyer J; Perucci CA. (2005). A case-crossover analysis of out-of-hospital coronary deaths and air
pollution in Rome, Italy. Am J Respir Grit Care Med, 172: 1549-1555. 086323
Forastiere F; Stafoggia M; Tasco C; Picciotto S; Agabiti N; Cesaroni G; Perucci CA. (2007). Socioeconomic status,
particulate air pollution, and daily mortality: differential exposure or differential susceptibility. Am J Ind Med, 50:
208-216. 090720
Forbes LJ; Patel MD; Rudnicka AR; Cook DG; Bush T; Stedman JR; Whincup PH; Strachan DP; Anderson RH. (2009).
Chronic exposure to outdoor air pollution and markers of systemic inflammation. Epidemiology, 20: 245-253.
190351
Forsberg B; Segerstedt B; Stjernberg N; Roemer W. (1998). Air pollution and respiratory health of children: the PEACE
panel study in Umea, Sweden. Eur Respir Rev, 8: 12-19. 051714
Franklin M; Koutrakis P; Schwartz J . (2008). The role of particle composition on the association between PM2.5 and
mortality. Epidemiology, 19: 680-689. 097426
Franklin M; Zeka A; Schwartz J. (2007). Association between PM2.5 and all-cause and specific-cause mortality in 27 US
communities. JExpo Sci Environ Epidemiol, 17: 279-287. 091257
Fuentes M; Song HR; Ghosh SK; Holland DM; Davis JM. (2006). Spatial association between speciated fine particles and
mortality. Biometrics, 62: 855-863. 097647
Fung KY; Khan S; Krewski D; Chen Y. (2006). Association between air pollution and multiple respiratory hospitalizations
among the elderly in Vancouver, Canada. Inhal Toxicol, 18: 1005-1011. 089789
Fung KY; Luginaah IKMG; Webster G. (2005). Air pollution and daily hospitalization rates for cardiovascular and
respiratory diseases in London, Ontario. Int J Environ Stud, 62: 677-685. 093262
Galan I; Tobias A; Banegas JR; Aranguez E. (2003). Short-term effects of air pollution on daily asthma emergency room
admissions. Eur Respir J, 22: 802-808. 087408
Gauderman WJ; Avol E; Gilliland F; Vora H; Thomas D; Berhane K; McConnell R; Kuenzli N; Lurmann F; Rappaport E;
Margolis H; Bates D; Peters J. (2004). The effect of air pollution on lung development from 10 to 18 years of age.
NEnglJMed, 351: 1057-1067. 056569
Gauderman WJ; Gilliland GF; Vora H; Avol E; Stram D; McConnell R; Thomas D; Lurmann F; Margolis HG; Rappaport
EB; Berhane K; Peters JM. (2002). Association between air pollution and lung function growth in southern
California children: results from a second cohort. Am J Respir Crit Care Med, 166: 76-84. 026013
December 2009 E-531
-------
Gauderman WJ; McConnell R; Gilliland F; London S; Thomas D; Avol E; Vora H; Berhane K; Rappaport EB; Lurmann F;
Margolis HG; Peters J. (2000). Association between air pollution and lung function growth in southern California
children. Am J Respir Grit Care Med, 162: 1383-1390. 012531
Gauderman WJ; Vora H; McConnell R; Berhane K; Gilliland F; Thomas D; Lurmann F; Avol E; Kunzli N. (2007). Effect
of exposure to traffic on lung development from 10 to 18 years of age: a cohort study. Lancet, 369: 571-577'.
090121
Gehring U; Cyrys J; Sedlmeir G; Brunekreef B; Bellander T; Fischer P; Bauer CP; Reinhardt D; Wichmann HE; Heinrich J.
(2002). Traffic-related air pollution and respiratory health during the first 2 yrs of life. Eur Respir J, 19: 690-698.
036250
Gent JF; Koutrakis P; Belanger K; Triche E; Holford TR; Bracken MB; Leaderer BP (2009). Symptoms and medication
use in children with asthma and traffic-related sources of fine particle pollution. Environ Health Perspect, 117:
1168-1174. 180399
Gent JF; Triche EW; Holford TR; Belanger K; Bracken MB; Beckett WS; Leaderer BP. (2003). Association of low-level
ozone and fine particles with respiratory symptoms in children with asthma. JAMA, 290: 1859-1867. 052885
Gilboa SM; Mendola P; Olshan AF; Langlois PH; Savitz DA; Loomis D; Herring AH; Fixler DE. (2005). Relation between
ambient air quality and selected birth defects, seven county study, Texas, 1997-2000. Am J Epidemiol, 162: 238-
252. 087892
Girardot SP; Ryan PB; Smith SM; Davis WT; Hamilton CB; Obenour RA; Renfro JR; Tromatore KA; Reed GD. (2006).
Ozone and PM25 exposure and acute pulmonary health effects: a study of hikers in the Great Smoky Mountains
National Park. Environ Health Perspect, 113: 612-617. 088271
Goldberg MS; Burnett RT; Valois M-F; Flegel K; Bailar JC III; Brook J; Vincent R; Radon K. (2003). Associations
between ambient air pollution and daily mortality among persons with congestive heart failure. Environ Res, 91:8-
20. 035202
Goldberg MS; Giannetti N; Burnett RT; Mayo NE; Valois MF; Brophy JM. (2008). A panel study in congestive heart
failure to estimate the short-term effects from personal factors and environmental conditions on oxygen saturation
and pulse rate. Occup Environ Med, 65: 659-666. 180380
Goncalves FLT; Carvalho LMV; Conde FC; Latorre MRDO; Saldiva PHN; Braga ALF. (2005). The effects of air pollution
and meteorological parameters on respiratory morbidity during the summer in Sao Paulo City. Environ Int, 31: 343-
349. 089884
Gordian ME; Choudhury AH. (2003). PM10 and asthma medication in schoolchildren. Arch Environ Occup Health, 58: 42-
47. 054842
Goss CH; Newsom SA; Schildcrout JS; Sheppard L; Kaufman JD. (2004). Effect of ambient air pollution on pulmonary
exacerbations and lung function in cystic fibrosis. Am J Respir Grit Care Med, 169: 816-821. 055624
Gouveia N; Bremner SA; Novaes HMD. (2004). Association between ambient air pollution and birth weight in Sao Paulo,
Brazil. J Epidemiol Community Health, 58: 11-17. 055613
Ha E-H; Lee J-T; Kim H; Hong Y-C; Lee. (2003). Infant susceptibility of mortality to air pollution in Seoul, South Korea.
Pediatrics, 111: 284-290. 042552
Hajat S; Anderson HR; Atkinson RW; Haines A. (2002). Effects of air pollution on general practitioner consultations for
upper respiratory diseases in London. Occup Environ Med, 59: 294-299. 030358
Halonen JI; Lanki T; Yli-Tuomi T; Kulmala M; Tiittanen P; Pekkanen J. (2008). Urban air pollution, and asthma and
COPD hospital emergency room visits. Thorax, 63: 635-641. 189507
Halonen JI; Lanki T; Yli-Tuomi T; Tiittanen P; Kulmala M; Pekkanen. (2009). Particulate air pollution and acute
cardiorespiratory hospital admissions and mortality among the elderly. Epidemiology, 20: 143-153. 180379
Hampson NB; Rudd RA; Hauff NM. (2009). Increased long-term mortality among survivors of acute carbon monoxide
poisoning*. Grit Care Med. 190271
Hanigan 1C; Johnston FH; Morgan GG (2008). Vegetation fire smoke, indigenous status and cardio respiratory hospital
admissions in Darwin, Australia, 1996-2005: a time-series study. Environ Health, 7: 42. 156518
December 2009 E-532
-------
Hansen C; Neller A; Williams G; Simpson R. (2006). Maternal exposure to low levels of ambient air pollution and preterm
birth in Brisbane, Australia. BJOG, 113: 935-941. 089818
Hansen C; Neller A; Williams G; Simpson R. (2007). Low levels of ambient air pollution during pregnancy and fetal
growth among term neonates in Brisbane, Australia. Environ Res, 103: 383-389. 090703
Hansen CA; Barnett AG; Jalaludin B; Morgan G. (2009). Ambient air pollution and birth defects in brisbane, Australia.
PLoS ONE, 4: e5408. 192362
Hapcioglu B; Issever H; Kocyigit E; Disci R; Vatansever S; Ozdilli K. (2006). The effect of air pollution and
meteorological parameters on chronic obstructive pulmonary disease at an Istanbul hospital. Indoor Built Environ,
15: 147-153.093263
Harre ESM; Price PD; Ayrey RB; Toop LJ; Martin IR; Town GI. (1997). Respiratory effects of air pollution in chronic
obstructive pulmonary disease: a three month prospective study. Thorax, 52: 1040-1044. 095726
Hastings DL; Jardine S. (2002). The relationship between air particulate levels and upper respiratory disease in soldiers
deployed to Bosnia (1997-1998). Mil Med, 167: 296-303. 030344
Henrotin JB; Besancenot JP; Bejot Y; Giroud M. (2007). Short-term effects of ozone air pollution on ischaemic stroke
occurrence: A case-crossover analysis from a 10-year population-based study in Dijon, France. Occup Environ
Med, 64: 439-445. 093270
Hertz-Picciotto I; Baker RJ; Yap PS; Dostal M; Joad JP; Lipsett M; Greenfield T; Herr CE; Benes I; Shumway RH;
Pinkerton KE; Sram R. (2007). Early childhood lower respiratory illness and air pollution.[see comment]. Environ
Health Perspect, 115: 1510-1518. 135917
Hertz-Picciotto I; Herr CE; Yap PS; Dostal M; Shumway RH; Ashwood P; Lipsett M; Joad JP; Pinkerton KE; Sram RJ.
(2005). Air pollution and lymphocyte phenotype proportions in cord blood. Environ Health Perspect, 113: 1391-
1398.088678
Hinwood AL; De Klerk N; Rodriguez C; Jacoby P; Runnion T; Rye P; Landau L; Murray F; Feldwick M; Spickett J.
(2006). The relationship between changes in daily air pollution and hospitalizations in Perth, Australia 1992-1998:
A case-crossover study. Int J Environ Health Res, 16: 27-46. 088976
Hirshon JM; Shardell M; Alles S; Powell JL; Squibb K; Ondov J; Blaisdell CJ. (2008). Elevated ambient air zinc increases
pediatric asthma morbidity. Environ Health Perspect, 116: 826-831. 180375
Ho W-C; Hartley WR; Myers L; Lin M-H;. (2007). Air pollution, weather, and associated risk factors related to asthma
prevalence and attack rate. Environ Res, 104: 402-409. 093265
Hoffmann B; Moebus S; Kroger K; Stang A; Mohlenkamp S; Dragano N; Schmermund A; Memmesheimer M; Erbel R;
Jockel K-H. (2009). Residential exposure to urban pollution, ankle-brachial index, and peripheral arterial disease.
Epidemiology, 20: 280-288. 190376
Hoffmann B; Moebus S; Mohlenkamp S; Stang A; Lehmann N; Dragano N; Schmermund A; Memmesheimer M; Mann K;
Erbel R; Jockel K-H; Heinz Nixdorf Recall Study Investigative Group. (2007). Residential exposure to traffic is
associated with coronary atherosclerosis. Circulation, 116: 489-496. 091163
Hoffmann B; Moebus S; Stang A; Beck E-M; Dragano N; Mohlenkamp S; Schmermund A; Memmesheimer M; Mann K;
Erbel R; Jockel K-H; Heinz Nixdorf RECALL Study Investigative Group. (2006). Residence close to high traffic
and prevalence of coronary heart disease. Eur Heart J, 27: 2696-2702. 091162
Hogervorst JG; de Kok TM; Briede JJ; Wesseling G; Kleinjans JC; van Schayck CP (2006). Relationship between radical
generation by urban ambient particulate matter and pulmonary function of school children. J Toxicol Environ
Health A, 69: 245-262. 156559
Holguin F; Flores S; Ross Z; Cortez M; Molina M; Molina L; Rincon C; Jerrett M; Berhane K; Granados A; Romieu I.
(2007). Traffic-related exposures, airway function, inflammation, and respiratory symptoms in children. Am J
Respir Grit Care Med, 176: 1236-1242. 099000
Holloman CH; Bortnick SM; Morara M; Strauss WJ; Calder CA. (2004). A Bayesian hierarchical approach for relating
PM2.5 exposure to cardiovascular mortliaty in North Carolina. Environ Health Perspect, 112: 1282-1288. 087375
Hong CY; Chia SE; Widjaja D; Saw SM; Lee J; Munoz C; Koh D. (2004). Prevalence of Respiratory Symptoms in
Children and Air Quality by Village inRural Indonesia. J Occup Environ Med, 46: 1174-1179. 156565
December 2009 E-533
-------
Hong Y-C; Hwang S-S; Kim JH; Lee K-H; Lee H-J; Lee K-H; Yu S-D; Kim D-S. (2007). Metals in participate pollutants
affect peak expiratory flow of schoolchildren. Environ Health Perspect, 115: 430-434. 091347
Hopke PK; Ito K; Mar T; Christensen WF; Eatough DJ; Henry RC; Kim E; Laden F; Lall R; Larson TV; Liu H; Neas L;
Pinto J; Stolzel M; Suh H; Paatero P; Thurston GD. (2006). PM source apportionment and health effects: 1
Intercomparison of source apportionment results. J Expo Sci Environ Epidemiol, 16: 275-286. 088390
Horak F Jr; Studnicka M; Gartner C; Spengler JD; Tauber E; Urbanek R; Veiter A; Frischer T. (2002). Particulate matter
and lung function growth in children: a 3-yr follow-up study in Austrian schoolchildren. Eur Respir J, 19: 838-845.
034792
Host S; Larrieu S; Pascal L; Blanchard M; Declercq C; Fabre P; Jusot JF; Chardon B; Le Tertre A; Wagner V; Prouvost H;
Lefranc A. (2007). Short-term Associations between Fine and Coarse Particles and Cardiorespiratory
Hospitalizations in Six French Cities. Occup Environ Med, 18: S107-S108. 155851
Host S; Larrieu S; Pascal L; Blanchard M; Declercq C; Fabre P; Jusot JF; Chardon B; Le Tertre A; Wagner V; Prouvost H;
Lefranc A. (2008). Short-term associations between fine and coarse particles and hospital admissions for
Cardiorespiratory diseases in six French cities. Occup Environ Med, 65: 544-551. 155852
Hwang B-F; Jaakkola JJK; Lee Y-L; Lin Y-C; Guo Y-LL. (2006). Relation between air pollution and allergic rhinitis in
Taiwanese schoolchildren. Respir Res, 7: 23. 088971
Hwang J-S; Chan C-C. (2002). Effects of air pollution on daily clinic visits for lower respiratory tract illness. Am J
Epidemiol, 155: 1-10. 023222
Hwang K-W; Lee J-H; Jeong D-Y; Lee C-H; Bhatnagar A; Park J-M; Kim S-H. (2008). Observation of difference in the
size distribution of carbon and major inorganic compounds of atmospheric aerosols after the long-range transport
between the selected days of winter and summer. Atmos Environ, 42: 1057-1063. 134420
Ibald-Mulli A; Timonen KL; Peters A; Heinrich J; Wolke G; Lanki T; Buzorius G; Kreyling WG; De Hartog J; Hoek G;
Ten Brink HM; Pekkanen J. (2004). Effects of particulate air pollution on blood pressure and heart rate in subjects
with cardiovascular disease: a multicenter approach. Environ Health Perspect, 112: 369-377. 087415
Ingle ST; Pachpande BG; Wagh ND; Patel VS; Attarde SB. (2005). Exposure to vehicular pollution and respiratory
impairment of traffic policemen in Jalgaon City, India. Ind Health, 43: 656-662. 089014
Islam T; Gauderman WJ; Berhane K; McConnell R; Avol E; Peters JM; Gilliland FD. (2007). The relationship between air
pollution, lung function and asthma in adolescents. Thorax, 62: 957-963. 090697
Issever H; Disci R; Hapcioglu B; Vatansever S; Karan M A; Akkaya V; Erk O. (2005). The effect of air pollution and
meteorological parameters in Istanbul on hospital admissions for acute coronary syndrome. Indoor Built Environ,
14: 157-164.097736
Ito K; Christensen WF; Eatough DJ; Henry RC; Kim E; Laden F; Lall R; Larson TV; Neas L; Hopke PK; Thurston GD.
(2006). PM source apportionment and health effects: 2 An investigation of intermethod variability in associations
between source-apportioned fine particle mass and daily mortality in Washington, DC. J Expo Sci Environ
Epidemiol, 16: 300-310. 088391
Jaffe DH; Singer ME; Rimm AA. (2003). Air pollution and emergency department visits for asthma among Ohio Medicaid
recipients, 1991-1996. Environ Res, 91: 21-28. 041957
Jalaludin B; Mannes T; Morgan G; Lincoln D; Sheppeard V; Corbett S. (2007). Impact of ambient air pollution on
gestational age is modified by season in Sydney, Australia. Environ Health, 6:16. 156601
Jalaludin B; Morgan G; Lincoln D; Sheppeard V; Simpson R; Corbett S. (2006). Associations between ambient air
pollution and daily emergency department attendances for cardiovascular disease in the elderly (65+ years),
Sydney, Australia. JExpo Sci Environ Epidemiol, 16: 225-237. 189416
Jalaludin BB; O'Toole BI; Leeder SR. (2004). Acute effects of urban ambient air pollution on respiratory symptoms,
asthma medication use, and doctor visits for asthma in a cohort of Australian children. Environ Res, 95: 32-42.
056595
Janes H; Dominici F; Zeger SL. (2007). Trends in air pollution and mortality: an approach to the assessment of unmeasured
confounding. Epidemiology, 18: 416-423. 090927
December 2009 E-534
-------
Jansen KL; Larson TV; Koenig JQ; Mar TF; Fields C; Stewart J; Lippmann M. (2005). Associations between health effects
and particulate matter and black carbon in subjects with respiratory disease. Environ Health Perspect, 113: 1741-
1746. 082236
Janssen NAH; Brunekreef B; van Vliet P; Aarts F; Maliefste K; Harssema H; Fischer P. (2003). The relationship between
air pollution from heavy traffic and allergic sensitization, bronchial byperresponsiveness, and respiratory symptoms
in Dutch schoolchildren. Epidemiology, 111: 1512-1518. 133555
Jedrychowski W; Masters E; Choi H; Sochacka E; Flak E; Mroz E; Pac A; Jacek R; Kaim I; Skolicki Z; Spengler JD;
Perera F. (2007). Pre-pregnancy dietary vitamin A intake may alleviate the adverse birth outcomes associated with
prenatal pollutant exposure: epidemiologic cohort study inPoland. Int J Occup Environ Health, 13: 175-180.
156607
Jerrett M; Burnett RT; Ma R; Pope CA III; Krewski D; Newbold KB; Thurston G; Shi Y; Finkelstein N; Calle EE; Thun
MJ. (2005). Spatial analysis of air pollution and mortality in Los Angeles. Epidemiology, 16: 727-736. 087600
Jerrett M; Burnett RT; Willis A; Krewski D; Goldberg MS; DeLuca P; Finkelstein N. (2003). Spatial analysis of the air
pollution-mortality relationship in the context of ecologic confounders. J Toxicol Environ Health A, 66: 1735-1777.
087380
Johnston FH; Bailie RS; Pilotto LS; Hanigan 1C. (2007). Ambient biomass smoke and cardio-respiratory hospital
admissions in Darwin, Australia. BMC Public Health, 7: 240. 155882
Johnston FH; Webby RJ; Pilotto LS; Bailie RS; Parry DL; Halpin SJ. (2006). Vegetation fires, particulate air pollution and
asthma: a panel study in the Australian monsoon tropics. Int J Environ Health Res, 16: 391-404. 091386
Just J; Segala C; Sahraoui F; Priol G; Grimfeld A; Neukirch F. (2002). Short-term health effects of particulate and
photochemical air pollution in asthmatic children. Eur Respir J, 20: 899-906. 035429
Kaiser R; Romieu I; Medina S; Schwartz J; Krzyzanowski M; Kunzli N. (2004). Air pollution attributable postneonatal
infant mortality in US metropolitan areas: a risk assessment study. Environ Health Global Access Sci Source, 3: 4.
076674
Kan H-D; Chen B-H. (2003). Air pollution and daily mortality in Shanghai: a time series study. Arch Environ Occup
Health, 58: 360-367. 087372
Kan H-D; Chen B-H; Fu C-W; Yu S-Z; Mu L-N. (2005). Relationship between ambient air pollution and daily mortality of
SARS in Beijing. Biomed Environ Sci, 18: 1-4. 087561
Kan H; Heiss G; Rose KM; Whitsel E; Lurmann F; London SJ. (2007). Traffic exposure and lung function in adults: the
Atherosclerosis Risk in Communities study. Thorax, 62: 873-879. 091383
Kan H; London SJ; Chen G; Zhang Y; Song G; Zhao N; Jiang L; Chen B. (2007). Differentiating the effects of fine and
coarse particles on daily mortality in Shanghai, China. Environ Int, 33: 376-384. 091267
Kan H; London SJ; Chen G; Zhang Y; Song G; Zhao N; Jiang L; Chen B. (2008). Season, sex, age, and education as
modifiers of the effects of outdoor air pollution on daily mortality in Shanghai, China: The Public Health and Air
Pollution in Asia (PAPA) Study. Environ Health Perspect, 116: 1183-1188. 156621
Karr C; Lumley T; Schreuder A; Davis R; Larson T; Ritz B; Kaufman J. (2007). Effects of subchronic and chronic
exposure to ambient air pollutants on infant bronchiolitis. Am J Epidemiol, 165: 553-560. 090719
Kasamatsu J; Shima M; Yamazaki S; Tamura K; Sun G. (2006). Effects of winter air pollution on pulmonary function of
school children in Shenyang, China. Int J Hyg Environ Health, 209: 435-444. 156627
Kaufman Y. (1987). Satellite sensing of aerosol absorption. J Geophys Res, 92: 4307-4317. 190960
Keatinge WR; Donaldson GC. (2006). Heat acclimatization and sunshine cause false indications of mortality due to ozone.
Environ Res, 100: 387-393. 087536
Kettunen J; Lanki T; Tiittanen P; Aalto PP; Koskentalo T; Kulmala M; Salomaa V; Pekkanen J. (2007). Associations of fine
and ultrafine particulate air pollution with stroke mortality in an area of low air pollution levels. Stroke, 38: 918-
922. 091242
Kim CG; Bell JNB; Power SA. (2003). Effects of soil cadmium on Pinus sylvestris L. seedlings. Plant Soil, 257: 443-449.
155899
December 2009 E-535
-------
Kim H; Lee J-T; Hong Y-C; Yi S-M; Kim Y (2004). Evaluating the effect of daily PM10 variation on mortality. Inhal
Toxicol, 1: 55-58. 087417
Kim JH; Lim DH; Kim JK; Jeong SJ; Son BK. (2005). Effects of particulate matter (PM10) on the pulmonary function of
middle-school children. J Korean Med Sci, 20: 42-45. 087418
Kim JJ; Smorodinsky S; Lipsett M; Singer BC; Hodgson AT; Ostro B. (2004). Traffic-related air pollution near busy roads:
the East Bay children's Respiratory Health Study. Am J Respir Grit Care Med, 170: 520-526. 087383
Kim OJ; Ha EH; Kim BM; Park HS; Jung WJ; Lee BE; Suh YJ; Kim YJ; Lee JT; Kim H; Hong YC. (2007). PM10 and
pregnancy outcomes: a hospital-based cohort study of pregnant women in Seoul. J Occup Environ Med, 49:13 94-
1402. 156642
Kim S Y; O'Neill MS; Lee JT; Cho Y; Kim J; Kim H. (2007). Air pollution, socioeconomic position, and emergency
hospital visits for asthma in Seoul, Korea. Int Arch Occup Environ Health, 80: 701-710. 092837
Klemm RJ; Lipfert FW; Wyzga RE; Gust C. (2004). Daily mortality and air pollution in Atlanta: two years of data from
ARIES. Inhal Toxicol, 16 Suppl 1: 131-141. 056585
Ko FWS; Tarn W; Wong TW; Chan DPS. (2007). Temporal relationship between air pollutants and hospital admissions for
chronic obstructive pulmonary disease in Hong Kong. Thorax, 62: 780-785. 091639
Ko FWS; Tarn W; Wong TW; Lai CKW;. (2007). Effects of air pollution on asthma hospitalization rates in different age
groups inHong Kong. ClinExp Allergy, 37: 1312-1319. 092844
Koenig JQ; Jansen K; Mar TF; Lumley T; Kaufman J; Trenga CA; Sullivan J; Liu LJ; Shapiro GG; Larson TV. (2003).
Measurement of offline exhaled nitric oxide in a study of community exposure to air pollution. Environ Health
Perspect, 111: 1625-1629. 156653
Koken PJM; Piver WT; Ye F; Elixhauser A; Olsen LM; Portier CJ. (2003). Temperature, air pollution, and hospitalization
for cardiovascular diseases among elderly people in Denver. Environ Health Perspect, 111: 1312-1317. 049466
Kongtip P; Thongsuk W; Yoosook W; Chantanakul S. (2006). Health effects of metropolitan traffic-related air pollutants on
street vendors. Atmos Environ, 40: 7138-7145. 096920
Krewski D; Jerrett M; Burnett RT; Ma R; Hughes E; Shi Y; Turner MC; Pope AC III; Thurston G; Calle EE; Thun MJ.
(2009). Extended follow-up and spatial analysis of the American Cancer Society study linking particulate air
pollution and mortality. Health Effects Institute. Cambridge, MA. Report Nr. 140. 191193
Kulkarni N; Pierse N; Rushton L; Grigg J. (2006). Carbon in airway macrophages and lung function in children. N Engl J
Med, 355: 21-30. 089257
Kumar R; Sharma M; Srivastva A; Thakur JS; Jindal SK; Parwana HK. (2004). Association of outdoor air pollution with
chronic respiratory morbidity in an industrial town in northern India. Arch Environ Occup Health, 59: 471-477.
089873
Kunzli N; Jerrett M; Mack WJ; Beckerman B; LaBree L; Gilliland F; Thomas D; Peters J; Hodis HN. (2005). Ambient air
pollution and atherosclerosis in Los Angeles. Environ Health Perspect, 113: 201-206. 087387
Kuo HW; Lai JS; Lee MC; Tai RC; Lee MC. (2002). Respiratory effects of air pollutants among asthmatics in central
Taiwan. Arch Environ Occup Health, 57: 194-200. 036310
Laden F; Schwartz J; Speizer FE; Dockery DW. (2006). Reduction in fine particulate air pollution and mortality: extended
follow-up of the Harvard Six Cities study. Am J Respir Grit Care Med, 173: 667-672. 087605
Lagorio S; Forastiere F; Pistelli R; lavarone I; Michelozzi P; Fano V; Marconi A; Ziemacki G; Ostro BD. (2006). Air
pollution and lung function among susceptible adult subjects: a panel study. Environ Health, 5:11. 089800
Langley-Turnbaugh SJ; Gordon NR; Lambert T. (2005). Airborne particulates and asthma: a Maine case study. Toxicol Ind
Health, 21: 75-92. 093269
Langrish JP; Mills NL; Chan JK; Leseman DL; Aitken RJ; Fokkens PH; Cassee FR; Li J; Donaldson K; Newby DE; Jiang
L. (2009). Beneficial cardiovascular effects of reducing exposure to particulate air pollution with a simple
facemask. Part Fibre Toxicol, 6: 8. 191908
Lanki T; De Hartog JJ; Heinrich J; Hoek G; Janssen NAH; Peters A; Stolzel M; Timonen KL; Vallius M; Vanninen E;
Pekkanen J. (2006). Can we identify sources of fine particles responsible for exercise-induced ischemia on days
with elevated air pollution? The ULTRA study. Environ Health Perspect, 114: 655-660. 088412
December 2009 E-536
-------
Lanki T; Hoek G; Timonen K; Peters A; Tiittanen P; Vanninen E; Pekkanen J. (2008). Hourly variation in fine particle
exposure is associated with transiently increased risk of ST segment depression. Br Med J, 65: 782-786. 191984
Lanki T; Pekkanen J; Aalto P; Elosua R; Berglind N; D'Ippoliti D; Kulmala M; Nyberg F; Peters A; Picciotto S; Salomaa V;
Sunyer J; Tiittanen P; Von Klot S; Forastiere F; for the HEAPSS Study Group. (2006). Associations of traffic-
related air pollutants with hospitalisation for first acute myocardial infarction: the HEAPSS study. Occup Environ
Med, 63: 844-851.089788
Larrieu S; Jusot J-F; Blanchard M; Prouvost H; Declercq C; Fabre P; Pascal L; Le Tertre A; Wagner V; Riviere S; Chardon
B; Borelli D; Cassadou S; Eilstein D; Lefranc A. (2007). Short term effects of air pollution on hospitalizations for
cardiovascular diseases in eight French cities: The PSAS program. Sci Total Environ, 387: 105-112. 093031
Laurent O; Pedrono G; Segala C; Filleul L; Havard S; Deguen S; Schillinger C; Riviere E; Bard D. (2008). Air pollution,
asthma attacks, and socioeconomic deprivation: a small-area case-crossover study. Am J Epidemiol, 168: 58-65.
156672
Le Tertre A; Schwartz J; Touloumi G. (2005). Empirical Bayes and adjusted estimates approach to estimating the relation of
mortality to eposure of PM10. Risk Anal, 25: 711-718. 087560
Lee BE; Ha EH; Park HS; Kim YJ; Hong YC; Kim H; Lee JT. (2003). Exposure to air pollution during different gestational
phases contributes to risks of low birth weight. Hum Reprod, 18: 638-643. 043202
Lee D; Shaddick G. (2007). Time-varying coefficient models for the analysis of air pollution and health outcome data.
Biometrics, 63: 1253-1261. 156685
Lee IM; Tsai SS; Ho CK; Chiu HF; Wu TN; Yang CY (2008). Air pollution and hospital admissions for congestive heart
failure: are there potentially sensitive groups?. Environ Res, 108: 348-353. 192076
Lee J-T; Kim H; Song H; Hong Y-C; Cho Y-S; Shin S-Y; Hyun Y-J; Kim Y-S. (2002). Air pollution and asthma among
children in Seoul, Korea. Epidemiology, 13: 481-484. 034826
Lee J-T; Son J-Y; Cho Y-S. (2007). A comparison of mortality related to urban air particles between periods with Asian dust
days and without Asian dust days in Seoul, Korea, 2000-2004. Environ Res, 105: 409-13. 093042
Lee JT; Kim H; Cho YS; Hong YC; Ha EH; Park H. (2003). Air pollution and hospital admissions for ischemic heart
diseases among individuals 64+years of age residing in Seoul, Korea. Arch Environ Health, 58: 617-623. 095552
Lee SL; Wong WHS; Lau YL. (2006). Association between air pollution and asthma admission among children in Hong
Kong. Clin Exp Allergy, 36: 1138-1146. 090176
Leem J-H; Kaplan BM; Shim YK; Pohl HR; Gotway CA; Bullard SM; Rogers JF; Smith MM; Tylenda CA. (2006).
Exposures to air pollutants during pregnancy andpreterm delivery. Environ Health Perspect, 114: 905-910. 089828
Leonardi GS; Houthuijs D; Steerenberg PA; Fletcher T; Armstrong B; Antova T; Lochman I; Lochmanova A; Rudnai P;
Erdei E; Musial J; Jazwiec-Kanyion B; Niciu EM; Durbaca S; Fabianova E; Koppova K; Lebret E; Brunekreef B;
Van Loveren H. (2000). Immune biomarkers in relation to exposure to particulate matter: a cross-sectional survey
in 17 cities of central Europe. Inhal Toxicol, 12:1-14. 010272
Letz AG; Quinn JM. (2005). Relationship of basic military trainee emergency department visits for asthma and San
Antonio air quality. Allergy Asthma Proc, 26: 463-467. 088752
Lewis TC; Robins TG; Dvonch JT; Keeler GJ; Yip FY; Mentz GB; Lin X; Parker EA; Israel BA; Gonzalez L; Hill Y
(2005). Air pollution-associated changes in lung function among asthmatic children in Detroit. Environ Health
Perspect, 113: 1068-1075. 081079
Le Tertre A; Medina S; Samoli E; Forsberg B; Michelozzi P; Boumghar A; Vonk JM; Bellini A; Atkinson R; Ayres JG;
Sunyer J; Schwartz J; Katsouyanni K. (2002). Short term effects of particulate air pollution on cardiovascular
diseases in eight European cities. J Epidemiol Community Health, 56: 773-779. 023746
Liao D; Duan Y; Whitsel EA; Zheng Z-J; Heiss G; Chinchilli VM; Lin H-M. (2004). Association of higher levels of
ambient criteria pollutants with impaired cardiac autonomic control: a population-based study. Am J Epidemiol,
159:768-777.056590
Liao D; Heiss G; Chinchilli VM; Duan Y; Folsom AR; Lin HM; Salomaa V (2005). Association of criteria pollutants with
plasma hemostatic/inflammatory markers: a population-based study. J Expo Sci Environ Epidemiol, 15: 319-328.
088677
December 2009 E-537
-------
Liao KJ; Tagaris E; Manomaiphiboon K; Napelenok SL; Woo JH; He S; Amar P; Russell AG. (2007). Sensitivities of
Ozone and Fine Particulate Matter Formation toEmissions under the Impact of Potential Future Climate Change.
Environ Sci Technol, 41: 8355-8361. 180272
Lichtenfels AJFC; Gomes IB; Fieri PC; Miraglia SGEK; Hallak J; Saldiva PFTN. (2007). Increased levels of air pollution
and a decrease in the human and mouse male-to-female ration in Sao Paulo, Brazil. Fertil Steril, 87: 230-232.
097041
Lin C-M; Li C-Y; Mao I-F. (2004). Increased risks of term low-birth-weight infants in a petrochemical industrial city with
high air pollution levels. Arch Environ Occup Health, 59: 663-668. 089827
Lin CA; Pereira LAA; Nishioka DC; Conceicao GMS; Graga ALF; Saldiva PHN. (2004). Air pollution and neonatal deaths
in Sao Paulo, Brazil. Braz J Med Biol Res, 37: 765-770. 095787
Lin M; Chen Y; Burnett RT; Villeneuve PJ; Krewski D. (2002). The influence of ambient coarse particulate matter on
asthma hospitalization in children: case-crossover and time-series analyses. Environ Health Perspect, 110: 575-581.
026067
Lin M; Stieb DM; Chen Y. (2005). Coarse particulate matter and hospitalization for respiratory infections in children
younger than 15 years in Toronto: a case-crossover analysis. Pediatrics, 116: 235-240. 087828
Lin Y-C; Lee C-F; Fang T. (2008). Characterization of particle size distribution from diesel engines fueled with palm-
biodiesel blends and paraffinic fuel blends. Atmos Environ, 42: 1133-1143. 126812
Linares C; Diaz J; Tob?as A; Migue JMDe; Otero A. (2006). Impact of urban air pollutants and noise levels over daily
hospital admissions in children in Madrid: a time series analysis. Int Arch Occup Environ Health, 79: 143-152.
092846
Lipfert FW; Baty JD; Miller JP; Wyzga RE. (2006). PM2.5 constituents and related air quality variables as predictors of
survival in a cohort of U.S. military veterans. Inhal Toxicol, 18: 645-657. 088756
Lipfert FW; Wyzga RE; Baty JD; Miller JP. (2006). Traffic density as a surrogate measure of environmental exposures in
studies of air pollution health effects: long-term mortality in a cohort of US veterans. Atmos Environ, 40: 154-169.
088218
Lipfert FW; Zhang J; Wyzga RE. (2000). Infant mortality and air pollution: a comprehensive analysis of US data for 1990.
J Air Waste Manag Assoc, 50: 1350-1366. 004103
Lippmann M; Ito K; Hwang JS; Maciejczyk P; Chen LC. (2006). Cardiovascular effects of nickel in ambient air. Environ
Health Perspect, 114: 1662-1669. 091165
Lipsett MJ; Tsai FC; Roger L; Woo M; Ostro BD. (2006). Coarse particles and heart rate variability among older adults
with coronary artery disease in the Coachella Valley, California. Environ Health Perspect, 114: 1215-1220. 088753
Lisabeth LD; Escobar JD; Dvonch JT; Sanchez BN; Majersik JJ; Brown DL; Smith MA; Morgenstern LB. (2008). Ambient
air pollution and risk for ischemic stroke and transient ischemic attack. Ann Neurol, 64: 53-59. 155939
Liu CC; Chen CC; Wu TN; Yang CY (2008). Association of brain cancer with residential exposure to petrochemical air
pollution in Taiwan. J Toxicol Environ Health A, 71:310-314. 156708
Liu CC; Tsai SS; Chiu HF; Wu TN; Chen CC; Yang CY. (2009). Ambient exposure to criteria air pollutants and risk of
death from bladder cancer in Taiwan. Inhal Toxicol, 21: 48-54. 190292
Liu L; Poon R; Chen L; Frescura AM; Montuschi P; Ciabattoni G; Wheeler A; Dales R. (2009). Acute effects of air
pollution on pulmonary function, airway inflammation, and oxidative stress in asthmatic children. Environ Health
Perspect, 117: 668-674. 192003
Liu L; Ruddy TD; Dalipaj M; Szyszkowicz M; You H; Poon R; Wheeler A; Dales R. (2007). Influence of personal
exposure to particulate air pollution on cardiovascular physiology and biomarkers of inflammation and oxidative
stress in subjects with diabetes. J Occup Environ Med, 49: 258-265. 156705
Liu S; Krewski D; Shi Y; Chen Y; Burnett R. (2007). Association between maternal exposure to ambient air pollutants
during pregnancy and fetal growth restriction. J Expo Sci Environ Epidemiol, 17: 426-432. 090429
Ljungman P; Bellander T; Schneider A; Breitner S; Forastiere F; Hampel R; Illig T; Jacquemin B; Katsouyanni K; von Klot
S. (2009). Modification of the interleukin-6 response to air pollution by interleukin-6 and fibrinogen
polymorphisms. Environ Health Perspect, 117: 1373-1379. 191983
December 2009 E-538
-------
Ljungman PLS; Berglind N; Holmgren C; Gadler F; Edvardsson N; Pershagen G; Rosenqvist M; Sjogren B; Bellander T.
(2008). Rapid effects of air pollution on ventricular arrhythmias. Eur Heart J, 29: 2894-2901. 180266
Llorca J; Salas A; Prieto-Salceda D; Chinchon-Bengoechea V; Delgado-Rodriguez M. (2005). Nitrogen dioxide increases
cardiorespiratory admissions in Torrelavega (Spain). J Environ Health, 68: 30-35. 087825
Loomis D; Castillejos M; Gold DR; McDonnell W; Borja-Aburto VH. (1999). Air pollution and infant mortality in Mexico
City. Epidemiology, 10: 118-123. 087288
Lubinski W; Toczynska I; Chcialowski A; Plusa T. (2005). Influence of air pollution on pulmonary function in healthy
young men from different regions of Poland. Ann Agric Environ Med, 12: 1-4. 087563
Luginaah IN; Fung K Y; Gorey KM; Webster G; Wills C. (2005). Association of ambient air pollution with respiratory
hospitalization in a government designated "area of concern": the case of Windsor, Ontario. Environ Health
Perspect, 113: 290-296. 057327
Luttmann-Gibson H; Suh HH; Coull BA; Dockery DW; Sarnet SE; Schwartz J; Stone PH; Gold DR. (2006). Short-term
effects of air pollution on heart rate variability in senior adults in Steubenville, Ohio. J Occup Environ Med, 48:
780-788. 089794
Magas OK; Gunter JT; Regens JL. (2007). Ambient air pollution and daily pediatric hospitalizations for asthma. Environ
Sci Pollut Res Int, 14: 19-23. 090714
Maheswaran R; Haining RP; Brindley P; Law J; Pearson T; Fryers PR; Wise S; Campbell MJ. (2005). Outdoor air pollution
and stroke in Sheffield, United Kingdom: A small-area level geographical study. Stroke, 36: 239-243. 088683
Maheswaran R; Haining RP; Brindley P; Law J; Pearson T; Fryers PR; Wise S; Campbell MJ. (2005). Outdoor air
pollution, mortality, and hospital admissions from coronary heart disease in Sheffield, UK: A small-area level
ecological study. Eur Heart J, 26: 2543-2549. 090769
Maisonet M; Bush TJ; Correa A; Jaakkola JJK. (2001). Relation between ambient air pollution and low birth weight in the
northeastern United States. Environ Health Perspect, 1093: 351-356. 016624
Mann JK; Tager IB; Lurmann F; Segal M; Quesenberry CP Jr; Lugg MM; Shan J; Van den Eeden SK. (2002). Air pollution
and hospital admissions for ischemic heart disease in persons with congestive heart failure or arrhythmia. Environ
Health Perspect, 110: 1247-1252. 036723
Marines T; Jalaludin B; Morgan G; Lincoln D; Sheppeard V; Corbett S. (2005). Impact of ambient air pollution on birth
weight in Sydney, Australia. Occup Environ Med, 62: 524-530. 087895
Mar TF; Jansen K; Shepherd K; Lumley T; Larson TV; Koenig JQ. (2005). Exhaled nitric oxide in children with asthma
and short-term PM25 exposure in Seattle. Environ Health Perspect, 113: 1791-1794. 088759
Mar TF; Koenig JQ; Jansen K; Sullivan J; Kaufman J; Trenga CA; Siahpush SH; Liu L-JS; Neas L. (2005). Fine particulate
air pollution and cardiorespiratory effects in the elderly. Epidemiology, 16: 681-687. 087566
Mar TF; Larson TV; Stier RA; Claiborn C; Koenig JQ. (2004). An analysis of the association between respiratory
symptoms in subjects with asthma and daily air pollution in Spokane, Washington. Inhal Toxicol, 16: 809-815.
057309
Martins LC; Latorre MRDO; Saldiva PHN; Braga ALF. (2002). Air pollution and emergency room visits due to chronic
lower respiratory diseases in the elderly: An ecological time-series study in Sao Paulo, Brazil. J Occup Environ
Med, 44: 622-627. 035059
Martins MCH; Fatigati FL; Vespoli TC; Martins LC; Martins MA; Saldiva PHN; Braga ALF. (2004). Influence of
socioeconomic conditions on air pollution effects in elderly people an analysis of six regions in Sao Paolo, Brazil. J
Epidemiol Community Health, 58: 41-46. 087457
Masjedi MR; Jamaati HR; Dokouhaki P; Ahmadzadeh Z; Taheri SA; Bigdeli M; Izadi S; Rostamian A; Aagin K; Ghavam
SM. (2003). The effects of air pollution on acute respiratory conditions. Respirology, 8: 213-230. 052100
McConnell R; Berhane K; Gilliland F; London SJ; Islam T; Gauderman WJ; Avol E; Margolis HG; Peters JM. (2002).
Asthma in exercising children exposed to ozone: a cohort study. Lancet, 359: 386-391. 023150
McConnell R; Berhane K; Gilliland F; London SJ; Vora H; Avol E; Gauderman WJ; Margolis HG; Lurmann F; Thomas
DC; Peters JM. (1999). Air pollution and bronchitic symptoms in southern California children with asthma. Environ
Health Perspect, 107: 757-760. 007028
December 2009 E-539
-------
McConnell R; Berhane K; Gilliland F; Molitor J; Thomas D; Lurmann F; Avol E; Gauderman WJ; Peters JM. (2003).
Prospective study of air pollution and bronchitic symptoms in children with asthma. Am J Respir Crit Care Med,
168: 790-797. 049490
McConnell R; Berhane K; Molitor J; Gilliland F; Kunzli N; thorne PS; Thomas D; Gauderman WJ; Avol E; Lurmann F;
Rappaport E; Jerrett M; Peters JM. (2006). Dog ownership enhances symptomatic responses to air pollution
inchildren with asthma. Environ Health Perspect, 114: 1910-1915. 180226
McCormack MC; Breysse PN; Matsui EC; Hansel NN; Williams D; Curtin-Brosnan J; Eggleston P; Diette GB. (2009). In-
home particle concentrations and childhood asthma morbidity. Environ Health Perspect, 117: 294-298. 199833
McCreanor J; Cullinan P; Nieuwenhuijsen MJ; Stewart-Evans J; Malliarou E; Jarup L; Harrington R; Svartengren M; Han
I-K; Ohman-Strickland P; Chung KF; Zhang J. (2007). Respiratory effects of exposure to diesel traffic in persons
with asthma. N Engl J Med, 357: 2348-2358. 092841
McDonnell WF; Nishino-Ishikawa N; Petersen FF; Chen LH; Abbey DE. (2000). Relationships of mortality with the fine
and coarse fractions of long-term ambient PM10 concentrations in nonsmokers. J Expo Sci Environ Epidemiol, 10:
427-436. 010319
McGowan JA; Hider PN; Chacko E; Town GI. (2002). Particulate air pollution and hospital admissions in Christchurch,
New Zealand. Aust N Z J Public Health, 26: 23-29. 030325
Medina-Ramon M; Zanobetti A; Schwartz J. (2006). The effect of ozone and PM10 on hospital admissions for pneumonia
and chronic obstructive pulmonary disease: a national multicity study. Am J Epidemiol, 163: 579-588. 087721
Meng YY; Wilhelm M; Rull RP; English P; Ritz B. (2007). Traffic and outdoor air pollution levels near residences and
poorly controlled asthma in adults. Ann Allergy Asthma Immunol, 98: 455-463. 093275
Metzger KB; Klein M; Flanders WD; Peel JL; Mulholland JA; Langberg JJ; Tolbert PE. (2007). Ambient air pollution and
cardiac arrhythmias in patients with implantable defibrillators. Epidemiology, 18: 585-592. 092856
Metzger KB; Tolbert PE; Klein M; Peel JL; Flanders WD; Todd KH; Mulholland JA; Ryan PB; Frumkin H. (2004).
Ambient air pollution and cardiovascular emergency department visits. Epidemiology, 15: 46-56. 044222
Michaud JP; Grove JS; Krupitsky DCEH. (2004). Emergency department visits and "vog"-related air quality inHilo,
Hawai'i. Environ Res, 95: 11-9. 188530
Middleton N; Yiallouros P; Kleanthous S; Kolokotroni O; Schwartz J; Dockery DW; Demokritou P; Koutrakis P. (2008). A
10-year time-series analysis of respiratory and cardiovascular morbidity in Nicosia, Cyprus: the effect of short-term
changes in air pollution and dust storms. Environ Health, 7:39. 156760
Migliaretti G; Cadum E; Migliore E; Cavallo F. (2005). Traffic air pollution and hospital admission for asthma: a case-
control approach in a Turin (Italy) population. Int Arch Occup Environ Health, 78: 164-169. 088689
Migliaretti G; Cavallo F. (2004). Urban air pollution and asthma in children. Pediatr Pulmonol, 38: 198-203. 087425
Miller KA; Siscovick DS; Sheppard L; Shepherd K; Sullivan JH; Anderson GL; Kaufman JD. (2007). Long-term exposure
to air pollution and incidence of cardiovascular events in women. N Engl J Med, 356: 447-458. 090130
Millstein J; Gilliland F; Berhane K; Gauderman WJ; McConnell R; Avol E; Rappaport EB; Peters JM. (2004). Effects of
ambient air pollutants on asthma medication use and wheezing among fourth-grade school children from 12
Southern California communities enrolled in The Children's Health Study. Arch Environ Occup Health, 59: 505-
514. 088629
Min KB; Min JY; Cho SI; Paek D. (2008). The relationship between air pollutants and heart-rate variability among
community residents in Korea. Inhal Toxicol, 4: 435-444. 191901
Mohr LB; Luo S; Mathias E; Tobing R; Homan S; Sterling D. (2008). Influence of season and temperature on the
relationship of elementalcarbon air pollution to pediatric asthma emergency room visits. J Asthma, 45: 936-943.
180215
Moore K; Neugebauer R; Lurmann F; Hall J; Brajer V; Alcorn S; Tager I. (2008). Ambient ozone concentrations cause
increased hospitalizations for asthma in children: An 18-year study in Southern California. Environ Health
Perspect, 116: 1063-1070. 196685
December 2009 E-540
-------
Morgenstern V; Zutavern A; Cyrys J; Brockow I; Gehring U; Koletzko S; Bauer CP; Reinhardt D; Wichmann H-E;
Heinrich J. (2007). Respiratory health and individual estimated exposure to traffic-related air pollutants in a cohort
of young children. Occup Environ Med, 64: 8-16. 090747
Mortimer K; Neugebauer R; Lurmann F; Alcorn S; Balmes J; Tager I. (2008). Early-Lifetime exposure to air pollution and
allergic sensitization in children with asthma. J Asthma, 45: 874-881. 187280
Mortimer KM; Neas LM; Dockery DW; Redline S; Tager IB. (2002). The effect of air pollution on inner-city children with
asthma. Eur Respir J, 19: 699-705. 030281
Moshammer H; Hutter H-P; Hauck H; Neuberger M. (2006). Low levels of air pollution induce changes of lung function in
a panel of schoolchildren. Eur Respir J, 27: 1138-1143. 090771
Moshammer H; Neuberger M. (2003). The active surface of suspended particles as a predictor of lung function and
pulmonary symptoms in Austrian school children. Atmos Environ, 37: 1737-1744. 041956
Murata A; Kida K; Hasunuma H; Kanegae H; Ishimaru Y; Motegi T; Yamada K; Yoshioka H; Yamamoto K; Kudoh S.
(2007). Environmental influence on the measurement of exhaled nitric oxide concentration in school children:
special reference to methodology. J Nippon Med Sch, 74: 30-6. 189159
Naess O; Nafstad P; Aamodt G; Claussen B; Rosland P. (2007). Relation between concentration of air pollution and cause-
specific mortality: four-year exposures to nitrogen dioxide and particulate matter pollutants in 470 neighborhoods
in Oslo, Norway. Am J Epidemiol, 165: 435-443. 090736
Nafstad P; Haheim LL; Wisloff T; Gram F; Oftedal B; Holme I; Hjermann I; Leren P. (2004). Urban air pollution and
mortality in a cohort of Norwegian men. Environ Health Perspect, 112: 610-605. 087949
Nascimento LF; Pereira LA; Braga AL; Modolo MC; Carvalho JA Jr. (2006). Effects of air pollution on children's health in
a city in southeastern Brazil. Rev Saude Publica, 40: 77-82. 093247
Nawrot TS; Torfs R; Fierens F; De Henauw S; Hoet PH; Van Kersschaever G; De Backer G; Nemery B. (2007). Stronger
associations between daily mortality and fine particulate air pollution in summer than in winter: evidence from a
heavily polluted region in western Europe. J Epidemiol Community Health, 61: 146-149. 098619
Nerriere E; Zmirou-Navier D; Desqueyroux P; Leclerc N; Momas I; Czernichow P. (2005). Lung cancer risk assessment in
relation with personal exposure to airborne particles in four French metropolitan areas. J Occup Environ Med, 47:
1211-1217.088630
Neuberger M; Schimek MG; Horak F Jr; Moshammer H; Kundi M; Frischer T; Gomiscek B; Puxbaum H; Hauck H;
AUPHEP-Team. (2004). Acute effects of particulate matter on respiratory diseases, symptoms and functions:
epidemiological results of the Austrian Projects on Health Effects of Particulate Matter (AUPHEP). Atmos Environ,
38: 3971-3981.093249
O'Connor GT; Neas L; Vaughn B; Kattan M; Mitchell H; Grain EF; Evans R 3rd; Gruchalla R; Morgan W; Stout J; Adams
GK; Lippmann M. (2008). Acute respiratory health effects of air pollution on children with asthma in US inner
cities. JAllergy Clinlmmunol, 121: 1133-1139. 156818
O'Neill MS; Bell ML; Ranjit N; Cifuentes LA; Loomis D; gouveia N; Borja-Aburto VH. (2008). Air pollution and
mortality in Latin America: The role of education. Epidemiology, 19: 810-819. 192314
O'Neill MS; Diez-Roux AV; Auchincloss AH; Franklin TG; Jacobs Jnr DR; Astor BC; Dvonch JT; Kaufman J. (2007).
Airborne particulate matter exposure and urinary albumin excretion: The Multi-Ethnic Study of Atherosclerosis.
Occup Environ Med, 65: 534-540. 156006
O'Neill MS; Hajat S; Zanobetti A; Ramirez-Aguilar M; Schwartz J. (2005). Impact of control for air pollution and
respiratory epidemics on the estimated associations of temperature and daily mortality. Int J Biometeorol, 50: 121-
129. 098094
O'Neill MS; Loomis D; Borja Aburto VH; Gold D; Hertz-Picciotto I; Castillejos M. (2004). Do associations between
airborne particles and daily mortality in Mexico City differ by measurement method, region, or modeling strategy?.
J Expo Sci Environ Epidemiol, 14: 429-439. 087429
O'Neill MS; Veves A; Sarnat JA; Zanobetti A; Gold DR; Economides PA; Horton ES; Schwartz J. (2007). Air pollution and
inflammation in type 2 diabetes: a mechanism for susceptibility. Occup Environ Med, 64: 373-379. 091362
December 2009 E-541
-------
O'Neill MS; Veves A; Zanobetti A; Sarnat JA; Gold DR; Economides PA; Horton ES; Schwartz J. (2005). Diabetes
enhances vulnerability to participate air pollution-associated impairment in vascular reactivity and endothelial
function. Circulation, 111: 2913-2920. 088423
Odajima H; Yamazaki S; Nitta H. (2008). Decline in peak expiratory flow according to hourly short-term concentration of
particulate matter in asthmatic children. Inhal Toxicol, 20: 1263-1272. 192005
Oftedal B; Brunekreef B; Nystad W; Madsen C; Walker S-E; Nafstad P. (2008). Residential outdoor air pollution and lung
function in schoolchildren. Epidemiology, 19: 129-137. 093202
Oftedal B; Nafstad P; Magnus P; Bjorkly S; Skrondal A. (2003). Traffic related air pollution and acute hospital admission
for respiratory diseases in Drammen, Norway 1995-2000. Eur J Epidemiol, 18: 671-675. 055623
Ostro B; Broadwin R; Green S; Feng W-Y; Lipsett M. (2006). Fine particulate air pollution and mortality in nine California
counties: results from CALFINE. Environ Health Perspect, 114: 29-33. 087991
Ostro B; Feng W-Y; Broadwin R; Green S; Lipsett M. (2007). The effects of components of fine particulate air pollution on
mortality in California: results from CALFINE. Environ Health Perspect, 115: 13-19. 091354
Ostro BD; Feng WY; Broadwin R; Malig BJ; Green RS; Lipsett MJ. (2008). The impact of components of fine particulate
matter on cardiovascular mortality in susceptible subpopulations. Occup Environ Med, 65: 750-756. 097971
Ozkaynak H; Thurston GD. (1987). Associations between 1980 US mortality rates and alternative measures of airborne
particle concentration. Risk Anal, 7: 449-461. 072960
Park SK; O'Neill MS; Vokonas PS; Sparrow D; Schwartz J. (2005). Effects of air pollution on heart rate variability: The VA
normative aging study. Environ Health Perspect, 113: 304-309. 057331
Park SK; O'Neill MS; Vokonas PS; Sparrow D; Spiro A 3rd; Tucker KL; Suh H; Hu H; Schwartz J. (2008). Traffic-related
particles are associated with elevated homocysteine: the VA normative aging study. Am J Respir Crit Care Med,
178: 283-289. 156845
Park SK; O'Neill MS; Wright RO; Hu H; Vokonas PS; Sparrow D; Suh H; Schwartz J. (2006). HFE genotype, particulate
air pollution, and heart rate variability; a gene-environmental interaction. Circulation, 114: 2798-2805. 091245
Parker JD; Akinbami LJ; Woodruff TJ. (2009). Air Pollution and Childhood Respiratory Allergies in the United States.
Environ Health Perspect, 117: 140-147. 192359
Parker JD; Woodruff TJ. (2008). Influences of study design and location on the relationship between particulate matter air
pollution and birthweight. Paediatr Perinat Epidemiol, 22: 214-227. 156846
Parker JD; Woodruff TJ; Basu R; Schoendorf KC. (2005). Air pollution and birth weight among term infants in California.
Pediatrics, 115: 121-128. 087462
Peacock JL; Symonds P; Jackson P; Bremner SA; Scarlett JF; Strachan DP; Anderson HR. (2003). Acute effects of winter
air pollution on respiratory function in schoolchildren in southern England. Occup Environ Med, 60: 82-89. 042026
Peel JL; Metzger KB; Klein M; Flanders WD; Mulholland JA; Tolbert PE. (2007). Ambient air pollution and
cardiovascular emergency department visits in potentially sensitive groups. Am J Epidemiol, 165: 625-633. 090442
Peel JL; Tolbert PE; Klein M; Metzger KB; Flanders WD; Knox T; Mulholland JA; Ryan PB; Frumkin H. (2005). Ambient
air pollution and respiratory emergency department visits. Epidemiology, 16: 164-174. 056305
Pekkanen J; Peters A; Hoek G; Tiittanen P; Brunekreef B; de Hartog J; Heinrich J; Ibald-Mulli A; Kreyling WG; Lanki T;
Timonen KL; Vanninen E. (2002). Particulate air pollution and risk of ST-segment depression during repeated
submaximal exercise tests among subjects with coronary heart disease: the exposure and risk assessment for fine
and ultrafine particles in ambient air (ULTRA) study. Circulation, 106: 933-938. 035050
Peled R; Friger M; Bolotin A; Bibi H; Epstein L; Pilpel D; Scharf S. (2005). Fine particles and meteorological conditions
are associated with lung function in children with asthma living near two power plants. Public Health, 119: 418-
425. 156015
Penard-Morand C; Charpin D; Raherison C; Kopferschmitt C; Caillaud D; Lavaud F; Annesi-Maesano I. (2005). Long-
term exposure to background air pollution related to respiratory and allergic health in schoolchildren. Clin Exp
Allergy, 35: 1279-1287. 087951
December 2009 E-542
-------
Peng RD; Chang HH; Bell ML; McDermott A; Zeger SL; Samet JM; Dominici F. (2008). Coarse participate matter air
pollution and hospital admissions for cardiovascular and respiratory diseases among Medicare patients. JAMA,
299: 2172-2179. 156850
Peng RD; Dominici F; Pastor-Barriuso R; Zeger SL; Samet JM. (2005). Seasonal analyses of air pollution and mortality in
100 US cities. Am J Epidemiol, 161: 585-594. 087463
Penttinen P; Tiittanen P; Pekkanen J. (2004). Mortality and air pollution in metropolitan Helsinki, 1988-1996. Scand J
Work Environ Health, 2: 19-27. 087432
Penttinen P; Vallius M; Tiittanen P; Ruuskanen J; Pekkanen J. (2006). Source-specific fine particles in urban air and
respiratory function among adult asthmatics. Inhal Toxicol, 18: 191-198. 087988
Pereira LAA; Loomis D; Conceicao GMS; Braga ALF; Areas RM; Kishi HS; Singer JM; Bohm GM; Saldiva PHN. (1998).
Association between air pollution and intrauterine mortality in Sao Paulo, Brazil. Environ Health Perspect, 106:
325-329. 007264
Perez L; Tobias A; Querol X; Kunzli N; Pey J; Alastuey A; Viana M; Valero N; Gonzalez-Cabre M; Sunyer J. (2008).
Coarse particles from Saharan dust and daily mortality. Epidemiology, 19: 800-807. 156020
Peters A; Greven S; Heid I; Baldari F; Breitner S; Bellander T; Chrysohoou C; Illig T; Jacquemin B; Koenig W. (2009).
Fibrinogen genes modify the fibrinogen response to ambient particulate matter. Am J Respir Crit Care Med, 179:
484-491. 191992
Peters A; von Klot S; Heier M; Trentinaglia I; Cyrys J; Hermann A; Hauptmann M; Wichmann HE; Lowel H. (2005).
Particulate air pollution and nonfatal cardiac events. Part I. Air pollution, personal activities, and onset of
myocardial infarction in a case-crossover study. Health Effects Institute. Boston, MA. 156859
Peters A; Wichmann HE; Lowel H; Meisinger C; Illig T; Holle R; John J; Keil U; Doring A; Filipiak B; Hense HW; Lowel
H; Stieber J. (2005). Partikel in der Aussenluft erhoren das Risiko fur Herz-Kreislauf-Erkrankungen [Aerosol
particles increase the risk of cardiovascular diseases]. Gesundheitswesen, 1: S79- S85. 095747
Peters JM; Avol E; Gauderman WJ; Linn WS; Navidi W; London SJ; Margolis H; Rappaport E; Vora H; Gong H Jr;
Thomas DC. (1999). A study of twelve southern California communities with differing levels and types of air
pollution II Effects on pulmonary function. Am J Respir Crit Care Med, 159: 768-775. 087237
Pierse N; Rushton L; Harris RS; Kuehni CE; Silverman M; Grigg J. (2006). Locally-generated particulate pollution and
respiratory symptoms in young children. Thorax, 61: 216-220. 088757
Pino P; Walter T; Oyarzun M; Villegas R; Romieu I. (2004). Fine particulate matter and wheezing illnesses in the first year
of life. Epidemiology, 15: 702-708. 050220
Pitard A; ZeghnounA; Courseaux A; Lamberty J; Delmas V; Fossard JL; VilletH. (2004). Short-term associations between
air pollution and respiratory drug sales. Environ Res, 95: 43-52. 087433
Pope C; Renlund D; Kfoury A; May H; Home B. (2008). Relation of heart failure hospitalization to exposure to fine
particulate air pollution. Am J Cardiol, 102: 1230-1234. 191969
Pope CAIII; Burnett RT. (2007). Confounding in air pollution epidemiology: the broader context. Epidemiology, 18: 424-
426. 090928
Pope CAIII; Rodermund DL; Gee MM. (2007). Mortality effects of a copper smelter strike and reduced ambient sulfate
particulate matter air pollution. Environ Health Perspect, 115: 679-683. 091256
Pope CA; Hansen ML; Long RW; Nielsen KR; Eatough NL; Wilson WE; Eatough DJ. (2004). Ambient particulate air
pollution, heart rate variability, and blood markers of inflammation in a panel of elderly subjects. Environ Health
Perspect, 112: 339-345. 055238
Pope CAIII; Burnett RT; Thun MJ; Calle EE; Krewski D; Ito K; Thurston GD. (2002). Lung cancer, cardiopulmonary
mortality, and long-term exposure to fine particulate air pollution. JAMA, 287: 1132-1141. 024689
Pope CAIII; Muhlestein JB; May HT; Renlund DG; Anderson JL; Home BD. (2006). Ischemic heart disease events
triggered by short-term exposure to fine particulate air pollution. Circulation, 114: 2443-2448. 091246
Pope III CA; Burnett RT; Thurston GD; Thun MJ; Calle EE; Krewski D; Godleski JJ. (2004). Cardiovascular mortality and
long-term exposure to particulate air pollution: epidemiological evidence of general pathophysiological pathways of
disease. Circulation, 109: 71-77. 055880
December 2009 E-543
-------
Pope III CA; Ezzati M; Dockery DW. (2009). Fine-particulate air pollution and life expectancy in the United States. N Engl
JMed, 360: 376-386. 190107
Preutthipan A; Udomsubpayakul U; Chaisupamongkollarp T; Pentamwa P. (2004). Effect of PM10 pollution in Bangkok on
children with and without asthma. Pediatr Pulmonol, 37: 187-192. 055598
Puett RC; Schwartz J; Hart JE; Yanosky JD; Speizer FE; Suh H; Paciorek CJ; Neas LM; Laden F. (2008). Chronic
particulate exposure, mortality, and coronary heart disease in the nurses' health study. Am J Epidemiol, 168: 1161-
1168. 156891
Qian Z; He Q; Lin H-M; Kong L; Liao D; Dan J; Bentley CM; Wang B. (2007). Association of daily cause-specific
mortality with ambient particle air pollution in Wuhan, China. Environ Res, 105: 380-389. 093054
Qian Z; He Q; Lin HM; Kong L; Bentley CM; Liu W; Zhou D. (2008). High temperatures enhanced acute mortality effects
of ambient particle pollution in the "oven" city of Wuhan, China. Environ Health Perspect, 116: 1172-1178. 156894
Qian Z; Liao D; Lin H-M; Whitsel EA; Rose KM; Duan Y. (2005). Lung function and long-term exposure to air pollutants
in middle-aged American adults. Arch Environ Occup Health, 60: 156-163. 093283
Rabinovitch N; Strand M; Gelfand EW (2006). Particulate levels are associated with early asthma worsening in children
with persistent disease. Am J Respir Grit Care Med, 173: 1098-1105. 088031
Rabinovitch N; Zhang LN; Murphy JR; Vedal S; Dutton SJ; Gelfand EW. (2004). Effects of wintertime ambient air
pollutants on asthma exacerbations in urban minority children with moderate to severe disease. J Allergy Clin
Immunol, 114: 1131-1137.096753
Rainham DG; Smoyer-Tomic KE; Sheridan SC; Burnett RT. (2005). Synoptic weather patterns and modification of the
association between air pollution and human mortality. Int J Environ Health Res, 15: 347-360. 088676
Ranzi A; Gambini M; Spattini A; Galassi C; Sesti D; Bedeschi M; Messori A; Baroni A; Cavagni G; Lauriola P. (2004). Air
pollution and respiratory status in asthmatic children: Hints for a locally based preventive strategy AIRE study. Eur
J Epidemiol, 19: 567-576. 089500
Ren C; Tong S. (2006). Temperature modifies the health effects of particulate matter in Brisbane, Australia. Int J
Biometeorol, 51: 87-96. 092824
Renzetti G; Silvestre G; D'Amario C; Bottini E; Gloria-Bottini F; Bottini N; Auais A; Perez MK; Piedimonte G. (2009).
Less air pollution leads to rapid reduction of airway inflammation and improved airway function in asthmatic
children. Pediatrics, 123: 1051-1058. 199834
Rich DQ; Demissie K; Lu SE; Kamat L; Wartenberg D; Rhoads GG (2009). Ambient air pollutant concentrations during
pregnancy and the risk of fetal growth restriction. J Epidemiol Community Health, 63: 488-496. 180122
Rich DQ; Freudenberger RS; Ohman-Strickland P; Cho Y; Kipen HM. (2008). Right heart pressure increases after acute
increases in ambient particulate concentration. Environ Health Perspect, 116: 1167-1171. 156910
Rich DQ; Kim MH; Turner JR; Mittleman MA; Schwartz J; Catalano PJ; Dockery DW. (2006). Association of ventricular
arrhythmias detected by implantable cardioverter defibrillator and ambient air pollutants in the St Louis, Missouri
metropolitan area. Occup Environ Med, 63: 591-596. 089814
Rich DQ; Mittleman MA; Link MS; Schwartz J; Luttmann-Gibson H; Catalano PJ; Speizer FE; Gold DR; Dockery DW.
(2006). Increased risk of paroxysmal atrial fibrillation episodes associated with acute increases in ambient air
pollution. Environ Health Perspect, 114: 120-123. 088427
Rich DQ; Schwartz J; Mittleman MA; Link M; Luttmann-Gibson H; Catalano PJ; Speizer FE; Dockery DW. (2005).
Association of short-term ambient air pollution concentrations and ventricular arrhythmias. Am J Epidemiol, 161:
1123-1132.079620
Rich KE; Petkau J; Vedal S; Brauer M. (2004). A case-crossover analysis of particulate air pollution and cardiac arrhythmia
in patients with implantable cardioverter defibrillators. Inhal Toxicol, 16: 363-372. 055631
Riediker M; Devlin RB; Griggs TR; Herbst MC; Bromberg PA; Williams RW; Cascio WE. (2004). Cardiovascular effects
in patrol officers are associated with fine particulate matter from brake wear and engine emissions. Part Fibre
Toxicol, 1: 2. 091261
December 2009 E-544
-------
Riojas-Rodriguez H; Escamilla-Cejudo JA; Gonzalez-Hermosillo JA; Tellez-Rojo MM; Vallejo M; Santos-Burgoa C;
Rojas-Bracho L. (2006). Personal PM2.5 and CO exposures and heart rate variability in subjects with known
ischemic heart disease in Mexico City. JExpo Sci Environ Epidemiol, 16: 131-137. 156913
Rios JLM; Boechat JL; Sant'Anna CC; Franca AT. (2004). Atmospheric pollution and the prevalence of asthma: study
among schoolchildren of 2 areas in Rio de Janeiro, Brazil. Ann Allergy Asthma Immunol, 92: 629-634. 087800
Ritz B; Wilhelm M; Hoggatt KJ; Ghosh JK. (2007). Ambient air pollution and preterm birth in the environment and
pregnancy outcomes study at the University of California, Los Angeles. Am JEpidemiol, 166: 1045-1052. 096146
Ritz B; Wilhelm M; Zhao Y. (2006). Air pollution and infant death in southern California, 1989-2000. Pediatrics, 118: 493-
502. 089819
Ritz B; Yu F; Chapa G; Fruin S. (2000). Effect of air pollution on preterm birth among children born in Southern California
between 1989 and 1993. Epidemiology, 11: 502-511. 012068
Ritz B; Yu F; Fruin S; Chapa G; Shaw GM; Harris JA. (2002). Ambient air pollution and risk of birth defects in Southern
California. Am J Epidemiol, 155: 17-25. 023227
Roberts S. (2004). Interactions between particulate air pollution and temperature in air pollution mortality time series
studies. Environ Res, 96: 328-337. 087924
Roberts S. (2005). Using moving total mortality counts to obtain improved estimates for the effect of air pollution on
mortality. Environ Health Perspect, 113: 1148-1152. 087992
Roberts S. (2006). A new model for investigating the mortality effects of multiple air pollutants in air pollution mortality
time-series studies. J Toxicol Environ Health A, 69: 417-435. 089762
Roberts S; Martin MA. (2006). Applying a moving total mortality count to the cities in the NMMAPS database to estimate
the mortality effects of particulate matter air pollution. Occup Environ Med, 63: 193-197. 088670
Roberts S; Martin MA. (2006). The question of nonlinearity in the dose-response relation between particulate matter air
pollution and mortality: can Akaike's Information Criterion be trusted to take the right turn?. Am J Epidemiol, 164:
1242-1250. 097799
Roberts S; Martin MA. (2007). A distributed lag approach to fitting non-linear dose-response models in particulate matter
air pollution time series investigations. Environ Res, 104: 193-200. 156916
Roberts S; Martin MA. (2007). Methods for bias reduction in time-series studies of particulate matter air pollution and
mortality. J Toxicol Environ Health A, 70: 665-675. 156917
Rodriguez C; Tonkin R; Heyworth J; Kusel M; De Klerk N; Sly PD; Franklin P; Runnion T; Blockley A; Landau L;
Hinwood AL. (2007). The relationship between outdoor air quality and respiratory symptoms in young children. Int
J Environ Health Res, 17: 351-360 . 092842
Rogers JF; Dunlop AL. (2006). Air pollution and very low birth weight infants: a target population?. Pediatrics, 118: 156-
164. 091232
Rojas-Martinez R; Perez-Padilla R; Olaiz-Fernandez G; Mendoza-Alvarado L; Moreno-Macias H; Fortoul T; McDonnell
W; Loomis D; Romieu I. (2007). Lung function growth in children with long-term exposure to air pollutants in
Mexico City. Am J Respir Grit Care Med, 176: 377-384. 091064
Roman HA; Walker KD; Walsh TL; Conner L; Richmond HM; Hubbell BJ; Kinney PL. (2008). Expert judgment
assessment of the mortality impact of changes in ambient fine particulate matter in the U. S. Environ Sci Technol,
42: 2268-2274. 156921
Romieu I; Garcia-Esteban R; Sunyer J; Rios C; Alcaraz-Zubeldia M; Velasco SR; Holguin F. (2008). The effect of
supplementation with omega-3 polyunsaturated fatty acids on markers of oxidative stress in elderly exposed to
PM(2.5). Environ Health Perspect, 116: 1237-1242. 156922
Romieu I; Ramirez-Aguilar M; Moreno-Macias H; Barraza-Villarreal A; Miller P; Hernandez-Cadena L; Carbajal-Arroyo
LA; Hernandez-Avila M. (2004). Infant mortality and air pollution: modifying effect by social class. J Occup
Environ Hyg, 46: 1210-1216. 093074
Romieu I; Tellez-Rojo MM; Lazo M; Manzano-Patino Cortez-Lugo M; Julien P; Belanger MC; Hernandez-Avila
MHolguin F. (2005). Omega-3 fatty acid prevents heart rate variability reductions associated with particulate
matter. Am J Respir Grit Care Med, 172: 1534-1540. 086297
December 2009 E-545
-------
Rosenlund M; Bellander T; Nordquist T; Alfredsson L. (2007). Long-Term Exposure to Air Pollution and Cancer.
Epidemiology, 18: S66. 114679
Rosenlund M; Bellander T; Nordqvist T; Alfredsson L. (2006). A register-based case-control study of air pollution and
myocardial infarction. Epidemiology, 17: S240. 114678
Rosenthal FS; Carney JP; Olinger ML. (2008). Out-of-hospital cardiac arrest and airborne fine particulate matter: a case-
crossover analysis of emergency medical services data in Indianapolis, Indiana. Environ Health Perspect, 116: 631 -
636. 156925
Ruckerl R; Greven S; Ljungman P; Aalto P; Antoniades C; Bellander T; Berglind N; Chrysohoou C; Forastiere F;
JacqueminB; vonKlot S; Koenig W; Kuchenhoff H; Lanki T; Pekkanen J; Perucci CA; Schneider A; Sunyer J;
Peters A. (2007). Air pollution and inflammation (interleukin-6, C-reactive protein, fibrinogen) in myocardial
infarction survivors. Environ Health Perspect, 115: 1072-1080. 156931
Ruckerl R; Ibald-Mulli A; Koenig WSchneider A; Woelke G; Cyrys J; Heinrich J; Marder V; Frampton M; Wichmann HE;
Peters A. (2006). Air pollution and markers of inflammation and coagulation in patients with coronary heart
disease. Environ Health Perspect, 173: 432-441.088754
Ruckerl R; Phipps RP; Schneider A; Frampton M; Cyrys J; Oberdorster G; Wichmann HE; Peters A. (2007). Ultrafine
particles and platelet activation in patients with coronary heart disease — results from a prospective panel study. Part
Fibre Toxicol, 4:1. 091379
Sagiv SK; Mendola P; Loomis D; Herring AH; Neas LM; Savitz DA; Poole C. (2005). A time-series analysis of air
pollution and preterm birth in Pennsylvania, 1997-2001. Environ Health Perspect, 113: 602-606. 087468
Sahsuvaroglu T; Jerrett M; Sears MR; McConnell R; Finkelstein N; Arain A; Newbold B; Burnett R. (2009). Spatial
analysis of air pollution and childhood asthma in Hamilton, Canada: comparing exposure methods in sensitive
subgroups. Environ Health, 8: Z. 190983
Sakai M; Sato Y; Sato S; Ihara S; Onizuka M; Sakakibara Y; Takahashi H. (2004). Effect of relocating to areas of reduced
atmospheric particulate matter levels on the human circulating leukocyte count. J Appl Physiol, 97: 1774-1780.
087435
Salam MT; Millstein J; Li Y-F; Lurmann FW; Margolis HG; Gilliland FD. (2005). Birth outcomes and prenatal exposure to
ozone, carbon monoxide, and particulate matter: results from the Children's Health Study. Environ Health Perspect,
113: 1638-1644. 087885
Samoli E; Analitis A; Touloumi G; Schwartz J; Anderson HR; Sunyer J; Bisanti L; Zmirou D; Vonk JM; Pekkanen J;
Goodman P; Paldy A; Schindler C; Katsouyanni K. (2005). Estimating the exposure-response relationships between
particulate matter and mortality within the APHEAmulticity project. Environ Health Perspect, 113: 88-95. 087436
Sanchez-Carrillo CI; Ceron-Mireles P; Rojas-Martinez MR; Mendoza-Alvarado L; Olaiz-Fernandez G; Borja-Aburto VH.
(2003). Surveillance of acute health effects of air pollution in Mexico City. Epidemiology, 14: 536-544. 098428
Santos U; Terra-Filho M; Lin C; Pereira L; Vieira T; Saldiva P; Braga A. (2008). Cardiac arrhythmia emergency room
visits and environmental air pollution in Sao Paulo, Brazil. J Epidemiol Community Health, 62: 267-272. 192004
Sarnat JA; Marmur A; Klein M; Kim E; Russell AG; Sarnat SE; Mulholland JA; Hopke PK; Tolbert PE. (2008). Fine
particle sources and cardiorespiratory morbidity: An application of chemical mass balance and factor analytical
source-apportionment methods. Environ Health Perspect, 116: 459-466. 097972
Sarnat SE; Suh HH; Coull BA; Schwartz J; Stone PH; Gold DR. (2006). Ambient particulate air pollution and cardiac
arrhythmia in a panel of older adults in Steubenville, Ohio. Occup Environ Med, 63: 700-706. 090489
Sauerzapf V; Jones AP; Cross J. (2009). Environmental factors and hospitalisation for chronic obstructive pulmonary
disease in a rural county of England. J Epidemiol Community Health, 63: 324-328. 180082
Schikowski T; Sugiri D; Ranft U; Gehring U; Heinrich J; Wichmann HE; Kramer U. (2005). Long-term air pollution
exposure and living close to busy roads are associated with COPD in women. Respir Res, 22: 152-161. 088637
Schildcrout JS; Sheppard L; Lumley T; Slaughter JC; Koenig JQ; Shapiro GG (2006). Ambient air pollution and asthma
exacerbations in children: An eight-city analysis. Am J Epidemiol, 164: 505-517. 089812
December 2009 E-546
-------
Schindler C; Keidel D; Gerbase MW; Zemp E; Bettschart R; Brandli O; Brutsche MH; Burdet L; Karrer W; Knopfli B;
Pons M; Rapp R; Bayer-Oglesby L; Kunzli N; Schwartz J; Liu L-JS; Ackermann-Liebrich U; Rochat T; the
SAPALDIA Team. (2009). Improvements in PM10 exposure and reduced rates of respiratory symptoms in a cohort
of swiss adults (SAPALDIA). Am J Respir Grit Care Med, 179: 579-587. 191950
Schneider A; Neas L; Herbst M; Case M; Williams R; Cascio W; Hinderliter A; Holguin F; Buse J; Dungan K. (2008).
Endothelial dysfunction: associations with exposure to ambient fine particles in diabetic individuals. Environ
Health Perspect, 116: 1666-1674. 191985
Schreuder AB; Larson TV; Sheppard L; Claiborn CS. (2006). Ambient woodsmoke and associated respiratory emergency
department visits in Spokane, Washington. Int J Occup Environ Health, 12: 147-153. 097959
Schwartz J. (2004). Is the association of airborne particles with daily deaths confounded by gaseous air pollutants? An
approach to control by matching. Environ Health Perspect, 112: 557-561. 053506
Schwartz J. (2004). The effects of particulate air pollution on daily deaths: a multi-city case crossover analysis. Occup
Environ Med, 61: 956-961. 078998
Schwartz J; Coull B; Laden F; Ryan L. (2008). The effect of dose and timing of dose on the association between airborne
particles and survival. Environ Health Perspect, 116: 64-69. 156963
Schwartz J; Laden F; Zanobetti A. (2002). The concentration-response relation between PM2.5 and daily deaths. Environ
Health Perspect, 110: 1025-1029. 025312
Schwartz J; Litonjua A; Suh H; Verrier M; Zanobetti A; Syring M; Nearing B; Verrier R; Stone P; MacCallum G; Speizer
FE; Gold DR. (2005). Traffic related pollution and heart rate variability in a panel of elderly subjects. Thorax, 60:
455-461.074317
Sekine K; Shima M; Nitta Y; Adachi M. (2004). Long term effects of exposure to automobile exhaust on the pulmonary
function of female adults in Tokyo, Japan. Occup Environ Med, 61: 350-357. 090762
Sharma M; Kumar VN; Katiyar SK; Sharma R; Shukla BP; Sengupta B. (2004). Effects of particulate air pollution on the
respiratory health of subjects who live in three areas in Kanpur, India. Arch Environ Health, 59: 348-358. 156974
Sharovsky R; Cesar LA; Ramires JA. (2004). Temperature, air pollution, and mortality from myocardial infarction in Sao
Paulo, Brazil. Braz J Med Biol Res, 37: 1651-1657. 156976
Silkoff PE; Zhang L; Dutton S; Langmack EL; Vedal S; Murphy J; Make B. (2005). Winter air pollution and disease
parameters in advanced chronic obstructive pulmonary disease panels residing in Denver, Colorado. J Allergy Clin
Immunol, 115: 337-344. 087471
Simpson R; Williams G; Petroeschevsky A; Best T; Morgan G; Denison L; Hinwood A; Neville G. (2005). The short-term
effects of air pollution on hospital admissions in four Australian cities. AustNZ J Public Health, 29: 213-221.
087438
Sinclair AH; Tolsma D. (2004). Associations and lags between air pollution and acute respiratory visits in an ambulatory
care setting: 25-month results from the aerosol research and inhalation epidemiological study. J Air Waste Manag
Assoc, 54: 1212-1218. 088696
Singh V; Khandelwal R; Gupta AB. (2003). Effect of air pollution on peak expiratory flow rate variability. J Asthma, 40:
81-86. 052686
Sivacoumar R; Jayabalou R; Swarnalatha S; Balakrishnan K. (2006). Particulate Matter from Stone Crushing Industry: Size
Distribution and Health Effects. J Environ Eng, 132: 405-414. 111115
Slama R; Morgenstern V; Cyrys J; Zutavern A; Herbarth O; Wichmann HE; Heinrich J; LISA Study Group. (2007). Traffic-
related atmospheric pollutants levels during pregnancy and offspring's term birth weight: a study relying on a land-
use regression exposure model. Environ Health Perspect, 115: 1283-1292. 093216
Slaughter JC; Kim E; Sheppard L; Sullivan JH; Larson TV; Claiborn C. (2005). Association between particulate matter and
emergency room visits, hospital admissions and mortality in Spokane, Washington. J Expo Sci Environ Epidemiol,
15: 153-159.073854
Slaughter JC; Lumley T; Sheppard L; Koenig JQ; Shapiro GG (2003). Effects of ambient air pollution on symptom
severity and medication use in children with asthma. Ann Allergy Asthma Immunol, 91: 346-353. 086294
December 2009 E-547
-------
Sokol RZ; Kraft P; Fowler IM; Mamet R; Kim E; Berhane KT. (2006). Exposure to environmental ozone alters semen
quality. Environ Health Perspect, 114: 360-5. 098539
Solomon C; Poole J; Jarup L; Palmer K; Coggon D. (2003). Cardio-respiratory morbidity and long-term exposure to
particulate air pollution. Int J Environ Health Res, 13: 327-335. 087441
Solomon P; Baumann K; Edgerton E; Tanner R; Eatough D; Modey W; Marin H; Savoie D; Natarajan S; Meyer MB.
(2003). Comparison of integrated samplers for mass and composition during the 1999 Atlanta supersites project. J
Geophys Res, 108: 8423. 156994
Sorensen M; Daneshvar B; Hansen M; Dragsted LO; Hertel O; Knudsen L; Loft S. (2003). Personal PM25 exposure and
markers of oxidative stress in blood. Environ Health Perspect, 111: 161-165. 042700
Sram R; Beskid O; Binkova B; Chvatalova I; Lnenickova Z; Milcova A; Solansky I; Tulupova E; Bavorova H; Ocadlikova
D. (2007). Chromosomal aberrations in environmentally exposed population in relation to metabolic and DNA
repair genes polymorphisms. Mutat Res Fund Mol Mech Mutagen, 620: 22-33. 188457
Stafoggia M; Schwartz J; Forastiere F; Perucci CA. (2008). Does temperature modify the association between air pollution
and mortality? A multicity case-crossover analysis in Italy. Am J Epidemiol, 167: 1476-1485. 157005
Staniswalis JG; Parks NJ; Bader JO; Maldonado YM. (2005). Temporal analysis of airborne particulate matter reveals a
dose-rate effect on mortality in El Paso: indications of differential toxicity for different particle mixtures. J Air
Waste Manag Assoc, 55: 893-902. 087473
Steinvil A; Kordova-Biezuner L; Shapira I; Berliner S; Rogowski O. (2008). Short-term exposure to air pollution and
inflammation-sensitive biomarkers. Environ Res, 106: 51-61. 188893
Stieb DM; Judek S; Burnett RT. (2002). Meta-analysis of time-series studies of air pollution and mortality: effects of gases
and particles and the influence of cause of death, age, and season. J Air Waste Manag Assoc, 52: 470-484. 025205
Stolzel M; Breitner S; Cyrys J; Pitz M; Wolke G; Kreyling W; Heinrich J; Wichmann H-E; Peters A. (2007). Daily
mortality and particulate matter in different size classes in Erfurt, Germany. J Expo Sci Environ Epidemiol, 17:
458-467. 091374
Strand M; Vedal S; Rodes C; Dutton SJ; Gelfand EW; Rabinovitch N. (2006). Estimating effects of ambient PM2.5
exposure on health using PM2.5 component measurements and regression calibration. J Expo Sci Environ
Epidemiol, 16: 30-38. 089203
Su TC; Chan CC; Liau CS; Lin LY; Kao HL; Chuang KJ. (2006). Urban air pollution increases plasma fibrinogen and
plasminogen activator inhibitor-1 levels in susceptible patients. Eur J Cardiovasc Prev Rehabil, 13: 849-852.
157022
Suglia SF; Gryparis A; Wright RO; Schwartz J; Wright RJ. (2008). Association of black carbon with cognition among
children in a prospective birth cohort study. Am J Epidemiol, 167: 280-286. 157027
Suh YJ; Kim BM; Park BH; Park H; Kim YJ; Kim H; Hong YC; Ha EH. (2007). Cytochrome P450IA1 polymorphisms
along with PM(10) exposure contribute to the risk of birth weight reduction. Reprod Toxicol, 24: 281-288. 157028
Sullivan J; Ishikawa N; Sheppard L; Siscovick D; Checkoway H; Kaufman J. (2003). Exposure to ambient fine particulate
matter and primary cardiac arrest among persons with and without clinically recognized heart disease. Am J
Epidemiol, 157: 501-509. 043156
Sullivan JH; Hubbard R; Liu SL; Shepherd K; Trenga CA; Koenig JQ; Chandler WL; Kaufman JD. (2007). A community
study of the effect of particulate matter on blood measures of inflammation and thrombosis in an elderly population.
Environ Health Perspect, 6:3. 100083
Sullivan JH; Schreuder AB; Trenga CA; Liu SL; Larson TV; Koenig JQ; Kaufman JD. (2005). Association between short
term exposure to fine particulate matter and heart rate variability in older subjects with and without heart disease.
Thorax, 60: 462-466. 109418
Sun HL; Chou MC; Lue KH. (2006). The relationship of air pollution to ED visits for asthma differ between children and
adults. Am J Emerg Med, 24: 709-713. 090768
Sunyer J; Basagana X; Belmonte J; Anto JM. (2002). Effect of nitrogen dioxide and ozone on the risk of dying in patients
with severe asthma. Thorax, 57: 687-693. 034835
December 2009 E-548
-------
Sunyer J; Jarvis D; Gotschi T; Garcia-Esteban R; Jacquemin B; Aguilera I; Ackerman U; De Marco R; Forsberg B;
Gislason T; Heinrich J; Norback D; Villani S; Kunzli N. (2006). Chronic bronchitis and urban air pollution in an
international study. Occup Environ Med, 63: 836-843. 089771
Symons JM; Wang L; Guallar E; Howell E; Dominici F; Schwab M; Ange BA; Samet J; Ondov J; Harrison D; Geyh A.
(2006). A case-crossover study of fine particulate matter air pollution and onset of congestive heart failure symptom
exacerbation leading to hospitalization. Am J Epidemiol, 164: 421-433. 091258
Szyszkowicz M. (2007). Air pollution and emergency department visits for depression in Edmonton, Canada. Int J Occup
Med Environ Health, 20: 241-245. 092829
S0rensen M; Loft S; Andersen HV; Raaschou-Nielsen O; Skovgaard LT; Knudsen LE; Nielsen IV; Hertel O. (2005).
Personal exposure to PM25, black smoke and NO2 in Copenhagen: relationship to bedroom and outdoor
concentrations covering seasonal variation. JExpo Sci Environ Epidemiol, 15: 413-422. 089428
Tager IB; Balmes J; Lurmann F; Ngo L; Alcorn S; Kunzli N. (2005). Chronic exposure to ambient ozone and lung function
in young adults. Epidemiology, 16: 751-759. 087538
Tainio M; Tuomisto JT; Hanninen O; Aarnio P; Koistinen KJ; Jantunen MJ; Pakkanen J. (2005). Health effects caused by
primary fine particulate matter (PM2.5) emitted from buses in the Helsinki metropolitan area, Finland. Risk Anal,
25: 151-160.087444
Tamura K; Jinsart W; Yano E; Karita K; Boudoung D. (2003). Particulate air pollution and chronic respiratory symptoms
among traffic policemen in Bangkok. Arch Environ Occup Health, 58: 201-207. 087445
Tang C-S; Chang L-T; Lee H-C; Chan C-C. (2007). Effects of personal particulate matter on peak expiratory flow rate of
asthmatic children. Sci Total Environ, 382: 43-51. 091269
Tarantini L; Bonzini M; Apostoli P; Pegoraro V; Bollati V; Marinelli B; Cantone L; Rizzo G; Hou L; Schwartz J; Bertazzi
PA; Baccarelli A. (2009). Effects of particulate matter on genomic DNA methylation content and iNOS promoter
methylation. Environ Health Perspect, 117: 217-222. 192010
Tecer LH; Alagha O; Karaca F; Tuncel G; Eldes N. (2008). Particulate Matter (PM2.5, PM10-2.5, and PM10) and
Children's Hospital Admissions for Asthma and Respiratory Diseases: A Bidirectional Case-Crossover Study. J
Toxicol Environ Health A, 71: 512-520. 180030
Thurston G; Ito K; Mar T; Christensen WF; Eatough DJ; Henry RC; Kim E; Laden F; Lall R; Larson TV; Liu H; Neas L;
Pinto J; Stolzel M; Suh H; Hopke PK. (2005). Results and implications of the workshop on the source
apportionment of PM health effects. Epidemiology, 16: S134-S135. 097949
Timonen KL; Hoek G; Heinrich J; Bernard A; Brunekreef B; De Hartog J; Hameri K; Ibald-Mulli A; Mirme A; Peters A;
Tiittanen P; Kreyling WG; Pekkanen J. (2004). Daily variation in fine and ultrafine particulate air pollution and
urinary concentrations of lung Clara cell protein CC16. Occup Environ Med, 61: 908-914. 087915
Timonen KL; Vanninen E; De Hartog J; Ibald-Mulli A; Brunekreef B; Gold DR; Henrich J; Hoek G; Lanki T; Peters A;
Tarkiainen T; Tiittanen P; Kreyling W; Pekkanen J. (2006). Effects of ultrafine and fine particulate and gaseous air
pollution on cardiac autonomic control in subjects with coronary artery disease: The ULTRA study. J Expo Sci
Environ Epidemiol, 16: 332-341.088747
Tolbert PE; Klein M; Peel JL; Sarnat SE; Sarnat JA. (2007). Multipollutant modeling issues in a study of ambient air
quality and emergency department visits in Atlanta. J Expo Sci Environ Epidemiol, 17: S29-S35. 090316
Touloumi G; Samoli E; Quenel P; Paldy A; Anderson RH; Zmirou D; Galan I; Forsberg B; Schindler C; Schwartz J;
Katsouyanni K. (2005). Short-term effects of air pollution on total and cardiovascular mortality: the confounding
effect of influenza epidemics. Epidemiology, 16: 49-57. 087477
Tovalin H; Valverde M; Morandi MT; Blanco S; Whitehead L; Rojas E. (2006). DNA damage in outdoor workers
occupationally exposed to environmental air pollutants. Occup Environ Med, 63: 230-236. 091322
Trenga CA; Sullivan JH; Schildcrout JS; Shepherd KP; Shapiro GG; Liu LJ; Kaufman JD; Koenig JQ. (2006). Effect of
particulate air pollution on lung function in adult and pediatric subjects in a Seattle panel study. Chest, 129: 1614-
1622. 155209
Tsai CJ; Chang CT; Huang CH. (2006). Direct field observation of the relative humidity effect on the beta-gauge readings.
J Air Waste Manag Assoc, 56: 834-40. 098312
December 2009 E-549
-------
Tsai S-S; Chen C-C; Hsieh H-J; Chang C-C; Yang C-Y. (2006). Air pollution and postneonatal mortality in a tropical city:
Kaohsiung, Taiwan. Inhal Toxicol, 18: 185-189. 090709
Tsai S-S; Cheng M-H; Chiu H-F; Wu T-N; Yang C-Y. (2006). Air pollution and hospital admissions for asthma in a tropical
city: Kaohsiung, Taiwan. Inhal Toxicol, 18: 549-554. 089768
Tsai S-S; Goggins WB; Chiu H-F; Yang C-Y. (2003). Evidence for an association between air pollution and daily stroke
admissions in Kaohsiung, Taiwan. Stroke, 34: 2612-2616. 080133
Tsai S-S; Huang C-H; Goggins WB; Wu T-N; Yang C-Y. (2003). Relationship between air pollution and daily mortality in a
tropical city: Kaohsiung, Taiwan. J Toxicol Environ Health A, 66: 1341-1349. 050480
Ulirsch GV; Ball LM; Kaye W; Shy CM; Lee CV; Crawford-Brown D; Symons M; Holloway T. (2007). Effect of
particulate matter air pollution on hospital admissions and medical visits for lung and heart disease in two southeast
Idaho cities. J Expo Sci Environ Epidemiol, 17: 478-487. 091332
Vajanapoom N; Shy CM; Neas LM; Loomis D. (2002). Associations of particulate matter and daily mortality in Bangkok,
Thailand. Southeast Asian J Trop Med Public Health, 33: 389-399. 042542
Vallejo M; Ruiz S; Hermosillo AG; Borja-Aburto VH; Cardenas M. (2006). Ambient fine particles modify heart rate
variability in young healthy adults. J Expo Sci Environ Epidemiol, 16: 125-130. 157081
Van Hee VC; Adar SD; Szpiro AA; Barr RG; Bluemke DA; Diez Roux AV; Gill EA; Sheppard L; Kaufman JD. (2009).
Exposure to traffic and left ventricular mass and function: the Multi-Ethnic Study of Atherosclerosis. Am J Respir
Grit Care Med, 179: 827-834. 192110
Vedal S; Brauer M; White R; Petkau J. (2003). Air pollution and daily mortality in a city with low levels of pollution.
Environ HealthPerspect, 111: 45-51. 039044
Vedal S; Rich K; Brauer M; White R; Petkau J. (2004). Air pollution and cardiac arrhythmias in patients with implantable
cardiovascular defibrillators. Inhal Toxicol, 16: 353-362. 055630
Vegni FE; Ros O. (2004). Hospital accident and emergency burden is unaffected by today's air pollution levels. Eur J
Emerg Med, 11: 86-88. 087448
Venners SA; Wang B; Xu Z; Schlatter Y; Wang L; Xu X. (2003). Particulate matter, sulfur dioxide, and daily mortality in
Chongqing, China. Environ Health Perspect, 111: 562-567. 089931
Vichit-VadakanN; Vajanapoom N; Ostro B. (2008). The Public Health and Air Pollution in Asia (PAPA) Project: estimating
the mortality effects of particulate matter in Bangkok, Thailand. Environ Health Perspect, 116: 1179-1182. 157095
Vigotti MA; Chiaverini F; Biagiola P; Rossi G (2007). Urban air pollution and emergency visits for respiratory complaints
in Pisa, Italy. J Toxicol Environ Health A, 70: 266-269. 090711
Villeneuve PJ; Burnett RT; Shi Y; Krewski D; Goldberg MS; Hertzman C; Chen Y; Brook J. (2003). A time-series study of
air pollution, socioeconomic status, and mortality in Vancouver, Canada. J Expo Sci Environ Epidemiol, 13: 427-
435. 055051
Villeneuve PJ; Chen L; Stieb D; Rowe BH. (2006). Associations between outdoor air pollution and emergency department
visits for stroke in Edmonton, Canada. Eur J Epidemiol, 21: 689-700. 090191
Villeneuve PJ; Goldberg MS; Krewski D; Burnett RT; Chen Y. (2002). Fine particulate air pollution and all-cause mortality
within the Harvard six-cities study: variations in risk by period of exposure. Ann Epidemiol, 12: 568-576. 042576
Vineis P; Hoek G; Krzyzanowski M; Vigna-Taglianti F; Veglia F; Airoldi L; Autrup H; Dunning A; Garte S; Hainaut P;
Malaveille C; Matullo G; Overvad K; Raaschou-Nielsen O; Clavel-Chapelon F; Linseisen J; Boeing H;
Trichopoulou A; Palli D; Peluso M; Krogh V; Tumino R; Panico S; Bueno-De-Mesquita HB; Peeters PH; Lund EE;
Gonzalez CA; Martinez C; Dorronsoro M; Barricarte A; Cirera L; Quiros JR; Berglund G; Forsberg B; Day NE;
Key TJ; Saracci R; Kaaks R; Riboli E. (2006). Air pollution and risk of lung cancer in a prospective study in
Europe. Int J Cancer, 119: 169-174. 192089
Von Klot S; Wolke G; Tuch T; Heinrich J; Dockery DW; Schwartz J; Kreyling WG; Wichmann HE; Peters A. (2002).
Increased asthma medication use in association with ambient fine and ultrafine particles. Eur Respir J, 20: 691-702.
034706
December 2009 E-550
-------
Von Klot S; Peters A; Aalto P; Bellander T; Berglind N; D'Ippoliti D; Elosua R; Hermann A; Kulmala M; Lanki T; Lowel
H; Pekkanen J; Picciotto S; Sunyer J; Forastiere F; Health Effects of Particles on Susceptible Subpopulations
(HEAPSS) Study Group. (2005). Ambient air pollution is associated with increased risk of hospital cardiac
readmissions of myocardial infarction survivors in five European cities. Circulation, 112: 3073-3079. 088070
Ward DJ; Roberts KT; Jones N; Harrison RM; Ayres JG; Hussain S; Walters S. (2002). Effects of daily variation in outdoor
particulates and ambient acid species in normal and asthmatic children. Thorax, 57: 489-502. 025839
Wei Y; Han I-K; Shao M; Hu M; Zhang J; Tang X. (2009). PM2.5 Constituents and Oxidative DNA Damage in Humans.
Environ Sci Technol, 43: 4757^762. 192361
Wellenius GA; Bateson TF; Mittleman MA; Schwartz J. (2005). Particulate air pollution and the rate of hospitalization for
congestive heart failure among medicare beneficiaries in Pittsburgh, Pennsylvania. Am J Epidemiol, 161: 1030-
1036. 087483
Wellenius GA; Schwartz J; Mittleman MA. (2005). Air pollution and hospital admissions for ischemic and hemorrhagic
stroke among medicare beneficiaries. Stroke, 36: 2549-2553. 088685
Wellenius GA; Schwartz J; Mittleman MA. (2006). Particulate air pollution and hospital admissions for congestive heart
failure in seven United States cities. Am J Cardiol, 97: 404-408. 088748
Wellenius GA; Yeh GY; Coull BA; Suh HH; Phillips RS; Mittleman MA. (2007). Effects of ambient air pollution on
functional status in patients with chronic congestive heart failure: a repeated-measures study. Environ Health, 6: 1-
7. 092830
Welty LJ; Peng RD; Zeger SL; Dominici F. (2008). Bayesian Distributed Lag Models: Estimating Effects of Particulate
Matter Air Pollution on Daily Mortality. Biometrics, 65: 282-291. 157134
Welty LJ; Zeger SL. (2005). Are the acute effects of particulate matter on mortality in the National Morbidity, Mortality,
and Air Pollution study the result of inadequate control for weather and season? A sensitivity analysis using flexible
distributed lag models. Am J Epidemiol, 162: 80-88. 087484
Wheeler A; Zanobetti A; Gold DR; Schwartz J; Stone P; Suh HH. (2006). The relationship between ambient air pollution
and heart rate variability differs for individuals with heart and pulmonary disease. Environ Health Perspect, 114:
560-566. 088453
Wheeler BW; Ben-Shlomo Y. (2005). Environmental equity, air quality, socioeconomic status, and respiratory health: a
linkage analysis of routine data from the Health Survey for England. J Epidemiol Community Health, 59: 948-954.
188766
Whitsel E; Quibrera P; Christ S; Liao D; Prineas R; Anderson G; Heiss G (2009). Heart rate variability, ambient particulate
matter air pollution, and glucose homeostasis: the environmental epidemiology of arrhythmogenesis in the Women's
Health Initiative. Am J Epidemiol, 169: 693-703. 191980
Wilhelm M; Ritz B. (2005). Local variations in CO and particulate air pollution and adverse birth outcomes in Los Angeles
County, California, USA. Environ Health Perspect, 113: 1212-1221. 088668
Willis A; Jerrett M; Burnett RT; Krewski D. (2003). The association between sulfate air pollution and mortality at the
county scale: an exploration of the impact of scale on a long-term exposure study. J Toxicol Environ Health A, 66:
1605-1624. 089922
Wilson WE; Mar TF; Koenig JQ. (2007). Influence of exposure error and effect modification by socioeconomic status on
the association of acute cardiovascular mortality with particulate matter in Phoenix. J Expo Sci Environ Epidemiol,
17: S11-S19. 157149
Wong C-M; Atkinson RW; Anderson HR; Hedley AJ; Ma S; Chau PY-K; Lam T-H. (2002). A tale of two cities: effects of
air pollution on hospital admissions in Hong Kong and London compared. Environ Health Perspect, 110: 61-11.
023232
Wong C-M; Ou C-Q; Thach T-Q; Chau Y-K; Chan K-P; Ho S-Y; Chung RY; Lam T-H; Hedley AJ. (2007). Does regular
exercise protect against air pollution-associated mortality?. Prev Med, 44: 386-392. 093278
Wong CM; Ou CQ; Chan KP; Chau YK; Thach TQ; Yang L; Chung RY; Thomas GN; Peiris JS; Wong TW; Hedley AJ;
Lam TH. (2008). The effects of air pollution on mortality in socially deprived urban areas in Hong Kong, China.
Environ Health Perspect, 116: 1189-1194. 157151
December 2009 E-551
-------
Wong CM; Ou CQ; Lee NW; Chan KP; Thach TQ; Chau YK; Ho SY; Hedley AJ; Lam TH. (2007). Short-term effects of
participate air pollution on male smokers and never-smokers. Epidemiology, 18: 593-598. 098391
Wong CM; Vichit-Vadakan N; Kan H; Qian Z. (2008). Public Health and Air Pollution in Asia (PAPA): a multicity study of
short-term effects of air pollution on mortality. Environ Health Perspect, 116: 1195-1202. 157152
Wong TW; Tarn W; Tak Sun Yu I; Wun TY; Wong AH; Wong CM. (2006). Association between air pollution and general
practitioner visits for respiratory diseases in Hong Kong. Thorax, 61: 585-591. 093266
Wong TW; Tarn WS; Yu TS; Wong AHS. (2002). Associations between daily mortalities from respiratory and
cardiovascular diseases and air pollution in Hong Kong, China. Occup Environ Med, 59: 30-35. 025436
Woodruff TJ; Darrow LA; Parker JD. (2008). Air pollution and postneonatal infant mortality in the United States, 1999-
2002. Environ Health Perspect, 116: 110-115. 098386
Woodruff TJ; Grillo J; Schoendorf KG. (1997). The relationship between selected causes of postneonatal infant mortality
and particulate air pollution in the United States. Environ Health Perspect, 105: 608-612. 084271
Woodruff TJ; Parker JD; Schoendorf KC. (2006). Fine particulate matter (PM25) air pollution and selected causes of
postneonatal infant mortality in California. Environ Health Perspect, 114: 785-790. 088758
Xirasagar S; Lin HC; Liu TC. (2006). Seasonality in pediatric asthma admissions: the role of climate and environmental
factors. Eur J Pediatr, 165: 747-752. 093267
Yamazaki S; Nitta H; Ono M; Green J; Fukuhara S. (2007). Intracerebral haemorrhage associated with hourly
concentration of ambient particulate matter: case-crossover analysis. Occup Environ Med, 64: 17-24. 090748
Yang C-Y; Chang C-C; Chuang H-Y; Tsai S-S; Wu T-N; Ho C-K. (2004). Relationship between air pollution and daily
mortality in a subtropical city: Taipei, Taiwan. Environ Int, 30: 519-523. 055603
Yang C-Y; Chen Y-S; Yang C-H; Ho S-C. (2004). Relationship between ambient air pollution and hospital admissions for
cardiovascular diseases in Kaohsiung, Taiwan. J Toxicol Environ Health A, 67: 483-493. 094376
Yang CY (2008). Air pollution and hospital admissions for congestive heart failure in a subtropical city: Taipei, Taiwan. J
Toxicol Environ Health A, 71: 1085-1090. 157160
Yang CY; Chen CJ. (2007). Air pollution and hospital admissions for chronic obstructive pulmonary disease in a
subtropical city: Taipei, Taiwan. J Toxicol Environ Health A, 70: 1214-1219. 092847
Yang Q; Chen Y; Krewski D; Shi Y; Burnett RT; McGrail KM. (2004). Association between particulate air pollution and
first hospital admission for childhood respiratory illness in Vancouver, Canada. Arch Environ Occup Health, 59: 14-
21.087488
Yeatts K; Svendsen E; Creason J; Alexis N; Herbst M; Scott J; Kupper L; Williams R; Neas L; Cascio W; Devlin RB;
Peden DB. (2007). Coarse particulate matter (PM25-10) affects heart rate variability, blood lipids, and circulating
eosinophils in adults with asthma. Environ Health Perspect, 115: 709-714. 091266
Yue W; Schneider A; Stolzel M; Ruckerl R; Cyrys J; Pan X; Zareba W; Koenig W; Wichmann HE; Peters A. (2007).
Ambient source-specific particles are associated with prolonged repolarization and increased levels of inflammation
in male coronary artery disease patients. Mutat Res Fund Mol Mech Mutagen, 621: 50-60. 097968
Zanobetti A; Canner MJ; Stone PH; Schwartz J; Sher D; Eagan-Bengston E; Gates KA; Hartley LH; Suh H; Gold DR.
(2004). Ambient pollution and blood pressure in cardiac rehabilitation patients. Circulation, 110: 2184-2189.
087489
Zanobetti A; Schwartz J. (2002). Cardiovascular damage by airborne particles: are diabetics more susceptible?.
Epidemiology, 13: 588-592. 034821
Zanobetti A; Schwartz J. (2003). Multicity assessment of mortality displacement within the APHEA2 project. Health
Effects Institute. Boston, MA. 042812
Zanobetti A; Schwartz J. (2005). The effect of particulate air pollution on emergency admissions for myocardial infarction:
A multicity case-crossover analysis. Environ Health Perspect, 113: 978-982. 088069
Zanobetti A; Schwartz J. (2006). Air pollution and emergency admissions in Boston, MA. J Epidemiol Community Health,
60: 890-895. 090195
December 2009 E-552
-------
Zanobetti A; Schwartz J. (2007). Particulate air pollution, progression, and survival after myocardial infarction. Environ
HealthPerspect, 115: 769-775. 091247
Zanobetti A; Schwartz J. (2009). The effect of fine and coarse particulate air pollution on mortality: A national analysis.
Environ Health Perspect, 117: 1-40. 188462
Zeger S; Dominici F; McDermott A; Samet J. (2008). Mortality in the Medicare population and chronic exposure to fine
particulate air pollution in urban centers (2000-2005). Environ Health Perspect, 116: 1614-1619. 191951
Zeger S; McDermott A; Dominici F; Samet J. (2007). Mortality in the medicare population and chronic exposure to fine
particulate air pollution. Johns Hopkins University. Baltimore.http://www.bepress.com/jhubiostat/paperl33. 157176
Zeka A; Sullivan JR; Vokonas PS; Sparrow D; Schwartz J. (2006). Inflammatory markers and particulate air pollution:
characterizing the pathway to disease. Int J Epidemiol, 35: 1347-1354. 157177
Zeka A; Zanobetti A; Schwartz J. (2005). Short term effects of particulate matter on cause specific mortality: effects of lags
and modification by city characteristics. Occup Environ Med, 62: 718-725. 088068
Zeka A; Zanobetti A; Schwartz J. (2006). Individual-level modifiers of the effects of particulate matter on daily mortality.
Am J Epidemiol, 163: 849-859. 088749
Zhang J; Hu W; Wei F; Wu G; Korn LR; Chapman RS. (2002). Children's respiratory morbidity prevalence in relation to
air pollution in four Chinese cities. Environ Health Perspect, 110: 961-967. 034814
Zhang Z; Whitsel E; Quibrera P; Smith R; Liao D; Anderson G; Prineas R. (2009). Ambient fine particulate matter
exposure and myocardial ischemia in the Environmental Epidemiology of Arrhythmogenesis in the Women's Health
Initiative (EEAWHI) study. Environ Health Perspect, 117: 751-756. 191970
Zhong W; Levin L; Reponen T; Hershey GK; Adhikari A; Shukla R; LeMasters G. (2006). Analysis of short-term
influences of ambient aeroallergens on pediatric asthma hospital visits. Sci Total Environ, 370: 330-336. 093264
December 2009 E-553
-------
Reference: Gent et al.
(2009, 180399)
Location: 2 monitors in
New Haven, CT/ 3.5 yr
Particle Size: PM2.5
Subjects: N: 149
Children with children
physician
diagnosed
asthma and
symptoms or
medication use
in previous 12
mo, and resided
within 30km of
New Haven
county monitor
Exposure: NR
Number of Grouping
Constituents method: PCA
considered for
grouping: 17 # of groups: 6
elements + EC
Groups/Factors/ Sources:
Vehicle (EC, Zn, Pb.Cu,
Se), road dust (Si, Fe.AI,
Ca.Ba.Ti), sulfur (S,P),
biomass burning, (K) oil (V,
Ni), sea salt (Na.CI)
In addition, effects of N02,
CO, S02, and 03 were
included in the health
outcomes model
PM variables used:
Groupings and
individual elements
Results: Overall: Trace elements originating from motor vehicle, road dust, biomass burning, and oil sources associated with
symptoms and/or medication use. No associations with S or sea salt.
Specific Results: PM2.5 mass from motor vehicle or road dust associated with increased odds of respiratory symptoms or
inhaler use. Reduced odds of wheeze or inhaler use with same day S. Significant reductions odds of wheeze with biomass
burning.
Co-pollutant: Positive effects of motor vehicles and road dust on wheeze were robust to the inclusion of gaseous copollutants.
However, N02 increases association with wheeze.
Reference: Ito et al. Subjects: NR N: NR
(2006, 188554)
Exposure: NR
Location: Washington,
DC
Particle Size: PM25
Number of
Constituents
considered for
grouping: NR
Grouping
method:
Comparison of:
PMF; (absolute)
PCA; UNMIX
# of groups: 6-10
Groups/ Factors/
Sources: Different
research groups
gave different
names to sources
Sources for which
association with health
was analyzed: Soil, traffic,
Secondary S04, N03
(Washington, DC only),
residual oil (Washington,
DC only), Wood smoke/
biomass combustion, Sea
salt, incinerator
(Washington, DC only),
primary coal (Washington,
DC only), Cu smelter
(Phoenix only)
PM variables used:
Mass contribution of
sources
Results: Overall, PM25 effects observed at lag 3. Lag structure of association varied across source types, but consistent
across investigators for total (nonaccidental mortality): soil factor - mostly positive at various lags (not significant); secondary
sulfate - strongest association at lag 3; nitrate - mostly negative except at lag 3; residual oil - strongest association at lag 2 (not
significant); wood-burning - increasing association as lag increases (not significant); incinerator - significant negative
associations at lag 0; primary coal- significant association at lag 3.
Reference: Laden et al.
(2000, 012102)
Location: Monitors in 6
Eastern US cities (Harvard
Six Cities Study)
Particle Size: NR
Subjects: NR N: NR
Exposure: NR
Number of Grouping
Constituents method: PCA
considered for
grouping: 15 # of groups: 8
elements
Groups/ Factors/ Sources: PM variables used:
Soil/crustal (PM fine), Tracers: Si, V, Cl, Pb,
mobile vehicle exhaust (PM Se
fine), coal (PM fine), fuel oil;
metals, salt manganese,
residual
Results: Lag 0-1 avg for all results. Overall 6 cities, mobile source factor (using Pb as tracer) had greatest association with
daily mortality (3.4%) with 10 ug/m3 increase. The greatest effects for mortality due to mobile sources were observed in
Madison (Portage), Knoxville (Kingston-Harriman), and St. Louis, although the Madison results were not statistically significant.
The coal source factor was only significant in Boston (Watertown) - 2.8% increase in mortality and the overall percent increase
was also significant (1.1%). Deaths from pneumonia attributable to coal combustion sources was 7.9% (Cl 3.1-12.7%) and
statistically significant. The crustal factor was not associated with mortality in any city, although this factor was not a significant
predictor in the regression model for Boston (Watertown) due to its low contribution to PM2.5 mass. For specific elements
included simultaneously, S, Pb, and Ni were significantly associated with overall mortality (3.0,1.6,1.5%, respectively). Boston
had the greatest percent increase in mortality for S (7.9%), Knoxville for Pb (15.0%), and Steubenville for Ni (8.2%), although
the CIs are all quite large.
Reanalysis results: (Schwartz, 2003, 042811) Effects changed slightly. New percent increases in mortality for combined cities
are 3.5 and 0.79 for traffic and coal, respectively. The coal factor in Boston decreased to 2.1 % increased mortality. A residual
oil factor in Boston and Steubenville resulted in at 22.9% increase in daily deaths (but was not significant in the original
analysis).
Reference: Lanki et al.
(2006, 088412)
Location: Monitors in
Helsinki, Finland,
Amsterdam,The
Netherlands and Erfurt,
Germany
Subjects: NR
Exposure: NR
N:NR
Number of
Constituents
considered for
grouping: 13
elements
Grouping
method: Absolute
PCA
# of groups: 5
Groups/ Factors/ Sources:
Crustal; long range
transported; oil combustion;
soil; traffic
Results: Highest observed effects were for crustal sources and salt at lag 3 (when analyzing sources),
PM variables used:
Tracers: Si (crusta I) ;S
(long-range transport);
Ni (oil combustion) ;CI
(salt) ;ABS (local
traffic).
but not consistent or
.._,_.. significant. In multipollutant models only ABS associated with ST-segment depression, but wide CIs. When examining indicator
Particle Size: UF/PM2 5 elements of a source, local traffic found to be the most toxic, but when examined per IQR long-range transport and traffic had
similar effects.
December 2009
F-2
-------
Results: All had significant associations with mortality. Traffic density and EC had the largest effects.
Reference: Lippmann et
al. (2006.091165)
Location: U.S.
Subjects: NR N: NR
Exposure: NR
Number of
Constituents
considered for
grouping: NR
Grouping
method: No
grouping was
performed
Groups/ Factors/ Sources:
NR
PM variables used:
Mass contribution of
16 constituents
Particle Size: PM10 for risk
estimates, PM2.5for
speciation data
# of groups: NR
Results: The strongest predictions of the variation in PM10 risk estimates across the 90 NMMAPs MSAs was for Ni and V.
Elevated, but nonsignificant increases were associated with EC, Zn, S042-, Cu, Pb, and OC. Al and Si had the lowest values.
Reference: Mar et al.
(2000, 001760)
Location: 1 monitor in
Phoenix, AZ
Particle Size: NR
Subjects: N: NR
Elderly only
Exposure: NR
Number of
Constituents
considered for
grouping: 10
elements, OC,
EC, CO, N02;
S02
Grouping
method:
Unspecified type of
factor analysis
# of groups: 3 or 5
Groups/ Factors/ Sources:
Motor exhaust/road dust,
soil, vegetative burning,
local S02, regional S04
PM variables used:
First used individual
constituents: S,Zn,
Pb, K,OC,EC,TC
(AL+Si+Ca+Fe+Ti),
then factor scores
Results: Cardiovascular mortality associated with PM25 mass on lag 1 and 4 (6 and 4%, respectively). EC and TC associated
with CV mortality for lag 1 (RR= 1.05);OC was weakly associated with CV mortality for lags 1 and 3. For total mortality,
regional sulfate was positively associated at lag 0, but negatively associated at lag 3. The local S02 and the soil factors were
negatively associated with total mortality. For CV mortality, secondary sulfate was positively associated at lag 0, motor vehicle
at lag 1, and vegetative burning at lag 3.
Reanalysis results (Mar, 2003): Similar associations were observed.
Reference: Mar et al. Subjects: NR N: NR
(2006, 086143)
Exposure: NR
Location: Phoenix, AZ
Particle Size: PM25
Number of
Constituents
considered for
grouping: NR
Grouping
method:
Comparison of:
PMF (absolute);
PCA; UNMIX
# of groups: 6-10
Groups/ Factors/
Sources: Different
labs gave different
names to sources
(see Hopke et al,
table 2)
Sources for which
association with health was
analyzed: Soil, Traffic,
secondary S04, N03,
(Washington, DC only),
residual oil (Washington,
DConly),woodsmoke/
biomass combustion, sea
S3 it, incinsrstor
(Washington, DC only);
primary coal (Washington,
DC only) ;Cu smelter
(Phoenix only)
PM variables used:
Mass contribution of
sources
Results: Using daily PM25 data found the following associations with cardiovascular mortality: Secondary sulfate - greatest
effect observed for all sources and at lag 0; traffic - associated at lag 1; copper smelter associated at lag 0; sea salt - had the
greatest statistical significance and observed at lag 5; biomass/wood burning - less consistent lag structure but greatest
association at lag 3; soil - did not show an association or consistent lag structure. For total (nonaccidental) mortality
associations were weaker and consistently observed for only: copper smelter - lag 0; sea salt - lag 5.
Reference: Ostro et al.
(2007, 091354)
Location: Monitors in 6
CA counties, some with 2
monitors, for 4 yr
Particle Size: PM25
Subjects NR N: NR
Exposure: NR
Number of
Constituents
considered for
grouping: 15
elements, EC,
OC;N03;S04,
PM2.5 mass
Grouping
method: No
grouping was
performed
# of groups: NA
Groups/ Factors/ Sources: PM variables used:
NR Mass contribution of
every constituent
Results: Effects were greater during the winter months. In the all year analysis, at 3-day lag associations observed for EC, OC,
N03 and Zn. During winter months (Oct -March) effects observed for most species for both all-cause and cardiovascular
mortality at lag 3 (EC, OC, S04, Ca, Fe, K, Mn, Pb, S, Si, Ti, Zn) and (OC, N03, S04, Fe, Mn, S, V, Zn), respectively.
Reference: Ostro et al.
(2009,191971)
Location: Monitors in 6
CA counties, some with 2
monitors/4 yr
Particle Size: PM25
Subjects NR N: NR
Exposure: NR
Number of
Constituents
considered for
grouping: 9
elements, EC, OC,
PM25mass,S04,
N03
Grouping
method: No
grouping was
performed
# of groups: NA
Groups/ Factors/ Sources:
NR
PM variables used:
Mass contribution of
every constituent
Results: The following associations were observed with cardiovascular mortality: PM25 (lag 3); EC (lag 2); N03 (lag 3); S04
(lag 3); Fe (lag 2); K (lag 2); S (lag 3); Ti (lag 2); Zn (lag 3).
December 2009
F-3
-------
Reference: Peng et al.
(2009,191998)
Location: 119 urban
communities STN
data/2000-2006
Particle Size: PM25
Subjects:
Medicare
enrollees 65 or
older
Exposure: NR
N:NR
Number of
Constituents
considered for
grouping: S04,
N03,Si, EC,
OCM, Na, NH4
Grouping
method: NR
# of groups: NR
Groups/Factors/ Sources: PM variables used:
Only suggested in Tracers
discussion
Results: CVD HAs: EC associated with same-day CVD HAs in single and multi-pollutant models. In single pollutant models
associations also observed for sulfate, nitrate, OCM, and ammonium. However, the sulfate, nitrate, OCM, and ammonium
associations were reduced in the multi-pollutant models.
Respiratory HAs: OCM associated with same-day respiratory HAs in single and multi-pollutant models. Some evidence for
sulfate associations at one and two-day lag.
Reference: Penttinen et
al. (2006, 087988)
Location: Helsinki
1996-1 997 (7 mo)
Particle Size: PM2.5
Subjects: Adult N: 78
asthma subjects,
max 2 km from
single monitor
Exposure: NR
Number of
Constituents
considered for
grouping:
Unknown
Grouping
method: PCA
# of groups: 6
Groups/Factors/ Sources:
Long range (PM mass, S,
K,Zn), local
combustion-traffic (Cu, Zn,
Mn,Fe),soil(Si,AI, Ca, Fe,
Mn), oil (V, Ni), salt (Na, Cl),
unidentified
PM variables used:
every component
individually, then
groupings
Results: Long range PM2.5 associated with decreased mean PEF in the morning at lag 1. Local combustion PM2.5 associated
with decreased mean PEF in the evening for lag 1. Local combustion PM2.5 associated with decreased mean PEF in the
afternoon and evening for 5-day mean lag. Negative significant association between long-range PM2 5 and asthma symptom
prevalence at lag 3. Sea-salt PM25 negatively associated with bronchodilator use at lag 3 and 5-day mean lag. Sea-salt PM25
negatively associated with corticosteroid use for 5-day mean lag. Unidentified PM25 negatively associated with corticosteroid
use at lag 1. Most consistent negative responses for local combustion, although not always significant. No consistent or
significant associations between 5-day avg concentrations of elements and PEF, cough, asthma symptoms, or medication use.
Reference: Riediker et al.
(2004, 091261)
Location: Inside 9 state
police patrol cars
Particle Size: PM25
Subjects: N: 9
Healthy male
young police
officers
Exposure: 4
consecutive days
Number of Grouping Groups/ Factors/ Sources:
Constituents method: PCA Soil; automotive steel wear;
considered for gasoline combustion;
grouping: 10 #of groups: 4 speed-changing traffic
elements; 3 when 13+2
gaseous constituents
pollutants; 2 included; 3 when
physical variables on|y9
"PM-associated"
constituents
included
PM variables used:
Mass contribution or
score of sources
Results: Using two different factor analysis models found most significant effects (MCL, SDNN, PNN50, supraventricular
ectopic beats, % neutrophils, % lymphocytes, MCV, von Willebrand Factor, and protein C) were for "speed-change factor" (i.e.,
Cu, S, aldehydes). Some associations observed for "crustal" and none for "steel wear" and "gasoline."
Reference: Sarnat et al.
(2008, 097972)
Location: 1 monitor in
Atlanta. GA for 2 yr
Particle Size: PM25
Subjects: NR N: NR
Exposure: NR
Number of Grouping Groups/ Factors/ Sources:
Constituents method: gasoline, diesel, wood
considered for Comparison of: smoke/ biomass burning,
grouping: NR PMF, CMB-LGO, "a soil, secondary
priori decision" S04/ammonium sulfate,
secondary nitrate/
# of groups: 9,11 ammonium nitrate, metal
(6 of them common processing, railroad, bus
between methods) anc| highway, cement kiln,
power plants, other OC,
ammonium bisulfate
PM variables used:
Mass contribution or
score of sources, and
tracers
Results: Sulfate secondary associated with 1.2-2.0% increase in RD visits, significant negative association RD visits and
primary emissions from coal-fired power plants. CVD significantly associated with other OC (1.014), biomass (1.033), diesel
and gas for CMB-LGO. For PMF and CVD visits: diesel (1.025), gas, wood smoke, metal processing (1.013). Year-long
associations: PMF diesel, EC, CMB-LGO gas, Zn and biomass combustion sources (CMB-LGO biomass burning, PMF wood
smoke, and K). Diesel and gas sources association with RD in the warm season (1.2-2.1% per IQR).
Reference: Schreuder et
al. (2006. 097959)
Location: 1 monitor in
Spokane, WA for 7 yr
Particle Size: PM25
Subjects: NR
Exposure: NR
N:NR
Number of
Constituents
considered for
grouping: 11
elements,TC,
N03
Grouping
method:
Comparison of:
PMF, UNMIX,
Multilinear Engine
# of groups: 8
Groups/ Factors/ Sources: PM variables used
Vegetative burning; As-rich Tracers:TC
Vehicle; S04;N03; Soil;
Cu-rich; Marine
(vegetative burning);
As (As-rich) ;Zn
(vehicle); Si (soil)
Results: Si, As, and Zn were not associated with any health outcomes; while an IQR increase in TC (vegetative burning) was
associated with a 2% increase in respiratory ED visits.
December 2009
F-4
-------
Reference: Tsai et al.
(2000,006251)
Location: 3 NJ sites for 2
summers (ATEOS study)
Particle Size: NR
Subjects: NR N: NR
Exposure: NR
Number of
Constituents
considered for
grouping: 8
metals, IPM,
FPM,S04,CX,
DCM.ACE.CO
Grouping
method:
Unspecified type of
factor analysis
# of groups: 5
Groups/ Factors/ Sources:
Oil burning, motor
emissions, resuspended
dust, secondary aerosol,
industrial sources
PM variables used:
individual constituents,
then factor scores,
then tracers
Results: RR associated with 10 ug/m3 increases: Newark -1.03 for industrial and total daily deaths; 1.02 for sulfate and total
daily deaths; 1.04 for sulfate and cardiopulmonary deaths. Camden-1.11 for oil burning sources and total daily deaths; 1.10
industrial and total daily deaths; 1.12 for oil burning sources and cardiopulmonary daily deaths; 1.02 for sulfate and
cardiopulmonary daily deaths
Reference: Yue et al.
(2007, 097968)
Location: 1 monitor in
German city, 30.000
samples
Particle Size: UF/PM25
Subjects: Adult
males
Exposure: CAD
n: 56, data
collected 12
times over 6
mo for every
subject, but
extended
period of
missing PM
data
Number of
Constituents
considered for
grouping:
Apportionment
based on particle
size distribution.
Grouping
method: PMF
# of groups: 5
Groups/ Factors/ Sources:
Airborne soil, local traffic,
local fuel combustion,
remote traffic (diesel),
secondary aerosols
PM variables used:
Mass contribution or
score of sources
Results: Overall, repolarization parameters influenced by traffic-related particles; vWF increased in response to traffic-related
particles and combustion-generated aerosols. All source factors contributed to increasing CRP levels.
December 2009
F-5
-------
Table F-2. Human clinical studies of ambient PM sources, factors, or constituents
Study: Gong et al.
(2003, 042106)
Location: Los
Angeles, CA
Particle Size: PM25
Study: Gong et al.
(2005, 087921)
Location: Los
Angeles, CA
Particle Size: PM2.5
Reference: Huang
etal.(2003,
087377)
Location: Chapel
Hill, NC
Particle Size: PM2.5
Reference: Urch et
al. (2004, 055629)
Location: Toronto,
Canada
Particle Size: PM25
Reference: Urch et
al. (2004, 055629)
Location: Toronto,
Canada
Particle Size: PM25
Subjects: Adult N: 12 healthy, 12 Constituents Grouping method: Groups/ Factors/ PM variables used:
18-45, healthy vs. asthmatic considered for PCA Sources: Crustal (Al Total mass, then
asthmatic grouping: 7 SiCAKFe),S(2 tracers: S04, EC, Fe
elements, EC, N03, #of groups: 4 metrics of S04 +
Exposure: CAPs, sc,4 (note :OC data was elemental S) Total
healthy and unavailable) Mass+N03, EC
asthmatic subjects
exposed at different
times
Results: Fe and EC associated with a decrease in ST-segment voltage 2 days post-exposure. EC associated with an increase in
ST-segment voltage immediately following exposure. Sulfate content associated with a decrease in systolic BP 4 h post-exposure.
Subjects: Elderly, N: 6 healthy, 18 Constituents Grouping method: Groups/ Factors/ PM variables used:
COPD vs. healthy/ COPD considered for PCA Sources: Crustal (Al Total mass, then
CAPs grouping: 7 SiCAKFe), tracers:, S04, Si, Fe,
elements # of groups: 3 S(=S04),Na EC
Exposure: N02 (full + EC (note: OC was
factorial) unavailable)
Results: Mass concentration of CAPs not observed to significantly affect lung function. However, sulfate content was associated with
a decrease lung function (FEV, and FVC), which was enhanced by coexposure to N02.
Subjects: Healthy N: 35 male; 2 female Constituents Grouping method: Groups/ Factors/ PM variables used:
adults considered for PCA Sources: Factor scores, then
grouping: 8 Fe/S04/Se/V/Zn/Cu mass contribution of
Exposure: CAPs elements and S04 # of groups: 2 all 9 constituents
Results: Associations observed between sulfate ,Zn, and Se content and increases in BAL neutrophils. Increases in fibrinogen
associated with Cu, Zn, and V content.
Subjects: Healthy N: 23 Constituents Grouping method: Groups/ Factors/ PM variables used:
adults considered for No grouping was Sources: NR Every constituent in
19-50 yr/CAPs grouping: unknown performed univariate analysis,
then OC and S04 in
Exposure: 03 # of groups: NA multivariate analysis
Results: CAPs-induced increase in diastolic BP significantly associated with carbon content of the particles.
Subjects: Healthy N: 24 Constituents Grouping method: Groups/ Factors/ PM variables used:
adults/CAPs considered for No grouping was Sources: NR Every constituent in
grouping: 14 performed univariate analysis,
Exposure: 03 elements, EC, OC then OC and S04 in
# of groups: NA multivariate analysis
Results: Both organic and EC content of CAPs associated with an increase in brachial artery vasoconstriction.
December 2009
F-6
-------
Table F-3. Toxicological studies of ambient PM sources, factors, or constituents
Reference:
Batalha et al.
(2002, 088109)
Location:
Boston, MA
Particle Size:
PM2.5
Reference:
Becker et al.
(2005, 088590)
Location:
Chapel Hill, NC;
repeated
sampling for 1 yr
Particle Size:
PM10
Reference:
Clarke et al.
(2000, 013252)
Location:
Boston, MA
Particle Size:
PM,,
Subjects: Rats N: 7-10 rats * Constituents Grouping Groups/ Factors/ Sources: V/Ni, PM variables
2 levels CAPs considered for method: S,AI/Si, Br/Pb used: 4 tracers (Si,
Exposure: CAPs x 2 levels S02 grouping: 20 Previous study S04,V, Pb) and
(3-day mean CAPs x 6 runs in elements ;OC; EC in same city EC.OCin
concentration range: different (Clarke et al., univariate step. 4
126.1 -481 .Oug/m3) seasons 2000.013252) tracers (Si, S04, V,
CAPs (3-day mean anc| RCA of this Pb) in multivariate
CAPs concentration experiment's step
range: data
126.1 -481 .Oug/m3)
# of groups: 4
Results: Univariate analyses for first day not significant for UW ratio. Univariate analyses for second and third day and second+third day
mean were similar. Presented second+third day mean regression data. CAPs mass, Si, Pb, S04, EC, OC significant for decreased L/W
ratio in normal+CB rats exposed to CAPs. Si, S04 significant for decreased L/W ratio in normal rats. Si, OC significant for decreased UW
ratio in CB rats. Multivariate analysis using normal+CB rats for Si, S04, V, Pb - only Si remained significant with decreased L/W ratio.
Subjects: Normal N: NR Constituents Grouping Groups/ Factors/ Sources: PM variables
human bronchial considered for method: PCA Cr/AI/Si/Ti/Fe/Cu ("crustal"), used:NR
epithelial and human grouping: 12 Zn/As/V/Ni/Pb/
AM elements # of groups: 2 Se
Exposure: (2-3X105
cells/mL;11 or 50
ug/mL)
Results: Cr/AI/Si/Ti/Fe/Cu associated with IL-8 release in normal human bronchial epithelial cells and IL-6 release in AM.
Zn/As/V/Ni/Pb/Se not associated with any endpoints. Stepwise linear regression with individual constituents Fe and Si associated with
IL-6 release in AM. Cr associated with IL-8 release in NHBE cells.
Subjects: Dogs N: 10 dogs, 20 Constituents Grouping Groups/ Factors/ Sources: V/Ni, PM variables
paired considered for method: PCA S,AI/Si, Br/Pb, S, Na/CI, Cr used: All elements,
Exposure: CAPs (avg exposures, 24 grouping: 19 then factor scores
for all studies, paired: crossover elements, black C # of groups: 4
203.4, crossover: for exposure in
360.8 ug/m3) repeated paired runs,6 for
exposure with several exposure in
weeks in between crossover runs
Results: No significant differences between baseline, sham, or CAPs group for BAL cell differential percentages. Total BAL protein
increased with CAPs compared to sham. No significant hematological effects with CAPs exposure. Mixed linear regression analyses
(statistics not provided): Al and Ti (3-day avg. concentrations) associated with dose-dependent decreases in BAL AM and increases in
BAL PMN percentages. Sulfate associated with increased WBC. BC, Al, Mn, Si, Zn, Ti, V, Fe, Ni associated with increased blood PMN.
Na associated with increased blood lymphocytes. Al, Mn, Si associated with decreased blood lymphocytes. CAPs mass and BC
associated with decreased blood eosinophils. CAPs mass associated with decreased platelet count. Regression using results of factor
analysis: None for 3-day avg. concentration for BAL parameters. V/Ni for increased AM percentage and Br/Pb for increased PMN
percentage for 3rd-day only concentration. V/Ni and Al/Si for increased blood PMN percentage and decreased blood lymphocyte
percentage. Al/Si also for increased WBC counts. Na/CI for increased blood lymphocyte percentage. S for decreased RBC and
hemoglobin.
Reference:
Duvall et al.
(2008, 097969)
Location: 5 US
cities
Particle Size:
PM,,
Subjects Primary N: NR
human airway
epithelial cells
(1 00,000 cells/mL;
dose not provided)
Exposure: NR
Constituents
considered for
grouping: NR
Grouping
method: CMB,
but not on
coarse and
ultrafine
# of groups: 6
or?
Groups/ Factors/ Sources:
Mobile, residual, oil, wood, soil,
secondary S04, secondary N03
PM variables
used: Mass
contribution of
constituents, then
mass contribution
of sources
Results: Linear regression with individual constituents: Sulfate associated with increased IL-8 mRNAexpression. Sr associated with
increased COX-2 and decreased HO-1 mRNA expressions. K associated with decreased HO-1 mRNA expression.
Linear regression with sources: Significance levels not provided.
December 2009
F-7
-------
Reference:
Godleski et al.
(2002, 156478)
Location:
Boston, MA
Particle Size:
NR
Reference:
Gurgueira et al.
(2002, 036535)
Location:
Boston, MA
Particle Size:
PM2.5
Reference:
Kodavanti et al.
(2005, 087946)
Location: RTP,
NC
Particle Size:
PM2.5
Reference:
Lippmannetal.
(2005, 087453)
Location: Rural
location upwind
from New York
City
Particle Size:
PM2.5
Reference:
Lippmann etal.
(2006, 091165)
Location: Rural
location upwind
from New York
City
Particle Size:
PM2.5
Subjects: Rats N: 7-10 rats * Constituents Grouping Groups/ Factors/ Sources: V/Ni, PM variables
2 levels CAPs considered for method: S,AI/Si/Ca, Br/Pb used: 4 tracers (I,
Exposure: CAPs x 2 levels S02 grouping: 20 Previous study S04,V, Pb) and EC,
(3-day mean CAPs x 6 runs in elements, OC, EC in same city OC
concentration range: different (Clarke et al.),
126.1 -481 .Oug/m3) seasons andPCAofthis
experiment's
data
# of groups: 4
Results: Increased percent of PMNs in BALF in CAPs-exposed rats at 24 h. CAPs affected lung tissue mRNA involved in
pro-inflammation, immune, and vascular endothelial responses. Linear regression: Increased PMN associated with CAPs mass, Br, Pb,
S04, EC, and OC.
Subjects Rats N: 13 Constituents Grouping Groups/ Factors/ Sources: NR PM variables
(Sprague Dawley) experiments considered for method: No used: Mass
(1 rat/group at grouping: 20 grouping was contribution of
Exposure: CAPs (avg. each time elements performed every constituent
mass concentration no'ml)
600 ug/m3); also #of groups: NA
carbon black and
ROFA
Results: Increased oxidative stress in heart and lungs following CAPs exposure (and ROFA exposure).
Univariate regression: Mn, Zn, Fe, Cu, and Ca most significant responses for lung (r2>0.40). Al, Si, Ti, Fe, and total mass most significant
response for heart (r >0.49).
Subjects Rats (SH and N: 6 1-day, Constituents Grouping Groups/ Factors/ Sources: NR PM variables
l/VKY) 1-strain runs, considered for method: No used: Mass
72-day, grouping : NR grouping was contribution of
Exposure: CAPs 2-strain runs, performed every constituent
(1 44-2758 ug/m3) 4-9 rats per
run # of groups: NA
Results: No significant correlations between biologic responses and exposure variables (i.e., CAP mass, OC, inorganic C, sulfate, and
other major elemental constituents). Al, Cu, Zn correlated with biologic responses when constituents normalized per unit mass of CAP
(ug/mg). Zn correlated with plasma fibrinogen in SH rats (p = 0.0023).
Subjects: Mice (C57 N:C57:3-6 Constituents Grouping Groups/ Factors/ Sources: PM variables
andApoE) mice/group considered for method: Regional S04(S/Si/OC); used: Mass
, grouping: 19 (Absolute) PCA Resuspended soil contribution of
Exposure: CAPs (avg. ApoE: 9-10 elements + OC, EC, (CA/Fe/AI/Si);RO power plants sources
mass concentration mice/group njo3 # of groups: 4 (V/Ni/SeV
1 1 3 ug/m3) traffic and unknown
Results: ApoE null mice: Resuspended soil associated with decreased HR during exposure, but increased HR after exposure.
Secondary sulfate associated with decreased HR after exposure. Residual oil associated with increased RMSSD and SDNN in afternoon
following exposure. Secondary sulfate associated with decreased RMSSD and SDNN in night following exposure. Resuspended soil
associated with increased RMSSD at night following exposure. PM mass associated with decreased HR during exposure and decreased
RMSSD at night following exposure.
C57 mice: Motor vehicle/other source category associated with decrease in RMSSD in afternoon following exposure
Subjects: Mice N: 12 ApoE''' Number of Grouping Groups/ Factors/ Sources: NR PM variables
(ApoE") mice (6/group) Constituents method: No used: Mass
considered for grouping was contribution of
Exposure: CAPs (avg. grouping: NR performed every constituent in
mass concentration CAPs portion of
85.6 ug/m ) # of groups: NR study contribution
of 16 constituents
in epi portion
Results: Lag for HR elevations on 14 days with wind from NWwas same day. Lag for SDNN reduction on 14 days with wind from NW
was 0, 1 and 2.
GAM analysis: B coefficient significant for Ni and HR (but not V, Cr, or Fe). B coefficient significant for Ni and log SDNN (but not V, Cr, or
Fe).
December 2009
F-8
-------
Reference:
Maciejczyk and
Chen (2005,
087456)
Location: Rural;
Subjects: NR
Exposure: CAPs
(90,000/well;300
ug/mL)
N:110
samples
Constituents
considered for
grouping: 19
elements + OC, EC,
N03
Grouping
method:
(Absolute) PCA
# of groups: 4
Groups/ Factors/ Sources:
Regional S04
soil; unknown
oil combustion
PM variables
used: Mass
contribution of
sources
upwind from New Results: Correlation: V and Ni positively correlated with NF-KB. Oil combustion correlated the greatest with NF-KB (0.316). Significance
upw
York
'ork City
not provided. Only 2% of mass contribution originates from this source.
Particle Size:
PM2.5
Reference: Subjects: Dogs
Nikolov et al.
(2008, 156808) Exposure:
Location:
Boston, MA
Particle Size:
NR
N: 8 dogs, 24
exposure-days
in 1997-98; 4
dogs, 21
exposure-days
in 2001 -2002
Constituents
considered for
grouping: 13
elements, BC, EC,
OC
Grouping
method:
Compared 3
factor-analytic
models within a
SEM model
# of groups: 4
Groups/ Factors/ Sources: Oil
Combustion V/Ni; power plants
S ;roaddustAI/Si ;motor vehicles
BC/OC/EC
PM variables
used: Mass
contribution of
every constituent
Results: Univariate response for respiratory outcomes: road dust and oil combustion associated with decreased respiratory frequency;
motor vehicles associated with increased respiratory frequency; motor vehicles associated with increased PEF; road dust associated
with decreased penh and motor vehicles associated with increased penh.
Multivariate responses for respiratory outcome: Road dust associated with decreased respiratory rate; Motor vehicles associated
with increased airway irritation.
Reference:
Rhoden et al.
(2004, 087969)
Location:
Boston, MA
Particle Size:
PM2.5
Reference:
Saldiva et al.
(2002, 025988)
Location:
Boston, MA
Particle Size:
PM2.5
Subjects: Rats N: 4-8 rats Constituents
(Sprague-Dawley) (1-2 per group considered for
-sham, CAPs, grouping: 20
Exposure: CAPs (avg. sham/NAC, elements
mass concentration CAP/NAC)
range 150-2520 10 exposures
ug/m3) acetylcysteine
full factorial
Grouping Groups/ Factors/ Sources: NR
method: No
grouping was
performed
# of groups: NA
PM variables
used: Mass
contribution of
every constituent
Results: Increased oxidative stress and inflammation in lungs of CAPs animals that was attenuated with NAC.
Univariate regression: Al, Si, Fe, K, Pb, and Cu most significantly correlated with lung TBARS. No significant correlations for lung
carbonyls or lung PMN.
Subjects: Rats N:7-10 Constituents
(Sprague-Dawley rats/group considered for
(air/sham, grouping: 15
Exposure: CAPs S02/sham, elements (used
(3-day avg. mass air/CAP, Clarke 2000 to
concentration range S02/CAP) x 6 select tracers)
126.1-481 ug/m3) mns in
different
seasons
Grouping Groups/ Factors/ Sources: V/Ni
method: S
Previous study Al/Si
in same city Br/Pb
(Clarke et al.
2000) Na/CI
Cr
# of groups: 6
PM variables
used: Mass
contribution of 8
elements in
univariate step.
Tracers (Si, S04, V,
Pb, Br, Cl) and EC,
OC in multivariate
step.
Results: Increased percent and number of PMN in majority of air and S02 rats exposed to CAPs, but significance levels not provided.
Other responses (protein, LDH, NAG) were variable and depended upon the CAPs exposure. No CAPs effect on histopathology.
Linear regression: V, Br, Pb, S04, EC, OC, Si, CAP mass associated with increased PMN and lymphocytes for normal+CB rats. Only V
not associated with PMN in normal rats. Lymphocyte response due to CB rats, but not observed for S04, Si, or mass in this group. Br,
Pb, S04, EC, OC, Si associated with increased total protein in CB rats. Cl and V associated with decreased LDH in CB rats. No BAL
effects in normal rats exposed to CAPs. V, Br, Pb, EC, OC, and Cl associated with increased neutrophil density in lungs of normal rats.
Reference:
Seagrave et al.
(2006, 091291)
Location: 4 SE
US sites for 2
seasons
Particle Size:
PM2.5
Subjects Rats (Fisher N: 5 rats/dose
344)
Exposure: 0.75, 1.5
and 3 mg/rat via
intratracheal instillation
Constituents
considered for
grouping: NR
Grouping Groups/ Factors/ Sources: PM variables
method: CMB secondary N03; secondary NH4; used: Mass
secondary S04; coke production; contribution of
# of groups: 1 3 vegetative detritus ; natural gas every constituent,
combust; road dust; wood then mass
combust; meat cooking gasoline; contribution of
diesel other OM; other mass sources
Results: Potency depended upon season and site of sample collection. In general, effects were greater in the winter.
PLS analysis: 2 major constituents identified (OC, Pb, hopanes/steranes, nitrate, As for first and major metal oxides for the second),
gasoline most important predictor for both constituents, with diesel influencing second constituent and nitrate influencing first constituent.
First constituent affected cytotoxic responses, second constituent affected inflammatory responses.
December 2009
F-9
-------
Reference: Subjects: BEAS-2B N: 6; 16 runs
Veranthetal. cells (35000 cells/cm2; over 6 mo
(2006. 087479) 10, 20,40,80 ug/cm2)
Location: 8 sites Exposure: Loose
in the western surface soil sweepings
US through mechanical
_ .. , _. tumbler and cascade
Particle Size: impactor
PM25
Constituents Grouping Groups/ Factors/ Sources: NR
considered for method: PLS
grouping: 13
elements, TC, 5 OC # of groups: NR
variables, 4 EC
variables, 2 ions,
EU, one ratio (Ca:
AI),OP,C03
PM variables
used: Mass
contribution every
constituent (?)
Results: Dose-related increase in IL-6 and decreases in cell viability for all soil types. IL-8 responses more variable and dependent upon
soil type. Univariate correlations. Low correlations for all constituents tested with IL-6. Highest correlations for EC1 (R = 0.50) and
pyrolyzed OC (R2 = 0.46), then Ca/AI (R2 = 0.21). Carbonate carbon, ECS, and Sr correlated with IL-8 (R2 = 0.27, 0.13, and 0.25,
respectively). EC and Ni correlated with IL-8 trend over the range of 10-80 ug/cm2 (R2= 0.39 and 0.27, respectively). Multivariate
redundancy analysis OC1, OC3, OC2, EC2, Br, EC1, Ni correlated with IL-8 release, decreased viability, and decreased IL-6 at low and
high doses. Ni, EC1, and EC2 correlated with IL-6 release at the high dose, decreased IL-6 at the low dose, decreased IL-8 release, and
decreased viability. Brwas negatively associated.
Reference:
Wellenius et al.
(2003, 055691)
Location:
Boston, MA
Particle Size:
PM2.5
Subjects: Dogs N: 6 dogs, 20
exposures
Exposure: CAPs (avg.
mass concentration
range 161 .3-957.3
ug/m3) repeated
exposure with several
weeks in between
Constituents
considered for
grouping: 15
elements (+EC
OC?) (used Clarke
etal.2000)
Grouping
method:
Previous study
in same city
(Clarke et al.
2000)
# of groups: 6
(but did not use
all in analysis of
health effects)
Groups/ Factors/ Sources: V/Ni
S
Al/Si
Br/Pb
Na/CI
Cr
PM variables
used: Univariate:
Mass
Number
Ni,S,Si,BC
Multivariate: Ni,S,
Si.BC
Results: ST-segment elevation increased with CAPs.
Univariate regression: Si and Pb associated with peak ST-segment elevation and integrated ST-segment change.CAPs mass or
number concentration were not associated with any change.
Multivariate regression: Si associated with peak ST-segment elevation and integrated ST-segment change.
Reference:
Zhang etal.
(2008, 192008)
Location: Metro
area of Denver,
CO/ 45 samples
through 1 yr
Particle Size:
2.5; filtered to
0.22 um
Subjects: Alveolar N: 45 PM
macrophage cell line samples, 3
(NR8383);1 xi06 runs
cells/ml
Exposure: Soluble
components exposure
concentration range
from 20-200 pg of
PM/cell
Results: Started with regression on 9
Constituents
considered for
grouping: 43 + EC,
OC
sources, then 3 (water-s
Grouping
method: PMF
# of groups: 9
Groups/Factors/ Sources:
Mobile, water soluble carbon,
sulfate, soil, iron, Cd and Zn point
source, Pb, pyrotechnics,
platinum
PM variables
used: Mass
contribution of
sources
;oluble carbon factor, soil dust source, iron source). Soil dust source was
not significant. Final regression model excluded 3 days of outliers (Fe source most significant, then water-soluble carbon factor, then soil
dust source) for ROS effects, with adjusted R2 of 0.774. Fe source likely associated with industrial source and includes high loadings of
water-soluble Fe and Ti (not identified); water-soluble C factor derived from both secondary organic aerosol and biomass smoke (largely
consists of polar organic compounds); soil dust source identified by water-soluble resuspended dust elements and contains Mgand Ca.
December 2009
F-10
-------
Annex F References
Andersen ZJ; Wahlin P; Raaschou-Nielsen O; Scheike T; Loft S. (2007). Ambient particle source apportionment and daily
hospital admissions among children and elderly in Copenhagen. J Expo Sci Environ Epidemiol, 17: 625-636.
093201
Batalha JR; Saldiva P H; Clarke RW; Coull BA; Stearns RC; Lawrence J; Murthy GG; Koutrakis P; Godleski JJ. (2002).
Concentrated ambient air particles induce vasoconstriction of small pulmonary arteries in rats. Environ Health
Perspect, 110: 1191-1197. 088109
Becker S; Mundandhara S; Devlin RB; Madden M. (2005). Regulation of cytokine production in human alveolar
macrophages and airway epithelial cells in response to ambient air pollution particles: further mechanistic studies.
Toxicol Appl Pharmacol, 207: 269-275. 088590
Bell ML; Ebisu K; Peng RD; Dominici F. (2009). Adverse health effects of particulate air pollution: modification by air
conditioning. Epidemiology, 20: 682-686. 191007
Cakmak S; Dales RE; Vida CB. (2009). Components of particulate air pollution and mortality in Chile. Int J Occup Environ
Health, 15: 152-158. 191995
Clarke RW; Coull B; Reinisch U; Catalano P; Killingsworth CR; Koutrakis P; Kavouras I; Murthy GGK; Lawrence J;
Lovett E; Wolfson JM; Verrier RL; Godleski JJ. (2000). Inhaled concentrated ambient particles are associated with
hematologic and bronchoalveolar lavage changes in canines. Environ Health Perspect, 108: 1179-1187. 013252
Duvall RM; Norris GA; Dailey LA; Burke JM; McGee JK; Gilmour MI; Gordon T; Devlin RB. (2008). Source
apportionment of particulate matter in the US and associations with lung inflammatory markers. Inhal Toxicol, 20:
671-683. 097969
Franklin M; Koutrakis P; Schwartz J. (2008). The role of particle composition on the association between PM2.5 and
mortality. Epidemiology, 19: 680-689. 155779
Gent JF; Koutrakis P; Belanger K; Triche E; Holford TR; Bracken MB; Leaderer BP (2009). Symptoms and medication
use in children with asthma and traffic-related sources of fine particle pollution. Environ Health Perspect, 117:
1168-1174. 180399
Godleski JJ; Clarke RW; Coull BA; Saldiva PHN; Jiang NF; Lawrence J; Koutrakis P. (2002). Composition of inhaled
urban air particles determines acute pulmonary responses. Ann Occup Hyg, 46: 419-424. 156478
Gong H Jr; Linn WS; Clark KW; Anderson KR; Geller MD; Sioutas C. (2005). Respiratory responses to exposures with
fine particulates and nitrogen dioxide in the elderly with and without COPD. Inhal Toxicol, 17: 123-132. 087921
Gong H Jr; Linn WS; Sioutas C; Terrell SL; Clark KW; Anderson KR; Terrell LL. (2003). Controlled exposures of healthy
and asthmatic volunteers to concentrated ambient fine particles in Los Angeles. Inhal Toxicol, 15: 305-325. 042106
Gurgueira SA; Lawrence J; Coull B; Murthy GGK; Gonzalez-Flecha B. (2002). Rapid increases in the steady-state
concentration of reactive oxygen species in the lungs and heart after particulate air pollution inhalation. Environ
Health Perspect, 110: 749-755. 036535
Huang Y-CT; Ohio AJ; Stonehuerner J; McGee J; Carter JD; Grambow SC; Devlin RB. (2003). The role of soluble
components in ambient fine particles-induced changes in human lungs and blood. Inhal Toxicol, 15: 327-342.
087377
Ito K; Christensen WF; Eatough DJ; Henry RC; Kim E; Laden F; Lall R; Larson TV; Neas L; Hopke PK; Thurston GD.
(2006). PM source apportionment and health effects: 2. An investigation of intermethod variability in associations
between source-apportioned fine particle mass and daily mortality in Washington, DC. J Expo Sci Environ
Epidemiol, 16: 300-10. 188554
Note: Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the process of
developing science assessments such as the Integrated Science Assessments (ISA) and the Integrated Risk Information System (IRIS).
December 2009 F-11
-------
Kodavanti UP; Schladweiler MC; Ledbetter AD; McGee JK; Walsh L; Gilmour PS; Highfill JW; Davies D; Pinkerton KE;
Richards JH; Crissman K; Andrews D; Costa DL. (2005). Consistent pulmonary and systemic responses from
inhalation of fine concentrated ambient particles: roles of rat strains used and physicochemical properties. Environ
Health Perspect, 113: 1561-1568. 087946
Laden F; Neas LM; Dockery DW; Schwartz J. (2000). Association of fine particulate matter from different sources with
daily mortality in six US cities. Environ Health Perspect, 108: 941-947. 012102
Lanki T; De Hartog JJ; Heinrich J; Hoek G; Janssen NAH; Peters A; Stolzel M; Timonen KL; Vallius M; Vanninen E;
Pekkanen J. (2006). Can we identify sources of fine particles responsible for exercise-induced ischemia on days
with elevated air pollution? The ULTRA study. Environ Health Perspect, 114: 655-660. 088412
Lippmann M; Hwang J; Maclejczyk P; Chen L. (2005). PM source apportionment for short-term cardiac function changes
inApoE-/- mice. Environ Health Perspect, 113: 1575-1579. 087453
Lippmann M; Ito K; Hwang JS; Maciejczyk P; Chen LC. (2006). Cardiovascular effects of nickel in ambient air. Environ
Health Perspect, 114: 1662-1669. 091165
Maciejczyk P; Chen LC. (2005). Effects of subchronic exposures to concentrated ambient particles (CAPs) in mice: VIII
source-related daily variations in in vitro responses to CAPs. Inhal Toxicol, 17: 243-253. 087456
Mar TF; Ito K; Koenig JQ; Larson TV; Eatough DJ; Henry RC; Kim E; Laden F; Lall R; Neas L; Stolzel M; Paatero P;
Hopke PK; Thurston GD. (2006). PM source apportionment and health effects: 3 Investigation of inter-method
variations in associations between estimated source contributions of PM25 and daily mortality in Phoenix, AZ. J
Expo Sci Environ Epidemiol, 16: 311-320. 086143
Mar TF; Norris GA; Koenig JQ; Larson TV. (2000). Associations between air pollution and mortality in Phoenix, 1995-
1997. Environ Health Perspect, 108: 347-353. 001760
Nikolov MC; Coull BA; Catalano PJ; Diaz E; Godleski JJ. (2008). Statistical methods to evaluate health effects associated
with major sources of air pollution: a case-study of breathing patterns during exposure to concentrated Boston air
particles. J Roy Stat Soc C Appl Stat, 57: 357-378. 156808
Ostro B; Feng W-Y; Broadwin R; Green S; Lipsett M. (2007). The effects of components of fine particulate air pollution on
mortality in California: results from CALFINE. Environ Health Perspect, 115: 13-19. 091354
Ostro B; Roth L; Malig B; Marty M. (2009). The effects of fine particle components on respiratory hospital admissions in
children. Environ Health Perspect, 117: 475-480. 191971
Peng R; Bell M; Geyh A; McDermott A; Zeger S; Samet J; Dominici F. (2009). Emergency admissions for cardiovascular
and respiratory diseases and the chemical composition of fine particle air pollution. Environ Health Perspect, 117:
957-963. 191998
Penttinen P; Vallius M; Tiittanen P; Ruuskanen J; Pekkanen J. (2006). Source-specific fine particles in urban air and
respiratory function among adult asthmatics. Inhal Toxicol, 18: 191-198. 087988
Rhoden CR; Lawrence J; Godleski JJ; Gonzalez-Flecha B. (2004). N-acetylcysteine prevents lung inflammation after
short-term inhalation exposure to concentrated ambient particles. Toxicol Sci, 79: 296-303. 087969
Riediker M; Devlin RB; Griggs TR; Herbst MC; Bromberg PA; Williams RW; Cascio WE. (2004). Cardiovascular effects
in patrol officers are associated with fine particulate matter from brake wear and engine emissions. Part Fibre
Toxicol, 1: 2. 091261
Saldiva PHN; Clarke RW; Coull BA; Stearns RC; Lawrence J; Krishna-Murthy GG; Diaz E; Koutrakis P; Suh H; Tsuda A;
Godleski JJ. (2002). Lung inflammation induced by concentrated ambient air particles is related to particle
composition. Am J Respir Grit Care Med, 165: 1610-1617. 025988
Sarnat JA; Marmur A; Klein M; Kim E; Russell AG; Sarnat SE; Mulholland JA; Hopke PK; Tolbert PE. (2008). Fine
particle sources and cardiorespiratory morbidity: An application of chemical mass balance and factor analytical
source-apportionment methods. Environ Health Perspect, 116: 459-466. 097972
Schreuder AB; Larson TV; Sheppard L; Claiborn CS. (2006). Ambient woodsmoke and associated respiratory emergency
department visits in Spokane, Washington. Int J Occup Environ Health, 12: 147-153. 097959
Schwartz J. (2003). Daily deaths associated with air pollution in six US cities and short-term mortality displacement in
Boston. 042811
December 2009 F-12
-------
Seagrave JC; McDonald JD; Bedrick E; Edgerton ES; Gigliotti AP; Jansen JJ; Ke L; Naeher LP; Seilkop SK; Zheng M;
Mauderley JL. (2006). Lung toxicity of ambient particulate matter from southeastern US sites with different
contributing sources: relationships between composition and effects. Environ Health Perspect, 114: 1387-93.
091291
Tsai FC; Apte MG; Daisey JM. (2000). An exploratory analysis of the relationship between mortality and the chemical
composition of airborne particulate matter. Inhal Toxicol, 12: 121-135. 006251
Urch B; Brook JR; Wasserstein D; Brook RD; Rajagopalan S; Corey P; Silverman F. (2004). Relative contributions of
PM2.5 chemical constituents to acute arterial vasoconstriction in humans. Inhal Toxicol, 16: 345-352. 055629
Veranth JM; Moss TA; Chow JC; Labban R; Nichols WK; Walton JC; Walton JG; Yost GS. (2006). Correlation of in vitro
cytokine responses with the chemical composition of soil-derived particulate matter. Environ Health Perspect, 114:
341-349. Q87479
Wellenius GA; Coull BA; Godleski JJ; Koutrakis P; Okabe K; Savage ST. (2003). Inhalation of concentrated ambient air
particles exacerbates myocardial ischemia in conscious dogs. Environ Health Perspect, 111: 402-408. 055691
Yue W; Schneider A; Stolzel M; Ruckerl R; Cyrys J; Pan X; Zareba W; Koenig W; Wichmann HE; Peters A. (2007).
Ambient source-specific particles are associated with prolonged repolarization and increased levels of inflammation
in male coronary artery disease patients. Mutat Res Fund Mol Mech Mutagen, 621: 50-60. 097968
Zhang Y; Schauer JJ; Shafer MM; Hannigan MP; Dutton SJ. (2008). Source apportionment of in vitro reactive oxygen
species bioassay activity from atmospheric particulate matter. Environ Sci Technol, 42: 7502-7509. 192008
December 2009 F-13
------- |